Harms of Dependence, Withdrawal Or Discontinuation

National Guideline Centre

Harms of dependence, withdrawal or discontinuation syndrome Review of the evidence on harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms from prescribed medicines. Rapid Evidence Assessment Methods, evidence and summary

Serena Carville, Kate Ashmore, Agnès Cuyàs, Nicole Downes, Lina Gulhane, Clare Jones, Eleanor Priestnall, Sharangini Rajesh, Eleanor Samarasekera and Maria Smyth

Developed by the National Guideline Centre

Harms

Contents

Project team:

Serena Carville, Associate Director, Project lead

Kate Ashmore, Project manager to December 2018

Agnès Cyàs, Information Specialist.

Nicole Downes, Systematic reviewer.

Lina Gulhane, Head of Information Specialists.

Clare Jones, Senior Systematic reviewer.

Eleanor Priestnall, Project manager from January 2019.

Sharangini Rajesh, Senior Systematic reviewer.

Eleanor Samarasekera, Senior Systematic reviewer.

Maria Smyth, Senior Systematic reviewer.

Declarations of interest:

We declare the following conflicts of interest: KA; no conflicts of interest, SC; no conflicts of interest, AC; no conflicts of interest, Nicole Downes; no conflicts of interest, Lina Gulhane, CJ; no conflicts of interest, EP; no conflicts of interest, SR; no conflicts of interest, ES; no conflicts of interest and MS; no conflicts of interest. All authors were employees of the Royal College of Physicians at the time of the project.

Acknowledgements:

The development of this rapid evidence assessment was greatly assisted by Jill Cobb, Kirsten Elliott, Kate Kelley, Elizabeth Pearton, Joseph Runicles and Beth Shaw.

Executive summary Contents 1 Executive summary ...... 6 1.1 Objectives ...... 6 1.2 Methods ...... 6 1.3 Findings ...... 6 1.4 Conclusion ...... 6 2 Introduction ...... 8 2.1 Review question to be addressed ...... 8 3 Methodology ...... 9 3.1 Searching for evidence ...... 9 3.2 Identifying and analysing evidence ...... 10 3.2.1 Inclusion and exclusion criteria ...... 10 3.2.2 Type of studies ...... 11 3.3 Methods of combining clinical studies...... 11 3.3.2 Heterogeneity ...... 12 3.4 Appraising the quality of evidence by outcomes ...... 12 3.4.1 Risk of bias ...... 13 3.4.2 Indirectness ...... 14 3.4.3 Inconsistency ...... 14 3.4.4 Imprecision ...... 15 3.4.5 Overall grading of the quality of clinical evidence ...... 16 4 Review question and findings ...... 18 4.1 Available evidence ...... 18 4.1.1 Summary of included studies ...... 19 4.1.2 Short term versus long term opioid use ...... 28 4.1.3 Short versus long term drug withdrawal ...... 28 4.2 Evidence summary ...... 31 4.2.1 Harms of dependency on a medication ...... 31 4.2.2 Harms/ side effects from stopping these medications over a short time frame ... 32 4.2.3 Short term versus long term opioid use ...... 43 4.2.4 Short term versus long term drug withdrawal ...... 44 4.3 Narrative findings ...... 52 4.3.1 Opioids ...... 52 4.3.2 Z drugs ...... 52 4.3.3 Antidepressants...... 52 5 Conclusions ...... 56 6 References ...... 58

Executive summary

7 Appendices ...... 126 Appendix A: Review protocols ...... 126 Appendix B: Study selection ...... 128 Appendix C: Search strategies ...... 129 Appendix D: Forest plots ...... 148 Appendix E: Evidence tables ...... 172 Appendix F: GRADE tables ...... 251 Appendix G: Excluded studies ...... 272 Appendix H: List of medicines to be included ...... 310

Executive summary 1 Executive summary

1.1 Objectives

The Parliamentary Under Secretary of State for Public Health and Primary Care has commissioned Public Health England (PHE) to conduct a public health focussed review of the evidence on dependence, and the short term discontinuation or longer term withdrawal symptoms associated with prescribed medicines. A rapid evidence assessment was subsequently commissioned focussing on 5 key areas reported in 5 associated papers. This report focusses on a review of the evidence on harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms from prescribed medicines.

1.2 Methods

We conducted searches of the Cochrane Library, Epistemonikos, Database of Promoting Health Effectiveness Reviews, Health Evidence, Medline, Epub Ahead of Print, In-Process & Other Non- Indexed Citations and Daily, Embase, PsycINFO, Health Technology appraisals, Trials Register of Promoting Health Interventions and Applied Social Sciences Index & Abstracts for systematic reviews, RCTs and non-randomised studies of dependence, short term discontinuation or longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life /palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and pregabalin (excluding epilepsy treatment), and antidepressants. Searches were limited to the last 10 years to capture the most recent literature in the area.

A systematic review was undertaken for all evidence relevant to the review protocol that assessed harms for the specified drug classes. Data were extracted into evidence tables from the included studies. Risk of bias assessment of individual studies was undertaken according to the Cochrane risk of bias tool for randomised and non-randomised studies as appropriate and quality of the evidence was assessed using the GRADE approach. When full analysis of the data that was identified was not possible, a narrative summary of the results is presented together with the risk of bias rating. Details of the results together with quality of evidence incorporating the risk of bias assessment are provided within the report.

1.3 Findings

Thirty studies were included in the review of which 6 studies reported on opioids, 1 on benzodiazepines, 1 on z-drugs, 1 compared gabapentinoids with benzodiazepines and 21 on antidepressants. There were no studies identified gabapentinoids in isolation of other drugs. The evidence was analysed in the following strata: harms of dependency, harms of short term withdrawal, short versus long term drug withdrawal and short versus long term drug use. Non randomised studies were considered for inclusion for this question as well as randomised controlled trials. Comparative non randomised studies were preferentially included but 1 non-comparative study was included where no other data were available, reporting harms of benzodiazepine withdrawal.

1.4 Conclusion

There is a lack of good quality quantitative evidence of the harms of dependency of the prescription medications included in this review. The only available evidence informing this area was comparing 2 different opioids and indicated that oxycodone may be more harmful than tapentadol in developing doctor shopping behaviour. Another study suggested that longer term use of opioids led to more

Executive summary incidences of depression, alcohol abuse, opioid abuse, other substance abuse, opioid overdose, other substance overdose, opioid dependence and other substance dependence. It is unclear whether these are due to dependence on the opioid however, as long term use was used as a proxy for dependence.

Good quality recent evidence was lacking for benzodiazepines for all areas covered in the review. All of this evidence was considered low to very low quality, meaning firm conclusions cannot be drawn from the effects observed and it is likely that more evidence would change the conclusions drawn from this. Similarly, there was only 1 study identified for z-drugs. This very limited recent evidence base for these drugs is insufficient for drawing conclusions.

Whether or not a gradual tapering schedule minimised the adverse effects experiences was inconclusive from this evidence, although there did appear to be some advantage compared to abruptly stopping the medication.

The majority of the available evidence was for harms from withdrawal or discontinuation of these medications. Most of these were from pharmaceutical funded trials designed to look at efficacy of medications rather than withdrawal from the drugs, and often of relatively small sample size to detect events such as harms. However, from these trials there was a clear suggestion that when compared to placebo, withdrawal from antidepressants may lead to more people experiencing withdrawal syndrome, including taper/post-study emergent adverse events such as vertigo, dizziness and nausea.

Further research would be beneficial to increase confidence in the findings from this review and to further clarify the harms experienced in withdrawing/discontinuing from long term opioid use and antidepressants, including longer term outcomes for all groups of drugs included.

Introduction 2 Introduction The Parliamentary Under Secretary of State for Public Health and Primary Care has commissioned Public Health England (PHE) to conduct a public health focussed review of the evidence on dependence, and the short term discontinuation or longer term withdrawal symptoms associated with prescribed medicines. Within scope of the review will be commonly prescribed medicines for adults who have chronic pain (excluding end of life/palliative care/cancer pain), anxiety, insomnia or depression.

The Review of the evidence on dependence, and the short term discontinuation and longer term withdrawal symptoms associated with prescribed medicines aims to answer the following three questionsa:

1. What is the scale and distribution of dependence, and the short term discontinuation or longer term withdrawal symptoms from prescribed benzodiazepines, z-drugs, gabapentinoids, opioid pain medicines and antidepressants?b 2. What are the factors that may contribute to the risk of dependence, and the short term discontinuation and longer term withdrawal symptoms? 3. How can dependence, and the short term discontinuation and longer term withdrawal symptoms be effectively prevented and treated?

The literature review was commissioned externally by PHE via an open tender process which started in June 2018. The specification for the literature review was drafted by the PHE project team and commented on by the expert reference group. The contract started in August 2018 and the work finished in February 2019. There were 9 bids in total; written bids were assessed by a panel, and 3 bidders were invited to a second stage interview/meeting. The National Guideline Centre (NGC) won the contract based on their skills, experience in undertaking this kind of review and understanding of the potentially contentious nature of the topic.

PHE worked closely with NGC formally in monitoring meetings, ensuring the work progressed, answering queries relating to the specification, or utilising PHE clinical advisers.

2.1 Review question to be addressed

This report focusses on the recent evidence of the harms of dependence, short term discontinuation and longer term withdrawal.

The review question was developed using a PICO framework (population, intervention, comparison and outcome) for intervention reviews.

This use of a framework guided the literature searching process, critical appraisal and synthesis of evidence. The review questions were agreed by PHE and the questions and protocols were reviewed and commented on by the expert reference group.

a The initial review commission sought the answers to three broad questions. During the development of the rapid evidence assessment (REA) methodology the broader questions were refined and the evidence was subsequently sought to answer five focussed research questions. Hereafter in this report we refer to the five review questions. b Prevalence and distribution will be covered in a separate data analysis exercise. However sociodemographic characteristics of patient groups disproportionately affected by prescribed medicine dependence/discontinuation and withdrawal is relevant to the REA.

Methodology 3 Methodology

3.1 Searching for evidence

Literature searches were undertaken to identify all published clinical evidence relevant to the review questions. The approach of running one larger search to cover all review questions was adopted, because it was more comprehensive and inclusive and avoided any overlaps.

The standard approach of applying a PICO structure would have resulted in excessive results. Instead a systematic review and citation-led searching was used in preference to a more traditional database-led PICO approach. The traditional approach resulted in the retrieval of an unmanageable number of references for example, Medline alone retrieved in excess of 28,000 references.

This approach to finding the evidence is consistent with the requirements of NICE’s (2014) methods manual.

Step 1: A search of existing systematic reviews on dependence, discontinuation, and prescribed medicines was undertaken on the following databases: • Cochrane Database of Systematic Reviews (The Cochrane Library) • Epistemonikos • Database of Promoting Health Effectiveness Reviews (DoPHER) • Health Evidence

Step 2: Relevant systematic reviews were ordered to check for inclusion, and bibliographies of relevant reviews identified during step 1 were checked and ordered as required.

Step 3: Databases were then searched to identify most recent evidence which would not have been incorporated in existing reviews due to publication dates. Searches were composed using relevant medical subject headings specific to individual databases, free-text terms and where appropriate study-type filters were applied. Where possible, searches were restricted to papers published in English. Studies published in languages other than English were not reviewed.

Date parameters were applied from 2008 to the date the search was run. Applying a 10-year publication date limit to the search ensured the inclusion of only the most up to date evidence. The limit was chosen based on the upward trend in the prescribing of most classes of medicines and the fact that the larger proportion of the literature on prescribed medicines has been published in the last decade. All searches were completed from 19th September to 3rd October 2018. No papers published after these dates were considered.

Database searches were conducted in: • Medline & Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Daily (OVID) • Embase (OVID) • Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library -Wiley) • PsycINFO (ProQuest) • Health Technology appraisals (Centre for Reviews and Dissemination) • Trials Register of Promoting Health Interventions (TRoPHI) • Applied Social Sciences Index & Abstracts (ASSIA) (ProQuest)

Step 4: Six key papers covering the 5 questions were identified from the main search above for preliminary citation searching. Citation searching was undertaken using Scopus. After examining these results it was decided we did not need to undertake further Scopus searches.

Methodology

Search strategies were quality assured by checking the reference lists of highly relevant studies to ensure that relevant studies were picked up by the searches, and analysing search strategies in other systematic reviews. Search strategies were also checked by a second information specialist using PRESS: Peer Review of Electronic Search Strategies before being run.

The questions, search strategies for the databases, study types and exclusions filters applied and the years covered can be found in appendix C.

The titles and abstracts of records retrieved by the searches were sifted for relevance by 2 independent reviewers. Discrepancies were resolved by a third party. Potentially significant publications were obtained in full text. These were assessed against the inclusion criteria as set out in the protocol.

3.2 Identifying and analysing evidence

3.2.1 Inclusion and exclusion criteria

The inclusion and exclusion of studies was based on the criteria defined in the review protocol, which can be found in Appendix A:. Protocols were published on PROSPERO prior to data extraction commencing. Excluded studies (excluded at the full text stage) with the reasons for their exclusion, are listed in Appendix G: As the search covered all of the 5 questions within this programme of work (see associated reports), if a study was excluded from one question, it would still be considered for all others. The excluded studies list only includes details of studies excluded from all 5 questions after consideration for each.

The key population inclusion criterion was: • Adults (18 years and over) • Dependence, and the short term discontinuation or longer term withdrawal symptoms associated with the prescribed medicines: Benzodiazepines, Z drugs, GABA-ergic medication, opioid pain medications, antidepressants and combinations of these medicines. (Full drug list provided in Appendix H:). • Prescribing in the community, and on discharge from secondary care.

The key population exclusion criterion was:

• People aged under 18 years • Over-the-counter medicines • Medicines prescribed/dispensed in prisons • Medicines prescribed for cancer pain/ palliative care/end of life patients • Gabapentinoids prescribed for epilepsy • Antipsychotic and stimulant medicines • Prescribed medicines to treat drug misuse disorders • The views, experiences and opinions of individual professionals, researchers, commentators or patients. • Non-NHS prescribed medicines.

Narrative literature reviews, posters, letters, editorials, comment articles, unpublished studies and studies not in English were excluded. Systematic reviews were assessed for inclusion, if not relevant to include, references were checked for relevance to the review question and ordered if deemed potentially relevant.

Methodology

3.2.2 Type of studies

Systematic reviews, randomised trials and non-randomised intervention studies were included in the evidence reviews as appropriate.

Parallel randomised controlled trials (RCTs) or systematic reviews of these were included preferentially because they are considered the most robust type of study design that can produce an unbiased estimate of the intervention effects. Non-randomised intervention studies were considered appropriate for inclusion if for example no randomised evidence was available for critical outcomes or for a specific intervention. Systematic reviews were included if they met the protocol and were conducted to the same methodological standard. If insufficient detail were reported to incorporate a relevant systematic review, it was used for citation searching only. Please refer to the review protocol in appendix A for full details on the study design of studies selected for this review question.

3.3 Methods of combining clinical studies

3.3.1.1 Data synthesis for intervention reviews

Where possible, meta-analyses were conducted using Cochrane Review Manager (RevMan5) software to combine the data given in all studies for each of the outcomes of interest for the review question.

Where possible, results were stratified by three elements of the review question: • Harms of dependency on a medication. • Harms / side effects from stopping these medications over a short time frame. • Harms / side effects from longer term withdrawal (or clinical features of long term withdrawal). • Harms of misdiagnosis of withdrawal symptoms as existing or new disorder.

3.3.1.1.1 Analysis of different types of data

Dichotomous outcomes

Fixed-effects (Mantel-Haenszel) techniques (using an inverse variance method for pooling) were used to calculate risk ratios (relative risk, RR) for binary outcomes.

The absolute risk difference was also calculated using GRADEpro(1) software, using the median event rate in the control arm of the pooled results.

For binary variables where there were zero events in either arm or a less than 1% event rate, Peto odds ratios, rather than risk ratios, were calculated. Peto odds ratios are more appropriate for data with a low number of events.

Where sufficient information was provided, hazard ratios were calculated in preference for outcomes such as mortality where the time to the event occurring was important for decision- making.

Continuous outcomes

Continuous outcomes were analysed using an inverse variance method for pooling weighted mean differences.

Where the studies within a single meta-analysis had different scales of measurement, standardised mean differences were used (providing all studies reported either change from baseline or final values rather than a mixture of both); each different measure in each study was ‘normalised’ to the

Methodology

standard deviation value pooled between the intervention and comparator groups in that same study.

The means and standard deviations of continuous outcomes are required for meta-analysis. However, in cases where standard deviations were not reported, the standard error was calculated if the p values or 95% confidence intervals (95% CI) were reported, and meta-analysis was undertaken with the mean and standard error using the generic inverse variance method in Cochrane Review Manager software. Where p values were reported as ‘less than’, a conservative approach was undertaken. For example, if a p value was reported as ‘p≤0.001’, the calculations for standard deviations were based on a p value of 0.001. If these statistical measures were not available then the methods described in section 16.1.3 of the Cochrane Handbook (version 5.1.0, updated March 2011) were applied.

3.3.1.1.2 Generic inverse variance

If a study reported only the summary statistic and 95% CI the generic-inverse variance method was used to enter data into RevMan5. If the control event rate was reported this was used to generate the absolute risk difference in GRADEpro.(1) If multivariate analysis was used to derive the summary statistic but no adjusted control event rate was reported no absolute risk difference was calculated.

3.3.2 Heterogeneity

Statistical heterogeneity was assessed for each meta-analysis estimate by considering the chi- squared test for significance at p<0.1 or an I-squared (I2) inconsistency statistic (with an I-squared value of more than 50% indicating significant heterogeneity) as well as the distribution of effects. Where significant heterogeneity was present, predefined subgrouping of studies was carried out as defined in the review protocol.

If the subgroup analysis resolved heterogeneity within all of the derived subgroups, then each of the derived subgroups were adopted as separate outcomes (providing at least 1 study remained in each subgroup. Assessments of potential differences in effect between subgroups were based on the chi- squared tests for heterogeneity statistics between subgroups. Any subgroup differences were interpreted with caution as separating the groups breaks the study randomisation and as such is subject to uncontrolled confounding.

If all predefined strategies of subgrouping were unable to explain statistical heterogeneity within each derived subgroup, then a random effects (DerSimonian and Laird) model was employed to the entire group of studies in the meta-analysis. A random-effects model assumes a distribution of populations, rather than a single population. This leads to a widening of the confidence interval around the overall estimate, thus providing a more realistic interpretation of the true distribution of effects across more than 1 population. If, the heterogeneity was considered so large that meta- analysis was inappropriate, then the results were described narratively.

3.4 Appraising the quality of evidence by outcomes

The evidence for outcomes from the included RCTs and, where appropriate, non-randomised intervention studies, were evaluated and presented using an adaptation of the ‘Grading of Recommendations Assessment, Development and Evaluation (GRADE) toolbox’ developed by the international GRADE working group (http://www.gradeworkinggroup.org/). The software (GRADEpro(1)) developed by the GRADE working group was used to assess the quality of each outcome, taking into account individual study quality and the meta-analysis results.

Each outcome was first examined for each of the quality elements listed and defined in Table 1.

Methodology

Table 1: Description of quality elements in GRADE for intervention studies Quality element Description Risk of bias Limitations in the study design and implementation may bias the estimates of the treatment effect. Major limitations in studies decrease the confidence in the estimate of the effect. Examples of such limitations are selection bias (often due to poor allocation concealment), performance and detection bias (often due to a lack of blinding of the patient, healthcare professional or assessor) and attrition bias (due to missing data causing systematic bias in the analysis). Indirectness Indirectness refers to differences in study population, intervention, comparator and outcomes between the available evidence and the review question. Inconsistency Inconsistency refers to an unexplained heterogeneity of effect estimates between studies in the same meta-analysis. Imprecision Results are imprecise when studies include relatively few patients and few events (or highly variable measures) and thus have wide confidence intervals around the estimate of the effect relative to clinically important thresholds. 95% confidence intervals denote the possible range of locations of the true population effect at a 95% probability, and so wide confidence intervals may denote a result that is consistent with conflicting interpretations (for example a result may be consistent with both clinical benefit AND clinical harm) and thus be imprecise. Publication bias Publication bias is a systematic underestimate or overestimate of the underlying beneficial or harmful effect due to the selective publication of studies. A closely related phenomenon is where some papers fail to report an outcome that is inconclusive, thus leading to an overestimate of the effectiveness of that outcome. Other issues Sometimes randomisation may not adequately lead to group equivalence of confounders, and if so this may lead to bias, which should be taken into account. Potential conflicts of interest, often caused by excessive pharmaceutical company involvement in the publication of a study, should also be noted.

Details of how the 4 main quality elements (risk of bias, indirectness, inconsistency and imprecision) were appraised for each outcome are given below. Publication or other bias was only taken into consideration in the quality assessment if it was apparent.

3.4.1 Risk of bias

Risk of bias was assessed using the Cochrane risk of bias tool for randomised and non-randomised studies as appropriate. The main domains of bias for RCTs are listed in Table 2. Each outcome had its risk of bias assessed within each study first. For each study, if there were no risks of bias in any domain, the risk of bias was given a rating of 0. If there was risk of bias in just 1 domain, the risk of bias was given a ‘serious’ rating of −1, but if there was risk of bias in 2 or more domains the risk of bias was given a ‘very serious’ rating of −2. A weighted average score was then calculated across all studies contributing to the outcome, by taking into account the weighting of studies according to study precision. For example if the most precise studies tended to each have a score of −1 for that outcome, the overall score for that outcome would tend towards −1.

Table 2: Principle domains of bias in randomised controlled trials Limitation Explanation Selection bias If those enrolling patients are aware of the group to which the next enrolled patient (sequence will be allocated, either because of a non-random sequence that is predictable, or generation and because a truly random sequence was not concealed from the researcher, this may allocation translate into systematic selection bias. This may occur if the researcher chooses not concealment) to recruit a participant into that specific group because of: • knowledge of that participant’s likely prognostic characteristics, and • a desire for one group to do better than the other.

Methodology

Limitation Explanation Performance and Patients, caregivers, those adjudicating or recording outcomes, and data analysts detection bias (lack should not be aware of the arm to which patients are allocated. Knowledge of the of blinding of group can influence: patients and • the experience of the placebo effect healthcare • performance in outcome measures professionals) • the level of care and attention received, and • the methods of measurement or analysis All of which can contribute to systematic bias. Attrition bias Attrition bias results from an unaccounted for loss of data beyond a certain level (a differential of 10% between groups). Loss of data can occur when participants are compulsorily withdrawn from a group by the researchers (for example, when a per- protocol approach is used) or when participants do not attend assessment sessions. If the missing data are likely to be different from the data of those remaining in the groups, and there is a differential rate of such missing data from groups, systematic attrition bias may result. Selective outcome Reporting of some outcomes and not others on the basis of the results can also lead reporting to bias, as this may distort the overall impression of efficacy. Other limitations For example: • Stopping early for benefit observed in randomised trials, in particular in the absence of adequate stopping rules. • Use of unvalidated patient-reported outcome measures. • Lack of washout periods to avoid carry-over effects in crossover trials. • Recruitment bias in cluster-randomised trials.

The assessment of risk of bias differs for non-randomised intervention studies, as they are inherently at high risk of selection bias. For this reason, GRADE requires that non-randomised evidence is initially downgraded on the basis of study design, starting with a rating of −2. This accounts for selection bias and so non-randomised intervention studies are not downgraded any further on that domain. Non-randomised evidence was assessed against the remaining domains used for RCTs in Table 2, and downgraded further as appropriate.

3.4.2 Indirectness

Indirectness refers to the extent to which the populations, interventions, comparisons and outcome measures are dissimilar to those defined in the inclusion criteria for the reviews. Indirectness is important when these differences are expected to contribute to a difference in effect size, or may affect the balance of harms and benefits considered for an intervention. As for the risk of bias, each outcome had its indirectness assessed within each study first. For each study, if there were no sources of indirectness, indirectness was given a rating of 0. If there was indirectness in just 1 source (for example, in terms of population), indirectness was given a ‘serious’ rating of −1, but if there was indirectness in 2 or more sources (for example, in terms of population and treatment) the indirectness was given a ‘very serious’ rating of −2. A weighted average score was then calculated across all studies contributing to the outcome by taking into account study precision. For example, if the most precise studies tended to have an indirectness score of −1 each for that outcome, the overall score for that outcome would tend towards −1.

3.4.3 Inconsistency

Inconsistency refers to an unexplained heterogeneity of results for an outcome across different studies. When estimates of the treatment effect across studies differ widely, this suggests true differences in the underlying treatment effect, which may be due to differences in populations,

Methodology

settings or doses. When heterogeneity existed within an outcome (chi-squared p<0.1, or I2>50%), but no plausible explanation could be found, the quality of evidence for that outcome was downgraded. Inconsistency for that outcome was given a ‘serious’ score of −1 if the I2 was 50–74%, and a ‘very serious’ score of −2 if the I2 was 75% or more.

If inconsistency could be explained based on pre-specified subgroup analysis (that is, each subgroup had an I2<50%), the evidence would be reported for each subgroup separately. In such a situation the quality of evidence was not downgraded for those emergent outcomes.

3.4.4 Imprecision

The criteria applied for imprecision were based on the 95% CIs for the pooled estimate of effect, and the minimal important differences (MID) for the outcome. The MIDs are the threshold for appreciable benefits and harms, separated by a zone either side of the line of no effect where there is assumed to be no clinically important effect. If either end of the 95% CI of the overall estimate of effect crossed 1 of the MID lines, imprecision was regarded as serious and a ‘serious’ score of −1 was given. This was because the overall result, as represented by the span of the confidence interval, was consistent with 2 interpretations as defined by the MID (for example, both no clinically important effect and clinical benefit were possible interpretations). If both MID lines were crossed by either or both ends of the 95% CI then imprecision was regarded as very serious and a ‘very serious’ score of −2 was given. This was because the overall result was consistent with all 3 interpretations defined by the MID (no clinically important effect, clinical benefit and clinical harm). This is illustrated in Figure 1. As for inconsistency, since the imprecision score was based on the meta-analysis results, the score represented the whole outcome and so weighted averaging across studies was not necessary.

The position of the MID lines is ideally determined by values reported in the literature. ‘Anchor- based’ methods aim to establish clinically meaningful changes in a continuous outcome variable by relating or ‘anchoring’ them to patient-centred measures of clinical effectiveness that could be regarded as gold standards with a high level of face validity. For example, a MID for an outcome could be defined by the minimum amount of change in that outcome necessary to make patients feel their quality of life had ‘significantly improved’. MIDs in the literature may also be based on expert clinician or consensus opinion concerning the minimum amount of change in a variable deemed to affect quality of life or health. For binary variables, any MIDs reported in the literature will inevitably be based on expert consensus, as such MIDs relate to all-or-nothing population effects rather than measurable effects on an individual, and so are not amenable to patient-centred ‘anchor’ methods.

In the absence of values identified in the literature, the alternative approach to deciding on MID levels is the ‘default’ method, as follows: • For categorical outcomes the MIDs were taken to be RRs of 0.75 and 1.25. For ‘positive’ outcomes such as ‘patient satisfaction’, the RR of 0.75 is taken as the line denoting the boundary between no clinically important effect and a clinically significant harm, whilst the RR of 1.25 is taken as the line denoting the boundary between no clinically important effect and a clinically significant benefit. For ‘negative’ outcomes such as ‘bleeding’, the opposite occurs, so the RR of 0.75 is taken as the line denoting the boundary between no clinically important effect and a clinically significant benefit, whilst the RR of 1.25 is taken as the line denoting the boundary between no clinically important effect and a clinically significant harm. • For mortality any change was considered to be clinically important and the imprecision was assessed on the basis of the whether the confidence intervals crossed the line of no effect; that is, whether the result was consistent with both benefit and harm. • For continuous outcome variables the MID was taken as half the median baseline standard deviation of that variable, across all studies in the meta-analysis. Hence the MID denoting the minimum clinically significant benefit was positive for a ‘positive’ outcome (for example, a quality of life measure where a higher score denotes better health), and negative for a ‘negative’

Methodology

outcome (for example, a visual analogue scale [VAS] pain score). Clinically significant harms will be the converse of these. If baseline values are unavailable, then half the median comparator group standard deviation of that variable will be taken as the MID. • If standardised mean differences have been used, then the MID will be set at the absolute value of +0.5. This follows because standardised mean differences are mean differences normalised to the pooled standard deviation of the 2 groups, and are thus effectively expressed in units of ‘numbers of standard deviations’. The 0.5 MID value in this context therefore indicates half a standard deviation, the same definition of MID as used for non-standardised mean differences.

For this review the default method was adopted.

Figure 1: Illustration of precise and imprecise outcomes based on the 95% CI of dichotomous outcomes in a forest plot (Note that all 3 results would be pooled estimates, and would not, in practice, be placed on the same forest plot).

MID indicating clinically MID indicating significant benefit clinically significant harm

precise

serious imprecision

very serious imprecision

0.5 1 2 Risk ratio (RR)

3.4.5 Overall grading of the quality of clinical evidence

Once an outcome had been appraised for the main quality elements, as above, an overall quality (or certainty) rating was calculated for that outcome. The scores (0, −1 or −2) from each of the main quality elements were summed to give a score that could be anything from 0 (the best possible) to −8 (the worst possible). However, scores were capped at −3. This final score was then applied to the starting rating that had originally been applied to the outcome by default, based on study design. All RCTs started as ‘high’ and the overall quality became ‘moderate’, ‘low’ or ‘very low’ if the overall score was −1, −2 or −3 points respectively. The significance of these overall ratings is explained in Table 3. The reasons for downgrading in each case were specified in the footnotes of the GRADE tables.

Non-randomised intervention studies started at ‘low’, and so a score of −1 would be enough to take the grade to the lowest level of ‘very low’. Non-randomised intervention studies could, however, be upgraded if there was a large magnitude of effect or a dose-response gradient.

Methodology

Table 3: Overall quality of outcome evidence in GRADE Level Description High Further research is very unlikely to change our confidence in the estimate of effect Moderate Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate Low Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate Very low Any estimate of effect is very uncertain

Review question and findings 4 Review question and findings

For full details see review protocol in appendix A.

Table 4: PICO characteristics of review question Population Adults using prescription medicines associated with dependence and withdrawal (see interventions below) Interventions • Opioids for chronic pain • Benzodiazepines • Z-drugs • Gabapentinoids • Antidepressants Outcomes Harms (as described by papers) Study design A hierarchy of evidence will be used. • Systematic review • RCT • Non-randomised comparative study • Non-randomised prospective cohort • Non-randomised retrospective cohort

4.1 Available evidence

Thirty studies were included in the review(2-32); these are summarised in Table 5 below. Evidence from these studies is summarised in the summary GRADE tables below (Table 13 - Table 25). See also the study selection flow chart in appendix B, forest plots in appendix D, study evidence tables in appendix E, GRADE tables in appendix F and excluded studies list in appendix G.

• Six studies reported harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms of opioids. • One study reported harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms of benzodiazepines. • One study reported harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms of z-drugs. • One study reported harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms of gabapentinoids compared to benzodiazepines. No studies were identified that were relevant to the review protocol that reported gabapentinoids in isolation of other drugs. • Twenty one studies reported harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms of antidepressants.

The evidence was analysed in the following strata: • Harms of dependency • Harms of short term withdrawal • Short versus long term drug withdrawal • Short versus long term drug use.

Review question and findings

Non randomised studies were considered for inclusion for this question due to the limited number of randomised controlled trials identified. Comparative non randomised studies were preferentially included consistent with the hierarchy of studies stated in the protocol. Only 1 non-comparative study was included, reporting harms of benzodiazepine withdrawal, due to the lack of evidence for this intervention. One matched cohort study(33) comparing benzodiazepines with z-drugs that reported mortality risk but was considered for inclusion, but subsequently excluded as it was unclear whether the population were dependent on the prescribed medications. See Appendix G for the list of excluded studies for further details. The main reasons that studies were excluded were that they were published prior to 2008 or had a non-comparative study design which was excluded when better quality evidence was available for that intervention.

4.1.1 Summary of included studies

The thirty included studies for this review are summarised in the table below. The table describes the intervention and comparison, population and outcomes. The treatment and discontinuation times are reported in the intervention column where appropriate. The comments column reports study design and source of funding. The table has been ordered by strata and within each strata by type of intervention.

Review question and findings and question Review Harms

Table 5: Summary of studies included in the review - Opioids Study Intervention and comparison Population Outcomes Comments Cepeda 2013(6) Tapentadol People exposed to tapentadol ‘Doctor shopping’ behaviour, Retrospective cohort, (analysis of a vs or oxycodone number of shopping episodes database) Oxycodone at any point during one year Not all subjects were using opioids n=155,761 follow-up for chronic pain. Unclear if doctor shopping was due USA to dependency of the prescribed opioid.

Funding: Pharma (Janssen)

4.1.1.2 Harms of short term withdrawal

Table 6: Summary of studies included in the review – Opioids Study Intervention and comparison Population Outcomes Comments Afilalo 2010(2) Tapentadol 100-250mg twice daily Chronic pain Clinical opioid withdrawal scale RCT vs (COWS) Funding: Pharma (Johnson and Oxycodone 20-50mg vs placebo n=1030 Johnson Pharmaceutical Research and Development) 12 week treatment and follow-up Multicentre: USA, Canada, New 14 days after last intake of study Zealand and Australia medication. Buynak 2010(5) Tapentadol 100-250 mg twice Low back pain COWS RCT daily Funding: Pharma (Johnson and vs n=981 Johnson Pharmaceutical Research Oxycodone 20-50 mg vs placebo and Development and Global Development) Multicentre: USA, Canada and 12 week treatment and follow up Australia period consisted of a visit 4 days

Review question and findings and question Review Harms © Royal © College of Physicians 2019 after end of treatment. Hale 2009(8) Tapentadol 50 or 100mg every 4-6 Low back pain COWS RCT

hours with a maximum dose 600 Funding: Pharma (Johnson and mg day n=878 Johnson Pharmaceutical Research vs and Development and Grunenthal GmbH) Oxycodone 10 or 15mg every 4-6 Multicentre: Canada and USA hours with a maximum daily dose of 90 mg

3 month treatment and 3 day follow-up period.

Wild, 2010(17) Tapentadol 100-250 mg twice Low back pain COWS RCT daily Funding: Pharma (Johnson and vs n=1121 Johnson Pharmaceutical Services, Oxycodone 20-50 mg and Global Development) Multicentre: USA and Europe

21 51 weeks treatment and follow-up

period consisting of a visit 4 days after treatment and a call 10-14 days after treatment.

Table 7: Summary of studies included in the review – Z-drugs Study Intervention and comparison Population Outcomes Comments Hajak 2009 (7) Zolpidem 10 mg Primary insomnia Rebound insomnia RCT vs Funding: Pharma (Lundbeck) Gaboxadol and placebo n=742

Study run over 2 weeks with a Multicentre: Canada and run-out phase over 1 week Europe

Review question and findings and question Review Harms

Table 8: Summary of studies included in the review – Benzodiazepines

Study Intervention and comparison Population Outcomes Comments Bourgeois, 2014 (4) GP-initiated discontinuation Nursing home residents with Successful discontinuation, Observational study chronic use of benzodiazepines withdrawal symptoms, Funding not stated or Z-drugs for insomnia symptom scores, sleep quality, quality of life n=38 Narrative outcome only

Belgium

Table 9: Summary of studies included in the review – Gabapentinoids and benzodiazepines Study Intervention and comparison Population Outcomes Comments Kasper 2014(10) Pregabalin 450-600mg/d vs Generalised anxiety disorder Rebound insomnia RCT pregabalin 150-300 mg/d Funding: Pharma (Pfizer)

22 vs n=615

Lorazepam 3-4 mg/d Multicentre: South America, Study duration was 2 periods of Europe, Asia and Russia 12 weeks treatments followed by 1 week taper and 1 week follow- up.

Table 10: Summary of studies included in the review - Antidepressants Study Intervention and comparison Population Outcomes Comments Boyer 2008(32) Desvenlafaxine 50mg/d People with major depressive Discontinuation emergent signs RCT vs disorder and symptoms checklist (DESS), Funding: Pharma (Wyeth Research) Desvenlafaxine 100mg/d taper/post study-emergent adverse events vs n=485 Placebo. Multicentre - Europe & South Treatment 8 weeks. Followed by a Africa

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Study Intervention and comparison Population Outcomes Comments 7 day taper with placebo except

for desvenlafaxine 10 mg/d that received 50mg/d dose Cutler 2009(21) Duloxetine 60 mg/d Major depressive disorder Adverse events after RCT vs discontinuation, treatment Funding Pharma (AstraZeneca) Placebo n=610 discontinuation signs and symptoms (score developed by (Quetiapine 150 mg/d and 300 AstraZeneca) mg/d arms but not included in the USA analysis).

6 weeks treatment. 2 weeks discontinuation period; duloxetine dose halved in first week of tapering and then discontinued in second week.

23 Krystal 2011(12) 3/6 mg doxepin. Followed by Primary insomnia. Sleep quality, adverse events, RCT

placebo again for discontinuation rebound insomnia, withdrawal Funding Pharma (Somaxon period. n=229 symptoms Pharmaceutical) vs Placebo. Followed by placebo for USA discontinuation period.

Treatment 5 weeks. Abrupt discontinuation with placebo for 2 days. Liebowitz 2009(22) Desvenlafaxine 50 mg/d Major depressive disorder Suicidal ideation RCT vs DESS Funding Pharma (Wyeth) Desvenlafaxine 100 mg/d n=474 % taper emergent adverse vs events Placebo USA

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Study Intervention and comparison Population Outcomes Comments

Treatment 8 weeks Tapering 1 week – Desvenlafaxine group tapered to 50mg/d and desvenlafaxine 50 mg/d went to placebo and placebo remained on placebo. Liebowitz 2009(28) Venlafaxine extended release Panic disorder Percentage of total taper/post RCT

flexible dose 75-225 mg/d study-emergent adverse Funding Pharma (Wyeth) vs placebo n=343 events, Percentage of specific taper/post study-emergent Treatment: 10 weeks. Canada & USA Taper period: Up to 2 weeks; adverse events higher dose tapered but if dose on 75 mg/d then straight to placebo. Mahableshwarkar Vortioxetine 15mg, vortioxetine Major depressive disorder Suicidal ideation RCT

24 2015(13) 20mg or duloxetine 60 mg Funding Pharma (Takeda

vs placebo n=614 Pharmaceutical company, and Lundbeck). Treatment 8 weeks. USA 2 week discontinuation period. Abrupt discontinuation for vortioxetine and tapered to 30mg for first week in duloxetine. Nishimura, 2018(14) Vortioxetine 5mg, 10mg or 20mg Major depressive disorder Symptoms after RCT daily discontinuation Funding Pharma (Takeda vs n=600 Pharmaceutical company, and Dose comparison and placebo Lundbeck) Multicentre: Europe, Japan and Treatment 8 weeks. Asia/Oceania 2 week discontinuation period on placebo.

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Study Intervention and comparison Population Outcomes Comments

Perahia 2009(27) Duloxetine 60-120 mg/d Major depressive disorder Percentage of discontinuation RCT vs placebo emergent adverse events Funding Pharma (Eli Lilly and n=288 Boehringer Ingelheim GmbH) Treatment phase: 52 weeks Optional discontinuation phase: 3 Multicentre: Europe, Russia weeks where duloxetine patients and USA were down titrated occurred over

a 2-3 week period depending on the dose taken. Raskin 2008(20) Duloxetine 60 mg/daily Elderly people with major Incidence of at least one RCT

vs placebo depressive disorder discontinuation-emergent Funding Pharma (Eli Lilly and adverse event, Incidence of Boehringer Ingelheim GmbH) most frequent discontinuation- Screening: 1 week n=311 emergent adverse events Double blind placebo phase: 1

25 week USA

Treatment: 8 weeks Double blind discontinuation phase: 1 week; dosage of the study drug was tapered. Rickels 2010(29) Desvenlafaxine 200 or 400 mg/d Major depressive disorder; DESS RCT vs placebo people that were responders to Funding Pharma (Wyeth) treatment Total taper/post therapy- Open label period of emergent adverse events Note: recommended dose of desvenlafaxine at 200 mg for 12 n=374 desvenlafaxine for MDD is 50 mg/d weeks. Treatment period for those that Multicentre: Europe, USA and responded to open label phase Taiwan. and randomised to drug v placebo (6 months). Discontinuation period of 1-2 weeks.

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Study Intervention and comparison Population Outcomes Comments

Rosenthal 2013 (25) Desvenlafaxine 50 mg/d Major depressive disorder; Percentage of taper/post study RCT vs placebo people that were responders to adverse events Funding Pharma (Pfizer) treatment Screening period 20 weeks. Serious adverse events during Responders eligible for treatment n=874 taper/post therapy period. Treatment period 6 months. Multicentre: USA, South Taper period 1 week where America, Europe and South desvenlafaxine patients received Africa. 25 mg/d for one week.

Follow up appointment 1 week after taper period. Rynn 2008(23) Duloxetine 60 mg Generalised anxiety disorder Discontinuation emergent RCT vs placebo adverse events and dizziness Funding Pharma (Eli Lily and Boehringer Ingelheim)

26 n=327

Treatment 20 weeks and 2 weeks discontinuation. USA Schagen 2008(24) Duloxetine 40 mg Women with symptoms of Discontinuation emergent RCT vs placebo stress urinary continence or adverse events Funding Pharma (Eli Lily and stress predominant mixed Boehringer Ingelheim) urinary continence Treatment 12 weeks and 2 week discontinuation n=265

Multicentre: Europe & South Africa Stein 2012(15) Agomelatine 25-50 mg/d Generaliseed anxiety disorder Discontinuation emergent signs RCT vs placebo and symptoms checklist Funding Pharma (Servier) n=227 Treatment 6 months. Randomised to one week abrupt Multicentre: Canada and

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Study Intervention and comparison Population Outcomes Comments discontinuation or one week Europe

continued treatment. Tourian 2009(30) Desvenlafaxine 50 mg/d Major depressive disorder DESS score RCT vs desvenlafaxine 100 mg/d Drug withdrawal syndrome, Funding Pharma vs n=638 adverse events (Wyeth) Duloxetine 60 mg/d vs placebo USA

Treatment 8 weeks and taper one week.

YaDeau, 2016(18) Duloxetine 60 mg one per day Osteoarthritis patients Changes after discontinuation RCT vs undergoing total knee Funding Non pharma (hospital for Placebo one per day arthroplasty special surgery anaesthesiology department research and education

27 fund) Treatment 2 weeks (starting day n=106 of surgery). Discontinuation symptoms USA evaluated 4-6 days after stopping medication. Yonkers, 2015(19) Sertraline 50-100 mg/d Premenstrual dysphoric Michelson SSRI withdrawal RCT vs placebo disorder scale Funding Non pharma (National Narrative outcome only Institute of Mental Health) but study 6 monthly menstrual cycles of n=252 drug were donated by Pfizer medication with abrupt discontinuation and symptoms USA reported 3 days after stopping.

Review question and findings and question Review Harms © Royal © College of Physicians 2019 4.1.2 Short term versus long term opioid use

Table 11: Summary of studies included in the review – Opioids Study Intervention and comparison Population Outcomes Comments Hoffman 2017 (9) Long term opioid therapy Patients with polyneuropathy Depression, Abuse (alcohol, Retrospective cohort vs opioid, other-substance), Funding Non Pharma (Mayo Shorter duration opioid therapy n=2892 Overdose (opioid, other foundation for medical education substance), Dependence and Research and NIH) (opioid, other substance), USA Mortality

4.1.3 Short versus long term drug withdrawal

Table 12: Summary of studies included in the review – Antidepressants Study Intervention and comparison Population Outcomes Comments

28 Baldessarini 2010(3) Managed withdrawal: Rapid Patients with a DSM-IV Time to first new illness – Observational (cohort) discontinuation of recurrent major depressive depressive/panic episode antidepressants (1-7 days) disorder, bipolar I or II disorder, Serious population indirectness: Not or panic disorder that all of the antidepressants are listed responded well to Managed withdrawal: Gradual (e.g. serotonin reuptake inhibitors discontinuation of antidepressant treatment with grouped together and individuals’ antidepressants (≥2 weeks) or without a mood stabiliser. drugs not listed), therefore could include some not on included list. n=398 Funding Non Pharma Italy Gallagher 2012(26) Treatment on desvenlafaxine 100 Premenopausal women with DESS RCT mg/d for 15 weeks. vasomotor symptoms Adverse events during taper Funding Pharma (Wyeth) Taper period 2 weeks: periods Desvenlafaxine 50 mg/d followed n=384 by 25mg/d vs Country not reported

Review question and findings and question Review Harms © Royal © College of Desvenlafaxine 50 mg/d every other day

vs Desvenlafaxine 50mg/d for one week followed by placebo vs

Physicians 2019 Placebo (no taper) Khan 2014(11) All participants received 24-week Primary diagnosis of single or Headache, nausea, dizziness RCT

open-label treatment with recurrent major depressive desvenlafaxine (50 mg/day). disease. Side effects and non-fatal Funding Pharma (Pfizer) Participants completing this overdose, suicide ideation and period were randomly assigned to

n=361 mortality three discontinuation regimens

for 4 weeks: USA Managed withdrawal: Taper - Desvenlafaxine at dose of 25

29 mg/day for 1 week followed by

placebo for 3 weeks. Control: Abrupt discontinuation - Placebo received for 4 weeks. Usual care: No discontinuation - Continued to receive 50 mg/day desvenlafaxine for 4 weeks. Tint, 2008 (16) Antidepressants: Major depressive disorder Discontinuation symptoms RCT 3 day taper Short report vs n=28 Funding – No external funding 14 day taper Setting not reported (assumed UK)

Harms

Review question and findings

Outcome scales reported:

Clinical Opioid Withdrawal Scale (COWS) is a clinician rated 11-item scale based on questions and clinical observations rated by the investigator on a Likert-type scale (0 to 4 or 0 to 5, depending on the item). The severity of opioid withdrawal is categorized according to the sum of the 11 items (maximum score 48) as follows: no withdrawal , <5; mild withdrawal, 5 to 12; moderate withdrawal, 13 to 24; moderately severe withdrawal, 25 to 36; or severe withdrawal, >36.

The subjective Opiate Withdrawal Scale (SOWS) questionnaire is a 16-item, self-reported measure of perceived opioid withdrawal.

The Objective Opiate Withdrawal Scale (OOWS) is an instrument to assess the presence and intensity of opioid withdrawal symptoms. It is based on an observer-based scale.

The Benzodiazepine Withdrawal Symptom Questionnaire (BWSQ) is a self-report questionnaire measuring 20 symptoms. The three answer possibilities (no-moderate-severe) are linked to a rating of 0, 1 or 2, with higher scores indicating more withdrawal symptoms (maximum score is 40).

The discontinuation-emergent signs and syndromes (DESS) checklist is a patient-rated instrument that is administered by a clinician. It queries for signs and symptoms associated with the discontinuation of an SSRI treatment. 43-item checklist with one point scored for each new or worsened symptom.

The Physician Withdrawal checklist (PWC) is a 20 item physician-rated interview (total score range 0-60) that measure the presence of anxiolytic drug discontinuation related signs and symptoms in the following areas: gastrointestinal, mood, sleep, motor, somatic, perception and cognition.

Review question and findings and question Review Harms

4.2 Evidence summary

4.2.1 Harms of dependency on a medication

Table 13: Opioids for chronic pain Outcomes No of Quality of the Relative Anticipated absolute effects Participants evidence effect

(studies) (GRADE) (95% CI) Risk with Control Risk difference with Harms of Follow up dependency - opioids (non RCTS) (95% CI) Developed doctor shopping behaviour (defined as more than 155761 ⊕⊝⊝⊝ RR 0.24 Moderate (1 study) VERY LOW1,2 (0.19 to 1 prescription by at least 2 different prescribers with at least 1 9 per 1000 7 fewer per 1000 1 years due to risk of 0.3) (from 6 fewer to 7 fewer) day of overlap and filled at 3 or more pharmacies) bias, indirectness OR3 3.5 (2.8 to 4.4) Number of shopping episodes per subject 155761 ⊕⊝⊝⊝ The mean number of shopping 1,2

31 (1 study) VERY LOW The mean number of shopping episodes per subject in the 1 years due to risk of episodes per subject in the intervention groups was

bias, indirectness control group was 0.02 0.02 lower (0.02 to 0.01 lower) 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 The majority of the evidence included an indirect population (downgrade by one increment) 3 Adjusted odds ratio – controlling for gender, benzodiazepine use and type of payment at first opioid exposure using a conditional logistic regression.

This table relates to forest plots figures 3-5 (Appendix D).

One non randomised study(34) retrospectively reviewed database records of 155,761 individuals who had been exposed to tapentadol or oxycodone. Not all the participants were using opioids for chronic pain and so this evidence was considered as indirect to the population specified in the review protocol and downgraded in quality for that reason. The results in the table above demonstrate that the evidence suggested fewer people taking tapentadol developed doctor shopping behaviour (defined as more than 1 prescription by at least 2 different prescribers with at least 1 day of overlap and filled at 3 or more pharmacies) at any point during the 1 year follow-up compared to oxycodone. However, the evidence suggested that there was no difference in number of shopping episodes per subject during the year follow up (with evidence rated as very low quality due to selection bias, risk of confounding and risk of deviations from intended intervention, based on 1 study). There was also no data available if people changed to a pharmacy that was not covered by this database or ‘doctor shopped’ outside of this database.

Review question and findings and question Review Harms © Royal © College of Physicians 2019

4.2.2 Harms/ side effects from stopping these medications over a short time frame

Table 14: Opioid versus placebo Outcomes No of Quality of the evidence Relative Anticipated absolute effects Participants (GRADE) effect (studies) (95% CI) Risk with Risk difference with Harms short Follow up Control time - opioids (95% CI)

Opioid vs control - no opioid withdrawal – Clinical Opioid Withdrawal 95 ⊕⊕⊕⊝ RR 0.87 Moderate 1 Scale (COWS) assessment 2-4 days after last intake of medication (1 study) MODERATE (0.77 to 1000 per 130 fewer per 1000 due to risk of bias 0.99) 1000 (from 10 fewer to 230 fewer) Opioid vs control - no opioid withdrawal - COWS assessment at 4 days 211 ⊕⊕⊕⊝ RR 1.04 Moderate 1 after last intake of medication (1 study) MODERATE (0.94 to 898 per 36 more per 1000 due to risk of bias 1.14) 1000 (from 54 fewer to 126 more) Opioid vs control - no opioid withdrawal - COWS assessment 5+ days 213 ⊕⊕⊕⊝ RR 1 Moderate 1 after last intake of medication (1 study) MODERATE (0.92 to 916 per 0 fewer per 1000 due to risk of bias 1.1) 1000 (from 73 fewer to 92 more) 32 Opioid vs control - mild or moderate opioid withdrawal - COWS 95 ⊕⊝⊝⊝ RR 4.05 Moderate 1,2 assessment 2-4 days after last intake of medication (1 study) VERY LOW (0.55 to 0 per 150 more per 1000 due to risk of bias, 29.59) 1000 (from 40 more to 270 more) imprecision Opioid vs control - mild or moderate opioid withdrawal - COWS 211 ⊕⊝⊝⊝ RR 0.69 Moderate 1,2 assessment at 4 days after last intake of medication (1 study) VERY LOW (0.26 to 102 per 32 fewer per 1000 due to risk of bias, 1.8) 1000 (from 75 fewer to 82 more) imprecision Opioid vs control - mild or moderate opioid withdrawal - COWS 213 ⊕⊝⊝⊝ RR 0.63 Moderate 1,2,3 assessment 5+ days after last intake of medication (1 study) VERY LOW (0.05 to 85 per 31 fewer per 1000 due to risk of bias, 8.48) 1000 (from 81 fewer to 636 more) inconsistency, imprecision

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Downgraded by 1 or 2 increments because heterogeneity, I2=75%, p=0.05, unexplained by sub-group analysis.

This table relates to forest plots figures 6-7 (Appendix D).

Two studies(2, 5) comprising 2011 participants’ evaluated opioids compared to placebo. Both of these studies had 3 arms (oxycodone, tapentadol and placebo) and this analysis includes comparison of oxycodone with placebo as well as tapentadol with placebo to measure the Clinical Opioid Withdrawal

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Scale (COWS). The results in the table above demonstrate that the evidence suggested that the placebo group had more people with no opioid withdrawal measured on the COWS score compared to those receiving opioids measured 2-4 days after last intake of medication. However, there was no difference

between groups with no opioid withdrawal measured on the COWS score on day 4 and day 5 onwards after last medication intake (evidence rated as moderate quality due to risk of bias of incomplete outcome data due to high attrition rates, based on 1 study). The evidence suggested that the opioid group had more participants with mild or moderate opioid withdrawal measured by COWS at days 2-4 after last intake of medication (evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study). However, there was no difference between opioid and placebo in mild or moderate opioid withdrawal measured by COWS at day 4 (evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition

rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study or day 5 plus after last intake of medication (evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition rates, wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie and inconsistency between the tapentadol versus placebo arm with the oxycodone versus placebo arm, based on 1 study).

Review question and findings and question Review Harms © Royal © College of Physicians 2019

Table 15: Opioid versus opioid Outcomes No of Quality of the evidence Relative Anticipated absolute effects Participants (GRADE) effect (studies) (95% CI) Risk with Risk difference with Harms short Follow up Control time - opioids (95% CI) 85 per 5 fewer per 1000 1000 (from 56 fewer to 137 more) Withdrawal syndrome 1117 ⊕⊝⊝⊝ RR 1.62 Moderate 1,2 (1 study) VERY LOW (0.37 to 9 per 1000 6 more per 1000 due to risk of bias, imprecision 7.13) (from 6 fewer to 55 more) Tapentadol vs oxycodone - no opioid withdrawal - COWS assessment 219 ⊕⊕⊕⊝ RR 1.01 Moderate 1 2-4 days after last intake of medication (2 studies) MODERATE (0.85 to 796 per 8 more per 1000 due to risk of bias 1.19) 1000 (from 119 fewer to 151 more) Tapentadol vs oxycodone - no opioid withdrawal - COWS assessment 152 ⊕⊕⊕⊝ RR 1.04 Moderate 1 at 4 days after last intake of medication (1 study) MODERATE (0.96 to 911 per 36 more per 1000 due to risk of bias 1.14) 1000 (from 36 fewer to 128 more) Tapentadol vs oxycodone - no opioid withdrawal - COWS assessment 370 ⊕⊕⊕⊝ RR 1.1 Moderate

34 1 5+ days after last intake of medication (2 studies) MODERATE (1.01 to 849 per 85 more per 1000 due to risk of bias 1.19) 1000 (from 8 more to 161 more) Tapentadol vs oxycodone - mild or moderate opioid withdrawal - 591 ⊕⊕⊝⊝ RR 0.70 Moderate 1,2 COWS assessment 2-4 days after last intake of medication (3 studies) LOW (0.48 to 273 per 82 fewer per 1000 due to risk of bias, imprecision 1.00) 1000 (from 142 fewer to 0 more) Tapentadol vs oxycodone - mild or moderate opioid withdrawal - 152 ⊕⊝⊝⊝ RR 0.54 Moderate 1,2 COWS assessment 4 days after last intake of medication (1 study) VERY LOW (0.15 to 89 per 41 fewer per 1000 due to risk of bias, imprecision 1.97) 1000 (from 76 fewer to 86 more) Tapentadol vs oxycodone - mild or moderate opioid withdrawal - 370 ⊕⊝⊝⊝ RR 0.33 Moderate 1,2,3 COWS assessment 5+ days after last intake of medication (2 studies) VERY LOW (0.04 to 151 per 101 fewer per 1000 due to risk of bias, 2.72) 1000 (from 145 fewer to 260 more) inconsistency, imprecision Drug withdrawal syndrome 1117 ⊕⊝⊝⊝ RR 2.24 Moderate 1,2 (1 study) VERY LOW (0.29 to 5 per 1000 6 more per 1000 due to risk of bias, imprecision 17.63) (from 4 fewer to 83 more)

Review question and findings and question Review Harms © Royal © College of Physicians 2019 This table relates to forest plots figures 8-11 (Appendix D).

Four studies(2, 5, 8, 17) comprising 4010 participants evaluated tapentadol compared to oxycodone. The results in the table above demonstrate that the evidence suggested that there was a benefit with tapentadol in more participants with no opioid withdrawal using COWS at 5 plus days after last intake of medication (evidence rated as moderate quality due to risk of bias of incomplete outcome data due to high attrition rates, based on 2 studies). The evidence suggested there was no difference between the opioids in COWS measure of no opioid withdrawal at day 2-4 days after last intake of medication (evidence rated as moderate quality due to risk of bias of incomplete outcome data due to high attrition rates, based on 2 studies) and day 4 after last intake of medication (evidence rated as moderate quality due to risk of bias of incomplete outcome data due to high attrition rates, based on 1 study), withdrawal syndrome 3 days after last dose (evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition and lack of investigator blinding; and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study) and drug withdrawal syndrome 3 days after last dose (evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition and lack of investigator blinding; and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study). The evidence suggested a benefit for tapentadol in fewer participants with mild or moderate opioid withdrawal measured with COWS at 2-4 days (evidence rated as low quality due to risk of bias of incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 3 studies, low quality evidence), 4 days (evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where

35 the true effect might lie, based on 1 study) and 5 days (evidence rated as very low quality due to and lack of investigator blinding; and wide confidence

intervals around the effect and therefore uncertainty of where the true effect might lie, based on 2 studies) last after intake of medication.

Table 16: Z-drugs Outcomes No of Quality of the Relative Anticipated absolute effects Participants evidence effect (studies) (GRADE) (95% CI) Risk with Risk difference with Z-drugs v Follow up Control placebo (95% CI) Rebound insomnia – proportion of patients with a lower self-reported total 193 ⊕⊕⊕⊕ RR 3.06 Moderate sleep time - Run out phase - day 1 (1 study) HIGH (1.33 to 7) 64 per 132 more per 1000 1000 (from 21 more to 384 more) Rebound insomnia - proportion of patients with a lower self-reported total 193 ⊕⊕⊝⊝ RR 1.34 Moderate 1 sleep time - Run out phase - day 2 (1 study) LOW (0.44 to 56 per 19 more per 1000 due to imprecision 4.07) 1000 (from 31 fewer to 172 more) Rebound insomnia – proportion of patients with a lower self-reported total 193 ⊕⊕⊝⊝ RR 1.88 Moderate 1 sleep time - Run out phase - day 3 (1 study) LOW (0.56 to 40 per 35 more per 1000 due to imprecision 6.27) 1000 (from 18 fewer to 211 more) Rebound insomnia – proportion of patients with a lower self-reported time 193 ⊕⊕⊕⊝ RR 2.82 Moderate 1 to sleep onset - Run out phase - day 1 (1 study) MODERATE (1.21 to 64 per 116 more per 1000 due to imprecision 6.56) 1000 (from 13 more to 356 more)

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Rebound insomnia – proportion of patients with a lower self-reported time 193 ⊕⊕⊝⊝ RR 1.88 Moderate 1 to sleep onset - Run out phase - day 2 (1 study) LOW (0.49 to 32 per 28 more per 1000 due to imprecision 7.28) 1000 (from 16 fewer to 201 more) Rebound insomnia – proportion of patients with a lower self-reported time 193 ⊕⊕⊝⊝ RR 1.88 Moderate 1 to sleep onset - Run out phase - day 3 (1 study) LOW (0.56 to 40 per 35 more per 1000 due to imprecision 6.27) 1000 (from 18 fewer to 211 more)

1 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

This table relates to forest plots figures 12-13 (Appendix D).

One RCT(7) comprising 193 participants evaluated zolpidem compared to placebo for insomnia. After the 2 week treatment phase there was a 1 week run- out phase where 50% of patients were randomised to placebo, while the other 50% remained on their existing treatment. Rebound insomnia reported in

the run-out phase compared the placebo group to the group that were on Z-drugs during the treatment phase but had been randomised to placebo during the run-out phase. The evidence in the table above demonstrates that the evidence suggested that there was an increase in rebound insomnia (proportion of the patients with a lower self-reported total sleep time value after the first night of the run-out phase compared to lowest value during the run-in phase) in the zolpidem group compared to placebo at day 1 of the run-out phase (evidence rated as high quality, based on 1 study). The evidence suggested that there was an increase in rebound insomnia (proportion of the patients with a lower self-reported time to sleep onset value after the first 36 night of the run-out phase compared to lowest value during the run-in phase) in the zolpidem group compared to placebo at day 1 of the run-out phase

(evidence rated as moderate quality due to wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study). However, the evidence suggested that there was no difference in rebound insomnia (self-reported lower total sleep time and self-reported time to sleep onset) on days 2 and 3 of the run-out phase (evidence rated as low quality evidence due to wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study).

Review question and findings and question Review Harms © Royal © College of Physicians 2019

Table 17: Benzodiazepines versus gabapentinoids Outcomes No of Quality of the Relative Anticipated absolute effects Participants evidence effect (studies) (GRADE) (95% CI) Risk with Control Risk difference with Benzodiazepines versus gabapentinoids (95% CI) Follow up Mean change in Physician Withdrawal Checklist 107 ⊕⊕⊝⊝ The mean change in PWC - higher dose gab - 1 week 1 (PWC) – after treatment period 1 (higher dose gab) - 1 (1 study) LOW The mean change in after taper in the intervention groups was due to risk of bias the control group was 0.4 lower week after taper

2.3 (3.09 lower to 2.29 higher) Mean change in PWC - higher dose gab – after 98 ⊕⊕⊝⊝ The mean change in PWC - higher dose gab - 2 week 1 treatment period 1 - 2 week after taper (1 study) LOW The mean change in after taper in the intervention groups was due to risk of bias the control group was 0.5 higher 1.6 (1.92 lower to 2.92 higher)

Mean change in PWC after treatment period 2 - higher 199 ⊕⊕⊝⊝ The mean change in PWC after treatment period 2 - 1 dose gab - 2 week after taper (1 study) LOW The mean change in higher dose gab - 2 week after taper in the due to risk of bias the control group was intervention groups was 2.2 0.6 higher (1.08 lower to 2.28 higher) Mean change in PWC after treatment period 2 - higher 208 ⊕⊕⊝⊝ The mean change in PWC after treatment period 2 - 37 (1 study) LOW1 The mean change in higher dose gab - 1 week after taper in the

dose gab - 1 week after taper due to risk of bias the control group was intervention groups was 3.0 1.3 lower (2.92 lower to 0.32 higher) Discontinuation emergent signs and syndromes 162 ⊕⊝⊝⊝ RR 1.06 Moderate 1,2 checklist (DESS) after period 1 (1 study) VERY LOW (0.66 to 327 per 1000 20 more per 1000 due to risk of bias, 1.69) (from 111 fewer to 226 more) imprecision DESS after period 2 303 ⊕⊝⊝⊝ RR 0.97 Moderate 1,2 (1 study) VERY LOW (0.66 to 280 per 1000 8 fewer per 1000 due to risk of bias, 1.42) (from 95 fewer to 118 more) imprecision Anxiety after period 1 162 ⊕⊝⊝⊝ RR 0.71 Moderate 1,2 (1 study) VERY LOW (0.12 to 39 per 1000 11 fewer per 1000 due to risk of bias, 4.12) (from 34 fewer to 122 more) imprecision Anxiety after period 2 303 ⊕⊝⊝⊝ RR 0.68 Moderate 1,2 (1 study) VERY LOW (0.28 to 80 per 1000 26 fewer per 1000 due to risk of bias, 1.63) (from 58 fewer to 50 more) imprecision Dizziness after period 1 162 ⊕⊝⊝⊝ Peto OR Moderate 1,2 (1 study) VERY LOW 4.44 0 per 1000 30 more per 1000

Review question and findings and question Review Harms

Headache after period 1 162 ⊕⊝⊝⊝ RR 3.31 Moderate 1,2 (1 study) VERY LOW (0.42 to 19 per 1000 44 more per 1000 due to risk of bias, 26.2) (from 11 fewer to 479 more) imprecision Headache after period 2 303 ⊕⊝⊝⊝ RR 1.97 Moderate 1,2 (1 study) VERY LOW (0.43 to 20 per 1000 19 more per 1000

Physicians 2019 due to risk of bias, 9.11) (from 11 fewer to 162 more) imprecision Insomnia after period 1 162 ⊕⊝⊝⊝ RR 0.47 Moderate

1,2 (1 study) VERY LOW (0.21 to 192 per 1000 102 fewer per 1000 due to risk of bias, 1.06) (from 152 fewer to 12 more) imprecision Insomnia after period 2 303 ⊕⊝⊝⊝ RR 1.72 Moderate 1,2 (1 study) VERY LOW (0.72 to 60 per 1000 43 more per 1000 due to risk of bias, 4.14) (from 17 fewer to 188 more) imprecision Nausea after period 1 162 ⊕⊝⊝⊝ RR 1.65 Moderate 1,2 (1 study) VERY LOW (0.36 to 39 per 1000 25 more per 1000 due to risk of bias, 7.69) (from 25 fewer to 261 more)

38 imprecision

Rebound anxiety after treatment period 1 158 ⊕⊝⊝⊝ RR 0.87 Moderate 1,2 (1 study) VERY LOW (0.17 to 42 per 1000 5 fewer per 1000 due to risk of bias, 4.6) (from 35 fewer to 151 more) imprecision Rebound anxiety after treatment period 2 303 ⊕⊝⊝⊝ RR 0.33 Moderate 1,2 (1 study) VERY LOW (0.09 to 60 per 1000 40 fewer per 1000 due to risk of bias, 1.14) (from 55 fewer to 8 more) imprecision

This table relates to forest plots figures 14-29 (Appendix D).

One RCT(10) comprising 615 participants evaluated pregabalin 450-600mg/d with pregabalin 150-300mg and lorazepam 3-4mg/d. The study had 2 treatment periods both 12 weeks long. The results from the 2 pregabalin doses were combined for comparison with placebo, except for the mean changes in discontinuation symptoms where the higher pregabalin dose was used. Period 1 had a flexible dose in the first 6 weeks and fixed dose in the last 6 weeks. Only participants who showed a clinical response (CGI-I score of 1 or 2) at week 6 continued treatment and the rest discontinued the study. Period 2 was a fixed dose phase of 12 weeks treatment. 25% of patients were randomised to discontinue active medication and received placebo. The results in

Review question and findings and question Review Harms © Royal © College of Physicians 2019 the table above demonstrated that the evidence suggested that there was no difference in drugs for mean change in discontinuation symptoms measured with the Physician Withdrawal Checklist after treatment period 1 or 2 after 1 or 2 weeks of tapering (evidence rated as low quality due to risk of bias of

selection bias and incomplete outcome data due to high attrition rates, based on 1 study). The evidence suggested that there was no difference in discontinuation emergent signs and syndromes checklist (DESS) after period 1 or 2, anxiety after treatment period 1 or 2, dizziness after treatment period 1 or 2, headache after treatment period 1 or 2, insomnia after treatment period 2, nausea after period 1 or rebound anxiety after treatment period 1 or 2 (evidence rated as very low quality due to risk of bias of selection bias and incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study). The evidence suggested a benefit for gabapentinoids in that there were fewer cases of insomnia after treatment period 1 (evidence rated as very low quality due to risk of bias of selection bias and incomplete

outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study).

Table 18: Antidepressants Outcomes No of Quality of the evidence Relative Anticipated absolute effects Participants (GRADE) effect (studies) (95% CI) Risk with Control Risk difference with Antidepressants Follow up (95% CI) Rebound insomnia 221 ⊕⊝⊝⊝ RR 1.97 Moderate 1,2 39 (1 study) VERY LOW (0.22 to 14 per 1000 14 more per 1000

due to risk of bias, 17.34) (from 11 fewer to 229 more) imprecision Benzodiazepine Withdrawal symptom questionnaire criteria 221 ⊕⊝⊝⊝ RR 0.49 Moderate 1,2 BWSQ (1 study) VERY LOW (0.03 to 14 per 1000 7 fewer per 1000 due to risk of bias, 7.77) (from 14 fewer to 95 more) imprecision Suicide attempts (one study after discontinuation; other study 918 ⊕⊝⊝⊝ Peto OR Moderate 2 time point not reported) (2 studies) LOW 1.02 3 per 1000 0 fewer per 1000 due to imprecision (0.09 to (from 10 fewer to 10 more) 12.12) Depression (after discontinuation; except one study time point 1633 ⊕⊝⊝⊝ RR 1.16 Moderate 1,2,3 not reported) (3 studies) VERY LOW (0.15 to 6 per 1000 1 more per 1000 due to risk of bias, 8.76) (from 5 fewer to 47 more) inconsistency, imprecision Suicide ideation - antidepressant v placebo (after 2124 ⊕⊕⊝⊝ Peto OR Moderate 6 1 discontinuation in two studies, one during study and other (4 studies ) LOW 4.24 0 per 1000 0 more per 1000 due to risk of bias, (0.82 to (from 0 fewer to 10 more) time point not reported) 21.90) imprecision DESS - taper week 1 1218 ⊕⊝⊝⊝ The mean DESS - taper week 1 (3 studies) VERY LOW1,2,4 The mean change in in the intervention groups was due to risk of bias, the control group was 0.56 higher (0.01 lower to 1.13 inconsistency, imprecision 1.1 higher)

findings and question Review Harms © Royal © College of Physicians 2019 DESS - taper week 2 1053 ⊕⊕⊕⊝ The mean DESS - taper week 2 (2 studies) MODERATE1 The mean change in in the intervention groups was the control group was due to risk of bias 0.48 higher (0.18 to 0.77 higher) 1.1 DESS - taper week 3 40 ⊕⊕⊝⊝ The mean DESS - taper week 3 (1 study) LOW1 The mean change in in the intervention groups was due to risk of bias the control group was 6 lower (9.56 to 2.45 lower) 7.3 Vertigo after discontinuation 485 ⊕⊕⊝⊝ Peto OR Moderate 1 (1 study) LOW 4.63 0 per 1000 40 more per 1000 due to risk of bias (1.37 to (from 10 fewer to 60 more) 15.6)

Change after discontinuation 106 ⊕⊕⊝⊝ RR 1.27 Moderate 2 (1 study) LOW (0.64 to 208 per 1000 56 more per 1000 due to imprecision 2.54) (from 75 fewer to 320 more)

Discontinuation syndrome 106 ⊕⊕⊝⊝ RR 1.8 Moderate 2 (1 study) LOW (0.65 to 94 per 1000 75 more per 1000 due to imprecision 5.02) (from 33 fewer to 378 more) Total taper/post study emergent AE 3041 ⊕⊕⊝⊝ RR 1.63 Moderate 1 (9 studies) LOW (1.44 to 205 per 1000 129 more per 1000 due to risk of bias 1.84)

40 (from 90 more to 172 more)

Vomiting after discontinuation 306 ⊕⊝⊝⊝ RR 0.35 Moderate 1,2 (1 study) VERY LOW (0.04 to 19 per 1000 12 fewer per 1000 due to risk of bias, 3.34) (from 18 fewer to 44 more) imprecision Dizziness after discontinuation 2479 ⊕⊕⊝⊝ RR 4.86 Moderate 1 (7 studies) LOW (2.91 to 25 per 1000 97 more per 1000 due to risk of bias 8.14) (from 48 more to 179 more) Nausea after discontinuation 2552 ⊕⊝⊝⊝ RR 2.78 Moderate 1,5 (6 studies) VERY LOW (1.36 to 25 per 1000 45 more per 1000 due to risk of bias, 5.69) (from 9 more to 117 more) inconsistency Headache after discontinuation 2329 ⊕⊝⊝⊝ RR 1.39 Moderate 1,2 (5 studies) VERY LOW (0.96 to 2) 44 per 1000 17 more per 1000 due to risk of bias, (from 2 fewer to 44 more) imprecision Insomnia after discontinuation 1166 ⊕⊝⊝⊝ RR 1.37 Moderate 1,2 (3 studies) VERY LOW (0.75 to 19 per 1000 7 more per 1000 due to risk of bias, 2.52) (from 5 fewer to 29 more) imprecision Diarrhoea after discontinuation 306 ⊕⊝⊝⊝ RR 1.05 Moderate

Review question and findings and question Review Harms

imprecision Serious adverse events during taper 859 ⊕⊝⊝⊝ Peto OR Moderate 1,2 (2 studies) VERY LOW 0.40 5 per 1000 0 more per 1000 due to risk of bias, (0.04 to (from 10 fewer to 10 more) imprecision 4.03) Withdrawal syndrome after discontinuation 615 ⊕⊕⊕⊕ RR 2.2 Moderate (1 study) HIGH (1.4 to 119 per 1000 143 more per 1000 3.46) (from 48 more to 293 more) 615 ⊕⊕⊝⊝ RR 3.17 Moderate Upper respiratory tract infection after discontinuation 2 (1 study) LOW (0.75 to 13 per 1000 28 more per 1000 due to imprecision 13.44) (from 3 fewer to 162 more) Suicide ideation (time point not reported) - antidepressant v 453 ⊕⊝⊝⊝ RR 0.5 Moderate 2 antidepressant (1 study) VERY LOW (0.07 to 13 per 1000 6 fewer per 1000 due to risk of bias, 3.55) (from 12 fewer to 33 more) imprecision 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Downgraded by one increment because heterogeneity, I2=65%, p=0.06, unexplained by subgroup analysis. 4 Downgraded by one increment because heterogeneity, I2=72%, p=0.01, unexplained by subgroup analysis.

41 5 Downgraded by one increment because heterogeneity, I2=53%, p=0.05, unexplained by subgroup analysis. 6 One study reported suicide ideation with depression.

This table relates to forest plots figures 30-51 (Appendix D).

Seventeen RCTs compromising 6729 participants’ evaluated antidepressants compared to placebo. The results in the table above suggested that the evidence suggested that the antidepressant group had more cases of withdrawal syndrome (with evidence rated as high quality, based on 1 study), vertigo (with evidence rated as low quality due to risk of bias of selection bias and incomplete outcome data due to high attrition rates, based on 1 study), dizziness (with evidence rated as low quality due to risk of bias of selection, incomplete outcome data due to high attrition rates and incomplete outcome reporting, based on 7 studies) and nausea (with evidence rated as low quality due to risk of bias of selection, incomplete outcome data due to high attrition rates and incomplete outcome reporting and inconsistency due to unexplained heterogeneity across different studies, based on 6 studies) than placebo after discontinuation of treatment.

The evidence suggested that the antidepressant group had more taper/post study emergent adverse events than placebo (with evidence rated as low quality due to risk of bias of selection and incomplete outcome data due to high attrition rates, based on 9 studies). Furthermore, the antidepressant group had a worse mean score on the Discontinuation Signs and Symptoms Scale (DESS) after taper week 1 (with evidence rated as very low quality due to risk of bias of selection and incomplete outcome data due to high attrition rates, inconsistency due to unexplained heterogeneity across different studies and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 3 studies) and week 2 (with

Review question question Review Harms © Royal © College of Physicians 2019 evidence rated as moderate quality due to risk of bias of selection, based on 2 studies), however, they had a better score after week 3 (with evidence rated as low quality due to high risk of bias of incomplete outcome data due to high attrition rates and lack of blinding for patients and investigators, based on 1

study) compared to placebo. At taper week 3 the placebo arm only had 8 participants remaining in the study so this result should be interpreted with caution. The evidence suggested that there were more participants in the antidepressant group who reported noticing a change after medication discontinuation (with evidence rated as low quality due to wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study) compared to the placebo group. findings and

The evidence suggested that there was no clear difference in rebound insomnia (with evidence rated as very low quality due to risk of bias of selection and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study), Benzodiazepine Withdrawal Symptom Questionnaire (BWSQ) criteria score (with evidence rated as very low quality due to risk of bias of selection and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study), suicide attempts (with evidence rated as low quality due wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 2 studies), depression (with evidence rated as very low quality due to risk of bias of selection, inconsistency of effect across the studies and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 3 studies), suicide ideation (with evidence rated as low quality due to risk of bias of selection or incomplete outcome data due to high attrition rates, based on 4 studies – of which 1 reported depression with suicide ideation), upper respiratory tract infection (with evidence rated as low quality due to wide confidence intervals around the effect and therefore uncertainty of where the

42 true effect might lie, based on 1 study), vomiting (with evidence rated as very low quality due to risk of bias of incomplete outcome data due to high

attrition rates, based on 1 study), headache (with evidence rated as very low quality due to risk of bias of selection and incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 5 studies), insomnia (with evidence rated as very low quality due to risk of bias of selection and incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 3 studies), diarrhoea (with evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study) and serious adverse events (with evidence rated very low quality due to risk of bias of incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 2 studies).

One of the RCTs(13) also compared two antidepressants vortioxetine and duloxetine. The results in the table above demonstrated that the evidence suggested that there was no difference in suicide ideation (time point not reported) between the2 antidepressants (with evidence rated as very low quality due to risk of bias of incomplete outcome data due to high attrition rates and wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie, based on 1 study).

Review question and findings and question Review Harms © Royal © College of Physicians 2019 4.2.3 Short term versus long term opioid use

Table 19: Opioids Outcomes No of Quality of the evidence Relative Anticipated absolute effects Participants (GRADE) effect (studies) (95% CI) Risk with Risk difference with Long term opioid use compared to short term opioid use Follow up Control (95% CI) Depression 2892 ⊕⊝⊝⊝ HR 1.53 Moderate 1,2 (1 study) VERY LOW (1.29 to 1.81) 0 per 1000 Not calculable due to risk of bias, indirectness Alcohol abuse 2892 (1 study) ⊕⊝⊝⊝ HR 1.38 Moderate 1,2,3 VERY LOW (0.9 to 2.12) 0 per 1000 Not calculable due to risk of bias, indirectness, imprecision Opioid abuse 2892 ⊕⊝⊝⊝ HR 3.97 Moderate 1,2,3 (1 study) VERY LOW (0.87 to 0 per 1000 Not calculable due to risk of bias, indirectness, 18.12) imprecision Other substance abuse 2892 ⊕⊝⊝⊝ HR 1.81 Moderate 1,2,3 (1 study) VERY LOW (0.92 to 3.56) 0 per 1000 Not calculable 43 due to risk of bias, indirectness,

imprecision Opioid overdose 2892 ⊕⊝⊝⊝ HR 5.12 Moderate 1,2 (1 study) VERY LOW (1.63 to 0 per 1000 Not calculable due to risk of bias, indirectness 16.08) Other substance overdose 2892 ⊕⊝⊝⊝ HR 1.82 Moderate 1,2,3 (1 study) VERY LOW (0.92 to 3.6) 0 per 1000 Not calculable due to risk of bias, indirectness, imprecision Opioid dependence 2892 ⊕⊝⊝⊝ HR 2.85 Moderate 1,2 (1 study) VERY LOW (1.54 to 5.27) 0 per 1000 Not calculable due to risk of bias, indirectness Other substance 2892 ⊕⊝⊝⊝ HR 1.73 Moderate 1,2,3 dependence (1 study) VERY LOW (1.21 to 2.47) 0 per 1000 Not calculable due to risk of bias, indirectness, imprecision Mortality 2892 ⊕⊝⊝⊝ HR 0.99 Moderate 1,2 (1 study) VERY LOW (0.84 to 1.17) 0 per 1000 Not calculable due to risk of bias, indirectness

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 or 2 increments because long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS.

Review question and findings and question Review Harms © Royal © College of Physicians 2019 3 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

This table relates to forest plots figures 52-60 (Appendix D).

One observational study(9) comprising of 2892 people receiving opioid prescriptions were reviewed retrospectively from a database. The study compared long term opioid use as a proxy for dependence (more than 90 days) with shorter term opioid use. The outcomes for this study were downgraded due to indirectness due to the following reasons: long term use opioids was used as a proxy for dependence, short term sue population included children and some opioids The results in the table above demonstrated that the evidence suggested there was a benefit in short term use of opioids in terms of fewer cases of depression, alcohol abuse, opioid abuse, other substance abuse, opioid overdose, other substance overdose, opioid dependence, other substance dependence (evidence rated as very low quality due to risk of bias of confounding, selection bias, classification of interventions and bias due to deviations from intended intervention and indirectness, based on 1 study. Additionally alcohol abuse, opioid abuse, other substance abuse, other substance overdose and other substance dependence had risk of bias of wide confidence intervals around the effect and therefore uncertainty of where the true defect lies). The evidence suggested that there was no difference in mortality (evidence rated as very low quality due to risk of bias of confounding, selection bias, classification of interventions was not defined at the start of intervention and bias due to deviations from intended interventions and indirectness, based on 1 study).

44 4.2.4 Short term versus long term drug withdrawal

Table 20: Antidepressants - Short (3 days) versus longer tapering (14 days) Outcomes No of Participants Quality of the evidence Relative effect Anticipated absolute effects (studies) (GRADE) (95% CI) Follow up Risk with Longer taper Risk difference with Short taper (95% CI) DESS symptoms 28 ⊕⊝⊝⊝ RR 1.01 Moderate 1,2 (1 study) VERY LOW (0.46 to 2.25) 462 per 1000 5 more per 1000 due to risk of bias, imprecision (from 249 fewer to 577 more)

This table relates to forest plots figures 61 (Appendix D).

One RCT(16) comprising 28 participants evaluated a short taper (3 days) of medication compared to a longer tapering period (14 days) for antidepressants. The results in the table above demonstrated that the evidence suggested no difference in DESS symptoms (evidence rated as very low quality due to risk of bias of selection and lack of blinding and wide confidence intervals around the effect and therefore uncertainty of where the true defect lies, based on 1 study).

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Table 21: Antidepressants - Abrupt versus taper of withdrawal (managed withdrawal over one week) Outcomes No of Participants Quality of the evidence Relative effect Anticipated absolute effects (studies) (GRADE) (95% CI) Follow up Risk with Taper Risk difference with Abrupt (95% CI) DESS score 285 ⊕⊝⊝⊝ RR 0.98 Moderate 1,2 (1 study) VERY LOW (0.63 to 1.54) 216 per 1000 4 fewer per 1000 due to risk of bias, imprecision (from 80 fewer to 117 more) Taper/post-therapy-emergent adverse events 285 ⊕⊝⊝⊝ RR 1.32 Moderate 1,2 (1 study) VERY LOW (1.02 to 1.72) 389 per 1000 124 more per 1000 due to risk of bias, imprecision (from 8 more to 280 more) Headache 285 ⊕⊝⊝⊝ RR 2.01 Moderate 1,2 (1 study) VERY LOW (0.94 to 4.29) 65 per 1000 66 more per 1000 due to risk of bias, imprecision (from 4 fewer to 214 more) 285 ⊕⊝⊝⊝ RR 1.43 Moderate Nausea 1,2 (1 study) VERY LOW (0.52 to 3.91) 43 per 1000 18 more per 1000 due to risk of bias, imprecision (from 21 fewer to 125 more) Dizziness 285 ⊕⊝⊝⊝ RR 1.67 Moderate 1,2 (1 study) VERY LOW (0.72 to 3.85) 58 per 1000 39 more per 1000 due to risk of bias, imprecision (from 16 fewer to 165 more)

45 285 ⊕⊝⊝⊝ RR 0.95 Moderate Suicide ideation 1,2 (1 study) VERY LOW (0.06 to 15.07) 7 per 1000 0 fewer per 1000 due to risk of bias, imprecision (from 7 fewer to 98 more) Suicide attempts 285 ⊕⊝⊝⊝ Peto OR 7.04 Moderate 1,2 (1 study) VERY LOW (0.14 to 355.37) 0 per 1000 10 fewer per 1000 due to risk of bias, imprecision (from 10 fewer to 30 more)

This table relates to forest plots figures 62-68 (Appendix D).

Another RCT(11) comprising 285 participants evaluated abrupt withdrawal with managed tapering of medication over 1 week. The evidence in the table above demonstrated that the evidence suggested a benefit for the taper of withdrawal in reduced number of taper/post-therapy emergent adverse events and cases of headache (evidence rated as very low quality due to risk of selection bias and outcome reporting bias and wide confidence intervals around the 3effect and therefore uncertainty of where the true effect lies, based on 1 study). The evidence suggested that there was no difference between groups in DESS score, cases of nausea, dizziness, suicide ideation and suicide attempts (evidence rated as very low quality due to risk of selection bias and outcome reporting bias and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study).

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Table 22: Antidepressants - Rapid (1-7 days) versus gradual withdrawal (2 weeks plus) Outcomes No of Participants Quality of the evidence Relative effect Anticipated absolute effects (studies) (GRADE) (95% CI) Follow up Risk with Gradual withdrawal Risk difference with Rapid withdrawal (95% CI) Time to first new illness 398 ⊕⊝⊝⊝ HR 1.5 Moderate 1,2 (1 study) VERY LOW (1.14 to 1.97) 0 per 1000 Not calculable due to risk of bias, imprecision, indirectness

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Downgraded by 1 increment because the majority of the evidence included drugs grouped together so they were not all listed and could have included drugs not listed on the included list.

This table relates to forest plot figure 69 (Appendix D).

One observational study(3) comprising 398 participants evaluated rapid (1-7 days) versus gradual withdrawal (2 weeks plus) of drug medication. The

evidence in the table above demonstrated that evidence suggested that there may be a benefit for the gradual withdrawal in reduced time to first new illness (depressive/panic episode) within 1 year (evidence rated as very low quality due to risk of bias including selection bias, classification of intervention status, deviations from intended interventions and risk of measurement bias. The outcome was downgraded for indirectness due to the majority of the evidence included drugs grouped together so they were not all listed and could have included drugs not on the included list for this review. Additionally, 46 there were wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study).

Table 23: Antidepressants – Abrupt versus taper of 50 mg alternate days for 2 weeks Outcomes No of Participants Quality of the evidence Relative effect Anticipated absolute effects (studies) (GRADE) (95% CI) Risk with Taper (alternate) Risk difference with Abrupt (95% CI) Follow up DESS after taper week 3 118 ⊕⊕⊝⊝ The mean DESS in the intervention groups was (1 study) LOW1,2 The mean change in the control group was 1.78 1.44 higher due to risk of bias, imprecision (0.04 lower to 2.92 higher) Any adverse events 203 ⊕⊕⊝⊝ RR 1.01 Moderate 1,2 (1 study) LOW (0.78 to 1.31) 520 per 1000 5 more per 1000 due to risk of bias, imprecision (from 114 fewer to 161 more) Asthenia 203 ⊕⊝⊝⊝ RR 1.62 Moderate 1,2 (1 study) VERY LOW (0.55 to 4.77) 49 per 1000 30 more per 1000 due to risk of bias, imprecision (from 22 fewer to 185 more) Diarrhoea 203 ⊕⊝⊝⊝ RR 0.84 Moderate 1,2 (1 study) VERY LOW (0.27 to 2.67) 59 per 1000 9 fewer per 1000 due to risk of bias, imprecision (from 43 fewer to 99 more) Dizziness 203 ⊕⊝⊝⊝ RR 1.26 Moderate

Review question and findings and question Review Harms

Royal College of Physicians 2019 due to risk of bias, imprecision (from 45 fewer to 184 more)

Emotional lability 203 ⊕⊝⊝⊝ RR 0.67 Moderate 1,2 (1 study) VERY LOW (0.2 to 2.31) 59 per 1000 19 fewer per 1000 due to risk of bias, imprecision (from 47 fewer to 77 more) Headache 203 ⊕⊝⊝⊝ RR 0.76 Moderate 1,2 (1 study) VERY LOW (0.33 to 1.72) 118 per 1000 28 fewer per 1000 due to risk of bias, imprecision (from 79 fewer to 85 more) Hypertension 203 ⊕⊝⊝⊝ RR 0.34 Moderate 1,2 (1 study) VERY LOW (0.07 to 1.63) 59 per 1000 39 fewer per 1000 due to risk of bias, imprecision (from 55 fewer to 37 more) Infection 203 ⊕⊝⊝⊝ RR 0.34 Moderate 1,2 (1 study) VERY LOW (0.07 to 1.63) 59 per 1000 39 fewer per 1000 due to risk of bias, imprecision (from 55 fewer to 37 more)

Insomnia 203 ⊕⊝⊝⊝ RR 0.84 Moderate 1,2 (1 study) VERY LOW (0.27 to 2.67) 59 per 1000 9 fewer per 1000 due to risk of bias, imprecision (from 43 fewer to 99 more) Nausea 203 ⊕⊝⊝⊝ RR 1.12 Moderate (1 study) VERY LOW1,2 (0.48 to 2.65)

47 88 per 1000 11 more per 1000 due to risk of bias, imprecision (from 46 fewer to 145 more)

Sweating 203 ⊕⊝⊝⊝ RR 0.43 Moderate 1,2 (1 study) VERY LOW (0.12 to 1.63) 69 per 1000 39 fewer per 1000 due to risk of bias, imprecision (from 61 fewer to 43 more) Vasodilation 203 ⊕⊝⊝⊝ RR 1.18 Moderate 1,2 (1 study) VERY LOW (0.41 to 3.38) 59 per 1000 11 more per 1000 due to risk of bias, imprecision (from 35 fewer to 140 more)

This table relates to forest plots figures 70-82 (Appendix D).

One RCT(26) comprising 384 participants evaluated abrupt withdrawal versus 3 different methods of tapering the withdrawal from desvenlafaxine 100 mg/d. The tapering methods were (1) to drop the dose to 50 mg/d for alternate days for two weeks, (2) to drop the dose to 50 mg/d for 7 days and then further to 25 mg/d for the second week and finally (3) to drop the dose to 50 mg/d for 7 days and then to give placebo for the remaining 7 days of the discontinuation period.

Review question and findings and question Review Harms © Royal © College of Physicians 2019 The results in the table above demonstrated that the evidence suggested that there was no difference in DESS score at week 3 after tapering, any adverse event (evidence rated as low quality due to risk of bias of selection bias and wide confidence intervals around the effect and therefore uncertainty of

where the true effect lies, based on 1 study) or any of the specific adverse events reported by 5% or more of the participants in any group (evidence rated as very low quality due to risk of bias of selection bias and very wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study) between abrupt tapering and tapering on alternate days for 2 weeks.

Table 24: Antidepressants – Abrupt versus taper of 50 mg for 7 days and the placebo for 7 days Outcomes No of Participants Quality of the evidence Relative effect Anticipated absolute effects (studies) (GRADE) (95% CI) Risk with Taper (50 then placebo) Risk difference with Abrupt (95% CI) Follow up DESS after taper week 106 ⊕⊕⊕⊝ The mean DESS in the intervention groups was (1 study) MODERATE1 The mean change in the control group was 1.78 0.08 lower due to risk of bias (1.3 lower to 1.14 higher) Any adverse events 189 ⊕⊝⊝⊝ RR 0.95 Moderate 1,2 (1 study) VERY LOW (0.72 to 1.26) 520 per 1000 26 fewer per 1000 due to risk of bias, imprecision (from 146 fewer to 135 more) Asthenia 189 ⊕⊝⊝⊝ RR 0.47 Moderate 1,2 (1 study) VERY LOW (0.09 to 2.36) 49 per 1000 26 fewer per 1000

48 due to risk of bias, imprecision (from 45 fewer to 67 more) Diarrhoea 189 ⊕⊝⊝⊝ RR 0.98 Moderate 1,2 (1 study) VERY LOW (0.31 to 3.09) 59 per 1000 1 fewer per 1000 due to risk of bias, imprecision (from 41 fewer to 123 more) Dizziness 189 ⊕⊝⊝⊝ RR 0.98 Moderate 1,2 (1 study) VERY LOW (0.44 to 2.15) 118 per 1000 2 fewer per 1000 due to risk of bias, imprecision (from 66 fewer to 136 more) Emotional lability 189 ⊕⊝⊝⊝ RR 0.78 Moderate 1,2 (1 study) VERY LOW (0.23 to 2.68) 59 per 1000 13 fewer per 1000 due to risk of bias, imprecision (from 45 fewer to 99 more) Headache 189 ⊕⊝⊝⊝ RR 0.68 Moderate 1,2 (1 study) VERY LOW (0.28 to 1.66) 118 per 1000 38 fewer per 1000 due to risk of bias, imprecision (from 85 fewer to 78 more) Hypertension 189 ⊕⊝⊝⊝ RR 0.39 Moderate 1,2 (1 study) VERY LOW (0.08 to 1.89) 59 per 1000 36 fewer per 1000 due to risk of bias, imprecision (from 54 fewer to 53 more) Infection 189 ⊕⊝⊝⊝ RR 0.59 Moderate 1,2 (1 study) VERY LOW (0.15 to 2.28) 59 per 1000 24 fewer per 1000 due to risk of bias, imprecision (from 50 fewer to 76 more)

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Insomnia 189 ⊕⊝⊝⊝ RR 0.98 Moderate 1,2 (1 study) VERY LOW (0.31 to 3.09) 59 per 1000 1 fewer per 1000 due to risk of bias, imprecision (from 41 fewer to 123 more) Nausea 189 ⊕⊝⊝⊝ RR 0.91 Moderate 1,2 (1 study) VERY LOW (0.35 to 2.35) 88 per 1000 8 fewer per 1000 due to risk of bias, imprecision (from 57 fewer to 119 more) Sweating 189 ⊕⊝⊝⊝ RR 0.67 Moderate 1,2 (1 study) VERY LOW (0.2 to 2.21) 69 per 1000 23 fewer per 1000 due to risk of bias, imprecision (from 55 fewer to 83 more) 189 ⊕⊝⊝⊝ RR 0.2 Moderate

Vasodilation 1,2 (1 study) VERY LOW (0.02 to 1.59) 59 per 1000 47 fewer per 1000 due to risk of bias, imprecision (from 58 fewer to 35 more)

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

This table relates to forest plots figures 83-95 (Appendix D).

One RCT(26) comprising 384 participants evaluated abrupt withdrawal versus 3 different methods of tapering the withdrawal from desvenlafaxine 100 49 mg/d. The tapering methods were (1) to drop the dose to 50 mg/d for alternate days for two weeks, (2) to drop the dose to 50 mg/d for 7 days and then

further to 25 mg/d for the second week and finally (3) to drop the dose to 50 mg/d for 7 days and then to give placebo for the remaining 7 days of the discontinuation period.

The results in the table above demonstrated that the evidence suggested that tapering desvenlafaxine to 50 mg/d for one week followed by placebo for a week may have a benefit in reduced cases of vasodilation compared to abrupt discontinuation of treatment (evidence rated as very low quality due to risk of bias of selection bias and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study). The evidence suggested that there was no difference between DESS score at week 3 of tapering, any adverse event or any of the specific adverse events reported by 5% or more of the participants in any group (evidence rated as very low quality due to risk of bias of selection bias and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study).

Table 25: Antidepressants – Abrupt versus taper of 50 mg for one week and then 25 mg for 7 days Outcomes No of Participants Quality of the evidence Relative effect Anticipated absolute effects (studies) (GRADE) (95% CI) Risk with Taper (50 then 25) Risk difference with Abrupt (95% CI) Follow up DESS after taper week 116 ⊕⊕⊝⊝ The mean DESS in the intervention groups was (1 study) LOW1,2 The mean change in the control group was 1.78 2.33 higher due to risk of bias, imprecision (0.62 to 4.04 higher)

Review question and findings and question Review Harms © Royal © College of Physicians 2019 Any adverse events 196 ⊕⊝⊝⊝ RR 0.96 Moderate 1,2 (1 study) VERY LOW (0.73 to 1.27) 520 per 1000 21 fewer per 1000 due to risk of bias, imprecision (from 140 fewer to 140 more) Asthenia 196 ⊕⊕⊝⊝ OR 0.14 Moderate 1,2 (1 study) LOW (0.02 to 0.83) 49 per 1000 42 fewer per 1000 due to risk of bias, imprecision (from 8 fewer to 48 fewer) Diarrhoea 196 ⊕⊝⊝⊝ RR 0.54 Moderate 1,2 (1 study) VERY LOW (0.14 to 2.11) 59 per 1000 27 fewer per 1000 due to risk of bias, imprecision (from 51 fewer to 65 more) 196 ⊕⊝⊝⊝ RR 0.72 Moderate

Dizziness 1,2 (1 study) VERY LOW (0.31 to 1.69) 118 per 1000 33 fewer per 1000 due to risk of bias, imprecision (from 81 fewer to 81 more) Emotional lability 196 ⊕⊝⊝⊝ RR 0.18 Moderate (1 study) 1,2 (0.02 to 1.47) VERY LOW 59 per 1000 48 fewer per 1000 due to risk of bias, imprecision (from 58 fewer to 28 more) Headache 196 ⊕⊝⊝⊝ RR 0.63 Moderate 1,2 (1 study) VERY LOW (0.26 to 1.54) 118 per 1000 44 fewer per 1000 due to risk of bias, imprecision (from 87 fewer to 64 more)

50 Hypertension 196 ⊕⊕⊕⊝ OR 0.14 Moderate (1 study) MODERATE1 (0.03 to 0.7) 59 per 1000 50 fewer per 1000 due to risk of bias (from 17 fewer to 57 fewer) Infection 196 ⊕⊝⊝⊝ RR 0.36 Moderate 1,2 (1 study) VERY LOW (0.07 to 1.75) 59 per 1000 38 fewer per 1000 due to risk of bias, imprecision (from 55 fewer to 44 more) Insomnia 196 ⊕⊝⊝⊝ RR 0.9 Moderate 1,2 (1 study) VERY LOW (0.29 to 2.87) 59 per 1000 6 fewer per 1000 due to risk of bias, imprecision (from 42 fewer to 110 more) Nausea 196 ⊕⊝⊝⊝ RR 1.09 Moderate 1,2 (1 study) VERY LOW (0.45 to 2.62) 88 per 1000 8 more per 1000 due to risk of bias, imprecision (from 48 fewer to 143 more) Sweating 196 ⊕⊝⊝⊝ RR 0.31 Moderate 1,2 (1 study) VERY LOW (0.07 to 1.46) 69 per 1000 48 fewer per 1000 due to risk of bias, imprecision (from 64 fewer to 32 more) Vasodilation 196 ⊕⊝⊝⊝ RR 0.9 Moderate 1,2 (1 study) VERY LOW (0.29 to 2.87) 59 per 1000 6 fewer per 1000 due to risk of bias, imprecision (from 42 fewer to 110 more)

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

Review question and findings and question Review Harms © Royal © College of Physicians 2019 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

This table relates to forest plots figures 96-108 (Appendix D).

One RCT(26) comprising 384 participants evaluated abrupt withdrawal versus three different methods of tapering the withdrawal from desvenlafaxine 100 mg/d. The tapering methods were (1) to drop the dose to 50 mg/d for alternate days for 2 weeks, (2) to drop the dose to 50 mg/d for 7 days and then further to 25 mg/d for the second week and finally (3) to drop the dose to 50 mg/d for 7 days and then to give placebo for the remaining 7 days of the discontinuation period.

The results from the table above demonstrated that the evidence suggested that tapering desvenlafaxine to 50 mg/d for 1 week followed by 25 mg/d for a week may have a benefit in fewer cases of hypertension (evidence rated as moderate quality due to risk of selection bias), sweating and emotional lability (evidence rated as very low quality due to risk of selection bias and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study). The evidence suggested that the abrupt discontinuation group may have a benefit in a lower DESS score after taper

week 3 than the taper group of 50mg/d for 1 week followed by 25 mg for 1 week (evidence rated as low quality due to risk of selection bias and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study). The evidence suggested that there was no difference between the groups for any of the specific adverse events reported by 5% or more of the participants in any group (evidence rated as very low quality except asthenia which was low quality due to risk of selection bias and wide confidence intervals around the effect and therefore uncertainty 51 of where the true effect lies, based on 1 study). or any adverse event (evidence rated as very low quality due to risk of selection bias and wide confidence intervals around the effect and therefore uncertainty of where the true effect lies, based on 1 study).

Harms

Review question and findings

4.3 Narrative findings

Thirteen studies reported outcome data that was in insufficient detail to be fully analysed and quality assessed within this report. It should be considered at very high risk of bias. The outcomes are listed below but please see evidence tables in Appendix E for full details.

4.3.1 Opioids

Two opioid studies(8, 17) reported additional narrative evidence on the SOWS scores (higher scores worse). The first study(8) reported mean total subjective opioid withdrawal scale (SOWS) score of 6.9 in tapentadol (n=308) and 8.7 in oxycodone (n=67) (1 study, high risk of bias). The other study(17) reported mean range of SOWS total scores from 2, 3, 4 and 5 or more days after discontinuation. The tapentadol group (n=164) range was 6.9-9.5 and the oxycodone group (n=46) was 7.5-12.3 (1 study, very high risk of bias).

4.3.2 Z drugs

One study(7) reported withdrawal signals after a 2 week trial of zolpidem compared to placebo. The proportion of patients (%) with withdrawal signal (defined as ≥3 new or worsened withdrawal symptoms according to the Benzodiazepine Withdrawal Symptom Questionnaire), using day 15 as baseline; analysis based on group of patients that were switched from active treatment to placebo during the run-out phase: • Day 1 of run-out phase: Zolpidem: 5.0%, Placebo: 11.4% • Day 2 of run-out phase: Zolpidem: 1.7%, Placebo: 12.8% • Day 3 of run-out phase: Zolpidem: 8.6%, Placebo: 14.5% • Day 1-3 of run-out phase: Zolpidem: 16.7%, Placebo: 29.9%

The study only provided percentages without the total number of patients.

Authors state that during the run-out phase approximately 25% of patients had adverse events. The incidence of adverse events in the run-out phase was similar in all treatment groups. Discontinuation of active treatment did not appear to induce additional adverse events (1 study, low risk of bias).

4.3.3 Antidepressants

Four studies comparing antidepressants and placebo reported the discontinuation emergent signs and symptoms scale (DESS).(13, 22, 29, 32)

The first study(32) reported DESS scores during taper period (7 days after the end of the treatment period).

Actual outcome: Mean (SD) DESS checklist – 1 week after treatment period

Desvenlafaxine 50mg/d (n=146): 2.14 (3.64), Desvenlafaxine 100mg/d (n=128): NR, Placebo (n=146): 0.86 (2.01)

Actual outcome: Mean (SD) DESS checklist – 2 week after treatment period

Desvenlafaxine 50mg/d (n=146): NR, Desvenlafaxine 100mg/d (n=128): 1.36 (2.72), Placebo (n=146): 0.62 (1.41)

Review question and findings

The authors only reported figures for groups with significant differences. No figures were provided for week 3 after treatment. Authors note that DESS scores were not significantly different between desvenlafaxine 50 mg/day and placebo groups at weeks 2 and 3 after treatment. Desvenlafaxine 100 mg/day group showed a significantly higher mean DESS score compared with the placebo group only during week 2 after treatment (at which point patients went from receiving 50mg/day to no active treatment); differences from placebo were not statistically significant at week 1 or 3 after treatment(1 study, very high risk of bias).

Another study486 reported DESS scores by completers (≥ 53days of on therapy study drug) at 1 week after taper – day 63 (n=316) and at follow up visit day 70 (n=300).

Narrative: There was a significant increase (+2.1, p=0.001) in mean DESS scores for the desvenlafaxine 50-mg dose group compared with the placebo group (+0.2) at the end of week 1 of the taper period after patients had received placebo for one week. There was no significant differences in DESS scores between the desvenlafaxine 100 mg/day dose group and the placebo group after 1 week at the 50-mg/day dose (+0.6) or 1 week after discontinuation from all active treatment (+1.6) (1 study, very high risk of bias).

The third study(13) comparing vortioxetine 15mg, vortioxetine 20mg, duloxetine and placebo reported DESS scores during the 2 week discontinuation period (1 study, high risk of bias): – Week 9: Vortioxetine 15mg: 1.8, Vortioxetine 20mg: 1.8, Duloxetine: 1.5, Placebo: 1.1 – Week 10: Vortioxetine 15mg: 2.5, Vortioxetine 20mg: 1.6, Duloxetine: 2.7, Placebo: 1.7

The fourth study(29) comparing desvenlafaxine with placebo reported mean (SD) DESS scores after the post open label period where all patients received desvenlafaxine. However, the study split the groups into two different doses of desvenlafaxine (400mg and 200mg) but did not give total number in each group so further analysis was not possible.

Actual outcome: Mean (SD) discontinuation Emergent signs and symptoms (DESS) score at end of open label period.

Week 3 of double blind treatment phase:

Group 1: 400 mg/d dose=1.15 (2.61) and 200 mg/d dose=0.73 (1.76), Group 2: 400 mg/d dose=2.16 (3.60) and 200 mg/d dose=2.28 (3.69).

After week 3 of double blind treatment phase:

Group 1: 400 mg/d dose=1.14 (2.37), Group 2: 400 mg/d dose=3.11 (4.51).

The study only reported figures for 400mg/d dose and stated that there was no significant difference for 200mg/d dose.

Actual outcome: Mean (SD) discontinuation Emergent signs and symptoms (DESS) score at the end of the double blind treatment phase – higher score worse.

Week 3 of taper:

Group 1: 400 mg/d dose=1.75 (3.35), Group 2: 400 mg/d dose=0.59 (1.80).

The authors reported that there were no significant differences in mean DESS scores in the 200 mg/d desvenlafaxine group compared with the placebo group at any point during the taper period (1 study, very high risk of bias).

Review question and findings

Other outcomes reported: One study comparing antidepressants and placebo reported some frequent post treatment emergent adverse events.(20) The study provided percentages but the crude figures could not be calculated for further analysis.

Headache: duloxetine: 3.1%, placebo: 1.2%

Dizziness: duloxetine: 1.9%, placebo: 1.2%

Fatigue: duloxetine: 2.5%, placebo: 0%

Nausea: duloxetine: 2.5%, placebo: 0%

(1 study, very high risk of bias).

One study comparing antidepressants and placebo reported a treatment discontinuation signs and symptoms assessment score (TDSS) which was developed by AstraZeneca.(21) It is an 18 item scale but it is unclear if higher or lower is better and what the maximum possible score is. The range of total mean scores was 2.4 to 4.3 for duloxetine and 1.6 to 2.9 for placebo (1 study, very high risk of bias).

Tapering duration

One study(11) compared abrupt withdrawal and tapering over one week from antidepressants reported narratively the adjusted mean difference for total discontinuation emergent signs and syndromes checklist (DESS), although it is unclear what has been adjusted for. Adjusted mean difference (95% CI) for total DESS scale score for taper vs. abrupt discontinuation group: -0.50 (-1.89 to 0.88). Baseline values for each group were 0.6 (1.2, n=139) for taper, 1.1 (2.41, n=146) for abrupt discontinuation. One study(16) compared short term (3 days) and longer term (14 days) tapering narratively reported the Montgomery Asberg Depression Rating Scale. Figures not reported but authors comment that MADRS values were numerically slightly higher in the short taper group at baseline and this difference increased at assessment 5-7 days after stopping medication but no statistically significant difference between the groups overall (p=0.497). Most common new or worsened DESS items after discontinuation were dizziness (42%), headache (42%), nervousness/anxiety (42%), panic/sudden anxiety (32%), agitation (32%), nausea (32%) and sudden worsening of mood (32%) across all patients (1 study, very high risk of bias).

One study(3) comparing rapid discontinuation (1-7 days) and gradual discontinuation (2 weeks or more) of antidepressants reported median time to new illness (depressive/panic episode) within one year. Median time to new illness was 3.77 months (95% CI, 1.99-5.55) for rapid discontinuation and 9.07 months (95% CI, 6.78-11.4) for gradual discontinuation (1 study, very high risk of bias).

One study(26) comparing abrupt discontinuation with 3 different tapering methods reported specific adverse events at tapering week 1, 2 and 3 and total number of adverse events (1 study, high risk of bias). This report has analysed outcomes for specific adverse events that were reported by ≥ 5% of participants after discontinuation. The overall adverse events were reported after discontinuation rather than at each time point. Please see the evidence table (Appendix E) for further information.

Benzodiazepines

One observational non comparative study(4) reported on 38 nursing home residents who were chronic users of benzodiazepines whose GP agreed to initiate discontinuation of benzodiazepines. The GP was responsible for the discontinuation, but a possible discontinuation schedule was proposed in the recruitment letter as a 25% reduction per week or per 2 weeks.

Review question and findings

Number of participants relapsed at 2 months: 6/38; very high risk of bias.

Number of participants relapsed at 8 months: 7/38; very high risk of bias.

Real increase of withdrawal symptoms – Scored by the Benzodiazepine Withdrawal Symptom Questionnaire (BWSQ): 0/38; very high risk of bias.

BWSQ score, mean (SD): Baseline - 3.9 (2.8); After discontinuation (time-point unknown) – 4.1 (2.6); very high risk of bias.

Muscle pain: Baseline 49%; After discontinuation 63%; very high risk of bias.

Muscle twitching: Baseline 41%; After discontinuation 37%; very high risk of bias.

Loss of memory: Baseline 41%; After discontinuation 32%; very high risk of bias.

Pins and needles: Baseline 35%; After discontinuation 39%; very high risk of bias.

Feeling depressed: Baseline 32%; After discontinuation 37%; very high risk of bias.

The evidence suggested that there was a high rate of relapse but no increase in withdrawal symptoms. Muscle pain, depression and pins and needles increased after discontinuation but muscle twitch and memory loss was reduced in this cohort of patients (1 study, very high risk of bias).

Antidepressants

One RCT study(19) reported on 252 women receiving sertraline hydrochloride 50-100 mg/d during symptomatic interval for 6 menstrual cycles (n=125) compared to placebo (n-127).

The study reported the Michelson SSRI withdrawal Scale where the mean of each item was summed for the 3 days after treatment ended for each menstrual cycle. Scores ranged from 0-51 with higher scores indicating more severe withdrawal symptoms.

Cycle 1: Group 1: 72.9 (29.3), Group 2: 68.1 (25.4); very high risk of bias.

Cycle 2: Group 1: 73.4 (29.9), Group 2: 75.2 (33.1); very high risk of bias.

Cycle 3: Group 1: 70.3 (27.9), Group 2: 71.7 (39.0); very high risk of bias.

Cycle 4: Group 1: 62.3 (25.1) Group 2: 65.4 (25.3); very high risk of bias.

Cycle 5: Group 1: 63.9 (19.9) Group 2: 62.1 (24.0); very high risk of bias.

Cycle 6: Group 1: 67.5 (22.1) Group 2: 76.2 (34.1); very high risk of bias.

Note: it was unclear from the study what units these outcomes were reported.

The narrative evidence reported a Michelson SSRI withdrawal Scale for the 3 days after treatment ended for each menstrual cycle. The scores were varied across the cycles and it was unclear what the units reported were (1 study, very high risk of bias).

Conclusions

5 Conclusions

There is a lack of good quality quantitative evidence of the harms of dependency of the prescription medications included in this review. The only available evidence informing this area was comparing 2 different opioids and looked at whether their use led to ‘Doctor shopping behaviour’ in terms of visiting multiple prescribers to obtain more opioids, and a retrospective database review of short versus long term opioid use. Although this provides evidence for the dependence forming nature of these medications (and indicated that oxycodone may be more harmful that tapentadol), only the later of the 2 studies gave any indication of the specific harms experienced; suggesting that longer term use led to more incidence of depression, alcohol abuse, opioid abuse, other substance abuse, opioid overdose, other substance overdose, opioid dependence and other substance dependence. It is unclear whether these are due to dependence on the opioid however, as long term use was used as a proxy for dependence.

There was limited evidence on short term opioid withdrawal compared to placebo, but there were more opioid users with mild or moderate opioid withdrawal using the Clinical Opioid Withdrawal Scale (COWS) than placebo at 2-4 days after last intake of medication. However, this difference was no longer evident by days 4 and 5 plus.

The majority of the available evidence was for harms from withdrawal or discontinuation of these medications. This evidence again was limited, with the most evidence being available for withdrawal from antidepressants.

Good quality recent evidence was lacking for benzodiazepines for all areas covered in the review. The only available evidence was 1 randomised trial comparing benzodiazepines to gabapentinoids, and 1 non-randomised study looking at GP led withdrawal of medication in a selected group of nursing home residents. All of this evidence was considered low to very low quality, meaning firm conclusions cannot be drawn from the effects observed and it is likely that more evidence would change the conclusions drawn from this. Similarly, there was only 1 study identified for z-drugs. This very limited recent evidence base for these drugs is insufficient for drawing conclusions.

Where evidence was available, it was primarily pharmaceutical funded trials designed to look at efficacy of medications rather than to study withdrawal or discontinuation from the drugs. Consequently, the follow-up period after stopping the medication was usually only short term or up until the end of the taper phase and no long term harms can be inferred from this evidence. Such RCTs are also often underpowered to detect harms that may have relatively low event rates. Furthermore, the reporting was often only of checklists or scores, or incomplete reporting of the symptoms.

Irrespective of the limitations of the available evidence, there was a clear suggestion that when compared to placebo, withdrawal from antidepressants may lead to more people experiencing withdrawal syndrome, including taper/post-study emergent adverse events such as vertigo, dizziness and nausea. Whether or not a gradual tapering schedule minimised the adverse effects experiences was inconclusive from this evidence, although there did appear to be some advantage compared to abruptly stopping the medication.

The evidence was frequently rated as low or very low quality by GRADE criteria. The most common reasons for downgrading were risk of bias due to incomplete outcome data (high or differential attrition rates in the studies) and imprecision; evidenced by wide confidence intervals around the effect and therefore uncertainty of where the true effect might lie. The implication of this rating is that it is very likely that further research would have an important impact on our confidence in the estimate of effect. Further research is required to provide robust conclusions on the specific harms of dependence, discontinuation and withdrawal from opioids, benzodiazepines, Z-drugs and

Conclusions gabapentinoids. It would also be beneficial to increase confidence in the findings from this review and to further clarify the harms experienced in withdrawing/discontinuing from long term opioid use and antidepressants, including longer term outcomes for all groups of drugs included.

References

6 References

1. GRADE Working Group. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group website 2011 [updated 10/1/2011. Available from: http://www.gradeworkinggroup.org/.

2. Afilalo M, Etropolski MS, Kuperwasser B, Kelly K, Okamoto A, Van Hove I, et al. Efficacy and safety of Tapentadol extended release compared with oxycodone controlled release for the management of moderate to severe chronic pain related to osteoarthritis of the knee: a randomized, double-blind, placebo- and active-controlled phase III study. Clin Drug Investig. 2010;30(8):489-505.

3. Baldessarini RJ, Tondo L, Ghiani C, Lepri B. Illness risk following rapid versus gradual discontinuation of antidepressants. The American Journal of Psychiatry. 2010;167(8):934-41.

4. Bourgeois J, Elseviers MM, Van Bortel L, Petrovic M, Vander Stichele RH. Feasibility of discontinuing chronic benzodiazepine use in nursing home residents: a pilot study. Eur J Clin Pharmacol. 2014;70(10):1251-60.

5. Buynak R, Shapiro DY, Okamoto A, Van Hove I, Rauschkolb C, Steup A, et al. Efficacy and safety of tapentadol extended release for the management of chronic low back pain: results of a prospective, randomized, double-blind, placebo- and active-controlled Phase III study.[Erratum appears in Expert Opin Pharmacother. 2010 Nov;11(16):2773]. Expert Opin Pharmacother. 2010;11(11):1787-804.

6. Cepeda MS, Fife D, Vo L, Mastrogiovanni G, Yuan Y. Comparison of opioid doctor shopping for tapentadol and oxycodone: a cohort study. J Pain. 2013;14(2):158-64.

7. Hajak G, Hedner J, Eglin M, Loft H, Storustovu SI, Lutolf S, et al. A 2-week efficacy and safety study of gaboxadol and zolpidem using electronic diaries in primary insomnia outpatients. Sleep Med. 2009;10(7):705-12.

8. Hale M, Upmalis D, Okamoto A, Lange C, Rauschkolb C. Tolerability of tapentadol immediate release in patients with lower back pain or osteoarthritis of the hip or knee over 90 days: a randomized, double-blind study. Curr Med Res Opin. 2009;25(5):1095-104.

9. Hoffman EM, Watson JC, St Sauver J, Staff NP, Klein CJ. Association of long-term opioid therapy with functional status, adverse outcomes, and mortality among patients with polyneuropathy. JAMA Neurology. 2017;74(7):773-9.

10. Kasper S, Iglesias-Garcia C, Schweizer E, Wilson J, DuBrava S, Prieto R, et al. Pregabalin long- term treatment and assessment of discontinuation in patients with generalized anxiety disorder. Int J Neuropsychopharmacol. 2014;17(5):685-95.

11. Khan A, Musgnung J, Ramey T, Messig M, Buckley G, Ninan PT. Abrupt discontinuation compared with a 1-week taper regimen in depressed outpatients treated for 24 weeks with desvenlafaxine 50 mg/d. J Clin Psychopharmacol. 2014;34(3):365-8.

12. Krystal AD, Lankford A, Durrence HH, Ludington E, Jochelson P, Rogowski R, et al. Efficacy and safety of doxepin 3 and 6 mg in a 35-day sleep laboratory trial in adults with chronic primary insomnia. Sleep. 2011;34(10):1433-42.

13. Mahableshwarkar AR, Jacobsen PL, Chen Y, Serenko M, Trivedi MH. A randomized, double- blind, duloxetine-referenced study comparing efficacy and tolerability of 2 fixed doses of vortioxetine in the acute treatment of adults with MDD. Psychopharmacology (Berl). 2015;232(12):2061-70.

References

14. Nishimura A, Aritomi Y, Sasai K, Kitagawa T, Mahableshwarkar AR. Randomized, double-blind, placebo-controlled 8-week trial of the efficacy, safety, and tolerability of 5, 10, and 20 mg/day vortioxetine in adults with major depressive disorder. Psychiatry Clin Neurosci. 2018;72(2):64-72.

15. Stein DJ, Ahokas A, Albarran C, Olivier V, Allgulander C. Agomelatine prevents relapse in generalized anxiety disorder: a 6-month randomized, double-blind, placebo-controlled discontinuation study. J Clin Psychiatry. 2012;73(7):1002-8.

16. Tint A, Haddad PM, Anderson IM. The effect of rate of antidepressant tapering on the incidence of discontinuation symptoms: a randomised study.[Erratum appears in J Psychopharmacol. 2009 Nov;23(8):1006]. Journal of Psychopharmacology. 2008;22(3):330-2.

17. Wild JE, Grond S, Kuperwasser B, Gilbert J, McCann B, Lange B, et al. Long-term safety and tolerability of tapentadol extended release for the management of chronic low back pain or osteoarthritis pain. Pain Pract. 2010;10(5):416-27.

18. YaDeau JT, Brummett CM, Mayman DJ, Lin Y, Goytizolo EA, Padgett DE, et al. Duloxetine and Subacute Pain after Knee Arthroplasty when Added to a Multimodal Analgesic Regimen: A Randomized, Placebo-controlled, Triple-blinded Trial. Anesthesiology. 2016;125(3):561-72.

19. Yonkers KA, Kornstein SG, Gueorguieva R, Merry B, Van Steenburgh K, Altemus M. Symptom- Onset Dosing of Sertraline for the Treatment of Premenstrual Dysphoric Disorder: A Randomized Clinical Trial. JAMA psychiatry. 2015;72(10):1037-44.

20. Raskin J, Wiltse CG, Dinkel JJ, Walker DJ, Desaiah D, Katona C. Safety and tolerability of duloxetine at 60 mg once daily in elderly patients with major depressive disorder. J Clin Psychopharmacol. 2008;28(1):32-8.

21. Cutler AJ, Montgomery SA, Feifel D, Lazarus A, Astrom M, Brecher M. Extended release quetiapine fumarate monotherapy in major depressive disorder: a placebo- and duloxetine- controlled study.[Erratum appears in J Clin Psychiatry. 2009 Dec;70(12):1729]. J Clin Psychiatry. 2009;70(4):526-39.

22. Liebowitz MR, Manley AL, Padmanabhan SK, Ganguly R, Tummala R, Tourian KA. Efficacy, safety, and tolerability of desvenlafaxine 50 mg/day and 100 mg/day in outpatients with major depressive disorder. Curr Med Res Opin. 2008;24(7):1877-90.

23. Rynn M, Russell J, Erickson J, Detke MJ, Ball S, Dinkel J, et al. Efficacy and safety of duloxetine in the treatment of generalized anxiety disorder: a flexible-dose, progressive-titration, placebo- controlled trial. Depress Anxiety. 2008;25(3):182-9.

24. Schagen van Leeuwen JH, Lange RR, Jonasson AF, Chen WJ, Viktrup L. Efficacy and safety of duloxetine in elderly women with stress urinary incontinence or stress-predominant mixed urinary incontinence. Maturitas. 2008;60(2):138-47.

25. Rosenthal JZ, Boyer P, Vialet C, Hwang E, Tourian KA. Efficacy and safety of desvenlafaxine 50 mg/d for prevention of relapse in major depressive disorder:a randomized controlled trial. J Clin Psychiatry. 2013;74(2):158-66.

26. Gallagher JC, Strzinek RA, Cheng RF, Ausmanas MK, Astl D, Seljan P. The effect of dose titration and dose tapering on the tolerability of desvenlafaxine in women with vasomotor symptoms associated with menopause. J Womens Health. 2012;21(2):188-98.

27. Perahia DG, Maina G, Thase ME, Spann ME, Wang F, Walker DJ, et al. Duloxetine in the prevention of depressive recurrences: a randomized, double-blind, placebo-controlled trial.[Erratum appears in J Clin Psychiatry. 2009 Dec;70(12):1729]. J Clin Psychiatry. 2009;70(5):706-16.

References

28. Liebowitz MR, Asnis G, Mangano R, Tzanis E. A double-blind, placebo-controlled, parallel- group, flexible-dose study of venlafaxine extended release capsules in adult outpatients with panic disorder. J Clin Psychiatry. 2009;70(4):550-61.

29. Rickels K, Montgomery SA, Tourian KA, Guelfi JD, Pitrosky B, Padmanabhan SK, et al. Desvenlafaxine for the prevention of relapse in major depressive disorder: results of a randomized trial. J Clin Psychopharmacol. 2010;30(1):18-24.

30. Tourian KA, Padmanabhan SK, Groark J, Brisard C, Farrington D. Desvenlafaxine 50 and 100 mg/d in the treatment of major depressive disorder: an 8-week, phase III, multicenter, randomized, double-blind, placebo-controlled, parallel-group trial and a post hoc pooled analysis of three studies. Clin Ther. 2009;31 Pt 1:1405-23.

31. Chappell AS, Littlejohn G, Kajdasz DK, Scheinberg M, D'Souza DN, Moldofsky H. A 1-year safety and efficacy study of duloxetine in patients with fibromyalgia. Clin J Pain. 2009;25(5):365-75.

32. Boyer P, Montgomery S, Lepola U, Germain JM, Brisard C, Ganguly R, et al. Efficacy, safety, and tolerability of fixed-dose desvenlafaxine 50 and 100 mg/day for major depressive disorder in a placebo-controlled trial. Int Clin Psychopharmacol. 2008;23(5):243-53.

33. Kiepek N, Hancock L, Toppozini D, Cromarty H, Morgan A, Kelly L. Facilitating medical withdrawal from opiates in rural Ontario. Rural Remote Health. 2012;12:2193.

34. Cepeda MS, Fife D, Ma Q, Ryan PB. Comparison of the risks of opioid abuse or dependence between tapentadol and oxycodone: Results from a cohort study. J Pain. 2013;14(10):1227-41.

35. Abrahms JL. A cognitive-behavioral versus nondirective group treatment program for opioid- addicted persons: an adjunct to methadone maintenance. Int J Addict. 1979;14(4):503-11.

36. Adi Y, Juarez-Garcia A, Wang D, Jowett S, Frew E, Day E, et al. Oral naltrexone as a treatment for relapse prevention in formerly opioid-dependent drug users: a systematic review and economic evaluation. Health Technol Assess. 2007;11(6).

37. Aguiluz J, Alvarez M, Pimentel E, Abarca C, Moore P. How to face a patient with benzodiazepine dependence in primary health care? Strategies for withdrawal. Medwave. 2018;18(1):e7159.

38. Ahmadi J. A randomized, clinical trial of buprenorphine maintenance treatment for Iranian patients with opioid dependency. Addictive disorders & their treatment. 2002;1(1):25‐7.

39. Ahmadi J, Ahmadi K, Ohaeri J. Controlled, randomized trial in maintenance treatment of intravenous buprenorphine dependence with naltrexone, methadone or buprenorphine: a novel study. Eur J Clin Invest. 2003;33(9):824-9.

40. Ahmadi J, Ahmadi K. Controlled trial of maintenance treatment of intravenous buprenorphine dependence. Ir J Med Sci. 2003;172(4):171-3.

41. Ahmadi J, Jahromi MS, Ehsaei Z. The effectiveness of different singly administered high doses of buprenorphine in reducing suicidal ideation in acutely depressed people with co-morbid opiate dependence: a randomized, double-blind, clinical trial. Trials. 2018;19(1):462.

42. Ahmadi J, Sefidfard Jahromi M. Ultrarapid Influence of Buprenorphine on Major Depression in Opioid-Dependent Patients: A Double Blind, Randomized Clinical Trial. Subst Use Misuse. 2018;53(2):286-9.

43. Ahmed M, Westra HA, Stewart SH. A Self-Help Handout for Benzodiazepine Discontinuation Using Cognitive Behavioral Therapy. Cognitive and Behavioral Practice. 2008;15(3):317-24.

References

44. Alford DP. Chronic back pain with possible prescription opioid misuse. JAMA. 2013;309(9):919-25.

45. Ali MM, Dowd WN, Classen T, Mutter R, Novak SP. Prescription drug monitoring programs, nonmedical use of prescription drugs, and heroin use: Evidence from the National Survey of Drug Use and Health. Addict Behav. 2017;69:65-77.

46. Alter TH, Ilyas AM. A Prospective Randomized Study Analyzing Preoperative Opioid Counseling in Pain Management After Carpal Tunnel Release Surgery. J Hand Surg [Am]. 2017;42(10):810-5.

47. Altman RD, Smith HS. Opioid therapy for osteoarthritis and chronic low back pain. Postgrad Med. 2010;122(6):87-97.

48. Amanti CA. Benzodiazepine taper guidelines for older adults in an inpatient geriatric/psychiatric unit [D.N.P.]. Arizona: The University of Arizona; 2018.

49. Amarasuriya UK, Myles PR, Sanders RD. Long-term benzodiazepine use and mortality: are we doing the right studies? Current drug safety. 2012;7(5):367-71.

50. Amato L, Davoli M, Minozzi S, Ferroni E, Ali R, Ferri M. Methadone at tapered doses for the management of opioid withdrawal. Cochrane Database of Systematic Reviews [Internet]. 2013; (2). Available from: http://dx.doi.org/10.1002/14651858.CD003409.pub4.

51. Amato L, Minozzi S, Davoli M, Vecchi S. Psychosocial and pharmacological treatments versus pharmacological treatments for opioid detoxification. Cochrane Database of Systematic Reviews [Internet]. 2011; (9). Available from: http://dx.doi.org/10.1002/14651858.CD005031.pub4.

52. Amato L, Minozzi S, Davoli M, Vecchi S. Psychosocial combined with agonist maintenance treatments versus agonist maintenance treatments alone for treatment of opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2011; (10). Available from: http://dx.doi.org/10.1002/14651858.CD004147.pub4.

53. Amiri S, Malek A, Tofighnia F, Asl BH, Seidy A. Amantadine as augmentation in managing opioid withdrawal with clonidine: A randomized controlled trial. Iranian Journal of Psychiatry. 2014;9(3):142-6.

54. Anonymous. Guidelines for the Psychosocially Assisted Pharmacological Treatment of Opioid Dependence. World Health Organization. 2009.

55. Anonymous. Gabapentin for Adults with Neuropathic Pain: A Review of the Clinical Evidence and Guidelines Review. 2014 09 26. Report No.

56. Anonymous. Gabapentin for Adults with Neuropathic Pain: A Review of the Clinical Efficacy and Safety. Review. 2015 04 14. Report No.

57. Anonymous. Crushed Buprenorphine or Buprenorphine-Naloxone for Opioid Dependency: A Review of the Clinical Effectiveness and Guidelines. Review. 2016 07 18. Report No.

58. Ansseau M, De Roeck J. Trazodone in benzodiazepine dependence. J Clin Psychiatry. 1993;54(5):189-91.

59. Apelt SM, Scherbaum N, Gölz J, Backmund M, Soyka M. Safety, effectiveness and tolerance of buprenorphine-naloxone in the treatment of opioid dependence: Results from a nationwide non- interventional study in routine care. Pharmacopsychiatry. 2013;46(3):94-107.

References

60. Argoff CE, Kahan M, Sellers EM. Preventing and managing aberrant drug-related behavior in primary care: systematic review of outcomes evidence. Journal of opioid management. 2014;10(2):119-34.

61. Asaad TA, Ghanem MH, Abdel Samee AM, El-habiby MM. Sleep profile in patients with chronic opioid abuse: A polysomnographic evaluation in an Egyptian sample. Addictive Disorders and their Treatment. 2011;10(1):21-8.

62. Ashrafioun L. Prescription Opioid Craving: Relationship With Pain and Substance Use-Related Characteristics. Subst Use Misuse. 2016;51(11):1512.

63. Ashton CH, Rawlins MD, Tyrer SP. A double-blind placebo-controlled study of buspirone in diazepam withdrawal in chronic benzodiazepine users. Br J Psychiatry. 1990;157:232-8.

64. Baandrup L, Fagerlund B, Glenthoj B. Neurocognitive performance, subjective well-being, and psychosocial functioning after benzodiazepine withdrawal in patients with schizophrenia or bipolar disorder: a randomized clinical trial of add-on melatonin versus placebo. Eur Arch Psychiatry Clin Neurosci. 2017;267(2):163-71.

65. Baandrup L, Glenthoj BY, Jennum PJ. Objective and subjective sleep quality: Melatonin versus placebo add-on treatment in patients with schizophrenia or bipolar disorder withdrawing from long- term benzodiazepine use. Psychiatry Res. 2016;240:163-169.(doi):10.1016/j.psychres.2016.04.031. Epub Apr 15.

66. Baandrup L, Lindschou J, Winkel P, Gluud C, Glenthoj BY. Prolonged-release melatonin versus placebo for benzodiazepine discontinuation in patients with schizophrenia or bipolar disorder: A randomised, placebo-controlled, blinded trial. World J Biol Psychiatry. 2016;17(7):514-24. doi: 10.3109/15622975.2015.1048725. Epub 2015 Jun 18.

67. Baandrup L, Fasmer OB, Glenthoj BY, Jennum PJ. Circadian rest-activity rhythms during benzodiazepine tapering covered by melatonin versus placebo add-on: data derived from a randomized clinical trial. BMC Psychiatry. 2016;16(1):348.

68. Baandrup L, Ebdrup B, Rasmussen J, Lindschou J, Gluud C, Glenthøj B. Pharmacological interventions for benzodiazepine discontinuation in chronic benzodiazepine users. Cochrane Database of Systematic Reviews [Internet]. 2018; (3). Available from: http://dx.doi.org/10.1002/14651858.CD011481.pub2.

69. Bachhuber MA, Maughan BC, Mitra N, Feingold J, Starrels JL. Prescription monitoring programs and emergency department visits involving benzodiazepine misuse: Early evidence from 11 United States metropolitan areas. The International Journal on Drug Policy. 2016;28:120.

70. Back SE, Lawson KM, Singleton LM, Brady KT. Characteristics and correlates of men and women with prescription opioid dependence. Addict Behav. 2011;36(8):829-34.

71. Back SE, Payne RL, Wahlquist AH, Carter RE, Stroud Z, Haynes L, et al. Comparative profiles of men and women with opioid dependence: results from a national multisite effectiveness trial. Am J Drug Alcohol Abuse. 2011;37(5):313-23.

72. Bagra I, Krishnan V, Rao R, Agrawal A. Does Cannabis Use Influence Opioid Outcomes and Quality of Life Among Buprenorphine Maintained Patients? A Cross-sectional, Comparative Study. J Addict Med. 2018;12(4):315-20.

73. Baewert A, Jagsch R, Winklbaur B, Kaiser G, Thau K, Unger A, et al. Influence of site differences between urban and rural American and Central European opioid-dependent pregnant women and neonatal outcome characteristics. Eur Addict Res. 2012;18(3):130-9.

References

74. Bagoien G, Bjorngaard JH, Ostensen C, Reitan SK, Romundstad P, Morken G. The effects of motivational interviewing on patients with comorbid substance use admitted to a psychiatric emergency unit - a randomised controlled trial with two year follow-up. BMC Psychiatry. 2013;13:93.

75. Baker A, Lee NK, Claire M, Lewin TJ, Grant T, Pohlman S, et al. Brief cognitive behavioural interventions for regular amphetamine users: a step in the right direction. Addiction. 2005;100(3):367-78.

76. Baldacchino A, Balfour DJK, Passetti F, Humphris G, Matthews K. Neuropsychological consequences of chronic opioid use: A quantitative review and meta-analysis. Neurosci Biobehav Rev. 2012;36(9):2056-68.

77. Baldacchino A, Tolomeo S, Khan F, Humphris G, Carra G. Acute risk factors in fatal opioid overdoses as a result of hypoxia and cardiotoxicity. A systematic review and critical appraisal. Heroin Addiction and Related Clinical Problems. 2016;18(4):33-42.

78. Baldacchino A, Crocamo C, Humphris G, Neufeind J, Frisher M, Scherbaum N, et al. Decision support in addiction: The development of an e-health tool to assess and prevent risk of fatal overdose. The ORION Project. Comput Methods Programs Biomed. 2016;133:207-16.

79. Baldini A, von Korff M, Lin EHB. A review of potential adverse effects of long-term opioid therapy: A practitioner's guide. Prim Care Companion J Clin Psychiatry. 2012;14(3).

80. Baldwin DS, den Boer JA, Lyndon G, Emir B, Schweizer E, Haswell H. Efficacy and safety of pregabalin in generalised anxiety disorder: A critical review of the literature. Journal of Psychopharmacology. 2015;29(10):1047-60.

81. Ballard C, Corbett A. Randomised controlled trial: A small proportion of people with dementia and neuropsychiatric symptoms experience clinically significant worsening when antidepressants are discontinued (Evidence-Based Medicine (2013) 18, (27-28)). Evid Based Med. 2013;18(3):120.

82. Balodimos S, Nikolaou K, Njau S, Karamouzis M, Kovatsi L. The effect of opioid dependence on conventional and novel biochemical parameters of bone metabolism. Am J Drug Alcohol Abuse. 2015;41(6):535-40.

83. Banta-Green CJ, Coffin PO, Merrill JO, Sears JM, Dunn C, Floyd AS, et al. Impacts of an opioid overdose prevention intervention delivered subsequent to acute care. Inj Prev. 2018;07:07.

84. Banta-Green CJ, Von Korff M, Sullivan MD, Merrill JO, Doyle SR, Saunders K. The Prescribed Opioids Difficulties Scale: A patient-centered assessment of problems and concerns. The Clinical Journal of Pain. 2010;26(6):489-97.

85. Barry DT, Savant JD, Beitel M, Cutter CJ, Moore BA, Schottenfeld RS, et al. Pain and Associated Substance Use among Opioid Dependent Individuals Seeking Office-Based Treatment with Buprenorphine-Naloxone: A Needs Assessment Study. The American Journal on Addictions. 2013;22(3):212-7.

86. Barry CL, Kennedy‐Hendricks A, Gollust SE, Niederdeppe J, Bachhuber MA, Webster DW, et al. Understanding Americans’ views on opioid pain reliever abuse. Addiction. 2016;111(1):85-93.

87. Bartholomaeus JH, Arkenau-Maric E, Galia E. Opioid extended-release tablets with improved tamper-resistant properties. Expert opinion on drug delivery. 2012;9(8):879-91.

References

88. Bartoli A, Michna E, He E, Wen W. Efficacy and safety of once-daily, extended-release hydrocodone in individuals previously receiving hydrocodone/acetaminophen combination therapy for chronic pain. Postgrad Med. 2015;127(1):5-12.

89. Baser O, Xie L, Mardekian J, Schaaf D, Wang L, Joshi AV. Prevalence of diagnosed opioid abuse and its economic burden in the Veterans Health Administration. Pain Practice. 2014;14(5):437- 45.

90. Bashir K, King M, Ashworth M. Controlled evaluation of brief intervention by general practitioners to reduce chronic use of benzodiazepines. The British journal of general practice : the journal of the Royal College of General Practitioners. 1994;44(386):408-12.

91. Bassiony MM, Salah El-Deen G, Abdelghani M, Youssef UM, Hassan MS, El-Gohari H. Addressing Prevalence and Correlates among a Sample of Egyptian University Students Who Suffer from Substance Use Disorders. Addictive Disorders and their Treatment. 2018;17(3):114-23.

92. Bassiony MM, Abdelghani M, Salah El-Deen GM, Hassan MS, El-Gohari H, Youssef UM. Opioid Use Disorders Attributed to Tramadol Among Egyptian University Students. J Addict Med. 2018;12(2):150-5.

93. Bautista CA, Iosif A-M, Wilsey BL, Melnikow JA, Crichlow A, Henry SG. Factors associated with opioid dose increases: A chart review of patients’ first year on long-term opioids. Pain Med. 2017;18(5):908-16.

94. Baxley C, Weinstock J, Lustman PJ, Garner AA. The influence of anxiety sensitivity on opioid use disorder treatment outcomes. Exp Clin Psychopharmacol. 2018;06:06.

95. Beaudoin FL, Merchant RC, Clark MA. Prevalence and Detection of Prescription Opioid Misuse and Prescription Opioid Use Disorder Among Emergency Department Patients 50 Years of Age and Older: Performance of the Prescription Drug Use Questionnaire, Patient Version. Am J Geriatr Psychiatry. 2016;24(8):627-36.

96. Beaudoin FL, Banerjee GN, Mello MJ. State-level and system-level opioid prescribing policies: The impact on provider practices and overdose deaths, a systematic review. Journal of Opioid Management. 2016;12(2):109-18.

97. Becka J. Efficacy of cognitive behavioral methods in decreasing drug use by methadone clinic patients compared with standard conditions. Slovenská Psychiatrie. 2004;100(Suppl 1):4-7.

98. Becker WC, Fiellin DA, Dept of Internal Medicine YUSoMNHCT. Provider satisfaction with office-based treatment of opioid dependence: a systematic review. Subst Abus. 2005;26(1):15-22.

99. Becker WC, Sullivan LE, Tetrault JM, Desai RA, Fiellin DA. Non-medical use, abuse and dependence on prescription opioids among U.S. adults: Psychiatric, medical and substance use correlates. Drug Alcohol Depend. 2008;94(1-3):38-47.

100. Becker WC, Fraenkel L, Edelman EJ, Holt SR, Glover J, Kerns RD, et al. Instruments to assess patient-reported safety, efficacy, or misuse of current opioid therapy for chronic pain: a systematic review. Pain. 2013;154(6):905-16.

101. Belaise C, Gatti A, Chouinard VA, Chouinard G. Persistent Postwithdrawal Disorders Induced by Paroxetine, a Selective Serotonin Reuptake Inhibitor, and Treated with Specific Cognitive Behavioral Therapy. Psychother Psychosom. 2014;83(4):247-8.

References

102. Belknap SM, Moore H, Lanzotti SA, Yarnold PR, Getz M, Deitrick DL, et al. Application of software design principles and debugging methods to an analgesia prescription reduces risk of severe injury from medical use of opioids. Clin Pharmacol Ther. 2008;84(3):385-92.

103. Bell JR, Ryan A, Mutch C, Batey R, Rea F. Optimising the benefits of unobserved dose administration for stable opioid maintenance patients: follow-up of a randomised trial. Drug Alcohol Depend. 2008;96(1-2):183-6.

104. Belleville G, Morin CM. Hypnotic discontinuation in chronic insomnia: impact of psychological distress, readiness to change, and self-efficacy. Health Psychol. 2008;27(2):239-48.

105. Beneitez MC, Gil-Alegre ME. Opioid Addiction: Social Problems Associated and Implications of Both Current and Possible Future Treatments, including Polymeric Therapeutics for Giving Up the Habit of Opioid Consumption. Biomed Res Int. 2017;2017:7120815.

106. Bentzley BS, Barth KS, Back SE, Book SW. Discontinuation of Buprenorphine Maintenance Therapy: Perspectives and Outcomes. J Subst Abuse Treat. 2015;52:48-57.

107. Bergman H, Bhoopathi P, Soares‐Weiser K. Benzodiazepines for antipsychotic‐induced tardive dyskinesia. Cochrane Database of Systematic Reviews [Internet]. 2018; (1). Available from: http://dx.doi.org/10.1002/14651858.CD000205.pub3.

108. Berna C, Kulich RJ, Rathmell JP. Tapering Long-term Opioid Therapy in Chronic Noncancer Pain: Evidence and Recommendations for Everyday Practice. Mayo Clin Proc. 2015;90(6):828-42.

109. Beswick T, Best D, Rees S, Bearn J, Gossop M, Strang J. Major disruptions of sleep during treatment of the opiate withdrawal syndrome: differences between methadone and lofexidine detoxification treatments. Addict Biol. 2003;8(1):49‐57.

110. Bet PM, Hugtenburg JG, Penninx BWJH, Hoogendijk WJG. Side effects of antidepressants during long-term use in a naturalistic setting. Eur Neuropsychopharmacol. 2013;23(11):1443-51.

111. Bickel WK, Marsch LA, Buchhalter AR, Badger GJ. Computerized behavior therapy for opioid- dependent outpatients: a randomized controlled trial. Exp Clin Psychopharmacol. 2008;16(2):132-43.

112. Bigal ME, Lipton RB. Excessive opioid use and the development of chronic migraine. Pain. 2009;142(3):179-82.

113. Biondi D, Xiang J, Vorsanger G, Moskovitz B, Ashworth J, Etropolski M. Tapentadol extended release (ER) versus oxycodone controlled release (CR) for management of chronic low back or osteoarthritis pain: influence of prior opioid experience on study discontinuations due to constipation, nausea, or vomiting. J Pain. 2010;1:S42.

114. Bisaga AMD, Sullivan MAMDP, Glass AMS, Mishlen KMA, Carpenter KMP, Mariani JJMD, et al. A placebo-controlled trial of memantine as an adjunct to injectable extended-release naltrexone for opioid dependence. J Subst Abuse Treat. 2014;46(5):546.

115. Bishop B, Gilmour J, Deering D. Readiness and recovery: Transferring between methadone and buprenorphine/naloxone for the treatment of opioid use disorder. Int J Ment Health Nurs. 2018;18:18.

116. Black R, Coplan P, Cassidy T, Chilcoat H, Budman S, Landau C, et al. Effects of reformulated OxyContin® among patients assessed for substance abuse treatment in the NAVIPPRO sentinel surveillance network. The Journal of Pain. 2012;13(4):S58.

117. Bleckwenn M, Heister L, Weckbecker M, Weckbecker K, Mucke M. Misuse of Substitution Drugs in the Substitution-Based Therapy. Eur Addict Res. 2016;22(6):322-8.

References

118. Blom MB, Spinhoven P, Hoffman T, Jonker K, Hoencamp E, Haffmans PM, et al. Severity and duration of depression, not personality factors, predict short term outcome in the treatment of major depression. J Affect Disord. 2007;104(1‐3):119‐26.

119. Blondell RD, Frydrych LM, Arber BC, Bashaw HL, Vexler A, Purdy CH, et al. A randomized trial of extended buprenorphine detoxification for opioid dependency. J Addict Med. 2008;2(3):139-46.

120. Blondell RD, Ashrafioun L, Dambra CM, Foschio EM, Zielinski AL, Salcedo DM. A Clinical Trial Comparing Tapering Doses of Buprenorphine with Steady Doses for Chronic Pain and Co-existent Opioid Addiction. J Addict Med. 2010;4(3):140-6.

121. Blondi D, Xiang J, Lange R, Etropolski M, Vorsanger G, Moskovitz B. Tolerability of tapentadol extended release versus oxycodone controlled release in elderly patients with chronic low back or osteoarthritis pain in a 1-year safety study. Der schmerz. 2010;24(S1):92.

122. Bobes J, Rubio G, Teran A, Cervera G, Lopez-Gomez V, Vilardaga I, et al. Pregabalin for the discontinuation of long-term benzodiazepines use: an assessment of its effectiveness in daily clinical practice. Eur Psychiatry. 2012;27(4):301-7.

123. Bockting CLH, Klein NS, Elgersma HJ, van Rijsbergen GD, Slofstra C, Ormel J, et al. Effectiveness of preventive cognitive therapy while tapering antidepressants versus maintenance antidepressant treatment versus their combination in prevention of depressive relapse or recurrence (DRD study): a three-group, multicentre, randomised controlled trial. Lancet Psychiatry. 2018;5(5):401-10.

124. Bogetto F, Bellino S, Revello RB, Patria L. Discontinuation syndrome in dysthymic patients treated with selective serotonin reuptake inhibitors: a clinical investigation. CNS Drugs. 2002;16(4):273-83.

125. Bohnert AS, Valenstein M, Bair MJ, Ganoczy D, McCarthy JF, Ilgen MA, et al. Association between opioid prescribing patterns and opioid overdose-related deaths. JAMA. 2011;305(13):1315- 21.

126. Bohnert AS, Bonar EE, Cunningham R, Greenwald MK, Thomas L, Chermack S, et al. A pilot randomized clinical trial of an intervention to reduce overdose risk behaviors among emergency department patients at risk for prescription opioid overdose. Drug Alcohol Depend. 2016;163:40-7.

127. Bonar EE, Ilgen MA, Walton M, Bohnert ASB. Associations among Pain, Non-Medical Prescription Opioid Use, and Drug Overdose History. The American Journal on Addictions. 2014;23(1):41-7.

128. Bondi CD, Kamal KM, Johnson DA, Witt-Enderby PA, Giannetti VJ. The Effect of Melatonin Upon Postacute Withdrawal Among Males in a Residential Treatment Program (M-PAWS): A Randomized, Double-blind, Placebo-controlled Trial. J Addict Med. 2018;12(3):201-6.

129. Bowman S, Eiserman J, Beletsky L, Stancliff S, Bruce RD. Reducing the health consequences of opioid addiction in primary care. Am J Med. 2013;126(7):565-71.

130. Boyuan Z, Yang C, Ke C, Xueyong S, Sheng L. Efficacy of acupuncture for psychological symptoms associated with opioid addiction: A systematic review and meta-analysis. Evid Based Complement Alternat Med. 2014;2014 (no pagination)(313549).

131. Braden JB, Sullivan MD, Ray GT, Saunders K, Merrill J, Silverberg MJ, et al. Trends in long- term opioid therapy for noncancer pain among persons with a history of depression. Gen Hosp Psychiatry. 2009;31(6):564-70.

References

132. Brady JE, Giglio R, Keyes KM, DiMaggio C, Li G. Risk markers for fatal and non-fatal prescription drug overdose: a meta-analysis. Injury Epidemiology. 2017;4 (1) (no pagination)(24).

133. Branstetter SA, Bower EH, Kamien J, Amass L. A history of sexual, emotional, or physical abuse predicts adjustment during opioid maintenance treatment. J Subst Abuse Treat. 2008;34(2):208-14.

134. Breen CL, Harris SJ, Lintzeris N, Mattick RP, Hawken L, Bell J, et al. Cessation of methadone maintenance treatment using buprenorphine: transfer from methadone to buprenorphine and subsequent buprenorphine reductions. Drug Alcohol Depend. 2003;71(1):49-55.

135. Brigham G, Harrer JM, Winhusen, Pelt A, Amass L. Integrating buprenorphine-naloxone tablet treatment for short-term withdrawal from opioids into a residential integrated addiction and mental health service. Proceedings of the 66th annual scientific meeting of college on problems of drug dependence; 2004 june 12-17; san juan, puerto rico. 2004.

136. Broderick M, Guilleminault C. Rebound cataplexy after withdrawal from antidepressants. Sleep Med. 2009;10(4):403-4.

137. Brown ES, Tirado C, Minhajuddin A, Hillhouse M, Adinoff B, Ling W, et al. Association of race and ethnicity with withdrawal symptoms, attrition, opioid use, and side-effects during buprenorphine therapy. Journal of Ethnicity in Substance Abuse. 2010;9(2):106-14.

138. Brown J, Setnik B, Lee K, Wase L, Roland CL, Cleveland JM, et al. Assessment, stratification, and monitoring of the risk for prescription opioid misuse and abuse in the primary care setting. Journal of Opioid Management. 2011;7(6):467-82.

139. Brown S-E, Altice FL. Self-Management of Buprenorphine/Naloxone Among Online Discussion Board Users. Subst Use Misuse. 2014;49(8):1017-24.

140. Brown R, Riley MR, Ulrich L, Kraly EP, Jenkins P, Krupa NL, et al. Impact of New York prescription drug monitoring program, I-STOP, on statewide overdose morbidity. Drug Alcohol Depend. 2017;178:348.

141. Bruehl S, Burns JW, Passik SD, Gupta R, Buvanendran A, Chont M, et al. The Contribution of Differential Opioid Responsiveness to Identification of Opioid Risk in Chronic Pain Patients. J Pain. 2015;16(7):666-75.

142. Brummett CM, Waljee JF, Goesling J, Moser S, Lin P, Englesbe MJ, et al. New Persistent Opioid Use After Minor and Major Surgical Procedures in US Adults. JAMA Surg. 2017;152(6):e170504.

143. Bucher Bartelson B, Le Lait MC, Green JL, Cepeda MS, Coplan PM, Maziere JY, et al. Changes in misuse and abuse of prescription opioids following implementation of Extended-Release and Long- Acting Opioid Analgesic Risk Evaluation and Mitigation Strategy. Pharmacoepidemiol Drug Saf. 2017;26(9):1061-70.

144. Bujold E, Huff J, Staton EW, Pace WD. Improving use of narcotics for nonmalignant chronic pain: A lesson from Community Care of North Carolina. Journal of Opioid Management. 2012;8(6):363-7.

145. Butler SF, Cassidy TA, Chilcoat H, Black RA, Landau C, Budman SH, et al. Abuse rates and routes of administration of reformulated extended-release oxycodone: initial findings from a sentinel surveillance sample of individuals assessed for substance abuse treatment. J Pain. 2013;14(4):351-8.

References

146. Butler SF, Black RA, Cassidy TA, Dailey TM, Budman SH. Abuse risks and routes of administration of different prescription opioid compounds and formulations. Harm Reduction Journal. 2011;8 (no pagination)(29).

147. Butler SF, Black RA, Severtson SG, Dart RC, Green JL. Understanding abuse of buprenorphine/naloxone film versus tablet products using data from ASI-MV® substance use disorder treatment centers and RADARS® System Poison Centers. J Subst Abuse Treat. 2018;84:42.

148. Buttram ME, Kurtz SP, Surratt HL, Levi-Minzi MA. Health and social problems associated with prescription opioid misuse among a diverse sample of substance-using MSM. Subst Use Misuse. 2014;49(3):277-84.

149. Buynak R, Rappaport SA, Rod K, Arsenault P, Heisig F, Rauschkolb C, et al. Long-term Safety and Efficacy of Tapentadol Extended Release Following up to 2 Years of Treatment in Patients With Moderate to Severe, Chronic Pain: Results of an Open-label Extension Trial. Clin Ther. 2015;37(11):2420-38.

150. Cadth. Watchful dosing of morphine or morphine equivalent dosing in the treatment of chronic non-cancer pain: a review of the clinical evidence. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2012.

151. Cadth. Sleep medications for adults diagnosed with insomnia: clinical evidence and harms. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2013.

152. Cadth. Buprenorphine/naloxone (Suboxone) film versus buprenorphine/naloxone tablets for the treatment of opioid addiction: comparative safety. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2014.

153. Cadth. Narcotics, benzodiazepines, stimulants, and gabapentin: policies, initiatives, and practices across Canada, 2014. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2014.

154. Cadth. Discontinuation strategies for patients with long-term benzodiazepine use: clinical evidence and guidelines. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2015.

155. Cadth. Discontinuation strategies for patients with long-term Z-drug use: clinical evidence and guidelines. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2015.

156. Cadth. Opioids for the management of tramadol withdrawal: clinical effectiveness. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH), 2015.

157. Calcaterra SL, Scarbro S, Hull ML, Forber AD, Binswanger IA, Colborn KL. Prediction of Future Chronic Opioid Use Among Hospitalized Patients. J Gen Intern Med. 2018;33(6):898-905.

158. Caldentey CV, Gelabert FF, Garrido EG, Ojeda EM, Sabater CM, Cánaves JL. Long-term effectiveness of an intervention to discontinue chronic benzodiazepine use. Actas Españolas de Psiquiatría. 2008;36(5):295-8.

159. Campbell CI, Kline-Simon AH, Von Korff M, Saunders KW, Weisner C. Alcohol and Drug Use and Aberrant Drug-Related Behavior Among Patients on Chronic Opioid Therapy. Subst Use Misuse. 2017;52(10):1283-91.

160. Candiotti KA, Gitlin MC. Review of the effect of opioid-related side effects on the undertreatment of moderate to severe chronic non-cancer pain: tapentadol, a step toward a solution? Curr Med Res Opin. 2010;26(7):1677-84.

References

161. Canham SL, Gallo J, Simoni-Wastila L. Perceptions of benzodiazepine dependence among women age 65 and older. J Gerontol Soc Work. 2014;57(8):872-88.

162. Canan C, Polinski JM, Alexander GC, Kowal MK, Brennan TA, Shrank WH. Automatable algorithms to identify nonmedical opioid use using electronic data: A systematic review. J Am Med Inform Assoc. 2017;24(6):1204-10.

163. Cardinali DP, Gvozdenovich E, Kaplan MR, Fainstein I, Shifis HA, Perez Lloret S, et al. A double blind-placebo controlled study on melatonin efficacy to reduce anxiolytic benzodiazepine use in the elderly. Neuro Endocrinol Lett. 2002;23(1):55-60.

164. Cardinali DP, Golombek DA, Rosenstein RE, Brusco LI, Vigo DE. Assessing the efficacy of melatonin to curtail benzodiazepine/Z drug abuse. Pharmacol Res. 2016;109:12-23.

165. Carroll KM, Ball SA, Nich C, Martino S, Frankforter TL, Farentinos C, et al. Motivational interviewing to improve treatment engagement and outcome in individuals seeking treatment for substance abuse: a multisite effectiveness study. Drug Alcohol Depend. 2006;81(3):301-12.

166. Casati A, Sedefov R, Pfeiffer-Gerschel T. Misuse of medicines in the European Union: a systematic review of the literature. Eur Addict Res. 2012;18(5):228-45.

167. Cassano G, Petracca A, Borghi C, Chiroli S, Didoni G, Garreau M. A randomized, double-blind study of alpidem vs placebo in the prevention and treatment of benzodiazepine withdrawal syndrome. Eur Psychiatry. 1996;11(2):93-9. doi: 10.1016/0924-9338(96)84786-9.

168. Cassidy TA, DasMahapatra P, Black RA, Wieman MS, Butler SF. Changes in prevalence of prescription opioid abuse after introduction of an abuse-deterrent opioid formulation. Pain Med. 2014;15(3):440-51.

169. Chabal C, Erjavec MK, Jacobson L, Mariano A, Chaney E. Prescription opiate abuse in chronic pain patients: clinical criteria, incidence, and predictors. Clin J Pain. 1997;13(2):150-5.

170. Chan BK, Tam LK, Wat CY, Chung YF, Tsui SL, Cheung CW. Opioids in chronic non-cancer pain. Expert Opin Pharmacother. 2011;12(5):705-20.

171. Chan KT, Trant J. The Relationship of Psychological Distress and Living with Children and Adolescents for Adult Non-medical Prescription Opioid Users. Child & Adolescent Social Work Journal. 2018;35(4):391-405.

172. Chand P, Murthy P. Short-term outcome of take-home prescriptions for opioid dependence: a clinic-based study. Journal of Substance Use. 2013;18(2):108-18.

173. Chaparro L, Furlan A, Deshpande A, Mailis‐Gagnon A, Atlas S, Turk D. Opioids compared to placebo or other treatments for chronic low‐back pain. Cochrane Database of Systematic Reviews [Internet]. 2013; (8). Available from: http://dx.doi.org/10.1002/14651858.CD004959.pub4.

174. Chaparro LE, Furlan AD, Deshpande A, Mailis-Gagnon A, Atlas S, Turk DC. Opioids compared with placebo or other treatments for chronic low back pain: an update of the cochrane review. Spine. 2014;39(7):556-63.

175. Chatterjee A, Yu EJ, Tishberg L. Exploring opioid use disorder, its impact, and treatment among individuals experiencing homelessness as part of a family. Drug Alcohol Depend. 2018;188:161.

176. Chattopadhyay AC, Shukla L, Kandasamy A, Benegal V. High-dose zolpidem dependence - Psychostimulant effects? A case report and literature review. Industrial psychiatry journal. 2016;25(2):222-4.

References

177. Chavoustie S, Frost M, Snyder O, Owen J, Darwish M, Dammerman R, et al. Buprenorphine implants in medical treatment of opioid addiction. Expert Rev Clin Pharmacol. 2017;10(8):799-807.

178. Cheatle MD, O'Brien CP, Mathai K, Hansen M, Grasso M, Yi P. Aberrant behaviors in a primary care-based cohort of patients with chronic pain identified as misusing prescription opioids. Journal of Opioid Management. 2013;9(5):315-24.

179. Chen W, Hong Y, Zou X, McLaughlin MM, Xia Y, Ling L. Effectiveness of prize-based contingency management in a methadone maintenance program in China. Drug Alcohol Depend. 2013;133(1):270-4.

180. Cheon E, Lee J, Choi J, Lee Y, Koo B. Effectiveness of duloxetine monotherapy compared to combination therapy with other antidepressants in patients with major depressive disorder: a short- term, retrospective study. Psychiatry investigation 13 (4) (pp 447-452), 2016 Date of publication: july 2016. 2016.

181. Cheung CW, Qiu Q, Choi SW, Moore B, Goucke R, Irwin M. Chronic opioid therapy for chronic non-cancer pain: a review and comparison of treatment guidelines. Pain physician. 2014;17(5):401- 14.

182. Chopra MP, Landes RD, Gatchalian KM, Jackson LC, Buchhalter AR, Stitzer ML, et al. Buprenorphine medication versus voucher contingencies in promoting abstinence from opioids and cocaine. Exp Clin Psychopharmacol. 2009;17(4):226-36.

183. Chou R, Fanciullo GJ, Fine PG, Adler JA, Ballantyne JC, Davies P, et al. Clinical guidelines for the use of chronic opioid therapy in chronic noncancer pain. J Pain. 2009;10(2):113-30.

184. Chou R, Turner JA, Devine EB, Hansen RN, Sullivan SD, Blazina I, et al. The effectiveness and risks of long-term opioid therapy for chronic pain: a systematic review for a National Institutes of Health Pathways to Prevention Workshop. Ann Intern Med. 2015;162(4):276-86.

185. Christensen DR, Landes RD, Jackson L, Marsch LA, Mancino MJ, Chopra MP, et al. Adding an Internet-delivered treatment to an efficacious treatment package for opioid dependence. J Consult Clin Psychol. 2014;82(6):964-72.

186. Cialdella P, Boissel JP, Belon P. [Homeopathic specialties as substitutes for benzodiazepines: double-blind vs. placebo study]. Therapie. 2001;56(4):397-402.

187. Cicero TJ, Surratt HL, Inciardi J. Use and misuse of buprenorphine in the management of opioid addiction. J Opioid Manag. 2007;3(6):302-8.

188. Cicero TJ, Lynskey M, Todorov A, Inciardi JA, Surratt HL. Co-morbid pain and psychopathology in males and females admitted to treatment for opioid analgesic abuse. Pain. 2008;139(1):127-35.

189. Cicero TJ, Ellis MS, Surratt HL. Effect of abuse-deterrent formulation of OxyContin. N Engl J Med. 2012;367(2):187-9.

190. Cicero TJ, Ellis MS. Understanding the demand side of the prescription opioid epidemic: Does the initial source of opioids matter? Drug Alcohol Depend. 2017;173.

191. Cicero TJ, Ellis MS. The prescription opioid epidemic: A review of qualitative studies on the progression from initial use to abuse. Dialogues Clin Neurosci. 2017;19(3):259-69.

192. Clancy Z, O'Connell K, Couto J. The use of urine drug monitoring in chronic opioid therapy: an analysis of current clinician behavior. Journal of Opioid Management. 2013;9(2):121-7.

References

193. Clark MR, Stoller KB, Brooner RK. Assessment and management of chronic pain in individuals seeking treatment for opioid dependence disorder. Canadian Journal of Psychiatry - Revue Canadienne de Psychiatrie. 2008;53(8):496-508.

194. Clemans-Cope L, Wishner JB, Allen EH, Lallemand N, Epstein M, Spillman BC. Experiences of three states implementing the Medicaid health home model to address opioid use disorder-Case studies in Maryland, Rhode Island, and Vermont. J Subst Abuse Treat. 2017;83:27.

195. Cochran G, Stitzer M, Campbell ANC, Hu MC, Vandrey R, Nunes EV. Web-based treatment for substance use disorders: Differential effects by primary substance. Addict Behav. 2015;45:191-4.

196. Cock V, Edmonds C, Cock C. Complications related to chronic supratherapeutic use of codeine containing compound analgesics in a cohort of patients presenting for codeine withdrawal. Drug and alcohol review. 2018;37(6):731-7.

197. Colquhoun R. Rapid methadone detoxification using morphine to reduce severity of withdrawal. Journal of Global Drug Policy and Practice. 2013;7(3):1-24.

198. Colson J, Helm S, Silverman SM. Office-based opioid dependence treatment. Pain physician. 2012;15(3 Suppl):ES231-6.

199. Cooper S, Campbell G, Larance B, Murnion B, Nielsen S. Perceived stigma and social support in treatment for pharmaceutical opioid dependence. Drug and alcohol review. 2018;37(2):262-72.

200. Conermann T, Gosalia AR, Kabazie AJ, Moore C, Miller K, Fetsch M, et al. Utility of oral fluid in compliance monitoring of opioid medications. Pain physician. 2014;17(1):63-70.

201. Conrardy M, Lank P, Cameron KA, McConnell R, Chevrier A, Sears J, et al. Emergency Department Patient Perspectives on the Risk of Addiction to Prescription Opioids. Pain Med. 2016;17(1):114-21.

202. Cornish R, Macleod J, Strang J, Vickerman P, Hickman M. Risk of death during and after opiate substitution treatment in primary care: prospective observational study in UK General Practice Research Database. Br Med J. 2010;341(7779):928.

203. Correa D, Farney RJ, Chung F, Prasad A, Lam D, Wong J. Chronic opioid use and central sleep apnea: a review of the prevalence, mechanisms, and perioperative considerations. Anesth Analg. 2015;120(6):1273-85.

204. Cossmann JC, Scherbaum N, Bonnet U. Substance addiction in old age: A cross-sectional study in a German Hospital. GeroPsych: The Journal of Gerontopsychology and Geriatric Psychiatry. 2016;29(1):17-27.

205. Courtney von H, Henry JD, Terrett G, Mercuri K, McAlear K, Rendell PG. Stereotype threat and social function in opioid substitution therapy patients. The British Journal of Clinical Psychology. 2017;56(2):160-71.

206. Couvee JE, Zitman FG. The Benzodiazepine Withdrawal Symptom Questionnaire: psychometric evaluation during a discontinuation program in depressed chronic benzodiazepine users in general practice. Addiction. 2002;97(3):337-45.

207. Couvee JE, Timmermans MA, Zitman FG. The long-term outcome of a benzodiazepine discontinuation programme in depressed outpatients. J Affect Disord. 2002;70(2):133-41.

208. Couvee JE, Bakker A, Zitman FG. The relevance of psychiatric and somatic comorbidity in depressed chronic benzodiazepine users. Psychother Psychosom. 2002;71(5):263-8. doi: 10.1159/000064810.

References

209. Coviello DM, Cornish JW, Lynch KG, Alterman AI, O'Brien CP. A randomized trial of oral naltrexone for treating opioid-dependent offenders. Am J Addict. 2010;19(5):422-32.

210. Coyle DT, Pratt CY, Ocran-Appiah J, Secora A, Kornegay C, Staffa J. Opioid analgesic dose and the risk of misuse, overdose, and death: A narrative review. Pharmacoepidemiol Drug Saf. 2018;27(5):464-72.

211. Cragg A, Hau JP, Woo SA, Liu C, Doyle-Waters MM, Hohl CM. Risk factors for addiction among patients receiving prescribed opioids: A systematic review protocol. Systematic Reviews. 2017;6 (1) (no pagination)(265).

212. Crawford-Faucher A. Escitalopram effective for reducing hot flashes in menopausal women. Am Fam Physician. 2011;84(11):1293-4.

213. Creac HC, Frappe P, Cancade M, Laurent B, Peyron R, Demarquay G, et al. In-patient versus out-patient withdrawal programmes for medication overuse headache: A 2-year randomized trial. Cephalalgia. 2011;31(11):1189-98.

214. Crits-Christoph PP, Markell HMBA, Gibbons MBCP, Gallop RP, Lundy CP, Stringer MMA, et al. A Naturalistic Evaluation of Extended-Release Naltrexone in Clinical Practice in Missouri. J Subst Abuse Treat. 2016;70:50-7.

215. Croissant B, Grosshans M, Diehl A, Mann K. Oxcarbazepine in Rapid Benzodiazepine Detoxification. The American Journal of Drug and Alcohol Abuse. 2008;34(5):534.

216. Crowe SF, Stranks EK. The Residual Medium and Long-term Cognitive Effects of Benzodiazepine Use: An Updated Meta-analysis. Arch Clin Neuropsychol. 2017;13:13.

217. Culberson JW, Ticker RL, Burnett J, Marcus MT, Pickens SL, Dyer CB. Prescription Medication Use Among Self Neglecting Elderly. J Addict Nurs. 2011;22(1-2):63-8.

218. Cunningham CO, Giovanniello A, Li X, Kunins HV, Roose RJ, Sohler NL. A comparison of buprenorphine induction strategies: Patient-centered home-based inductions versus standard-of- care office-based inductions. J Subst Abuse Treat. 2011;40(4):349-56.

219. da CB, Nüesch E, Kasteler R, Husni E, Welch V, Rutjes A, et al. Oral or transdermal opioids for osteoarthritis of the knee or hip. Cochrane Database of Systematic Reviews [Internet]. 2014; (9). Available from: http://dx.doi.org/10.1002/14651858.CD003115.pub4.

220. D’Agostino AR, Optican AR, Sowles SJ, Krauss MJ, Lee KE, Cavazos-Rehg PA. Social networking online to recover from opioid use disorder: A study of community interactions. Drug Alcohol Depend. 2017;181:5.

221. Dakwar E, Kleber HD. Naltrexone-facilitated buprenorphine discontinuation: a feasibility trial. J Subst Abuse Treat. 2015;53:60-3.

222. Daniell HW. Opioid endocrinopathy in women consuming prescribed sustained-action opioids for control of nonmalignant pain. The Journal of Pain. 2008;9(1):28-36.

223. Daniulaityte R, Falck R, Carlson RG. "I'm not afraid of those ones just 'cause they've been prescribed": perceptions of risk among illicit users of pharmaceutical opioids. The International journal on drug policy. 2012;23(5):374-84.

224. Daniulaityte R, Carlson RG, Kenne DR. Initiation to Pharmaceutical Opioids and Patterns of Misuse: Preliminary Qualitative Findings Obtained by the Ohio Substance Abuse Monitoring Network. Journal of Drug Issues. 2006;36(4):787-808.

References

225. Darker C, Sweeney B, Barry J, Farrell M, Donnelly‐Swift E. Psychosocial interventions for benzodiazepine harmful use, abuse or dependence. Cochrane Database of Systematic Reviews [Internet]. 2015; (5). Available from: http://dx.doi.org/10.1002/14651858.CD009652.pub2.

226. Darnall BD, Stacey BR, Chou R. Medical and psychological risks and consequences of long‐term opioid therapy in women. Pain Med. 2012;13(9):1181-211.

227. Dart RC, Cicero TJ, Surratt HL, Rosenblum A, Bartelson BB, Adams EH. Assessment of the abuse of tapentadol immediate release: the first 24 months. J Opioid Manag. 2012;8(6):395-402.

228. Davies J, Read J. A systematic review into the incidence, severity and duration of antidepressant withdrawal effects: Are guidelines evidence-based? Addict Behav. 2018.

229. Davis D, Kurti A, Skelly J, Redner R, White T, Higgins S. A review of the literature on contingency management in the treatment of substance use disorders, 2009-2014. Prev Med. 2016;92:36-46.

230. Davis J, Smith D, Briley D. Substance use prevention and treatment outcomes for emerging adults in non-college settings: A meta-analysis. Psychol Addict Behav. 2017;31(3):242-54.

231. Davis JP, Berry D, Dumas TM, Ritter E, Smith DC, Menard C, et al. Substance use outcomes for mindfulness based relapse prevention are partially mediated by reductions in stress: Results from a randomized trial. J Subst Abuse Treat. 2018;91:37-48.

232. Day E, Ison J, Strang J. Inpatient versus other settings for detoxification for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2005; (2). Available from: http://dx.doi.org/10.1002/14651858.CD004580.pub2.

233. De S, Jain R, Ray R, Dhawan A, Varghese ST. Assessment of differential doses of buprenorphine for long term pharmacotherapy among opiate dependent subjects. Indian J Physiol Pharmacol. 2008;52(1):53-63.

234. DeFulio A, Everly JJ, Leoutsakos JM, Umbricht A, Fingerhood M, Bigelow GE, et al. Employment-based reinforcement of adherence to an FDA approved extended release formulation of naltrexone in opioid-dependent adults: a randomized controlled trial. Drug Alcohol Depend. 2012;120(1-3):48-54.

235. Degenhardt L, Bucello C, Mathers B, Briegleb C, Ali H, Hickman M, et al. Mortality among regular or dependent users of heroin and other opioids: a systematic review and meta-analysis of cohort studies. Addiction. 2011;106(1):32-51.

236. Degenhardt L, Bruno R, Ali R, Lintzeris N, Farrell M, Larance B. The introduction of a potentially abuse deterrent oxycodone formulation: Early findings from the Australian National Opioid Medications Abuse Deterrence (NOMAD) study. Drug Alcohol Depend. 2015;151:56-67.

237. de Gier NA, Gorgels WJ, Lucassen PL, Oude Voshaar R, Mulder J, Zitman F. Discontinuation of long-term benzodiazepine use: 10-year follow-up. Fam Pract. 2011;28(3):253-9.

238. Dehghani-Arani F, Rostami R, Nadali H. Neurofeedback training for opiate addiction: improvement of mental health and craving. Appl Psychophysiol Biofeedback. 2013;38(2):133-41.

239. Delcher C. Monitoring opioid abuse and doctor shopping: An epidemiologic perspective in person, place, and time [Ph.D.]: University of Florida; 2017.

240. del Rio M, Mino A, Perneger TV. Predictors of patient retention in a newly established methadone maintenance treatment programme. Addiction. 1997;92(10):1353-60.

References

241. De Maeyer J, Vanderplasschen W, Broekaert E. Quality of life among opiate-dependent individuals: A review of the literature. The International Journal on Drug Policy. 2010;21(5):364.

242. De Maricourt P, Gorwood P, Hergueta T, Galinowski A, Salamon R, Diallo A, et al. Balneotherapy Together with a Psychoeducation Program for Benzodiazepine Withdrawal: a Feasibility Study. Evid Based Complement Alternat Med. 2016;2016(no pagination).

243. Demidenko MI, Dobscha SK, Morasco BJ, Meath THA, Ilgen MA, Lovejoy TI. Suicidal ideation and suicidal self-directed violence following clinician-initiated prescription opioid discontinuation among long-term opioid users. Gen Hosp Psychiatry. 2017;47:29-35.

244. Dengler J, Sturesson B, Kools D, Prestamburgo D, Pflugmacher R, Vajkoczy P. Opioid use in relation to treatment outcome in patients with low back pain. Eur Spine J. 2016;Conference: 11th German Spine Conference ‐ Annual Meeting of the German Spine Society. Germany. Conference Start: 20161201. Conference End: 20161203. 25(11):3830‐1.

245. Dennis BB, Naji L, Bawor M, Bonner A, Varenbut M, Daiter J, et al. The effectiveness of opioid substitution treatments for patients with opioid dependence: a systematic review and multiple treatment comparison protocol. Systematic Reviews. 2014;3:105.

246. de Oliveira C, Carvalho L, Carlos K, Conti C, de OM, Prado L, et al. Opioids for restless legs syndrome. Cochrane Database of Systematic Reviews [Internet]. 2016; (6). Available from: http://dx.doi.org/10.1002/14651858.CD006941.pub2.

247. Derry S, Stannard C, Cole P, Wiffen P, Knaggs R, Aldington D, et al. Fentanyl for neuropathic pain in adults. Cochrane Database of Systematic Reviews [Internet]. 2016; (10). Available from: http://dx.doi.org/10.1002/14651858.CD011605.pub2.

248. Dhawan A, Chopra A, Jain R, Yadav D, Vedamurthachar. Effectiveness of yogic breathing intervention on quality of life of opioid dependent users. Int. 2015;8(2):144-7.

249. Di Costanzo E, Rovea A. [The prophylaxis of benzodiazepine withdrawal syndrome in the elderly: the effectiveness of carbamazepine. Double-blind study vs. placebo]. Minerva Psichiatr. 1992;33(4):301-4.

250. Diener HC, Holle D, Dresler T, Gaul C. Chronic Headache Due to Overuse of Analgesics and Anti-Migraine Agents. Dtsch. 2018;115(22):365-70.

251. Dijkstra BA, Krabbe PF, De Jong CA, van der Staak CP. Prediction of withdrawal symptoms during opioid detoxification. Journal of Opioid Management. 2008;4(5):311-9.

252. Dijkstra BAG, De Jong CAJ, Wensing M, Krabbe PFM, Van Der Staak CPF. Opioid detoxification: From controlled clinical trial to clinical practice. The American Journal on Addictions. 2010;19(3):283-90.

253. Doolittle B, Becker W. A case series of buprenorphine/naloxone treatment in a primary care practice. Subst Abus. 2011;32(4):262-5.

254. D'Onofrio G, O'Connor PG, Pantalon MV, Chawarski MC, Busch SH, Owens PH, et al. Emergency department-initiated buprenorphine/naloxone treatment for opioid dependence: a randomized clinical trial. JAMA. 2015;313(16):1636-44. doi: 10.001/jama.2015.3474.

255. D'Onofrio G, Chawarski MC, O'Connor PG, Pantalon MV, Busch SH, Owens PH, et al. Emergency Department-Initiated Buprenorphine for Opioid Dependence with Continuation in Primary Care: Outcomes During and After Intervention. J Gen Intern Med. 2017;32(6):660-6.

References

256. Donohue JM, Barry CL, Stuart EA, Greenfield SF, Song Z, Chernew ME, et al. Effects of global payment and accountable care on medication treatment for alcohol and opioid use disorders. J Addict Med. 2018;12(1):11-8.

257. Dou C, Rebane J, Bardal S. Interventions to improve benzodiazepine tapering success in the elderly: a systematic review. Aging Ment Health. 2018:1-6.

258. Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain-- United States, 2016. JAMA. 2016;315(15):1624-45.

259. Dunn KM, Saunders KW, Rutter CM, Banta-Green CJ, Merrill JO, Sullivan MD, et al. Opioid prescriptions for chronic pain and overdose: a cohort study. Ann Intern Med. 2010;152(2):85-92.

260. Dunn KE, Sigmon SC, Strain EC, Heil SH, Higgins ST. The association between outpatient buprenorphine detoxification duration and clinical treatment outcomes: a review. Drug Alcohol Depend. 2011;119(1-2):1-9.

261. Dunn KE, Saulsgiver KA, Miller ME, Nuzzo PA, Sigmon SC. Characterizing opioid withdrawal during double-blind buprenorphine detoxification. Drug Alcohol Depend. 2015;151:47-55.

262. Dunn KE, Tompkins DA, Bigelow GE, Strain EC. "Efficacy of tramadol extended-release for opioid withdrawal: A randomized clinical trial": Correction. JAMA psychiatry. 2017;74(9):975.

263. Dunn KE, Tompkins DA, Bigelow GE, Strain EC. Efficacy of Tramadol Extended-Release for Opioid Withdrawal: A Randomized Clinical Trial. JAMA psychiatry. 2017;74(9):885-93.

264. Dreyer TRF, Michalski T, Williams BC. Patient outcomes in a medicaid managed care lock-in program. Journal of Managed Care and Specialty Pharmacy. 2015;21(11):1006-12.

265. Dutra L, Stathopoulou G, Basden SL, Leyro TM, Powers MB, Otto MW. A meta-analytic review of psychosocial interventions for substance use disorders. Am J Psychiatry. 2008;165(2):179-87.

266. Dzialdowski A, London M, Tilbury J. A controlled comparison of cognitive behavioural and traditional counselling in a methadone tapering programme. Clin Psychol Psychother. 1998;5(1):47‐53.

267. Eccleston C, Fisher E, Thomas K, Hearn L, Derry S, Stannard C, et al. Interventions for the reduction of prescribed opioid use in chronic non‐cancer pain. Cochrane Database of Systematic Reviews [Internet]. 2017; (11). Available from: http://dx.doi.org/10.1002/14651858.CD010323.pub3.

268. Edlund MJ, Steffick D, Hudson T, Harris KM, Sullivan M. Risk factors for clinically recognized opioid abuse and dependence among veterans using opioids for chronic non-cancer pain. Pain. 2007;129(3):355-62.

269. Edlund MJ, Martin BC, Devries A, Fan MY, Braden JB, Sullivan MD. Trends in use of opioids for chronic noncancer pain among individuals with mental health and substance use disorders: the TROUP study. Clin J Pain. 2010;26(1):1-8.

270. Edlund MJ, Martin BC, Fan M-Y, Braden JB, Devries A, Sullivan MD. An Analysis of Heavy Utilizers of Opioids for Chronic Noncancer Pain in the TROUP Study. J Pain Symptom Manage. 2010;40(2):279-89.

271. Ehrich E, Turncliff R, Du Y, Leigh-Pemberton R, Fernandez E, Jones R, et al. Evaluation of opioid modulation in major depressive disorder. Neuropsychopharmacology. 2015;40(6):1448-55.

References

272. Eibl JK, Gauthier G, Pellegrini D, Daiter J, Varenbut M, Hogenbirk JC, et al. The effectiveness of telemedicine-delivered opioid agonist therapy in a supervised clinical setting. Drug Alcohol Depend. 2017;176:133-8.

273. Eilender P, Ketchen B, Maremmani I, Saenger M, Fareed A. Treatment Approaches for Patients With Opioid Use Disorder and Chronic Noncancer Pain: a Literature Review. Addictive Disorders & Their Treatment. 2016;15(2):85-98.

274. El-Aneed A, Alaghehbandan R, Gladney N, Collins K, MacDonald D, Fischer B. Prescription drug abuse and methods of diversion: The potential role of a pharmacy network. Journal of Substance Use. 2009;14(2):75-83.

275. Elefritz JL, Murphy CV, Papadimos TJ, Lyaker MR. Methadone analgesia in the critically ill. J Crit Care. 2016;34:84-8.

276. El-Hadidy MA, Helaly AMN. Medical and Psychiatric Effects of Long-Term Dependence on High Dose of tramadol. Subst Use Misuse. 2015;50(5):582.

277. Els C, Jackson T, Kunyk D, Lappi V, Sonnenberg B, Hagtvedt R, et al. Adverse events associated with medium‐ and long‐term use of opioids for chronic non‐cancer pain: an overview of Cochrane Reviews. Cochrane Database of Systematic Reviews [Internet]. 2017; (10). Available from: http://dx.doi.org/10.1002/14651858.CD012509.pub2.

278. Elsesser K, Sartory G, Maurer J. The efficacy of complaints management training in facilitating benzodiazepine withdrawal. Behav Res Ther. 1996;34(2):149-56.

279. Enriquez-Puga A, Baker R, Paul S, Villoro-Valdes R. Effect of educational outreach on general practice prescribing of antibiotics and antidepressants: a two-year randomised controlled trial. Scand J Prim Health Care. 2009;27(4):195-201.

280. Esposito E, Barbui C, Patten SB. Patterns of benzodiazepine use in a Canadian population sample. Epidemiol Psichiatr Soc. 2009;18(3):248-54.

281. Evans EA, Sullivan MA. Abuse and misuse of antidepressants. Subst. 2014;5:107-20.

282. Everitt H, Baldwin D, Stuart B, Lipinska G, Mayers A, Malizia A, et al. Antidepressants for insomnia in adults. Cochrane Database of Systematic Reviews [Internet]. 2018; (5). Available from: http://dx.doi.org/10.1002/14651858.CD010753.pub2.

283. Everly JJ, DeFulio A, Koffarnus MN, Leoutsakos JM, Donlin WD, Aklin WM, et al. Employment- based reinforcement of adherence to depot naltrexone in unemployed opioid-dependent adults: a randomized controlled trial. Addiction. 2011;106(7):1309-18.

284. Facey C, Brooks D, Boland JW. Assessment of the appropriateness of naloxone administration to patients receiving long-term opioid therapy. Hosp Pract (Minneap). 2016;44(2):86-91.

285. Faggiano F, Vigna‐Taglianti F, Versino E, Lemma P. Methadone maintenance at different dosages for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2003; (3). Available from: http://dx.doi.org/10.1002/14651858.CD002208.

286. Fareed A, Vayalapalli S, Casarella J, Drexler K. Treatment Outcome for Flexible Dosing Buprenorphine Maintenance Treatment. The American Journal of Drug and Alcohol Abuse. 2012;38(2):155-60.

287. Fatseas M, Serre F, Swendsen J, Auriacombe M. Effects of anxiety and mood disorders on craving and substance use among patients with substance use disorder: An ecological momentary assessment study. Drug Alcohol Depend. 2018;187:242.

References

288. Fava GA, Benasi G, Lucente M, Offidani E, Cosci F, Guidi J. Withdrawal Symptoms after Serotonin-Noradrenaline Reuptake Inhibitor Discontinuation: Systematic Review. Psychother Psychosom. 2018;87(4):195-203.

289. Fava GA, Gatti A, Belaise C, Guidi J, Offidani E. Withdrawal Symptoms after Selective Serotonin Reuptake Inhibitor Discontinuation: A Systematic Review. Psychother Psychosom. 2015;84(2):72-81.

290. Fava GA, Offidani E. The mechanisms of tolerance in antidepressant action. Prog Neuropsychopharmacol Biol Psychiatry. 2011;35(7):1593-602.

291. Feelemyer JP, Jarlais DCD, Arasteh K, Phillips BW, Hagan H. Changes in quality of life (WHOQOL-BREF) and addiction severity index (ASI) among participants in opioid substitution treatment (OST) in low and middle income countries: an international systematic review. Drug Alcohol Depend. 2014;134:251-8.

292. Feingold D, Goor-Aryeh I, Bril S, Delayahu Y, Lev-Ran S. Problematic use of prescription opioids and medicinal cannabis among patients suffering from chronic pain. Pain Med. 2017;18(2):294-306.

293. Fenton MC. An investigation into problem benzodiazepine use among individuals with a prescription [Ph.D.]: Columbia University; 2016.

294. Fergusson DM, Boden JM, Horwood LJ. Exposure to childhood sexual and physical abuse and adjustment in early adulthood. Child Abuse Negl. 2008;32(6):607-19.

295. Ferguson JM, Tourian KA, Rosas GR. High-dose desvenlafaxine in outpatients with major depressive disorder. CNS Spectr. 2012;17(3):121-30.

296. Fernandes LC, Kilicarslan T, Kaplan HL, Tyndale RF, Sellers EM, Romach MK. Treatment of codeine dependence with inhibitors of cytochrome P450 2D6. J Clin Psychopharmacol. 2002;22(3):326-9.

297. Ferri M, Minozzi S, Bo A, Amato L. Slow‐release oral morphine as maintenance therapy for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2013; (6). Available from: http://dx.doi.org/10.1002/14651858.CD009879.pub2.

298. Fibbi M, Silva K, Johnson K, Langer D, Lankenau SE. Denial of prescription opioids among young adults with histories of opioid misuse. Pain Med. 2012;13(8):1040-8.

299. Fiellin DA, Moore BA, Sullivan LE, Becker WC, Pantalon MV, Chawarski MC, et al. Long-term treatment with buprenorphine/naloxone in primary care: results at 2-5 years. Am J Addict. 2008;17(2):116-20.

300. Fiellin DA, Schottenfeld RS, Cutter CJ, Moore BA, Barry DT, O'Connor PG. Primary care-based buprenorphine taper vs maintenance therapy for prescription opioid dependence: a randomized clinical trial. JAMA Intern Med. 2014;174(12):1947-54. doi: 10.001/jamainternmed.2014.5302.

301. Fine PG, Messina J, Xie F, Rathmell J. Long-term safety and tolerability of fentanyl buccal tablet for the treatment of breakthrough pain in opioid-tolerant patients with chronic pain: an 18- month study. J Pain Symptom Manage. 2010;40(5):747-60.

302. Finkelman MD, Jamison RN, Kulich RJ, Butler SF, Jackson WC, Smits N, et al. Cross-validation of short forms of the Screener and Opioid Assessment for Patients with Pain-Revised (SOAPP-R). Drug Alcohol Depend. 2017;178:94.

References

303. Finley EP, Garcia A, Rosen K, McGeary D, Pugh MJ, Potter JS. Evaluating the impact of prescription drug monitoring program implementation: a scoping review. BMC Health Serv Res. 2017;17(1):420.

304. Fischer B, Ialomiteanu A, Boak A, Adlaf E, Rehm J, Mann RE. Prevalence and key covariates of non-medical prescription opioid use among the general secondary student and adult populations in Ontario, Canada. Drug & Alcohol Review. 2013;32(3):276-87.

305. Fishbain DA, Lewis JE, Gao J. Are psychoactive substance (opioid)‐dependent chronic pain patients hyperalgesic? Pain Practice. 2011;11(4):337-43.

306. Fishbain DA, Cole B, Lewis JE, Gao J. Is smoking associated with alcohol‐drug dependence in patients with pain and chronic pain patients? An evidence‐based structured review. Pain Med. 2012;13(9):1212-26.

307. Fister K. Tackling misuse of prescription opioids. Br Med J. 2010;340(7759):1273.

308. Fixsen AM. "I'm Not Waving, I'm Drowning": An Autoethnographical Exploration of Biographical Disruption and Reconstruction During Recovery From Prescribed Benzodiazepine Use. Qual Health Res. 2016;26(4):466.

309. Fixsen AM, Ridge D. Stories of Hell and Healing: Internet Users' Construction of Benzodiazepine Distress and Withdrawal. Qual Health Res. 2017;27(13):2030-41.

310. Fluyau D, Revadigar N, Manobianco BE. Challenges of the pharmacological management of benzodiazepine withdrawal, dependence, and discontinuation. Therapeutic advances in psychopharmacology. 2018;8(5):147-68.

311. Frandsen R, Baandrup L, Kjellberg J, Ibsen R, Jennum P. Increased all-cause mortality with psychotropic medication in Parkinson's disease and controls: A national register-based study. Parkinsonism Relat Disord. 2014;20(11):1124-8.

312. Franklin GM, Mai J, Turner J, Sullivan M, Wickizer T, Fulton-Kehoe D. Bending the prescription opioid dosing and mortality curves: impact of the Washington State opioid dosing guideline. Am J Ind Med. 2012;55(4):325-31.

313. Frauger E, Pauly V, Thirion X, Natali F, Pradel V, Reggio P, et al. Estimation of clonazepam abuse liability: A new method using a reimbursed drug database. Int Clin Psychopharmacol. 2009;24(6):318-24.

314. Fredheim OM, Moksnes K, Borchgrevink PC, Kaasa S, Dale O. Clinical pharmacology of methadone for pain. Acta Anaesthesiol Scand. 2008;52(7):879-89.

315. Freire RC, Hallak JE, Crippa JA, Nardi AE. New treatment options for panic disorder: clinical trials from 2000 to 2010. Expert Opin Pharmacother. 2011;12(9):1419-28.

316. Freire RC, Cosci F, Nardi AE. Update on pharmacological treatment of panic disorder. Minerva Psichiatr. 2011;52(3):145-55.

317. Freynhagen R, Backonja M, Schug S, Lyndon G, Parsons B, Watt S, et al. Pregabalin for the Treatment of Drug and Alcohol Withdrawal Symptoms: A Comprehensive Review. CNS drugs. 2016;30(12):1191-200.

318. Fullagar S. Negotiating the neurochemical self: Anti-depressant consumption in women’s recovery from depression. Health: An Interdisciplinary Journal for the Social Study of Health, Illness and Medicine. 2009;13(4):389-406.

References

319. Furukawa T, Streiner D, Young L, Kinoshita Y. Antidepressants plus benzodiazepines for major depression. Cochrane Database of Systematic Reviews [Internet]. 2001; (3). Available from: http://dx.doi.org/10.1002/14651858.CD001026.

320. Gafoor R, Booth HP, Gulliford MC. Antidepressant utilisation and incidence of weight gain during 10 years' follow-up: population based cohort study. BMJ. 2018;361:k1951.

321. Gahr M, Schönfeldt-Lecuona C, Kölle MA, Freudenmann RW. Withdrawal and discontinuation phenomena associated with tranylcypromine: a systematic review. Pharmacopsychiatry. 2013;46(4):123-9.

322. Implementing Evidence-Based Opioid Prescription Practices in a Primary Care Setting. The Journal for Nurse Practitioners. 2018;14(7):e143-e7.

323. Garcia-Guix A, Mestre-Pinto JI, Tirado-Muñoz J, Domingo-Salvany A, Torrens M. Psychiatric co-morbidity among women with substance use disorders. Advances in Dual Diagnosis. 2018;11(1):1- 3.

324. Garg RK, Fulton-Kehoe D, Turner JA, Bauer AM, Wickizer T, Sullivan MD, et al. Changes in opioid prescribing for Washington workers' compensation claimants after implementation of an opioid dosing guideline for chronic noncancer pain: 2004 to 2010. J Pain. 2013;14(12):1620-8.

325. Garland EL, Froeliger BE, Passik SD, Howard MO. Attentional bias for prescription opioid cues among opioid dependent chronic pain patients. J Behav Med. 2013;36(6):611-20.

326. Garland EL, Froeliger B, Howard MO. Effects of mindfulness-oriented recovery enhancement on reward responsiveness and opioid cue-reactivity. Psychopharmacology (Berl). 2014;231(16):3229- 38.

327. Garland EL, Thomas E, Howard MO. Mindfulness-oriented recovery enhancement ameliorates the impact of pain on self-reported psychological and physical function among opioid- using chronic pain patients. J Pain Symptom Manage. 2014;48(6):1091-9.

328. Garland EL, Howard MO. Opioid attentional bias and cue-elicited craving predict future risk of prescription opioid misuse among chronic pain patients. Drug Alcohol Depend. 2014;144:283-7.

329. Garland EL, Froeliger B, Howard MO. Neurophysiological evidence for remediation of reward processing deficits in chronic pain and opioid misuse following treatment with Mindfulness-Oriented Recovery Enhancement: exploratory ERP findings from a pilot RCT. J Behav Med. 2015;38(2):327-36.

330. Garland EL, Baker AK, Howard MO. Mindfulness-oriented recovery enhancement reduces opioid attentional bias among prescription opioid-treated chronic pain patients. Journal of the Society for Social Work and Research. 2017;8(4):493-509.

331. Gaughwin M, Solomon P, Ali R. Correlates of retention on the South Australian Methadone Program 1981-91. Aust N Z J Public Health. 1998;22(7):771-6.

332. Gerardin M, Victorri-Vigneau C, Guerlais M, Guillou-Landreat M, Grall-Bronnec M, Jolliet P. Benzodiazepines Consumption: Does Dependence Vary with Age? Subst Use Misuse. 2014;49(11):1417-25.

333. Gerra G, Zaimovic A, Giusti F, Moi G, Brewer C. Intravenous flumazenil versus oxazepam tapering in the treatment of benzodiazepine withdrawal: a randomized, placebo-controlled study. Addict Biol. 2002;7(4):385-95.

334. Ghaemi SN, Ostacher MM, El-Mallakh RS, Borrelli D, Baldassano CF, Kelley ME, et al. Antidepressant discontinuation in bipolar depression: a Systematic Treatment Enhancement Program

References for Bipolar Disorder (STEP-BD) randomized clinical trial of long-term effectiveness and safety. J Clin Psychiatry. 2010;71(4):372-80.

335. Ghalehney ZS, Ilbeigi S, Arshadi HR, Afshari R. Superiority of buprenorphine over suboxone in preventing addiction relapse in opioid addicts under maintenance therapy: A double-blind clinical trial. Asia Pacific Journal of Medical Toxicology. 2018;7(1):1-6.

336. Ghitza UE. Overlapping mechanisms of stress-induced relapse to opioid use disorder and chronic pain: Clinical implications. Frontiers in Psychiatry. 2016;7 (MAY) (no pagination)(80).

337. Gibiino S, Mori E, De Ronchi D, Serretti A. Potential benefits of slow titration of paroxetine treatment in an elderly population: Eight-week results from a naturalistic setting. J Clin Psychopharmacol. 2013;33(4):565-9.

338. Gibson K, Cartwright C, Read J. ‘In my life antidepressants have been…': A qualitative analysis of users’ diverse experiences with antidepressants. BMC Psychiatry. 2016;16:7.

339. Gibson L, Bowden K, Hill D, Towle I. The benzodiazepine prescribing support tool - a small scale pilot study. Heroin Addiction and Related Clinical Problems. 2017;19(5):15-22.

340. Gilbert A, Owen N, Innes JM, Sansom L. Trial of an intervention to reduce chronic benzodiazepine use among residents of aged-care accommodation. Aust N Z J Med. 1993;23(4):343- 7.

341. Gillaizeau F, Chan E, Trinquart L, Colombet I, Walton R, Rège‐Walther M, et al. Computerized advice on drug dosage to improve prescribing practice. Cochrane Database of Systematic Reviews [Internet]. 2013; (11). Available from: http://dx.doi.org/10.1002/14651858.CD002894.pub3.

342. Gilson AM, Fishman SM, Wilsey BL, Casamalhuapa C, Baxi H. Time series analysis of California’s prescription monitoring program: Impact on prescribing and multiple provider episodes. The Journal of Pain. 2012;13(2):103-11.

343. GlaxoSmithKline. A double-blind comparative study of withdrawal effects following abrupt discontinuation of treatment with paroxetine in low or high dose or imipramine. GSK - clinical study register [wwwgsk-clinicalstudyregistercom]. 1992.

344. GlaxoSmithKline. A double-blind, short-term comparison of paroxetine and amitriptyline with a 12-month, rater-blind maintenance phase comparing paroxetine, amitriptyline and lithium in the prevention of depressive relapse in patients with major depression, recurrent. GSK - clinical study register [wwwgsk-clinicalstudyregistercom]. 1993.

345. GlaxoSmithKline. A Double-Blind Study to Compare the Maintenance of Efficacy and Relapse Rates in Patients with Obsessive Compulsive Disorder who Responded to Paroxetine, Clomipramine or Placebo in the Short Term Study 136. GSK - clinical study register [wwwgsk- clinicalstudyregistercom]. 1994.

346. GlaxoSmithKline. Long-term treatment with paroxetine of outpatients with obsessive- compulsive disorder: an extension of the companion study. GSK - clinical study register [wwwgsk- clinicalstudyregistercom]. 1994.

347. GlaxoSmithKline. A Long-Term Extension of a Double-Blind, Placebo -Controlled Comparative Study of Paroxetine and Clomipramine in the Treatment of Panic Disorder. GSK - clinical study register [wwwgsk-clinicalstudyregistercom]. 1994.

348. GlaxoSmithKline. A Double-Blind, Placebo-Controlled, Continuation of Study 29060/120 to Assess the Long Term Safety and Efficacy of Paroxetine in the Treatment of Panic Disorder and its

References

Role in the Prevention of Relapse/Recurrence. GSK - clinical study register [wwwgsk- clinicalstudyregistercom]. 1994.

349. GlaxoSmithKline. A double-blind study comparing the efficacy and tolerability of paroxetine and placebo in hospital out-patients with depressive symptoms associated with chronic back pain. GSK - clinical study register [wwwgsk-clinicalstudyregistercom]. 1995.

350. GlaxoSmithKline. An Extension Trial Comparing Paroxetine and Placebo in the Long Term Treatment of Generalized Social Phobia. GSK - clinical study register [wwwgsk- clinicalstudyregistercom]. 1996.

351. GlaxoSmithKline. A Study of the Maintained Efficacy and Safety of Paroxetine versus Placebo in the Long-Term Treatment of Social Phobia. GSK - clinical study register [wwwgsk- clinicalstudyregistercom]. 1999.

352. GlaxoSmithKline. A Study of the Maintained Efficacy and Safety of Paroxetine Versus Placebo in the Long-Term Treatment of Posttraumatic Stress Disorder. GSK - clinical study register [wwwgsk- clinicalstudyregistercom]. 2001.

353. Godfrey C, Heather N, Bowie A, Brodie J, Parrott S, Ashton H, et al. Randomised controlled trial of two brief interventions against long-term benzodiazepine use: Cost-effectiveness. Addiction Research & Theory. 2008;16(4):309-17.

354. Gonzalez G, DiGirolamo G, Romero-Gonzalez M, Smelson D, Ziedonis D, Kolodziej M. Memantine improves buprenorphine/naloxone treatment for opioid dependent young adults. Drug Alcohol Depend. 2015;156:243-53.

355. Goldstein A, Judson BA. Proceedings: Efficacy and side effects of three widely different methadone doses. Proc Natl Conf Methadone Treat. 1973;1:21-44.

356. Gomes T, Juurlink DN, Dhalla IA, Mailis-Gagnon A, Paterson JM, Mamdani MM. Trends in opioid use and dosing among socio-economically disadvantaged patients. Open Med. 2011;5(1):e13- 22.

357. Gomes T, Juurlink DN, Antoniou T, Mamdani MM, Paterson JM, van den Brink W. Gabapentin, opioids, and the risk of opioid-related death: A population-based nested case-control study. PLoS Med. 2017;14(10):e1002396.

358. Gorgels WJ, Oude Voshaar RC, Mol AJ, van de Lisdonk EH, van Balkom AJ, van den Hoogen HJ, et al. Discontinuation of long-term benzodiazepine use by sending a letter to users in family practice: a prospective controlled intervention study. Drug Alcohol Depend. 2005;78(1):49-56.

359. Gorgels WJ, Oude Voshaar RC, Mol AJ, van de Lisdonk EH, Mulder J, van den Hoogen H, et al. Consequences of a benzodiazepine discontinuation programme in family practice on psychotropic medication prescription to the participants. Fam Pract. 2007;24(5):504‐10.

360. Gorgels W, Oude Voshaar R, Mol A, Van De Lisdonk E, Mulder J, Van Den Hoogen H, et al. A benzodiazepine discontinuation programme does not increase the frequency of contacts with the family practice. Scand J Prim Health Care. 2008;26(2):74-9.

361. Gould RL, Coulson MC, Patel N, Highton-Williamson E, Howard RJ. Interventions for reducing benzodiazepine use in older people: meta-analysis of randomised controlled trials. Br J Psychiatry. 2014;204(2):98-107.

References

362. Gowing L, Farrell M, Ali R, White J. Alpha2‐adrenergic agonists for the management of opioid withdrawal. Cochrane Database of Systematic Reviews [Internet]. 2016; (5). Available from: http://dx.doi.org/10.1002/14651858.CD002024.pub5.

363. Gowing L, Ali R, White J, Mbewe D. Buprenorphine for managing opioid withdrawal. Cochrane Database of Systematic Reviews [Internet]. 2017; (2). Available from: http://dx.doi.org/10.1002/14651858.CD002025.pub5.

364. Gowing L, Ali R, White JM. Buprenorphine for the management of opioid withdrawal. Cochrane Database of Systematic Reviews. 2009(3):CD002025.

365. Gowing L, Ali R, White J. Opioid antagonists under heavy sedation or anaesthesia for opioid withdrawal. Cochrane Database of Systematic Reviews [Internet]. 2010; (1). Available from: http://dx.doi.org/10.1002/14651858.CD002022.pub3.

366. Gowing L, Ali R, White J. Opioid antagonists with minimal sedation for opioid withdrawal. Cochrane Database of Systematic Reviews [Internet]. 2017; (5). Available from: http://dx.doi.org/10.1002/14651858.CD002021.pub4.

367. Graham JK, Hanzlick R. Accidental drug deaths in Fulton county, Georgia, 2002: Characteristics, case management and certification issues. Am J Forensic Med Pathol. 2008;29(3):224-30.

368. Garland EL, Howard MO, Zubieta JK, Froeliger B. Restructuring Hedonic Dysregulation in Chronic Pain and Prescription Opioid Misuse: Effects of Mindfulness-Oriented Recovery Enhancement on Responsiveness to Drug Cues and Natural Rewards. Psychother Psychosom. 2017;86(2):111-2.

369. Green TC, Bowman S, Davis C, Los C, McHugh K, Friedmann PD. Discrepancies in addressing overdose prevention through prescription monitoring programs. Drug Alcohol Depend. 2015;153:355-8.

370. Green J. Epidemiology of Opioid Abuse and Addiction. J Emerg Nurs. 2017;43(2):106-13.

371. Griesbach S, Lustig A, Malsin L, Carley B, Westrich KD, Dubois RW. Best practices: An electronic drug alert program to improve safety in an accountable care environment. J Manag Care Pharm. 2015;21(4):330-6.

372. Griffin ML, Dodd DR, Potter JS, Rice LS, Dickinson W, Sparenborg S, et al. Baseline characteristics and treatment outcomes in prescription opioid dependent patients with and without co-occurring psychiatric disorder. Am J Drug Alcohol Abuse. 2014;40(2):157-62.

373. Griffin ML, Bennett HE, Fitzmaurice GM, Hill KP, Provost SE, Weiss RD. Health-related quality of life among prescription opioid-dependent patients: Results from a multi-site study. Am J Addict. 2015;24(4):308-14.

374. Griffin ML, McDermott KA, McHugh RK, Fitzmaurice GM, Jamison RN, Weiss RD. Longitudinal association between pain severity and subsequent opioid use in prescription opioid dependent patients with chronic pain. Drug Alcohol Depend. 2016;163:216-21.

375. Griffin PM. Resilience in police: Opioid use and the double-edged sword [Ph.D.]. Philadelphia, PH: Temple University; 2018.

376. Griswold LM. Neurocognitive function in opioid and non-opioid using chronic pain patients as compared to healthy controls. California: California School of Forensic Studies - Alliant International Univeristy; 2013.

References

377. Gryczynski J, Mitchell SG, Jaffe JH, O'Grady KE, Olsen YK, Schwartz RP. Leaving buprenorphine treatment: patients' reasons for cessation of care. J Subst Abuse Treat. 2014;46(3):356-61.

378. GSK. Clinical comparison of paroxetine and placebo on the symptoms emerging during the taper phase of a chronic benzodiazepine treatment, in patients suffering from a variety of anxiety disorders: GSK; 2002 [Available from: https://www.gsk- clinicalstudyregister.com/study/29060/730#rs.

379. Guaiana G, Barbui C. Discontinuing benzodiazepines: best practices. Epidemiol Psychiatr Sci. 2016;25(3):214-6.

380. Gudin JA, Brennan MJ, Harris ED, Hurwitz PL, Dietze DT, Strader JD. Reduction of opioid use and improvement in chronic pain in opioid-experienced patients after topical analgesic treatment: an exploratory analysis. Postgrad Med. 2018;130(1):42-51.

381. Guina J, Rossetter SR, Derhodes BJ, Nahhas RW, Welton RS. Winner of resident paper award 2014: Benzodiazepines for PTSD: A systematic review and meta-analysis. J Psychiatr Pract. 2015;21(4):281-303.

382. Gunderson EW, Hjelmstrom P, Sumner M, Study I. Effects of a higher-bioavailability buprenorphine/naloxone sublingual tablet versus buprenorphine/naloxone film for the treatment of opioid dependence during induction and stabilization: a multicenter, randomized trial. Clin Ther. 2015;37(10):2244-55.

383. Gunderson EW, Sumner M. Efficacy of Buprenorphine/Naloxone Rapidly Dissolving Sublingual Tablets (BNX-RDT) After Switching From BNX Sublingual Film. J Addict Med. 2016;10(2):124-30.

384. Guo S, Manning V, Yang Y, Koh PK, Chan E, de Souza NN, et al. Lofexidine versus diazepam for the treatment of opioid withdrawal syndrome: A double-blind randomized clinical trial in Singapore. J Subst Abuse Treat. 2018;91:1-11.

385. Hadjistavropoulos T, Kaasalainen S, Williams J, Zacharias R. Improving pain assessment practices and outcomes in long-term care facilities: A mixed methods investigation. Pain Manag Nurs. 2014;15(4):748-59.

386. Hadley SJ, Mandel FS, Schweizer E. Switching from long-term benzodiazepine therapy to pregabalin in patients with generalized anxiety disorder: a double-blind, placebo-controlled trial. J Psychopharmacol. 2012;26(4):461-70. doi: 10.1177/0269881111405360. Epub 2011 Jun 21.

387. Haegerich TM, Paulozzi LJ, Manns BJ, Jones CM. What we know, and don't know, about the impact of state policy and systems-level interventions on prescription drug overdose. Drug Alcohol Depend. 2014;145:34-47.

388. Hajak G, Müller WE, Wittchen HU, Pittrow D, Kirch W. Abuse and dependence potential for the non-benzodiazepine hypnotics zolpidem and zopiclone: a review of case reports and epidemiological data. Addiction (Abingdon, England). 2003;98(10):1371-8.

389. Hajak G, Lemme K, Zisapel N. Lasting treatment effects in a postmarketing surveillance study of prolonged-release melatonin. Int Clin Psychopharmacol. 2015;30(1):36-42.

390. Hale ME, Zimmerman TR, Ma Y, Malamut R. Twelve-month, open-label assessment of long- term safety and abuse potential of hydrocodone extended-release formulated with abuse- deterrence technology in patients with chronic pain. Journal of Opioid Management. 2015;11(5):425- 34.

References

391. Hall MTPM, Wilfong JMSW, Huebner RAP, Posze LMA, Willauer TMPA. Medication-Assisted Treatment Improves Child Permanency Outcomes for Opioid-Using Families in the Child Welfare System. J Subst Abuse Treat. 2016;71:63-7.

392. Hamann S, Sloan P. Oral naltrexone to enhance analgesia in patients receiving continuous intrathecal morphine for chronic pain: a randomized, double-blind, prospective pilot study. Journal of opioid management. 2007;3(3):137‐44.

393. Hammig R, Kohler W, Bonorden-Kleij K, Weber B, Lebentrau K, Berthel T, et al. Safety and tolerability of slow-release oral morphine versus methadone in the treatment of opioid dependence. J Subst Abuse Treat. 2014;47(4):275-81.

394. Hansen L, Penko J, Guzman D, Bangsberg DR, Miaskowski C, Kushel MB. Aberrant Behaviors With Prescription Opioids and Problem Drug Use History in a Community-Based Cohort of HIV- Infected Individuals. J Pain Symptom Manage. 2011;42(6):893-902.

395. Hantouche EG, Guelfi JD, Comet D. Discontinuation of long-term benzodiazepine use: double-blind controlled study of (alpha)-(beta) L-aspartate magnesium versus placebo in 144 chronic users of BZD. Encephale-revue de psychiatrie clinique biologique ET therapeutique. 1998;24(5):469‐79.

396. Hantouche EG, Guelfi JD, Comet D. [alpha-beta L-aspartate magnesium in treatment of chronic benzodiazepine abuse: controlled and double-blind study versus placebo]. Encephale. 1998;24(5):469-79.

397. Harrison-Read PE, Tyrer P, Lawson C, Lack S, Fernandes C, File SE. Flumazenil-precipitated panic and dysphoria in patients dependent on benzodiazepines: a possible aid to abstinence. J Psychopharmacol. 1996;10(2):89-97. doi: 10.1177/026988119601000201.

398. Hartung DM, Kim H, Ahmed SM, Middleton L, Keast S, Deyo RA, et al. Effect of a high dosage opioid prior authorization policy on prescription opioid use, misuse, and overdose outcomes. Subst Abus. 2018.

399. Hartung DM, Evans D, Haxby DG, Kraemer DF, Andeen G, Fagnan LJ. Effect of drug sample removal on prescribing in a family practice clinic. Ann Fam Med. 2010;8(5):402-9.

400. Hartung DM, McCarty D, Fu R, Wiest K, Chalk M, Gastfriend DR. Extended-release naltrexone for alcohol and opioid dependence: a meta-analysis of healthcare utilization studies. J Subst Abuse Treat. 2014;47(2):113-21.

401. Hartzler B, Cotton AJ, Calsyn DA, Guerra R, Gignoux E. Dissolution of a harm reduction track for opiate agonist treatment: Longitudinal impact on treatment retention, substance use and service utilization. The International Journal on Drug Policy. 2010;21(1):82.

402. Hassan AN, Foll BL, Imtiaz S, Rehm J. The effect of post-traumatic stress disorder on the risk of developing prescription opioid use disorder: Results from the National Epidemiologic Survey on Alcohol and Related Conditions III. Drug Alcohol Depend. 2017;179:260.

403. Hatcher AE, Mendoza S, Hansen H. At the Expense of a Life: Race, Class, and the Meaning of Buprenorphine in Pharmaceuticalized "Care". Subst Use Misuse. 2018;53(2):301-10.

404. Hawk K, D'Onofrio G, Fiellin DA, Chawarski MC, O'Connor PG, Owens PH, et al. Past-year Prescription Drug Monitoring Program Opioid Prescriptions and Self-reported Opioid Use in an Emergency Department Population With Opioid Use Disorder. Acad Emerg Med. 2018;25(5):508-16.

405. Hay PJ, Bacaltchuk J. Bulimia nervosa. Clin Evid. 2008;12:12.

References

406. Hay JL, White JM, Bochner F, Somogyi AA, Semple TJ, Rounsefell B. Hyperalgesia in opioid- managed chronic pain and opioid-dependent patients. The Journal of Pain. 2009;10(3):316-22.

407. Hay PJ, Claudino AM. Bulimia nervosa. Clin Evid. 2010;19:19.

408. Helm S, Trescot AM, Colson J, Sehgal N, Silverman S. Opioid antagonists, partial agonists, and agonists/antagonists: the role of office-based detoxification. Pain physician. 2008;11(2):225-35.

409. Hengartner MP, Angst J, Rossler W. Antidepressant Use Prospectively Relates to a Poorer Long-Term Outcome of Depression: Results from a Prospective Community Cohort Study over 30 Years. Psychother Psychosom. 2018;87(3):181-3.

410. Henriksen K, Waal H, Krajci P. Will Non-Compliant 'Hard-To-Treat' Opioid-Dependent patients profit from low threshold methadone treatment? A prospective 15-month evaluation of patients on low dosage methadone treatment at oslo university hospital. Heroin addiction and related clinical problems. 2016;18(3):23‐8.

411. Hesse M, Vanderplasschen W, Rapp RC, Broekaert E, Fridell M. Case management for persons with substance use disorders. Cochrane Database of Systematic Reviews [Internet]. 2007; (4).

412. Heser K, Luck T, Röhr S, Wiese B, Kaduszkiewicz H, Oey A, et al. Potentially inappropriate medication: Association between the use of antidepressant drugs and the subsequent risk for dementia. J Affect Disord. 2018;226:28-35.

413. Heslin KC, Stein JA, Heinzerling KG, Pan D, Magladry C, Hays RD. Clinical correlates of health- related quality of life among opioid-dependent patients. Quality of Life Research: An International Journal of Quality of Life Aspects of Treatment, Care & Rehabilitation. 2011;20(8):1205-13.

414. Hewitt AL. Predictive utility of the Screener and Opioid Assessment for Patients with Pain- Revised (SOAPP-R) for documented aberrant drug-related behaviors in medical charts: Response patterns in false negative scores and important consideration of past medication-related behaviors 2015.

415. Hill D, McAuley A. A comparative study of stakeholder views on take-home naloxone services. Journal of Substance Use. 2012;17(5-6):430-11.

416. Hillhouse MP, Fahey J, Jenkins J, Ling W, Rawson R. Variation in retention by behavioural condition in buprenorphine treatment for opioid dependence. Proceedings of the 71th annual scientific meeting of the college on problems of drug dependence; 2009 june 20-25; reno/sparks, nevada, USA. 2009.

417. Hillhouse MP, Thomas C, Jenkins J, Fahey J, Thornton B, Schaper E. Comparison of contingency management reinforcement schedules provided with buprenorphine for treatment of opioid dependece. Proceedings of the 74th annual scientific meeting of the college on problems of drug dependence; 2012 june 9-14, palm springs, CA. 2012:Abstract no 263.

418. Hindmarch I, Kimber S, Cockle SM. Abrupt and brief discontinuation of antidepressant treatment: effects on cognitive function and psychomotor performance. Int Clin Psychopharmacol. 2000;15(6):305‐18.

419. Hitzeman N, Athale N. Opioids for osteoarthritis of the knee or Hip. Am Fam Physician. 2010;81(9):1094-6.

420. Hojsted J, Ekholm O, Kurita GP, Juel K, Sjogren P. Addictive behaviors related to opioid use for chronic pain: a population-based study. Pain. 2013;154(12):2677-83.

References

421. Holbrook AM, Jones HE, Heil SH, Martin PR, Stine SM, Fischer G, et al. Induction of pregnant women onto opioid-agonist maintenance medication: an analysis of withdrawal symptoms and study retention. Drug Alcohol Depend. 2013;132(1-2):329-34.

422. Holland RL. Panic disorder and treatment withdrawal: a case series of seventy patients withdrawn double-blind from fluvoxamine or placebo. Human psychopharmacology. 1995;10:240‐2.

423. Holliday SM, Hayes C, Dunlop AJ, Morgan S, Tapley A, Henderson KM, et al. Does brief chronic pain management education change opioid prescribing rates? A pragmatic trial in Australian early-career general practitioners. Pain. 2017;158(2):278-88.

424. Horspool MJ, Seivewright N, Armitage CJ, Mathers N. Post-treatment outcomes of buprenorphine detoxification in community settings: a systematic review. Eur Addict Res. 2008;14(4):179-85.

425. Houtsmuller EJ, Walsh SL, Schuh KJ, Johnson RE, Stitzer ML, Bigelow GE. Dose-response analysis of opioid cross-tolerance and withdrawal suppression during LAAM maintenance. J Pharmacol Exp Ther. 1998;285(2):387‐96.

426. Howard HPMSWL. Experiences of opioid-dependent women in their prenatal and postpartum care: Implications for social workers in health care. Soc Work Health Care. 2016;55(1):61- 85.

427. Howe CQ, Sullivan MD, Saunders KW, Merrill JO, Banta-Green CJ, Weisner C, et al. Depression and ambivalence toward chronic opioid therapy for chronic noncancer pain. The Clinical Journal of Pain. 2012;28(7):561-6.

428. Hser YI, Li J, Jiang H, Zhang R, Du J, Zhang C, et al. Effects of a randomized contingency management intervention on opiate abstinence and retention in methadone maintenance treatment in China. Addiction. 2011;106(10):1801-9.

429. Hser YI, Evans E, Huang D, Weiss R, Saxon A, Carroll KM, et al. Long-term outcomes after randomization to buprenorphine/naloxone versus methadone in a multi-site trial. Addiction. 2016;111(4):695-705.

430. Hser YI, Huang D, Saxon AJ, Woody G, Moskowitz AL, Matthews AG, et al. Distinctive Trajectories of Opioid Use Over an Extended Follow-up of Patients in a Multisite Trial on Buprenorphine + Naloxone and Methadone. J Addict Med. 2017;11(1):63-9.

431. Hsu DJ, Ghaemi SN, El-Mallakh RS, Baldassano CF, Ostacher MJ, Goodwin FK, et al. Antidepressant discontinuation and mood episode relapse in bipolar disorder. 157th annual meeting of the american psychiatric association; 2004 may 1-6; new york, NY. 2004:NR26.

432. Hu L, Matthews A, Shmueli-Blumberg D, Killeen TK, Tai B, VanVeldhuisen P. Prevalence of obesity for opioid- and stimulant-dependent participants in substance use treatment clinical trials. Drug Alcohol Depend. 2018;190:255.

433. Huang WY, Xiao XS, Liu Y, Liao XQ, Zhou DW, Dai H. The comparison of effects of rapid opiate detoxification with ketamine complex and with tramadol and naltrexone under general anesthesia with propofol. Chinese journal of clinical rehabilitation: chin j clin rehab. 2002;6(23):3625‐6.

434. Hunt G, Siegfried N, Morley K, Sitharthan T, Cleary M. Psychosocial interventions for people with both severe mental illness and substance misuse. Cochrane Database of Systematic Reviews [Internet]. 2013; (10). Available from: http://dx.doi.org/10.1002/14651858.CD001088.pub3.

References

435. Hussain SS, Farhat S, Rather YH, Abbas Z. Comparative trial to study the effectiveness of clonidine hydrochloride and buprenorphine-naloxone in opioid withdrawal - a hospital based study. J Clin Diagn Res. 2015;9(1):FC01-4.

436. Hussain S, Najmi A, Anil D. Risk of Parkinson's disease in zolpidem user: a systematic review and meta-analysis. Movement disorders Conference: 1st pan american parkinson's disease and movement disorders congress United states. 2017;32(no pagination).

437. Iafrati S. The investment and regenerative value of addiction treatment. Drugs and Alcohol Today. 2015;15(1):12-20.

438. Iannone A, Pereira DA, de Oliveira Silva Naves J, Moreira VAM, Da-Silva SL. Use of antidepressant medication during cognitive and depressive treatment in Brazilian elderly patients: A 180 day program. Clinical Neuropsychiatry: Journal of Treatment Evaluation. 2013;10(2):55-62.

439. Ilgen MA, Haas E, Czyz E, Webster L, Sorrell JT, Chermack S. Treating Chronic Pain in Veterans Presenting to an Addictions Treatment Program. Cognitive and Behavioral Practice. 2011;18(1):149- 60.

440. Inciardi JA, Surratt HL, Cicero TJ, Beard RA. Prescription opioid abuse and diversion in an urban community: the results of an ultrarapid assessment. Pain Med. 2009;10(3):537-48.

441. Imani S, Atef Vahid MK, Gharraee B, Habibi M, Bowen S, Noroozi A. Comparing Mindfulness- Based Group Therapy With Treatment as Usual for Opioid Dependents: A Pilot Randomized Clinical Trial Study Protocol. Iran. 2015;9(1):e216.

442. Isaka M. Withdrawal symptoms of benzodiazepines in panic disorder patients' pharmacotherapy. Journal of the osaka city medical center. 2009;58(1‐2):11‐20.

443. Ives TJ, Chelminski PR, Hammett-Stabler CA, Malone RM, Perhac JS, Potisek NM, et al. Predictors of opioid misuse in patients with chronic pain: a prospective cohort study. BMC Health Serv Res. 2006;6:46.

444. Iwanicki JL, Severtson SG, McDaniel H, Rosenblum A, Fong C, Cicero TJ, et al. Abuse and diversion of immediate release opioid analgesics as compared to extended release formulations in the United States. PLoS One. 2016;11(12):12.

445. Iyer S, Naganathan V, McLachlan AJ, Le Couteur DG. Medication withdrawal trials in people aged 65 years and older: a systematic review. Drugs Aging. 2008;25(12):1021-31.

446. Jackman RP, Purvis JM, Mallett BS. Chronic nonmalignant pain in primary care. Am Fam Physician. 2008;78(10):1155-62.

447. Jacob SA, Rahman AFA, Hassali MAA. Attitudes and beliefs of patients with chronic depression toward antidepressants and depression. Neuropsychiatr Dis Treat. 2015;11:1339-47.

448. Jacobsen PL, Harper L, Chrones L, Chan S, Mahableshwarkar AR. Safety and tolerability of vortioxetine (15 and 20 mg) in patients with major depressive disorder: results of an open-label, flexible-dose, 52-week extension study. Int Clin Psychopharmacol. 2015;30(5):255‐64.

449. Jain NMD, Bagwell ELL, O'Mahony SMDMS. When the Apple Falls Far from the Tree: A Systematic Approach to Providing Quality Care for Patients at High-Risk for Opioid-Dependence (TH334). J Pain Symptom Manage. 2017;53(2):331.

450. James NM. A double blind trial of nitrazepam and flurazepam as sedatives. N Z Med J. 1975;81(535):246‐8.

References

451. Jamison RN, Ross EL, Michna E, Chen LQ, Holcomb C, Wasan AD. Substance misuse treatment for high-risk chronic pain patients on opioid therapy: a randomized trial. Pain. 2010;150(3):390-400. doi: 10.1016/j.pain.2010.02.033. Epub Mar 23.

452. Jamison RN, Serraillier J, Michna E. Assessment and treatment of abuse risk in opioid prescribing for chronic pain. Pain research and treatment. 2011;2011:941808.

453. Jamison RN, Scanlan E, Matthews ML, Jurcik DC, Ross EL. Attitudes of Primary Care Practitioners in Managing Chronic Pain Patients Prescribed Opioids for Pain: A Prospective Longitudinal Controlled Trial. Pain Med. 2016;17(1):99-113.

454. Jarvis BP, Holtyn AF, Subramaniam S, Tompkins DA, Oga EA, Bigelow GE, et al. Extended- release injectable naltrexone for opioid use disorder: a systematic review. Addiction. 2018;113(7):1188-209.

455. Jean-Bart E, Moutet C, Dauphinot V, Krolak-Salmon P, Mouchoux C. Exposure to anticholinergic and sedative medicines as indicators of high-risk prescriptions in the elderly. Int J Clin Pharm. 2017;39(6):1237‐47.

456. Jiang K, Li L, Wang X, Fang M, Shi J, Cao Q, et al. Efficacy and tolerability of escitalopram in treatment of major depressive disorder with anxiety symptoms: a 24-week, open-label, prospective study in Chinese population. Neuropsychiatr Dis Treat. 2017;13:515‐26.

457. Joffe H, Guthrie KA, Cohen L, Carpenter J, Larson J, La Croix A. Hot flashes recur rapidly after discontinuation of the ssri escitalopram: results from the msflash research network. Menopause (new york, NY). 2011;18(12):1347 S‐15.

458. Jo‐Hanna I, Zgaga L, Sweeney B, Keenan E, Darker C, Smyth BP, et al. A naturalistic longitudinal analysis of post‐detoxification outcomes in opioid‐dependent patients. Drug and alcohol review. 2018;37(S1):S339-S47.

459. Johnson H, Paulozzi L, Porucznik C, Mack K, Herter B. Decline in drug overdose deaths after state policy changes - Florida, 2010-2012. MMWR Morb Mortal Wkly Rep. 2014;63(26):569-74.

460. Jones RT, Mendelson J. Buprenorphine and naloxone interactions in opite dependente volunteers. NIDA Res Monogr. 1994;141:144.

461. Jones HE, Heil SH, Baewert A, Arria AM, Kaltenbach K, Martin PR, et al. Buprenorphine treatment of opioid-dependent pregnant women: a comprehensive review. Addiction. 2012;107 Suppl 1:5-27.

462. Julius D, Sturesson B, Kools D, Prestamburgo D, Pflugmacher R, Vajkoczy P. Opioid use in relation to treatment outcome in patients with low back pain: data from a prospective randomized controlled trial. Global spine journal. 2017;Conference: 2017 Global Spine Congress. Italy. 7(2 Supplement 1):109S.

463. Kahan M, Mailis-Gagnon A, Wilson L, Srivastava A. Canadian guideline for safe and effective use of opioids for chronic noncancer pain - Clinical summary for family physicians. Part 1: General population. [French, English]. Can Fam Physician. 2011;57(11):1257-66.

464. Kahan M, Wilson L, Mailis-Gagnon A, Srivastava A. Canadian guideline for safe and effective use of opioids for chronic noncancer pain - Clinical summary for family physicians. Part 2: Special populations. [French, English]. Can Fam Physician. 2011;57(11):1269-76.

References

465. Kahler ZP, Musey PI, Schaffer JT, Johnson AN, Strachan CC, Shufflebarger CM. Effect of A "no Superuser Opioid Prescription" Policy on ED visits and statewide opioid prescription. West J Emerg Med. 2017;18(5):894-902.

466. Kamara SG, Van der Hyde VA. Outcomes of regular versus extended outpatient alcohol/drug treatment. Part II. Medical, psychiatric, legal and social problems. Med Law. 1998;17(1):131‐42.

467. Karow A, Verthein U, Pukrop R, Reimer J, Haasen C, Krausz M, et al. Quality of life profiles and changes in the course of maintenance treatment among 1,015 patients with severe opioid dependence. Subst Use Misuse. 2011;46(6):705-15.

468. Kasteenpohja T, Marttunen M, Aalto-Setala T, Perala J, Saarni SI, Suvisaari J. Treatment received and treatment adequacy of depressive disorders among young adults in Finland. BMC Psychiatry. 2015;15 (1) (no pagination)(47).

469. Kasteenpohja T, Marttunen M, Aalto-Setala T, Perala J, Saarni SI, Suvisaari J. Treatment adequacy of anxiety disorders among young adults in Finland. BMC Psychiatry. 2016;16 (1) (no pagination)(63).

470. Katz EC, Schwartz RP, King S, Highfield DA, O'Grady KE, Billings T, et al. Brief vs. extended buprenorphine detoxification in a community treatment program: engagement and short-term outcomes. Am J Drug Alcohol Abuse. 2009;35(2):63-7.

471. Kavukcu E, Akdeniz M, Avci HH, Altug M, Oner M. Chronic noncancer pain management in primary care: Family medicine physicians' risk assessment of opioid misuse. Postgrad Med. 2015;127(1):22-6.

472. Kawai K, Yoshizawa K, Fujie M, Kobayashi H, Ogawa Y, Yajima T. Use of Fentanyl Patch for Treatment of Moderate-to-severe Chronic Noncancer Pain: postmarketing Surveillance of Medical Practice in Japan Using a Risk Minimization Action Plan. Pain practice. 2017;17(2):239‐48.

473. Kawasaki SS, Jacapraro JS, Rastegar DA. Safety and effectiveness of a fixed-dose phenobarbital protocol for inpatient benzodiazepine detoxification. J Subst Abuse Treat. 2012;43(3):331-4.

474. Keast SL, Nesser N, Farmer K. Strategies aimed at controlling misuse and abuse of opioid prescription medications in a state Medicaid program: a policymaker's perspective. The American Journal of Drug and Alcohol Abuse. 2015;41(1):1.

475. Keast SL, Kim H, Deyo RA, Middleton L, McConnell KJ, Zhang K, et al. Effects of a prior authorization policy for extended‐release/long‐acting opioids on utilization and outcomes in a state medicaid program. Addiction. 2018.

476. Kelly SM, Brown BS, Katz EC, O'Grady KE, Mitchell SG, King S, et al. A comparison of attitudes toward opioid agonist treatment among short-term buprenorphine patients. Am J Drug Alcohol Abuse. 2012;38(3):233-8.

477. Kelly TM, Daley DC, Douaihy AB. Contingency management for patients with dual disorders in intensive outpatient treatment for addiction. J Dual Diagn. 2014;10(3):108-17.

478. Kendall SE, Sjøgren P, De Mattos Pimenta CA, Højsted J, Kurita GP. The cognitive effects of opioids in chronic non-cancer pain. Pain. 2010;150(2):225-30.

479. Kendrick T. Long-term antidepressant treatment: time for a review? Prescriber. 2015;26(19):7-10.

References

480. Kern AM, Akerman SC, Nordstrom BR. Opiate dependence in schizophrenia: case presentation and literature review. Journal of dual diagnosis. 2014;10(1):52-7.

481. Khalifa N, Duggan C, Stoffers J, Huband N, Völlm B, Ferriter M, et al. Pharmacological interventions for antisocial personality disorder. Cochrane Database of Systematic Reviews [Internet]. 2010; (8). Available from: http://dx.doi.org/10.1002/14651858.CD007667.pub2.

482. Khatami M, Woody G, O'Brien C, Mintz J. Biofeedback treatment of narcotic addiction: a double-blind study. Drug Alcohol Depend. 1982;9(2):111-7.

483. Kheirabadi GR, Salehi M, Bahrami M, Maracy MR. Gabapentin, Pregabalin, and Placebo in Reducing Opioid Withdrawal Symptoms in Opioid-dependent Individuals: A Randomized-controlled Trial. Addictive Disorders and their Treatment. 2018;17(2):55-64.

484. Kidorf M, Brooner RK, Gandotra N, Antoine D, King VL, Peirce J, et al. Reinforcing integrated psychiatric service attendance in an opioid-agonist program: a randomized and controlled trial. Drug Alcohol Depend. 2013;133(1):30-6.

485. King VL, Stoller KB, Kidorf M, Kindbom K, Hursh S, Brady T, et al. Assessing the effectiveness of an Internet-based videoconferencing platform for delivering intensified substance abuse counseling. J Subst Abuse Treat. 2009;36(3):331-8.

486. King VL, Brooner RK, Peirce JM, Kolodner K, Kidorf MS. A randomized trial of Web-based videoconferencing for substance abuse counseling. J Subst Abuse Treat. 2014;46(1):36-42.

487. Kitajima R, Miyamoto S, Tenjin T, Ojima K, Ogino S, Miyake N, et al. Effects of tapering of long-term benzodiazepines on cognitive function in patients with schizophrenia receiving a second- generation antipsychotic. Prog Neuropsychopharmacol Biol Psychiatry. 2012;36(2):300-6.

488. Klein E, Colin V, Stolk J, Lenox RH. Alprazolam withdrawal in patients with panic disorder and generalized anxiety disorder: vulnerability and effect of carbamazepine. Am J Psychiatry. 1994;151(12):1760-6. doi: 10.176/ajp.151.12..

489. Klemperer EM, Hughes J. Does the efficacy of medications for substance abuse treatment decrease over time? Drug Alcohol Depend. 2015;146((Klemperer E.M.; Hughes J.) Psychiatry, University of Vermont, Burlington, United States):e160.

490. Knisely JSP, Wunsch MJMD, Cropsey KLP, Campbell EMMS. Prescription Opioid Misuse Index: A brief questionnaire to assess misuse. J Subst Abuse Treat. 2008;35(4):380-6.

491. Kondoni T, Kouimtsidis C. What stops people on long-term opioid substitution from completing treatment? Drugs and Alcohol Today. 2017;17(4):232-41.

492. Kornowski J. [The comparison of tianeptine and carbamazepine in benzodiazepines withdrawal symptoms]. Psychiatr Pol. 2002;36(6 Suppl):311-8.

493. Kouvonen A, Vahtera J, Pentti J, Korhonen MJ, Oksanen T, Salo P, et al. Antidepressant use and work-related injuries. Psychol Med. 2016;46(7):1391-9.

494. Kovitwanichkanont T, Day CA. Prescription Opioid Misuse and Public Health Approach in Australia. Subst Use Misuse. 2018;53(2):200-5.

495. Kripke DF, Langer RD, Kline LE. Hypnotics' association with mortality or cancer: a matched cohort study. BMJ Open. 2012;2(1):e000850.

References

496. Krupitsky E, Nunes EV, Ling W, Illeperuma A, Gastfriend DR, Silverman BL. Injectable extended-release naltrexone for opioid dependence: A double-blind, placebo-controlled, multicentre randomised trial. The Lancet. 2011;377(9776):1506-13.

497. Patricia Biller K. Evaluation of CDC Opioid Prescribing and Medisafe Apps. The Journal for Nurse Practitioners. 2017;13(6):e293-e5.

498. Krupitsky E, Zvartau E, Blokhina E, Verbitskaya E, Wahlgren V, Tsoy-Podosenin M, et al. Randomized trial of long-acting sustained-release naltrexone implant vs oral naltrexone or placebo for preventing relapse to opioid dependence. JAMA psychiatry. 2012;69(9):973-81.

499. Krupitsky E, Nunes EV, Ling W, Gastfriend DR, Memisoglu A, Silverman BL. Injectable extended‐release naltrexone (XR-NTX) for opioid dependence: Long‐term safety and effectiveness. Addiction. 2013;108(9):1628-37.

500. Krupitsky EM. Injectable extended-release naltrexone for the prevention of relapse to opioid dependence following opioid detoxification. Zhurnal Nevrologii i Psihiatrii imeni SS. 2014;Korsakova. 2014(5):64-72.

501. Krupitsky EM, Blokhina EA, Zvartau EE, Verbitskaya VE, Bushara EM, Tiurina AA, et al. A double-blind randomized placebo-controlled study of the efficacy of the combined treatment with naltrexone and guanfacine for relapse prevention in opiate dependence. Zh Nevrol Psikhiatr Im S S Korsakova. 2015;115(10):39‐46.

502. Krupitsky EM, Ilyuk RD, Mikhailov AD, Kazankov KA, Rybakova KV, Skurat EP, et al. A randomized single blind study of the efficacy of pregabalin in the treatment of opioid withdrawal syndrome. Zh Nevrol Psikhiatr Im S S Korsakova. 2016;116(7):29‐36.

503. Krupitsky EMDPD, Zvartau EMDPD, Blokhina EMDP, Verbitskaya EP, Wahlgren VMDP, Tsoy- Podosenin MMDP, et al. Anhedonia, depression, anxiety, and craving in opiate dependent patients stabilized on oral naltrexone or an extended release naltrexone implant. Early Childhood Research Quarterly. 2016;42(5):614.

504. Krupitskii EM, Ilyuk RD, Mikhailov AD, Kazankov KA, Rybakova KV, Skurat EP, et al. A Randomized Controlled Study of the Efficacy of Pregabalin in the Treatment of Opiate Withdrawal Syndrome. Neurosci Behav Physiol. 2017;47(9):1094-101.

505. Kua EH. "Interventions for reducing benzodiazepine use in older people: Meta-analysis of randomised controlled trials": Correction. The British Journal of Psychiatry. 2014;205(4):331.

506. Kumar P, Jain MK. Gabapentin in the management of pentazocine dependence: a potent analgesic--anticraving agent. J Assoc Physicians India. 2004;51:673‐6.

507. Kunoe N, Lobmaier P, Vederhus JK, Hjerkinn B, Hegstad S, Gossop M, et al. Retention in naltrexone implant treatment for opioid dependence. Drug Alcohol Depend. 2010;111(1-2):166-9.

508. Kunz D, Bineau S, Maman K, Milea D, Toumi M. Benzodiazepine discontinuation with prolonged-release melatonin: hints from a German longitudinal prescription database. Expert Opin Pharmacother. 2012;13(1):9-16.

509. Kuperwasser B, Häufel T, Kelly K, Etropolski M, Laschewski F, Okamoto A. Incidence and severity of gastrointestinal treatment-emergent adverse events in patients treated with tapentadol extended release (ER) or oxycodone controlled release (CR) for relief of chronic osteoarthritis knee pain. Osteoarthritis Cartilage. 2009;17:S178.

References

510. Kurita GP, Hojsted J, Sjogren P. Tapering off long-term opioid therapy in chronic non-cancer pain patients: A randomized clinical trial. European Journal of Pain (United Kingdom). 2018;22(8):1528-43.

511. Kurokawa T, Katayama S, Amino K, Iimori M. Modified electroconvulsive therapy to control delirium caused by benzodiazepine withdrawal. Journal of tokyo medical university. 2011;69(2):243‐8.

512. Kuyken W, Byford S, Taylor RS, Watkins E, Holden E, White K, et al. Mindfulness-based cognitive therapy to prevent relapse in recurrent depression. J Consult Clin Psychol. 2008;76(6):966- 78.

513. Kuyken W, Hayes R, Barrett B, Byng R, Dalgleish T, Kessler D, et al. The effectiveness and cost-effectiveness of mindfulness-based cognitive therapy compared with maintenance antidepressant treatment in the prevention of depressive relapse/recurrence: results of a randomised controlled trial (the PREVENT study). Comparative Study

Randomized Controlled Trial

Research Support, Non-U.S. Gov't. 2015 Sep. Report No.: 2046-4924 Contract No.: 73.

514. LaBelle CTBSNRNBCC, Han SCBA, Bergeron AMPHL, Samet JHMDMAMPH. Office-Based Opioid Treatment with Buprenorphine (OBOT-B): Statewide Implementation of the Massachusetts Collaborative Care Model in Community Health Centers. J Subst Abuse Treat. 2016;60:6.

515. Lader M, Olajide D. A comparison of buspirone and placebo in relieving benzodiazepine withdrawal symptoms. J Clin Psychopharmacol. 1987;7(1):11-5.

516. Lader M, Farr I, Morton S. A comparison of alpidem and placebo in relieving benzodiazepine withdrawal symptoms. Int Clin Psychopharmacol. 1993;8(1):31-6.

517. Ladewig D. Buprenorphine in medium term substitutions: results of a multicentre, double- blind study. Sucht. 1998;44(5):354.

518. Lander LR, Gurka KK, Marshalek PJ, Riffon M, Sullivan CR. A comparison of pregnancy-only versus mixed-gender group therapy among pregnant women with opioid use disorder. Soc Work Res. 2015;39(4):235-44.

519. Landreat MG, Vigneau CV, Hardouin JB, Bronnec MG, Marais M, Venisse JL, et al. Can We Say that Seniors Are Addicted to Benzodiazepines? Subst Use Misuse. 2010;45(12):1988-99.

520. Lankford A. Low-dose doxepin (3 and 6 mg) for the treatment of insomnia. Future Neurol. 2011;6(2):143-54.

521. Lanier WA, Johnson EM, Rolfs RT, Friedrichs MD, Grey TC. Risk factors for prescription opioid‐related death, Utah, 2008–2009. Pain Med. 2012;13(12):1580-9.

522. Lanzillotta JA, Clark A, Starbuck E, Kean EB, Kalarchian M. The Impact of Patient Characteristics and Postoperative Opioid Exposure on Prolonged Postoperative Opioid Use: An Integrative Review. Pain management nursing : official journal of the American Society of Pain Management Nurses. 2018.

523. Larney S, Gowing L, Mattick RP, Farrell M, Hall W, Degenhardt L. A systematic review and meta-analysis of naltrexone implants for the treatment of opioid dependence. Drug & Alcohol Review. 2014;33(2):115-28.

References

524. LaRowe LR, Chilcott LN, Zvolensky MJ, Vanable PA, Flood K, Ditre JW. Associations between Pain-Related Anxiety, Gender, and Prescription Opioid Misuse among Tobacco Smokers Living with HIV/AIDS. Subst Use Misuse. 2018;53(13):2210-9.

525. Lasser KE, Shanahan C, Parker V, Beers D, Xuan Z, Heymann O, et al. A Multicomponent Intervention to Improve Primary Care Provider Adherence to Chronic Opioid Therapy Guidelines and Reduce Opioid Misuse: A Cluster Randomized Controlled Trial Protocol. J Subst Abuse Treat. 2016;60:101-9.

526. Lauer JW. The effect of tricyclic antidepressant compounds on patients with passive- dependent personality traits. Current therapeutic research. 1976;19(5):495‐505.

527. Lavie E, Fatseas M, Denis C, Auriacombe M. Benzodiazepine use among opiate-dependent subjects in buprenorphine maintenance treatment: Correlates of use, abuse and dependence. Drug Alcohol Depend. 2009;99(1-3):338-44.

528. Leas B, Goldmann D. Controlled medication agreements for opioid use in chronic pain. Philadelphia: Center for Evidence-based Practice (CEP), 2010.

529. Lecrubier Y, Fessard N. Arrêt des benzodiazépines chez des consommateurs chroniques : un essai en double insu du gluconate de lithium vs placebo. Annales Médico-psychologiques, revue psychiatrique. 2005;163(1):24-9.

530. Lee JH. The effects of music on pain: A meta-analysis. J Music Ther. 2016;53(4):430-77.

531. Leonard R, Kourlas H. Too much of a good thing? Treating the emerging syndrome of opioid- induced hyperalgesia. J Pharm Pract. 2008;21(2):165-8.

532. Leslie DL, Milchak W, Gastfriend DR, Herschman PL, Bixler EO, Velott DL, et al. Effects of injectable extended‐release naltrexone (XR‐NTX) for opioid dependence on residential rehabilitation outcomes and early follow‐up. The American Journal on Addictions. 2015;24(3):265-70.

533. Levitt AJ. Discontinuation Study Of Citalopram (Antidepressant) in Depressed Adolescents. Clinicaltrialsgov [wwwclinicaltrialsgov]. 2005.

534. Li L. Safety of opioids for non-cancer pain 2012.

535. Li X, Fan H, Sun L, Liang J, Deng Y. Changes and the impact on immune function of opioid- dependent subjects by Jitai tabelets during the withdrawal stage. Chinese journal of endemiology. 2017;38(4):531‐6.

536. Liebschutz JM, Lange AV, Heymann OD, Lasser KE, Corey P, Shanahan CW, et al. Communication between nurse care managers and patients who take opioids for chronic pain: Strategies for exploring aberrant behavior. Journal of Opioid Management. 2018;14(3):191-202.

537. Lin CW, McLachlan AJ, Latimer J, Day RO, Billot L, Koes BW, et al. OPAL: a randomised, placebo-controlled trial of opioid analgesia for the reduction of pain severity in people with acute spinal pain. Trial protocol. BMJ Open. 2016;6(8):e011278.

538. Ling W, Hillhouse M, Ang A, Jenkins J, Fahey J. Comparison of behavioral treatment conditions in buprenorphine maintenance. Addiction. 2013;108(10):1788-98.

539. Ling W, Hillhouse M, Jenkins J, Miotto K, Torrington M, Chapleo C. Comparisons of analgesic potency and side effects of buprenorphine and buprenorphine with ultra-low-dose naloxone. J Addict Med. 2012;6(2):118-23.

References

540. Ling W, Wesson DR, Charuvastra C, Klett CJ. A controlled trial comparing buprenorphine and methadone maintenance in opioid dependence. Arch Gen Psychiatry. 1996;53(5):401-7.

541. Lintzeris N, Leung SY, Dunlop AJ, Larance B, White N, Rivas GR, et al. A randomised controlled trial of sublingual buprenorphine-naloxone film versus tablets in the management of opioid dependence. Drug Alcohol Depend. 2013;131(1-2):119-26.

542. Liu TT, Shi J, Epstein DH, Bao YP, Lu L. A meta-analysis of acupuncture combined with opioid receptor agonists for treatment of opiate-withdrawal symptoms. Cell Mol Neurobiol. 2009;29(4):449- 54.

543. Llorca PM, Bodkin JA, Spann M, Ball SG, Russell JM, Ball SG. A continuous responder analysis from a long-term trial of duloxetine treatment for generalized anxiety disorder: implications for the randomized withdrawal relapse prevention study design. J Clin Psychopharmacol. 2009;29(1):96-7.

544. Lobmaier PP, Kunøe N, Gossop M, Waal H. Naltrexone depot formulations for opioid and alcohol dependence: a systematic review. CNS Neurosci Ther. 2011;17(6):629-36.

545. Lobmaier P, Kornor H, Kunoe N, Bjørndal A. Sustained‐Release Naltrexone For Opioid Dependence. Cochrane Database of Systematic Reviews [Internet]. 2008; (2). Available from: http://dx.doi.org/10.1002/14651858.CD006140.pub2.

546. Lofwall MR, Babalonis S, Nuzzo PA, Elayi SC, Walsh SL. Opioid withdrawal suppression efficacy of oral dronabinol in opioid dependent humans. Drug Alcohol Depend. 2016;164:143-50.

547. Lopez-Peig C, Mundet X, Casabella B, del Val JL, Lacasta D, Diogene E. Analysis of benzodiazepine withdrawal program managed by primary care nurses in Spain. BMC Res Notes. 2012;5:684.

548. López-Torres J, Párraga I, Del Campo JM, Villena A. Follow up of patients who start treatment with antidepressants: Treatment satisfaction, treatment compliance, efficacy and safety. BMC Psychiatry. 2013;13:9.

549. Lu L, Liu Y, Zhu W, Shi J, Liu Y, Ling W, et al. Traditional medicine in the treatment of drug addiction. Am J Drug Alcohol Abuse. 2009;35(1):1-11.

550. Lucas GM, Chaudhry A, Hsu J, Woodson T, Lau B, Olsen Y, et al. Clinic-based treatment of opioid-dependent HIV-infected patients versus referral to an opioid treatment program: A randomized trial. Ann Intern Med. 2010;152(11):704-11.

551. Lucas P, Walsh Z, Crosby K, Callaway R, Belle-Isle L, Kay R, et al. Substituting cannabis for prescription drugs, alcohol and other substances among medical cannabis patients: The impact of contextual factors. Drug and alcohol review. 2016;35(3):326-33.

552. Lucas P, Walsh Z. Medical cannabis access, use, and substitution for prescription opioids and other substances: A survey of authorized medical cannabis patients. The International Journal on Drug Policy. 2017;42:30.

553. Lum PJ, Little S, Botsko M, Hersh D, Thawley RE, Egan JE, et al. Opioid-prescribing practices and provider confidence recognizing opioid analgesic abuse in HIV primary care settings. J Acquir Immune Defic Syndr. 2011;56 Suppl 1:S91-7.

554. Lovejoy TI, Morasco BJ, Demidenko MI, Meath TH, Frank JW, Dobscha SK. Reasons for discontinuation of long-term opioid therapy in patients with and without substance use disorders. Pain. 2017;158(3):526-34.

References

555. Ma K, Jiang W, Zhou Q, Du DP. The efficacy of oxycodone for management of acute pain episodes in chronic neck pain patients. Int J Clin Pract. 2008;62(2):241-7.

556. Mackie SEMD, McHugh RKP, McDermott KBA, Griffin MLP, Winkelman JWMDP, Weiss RDMD. Prevalence of restless legs syndrome during detoxification from alcohol and opioids. J Subst Abuse Treat. 2017;73:35-9.

557. Madden ME, Shapiro SL. The methadone epidemic: methadone-related deaths on the rise in Vermont. Am J Forensic Med Pathol. 2011;32(2):131-5.

558. Maddux JF, Prihoda TJ, Vogtsberger KN. The relationship of methadone dose and other variables to outcomes of methadone maintenance. Am J Addict. 1997;6(3):246-55.

559. Main F, Kelly L. Systematic literature review on buprenorphine/naloxone use in outpatient opioid dependence treatment. Canadian Journal of Addiction. 2016;7(1):12-8.

560. Mammen K, Bell J. The clinical efficacy and abuse potential of combination buprenorphine- naloxone in the treatment of opioid dependence. Expert Opin Pharmacother. 2009;10(15):2537-44.

561. Mannelli P, Wu LT, Peindl KS, Gorelick DA. Smoking and opioid detoxification: behavioral changes and response to treatment. Nicotine Tob Res. 2013;15(10):1705-13.

562. Mannelli P, Wu LT, Peindl KS, Swartz MS, Woody GE. Extended release naltrexone injection is performed in the majority of opioid dependent patients receiving outpatient induction: a very low dose naltrexone and buprenorphine open label trial. Drug Alcohol Depend. 2014;138:83-8.

563. Mannelli P, Swartz M, Wu LT. Withdrawal severity and early response to treatment in the outpatient transition from opioid use to extended release naltrexone. The American Journal on Addictions. 2018.

564. Manning V, Best D, Faulkner N, Titherington E, Morinan A, Keaney F, et al. Does active referral by a doctor or 12-Step peer improve 12-Step meeting attendance? Results from a pilot randomised control trial. Drug Alcohol Depend. 2012;126(1-2):131-7.

565. Manubay J, Davidson J, Vosburg S, Jones J, Comer S, Sullivan M. Sex differences among opioid-abusing patients with chronic pain in a clinical trial. J Addict Med. 2015;9(1):46-52.

566. Marcovitz D, McHugh K, Volpe J, Votaw V, Connery HS. Predictors of early dropout in outpatient buprenorphine/naloxone treatment. Am J Addict. 2016;25(6):472-7.

567. Maree RD, Marcum ZA, Saghafi E, Weiner DK, Karp JF. A Systematic Review of Opioid and Benzodiazepine Misuse in Older Adults. Am J Geriatr Psychiatry. 2016;24(11):949-63.

568. Mariani JJ, Malcolm RJ, Mamczur AK, Choi JC, Brady R, Nunes E, et al. Pilot trial of gabapentin for the treatment of benzodiazepine abuse or dependence in methadone maintenance patients. Am J Drug Alcohol Abuse. 2016;42(3):333-40. doi: 10.3109/00952990.2015.1125493. Epub 2016 Mar 10.

569. Marschall U, L'Hoest H, Radbruch L, Hauser W. Long-term opioid therapy for chronic noncancer pain in Germany. European Journal of Pain. 2016;20(5):767-76.

570. Martel MO, Dolman AJ, Edwards RR, Jamison RN, Wasan AD. The association between negative affect and prescription opioid misuse in patients with chronic pain: the mediating role of opioid craving. J Pain. 2014;15(1):90-100.

571. Martin P, Tamblyn R, Ahmed S, Benedetti A, Tannenbaum C. A consumer-targeted, pharmacist-led, educational intervention to reduce inappropriate medication use in community older

References adults (D-PRESCRIBE trial): study protocol for a cluster randomized controlled trial. Trials. 2015;16:266.

572. Martin P, Tannenbaum C. A realist evaluation of patients' decisions to deprescribe in the EMPOWER trial. BMJ Open. 2017;7(4):e015959.

573. Martin P, Tannenbaum C. Use of the EMPOWER brochure to deprescribe sedative-hypnotic drugs in older adults with mild cognitive impairment. BMC Geriatr. 2017;17(1):37.

574. Martin CE, Terplan M, O'Grady KE, Jones HE. Brief Report: Pregnancy intention and opioid use disorder treatment retention in the MOTHER study. Am J Addict. 2018;27(5):372-4.

575. Martinez V, Attal N, Vanzo B, Vicaut E, Gautier JM, Bouhassira D, et al. Adherence of French GPs to chronic neuropathic pain clinical guidelines: results of a cross-sectional, randomized, "e" case- vignette survey. PLoS ONE [Electronic Resource]. 2014;9(4):e93855.

576. Maslej MM, Bolker BM, Russell MJ, Eaton K, Durisko Z, Hollon SD, et al. The Mortality and Myocardial Effects of Antidepressants Are Moderated by Preexisting Cardiovascular Disease: A Meta- Analysis. Psychother Psychosom. 2017;86(5):268-82.

577. Mastropietro DJ, Omidian H. Current approaches in tamper-resistant and abuse-deterrent formulations. Drug Dev Ind Pharm. 2013;39(5):611-24.

578. Matthias MS, Donaldson MT, Jensen AC, Krebs EE. "I Was a Little Surprised": Qualitative Insights From Patients Enrolled in a 12-Month Trial Comparing Opioids With Nonopioid Medications for Chronic Musculoskeletal Pain. J Pain. 2018;19(9):1082-90.

579. Mattick R, Breen C, Kimber J, Davoli M. Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2014; (2). Available from: http://dx.doi.org/10.1002/14651858.CD002207.pub4.

580. Mattick R, Breen C, Kimber J, Davoli M. Methadone maintenance therapy versus no opioid replacement therapy for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2009; (3). Available from: http://dx.doi.org/10.1002/14651858.CD002209.pub2.

581. Mattila MJ, Vanakoski J, Kalska H, Seppälä T. Effects of alcohol, zolpidem, and some other sedatives and hypnotics on human performance and memory. Pharmacology, biochemistry, and behavior. 1998;59(4):917‐23.

582. Matusow H, Rosenblum A, Parrino M. Prescription drug monitoring program utilization among 15 US opioid treatment programs. J Subst Abuse Treat. 2018;85:17.

583. Mauger S, Fraser R, Gill K. Utilizing buprenorphine-naloxone to treat illicit and prescription- opioid dependence. Neuropsychiatr Dis Treat. 2014;10:587-98.

584. Maughan DL, Patel A, Parveen T, Braithwaite I, Cook J, Lillywhite R, et al. Primary-care-based social prescribing for mental health: An analysis of financial and environmental sustainability. Primary Health Care Research and Development. 2016;17(2):114-21.

585. Mayet S, Farrell MF, Ferri M, Amato L, Davoli M. Psychosocial treatment for opiate abuse and dependence. Cochrane Database of Systematic Reviews. 2014;2014 (4) (no pagination)(CD004330).

586. Mbaba M, Brown S-E, Wooditch A, Kiss M, Murphy A, Kumari S, et al. Prevalence, Diagnosis, and Treatment Rates of Mood Disorders among Opioid Users under Criminal Justice Supervision. Subst Use Misuse. 2018;53(9):1519-28.

References

587. McAdam-Marx C, Roland CL, Cleveland J, Oderda GM. Costs of opioid abuse and misuse determined from a Medicaid database. J Pain Palliat Care Pharmacother. 2010;24(1):5-18.

588. McAuley A, Best D, Taylor A, Hunter C, Robertson R. From evidence to policy: The Scottish national naloxone programme. Drugs: Education, Prevention and Policy. 2012;19(4):309-19.

589. McCabe SE, Boyd CJ, Teter CJ. Subtypes of nonmedical prescription drug misuse. Drug Alcohol Depend. 2009;102(1-3):63-70.

590. McCarthy DM, Wolf MS, McConnell R, Sears J, Chevrier A, Ahlstrom E, et al. Improving patient knowledge and safe use of opioids: a randomized controlled trial. Acad Emerg Med. 2015;22(3):331-9.

591. McDermott KA, Griffin ML, Connery HS, Hilario EY, Fiellin DA, Fitzmaurice GM, et al. Initial response as a predictor of 12-week buprenorphine-naloxone treatment response in a prescription opioid-dependent population. J Clin Psychiatry. 2015;76(2):189-94.

592. McDonald R, Strang J. Are take-home naloxone programmes effective? Systematic review utilizing application of the Bradford Hill criteria. Addiction. 2016;111(7):1177-87.

593. McHugh RK, Devito EE, Dodd D, Carroll KM, Potter JS, Greenfield SF, et al. Gender differences in a clinical trial for prescription opioid dependence. J Subst Abuse Treat. 2013;45(1):38-43.

594. McHugh RK, Fitzmaurice GM, Carroll KM, Griffin ML, Hill KP, Wasan AD, et al. Assessing craving and its relationship to subsequent prescription opioid use among treatment-seeking prescription opioid dependent patients. Drug Alcohol Depend. 2014;145:121-6.

595. McMahon CG, Porst H. Oral agents for the treatment of premature ejaculation: review of efficacy and safety in the context of the recent International Society for Sexual Medicine criteria for lifelong premature ejaculation. J Sex Med. 2011;8(10):2707-25.

596. McNicol E, Midbari A, Eisenberg E. Opioids for neuropathic pain. Cochrane Database of Systematic Reviews [Internet]. 2013; (8). Available from: http://dx.doi.org/10.1002/14651858.CD006146.pub2.

597. Meade CS, Lukas SE, McDonald LJ, Fitzmaurice GM, Eldridge JA, Merrill N, et al. A randomized trial of transcutaneous electric acupoint stimulation as adjunctive treatment for opioid detoxification. J Subst Abuse Treat. 2010;38(1):12-21.

598. Mechcatie EMABSN. Duration of Postoperative Opioid Use Predicts Misuse More Than Dosage. The American Journal of Nursing. 2018;118(4):15.

599. Merchant RC, Baird JR, Liu T. Short-term Efficacy of a Brief Intervention to Reduce Drug Misuse and Increase Drug Treatment Utilization Among Adult Emergency Department Patients. Acad Emerg Med. 2015;22(10):1172-80.

600. Merrill JO, Von Korff M, Banta-Green CJ, Sullivan MD, Saunders KW, Campbell CI, et al. Prescribed opioid difficulties, depression and opioid dose among chronic opioid therapy patients. Gen Hosp Psychiatry. 2012;34(6):581-7.

601. Michael SS, Babu KM, Androski C, Jr., Reznek MA. Effect of a Data-driven Intervention on Opioid Prescribing Intensity Among Emergency Department Providers: A Randomized Controlled Trial. Acad Emerg Med. 2018;25(5):482-93.

602. Michna E, Ross EL, Hynes WL, Nedeljkovic SS, Soumekh S, Janfaza D, et al. Predicting aberrant drug behavior in patients treated for chronic pain: importance of abuse history. J Pain Symptom Manage. 2004;28(3):250-8.

References

603. Middleton C, Harden J. Acquired pharmaco‐dynamic opioid tolerance: A concept analysis. J Adv Nurs. 2014;70(2):272-81.

604. Midmer D, Kahan M, Marlow B. Effects of a distance learning program on physicians' opioid- and benzodiazepine-prescribing skills. J Contin Educ Health Prof. 2006;26(4):294-301.

605. Minegishi T, Garrido MM, Pizer SD, Frakt AB. Effectiveness of policy and risk targeting for opioid-related risk mitigation: A randomised programme evaluation with stepped-wedge design. BMJ Open. 2018;8 (6) (no pagination)(e020097).

606. Minozzi S, Amato L, Bellisario C, Ferri M, Davoli M. Maintenance agonist treatments for opiate‐dependent pregnant women. Cochrane Database of Systematic Reviews [Internet]. 2013; (12). Available from: http://dx.doi.org/10.1002/14651858.CD006318.pub3.

607. Minozzi S, Amato L, Davoli M. Development of dependence following treatment with opioid analgesics for pain relief: a systematic review. Addiction. 2013;108(4):688-98.

608. Minozzi S, Amato L, Vecchi S, Davoli M, Kirchmayer U, Verster A. Oral naltrexone maintenance treatment for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2011; (4). Available from: http://dx.doi.org/10.1002/14651858.CD001333.pub4.

609. Miotto K, Hillhouse M, Donovick R, Cunningham-Rathner J, Charuvastra C, Torrington M, et al. Comparison of buprenorphine treatment for opioid dependence in 3 settings. J Addict Med. 2012;6(1):68-76.

610. Miranda AMD, Taca AMD. Neuromodulation with percutaneous electrical nerve field stimulation is associated with reduction in signs and symptoms of opioid withdrawal: a multisite, retrospective assessment. The American Journal of Drug and Alcohol Abuse. 2018;44(1):56-63.

611. Mitchell TB, White JM, Somogyi AA, Bochner F. Slow release oral morphine as a maintenance pharmacotherapy for opioid dependence. Drug Alcohol Depend. 2002;66 Suppl 1.

612. Mol AJ, Oude Voshaar RC, Gorgels WJ, Breteler MH, van Balkom AJ, van de Lisdonk EH, et al. The role of craving in relapse after discontinuation of long-term benzodiazepine use. J Clin Psychiatry. 2007;68(12):1894-900.

613. Mogri M, Nadler J, Khan T, Mador MJ. Complex sleep apnea in patients with obstructive sleep apnea on opioids for chronic pain. Sleep Biol Rhythms. 2014;12(2):127-34.

614. Mohlman MK, Tanzman B, Finison K, Pinette M, Jones C. Impact of medication-assisted treatment for opioid addiction on Medicaid expenditures and health services utilization rates in Vermont. J Subst Abuse Treat. 2016;67:9-14.

615. Moisset X, Trouvin AP, Tran VT, Authier N, Vergne-Salle P, Piano V, et al. Use of strong opioids in chronic non-cancer pain in adults. Evidence-based recommendations from the French Society for the Study and Treatment of Pain. Presse Med. 2016;45(4):447-62.

616. Mokri A, Chawarski MC, Taherinakhost H, Schottenfeld RS. Medical treatments for opioid use disorder in Iran: a randomized, double-blind placebo-controlled comparison of buprenorphine/naloxone and naltrexone maintenance treatment. Addiction. 2016;111(5):874-82.

617. Mol AJ, Oude Voshaar RC, Gorgels WJ, Breteler MH, van Balkom AJ, van de Lisdonk EH, et al. The absence of benzodiazepine craving in a general practice benzodiazepine discontinuation trial. Addict Behav. 2006;31(2):211-22. doi: 10.1016/j.addbeh.2005.04.024. Epub Jun 13.

References

618. Molfenter T, Knudsen HK, Brown R, Jacobson N, Horst J, Van Etten M, et al. Test of a workforce development intervention to expand opioid use disorder treatment pharmacotherapy prescribers: protocol for a cluster randomized trial. Implement Sci. 2017;12(1):135.

619. Moller HJ, Ulm K, Gloggler A. Kavain as an aid in the withdrawal of benzodiazepines (Therapy study). Munchener medizinische wochenschrift (1950). 1992;134(37):41‐4.

620. Momper SL, Delva J, Reed BG. OxyContin Misuse on a Reservation: Qualitative Reports by American Indians in Talking Circles. Subst Use Misuse. 2011;46(11):1372-9.

621. Montgomery SA, Fava M, Padmanabhan SK, Guico-Pabia CJ, Tourian KA. Discontinuation symptoms and taper/poststudy-emergent adverse events with desvenlafaxine treatment for major depressive disorder. Int Clin Psychopharmacol. 2009;24(6):296-305.

622. Moore BA, Barry DT, Sullivan LE, O'Connor P G, Cutter CJ, Schottenfeld RS, et al. Counseling and directly observed medication for primary care buprenorphine maintenance: a pilot study. J Addict Med. 2012;6(3):205-11.

623. Moore SK, Guarino H, Marsch LA. "This is not who I want to be:" experiences of opioid- dependent youth before, and during, combined buprenorphine and behavioral treatment. Subst Use Misuse. 2014;49(3):303-14.

624. Moore BA, Fiellin DA, Cutter CJ, Buono FD, Barry DT, Fiellin LE, et al. Cognitive Behavioral Therapy Improves Treatment Outcomes for Prescription Opioid Users in Primary Care Buprenorphine Treatment. J Subst Abuse Treat. 2016;71:54-7.

625. Morasco BJ, Cavanagh R, Gritzner S, Dobscha SK. Care management practices for chronic pain in veterans prescribed high doses of opioid medications. Fam Pract. 2013;30(6):671-8.

626. Morgan JR, Schackman BR, Leff JA, Linas BP, Walley AY. Injectable naltrexone, oral naltrexone, and buprenorphine utilization and discontinuation among individuals treated for opioid use disorder in a United States commercially insured population. J Subst Abuse Treat. 2018;85:90.

627. Morley KI, Ferris JA, Winstock AR, Lynskey MT. Polysubstance use and misuse or abuse of prescription opioid analgesics: a multi-level analysis of international data. Pain. 2017;158(6):1138-44.

628. Morris BJ, Zumsteg JW, Archer KR, Cash B, Mir HR. Narcotic use and postoperative doctor shopping in the orthopaedic trauma population. Journal of Bone and Joint Surgery - American Volume. 2014;96(15):1257-62.

629. Morton S, Lader M. Buspirone treatment as an aid to benzodiazepine withdrawal. J Psychopharmacol. 1995;9(4):331-5. doi: 10.1177/026988119500900407.

630. Moselhy HFMMMDMRCP, Fahmy EMMMD, Mikhael VSMMMD, El-Sheikh HMMMD. Impact of Self-Reported Life Events on the Profile of Opioid Dependence Syndrome. The American Journal of Drug and Alcohol Abuse. 2010;36(1):7.

631. Moselhy HF, Fahmy E, Mikhael VS, El-Sheikh H. Emotional control inpatients with opioid dependence syndrome and reported history of negative life events. Addictive Disorders & Their Treatment. 2012;11(2):93-100.

632. Moy I, Crome P, Crome I, Fisher M. Systematic and narrative review of treatment for older people with substance problems. Eur Geriatr Med. 2011;2(4):212-36.

633. Mugunthan. Erratum: Minimal interventions to decrease long-term use of benzodiazepines in primary care: A systematic review and meta-analysis (British Journal of General Practice (2011) DOI: 10.3399/bjgp11X593857). Br J Gen Pract. 2013;63(606):12.

References

634. Munshi O. The role of Embeda (morphine sulfate-naltrexone hydrochloride) in opioid abuse: A systematic review of literature. J Pain. 2013:S11.

635. Murphy Y, Wilson E, Goldner EM, Fischer B. Benzodiazepine Use, Misuse, and Harm at the Population Level in Canada: A Comprehensive Narrative Review of Data and Developments Since 1995. Clin Drug Investig. 2016;36(7):519-30.

636. Murphy SM, Howell D, McPherson S, Grohs R, Roll J, Neven D. A Randomized Controlled Trial of a Citywide Emergency Department Care-Coordination Program to Reduce Prescription Opioid- Related Visits: An Economic Evaluation. J Emerg Med. 2017;53(2):186-94.

637. Mutlu E, Bilici R. Outcomes of buprenorphine-naloxone maintenance therapy: one-year follow up study from Turkey. Heroin addiction and related clinical problems. 2015;17(1):79‐90.

638. Mysels DJMDMBA, Cheng WYMPH, Nunes EVMD, Sullivan MAMDP. The association between naltrexone treatment and symptoms of depression in opioid-dependent patients. The American Journal of Drug and Alcohol Abuse. 2011;37(1):22.

639. Nadpara PA, Joyce AR, Murrelle EL, Carroll NW, Carroll NV, Barnard M, et al. Risk factors for serious prescription opioid-induced respiratory depression or overdose: Comparison of commercially insured and veterans health affairs populations. Pain Med. 2018;19(1):79-96.

640. Nalamachu SR, Narayana A, Janka L. Long-term dosing, safety, and tolerability of fentanyl buccal tablet in the management of noncancer-related breakthrough pain in opioid-tolerant patients. Curr Med Res Opin. 2011;27(4):751-60.

641. Naliboff BD, Wu SM, Schieffer B, Bolus R, Pham Q, Baria A, et al. A randomized trial of 2 prescription strategies for opioid treatment of chronic nonmalignant pain. J Pain. 2011;12(2):288-96.

642. Nardi AE, Freire RC, Valenca AM, Amrein R, de Cerqueira AC, Lopes FL, et al. Tapering clonazepam in patients with panic disorder after at least 3 years of treatment. J Clin Psychopharmacol. 2010;30(3):290-3. doi: 10.1097/JCP.0b013e3181dcb2f3.

643. Narayan V, Haddad PM. Antidepressant discontinuation manic states: a critical review of the literature and suggested diagnostic criteria. Journal of psychopharmacology (Oxford, England). 2011;25(3):306-13.

644. Nathan RG, Robinson D, Cherek DR, Sebastian CS, Hack M, Davison S. Alternative treatments for withdrawing the long-term benzodiazepine user: a pilot study. Int J Addict. 1986;21(2):195-211.

645. Addiction to medicine: an investigation into the configuration and commissioning of treatment services to support those who develop problems with prescription-only or over-the- counter medicine. London: NTA, 2011 2011. Report No.

646. Naumann RB. Assessing sustained and differential impacts of North Carolina's Medicaid "Lock-in" program [Ph.D.]: The University of North Carolina at Chapel Hill; 2018.

647. Nct. Neurobiology of Opioid Dependence: 3 - 3. Https://clinicaltrialsgov/show/nct00000194. 1999.

648. Nct. Neurobiology of Opioid Dependence: 1 - 1. Https://clinicaltrialsgov/show/nct00000192. 1999.

649. Nct. Neurobiology of Opioid Dependence: 2 - 2. Https://clinicaltrialsgov/show/nct00000193. 1999.

References

650. Nct. Buprenorphine/Naloxone for Treatment of Opiate Dependence - 9. Https://clinicaltrialsgov/show/nct00000344. 1999.

651. Nct. Buprenorphine Dose Alteration Study - 1. Https://clinicaltrialsgov/show/nct00000219. 1999.

652. Nct. Gradual Vs. Rapid Buprenorphine Detoxification - 2. Https://clinicaltrialsgov/show/nct00000220. 1999.

653. Nct. Buprenorphine Detox With Two Types of Treatment. BBD II - 11. Https://clinicaltrialsgov/show/nct00000229. 1999.

654. Nct. Buprenorphine/Naloxone Versus Clonidine for Inpatient Opiate Detoxification - 1. Https://clinicaltrialsgov/show/nct00032955. 2002.

655. Nct. Dextromethorphan, Gabapentin, and Oxycodone to Treat Opioid-Induced Hyperalgesia. Https://clinicaltrialsgov/show/nct00218374. 2005.

656. Nct. Opiate Dependence: combined Naltrexone/Behavior Therapy - 1. Https://clinicaltrialsgov/show/nct00238914. 2005.

657. Nct. Buprenorphine as a Treatment for Individuals Dependent on Analgesic Opioids. Https://clinicaltrialsgov/show/nct00218101. 2005.

658. Nct. Tramadol to Reduce Opioid Withdrawal Symptoms. Https://clinicaltrialsgov/show/nct00142896. 2005.

659. Nct. Blockade Efficacy of Buprenorphine/Naloxone For Opioid Dependence. Https://clinicaltrialsgov/show/nct00134888. 2005.

660. Nct. Long Term Treatment With Zolpidem: nightly and Intermittent Dosing. Https://clinicaltrialsgov/show/nct00156533. 2005.

661. Nct. Medication Adherence Therapy for Opioid Abusing Pain Patients. Https://clinicaltrialsgov/show/nct00249587. 2005.

662. Nct. Opiate-Induced Tolerance & Hyperalgesia in Pain Patients. Https://clinicaltrialsgov/show/nct00246532. 2005.

663. Nct. A Study of the Effect on Pain Control of Treatment With Fentanyl, Administered Through the Skin, Compared With Placebo in Patients With Osteoarthritis. Https://clinicaltrialsgov/show/nct00236366. 2005.

664. Nct. Applying Web Technology to Buprenorphine Treatment. Https://clinicaltrialsgov/show/nct00182598. 2005.

665. Nct. Assessment of Tramadol as a Treatment for Opioid Addiction. Https://clinicaltrialsgov/show/nct00301210. 2006.

666. Nct. Evolution of Analgesic Tolerance With Opioids. Https://clinicaltrialsgov/show/nct00275249. 2006.

667. Nct. Opioid Rotation From Morphine to Methadone in Treatment of Non-malignant Pain. Https://clinicaltrialsgov/show/nct00356083. 2006.

668. Nct. Efficacy and Safety of Pregabalin vs Placebo for Generalized Anxiety Disorder (GAD) Symptoms in Subjects Discontinuing Benzodiazepine Treatment and Remaining 6 Weeks on Study Medication, Free From Benzodiazepine Use. Https://clinicaltrialsgov/show/nct00368745. 2006.

References

669. Nct. The Effect Of Dose Titration And Dose Tapering On The Tolerability Of DVS SR In Women With Vasomotor Symptoms. Https://clinicaltrialsgov/show/nct00401245. 2006.

670. Nct. The Discriminative Effects of Tramadol in Humans. Https://clinicaltrialsgov/show/nct00499746. 2007.

671. Nct. The Effect of Valproate on Benzodiazepine Withdrawal Severity. Https://clinicaltrialsgov/show/nct00570219. 2007.

672. Nct. Buprenorphine as a Treatment in Opiate Dependent Pain Patients. Https://clinicaltrialsgov/show/nct00552578. 2007.

673. Nct. Efficacy and Safety of Zolpidem-MR Administered Over a Long Term Period as Needed for Insomnia (ZOLONG). Https://clinicaltrialsgov/show/nct00425243. 2007.

674. Nct. Buprenorphine Maintenance Treatment of Opioid Dependence in Primary Care: a Randomized Clinical Trial. Https://clinicaltrialsgov/show/nct00684554. 2008.

675. Nct. Testing a Primary Care Model for the Treatment of Opioid Dependence Using Long- acting Injectable Naltrexone (Vivitrol). Https://clinicaltrialsgov/show/nct00675532. 2008.

676. Nct. Clinical Trial of Integrated Treatment for Pain and Opioid Dependence. Https://clinicaltrialsgov/show/nct00634803. 2008.

677. Nct. Prospective Cohort of Opiate Dependent Patients on Buprenorphine/Naloxone for Maintenance. Https://clinicaltrialsgov/show/nct00622596. 2008.

678. Nct. Randomized, Double-Blind Cross-Over Trial Of Buprenorphine and Low-Dose Naloxone Versus Buprenorphine. Https://clinicaltrialsgov/show/nct00679458. 2008.

679. Nct. A Study of Duloxetine in Patients With Osteoarthritis Knee Pain. Https://clinicaltrialsgov/show/nct00945945. 2009.

680. Nct. Duloxetine in Patients With Diabetic in Peripheral Neuropathic Pain With or Without Co- morbid Major Depressive Disorder. Https://clinicaltrialsgov/show/nct00844194. 2009.

681. Nct. Efficacy of Extended Release Tramadol for Treating Prescription Opioid Withdrawal. Https://clinicaltrialsgov/show/nct00980044. 2009.

682. Nct. Multicentre Trial of Suboxone in Opiate-dependent Subjects in Taiwan. Https://clinicaltrialsgov/show/nct00901875. 2009.

683. Nct. A Study to Evaluate the Efficacy, Safety, and Tolerability of Tapentadol ER Compared With Placebo in Patients With Chronic, Painful Diabetic Peripheral Neuropathy. Https://clinicaltrialsgov/show/nct01041859. 2009.

684. Nct. Prospective Study About Clinical and Pharmacogenetic Safety of Opioid Use for Chronic Pain. Https://clinicaltrialsgov/show/nct00916890. 2009.

685. Nct. Efficacy and Safety Study of Vortioxetine (Lu AA21004) for Treatment of Major Depressive Disorder. Https://clinicaltrialsgov/show/nct01255787. 2010.

686. Nct. A Long-term Safety Study With Tapentadol ER and Oxycodone CR in Patients With Moderate to Severe Pain Due to Chronic, Painful Diabetic Peripheral Neuropathy (DPN). Https://clinicaltrialsgov/show/nct01063868. 2010.

687. Nct. An Efficacy and Safety Study for Tapentadol Extended Release (JNS024ER) in Chronic Pain Participants. Https://clinicaltrialsgov/show/nct01124604. 2010.

References

688. Nct. Study to Evaluate the Safety and Efficacy of Naldemedine (S-297995) for the Treatment of Opioid-Induced Bowel Dysfunction in Subjects With Chronic Pain. Https://clinicaltrialsgov/show/nct01122030. 2010.

689. Nct. A Phase 2 Study of Tapentadol Extended-Release (JNS024ER) ) in Japanese Participants With Chronic Pain Due to Diabetic Neuropathic Pain or Postherpetic Neuralgia. Https://clinicaltrialsgov/show/nct01124617. 2010.

690. Nct. Efficacy and Safety of Oxycodone/Naloxone Controlled-release Tablets (OXN) Compared to Placebo in Opioid-experienced Subjects With Moderate to Severe Chronic Low Back Pain. Https://clinicaltrialsgov/show/nct01358526. 2011.

691. Nct. Comparison of Increasing Doses of Tapentadol Versus a Combination of Tapentadol and Pregabalin. Https://clinicaltrialsgov/show/nct01352741. 2011.

692. Nct. Efficacy and Safety Study of Pregabalin in the Treatment of Pain on Walking in Patients With Diabetic Peripheral Neuropathy (DPN). Https://clinicaltrialsgov/show/nct01474772. 2011.

693. Nct. Buprenorphine and Methadone for Opioid-dependent Chronic Back Pain Patients. Https://clinicaltrialsgov/show/nct01559454. 2012.

694. Nct. 5HT3 Antagonists to Treat Opioid Withdrawal and to Prevent the Progression of Physical Dependence. Https://clinicaltrialsgov/show/nct01549652. 2012.

695. Nct. Efficacy Study to Evaluate Buprenorphine HCl Buccal Film in Opioid-Experienced Subjects. Https://clinicaltrialsgov/show/nct01675167. 2012.

696. Nct. Safety Study of CB-5945 for the Treatment of Opioid-Induced Constipation. Https://clinicaltrialsgov/show/nct01696643. 2012.

697. Nct. Efficacy and Safety of Hydrocodone Bitartrate Extended-Release Tablets for Moderate to Severe Chronic Low Back Pain. Https://clinicaltrialsgov/show/nct01789970. 2013.

698. Nct. Tapentadol Prolonged Release (PR) Versus Oxycodone/Naloxone Prolonged Release in Severe Chronic Low Back Pain With a Neuropathic Component. Https://clinicaltrialsgov/show/nct01838616. 2013.

699. Nct. Improving Buprenorphine Detoxification Outcomes With Isradipine. Https://clinicaltrialsgov/show/nct01895270. 2013.

700. Nct. Gabapentin Treatment of Benzodiazepine Dependence. Https://clinicaltrialsgov/show/nct01893632. 2013.

701. Nct. Lidocaine Infusion for Chronic Pain in Opioid Dependent Patients. Https://clinicaltrialsgov/show/nct02219321. 2014.

702. Nct. Effectiveness of Mindfulness Based Relapse Prevention for Chronic Users of Benzodiazepines. Https://clinicaltrialsgov/show/nct02127411. 2014.

703. Nct. A Pragmatic Randomized Control Trial Comparing Models of Care in the Management of Prescription Opioid Misuse. Https://clinicaltrialsgov/show/nct03033732. 2017.

704. Nelson L, Schwaner R. Transdermal fentanyl: Pharmacology and toxicology. J Med Toxicol. 2009;5(4):230-41.

References

705. Neumann AM, Blondell RD, Jaanimagi U, Giambrone AK, Homish GG, Lozano JR, et al. A preliminary study comparing methadone and buprenorphine in patients with chronic pain and coexistent opioid addiction. J Addict Dis. 2013;32(1):68-78. doi: 10.1080/10550887.2012.759872.

706. Neven D, Paulozzi L, Howell D, McPherson S, Murphy SM, Grohs B, et al. A Randomized Controlled Trial of a Citywide Emergency Department Care Coordination Program to Reduce Prescription Opioid Related Emergency Department Visits. J Emerg Med. 2016;51(5):498-507.

707. Nielsen S, Hillhouse M, Mooney L, Fahey J, Ling W. Comparing buprenorphine induction experience with heroin and prescription opioid users. J Subst Abuse Treat. 2012;43(3):285-90.

708. Nielsen M, Hansen EH, Gotzsche PC. What is the difference between dependence and withdrawal reactions? A comparison of benzodiazepines and selective serotonin re-uptake inhibitors. Addiction. 2012;107(5):900-8.

709. Nielsen S, Hillhouse M, Thomas C, Hasson A, Ling W. A comparison of buprenorphine taper outcomes between prescription opioid and heroin users. J Addict Med. 2013;7(1):33-8.

710. Nielsen S, Hillhouse M, Weiss RD, Mooney L, Sharpe Potter J, Lee J, et al. The relationship between primary prescription opioid and buprenorphine-naloxone induction outcomes in a prescription opioid dependent sample. Am J Addict. 2014;23(4):343-8.

711. Nielsen S, Hillhouse M, Mooney L, Ang A, Ling W. Buprenorphine pharmacotherapy and behavioral treatment: comparison of outcomes among prescription opioid users, heroin users and combination users. J Subst Abuse Treat. 2015;48(1):70-6.

712. Nielsen S, Larance B, Degenhardt L, Gowing L, Kehler C, Lintzeris N. Opioid agonist treatment for pharmaceutical opioid dependent people. Cochrane Database of Systematic Reviews [Internet]. 2016; (5). Available from: http://dx.doi.org/10.1002/14651858.CD011117.pub2.

713. Nielsen S, Bruno R, Murnion B, Dunlop A, Degenhardt L, Demirkol A, et al. Treating codeine dependence with buprenorphine: Dose requirements and induction outcomes from a retrospective case series in New South Wales, Australia. Drug and alcohol review. 2016;35(1):70-5.

714. Nielsen RV, Fomsgaard JS, Nikolajsen L, Dahl JB, Mathiesen O. Intraoperative S-ketamine for the reduction of opioid consumption and pain one year after spine surgery: A randomized clinical trial of opioid dependent patients. European Journal of Pain. 2018;23:23.

715. Nielsen S, MacDonald T, Johnson JL. Identifying and treating codeine dependence: a systematic review. The Medical journal of Australia. 2018;208(10):451-61.

716. Nihr H. CAM2038 for opioid dependence. Birmingham: NIHR Horizon Scanning Research&Intelligence Centre, 2016.

717. Nikulina V, Guarino H, Acosta MC, Marsch LA, Syckes C, Moore SK, et al. Patient vs provider reports of aberrant medication-taking behavior among opioid-treated patients with chronic pain who report misusing opioid medication. Pain. 2016;157(8):1791-8.

718. Nilsen HK, Stiles TC, Landro NI, Fors EA, Kaasa S, Borchgrevink PC. Patients with problematic opioid use can be weaned from codeine without pain escalation. Acta Anaesthesiol Scand. 2010;54(5):571-9.

719. Noble M, Tregear SJ, Treadwell JR, Schoelles K. Long-term opioid therapy for chronic noncancer pain: a systematic review and meta-analysis of efficacy and safety. J Pain Symptom Manage. 2008;35(2):214-28.

References

720. Noble M, Treadwell J, Tregear S, Coates V, Wiffen P, Akafomo C, et al. Long‐term opioid management for chronic noncancer pain. Cochrane Database of Systematic Reviews [Internet]. 2010; (1). Available from: http://dx.doi.org/10.1002/14651858.CD006605.pub2.

721. Noller GEP, Frampton CMP, Yazar-Klosinski BP. Ibogaine treatment outcomes for opioid dependence from a twelve-month follow-up observational study. The American Journal of Drug and Alcohol Abuse. 2018;44(1):37-46.

722. Nordfjærn T. Prospective associations between benzodiazepine use and later life satisfaction, somatic pain and psychological health among the elderly. Human Psychopharmacology: Clinical and Experimental. 2013;28(3):248-57.

723. Norman IJ, Bergin M, Parry CD, Van Hout MC. Best practices and innovations for managing codeine misuse and dependence. Journal of Pharmacy and Pharmaceutical Sciences. 2016;19(3):367- 81.

724. Northrup TF, Stotts AL, Green C, Potter JS, Marino EN, Walker R, et al. Opioid withdrawal, craving, and use during and after outpatient buprenorphine stabilization and taper: a discrete survival and growth mixture model. Addict Behav. 2015;41:20-8.

725. Nosyk B, Guh DP, Sun H, Oviedo-Joekes E, Brissette S, Marsh DC, et al. Health related quality of life trajectories of patients in opioid substitution treatment. Drug Alcohol Depend. 2011;118(2- 3):259-64.

726. Nosyk B, Bray JW, Wittenberg E, Aden B, Eggman AA, Weiss RD, et al. Short term health- related quality of life improvement during opioid agonist treatment. Drug Alcohol Depend. 2015;157:121-8.

727. Novak SP, Herman-Stahl M, Flannery B, Zimmerman M. Physical pain, common psychiatric and substance use disorders, and the non-medical use of prescription analgesics in the United States. Drug Alcohol Depend. 2009;100(1-2):63-70.

728. Nuesch E, Rutjes AWS, Husni E, Welch V, Juni P. Oral or transdermal opioids for osteoarthritis of the knee or hip. Sao Paulo Med J. 2009;127(6):388.

729. Nunes EV, Gordon M, Friedmann PD, Fishman MJ, Lee JD, Chen DT, et al. Relapse to opioid use disorder after inpatient treatment: Protective effect of injection naltrexone. J Subst Abuse Treat. 2018;85:49-55.

730. Nurminen J, Puustinen J, Lahteenmaki R, Vahlberg T, Lyles A, Partinen M, et al. Handgrip strength and balance in older adults following withdrawal from long-term use of temazepam, zopiclone or zolpidem as hypnotics. BMC Geriatr. 2014;14:121.

731. O'Connor K, Marchand A, Brousseau L, Aardema F, Mainguy N, Landry P, et al. Cognitive- behavioural, pharmacological and psychosocial predictors of outcome during tapered discontinuation of benzodiazepine. Clin Psychol Psychother. 2008;15(1):1-14.

732. O'Connor K, Belanger L, Aardema F, Perodeau G, Harel F. The predictive utility of health preoccupations and outcome expectancies prior to benzodiazepine withdrawal. Journal of Substance Use. 2012;17(2):122-33.

733. Ogawa S, Arakawa A, Yoshiyama T, Suzuki M. Long-term efficacy and safety of pregabalin in patients with postherpetic neuralgia: results of a 52-week, open-label, flexible-dose study. Japanese journal of anesthesiology. 2010;59(8):961‐70.

References

734. Ogle NR, Akkerman SR. Guidance for the discontinuation or switching of antidepressant therapies in adults. J Pharm Pract. 2013;26(4):389-96.

735. Oliva EM, Bowe T, Tavakoli S, Martins S, Lewis ET, Paik M, et al. Development and applications of the Veterans Health Administration's Stratification Tool for Opioid Risk Mitigation (STORM) to improve opioid safety and prevent overdose and suicide. Psychol Serv. 2017;14(1):34-49.

736. O'Mullan C, Doherty M, Coates R, Tilley PJM. ‘Accepting what is’: An approach for managing the long-term sexual side effects of selective serotonin reuptake inhibitors (SSRIs) in women. Sex Relation Ther. 2015;30(3):325-37.

737. Onyett Stephen R, Turpin G. Benzodiazepine withdrawal in primary care: a comparison of behavioural group training and individual sessions. Behavioural psychotherapy. 1988;16(4):297‐312.

738. Osmun WE, Copeland J, Parr J, Boisvert L. Characteristics of chronic pain patients in a rural teaching practice. Can Fam Physician. 2011;57(11):e436-e40.

739. Ostini R, Jackson C, Hegney D, Tett SE. How is medication prescribing ceased? A systematic review. Med Care. 2011;49(1):24-36.

740. Otto MW, Pollack MH, Sachs GS, Reiter SR, Meltzer-Brody S, Rosenbaum JF. Discontinuation of benzodiazepine treatment: efficacy of cognitive-behavioral therapy for patients with panic disorder. Am J Psychiatry. 1993;150(10):1485-90.

741. Otto MW, McHugh RK, Simon NM, Farach FJ, Worthington JJ, Pollack MH. Efficacy of CBT for benzodiazepine discontinuation in patients with panic disorder: Further evaluation. Behav Res Ther. 2010;48(8):720-7.

742. Oulis P, Masdrakis VG, Karapoulios E, Karakatsanis NA, Kouzoupis AV, Papadimitriou GN. For publication: Tiagabine in the discontinuation of long-term benzodiazepine use. Psychiatry Clin Neurosci. 2009;63(1):122.

743. Oulis P, Konstantakopoulos G. Pregabalin in the treatment of alcohol and benzodiazepines dependence. CNS Neurosci Ther. 2010;16(1):45-50.

744. Pade PA, Cardon KE, Hoffman RM, Geppert CMA. Prescription opioid abuse, chronic pain, and primary care: A Co-Occurring Disorders Clinic in the chronic disease model. J Subst Abuse Treat. 2012;43(4):446-50.

745. Page J, Traver R, Patel S, Saliba C. Implementation of a Proactive Pilot Health Plan-Driven Opioid Tapering Program to Decrease Chronic Opioid Use for Conditions of the Back and Spine in a Medicaid Population. J Manag Care Spec Pharm. 2018;24(3):191-6.

746. Page MG, Kudrina I, Zomahoun HTV, Ziegler D, Beaulieu P, Charbonneau C, et al. Relative frequency and risk factors for long-term opioid therapy following surgery and trauma among adults: A systematic review protocol. Systematic Reviews. 2018;7 (1) (no pagination)(97).

747. Pani P, Vacca R, Trogu E, Amato L, Davoli M. Pharmacological treatment for depression during opioid agonist treatment for opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2010; (9). Available from: http://dx.doi.org/10.1002/14651858.CD008373.pub2.

748. Pani P, Trogu E, Maremmani I, Pacini M. QTc interval screening for cardiac risk in methadone treatment of opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2013; (6). Available from: http://dx.doi.org/10.1002/14651858.CD008939.pub2.

References

749. Paolucci M, Altamura C, Brunelli N, Rizzo AC, Assenza F, Pasqualetti P, et al. Methylprednisolone plus diazepam i.v. as bridge therapy for medication overuse headache. Neurol Sci. 2017;38(11):2025-9.

750. Papaleontiou M, Henderson CR, Jr., Turner BJ, Moore AA, Olkhovskaya Y, Amanfo L, et al. Outcomes associated with opioid use in the treatment of chronic noncancer pain in older adults: a systematic review and meta-analysis. J Am Geriatr Soc. 2010;58(7):1353-69.

751. Papazisis G, Tzellos T, Toulis K, Goulis D, Sardeli C, Chourdakis M, et al. Safety/tolerability of alpha2delta ligand pregabalin in the treatment of fibromyalgia: An evidence based evaluation. Review of Clinical Pharmacology and Pharmacokinetics, International Edition. 2010;24(2):181-2.

752. Paquin AM, Zimmerman K, Rudolph JL. Risk versus risk: a review of benzodiazepine reduction in older adults. Expert opinion on drug safety. 2014;13(7):1-16.

753. Pardo B. Do more robust prescription drug monitoring programs reduce prescription opioid overdose? Addiction. 2017;112(10):1773-83.

754. Park J, Lavin R. Risk factors associated with opioid medication misuse in community-dwelling older adults with chronic pain. Clin J Pain. 2010;26(8):647-55.

755. Parker MA, Streck JM, Sigmon SC. Associations between opioid and nicotine dependence in nationally representative samples of United States adult daily smokers. Drug Alcohol Depend. 2018;186:167.

756. Parr JM, Kavanagh DJ, Cahill L, Mitchell G, Mc DYR. Effectiveness of current treatment approaches for benzodiazepine discontinuation: a meta-analysis. Addiction. 2009;104(1):13-24.

757. Parr JM, Kavanagh DJ, Young RM, Mitchell G. Acceptability of cognitive-behaviour therapy via the Internet for cessation of benzodiazepine use. Drug and alcohol review. 2011;30(3):306-14.

758. Parran TV, Adelman CA, Merkin B, Pagano ME, Defranco R, Ionescu RA, et al. Long-term outcomes of office-based buprenorphine/naloxone maintenance therapy. Drug Alcohol Depend. 2010;106(1):56-60.

759. Passik SD, Hays L, Eisner N, Kirsh KL. Psychiatric and pain characteristics of prescription drug abusers entering drug rehabilitation. J Pain Palliat Care Pharmacother. 2006;20(2):5-13.

760. Passik SD, Kirsh KL, Casper D. Addiction-related assessment tools and pain management: Instruments for screening, treatment planning, and monitoring compliance. Pain Med. 2008;9(SUPPL. 2):S145-S66.

761. Passik SD. Issues in long-term opioid therapy: Unmet needs, risks, and solutions. Mayo Clin Proc. 2009;84(7):593-601.

762. Passik SD, Messina J, Golsorkhi A, Xie F. Aberrant drug-related behavior observed during clinical studies involving patients taking chronic opioid therapy for persistent pain and fentanyl buccal tablet for breakthrough pain. J Pain Symptom Manage. 2011;41(1):116-25.

763. Passik SD, Narayana A, Yang R. Aberrant drug-related behavior observed during a 12-week open-label extension period of a study involving patients taking chronic opioid therapy for persistent pain and fentanyl buccal tablet or traditional short-acting opioid for breakthrough pain. Pain Med. 2014;15(8):1365-72.

764. Pat-Horenczyk R, Hacohen D, Herer P, Lavie P. The effects of substituting zopiclone in withdrawal from chronic use of benzodiazepine hypnotics. Psychopharmacology (Berl). 1998;140(4):450-7.

References

765. Patrick ME, Dunn KE, Badger GJ, Heil SH, Higgins ST, Sigmon SC. Spontaneous reductions in smoking during double-blind buprenorphine detoxification. Addict Behav. 2014;39(9):1353-6.

766. Paulozzi LJ, Budnitz DS, Xi Y. Increasing deaths from opioid analgesics in the United States. Pharmacoepidemiol Drug Saf. 2006;15(9):618-27.

767. Paulozzi LJ, Kilbourne EM, Shah NG, Nolte KB, Desai HA, Landen MG, et al. A history of being prescribed controlled substances and risk of drug overdose death. Pain Med. 2012;13(1):87-95.

768. Pauly NJ, Slavova S, Delcher C, Freeman PR, Talbert J. Features of prescription drug monitoring programs associated with reduced rates of prescription opioid-related poisonings. Drug Alcohol Depend. 2018;184:26.

769. Peckham AM, Fairman KA, Sclar DA. All-Cause and Drug-Related Medical Events Associated with Overuse of Gabapentin and/or Opioid Medications: A Retrospective Cohort Analysis of a Commercially Insured US Population. Drug Saf. 2018;41(2):213-28.

770. Peirce GL, Smith MJ, Abate MA, Halverson J. Doctor and pharmacy shopping for controlled substances. Med Care. 2012;50(6):494-500.

771. Peles E, Hetzroni T, Bar-Hamburger R, Adelson M, Schreiber S. Melatonin for perceived sleep disturbances associated with benzodiazepine withdrawal among patients in methadone maintenance treatment: a double-blind randomized clinical trial. Addiction. 2007;102(12):1947-53. doi: 10.111/j.360-0443.2007.02007.x. Epub 2007 Oct 4.

772. Penninga EI, Graudal N, Ladekarl MB, Jurgens G. Adverse Events Associated with Flumazenil Treatment for the Management of Suspected Benzodiazepine Intoxication--A Systematic Review with Meta-Analyses of Randomised Trials. Basic Clin Pharmacol Toxicol. 2016;118(1):37-44.

773. Perahia DG, Pritchett YL, Kajdasz DK, Bauer M, Jain R, Russell JM, et al. A randomized, double-blind comparison of duloxetine and venlafaxine in the treatment of patients with major depressive disorder. J Psychiatr Res. 2008;42(1):22-34.

774. Pergolizzi JV, Raffa RB, Taylor R, Vacalis S. Abuse-deterrent opioids: an update on current approaches and considerations. Curr Med Res Opin. 2018;34(4):711-23.

775. Pergolizzi JV, Jr., Gharibo C, Passik S, Labhsetwar S, Taylor R, Jr., Pergolizzi JS, et al. Dynamic risk factors in the misuse of opioid analgesics. J Psychosom Res. 2012;72(6):443-51.

776. Pietrasanta L. Therapeutic management and comorbidities in opiate-dependent patients undergoing a replacement therapy programme in Spain: the PROTEUS study. Heroin addiction and related clinical problems. 2011;13(3):5-16.

777. Pikovsky M, Peacock A, Larney S, Larance B, Conroy E, Nelson E, et al. Alcohol use disorder and associated physical health complications and treatment amongst individuals with and without opioid dependence: A case-control study. Drug Alcohol Depend. 2018;188:304.

778. Pilkonis PA, Yu L, Dodds NE, Johnston KL, Lawrence SM, Hilton TF, et al. An item bank for abuse of prescription pain medication from the Patient- Reported Outcomes Measurement Information System (PROMIS®). Pain Med. 2017;18(8):1516-27.

779. Pilowsky DJ, Wu L-T, Burchett B, Blazer DG, Ling W. Depressive symptoms, substance use, and HIV-related high-risk behaviors among opioid-dependent individuals: Results from the Clinical Trials Network. Subst Use Misuse. 2011;46(14):1716-25.

References

780. Pimlott NJG, Hux JE, Wilson LM, Kahan M, Li C, Rosser WW. Educating physicians to reduce benzodiazepine use by elderly patients: a randomized controlled trial. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2003;168(7):835-9.

781. Pink LR, Smith AJ, Peng PW, Galonski MJ, Tumber PS, Evans D, et al. Intake assessment of problematic use of medications in a chronic noncancer pain clinic. Pain research & management. 2012;17(4):276-80.

782. Pohjanoksa-Mantyla M, Saari JK, Narhi U, Karjalainen A, Pylkkanen K, Airaksinen MS, et al. How and why do people with depression access and utilize online drug information: A qualitative study. J Affect Disord. 2009;114(1-3):333-9.

783. Pollack M, Kinrys G, Krystal A, McCall WV, Roth T, Schaefer K, et al. Eszopiclone coadministered with escitalopram in patients with insomnia and comorbid generalized anxiety disorder. Arch Gen Psychiatry. 2008;65(5):551-62.

784. Pollmann AS, Murphy AL, Bergman JC, Gardner DM. Deprescribing benzodiazepines and Z- drugs in community-dwelling adults: a scoping review. BMC Pharmacol Toxicol. 2015;16:19.

785. Polsky D, Glick HA, Yang J, Subramaniam GA, Poole SA, Woody GE. Cost-effectiveness of extended buprenorphine naloxone treatment for opioid-dependent youth: Data from a randomized trial. Addiction. 2010;105(9):1616-24.

786. Popovici I, Maclean JC, Hijazi B, Radakrishnan S. The effect of state laws designed to prevent nonmedical prescription opioid use on overdose deaths and treatment. Health Econ. 2018;27(2):294- 305.

787. Posadzki P, Mastellos N, Ryan R, Gunn L, Felix L, Pappas Y, et al. Automated telephone communication systems for preventive healthcare and management of long‐term conditions. Cochrane Database of Systematic Reviews [Internet]. 2016; (12). Available from: http://dx.doi.org/10.1002/14651858.CD009921.pub2.

788. Posadzki P, Khalil MMK, AlBedah AMN, Zhabenko O, Car J. Complementary and alternative medicine for addiction: an overview of systematic reviews. Focus on Alternative and Complementary Therapies. 2016;21(2):69-81.

789. Potter JS, Chakrabarti A, Domier CP, Hillhouse MP, Weiss RD, Ling W. Pain and continued opioid use in individuals receiving buprenorphine-naloxone for opioid detoxification: secondary analyses from the Clinical Trials Network. J Subst Abuse Treat. 2010;38 Suppl 1:S80-6.

790. Potter JS, Dreifuss JA, Marino EN, Provost SE, Dodd DR, Rice LS, et al. The multi-site prescription opioid addiction treatment study: 18-month outcomes. J Subst Abuse Treat. 2015;48(1):62-9.

791. Pottie K, Thompson W, Davies S, Grenier J, Sadowski CA, Welch V, et al. Deprescribing benzodiazepine receptor agonists: Evidence-based clinical practice guideline. Can Fam Physician. 2018;64(5):339-51.

792. Prakash S, Balhara Y. Perceptions Related to Pharmacological Treatment of Opioid Dependence Among Individuals Seeking Treatment at a Tertiary Care Center in Northern India: A Descriptive Study. Subst Use Misuse. 2016;51(7):861.

793. Prendergast ML, Urada D, Podus D. Meta-analysis of HIV risk-reduction interventions within drug abuse treatment programs. J Consult Clin Psychol. 2001;69(3):389-405.

References

794. Priddy SE, Hanley AW, Riquino MR, Platt KA, Baker AK, Garland EL. Dispositional mindfulness and prescription opioid misuse among chronic pain patients: Craving and attention to positive information as mediating mechanisms. Drug Alcohol Depend. 2018;188:86.

795. Prommer EMDHF, McPherson MLPMABCPE, Walker KPBCPE, Davis MMDFF. Hey, Doc, Can We Cut Down on the Pain Pills? Current Status of Opioid Tapering Guidelines and How to Do It (SA525). J Pain Symptom Manage. 2017;53(2):406.

796. Puangkot S, Laohasiriwong W, Saengsuwan J, Chiawiriyabunya I. Prevalence of benzodiazepines misuse in Ubon Ratchathani province Thailand. J Med Assoc Thai. 2011;94(1):118- 22.

797. Commissioning treatment for dependence on prescription and over-the-counter medicines: a guide for NHS and local authority commissioners. London: PHE, 2013 2013. Report No.

798. Puustinen J, Lahteenmaki R, Polo-Kantola P, Salo P, Vahlberg T, Lyles A, et al. Effect of withdrawal from long-term use of temazepam, zopiclone or zolpidem as hypnotic agents on cognition in older adults. Eur J Clin Pharmacol. 2014;70(3):319-29.

799. Quagliato LA, Freire RC, Nardi AE. Risks and benefits of medications for panic disorder: a comparison of SSRIs and benzodiazepines. Expert Opinion on Drug Safety. 2018;17(3):315-24.

800. Raheem OA, Patel SH, Sisul D, Furnish TJ, Hsieh T-C. The role of testosterone supplemental therapy in opioid-induced hypogonadism: A retrospective pilot analysis. American Journal of Men's Health. 2017;11(4):1208-13.

801. Rahimi‐Movaghar A, Amin‐Esmaeili M, Hefazi M, Yousefi‐Nooraie R. Pharmacological therapies for maintenance treatments of opium dependence. Cochrane Database of Systematic Reviews [Internet]. 2013; (1). Available from: http://dx.doi.org/10.1002/14651858.CD007775.pub2.

802. Raisch DW, Campbell HM, Garnand DA, Jones MA, Sather MR, Naik R, et al. Health-related quality of life changes associated with buprenorphine treatment for opioid dependence. Quality of Life Research: An International Journal of Quality of Life Aspects of Treatment, Care & Rehabilitation. 2012;21(7):1177-83.

803. Raknes G, Smabrekke L. Low-dose naltrexone and opioid consumption: a drug utilization cohort study based on data from the Norwegian prescription database. Pharmacoepidemiol Drug Saf. 2017;26(6):685-93.

804. Raoof M, Nawaz H, Nusrat R, Pabaney AH, Randhawa AR, Rehman R, et al. Awareness and use of benzodiazepines in healthy volunteers and ambulatory patients visiting a tertiary care hospital: A cross sectional survey. PLoS One. 2008;3 (3) (no pagination)(e1804).

805. Raphael JH, Duarte RV, Southall JL, Nightingale P, Kitas GD. Randomised, double-blind controlled trial by dose reduction of implanted intrathecal morphine delivery in chronic non-cancer pain. BMJ Open. 2013;3(7):31.

806. Raskin P, Huffman C, Toth C, Asmus MJ, Messig M, Sanchez RJ, et al. Pregabalin in patients with inadequately treated painful diabetic peripheral neuropathy: a randomized withdrawal trial. Clin J Pain. 2014;30(5):379-90.

807. Rathlev N, Almomen R, Deutsch A, Smithline H, Li H, Visintainer P. Randomized Controlled Trial of Electronic Care Plan Alerts and Resource Utilization by High Frequency Emergency Department Users with Opioid Use Disorder. The Western Journal of Emergency Medicine. 2016;17(1):28-34.

References

808. Reece AS. Favorable Mortality Profile of Naltrexone Implants for Opiate Addiction. J Addict Dis. 2010;29(1):30-50.

809. Reece AS. Comparative treatment and mortality correlates and adverse event profile of implant naltrexone and sublingual buprenorphine. J Subst Abuse Treat. 2009;37(3):256-65.

810. Reeve E, Ong M, Wu A, Jansen J, Petrovic M, Gnjidic D. A systematic review of interventions to deprescribe benzodiazepines and other hypnotics among older people. Eur J Clin Pharmacol. 2017;73(8):927-35.

811. Reid MC, Engles-Horton LL, Weber MB, Kerns RD, Rogers EL, O'Connor PG. Use of opioid medications for chronic noncancer pain syndromes in primary care. J Gen Intern Med. 2002;17(3):173-9.

812. Reifler LM, Droz D, Bailey JE, Schnoll SH, Fant R, Dart RC, et al. Do prescription monitoring programs impact state trends in opioid abuse/misuse? Pain Med. 2012;13(3):434-42.

813. Reimer J, Wright N, Somaini L, Roncero C, Maremmani I, McKeganey N, et al. The impact of misuse and diversion of opioid substitution treatment medicines: Evidence review and expert consensus. Eur Addict Res. 2016;22(2):99-106.

814. Reisman RM, Shenoy PJ, Atherly AJ, Flowers CR. Prescription opioid usage and abuse relationships: an evaluation of state prescription drug monitoring program efficacy. Substance abuse : research and treatment. 2009;3:41-51.

815. Rickels K, Schweizer E, Garcia Espana F, Case G, DeMartinis N, Greenblatt D. Trazodone and valproate in patients discontinuing long-term benzodiazepine therapy: effects on withdrawal symptoms and taper outcome. Psychopharmacology (Berl). 1999;141(1):1-5.

816. Rigg KK, Ibanez GE. Motivations for non-medical prescription drug use: A mixed methods analysis. J Subst Abuse Treat. 2010;39(3):236-47.

817. Rikala M, Korhonen MJ, Sulkava R, Hartikainen S. Psychotropic drug use in community- dwelling elderly people - Characteristics of persistent and incident users. Eur J Clin Pharmacol. 2011;67(7):731-9.

818. Ringwalt C, Gugelmann H, Garrettson M, Dasgupta N, Chung AE, Proescholdbell SK, et al. Differential prescribing of opioid analgesics according to physician specialty for Medicaid patients with chronic noncancer pain diagnoses. Pain research & management. 2014;19(4):179-85.

819. Ristanovic RK, Liang H, Hornfeldt CS, Lai C. Exacerbation of cataplexy following gradual withdrawal of antidepressants: manifestation of probable protracted rebound cataplexy. Sleep Med. 2009;10(4):416-21.

820. Robinson JP, Dansie EJ, Wilson HD, Rapp S, Turk DC. Attitudes and beliefs of working and work-disabled people with chronic pain prescribed long-term opioids. Pain Med. 2015;16(7):1311-24.

821. Roehrs TA, Randall S, Harris E, Maan R, Roth T. Twelve months of nightly zolpidem does not lead to rebound insomnia or withdrawal symptoms: a prospective placebo-controlled study. Journal of Psychopharmacology. 2012;26(8):1088-95.

822. Rojo-Mota G, Pedrero-Pérez EJ, Huertas-Hoyas E. Systematic Review of Occupational Therapy in the Treatment of Addiction: Models, Practice, and Qualitative and Quantitative Research. The American Journal of Occupational Therapy. 2017;71(5):1-11.

823. Romach MK, Kaplan HL, Busto UE, Somer G, Sellers EM. A controlled trial of ondansetron, a 5-HT3 antagonist, in benzodiazepine discontinuation. J Clin Psychopharmacol. 1998;18(2):121-31.

References

824. Romach MK, Otton SV, Somer G, Tyndale RF, Sellers EM. Cytochrome P450 2D6 and treatment of codeine dependence. J Clin Psychopharmacol. 2000;20(1):43-5.

825. Pregabalin (Lyrica ®, Pfizer, Inc.). Issues Ment Health Nurs. 2009;30(2):138-9.

826. Rosenblum A, Cruciani RA, Strain EC, Cleland CM, Joseph H, Magura S, et al. Sublingual buprenorphine/naloxone for chronic pain in at-risk patients: development and pilot test of a clinical protocol. J Opioid Manag. 2012;8(6):369-82.

827. Rosenthal RN, Ling W, Casadonte P, Vocci F, Bailey GL, Kampman K, et al. Buprenorphine implants for treatment of opioid dependence: randomized comparison to placebo and sublingual buprenorphine/naloxone. Addiction. 2013;108(12):2141-9.

828. Rounsaville BJ, Glazer W, Wilber CH, Weissman MM, Kleber HD. Short-term interpersonal psychotherapy in methadone-maintained opiate addicts. Arch Gen Psychiatry. 1983;40(6):629-36.

829. Roussin A, Bouyssi A, Pouche L, Pourcel L, Lapeyre-Mestre M. Misuse and dependence on non-prescription codeine analgesics or sedative H1 antihistamines by adults: a cross-sectional investigation in France. PLoS ONE [Electronic Resource]. 2013;8(10):e76499.

830. Roux P, Sullivan MA, Cohen J, Fugon L, Jones JD, Vosburg SK, et al. Buprenorphine/naloxone as a promising therapeutic option for opioid abusing patients with chronic pain: reduction of pain, opioid withdrawal symptoms, and abuse liability of oral oxycodone. Pain. 2013;154(8):1442-8.

831. Rudolf GMD, Walsh JMD, Plawman AMD, Gianutsos PMDMPH, Alto WMD, Mancl LP, et al. A novel non-opioid protocol for medically supervised opioid withdrawal and transition to antagonist treatment. The American Journal of Drug and Alcohol Abuse. 2018;44(3):302-9.

832. Ruetsch C. Treating prescription opioid dependence. JAMA - Journal of the American Medical Association. 2014;312(11):1145-6.

833. Ruetsch C, Tkacz J, McPherson TL, Cacciola J. The effect of telephonic patient support on treatment for opioid dependence: outcomes at one year follow-up. Addict Behav. 2012;37(5):686-9.

834. Ruetsch C, Cacciola J, Tkacz J. A national study of a telephone support service for patients receiving office-based buprenorphine medication-assisted treatment: study feasibility and sample description. J Subst Abuse Treat. 2010;39(4):307-17.

835. Sabzghabaee AM, Eizadi-Mood N, Gheshlaghi F, Javani A, Shirani S, Aghaabdollahian S. Role of Benzodiazepines in the management of agitation due to inappropriate use of naltrexone. Iran J Nurs Midwifery Res. 2012;17(5):365-9.

836. Safer DJ. Differing antidepressant maintenance methodologies. Contemp Clin Trials. 2017;61:87-95.

837. Saleh MI. Modeling longitudinal changes in buprenorphine treatment outcome for opioid dependence. Pharmacopsychiatry. 2014;47(7):251-8.

838. Saleh MI. Predictors of long term opioid withdrawal outcome after short-term stabilization with buprenorphine. Eur Rev Med Pharmacol Sci. 2014;18(24):3935-42.

839. Salehi M, Barekatain M, Faghani F, Karimian N, Molaeinezhad M, Asadalloahi GA, et al. Bupropion efficacy on sexual dysfunction among male patients on methadone maintenance therapy: A double-blind placebo-controlled trial. Sex Relation Ther. 2015;30(3):364-75.

References

840. Salminen M, Vahlberg T, Kivela SL. The long-term effect of a multifactorial fall prevention programme on the incidence of falls requiring medical treatment. Public Health. 2009;123(12):809- 13.

841. Sanchez K, Greer TL, Walker R, Carmody T, Rethorst CD, Trivedi MH. Racial and ethnic differences in treatment outcomes among adults with stimulant use disorders after a dosed exercise intervention. Journal of Ethnicity in Substance Abuse. 2017;16(4):495-510.

842. Saul PA, Korlipara K, Presley P. A randomised, multicentre, double-blind, comparison of atenolol and placebo in the control of benzodiazepine withdrawal symptoms. Acta Ther. 1989;15(2):117-23.

843. Saulle R, Vecchi S, Gowing L. Supervised dosing with a long‐acting opioid medication in the management of opioid dependence. Cochrane Database of Systematic Reviews [Internet]. 2017; (4). Available from: http://dx.doi.org/10.1002/14651858.CD011983.pub2.

844. Saunders KW, Shortreed S, Thielke S, Turner JA, LeResche L, Beck R, et al. Evaluation of Health Plan Interventions to Influence Chronic Opioid Therapy Prescribing. The Clinical journal of pain. 2015;31(9):820-9.

845. Saxon AJ, Akerman SC, Liu CC, Sullivan MA, Silverman BL, Vocci FJ. Extended-release naltrexone (XR-NTX) for opioid use disorder in clinical practice: Vivitrol's Cost and Treatment Outcomes Registry. Addiction. 2018;113(8):1477-87.

846. Saxon AJ, Ling W, Hillhouse M, Thomas C, Hasson A, Ang A, et al. Buprenorphine/Naloxone and methadone effects on laboratory indices of liver health: a randomized trial. Drug Alcohol Depend. 2013;128(1-2):71-6. doi: 10.1016/j.drugalcdep.2012.08.002. Epub Aug 22.

847. Saxon L, Borg S, Hiltunen AJ. Reduction of aggression during benzodiazepine withdrawal: effects of flumazenil. Pharmacol Biochem Behav. 2010;96(2):148-51.

848. Schaffer CB, Schaffer LC, Miller AR, Hang E, Nordahl TE. Preliminary communication: Efficacy and safety of nonbenzodiazepine hypnotics for chronic insomnia in patients with bipolar disorder. J Affect Disord. 2011;128(3):305-8.

849. Schepis TS, Teter CJ, McCabe SE. Prescription drug use, misuse and related substance use disorder symptoms vary by educational status and attainment in U.S. adolescents and young adults. Drug Alcohol Depend. 2018;189:172-7.

850. Scherbaum N, Kluwig J, Specka M, Krause D, Merget B, Finkbeiner T, et al. Group psychotherapy for opiate addicts in methadone maintenance treatment--a controlled trial. Eur Addict Res. 2005;11(4):163-71.

851. Schieffer BM, Pham Q, Labus J, Baria A, Van Vort W, Davis P, et al. Pain medication beliefs and medication misuse in chronic pain. J Pain. 2005;6(9):620-9.

852. Schifano F, Chiappini S, Corkery JM, Guirguis A. Abuse of Prescription Drugs in the Context of Novel Psychoactive Substances (NPS): A Systematic Review. Brain sciences. 2018;8(4).

853. Schroeder JR, Phillips KA, Epstein DH, Jobes ML, Furnari MA, Kennedy AP, et al. Assessment of pioglitazone and proinflammatory cytokines during buprenorphine taper in patients with opioid use disorder. Psychopharmacology (Berl). 2018;235(10):2957-66.

854. Schuman-Olivier Z, Weiss RD, Hoeppner BB, Borodovsky J, Albanese MJ. Emerging adult age status predicts poor buprenorphine treatment retention. J Subst Abuse Treat. 2014;47(3):202-12.

References

855. Schuman-Olivier Z, Claire Greene M, Bergman BG, Kelly JF. Is residential treatment effective for opioid use disorders? A longitudinal comparison of treatment outcomes among opioid dependent, opioid misusing, and non-opioid using emerging adults with substance use disorder. Drug Alcohol Depend. 2014;144:178-85.

856. Schwarzer A, Aichinger-Hinterhofer M, Maier C, Vollert J, Walther JW. Sleep-disordered breathing decreases after opioid withdrawal: results of a prospective controlled trial. Pain. 2015;156(11):2167-74.

857. Schweizer E, Case WG, Garcia-Espana F, Greenblatt DJ, Rickels K. Progesterone co- administration in patients discontinuing long-term benzodiazepine therapy: effects on withdrawal severity and taper outcome. Psychopharmacology (Berl). 1995;117(4):424-9.

858. Schweizer E, Rickels K, Case WG, Greenblatt DJ. Long-term therapeutic use of benzodiazepines. II. Effects of gradual taper. Arch Gen Psychiatry. 1990;47(10):908-15.

859. Seay KD, Iachini AL, DeHart DD, Browne T, Clone S. Substance abuse treatment engagement among mothers: Perceptions of the parenting role and agency-related motivators and inhibitors. Journal of Family Social Work. 2017;20(3):196-212.

860. Segal ZV, Bieling P, Young T, MacQueen G, Cooke R, Martin L, et al. Antidepressant monotherapy vs sequential pharmacotherapy and mindfulness-based cognitive therapy, or placebo, for relapse prophylaxis in recurrent depression. In: Gaudiano BA, editor. Mindfulness: Clinical applications of mindfulness and acceptance: Specific interventions for psychiatric, behavioural, and physical health conditions (Vol III). Major themes in mental health: Routledge/Taylor & Francis Group, New York, NY; 2017. p. 137-56, Chapter xii, 648 Pages.

861. Sekhon R, Aminjavahery N, Davis CN, Roswarski MJ, Robinette C. Compliance with Opioid Treatment Guidelines for Chronic Non-Cancer Pain (CNCP) in Primary Care at a Veterans Affairs Medical Center (VAMC). Pain Medicine (United States). 2013;14(10):1548-56.

862. Setnik B, Roland CL, Goli V, Sommerville K, Webster L. A clinical trial to determine if corelease of morphine and naltrexone from crushed extended-release capsules induces withdrawal in opioid- dependent patients: a descriptive analysis of six patients. Journal of Opioid Management. 2013;9(2):139-50.

863. Setnik B, Roland CL, Sommerville KW, Pixton GC, Berke R, Calkins A, et al. A multicenter, primary care-based, open-label study to identify behaviors related to prescription opioid misuse, abuse, and diversion in opioid-experienced patients with chronic moderate-to-severe pain. J Pain Res. 2015;8:361-73.

864. Setnik B, Roland CL, Pixton GC, Sommerville KW. Prescription opioid abuse and misuse: gap between primary-care investigator assessment and actual extent of these behaviors among patients with chronic pain. Postgrad Med. 2017;129(1):5-11.

865. Shigemura J, Sato Y, Yoshino A, Nomura S. Research report: Patient satisfaction with antidepressants: An Internet-based study. J Affect Disord. 2008;107(1-3):155-60.

866. Shokrzadeh M, Hajimohammadi A, Delaram A, Shayeste Y. Characteristics of patients hospitalized with tramadol intoxication in Gorgan, Iran, 2008 - 2015. Journal of mazandaran university of medical sciences. 2017;26(146):185‐90.

867. Shorey RC, Elmquist J, Gawrysiak MJ, Strauss C, Haynes E, Anderson S, et al. A Randomized Controlled Trial of a Mindfulness and Acceptance Group Therapy for Residential Substance Use Patients. Subst Use Misuse. 2017;52(11):1400-10.

References

868. Siassi I, Thomas M, Vanov SK. Evaluation of the safety and therapeutic effects of lorazepam on long-term use. Curr Ther Res Clin Exp. 1975;18(1 pt 2):163‐71.

869. Sigmon SC, Dunn KE, Badger GJ, Heil SH, Higgins ST. Brief buprenorphine detoxification for the treatment of prescription opioid dependence: A pilot study. Addict Behav. 2009;34(3):304-11.

870. Sigmon SC, Dunn KE, Saulsgiver K, Patrick ME, Badger GJ, Heil SH, et al. A randomized, double-blind evaluation of buprenorphine taper duration in primary prescription opioid abusers. JAMA psychiatry. 2013;70(12):1347-54.

871. Simon JA, Portman DJ, Kaunitz AM, Mekonnen H, Kazempour K, Bhaskar S, et al. Low-dose paroxetine 7.5 mg for menopausal vasomotor symptoms: two randomized controlled trials. Menopause. 2013;20(10):1027-35.

872. Simonsen AL, Danborg PB, Gotzsche PC. Persistent sexual dysfunction after early exposure to SSRIs: Systematic review of animal studies. Int J Risk Saf Med. 2016;28(1):1-12.

873. Sinnemaki J, Sihvo S, Isojarvi J, Blom M, Airaksinen M, Mantyla A. Automated dose dispensing service for primary healthcare patients: a systematic review. Systematic Reviews. 2013;2:1.

874. Slofstra C, Klein NS, Nauta MH, Wichers M, Batalas N, Bockting CLH. Imagine your mood: Study design and protocol of a randomized controlled micro-trial using app-based experience sampling methodology to explore processes of change during relapse prevention interventions for recurrent depression. Contemp Clin Trials Commun. 2017;7:172-8.

875. Sirdifield C, Chipchase SY, Owen S, Siriwardena AN. A Systematic Review and Meta-Synthesis of Patients' Experiences and Perceptions of Seeking and Using Benzodiazepines and Z-Drugs: Towards Safer Prescribing. Patient. 2017;10(1):1-15.

876. Smith AJ, Tett SE. An intervention to improve benzodiazepine use-a new approach. Fam Pract. 2010;27(3):320-7.

877. Socias ME, Ahamad K, Le Foll B, Lim R, Bruneau J, Fischer B, et al. The OPTIMA study, buprenorphine/naloxone and methadone models of care for the treatment of prescription opioid use disorder: Study design and rationale. Contemp Clin Trials. 2018;69:21-7.

878. Soeffing JM, Martin LD, Fingerhood MI, Jasinski DR, Rastegar DA. Buprenorphine maintenance treatment in a primary care setting: Outcomes at 1 year. J Subst Abuse Treat. 2009;37(4):426-30.

879. Sohler NL, Li X, Kunins HV, Sacajiu G, Giovanniello A, Whitley S, et al. Home- versus office- based buprenorphine inductions for opioid-dependent patients. J Subst Abuse Treat. 2010;38(2):153- 9.

880. Sokol R, LaVertu AE, Morrill D, Albanese C, Schuman-Olivier Z. Group-based treatment of opioid use disorder with buprenorphine: A systematic review. J Subst Abuse Treat. 2018;84:78-87.

881. Soldatos CR, Dikeos DG, Whitehead A. Tolerance and rebound insomnia with rapidly eliminated hypnotics: a meta-analysis of sleep laboratory studies. Int Clin Psychopharmacol. 1999;14(5):287-303.

882. Soyka M, Trader A, Klotsche J, Backmund M, Buhringer G, Rehm JT, et al. Mortality in long- term opioid maintenance treatment in Germany: frequency, causes and predictors. Suchtmedizin in forschung und praxis. 2011;13(5):247‐52.

883. Spiegel DA, Bruce TJ, Gregg SF, Nuzzarello A. Does cognitive behavior therapy assist slow- taper alprazolam discontinuation in panic disorder? Am J Psychiatry. 1994;151(6):876-81.

References

884. Spiller H, Lorenz DJ, Bailey EJ, Dart RC. Epidemiological Trends in Abuse and Misuse of Prescription Opioids. J Addict Dis. 2009;28(2):130-6.

885. Stannard C, Gaskell H, Derry S, Aldington D, Cole P, Cooper T, et al. Hydromorphone for neuropathic pain in adults. Cochrane Database of Systematic Reviews [Internet]. 2016; (5). Available from: http://dx.doi.org/10.1002/14651858.CD011604.pub2.

886. Stanton MD, Shadish WR. Outcome, attrition, and family-couples treatment for drug abuse: a meta-analysis and review of the controlled, comparative studies. Psychol Bull. 1997;122(2):170-91.

887. Starrels JL, Becker WC, Alford DP, Kapoor A, Williams AR, Turner BJ. Systematic review: treatment agreements and urine drug testing to reduce opioid misuse in patients with chronic pain. Ann Intern Med. 2010;152(11):712-20.

888. Steele A, Cunningham P. A comparison of suboxone and clonidine treatment outcomes in opiate detoxification. Arch Psychiatr Nurs. 2012;26(4):316-23.

889. Stein MDMD, Anderson BJP, Thurmond PMD, Bailey GLMD. Comparing the life concerns of prescription opioid and heroin users. J Subst Abuse Treat. 2015;48(1):43.

890. Stein DJ, Ahokas AA, de Bodinat C. Efficacy of agomelatine in generalized anxiety disorder: a randomized, double-blind, placebo-controlled study. J Clin Psychopharmacol. 2008;28(5):561-6.

891. Stewart R, Niessen WJ, Broer J, Snijders TA, Haaijer-Ruskamp FM, Meyboom-De Jong B. General Practitioners reduced benzodiazepine prescriptions in an intervention study: a multilevel application. J Clin Epidemiol. 2007;60(10):1076‐84.

892. Strain EC, Harrison JA, Bigelow GE. Induction of opioid-dependent individuals onto buprenorphine and buprenorphine/naloxone soluble-films. Clin Pharmacol Ther. 2011;89(3):443-9.

893. Strang J, Reed K, Bogdanowicz K, Bell J, van der Waal R, Keen J, et al. Randomised Comparison of a Novel Buprenorphine Oral Lyophilisate versus Existing Buprenorphine Sublingual Tablets in Opioid-Dependent Patients: A First-in-Patient Phase II Randomised Open Label Safety Study. Eur Addict Res. 2017;23(2):61-70.

894. Strick V. Management of severe chronic pain with tapentadol prolonged release - long-term data from pain specialists. Curr Med Res Opin. 2014;30(10):2085-92.

895. Su MK, Lopez JH, Crossa A, Hoffman RS. Low dose intramuscular methadone for acute mild to moderate opioid withdrawal syndrome. Am J Emerg Med. 2018;02:02.

896. Subodh BN, Hazari N, Elwadhi D, Basu D. Prevalence of dual diagnosis among clinic attending patients in a de-addiction centre of a tertiary care hospital. Asian J Psychiatr. 2017;25:169-74.

897. Subramaniam S, DeFulio A, Jarvis BP, Holtyn AF, Silverman K. Earning, Spending, and Drug Use in a Therapeutic Workplace. The Psychological Record. 2017;67(2):273-83.

898. Sullivan MD, Von Korff M, Banta-Green C, Merrill JO, Saunders K. Problems and concerns of patients receiving chronic opioid therapy for chronic non-cancer pain. Pain. 2010;149(2):345-53.

899. Sullivan M, Bisaga A, Pavlicova M, Choi CJ, Mishlen K, Carpenter KM, et al. Long-Acting Injectable Naltrexone Induction: A Randomized Trial of Outpatient Opioid Detoxification With Naltrexone Versus Buprenorphine. Am J Psychiatry. 2017;174(5):459-67.

900. Sun EC, Darnall BD, Baker LC, Mackey S. Incidence of and Risk Factors for Chronic Opioid Use Among Opioid-Naive Patients in the Postoperative Period.[Erratum appears in JAMA Intern Med. 2016 Sep 1;176(9):1412; PMID: 27533349]. JAMA Intern Med. 2016;176(9):1286-93.

References

901. Syed YY, Keating GM. Extended-release intramuscular naltrexone (VIVITROL): a review of its use in the prevention of relapse to opioid dependence in detoxified patients. CNS Drugs. 2013;27(10):851-61.

902. Syed UAM, Aleem AW, Wowkanech C, Weekes D, Freedman M, Tjoumakaris F, et al. Neer Award 2018: the effect of preoperative education on opioid consumption in patients undergoing arthroscopic rotator cuff repair: a prospective, randomized clinical trial. J Shoulder Elbow Surg. 2018;27(6):962-7.

903. Sylvestre MP, Abrahamowicz M, Capek R, Tamblyn R. Assessing the cumulative effects of exposure to selected benzodiazepines on the risk of fall-related injuries in the elderly. Int Psychogeriatr. 2012;24(4):577-86.

904. Szczypa P, Hadley SJ, Donevan S, Mandel FS, Leon T. P.4.a.016 Switching from long-term benzodiazepine therapy to pregabalin in patients with generalised anxiety disorder (GAD). Eur Neuropsychopharmacol. 2009;19:S594-S5.

905. Tavakolian E, Abolghasemi A. Effects of cognitive restructuring training on neurocognitive functions in opioid addicts. Archives of Psychiatry and Psychotherapy. 2016;18(1):14-21.

906. Tetrault JM, Moore BA, Barry DT, O'Connor PG, Schottenfeld R, Fiellin DA, et al. Brief versus extended counseling along with buprenorphine/naloxone for HIV-infected opioid dependent patients. J Subst Abuse Treat. 2012;43(4):433-9.

907. Thakral M, Walker RL, Saunders K, Shortreed SM, Parchman M, Hansen RN, et al. Comparing pain and depressive symptoms of chronic opioid therapy patients receiving dose reduction and risk mitigation initiatives with usual care. The Journal of Pain. 2018;19(1):111-20.

908. Thorlund K, Sun-Edelstein C, Druyts E, Kanters S, Ebrahim S, Bhambri R, et al. Risk of medication overuse headache across classes of treatments for acute migraine. J Headache Pain. 2016;17 (1) (no pagination)(107).

909. Tint A, Haddad PM, Anderson IM. Erratum: The effect of rate of antidepressant tapering on the incidence of discontinuation symptoms: A randomised study (Journal of Psychopharmacology (2008) 22:3 DOI: 10.1177/0269881107081550). Journal of Psychopharmacology. 2009;23(8):1006.

910. Tkacz J, Severt J, Kassed C, Ruetsch C. Clinical Differences Between Opioid Abuse Classes Ameliorated After 1 Year of Buprenorphine-Medication Assisted Treatment. J Addict Dis. 2012;31(2):100-11.

911. Tkacz J, Severt J, Cacciola J, Ruetsch C. Compliance with Buprenorphine Medication-Assisted Treatment and Relapse to Opioid Use. The American Journal on Addictions. 2012;21(1):55-62.

912. Tofighi B, Grazioli F, Bereket S, Grossman E, Aphinyanaphongs Y, Lee JD. Text message reminders for improving patient appointment adherence in an office-based buprenorphine program: A feasibility study. Am J Addict. 2017;26(6):581-6.

913. Tompkins DA, Bigelow GE, Harrison JA, Johnson RE, Fudala PJ, Strain EC. Concurrent validation of the Clinical Opiate Withdrawal Scale (COWS) and single-item indices against the Clinical Institute Narcotic Assessment (CINA) opioid withdrawal instrument. Drug Alcohol Depend. 2009;105(1-2):154-9.

914. Tompkins DA, Smith MT, Mintzer MZ, Campbell CM, Strain EC. A double blind, within subject comparison of spontaneous opioid withdrawal from buprenorphine versus morphine. J Pharmacol Exp Ther. 2014;348(2):217-26.

References

915. Tonks A. Combined drug treatment for opioid addiction works, while it lasts. Br Med J. 2008;337(7679):1136.

916. Tonks A. Implants for opiate addiction are effective in preliminary trials. Br Med J. 2010;341(7777):803.

917. Tonks A. Long term opioids are a risky treatment for non-cancer pain. Br Med J. 2010;340(7740):239.

918. Tonks A. Higher doses mean higher risks for patients taking prescription opioids. Br Med J. 2011;342(7801):791.

919. Tourian KA, Pitrosky B, Padmanabhan SK, Rosas GR. A 10-month, open-label evaluation of desvenlafaxine in outpatients with major depressive disorder. The Primary Care Companion to CNS Disorders. 2011;13(2).

920. Tournebize J, Gibaja V, Muszczak A, Kahn JP. Are Physicians Safely Prescribing Opioids for Chronic Noncancer Pain? A Systematic Review of Current Evidence. Pain practice : the official journal of World Institute of Pain. 2016;16(3):370-83.

921. Townsend CO, Kerkvliet JL, Bruce BK, Rome JD, Hooten WM, Luedtke CA, et al. A longitudinal study of the efficacy of a comprehensive pain rehabilitation program with opioid withdrawal: comparison of treatment outcomes based on opioid use status at admission. Pain. 2008;140(1):177- 89.

922. Tribl GG, Wetter TC, Schredl M. Dreaming under antidepressants: a systematic review on evidence in depressive patients and healthy volunteers. Sleep Med Rev. 2013;17(2):133-42.

923. Trivedi MH, Dunner DL, Kornstein SG, Thase ME, Zajecka JM, Rothschild AJ, et al. Psychosocial outcomes in patients with recurrent major depressive disorder during 2 years of maintenance treatment with venlafaxine extended release. J Affect Disord. 2010;126(3):420-9.

924. Turk DC, Swanson KS, Gatchel RJ. Predicting opioid misuse by chronic pain patients: a systematic review and literature synthesis. Clin J Pain. 2008;24(6):497-508.

925. Tyrer P, Rutherford D, Huggett T. Benzodiazepine withdrawal symptoms and propranolol. Lancet. 1981;1(8219):520-2.

926. Tyrer P, Ferguson B, Hallstrom C, Michie M, Tyrer S, Cooper S, et al. A controlled trial of dothiepin and placebo in treating benzodiazepine withdrawal symptoms. Br J Psychiatry. 1996;168(4):457-61.

927. Udelman HD, Udelman DL. Concurrent use of buspirone in anxious patients during withdrawal from alprazolam therapy. J Clin Psychiatry. 1990;51(Suppl):46-50.

928. Uebelacker LA, Weisberg RB, Herman DS, Bailey GL, Pinkston-Camp MM, Stein MD. Chronic Pain in HIV-Infected Patients: Relationship to Depression, Substance Use, and Mental Health and Pain Treatment. Pain Med. 2015;16(10):1870-81.

929. Urton MS, Rohlik E, Farrell M, Ng W, Woodard EK. Decreasing Opioid Utilization in Rehabilitation Patients Using a Clinical Nurse Specialist Pain Consultant Program. Arch Phys Med Rehabil. 2017;98(12):2491-7.

930. Uzun S, Kozumplik O, Jakovljevic M, Sedic B. Side effects of treatment with benzodiazepines. Psychiatr. 2010;22(1):90-3.

References

931. Vaapio S, Puustinen J, Salminen MJ, Vahlberg T, Salonoja M, Lyles A, et al. Symptoms associated with long-term benzodiazepine use in elderly individuals aged 65 years and older: A longitudinal descriptive study. International Journal of Gerontology. 2015;9(1):34-9.

932. Vankova H, Holmerova I, Machacova K, Volicer L, Veleta P, Celko AM. The effect of dance on depressive symptoms in nursing home residents. J Am Med Dir Assoc. 2014;15(8):582-7.

933. Veal FC, Thompson AJ, Bereznicki LR, Peterson GM. Persistent pain management in Australian general practice. Australian Journal of Primary Health. 2017;23(2):105-6.

934. Veiraiah A, Dyas J, Cooper G, Routledge PA, Thompson JP. Flumazenil use in benzodiazepine overdose in the UK: a retrospective survey of NPIS data. Emerg Med J. 2012;29(7):565-9.

935. Velert Vila J, Moreno Royo L, Velert Vila MM, Salar Ibanez L. Pharmacist can improve the use of benzodiazepines. Pharmaceutical care espana. 2012;14(3):94‐101.

936. Vicens C, Fiol F, Llobera J, Campoamor F, Mateu C, Alegret S, et al. Withdrawal from long- term benzodiazepine use: randomised trial in family practice. Br J Gen Pract. 2006;56(533):958-63.

937. Vicens C, Socias I, Mateu C, Leiva A, Bejarano F, Sempere E, et al. Comparative efficacy of two primary care interventions to assist withdrawal from long term benzodiazepine use: a protocol for a clustered, randomized clinical trial. BMC Fam Pract. 2011;12:23.

938. Victorri-Vigneau C, Sebille V, Gerardin M, Simon D, Pivette J, Jolliet P. Epidemiological Characterization of Drug Overconsumption: The Example of Antidepressants. J Addict Dis. 2011;30(4):342-50.

939. Victorri-Vigneau C, Dailly E, Veyrac G, Jolliet P. Evidence of zolpidem abuse and dependence: results of the French Centre for Evaluation and Information on Pharmacodependence (CEIP) network survey. Br J Clin Pharmacol. 2007;64(2):198-209.

940. Vittengl JR. Poorer Long-Term Outcomes among Persons with Major Depressive Disorder Treated with Medication. Psychother Psychosom. 2017;86(5):302-4.

941. Vo HT, Burgower R, Rozenberg I, Fishman M. Home-based delivery of XR-NTX in youth with opioid addiction. J Subst Abuse Treat. 2018;85:84.

942. Voderholzer U, Riemann D, Hornyak M, Backhaus J, Feige B, Berger M, et al. A double-blind, randomized and placebo-controlled study on the polysomnographic withdrawal effects of zopiclone, zolpidem and triazolam in healthy subjects. Eur Arch Psychiatry Clin Neurosci. 2001;251(3):117-23.

943. Vogel M, Dursteler-MacFarland KM, Walter M, Strasser J, Fehr S, Prieto L, et al. Prolonged use of benzodiazepines is associated with childhood trauma in opioid-maintained patients. Drug Alcohol Depend. 2011;119(1-2):93-8.

944. Vogel M, Dürsteler KM, Walter M, Herdener M, Nordt C. Rethinking retention in treatment of opioid dependence-The eye of the beholder. The International Journal on Drug Policy. 2017;39:109- 33.

945. Von Korff M, Walker RL, Saunders K, Shortreed SM, Thakral M, Parchman M, et al. Prevalence of prescription opioid use disorder among chronic opioid therapy patients after health plan opioid dose and risk reduction initiatives. The International Journal on Drug Policy. 2017;46:90.

946. Voon P, Karamouzian M, Kerr T. Chronic pain and opioid misuse: a review of reviews. Subst Abuse Treat Prev Policy. 2017;12(1):36.

References

947. Voris John C, Barchiesi A. Experimental withdrawal of amitriptyline. Psychiatric forum. 1989;15(1):53‐6.

948. Vorma H, Naukkarinen H, Sarna S, Kuoppasalmi K. Treatment of out-patients with complicated benzodiazepine dependence: comparison of two approaches. Addiction. 2002;97(7):851-9.

949. Vorma H, Katila H. Effect of valproate on benzodiazepine withdrawal severity in opioid- dependent subjects: a pilot study. Heroin Addiction and Related Clinical Problems. 2010;13(1):15-20.

950. Vosburg SK, Jones JD, Manubay JM, Ashworth JB, Benedek IH, Comer SD. Assessment of a formulation designed to be crush-resistant in prescription opioid abusers. Drug Alcohol Depend. 2012;126(1-2):206-15.

951. Vosburg SK, Severtson SG, Dart RC, Cicero TJ, Kurtz SP, Parrino MW, et al. Assessment of Tapentadol API Abuse Liability With the Researched Abuse, Diversion and Addiction-Related Surveillance System. J Pain. 2018;19(4):439-53.

952. Voshaar RC, Gorgels WJ, Mol AJ, van Balkom AJ, Mulder J, van de Lisdonk EH, et al. Predictors of long-term benzodiazepine abstinence in participants of a randomized controlled benzodiazepine withdrawal program. Can J Psychiatry. 2006;51(7):445-52. doi: 10.1177/070674370605100706.

953. Vowles KE, McEntee ML, Julnes PS, Frohe T, Ney JP, van der Goes DN. Rates of opioid misuse, abuse, and addiction in chronic pain: a systematic review and data synthesis. Pain. 2015;156(4):569- 76.

954. Voyer P, Preville M, Cohen D, Berbiche D, Beland SG. The prevalence of benzodiazepine dependence among community-dwelling older adult users in Quebec according to typical and atypical criteria. Can J Aging. 2010;29(2):205-13.

955. Wakeland WP, Nielsen ABS, Schmidt TDMA, McCarty DP, Webster LRMD, Fitzgerald JP, et al. Modeling the Impact of Simulated Educational Interventions on the Use and Abuse of Pharmaceutical Opioids in the United States: A Report on Initial Efforts. Health Educ Behav. 2013;40(10):74S-86S.

956. Walsh SL, Comer SD, Lofwall MR, Vince B, Levy-Cooperman N, Kelsh D, et al. Effect of Buprenorphine Weekly Depot (CAM2038) and Hydromorphone Blockade in Individuals With Opioid Use Disorder: A Randomized Clinical Trial. JAMA psychiatry. 2017;74(9):894-902.

957. Wang EA, Moore BA, Sullivan LE, Fiellin DA. Effect of incarceration history on outcomes of primary care office-based buprenorphine/naloxone. J Gen Intern Med. 2010;25(7):670-4.

958. Wang H, Akbar M, Weinsheimer N, Gantz S, Schiltenwolf M. Longitudinal Observation of Changes in Pain Sensitivity during Opioid Tapering in Patients with Chronic Low-Back Pain. Pain Med. 2011;12(12):1720-6.

959. Wang MT, Chu CL, Yeh CB, Chang LC, Malone DC, Liou JT. Antidepressant use and risk of recurrent stroke: a population-based nested case-control study. J Clin Psychiatry. 2015;76(7):e877- 85.

960. Watson CP, Watt-Watson J, Chipman M. The long-term safety and efficacy of opioids: a survey of 84 selected patients with intractable chronic noncancer pain. Pain research & management. 2010;15(4):213-7.

961. Weber J, Siddiqui MA, Wagstaff AJ, McCormack PL. Low-dose doxepin: in the treatment of insomnia. CNS Drugs. 2010;24(8):713-20.

References

962. Webster LR, Webster RM. Predicting aberrant behaviors in opioid-treated patients: preliminary validation of the Opioid Risk Tool. Pain Med. 2005;6(6):432-42.

963. Webster LR, Brewer R, Wang C, Sekora D, Johnson FK, Morris D, et al. Long-Term Safety and Efficacy of Morphine Sulfate and Naltrexone Hydrochloride Extended Release Capsules, a Novel Formulation Containing Morphine and Sequestered Naltrexone, in Patients with Chronic, Moderate to Severe Pain. J Pain Symptom Manage. 2010;40(5):734-46.

964. Webster LR, Smith MD, Unal C, Finn A. Low-dose naloxone provides an abuse-deterrent effect to buprenorphine. J Pain Res. 2015;8:791-8.

965. Webster L, Hjelmstrom P, Sumner M, Gunderson EW. Efficacy and safety of a sublingual buprenorphine/naloxone rapidly dissolving tablet for the treatment of adults with opioid dependence: A randomized trial. J Addict Dis. 2016;35(4):325-38.

966. Webster L, Gruener D, Kirby T, Xiang Q, Tzanis E, Finn A. Evaluation of the Tolerability of Switching Patients on Chronic Full mu-Opioid Agonist Therapy to Buccal Buprenorphine. Pain Med. 2016;25:25.

967. Weiss RD, Potter JS, Fiellin DA, Byrne M, Connery HS, Dickinson W, et al. Adjunctive counseling during brief and extended buprenorphine-naloxone treatment for prescription opioid dependence: a 2-phase randomized controlled trial. Arch Gen Psychiatry. 2011;68(12):1238-46.

968. Weiss RD, Potter JS, Provost SE, Huang Z, Jacobs P, Hasson A, et al. A multi-site, two-phase, Prescription Opioid Addiction Treatment Study (POATS): rationale, design, and methodology. Contemp Clin Trials. 2010;31(2):189-99.

969. Weiss RD, Griffin ML, Jaffee WB, Bender RE, Graff FS, Gallop RJ, et al. A "community-friendly" version of integrated group therapy for patients with bipolar disorder and substance dependence: a randomized controlled trial. Drug Alcohol Depend. 2009;104(3):212-9.

970. Weiss RD, Potter JS, Griffin ML, McHugh RK, Haller D, Jacobs P, et al. Reasons for opioid use among patients with dependence on prescription opioids: the role of chronic pain. J Subst Abuse Treat. 2014;47(2):140-5.

971. Weiss RD, Potter JS, Griffin ML, Provost SE, Fitzmaurice GM, McDermott KA, et al. Long-term outcomes from the National Drug Abuse Treatment Clinical Trials Network Prescription Opioid Addiction Treatment Study. Drug Alcohol Depend. 2015;150:112-9.

972. Weiss RD, Rao V. The Prescription Opioid Addiction Treatment Study: What have we learned. Drug Alcohol Depend. 2017;173 Suppl 1:S48-S54.

973. Welsh JW, Tretyak V, McHugh RK, Weiss RD, Bogunovic O. Review: Adjunctive pharmacologic approaches for benzodiazepine tapers. Drug Alcohol Depend. 2018;189:96-107.

974. Westanmo A, Marshall P, Jones E, Burns K, Krebs EE. Opioid Dose Reduction in a VA Health Care System--Implementation of a Primary Care Population-Level Initiative. Pain Med. 2015;16(5):1019-26.

975. Westbury J, Tichelaar L, Peterson G, Gee P, Jackson S. A 12-month follow-up study of "RedUSe": a trial aimed at reducing antipsychotic and benzodiazepine use in nursing homes. Int Psychogeriatr. 2011;23(8):1260-9.

976. Whitehead PB. Interventions for the reduction of prescribed opioid use in chronic non‐cancer pain. Res Nurs Health. 2018;41(3):329-30.

References

977. Whiteside LK, Darnell D, Jackson K, Wang J, Russo J, Donovan DM, et al. Collaborative care from the emergency department for injured patients with prescription drug misuse: An open feasibility study. J Subst Abuse Treat. 2017;82:12.

978. Wightman R, Perrone J, Portelli I, Nelson L. Likeability and abuse liability of commonly prescribed opioids. Journal of medical toxicology : official journal of the American College of Medical Toxicology. 2012;8(4):335-40.

979. Willems IAT, Gorgels WJMJ, Oude Voshaar RC, Mulder J, Lucassen PLBJ. Tolerance to benzodiazepines among long-term users in primary care. Fam Pract. 2013;30(4):404-10.

980. Wilsey BL, Fishman S, Li CS, Storment J, Albanese A. Markers of abuse liability of short- vs long-acting opioids in chronic pain patients: a randomized cross-over trial. Pharmacol Biochem Behav. 2009;94(1):98-107. doi: 10.1016/j.pbb.2009.07.014. Epub Aug 4.

981. Wilson E, Lader M. A review of the management of antidepressant discontinuation symptoms. Therapeutic Advances in Psychopharmacology. 2015;5(6):357-68.

982. Wittchen HU, Buhringer G, Rehm JT, Klotsche J. The stability of 6-year course and outcome data about opiate maintenance treatment: a comparison of 6 to 7-year data. Suchtmedizin in forschung und praxis. 2011;13(5):269‐71.

983. Wobrock T, Soyka M. Pharmacotherapy of schizophrenia with comorbid substance use disorder--reviewing the evidence and clinical recommendations. Prog Neuropsychopharmacol Biol Psychiatry. 2008;32(6):1375-85.

984. Wobrock T, Soyka M. Pharmacotherapy of patients with schizophrenia and substance abuse. Expert Opin Pharmacother. 2009;10(3):353-67.

985. Wong E, Nguyen TV. Zolpidem Use in the Elderly and Recent Safety Data. The Journal for Nurse Practitioners. 2014;10(2):140-1.

986. Woody GE, Poole SA, Subramaniam G, Dugosh K, Bogenschutz M, Abbott P, et al. Extended vs short-term buprenorphine-naloxone for treatment of opioid-addicted youth: a randomized trial. JAMA. 2008;300(17):2003-11. doi: 10.1001/jama.2008.574.

987. Woody GE, Luborsky L, McLellan AT, O'Brien CP, Beck AT, Blaine J, et al. Psychotherapy for opiate addicts. Does it help? Arch Gen Psychiatry. 1983;40(6):639-45.

988. Worley MJ, Heinzerling KG, Shoptaw S, Ling W. Pain volatility and prescription opioid addiction treatment outcomes in patients with chronic pain. Exp Clin Psychopharmacol. 2015;23(6):428-35.

989. Worley MJ, Shoptaw SJ, Bickel WK, Ling W. Using behavioral economics to predict opioid use during prescription opioid dependence treatment. Drug Alcohol Depend. 2015;148:62-8.

990. Worley MJ, Heinzerling KG, Shoptaw S, Ling W. Volatility and change in chronic pain severity predict outcomes of treatment for prescription opioid addiction. Addiction. 2017;112(7):1202-9.

991. Wright A, Diebold J, Otal J, Stoneman C, Wong J, Wallace C, et al. The Effect of Melatonin on Benzodiazepine Discontinuation and Sleep Quality in Adults Attempting to Discontinue Benzodiazepines: A Systematic Review and Meta-Analysis. Drugs Aging. 2015;32(12):1009-18.

992. Wu CS, Ting TT, Wang SC, Chang IS, Lin KM. Effect of benzodiazepine discontinuation on dementia risk. Am J Geriatr Psychiatry. 2011;19(2):151-9.

References

993. Xiang YT, Weng YZ, Leung CM, Tang WK, Ungvari GS. Clinical and social determinants of long- term use of benzodiazepines and its impact on quality of life of Chinese schizophrenia patients. Pharmacopsychiatry. 2007;40(6):269‐74.

994. Xie L, Joshi AV, Schaaf D, Mardekian J, Harnett J, Shah ND, et al. Differences in Healthcare Utilization and Associated Costs Between Patients Prescribed vs. Nonprescribed Opioids During an Inpatient or Emergency Department Visit. Pain Practice. 2014;14(5):446-56.

995. Xue CC, Helme RD, Gibson S, Hogg M, Arnold C, Somogyi AA, et al. Effect of electroacupuncture on opioid consumption in patients with chronic musculoskeletal pain: protocol of a randomised controlled trial. Trials. 2012;13:169.

996. Yang L, Qian S, Liu H, Liu L, Pu C, Han P, et al. Role of Tramadol in Premature Ejaculation: A Systematic Review and Meta-Analysis. Urol Int. 2013;91(2):197-205.

997. Yang DD. Clinical observation of tranquilizing and regulating with benzodiazepine withdrawal syndrome. Chinese journal of information on traditional chinese medicine. 2006;13(10):60‐1.

998. Yarbrough CR. Prescription Drug Monitoring Programs Produce a Limited Impact on Painkiller Prescribing in Medicare Part D. Health Serv Res. 2018;53(2):671-89.

999. Yasui-Furukori N, Hashimoto K, Tsuchimine S, Tomita T, Sugawara N, Ishioka M, et al. Characteristics of Escitalopram Discontinuation Syndrome: A Preliminary Study. Clin Neuropharmacol. 2016;39(3):125-7.

1000. Yen C-F, Yen C-N, Ko C-H, Hwang T-J, Chen C-S, Chen T-T, et al. Correlates of Dependence and Beliefs About the Use of Hypnotics Among Zolpidem and Zopiclone Users. Subst Use Misuse. 2015;50(3):350.

1001. Yeo SH, Lim ZI, Mao J, Yau WP. Effects of Central Nervous System Drugs on Recovery After Stroke: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Clin Drug Investig. 2017;37(10):901-28.

1002. Yeung WF, Chung KF, Zhang ZJ, Chan WC, Zhang SP, Ng RM, et al. Electroacupuncture for tapering off long-term benzodiazepine use: study protocol of randomized controlled trial. BMC Complement Altern Med. 2017;17(1):183.

1003. Yovell Y, Bar G, Mashiah M, Baruch Y, Briskman I, Asherov J, et al. Ultra-Low-Dose Buprenorphine as a Time-Limited Treatment for Severe Suicidal Ideation: A Randomized Controlled Trial. Am J Psychiatry. 2016;173(5):491-8.

1004. Zacny JP, Gutierrez S. Within-subject comparison of the psychopharmacological profiles of oral hydrocodone and oxycodone combination products in non-drug-abusing volunteers. Drug Alcohol Depend. 2009;101(1-2):107-14. doi: 10.1016/j.drugalcdep.2008.11.013. Epub Dec 31.

1005. Zacny JP, Gutierrez S. Characterizing the subjective, psychomotor, and physiological effects of oral oxycodone in non-drug-abusing volunteers. Psychopharmacology (Berl). 2003;170(3):242-54.

1006. Zacny JP. Characterizing the subjective, psychomotor, and physiological effects of a hydrocodone combination product (Hycodan) in non-drug-abusing volunteers. Psychopharmacology (Berl). 2003;165(2):146-56. doi: 10.1007/s00213-002-1245-5. Epub 2002 Oct 29.

1007. Zacny JP, Lichtor SA. Within-subject comparison of the psychopharmacological profiles of oral oxycodone and oral morphine in non-drug-abusing volunteers. Psychopharmacology (Berl). 2008;196(1):105-16. doi: 10.1007/s00213-007-0937-2. Epub 2007 Sep 27.

References

1008. Zahradnik A, Otto C, Crackau B, Lohrmann I, Bischof G, John U, et al. Randomized controlled trial of a brief intervention for problematic prescription drug use in non-treatment-seeking patients. Addiction. 2009;104(1):109-17.

1009. Zajecka J, Fawcett J, Amsterdam J, Quitkin F, Reimherr F, Rosenbaum J, et al. Safety of abrupt discontinuation of fluoxetine: a randomized, placebo-controlled study. J Clin Psychopharmacol. 1998;18(3):193‐7.

1010. Zarghami M, Masoum B, Shiran MR. Tramadol versus methadone for treatment of opiate withdrawal: a double-blind, randomized, clinical trial. J Addict Dis. 2012;31(2):112-7.

1011. Zedler B, Xie L, Wang L, Joyce A, Vick C, Kariburyo F, et al. Risk factors for serious prescription opioid-related toxicity or overdose among Veterans Health Administration patients. Pain Med. 2014;15(11):1911-29.

1012. Zedler B, Xie L, Wang L, Joyce A, Vick C, Brigham J, et al. Development of a Risk Index for Serious Prescription Opioid-Induced Respiratory Depression or Overdose in Veterans' Health Administration Patients. Pain Medicine (United States). 2015;16(8):1566-79.

1013. Zgierska AE, Burzinski CA, Cox J, Kloke J, Stegner A, Cook DB, et al. Mindfulness Meditation and Cognitive Behavioral Therapy Intervention Reduces Pain Severity and Sensitivity in Opioid- Treated Chronic Low Back Pain: Pilot Findings from a Randomized Controlled Trial. Pain Med. 2016;17(10):1865-81.

1014. Zgierska AE, Burzinski CA, Cox J, Kloke J, Singles J, Mirgain S, et al. Mindfulness Meditation- Based Intervention Is Feasible, Acceptable, and Safe for Chronic Low Back Pain Requiring Long-Term Daily Opioid Therapy. J Altern Complement Med. 2016;22(8):610-20.

1015. Zhang HJ, Jiang XF, Ma MM, Zhang JW. A control study on treatment for benzodiazepine dependence with trazodone. Chinese journal of contemporary neurology and neurosurgery. 2013;13(5):411‐5.

1016. Zheng Z, Guo RJ, Helme RD, Muir A, Da Costa C, Xue CC. The effect of electroacupuncture on opioid-like medication consumption by chronic pain patients: a pilot randomized controlled clinical trial. Eur J Pain. 2008;12(5):671-6.

1017. Ziaaddini H, Nasirian M, Nakhaee N. Comparison of the efficacy of buprenorphine and clonidine in detoxification of opioid-dependents. Addict. 2012;4(3-4):79-86.

1018. Ziedonis DM, Amass L, Steinberg M, Woody G, Krejci J, Annon JJ, et al. Predictors of outcome for short-term medically supervised opioid withdrawal during a randomized, multicenter trial of buprenorphine-naloxone and clonidine in the NIDA clinical trials network drug and alcohol dependence. Drug Alcohol Depend. 2009;99(1-3):28-36.

1019. Zippel-Schultz B, Specka M, Cimander K, Eschenhagen T, Gölz J, Maryschok M, et al. Outcomes of Patients in Long-Term Opioid Maintenance Treatment. Subst Use Misuse. 2016;51(11):1493.

1020. Zisook S, Lesser IM, Lebowitz B, Rush AJ, Kallenberg G, Wisniewski S, et al. Effect of Antidepressant Medication Treatment on Suicidal Ideation and Behavior in a Randomized Trial: An Exploratory Report From the Combining Medications to Enhance Depression Outcomes Study. J Clin Psychiatry. 2011;72(10):1322-32.

1021. Zullich SG, Grasela TH, Jr., Fiedler-Kelly JB, Gengo FM. Impact of triplicate prescription program on psychotropic prescribing patterns in long-term care facilities. Ann Pharmacother. 1992;26(4):539-46.

References

1022. Zullig KJ, Lander LR, Sloan S, Brumage MR, Hobbs GR, Faulkenberry L. Mindfulness-based relapse prevention with individuals receiving medication-assisted outpatient treatment for opioid use disorder. Mindfulness. 2018;9(2):423-9.

Appendices 7 Appendices Appendix A: Review protocols

Table 26: Review protocol: Harms (PROSPERO Registration: CRD42018111310) What are the harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life /palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and Review question pregabalin (excluding epilepsy treatment), and antidepressants. Objectives To identify the physical, mental and social harms associated with dependence, discontinuation and withdrawal of prescription medications. Review population Adults using prescription medicines associated with dependence and withdrawal (see interventions below) Major age categories e.g. Adults ≥ 18 years adults Interventions and • Opioids for chronic pain comparators: • Benzodiazepines generic/class; • Z-drugs specific/drug • Gabapentinoids • Antidepressants

Compared to each other, placebo or usual care (as defined by studies). Non-comparative trials will also be included if no comparative data available.

If available, comparisons will be made of the harms from stopping medications rapidly compared to a longer tapering of treatment Outcomes Harms (as described by papers) Study design A hierarchy of evidence will be used: • Systematic review • RCT • Non-randomised comparative study • Non-randomised prospective cohort • Non-randomised retrospective cohort If a high quality systematic review is identified that matches the protocol, analysis will be incorporated, otherwise relevant systematic reviews will be used for citation searching only. Other exclusions Non-NHS prescribed medicines Medicines prescribed for end of life / palliative care / cancer pain. Over-the-counter medicines. Medicines dispensed/prescribed in prisons. Gabapentinoids prescribed for epilepsy. Antipsychotic and stimulant medicines. Medicines to treat drug misuse disorders. Review strategy Drugs will be pooled within classes (as per the grouping in ‘interventions’ above) If possible, results will be reported under four separate headings: • Harms of dependency on a medication.

Appendices

What are the harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life /palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and Review question pregabalin (excluding epilepsy treatment), and antidepressants. • Harms / side effects from stopping these medications over a short time frame. • Harms / side effects from longer term withdrawal (or clinical features of long term withdrawal). • Harms of misdiagnosis of withdrawal symptoms as existing or new disorder. Side effects of use of medications will not be listed, although noting that some of these may overlap with those relevant to the review.

The risk of bias of included studies will be assessed using the appropriate study design checklist according to the study type: Systematic review: ROBIS RCT: Cochrane RoB 2.0 Non-randomised study: Cochrane ROBINS-I Subgroup analyses if Antidepressants will be pooled as a class unless heterogeneity is observed. there is heterogeneity Gabapentin and pregabalin will be pooled in the analysis as ‘gabapentinoids’ unless heterogeneity is observed. Higher potency/shorter half-life and lower potency/longer half-life benzodiazepines will be pooled unless heterogeneity is observed. Socio-demographic groups will be analysed if differences are observed where it is possible to separate this data, e.g. age, sex, ethnicity, education level, income, area of residence etc.

Appendices

Appendix B: Study selection

Figure 2: Flow chart of study selection for the review

Records identified through database Additional records identified through searching, n=39087 other sources, n=249

Records screened, n=39336

Records excluded, n=38269 (113 requested but unobtainable) Full-text papers assessed for eligibility, n=1067

Papers included in review Papers excluded from full text sift, • Harms n=30 n=990 • Risk factors n=17 • Interventions n=28 Reasons for exclusion: see appendix G • Patients’ experience n=2 (10 additional identified from call for evidence, leading to 12 total) • Current practice n=0 (3 included identified from call for evidence, 1 from grey literature, leading to 4 total)

NB. One search was carried out to cover all 5 questions, therefore the PRISMA flowchart covers all 5. The final box of included studies details the number relevant to each question.

Appendices Appendix C: Search strategies

Systematic Literature search

The literature searches for these reviews are detailed below and complied with the methodology outlined in Developing NICE guidelines: the manual 2014, updated 2017 https://www.nice.org.uk/guidance/pmg20/resources/developing-nice-guidelines-the-manual-pdf- 72286708700869;

1. What are the harms associated with dependence, and the short term discontinuation and longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life /palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and pregabalin (excluding epilepsy treatment), and antidepressants? 2. What are the factors that contribute to the risk of harms associated with dependence and the short term discontinuation or longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life /palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and pregabalin (excluding epilepsy treatment), and antidepressants? 3. What are the most effective and cost effective approaches to the prevention and treatment of dependence and the short term discontinuation or longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life /palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and pregabalin (excluding epilepsy treatment), and antidepressants? 4. What is the current evidence about patients’ own experiences of the harms caused by prescribed medicines specifically relating to dependence and the short term discontinuation or longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life/palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and pregabalin (excluding epilepsy treatment), and antidepressants and experiences of accessing and engaging in treatment? 5. What are the current existing examples of services providing withdrawal support and what is the effectiveness and cost-effectiveness of the health/social service delivery models that prevent or treat dependence and the short term discontinuation or longer term withdrawal symptoms from the following prescribed medicines: opioids for chronic pain (excluding end of life/palliative care/cancer pain), benzodiazepines, Z-drugs, gabapentin and pregabalin (excluding epilepsy treatment), and antidepressants? (In England, as well as health service delivery models in other countries that might inform provision in England).

For more detailed information, please see the Methodology Review.

Appendices

C.1 Databases date parameters and filters used

C.1.1 Step 1: Existing systematic reviews Database Dates searched Search filter used Cochrane Database of Systematic All years to Issue 9 of 12, None Reviews (The Cochrane Library - September 2018 Wiley)

Epistemonikos All years to 24 September 2018 Systematic review studies

Database of promoting health All years to 19 September 2018 None effectiveness reviews (DoPHER)

HealthEvidence All years to 19 September 2018 None

Cochrane Library (Wiley) search terms #1. MeSH descriptor: [Substance-Related Disorders] explode all trees #2. MeSH descriptor: [Substance Withdrawal Syndrome] explode all trees #3. MeSH descriptor: [Inappropriate Prescribing] explode all trees #4. MeSH descriptor: [Medical Overuse] explode all trees #5. MeSH descriptor: [Deprescriptions] explode all trees #6. (abstinen* or abstain* or cessat* or detox* or discontinu* or reduc* or stop* or taper* or withdraw* or substitut* or depend* or addict* or abuse* or abusing or chronic or long* term or longterm or short* term or short term or misus* or overus* or deprescrib*):ti,ab #7. (over* near/3 use* or using or utlisat* or utilizat*) near/3 (prescription* or prescrib* or drug* or medicine* or medication* or pharm*):ti,ab #8. inappropriate near/3 (prescription or prescrib*):ti,ab #9. (OR #1-#8) #10. MeSH descriptor: [Narcotics] explode all trees #11. MeSH descriptor: [Analgesics, Opioid] explode all trees #12. (analgesic* near/3 opioid* or narcotic near/3 agent*):ti,ab #13. MeSH descriptor: [Buprenorphine] explode all trees #14. MeSH descriptor: [Codeine] explode all trees #15. MeSH descriptor: [Dextromoramide] explode all trees #16. MeSH descriptor: [Heroin] explode all trees #17. MeSH descriptor: [Fentanyl] explode all trees #18. MeSH descriptor: [Hydromorphone] explode all trees #19. MeSH descriptor: [Meptazinol] explode all trees #20. MeSH descriptor: [Methadone] explode all trees #21. MeSH descriptor: [Morphine] explode all trees #22. MeSH descriptor: [Oxycodone] explode all trees #23. MeSH descriptor: [Opium] explode all trees

Appendices

#24. MeSH descriptor: [Pentazocine] explode all trees #25. MeSH descriptor: [Meperidine] explode all trees #26. MeSH descriptor: [Tramadol] explode all trees #27. (buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin):ti,ab #28. (z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem):ti,ab #29. (generation near/3 hypnotic*):ti,ab #30. MeSH descriptor: [Benzodiazepines] explode all trees #31. (benzodiazepin* or bzd or flurazepam or loprazolam or lormetazepam or nitrazepam or temazepam or diazepam or chlordiazepoxide or lorazepam or oxazepam):ti,ab #32. MeSH descriptor: [Pregabalin] explode all trees #33. (gabapentin* or pregabalin*):ti,ab #34. MeSH descriptor: [Antidepressive Agents] explode all trees #35. (antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or MAOI* or "monoamine oxidase inhibit*" or RIMA* or tricyclic* or SSRI* or SNRI* or SNORI*):ti,ab #36. MeSH descriptor: [Amitriptyline] explode all trees #37. MeSH descriptor: [Amoxapine] explode all trees #38. MeSH descriptor: [Clomipramine] explode all trees #39. MeSH descriptor: [Dothiepin] explode all trees #40. MeSH descriptor: [Doxepin] explode all trees #41. MeSH descriptor: [Imipramine] explode all trees #42. MeSH descriptor: [Lofepramine] explode all trees #43. MeSH descriptor: [Maprotiline] explode all trees #44. MeSH descriptor: [Mianserin] explode all trees #45. MeSH descriptor: [Nortriptyline] explode all trees #46. MeSH descriptor: [Protriptyline] explode all trees #47. MeSH descriptor: [Trazodone] explode all trees #48. MeSH descriptor: [Trimipramine] explode all trees #49. MeSH descriptor: [Isocarboxazid] explode all trees #50. MeSH descriptor: [Moclobemide] explode all trees #51. MeSH descriptor: [Phenelzine] explode all trees #52. MeSH descriptor: [Tranylcypromine] explode all trees #53. MeSH descriptor: [Citalopram] explode all trees #54. MeSH descriptor: [Fluoxetine] explode all trees #55. MeSH descriptor: [Fluvoxamine] explode all trees #56. MeSH descriptor: [Paroxetine] explode all trees #57. MeSH descriptor: [Sertraline] explode all trees #58. MeSH descriptor: [5-Hydroxytryptophan] explode all trees #59. MeSH descriptor: [Duloxetine Hydrochloride] explode all trees #60. MeSH descriptor: [Flupenthixol] explode all trees #61. MeSH descriptor: [Tryptophan] explode all trees #62. MeSH descriptor: [Venlafaxine Hydrochloride] explode all trees #63. (amitriptyline or amoxapine or clomipramine or dosulepin or doxepin or imipramine or lofepramine or maprotiline or mianserin or nortriptyline or protriptyline or trazodone or

Appendices

trimipramine or isocarboxazid or moclobemide or phenelzine or tranylcypromine or citalopram or escitalopram or fluoxetine or fluvoxamine or paroxetine or sertraline or agomelatine or duloxetine or flupentixol or mirtazapine or nefazodone or oxitriptan or reboxetine or tryptophan or venlafaxine or vortioxetine):ti,ab #64. (OR #10-#63) #65. #9 and #64

Epistemonikos search terms 1. "substance-related disorders" OR "substance withdrawal syndrome" OR "inappropriate prescribing" OR "medical overuse" OR "deprescriptions" OR "abstinen*" or "abstain*" or "cessat*" or "detox*" or "discontinu*" or "reduc*" or "stop*" or "taper*" or "withdraw*" or "substitut*" or "depend*" or "addict*" or "abuse*" or "abusing" or "chronic" or "long* term" or "longterm" or "short* term" or "short term" or "misus*" or "overus*" OR "deprescrib*" OR "inappropriate prescription" 2. "buprenorphine*" or "codeine*" or "dextromoramide*" or "diamorphine*" or "dihydrocodeine*" or "dipipanone*" or "fentanyl" or "hydromorphone*" or "meptazinol" or "methadone*" or "morphine*" or "oxycodone" or "papaveretum" or "pentazocine" or "pethidine" or "tapentadol" or "tramadol" or "heroin" OR "z drug*" or "z hypnotic*" or "nonbenzodiazepin*" or "nonbenzodiazepin*" or "zaleplon" or "zopiclone" or "zolpidem" OR "generation hypnotic" OR "benzodiazepin*" or "bzd" or "flurazepam" or "loprazolam" or "lormetazepam" or "nitrazepam" or "temazepam" or "diazepam" or "chlordiazepoxide" or "lorazepam" or "oxazepam" OR "gabapentin*" or "pregabalin*" OR "antidepress*" or "anti depress*" or "thymoanaleptic*" or "thymoleptic*" or "MAOI*" or "monoamine oxidase inhibit*" or "RIMA*" or "tricyclic*" or "SSRI*" or "SNRI*" or "SNORI*" OR "amitriptyline" or "amoxapine" or "clomipramine" or "dosulepin" or "doxepin" or "imipramine" or "lofepramine" or "maprotiline" or "mianserin" or "nortriptyline" or "protriptyline" or "trazodone" or "trimipramine" or "isocarboxazid" or "moclobemide" or "phenelzine" or "tranylcypromine" or "citalopram" or "escitalopram" or "fluoxetine" or "fluvoxamine" or "paroxetine" or "sertraline" or "agomelatine" or "duloxetine" or "flupentixol" or "mirtazapine" or "nefazodone" or "oxitriptan" or "reboxetine" or "tryptophan" or "venlafaxine" or "vortioxetine" 3. (title:("substance-related disorders" OR "substance withdrawal syndrome" OR "inappropriate prescribing" OR "medical overuse" OR "deprescriptions" OR "abstinen*" OR "abstain*" OR "cessat*" OR "detox*" OR "discontinu*" OR "reduc*" OR "stop*" OR "taper*" OR "withdraw*" OR "substitut*" OR "depend*" OR "addict*" OR "abuse*" OR "abusing" OR "chronic" OR "long* term" OR "longterm" OR "short* term" OR "short term" OR "misus*" OR "overus*" OR "deprescrib*" OR "inappropriate prescription") OR abstract:("substance-related disorders" OR "substance withdrawal syndrome" OR "inappropriate prescribing" OR "medical overuse" OR "deprescriptions" OR "abstinen*" OR "abstain*" OR "cessat*" OR "detox*" OR "discontinu*" OR "reduc*" OR "stop*" OR "taper*" OR "withdraw*" OR "substitut*" OR "depend*" OR "addict*" OR "abuse*" OR "abusing" OR "chronic" OR "long* term" OR "longterm" OR "short* term" OR "short term" OR "misus*" OR "overus*" OR "deprescrib*" OR "inappropriate prescription")) AND (title:("buprenorphine*" OR "codeine*" OR "dextromoramide*" OR "diamorphine*" OR "dihydrocodeine*" OR "dipipanone*" OR "fentanyl" OR "hydromorphone*" OR "meptazinol" OR "methadone*" OR "morphine*" OR "oxycodone" OR "papaveretum" OR "pentazocine" OR "pethidine" OR "tapentadol" OR "tramadol" OR "heroin" OR "z drug*" OR "z hypnotic*" OR "non-benzodiazepin*" OR "nonbenzodiazepin*" OR "zaleplon" OR "zopiclone" OR "zolpidem" OR "generation hypnotic" OR "benzodiazepin*" OR "bzd" OR "flurazepam" OR "loprazolam" OR "lormetazepam" OR "nitrazepam" OR "temazepam" OR "diazepam" OR "chlordiazepoxide" OR "lorazepam" OR "oxazepam" OR "gabapentin*" OR "pregabalin*" OR "antidepress*" OR "anti depress*" OR "thymoanaleptic*" OR "thymoleptic*" OR "MAOI*" OR "monoamine oxidase inhibit*" OR "RIMA*" OR "tricyclic*" OR "SSRI*" OR "SNRI*" OR "SNORI*" OR "amitriptyline" OR "amoxapine" OR "clomipramine" OR "dosulepin" OR "doxepin" OR "imipramine" OR "lofepramine" OR "maprotiline" OR "mianserin" OR "nortriptyline" OR "protriptyline" OR "trazodone" OR "trimipramine" OR "isocarboxazid" OR "moclobemide" OR "phenelzine" OR "tranylcypromine" OR "citalopram"

Appendices

OR "escitalopram" OR "fluoxetine" OR "fluvoxamine" OR "paroxetine" OR "sertraline" OR "agomelatine" OR "duloxetine" OR "flupentixol" OR "mirtazapine" OR "nefazodone" OR "oxitriptan" OR "reboxetine" OR "tryptophan" OR "venlafaxine" OR "vortioxetine") OR abstract:("buprenorphine*" OR "codeine*" OR "dextromoramide*" OR "diamorphine*" OR "dihydrocodeine*" OR "dipipanone*" OR "fentanyl" OR "hydromorphone*" OR "meptazinol" OR "methadone*" OR "morphine*" OR "oxycodone" OR "papaveretum" OR "pentazocine" OR "pethidine" OR "tapentadol" OR "tramadol" OR "heroin" OR "z drug*" OR "z hypnotic*" OR "non-benzodiazepin*" OR "nonbenzodiazepin*" OR "zaleplon" OR "zopiclone" OR "zolpidem" OR "generation hypnotic" OR "benzodiazepin*" OR "bzd" OR "flurazepam" OR "loprazolam" OR "lormetazepam" OR "nitrazepam" OR "temazepam" OR "diazepam" OR "chlordiazepoxide" OR "lorazepam" OR "oxazepam" OR "gabapentin*" OR "pregabalin*" OR "antidepress*" OR "anti depress*" OR "thymoanaleptic*" OR "thymoleptic*" OR "MAOI*" OR "monoamine oxidase inhibit*" OR "RIMA*" OR "tricyclic*" OR "SSRI*" OR "SNRI*" OR "SNORI*" OR "amitriptyline" OR "amoxapine" OR "clomipramine" OR "dosulepin" OR "doxepin" OR "imipramine" OR "lofepramine" OR "maprotiline" OR "mianserin" OR "nortriptyline" OR "protriptyline" OR "trazodone" OR "trimipramine" OR "isocarboxazid" OR "moclobemide" OR "phenelzine" OR "tranylcypromine" OR "citalopram" OR "escitalopram" OR "fluoxetine" OR "fluvoxamine" OR "paroxetine" OR "sertraline" OR "agomelatine" OR "duloxetine" OR "flupentixol" OR "mirtazapine" OR "nefazodone" OR "oxitriptan" OR "reboxetine" OR "tryptophan" OR "venlafaxine" OR "vortioxetine")) 4. Limit 3 to systematic reviews

Database of promoting health effectiveness reviews (DoPHER) search terms 1. Freetext (All but Authors): "substance-related disorders" OR "substance withdrawal syndrome" OR "inappropriate prescribing" OR "medical overuse" OR "deprescriptions" OR "abstinen*" or "abstain*" or "cessat*" or "detox*" or "discontinu*" or "reduc*" or "stop*" or "taper*" or "withdraw*" or "substitut*" or "depend*" or "addict*" or "abuse*" or "abusing" or "chronic" or "long* term" or "longterm" or "short* term" or "short term" or "misus*" or "overus*" OR "deprescrib*" 2. Freetext (All but Authors): "over*" near "use*" near "prescri*" 3. Freetext (All but Authors):"inappropriate" near "prescri*" 4. 1 or 2 or 3 5. Freetext (All but Authors): "buprenorphine*" or "codeine*" or "dextromoramide*" or "diamorphine*" or "dihydrocodeine*" or "dipipanone*" or "fentanyl" or "hydromorphone*" or "meptazinol" or "methadone*" or "morphine*" or "oxycodone" or "papaveretum" or "pentazocine" or "pethidine" or "tapentadol" or "tramadol" or "heroin" 6. Freetext (All but Authors): "z drug*" or "z hypnotic*" or "non-benzodiazepin*" or "nonbenzodiazepin*" or "zaleplon" or "zopiclone" or "zolpidem" 7. Freetext (All but Authors): "generation" near "hypnotic" 8. Freetext (All but Authors): "benzodiazepin*" or "bzd" or "flurazepam" or "loprazolam" or "lormetazepam" or "nitrazepam" or "temazepam" or "diazepam" or "chlordiazepoxide" or "lorazepam" or "oxazepam" 9. Freetext (All but Authors): "gabapentin*" or "pregabalin*" 10. Freetext (All but Authors): "antidepress*" or "anti depress*" or "thymoanaleptic*" or "thymoleptic*" or "MAOI*" or "monoamine oxidase inhibit*" or "RIMA*" or "tricyclic*" or "SSRI*" or "SNRI*" or "SNORI*" 11. Freetext (All but Authors): "amitriptyline" or "amoxapine" or "clomipramine" or "dosulepin" or "doxepin" or "imipramine" or "lofepramine" or "maprotiline" or "mianserin" or "nortriptyline" or "protriptyline" or "trazodone" or "trimipramine" or "isocarboxazid" or "moclobemide" or "phenelzine" or "tranylcypromine" or "citalopram" or "escitalopram" or "fluoxetine" or "fluvoxamine" or "paroxetine" or "sertraline" or "agomelatine" or "duloxetine" or "flupentixol" or "mirtazapine" or "nefazodone" or "oxitriptan" or "reboxetine" or "tryptophan" or

Appendices

"venlafaxine" or "vortioxetine" 12. 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 13. 4 AND 12

HealthEvidence search terms 1. [((substance-related disorders or substance withdrawal syndrome or inappropriate prescribing OR medical overuse OR deprescriptions or abstinen* or abstain* or cessat* or detox* or discontinu* or reduc* or stop* or taper* or withdraw* or substitut* or depend* or addict* or abuse* or abusing or chronic or long* term or longterm or short* term or short term or misus* or overus* or deprescrib*)) AND ((narcotics or analgesics or buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin or z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem or benzodiazepin* or bzd or flurazepam or loprazolam or lormetazepam or nitrazepam or temazepam or diazepam or chlordiazepoxide or lorazepam or oxazepam or gabapentin* or pregabalin* or antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or MAOI* or "monoamine oxidase inhibit*" or RIMA* or tricyclic* or SSRI* or SNRI* or SNORI*or amitriptyline or amoxapine or clomipramine or dosulepin or doxepin or imipramine or lofepramine or maprotiline or mianserin or nortriptyline or protriptyline or trazodone or trimipramine or isocarboxazid or moclobemide or phenelzine or tranylcypromine or citalopram or escitalopram or fluoxetine or fluvoxamine or paroxetine or sertraline or agomelatine or duloxetine or flupentixol or mirtazapine or nefazodone or oxitriptan or reboxetine or tryptophan or venlafaxine or vortioxetine))]

C.1.2 Step 2: Recent evidence Database Dates searched Search filter used Medline (OVID) 1st January 2008 – 3 October 2018 Exclusions Randomised controlled trials Systematic review studies

Embase (OVID) 1st January 2008 – 3 October 2018 Exclusions Randomised controlled trials Systematic review studies

Cochrane Central Register of 1st January 2008 to Issue 8 of 12, None Controlled Trials (CENTRAL) (The August 2018 Cochrane Library -Wiley)

PsycINFO (ProQuest) 1st January 2008 – 3 October 2018 Exclusions Randomised controlled trials Systematic review studies

Health Technology Appraisals 1st January 2008 – 3 October 2018 None (Centre for Reviews and Dissemination) Trials Register of Promoting All years to 3 October 2018 None Health Interventions (TRoPHI)

Appendices

Database Dates searched Search filter used ASSIA (Proquest) 1st January 2008 – 3 October 2018 None

C.1.2.1 Medline (Ovid) search terms 1. exp substance-related disorders/ 2. exp substance withdrawal syndrome/ 3. exp inappropriate prescribing/ 4. exp medical overuse/ 5. exp deprescriptions/ 6. (abstinen* or abstain* or cessat* or detox* or discontinu* or reduc* or stop* or taper* or withdraw* or substitut* or depend* or addict* or abuse* or abusing or chronic or long* term or longterm or short* term or short term or misus* or overus* or deprescrib*).ti,ab. 7. (over* adj3 (use* or using or utlisat* or utilizat*) adj3 (prescription* or prescrib* or drug* or medicine* or medication* or pharm*)).ti,ab. 8. (inappropriate adj3 (prescription or prescrib*)).ti,ab. 9. or/1-8 10. letter/ 11. editorial/ 12. news/ 13. exp historical article/ 14. anecdotes as topic/ 15. comment/ 16. case report/ 17. (letter or comment*).ti. 18. or/10-17 19. randomized controlled trial/ or random*.ti,ab. 20. 18 not 19 21. animals/ not humans/ 22. exp animals, laboratory/ 23. exp animal experimentation/ 24. exp models, animal/ 25. exp rodentia/ 26. (rat or rats or mouse or mice).ti. 27. or/20-26 28. 9 not 27 29. limit 28 to English language 30. exp narcotics/ 31. exp analgesics, opioid/ 32. (analgesic* adj3 (opioid* or narcotic) adj3 agent*).ti,ab. 33. exp buprenorphine/ 34. exp codeine/ 35. exp dextromoramide/ 36. exp heroin/

Appendices

37. exp fentanyl/ 38. exp hydromorphone/ 39. exp meptazinol/ 40. exp methadone/ 41. exp morphine/ 42. exp oxycodone/ 43. exp opium/ 44. exp pentazocine/ 45. exp meperidine/ 46. exp tramadol/ 47. (buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin).ti,ab. 48. (z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem).ti,ab. 49. (generation adj3 hypnotic*).ti,ab. 50. exp benzodiazepines/ 51. (benzodiazepin* or bzd or flurazepam or loprazolam or lormetazepam or nitrazepam or temazepam or diazepam or chlordiazepoxide or lorazepam or oxazepam).ti,ab. 52. exp pregabalin/ 53. (gabapentin* or pregabalin*).ti,ab. 54. exp antidepressive agents/ 55. (antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or maoi* or "monoamine oxidase inhibit*" or rima* or tricyclic* or ssri* or snri* or snori*).ti,ab. 56. exp amitriptyline/ 57. exp amoxapine/ 58. exp clomipramine/ 59. exp dothiepin/ 60. exp doxepin/ 61. exp imipramine/ 62. exp lofepramine/ 63. exp maprotiline/ 64. exp mianserin/ 65. exp nortriptyline/ 66. exp protriptyline/ 67. exp trazodone/ 68. exp trimipramine/ 69. exp isocarboxazid/ 70. exp moclobemide/ 71. exp phenelzine/ 72. exp tranylcypromine/ 73. exp citalopram/ 74. exp fluoxetine/ 75. exp fluvoxamine/ 76. exp paroxetine/

Appendices

77. exp sertraline/ 78. exp 5-hydroxytryptophan/ 79. exp duloxetine hydrochloride/ 80. exp flupenthixol/ 81. exp tryptophan/ 82. exp venlafaxine hydrochloride/ 83. (amitriptyline or amoxapine or clomipramine or dosulepin or doxepin or imipramine or lofepramine or maprotiline or mianserin or nortriptyline or protriptyline or trazodone or trimipramine or isocarboxazid or moclobemide or phenelzine or tranylcypromine or citalopram or escitalopram or fluoxetine or fluvoxamine or paroxetine or sertraline or agomelatine or duloxetine or flupentixol or mirtazapine or nefazodone or oxitriptan or reboxetine or tryptophan or venlafaxine or vortioxetine).ti,ab. 84. or/30-83 85. 29 and 84 86. randomized controlled trial.pt. 87. controlled clinical trial.pt. 88. randomi#ed.ab. 89. placebo.ab. 90. randomly.ab. 91. clinical trials as topic.sh. 92. trial.ti. 93. or/86-92 94. meta-analysis/ 95. meta-analysis as topic/ 96. (meta analy* or metanaly* or metaanaly* or meta regression).ti,ab. 97. ((systematic* or evidence*) adj2 (review* or overview*)).ti,ab. 98. (reference list* or bibliograph* or hand search* or manual search* or relevant journals).ab. 99. (search strategy or search criteria or systematic search or study selection or data extraction).ab. 100. (search* adj4 literature).ab. 101. (medline or pubmed or cochrane or embase or psychlit or psyclit or psychinfo or psycinfo or cinahl or science citation index or bids or cancerlit).ab. 102. cochrane.jw. 103. ((multiple treatment* or indirect or mixed) adj2 comparison*).ti,ab. 104. or/94-103 105. 85 and (93 or 104)

C.1.2.2 Embase (Ovid) search terms 1. *drug dependence/ 2. *withdrawal syndrome/ or *alcohol withdrawal syndrome/ or *neonatal abstinence syndrome/ 3. *inappropriate prescribing/ 4. *deprescription/ 5. (abstinen* or abstain* or cessat* or detox* or discontinu* or reduc* or stop* or taper* or withdraw* or substitut* or depend* or addict* or abuse* or abusing or chronic or long* term or longterm or short* term or short term or misus* or overus* or deprescrib*).ti,ab.

Appendices

6. (over* adj3 (use* or using or utlisat* or utilizat*) adj3 (prescription* or prescrib* or drug* or medicine* or medication* or pharm*)).ti,ab. 7. (inappropriate adj3 (prescription or prescrib*)).ti,ab. 8. or/1-7 9. letter.pt. or letter/ 10. note.pt. 11. editorial.pt. 12. case report/ or case study/ 13. (letter or comment*).ti. 14. or/9-13 15. randomized controlled trial/ or random*.ti,ab. 16. 14 not 15 17. animal/ not human/ 18. nonhuman/ 19. exp Animal Experiment/ 20. exp Experimental Animal/ 21. animal model/ 22. exp Rodent/ 23. (rat or rats or mouse or mice).ti. 24. or/16-23 25. 8 not 24 26. limit 25 to English language 27. *narcotic agent/ 28. *narcotic analgesic agent/ 29. (analgesic* adj3 (opioid* or narcotic) adj3 agent*).ti,ab. 30. *buprenorphine/ 31. *codeine/ 32. *dextromoramide/ 33. *diamorphine/ 34. *fentanyl/ 35. *hydromorphone/ 36. *meptazinol/ 37. *methadone/ 38. *morphine/ 39. *oxycodone/ 40. *opiate/ 41. *pentazocine/ 42. *pethidine/ 43. *tramadol/ 44. (buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin).ti,ab. 45. (z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem).ti,ab. 46. (generation adj3 hypnotic*).ti,ab.

Appendices

47. *benzodiazepine derivative/ 48. (benzodiazepin* or bzd or flurazepam or loprazolam or lormetazepam or nitrazepam or temazepam or diazepam or chlordiazepoxide or lorazepam or oxazepam).ti,ab. 49. *pregabalin/ 50. (gabapentin* or pregabalin*).ti,ab. 51. *antidepressant agent/ 52. (antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or MAOI* or "monoamine oxidase inhibit*" or RIMA* or tricyclic* or SSRI* or SNRI* or SNORI*).ti,ab. 53. *amitriptyline/ 54. *amoxapine/ 55. *clomipramine/ 56. *dosulepin/ 57. *doxepin/ 58. *imipramine/ 59. *lofepramine/ 60. *maprotiline/ 61. *mianserin/ 62. *nortriptyline/ 63. *protriptyline/ 64. *trazodone/ 65. *trimipramine/ 66. *isocarboxazid/ 67. *moclobemide/ 68. *phenelzine/ 69. *tranylcypromine/ 70. *citalopram/ 71. *fluoxetine/ 72. *fluvoxamine/ 73. *paroxetine/ 74. *sertraline/ 75. *5 hydroxytryptophan/ 76. *duloxetine/ 77. *flupentixol/ 78. *tryptophan/ 79. *venlafaxine/ 80. (amitriptyline or amoxapine or clomipramine or dosulepin or doxepin or imipramine or lofepramine or maprotiline or mianserin or nortriptyline or protriptyline or trazodone or trimipramine or isocarboxazid or moclobemide or phenelzine or tranylcypromine or citalopram or escitalopram or fluoxetine or fluvoxamine or paroxetine or sertraline or agomelatine or duloxetine or flupentixol or mirtazapine or nefazodone or oxitriptan or reboxetine or tryptophan or venlafaxine or vortioxetine).ti,ab. 81. or/27-80 82. 26 and 81 83. random*.ti,ab. 84. factorial*.ti,ab. 85. (crossover* or cross over*).ti,ab. 86. ((doubl* or singl*) adj blind*).ti,ab.

Appendices

87. (assign* or allocat* or volunteer* or placebo*).ti,ab. 88. crossover procedure/ 89. single blind procedure/ 90. randomized controlled trial/ 91. double blind procedure/ 92. or/83-91 93. systematic review/ 94. Meta-Analysis/ 95. (meta analy* or metanaly* or metaanaly* or meta regression).ti,ab. 96. ((systematic* or evidence*) adj2 (review* or overview*)).ti,ab. 97. (reference list* or bibliograph* or hand search* or manual search* or relevant journals).ab. 98. (search strategy or search criteria or systematic search or study selection or data extraction).ab. 99. (search* adj4 literature).ab. 100. (medline or pubmed or cochrane or embase or psychlit or psyclit or psychinfo or psycinfo or cinahl or science citation index or bids or cancerlit).ab. 101. cochrane.jw. 102. ((multiple treatment* or indirect or mixed) adj2 comparison*).ti,ab. 103. or/93-102 104. 82 and (92 or 103)

C.1.2.3 Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library -Wiley) search terms #1. MeSH descriptor: [Substance-Related Disorders] explode all trees #2. MeSH descriptor: [Substance Withdrawal Syndrome] explode all trees #3. MeSH descriptor: [Inappropriate Prescribing] explode all trees #4. MeSH descriptor: [Medical Overuse] explode all trees #5. MeSH descriptor: [Deprescriptions] explode all trees #6. (abstinen* or abstain* or cessat* or detox* or discontinu* or reduc* or stop* or taper* or withdraw* or substitut* or depend* or addict* or abuse* or abusing or chronic or long* term or longterm or short* term or short term or misus* or overus* or deprescrib*):ti,ab #7. (over* near/3 use* or using or utlisat* or utilizat*) near/3 (prescription* or prescrib* or drug* or medicine* or medication* or pharm*):ti,ab #8. inappropriate near/3 (prescription or prescrib*):ti,ab #9. (OR #1-#8) #10. MeSH descriptor: [Narcotics] explode all trees #11. MeSH descriptor: [Analgesics, Opioid] explode all trees #12. (analgesic* near/3 opioid* or narcotic near/3 agent*):ti,ab #13. MeSH descriptor: [Buprenorphine] explode all trees #14. MeSH descriptor: [Codeine] explode all trees #15. MeSH descriptor: [Dextromoramide] explode all trees #16. MeSH descriptor: [Heroin] explode all trees #17. MeSH descriptor: [Fentanyl] explode all trees #18. MeSH descriptor: [Hydromorphone] explode all trees #19. MeSH descriptor: [Meptazinol] explode all trees #20. MeSH descriptor: [Methadone] explode all trees

Appendices

#21. MeSH descriptor: [Morphine] explode all trees #22. MeSH descriptor: [Oxycodone] explode all trees #23. MeSH descriptor: [Opium] explode all trees #24. MeSH descriptor: [Pentazocine] explode all trees #25. MeSH descriptor: [Meperidine] explode all trees #26. MeSH descriptor: [Tramadol] explode all trees #27. (buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin):ti,ab #28. (z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem):ti,ab #29. (generation near/3 hypnotic*):ti,ab #30. MeSH descriptor: [Benzodiazepines] explode all trees #31. (benzodiazepin* or bzd or flurazepam or loprazolam or lormetazepam or nitrazepam or temazepam or diazepam or chlordiazepoxide or lorazepam or oxazepam):ti,ab #32. MeSH descriptor: [Pregabalin] explode all trees #33. (gabapentin* or pregabalin*):ti,ab #34. MeSH descriptor: [Antidepressive Agents] explode all trees #35. (antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or MAOI* or "monoamine oxidase inhibit*" or RIMA* or tricyclic* or SSRI* or SNRI* or SNORI*):ti,ab #36. MeSH descriptor: [Amitriptyline] explode all trees #37. MeSH descriptor: [Amoxapine] explode all trees #38. MeSH descriptor: [Clomipramine] explode all trees #39. MeSH descriptor: [Dothiepin] explode all trees #40. MeSH descriptor: [Doxepin] explode all trees #41. MeSH descriptor: [Imipramine] explode all trees #42. MeSH descriptor: [Lofepramine] explode all trees #43. MeSH descriptor: [Maprotiline] explode all trees #44. MeSH descriptor: [Mianserin] explode all trees #45. MeSH descriptor: [Nortriptyline] explode all trees #46. MeSH descriptor: [Protriptyline] explode all trees #47. MeSH descriptor: [Trazodone] explode all trees #48. MeSH descriptor: [Trimipramine] explode all trees #49. MeSH descriptor: [Isocarboxazid] explode all trees #50. MeSH descriptor: [Moclobemide] explode all trees #51. MeSH descriptor: [Phenelzine] explode all trees #52. MeSH descriptor: [Tranylcypromine] explode all trees #53. MeSH descriptor: [Citalopram] explode all trees #54. MeSH descriptor: [Fluoxetine] explode all trees #55. MeSH descriptor: [Fluvoxamine] explode all trees #56. MeSH descriptor: [Paroxetine] explode all trees #57. MeSH descriptor: [Sertraline] explode all trees #58. MeSH descriptor: [5-Hydroxytryptophan] explode all trees #59. MeSH descriptor: [Duloxetine Hydrochloride] explode all trees #60. MeSH descriptor: [Flupenthixol] explode all trees

Appendices

#61. MeSH descriptor: [Tryptophan] explode all trees #62. MeSH descriptor: [Venlafaxine Hydrochloride] explode all trees #63. (amitriptyline or amoxapine or clomipramine or dosulepin or doxepin or imipramine or lofepramine or maprotiline or mianserin or nortriptyline or protriptyline or trazodone or trimipramine or isocarboxazid or moclobemide or phenelzine or tranylcypromine or citalopram or escitalopram or fluoxetine or fluvoxamine or paroxetine or sertraline or agomelatine or duloxetine or flupentixol or mirtazapine or nefazodone or oxitriptan or reboxetine or tryptophan or venlafaxine or vortioxetine):ti,ab #64. (OR #10-#63) #65. #9 and #64

C.1.2.4 PsycINFO (ProQuest) search terms 1. (((((MAINSUBJECT.EXACT("Drug Withdrawal") OR MAINSUBJECT.EXACT("Substance Use Disorder")) OR ti,ab(abstinen* OR abstain* OR cessat* OR detox* OR discontinu* OR reduc* OR stop* OR taper* OR withdraw* OR substitut* OR depend* OR addict* OR abuse* OR abusing OR chronic OR long* term OR longterm OR short* term OR short term OR misus* OR overus* OR deprescrib*) OR ti,ab(over* NEAR/3 (use* OR using OR utlisat* OR utilizat*) NEAR/3 (prescription* OR prescrib* OR drug* OR medicine* OR medication* OR pharm*)) OR ti,ab(inappropriate NEAR/3 (prescription OR prescrib*))) AND ((MAINSUBJECT.EXACT.EXPLODE("Analgesic Drugs") OR MAINSUBJECT.EXACT.EXPLODE("Narcotic Drugs")) OR ti,ab(analgesic* NEAR/3 (opioid* OR narcotic) NEAR/3 agent*) OR (MAINSUBJECT.EXACT.EXPLODE("Buprenorphine") OR MAINSUBJECT.EXACT.EXPLODE("Heroin") OR MAINSUBJECT.EXACT.EXPLODE("Methadone") OR MAINSUBJECT.EXACT.EXPLODE("Pentazocine") OR MAINSUBJECT.EXACT.EXPLODE("Morphine") OR MAINSUBJECT.EXACT.EXPLODE("Tramadol") OR MAINSUBJECT.EXACT.EXPLODE("Codeine") OR MAINSUBJECT.EXACT.EXPLODE("Fentanyl") OR MAINSUBJECT.EXACT.EXPLODE("Meperidine")) OR ti,ab(buprenorphine* OR codeine* OR dextromoramide* OR diamorphine* OR dihydrocodeine* OR dipipanone* OR fentanyl OR hydromorphone* OR meptazinol OR methadone* OR morphine* OR oxycodone OR papaveretum OR pentazocine OR pethidine OR tapentadol OR tramadol OR heroin) OR ti,ab(z drug* OR z hypnotic* OR non-benzodiazepin* OR nonbenzodiazepin* OR zaleplon OR zopiclone OR zolpidem) OR ti,ab(generation NEAR/3 hypnotic*) OR MAINSUBJECT.EXACT.EXPLODE("Benzodiazepines") OR ti,ab(benzodiazepin* OR bzd OR flurazepam OR loprazolam OR lormetazepam OR nitrazepam OR temazepam OR diazepam OR chlordiazepoxide OR lorazepam OR oxazepam) OR (MAINSUBJECT.EXACT.EXPLODE("Pregabalin") OR MAINSUBJECT.EXACT.EXPLODE("Gabapentin")) OR ti,ab(gabapentin* OR pregabalin*) OR MAINSUBJECT.EXACT.EXPLODE("Antidepressant Drugs") OR ti,ab(antidepress* OR anti depress* OR thymoanaleptic* OR thymoleptic* OR MAOI* OR "monoamine oxidase inhibit*" OR RIMA* OR tricyclic* OR SSRI* OR SNRI* OR SNORI*) OR (MAINSUBJECT.EXACT.EXPLODE("Doxepin") OR MAINSUBJECT.EXACT.EXPLODE("Tranylcypromine") OR MAINSUBJECT.EXACT.EXPLODE("Sertraline") OR MAINSUBJECT.EXACT.EXPLODE("Isocarboxazid") OR MAINSUBJECT.EXACT.EXPLODE("Tryptophan") OR MAINSUBJECT.EXACT.EXPLODE("Fluoxetine") OR MAINSUBJECT.EXACT.EXPLODE("Hydroxytryptophan (5-)") OR MAINSUBJECT.EXACT.EXPLODE("Nortriptyline") OR MAINSUBJECT.EXACT.EXPLODE("Citalopram") OR MAINSUBJECT.EXACT.EXPLODE("Phenelzine") OR MAINSUBJECT.EXACT.EXPLODE("Imipramine") OR MAINSUBJECT.EXACT.EXPLODE("Mianserin") OR MAINSUBJECT.EXACT.EXPLODE("Paroxetine") OR MAINSUBJECT.EXACT.EXPLODE("Moclobemide") OR MAINSUBJECT.EXACT.EXPLODE("Amitriptyline") OR

Appendices

MAINSUBJECT.EXACT.EXPLODE("Maprotiline") OR MAINSUBJECT.EXACT.EXPLODE("Trazodone") OR MAINSUBJECT.EXACT.EXPLODE("Fluvoxamine") OR MAINSUBJECT.EXACT.EXPLODE("Chlorimipramine")) OR ti,ab(amitriptyline OR amoxapine OR clomipramine OR dosulepin OR doxepin OR imipramine OR lofepramine OR maprotiline OR mianserin OR nortriptyline OR protriptyline OR trazodone OR trimipramine OR isocarboxazid OR moclobemide OR phenelzine OR tranylcypromine OR citalopram OR escitalopram OR fluoxetine OR fluvoxamine OR paroxetine OR sertraline OR agomelatine OR duloxetine OR flupentixol OR mirtazapine OR nefazodone OR oxitriptan OR reboxetine OR tryptophan OR venlafaxine OR vortioxetine))) AND ((su.exact.explode("clinical trials") OR ti,ab((clinical OR control*) NEAR/3 trial*) OR ti,ab((singl* OR doubl* OR trebl* OR tripl*) NEAR/5 (blind* OR mask*)) OR ti,ab(volunteer* OR control-group OR controls) OR su.exact("placebo") OR ti,ab(placebo*)) OR (((SU.EXACT("Literature Review") OR RTYPE(review) OR ti(review) OR me(literature review)) AND (ti,ab(systematic OR evidence OR methodol* OR quantitative*))) OR (SU.EXACT("Meta Analysis") OR ti,ab(meta-analys* OR metanalys* OR metaanalys* OR meta analys*) OR ti,ab((systematic OR evidence* OR methodol* OR quantitative*) NEAR/3 (review* OR overview*)) OR ti,ab((pool* OR combined OR combining) NEAR/2 (data OR trials OR studies OR results)) OR RTYPE(systematic OR meta*) OR ME(meta analysis OR systematic review))))) NOT (su.exact.explode("rodents") OR su.exact.explode("mice") OR (su.exact("animals") NOT (su.exact("human males") OR su.exact("human females"))) OR ti(rat OR rats OR mouse OR mice))) AND (la.exact("English"))

C.1.2.5 Health Technology appraisals (HTA) (Centre for Reviews and Disseminations) search terms 1. MeSH DESCRIPTOR Substance-Related Disorders EXPLODE ALL TREES 2. (MeSH descriptor Substance Withdrawal Syndrome explode all trees) 3. (MeSH descriptor Inappropriate Prescribing explode all trees) 4. (MeSH descriptor Medical Overuse explode all trees) 5. (MeSH descriptor Deprescriptions explode all trees) 6. ((abstinen* or abstain* or cessat* or detox* or discontinu* or reduc* or stop* or taper* or withdraw* or substitut* or depend* or addict* or abuse* or abusing or chronic or long* term or longterm or short* term or short term or misus* or overus* or deprescrib*)) 7. ((over* adj3 (use* or using or utlisat* or utilizat*) adj3 (prescription* or prescrib* or drug* or medicine* or medication* or pharm*))) 8. ((inappropriate adj3 (prescription or prescrib*))) 9. #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 10. (MeSH descriptor Narcotics explode all trees) 11. (MeSH descriptor: [Analgesics, Opioid] explode all trees) 12. ((analgesic* adj3 (opioid* or narcotic) adj3 agent*)) 13. (MeSH descriptor: [Buprenorphine] explode all trees) 14. (MeSH descriptor: [Codeine] explode all trees) 15. (MeSH descriptor: [Dextromoramide] explode all trees) 16. (MeSH descriptor: [Heroin] explode all trees) 17. (MeSH descriptor: [Fentanyl] explode all trees) 18. (MeSH descriptor: [Hydromorphone] explode all trees) 19. (MeSH descriptor: [Meptazinol] explode all trees) 20. (MeSH descriptor: [Methadone] explode all trees) 21. (MeSH descriptor: [Morphine] explode all trees)

Appendices

22. (MeSH descriptor: [Oxycodone] explode all trees) 23. (MeSH descriptor: [Opium] explode all trees) 24. (MeSH descriptor: [Pentazocine] explode all trees) 25. (MeSH descriptor: [Meperidine] explode all trees) 26. (MeSH descriptor: [Tramadol] explode all trees) 27. (buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin) 28. (z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem) 29. (generation adj3 hypnotic*) 30. (MeSH descriptor: [Benzodiazepines] explode all trees) 31. (benzodiazepin* or bzd or flurazepam or loprazolam or lormetazepam or nitrazepam or temazepam or diazepam or chlordiazepoxide or lorazepam or oxazepam) 32. (MeSH descriptor: [Pregabalin] explode all trees) 33. (gabapentin* or pregabalin*) 34. (MeSH descriptor: [Antidepressive Agents] explode all trees) 35. (antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or MAOI* or "monoamine oxidase inhibit*" or RIMA* or tricyclic* or SSRI* or SNRI* or SNORI*) 36. (MeSH descriptor: [Amitriptyline] explode all trees) 37. (MeSH descriptor: [Amoxapine] explode all trees) 38. (MeSH descriptor: [Clomipramine] explode all trees) 39. (MeSH descriptor: [Dothiepin] explode all trees) 40. (MeSH descriptor: [Doxepin] explode all trees) 41. (MeSH descriptor: [Imipramine] explode all trees) 42. (MeSH descriptor: [Lofepramine] explode all trees) 43. (MeSH descriptor: [Maprotiline] explode all trees) 44. (MeSH descriptor: [Mianserin] explode all trees) 45. (MeSH descriptor: [Nortriptyline] explode all trees) 46. (MeSH descriptor: [Protriptyline] explode all trees) 47. (MeSH descriptor: [Trazodone] explode all trees) 48. (MeSH descriptor: [Trimipramine] explode all trees) 49. (MeSH descriptor: [Isocarboxazid] explode all trees) 50. (MeSH descriptor: [Moclobemide] explode all trees) 51. (MeSH descriptor: [Phenelzine] explode all trees) 52. (MeSH descriptor: [Tranylcypromine] explode all trees) 53. (MeSH descriptor: [Citalopram] explode all trees) 54. (MeSH descriptor: [Fluoxetine] explode all trees) 55. (MeSH descriptor: [Fluvoxamine] explode all trees) 56. (MeSH descriptor: [Paroxetine] explode all trees) 57. (MeSH descriptor: [Sertraline] explode all trees) 58. (MeSH descriptor: [5-Hydroxytryptophan] explode all trees) 59. (MeSH descriptor: [Duloxetine Hydrochloride] explode all trees)

Appendices

60. (MeSH descriptor: [Flupenthixol] explode all trees) 61. (MeSH descriptor: [Tryptophan] explode all trees) 62. (MeSH descriptor: [Venlafaxine Hydrochloride] explode all trees) 63. (amitriptyline or amoxapine or clomipramine or dosulepin or doxepin or imipramine or lofepramine or maprotiline or mianserin or nortriptyline or protriptyline or trazodone or trimipramine or isocarboxazid or moclobemide or phenelzine or tranylcypromine or citalopram or escitalopram or fluoxetine or fluvoxamine or paroxetine or sertraline or agomelatine or duloxetine or flupentixol or mirtazapine or nefazodone or oxitriptan or reboxetine or tryptophan or venlafaxine or vortioxetine) 64. #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 OR #38 OR #39 OR #40 OR #41 OR #42 OR #43 OR #44 OR #45 OR #46 OR #47 OR #48 OR #49 OR #50 OR #51 OR #52 OR #53 OR #54 OR #55 OR #56 OR #57 OR #58 OR #59 OR #60 OR #61 OR #62 OR #63 65. #9 AND #64

C.1.2.6 Trials Register of Promoting Health Interventions (TRoPHI) search terms 1. Freetext (All but Authors): "substance-related disorders" OR "substance withdrawal syndrome" OR "inappropriate prescribing" OR "medical overuse" OR "deprescriptions" OR "abstinen*" or "abstain*" or "cessat*" or "detox*" or "discontinu*" or "reduc*" or "stop*" or "taper*" or "withdraw*" or "substitut*" or "depend*" or "addict*" or "abuse*" or "abusing" or "chronic" or "long* term" or "longterm" or "short* term" or "short term" or "misus*" or "overus*" OR "deprescrib*" 2. Freetext (All but Authors): "over*" near "use*" near "prescri*" 3. Freetext (All but Authors): "inappropriate" near "prescri*" 4. 1 OR 2 OR 3 5. Freetext (All but Authors): "buprenorphine*" or "codeine*" or "dextromoramide*" or "diamorphine*" or "dihydrocodeine*" or "dipipanone*" or "fentanyl" or "hydromorphone*" or "meptazinol" or "methadone*" or "morphine*" or "oxycodone" or "papaveretum" or "pentazocine" or "pethidine" or "tapentadol" or "tramadol" or "heroin" 6. Freetext (All but Authors): "z drug*" or "z hypnotic*" or "non-benzodiazepin*" or "nonbenzodiazepin*" or "zaleplon" or "zopiclone" or "zolpidem" 7. Freetext (All but Authors): "generation" near "hypnotic" 8. Freetext (All but Authors): "benzodiazepin*" or "bzd" or "flurazepam" or "loprazolam" or "lormetazepam" or "nitrazepam" or "temazepam" or "diazepam" or "chlordiazepoxide" or "lorazepam" or "oxazepam" 9. Freetext (All but Authors): "gabapentin*" or "pregabalin*" 10. Freetext (All but Authors): "antidepress*" or "anti depress*" or "thymoanaleptic*" or "thymoleptic*" or "MAOI*" or "monoamine oxidase inhibit*" or "RIMA*" or "tricyclic*" or "SSRI*" or "SNRI*" or "SNORI*" 11. Freetext (All but Authors): "amitriptyline" or "amoxapine" or "clomipramine" or "dosulepin" or "doxepin" or "imipramine" or "lofepramine" or "maprotiline" or "mianserin" or "nortriptyline" or "protriptyline" or "trazodone" or "trimipramine" or "isocarboxazid" or "moclobemide" or "phenelzine" or "tranylcypromine" or "citalopram" or "escitalopram" or "fluoxetine" or "fluvoxamine" or "paroxetine" or "sertraline" or "agomelatine" or "duloxetine" or "flupentixol" or "mirtazapine" or "nefazodone" or "oxitriptan" or "reboxetine" or "tryptophan" or "venlafaxine" or "vortioxetine" 12. 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 13. 4 AND 12

Appendices

C.1.2.7 ASSIA (Proquest) search terms 1. ((MAINSUBJECT.EXACT.EXPLODE("Substance abuse disorders") OR ti,ab(abstinen* OR abstain* OR cessat* OR detox* OR discontinu* OR reduc* OR stop* OR taper* OR withdraw* OR substitut* OR depend* OR addict* OR abuse* OR abusing OR chronic OR long* term OR longterm OR short* term OR short term OR misus* OR overus* OR deprescrib*) OR ti,ab(over* NEAR/3 (use* OR using OR utlisat* OR utilizat*) NEAR/3 (prescription* OR prescrib* OR drug* OR medicine* OR medication* OR pharm*)) OR ti,ab(inappropriate NEAR/3 (prescription OR prescrib*))) AND ((MAINSUBJECT.EXACT.EXPLODE("Analgesics") OR MAINSUBJECT.EXACT.EXPLODE("Narcotics")) OR ti,ab(analgesic* NEAR/3 (opioid* or narcotic) NEAR/3 agent*) OR (MAINSUBJECT.EXACT.EXPLODE("Methadone") OR MAINSUBJECT.EXACT.EXPLODE("Heroin") OR MAINSUBJECT.EXACT.EXPLODE("Buprenorphine") OR MAINSUBJECT.EXACT.EXPLODE("Codeine") OR MAINSUBJECT.EXACT.EXPLODE("Hydromorphone") OR MAINSUBJECT.EXACT.EXPLODE("Tramadol") OR MAINSUBJECT.EXACT.EXPLODE("Morphine") OR MAINSUBJECT.EXACT.EXPLODE("Opium")) OR ti,ab(buprenorphine* or codeine* or dextromoramide* or diamorphine* or dihydrocodeine* or dipipanone* or fentanyl or hydromorphone* or meptazinol or methadone* or morphine* or oxycodone or papaveretum or pentazocine or pethidine or tapentadol or tramadol or heroin) OR ti,ab(z drug* or z hypnotic* or non-benzodiazepin* or nonbenzodiazepin* or zaleplon or zopiclone or zolpidem) OR ti,ab(generation NEAR/3 hypnotic*) OR MAINSUBJECT.EXACT.EXPLODE("Benzodiazepines") OR ti,ab(benzodiazepin* OR bzd OR flurazepam OR loprazolam OR lormetazepam OR nitrazepam OR temazepam OR diazepam OR chlordiazepoxide OR lorazepam OR oxazepam) OR MAINSUBJECT.EXACT.EXPLODE("Gabapentin") OR ti,ab(gabapentin* or pregabalin*) OR MAINSUBJECT.EXACT.EXPLODE("Antidepressant drugs") OR ti,ab(antidepress* or anti depress* or thymoanaleptic* or thymoleptic* or MAOI* or "monoamine oxidase inhibit*" or RIMA* or tricyclic* or SSRI* or SNRI* or SNORI*) OR (MAINSUBJECT.EXACT.EXPLODE("Imipramine") OR MAINSUBJECT.EXACT.EXPLODE("Amitriptyline") OR MAINSUBJECT.EXACT.EXPLODE("Clomipramine") OR MAINSUBJECT.EXACT.EXPLODE("Moclobemide") OR MAINSUBJECT.EXACT.EXPLODE("Sertraline") OR MAINSUBJECT.EXACT.EXPLODE("Paroxetine") OR MAINSUBJECT.EXACT.EXPLODE("Venlafaxine") OR MAINSUBJECT.EXACT.EXPLODE("Fluoxetine") OR MAINSUBJECT.EXACT.EXPLODE("Citalopram") OR MAINSUBJECT.EXACT.EXPLODE("Tryptophan")) OR ti,ab(amitriptyline OR amoxapine OR clomipramine OR dosulepin OR doxepin OR imipramine OR lofepramine OR maprotiline OR mianserin OR nortriptyline OR protriptyline OR trazodone OR trimipramine OR isocarboxazid OR moclobemide OR phenelzine OR tranylcypromine OR citalopram OR escitalopram OR fluoxetine OR fluvoxamine OR paroxetine OR sertraline OR agomelatine OR duloxetine OR flupentixol OR mirtazapine OR nefazodone OR oxitriptan OR reboxetine OR tryptophan OR venlafaxine OR vortioxetine))) AND (la.exact("English"))

C.1.3 Step 3: Citation searching

Database Dates searched Search filter used Scopus 6 key references None

C.1.3.1 Scopus

The following papers were used for citation searching:

1. Siegfried Kasper, Celso Iglesias-García, Edward Schweizer, Jacquelyn Wilson, Sarah DuBrava, Rita Prieto, Verne W. Pitman and Lloyd Knapp. Pregabalin long-term treatment and assessment of discontinuation in patients with generalized anxiety disorder. International Journal of Neuropsychopharmacology (2014),

Appendices

17, 685–695.

2. Cook B, Creedon T, Wang Y, Lu C, Carson N, Jules P et al. Examining racial/ethnic differences in patterns of benzodiazepine prescription and misuse. Drug and Alcohol Dependence. 2018; 187:29.

3. Zahradnik A, Otto C, Crackau B, Löhrmann I, Bischof G, John U, Rumpf HJ (2009). Randomized controlled trial of a brief intervention for problematic prescription drug use in non-treatment-seeking patients. Addiction. Jan;104(1):109-17. doi: 10.1111/j.1360-0443.2008.02421.x.

4. Tannenbaum et al. 2014 Reduction of inappropriate benzodiazepine prescriptions among older adults through direct patient education: the EMPOWER cluster randomized trial. JAMA Intern Med. 2014;174(6):890-898. doi:10.1001/jamainternmed.2014.949.

5. Davies 2018: Antidepressant withdrawal: a survey of patients' experience by the All-Party Parliamentary Group for Prescribed Drug Dependence

6. Guy 2018: The Patient Voice: an analysis of personal accounts of prescribed drug dependence and withdrawal submitted to persons in Scotland and Wales.

Appendices

Appendix D: Forest plots

D.1 Harms of dependency on a medication

Opioids for chronic pain

Figure 3: Proportion of subjects who developed shopping behaviour during any point of 1 year follow-up: tapentadol versus oxycodone Tapentadol Oxycodone Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Cepeda 2013 88 42940 967 112821 100.0% 0.24 [0.19, 0.30]

Total (95% CI) 42940 112821 100.0% 0.24 [0.19, 0.30] Total events 88 967 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 12.87 (P < 0.00001) Tapentadol Oxycodone

Figure 4: Proportion of subjects who developed shopping behaviour during any point of 1 year follow-up: tapentadol versus oxycodone (adjusted odds ratio) Odds Ratio Odds Ratio Study or Subgroup log[Odds Ratio] SE Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Cepeda 2013 1.2528 0.1115 100.0% 3.50 [2.81, 4.36]

Total (95% CI) 100.0% 3.50 [2.81, 4.36] Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 11.24 (P < 0.00001) Oxycodone Tapentadol

Note: Adjusted odds ratio reported by authors – controlling for gender, benzodiazepine use and type of payment at first opioid exposure using a conditional logistic regression.

Figure 5: Number of shopping episodes per subject during the year of follow-up: tapentadol versus oxycodone Tapentadol Oxycodone Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Cepeda 2013 0.004 0.1 42940 0.02 0.3 112821 100.0% -0.02 [-0.02, -0.01]

Total (95% CI) 42940 112821 100.0% -0.02 [-0.02, -0.01] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 15.76 (P < 0.00001) Tapentadol Oxycodone

Appendices

D.2 Harms / side effects from stopping these medications over a short time frame

Opioids for chronic pain

Figure 6: Clinical Opioid Withdrawal Scale (COWS) no opioid withdrawal – opioid vs placebo Opioid Control Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI 1.1.1 COWS assessment 2-4 days after last intake of medication Afilalo 2010 - Oxycodone vs placebo 32 37 12 12 51.9% 0.89 [0.75, 1.05] Afilalo 2010 - Tapentadol vs placebo 29 35 11 11 48.1% 0.86 [0.70, 1.04] Subtotal (95% CI) 72 23 100.0% 0.87 [0.77, 0.99] Total events 61 23 Heterogeneity: Chi² = 0.09, df = 1 (P = 0.76); I² = 0% Test for overall effect: Z = 2.08 (P = 0.04)

1.1.2 COWS assessment at 4 days after last intake of medication Buynak 2010 - Oxycodone vs placebo 82 90 26 29 51.9% 1.02 [0.88, 1.17] Buynak 2010 - Tapentadol vs placebo 59 62 27 30 48.1% 1.06 [0.93, 1.21] Subtotal (95% CI) 152 59 100.0% 1.04 [0.94, 1.14] Total events 141 53 Heterogeneity: Chi² = 0.17, df = 1 (P = 0.68); I² = 0% Test for overall effect: Z = 0.72 (P = 0.47)

1.1.3 COWS assessment 5+ days after last intake of medication Afilalo 2010 - Oxycodone vs placebo 72 84 27 30 51.0% 0.95 [0.82, 1.10] Afilalo 2010 - Tapentadol vs placebo 69 70 27 29 49.0% 1.06 [0.96, 1.17] Subtotal (95% CI) 154 59 100.0% 1.00 [0.92, 1.10] Total events 141 54 Heterogeneity: Chi² = 1.50, df = 1 (P = 0.22); I² = 33% Test for overall effect: Z = 0.10 (P = 0.92)

0.5 0.7 1 1.5 2 Favours placebo Favours opioid Test for subgroup differences: Chi² = 4.65, df = 2 (P = 0.10), I² = 57.0%

Figure 7: COWS - mild or moderate opioid withdrawal: opioid vs placebo Opioid Control Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI 1.2.1 COWS assessment 2-4 days after last intake of medication Afilalo 2010 - Oxycodone vs placebo 5 37 0 12 49.6% 3.76 [0.22, 63.50] Afilalo 2010 - Tapentadol vs placebo 6 35 0 11 50.4% 4.33 [0.26, 71.35] Subtotal (95% CI) 72 23 100.0% 4.04 [0.55, 29.54] Total events 11 0 Heterogeneity: Tau² = 0.00; Chi² = 0.00, df = 1 (P = 0.94); I² = 0% Test for overall effect: Z = 1.38 (P = 0.17)

1.2.2 COWS assessment at 4 days after last intake of medication Buynak 2010 - Oxycodone vs placebo 8 90 3 30 59.8% 0.89 [0.25, 3.14] Buynak 2010 - Tapentadol vs placebo 3 62 3 29 40.2% 0.47 [0.10, 2.18] Subtotal (95% CI) 152 59 100.0% 0.69 [0.26, 1.82] Total events 11 6 Heterogeneity: Tau² = 0.00; Chi² = 0.40, df = 1 (P = 0.53); I² = 0% Test for overall effect: Z = 0.76 (P = 0.45)

1.2.3 COWS assessment 5+ days after last intake of medication Afilalo 2010 - Oxycodone vs placebo 12 84 2 29 55.2% 2.07 [0.49, 8.71] Afilalo 2010 - Tapentadol vs placebo 1 70 3 30 44.8% 0.14 [0.02, 1.32] Subtotal (95% CI) 154 59 100.0% 0.63 [0.05, 8.48] Total events 13 5 Heterogeneity: Tau² = 2.66; Chi² = 3.92, df = 1 (P = 0.05); I² = 75% Test for overall effect: Z = 0.35 (P = 0.72)

0.001 0.1 1 10 1000 Favours opioid Favours placebo Test for subgroup differences: Chi² = 2.55, df = 2 (P = 0.28), I² = 21.7%

Appendices

Figure 8: Withdrawal syndrome: Tapentadol vs oxycodone Tapentadol Oxycodone Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Wild 2010 13 894 2 223 100.0% 1.62 [0.37, 7.13]

Total (95% CI) 894 223 100.0% 1.62 [0.37, 7.13] Total events 13 2 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.64 (P = 0.52) Tapentadol Oxycodone

Figure 9: COWS – no clinical withdrawal: Tapentadol vs oxycodone Tapentadol Oxycodone Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI 1.4.1 COWS assessment 2-4 days after last intake of medication Afilalo 2010 - Tapentadol vs oxycodone 29 35 32 37 53.3% 0.96 [0.79, 1.17] Wild 2010 97 125 16 22 46.7% 1.07 [0.81, 1.40] Subtotal (95% CI) 160 59 100.0% 1.01 [0.85, 1.19] Total events 126 48 Heterogeneity: Chi² = 0.43, df = 1 (P = 0.51); I² = 0% Test for overall effect: Z = 0.10 (P = 0.92)

1.4.2 COWS assessment at 4 days after last intake of medication Buynak 2010 - Tapentadol vs oxycodone 59 62 82 90 100.0% 1.04 [0.96, 1.14] Subtotal (95% CI) 62 90 100.0% 1.04 [0.96, 1.14] Total events 59 82 Heterogeneity: Not applicable Test for overall effect: Z = 1.00 (P = 0.32)

1.4.3 COWS assessment 5+ days after last intake of medication Afilalo 2010 - Tapentadol vs oxycodone 69 70 72 84 50.3% 1.15 [1.05, 1.26] Wild 2010 146 166 42 50 49.7% 1.05 [0.92, 1.20] Subtotal (95% CI) 236 134 100.0% 1.10 [1.01, 1.19] Total events 215 114 Heterogeneity: Chi² = 1.45, df = 1 (P = 0.23); I² = 31% Test for overall effect: Z = 2.31 (P = 0.02)

0.5 0.7 1 1.5 2 Favours oxycodone Favours tapentadol Test for subgroup differences: Chi² = 1.19, df = 2 (P = 0.55), I² = 0%

Figure 10: COWS - mild or moderate opioid withdrawal: Tapentadol vs oxycodone Tapentadol Oxycodone Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI 1.5.1 COWS assessment 2-4 days after last intake of medication Afilalo 2010 - Tapentadol vs oxycodone 6 35 5 37 11.2% 1.27 [0.43, 3.78] Hale 2009 52 306 19 66 65.3% 0.59 [0.38, 0.93] Wild 2010 28 125 6 22 23.4% 0.82 [0.39, 1.75] Subtotal (95% CI) 466 125 100.0% 0.70 [0.48, 1.00] Total events 86 30 Heterogeneity: Tau² = 0.00; Chi² = 1.87, df = 2 (P = 0.39); I² = 0% Test for overall effect: Z = 1.95 (P = 0.05)

1.5.2 COWS assessment 4 days after last intake of medication Buynak 2010 - Tapentadol vs oxycodone 3 62 8 90 100.0% 0.54 [0.15, 1.97] Subtotal (95% CI) 62 90 100.0% 0.54 [0.15, 1.97] Total events 3 8 Heterogeneity: Not applicable Test for overall effect: Z = 0.93 (P = 0.35)

1.5.3 COWS assessment 5+ days after last intake of medication Afilalo 2010 - Tapentadol vs oxycodone 1 70 12 84 40.5% 0.10 [0.01, 0.75] Wild 2010 20 166 8 50 59.5% 0.75 [0.35, 1.60] Subtotal (95% CI) 236 134 100.0% 0.33 [0.04, 2.72] Total events 21 20 Heterogeneity: Tau² = 1.78; Chi² = 3.96, df = 1 (P = 0.05); I² = 75% Test for overall effect: Z = 1.03 (P = 0.30)

0.01 0.1 1 10 100 Favours tapentadol Favours oxycodone Test for subgroup differences: Chi² = 0.57, df = 2 (P = 0.75), I² = 0%

Appendices

Figure 11: Drug withdrawal syndrome: Tapentadol vs oxycodone Tapentadol Oxycodone Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Wild 2010 9 894 1 223 100.0% 2.24 [0.29, 17.63]

Total (95% CI) 894 223 100.0% 2.24 [0.29, 17.63] Total events 9 1 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.77 (P = 0.44) Tapentadol Oxycodone

Z-drugs

Figure 12: Rebound insomnia (proportion of patients with a lower self-reported total sleep time (sTST) value after first night of run out phase compared to lowest sTST value during run-in phase): Zolpidem versus placebo in people with insomnia (based on patients switched from active treatment to placebo during run-out phase) Z-drug Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI 4.1.1 Run out phase - day 1 Hajak 2018 13 67 8 126 100.0% 3.06 [1.33, 7.00] Subtotal (95% CI) 67 126 100.0% 3.06 [1.33, 7.00] Total events 13 8 Heterogeneity: Not applicable Test for overall effect: Z = 2.64 (P = 0.008)

4.1.2 Run out phase - day 2 Hajak 2018 5 67 7 126 100.0% 1.34 [0.44, 4.07] Subtotal (95% CI) 67 126 100.0% 1.34 [0.44, 4.07] Total events 5 7 Heterogeneity: Not applicable Test for overall effect: Z = 0.52 (P = 0.60)

4.1.3 Run out phase - day 3 Hajak 2018 5 67 5 126 100.0% 1.88 [0.56, 6.27] Subtotal (95% CI) 67 126 100.0% 1.88 [0.56, 6.27] Total events 5 5 Heterogeneity: Not applicable Test for overall effect: Z = 1.03 (P = 0.30)

0.01 0.1 1 10 100 Z-drug Placebo Test for subgroup differences: Chi² = 1.43, df = 2 (P = 0.49), I² = 0% Note: Crude numbers estimated from the percentages and total numbers provided in study.

Appendices

Figure 13: Rebound insomnia - proportion of patients with a lower self-reported time to sleep onset (sTSO) value after first night of run out phase compared to lowest sTSO value during run-in phase): Zolpidem versus placebo in people with insomnia (based on patients switched from active treatment to placebo during run-out phase) Z-drug Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI 4.2.1 Run out phase - day 1 Hajak 2018 12 67 8 126 100.0% 2.82 [1.21, 6.56] Subtotal (95% CI) 67 126 100.0% 2.82 [1.21, 6.56] Total events 12 8 Heterogeneity: Not applicable Test for overall effect: Z = 2.41 (P = 0.02)

4.2.2 Run out phase - day 2 Hajak 2018 4 67 4 126 100.0% 1.88 [0.49, 7.28] Subtotal (95% CI) 67 126 100.0% 1.88 [0.49, 7.28] Total events 4 4 Heterogeneity: Not applicable Test for overall effect: Z = 0.91 (P = 0.36)

4.2.3 Run out phase - day 3 Hajak 2018 5 67 5 126 100.0% 1.88 [0.56, 6.27] Subtotal (95% CI) 67 126 100.0% 1.88 [0.56, 6.27] Total events 5 5 Heterogeneity: Not applicable Test for overall effect: Z = 1.03 (P = 0.30)

0.01 0.1 1 10 100 Z-drug Placebo Test for subgroup differences: Chi² = 0.42, df = 2 (P = 0.81), I² = 0% Note: Crude numbers estimated from the percentages and total numbers provided in study.

Benzodiazepines vs gabapentinoids

Figure 14: Mean change in discontinuation symptoms measured with the Physician Withdrawal Checklist (PWC) following treatment period 1 – 1 week after taper (using higher dose of gabapentinoids – 450-600 mg/d pregabalin): gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Kasper 2013 1.9 7.6064 58 2.3 6.6148 49 100.0% -0.40 [-3.09, 2.29]

Total (95% CI) 58 49 100.0% -0.40 [-3.09, 2.29] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 0.29 (P = 0.77) Gabapentinoids Benzodiazepines

Figure 15: Mean change in discontinuation symptoms measured with the Physician Withdrawal Checklist (PWC) after treatment period 1 – 2 weeks after taper (using higher dose of gabapentinoids – 450-600 mg/d pregabalin): gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Kasper 2013 2.1 5.8619 54 1.6 6.2494 44 100.0% 0.50 [-1.92, 2.92]

Total (95% CI) 54 44 100.0% 0.50 [-1.92, 2.92] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 0.41 (P = 0.69) Gabapentinoids Benzodiazepines

Appendices

Figure 16: Mean change in discontinuation symptoms measured with the physician Withdrawal Checklist (PWC) after treatment period 2 – 1 weeks after taper (using higher dose of gabapentinoids – 450-600 mg/d pregabalin): gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Kasper 2013 1.7 5.2671 109 3 6.518 99 100.0% -1.30 [-2.92, 0.32]

Total (95% CI) 109 99 100.0% -1.30 [-2.92, 0.32] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 1.57 (P = 0.12) Gabapentinoids Benzodiazepines

Figure 17: Mean change in discontinuation symptoms measured with the Physician Withdrawal Checklist (PWC) after treatment period 2 – 2 weeks after taper (using higher dose of gabapentinoids – 450-600 mg/d pregabalin): gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Kasper 2013 2.8 6.2309 106 2.2 5.8267 93 100.0% 0.60 [-1.08, 2.28]

Total (95% CI) 106 93 100.0% 0.60 [-1.08, 2.28] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 0.70 (P = 0.48) Gabapentinoids Benzodiazepines

Figure 18: Discontinuation emergent signs and symptoms (DESS) occurring in 5% or more patients during 2 weeks following taper initiation after treatment period 1 (12 weeks treatment): Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 38 110 17 52 100.0% 1.06 [0.66, 1.69]

Total (95% CI) 110 52 100.0% 1.06 [0.66, 1.69] Total events 38 17 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.23 (P = 0.82) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 19: DESS occurring in 5% or more patients during 2 weeks following taper initiation after treatment period 2 (patients given active treatment in period 2): Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 55 203 28 100 100.0% 0.97 [0.66, 1.42]

Total (95% CI) 203 100 100.0% 0.97 [0.66, 1.42] Total events 55 28 Heterogeneity: Not applicable 0.5 0.7 1 1.5 2 Test for overall effect: Z = 0.17 (P = 0.87) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Appendices

Figure 20: Anxiety after treatment period 1: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 3 110 2 52 100.0% 0.71 [0.12, 4.12]

Total (95% CI) 110 52 100.0% 0.71 [0.12, 4.12] Total events 3 2 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.38 (P = 0.70) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 21: Anxiety after period 2 (patients given active treatment in period 2): Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 11 203 8 100 100.0% 0.68 [0.28, 1.63]

Total (95% CI) 203 100 100.0% 0.68 [0.28, 1.63] Total events 11 8 Heterogeneity: Not applicable 0.2 0.5 1 2 5 Test for overall effect: Z = 0.87 (P = 0.38) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 22: Dizziness after period 1: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Kasper 2013 3 110 0 52 100.0% 4.44 [0.39, 50.92]

Total (95% CI) 110 52 100.0% 4.44 [0.39, 50.92] Total events 3 0 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.20 (P = 0.23) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 23: Headache after period 1: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 7 110 1 52 100.0% 3.31 [0.42, 26.20]

Total (95% CI) 110 52 100.0% 3.31 [0.42, 26.20] Total events 7 1 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.13 (P = 0.26) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Appendices

Figure 24: Headache after period 2 (patients given active treatment in period 2): Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 8 203 2 100 100.0% 1.97 [0.43, 9.11]

Total (95% CI) 203 100 100.0% 1.97 [0.43, 9.11] Total events 8 2 Heterogeneity: Not applicable 0.05 0.2 1 5 20 Test for overall effect: Z = 0.87 (P = 0.39) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 25: Insomnia after period 1: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 10 110 10 52 100.0% 0.47 [0.21, 1.06]

Total (95% CI) 110 52 100.0% 0.47 [0.21, 1.06] Total events 10 10 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.81 (P = 0.07) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 26: Insomnia after period 2 (patients given active treatment in period 2): Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 21 203 6 100 100.0% 1.72 [0.72, 4.14]

Total (95% CI) 203 100 100.0% 1.72 [0.72, 4.14] Total events 21 6 Heterogeneity: Not applicable 0.2 0.5 1 2 5 Test for overall effect: Z = 1.22 (P = 0.22) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 27: Nausea after period 1: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 7 110 2 52 100.0% 1.65 [0.36, 7.69]

Total (95% CI) 110 52 100.0% 1.65 [0.36, 7.69] Total events 7 2 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.64 (P = 0.52) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Appendices

Figure 28: Rebound anxiety after treatment period 1: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 4 110 2 48 100.0% 0.87 [0.17, 4.60]

Total (95% CI) 110 48 100.0% 0.87 [0.17, 4.60] Total events 4 2 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.16 (P = 0.87) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Figure 29: Rebound anxiety after treatment period 2: Gabapentinoids vs benzodiazepines Gabapentinoids Benzodiazepines Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Kasper 2013 4 203 6 100 100.0% 0.33 [0.09, 1.14]

Total (95% CI) 203 100 100.0% 0.33 [0.09, 1.14] Total events 4 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.76 (P = 0.08) Gabapentinoids Benzodiazepines Note: Gabapentinoid doses combined

Benzodiazepines

No comparative studies for this intervention.

Antidepressants

Figure 30: Rebound insomnia (based on wake time after sleep onset criteria experience over the two nights after discontinuation): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Krystal 2011 4 148 1 73 100.0% 1.97 [0.22, 17.34]

Total (95% CI) 148 73 100.0% 1.97 [0.22, 17.34] Total events 4 1 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.61 (P = 0.54) Antidepressant Placebo NB – actual numbers assumed by NGC calculations, % only provided in study.

Figure 31: Benzodiazepine Withdrawal Symptom Questionnaire Criteria BWSQ (based on 3 more new symptoms in the score): antidepressant vs placebo Antidepresant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Krystal 2011 1 148 1 73 100.0% 0.49 [0.03, 7.77]

Total (95% CI) 148 73 100.0% 0.49 [0.03, 7.77] Total events 1 1 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.50 (P = 0.62) Antidepressant Placebo

Appendices

Figure 32: Suicide attempts: antidepressant vs placebo Antidepressant Placebo Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Nishimura 2018 1 448 1 152 60.2% 0.27 [0.01, 6.57] Tourian 2009 - Duloxetine vs placebo 1 157 0 161 39.8% 7.58 [0.15, 382.11]

Total (95% CI) 605 313 100.0% 1.02 [0.09, 12.12] Total events 2 1 Heterogeneity: Chi² = 1.67, df = 1 (P = 0.20); I² = 40% 0.002 0.1 1 10 500 Test for overall effect: Z = 0.02 (P = 0.99) Antidepressant Placebo Note: Nishimura 2018 reported as a serious adverse event but time point not reported whereas Tourian 2009 reported during taper phase. Tourian 2009 had two other study arms (desvenlafaxine 50 mg/d and 100 mg/d) that did not report any suicide attempts.

Figure 33: Depression: antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI Boyer 2008 12 324 1 161 34.2% 5.96 [0.78, 45.46] Nishimura 2018 1 448 0 152 22.7% 1.02 [0.04, 24.96] Rosenthal 2012 3 272 9 276 43.1% 0.34 [0.09, 1.24]

Total (95% CI) 1044 589 100.0% 1.16 [0.15, 8.76] Total events 16 10 Heterogeneity: Tau² = 2.03; Chi² = 5.78, df = 2 (P = 0.06); I² = 65% 0.01 0.1 1 10 100 Test for overall effect: Z = 0.14 (P = 0.88) Antidepressant Placebo Note: Liebowitz 2008 – one case of depression with suicide ideation, which is reported for suicide ideation below Nishimura 2018 time point not reported; other studies reported during taper period

Figure 34: Suicide ideation: antidepressant vs placebo Antidepressant Placebo Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI

Liebowitz 2008 1 299 0 152 15.7% 4.52 [0.07, 285.64] Mahableshwarkar 2015 - Duloxetine vs placebo 2 152 0 80 31.6% 4.63 [0.25, 86.06] Mahableshwarkar 2015 - Vortioxetine vs placebo 2 301 0 81 23.4% 3.57 [0.12, 106.43] Nishimura 2018 1 448 0 152 13.3% 3.82 [0.04, 345.77] Tourian 2009 - Desvenlafaxine vs placebo 1 298 0 161 16.0% 4.67 [0.08, 283.56]

Total (95% CI) 1498 626 100.0% 4.24 [0.82, 21.90]

Total events 7 0 Heterogeneity: Chi² = 0.02, df = 4 (P = 1.00); I² = 0% 0.01 0.1 1 10 100 Test for overall effect: Z = 1.72 (P = 0.09) Antidepressant Placebo

Note: Nishimura 2018 time point not reported and Mahableshwarkar 2015 time point reported as during study; other studies reported during taper period. Liebowitz 2008 – depression with suicide ideation. Tourian 2009 – desvenlafaxine doses combined; study also had another arm of duloxetine which did not report any cases of suicide ideation.

Figure 35: Mean DESS at 1 week 1 after discontinuation (higher worse) – Antidepressant vs placebo Antidepressant Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI Nishimura 2018 1.199 2.426 374 1 1.93 128 29.9% 0.20 [-0.22, 0.61] Stein 2012 0.9 1.6 89 0.9 1.9 75 26.9% 0.00 [-0.54, 0.54] Tourian 2009 - Desvenlafaxine vs placebo 2.514 4.216 219 1.3 2.4 111 22.8% 1.21 [0.50, 1.93] Tourian 2009 - Duloxetine vs placebo 2.4 3.7 111 1.3 2.4 111 20.4% 1.10 [0.28, 1.92]

Total (95% CI) 793 425 100.0% 0.56 [-0.01, 1.13] Heterogeneity: Tau² = 0.24; Chi² = 10.75, df = 3 (P = 0.01); I² = 72% -10 -5 0 5 10 Test for overall effect: Z = 1.94 (P = 0.05) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo arm reported twice in forest plot for comparison with two different antidepressants.

Appendices

Figure 36: Mean DESS at 2 weeks after discontinuation (higher worse) – Antidepressant vs placebo Antidepressant Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Nishimura 2018 0.869 2.169 376 0.5 1.38 125 82.3% 0.37 [0.04, 0.70] Tourian 2009 - Desvenlafaxine vs placebo 2.298 4.244 215 1.6 3.5 117 12.1% 0.70 [-0.15, 1.55] Tourian 2009 - Duloxetine vs placebo 3.2 5.6 103 1.6 3.5 117 5.6% 1.60 [0.35, 2.85]

Total (95% CI) 694 359 100.0% 0.48 [0.18, 0.77] Heterogeneity: Chi² = 3.76, df = 2 (P = 0.15); I² = 47% -10 -5 0 5 10 Test for overall effect: Z = 3.16 (P = 0.002) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo arm reported twice in forest plot for comparison with two different antidepressants.

Figure 37: DESS taper week 3: antidepressant vs placebo Antidepressant Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Tourian 2009 - Desvenlafaxine vs placebo 1.589 3.164 19 7.3 4.9 4 50.3% -5.71 [-10.72, -0.70] Tourian 2009 - Duloxetine vs placebo 1 2.8 13 7.3 4.9 4 49.7% -6.30 [-11.34, -1.26]

Total (95% CI) 32 8 100.0% -6.00 [-9.56, -2.45] Heterogeneity: Chi² = 0.03, df = 1 (P = 0.87); I² = 0% -20 -10 0 10 20 Test for overall effect: Z = 3.31 (P = 0.0009) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo arm reported twice in forest plot for comparison with two different antidepressants.

Figure 38: Vertigo – Antidepressant vs placebo (during discontinuation period) Antidepressant Placebo Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Boyer 2008 12 324 0 161 100.0% 4.63 [1.37, 15.60]

Total (95% CI) 324 161 100.0% 4.63 [1.37, 15.60] Total events 12 0 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 2.47 (P = 0.01) Antidepressant Placebo

Figure 39: Change (defined from question ‘Have you noticed any changes since you stopped taking the study drug?’) after discontinuation: antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI YaDeau 2016 14 53 11 53 100.0% 1.27 [0.64, 2.54]

Total (95% CI) 53 53 100.0% 1.27 [0.64, 2.54] Total events 14 11 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.68 (P = 0.49) Antidepressant Placebo

Figure 40: Discontinuation syndrome: antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI YaDeau 2016 9 53 5 53 100.0% 1.80 [0.65, 5.02]

Total (95% CI) 53 53 100.0% 1.80 [0.65, 5.02] Total events 9 5 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.12 (P = 0.26) Antidepressant Placebo

Appendices

Figure 41:Total taper/post study emergent adverse events: antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Boyer 2008 138 324 33 161 14.6% 2.08 [1.49, 2.89] Liebowitz 2008 147 299 54 152 23.7% 1.38 [1.09, 1.76] Liebowitz 2009 70 164 46 159 15.5% 1.48 [1.09, 1.99] Perahia 2009 14 61 4 48 1.5% 2.75 [0.97, 7.83] Raskin 2008 36 207 12 104 5.3% 1.51 [0.82, 2.77] Rickels 2010 101 190 52 185 17.5% 1.89 [1.45, 2.47] Rosenthal 2012 48 272 38 276 12.5% 1.28 [0.87, 1.90] Rynn 2008 21 93 19 109 5.8% 1.30 [0.74, 2.26] Schagen 2008 25 118 11 119 3.6% 2.29 [1.18, 4.44]

Total (95% CI) 1728 1313 100.0% 1.63 [1.44, 1.84] Total events 600 269 Heterogeneity: Chi² = 9.60, df = 8 (P = 0.29); I² = 17% 0.01 0.1 1 10 100 Test for overall effect: Z = 7.78 (P < 0.00001) Antidepressant Placebo

Figure 42: Vomiting (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Cutler 2009 1 149 3 157 100.0% 0.35 [0.04, 3.34]

Total (95% CI) 149 157 100.0% 0.35 [0.04, 3.34] Total events 1 3 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.91 (P = 0.36) Antidepressant Placebo

Figure 43: Dizziness (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Boyer 2008 37 324 4 161 28.7% 4.60 [1.67, 12.67] Cutler 2009 8 149 1 157 5.2% 8.43 [1.07, 66.58] Liebowitz 2009 24 164 4 159 21.8% 5.82 [2.06, 16.39] Rosenthal 2012 13 272 1 276 5.3% 13.19 [1.74, 100.14] Rynn 2008 6 93 3 109 14.8% 2.34 [0.60, 9.11] Tourian 2009 - Desvenlafaxine vs placebo 20 298 2 80 16.9% 2.68 [0.64, 11.25] Tourian 2009 - Duloxetine vs placebo 9 157 1 80 7.1% 4.59 [0.59, 35.57]

Total (95% CI) 1457 1022 100.0% 4.86 [2.91, 8.14] Total events 117 16 Heterogeneity: Chi² = 3.10, df = 6 (P = 0.80); I² = 0% 0.01 0.1 1 10 100 Test for overall effect: Z = 6.02 (P < 0.00001) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo outcomes halved to provide correct total scores over both arms reported.

Figure 44: Nausea (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI Boyer 2008 34 324 4 161 18.6% 4.22 [1.53, 11.70] Cutler 2009 2 149 1 157 6.9% 2.11 [0.19, 23.00] Liebowitz 2009 11 64 1 159 8.8% 27.33 [3.60, 207.34] Rickels 2010 26 190 9 185 22.7% 2.81 [1.35, 5.84] Rosenthal 2012 7 272 1 276 8.4% 7.10 [0.88, 57.35] Tourian 2009 - Desvenlafaxine vs placebo 14 298 4 80 17.7% 0.94 [0.32, 2.78] Tourian 2009 - Duloxetine vs placebo 9 157 4 80 16.9% 1.15 [0.36, 3.61]

Total (95% CI) 1454 1098 100.0% 2.78 [1.36, 5.69] Total events 103 24 Heterogeneity: Tau² = 0.45; Chi² = 12.82, df = 6 (P = 0.05); I² = 53% 0.01 0.1 1 10 100 Test for overall effect: Z = 2.80 (P = 0.005) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo outcomes halved to provide correct total scores over both arms reported.

Appendices

Figure 45: Headache (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Boyer 2008 22 324 10 161 28.7% 1.09 [0.53, 2.25] Cutler 2009 9 149 6 157 12.5% 1.58 [0.58, 4.33] Rickels 2010 23 190 13 185 28.3% 1.72 [0.90, 3.30] Rosenthal 2012 7 272 4 276 8.5% 1.78 [0.53, 6.00] Tourian 2009 - Desvenlafaxine vs placebo 15 298 4 80 13.5% 1.01 [0.34, 2.95] Tourian 2009 - Duloxetine vs placebo 7 157 3 80 8.5% 1.19 [0.32, 4.48]

Total (95% CI) 1390 939 100.0% 1.39 [0.96, 2.00] Total events 83 40 Heterogeneity: Chi² = 1.46, df = 5 (P = 0.92); I² = 0% 0.01 0.1 1 10 100 Test for overall effect: Z = 1.75 (P = 0.08) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo outcomes halved to provide correct total scores over both arms reported.

Figure 46: Insomnia (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Boyer 2008 15 324 3 161 22.2% 2.48 [0.73, 8.46] Cutler 2009 2 149 3 157 16.2% 0.70 [0.12, 4.15] Rickels 2010 13 190 11 185 61.7% 1.15 [0.53, 2.50]

Total (95% CI) 663 503 100.0% 1.37 [0.75, 2.52] Total events 30 17 Heterogeneity: Chi² = 1.65, df = 2 (P = 0.44); I² = 0% 0.01 0.1 1 10 100 Test for overall effect: Z = 1.03 (P = 0.30) Antidepressant Placebo

Figure 47: Diarrhoea (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Cutler 2009 4 149 4 157 100.0% 1.05 [0.27, 4.14]

Total (95% CI) 149 157 100.0% 1.05 [0.27, 4.14] Total events 4 4 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.07 (P = 0.94) Antidepressant Placebo

Figure 48: Serious adverse events (during taper): antidepressant vs placebo Antidepressant Placebo Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Raskin 2008 0 207 1 104 30.8% 0.05 [0.00, 3.20] Rosenthal 2012 1 272 1 276 69.2% 1.01 [0.06, 16.27]

Total (95% CI) 479 380 100.0% 0.40 [0.04, 4.03] Total events 1 2 Heterogeneity: Chi² = 1.39, df = 1 (P = 0.24); I² = 28% 0.01 0.1 1 10 100 Test for overall effect: Z = 0.77 (P = 0.44) Antidepressant Placebo

Figure 49: Withdrawal syndrome: antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Tourian 2009 - Desvenlafaxine vs placebo 74 298 9 80 51.7% 2.21 [1.16, 4.21] Tourian 2009 - Duloxetine vs placebo 43 157 10 80 48.3% 2.19 [1.16, 4.13]

Total (95% CI) 455 160 100.0% 2.20 [1.40, 3.46] Total events 117 19 Heterogeneity: Chi² = 0.00, df = 1 (P = 0.99); I² = 0% 0.01 0.1 1 10 100 Test for overall effect: Z = 3.41 (P = 0.0007) Antidepressant Placebo

Appendices

Note: Tourian 2009 – desvenlafaxine doses combined; and placebo outcomes halved to provide correct total scores over both arms reported.

Figure 50: Upper respiratory tract infection (during discontinuation period): antidepressant vs placebo Antidepressant Placebo Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Tourian 2009 - Desvenlafaxine vs placebo 9 298 1 80 54.3% 2.42 [0.31, 18.79] Tourian 2009 - Duloxetine vs placebo 8 157 1 80 45.7% 4.08 [0.52, 32.03]

Total (95% CI) 455 160 100.0% 3.17 [0.75, 13.44] Total events 17 2 Heterogeneity: Chi² = 0.12, df = 1 (P = 0.72); I² = 0% 0.01 0.1 1 10 100 Test for overall effect: Z = 1.57 (P = 0.12) Antidepressant Placebo Note: Tourian 2009 – desvenlafaxine doses combined; and placebo outcomes halved to provide correct total scores over both arms reported.

Figure 51: Suicide ideation (time point not reported): antidepressant vs antidepressant Vortioxetine Duloxetine Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Mahableshwarkar 2015 - Vortioxetine vs duloxetine 2 301 2 152 100.0% 0.50 [0.07, 3.55]

Total (95% CI) 301 152 100.0% 0.50 [0.07, 3.55] Total events 2 2 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.69 (P = 0.49) Vortioxetine Duloxetine D.3 Short term versus long term opioid use

Figure 52: Depression

Figure 53: Alcohol abuse

Figure 54: Opioid abuse

Appendices

Figure 55: Other substance abuse

Figure 56: Opioid overdose

Figure 57: Other substance overdose

Figure 58: Opioid dependence

Figure 59: Other substance dependence

Appendices

Figure 60: Mortality

D.4 Short term versus long term tapering

Antidepressants

Figure 61: Discontinuation syndrome (3 or more new items on DESS score at 5-7 days after stopping medication): short taper vs longer taper

3 day taper 14 day taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Tint 2008 7 15 6 13 100.0% 1.01 [0.46, 2.25]

Total (95% CI) 15 13 100.0% 1.01 [0.46, 2.25] Total events 7 6 Heterogeneity: Not applicable 0.05 0.2 1 5 20 Test for overall effect: Z = 0.03 (P = 0.98) 3 day taper 14 day taper

Figure 62: Presence of discontinuation syndrome based on DESS scores (based on 4 or more on the DESS scale during the intervention baseline): abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Khan 2014 31 146 30 139 100.0% 0.98 [0.63, 1.54]

Total (95% CI) 146 139 100.0% 0.98 [0.63, 1.54] Total events 31 30 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.07 (P = 0.94) Abrupt Taper

Figure 63: Taper/post-therapy-emergent adverse events: abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Khan 2014 75 146 54 139 100.0% 1.32 [1.02, 1.72]

Total (95% CI) 146 139 100.0% 1.32 [1.02, 1.72] Total events 75 54 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 2.09 (P = 0.04) Abrupt Taper

Figure 64: Headache: abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Khan 2014 19 146 9 139 100.0% 2.01 [0.94, 4.29]

Total (95% CI) 146 139 100.0% 2.01 [0.94, 4.29] Total events 19 9 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.80 (P = 0.07) Abrupt Taper

Appendices

Figure 65: Nausea: abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Khan 2014 9 146 6 139 100.0% 1.43 [0.52, 3.91]

Total (95% CI) 146 139 100.0% 1.43 [0.52, 3.91] Total events 9 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.69 (P = 0.49) Abrupt Taper

Figure 66: Dizziness: abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Khan 2014 14 146 8 139 100.0% 1.67 [0.72, 3.85]

Total (95% CI) 146 139 100.0% 1.67 [0.72, 3.85] Total events 14 8 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.20 (P = 0.23) Abrupt Taper

Figure 67: Suicide ideation (C-SSRS): abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Khan 2014 1 146 1 139 100.0% 0.95 [0.06, 15.07]

Total (95% CI) 146 139 100.0% 0.95 [0.06, 15.07] Total events 1 1 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.03 (P = 0.97) Abrupt Taper

Figure 68: Suicide attempts: abrupt vs taper (managed taper of 25 mg/day for 1 week)

Abrupt Taper Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Khan 2014 1 146 0 139 100.0% 7.04 [0.14, 355.37]

Total (95% CI) 146 139 100.0% 7.04 [0.14, 355.37] Total events 1 0 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.98 (P = 0.33) Abrupt Taper

Figure 69: Time to first new illness (depressive/panic episode) within 1 year: rapid withdrawal (1-7 days) vs gradual withdrawal (2 weeks or more)

Hazard Ratio Hazard Ratio Study or Subgroup log[Hazard Ratio] SE Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Baldessarini 2010 0.4055 0.14 100.0% 1.50 [1.14, 1.97]

Total (95% CI) 100.0% 1.50 [1.14, 1.97] Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 2.90 (P = 0.004) Rapid withdrawal Gradual withdrawal

Appendices

Cox multivariate modelling of survival functions, with covariates including factors likely to reflect illness severity, such as diagnosis prior recurrence rate, concomitant treatment with a mood stabiliser or other drugs and duration of the first new episode of illness, as well as aspects of treatment including total daily antidepressant dose and duration of index treatment.

Figure 70: DESS after taper week 3: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 3.22 4.82 59 1.78 3.2 59 100.0% 1.44 [-0.04, 2.92]

Total (95% CI) 59 59 100.0% 1.44 [-0.04, 2.92] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 1.91 (P = 0.06) Taper Abrupt

Figure 71: Any adverse event: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 53 101 53 102 100.0% 1.01 [0.78, 1.31]

Total (95% CI) 101 102 100.0% 1.01 [0.78, 1.31] Total events 53 53 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.07 (P = 0.94) Taper Abrupt

Figure 72: Asthenia: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 8 101 5 102 100.0% 1.62 [0.55, 4.77]

Total (95% CI) 101 102 100.0% 1.62 [0.55, 4.77] Total events 8 5 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.87 (P = 0.39) Taper Abrupt

Figure 73: Diarrhoea: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 5 101 6 102 100.0% 0.84 [0.27, 2.67]

Total (95% CI) 101 102 100.0% 0.84 [0.27, 2.67] Total events 5 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.29 (P = 0.77) Taper Abrupt

Figure 74: Dizziness: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 15 101 12 102 100.0% 1.26 [0.62, 2.56]

Total (95% CI) 101 102 100.0% 1.26 [0.62, 2.56] Total events 15 12 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.65 (P = 0.52) Taper Abrupt

Figure 75: Emotional lability: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 4 101 6 102 100.0% 0.67 [0.20, 2.31]

Total (95% CI) 101 102 100.0% 0.67 [0.20, 2.31] Total events 4 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.63 (P = 0.53) Taper Abrupt

Appendices

Figure 76: Headache: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 9 101 12 102 100.0% 0.76 [0.33, 1.72]

Total (95% CI) 101 102 100.0% 0.76 [0.33, 1.72] Total events 9 12 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.66 (P = 0.51) Taper Abrupt

Figure 77: Hypertension: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 2 101 6 102 100.0% 0.34 [0.07, 1.63]

Total (95% CI) 101 102 100.0% 0.34 [0.07, 1.63] Total events 2 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.35 (P = 0.18) Taper Abrupt

Figure 78: Infection: abrupt vs taper (50 mg/d alternate days for two weeks)

Figure 79: Insomnia: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 5 101 6 102 100.0% 0.84 [0.27, 2.67]

Total (95% CI) 101 102 100.0% 0.84 [0.27, 2.67] Total events 5 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.29 (P = 0.77) Taper Abrupt

Figure 80: Nausea: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 10 101 9 102 100.0% 1.12 [0.48, 2.65]

Total (95% CI) 101 102 100.0% 1.12 [0.48, 2.65] Total events 10 9 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.26 (P = 0.79) Taper Abrupt

Figure 81: Sweating: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 3 101 7 102 100.0% 0.43 [0.12, 1.63]

Total (95% CI) 101 102 100.0% 0.43 [0.12, 1.63] Total events 3 7 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.24 (P = 0.22) Taper Abrupt

Appendices

Figure 82: Vasodilation: abrupt vs taper (50 mg/d alternate days for two weeks)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50 mg alternate days 7 101 6 102 100.0% 1.18 [0.41, 3.38]

Total (95% CI) 101 102 100.0% 1.18 [0.41, 3.38] Total events 7 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.30 (P = 0.76) Taper Abrupt

Figure 83: DESS after taper week 3: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 1.7 3.16 47 1.78 3.2 59 100.0% -0.08 [-1.30, 1.14]

Total (95% CI) 47 59 100.0% -0.08 [-1.30, 1.14] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 0.13 (P = 0.90) Taper Abrupt

Figure 84: Any adverse event: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 43 87 53 102 100.0% 0.95 [0.72, 1.26]

Total (95% CI) 87 102 100.0% 0.95 [0.72, 1.26] Total events 43 53 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.35 (P = 0.73) Taper Placebo

Figure 85: Asthenia: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 2 87 5 102 100.0% 0.47 [0.09, 2.36]

Total (95% CI) 87 102 100.0% 0.47 [0.09, 2.36] Total events 2 5 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.92 (P = 0.36) Taper Placebo

Figure 86: Diarrhoea: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 5 87 6 102 100.0% 0.98 [0.31, 3.09]

Total (95% CI) 87 102 100.0% 0.98 [0.31, 3.09] Total events 5 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.04 (P = 0.97) Taper Placebo

Figure 87: Dizziness: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 10 87 12 102 100.0% 0.98 [0.44, 2.15]

Total (95% CI) 87 102 100.0% 0.98 [0.44, 2.15] Total events 10 12 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.06 (P = 0.95) Taper Placebo

Appendices

Figure 88: Emotional lability: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 4 87 6 102 100.0% 0.78 [0.23, 2.68]

Total (95% CI) 87 102 100.0% 0.78 [0.23, 2.68] Total events 4 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.39 (P = 0.70) Taper Placebo

Figure 89: Headache: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 7 87 12 102 100.0% 0.68 [0.28, 1.66]

Total (95% CI) 87 102 100.0% 0.68 [0.28, 1.66] Total events 7 12 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.84 (P = 0.40) Taper Placebo

Figure 90: Hypertension: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 2 87 6 102 100.0% 0.39 [0.08, 1.89]

Total (95% CI) 87 102 100.0% 0.39 [0.08, 1.89] Total events 2 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.17 (P = 0.24) Taper Placebo

Figure 91: Infection: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 3 87 6 102 100.0% 0.59 [0.15, 2.28]

Total (95% CI) 87 102 100.0% 0.59 [0.15, 2.28] Total events 3 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.77 (P = 0.44) Taper Placebo

Figure 92: Insomnia: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 5 87 6 102 100.0% 0.98 [0.31, 3.09]

Total (95% CI) 87 102 100.0% 0.98 [0.31, 3.09] Total events 5 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.04 (P = 0.97) Taper Placebo

Figure 93: Nausea: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 7 87 9 102 100.0% 0.91 [0.35, 2.35]

Total (95% CI) 87 102 100.0% 0.91 [0.35, 2.35] Total events 7 9 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.19 (P = 0.85) Taper Placebo

Appendices

Figure 94: Sweating: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 4 87 7 102 100.0% 0.67 [0.20, 2.21]

Total (95% CI) 87 102 100.0% 0.67 [0.20, 2.21] Total events 4 7 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.66 (P = 0.51) Taper Placebo

Figure 95: Vasodilation: abrupt vs taper (50 mg/d for 7 days then placebo for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then placebo 1 87 6 102 100.0% 0.20 [0.02, 1.59]

Total (95% CI) 87 102 100.0% 0.20 [0.02, 1.59] Total events 1 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.53 (P = 0.13) Taper Placebo

Figure 96: DESS after taper week 3: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 4.11 5.77 57 1.78 3.2 59 100.0% 2.33 [0.62, 4.04]

Total (95% CI) 57 59 100.0% 2.33 [0.62, 4.04] Heterogeneity: Not applicable -100 -50 0 50 100 Test for overall effect: Z = 2.68 (P = 0.007) Taper Abrupt

Figure 97: Any adverse event: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 47 94 53 102 100.0% 0.96 [0.73, 1.27]

Total (95% CI) 94 102 100.0% 0.96 [0.73, 1.27] Total events 47 53 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.27 (P = 0.78) Taper Placebo

Figure 98: Asthenia: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 0 94 5 102 100.0% 0.14 [0.02, 0.83]

Total (95% CI) 94 102 100.0% 0.14 [0.02, 0.83] Total events 0 5 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 2.17 (P = 0.03) Taper Placebo

Figure 99: Diarrhoea: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 3 94 6 102 100.0% 0.54 [0.14, 2.11]

Total (95% CI) 94 102 100.0% 0.54 [0.14, 2.11] Total events 3 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.88 (P = 0.38) Taper Placebo

Appendices

Figure 100: Dizziness: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 8 94 12 102 100.0% 0.72 [0.31, 1.69]

Total (95% CI) 94 102 100.0% 0.72 [0.31, 1.69] Total events 8 12 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.75 (P = 0.46) Taper Placebo

Figure 101: Emotional lability: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 1 94 6 102 100.0% 0.18 [0.02, 1.47]

Total (95% CI) 94 102 100.0% 0.18 [0.02, 1.47] Total events 1 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.60 (P = 0.11) Taper Placebo

Figure 102: Headache: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 7 94 12 102 100.0% 0.63 [0.26, 1.54]

Total (95% CI) 94 102 100.0% 0.63 [0.26, 1.54] Total events 7 12 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.01 (P = 0.31) Taper Placebo

Figure 103: Hypertension: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Peto Odds Ratio Peto Odds Ratio Study or Subgroup Events Total Events Total Weight Peto, Fixed, 95% CI Peto, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 0 94 6 102 100.0% 0.14 [0.03, 0.70]

Total (95% CI) 94 102 100.0% 0.14 [0.03, 0.70] Total events 0 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 2.38 (P = 0.02) Taper Placebo

Figure 104: Infection: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 2 94 6 102 100.0% 0.36 [0.07, 1.75]

Total (95% CI) 94 102 100.0% 0.36 [0.07, 1.75] Total events 2 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.27 (P = 0.21) Taper Placebo

Figure 105: Insomnia: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 5 94 6 102 100.0% 0.90 [0.29, 2.87]

Total (95% CI) 94 102 100.0% 0.90 [0.29, 2.87] Total events 5 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.17 (P = 0.86) Taper Placebo

Appendices

Figure 106: Nausea: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 9 94 9 102 100.0% 1.09 [0.45, 2.62]

Total (95% CI) 94 102 100.0% 1.09 [0.45, 2.62] Total events 9 9 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.18 (P = 0.86) Taper Placebo

Figure 107: Sweating: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 2 94 7 102 100.0% 0.31 [0.07, 1.46]

Total (95% CI) 94 102 100.0% 0.31 [0.07, 1.46] Total events 2 7 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 1.48 (P = 0.14) Taper Placebo

Figure 108: Vasodilation: abrupt vs taper (50 mg/d for 7 days then 25 mg/d for 7 days)

Taper Abrupt Risk Ratio Risk Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Gallagher 2012 - 50mg then 25mg 5 94 6 102 100.0% 0.90 [0.29, 2.87]

Total (95% CI) 94 102 100.0% 0.90 [0.29, 2.87] Total events 5 6 Heterogeneity: Not applicable 0.01 0.1 1 10 100 Test for overall effect: Z = 0.17 (P = 0.86) Taper Placebo

Evidence tables Evidence Harms

Appendix E: Evidence tables

Harms of dependency on a medication

Opioids Study Cepeda, 2013(6) Study type Retrospective cohort Number of studies (number of participants) 1 (n=155,761)

Countries and setting USA, IMS LRx database Duration of study 1 year Inclusion criteria People exposed to tapentadol immediate release (IR) or oxycodone IR from July 2009 to December 2010 who had not

172 received an opioid of any type in the 3 months before the index date. Index date was defined as the date of the first

prescription for tapentadol IR or oxycodone IR after June 30, 2009. Exclusion criteria People who filled a prescription for an opioid other than tapentadol IR or oxycodone IR on the index date on in the next 3 days. Intended to ensure that subjects did not have the outcome of interest before the study started, and to isolate the exposure to the indexed opioid Recruitment/selection of patients Retrospective review of IMS LRx database. Each tapentadol IT-exposed subject was matched to up to 4 randomly selected oxycodone IR-exposed subjects by: calendar quarter and year of initial exposure (index date); fist 3 digits of the zip code of the pharmacy dispensing the opioid at the index date; age ±5 years ; and specialty of prescriber. A matching ration of 1:4 was used. Age, gender and ethnicity Age, mean (SD), y: Oxycodone 51.1 (14.9), Tapentadol 50.1 (14.9) Sex, men, No. (%): Oxycodone 54,450 (48.3), Tapentadol 16,002 (37.3) History of benzodiazepine use, no. (%): Oxycodone 19,324 (17.1), Tapentadol 5,171 (12.0) Prescriber specialty, No (%): Addiction medicine: Oxycodone 0 (0), Tapentadol 0 (0) Dentistry: Oxycodone 72 (0.06), Tapentadol 47 (0.1) Emergency medicine: Oxycodone 2,478 (2.2), Tapentadol 946 (2.2) Surgery: Oxycodone: 17,559 (15.6), Tapentadol 8,203 (19.1) Pain medicine: Oxycodone 6,406 (5.7), Tapentadol 3,363 (7.8)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Primary care medicine: Oxycodone 44,769 (39.7), Tapentadol 15,052 (35.0) Other: Oxycodone 41,537 (36.8), Tapentadol 15,329(35.7)

Extra comments Classification by specialist was used to differentiate subjects having low and high risk of opioid shopping behavior. Shopping behavior defined as more than 1 prescription by at least 2 different prescribers with at least 1 day of overlap

and filled at 3 or more pharmacies. Gender, any exposure to benzodiazepines during the 3 months before the index date, and type of payment at the index date were also considered as potential confounders in the analysis. Indirectness of population Likely that not all participants were using opioids for chronic pain. Specialties covered included: Primary care, Orthopaedic surgery or general surgery, pain medicine, dentistry, emergency medicine, addiction medicine, and other. Length of opioid use also not defined, therefore cannot assume all participants were dependent. (NB None were listed as addiction medicine)

Interventions & comparators Tapentadol (n=42,940) Oxycodone (n=112,821) Funding Janssen Research & Development, LLC funded the study including collection and analysis of the data. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: hypnotic users vs controls

173 Protocol outcome 1: Harm – Opioid shopping behaviour

Actual outcome: Proportion of subjects who developed shopping behaviour during any point of 1 year follow-up, No. (%, 95% CI): Oxycodone 967 (0.87, 0.80-0.91), Tapentadol 88 (0.20, 0.16 – 0.25) Risk of bias: All domain – Very High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - High, Outcome reporting - Low, Measurement – High, Serious indirectness; Unclear if shopping episodes were due to dependence of prescribed drug. Not all using opioids for chronic pain. No. missing not reported only time in database.

Actual outcome: Proportion of subjects who developed shopping behaviour during any point of 1 year follow-up, Adjusted OR (95% CI): 3.5 (2.8-4.4) Risk of bias: All domain – Very High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - High, Outcome reporting - Low, Measurement – High, Serious indirectness; Unclear if shopping episodes were due to dependence of prescribed drug. Not all using opioids for chronic pain. No. missing not reported only time in database.

Actual outcome: Harm – Number of shopping episodes per subject during the year of follow-up, mean (SD): Oxycodone, 0.02 (0.3), Tapentadol 0.004 (0.1) Risk of bias: All domain – Very High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - High, Outcome reporting - Low, Measurement – High, Serious indirectness; Unclear if shopping episodes were due to dependence of prescribed drug. Not all using opioids for

Evid Harms © Royal © College of Physicians 2019 chronic pain. No. missing not reported only time in database. ence tables

Actual outcome: Harm – Number of shopping episodes per subject during the year of follow-up, mean difference (95% CI): 0.15 (0.01-0.02) Risk of bias: All domain – Very High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - High,

Outcome reporting - Low, Measurement – High, Serious indirectness; Unclear if shopping episodes were due to dependence of prescribed drug. Not all using opioids for chronic pain. No. missing not reported only time in database.

Also reported mean shopping events per subject who developed shopping behaviour, time to first shopping episode, whether the dispensing in the shopping episodes was for the same opioid or not and a post-hoc analysis of risk of developing heavy shopping behaviour.

Harms from stopping medications over a short time frame

Opioids Study Afilalo 2010(2)

174 Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n=1030 randomised). Countries and setting 87 sites in the US and 15 sites in Canada, 6 sites in New Zealand and 4 sites in Australia. Duration of study Titration 3 weeks, maintenance 12 weeks and follow-up 2 weeks after last intake of medication. Conducted from 7 February 2007 to 4 June 2008. Inclusion criteria Men and women ≥ 40years of age with a diagnosis of osteoarthritis for the knee, functional capacity class I –III, and apian at the reference joint requiring the use of analgesics (non-opioids or opioids at doses equivalent to ≤ 160 mg oral morphine/day) for ≥ 3 months prior to screening. A patient rated 11-point numerical rating scale (0=no pain, 10=pain as bad as you can imagine) was used to assess pain intensity twice daily. Patients were dissatisfied with their current analgesic therapy and had average pain intensity numerical rating scale score of ≥ 5 during the 3 days preceding randomization. Exclusion criteria Presence of clinically significant or unstable medical or psychiatric disease, requirement for painful procedures during the study that could influence efficacy or safety assessments, and history of substance abuse, epilepsy/seizure disorder, stroke; transient ischemic attach, malignancy, HIV, chronic hepatitis B or C, uncontrolled hypertension, severe renal impairment, moderate or severe hepatic impairment, ALT or AST concentrations over 3 times the upper limit of normal, and hypersensitivity to study medications or their excipients. Patients with conditions potentially influencing g the

Evidence tables Evidence Harms © Royal © College of Physicians 2019 assessment of osteoarthritis pain were excluded. The use of concomitant analgesics was prohibited. Neuroleptics, tricyclic antidepressants, anticonvulsants, antiparkinsonian drugs and serotonin norepinephrine reuptake inhibitors were prohibited within 14 days prior to screening ad during the study because their use could confound efficacy or safety assessment. Medications other than those listed above such as SSRI were allowed for patients with diagnosed

controlled psychiatric or neurological conditions if taken at a stable dose for 3 or more months prior to randomization. Monoamine oxidase inhibitors were prohibited within 14 days prior to screening and during the std. Corticosteroids were prohibited during the trial and within 4 weeks to 6 months prior to screening, depending on route of administration. Recruitment/selection of patients Recruited patients with moderate to severe chronic pain of osteoarthritis related to the knee. Age, gender and ethnicity Age, y, Mean (SD): Group 1: 58.4 (10.09), Group 2: 58.2 (10.29), Group 3: 58.2 (9.15) Male %: Group 1: 37.2%, Group 2: 40.9%, Group 3: 40.7%

Race White: Group 1: 75.6%, Group 2: 71.6%, Group 3: 79.2% Black: Group 1: 14.2%, Group 2: 13.2%, Group 3: 11.3% Hispanic: Group 1: 6.1%, Group 2: 10.8%, Group 3: 5.9%

175 Other: Group 1: 4.1%, Group 2: 4.4%, Group 3: 3.6%

Authors reported that demographic and baseline characteristics were balanced across groups. Extra comments Efficacy, quality of life and treatment emergent adverse events were reported. Indirectness of population No indirectness. Interventions & comparators Period 1: Screening (≤ 14 days) Period 2: Washout (3-7 days, during which patients were to discontinue all analgesic medication) Period 3: Titration (3 weeks) Period 4: Maintenance (12 weeks) Period 5: Follow-up (14 days after last intake of study medication).

(n=346) Group 1: Tapentadol, twice daily controlled, adjustable, oral doses of tapentadol ER 100-250 mg. Started with twice daily dose of tapentadol ER 50 mg. After the first 3 days, doses were increased to 100 mg twice daily; these were the minimum doses for the remainder of the study. At 3-day intervals, paints could increase their doses in consultation with a study investigator in twice –daily increments of tapentadol ER 50 mg (maximum twice daily doses of tapentadol ER 250 mg); downward titration was possible in twice daily decrements of tapentadol R 50 mg without a time restriction. All doses taken in the morning and evening. Paracetamol could be taken up to 3 days before the

Evidence tables Evidence Harms © Royal © College of Physicians 2019 conclusion of the titration period. During the maintenance period patients were encouraged to remain on a steady dose of study medication but could request additional dose adjustment to maintain their optimal balance.

(n=345) Group 2: Oxycodone HCI CR 20-50 mg Started with twice daily dose of oxycodone HCI CR 10mg. After the first 3 day s, doses were increased to 20 mg twice daily; these were the minimum doses for the remainder of the study. At 3-day intervals, paints could increase their doses in consultation with a study investigator in twice –daily increments of oxycodone HCI CR 10 mg (maximum twice daily doses of oxycodone HCI CR 50 mg); downward titration was possible in twice daily decrements of oxycodone HCI CR 10mg without a time restriction. All doses taken in the morning and evening. Paracetamol could be taken up to 3 days before the conclusion of the titration period. During the maintenance period patients were encouraged to remain on a steady dose of study medication but could request additional dose adjustment to maintain their optimal balance.

(n=339) Group 3: Placebo Funding Johnson & Johnson Pharmaceutical Research and Development, L.L.C. Afilalo received funding for study support from Johnson and Johnson Pharmaceutical Research and Development, L.L.C. Most other authors are employees of 176 Grunenthal GmbH.

RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: tapentadol versus oxycodone versus placebo

Protocol outcome 1: Harm – Withdrawal symptoms Actual outcome: COWS scores for all treatment group for all time periods (in patients who did not use opioids following discontinuation of study medication)

COWS assessments completed ≥ 2 days to <5 days after last intake of medication: No opioid withdrawal: Group 1: 29/35, Group 2: 32/37, Group 3: 23/23. Mild opioid withdrawal: Group 1: 6/35, Group 2: 5/37, Group 3: 0/23.

COWS assessments completed ≥ 5 days after last intake of study medication: No opioid withdrawal: Group 1: 69/70, Group 2: 72/84, Group 3: 54/59 Mild opioid withdrawal: Group 1: 1/70, Group 2: 10/84, Group 3: 5/59 Moderate opioid withdrawal: Group 1: 0/70, Group 2: 2/84, Group 0/59

Risk of bias:

Evidence tables Evidence Harms © Royal © College of Physicians 2019 All domain – High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 163; Group 2 Number missing: 224, Group 3 Number missing: 134, Reason*: patient choice (141), lost to follow-up (8), adverse event (223), lack of efficacy (57), study drug noncompliance (17), other (74), no drug received (7).

Narrative results: In the subgroup of patients eligible for SOWS assessments, there were no statistically significant differences in LSM SOWS total scores in the tapentadol ER and placebo groups at 3, 4 or ≥ 5 days after the last dose of study drug.

Study Buynak 2010(5)

Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=981 randomised). Countries and setting Conducted in US (85 sites), Canada (15 sites), and Australia (3 sites).

177 Duration of study Conducted from 21 February 2007 to 12 March 2008.

15 weeks (3 week titration period and 12 week maintenance period). Inclusion criteria Men and women aged 18 years and over, with a history of nonmalignant low back pain for 3 or more months prior to the study were eligible. Patients were required to have been taking analgesics for low back pain for ≥ 3 months prior to screening, to be dissatisfied with their current treatment and to have a baseline pain intensity of ≥ 5 on an 11-pint numerical rating scale after a 3 to 7 day washout period of all previous analgesics. For patients taking opioid analgesics, daily doses of opioids had to be equivalent to ≤ 160 mg of oral morphine. Exclusion criteria Patients in another clinical study or who had previously participated in other tapentadol studies were excluded. Neuroleptics, tricyclic antidepressants, anticonvulsants, antiparkinsonian drugs, and serotonin norepinephrine reuptake inhibitors were prohibited within in 14 days prior to screening and during the study. Patients with diagnosed psychiatric or neurologic conditions could participate if they were treated with medications not listed above. Corticosteroids were also prohibited during the trial and within timelines prior to screening. The use of concomitant analgesics, with the exception of allowed doses of acetaminophen was prohibited during the study. Transcutaneous Electrical Nerve Stimulation, acupuncture, physical therapy, packs, massages, and other interventional adjunctive therapy were permitted during the study if patients started the treatment more than 14 days prior to enrollment and continued on the same regimen during the study. Other reasons for exclusion were the presence of a clinically significant medical or psychiatric disease, requirement for painful procedures, surgery in the low back area within 3 months of screening, history of substance abuse, epilepsy/seizure disorder, stroke/transient ischemic attack, HIV, chronic hepatitis B or C,

Evidence tables Evidence Harms © Royal © College of Physicians 2019 malignancy, uncontrolled hypertension, server renal impairment, moderate or severe hepatic impairment, abnormal baseline laboratory values that might affect patient safety and hypersensitivity to the study medications or their excipients. Patients with conditions potentially influencing the self-assessment of blowback pain were also excluded. Recruitment/selection of patients Patients with moderate to severe chronic low back pain. Age, gender and ethnicity Age, y, Mean (SD): Group 1: 49.4 (13.21), Group 2: 50 (14.21), Group 3: 50.4 (14.05) Male %: Group 1: 39.0%, Group 2: 44.8%, Group 3: 42.3%

Race White: Group 1: 72.0%, Group 2: 73.5%, Group 3: 74.3% Black: Group 1: 19.5%, Group 2: 16.8%, Group 3: 15.7% Hispanic: Group 1: 5.7%, Group 2:6.4%, Group 3: 6.6% Other: Group 1: 2.8%, Group 2: 3.4%, Group 3: 3.4%

Authors state that the demographic and baseline characteristics were similar between treatment groups. Extra comments Efficacy, quality of life and treatment emergent adverse events reported. Indirectness of population No indirectness.

178 Interventions & comparators Titration (3 weeks)

Maintenance period (12 weeks) Follow up period consisted of a visit 4 days after end of treatment and a post-study telephone call. Patients who completed the study were given the option to continue treatment in an open label extension study.

(n=321) Group 1: Tapentadol: controlled adjustable doses of tapentadol ER 100 – 250 mg b.i.d. (n=334) Group 2: Oxycodone HCI controlled release 20 – 50 mg b.i.d. (n=326) Group 3: Placebo b.i.d.

Funding Funded by Johnson and Johnson Pharmaceutical Research and Development L.L.C., and Global Development, Grunenthal GmbH, Aachen, Germany. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: tapentadol versus oxycodone versus placebo

Protocol outcome 1: Harm – Withdrawal symptoms Actual outcome: COWS; at 4 day follow-up visit for patients that had abrupt discontinuation without tapering and did not enter the open label study (COWS scores ranged from 0 to 48; with a higher score indicating more severe opioid withdrawal) No opioid withdrawal symptoms: Group (n=62): 59/62 (95.2%), Group 2 (n=90): 82/90 (91.1%), Group 3 (n=59): 53/59 (89.8%)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Mild/moderate opioid withdrawal symptoms: Group 1 (n=62): 3/62 (4.8%), Group 2 (n=90): 8/90 (8.9%), Group 3 (n=59): 6/59 (10.2%)

Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of

outcome: No indirectness; At end of study - Group 1 Number missing: 152; Group 2 Number missing: 195, Group 3 Number missing: 167, Reason*: Study drug noncompliant (36), patient choice (141), lost to follow-up (33), adverse event (173), lack of efficacy (70), other (61), not taken single dose (14).

Study Hale 2009(8) Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n=878 randomised). Countries and setting Multi-centre study in Canada and US. Duration of study Conducted from 31 July 2006 to 17 July 2007. Screening period 4 weeks, double-blind treatment period 3 months and a 3-day follow up period.

179 Inclusion criteria Men and women 18 years and over with a clinical diagnosis and at least a 3-month history of lower back pain of non-

malignant origin or osteoarthritis (OA) pain of the knee or hip were eligible. All patients demonstrated a qualifying pain intensity (PI) score of at least 4 on an 11-point numerical rating scale (NRS; 0=no pain, 10=worst pain imaginable) while taking non-opioid analgesics or following a 24-hour washout of opioid analgesics. Prior treatment consisted of NSAIDS or opioids (equivalent to no more than 80 mg morphine per day) taken alone or in combination for at least 30 days prior to enrollment and stable regimens of non-opioid analgesics were continued throughout the study. Exclusion criteria Patients with uncontrolled hypertension, uncontrolled or poorly controlled psychiatric disorders, or other medical conditions. Also patients with a history of chronic hepatitis B or C or HIV infection. Patents with a history of malignancy, seizure disorders, epilepsy, or alcohol or drug abuse were also excluded. The presence of concomitant autoimmune inflammatory conditions, acute crystal-induced arthropathy or moderate to severe renal or hepatic insufficiency was also reason for exclusion. Patients who were receiving systemic steroid therapy within 3 months of study were prohibited. Other prohibited medications included monoamine oxidase inhibitors, tricyclic antidepressants, or neuroleptics. Selective serotonin reuptake inhibitors and serotonin norepinephrine reuptake inhibitors were allowed if taken for at least 30 days at a stable dose prior to study screening. Patient who had received an experimental drug or used an experimental device within 30 days of study entry were excluded form participation, as were patients with a history of sensitivity to tapentadol, oxycodone, hydromorphone, morphine or fentanyl or their excipients. Pregnant or nursing women were excluded.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Recruitment/selection of patients Patients with a clinical diagnosis of low back pain or OA pain of the knee or hip for at least 3 months. Recruited from multiple primary and specialty care treatment centres.

Age, gender and ethnicity Age, y - Mean (SD): Group 1: 55.9 (11.96), Group 2: 56.4 (11.86) Female: Group 1: 54.6%, Group 2: 55.9%

Pain intensity score, Mean (SD): Group 1: 7.0 (1.60), Group 2: 7.2 (1.40) Opioid experienced, n (%) with yes: Group 1: 49.0%, Group 2: 48.2% Authors Baseline characteristics were similar between treatment group. Extra comments Efficacy and treatment adverse events also reported. Indirectness of population No indirectness. Interventions & comparators Screening period (days -28 to -1) Double-blind treatment period (3 months, day 1-91) Follow up period (3 days from day 93-95).

(n=679) Group 1: Tapentadol immediate release (IR) (flexible oral dose of 50 or 100 mg, every 4-6 hours as needed to a maximum dose of 600mg per day.

180 (n=170) Group 2: Oxycodone IR (flexile oral dose of 10 or 15 mg, every 4-6 hours as needed to a maximum dose of 90 mg per day.

All patients: patients receiving non-opioid analgesics at study entry were allowed to continue that regimen during the study, provided that it did not change. In addition, rescue analgesics were allowed for no more than 3 days each week while on study drug. Opioid rescue medication was not permitted. Funding Sponsored by Johnson and Johnson Pharmaceutical Research and Development, L.L.C., Raritan, NH, USA and Grunenthal GmbH, Aachen, Germany. 3 of the authors are employees and stockholders of Johnson and Johnson, while another is an employee of Grunenthal GmbH. The coordinating investigator of the clinical trial was contracted by, but no an employee of, Gold Coast Research. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Tapentadol versus oxycodone

Protocol outcome 1: Harm – Withdrawal symptoms - Actual outcome: % of patients detecting Clinical Opioid Withdrawal Scale (COWS) after cessation of study drug. COWS is a clinician rated 11-item scale evaluating physical components of opioid withdrawal based on question and clinical observations. Total score form 0-48 with high scores indication more severe symptoms of opioid withdrawal. Group 1: (n=306) 17%, Group 2: (n=66) 29%; indicated withdrawal symptoms which were only of mild to moderate intensity

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 397; Group 2 Number missing: 109, Reason*: Withdrawn (372), patient choice (89), lost to follow-up (29), adverse event (197), other (65).

Protocol outcome 1: Harm – Withdrawal symptoms - Actual outcome: Mean total Subjective Opioid Withdrawal Scale (SOWS) score. SOWS is a 16-item self-reported measure of the severity of symptoms associated with rapid opioid detoxification. Higher scores indicating greater severity of withdrawal symptoms. - Group 1: (n=308) 6.9, Group 2: (n=67) 8.7 Risk of bias: All domain – High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 397; Group 2 Number missing: 109, Reason*: Withdrawn (372), patient choice (89), lost to follow-up (29), adverse event (197, other (65).

* Reasons for participants missing were provided for those that did not complete study. However, not all participants that completed the trial reported withdrawal symptoms.

181 Study Wild 2010(17)

Study type RCT (Patient randomised; Parallel) Open label. Number of studies (number of participants) 1 (n=1121 randomised; n=1117 received study drug). Countries and setting 53 study sites in North America and 36 sites in Europe. Duration of study 1 year. November 2006 to July 2008. Inclusion criteria Men and non-pregnant, nonlactating women 18 years or older with a clinical diagnosis of moderate to severe knee or hip osteoarthritis pain or low back pain of nonmalignant origin, with at least a 3-month history of pain prior to screening, and who were dissatisfied with their current analgesic therapy were eligible. Patients were require to have a pain intensity score of at least 4 on a 11-point numerical rating scale at baseline following a 3-to 7-day washout of all prior analgesic treatments. Exclusion criteria Lifelong history of seizure disorder or epilepsy; mild or moderate traumatic brain injury, stroke, transient ischemic attack, or brain neoplasm within 1 year of screening; severe traumatic brain injury within 15 years of screening; residual sequelae suggesting transient changes in consciousness; history of malignancy within 2 years of screening, with the exception of successfully treated basal cell carcinoma; history of alcohol or drug abuse; history of chronic hepatitis B or C

Evidence tables Evidence Harms © Royal © College of Physicians 2019 or active hepatitis B or C within 3 months of screening; history of HIV; a clinically relevant history of hypersensitivity, Allery, or contraindication to oxycodone or acetaminophen; or previous participation in this study or other studies of tapentadol. Patients requiring major surgery during the study or who had surgery of the back or reference joint within 3 months of screening were also excluded. Patients with moderately or severely impaired hepatic function, severely

impaired renal function, uncontrolled hypertension, clinically significant disease that could affect efficacy or safety assessment or significant pain associated with conditions other than osteoarthritis or low back pain were excluded. Recruitment/selection of patients Men and non-pregnant, nonlactating women 18 years or older with a clinical diagnosis of moderate to severe knee or hip osteoarthritis pain or low back pain of nonmalignant origin, with at least a 3-month history of pain prior to screening, and who were dissatisfied with their current analgesic therapy were eligible. Age, gender and ethnicity Age, y - Mean (SD): Group 1: 56.8 (12.51), Group 2: 58.1 (11.83) Female: Group 1: 57.6%, Group 2: 56.1%

Race: White: Group 1: 88.6%, Group 2: 91.0% Black: Group 1: 6.7%, Group 2: 5.8% Hispanic: Group 1: 2.9%, Group 2: 1.8%

182 Other: Group 1: 1.8%, Group 2: 1.3%

Prior opioid experience: Group 1: 52.9%; Group 2: 50.2%

Authors reported demographic and baseline characteristics were similar. Extra comments Efficacy and treatment adverse events reported. Indirectness of population No indirectness. Interventions & comparators Screening period (up to 14 days) Washout period (3 -7 days prior to randomization) titration period (1 week) Maintenance period (51 weeks) Follow-up period (visit 4 days post treatment and a telephone call 10-14 days post treatment).

(n=894) Group 1: Controlled, adjustable twice daily doses of tapentadol ER (100-250mg) Titration period in first 3 days – patients received twice daily oral doses of ER 50 mg tapentadol. Doses then increased to ER 100 mg bid for the next 4 days, which was the minimum therapeutic dose for the study. Under the supervision of a study physician, patients could adjust their doses during the 51-week maintenance period to maintain an optimal balance of efficacy and tolerability. Doses could be titrated upward in twice daily increments of

Evidence tables Evidence Har © Royal © College of Physicians 2019 tapentadol ER 50mg at a minimum of 3-day intervals to a maximum therapeutic dose of tapentadol ER 250 mg bid.

ms Downward titration was possible in twice daily decrements of tapentadol ER 50mg with no time restriction, but doses could not fall below the minimum therapeutic doses.

(n=223) Group 2: Oxycodone HCI controlled release (CR; 20-50 mg). Titration period in first 3 days – patients received twice daily oral doses of oxycodone HCI CR 10 mg. Doses increased to 20 mg bid for the next 4 days, which was the minimum therapeutic dose for the study. Under the supervision of a study physician, patients could adjust their doses during the 51-week maintenance period to maintain an optimal balance of efficacy and tolerability. Doses could be titrated upward in twice daily increments of oxycodone HCI CR 10 mg at a minimum of 3-day intervals to a maximum therapeutic dose of oxycodone HCI CR 50 mg bid. Downward titration was possible in twice daily decrements of oxycodone HCI CR 10 mg with no time restriction, but doses could not fall below the minimum therapeutic doses.

Funding Funded by Johnson and Johnson Pharmaceutical Services, L.L.C and Global Development, Grunenthal GmbH, Aachen, Germany. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Tapentadol versus oxycodone

183

Protocol outcome 1: Harm – Withdrawal syndrome - Actual outcome: Treatment emergent adverse event – withdrawal syndrome within 3 days of last dose (reported by investigator as withdrawal symptoms, withdrawal syndrome or related unspecified symptoms such as anxiety, nausea, shivering agitation, tremor, tachycardia, or sweating). Group 1: 13/894 (1.5%), Group 2: 2/223 (0.9%) Risk of bias: All domain – Very high, Selection - Low, Blinding - High, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 481; Group 2 Number missing: 145, Reason*: Adverse event (285), patient choice (125), noncompliant (57), lack of efficacy (79), lost to follow-up (47), resolution of pain (2), other (31).

Protocol outcome 2: Harm – Drug withdrawal syndrome - Actual outcome: Treatment emergent adverse event – Drug withdrawal syndrome Group 1: 9/894 (1.0%), Group 2: 1/223 (0.4%) Risk of bias: All domain – Very high, Selection - Low, Blinding - High, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 481; Group 2 Number missing: 145, Reason*: Adverse event (285), patient choice (125), noncompliant (57), lack of efficacy (79), lost to follow-up (47), resolution of pain (2), other (31).

Note: Of the patients in group 1 that reported either withdrawal syndrome or drug withdrawal syndrome, 8 patients had corresponding COWS assessments at the end of

Evidence tables Evidence Harms © Royal © College of Physicians 2019 treatment that indicated mild opioid withdrawal and 3 patients had COWS assessments that indicated moderate opioid withdrawal. Nine of the patients that reported withdrawal had corresponding COWS assessments that indicated no opioid withdrawal. One patient in group 2 and reported withdrawal syndrome or drug withdrawal syndrome had corresponding COWS data that showed mild opioid withdrawal.

Protocol outcome 3: Harm – Withdrawal symptoms - Actual outcome: Clinical Opioid Withdrawal Scale (COWS) ≥ 2 to 4 days after treatment discontinuation for patients who did not take opioids following discontinuation: No opioid withdrawal: Group 1: 97/125 (77.6%), Group 2: 16/22 (72.7%) Mild opioid withdrawal: Group 1: 22/125 (17.6%), Group 2: 5/22 (22.7%) Moderate opioid withdrawal: Group 1: 6/125 (4.8%), Group 2: 1/22 (4.5%) Moderately severe or severe opioid withdrawal: Group 1: 0/125 (0%), Group 2: 0/22 (0%) Risk of bias: All domain – Very high, Selection - Low, Blinding - High, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 481; Group 2 Number missing: 145, Reason*: Adverse event (285), patient choice (125), noncompliant (57), lack of efficacy (79), lost to follow-up (47), resolution of pain (2), other (31).

184 Protocol outcome 4: Harm – Withdrawal symptoms - Actual outcome: Clinical Opioid Withdrawal Scale (COWS) ≥ 5 days after treatment discontinuation for patients who did not take opioids following discontinuation:

No opioid withdrawal: Group 1: 146/166 (88.0%), Group 2: 42/50 (84.0%) Mild opioid withdrawal: Group 1: 18/166 (10.8%), Group 2: 7/50 (14.0%) Moderate opioid withdrawal: Group 1: 2/166 (1.2%), Group 2: 1/50 (2.0%) Risk of bias: All domain – Very high, Selection - Low, Blinding - High, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 481; Group 2 Number missing: 145, Reason*: Adverse event (285), patient choice (125), noncompliant (57), lack of efficacy (79), lost to follow-up (47), resolution of pain (2), other (31).

Protocol outcome 5: Harm – Withdrawal symptoms - Risk of bias: Harm –SOWs – abbreviated 15 item version with assessments scored form 0-60, with a score of 60 indicating extremely severe opioid withdrawal and reported after last dose of study medication. Mean range SOWS total scores from 2, 3, 4 and 5 or more days after discontinuation: Group 1: (n=164): 6.9-9.5, Group 2: (n=46): 7.5-12.3 Risk of bias: All domain – Very high, Selection - Low, Blinding - High, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 481; Group 2 Number missing: 145, Reason*: Adverse event (285), patient choice (125), noncompliant (57), lack of efficacy (79), lost to follow-up (47), resolution of pain (2), other (31).

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Z-drugs

Study Hajak(7) Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n=742 randomised) Countries and setting Conducted at 73 centres in Europe and Canada. Duration of study Patients enrolled between 29 January 2003 and 1 September 2004. Study run over 2 weeks with a run-out phase over 1 week where 50% of patients previously assigned to active treatment were randomised to placebo, while the other 50% of patients remained on their existing treatment. There was a 1-week safety follow-up period after the end of the run-out phase. Inclusion criteria Outpatients aged 18-65 years who met Diagnostic and Statistical manual of Mental disorder (DSM) -IV criteria for

primary insomnia. Patients also had to have a usual bedtime between 20:00 and 24:00 h and had to meet the following criteria: self-reported total sleep time less than 6 hours and self-reported time-to-sleep onset 45 minutes or more on at least 4 out of 7 nights. Exclusion criteria Patients with a history or current abuse or dependence on any substance with abuse potential, any psychotic disorder,

185 or any other current Axis I diagnosis other than primary insomnia were excluded. The following medications were prohibited prior to the screening visit: within 1 week – any psychoactive or hypnotic drug, benzodiazepines, any opiate or opiate derivative containing drug, itraconazol or rifampicin; within 2 weeks – sedating antihistamines, antiepileptics, flurazepam, or diazepam; within 5 weeks – fluoxetine. Patients were excluded if they consumed more than 14 units (women) or 21 units (men) of alcohol a week, or consumed more than 5 caffeine-containing beverages a day, or consumed more than the nicotine-equivalent of 15 cigarettes a day. Recruitment/selection of patients Primary insomnia outpatients. Age, gender and ethnicity Age (y) – mean (SD): Group 1: 45.9 (11.1), Group 2: 49.7 (10.6). Gender female %: Group 1: 66.7%, Group 2: 64.1%. Caucasians %: Group 1: 95.3%, Group 2: 99.3%. Extra comments Main purpose of study was to assess efficacy of self-reported sleep quality with gaboxadol which is not reported here as this intervention did not match the protocol. Indirectness of population No indirectness. Interventions & comparators Study run over 2 weeks with a run-out phase over 1 week.

(n=150) Group 1: Zolpidem 10mg (included as an active reference) Indirectness: No indirectness

Evidence tables Evidence Harms © Royal © College of Physi (n=142) Group 2: Placebo Indirectness: No indirectness

(n=145) Group 3: Gaboxadol 5 mg

Indirectness: Indirect

(n=151) Group 4 Gaboxadol 10 mg Indirectness: Indirect

cians 2019 (n=153) Group 5: Gaboxadol 15 mg Indirectness: Indirect

Funding Funded by H. Lundbeck A/S. Assistance of Doctor Christopher Lines (Merck Research) in helping draft the manuscript. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Zolpidem versus placebo

186 Protocol outcome 1: Harm – withdrawal signals - Actual outcome: proportion of patients (%) with withdrawal signals (defined as ≥3 new or worsened withdrawal symptoms according to the Benzodiazepine

Withdrawal Symptom Questionnaire), using day 15 as baseline; analysis based on group of patients that were switched from active treatment to placebo during the run-out phase: Day 1 of run-out phase: Group 1: 5.0% , Group 2: 11.4% Day 2 of run-out phase: Group 1: 1.7% , Group 2: 12.8% Day 3 of run-out phase: Group 1: 8.6% , Group 2: 14.5% Day 1-3 of run-out phase: Group 1: 16.7%, Group 2: 29.9% Note: NB – % only provided in study, actual numbers could not be calculated by NGC.

Risk of bias: All domain – Low risk of bias, Selection - Low, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement - Low, Crossover - Low; Indirectness of outcome: No indirectness ; Group 1 Number missing: 7%; Group 2 Number missing: 6%, Most frequent reason: adverse events with zolpidem (7%) and protocol violations in placebo group (6%).

Protocol outcome 2: Harm – Rebound insomnia using self-reported total sleep time (sTST) - Actual outcome: Rebound insomnia defined as the proportion of the patients (%) with a lower sTST value after the first night of the run-out phase compared to the lowest sTST value during the run-in phase; analysis based on group of patients that were switched from active treatment to placebo during the run-out phase:

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Day 1 of run-out phase: Group 1 (n=67): 19.4% , Group 2 (n=126): 6.3%

Day 2 of run-out phase: Group 1 (n=67): 7.6% , Group 2 (n=126): 5.5% Day 3 of run-out phase: Group 1 (n=67): 7.5% , Group 2 (n=126): 3.9% Note: NB – actual numbers assumed by NGC calculations, % only provided in study.

Protocol outcome 3: Harm – Rebound insomnia self-reported time to sleep onset (sSTO) - Actual outcome: Rebound insomnia defined as the proportion of the patients (%) with a lower sTSO value after the first night of the run-out phase compared to the lowest sTSO value during the run-in phase; analysis based on group of patients that were switched from active treatment to placebo during the run-out phase:

Day 1 of run-out phase: Group 1 (n=67): 17.9% , Group 2 (n=126): 6.3% Day 2 of run-out phase: Group 1 (n=67): 6.1% , Group 2 (n=126): 3.2% Day 3 of run-out phase: Group 1 (n=67): 7.5% , Group 2 (n=126): 3.9%

187 Risk of bias: All domain – Low risk of bias, Selection - Low, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement - Low, Crossover - Low;

Indirectness of outcome: No indirectness ; Group 1 Number missing: 7%; Group 2 Number missing: 6%, Most frequent reason: adverse events with zolpidem (7%) and protocol violations in placebo group (6%).

Narrative outcomes: Authors state that during the run-out phase approximately 25% of patients had adverse events. The incidence of adverse events in the run-out phase was similar in all treatment groups. Discontinuation of active treatment did not appear to induce additional adverse events.

Benzodiazepines Study Bourgeois 2014(4) Study type Observational study Number of studies (number of participants) 1 (n=38 included). Countries and setting 5 nursing homes, Belgium Duration of study 8 months (follow-up)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Inclusion criteria Cognitively competent nursing home residents with established chronic use of benzodiazepines or Z drugs for insomnia (administered daily at bedtime for at least 3 months). Willingness of the GP and resident was also part of the inclusion criteria. Exclusion criteria Residents who were critically ill or had a fatal diagnosis with short life expectancy and residents who used a

benzodiazepine during the day (for the indication anxiety) or used a sedative antidepressant (trazodone, amitriptyline, mirtazapine) or phytotherapy as co-medication. Recruitment/selection of patients All cognitively competent residents were screened for inclusion (between September 2012 – July 2013). Recruitment letters were sent to the GPs of eligible residents enquiring if they were willing to initiate a discontinuation in the particular resident. All residents whose GP agreed to initiate discontinuation were informed and asked for their willingness to discontinue their benzodiazepine / Z drug (informed consent). Age, gender and ethnicity Age, y - Mean (range): 84.3 (65-97)

Gender (Female %): 71.1 Length of stay in months - mean (range): 34.1 (3-128)

Most frequent benzodiazepine / Z-drug use: 36.8% lormetazepam, 21.1% lorazepam 37 residents used a short-acting benzodiazepine / z drug

188 Median prescribed dose: one DDD (range 0.10-2.00)

Extra comments Willingness to stop, at GP or patient level also reported: 135 residents were deemed eligible, of those GPs indicated that discontinuation was feasible in 51 residents. For 13 the GP did not respond. More than half GPs (n=71) were not willing to initiate a discontinuation due to lack of motivation of the resident themselves (56%). In 18% the failure of a previous attempt was the reason not to discontinue. Of the 51 residents, 13 refused discontinuation. The main reason (12/13) was the reluctance towards change. Indirectness of population No indirectness. Interventions & comparators There was no strict discontinuation schedule and no substitution or alternatives were used. The GP was responsible for the discontinuation, but a possible discontinuation schedule was proposed in the recruitment letter: 25% reduction per week or per 2 weeks. Funding Funding not stated RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: no comparison

Protocol outcome 1: Harm – Relapse of intervention - Actual outcome: No. participants relapsed at 2 months: 6/38 Risk of bias: All domain –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low,

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data.

- Actual outcome: No. participants relapsed at 8 months: 7/38 Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low,

Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data.

Protocol outcome 2: Harm – Withdrawal symptoms - Actual outcome: Real increase of withdrawal symptoms – Scored by the Benzodiazepine Withdrawal Symptom Questionnaire (BWSQ): 0/38 Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data.

- Actual outcome: BWSQ score, mean (SD): Baseline - 3.9 (2.8) After discontinuation – 4.1 (2.6) Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data.

189

Protocol outcome 2: Harm – Most commonly reported withdrawal symptoms - Actual outcome: Muscle pain: Baseline 49% After discontinuation 63% Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data. - Actual outcome: Muscle twitching: Baseline 41% After discontinuation 37% Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data. - Actual outcome: Loss of memory: Baseline 41% After discontinuation 32% Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data. - Actual outcome: Pins and needles: Baseline 35% After discontinuation 39% Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data.

Evidence tables Evidence Harms

Risk of bias: All domains –Very High, Confounding – High, Selection - High, Classification of intervention – High, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, No indirectness. No. missing data.

The study also reported the Pittsburgh Sleep Quality Index (PSQI), ADL score, number of chronic medications not including benzodiazepines or z-drugs and EQ5D-3L for participants that were successful in discontinuing drug and for those that relapsed separately.

Benzodiazepines and Gabapentinoids Study Kasper 2014(10)

Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=615 randomised) Countries and setting Multicentre trial conducted in 60 centres in 16 countries (Argentina, Austria, Costa Rica, Croatia, Czech Republic,

190 Finland, Greece, Guatemala, Indonesia, Lithuania, Mexico, Russian Federation, Serbia, Slovenia, Spain and Turkey).

Duration of study Study conducted from 13 May 2009 to 2 April 2012. Study duration was 2, 12 weeks treatment periods followed by 1 week taper and 1 week of follow-up. Inclusion criteria Patients were aged 18-65 years with a primary diagnosis of generalized anxiety disorder (GAD) at baseline. In addition: Hamilton Anxiety Rating Scale (HAM-A) total score ≥ 14and a Hamilton Rating Scale for Depression (HAM-D) item 1 score ≤2 at both screening and baseline visits (baseline assessment occurred ~ 4-10 days following screening). Exclusion criteria Patients with a current or past diagnosis of any other Diagnostic and Statistical manual of Mental disorder (DSM)-IV Axis I disorder besides GAD were excluded (with the exception of current or past diagnosis of depression not otherwise specified, specific phobia, somatization disorder, nicotine or caffeine abuse/dependence or past history of major depressive disorder, social phobia, panic disorder or eating disorder). Individual excluded if they reported daily use of benzodiazepines for treating GAD during the 3 months prior to screening, a history of failed treatment with any benzodiazepine or any reported prior exposure to pregabalin. Those individuals taking benzodiazepine for less than 5d/week could be included if they stopped taking the benzodiazepine 2 weeks prior to baseline. O benzodiazepine use was allowed during the study. Additional exclusion criteria were pregnancy/lactation, suicide risk, current use of psychotropic medication that could not be discontinued prior to baseline, positive urine test results at screening for potential drug abuse or illegal drugs, positive alcohol breathalyzer test at screening or any serious or unstable medical condition assessed at screening.

Ev Harms

Age, gender and ethnicity Baseline from treatment period 1: Male, N (%): Group 1: 87 (42.2), Group 2: 73 (35.4), Group 3: 81 (39.9) Age, years – Mean (SD): Group 1: 42.4 (11.5), Group 2: 40.5 (12.3), Group 3: 42.6 (11.2).

Duration of illness, years – Mean (SD): Group 1: 2.2 (4.4), Group 2: 2.1 (4.3), Group 3: 2.4 (4.3). Treatment duration, days – Median (SD): Group 1: 139.4 (55.1), Group 2: 133.2 (58.3), Group 3: 136.7 (59.4).

Authors reported that the six treatment groups in treatment period 2 did not differ significantly on available baseline characteristics or median treatment duration. Extra comments Treatment efficacy and treatment emergent adverse events were also reported in study. Indirectness of population No indirectness.

Interventions & comparators All Patients Study Schedule: Screen: 1 week Period 1: Flexible dose (week 1-6), Fixed dose (weeks 7-12), Double blind, 12 weeks. Patients who showed a clinical response with a Clinical Global Impressions-Improvement (CGI-I) score of 1 or 2 at week 6 continued treatment; those who had a CGI-I score >2 at week 6 were discontinued from the study. During the second

191 half of treatment period 1, patients were maintained on a fixed-dose treatment at the final dosage achieved during the initial 6-week flexible dosage phase. Period 2: Fixed dose, double-blind, 12 weeks. Patients who showed a clinical response (CGI-I score of 1 or 2) at week 6 of period 1 continued treatment. 25% of patients from each medication group were randomised to discontinue active medication and received placebo during treatment period 2. 75% continued on the active treatment. Taper (double blind): 1 week. Generally consistent with product labelling and was intended to minimize the risk that patients could potentially experience severe drug discontinuation symptoms. Any patients experiencing severe discontinuation symptoms during the taper periods and up to 7 days afterwards could be provided with a more gradual rescue taper extending the taper to 4 weeks while maintaining the blind. This same taper schedule and rescue taper protocol was used for all patients, regardless of the when treatment was discontinued. Follow-up: 1 week.

(n=206) Group 1: Pregabalin 450-600 mg/d Treatment initiated with a 150 mg/d starting dose. Upward dose escalation occurred during the first 3 week. Following dose escalation, patients received pregabalin 450-600 mg/d (high dose). Flexible-dose treatment within the specified ranges during the first 6 week based on tolerability and clinical improvement. Study drug administered twice a day (in equal doses) and was blinded using a double dummy method. Patients entering treatment period 2 either continued on the same fixed dose or switched to placebo, according to the randomization

(n=206) Group 2: Pregabalin 150-300 mg/d Treatment initiated with a 150 mg/d starting dose. Upward dose escalation occurred during the first 3 week.

Following dose escalation, patients received pregabalin 150-300 mg/d (low dose). Flexible-dose treatment within the specified ranges during the first 6 week based on tolerability and clinical improvement. Study drug administered twice a day (in equal doses) and was blinded using a double dummy method. Patients entering treatment period 2 either continued on the same fixed dose or switched to placebo, according to the randomization scheme.

(n=203) Group 3: Lorazepam 3-4 mg/d Starting dose was 2mg/d. Upward dose escalation occurred during the first 3 week. Following dose escalation, patients received lorazepam 3-4 mg/d. Flexible-dose treatment within the specified ranges

during the first 6 week based on tolerability and clinical improvement. Study drug administered twice a day (in equal doses) and was blinded using a double dummy method. Patients entering treatment period 2 either continued on the same fixed dose or switched to placebo, according to the randomization scheme.

192 Funding Funded by Pfizer Inc. Editorial support to prepared this manuscript for submission provided by Lorna Forse, PhD of

Engage Scientific solutions and funded by Pfizer Inc. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Pregabalin high dose versus Pregabalin low dose versus Lorazepam

Protocol outcome 1: Harm –Discontinuation symptoms measured with the Physician Withdrawal Checklist (PWC) that is a 20 item physician-rated interview (total score range 0-60) that measure the presence of anxiolytic drug discontinuation related signs and symptoms in the following areas: gastrointestinal, mood, sleep, motor, somatic, perception and cognition. - Actual outcome: Discontinuation symptoms following treatment period 1 and treatment period 2: mean change in PWC (95% CI): Following treatment period 1: Week 1 after initiating taper: Group 1 (n=58): 1.9 (-0.1, 3.8), Group 2 (n=52): 1.4 (0.2, 2.7), Group 3 (n=49): 2.3 (0.4, 4.2). Week 2 after initiating taper: Group 1 (n=54): 2.1 (0.4, 3.7), Group 2 (n=49): 2.0 (0.5, 3.6), Group 3 (n=44): 1.6 (-0.3, 3.6).

Following treatment period 2: Week 1 after initiating taper: Group 1 (n=109): 1.7 (0.7, 2.6), Group 2 (n=93): 1.1 (0.4, 1.9), Group 3 (n=99): 3.0 (1.7, 4.4). Week 2 after initiating taper: Group 1 (n=106): 2.8 (1.6, 3.9), Group 2 (n=84): 1.7 (0.7, 2.8), Group 3 (n=93): 2.2 (1.0, 3.5). Risk of bias: All domain – Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low;

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 77; Group 2 Number missing: 91, Group 3 Number missing: 81, Reason: discontinued (249), adverse events (70), lack of efficacy (50), miscellaneous (129).

Protocol outcome 2: Harm – Discontinuation emergent signs and symptoms (DESS) occurring in ≥ 5% of patients after 12 and 24 weeks of treatment. - Actual outcome: DESS adverse events are a subset to Treatment Emergent Signs and symptoms and are defined as those spontaneously reported adverse events that developed or existed prior to but worsened during the 2 week following taper initiation.

DESS during 2 weeks following taper initiation after treatment period 1: Patients with DESS, n (%): Group 21/58 (36.2), Group 2: 17/52 (32.7), Group 3: 17/52 (32.7) Anxiety, n (%): Group 1: 3/58 (5.2), Group 2: 0/52 (0), Group 3: 2/52 (3.8) Dizziness, n (%): Group 1: 3/58 (5.2), Group 2: 0/52 (0), Group 3: 0/52 (0)

Headache, n (%): Group 1: 3/58 (5.2), Group 2: 4/52 (7.7), Group 3: 1/52 (1.9) Insomnia, n (%): Group 1: 6/58 (10.3), Group 2: 4/52 (7.7), Group 3: 10/52 (19.2) Nausea, n (%): Group 1: 4/58 (6.9), Group 2: 3/52 (5.8), Group 3: 2/52 (3.8)

193 DESS during 2 weeks following taper initiation after treatment period 2:

Patients with DESS, n (%): Group 1 (active treatment in period 2): 34/109 (31.2), Group 1 (placebo treatment in period 2): 4/30 (13.3), Group 2 (active): 21/94 (22.3), Group 2 (placebo): 9/29 (31.0), Group 3 (active): 28/100 (28.0), Group 3 (placebo): 4/30 (13.3) Anxiety, n (%): Group 1 (active treatment in period 2): 7/109 (6.4), Group 1 (placebo treatment in period 2): 1/30 (3.3), Group 2 (active): 4/94 (4.3), Group 2 (placebo): 0/29 (0), Group 3 (active): 8/100 (8.0), Group 3 (placebo): 0/30 (0) Headache, n (%): Group 1 (active treatment in period 2): 5/109 (4.6), Group 1 (placebo treatment in period 2): 0/30 (0), Group 2 (active): 3/94 (3.2), Group 2 (placebo): 2/29 (6.9), Group 3 (active): 2/100 (2.0), Group 3 (placebo): 0/30 (0) Insomnia, n (%): Group 1 (active treatment in period 2): 13/109 (11.9), Group 1 (placebo treatment in period 2): 1/30 (3.3), Group 2 (active): 8/94 (8.5), Group 2 (placebo): 2/29 (6.9), Group 3 (active): 6/100 (6.0), Group 3 (placebo): 2/30 (6.7) Risk of bias: All domain – Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 77; Group 2 Number missing: 91, Group 3 Number missing: 81, Reason: discontinued (249), adverse events (70), lack of efficacy (50), miscellaneous (129).

Protocol outcome 3: Harm - Rebound anxiety - Actual outcome n/N (%) rebound anxiety defined as a HAM-A total score higher during the discontinuation assessment (Weeks 13, 14, 25 and 26) compared with baseline: After treatment period 1: Group 1: 3/58 (5.2), Group 2: 1/52 (1.9), Group 3: 2/48 (4.2)

Eviden Harms © Royal © College of Physicians 2019 After treatment period 2: Group 1: 4/109 (3.7), Group 2: 0/94 (0), Group 3: 6/100 (6.0)

Risk of bias: All domain – Very high, Selection - High, Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness ce tables

of outcome: No indirectness; At end of study - Group 1 Number missing: 77; Group 2 Number missing: 91, Group 3 Number missing: 81, Reason: discontinued (249), adverse events (70), lack of efficacy (50), miscellaneous (129).

NB For analysis in this review Group 1 and Group 2 (Pregabalin) will be pooled.

Antidepressants Study Boyer 2008(32) Study type RCT phase 3 trial (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n=565 screened; n=485 randomised) Countries and setting 44 centres in Europe (Croatia, Estonia, Finland, France, Latvia, Lithuania, Poland, Romania and Slovakia) and South Africa. Duration of study Treatment 8 weeks and tapering period of one week.

194 Conducted from March 2006 to January 2007.

Inclusion criteria Male and female outpatients aged 18 years or older with a primary diagnosis of major depressive disorder (MDD). Participants were new, referred, or existing patients at outpatient clinics in Europe and South Africa. Sites in the study included primary care clinics, where patients received treatment from GPs with training and demonstrated experience in psychiatry, and outpatient psychiatric clinics, where patients were treated by psychiatrists. At screening, a psychiatric assessment was performed by the investigator using a modified Mini-International Neuropsychiatric interview, and a diagnosis of MDD – single or recurrent episode without psychotic features – was confirmed according to the Diagnostic and Statistical Manual of mental disorders, fourth Edition, criteria. Patients were required to have depressive symptoms for at least 30 days before the screening visit. The following minimum scores were required at screening and baseline: 17-item Hamilton Rating Scale for Depression (HAM-D17) total score of 20 or more; HAM-D17 item 1 (depressed mood) score ≥ 2; and Clinical Global Impressions-Severity (CGI-S) scale score ≥4 (moderately ill). All sexually active patients in the study were required to use medically acceptable contraception. Exclusion criteria Excluded if ever received earlier treatment with desvenlafaxine; had known hypersensitivity to venlafaxine; had significant risk of suicide based on clinical judgment; were pregnant or breastfeeding; had current psychoactive substance abuse or dependence, manic episodes, post-traumatic stress disorder, obsessive compulsive disorder, or a lifetime diagnosis of bipolar or compulsive disorder, or a lifetime diagnosis of bipolar or psychotic disorder; had current generalized anxiety disorder, panic disorder, social anxiety disorder or clinically important personality disorder; had depression associated with the presence of an organic mental disorder; had a history of seizure disorder; had clinically

Evidence tables Evidence Harms © Royal © College of Physicians 2019 important medical disease; or had used prohibited treatments.

Recruitment/selection of patients Outpatients with MDD recruited from centres. In Finland and South Africa new patients were recruited by advertising. Age, gender and ethnicity Age, y, Mean (SD): Group 1: 44 (14), Group 2: 46 (13), Group 3: 46 (12) Female: Group 1: 70%, Group 2: 71%, Group 3: 68%

Race/Ethnicity: White: Group 1: 99%, Group 2: 99%, Group 3: 98% Black: Group 1: 0%, Group 2: 0%, Group 3: 1% Hispanic: Group 1: 1%, Group 2: 1%, Group 3: 1% Duration of current episode, mean (SD), months: Group 1: 7 (12), Group 2: 9 (17), Group 3: 8 (22) Baseline HAM-D17 total score, mean (SD): Group 1: 24 (2), Group 2: 24 (3), Group 3: 24 (3)

Authors reported that no significant differences among treatment groups in pretreatment demographic and clinical characteristics were present.

Extra comments Efficacy outcomes reported (including primary outcome HAM-D17). Quality of life and clinical signs reported. Indirectness of population No indirectness.

195

Interventions & comparators Screening period: 6-14 days Treatment period: 8 weeks Discontinuation period: 1 week tapering study medication (7-day taper recommended but could be omitted, shortened, or lengthened at the discretion of the investigator). During the taper period Group 1 received placebo and group 2 received 50 mg/day. Patients in the group 3 received placebo during the taper period. Follow-up visit: 7 days after last dose of medication

(n=166) Group 1: Desvenlafaxine succinate 50mg/day; received assigned maintenance dose on day 1 until day 56 or early withdrawal). (n=158) Group 2: Desvenlafaxine succinate 100mg/day; doses titrated to their maintenance dose (days 1-7 received 50mg/day; on day 8 received assigned dose of 100mg/d and continued on this regimen until day 56 or early withdrawal) (n=161) Group 3: Placebo for 8 weeks

Funding Sponsored by Wyeth Research, Collegeville, Pennsylvania, USA. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Desvenlafaxine 50 mg/d vs desvenlafaxine 100mg/d vs placebo

Protocol outcome 1: Harm – Discontinuation-Emergent Signs and Symptoms (DESS) checklist (summation of symptoms that first appeared (new symptoms) or were

Evidence tables Evidence Harms © Royal © College of Physicians 2019 exacerbated (old symptoms but worse) during taper period (7-day period after the end of the treatment period) – checklist completed by 420/423 patients who completed least 53 days of on-therapy study drug

Actual outcome: Mean (SD) DESS checklist – 1 week after treatment period Group 1 (n=146): 2.14 (3.64), Group 2 (n=128): NR, Group 3 (n=146): 0.86 (2.01)

Actual outcome: Mean (SD) DESS checklist – 2 week after treatment period Group 1 (n=146): NR, Group 2 (n=128): 1.36 (2.72), Group 3 (n=146): 0.62 (1.41)

Note: Authors only reported figures for groups with significant differences. No figures were provided for outcomes at 3 weeks after treatment period. This has been added to narrative section and no analysis has been conducted due to incomplete data. Narrative: Desvenlafaxine 50mg/day showed a significantly higher mean DESS score compared with the placebo group during the 7 day taper period.

DESS scores were not significantly different between desvenlafaxine 50 mg/day and placebo groups at weeks 2 and 3 after treatment. Desvenlafaxine 100 mg/day group showed a significantly higher mean DESS score compared with the placebo group only during week 2 after treatment (at which point patients went from receiving 50mg/day to no active treatment); differences from placebo were not statistically significant at week 1 or 3 after treatment.

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - high, Outcome reporting - high, Measurement – Low, Crossover - Low;

196 Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 20; Group 2 Number missing: 30, Group 3 Number missing; 15. Reason: Reported

420/423 who completed at least 53 days; Reason*: Adverse event (24), failed to return (2), investigator request (2), unsatisfactory response (8), protocol violation (5), other event (1), patient request not study related (8), additional 12 must have been included in those that did not complete trial (53 days of exposure to study drug), questionnaire not given to (3).

Protocol outcome 2: Harm – Taper / post study-emergent adverse events Actual outcome: Total number (%) taper / post study-emergent adverse events Group 1: 74/166 (45%), Group 2: 64/158 (41%), Group 3: 33/161 (21%)

Actual outcome: Taper / post study-emergent adverse events reported by at least 5% of patients in either active treatment group during the taper/post study period Dizziness: Group 1: 22/166 (13%), Group 2: 15/158 (10%), Group 3: 4/161 (3%) Nausea: Group 1: 19/166 (11%), Group 2: 15/158 (10%), Group 3: 4/161 (3%) Headache: Group 1: 11/166 (7%), Group 2: 11/158 (7%), Group 3: 10/161 (6%) Insomnia: Group 1: 9/166 (5%), Group 2: 6/158 (4%), Group 3: 3/161 (2%) Depression: Group 1: 4/166 (2%), Group 2: 8/158 (5%), Group 3: 1/161 (1%) Vertigo: Group 1: 4/166 (2%), Group 2: 8/158 (5%), Group 3: 0/161 (0%)

Evidence tables Evidence Harms © Royal © Coll Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 18; Group 2 Number missing: 30, Group 3 Number missing; 14. Reason*: Adverse event (24), failed to return (2), investigator request (2), unsatisfactory response (8), protocol violation (5), other event (1), patient request not study related (8), additional 12 must have been included in those that did not complete trial (53 days of exposure to study drug).

ege of Physicians 2019

NB For analysis in this review Group 1 and Group 2 (Desvenlafaxine) will be pooled.

Study Cutler 2009(21) Study type RCT phase 3 (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=912 screened; n=612 randomised of which 610 received treatment and were included in the safety analysis set)

Countries and setting Multicentre: 38 centres in the United States Duration of study 6 weeks treatment and two weeks taper. Conducted between April 2006 and May 2007. Inclusion criteria Male or female outpatients, aged 18 to 65 years inclusive, with a Diagnostic and Statistical Manual of Mental Disorders,

197 fourth edition (DSM-IV) diagnosis of major depressive disorder (single episode or recurrent) were eligible for inclusion in the study. The diagnosis was confirmed by the Mini-International Neuropsychiatric interview. Patients were required to have a 17-item Hamilton Rating Scale for Depression (HAM-D) total score ≥ 22 and HAM-D item 1 (depressed mood) score of ≥ 2 at enrollment and randomization. Exclusion criteria Excluded if they were diagnosed with a DSM-IV Axis I disorder other than major depressive disorder within 6 months prior to enrollment, if they had any DSM-IV Axis II disorder that would significantly impact on the patients current psychiatric status, if the duration of their current major depressive disorder episode exceeded 12 months or was less than 4 weeks, or if they had had an inadequate response to at least 6 weeks treatment with 2 or more classes of antidepressants during the current episode. Psychotherapy was allowed only if it had been ongoing for at least 3 months prior to randomization. Other exclusion criteria included a clinically significant medical illness or any clinically significant findings on physical examination, laboratory tests, or electrocardiogram. Patients who posed a current serious suicidal or homicidal risk were also excluded. Patients were not permitted to take antipsychotic, mood stabilizer or antidepressant drugs within 7 days before randomization. Fluoxetine was prohibited within 28 days before randomization and use of MAOIs, anxiolytics, and hypnotics was prohibited within 14 days before randomization. Recruitment/selection of patients Patients with MDD. Age, gender and ethnicity Gender, Female: Group 1: 62.4%, Group 2: 64.5% Mean (SD) age, years: Group 1: 40.2 (12.5) Group 2: 42.3 (11.5)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Ethnicity White: Group 1: 75.9%, Group 2: 69.1% Black: Group 1: 17.7%, Group 2: 25.7%

Asian: Group 1: 0.7%, Group 2: 1.3% Other: Group 1: 5.7%, Group 2: 3.9% Extra comments Purpose of study was to determine the effectiveness of quetiapine fumarate compared to placebo and duloxetine (active placebo). Quetiapine outcomes were not reported as they are not one of the drugs of interest in this review. Effectiveness outcomes were the primary outcomes reported. Clinical outcomes were also reported. Indirectness of population No indirectness of population. Interventions & comparators Enrollment period: Enrollment and wash out of any prior psychotropic medications; 7-28 days

Treatment phase: 6 weeks active treatment Post treatment drug-discontinuation/drug tapering phase: 2 weeks. Duloxetine and quetiapine 300mg/d groups had their doses halved during the first week of the tapering period. Follow up period: 2 weeks

198

(n=151) Group 1: duloxetine 60 mg/day orally in a single dose, once daily in the evening. (n= 157) Group 2: Placebo Group 3: Quetiapine fumarate150mg/d (not extracted as not on our drug list) Group 4: Quetiapine fumarate 300mg/d (not extracted as not on our drug list) Funding Sponsored by AstraZeneca. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Duloxetine vs Placebo

Protocol outcome 1: Harm – Adverse events Actual outcome: Range of total mean scores treatment discontinuation signs and symptoms (TDSS) assessment – 18 item scale developed by AstraZeneca as a hybrid of two other scores (unclear if higher or lower better and maximum scores possible). Group 1: 2.4 to 4.3, Group 2: 1.6 to 2.9 Risk of bias: All domain –Very high, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - High, Measurement – High, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 78; Group 2 Number missing: 57, Reason*: Adverse event (27), discontinued study (79), condition under investigation worsened (5), development of study-specific discontinuation criteria (2), other (3), severe non adherence to protocol (3), lost to follow up (16), not willing to continue (16). Note: Outcomes calculated using safety analysis population including all patients randomised with the exception of two patients in group 1 that were reported under the discontinued from study treatment.

Actual outcome: Number (%) Adverse events (>3%) in any treatment group during the 2 week discontinuation/tapering phase (safety population) Nausea: Group 1: 2/149 (1.3), Group 2: 1/157 (0.6)

Headache: Group 1: 9/149 (6.0), Group 2: 6/157 (3.8) Insomnia: Group 1: 2/149 (1.3), Group 2: 3/157 (1.9) Diarrhea: Group 1: 4/149 (2.7), Group 2: 4/157 (2.5) Dizziness: Group 1: 8/149 (5.4), Group 2: 1/157 (0.6) Vomiting: Group 1: 1/149 (0.7), Group 2: 3/157 (1.9)

(79), condition under investigation worsened (5), development of study-specific discontinuation criteria (2), other (3), severe non adherence to protocol (3), lost to follow up (16), not willing to continue (16). Note: Outcomes calculated using safety analysis population including all patients randomised with the exception of two patients in group 1 that were reported under the discontinued from study treatment.

199 Study Krystal 2011(12) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=229 randomised; 8 patients discontinued after randomization but before receiving a dose of double-blind study drug and these patients were not included in ITT or safety analyses data sets). Countries and setting 22 sleep centres in the US. Duration of study 5 week sleep study. Inclusion criteria Men and women between the ages of 18 and 64 years with a DSM-IV-TR diagnosis of primary insomnia who reported sleep maintenance difficulty were eligible. Patient screening for general eligibility and sleep history was conducted during an initial clinical visit and involved a medical, sleep and psychiatric history; physical examination; vital sign measurements; clinical laboratory tests; and an electrocardiogram. Patients meeting genital screening criteria were asked to record their sleep patterns onto a daily sleep diary prior to PSG screening (≥7 days of assessment). The initial screening results and sleep diary data were used to verify a DSM-IV-TR diagnosis of insomnia for at least the last 3 months. During the first 2 nights of placebo the patients spent it in the sleep laboratory to determine whether they met the PSG screening criteria. Patients were required to meet the following polysomnographic entry criteria in order to be eligible for randomization: latency to persistent sleep (LPS) > 10 minutes on both PSG screening nights; mean wake time

Evidence tables Evidence Harms © Royal © College of Physicians 2019 during sleep (WTDS) ≥ 60 min on both PSG screening nights, with no night < 45 minutes; and TST > 240 and ≤ 400 minutes on both screening nights. Patients were excluded from the study if the y had 10 or more apnea/hypopnea events or periodic leg movements with arousals/hour of sleep/ Exclusion criteria Excessive use of alcohol, nicotine, or caffeinated beverages, intentional napping more than twice per week; having a

variation in bedtime > 2 hours on 5 of 7 nights; or use of a hypnotic or any other medication known to affect sleep. Recruitment/selection of patients Primary insomnia patients from sleep centers. Age, gender and ethnicity Age, y, Mean (SD): Group 1: 45.5 (10.6), Group 2: 44.2 (11.1), Group 3: 43.6 (12.3) Female: Group 1: 77%, Group 2: 71%, Group 3: 70%

Race/Ethnicity: Caucasian: Group 1: 44%, Group 2: 53%, Group 3: 48%

African American: Group 1: 35%, Group 2: 29%, Group 3: 34% Hispanic: Group 1: 20%, Group 2: 14%, Group 3: 15% Other: Group 1: 1%, Group 2: 4%, Group 3: 2%

Authors reported that baseline characteristics were comparable across groups.

200 Extra comments Sleep maintenance and duration reported.

Indirectness of population No indirectness. Interventions & comparators Phase A: 2 weeks of single blind placebo dosing. Patients spent first 2 nights in a sleep laboratory to determine whether they met PSG screening criteria and then took the placebo for 5 nights at home. After completing this, patients participated in 2 consecutive nights of 8-h continuous PSG recordings in a sleep laboratory.

(n=75) Group 1: Doxepin 3mg, 35 days of nightly treatment and two days of placebo to evaluate discontinuation effects. (n=73) Group 2: Doxepin 6mg, 35 days of nightly treatment and two days of placebo to evaluate discontinuation effects. (n=73) Group 3: Placebo for 35 days and an additional two days of placebo to evaluate discontinuation effects.

Funding Provided by Somaxon Pharmaceuticals, Inc. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Doxepin 3mg vs doxepin 6mg vs placebo

Protocol outcome 1: Harm – Rebound insomnia Actual outcome: Rebound insomnia based on wake time after sleep onset (WASO) criteria experienced over the 2 nights after discontinuation:

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Group 1: 1/75 (1%), Group 2: 3/73 (4%), and Group 3: 1 /73 (1%)

Note: actual numbers assumed by NGC calculations, % only provided in study.

Risk of bias: All domain –High, Selection - High, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low;

Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 9; Group 2 Number missing: 8, Group 3 Number missing; 9. Reason*: Adverse event (7%), consent withdrawn (7%), protocol violation (1%), noncompliance (9%), other (11%).

Protocol outcome 1: Harm – Withdrawal signs Actual outcome: Benzodiazepine Withdrawal Symptom Questionnaire (BWSQ) criteria of 3 more new symptoms in the BWSQ, the predetermined withdrawal criteria. Group 1 (n=75): 1 , Group 2 (n=73): 0, Group 3 (n=73): 1

Mean BWSQ scores from day 38: Group 1 (n=75): 0.8 , Group 2 (n=73): 0.4, Group 3 (n=73): 0.6

Risk of bias: All domain –High, Selection - High, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low; 201 Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 9; Group 2 Number missing: 8, Group 3 Number missing; 9. Reason*: Adverse event

(7%), consent withdrawn (7%), protocol violation (1%), noncompliance (9%), other (11%).

Narrative information: Overall there was a low incidence of AEs reported during the discontinuation period. Approximately 8% in each of the 3 groups experienced an AE during the discontinuation period. A review of these adverse events revealed no evidence of physical dependence, withdrawal syndrome, or worsening insomnia. Additionally, BWSQ data indicated no evidence of withdrawal syndrome.

Study Liebowitz 2008(22) Study type RCT phase 3 (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=703 screened; n=474 randomised) Countries and setting 25 centres in the United States. Duration of study 8 weeks treatment followed by 1 week taper. Conducted from February 2006 to January 2007. Inclusion criteria Male and female outpatients 18 years of age or older who met the Diagnostic and Statistical Manual of Mental Disorders, Fourth edition (DSM-IV) criteria for a single or recurrent episode of major depressive disorder (MDD), without

Evidence tables Evidence Harms © Royal © College of Physicians 2019 psychotic features, were enrolled. MDD had to have been the predominant psychiatric disorder and depressive symptoms had to have been present for at least 30 days before the screening visit. The following scores were required

at screening and baseline (study day -1): 17-item Hamilton Rating Scale for Depression (HAM-D17) total score ≥ 20; HAM- D17 item 1 (depressed mood) score ≥ 2; and clinical Global Impressions-severity (CGI-S) scale score ≥ 4 (moderately ill).

All sexually active patients agreed to use a medically acceptable form of contraception for the duration of the study and at least 15 days after the last dose of study medication. Exclusion criteria Excluded if received treatment with desvenlafaxine in the past; known hypersensitivity to venlafaxine; significant risk of suicide based on clinical judgment; was pregnant or breastfeeding; current psychoactive substance abuse or dependence; manic episode, post-traumatic stress disorder, obsessive-compulsive disorder, or a lifetime diagnosis of bipolar or psychotic disorder; current primary generalized anxiety disorder, panic disorder, social anxiety disorder, or clinically important personality disorder; depression associated with the presence of an organic mental disorder; used prohibited treatments; history of seizure disorder; or clinically important medical disease, clinical abnormalities on

screening physical examination, electrocardiogram (ECG), laboratory tests, or urine drug screen. Nonpsychopharmacologic drugs with psychotropic effects, at a stable dose for at least 90 days before study day 1 and without dose changes throughout the study, were permitted. Washout periods for prohibited therapies were: venlafaxine, 90 days; fluoxetine and antipsychotics, 30 days; sumatriptan, naratriptan, and zolmitriptan, anxiolytics, MAOIs and other antidepressants, 14 days; sedative hypnotics other than zaleplon or zolpidem, herbal products 202 intended to treat anxiety, insomnia, and depression and other psychotropic drugs and substances, 7 days.

Investigational drugs and procedures were prohibited within 30 days before study day 1. Recruitment/selection of patients Outpatients with major depressive disorder patients. Age, gender and ethnicity Authors reported a difference in baseline characteristic weight between desvenlafaxine 50 mg dose and the placebo group but this was not expected to have an impact on the efficacy results.

Age, mean (SD) years: Group 1: 43 (15), Group 2: 43 (13), Group 3: 42 (14) Female %: Group 1: 62, Group 2: 53%, Group 3: 63

Ethnic origin, n (%): White: Group 1: 110 (73), Group 2: 31 (21), Group 3: 29 (19) Black: Group 1: 26 (17), Group 2: 102 (69), Group 3: 106 (71) Hispanic: Group 1: 7 (5), Group 2: 8 (5), Group 3: 9 (6) Other: Group 1: 7 (5), Group 2: 6 (4), Group 3: 6 (4) Extra comments Primary outcomes reported were efficacy of treatment. Clinical outcomes reported. Indirectness of population No indirectness of population.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Interventions & comparators Treatment: 8 weeks Tapering: 1 week recommended for all patients after the end of the double blind treatment but may have been omitted, shortened or lengthened at the discretion of the investigator. Taper visit day 63 and a post study follow up visit on day 70, a week after treatment. Group 2 tapered from 100 mg/d to 50 mg/d; whereas the other two groups went or

remained on placebo)

(n=158) Group 1: Desvenlafaxine 50 mg/d (n=157) Group 2: Desvenlafaxine 100 mg/d; began on 50mg/d and titrated to their assigned dose on day 8. (n=159) Group 3: Placebo Funding Wyeth Research. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Desvenlafaxine 50 mg/d vs desvenlafaxine 100 mg/d vs placebo

Protocol outcome 1: Harm – Adverse events

Actual outcome: Taper emergent adverse events reported Group 1: 50% (75 /151), Group 2: 49% (72 /148), Group 3: 36% (54 /152)

203 Note: NB – actual numbers assumed by NGC calculations, % only provided in study. However, the number could have been two different figures due to rounding of the

overall percentages so we have used the lower number for each group to give a conservative estimate.

Actual outcome: Depression with suicide ideation during post study period Group 1: 1/151, Group 2: 0/148, Group 3: 0/152 Note: Serious adverse events in treatment phase also reported including intentional overdoses and suicide attempt.

Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 36; Group 2 Number missing: 43, Group 3 Number missing; 42. Reason*: Adverse event (20), failed to return (32) investigator request (1), unsatisfactory response (6), protocol violation (3), other (8), patient request not study related (20).Note: figures given in study for drop outs does not agree with figures given for completers.

Actual outcome: DESS scores reported by completers (≥ 53days of on therapy study drug) at 1 week after taper – day 63 (n=316) and at follow up visit day 70 (n=300) Narrative: There was a significant increase (+2.1, p=0.001) in mean DESS scores for the desvenlafaxine 50-mg dose group compared with the placebo group (+0.2) at the end of week 1 of the taper period after patients had received placebo for one week. There was no significant differences in DESS scores between the desvenlafaxine 100 mg/day dose group and the placebo group after 1 week at the 50-mg/day dose (+0.6) or 1 week after discontinuation from all active treatment (+1.6). Risk of bias: All domain – Very high, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - High, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 36; Group 2 Number missing: 43, Group 3 Number missing; 42. Reason*: Adverse

Evidence tables Evidence Harms © Royal © College of Physicians 2019 event (20), failed to return (32) investigator request (1), unsatisfactory response (6), protocol violation (3), other (8), patient request not study related (20).Note: figures given in study for drop outs does not agree with figures given for completers.

Study Liebowitz 2009(28) Study type RCT phase 3 (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=666 screened; n=343 randomised) Countries and setting Multicentre: 56 sites (7 in Canada and 49 in the United States). Duration of study 10 week treatment and 2 week taper.

Conducted April 2001 to December 2002. Inclusion criteria Healthy adult outpatients aged 18 years and older who met criteria who met DSM-IV criteria for panic disorder (with or without agoraphobia) for at least 3 months before study day 1 and had sufficient symptoms to require anxiolytic drug therapy. Patients were required to have a score ≥ 4 on the Clinical Global Impressions-Severity of Illness (CGI-S) scale, at least 8 full-symptom panic attacks within 4 weeks of screening, a minimum o f4 full symptom panic attacks during the 14

204 (±3)-day placebo lead-in period between screening and baseline, and a Covi Anxiety Scale total score greater than the Raskin Depression Scale total score. Exclusion criteria Excluded from the trial if they had DSM-IV-diagnosed GAD or MDD considered by the investigator as primary (i.e., causing a higher degree of distress or impairment than panic disorder); patients with a secondary major depression or GAD were eligible provided that other exclusionary requirements were not met. Patients were excluded if they had any other clinically significant Axis I or Axis II disorders, other than panic disorder (with or without agoraphobia), current or predominant within 6 months of baseline; had a history or presence of bipolar affective disorder, organic brain disorder, seizure disorder, or any psychotic illness; or were acutely suicidal or had a history of drug or alcohol dependence or abuse within 1 year of baseline. Also excluded were patients with a screening or baseline 17-item 1 (depressed mood) score > 2, a screening or baseline Raskin Depression Scale single item score > 3, or a Raskin total score > 9. In addition, patients were excluded if they had received prior treatment with venlafaxine or venlafaxine ER within 6 months of baseline or had a known hypersensitivity to venlafaxine; had taken investigational drugs, antipsychotics, fluoxetine, sumatriptan, naratriptan, or zolmitriptan within 30 days of baseline; had regularly used benzodiazepines within 14 days of screening; had taken herbal products, antidepressants, monoamine oxidase inhibitors, nonbenzodiazepine anxiolytics or psychopharmacological drugs within 14 days of baseline; had taken nonpsychopharmacologic drugs with psychotropic effects within 7 days of baseline, unless taken at a stable dose for at least 3 months before baseline; had received electroconvulsive therapy within 6 months of baseline ; had had cognitive behavioral therapy within 30 days of baseline or had initiated or changed the intensity of formal psychotherapy within

Evidence tables Evidence H

© Royal © College of Physicians 2019 60 days of baseline. Presence of clinically significant abnormal findings o lab test, physical examination, arms electrocardiogram, or vital signs or a history or presence of clinically important medical conditions. Women who were pregnant or lactation were excluded as well as women of childbearing age that were not on medically acceptable contraception. .

Recruitment/selection of patients Adult outpatients panic disorder patients. Age, gender and ethnicity Reported for intention to treat population: Age, mean (SD), y: Group 1: 36 (12.4) , Group 2: 36.7 (12.0) Male, %: Group 1: 28, Group 2: 41 Women, %: Group 1: 72, Group 2: 59 Only baseline significant difference reported was the distribution of men and women. Extra comments Efficacy, quality of life, clinical outcomes and adverse events reported for treatment phase.

Indirectness of population No indirectness of population. Interventions & comparators Lead-in period: 14 day placebo lead in period Treatment: 10 weeks. Taper period: Up to 2 weeks, taper period could have been omitted or prolonged if medically indicated.

205 Evaluations: 4 to 10 days study medication had discontinued or after the completion for the taper period for all patients who received study treatment.

(n=175) Venlafaxine extended release (ER) – flexible dose 75-225 mg/d On study days 1-4 received 37.5mg/d (1 capsule) and the dose was increased to 75 mg/d (1 capsule) on days 5 to 14. If clinically indicated to improve response, the daily dose could be increased to 150 mg/d (3 capsules) on day 22. Dosage could have been reduced at any time during the study to improve tolerance, but after study day 7, the minimum daily dose allowed was 75 mg/d (1 capsule). Continued active treatment until day 70 or early withdrawal. During taper period doses were reduced at weekly intervals. Patients who had received 2 capsules or 3 capsules (150 or 225 mg/d) during the on therapy period tapered their dose by taking 1 less capsule each week; patients who had received 75 mg/d during the treatment period did not need to have their dose tapered.

(n=168) Placebo for 70 days or early withdrawal. During taper period continued to receive placebo for 14 days. Funding Sponsored by Wyeth Research. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Venlafaxine vs Placebo

Protocol outcome 1: Harm – total % of adverse events Actual outcome: % of total taper/post study-emergent adverse events reported 4-10 days after last dose of double blind or taper study medication

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 43; Group 2 Number missing: 55. Reason*: Adverse event (16), failed to return (27),

unsatisfactory response (17), protocol violation (19), patient request not study related (13), other event (6).Note: total reason missing correct but study flow diagram figures report an extra discontinuation reason for venlafaxine and one less in placebo group.

Protocol outcome 2: Harm – specific adverse events Narrative: Events reported by 5% or more of venlafaxine patients were headache, dizziness and nausea. Headache was reported by 5% or more of placebo patients during the same period. Note: Study only reported results for adverse events with significant outcomes (below) and not for headache which was reported by ≥ 5% in both groups.

Actual outcome: Nausea – taper/post emergent adverse events Group 1: 11/64 (7%), Group 2: 1/159 (1%) NB – actual numbers assumed by NGC calculations, % only provided in study. However, the number could have been two different figures due to rounding of the overall percentages so we have used the lower number for each group to give a conservative estimate.

206

Actual outcome: Dizziness – taper/post emergent adverse events Group 1: 24/164 (15%), Group 2: 4/159 (3%) NB – actual numbers assumed by NGC calculations, % only provided in study. However, the number could have been two different figures due to rounding of the overall percentages so we have used the lower number for each group to give a conservative estimate.

Note: Study reported ‘of the 13 patients (6 placebo, 7 venlafaxine) considered to have serious AEs, 6 discontinued treatment: 3 placebo patients (1 with depression, 2 with unintended pregnancies) and 3 venlafaxine ER patients (1 each suicide attempt, depression and convulsion). The remaining 7 patients did not discontinue treatment: 3 placebo patients (1 each with chest pain, basal cell carcinoma, and false-positive urine pregnancy test) and four venlafaxine patients (3 with unintended pregnancies, 1 with gastroenteritis)’. NB – NGC assumed serious adverse events during treatment not taper period.

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - High, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 43; Group 2 Number missing: 55. Reason*: Adverse event (16), failed to return (27), unsatisfactory response (17), protocol violation (19), patient request not study related (13), other event (6).Note: total reason missing correct but study flow diagram figures report an extra discontinuation reason for venlafaxine and one less in placebo group.

Study Mahableshwarkar 2015(13) Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n= 614 randomised). Countries and setting 58 sites in the USA. Duration of study 8 week treatment and 2 week taper period. Enrolment from June 2010 and study completed in February 2012. Inclusion criteria Adult men and women (aged 18-75years inclusive) were included if they had a primary diagnosis of recurrent major depressive disorder as defined by the Diagnostic and Statistical Manual of Mental Disorders with a reported duration of current major depressive episode ≥ 3 months. Patients were required to have a Montgomery-Asberg Depression Rating

Scale (MADRS) total score ≥26 at screen and baseline and a clinical global Impression severity total score ≥4. Exclusion criteria Treatment with any investigational compound more than 30 days before screening or five half-lives prior to screening; treatment with vortioxetine in a previous clinical study; a lack of response to previous adequate treatment with duloxetine for any major depressive episode (MDE); symptoms considered resistant to two or more antidepressant rails;

207 any concurrent psychiatric disorder than major depressive disorder or prior history of psychiatric disorders such as manic or hypomanic episode, schizophrenia, or substance abuse disorder; significant risk of suicide in the opinion of the investigator or a score of ≥5 on item 10 of MADRS; or a history of neurological disorders or medically unstable conditions. Patients were prohibited from receiving formal cognitive or behavioral therapy, systematic psychotherapy, or taking any medication deemed to potentially affect the outcomes of the study. All subjects required to have a 2 week washout period for nay psychoactive medications prior to screening. Recruitment/selection of patients Multicenter recruitment of people with a primary diagnosis of recurrent major depressive disorder. Age, gender and ethnicity Age years, mean (SD): Group 1: 43.1 (12.28), Group 2: 42.8 (12.40) , Group 3: 43.4 (12.24), Group 4: 42.4 (12.55) Female: Group 1: 70.7%, Group 2: 74.0% , Group 3:78.3%, Group 4: 72.0%

Race: White: Group 1: 77.6%, Group 2: 74.7%, Group 3: 78.3%, Group 4: 75.8% Black: Group 1: 21.1%, Group 2: 23.4%, Group 3: 21.1%, Group 4: 23.0% Asian: Group 1: 1.4%, Group 2: 1.9%, Group 3: 0.7%, Group 4: 0.6% Native American/Alaskan native: Group 1: 0%, Group 2: 0%, Group 3: 0, Group 4: 0.6%

Authors concluded that demographics and baseline clinical characteristics were balanced across treatment groups. Extra comments Efficacy, clinical outcomes and safety data was reported.

Interventions & comparators Treatment phase (8 weeks) Discontinuation phase (2 week period) following an abrupt discontinuation of vortioxetine treatment. Patients in the duloxetine group had their dose tapered to 30 mg for the first week.

(n=147) Group 1: Vortioxetine 15 mg once daily; doses up-titrated after the first week; received 10 mg dose for the first week for the 8-week study

sicians 2019 (n=154) Group 2: Vortioxetine 20 mg once daily; doses up-titrated after the first week; received 10 mg dose for the first week for the 8-week study

(n=152) Group 3: Duloxetine 60 mg once daily (active comparator); doses up-titrated after the first week; received 30

mg dose for the first week for the 8-week study

(n=161) Group 4: Placebo once daily Funding Supported by Takeda Pharmaceutical company, Ltd and H. Lundbeck A/S. Takeda Pharmaceutical company was involved

208 in the design, investigator selection, conduct of the trial, collection of data, analysis and interpretation and wiring of the final study report.

RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Vortioxetine 15 mg vs vortioxetine 20mg vs duloxetine vs placebo

Protocol outcome 1: Harm – Discontinuation symptoms - Actual outcome: Discontinuation –Emergent Signs and Symptoms (DESS) scale during the 2-week discontinuation period: Week 9: Group 1: 1.8, Group 2: 1.8, Group 3: 1.5, Group 4: 1.1 Week 10: Group 1: 2.5, Group 2: 1.6, Group 3: 2.7, Group 4: 1.7 Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 34; Group 2 Number missing: 41, Group 3: Number missing: 37, Group 4 Number missing: 32, Reason*: Adverse event (42), lack of efficacy (12), noncompliant (9), protocol deviation (12), withdrawal of consent (25), lost to follow-up (42), other (7).

Protocol outcome 2: Harm – Discontinuation symptoms - Actual outcome: Patients with positive Columbia-Suicide Severity Rating Scale (C-SSRS) reports: Baseline: Group 1: 17%, Group 2: 27.3%, Group 3: 21.3%, Group 4: 23.3% Active suicidal ideation during study: Group 1: 1, Group 2: 1, Group 3: 2, Group 4: 0 No suicidal behavioral-related events were reported during the study.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 34; Group 2 Number missing: 41, Group 3: Number missing: 37, Group 4 Number missing: 32, Reason*: Adverse event (42), lack of efficacy (12), noncompliant (9), protocol deviation (12), withdrawal of consent (25), lost to follow-up (42), other (7).

Study Nishimura 2018(14) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=600 randomised; n=received study drug). Countries and setting 90 sites from 14 countries comprising 44 in Europe, 31 in Japan, and an additional 15 sites in Asia/Oceania.

Duration of study 8 weeks pf treatment and 2 weeks discontinuation. Inclusion criteria Patients had a primary diagnosis of major depressive disorder (MDD) according to the DSM-IV-TR criteria, an MADRS total score ≥ 26, a clinical Global Impression Scale – Severity (CGI-S) score ≥ 4, and had the current major depressive episode for ≥ 3 months at baseline. Additional eligibility criteria included age ≥ 20 and ≤ 64 year; capacity to understand and comply with protocol requirements; signed informed consent form; and agreement from female patients to

209 routinely use adequate contraception throughout the duration of the study.

Exclusion criteria Patients ineligible for enrollment if they had any current psychiatric or neurological disorder other than MDD as defined in the DSM-IV-TR; had a significant health related issue, or an abnormal test result, or taken any disallowed medication; had current depressives symptoms resistant to two adequate antidepressant treatments of at least 6 weeks; duration each; had significant risk of suicide, a score of 5 or more on Item 10 (suicidal thought) of the MADRS, or had attempted suicide within 6 months, and had received vortioxetine or a disallowed treatment. Recruitment/selection of patients Primary diagnosis of major depressive disorder. Age, gender and ethnicity Mean age, y (SD): Group 1: 44.2 (11.89), Group 2: 45.7 (10.9), Group 3: 44 (11.79), Group 4: 43.6 (11.57) Male %: Group 1: 31.9%, Group 2: 38%, Group 3: 39.6%, Group 4: 40.1% Race Caucasian %: Group 1: 70.1%, Group 2: 69.3%, Group 3: 68.2%, Group 4: 68.4 Asian %: Group 1: 29.9%, Group 2: 30.7%, Group 3: 31.8%, Group 4: 31.6%

Authors report no meaningful differences at baseline between the treatment groups in patient demographics or disease characteristics, although the proportion of male patients was smaller in the vortioxetine 5 mg group. Extra comments Montgomery-Asberg Depression Rating scale (MADRS) and treatment emergent adverse events reported.

Interventions & comparators 1 week screening. 8 week treatment period. 2 week discontinuation period.

(n=144) Group 1: Vortioxetine 5 mg q.d. (n=150) Group 2: Vortioxetine 10 mg q.d. (n=154)Group 3: Vortioxetine 20 mg q.d. –received 10 mg q.d. for the first week and was then titrated up to 20 mg q.d. for the remaining 7 weeks. (n=152) Group 4: Placebo Funding Funded by Takeda Pharmaceutical Company, Ltd, and H Lundbeck A/S. Authors AN, YA, KS and TK are employees of Takeda Development Center Japan. AM was employed of Takeda Development Center Americas, Inc at time of study. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Vortioxetine 5mg vs 10mg vs 20 mg vs placebo

Protocol outcome 1: Harm – Serious adverse events

210

- Actual outcome: Suicide attempts (unclear at what time point) Group 1: n=0, Group 2: n=0, Group 3: n=1, Group 4: n=1

- Actual outcome: Depression (unclear at what time point) Group 1: n=0, Group 2: n=1, Group 3: n=0, Group 4: n=0

- Actual outcome: suicidal ideation (unclear at what time point) Group 1: n=0, Group 2: n=0, Group 3: n=1, Group 4: n=0

Risk of bias: All domain – High, Selection - High, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 17; Group 2 Number missing: 18, Group 3 Number missing: 22, Group 4 Number missing: 16, Reason*: Pre-treatment event or adverse event (26), noncompliant (4), lack of efficacy (8), lost to follow-up (9), withdrawal of consent (19), major protocol deviation (6) and pregnancy (1).

Protocol outcome 2: Harm – Discontinuation symptoms

Eviden Harms © Royal © College of Physicians 2019 - Actual outcome: Mean (SD) scores of Discontinuation Emergent signs and symptoms (DESS)

ce tables

Week 9: Group 1 (n=122): 1.2 (2.73), Group 2 (n=127): 1.0 (2.08), Group 3 (n=125): 1.4 (2.42), Group 4 (n=128): 1.0 (1.93) Week 10: Group 1 (n=122): 0.8 (2.21), Group 2 (n=128): 1.1 (2.49), Group 3 (n=126): 0.7 (1.71), Group 4 (n=125): n=0.5 (1.38)

Study Perahia 2009

Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=579 screened; n=288 randomised) Countries and setting 43 study centers in 5 European countries (France, Germany, Italy, Russia and Sweden) and the United States.

211 Duration of study 10 weeks open label acute treatment phase followed by 24 week open label continuation treatment phase for patients given active treatment and only responders entered the next phase. Treatment period of 52 weeks with an optional follow-up phase. Conducted from March 2005 to January 2008. Inclusion criteria Patient were male and female outpatients of at least 18 years of age who met criteria for recurrent major depressive disorder as defined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) and confirmed via the MINI-International Neuropsychiatric Interview. Patients had to have a HAM-D-17 score ≥ 18 and a CGI-S score ≥ 4 at the screening visit and the beginning of the acute phase and must have had at least 3 episodes of depression (including the presenting episode) within the past 5 years. Patients also had to have been in remission between these 3 episodes of depression and had to have been stable and off antidepressant medication for at least 2 months prior to the onset of the presenting episode. Exclusion criteria If they met any of the following criteria: a current and primary Axis I disorder other than major depressive disorder, including but not limited to dysthymia; a previous diagnosis of bipolar disorder, schizophrenia, or other psychotic disorders; any anxiety disorder as a primary diagnosis within the past year; an Axis II disorder that in the judgement of the investigator would interfere with compliance with the study protocol; a DSM-IV-defined history of substance abuse or dependence with in the past year, excluding nicotine and caffeine; a positive urine drug screen for any substances of abuse, including benzodiazepines; taking any excluded medications within 7 days prior to visit 2; treatment with a

Evidence tables Evidence Harms © Royal © College of Physicians 2019 monoamine oxidase inhibitor within 14 days prior to study onset; and treatment with fluoxetine within 30 days prior to study onset. Patients who had a prior treatment history with duloxetine, who were judged to be at serious suicide risk or who had had a serious medical illness likely to require hospitalization and/or the use of prohibited medications were also excluded, as were women who were breastfeeding or pregnant. Women of childbearing potential were required to

use reliable methods of birth control. Recruitment/selection of patients Patient were male and female outpatients of at least 18 years of age who met criteria for recurrent major depressive disorder as defined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) and confirmed via the MINI-International Neuropsychiatric Interview. Age, gender and ethnicity Author’s report that there were no significant differences between treatment groups on any measure of baseline demographics at the start of the double-blind maintenance phase.

At stat of maintenance phase: Age, mean (SD), y: Group 47.1 (12.8), Group 2: 48.0 (12.3) Women %: Group 68.5, Group 2: 74.6

Race, n (%):

212 White: Group 1: 143 (97.9), Group 2: 139 (97.9) Black: Group 1: 1 (0.7), Group 2: 1 (0.7) Hispanic: Group 1: 1 (0.7), Group 2: 2 (1.4) East Asian: Group 1: 1 (0.7), Group 2: 0 (0) South Asian: Group 1: 0 (0), Group 2: 0 (0) Extra comments Time to depressive recurrence was primary outcome. Efficacy, clinical outcomes and adverse events (including suicide) of treatment were reported. Indirectness of population No indirectness of population. Interventions & comparators Screening phase (3-9 days): assessed for study eligibility Open label acute treatment phase (4 to 10 weeks): all patients received duloxetine, initially at a dose of 60 mg/d day. In event of nonresponse after 4 weeks of treatment, the duloxetine dose was increased to 90 mg/d, and if response criteria had still not been met after 6 weeks of treatment a further dose increase to 120mg/day occurred. Open label continuation treatment phase (24 weeks): patients meeting response criteria were able to enter this phase; those that did not were discontinued from the study. Continued on same dose of duloxetine to which they had responded during the acute phase. Those that continued to meet response criteria after 24 weeks were able to enter next phase. Double blind maintenance period (52 weeks or until experienced a depressive recurrence): if randomised to duloxetine they continued on the same dose. If randomised to placebo they had their duloxetine treatment gradually down-titrated

Evidence tables Evidence Harms © Royal © College of Physicians 2019 over a 4 week period in order to minimize any antidepressant discontinuation syndrome that might have accompanied the transition. Optional follow up phase (3 weeks): down-titration occurred over a 2-to 3-week period depending on the dose the patient had been taking previously.

Definition of responders to treatment: if they met all of the following criteria: had a 17-item Hamilton Rating Scale for Depression total score ≤ 9, had a Clinical Global Impressions-Severity of illness scale score ≤ 2 and did not meet the DSM- IV criteria for a major depressive episode as assessed by the Mini-International Neuropsychiatric Interview for the DSM.

(n=146) Group 1: Duloxetine 60-120 mg once daily – doses taken were 60mg once daily by 44%, 90 mg once daily by 31% and 120mg once daily by 25%. (n=142) Group 2: Placebo

Funding Funded by Eli Lilly and company and Boehringer Ingelheim GmbH RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Duloxetine vs Placebo

213 Protocol outcome 1: Harm – Discontinuation emergent adverse events Actual outcome: Number (%) of patients with at least one discontinuation-emergent adverse events (DEAEs) – available case analysis

Group 1: 14/61 (23%) Group 2: 4/48 (8.3)

One placebo treated patient experienced a serious adverse event (uterine leiomyoma) whereas 1 duloxetine treated patient experience 3 serious adverse events (convulsion, hypertension, and hypertensive crisis).

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 69 from end of maintenance period (85 from follow-up phase); Group 2 Number missing: 50 (94). Reason given for end of maintenance/treatment phase: Recurrence criteria met (57), Adverse event (9), patient decision (35), protocol violation (9), physician decision (5), lost to follow-up (4). Reason for patient’s not entering follow-up phase not given (60).

Study Raskin 2008(20) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=311 randomised)

Duration of study 8 weeks treatment, 1 week discontinuation phase. Inclusion criteria All patients were 65 years and older. All patients met diagnostic criteria for MDD as defined in the Diagnostic and

ege of Physicians 2019 Statistical Manual of Mental disorders, Fourth Edition. The diagnosis was confirmed by the Mini International

Neuropsychiatric Interview. Baseline diseased severity was defined by patient’s scores on the HAMD17. Patients were required to have HAMD17 total score ≥ 18 at visits 1 and 2, mini-Mental State Examination score ≥ 20 with or without mild dementia; and at least one previous episode of MDD. Exclusion criteria Patients were excluded for the following reasons: current primary axis I diagnosis other than MDD or mild dementia; previous diagnosis of psychotic disorder; organic mental disorder, moderate to severe dementia or mental retardation diagnosis; and serious or unstable medical illness. Recruitment/selection of patients Elderly patients (≥ 65 years) with major depressive disorder.

Age, gender and ethnicity Sex, female, n (%): Group 1: 125 (60.4), Group 2: 60 (57.7) Mean age, yr (SD): Group 1: 72.6 (5.7), Group 2: 73.3 (5.7)

Age range: Group 1:65-90 , Group 2: 65-89

214 Ethnicity:

White: Group 1: 161 (77.8) , Group 2:82 (78.8) Hispanic: Group 1: 31 (15.0) , Group 2: 17 (16.3) African descent: Group 1: 13 (6.3), Group 2: 4 (3.8) Other: Group 1: 2 (1.0), Group 2: 1 (1.0) Extra comments Efficacy, clinical outcomes and treatment emergent adverse events also reported. Indirectness of population No indirectness of population. Interventions & comparators Screening phase (1 week) Double blind placebo phase (1 week) Treatment (8 weeks). If a patient could not tolerate the drug, the dose could be decreased from 60 to 30 mg/d but had to be at 60 mg/d by the end of the second week after randomization; otherwise the patient was discontinued from the study. Double blind discontinuation phase (1 week): dosage of the study drug was tapered.

(n=207) Group 1: duloxetine 60 mg once daily (n=104) Group 2: Placebo

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Funding Eli Lilly and Company and Boehringer Ingelheim GmbH.

RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Duloxetine vs placebo

Protocol outcome 1: Harm – Discontinuation-emergent adverse events

Actual outcome: Incidence of at least one discontinuation-emergent adverse event (defined as events that first occurred or worsened during the discontinuation phase of the study as compared with maximum severity at weeks 4 and 8). Group 1: 36/207 (17.3%), Group 2: 12/104 (11.3%) NB – actual numbers assumed by NGC calculations, % only provided in study.

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data – Very high, Outcome reporting - Low, Measurement – Low, Crossover -

Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 45; Group 2 Number missing: 24. Reason*: Adverse event (29), Lack of efficacy (16), not reported (24).

Actual outcome: Incidence of most frequent discontinuation-emergent adverse events Headache: Group 1: 3.1%, group 2: 1.2%

215 Dizziness: Group 1: 1.9%, Group 2: 1.2%

Fatigue: Group 1: 2.5%, Group 2: 0% Nausea: Group 1: 2.5%, Group 2: 0% Note: Study only provided percentage but NGC could not work out crude figures as did not match

Actual outcome: Serious adverse events during taper Group 1: 0, Group: 2: 1

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data – Very high, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 45; Group 2 Number missing: 24. Reason*: Adverse event (29), Lack of efficacy (16), not reported (24).

Study Rickels 2010(29) Study type RCT phase 3 (Patient randomised; Parallel)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Number of studies (number of participants) 1 (n=screened; n=374 randomised)

Countries and setting Multicentre – 49 sites: 31 in Europe, 15 in United States, and 3 in Taiwan. Duration of study 12 weeks open label treatment to assess responders for the 6 months treatment.

Inclusion criteria Male and female outpatients, 18 to 75 years of age, with a primary diagnosis of major depressive disorder (MDD) using the Diagnostic and Statistical Manual of mental Disorders, Fourth edition criteria, single or recurrent episode, without psychotic features, and who had symptoms for at least 30 days before screening, were eligible or study participation. The modified Mini-International Neuropsychiatric interview was used to indicate the primary diagnosis of MDD and any comorbid psychiatric disorders that may have been present, with confirmation by psychiatric interview. At screening and baseline, patients were also required to have a minimum HAM-D17 total score of 20, score at least 2 on item 1 (depressed mood) of the HAM-D17, and a clinical Global Impression – severity (CGI-S) score of at least 4 on a scale of 7 (1, normal, and 7, extremely ill).

Exclusion criteria Medical and psychiatric histories were obtained at screening. Patients with current comorbid substance use disorders were excluded; however, patients with comorbid generalized anxiety, panic or social anxiety disorder were allowed to participate as long as MDD was the primary diagnosis (i.e. the comorbid disorder did not causes a higher degree of distress or impairment than MDD), based on the opinion of the investigator. Other exclusion criteria included treatment with desvenlafaxine at any time in the past, treatment with venlafaxine within 90 days, or known hypersensitivity to

216 venlafaxine; risk of suicide based on clinical judgment; pregnant, breast-feeding , or planning to become pregnant during the study; current (within 12 months from baseline) manic episodes, posttraumatic stress disorder, obsessive – compulsive disorder, or clinically important personality disorder; depression associated with an organic mental disorder due to a general medical condition or neurological disorder; history of a seizure disorder; or clinically important medical disease. Recruitment/selection of patients Recruitment began in June 2003 and study ended August 2005. Age, gender and ethnicity Age, mean (SD), years: Group 1: 42.7 (12.3), Group 2: 42.8 (11.8) Women, n (%): Group 1: 127 (67), Group 2: 126 (68)

Ethnic origin, n (%); Black: Group 1: 11 (6), Group 2: 8 (4) Hispanic: Group 1: 7 (4), Group 2: 7 (4) Other: Group 1: 3 (2), Group 2: 9 (5) White: Group 1: 169 (89), Group 2: 161 (87) Extra comments Relapse rate was primary outcome. Clinical outcomes also reported. Author comments that this trial conducted before the current guidelines that a dose of 50mg/d of desvenlafaxine should be given to MDD patients.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Indirectness of population No indirectness of population.

Interventions & comparators Open label treatment phase (12 weeks): Desvenlafaxine at 200mg (2 tablets of 100mg). In the case of tolerability issues the dosage could be adjusted to desvenlafaxine at 100 mg from days 2 through 6. Patients were required to tolerate a daily dose of 200 mg by study day 7 to continue in the study. From weeks 2 through to 12 patients were treated with

desvenlafaxine at either 200 or 400 mg/d. Non responders were not eligible to enter next period and their dosages were tapered for a 1 to 2 week period. The taper period would be extended shortened, or omitted at the discretion of the investigator.

Double blind relapse prevention phase (6 months): Patients that responded to treatment in open label period (17-item Hamilton Rating Scale for Depression total score ≤ 11) were eligible to enter this phase. Patients were randomised to desvenlafaxine or placebo.

Taper period (1 to 2 weeks): taper period of 1 to 2 weeks could be extended, shortened, or omitted at discretion of investigator.

(n=189) Group 1: Desvenlafaxine succinate 200 or 400 mg/d (received same dose as they did at the end of the open 217 label phase). Dosages could not be adjusted during this period other than a decrease from 400 to 200 g/d for tolerability

reasons.

(n=185) Group 2: Placebo; patients were tapered off of desvenlafaxine; these patients received 200 mg/d during the first week and 100 mg/d during the second week of this phase, regardless of dosage of desvenlafaxine at the end of the open label treatment.

Funding This clinical trial and analysis were supported by Wyeth Research which was acquired by Pfizer Inc (October 2009). One author received support for conduct of the study via a grant to the University of Pennsylvania. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Desvenlafaxine vs placebo

Protocol outcome 1: Harm – Discontinuation Emergent signs and symptoms (DESS) score

Actual outcome: Mean (SD) DESS score after the post open label period – higher score worse Week 3 of double blind treatment phase: Group 1: 400 mg/d dose=1.15 (2.61) and 200 mg/d dose=0.73 (1.76), Group 2: 400 mg/d dose=2.16 (3.60) and 200 mg/d dose=2.28 (3.69) Note: Analysis not possible as total number in each dose sub-group not provided.

After week 3 of double blind treatment phase: Group 1: 400 mg/d dose=1.14 (2.37), Group 2: 400 mg/d dose=3.11 (4.51) Note: Analysis not possible as total number in each dose sub-group not provided. Also 200mg/d figures were not reported and authors stated that there was no significant difference.

Actual outcome: Mean (SD) discontinuation Emergent signs and symptoms (DESS) score at the end of the double blind treatment phase – higher score worse

Week 3 of taper: Group 1: 400 mg/d dose=1.75 (3.35), Group 2: 400 mg/d dose=0.59 (1.80) Note: Authors reported that there were no significant differences in mean DESS scores in the 200 mg/d desvenlafaxine group compared with the placebo group at any point during the taper period.

218 Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - High, Measurement – Low, Crossover - Low;

Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 58; Group 2 Number missing: 101. Reason*: unsatisfactory response (88), adverse events (54), remainder not reported (17).

Protocol outcome 2: Harm – Adverse events

Actual outcome: Total taper/post therapy-emergent adverse events Group 1: 101/190 (53%), Group 2: 52/185 (28%) Note: Unsure why authors reported n=190 for Group 1 when randomised number given was 189.

Actual outcome: taper/post therapy-emergent adverse events reported by at least 5% of patients: Dizziness: Group 1: 22%, Group 2: Not reported as less than 5% of patients Irritability: Group 1: 10%, Group 2: Not reported as less than 5% of patients Diarrhea: Group 1: 7%, Group 2: Not reported as less than 5% of patients Anxiety: Group 1: 6%, Group 2: Not reported as less than 5% of patients Fatigue: Group 1: 5%, Group 2: Not reported as less than 5% of patients Abnormal dreams: Group 1: 5%, Group 2: Not reported as less than 5% of patients Hyperhidrosis: Group 1: 5%, Group 2: Not reported as less than 5% of patients

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Headache: Group 1: 23/190 (12%), Group 2: 13/185 (7%) Insomnia: Group 1: 13/190 (7%), Group 2: 11/185 (6%) Nausea: Group 1: 26/190 (14%), Group 2: 9/185 (5%)

NB – actual numbers assumed by NGC calculations, % only provided in study.

Study Rosenthal 2013(25)

Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=1072 screened, n=874 entered open label response phase; n=752 completed the response phase and n=659 entered open label stability phase, n= 576 completed open label stability phase and n=548 randomised in treatment phase)

219 Countries and setting Multicentre: 87 sites in 14 countries worldwide (North America, 30; South America, 10; Europe, 44; South Africa, 3).

Duration of study 20 week screening period and 6 months treatment followed by a one week taper period. Inclusion criteria Male and female adult outpatients (≥18 years) with a primary diagnosis of single-episode or recurrent major depressive disorder (MDD) without psychotic features, based on criteria from the Diagnostic and Statistical Manual of mental disorders, Fourth Edition. Comorbid generalized anxiety, panic or social anxiety disorders were allowed if MDD was the primary diagnosis. Eligible patients had depressive symptoms for at least 30 days before the screening visit, a 17-item Hamilton Depression Rating Scale (HDRS17) total score ≥ 20, a HAM-D17 item 1 (depressed mood) score ≥ 2, and a clinical Global Impressions-Severity illness scale score ≥ 4 at screening and baseline visits. Exclusion criteria Excluded if they had been treated with desvenlafaxine at any time in the past or if they had a significant risk of suicide based on clinical judgement or an HDRS17 item 3 (suicide) score greater than 3 at screening. Other major exclusion criteria included current comorbid substance use disorder, manic episode, posttraumatic stress disorder, obsessive- compulsive disorder, clinically important personality disorder (as assessed by the modified Mini-International Neuropsychiatric Interview and a psychiatric interview), or clinically important medical disease. Recruitment/selection of patients MDD patients recruited from June 2009 and completed in March 2011. Age, gender and ethnicity Age, mean (SD), y: Group 1: 46.6 (13.0), Group 2: 45.3 (13.0) Sex Female, n (%): Group 1: 193 (71.0), Group 2: 198 (71.7)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Race, n (%) White: Group 1:240 (88.2), Group 2:230 (83.3) Black or African-American: Group 1:12 (4.4), Group 2: 15 (5.4)

Other: Group 1:20 (7.4), Group 2: 31 (11.2) Extra comments Primary outcome was time to relapse. Safety outcomes were reported. Indirectness of population No indirectness of population. Interventions & comparators Screening period (20 week): open label treatment period consisting of an 8-week response phase, followed by a 12- week stability phase. Only patients who had responded to 8-week open label treatment with desvenlafaxine 50 mg/d at week 8, as defined by an HDRS17 total score ≤ 11 and a clinical Global Impressions-Improvement scale score ≤ 2, were entered into the stability phase. Patients eligible for the stability phase received open label desvenlafaxine 50 mg/d for

an additional 12 weeks. Patients with continued stable response at the end of week 20 who did not have HDRS17 total score ≥ 16 or a CGI-I score ≥ 4 at any visit during the stability phase were eligible to enter the double blind treatment period.

Treatment period (6 months): Eligible patients randomly assigned to received desvenlafaxine or placebo. Patients who 220 were randomly assigned to receive placebo were tapered to double blind treatment with desvenlafaxine 25 mg/d during

the first week of the double blind period. Continued treatment until completion at 6 months or until relapse. Taper period (1 week): Patients who were administered desvenlafaxine 50 mg/d for 14 days or more during the study received desvenlafaxine 25 mg/d for a 7 day taper period at the study conclusion or at early discontinuation. Follow-up (1 week): Appointment scheduled approximately 7 days after the last tapered dose.

(n=272) Group 1: Desvenlafaxine succinate 50 mg/d (n=276) Group 2: Placebo Funding Pfizer. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Desvenlafaxine vs placebo

Protocol outcome 1: Harm – Adverse events Actual outcome: taper/post study-emergent adverse events (any adverse events that started or worsened the day after the last full dose of study medication, excluding taper dose) after double blind treatment: Group 1: 48/272 (17.6%), Group 2: 38/276 (13.8%)

Evidence tables Evidence Harms

Dizziness: Group 1: 13 /272 (4.8%), Group 2:1 /276 (0.4%) Depression: Group 1: 3/272 (1.1%), Group 2: 9/276 (3.3%) Headache: Group 1: 7/272 (2.6%), Group 2: 4/276 (1.4%)

Nausea: Group 1: 7/272 (2.6%), Group 2: 1/276 (0.4%) NB – actual numbers assumed by NGC calculations, % only provided in study.

Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 62; Group 2 Number missing: 100. Reason: Adverse event (9), failed to return (4) lack of efficacy (100), lost to follow-up (16), protocol violation (13), withdrawal by patient (18), other event (2).

Actual outcome: Number (%) serious adverse events during open-label taper/post therapy (2 week): Any serious adverse event: Group 1: 4/326 (1.2%); no comparator Suicidal ideation Group 1: 1/326 (0.4%) ; no comparator

Actual outcome: Number (%) serious adverse events during double blind taper/post therapy (2 week):

221 Any serious adverse event: Group 1: 1/272 (0.4%), Group 2: 1/276 (0.4%)

Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 62; Group 2 Number missing: 100. Reason: Adverse event (9), failed to return (4) lack of efficacy (100), lost to follow-up (16), protocol violation (13), withdrawal by patient (18), other event (2).

Study Rynn 2008 (23) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (515=screened; 327=randomised) Countries and setting USA, multicenter (27 outpatient departments) Duration of study 1 week placebo lead-in, 19 week acute therapy, 2 week taper / discontinuation Inclusion criteria 18 years or older, primary diagnosis of DSM-IV-DSM defined generalized anxiety disorder (GAD), severity ratings on the following measures: at least 4 (moderate) on the Clinical Global Impressions Severity Scale (CGI-S), a HADS anxiety subscale of ≥10, a Covi Anxiety Scale (CAS) total score ≥9, and the CAS total score had to be greater than the Raskin Depression Scale (RDS) total score. Patients were required to be medically healthy, as determined by physical exam,

Evidence tables Evidence H

© Royal © College of Physicians 2019 electrocardiogram, and laboratory results (renal, liver and thyroid function tests). They also had to undergo urine arms screens for benzodiazepine and illicit drug use. Females of potential childbearing status were required to use adequate contraceptive precautions. Exclusion criteria Those meeting the criteria for a recent (6 month0 diagnosis of MDD or substance abuse/dependence; a past history of

panic disorder, post-traumatic stress disorder, or eating disorder; or a lifetime history of psychotic, bipolar, obsessive- compulsive disorder, or psychosis. Patients were required to be free of psychotropic medications at least 2 weeks prior to randomization, with the exception of 4 weeks for those receiving fluoxetine. Lack of response of the GAD to two prior adequate trials of antidepressant or benzodiazepine treatments; any medical illness that would contraindicate the use of duloxetine; psychotherapy that was initiated within 6 weeks prior to enrollment; and the use of any concomitant medications that were excluded due to concerns about safety or psychotropic effects. Recruitment/selection of patients Unclear. Following the single-blind placebo lead-in week, patients were randomly assigned to receive either duloxetine or placebo. Randomisation procedure and allocation concealment not stated.

Age, gender and ethnicity Age, mean (SD), years: Duloxetine, 42.2 (13.9) Placebo, 41.0 (14.2) Sex, no (%) Female: Duloxetine, 103 (61.3) Placebo, 99 (62.3) Race, no (%), Caucasian: Duloxetine, 134 (79.8) Placebo, 124 (78) Race, no (%), African descent: Duloxetine, 20 (11.9) Placebo, 21 (13.2)

222 Race, no (%), Hispanic: Duloxetine, 7 (4.2) Placebo, 12 (7.6)

Race, no (%), Asian: Duloxetine, 7 (4.1) Placebo, 2 (1.3) Prior benzodiazepine treatment, no (%): Duloxetine, 29 (11.9) Placebo, 22 (13.8)

Study states no statistically significant differences in demographics, baseline symptoms or illness severity measures. Extra comments Efficacy, laboratory values and safety measures were also reported. Indirectness of population None Interventions & comparators Group 1: Duloxetine 60mg (n=168, completers=93) The starting dose was 60mg , but a dose decrease to 30mg was possible during the first two weeks to allow patients to acclimate to the medication. By week 2, patients in the duloxetine group were required to take a minimum dose of 60mg/day. Patients’ doses were progressively titrated at each subsequent visit; that is, they were required to increase their dose (30mg/day) if their CGI=I rating was ≥3 (minimal improvement, no change, or worsening) and they were able to tolerate a dose increase. Patients could be increased to a maximum dose of 120mg/day.

Group 2: Placebo (n=159, completers=109) Study visits conducted at 1, 2, 4, 7 and 19 weeks of double-blind treatment. Tapering at the end of treatment occurred over a 2-week period.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Funding Eli Lily and Company and Boehringer Ingelheim

RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: duloxetine vs placebo

Protocol outcome 1: Harm – Discontinuation-emergent adverse events (DEAE)

Actual outcome: DEAEs Duloxetine: n=21 (22.1%), Placebo: n=19 (17.3%)

Risk of bias: All domain – Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 7; Group 2 Number missing: 50. Reason: Adverse event (47), Patient decision (29), Lost to follow up (26), Physician decision (9), Lack of efficacy (10), Protocol violation (4).

Protocol outcome 1: Harm – Dizziness Actual outcome: Study states that only one DEAE, dizziness, occurred at a frequency of greater than 5% in duloxetine-treated patients. Duloxetine: n=6 (6.3%), Placebo: n=3 (2.7%) NB – actual numbers assumed by NGC calculations, % only provided in study.

Risk of bias: All domain – Very high, Selection - High, Blinding - Low, Incomplete outcome data - High, Outcome reporting - High, Measurement – Low, Crossover - Low;

223 Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 7; Group 2 Number missing: 50. Reason: Adverse event (47), Patient decision (29), Lost to follow up (26), Physician decision (9), Lack of efficacy (10), Protocol violation (4).

Study Schagen 2008 (24) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (341=screened; 265=randomised) Countries and setting Multicenter, multiple countries (33 study centres in Europe and South Africa) Duration of study 12 weeks active treatment (2 weeks dose escalation, 10 weeks stable full dose), 2 week de-escalation / discontinuation & 2 week follow-up Inclusion criteria Community-dwelling women of ≥65 years with symptoms of stress urinary continence or stress-predominant mixed urinary continence (S-MUI) for ≥3 consecutive months and ≥7 incontinence episodes per week as determined by the stress/urge incontinence questionnaire. For people in the S-MUI group, ≥50% of incontinence episodes had to be due to stress. Other inclusion criteria were a post-void residual ≤100mL identified by either catheterisation or ultrasound within 15 min of a spontaneous void and no language or significant cognitive barriers (modified min-mental state exam

Exclusion criteria 4 or more urinary tract infections in the preceding year or a positive urine culture at visit 1, any non-pharmacological intervention (surgery, bulking agents, initiation of pelvic floor muscle training) for incontinence or prolapse within 3 months before study entry or throughout the study, increased suicidal risk (score ≥2 on question 9 of the Beck

depression inventory), history of syncopal episodes, or hepatic dysfunction. Recruitment/selection of patients No details provided. Age, gender and ethnicity Age, mean (SD), years: Duloxetine, 70.63 (5.08), Placebo 71.13 (4.99) Race, %: Duloxetine, 99.2% Caucasian, 0.7% Hispanic. Placebo 100% Caucasian.

Study states demographic and baseline characteristics were comparable at baseline. Indirectness of population No serious indirectness – but NB 100% female

Interventions & comparators Group 1: Duloxetine 40 mg (n=134, Continued to discontinuation: 118) Patients received duloxetine 20mg BID for 2 weeks escalating to duloxetine 40mg BID for an additional 10 weeks followed by a 2 week double-blind de-escalation/discontinuation phase during which the duloxetine patients received 40mg once daily and a 2 week follow-up phase during which no medication was given to the patients

224 Group 2: Placebo (n=131, Continued to discontinuation: 119)

12 week active treatment, followed by 2 week double-blind de-escalation phase where patients continued to receive placebo. In the final 2 weeks follow up, no study medication was given. Funding Eli Lily and Company and Boehringer Ingelheim RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: duloxetine vs placebo

Protocol outcome 1: Harm – Discontinuation Emergent Adverse Events (DEAE) Duloxetine: n=25 (21.2%), Placebo: n=11 (9.2%) NB Actual numbers assumed from calculations by the NGC, only percentages reported in the study.

Risk of bias: All domain – Very high, Selection - High, Blinding - Low, Incomplete outcome data - Low, Outcome reporting – Very high, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 26; Group 2 Number missing: 18. Reason: Adverse event (28), Lack of response (3), Personal conflict/patient decision (5), Lost to follow-up (1), Protocol violation (3), Other (5).

No further detail provided on what the adverse events were – study states that there were no statistically significant differences for any individual DEAE.

Evidence tables Evidence Harms

Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n=227 randomised). Countries and setting 31 centres in 6 countries (Canada, Denmark, Estonia, Finland, Hungary and Sweden) from November 2007 to September 2009. Duration of study 6 month maintenance period. Inclusion criteria Male or female outpatients with a primary clinical diagnosis of generalized anxiety disorder (according DSM-IV-TR criteria) were eligible. Patients aged at least 18 years were required to have a screening 14-item HARS total score of ≥22, with HARS item 1≥2 and item 2≥2, with the sum of HARS item 1 and item 2 ≥5; a Montgomery Asberg Depression Rating

Scale totals core ≤ 16; and a decrease of <20%in HARS total score between screening and baseline. Patients were required to be physically healthy or to have stabilized somatic illnesses. Patients received agomelatine 25 mg/d during the first 4 weeks, then agomelatine was continued to the end of the open-label period (week 16) at a dose either maintained at 25 mg/d or increased to 50 mg/d in patients with insufficient improvement. Those patients who had improved sufficiently were eligible to continue in the study. Only patients who 225 met criteria for a clinical response (Hamilton anxiety Rating Scale [HARS] score ≤ 10 at week 16 and HARS score at week

16 minus HARS score at week 12 ≤ 4) were accepted to randomization stage.

Exclusion criteria Patients with any of the following disorders form DSM-IV-TR identified with the MINI, were excluded: panic disorder, posttraumatic stress disorder, agoraphobia, social phobia, obsessive compulsive disorder, or psychiatric disorder other than GAD. Patients were excluded if there was evidence of ongoing alcohol or drug abuse or any personality disorder that might compromise the study if they were at risk of suicide, had a score over 3 on item 10 of MADRS, or reported a suicide attempt with the pat year. Patients with a current generalized anxiety disorder episode resistant to at least 2 different previous antianxiety treatments during the past 12 months were also excluded. Patients were excluded if they had received any of the following recent/concomitant therapies: psychotherapy of any type started with 30 days prior to the selection visit and during the study or anxiolytics or antipsychotics taken within 1 week prior to the selection visit. Washout time was 5 weeks for fluoxetine; 4 weeks for clomipramine and imipramine; 2 weeks for paroxetine, sertraline, fluvoxamine, citalopram, escitalopram, duloxetine, and venlafaxine; and 1 week for mirtazapine. Benzodiazepines had to be stopped 2 weeks prior to screening in case of intermittent use and at least 6 months before screening visit in case of daily use. Other hypnotics/anxiolytics had to be discontinued at least one week prior to selection. Recruitment/selection of patients Recruited by advertisement (64.8%), self-referral (17.6%), or referral by a psychiatrist (5.2%), a psychologist (0.4%), GP or other specialist (12%).

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Age, gender and ethnicity Age, mean (SD), y: Group 1: 45.9 (14), Group 2: 47.0 (15.1) Female: Group 1: 62.8 %, Group 2: 61.4%

Authors report no significant demographic or clinical differences between the 2 groups.

Extra comments Primary outcome was risk of relapse. Other outcomes reported Hamilton Anxiety Rating Scale (HARS), Clinical Global Impressions scale (CGI), and hospital Anxiety and Depression scale (HAD), Leeds Sleep Evaluation Questionnaire (LSEQ) and Sheehan Disability Scale (SDS)and treatment emergent adverse events. Indirectness of population No indirectness. Interventions & comparators Week 1 to 16 - open label treatment period: Patients received agomelatine 25 mg/d during the first 4 weeks, then agomelatine was continued to the end of the open-label period (week 16) at a dose either maintained at 25 mg/d or increased to 50 mg/d in patients with insufficient improvement. Those patients who had improved sufficiently were

eligible to continue in the study.

Week 16 to 42 - maintenance period: (n=113) Group 1: Agomelatine orally 2 tablets taken in the evening

226 (n=114) Group 2: placebo

Week 42 – agomelatine treated patients (group 1) randomized to receive placebo or to continue with agomelatine (same dose) during 1 week to evaluate discontinuation symptoms. (n=89) Group 1: Agomelatine (n=75) Group 2: Placebo

Funding Sponsored by Servier (Suresnes, France). RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Agomelatine versus placebo

Protocol outcome 1: Harm - Discontinuation Emergent Signs and symptoms checklist (DESS) – evaluates discontinuation symptoms on a 43-item checklist: - Actual outcome: Mean (SD) DESS at week 43: Group 1: 0.9 (1.6), Group 2: 0.9 (1.9) Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study – Week 42 missing: 63/127, Group 1 missing: not reported; Group 2 Number missing: not reported, Reason: not reported.

Study Tourian 2009(30) Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (925=screened; 638=randomised) Countries and setting USA, multicenter (21 outpatient departments) Duration of study 8 weeks treatment, 1 week taper Inclusion criteria Eligible participants were male and female out-patients aged ≥18 years with a primary diagnosis of major depressive disorder (confirmed at screening by a psychiatrist using a modified Mini-International Neuropsychiatric Interview). Patients were required to have depressive symptoms for ≥30 days before the screening visit. The following minimum scores were required at screening and baseline: 17-item Hamilton Rating Scale for Depression (HAM-Dp) total score ≥20;

HAM-D17 item 1 (depressed mood) score ≥2; and Clinical Global Impressions-Severity (CGI-S) Scale score ≥4 (moderately ill). All sexually active patients in the study were required to use medically acceptable contraception. Exclusion criteria Previous treatment with desvenlafaxine; known hypersensitivity to venlafaxine; significant risk of suicide based on clinical judgment; pregnancy or breastfeeding; current psychoactive substance abuse or dependence, manic episode,

227 posttraumatic stress disorder, obsessive-compulsive disorder, or a lifetime diagnosis of bipolar or psychotic disorder; current primary generalized anxiety disorder, panic disorder, social anxiety disorder, or clinically important personality disorder; depression associated with the presence of an organic mental disorder; history of seizure disorder; clinically important medical disease; or use of prohibited treatments, including use of venlafaxine (immediate or extended release) and duloxetine within 90 days of baseline, investigational drugs, antipsychotics, fluoxetine within 30 days of baseline, or anxiolytics and other antidepressants within 14 days of baseline. Recruitment/selection of patients Not clearly stated apart from noting they were outpatients. After an initial screening period of 6 to 14 days, eligible patients were randomly assigned to treatment for 8 weeks plus 1 additional week for tapering. Age, gender and ethnicity Age, mean (SD), years: Desvenlafaxine 50 mg/d 41 (13), Desvenlafaxine 100mg/d 39 (12), Duloxetine 60mg/d: 39 (12), Placebo 39 (13). Sex, no (%) Female: Desvenlafaxine 50 mg/d 102 (69), Desvenlafaxine 100mg/d 99 (66), Duloxetine 60mg/d: 104 (66), Placebo 93 (58). Race, no (%), White: Desvenlafaxine 50 mg/d 111 (75), Desvenlafaxine 100mg/d 109 (73), Duloxetine 60mg/d: 117 (75), Placebo 121 (76). Race, no (%), Black: Desvenlafaxine 50 mg/d 29 (20), Desvenlafaxine 100mg/d 26 (17), Duloxetine 60mg/d: 25 (16), Placebo 22 (14). Race, no (%), Other: Desvenlafaxine 50 mg/d 8 (5), Desvenlafaxine 100mg/d 15 (10), Duloxetine 60mg/d: 15 (10), Placebo 17 (11).

Interventions & comparators (n=148) Group 1: Desvenlafaxine 50 mg/d (n=150) Group 2: Desvenlafaxine 100mg/d (n=157) Group 3: Duloxetine 60 mg/d

(n=160) Group 4: Placebo

Patients in the desvenlafaxine 50-mg and duloxetine 60-mg dose groups received their assigned treatment doses on study day 1 and continued the regimen until day 56 or early withdrawal. Patients in the desvenlafaxine 100-mg dose group had their doses titrated to their treatment dose. On days 1 to 7, they received desvenlafaxine 50 mg/d; on day 8, ans 2019 they received their assigned dose of desvenlafaxine 100 mg/d and continued on this regimen until day 56 or early withdrawal. For the 7-day taper period, the desvenlafaxine 50-mg/d dose group received placebo, the desvenlafaxine 100-mg/d dose group received desvenlafaxine 50 mg/d, and the duloxetine 60-mg/d dose group received duloxetine 30

mg/d. Patients assigned to placebo received placebo from study day 1 through day 56, or early withdrawal; they also received placebo during the taper period.

Funding Wyeth Pharmaceuticals sponsored the trial and analysis

228 RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Desvenlafaxine 50 mg/d vs desvenlafaxine 100mg/d vs duloxetine 60 mg/d vs placebo

Protocol outcome 1: Harm – Taper-emergent adverse events (classified on the Medical Dictionary for Regulatory Activities) NB Study notes that specific AEs were queried, so list might not include all experienced.

• Actual outcome: Drug withdrawal syndrome, n (%) Desvenlafaxine 50mg/d: 33 (22), Desvenlafaxine 100mg/d: 41 (27), Duloxetine 60 mg/d: 43 (27), Placebo 19 (12) • Actual outcome: Headache, n (%) Desvenlafaxine 50mg/d: 7 (5), Desvenlafaxine 100mg/d: 8 (5), Duloxetine 60 mg/d: 7 (5), Placebo 7 (4) • Actual outcome: Nausea, n (%) Desvenlafaxine 50mg/d: 7 (5), Desvenlafaxine 100mg/d: 7 (5), Duloxetine 60 mg/d: 9 (6), Placebo 8 (5) • Actual outcome: Upper respiratory tract infection, n (%) Desvenlafaxine 50mg/d: 6 (4), Desvenlafaxine 100mg/d: 3 (2), Duloxetine 60 mg/d: 8 (5), Placebo 2 (1) • Actual outcome: Dizziness, n (%) Desvenlafaxine 50mg/d: 5 (3), Desvenlafaxine 100mg/d: 15 (10), Duloxetine 60 mg/d: 9 (6), Placebo 3 (2)

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Risk of bias: All domain – Low, Selection - Low, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - High, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 7; Group 2 Number missing: 10, Group 3 Number missing; 2, Group 4 Number missing; 3. Reason: Lost to follow-up (32), Adverse event (49), Failed to return (8), Patient request not study related (15), Unsatisfactory response (14), Protocol violation (16), Investigator request (1), Other event (2).

Protocol outcome 2: Harm – DESS score Actual outcome: Discontinuation-emergent signs and symptoms (DESS) in the completer population, mean score (SD). NB Desvenlafaxine 50mg stopped medication at the final evaluation, desvenlafaxine 100mg and duloxetine 60mg were both on half dose for week 1 post evaluation. • Final double blind evaluation: Desvenlafaxine 50mg/d, n=117: 0.9 (2.0), Desvenlafaxine 100mg/d, n=108: 0.7 (1.6), Duloxetine 60 mg/d, n=117: 1.0 (2.3), Placebo, n=118: 0.7 (1.8) • Week 1 post evaluation: Desvenlafaxine 50mg/d, n=112: 3.1 (5.1), Desvenlafaxine 100mg/d, n=107: 1.9 (2.9), Duloxetine 60 mg/d, n=111: 2.4 (3.7), Placebo, n=111: 1.3 (2.4)

• Week 2 post evaluation: Desvenlafaxine 50mg/d, n=115: 1.6 (3.3), Desvenlafaxine 100mg/d, n=100: 3.1 (5.0), Duloxetine 60 mg/d, n=103: 3.2 (5.6), Placebo, n=117: 1.6 (3.5) • Week 3 post evaluation: Desvenlafaxine 50mg/d, n=7: 1.4 (2.7), Desvenlafaxine 100mg/d, n=12: 1.7 (3.4), Duloxetine 60 mg/d, n=13: 1.0 (2.8), Placebo, n=4: 7.3 (4.9)

229 • >Week 3 post evaluation: Desvenlafaxine 50mg/d, n=3: NR, Desvenlafaxine 100mg/d, n=6: 1.3 (2.4), Duloxetine 60 mg/d, n=1: 3.0 (NA), Placebo, n=3: NR

Risk of bias: All domain – Very high, Selection - Low, Blinding – Very high, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study – No. missing varies over weeks (n reported with results above). Reason: Lost to follow-up - Not reported.

Protocol outcome 3: Harm – Serious adverse events Actual outcome: Suicidal ideation in post study period Group 1: 1/148, Group 2: 0/150, Group 3: 0/157, Group 4: 0/161

Actual outcome: Suicidal attempt in post study period Group 1: 0/148, Group 2: 0/150, Group 3: 1/157, Group 4: 0/161

Additional serious adverse events reported: Group 1: accidental injury, Group 2: gastrointestinal haemorrhage and menorrhagia; Group 3: nothing additional, Group 4: accidental injury and asthma.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 NB For analysis in this review Group 1 and Group 2 (Desvenlafaxine) will be pooled.

Study YaDeau 2016(18) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=106 randomised). Countries and setting Hospital for special surgery, Weill Cornell medical School – affiliated teaching hospital that specializes in musculoskeletal disease. USA. Duration of study 15 days.

Recruited from November 2013 to May 2015. Inclusion criteria English speakers, age 25 to 75 years, who were judged able to follow the protocol and who planned to have regional anesthesia and to be discharged either to home or to a participating rehabilitation center.

230 Exclusion criteria Planned general anesthesia, allergy or intolerance to one of the study medications, major previous ipsilateral open knee surgery, American Society of Anesthesiologists physical status IV, hepatic failure, renal failure, contraindication to using dexamethasone in the peripheral nerve block, difficult to manage diabetes mellitus, chronic gabapentin or pregabalin use, chronic opioid use. Concurrent use of duloxetine or other SNRIs, monoamine oxidase inhibitors, tricyclic antidepressants, triptans, lithium, busprione, or St John’s Wort. Recruitment/selection of patients People with osteoarthritis scheduled for primary total knee arthroplasty. Age, gender and ethnicity Age, y (range): Group 1: 67 (61-71), Group 2: 63 (57-67) Male %: Group 1: 47.2%, Group 2: 49.1%

Race Asian: Group 1: 1.9%, Group 2: 0% Black/African American: Group 1: 3.8%, Group 2: 7.5% White: Group 1: 90.6%, Group 2: 90.6% Other: Group 1: 1.9%, Group 2: 0% Declined: Group 1: 1.9%, Group 2: 1.9%

Extra comments Efficacy and opioid use following operation was reported. Anxiety and depression scores were reported during the 14 days treatment.

Interventions & comparators (n=53) Group 1: Duloxetine 60 mg oral daily for 15 days. A capsule (60mg) given approximately 30 minutes before transfer to the operating room. Patients received one capsule per day up to and including postoperative day 14. (n=53) Group 2: Placebo

Anesthetic and Analgesic protocol: patients received a standardized anesthetic and multimodal analgesic protocol. Included a combined spinal epidural anesthetic, adductor canal block, intravenous sedation with midazolam and propofol; intraoperative intravenous dexamethasone, intravenous ketorolac, pubivacaine/hydromorphone, meloxicam, and oxycodone. Patients were discharged with meloxicam and oxycodone/acetaminophen. Funding Supported by Hospital for Special Surgery Anesthesiology Department Research and Education Fund, New York. The REDCap electronic data capture tools funded by the Weill Cornell clinical and Translational Science Center grant from National Center for Advancing Translational Sciences, National Institutes of Health, Bethesda, Maryland.

RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: duloxetine versus placebo

Protocol outcome 1: Harm Actual outcome: Change (defined from question ‘Have you noticed any changes since you stopped taking the study drug?’. If yes the symptoms were noted and 231 assessed for either increased pain and /or possible evidence of duloxetine discontinuation syndrome.

Group 1: 14/53 (26.4%), Group 2: 11/53 (20.8%); Wilcoxon-Mann-Whitney Odds or Relative Risk (95% CI: 0.78 (0.36 to 1.60), Risk difference: -6.3% (-23.3 to 11.2), p=0.485 Risk of bias: All domain –Low, Selection - Low, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 0; Group 2 Number missing: 0, Reason*: not applicable.

Protocol outcome 1: Harm Actual outcome: Discontinuation syndrome at day 18-20 (defined as new-onset symptoms including dizziness, nausea, headache, fatigue, irritability, insomnia and anxiety). Group 1: 9/53 (17%), Group 2: 5/53 (9.4%); Wilcoxon-Mann-Whitney Odds or Relative Risk (95% CI: 0.56 (0.20 to 1.54), Risk difference: -8.3% (-22.8 to 6.2), p=0.247

Risk of bias: All domain –Low, Selection - Low, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 0; Group 2 Number missing: 0, Reason*: not applicable. Narrative information: Seven adverse events reported to the institutional review board: three in placebo group (atrial fibrillation; nausea, insomnia and hypervigilance; and headache) and four in the duloxetine group (nausea, drowsiness, and fatigue; somnolence and dizziness; confusion and syndrome of inappropriate antidiuretic hormone secretion; and nausea, drowsiness and hypervigilance. This last patient was diagnosed with duloxetine discontinuation syndrome by an outside physician and was treated with a tapering course of duloxetine with resolved the symptoms. The patient expressed a desire to continue duloxetine as it improved her mood and reduced pain.

Study Yonkers 2015(19) Study type RCT (Patient randomised; Parallel)

Number of studies (number of participants) 1 (n= 252 randomised). Countries and setting 3 University medical centres, Connecticut, New York and Virginia in the USA.

Duration of study 6 months (consisting of 6 menstrual cycles) Conducted from 1 September 2007 to 29 February 2012. Inclusion criteria Women aged 18 to 48 years of age, had menstrual cycles of 21 to 35 day, and met criteria for premenstrual dysphoric disorder (PDD) were eligible.

232 Exclusion criteria Ineligible if met criteria for a major depressive episode, bulimia, or a substance use condition other than tobacco use; had lifetime bipolar disorder or a psychotic illness; had severe suicidal thoughts; were undergoing treatment with a psychotropic medication; were using an oral contraceptive that included drospirenone; were receiving a depot hormonal preparation or using an intrauterine device that could stop menses; used any oral contraceptive for less than 6 months before screening or did not plan to continue the same hormonal contraceptive use through the study; used an inadequate birth control method; had a history of hypersensitivity to sertraline; were pregnant or lactating; were planning on relocating during the study period; or were unable or unwilling to provide informed consent. Recruitment/selection of patients Recruited by flyers, newspaper advertisement and direct mail to women aged 18 to 40 years in local zip codes. Age, gender and ethnicity Age, mean (SD) y, Group 1: 33.7 (6.7), Group 2: 34.6 (6.9) Race: White: Group 1: 68.8%, Group 2: 70.1% Black: Group 1: 15.2%, Group 2: 15.7% Hispanic: Group 1: 12.0%, Group 2: 10.2% Asian, mixed or other: Group 1: 4.0%, Group 2: 3.9%

Educational level of some college is higher in group 1 (32%) compared to group 2 (17.3%). Extra comments Efficacy (Premenstrual tension scale, Inventory of Depressive Symptomatology-Clinical Rated, daily record of severity problems) and adverse events were reported.

ence tables

Interventions & comparators Pretrial assessment of at least 2 months to confirm a diagnosis of PMDD. Women who did not achieve a clinical Global Impression Scale Severity rating of 2 or less, indicating a reduction in severity of symptoms after 2 months at a dosage of 100mg/d or the highest tolerate dosage were offered removal from

the trial and daily sertraline rescue treatment. Participants who completed the trial were also offered 3 months of open label daily continuation treatments.

(n=125) Group 1: Sertraline hydrochloride, 50 to 100 mg/d during symptomatic interval for 6 menstrual cycles. Starting dose was 50 mg (two 25 mg capsules) taken once per day during the symptomatic interval. Participants were instructed to begin taking sertraline when they first noticed onset of their typical premenstrual symptoms and were asked to cease taking pills within a few days of their menstrual flow, around the time that these symptoms typically

ended. Participants with an inadequate response were instructed to increase their daily dose to a maximum of 4 capsules (100mg) Participants were instructed to titrate the dose by 2 capsules every 2 days to the final dose of 4 capsules and follow the reverse schedule to end dosing. Women who reported moderate to severe adverse effects were allowed to reduce the daily dose to 1 capsule but to increase the dose after the next cycle unless rate limiting adverse effects continued. 233

(n=127) Group 2: Placebo Funding Supported by grants from the National Institute of Mental Health. The study drug and matching placebo were donated by Pfizer, Inc. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Sertraline versus placebo

Protocol outcome 1: Harm – Withdrawal symptoms - Actual outcome: Michelson SSRI withdrawal Scale – the mean of each item was summed for the 3 days after treatment ended for each menstrual cycle. Scores ranged from 0-51 with higher scores indicating more severe withdrawal symptoms. Cycle 1: Group 1: 72.9 (29.3), Group 2: 68.1 (25.4) Cycle 2: Group 1: 73.4 (29.9), Group 2: 75.2 (33.1) Cycle 3: Group 1: 70.3 (27.9), Group 2: 71.7 (39.0) Cycle 4: Group 1: 62.3 (25.1) Group 2: 65.4 (25.3) Cycle 5: Group 1: 63.9 (19.9) Group 2: 62.1 (24.0) Cycle 6: Group 1: 67.5 (22.1) Group 2: 76.2 (34.1) Note: it was unclear from the study what units these outcomes were reported in.

Evidence tables Evidence Harms © Royal © College of Physicians 2019 No further details provided. Risk of bias: All domain –High, Selection - Low, Blinding - Low, Incomplete outcome data - High, Outcome reporting - Low, Measurement – Low, Crossover - Low;

Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 37; Group 2 Number missing: 39, Reason*: out of contact (29), withdrawal of consent (28), adverse event (4), rescue therapy (12), pregnancy (2), protocol violation (1).

Short term versus long term opioid use Study Hoffman, 2017 (9) Study type Retrospective cohort Number of studies (number of participants) 1 (n=2892) Countries and setting USA, Rochester Epidemiology Project (REP) database

Duration of study Census data from January 1, 2006 – December 31, 2010. Inclusion criteria People receiving opioid prescriptions (codeine sulfate, fentanyl citrate, hydrocodone bitartrate, hydromorphone hydrochloride, meperidine hydrochloride, methadone hydrochloride, morphine sulfate, oxycodone hydrochloride,

234 oxymorphone hydrochloride, propoxyphene hydrochloride or napsylate, tapentadol hydrochloride, and tramadol hydrochloride).

Exclusion criteria Not reported Recruitment/selection of patients Review of a comprehensive prescription database from all ambulatory practice professionals for prescriptions issued in the study period. Indication for opioid prescriptions and the prescribing clinician’s specialty among those receiving 90 or more consecutive days of opioid therapy were obtained by reviewing clinical documentation around the time that the prescription was issued. Age, gender and ethnicity Age, median (range), y: Long term use 70 (26-98), Short term 69 (3 – 101) Sex, female, No. (%): Long term 308 (56.9), Short term 674 (46.4) Duration of consecutive opioids, median (IQR), d Long term 228 (133-392), Short term 17 (8-34) Oxycodone was the most common opioid prescribed (45.9%) Extra comments None Indirectness of population Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. Interventions & comparators Long term opioid use (>90 days) n= 545 vs shorter term opioid use n= 1464. Intervals of continuous opioid exposure were defined as either the duration of a single prescription or the summed

Evidence tables Evidence Harms © Royal © College of Physicians 2019 durations of multiple prescriptions when there was less than or equal to 10 days between the expected end date of 1 prescription and the start date of the next prescription. Long-term opioid therapy was defined as greater than or equal to 90 continuous days based on the length of use to treat chronic instead of acute pain.

Study also included a matched healthy control group (not included here) Funding Mayo foundation for Medical Education and Research & the NIH. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: hypnotic users vs controls

Protocol outcome 1: Harm – Harms of dependence (long term use) Actual outcome: No. with depression (%) & HR (95% CI) Long term: 341 (63.0), Short term: 633 (43.6) HR 1.53 (1.29-1.82) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids

received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Alcohol abuse n (%) and HR: Long term: 54 (10), Short term 109 (7.5) HR 1.38 (0.90-2.11) 235 Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome

reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Opioid abuse n (%) and HR: Long term: 9 (1.7), Short term 2 (0.1) HR 3.97 (0.87-28.9) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Other substance abuse n (%) and HR: Long term: 25 (4.6), Short term 27 (1.9) HR 1.81 (0.92-3.58) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Evidence tables Evidence Harms

Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS.

llege of Physicians 2019

No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Other substance overdose n (%) and HR: Long term: 22 (4.1), Short term 24 (1.7) HR 1.82 (0.92-3.6) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Opioid dependence n (%) and HR: Long term: 39 (7.2), Short term 20 (1.4) HR 2.85 (1.54-5.47) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS.

236 No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Other substance dependence n (%) and HR: Long term: 95 (17.6), Short term 129 (8.9) HR 1.73 (1.21-2.49) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Actual outcome: Harm – Mortality by 11/25/16 n (%) and HR: Long term: 256 (47.3), Short term 530 (36.5) HR 0.99 (0.84-1.16) Risk of bias: All domain –High, Confounding – High, Selection - High, Classification of intervention – Low, Deviations from intervention – High, Attrition - Low, Outcome reporting - Low, Measurement – Low, Very serious indirectness; Long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. No. missing not reported – but analysed with cox proportional HRs censoring for missing data.

Study Baldessarini 2010(3) Study type Observational (cohort)

Number of studies (number of 1 (n=398) participants) Countries and setting Conducted in Italy; Setting, outpatients Duration of study Intervention + follow-up, mean (SD) follow-up of 2.81 (3.63) years after discontinuation. Includes participants discontinuing between 1975 and 2008. Inclusion criteria Clinical diagnosis of DSM-based recurrent major depressive disorder, bipolar I or II disorder, or panic disorder; received a tricyclic antidepressant (or tricyclic-like tetracyclics maprotiline and mianserin), a modern antidepressant (serotonin reuptake inhibitors or

bupropion, duloxetine or venlafaxine), or more than one antidepressant, with or without a mood stabiliser, following standard clinical practices regarding drug selection and dosing in the study community; recovered from an antidepressant-treated index episode of major depression or panic disorder, based on clinical euthymia and a score ≤7 on the Hamilton Depression Rating Scale sustained for at least 30 days; discontinued medication electively for clinical or personal reasons over a known period of time, allowing catergorisation into groups based on rapid (1-7 days) or gradual (≥2 weeks) discontinuation; remained clinically stable or

237 euthymic for at least 1 week after discontinuing treatment; and remained under prospective observation for at least 1 year,

during initial treatment and through a first new episode of major depression or panic disorder that met DSM-IV diagnostic criteria at clinical assessment. Exclusion criteria Patients who were moderately clinically depressed, anxious or hypomanic at the time of medication discontinuation; patients whose rate of discontinuation was uncertain. Recruitment/selection of patients Consecutive patients meeting inclusion criteria. Age, gender and ethnicity Age – mean (SD) at drug discontinuation: Rapid discontinuation, 39.5 (14.5) years; Gradual discontinuation, 44.1 (15.4) years. Age – mean (SD) at illness onset: Rapid discontinuation, 29.9 (11.7) years; Gradual discontinuation, 34.3 (15.3) years. Gender (M:F): Rapid discontinuation, 72/116; Gradual discontinuation, 65/145. Ethnicity: Not reported. Extra comments Note that antidepressants may have been taken alongside mood stabilisers, anticonvulsants, atypical antipsychotics or sedative- anxiolytic benzodiazepines as adjunctive medications. Adjunctive psychotropic medications were continued unchanged after discontinuation of antidepressants.

Reasons for discontinuation were clinical, not experimental. Decisions to discontinue were made by the patient in 80.7% of cases and at the advice of the prescribing physician in 19.3% of cases. Rapid discontinuation was the patient decision in 94.1% of cases. Indirectness of population Serious indirectness. Not all of the antidepressants are listed (e.g. serotonin reuptake inhibitors grouped together and individuals drugs not listed), therefore could include some not on included list.

Interventions & comparators Treatment was determined clinically and included use of standard antidepressants, alone or with mood stabilisers (lithium carbonate or anticonvulsants that have received regulatory approval), atypical antipsychotics, or sedative-anxiolytic benzodiazepines as adjunctive medications. Adjunctive psychotropic medications were continued unchanged after

discontinuation of antidepressants. Antidepressant dosages were standardised based on the relative potency of the various agents modified by median daily doses of specific antidepressants based on experience with ~2000 patient with major depressive disorder treated at the study centre.

All patients were clinically well at the time of discontinuing antidepressant treatment and had no clinical evidence of even mild depression or anxiety disorder.

(n=188) Intervention 1: Managed withdrawal. Rapid withdrawal of antidepressants (1-7 days).

Concurrent medication/care: Any patients taking adjunctive medications alongside antidepressant treatment continued the adjunctive medications whilst discontinuing antidepressants.

(n=210) Intervention 2: Managed withdrawal. Gradual withdrawal of antidepressants (≥2 weeks).

238 Concurrent medication/care: Any patients taking adjunctive medications alongside antidepressant treatment continued the adjunctive medications whilst discontinuing antidepressants.

Funding Supported in part by a National Institutes of Health grant, the Lucio Bini Private Donors Mood Disorders Research Fund and a grant from Bruce J. Anderson Foundation and the McLean Private Donors Psychopharmacology Research Fund. Dr. Baldessarini has been a consultant to or received research support from AstraZeneca, Auritex, Biotrofix, Eli Lilly, IFI, Janssen, JDS-Noven, Luitpold, Merck, NeuroHealing, Novartis, Pfizer, SK Bio-Pharmaceuticals and Solvay Corporation and has taught prepared CME materials for New England Educational Institute and Pri-Med. Dr. Tondo has been a consultant to or received research support from IFI, Janssen, Eli Lilly, Merck, and Solvay. RESULTS AND RISK OF BIAS FOR COMPARISON: Rapid vs gradual withdrawal

Protocol outcome 1: Intensity and duration of signs and symptoms an overall withdrawal syndrome. -Actual outcome: Time to first new illness (major depressive or panic disorder episode) within 1 year of rapid or gradual antidepressant discontinuation. All patients underwent initial diagnostic assessments, treatment and follow-up evaluations by the same mood disorders expert, based on semi-structured interviews that followed the mood disorder components of the Research Diagnostic Criteria and Structured Clinical Interview for DSM-IV research assessment procedures, in addition to extensive follow-up clinical assessments and repeated assessments with standard mood disorder rating scales during systematic prospective follow-up every 2-4 months.

HR obtained from Cox multivariate modelling of survival functions, with covariates including factors likely to reflect illness severity, such as diagnosis, prior recurrence rate, concomitant treatment with a mood stabiliser or other drugs, and duration of the first new episode of illness, as well as aspects of treatment including total daily antidepressant dose and duration of index treatment. Shorter latency observed after rapid discontinuation compared with gradual discontinuation: HR, 1.50 (95% CI,

1.14-2.17; X2=38.3; df=1; P<0.0001). Median time to new illness was 3.77 months (95% CI, 1.99-5.55) for rapid discontinuation and 9.07 months (95% CI, 6.78-11.4) for gradual discontinuation. Control group (gradual discontinuation) risk at 12 months after discontinuing antidepressant treatment: 124/210 had experienced the event at 12 month time-point (vs. 146/188 in rapid discontinuation group). Risk of bias: All domain – Very high, Confounding – Low, Selection – High, Classification of interventions – High, Deviations from interventions – High, Attrition – Low, Outcome reporting – Low, Measurement – High, Indirectness of outcome: No indirectness. Group 1 Number missing: 0, Reason missing: NA; Group 2 Number missing: 0, Reason missing: NA.

Study Gallagher 2012(26) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=500 randomised for titration phase, n=384 randomised for taper stage).

239 Countries and setting Not reported.

Duration of study Titration period 1 week and then 14 week treatment; and 2 week taper period followed by 8 weeks follow-up. Inclusion criteria Health, postmenopausal women who experienced ≥ 50 moderate to severe hot flashes per week during each of the 2 weeks immediately preceding randomization. Exclusion criteria Excluded if had taken hormone-containing products or prohibited medications within 2 to 26 weeks before study initiation, had experienced major depressive disorder or generalized anxiety disorder requiring treatment within the previous 6 months or had a history of bipolar disorder or psychotic disorder. Recruitment/selection of patients Post-menopausal women with vasomotor symptoms. Age, gender and ethnicity Baseline demographic and clinical characteristics were reported as similar across the titration phase and taper phase. Mean (SD) age: Group 1: 54.52 (5.01), Group 2: 54.40 (6.37), Group 3: 53.98 (5.16), Group 4: 53.48 (5.27)

Race, n (%): Asian: Group 1: 2 (1.6), Group 2: 1 (0.8), Group 3: 2 (1.6), Group 4: 1 (0.8) Black: Group 1: 12 (9.5), Group 2: 12 (9.9), Group 3: 8 (6.5), Group 4: 19 (15.6) White: Group 1: 109 (86.5), Group 2: 107 (88.4), Group 3: 112 (90.3), Group 4: 97 (79.5) Other: Group 1: 3 (2.4), Group 2: 1 (0.8), Group 3: 2 (1.6), Group 4: 5 (4.1)

Extra comments Adverse events during initial titration phase also reported. Indirectness of population No indirectness of population. Interventions & comparators Titration phase (1 week): participants randomised to receive desvenlafaxine 100mg/d (no titration), desvenlafaxine 50

mg/d, desvenlafaxine 25 mg/d (4 days) then 50 mg/d (3 days) or desvenlafaxine 25 mg/d Treatment (15 weeks): Participants received open-label desvenlafaxine 100 mg/d Taper phase (2 weeks): Participants who had complete the open label phase or who had taken 5 weeks or more of treatment at the time of early discontinuation were randomised to one of the taper regimens below Follow-up (8 weeks): Follow up for discontinuation symptoms

(n=94) Group 1: Desvenlafaxine succinate 50mg/d for 7 days followed by 25 mg/d for 7 days

(n=101) Group 2: Desvenlafaxine succinate 50 mg/d every other day for 14 days (n=87) Group 3: Desvenlafaxine succinate 50 mg/d for 7 days followed by placebo for 7 days (n=102) Group 4: Placebo Funding Wyeth Research, which was acquired by Pfizer Inc in 2009.

240 RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: Desvenlafaxine 50mg/d followed by 25mg/d vs Desvenlafaxine 50 mg/d every other day vs desvenlafaxine 50 mg/d for 7 days then placebo vs placebo (no tapering)

Protocol outcome 1: Harm –Discontinuation-Emergent Signs and Symptoms (DESS) checklist (higher scores indicate more symptoms experienced) Actual outcome: Mean (SD) DESS checklist total scores after taper baseline (end of open label treatment) Group 1 (n=68): 1.43 (2.51), Group 2 (n=66): 1.11 (1.76), Group 3 (n=60): 1.43 (2.82), Group 4 (n=74): 1.66 (3.19) Actual outcome: Mean (SD) DESS checklist total scores after taper week 1 Group 1 (n=86): 2.28 (3.59), Group 2 (n=100): 2.26 (3.46), Group 3 (n=83): 1.84 (3.36), Group 4 (n=98): 7.07 (7.13) Actual outcome: Mean (SD) DESS checklist total scores after taper week 2 Group 1 (n=87): 2.44 (5.53), Group 2 (n=89): 1.19 (1.96), Group 3 (n=79): 4.46 (6.44), Group 4 (n=91): 2.44 (4.96) Actual outcome: Mean (SD) DESS checklist total scores after taper week 3 Group 1 (n=57): 4.11 (5.77), Group 2 (n=59): 3.22 (4.82), Group 3 (n=47): 1.70 (3.16), Group 4 (n=59): 1.78 (3.20)

Note: Also reported taper week 1 and 2 combined and week 1, 2 and 3 combined.

Protocol outcome 2: Harm –Incidences of specific discontinuation-emergent signs and symptoms Actual outcome: Number (% - based on number of participants who had a least one assessment during the tapering phase) of harms at baseline (end of treatment phase): Dizziness light-headedness, sensation of spinning: Group 1: 6 (8.8), Group 2: 2 (3.0), Group 3: 5 (8.3), Group 4: 3 (4.1) Headaches: Group 1: 4 (5.9), Group 2: 2 (3.0), Group 3: 4 (6.7), Group 4: 3 (4.1) Increased dreaming/nightmare: Group 1: 5 (7.4), Group 2: 6 (9.1), Group: 8 (13.3), Group 4: 7 (9.5) Irritability: Group 1: 2 (2.9.8), Group 2: 0, Group 3: 3 (53), Group 4: 2 (2.7) Nausea: Group 1: 3 (4.4), Group 2: 3 (4.6), Group 3: 2 (3.3), Group 4: 2 (2.7)

Sudden worsening of mood: Group 1: 0, Group 2: 0, Group 3: 0, Group 4: 4 (5.4) Sweating more than usual: Group 1: 7 (10.3), Group 2: 6 (9.1), Group 3: 9 (15), Group 4: 8 (10.8) Trouble sleeping, insomnia: Group 1: 2 (2.9), Group 2: 10 (15.2), Group 3: 4 (6.7), Group 4: 3 (4.1)

241 Actual outcome: Number (% - based on number of participants who had a least one assessment during the tapering phase) of harms after taper week 1:

Dizziness light-headedness, sensation of spinning: Group 1: 7 (8.1), Group 2: 18 (18.0), Group 3: 9 (10.8), Group 4: 41 (41.8) Headaches: Group 1: 5 (5.8), Group 2: 8 (8.0), Group 3: 9 (10.8), Group 4: 28 (28.6) Increased dreaming/nightmare: Group 1: 13 (15.1), Group 2: 14 (14), Group 3: 12 (14.5), Group 4: 35 (35.7) Irritability: Group 1: 6 (7), Group 2: 6 (6.0), Group 3: 6 (7.2), Group 4: 26 (26.5) Nausea: Group 1: 7 (8.1), Group 2: 14 (14.0), Group 3: 5 (6.0), Group 4: 28 (28.6) Sudden worsening of mood: Group 1: 6 (7.0), Group 2: 3 (3.0), Group 3: 4 (4.8), Group 4: 22 (22.5) Sweating more than usual: Group 1: 19 (22.1), Group 2: 23 (23.0), Group 3: 14 (16.9), Group 4: 44 (44.9) Trouble sleeping, insomnia: Group 1: 11 (12.8), Group 2: 13 (13.0), Group 3: 8 (9.6), Group 4: 37 (37.8)

Actual outcome: Number (%) of harms after taper week 2 Dizziness light-headedness, sensation of spinning: Group 1: 12 (13.8), Group 2: 2 (2.3), Group 3: 21 (26.6), Group 4: 7 (7.7) Headaches: Group 1: 5 (5.8), Group 2: 4 (4.5), Group 3: 8 (10.1), Group 4: 4 (4.4) Increased dreaming/nightmare: Group 1: 5 (5.8), Group 2: 4 (4.5), Group 3: 10 (12.7), Group 4: 8 (8.8) Irritability: Group 1: 10 (11.5), Group 2: 5 (5.6), Group 3: 18 (22.8), Group 4: 13 (14.3) Nausea: Group 1: 6 (6.9), Group 2: 3 (3.4), Group 3: 15 (19.0), Group 4: 4 (4.4)

Sudden worsening of mood: Group 1: 5 (5.8), Group 2: 6 (6.7), Group 3: 18 (22.8), Group 4: 8 (8.8) Sweating more than usual: Group 1: 17 (19.5), Group 2: 19 (21.4), Group 3: 23 (29.1), Group 4: 23 (25.3) Trouble sleeping, insomnia: Group 1: 9 (10.3), Group 2: 8 (9.0), Group 3: 22 (27.9), Group 4: 19 (20.9)

Actual outcome: Number (%) of harms after taper week 3 Dizziness light-headedness, sensation of spinning: Group 1: 17 (29.8), Group 2: 13 (22.0), Group 3: 2 (4.3), Group 4: 0 Headaches: Group 1: 11 (19.3), Group 2: 4 (6.8), Group 3: 3 (6.4), Group 4: 5 (8.5) Increased dreaming/nightmare: Group 1: 10 (17.5), Group 2: 7 (11.9), Group: 0, Group 4: 3 (5.1) Irritability: Group 1: 9 (15.8), Group 2: 11 (18.6), Group 3: 17 (14.9), Group 4: 6 (10.2) Nausea: Group 1: 5 (8.8), Group 2: 5 (8.5), Group 3: 0, Group 4: 1 (1.7) Sudden worsening of mood: Group 1: 4 (7.0), Group 2: 7 (11.9), Group 3: 5 (10.6), Group 4: 5 (8.5) Sweating more than usual: Group 1: 19 (33.3), Group 2: 21 (35.6), Group 3: 4 (8.5), Group 4: 12 (20.3) Trouble sleeping, insomnia: Group 1: 13 (22.8), Group 2: 12 (20.3), Group 3: 5 (10.6), Group 4: 7 (11.9)

242 Protocol outcome 3: Harm – adverse events ≥ 5% in any group Actual outcome: Number (%) of women reporting adverse events during the taper phase Any adverse events: Group 1: 47/94, Group 2: 53/101, Group 3: 43/87, Group 4: 53/102 Asthenia: Group 1: 0/94, Group 2: 8/101, Group 3: 2/87, Group 4: 5/102 Diarrhoea: Group 1: 3/94, Group 2: 5/101, Group 3: 5/87, Group 4: 6/102 Dizziness: Group 1: 8/94, Group 2: 15/101, Group 3: 10/87, Group 4: 12/102 Emotional lability: Group 1: 1/94, Group 2: 4/101, Group 3: 4/87, Group 4: 6/102 Headache: Group 1: 7/94, Group 2: 9/101, Group 3: 7/87, Group 4: 12/102 Hypertension: Group 1: 0/94, Group 2: 2/101, Group 3: 2/87, Group 4: 6/102 Infection: Group 1: 2/94, Group 2: 2/101, Group 3: 3/87, Group 4: 6/102 Insomnia: Group 1: 5/94, Group 2: 5/101, Group 3: 5/87, Group 4: 6/102 Nausea: Group 1: 9/94, Group 2: 10/101, Group 3: 7/87, Group 4: 9/102 Sweating: Group 1: 2/94, Group 2: 3/101, Group 3: 4/87, Group 4: 7/102 Vasodilation: Group 1: 5/94, Group 2: 7/101, Group 3: 1/87, Group 4: 6/102

Risk of bias: All domain –High, Selection - High, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low;

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Study Gallagher 2012(26) Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 4; Group 2 Number missing: 4, Group 3 Number missing; 7, Group 4 Number

missing; 7. Reason*: Withdrew (22).

Protocol outcome 4: Harm – incidence of taper emergent adverse events (post hoc analysis) Actual outcome: % Adverse events that developed during the taper phase or were pre-existing adverse events that became worse during the taper period week 1. Group 1: 31.9%, Group 2: 35.6%, Group 3: 35.6%, Group 4: 48%

Actual outcome: % Adverse events that developed during the taper phase or were pre-existing adverse events that became worse during the taper period week 2. Group 1: 30.4%, Group 2: 30.9%, Group 3: 19.8%, Group 4: 15.6%

Actual outcome: % Adverse events that developed during the follow-up week or were pre-existing adverse events that became worse during the follow up week 1 (week 3 after discontinuation of treatment). Ranged from 11.1-15.4%; actual figures not reported.

243 Actual outcome: % Adverse events that developed during the follow-up week or were pre-existing adverse events that became worse during the follow up week 2

(week 4 after discontinuation of treatment). Group 1: 9.9%, Group 2: 14.1%, Group 3: 4.8%, Group 4: 7.4%

Risk of bias: All domain –Very high, Selection - High, Blinding - Low, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – High, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 4; Group 2 Number missing: 4, Group 3 Number missing; 7, Group 4 Number missing; 7. Reason*: Withdrew (22).

Study Khan 2014(11) Study type RCT, discontinuation study Number of studies (number of participants) 1 (n=361) Countries and setting Conducted in USA; 38 research centres across USA, outpatients. Duration of study 28-weeks intervention + follow-up at 1 and 2 weeks following last treatment day Inclusion criteria 18 years of age or older; primary diagnosis of single or recurrent major depressive disease without psychotic features

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Study Khan 2014(11) based on criteria from 4th edition of Diagnostic and Statistical Manual of Mental Disorders using modified Mini

International Neuropsychiatric Interview; presence of depressive symptoms for at least 30 days prior to screening visit; 17-item Hamilton Rating Scale for Depression total score ≥14 at baseline.

Exclusion criteria Current diagnosis of an anxiety disorder that was considered to be primary; at significant risk of suicide based on responses to questions 4 or 5 of the Columbia Suicide Severity Rating Scale at screening or baseline; current substance abuse or dependence; unstable hepatic, renal, pulmonary, cardiovascular (including uncontrolled hypertension, unstable angina or recent myocardial infarction); ophthalmologic or neurologic disorder; other clinically important medical disease (including uncontrolled diabetes). Recruitment/selection of patients Recruitment between March 2010 and February 2011. Outpatients. Age, gender and ethnicity Age – mean (SD): Taper, 46.7 (11.2) years; Abrupt discontinuation, 47.8 (13.7) years; No discontinuation, 46.7 (11.3)

years. Gender (M:F): Taper, 37/103; Abrupt discontinuation, 48/100; No discontinuation, 24/49. Ethnicity: White, 80.7% for Taper, 81.8% for Abrupt discontinuation and 87.7% for No discontinuation; African-American, 17.1 % for Taper, 16.2% for Abrupt discontinuation and 9.6% for No discontinuation; American Indian/Alaskan Native, 0% for Taper, 0.7% for Abrupt discontinuation and 0% for No discontinuation; Other, 2.1% for Taper, 1.4% for Abrupt discontinuation and 2.7% for No discontinuation. 244 Extra comments Only those with at least 1 post-randomisation DESS score assessment were included in the analysis performed by the

authors. Indirectness of population No indirectness. Interventions & comparators All participants received 24-week open-label treatment with desvenlafaxine (50 mg/day). Participants completing this 24-week period were randomly assigned to one of three discontinuation regimens for 4 weeks. All participants followed up at 1 and 2 weeks after the final dose of each discontinuation regimen. (n=140) Intervention 1: Managed withdrawal. Taper. Received desvenlafaxine treatment at a dose of 25 mg/day for 1 week, followed by placebo for 3 weeks. (n=148) Intervention 2: Control. Abrupt discontinuation. Placebo received for 4 weeks. (n=73) Intervention 3: Usual care. No discontinuation. Continued desvenlafaxine treatment for 4 weeks at the dose of 50 mg/day. Funding Funding provided by Pfizer. RESULTS AND RISK OF BIAS FOR COMPARISON: Managed versus abrupt tapering

Protocol outcome 1: Intensity of signs and symptoms and overall withdrawal syndrome experienced -Actual outcome: Discontinuation-Emergent Signs and Symptoms (DESS) scale total score during first 2 weeks of discontinuation. Study doesn’t describe this outcome and scale very well – other publications mention a scale of 0-43, with higher values indicating a higher number of new or worsened symptoms in the past 7 days, as

Evidence tables Evidence Harms © Royal © College of Physicians 2019 Study Khan 2014(11) reported by participants. Adjusted values given but unclear what has been adjusted for, potentially baseline values. No further details provided.

Adjusted mean difference (95% CI) for total DESS scale score – Taper vs. Abrupt discontinuation group: -0.50 (-1.89 to 0.88); Taper vs. No discontinuation: 0.66 (-1.03 to 2.35); Abrupt discontinuation vs. No discontinuation: 1.16 (-0.51 to 2.83). Baseline values for each group were 0.6 (1.2, n=139) for Taper, 1.1 (2.41, n=146) for Abrupt

discontinuation and 1.0 (2.01, n=72) for No discontinuation. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA -Group 2 vs group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA

Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

245 Protocol outcome 2: Intensity and duration of symptoms and overall withdrawal syndrome experienced

-Actual outcome: Presence of discontinuation syndrome based on DESS scores. Defined as an increase of ≥4 on the DESS scale during the intervention baseline (when randomised interventions began). Taper, 30/139; Abrupt discontinuation, 31/146; No discontinuation, 8/72. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA -Group 2 vs group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Protocol outcome 3: Intensity and duration of signs and symptoms and overall withdrawal syndrome experienced -Actual outcome: Taper/post-therapy-emergent adverse events. Defined as adverse events that that started or increased in severity during the randomised intervention phase. Taper, 54/139; Abrupt discontinuation, 75/146; No discontinuation, 26/72. Risk of bias:

-Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High,

Crossover – NA; -Group 2 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Protocol outcome 4: Intensity and duration of signs and symptoms and overall withdrawal syndrome experienced -Actual outcome: Headache. Study gave this outcome separately as at least 5% of participants reported this adverse event. Taper, 9/139; Abrupt discontinuation,

19/146; No discontinuation, 3/72. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; 246 -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High,

Crossover – NA; -Group 2 vs group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Protocol outcome 5: Intensity and duration of signs and symptoms and overall withdrawal syndrome experienced -Actual outcome: Nausea. Study gave this outcome separately as at least 5% of participants reported this adverse event. Taper, 6/139; Abrupt discontinuation, 9/146; No discontinuation, 5/72. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; -Group 2 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA;

Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Protocol outcome 6: Intensity and duration of signs and symptoms and overall withdrawal syndrome experienced -Actual outcome: Dizziness. Study gave this outcome separately as at least 5% of participants reported this adverse event. Taper, 8/139; Abrupt discontinuation, 14/146; No discontinuation, 1/72. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA;

-Group 2 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear. 247

Protocol outcome 7: Side effects and non-fatal overdose -Actual outcome: Suicide attempts. Taper, 0/139; Abrupt discontinuation, 1/146; No discontinuation, 0/72. Only one event in the Abrupt discontinuation arm. Patient attempted suicide using non-study drug overdose 2 days following completion of the Abrupt discontinuation protocol. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; -Group 2 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – Low, Measurement – High, Crossover – NA; Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Protocol outcome 8: Intensity and duration of signs and symptoms and overall withdrawal syndrome experienced -Actual outcome: Suicidal ideation based on Columbia Suicide Severity Rating Scale (C-SSRS) during randomised intervention phase. Taper, 1/139; Abrupt discontinuation, 1/146; No discontinuation, 0/72.

Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – High, Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – High, Crossover – NA; -Group 2 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – High, Crossover – NA; Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Protocol outcome 9: Mortality

-Actual outcome: Mortality during open-label and randomised phases. Taper, 0/139; Abrupt discontinuation, 0/146; No discontinuation, 0/72. No mortality reported for either phase. Risk of bias: -Group 1 vs. group 2: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low,

248 Crossover – NA; -Group 1 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA; -Group 2 vs. group 3: All domain – Very high, Selection – High, Blinding – Low, Incomplete outcome data – Low, Outcome reporting – High, Measurement – Low, Crossover – NA; Indirectness of outcome: No indirectness. Group 1 Number missing: 1, Reason missing: unclear; Group 2 Number missing: 2, Reason missing: unclear; Group 3 Number missing: 1, Reason missing: unclear.

Note: Analysis completed on managed withdrawal compared to abrupt withdrawal; while arm that continued on treatment was not analysed.

Study Tint 2008(16) Study type RCT (Patient randomised; Parallel) Number of studies (number of participants) 1 (n=28 randomised). Countries and setting Not reported. Authors from UK.

Inclusion criteria Patients with a clinical diagnosis of major depressive disorder treated with selective serotonin reuptake inhibitors (SSRI) or venlafaxine for 6 weeks or more and in whom the treating clinician wanted to switch antidepressant were recruited. Exclusion criteria Not reported.

Recruitment/selection of patients Inpatient and outpatient in whom the treating clinician wanted to switch antidepressant. Age, gender and ethnicity N=28 (26 outpatients and 2 inpatients) 17 female and 11 male. Mean (SD) age: 39 years (12) Mean duration of current anti-depressant treatments 62.4 (11.5) weeks

No statistical differences between groups at baseline.

Extra comments None. Indirectness of population No indirectness. Interventions & comparators (n=15) Group 1: three-day taper of existing antidepressant with the taper individualized according to antidepressant,

249 dose and table formulation. (n=13) Group 2: 14-day taper of existing antidepressant with the taper individualized according to antidepressant, dose

and table formulation.

T2 - Assessed 5-7 days after stopping medication and then started a new antidepressant of the clinician’s choice. T3 – 7 days after start of new antidepressant. Funding No external funding was involved in this study. Authors received reimbursement for lecturing and consultancy, research grants and support to attend educational meetings from pharmaceutical companies that manufacture antidepressants. RESULTS (NUMBER ANALYSED) AND RISK OF BIAS FOR COMPARISON: three day taper compared to 14 day taper of antidepressants.

Protocol outcome 1: Harm – Discontinuation signs - Actual outcome: Presence of a discontinuation syndrome defined as ≥3 new DESS items at T2. DESS symptoms; scored as absent (0), mild (1), moderate (2) or severe (3) and summed to give a total score. Group 1:7/15 (47%), Group 2: 6/13 (46%)

Additional information from paper:

Evidence tables Evidence Harms © Royal © College of Physi The ANOVA showed a significant effect of time on DESS symptoms [F (2,52)=9.893, p<0.001] due to an increase after taper at T2, which had returned to near baseline at T3. No further details provided.

Most common new or worsened DESS items at T2 were dizziness (42%), headache (42%), nervousness/anxiety (42%), panic/sudden anxiety (32%), agitation (32%), nausea (32%) and sudden worsening of mood (32%).

Of the 13 above with discontinuation syndrome 5/8 were on paroxetine, 3/5 venlafaxine, 3/5 citalopram, 0/3 fluvoxamine and 2/7 on fluoxetine.

Risk of bias: All domain – Very high, Selection - High, Blinding - High, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low; Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 0; Group 2 Number missing: 0, Reason*: Not relevant. cians 2019

Protocol outcome 2: Harm – Depression - Actual outcome: Montgomery Asberg Depression Rating Scale (MADRS) – see below

Narrative findings: Figures not reported but authors comment ‘MADRS values were numerically slightly higher in the short taper group at baseline and this difference increased at T2 but no statistically significant difference between the groups overall (p=0.497). Authors stated that four patients all on paroxetine developed emergent suicidal ideation (defined as a MADRS suicide item score of two or less at baseline and increasing to four or more at T2. 250 Risk of bias: All domain – Very high, Selection - High, Blinding - High, Incomplete outcome data - Low, Outcome reporting - Low, Measurement – Low, Crossover - Low;

Indirectness of outcome: No indirectness; At end of study - Group 1 Number missing: 0; Group 2 Number missing: 0, Reason*: Not relevant.

Appendix F: GRADE tables

Harms of dependency on a medication

Table 27: Evidence profile: Tapentadol vs oxycodone

Quality assessment No of patients Effect

Quality Importance

No of Risk of Other Relative Design Inconsistency Indirectness Imprecision Tapentadol Oxycodone Absolute studies bias considerations (95% CI)

Developed shopping behaviour (follow-up 1 years)

251 1 observational very no serious serious2 no serious none 88/42940 0.9% RR 0.24 (0.19 to 7 fewer per 1000  CRITICAL

studies serious1 inconsistency imprecision (0.2%) 0.3) (from 6 fewer to 7 VERY fewer) LOW Adjusted OR3: 3.5 (2.84 to 4.4)

Number of shopping episodes per subject (follow-up 1 years; Better indicated by lower values)

1 observational very no serious serious2 no serious none 42940 112821 - MD 0.02 lower (0.02  CRITICAL studies serious1 inconsistency imprecision to 0.01 lower) VERY LOW

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 The majority of the evidence included an indirect population (downgrade by one increment) 3 Adjusted odds ratio – controlling for gender, benzodiazepine use and type of payment at first opioid exposure using a conditional logistic regression.

Harms / side effects from stopping these medications over a short time frame

Table 28: Evidence profile: Opioids versus control

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Relative Design Inconsistency Indirectness Imprecision Opioids Control Absolute studies bias considerations (95% CI)

Opioid vs control - no opioid withdrawal - COWS assessment 2-4 days after last intake of medication

1 252 1 randomised serious no serious no serious no serious none 61/72 100% RR 0.87 (0.77 130 fewer per 1000 (from  CRITICAL trials inconsistency indirectness imprecision (84.7%) to 0.99) 10 fewer to 230 fewer) MODERATE

Opioid vs control - no opioid withdrawal - COWS assessment at 4 days after last intake of medication

1 randomised serious1 no serious no serious no serious none 141/152 89.8% RR 1.04 (0.94 36 more per 1000 (from  CRITICAL trials inconsistency indirectness imprecision (92.8%) to 1.14) 54 fewer to 126 more) MODERATE

Opioid vs control - no opioid withdrawal - COWS assessment 5+ days after last intake of medication

1 randomised serious1 no serious no serious no serious none 141/154 91.6% RR 1 (0.92 to 0 fewer per 1000 (from 73  trials inconsistency indirectness imprecision (91.6%) 1.1) fewer to 92 more) MODERATE

Opioid vs control - mild or moderate opioid withdrawal - COWS assessment 2-4 days after last intake of medication

1 randomised serious1 no serious no serious very serious2 none 11/72 0% RR 4.04 (0.55 150 more per 1000 (from  CRITICAL trials inconsistency indirectness (15.3%) to 29.59) 40 more to 270 more) VERY LOW

Opioid vs control - mild or moderate opioid withdrawal - COWS assessment at 4 days after last intake of medication

1 randomised serious1 no serious no serious very serious2 none 11/152 10.2% RR 0.69 (0.26 32 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (7.2%) to 1.82) 75 fewer to 82 more) VERY LOW

GRADE tables Harms

1 randomised serious1 very serious3 no serious very serious2 none 13/154 8.5% RR 0.63 (0.05 31 fewer per 1000 (from  CRITICAL trials indirectness (8.4%) to 8.48) 81 fewer to 636 more) VERY LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Downgraded by 1 or 2 increments because of heterogeneity, I2=75%, p=0.05.

Table 29: Evidence profile: Opioid versus opioid

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Relative Design Inconsistency Indirectness Imprecision Opioids Control Absolute studies bias considerations (95% CI)

Withdrawal syndrome

253 1 randomised very no serious no serious very serious2 none 13/894 0.9% RR 1.62 (0.37 6 more per 1000 (from 6  CRITICAL 1 trials serious inconsistency indirectness (1.5%) to 7.13) fewer to 55 more) VERY LOW

Tapentadol vs oxycodone - no opioid withdrawal - COWS assessment 2-4 days after last intake of medication

2 randomised serious1 no serious no serious no serious none 126/160 79.6% RR 1.01 (0.85 8 more per 1000 (from  trials inconsistency indirectness imprecision (78.8%) to 1.19) 119 fewer to 151 more) MODERATE

Tapentadol vs oxycodone - no opioid withdrawal - COWS assessment at 4 days after last intake of medication

1 randomised serious1 no serious no serious no serious none 59/62 91.1% RR 1.04 (0.96 36 more per 1000 (from  CRITICAL trials inconsistency indirectness imprecision (95.2%) to 1.14) 36 fewer to 128 more) MODERATE

Tapentadol vs oxycodone - no opioid withdrawal - COWS assessment 5+ days after last intake of medication

2 randomised serious1 no serious no serious no serious none 215/236 84.9% RR 1.1 (1.01 85 more per 1000 (from 8  CRITICAL trials inconsistency indirectness imprecision (91.1%) to 1.19) more to 161 more) MODERATE

Tapentadol vs oxycodone - mild or moderate opioid withdrawal - COWS assessment 2-4 days after last intake of medication

3 randomised serious1 no serious no serious serious2 none 86/466 27.3% RR 0.70 (0.48 82 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (18.5%) to 1.00) 142 fewer to 0 more) LOW

GRADE tables Harms

1 randomised serious1 no serious no serious very serious2 none 3/62 8.9% RR 0.54 (0.15 41 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (4.8%) to 1.97) 76 fewer to 86 more) VERY LOW

Tapentadol vs oxycodone - mild or moderate opioid withdrawal - COWS assessment 5+ days after last intake of medication

2 randomised very very serious4 no serious very serious2 none 21/236 15.1% RR 0.33 (0.04 101 fewer per 1000 (from  CRITICAL trials serious1 indirectness (8.9%) to 2.72) 145 fewer to 260 more) VERY LOW

Drug withdrawal syndrome

1 randomised very no serious no serious very serious2 none 9/894 0.5% RR 2.24 (0.29 95 fewer per 1000 (from  CRITICAL trials serious1 inconsistency indirectness (1%) to 17.63) 146 fewer to 465 more) VERY LOW

Table 30: Evidence profile: Z-drugs versus placebo

254

Quality assessment No of patients Effect Quality Importance

No of Other Z- Relative Design Risk of bias Inconsistency Indirectness Imprecision Control Absolute studies considerations drugs (95% CI)

Rebound insomnia – proportion of patients with a lower self-reported total sleep time - Run out phase - day 1 (follow-up 1 weeks)

1 randomised no serious no serious no serious no serious none 13/67 6.4% RR 3.06 (1.33 132 more per 1000 (from  CRITICAL trials risk of bias inconsistency indirectness imprecision (19.4%) to 7) 21 more to 384 more) HIGH

Rebound insomnia - proportion of patients with a lower self-reported total sleep time - Run out phase - day 2

1 randomised no serious no serious no serious very serious1 none 5/67 5.6% RR 1.34 (0.44 19 more per 1000 (from  CRITICAL trials risk of bias inconsistency indirectness (7.5%) to 4.07) 31 fewer to 172 more) LOW

Rebound insomnia - proportion of patients with a lower self-reported total sleep time - Run out phase - day 3

1 randomised no serious no serious no serious very serious1 none 5/67 4% RR 1.88 (0.56 35 more per 1000 (from  CRITICAL trials risk of bias inconsistency indirectness (7.5%) to 6.27) 18 fewer to 211 more) LOW

GRADE tables Harms

1 randomised no serious no serious no serious serious1 none 12/67 6.4% RR 2.82 (1.21 116 more per 1000 (from  CRITICAL trials risk of bias inconsistency indirectness (17.9%) to 6.56) 13 more to 356 more) MODERATE

Rebound insomnia – proportion of patients with a lower self-reported time to sleep onset - Run out phase - day 2

1 randomised no serious no serious no serious very serious1 none 4/67 3.2% RR 1.88 (0.49 28 more per 1000 (from  CRITICAL trials risk of bias inconsistency indirectness (6%) to 7.28) 16 fewer to 201 more) LOW

Rebound insomnia -– proportion of patients with a lower self-reported time to sleep onset - Run out phase - day 3

ns 2019 1 randomised no serious no serious no serious very serious1 none 5/67 4% RR 1.88 (0.56 35 more per 1000 (from  CRITICAL trials risk of bias inconsistency indirectness (7.5%) to 6.27) 18 fewer to 211 more) LOW

1 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs

Table 31: Evidence profile: Benzodiazepines versus gabapentinoids

255 Quality assessment No of patients Effect

Quality Importance

No of Risk of Other Benzodiazepines Relative Design Inconsistency Indirectness Imprecision Control Absolute studies bias considerations versus gabapentinoids (95% CI)

Mean change in Physician Withdrawal Checklist (PWC) - higher dose gab - 1 week after taper (Better indicated by lower values)

1 randomised very no serious no serious no serious none 58 49 - MD 0.4 lower (3.09  CRITICAL trials serious1 inconsistency indirectness imprecision lower to 2.29 higher) LOW

Mean change in PWC - higher dose gab - 2 week after taper (Better indicated by lower values)

1 randomised very no serious no serious no serious none 54 44 - MD 0.5 higher (1.92  CRITICAL trials serious1 inconsistency indirectness imprecision lower to 2.92 higher) LOW

Mean change in PWC after treatment period 2 - higher dose gab - 2 week after taper (Better indicated by lower values)

1 randomised very no serious no serious no serious none 106 93 - MD 0.6 higher (1.08  CRITICAL trials serious1 inconsistency indirectness imprecision lower to 2.28 higher) LOW

Mean change in PWC after treatment period 2 - higher dose gab - 1 week after taper (Better indicated by lower values)

GRADE tables Harms

© Royal © College of Physicians 2019 1 randomised very no serious no serious no serious none 109 99 - MD 1.3 lower (2.92  CRITICAL trials serious1 inconsistency indirectness imprecision lower to 0.32 higher) LOW

Discontinuation emergent signs and syndromes (DESS) after period 1

1 randomised very no serious no serious very serious2 none 38/110 32.7% RR 1.06 (0.66 20 more per 1000  CRITICAL trials serious1 inconsistency indirectness (34.5%) to 1.69) (from 111 fewer to VERY 226 more) LOW

DESS after period 2

1 randomised very no serious no serious very serious2 none 55/203 28% RR 0.97 (0.66 8 fewer per 1000  CRITICAL trials serious1 inconsistency indirectness (27.1%) to 1.42) (from 95 fewer to VERY 118 more) LOW

Anxiety after period 1

1 randomised very no serious no serious very serious2 none 3/110 3.9% RR 0.71 (0.12 11 fewer per 1000  CRITICAL trials serious1 inconsistency indirectness (2.7%) to 4.12) (from 34 fewer to VERY 122 more) LOW

256 Anxiety after period 2

1 randomised very no serious no serious very serious2 none 11/203 8% RR 0.68 (0.28 26 fewer per 1000  CRITICAL trials serious1 inconsistency indirectness (5.4%) to 1.63) (from 58 fewer to 50 VERY more) LOW

Dizziness after period 1

1 randomised very no serious no serious very serious2 none 3/110 0% Peto OR 4.44 30 more (from 10  CRITICAL trials serious1 inconsistency indirectness (2.7%) (0.39 to fewer to per 1000 to VERY 50.92) 70 more) LOW

Headache after period 1

1 randomised very no serious no serious very serious2 none 7/110 1.9% RR 3.31 (0.42 44 more per 1000  CRITICAL trials serious1 inconsistency indirectness (6.4%) to 26.2) (from 11 fewer to VERY 479 more) LOW

Headache after period 2

1 randomised very no serious no serious very serious2 none 8/203 2% RR 1.97 (0.43 19 more per 1000  CRITICAL trials serious1 inconsistency indirectness (3.9%) to 9.11) (from 11 fewer to VERY

GRADE tables Harms

Insomnia after period 1

1 randomised very no serious no serious serious2 none 10/110 19.2% RR 0.47 (0.21 102 fewer per 1000  CRITICAL trials serious1 inconsistency indirectness (9.1%) to 1.06) (from 152 fewer to VERY 12 more) LOW

Insomnia after period 2

1 randomised very no serious no serious serious2 none 21/203 6% RR 1.72 (0.72 43 more per 1000  CRITICAL trials serious1 inconsistency indirectness (10.3%) to 4.14) (from 17 fewer to VERY 188 more) LOW

Nausea after period 1

1 randomised very no serious no serious very serious2 none 7/110 3.9% RR 1.65 (0.36 25 more per 1000  CRITICAL trials serious1 inconsistency indirectness (6.4%) to 7.69) (from 25 fewer to VERY 261 more) LOW

Rebound anxiety after treatment period 1

257 1 randomised very no serious no serious very serious2 none 4/110 4.2% RR 0.87 (0.17 5 fewer per 1000  CRITICAL trials serious1 inconsistency indirectness (3.6%) to 4.6) (from 35 fewer to VERY 151 more) LOW

Rebound anxiety after treatment period 2

1 randomised very no serious no serious serious2 none 4/203 6% RR 0.33 (0.09 40 fewer per 1000  CRITICAL trials serious1 inconsistency indirectness (2%) to 1.14) (from 55 fewer to 8 VERY more) LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 32: Evidence profile: Antidepressants versus placebo

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Relative Design Inconsistency Indirectness Imprecision Antidepressants Control Absolute studies bias considerations (95% CI)

GRADE tables Harms

1 randomised serious1 no serious no serious very serious2 none 4/148 1.4% RR 1.97 (0.22 14 more per 1000  CRITICAL trials inconsistency indirectness (2.7%) to 17.34) (from 11 fewer to VERY LOW

229 more)

Benzodiazepine Withdrawal symptom questionnaire criteria BWSQ

1 randomised serious1 no serious no serious very serious2 none 1/148 1.4% RR 0.49 (0.03 7 fewer per 1000  CRITICAL trials inconsistency indirectness (0.68%) to 7.77) (from 14 fewer to 95 VERY LOW more)

Suicide attempts (one study after discontinuation; other study time point not reported)

2 randomised no serious no serious no serious very serious2 none 2/605 0.3% Peto OR 1.02 0 fewer per 1000  CRITICAL trials risk of bias inconsistency indirectness (0.33%) (0.09 to 12.12) (from 10 fewer to 10 LOW more)6

Depression (after discontinuation; except one study time point not reported)

3 randomised serious1 serious3 no serious very serious2 none 16/1044 0.6% RR 1.16 (0.15 1 more per 1000 CRITICAL

258  trials indirectness (1.5%) to 8.76) (from 5 fewer to 47 VERY LOW

more)

Suicide ideation (after discontinuation in two studies, one during study and other time point not reported)

4 randomised serious1 no serious no serious serious2 none 7/1498 0% Peto OR 4.24 0 more per 10000  CRITICAL trials inconsistency indirectness (0.47%) (0.82 to 21.90) (from 0 fewer to 10 LOW more)6

DESS - taper week 1 (Better indicated by lower values)

3 randomised serious1 serious4 no serious serious2 none 793 425 - MD 0.56 higher  CRITICAL trials indirectness (0.01 lower to 1.13 VERY LOW higher)

DESS - taper week 2 (Better indicated by lower values)

2 randomised serious1 no serious no serious no serious none 694 359 - MD 0.48 higher  CRITICAL trials inconsistency indirectness imprecision (0.18 to 0.77 higher) MODERATE

DESS - taper week 3 (Better indicated by lower values)

GRADE tables Harms

© Royal © College of Physicians 2019 1 randomised very no serious no serious no serious none 32 8 - MD 6 lower (9.56 to  CRITICAL trials serious1 inconsistency indirectness imprecision 2.45 lower) LOW

Vertigo (after discontinuation)

1 randomised very no serious no serious no serious none 12/324 0% Peto OR 4.63 40 more per 1000  CRITICAL trials serious1 inconsistency indirectness imprecision (3.7%) (1.37 to 15.6) (from 10 more to 60 LOW more)6

Change after discontinuation

1 randomised no serious no serious no serious very serious2 none 14/53 20.8% RR 1.27 (0.64 56 more per 1000  CRITICAL trials risk of bias inconsistency indirectness (26.4%) to 2.54) (from 75 fewer to LOW 320 more)

Discontinuation syndrome (after discontinuation)

1 randomised no serious no serious no serious very serious2 none 9/53 9.4% RR 1.8 (0.65 75 more per 1000  CRITICAL trials risk of bias inconsistency indirectness (17%) to 5.02) (from 33 fewer to LOW 378 more)

259 Total taper/post study emergent AE

9 randomised very no serious no serious no serious none 600/1728 20.5% RR 1.63 (1.44 129 more per 1000  CRITICAL trials serious1 inconsistency indirectness imprecision (34.7%) to 1.84) (from 90 more to 172 LOW more)

Vomiting (after discontinuation)

1 randomised serious1 no serious no serious very serious2 none 1/149 1.9% RR 0.35 (0.04 12 fewer per 1000  CRITICAL trials inconsistency indirectness (0.67%) to 3.34) (from 18 fewer to 44 VERY LOW more)

Dizziness (after discontinuation)

7 randomised very no serious no serious no serious none 117/1457 2.5% RR 4.86 (2.91 97 more per 1000  CRITICAL trials serious1 inconsistency indirectness imprecision (8%) to 8.14) (from 48 more to 179 LOW more)

Nausea (after discontinuation)

6 randomised very serious5 no serious no serious none 103/1454 2.5% RR 2.78 (1.36 45 more per 1000  CRITICAL trials serious1 indirectness imprecision (7.1%) to 5.69) (from 9 more to 117 VERY LOW more)

GRADE tables Harms

Royal College of Physicians 2019 Headache (after discontinuation)

5 randomised very no serious no serious serious2 none 83/1390 4.4% RR 1.39 (0.96 17 more per 1000  CRITICAL trials serious1 inconsistency indirectness (6%) to 2) (from 2 fewer to 44 VERY LOW

more)

Insomnia (after discontinuation)

3 randomised very no serious no serious very serious2 none 30/663 1.9% RR 1.37 (0.75 7 more per 1000  CRITICAL trials serious1 inconsistency indirectness (4.5%) to 2.52) (from 5 fewer to 29 VERY LOW more)

Diarrhoea (after discontinuation)

1 randomised serious1 no serious no serious very serious2 none 4/149 2.6% RR 1.05 (0.27 1 more per 1000  CRITICAL trials inconsistency indirectness (2.7%) to 4.14) (from 19 fewer to 82 VERY LOW more)

Serious adverse events during taper

2 randomised serious1 no serious no serious very serious2 none 1/479 0.53% RR 0.40 (0.04 0 fewer per 1000 CRITICAL

260  trials inconsistency indirectness (0.21%) to 4.03) (from 10 fewer to 10 VERY LOW 6

more)

Withdrawal syndrome (after discontinuation)

1 randomised no serious no serious no serious no serious none 117/455 11.9% RR 2.2 (1.4 to 143 more per 1000  CRITICAL trials risk of bias inconsistency indirectness imprecision (25.7%) 3.46) (from 48 more to 293 HIGH more)

Upper respiratory tract infection (after discontinuation)

1 randomised no serious no serious no serious very serious2 none 17/455 1.3% RR 3.17 (0.75 28 more per 1000  CRITICAL trials risk of bias inconsistency indirectness (3.7%) to 13.44) (from 3 fewer to 162 LOW more) 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Downgraded by one increment because heterogeneity, I2=65%, p=0.06, unexplained by subgroup analysis. 4 Downgraded by one increment because heterogeneity, I2=72%, p=0.01, unexplained by subgroup analysis. 5 Downgraded by one increment because heterogeneity, I2=53%, p=0.05, unexplained by subgroup analysis. 6 Zero events in one or more arms so absolute effect calculated from risk difference. 7 One study reported suicide ideation with depression.

Table 33: Evidence profile: Antidepressants versus antidepressants

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Relative Design Inconsistency Indirectness Imprecision Antidepressant Antidepressant Absolute studies bias considerations (95% CI)

Suicide ideation - antidepressant v antidepressant (time point not reported)

1 randomised serious1 no serious no serious very none 2/301 1.3% RR 0.5 (0.07 6 fewer per 1000  CRITICAL trials inconsistency indirectness serious2 (0.66%) to 3.55) (from 12 fewer to 33 VERY more) LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

261 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Short versus long term opioid use

Table 34: Evidence profile: Short versus long term opioid use

Quality assessment No of patients Effect

Quality Importance Long term opioid use No of Risk of Other Relative Design Inconsistency Indirectness Imprecision compared to short term Control Absolute studies bias considerations (95% CI) opioid use

Depression

1 observational very no serious very serious2 no serious none - 0% HR 1.53 (1.29 -  CRITICAL studies serious1 inconsistency imprecision to 1.81) VERY LOW

GRADE tables Harms

1 observational very no serious very serious2 serious3 none - 0% HR 1.38 (0.9 -  CRITICAL studies serious1 inconsistency to 2.12) VERY

LOW

Opioid abuse

1 observational very no serious very serious2 serious3 none - 0% HR 3.97 (0.87 -  CRITICAL studies serious1 inconsistency to 18.12) VERY LOW

Other substance abuse

1 observational very no serious very serious2 serious3 none - 0% HR 1.81 (0.92 -  CRITICAL studies serious1 inconsistency to 3.56) VERY LOW

Opioid overdose

262 1 observational very no serious very serious no serious none - 0% HR 5.12 (1.63 -  CRITICAL studies serious1 inconsistency imprecision to 16.08) VERY

LOW

Other substance overdose

1 observational very no serious very serious2 serious3 none - 0% HR 1.82 (0.92 -  CRITICAL studies serious1 inconsistency to 3.6) VERY LOW

Opioid dependence

1 observational very no serious very serious2 no serious none - 0% HR 2.85 (1.54 -  CRITICAL studies serious1 inconsistency imprecision to 5.27) VERY LOW

Other substance dependence

1 observational very no serious very serious2 serious3 none - 0% HR 1.73 (1.21 -  CRITICAL studies serious1 inconsistency to 2.47) VERY LOW

Mortality

GRADE tables Harms

© Royal © College of Physicians 2019 1 observational very no serious very serious2 no serious none - 0% HR 0.99 (0.84 -  CRITICAL studies serious1 inconsistency imprecision to 1.17) VERY LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 or 2 increments because long term use used as proxy for dependence. Short term use population includes children. Some opioids received are not available on the NHS. 3 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Short versus long term withdrawal

Table 35: Evidence profile: 3 day taper vs 14 day taper of antidepressants

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Short taper Longer taper Relative Design Inconsistency Indirectness Imprecision Absolute studies bias considerations – 3 day – 14 day (95% CI)

263 DESS symptoms

1 randomised very no serious no serious very none 7/15 46.2% RR 1.01 (0.46 5 more per 1000 (from  CRITICAL trials serious1 inconsistency indirectness serious2 (46.7%) to 2.25) 249 fewer to 577 more) VERY LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 36: Evidence profile: abrupt vs taper of antidepressant withdrawal (managed taper of 25mg/d for one week)

No of Quality assessment Effect patients Quality Importance

No of Risk of Other Relative Design Inconsistency Indirectness Imprecision Abrupt Taper Absolute studies bias considerations (95% CI)

DESS score

1 randomised very no serious no serious very none 31/146 21.6% RR 0.98 (0.63 to 4 fewer per 1000 (from 80  CRITICAL

GRADE tables Harms

Taper/post-therapy-emergent adverse events

1 randomised very no serious no serious serious2 none 75/146 38.9% RR 1.32 (1.02 to 124 more per 1000 (from 8  CRITICAL trials serious1 inconsistency indirectness (51.4%) 1.72) more to 280 more) VERY LOW

Headache

1 randomised very no serious no serious serious2 none 19/146 6.5% RR 2.01 (0.94 to 66 more per 1000 (from 4  CRITICAL trials serious1 inconsistency indirectness (13%) 4.29) fewer to 214 more) VERY LOW

Nausea

1 randomised very no serious no serious very none 9/146 4.3% RR 1.43 (0.52 to 18 more per 1000 (from 21  CRITICAL trials serious1 inconsistency indirectness serious2 (6.2%) 3.91) fewer to 125 more) VERY LOW

264 Dizziness

1 randomised very no serious no serious very none 14/146 5.8% RR 1.67 (0.72 to 39 more per 1000 (from 16  CRITICAL trials serious1 inconsistency indirectness serious2 (9.6%) 3.85) fewer to 165 more) VERY LOW

Suicide ideation

1 randomised very no serious no serious very none 1/146 0.7% RR 0.95 (0.06 to 0 fewer per 1000 (from 7  CRITICAL trials serious1 inconsistency indirectness serious2 (0.68%) 15.07) fewer to 98 more) VERY LOW

Suicide attempts

1 randomised very no serious no serious very none 1/146 0% Peto OR 7.04 (0.14 10 fewer per 1000 (from 10  CRITICAL trials serious1 inconsistency indirectness serious2 (0.68%) to 355.37) fewer to 30 more) VERY LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

al College of Physicians 2019

Table 37: Evidence profile: rapid (1-7 days) versus gradual withdrawal of antidepressants (2 weeks or more)

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Rapid Gradual Relative Design Inconsistency Indirectness Imprecision Absolute studies bias considerations withdrawal withdrawal (95% CI)

Time to first new illness

1 observational very no serious serious3 very none - 0% HR 1.5 (1.14 to -  CRITICAL studies serious1 inconsistency serious2 1.97) VERY LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

265 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Downgraded by 1 increment because the majority of the evidence included drugs grouped together so they were not all listed and could have included drugs not listed on the included list.

Table 38: Evidence profile: Abrupt withdrawal vs tapered withdrawal of antidepressants (50 mg alternate days for two weeks)

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Taper Relative Design Inconsistency Indirectness Imprecision Abrupt Absolute studies bias considerations (alternate) (95% CI)

DESS after taper week 3 (Better indicated by lower values)

1 randomised serious1 no serious no serious serious2 none 59 59 - MD 1.44 higher (0.04  CRITICAL trials inconsistency indirectness lower to 2.92 higher) LOW

Any adverse events

1 randomised serious1 no serious no serious serious2 none 53/101 52% RR 1.01 (0.78 5 more per 1000 (from  CRITICAL trials inconsistency indirectness (52.5%) to 1.31) 114 fewer to 161 more) LOW

GRADE tables Harms

1 randomised serious1 no serious no serious very none 8/101 4.9% RR 1.62 (0.55 30 more per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (7.9%) to 4.77) 22 fewer to 185 more) VERY

LOW

Diarrhoea

1 randomised serious1 no serious no serious very none 5/101 5.9% RR 0.84 (0.27 9 fewer per 1000 (from 43  CRITICAL trials inconsistency indirectness serious2 (5%) to 2.67) fewer to 99 more) VERY LOW

Dizziness

1 randomised serious1 no serious no serious very none 15/101 11.8% RR 1.26 (0.62 31 more per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (14.9%) to 2.56) 45 fewer to 184 more) VERY LOW

Emotional lability

266 1 randomised serious no serious no serious very none 4/101 5.9% RR 0.67 (0.2 19 fewer per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (4%) to 2.31) 47 fewer to 77 more) VERY

LOW

Headache

1 randomised serious1 no serious no serious very none 9/101 11.8% RR 0.76 (0.33 28 fewer per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (8.9%) to 1.72) 79 fewer to 85 more) VERY LOW

Hypertension

1 randomised serious1 no serious no serious very none 2/101 5.9% RR 0.34 (0.07 39 fewer per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (2%) to 1.63) 55 fewer to 37 more) VERY LOW

Infection

Insomnia

GRADE tables Harms

© Royal © College of Physicians 2019 1 randomised serious1 no serious no serious very none 5/101 5.9% RR 0.84 (0.27 9 fewer per 1000 (from 43  CRITICAL trials inconsistency indirectness serious2 (5%) to 2.67) fewer to 99 more) VERY LOW

Nausea

1 randomised serious1 no serious no serious very none 10/101 8.8% RR 1.12 (0.48 11 more per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (9.9%) to 2.65) 46 fewer to 145 more) VERY LOW

Sweating

1 randomised serious1 no serious no serious very none 3/101 6.9% RR 0.43 (0.12 39 fewer per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (3%) to 1.63) 61 fewer to 43 more) VERY LOW

Vasodilation

1 randomised serious1 no serious no serious very none 7/101 5.9% RR 1.18 (0.41 11 more per 1000 (from  CRITICAL trials inconsistency indirectness serious2 (6.9%) to 3.38) 35 fewer to 140 more) VERY

267 LOW 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 39: Evidence profile: Abrupt withdrawal vs tapered withdrawal of antidepressants (50 mg for one week and then placebo for one week)

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Taper (50 Relative Design Inconsistency Indirectness Imprecision Abrupt Absolute studies bias considerations then placebo) (95% CI)

DESS after taper week 3 (Better indicated by lower values)

1 randomised serious1 no serious no serious no serious none 47 59 - MD 0.08 lower (1.3  CRITICAL trials inconsistency indirectness imprecision lower to 1.14 higher) MODERATE

Any adverse events

GRADE tables Harms

© Royal © College of Physicians 2019 1 randomised serious1 no serious no serious very serious2 none 43/87 52% RR 0.95 26 fewer per 1000  CRITICAL trials inconsistency indirectness (49.4%) (0.72 to 1.26) (from 146 fewer to 135 VERY LOW more)

Asthenia

1 randomised serious1 no serious no serious very serious2 none 2/87 4.9% RR 0.47 26 fewer per 1000  CRITICAL trials inconsistency indirectness (2.3%) (0.09 to 2.36) (from 45 fewer to 67 VERY LOW more)

Diarrhoea

1 randomised serious1 no serious no serious very serious2 none 5/87 5.9% RR 0.98 1 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (5.7%) (0.31 to 3.09) 41 fewer to 123 more) VERY LOW

Dizziness

1 randomised serious1 no serious no serious very serious2 none 10/87 11.8% RR 0.98 2 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (11.5%) (0.44 to 2.15) 66 fewer to 136 more) VERY LOW

268 Emotional lability

1 randomised serious1 no serious no serious very serious2 none 4/87 5.9% RR 0.78 13 fewer per 1000  CRITICAL trials inconsistency indirectness (4.6%) (0.23 to 2.68) (from 45 fewer to 99 VERY LOW more)

Headache

1 randomised serious1 no serious no serious very serious2 none 7/87 11.8% RR 0.68 38 fewer per 1000  CRITICAL trials inconsistency indirectness (8%) (0.28 to 1.66) (from 85 fewer to 78 VERY LOW more)

Hypertension

1 randomised serious1 no serious no serious very serious2 none 2/87 5.9% RR 0.39 36 fewer per 1000  CRITICAL trials inconsistency indirectness (2.3%) (0.08 to 1.89) (from 54 fewer to 53 VERY LOW more)

Infection

1 randomised serious1 no serious no serious very serious2 none 3/87 5.9% RR 0.59 24 fewer per 1000  CRITICAL trials inconsistency indirectness (3.4%) (0.15 to 2.28) (from 50 fewer to 76 VERY LOW more)

GRADE tables Harms

Nausea

1 randomised serious1 no serious no serious very serious2 none 7/87 8.8% RR 0.91 8 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (8%) (0.35 to 2.35) 57 fewer to 119 more) VERY LOW

Sweating

1 randomised serious1 no serious no serious very serious2 none 4/87 6.9% RR 0.67 (0.2 23 fewer per 1000  CRITICAL trials inconsistency indirectness (4.6%) to 2.21) (from 55 fewer to 83 VERY LOW more)

Vasodilation

1 randomised serious1 no serious no serious very serious2 none 1/87 5.9% RR 0.2 (0.02 47 fewer per 1000  CRITICAL trials inconsistency indirectness (1.1%) to 1.59) (from 58 fewer to 35 VERY LOW

269 more) 1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 40: Evidence profile: Abrupt withdrawal vs tapered withdrawal of antidepressant (50 mg for one week and then 25 mg for one week)

Quality assessment No of patients Effect Quality Importance

No of Risk of Other Taper (50 Relative Design Inconsistency Indirectness Imprecision Abrupt Absolute studies bias considerations then 25) (95% CI)

DESS after taper week 3 (Better indicated by lower values)

1 randomised serious1 no serious no serious serious2 none 57 59 - MD 2.33 higher (0.62 to  CRITICAL trials inconsistency indirectness 4.04 higher) LOW

Any adverse events

GRADE tables Harms

© Royal © 1 randomised serious1 no serious no serious very serious2 none 47/94 52% RR 0.96 (0.73 21 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (50%) to 1.27) 140 fewer to 140 more) VERY LOW

College of Physicians 2019 Asthenia

1 randomised serious1 no serious no serious serious2 none 0/94 4.9% Peto OR 0.14 42 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (0%) (0.02 to 0.83) 8 fewer to 48 fewer) 3 LOW

Diarrhoea

1 randomised serious1 no serious no serious very serious2 none 3/94 5.9% RR 0.54 (0.14 27 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (3.2%) to 2.11) 51 fewer to 65 more) VERY LOW

Dizziness

1 randomised serious1 no serious no serious very serious2 none 8/94 11.8% RR 0.72 (0.31 33 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (8.5%) to 1.69) 81 fewer to 81 more) VERY LOW

Emotional lability

270 1 randomised serious1 no serious no serious very serious2 none 1/94 5.9% RR 0.18 (0.02 48 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (1.1%) to 1.47) 58 fewer to 28 more) VERY LOW

Headache

1 randomised serious1 no serious no serious very serious2 none 7/94 11.8% RR 0.63 (0.26 44 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (7.4%) to 1.54) 87 fewer to 64 more) VERY LOW

Hypertension

1 randomised serious1 no serious no serious no serious none 0/94 5.9% Peto OR 0.14 60 fewer per 1000 (from  CRITICAL trials inconsistency indirectness imprecision (0%) (0.03 to 0.7) 110 fewer to 10 fewer) 3 MODERATE

Infection

1 randomised serious1 no serious no serious very serious2 none 2/94 5.9% RR 0.36 (0.07 38 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (2.1%) to 1.75) 55 fewer to 44 more) VERY LOW

Insomnia

1 randomised serious1 no serious no serious very serious2 none 5/94 5.9% RR 0.9 (0.29 to 6 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (5.3%) 2.87) 42 fewer to 110 more) VERY LOW

GRADE tables Harms

1 randomised serious1 no serious no serious very serious2 none 9/94 8.8% RR 1.09 (0.45 8 more per 1000 (from  CRITICAL trials inconsistency indirectness (9.6%) to 2.62) 48 fewer to 143 more) VERY LOW

Sweating

1 randomised serious1 no serious no serious very serious2 none 2/94 6.9% RR 0.31 (0.07 48 fewer per 1000 (from  CRITICAL trials inconsistency indirectness (2.1%) to 1.46) 64 fewer to 32 more) VERY LOW

Vasodilation

1 Downgraded by 1 increment if the majority of the evidence was at high risk of bias and downgraded by 2 increments if the majority of the evidence was at very high risk of bias. 2 Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs. 3 Zero events in one or more arms so absolute effect calculated from risk difference.

271

Harms

Excluded studies

Appendix G: Excluded studies

The excluded studies list only includes details of studies excluded from all 5 questions after consideration for each.

Where population/study design/intervention etc. are listed as ‘wrong’, this is in reference to the review protocol for this rapid evidence assessment (appendix A) – i.e. not the population (or other element) of interest for this review, for example a study of harms from heroin use, or opioids obtained by illicit means, would be listed as wrong population. Where a study is excluded based on study design alone applies to the review question it was deemed most appropriate for, as the types of study design differ per protocols.

‘No relevant outcomes’ also refers to outcomes specified in the review protocol for this rapid evidence assessment. Study Exclusion reason Abrahms Published prior to 2008. Wrong population. Illicit drug use – heroin. 1979(35) Adi 2007(36) Wrong population. Illicit opioid use and not for chronic pain. Not prescribed opioids and also already detoxified. Aguiluz, 2018 (37) Wrong study type. Narrative review. Ahmadi, 2002 Wrong population. Not prescription drug use – OUD. (38) Ahmadi, 2003(39) Published prior to 2008. Ahmadi, 2003(40) Published prior to 2008. Ahmadi, 2018 Wrong population. No mention of prescription drug use, includes people with Opiate (41) Use Disorder – therefore likely all illicit drug use. Ahmadi, 2018, Wrong population. No mention of the dependence being on prescription opioids or (42) for chronic pain. Ahmed 2008 (43) No relevant outcomes. Alford, 2013 (44) Wrong study design. Narrative review / case study. Ali, 2017 (45) Wrong population. Non-medical/illicit use of prescription opioids - without a prescription or for euphoria. Alter 2017 (46) Wrong population. Surgical patients (not chronic pain). Altman, 2010 (47) Study design. Literature review. Amanti, 2018 (48) Wrong study type / not relevant to review questions. Development of a guideline, not assessment of the guideline being used in practice. Amarasuriya, Systematic review with different protocol to this review. Risk of mortality with long- 2012 (49) term benzodiazepine use. Amato 2013(50) Wrong population. Illicit opioid use and not for chronic pain. Amato 2011(51) Wrong population. Illicit opioid use. Not prescribed opioids for chronic pain. Amato 2011(52) Wrong population. Illicit opioid use. Not prescribed opioids for chronic pain. Amiri, 2014 (53) Wrong population. No mention of prescription opioids. Population of study is unclear, only references to opioids are illicit. Anonymous, 2009 Wrong study type / not relevant to review question. WHO guidelines, not assessment (54) of their effectiveness. Anonymous, 2014 Not relevant to review questions. Review of clinical evidence and guidelines for (55) gabapentin, only mention of dependence is re. NICE guideline consensus recommendation.

Excluded studies

Study Exclusion reason Anonymous, 2015 Wrong outcomes. Policy report looking at safety and efficacy, not dependence, (56) discontinuation or withdrawal. Anonymous, 2016 Wrong population / study type. Rapid response report. No mention of prescription (57) opioids, assume illicit use. Ansseau, Published prior to 2008. 1993(58) Apelt, 2013 (59) Wrong population. Not prescription drug use – OUD. Argoff 2014(60) Quality of systematic review inadequate. No quality assessment, no details of protocol Asaad, 2011 (61) Wrong population. Illicit opioids users - majority heroin with small proportion tramadol. Ashrafioun, 2016 Wrong population. Illicit use. (62) Ashton, 1990 (63) Published prior to 2008. Baandrup Wrong population. Majority of participants taking BZDs not mentioned on our drugs 2017(64) list - only 24% using oxazepam which is on list. Outcomes not given separately for those drugs on the list. Baandrup Wrong population. Majority of participants (>50%) taking a medication not included in 2016(65) the scope for this review. Outcomes not reported separately for those drugs that are on the pre-specified list. Baandrup Wrong population. Majority of participants (>50%) taking a medication not included in 2016(66) the scope for this review. Outcomes not reported separately for those drugs that are on the pre-specified list. Baandrup, 2016 Not relevant to review questions. Focus is whether melatonin restores circadian (67) rhythm in this group. Baandrup Wrong population. Includes studies beyond the scope of this review and illicit use. 2018(68) Bachhuber, 2016 Wrong population. Looks at whether introduction of pain management programmes (69) is associated with BZD misuse emergency department admissions. Mixed illicit and prescription misuse, unable to separate for analysis. Back 2011(70) Wrong population. Many were not introduced to the drug as a prescription from a doctor. Many using recreationally. Back, 2011 (71) Wrong population. Mainly illicit drug use. Does include prescription methadone, but results for illicit methadone and prescription are reported pooled. Bagra, 2018, (72) Wrong population. Majority using illicit opioids (heroin). Baewert, 2012 Wrong population. Not prescription drug use. (73) Bagoien 2013(74) Wrong population. Majority illicit drugs/alcohol. Only included BZDs that were not used as prescribed by physician. Baker 2005 (75) Wrong population. Illicit drug use – amphetamines. Baldacchino, 2012 Wrong population. Illicit drug use – heroin or methadone. (76) Baldacchino, 2016 Systematic review with different protocol. (77) Baldacchino, 2016 Study design and population. Only reporting on design and feasibility. (78) Baldini, 2012 (79) Systematic review – quality assessment and methods inadequate. Baldwin, 2015 Wrong outcomes. Safety and efficacy not harms. Literature review of safety of (80) pregabalin. References checked re dependence / withdrawal.

Excluded studies

Study Exclusion reason Ballard, 2013 (81) Wrong study design. Correction due to omitted text from original article. Balodimos, 2015 Wrong population. Heroin users. (82) Banta-green 2018 Wrong population. Mostly heroin users. (83) Banta-Green, Wrong population. Outcomes not relevant to review question. Although some people 2010 (84) may be dependent, this can't be clearly determined from the paper. Quantitative analysis of more qualitative data, relates to side effects. Barry, 2013 (85) Wrong population. Unclear if any are dependent on prescription opioids. Barry, 2016 (86) No relevant outcomes. Bartholomaeus Not relevant to review questions. Evaluates tamper resistant properties of two 2012(87) extended release drugs - chemical testing on the drugs. Bartoli, 2015 (88) No relevant outcomes. Reports side effects from use, not harms from dependence. Baser, 2014 (89) Not relevant to review questions. Determining prevalence and costs of prescription opioid dependence in US veterans. Bashir, 1994(90) Published prior to 2008. Bassiony, 2018 Wrong population. Substance Use Disorder. Some taking tramadol, but appears to not (91) be prescribed. Bassiony, 2018, Wrong population. Includes people obtaining tramadol without prescription and using (92) for conditions other than chronic pain. Bautista, 2017 Wrong population / Not relevant to review question. Looks an analysis of reasons for (93) increase in dose, not dependence, discontinuation or withdrawal. Baxley 2018 (94) Wrong population. 97% reported heroin as opioid of choice. Beaudoin, 2016 Not relevant to review questions. Assessing ability of this instrument to identify those (95) misusing/abusing prescription opioids. Not individual factors that may lead to dependence/harms of dependence or methods of treatment. Beaudoin, 2016 Systematic review – Definition of misuse unclear – references screened and relevant (96) studies ordered. Becka 2004(97) Article not in English Becker 2005(98) Wrong intervention. Intervention is to assess provider satisfaction of treatment in offices. Unclear if correct population as for treatment of opioids dependency but cause and if still prescribed unknown Becker, 2008 (99) Wrong population. Illicit drug use - non medical use. Becker, 2013 Not relevant to review questions. Review of tools for patient reporting of safety, (100) efficacy or harms. Belaise, 2014 Wrong study type. Letter to editor reporting case studies. (101) Belknap 2008 Wrong population. Hospitalised surgical patients, no dependency/withdrawal (102) outcomes. Bell 2008 (103) Wrong population. Illicit drug use. Belleville 2008 Wrong study design - unadjusted data. (104) Beneitez, 2017 Wrong population. No mention of prescription drug use. (105) Bentzley Systematic review – inadequate methodology and quality assessment. 2015(106) Bergman Wrong population - not all using drugs of dependence. Focus is treating the adverse 2018(107) event of the antipsychotic drugs.

Excluded studies

Study Exclusion reason Berna, 2015 (108) Systematic review – different protocol. Inadequate risk of bias and quality assessment. Focus is to develop recommendations, no analysis of outcomes. Beswick, 2003 Wrong population. Illicit drug use. (109) Bet, 2013 (110) No relevant outcomes. Adverse events rather than withdrawal and dependence harms. Bickel, 2008 (111) Wrong population. Illicit drug use. Bigal, 2009 (112) Wrong study design – literature review. Biondi, 2010 Wrong study type / not relevant to review question. Abstract on safety and (113) tolerability, not harms. Bisaga, 2014 Wrong population. Includes a mix of heroin dependent and prescription opioid (114) dependent - but people excluded if they were still being prescribed opioids for pain or illness. Only 18% included were dependent on prescription opioids. Bishop, 2018 Wrong population. Users had been taking methadone and switched to buprenorphine (115) (illicit drug use, not mention of prescription use). Black 2012(116) Wrong study type. Abstract. Bleckwenn, 2016 Not relevant to review questions. Focus is misuse of substitution medications for illicit (117) drug use. Blom, 2007 (118) No relevant outcomes. Blondell 2008 Wrong population. Illicit drug use. (119) Blondell Wrong population. Majority of patients taking hydrocodone - not on pre-specified 2010(120) drug list. Blondi, 2010 Order cancelled - article not in English. (121) Bobes 2012 (122) Wrong study design (higher quality evidence available). Bockting 2018 Wrong intervention/study aim. Aim of the study is preventing relapse of depression (123) with no mention of dependence. Bogetto, 2002 Published prior to 2008. (124) Bohnert Wrong study design (case control study). 2011(125) Bohnert 2016 Wrong population. Non-medical use of prescription opioids. (126) Bonar, 2014 (127) Wrong population. Non-prescription opioid use. Bondi, 2018, Wrong population. Includes those using illicit opioids. No mention of prescription (128) opioids. Bourgeois, 2014 Wrong study design (higher quality evidence available). (4) Bowman, 2013 Wrong study design. Narrative review of strategies. (129) Boyuan, 2014 Wrong population. Heroin dependence. (130) Braden, 2009 Wrong population / not relevant to review questions. Looks at trends in long term (131) opioid use based on whether used for depression or not. Not people who are dependent. Brady, 2017 (132) Wrong population. Not on prescription opioids as looking at abuse of drugs. Branstetter 2008 Wrong population. Illicit drug use.

Excluded studies

Study Exclusion reason (133) Breen 2003(134) Wrong intervention / population. Methadone being used in this scenario to treat drug misuse disorder. Brigham, 2004 Order cancelled - conference abstract. (135) Broderick, 2009 Wrong study type. Editorial. (136) Brown, 2010 Wrong population. Mainly heroin use - unclear if people were non-prescription opioid (137) users who then occasionally use prescription opioids, or a mix initially. Brown, 2011 Not relevant to review questions. Looking at one cohort and selecting risk categories (138) and checking urine for drug use. Checking to see if investigators following guidelines for risk selection and changing risk according to urine results. Brown, 2014 Wrong population. Excludes people who need to take opioids for chronic pain. (139) Brown, 2017 No relevant outcomes. (140) Bruehl, 2015 Not relevant to review questions. Prevalence of benzodiazepine use. (141) Brummett, 2017 Wrong population. Opioid use post-surgery not chronic pain. (142) Bucher Bartelson, No relevant outcomes. 2017 (143) Bujold, 2012 Not relevant to review questions. Development and implementation of a guideline to (144) reduce drug diversion. Butler 2013(145) Not relevant to review questions. Focuses on non-medical use/abuse of prescription opioids - e.g. abuse by tampering. Butler, 2011 (146) Not relevant to review questions. Risk of abuse of different drug preparations, not risk of dependence etc. Butler, 2018 (147) Not relevant to review questions. Looks at prescription-adjusted rates of abuse, but not risk factors for abuse. Buttram, 2014 Wrong population - opioids not said to be prescribed for chronic pain. Very specific (148) population - MSM, but does report risk of prescription opioid abuse. Buynak, 2015 Wrong study design. Open label extension - non-comparative. Better quality evidence (149) available for this question. Not comparative – includes people who previously received tapentadol, oxycodone or placebo, but when all are given tapentadol. Results are reported according to prior treatment group. Cadth, 2012 (150) Not relevant to review question / no relevant outcomes. Cadth, 2013 (151) Not relevant to review question / wrong study type. List of references only. Cadth, 2014 (152) Not relevant to review question / wrong study type - reference list. Compares safety of two preparations of the same treatment. Cadth, 2014 (153) No relevant outcomes. Cadth, 2015 (154) Systematic review - inadequate quality assessment. Review - strategies for BZD discontinuation. Cadth, 2015 (155) Systematic review - inadequate quality assessment. Cadth, 2015 (156) Systematic review - inadequate quality assessment. Calcaterra, 2018, Wrong population. Includes those given opioids for cancer and acute pain. (157) Caldentey, 2008 Order cancelled - not in English. (158)

Excluded studies

Study Exclusion reason Campbell, 2017 Wrong population / not relevant to review population. Looks at risks of aberrant drug (159) related behaviour (diverting drugs to family members / receiving from family members) not risk of becoming dependent. Candiotti, 2010 Wrong study design. Literature review, not systematic. (160) Canham 2014 Non-UK study (excluded from patients experience question). (161) Canan, 2017 (162) Wrong population / not relevant to review question. Review of algorithms to detect non-medical opioid use. Cardinali Published prior to 2008. 2002(163) Cardinali 2016 Systematic review – inadequate methodology. (164) Carroll 2006(165) Wrong population. Alcohol or illicit drug use within last 28 days as inclusion criterion. Casati, 2012 (166) Not relevant to review questions. More focus on prevalence. Cassano Published prior to 2008. 1996(167) Cassidy 2014(168) Wrong population. Adults assessed for substance abuse problems, not clear whether abuse was a result of becoming dependent on prescription opioid. Cepeda, 2013 (34) Wrong study design - comparing 2 different opioids. Chabal 1997(169) Published prior to 2008. Chan 2011 (170) Wrong study design (narrative description). Chan, 2018 (171) No relevant outcomes. Chand, 2013 Wrong population. No indication that abuse began after dependence on prescription (172) opioids. Chaparro Not relevant to review questions. Not approaches to prevent or treat dependence, 2013(173) but the effectiveness of opioids for chronic pain. Chaparro Not relevant to review questions. Not approaches to prevent or treat dependence, 2014(174) but the effectiveness of opioids for chronic pain Chatterjee, 2018 Wrong population. Homelessness and opioid use disorder. (175) Chattopadhyay Wrong study design. Case report and literature review. 2016 (176) Chavoustie, 2017, Wrong study design. Literature review. (177) Cheatle, 2013 Wrong study design. (178) Chen 2013(179) Wrong population. Illicit drug use. Cheon, 2016 No relevant outcomes. Just reports on effectiveness. (180) Cheung, 2014 Not relevant to review questions. Systematic review of existing guidelines, not an (181) assessment of those guidelines being implemented. Chopra, 2009 Wrong population. Abuse of cocaine and opioids, likely to have been obtained illicitly (182) without prescription. Intravenous and intranasal routes of administration. Chou, 2009 (183) Systematic review and guideline – analysis not reported. Quality assessment reported insufficiently. References checked re risk factors. Chou, 2015 (184) Systematic review. Risk of bias and quality assessment inadequate. Christensen, 2014 Wrong population. Not prescription opioid use, may contain some prescription opioid

Excluded studies

Study Exclusion reason (185) users, but not reported separately, or a percentage of population. Cialdella Published prior to 2008. 2001(186) Cicero 2007(187) Wrong population. People using opioids to get high. Cicero 2008(188) Wrong population. Includes those using prescription opioids to get high - abuse rather than dependence. Also includes large proportion obtaining illicitly from sources other than doctors. Cicero 2012(189) Wrong population. Those abusing oxycodone (using it to get high through alternative administration routes, e.g. nasal). Cicero, 2017 (190) Wrong intervention. Not interested in illicit use or comparison with illicit use. Cicero, 2017 (191) Systematic review - different protocol. Screened for relevant references. No relevant papers identified - studies on prescription opioid misusers. Clancy 2013 (192) No relevant outcomes. Clark, 2008 (193) Wrong study design. Narrative review only. Clemans-Cope, Wrong population. Not prescription drug use. 2017 (194) Cochran, 2015 Wrong population. Only 21.7% opioid abusers (not stated whether prescription use or (195) other). Cock, 2018 (196) Wrong population. Codeine dependent misusers. Unclear if prescribed or over the counter. People using supratherapeutic doses. Colquhoun, 2013 Order cancelled - not available, inspection of abstract appears to be illicit use. (197) Colson, 2012 Wrong study type. Narrative review only. (198) Cooper, 2018 No relevant outcomes. No qualitative outcomes (patient experience question). (199) Conermann, 2014 Not relevant to review questions. Comparing methods of compliance monitoring. (200) Conrardy, 2016 Not relevant to review questions. Patient views on whether the drugs are addictive or (201) not, rather than their experiences of harms. Cornish, 2010 Wrong population. Not prescription drugs. (202) Correa, 2015 Not relevant to review questions. Sleep apnoea association, not harm from (203) dependence. Cossmann, 2016 No relevant outcomes. Describes numbers and demographics of patients with opioid (204) addiction but not patient views. Courtney von, Wrong population. Not prescription drug use. 2017 (205) Couvee 2002(206) Published prior to 2008. Couvee 2002(207) Published prior to 2008. Couvee 2002(208) Published prior to 2008. Coviello, 2010 Wrong population. Opioid dependent offenders. (209) Coyle, 2018 (210) No relevant outcomes. Cragg, 2017 (211) Wrong study type. Systematic review protocol. Crawford- Not relevant to review question / wrong study type. Summary of study reported Faucher, 2011 elsewhere, just looking at efficacy of escitalopram for hot flushes. (212)

Excluded studies

Study Exclusion reason Creac, 2011 (213) Wrong population. Medication overuse headache - only 1 participant used opioids, other drugs used not relevant to protocol. Crits-Christoph, Wrong population. Appears to be illicit use; opioid dependence, no mention of 2016 (214) prescription. Croissant, 2008 Wrong study design (higher quality evidence available). (215) Crowe 2017 (216) Systematic review with different protocol. Culberson, 2011 No relevant outcomes. (217) Cunningham, Wrong population. Some people were dependent on opioid analgesics, but unclear if 2011 (218) that was due to prescription. Majority illicit drugs. Da Costa Wrong intervention/comparison. Not approaches to prevent or treat dependence, but 2014(219) the effectiveness of opioids for chronic pain. D’Agostino, 2017 Wrong population. Assumed illicit as online comment from chatrooms; actual (220) comments not included just themes. Dakwar2015 Wrong interventions. Discontinuation of medication used to treat opioid dependency. (221) Daniell, 2008 No relevant outcomes. Potential harms of medication rather than dependence (222) outcomes. Daniulaityte 2012 Wrong population. Misusers of various drugs. Not necessarily prescribed. (223) Daniulaityte Published prior to 2008. 2006(224) Darker 2015(225) Wrong population. Harmful benzodiazepine use, abuse or dependence in opiate dependent and non-dependent groups. Darnall, 2012 Not relevant to review question. Review of risks and consequences. (226) Dart, 2012 (227) Wrong study design (higher quality evidence available). Davies 2018(228) Systematic review. Inadequate methodology – no risk of bias or quality assessment. Davis 2016(229) Wrong population. Various drugs of abuse not listed in this review protocol. Davis 2017(230) Wrong population. Illicit drug use. Davis, 2018 (231) Wrong population. 'Substance use disorder'. Day 2005(232) Wrong population. Not prescribed opioids for chronic pain (illicit use). De, 2008 (233) Wrong population. Heroin dependence only. De Fulio Wrong population. Use of heroin at least 21 of last 30 days as inclusion criterion. 2012(234) Degenhardt 2010 Wrong population. Heroin as well as opioid addicts (unclear if prescription opioids). (235) Degenhardt Wrong population. Illicit drug use. 2015(236) De Gier, 2011 No relevant outcomes. Looks at determinants of successful discontinuation, not risk (237) factors for the harms. Dehghani-Arani, Wrong population. Illicit drug use. 2013 (238) Delcher, 2017 Wrong study type. Dissertation only. (239) Del Rio 1997(240) Wrong population. Illicit drug use. De Maeyer, 2010 Systematic review- screened for relevant references. Most studies on heroin users

Excluded studies

Study Exclusion reason (241) and some were studies on opiate dependent individuals but did not meet our protocol criteria (participants who were dependent on illicit opioids drugs in receipt of substitute medication; non-UK studies) De Maricourt, Wrong study design (higher quality evidence available). 2016 (242) Demidenko, 2017 Wrong study design. (243) Dengler, 2016 Order cancelled - conference abstract (244) Dennis, 2014 Wrong population. Illicit drug use. (245) de Oliveira, Wrong intervention/comparison. Not approaches to prevent or treat dependence, but 2016(246) the effectiveness of opioids for chronic pain. Derry 2016(247) Wrong intervention/comparison. Not approaches to prevent or treat dependence, but the effectiveness of opioids for chronic pain. Dhawan, 2015 Wrong population. Illicit drug use (primarily heroin). (248) Di Costanzo Order cancelled. Not available in English. 1992(249) Diener, 2018 Wrong population. Review of medication overuse headache, including prevalence. (250) Relates to drugs outside of our scope, some opioid use, but for acute pain not chronic. Dijkstra 2008 Wrong population. Illicit drug use. (251) Dijkstra, 2010 Wrong population. Not prescription drug use. (252) Doolittle 2011 Wrong population. Illicit drug use. (253) D'Onofrio Wrong population. Non-medical use of prescription opioids - those requiring opioids 2015(254) for a pain condition were excluded. D'Onofrio, 2017 Wrong population. Participants were excluded if opioids were needed for pain. (255) Donohue, 2018 Wrong study design (higher quality evidence available). (256) Dou 2018 (257) Systematic review - inadequate methodology. Dowell, 2016 Wrong study type / not relevant to review question. Guideline on prescribing opioids, (258) no assessment of the effect of the guideline. Dunn 2010(259) Wrong population / outcomes. Overdoses not always the result of dependence - includes accidental ingestion and suicide attempts. Not all drugs used on pre-specified list. Dunn 2011(260) Systematic review – inadequate methodology. No risk of bias or quality assessment. Dunn 2015(261) Wrong population. Using prescription opioids illicitly - without a valid prescription. Dunn, 2017 (262) Order cancelled - erratum/corrections Dunn, 2017 (263) Wrong population. Primarily heroin users. Dreyer, 2015 Not relevant to UK context. (264) Dutra 2008(265) Wrong population - illicit drug use Dzialdowski, 1998 Wrong population. Illicit drug use. (266) Eccleston, Systematic review – inadequate methodology. No quality assessment undertaken.

Excluded studies

Study Exclusion reason 2017(267) Edlund, Published prior to 2008. 2007(268) Edlund, Wrong outcomes. Does not compare any relevant outcomes between different 2010(269) interventions/risk factors - gives incidence of chronic opioid use in certain subsets of patients but does not indicate dependence. Edlund, 2010 Not relevant to review questions. Looks at reasons for heavy opioid use, e.g. higher (270) doses, particularly related to insurance status (US based). Ehrich, 2015 (271) Not relevant to review questions. Looks at opioids as a treatment alternative to antidepressants. Doesn’t report about dependence etc. Eibl, 2017 (272) Wrong population. Not prescription drug use. Eilender Systematic review – Inadequate methodology. No quality assessment, no data or 2016(273) analysis that can be reused. El-Aneed 2009 Wrong population. Abusers/diversion of prescriptions. (274) Elefritz, 2016 Wrong population. ‘Critically ill’ including people with cancer pain. (275) El-Hadidy, 2015 Wrong population. Illicit drug use. (276) Els 2017(277) Wrong intervention/comparison. Not approaches to prevent or treat dependence, but the effectiveness of opioids for chronic pain. Elsesser Published prior to 2008. 1996(278) Enriquez-Puga, Wrong population. Aim of intervention was actually to increase prescribing of 2009 (279) particular antidepressants. Espostito 2009 No relevant outcomes. Prevalence of benzodiazepine use. (280) Evans, 2014 (281) Not related to review questions. Review of misuse and abuse of prescription antidepressants. Not reporting harms. Everitt 2018(282) Wrong intervention/comparison. Not approaches to prevent or treat dependence, but the effectiveness of antidepressants for insomnia. Everly 2011(283) Wrong population. Illicit opioids - use of heroin at least 21 of the last 30 days as an inclusion criterion. Facey 2016 (284) Not a review question. 16 cases of patients received naloxone but were deemed to not follow guidelines for use. Faggiano Wrong population - illicit opioid use 2003(285) Fareed, 2012 Wrong study design (higher quality evidence available). (286) Fatseas, 2018 Order cancelled - not available. Inspection of the abstract suggests not relevant to (287) review questions. Fava, 2018 (288) Systematic review – inadequate reporting of risk of bias. No quality assessment. Different inclusion criteria. Fava, 2015 (289) Systematic review – inadequate reporting of risk of bias. No quality assessment. Different inclusion criteria. Fava, 2011 (290) Not relevant to review question. Feelemyer, 2014 Wrong population - illicit drug use. (291) Feingold, 2017 No relevant outcomes. Factors associated with problematic use of prescription

Excluded studies

Study Exclusion reason (292) opioids, not risk of dependence. Fenton, 2016 Wrong study type. PhD thesis. (293) Fergusson, 2008 Wrong population. Childhood abuse links to mental health issues including substance (294) dependence (likely illicit). Ferguson, Wrong study design. Non-comparative, so only include if no better data. 2012(295) Fernandes Published prior to 2008. 2002(296) Ferri, 2013(297) Wrong population. Not prescribed opioids for chronic pain (illicit use). Fibbi, 2012 (298) Wrong population. Includes substantial proportion under age of 18. Also includes proportion who may not have been prescribed prescription opioids initially prior to misusing. Fiellin, 2008 (299) Wrong population. Mix of prescription and heroin addicts, results not reported separately. Fiellin, 2014(300) Wrong population. Patients dependent on prescription opioids but those taking prescription opioids for a pain condition were excluded. Fine, 2010 (301) Wrong study design. Non-comparative cohort. Better quality evidence available for opioids for this question (Harms). Finkelman, 2017 Order cancelled - not available. Inspection of abstract suggests not relevant to review (302) questions. Finley, 2017 (303) Wong study type. Qualitative review of prescription drug monitoring programmes. Fischer 2013 No relevant outcomes. Risk factors for non-medical prescription opioid use defined as (304) pain relief pills without a prescription or a doctor telling them to take them. Fishbain 2011 Wrong study design, risk factors and outcomes. (305) Fishbain, 2012 Wrong population. Not all relating to prescribed opioid use. (306) Fister, 2010 (307) Wrong study type. Summary of articles to read in other journals. Fixsen, 2016 (308) Wrong study design. Personal account. Fixsen, 2017 (309) Wrong population. No distinction made between prescribed and illicit use. Fluyau 2018(310) Wrong study type. Literature review, not systematic. Frandsen, 2014 No relevant outcomes. Potential harms of medication rather than (311) dependence/withdrawal. Franklin Wrong study design (higher quality evidence available). 2012(312) Frauger, 2009 Wrong intervention. Drug not included in protocol. (313) Fredheim 2008 No relevant outcomes. (314) Freire 2011 (315) No relevant outcomes. Harm of dependency/withdrawal not discussed. Only that benzodiazepines known to have problems with dependency and should be second choice. Freire 2011 (316) No relevant outcomes. Freynhagen Wrong study type. Qualitative review of pregabalin as a treatment for withdrawal. 2016(317) Fullagar 2009 Wrong study design. Article not a study. (318)

Excluded studies

Study Exclusion reason Furukawa Wrong intervention/comparison. Not approaches to prevent or treat dependence, but 2001(319) the effectiveness of opioids for chronic pain. Gafoor 2018 (320) Not relevant to review question. Looks at relationship between antidepressant prescribing and weight gain. Gahr 2013 (321) Wrong comparison. Cases of withdrawal and discontinuation phenomena from different studies placed in cohorts and compared. Systematic review - did not reference included studies. Gaiennie 2018 Wrong study design (higher evidence available) (322) Garcia-Guix, 2018 Wrong study type / wrong population. Guest editorial on psychiatric comorbidities in (323) people with substance use disorder (illicit use). Garg 2013 (324) Wrong study design (higher quality evidence available). Garland, 2013 Not relevant to review question / no relevant outcomes. Not risk factors. (325) Garland, 2014 Not relevant to review questions. Reanalysis of another study of two withdrawal (326) interventions looking at patterns in outcomes. Garland, 2014 Not relevant to review questions. Looks at the effectiveness of the intervention on (327) pain and function, not related to dependence etc. Garland, 2014 Wrong risk factors. (328) Garland, 2015 Not relevant to review questions. Focus is on reward processing. (329) Garland, 2017 Wrong population. Opioid users but not dependent. (330) Gaughwin Wrong population. Illicit opioid users on methadone programmes. 1998(331) Gerardin, 2014 Wrong population. No relevant outcomes - just used questionnaire to work out (332) numbers they considered dependent; no harms reported. Gerra, 2002(333) Published prior to 2008. Ghaemi, 2010 Wrong population. Shows that no benefit to continue on antidepressants once (334) stabilised from episode. Ghalehney, 2018 Wrong population. Illicit drug use. (335) Ghitza, 2016 Wrong population – not specific to prescription drug use. Narrative review only. (336) Gibiino 2013 Wrong population. No harms associated with dependence/withdrawal just 8 weeks (337) adverse events from drug. Gibson, 2016 Non-UK study (excluded from patients experience question). (338) Gibson, 2017 No relevant outcomes. Study reports development of a benzodiazepine prescribing (339) support tool and reviews in how many previous prescriptions the tool had been adhered to the tool. Gilbert, Published prior to 2008. 1993(340) Gillaizeau, Not relevant to review question. Not relating to drugs on protocol. Review of 2013(341) computerised advice on drug dose to improve prescribing. Gilson, 2012 (342) Wrong study design (higher quality evidence available). GlaxoSmithKline, Cancelled - clinical trial webpage. 1992 (343)

Excluded studies

Study Exclusion reason GlaxoSmithKline, Cancelled - clinical trial webpage. 1993 (344) GlaxoSmithKline, Cancelled - clinical trial webpage. 1994 (345) GlaxoSmithKline, Cancelled - clinical trial webpage. 1994 (346) GlaxoSmithKline, Cancelled - clinical trial webpage. 1994 (347) GlaxoSmithKline, Cancelled - clinical trial webpage. 1994 (348) GlaxoSmithKline, Cancelled - clinical trial webpage. 1995 (349) GlaxoSmithKline, Cancelled - clinical trial webpage. 1996 (350) GlaxoSmithKline, Cancelled - clinical trial webpage. 1999 (351) GlaxoSmithKline, Cancelled - clinical trial webpage 2001 (352) Godfrey, 2008 No relevant effectiveness outcomes. Cost analysis of an intervention trialpublished in (353) 2006. Costs are based on 2005 prices. Costs have not been included as these are uninformative without corresponding effectiveness data. Gonzalez, 2015 Wrong population. Opioid prescription drugs but not with a prescription necessarily - (354) using intra-nasally and also includes those dependent on heroin. Goldstein Wrong population. Illicit drug use. 1973(355) Gomes 2011(356) Wrong population/outcomes. Includes those under 18 years of age. Unclear if deaths due to dependence issue - includes those where death was accidental or due to suicide. Gomes 2017(357) Wrong outcome - not to do with dependence. Case-control study. Looks at co- prescription of gabapentin and opioids and opioid overdose. Not sure it's relevant to the outcome we want? Gorgels 2005 Wrong study design (higher quality evidence available). (358) Gorgels, 2007 No relevant outcomes. Looks at rates of prescribing other drugs when people stop (359) taking BZDs. Gorgels, 2008 No relevant outcomes. Looks at practice visits before and after discontinuation. (360) Gould, 2014 (361) Systematic review. Majority of included studies prior to 2008, others checked for inclusion. Gowing Wrong population. Not prescribed opioids for chronic pain (illicit use). 2016(362) Gowing Wrong population. Not prescribed opioids for chronic pain (illicit use). 2017(363) Gowing, 2009 Wrong population. Systematic review, mainly including illicit drug dependence. (364) Gowing Wrong population. Not prescribed opioids for chronic pain (illicit use). 2010(365) Gowing Wrong population. Not prescribed opioids for chronic pain (majority illicit use). 2017(366)

Excluded studies

Study Exclusion reason Graham, 2008 Not relevant to review questions. Focus on accidental drug deaths. (367) Green, 2011 (368) Not relevant to review question. Trialling people for intrathecal opioids. Green 2015 (369) No relevant outcomes. Systematic review of prescription monitoring programmes web based materials content. Green, 2017 (370) Not relevant to review question. Epidemiological review. Griesbach, 2015 Not relevant to review question. Programme to improve safety of prescribing - no (371) focus on dependence withdrawal etc. Mainly drugs outside of our protocol. Griffin, 2014 Wrong study type: unadjusted data and no clear risk factors. (372) Griffin, 2015 Not relevant to review questions. Looks at quality of life in opioid dependent people (373) compared to those who aren't dependent. Griffin, 2016 Not relevant to review questions. Study describes link between pain and relapse to (374) opioid use. Not risk factor for dependence. Griffin, 2018 Wrong study type. Dissertation thesis. (375) Griswold, 2013 Cancelled - thesis not open access. (376) Gryczynski, 2014 No relevant outcomes. Patients’ reasons for stopping BZD treatment (from (377) questionnaire not views on harms. GSK 2002(378) Citation only. Guaiana, 2016 Systematic review with different protocol. Not relevant to any of our review questions (379) looks at link with PTSD, varies whether people had PTSD at baseline or not. Gudin, 2018, Wrong study design (higher quality evidence available). (380) Guina, 2015 (381) Not relevant to review question. Review of efficacy of benzodiazepines - does include a section on discontinuation / withdrawal but references prior to 2008. Gunderson, 2015 Wrong population. Illicit drug use. (382) Gunderson, 2016 Wrong population. Illicit drug use. (383) Guo 2018 (384) Wrong population. Illicit drug use. Hadjistavropoulos No relevant outcomes. Dependence only mentioned once in vague comment reported , 2014 (385) from a nurse. Hadley 2012(386) Wrong population. Majority using a BZD not on the pre-specified list. Haegerich, 2014 Wrong study design. Narrative review with limited information/data given. (387) Hajak 2003(388) Wrong study design. Review of case reports only. Hajak, 2015 (389) Wrong study design (higher quality evidence available). Hale, 2015 (390) Wrong study design (higher quality evidence available). Hall, 2016 (391) Wrong population. Opioid use, no mention of prescription. Hamann, 2007 No relevant outcomes / not relevant to review question. Doesn't report harms, just (392) that there were no cases of withdrawal. Hammig 2014 Wrong population. Illicit drug use. (393) Hansen, 2011 Wrong study type, wrong population. Very specific population - interviews these (394) people / give questionnaires to determine aberrant use of prescription opioids, but does not report on risks of misusing prescription opioids or harms etc.

Excluded studies

Study Exclusion reason Hantouche, 1998 Cancelled - article not in English (395) Hantouche Published prior to 2008 1998(396) Harrison-Read, Published prior to 2008. 1996(397) Hartung 2018 Wrong study design – higher quality evidence available. (398) Hartung, 2010 Not relevant to review question. Looks at change in prescribing practice of not giving (399) free-prescription samples and effect it has on prescribing. Covers all drugs, not specific to those on our protocol or to dependence etc. Hartung, 2014 Wrong population. Not prescription drug use. (400) Hartzler, 2010 Wrong population. Methadone maintained. (401) Hassan, 2017 No relevant outcomes. Study looked at whether people with PTSD were more likely to (402) have opioid use disorder (no mention of prescription) than non PTSD. Hatcher, 2018 Unclear population. No mention of prescription - appears to be illicit use. (403) Hawk, 2018 (404) Wrong population. Illicit drug use. Hay, 2008 (405) Not relevant to review questions. Reports relapse of bulimia in people discontinuing antidepressants when taking those as a treatment for bulimia. Systematic review and only identifies one RCT that they state is not directly relevant. Not dependence, withdrawal or discontinuation harms. Hay, 2009 (406) No relevant outcomes. Study compared levels of hyperalgesia in people taking opioids for chronic pain vs. methadone maintained people and healthy controls. Hay, 2010 (407) Not relevant to review questions. Reports relapse of bulimia in people discontinuing antidepressants when taking those as a treatment for bulimia. Systematic review and only identifies one RCT that they state is not directly relevant. Not dependence, withdrawal or discontinuation harms. Helm, 2008 (408) Wrong population. Review, no mention of prescription drugs. Focus on illicit use. Hengartner 2018 Wrong study design. Letter. (409) Henriksen, 2016 Wrong population. Illicit drug use. (410) Hesse 2007(411) Review withdrawn from publication. Published prior to 2008. Heser, 2018 (412) No relevant outcomes / not relevant to review questions. Dementia not harm of dependency/withdrawal but long term association with drug. Heslin, 2011 (413) Wrong population / not relevant to review question. Opioid misusers withdrawing from treatment (no mention of prescription opioids - likely to be illicit), looking at correlates with quality of life. Hewitt, 2015 Cancelled - thesis not open access. (414) Hill, 2012 (415) Wrong intervention. Quantitative questionnaire; programme aimed to reduce drug overdose deaths rather than treat dependence. Hillhouse, 2009 Cancelled - conference abstract. (416) Hillhouse, 2012 Cancelled - conference abstract. (417)

Excluded studies

Study Exclusion reason Hindmarch, 2000 Published prior to 2008 (418) Hitzeman, 2010 Not relevant to review questions. Clinical opinion about prescribing opioids for people (419) with osteoarthritis. Hojsted, 2013 Not relevant to review questions. Assessing presence of various addictive behaviours (420) in different groups of patients - those with/without chronic pain treated with/without opioids. Holbrook, 2013 Wrong population. Not prescription drug use. (421) Holland, 1995 Wrong study design. Letter. (422) Holliday2017 Not relevant to review questions. Study to assess whether training would mean fewer (423) prescriptions. Horspool 2008 Systematic review with different protocol. Buprenorphine reducing regime for (424) withdrawal from illicit use or detox following methadone maintenance. Houtsmuller, Wrong population. All heroin users. 1998 (425) Howard, 2016 Wrong population. Illicit drug use. (426) Howe, 2012 (427) Not relevant to review question / wrong population. Looks at people who have desire to stop taking opioids, but survey is closed questions on opioids. Not people's views of harms. Also not necessarily people dependent. Hser 2011(428) Wrong population. Illicit drug users. Hser, 2016 (429) Wrong population. Not prescription drug use. Hser, 2017 (430) Wrong population. Illicit drug use. Hsu, 2004 (431) Cancelled - conference abstract. Hu, 2018 (432) Wrong population. Illicit drug use. Huang, 2002 Cancelled - article not in English. (433) Hunt 2013(434) Wrong population. Substance misuse - not prescription drug use and majority not drugs listed in our protocol. Hussain, 2015 Wrong population. No mention of prescription drugs, appears to be illicit opioid use (435) (including heroin). Hussain, 2017 Cancelled - conference abstract. (436) Iafrati, 2015 (437) Wrong population. Drug addiction, type not specified Iannone, 2013 Not relevant to review questions. Looks at effect of antidepressants on cognitive (438) function. Ilgen 2011(439) Wrong population. Includes those taking non-opioid pain medications and also those with abuse of various substances (not limited to prescription opioids). Inciardi 2009 Wrong population. Abusers of prescription opioids. Not all obtaining legitimately from (440) prescribers. Imani, 2015 (441) Wrong population. Illicit drug use. Isaka, 2009 (442) Cancelled - article not in English Ives 2006(443) Wrong outcomes and pre-2008. Misuse is not limited to dependence/withdrawal symptoms - includes those involved in diversion of drugs, use of illicit stimulants, prescription forgery, etc. Iwanicki, 2016 Not relevant to review question. Looks at rates of abuse and diversion, not harms.

Excluded studies

Study Exclusion reason (444) Iyer, 2008 (445) Systematic review – inadequate methodology (including no risk of bias or quality assessment). Jackman, 2008 Wrong study type / not relevant to review question. Narrative review on approaches (446) to chronic non-malignant pain. Jacob, 2015 (447) Wrong population / not relevant to review question. People's attitudes towards antidepressants - whether taking them or not. Not people dependent / withdrawing from them. Jacobsen, 2015 Wrong study type. Non-comparative open label extension - better quality evidence (448) available for this question. Jain, 2017 (449) Wrong study type / not relevant to review question. Abstract for a workshop about substance misuse. James, 1975 (450) No relevant outcomes. Efficacy outcomes only. Jamison Wrong study design for prognostic review. 2010(451) Jamison Wrong study design. Literature review. 2011(452) Jamison, 2016 Not relevant to review questions. Attitudes of practitioners prescribing opioids. Not (453) risk factors or an intervention. Jarvis, 2018 (454) Wrong population. Not prescription drug use. Jean-Bart, 2017 No relevant outcomes. Study looked at whether certain prescription medications (455) were associated with falls (adverse events of the medication), no indication of length of medication use or dependence Jiang, 2017 (456) No relevant outcomes. Adverse events for drug treatment and not due to withdrawal or dependency. Joffe, 2011 (457) Wrong study type. Abstract. Jo‐Hanna, 2018 Wrong population. Assumed illicit drug use as no mention of prescription. (458) Johnson Wrong population. Mixture of illicit and prescribed use. Diversion included. 2014(459) Jones, 1994 (460) Cancelled - conference abstract. Jones, 2012 (461) Wrong population. Systematic review of buprenorphine for illicit drug use. Julius, 2017 (462) Cancelled - conference abstract. Kahan, 2011 (463) Wrong study type / not relevant to review protocol. Summary of guidelines for safe prescribing, not assessment of the guideline. Kahan, 2011 (464) Wrong study type / not relevant to review protocol. Summary of guidelines for safe prescribing, not assessment of the guideline. Kahler, 2017 Wrong study design (higher quality evidence available). (465) Kamara, 1998 Wrong population. Mixed drugs outside of the protocol. (466) Karow 2011 (467) Wrong population. Heroin addicted patients. Kasteenpohja, Wrong study type / not relevant to review protocol. Questionnaire on treatment 2015 (468) received for depressive disorders. Kasteenpohja, Wrong study type / wrong population / not relevant to review questions. 2016 (469) Questionnaire on treatment received for anxiety disorders. Katz, 2009 (470) Wrong population. Illicit use (heroin). Kavukcu, 2015 No relevant outcomes.

Excluded studies

Study Exclusion reason (471) Kawai 2017(472) Wrong study design. Non-comparative, so only include if no better data. Kawasaki Wrong population. Prescription drug abuse (no indication that it started from 2012(473) prescribed use). Some patients were also on methadone maintenance therapy and had previous benzo withdrawal treatment. Keast, 2015 (474) Wrong study type. Narrative review of policies. Keast, 2018 (475) Wrong study design (higher quality evidence available). Kelly, 2012 (476) Wrong population. Illicit drug use. Kelly 2014(477) Wrong population. Majority non-opioid substances. 20% that were opioid substance abusers not necessarily prescription opioid abusers. Kendall, 2010 Systematic review with different protocol. Cognitive effects of opioids, no (478) dependence outcomes. Kendrick, 2015 Wrong study type. Editorial. (479) Kern 2014 (480) Wrong study design. Case report and literature review. Khalifa 2010(481) Wrong population. Substance misuse - not prescription drug use and majority not drugs listed in our protocol. Khatami Wrong population. Almost all illicit drug use (intervention group 100% heroin). 1982(482) Kheirabadi 2018 Wrong population. Not necessarily prescription opioid dependent. No mention of (483) prescription and includes those smoking opioids. Kidorf 2013(484) Wrong population. Likely to be illicit opioids as all in methadone programme - is not clear but no mention of any being prescription opioid dependent. Kiepek, 2012 (33) Wrong study type. Narrative review only. King 2009(485) Wrong population. Illicit drug users. King, 2014 (486) Wrong population. No mention of prescription drug use, appears to be illicit use. Kitajima 2012 Wrong study design (higher quality evidence available). (487) Klein, 1994 (488) Published prior to 2008. Klemperer Wrong study design and population (drugs mentioned not on protocol (also tobacco 2015(489) and alcohol)). Knisely, 2008 No relevant outcomes. Study tested the sensitivity and specificity of an abuse (490) assessment tool in abusers and not abusers. Kondoni, 2017 Wrong population. Appears to be illicit use (methadone). (491) Kornowski Published prior to 2008. 2002(492) Kouvonen, 2016 Not relevant to review question. Looking at whether antidepressant use increases (493) injuries at work. Kovitwanichkano Wrong study design. Descriptive article. nt, 2018 (494) Kripke 2012(495) Wrong population. Not dependent. Krupitsky, 2011 Wrong population. Not prescription drug use, predominantly heroin. (496) Krauskopf 2017 Wrong study design. Not a clinical study, no outcomes. (497) Krupitsky, 2012 Wrong population. Illicit drug use – heroin. (498)

Excluded studies

Study Exclusion reason Krupitsky, 2013 Wrong population. Illicit drug use – heroin. (499) Krupitsky, 2014 Cancelled - not in English. (500) Krupitsky, 2015 Cancelled - not in English. (501) Krupitsky, 2016 Cancelled - not in English. (502) Krupitsky, 2016 Wrong population. Illicit drug use – heroin. (503) Krupitsky, 2017 Wrong population. Illicit drug use – heroin. (504) Kua, 2014 (505) Cancelled - erratum/corrections. Kumar, 2004 Wrong population / Published prior to 2008. Not prescription drug use. (506) Kunoe, 2010 Wrong population - illicit drug use. (507) Kunz, 2012 (508) Wrong study design (higher quality evidence available). Kuperwasser, Cancelled - conference abstract. 2009 (509) Kurita, 2018 (510) No relevant outcomes. Unclear if people are dependent on opioids, but have been prescribed them for >6 months. Kurokawa, 2011 Wrong population. People who abused BZD; no mention of prescription. (511) Kuyken, 2008 Wrong intervention. Aim of the paper is preventing relapse of depression with no (512) mention of dependence. Kuyken, 2015 Wrong intervention. Aim of the paper is preventing relapse of depression with no (513) mention of dependence. LaBelle, 2016 Wrong population. Not prescription drug addiction. (514) Lader, 1987(515) Published prior to 2008. Lader, 1993(516) Published prior to 2008. Ladewig, 1998 Cancelled - article not in English. (517) Lander, 2015 Wrong population. Not limited to prescription opioids - includes illicit opioids. (518) Landreat, 2010 Wrong study design: Prediction factors are part of the outcome measure; and Wrong (519) population: at least 33% taking non-listed drugs. Lankford 2011 Not relevant to review questions. Narrative review. (520) Lanier, 2012 (521) Wrong study design: unadjusted data. Risk factors from death for prescription opioids. Lanzillotta 2018 Wrong population. Review of postoperative opioid use. (522) Larney 2014 (523) Systematic review with different protocol. Focus on illicit use. LaRowe, 2018 Wrong population: only 57% were prescribed opioids. (524) Lasser, 2016 (525) Wrong study type. Protocol of a multicomponent intervention trial.

Excluded studies

Study Exclusion reason Lauer, 1976 (526) No relevant outcomes. Efficacy outcomes only. Lavie, 2009 (527) Wong population. Assumed illicit drug use as no mention of prescription. Leas, 2010 (528) Cancelled - not available. Inspection of abstract suggests not relevant to review questions. Lecrubier Published prior to 2008. 2005(529) Lee, 2016 (530) Not relevant to review questions. The effect of music on pain. Leonard, 2008 Not relevant to review questions. About opioid induced hyperalgesia, not dependence (531) / withdrawal etc. Leslie, 2015 (532) Wrong study design (higher quality evidence available). Levitt, 2005 (533) Cancelled - clinical trial webpage. Li, 2012 (534) Cancelled - thesis not open access. Li, 2017 (535) Cancelled - article not in English. Liebschutz 2018 No relevant outcomes/ not relevant to review questions. Focus on communication (536) strategies to avoid aberrant drug behaviour. Lin, 2016 (537) Wrong study type / not relevant to review questions. Trial protocol for opioids to reduce pain. Ling, 2013 (538) Wrong population. 57-62% heroin users in each group, the rest are prescription opioid users but no indication that abuse is a result of dependence on opioids prescribed for pain. Ling, 2012 (539) No relevant outcomes. No harms associated with withdrawal/dependency. Short term - 5 days. Ling 1996(540) Wrong population. All illicit drug use (heroin, cocaine, crack). Lintzeris, 2013 Wrong population; wrong comparison. Opioid dependence, no mention of (541) prescription; wrong comparison drug formulation. Liu, 2009 (542) Wrong population / study type. All heroin users. Llorca, 2009 (543) Wrong study design. Abstract. Lobmaier Wrong population. Alcohol dependence and illicit drug use. 2011(544) Lobmaier Wrong population. Substance misuse - not prescription drug use and includes drugs 2008(545) not listed in our protocol. Lofwall, 2016 Wrong population. Non-medical use of prescription drug use and heroin. (546) Lopez-Peig Wrong study design (higher quality evidence available). 2012(547) López-Torres, Wrong study type. Protocol only. 2013 (548) Lu, 2009 (549) Wrong study design. Literature review, not systematic. Lucas, 2010 (550) Wrong population. People taking medical cannabis. Lucas, 2016 (551) Wrong population. Cannabis users. Lucas, 2017 (552) No relevant outcomes. Lum 2011(553) Wrong outcomes. Does not report experiences of harms/treatment of dependence. Misuse not specifically for prevention of opioid withdrawal - includes diversion, for euphoria, etc. Addiction defined as physical dependence and loss of control over use. Lovejoy, 2017 Wrong study design. Not match any of the review questions. Population comparing (554) those with substance use disorders. Ma, 2008 (555) No relevant outcomes. No withdrawal symptoms occurred.

Excluded studies

Study Exclusion reason Mackie, 2017 Wrong population / and outcomes. People with substance misuse disorder (80% (556) heroin) and alcohol misuse. Looking at prevalence of restless leg syndrome. Madden, 2011 Wrong comparison, population. Does not fit any of the questions, wrong population. (557) Maddux Wrong population. Illicit drug use. 1997(558) Main, 2016 (559) Systematic review with different protocol. Not prescription opioid use. Mammen, 2009 Wrong study design / population. Not prescription opioid use. (560) Mannelli, 2013 Wrong population / not relevant. Effect of smoking on detoxification / not (561) prescription opioids. Mannelli, 2014 Wrong population. Not prescription drug use - use of opioids for chronic pain was (562) excluded. Mannelli, 2018 Wrong population. Not prescription opioid use. (563) Manning, 2012 Wrong population. (564) Manubay, 2015 Not relevant to review question / no relevant outcomes. Looks at differences (565) between sexes in opioid misuse, not risk factors. Marcovitz, 2016 Wrong population / no relevant outcomes. Only 16.7% prescription opioid use, looks (566) at predictors for dropping out of treatment programme. Maree, 2016 Systematic review with different protocol - not all prescription use, or relevant to (567) review question. Mariani Wrong population. Not all were prescription BZDs/opioids mixture of those obtained 2016(568) via prescriptions and those from elsewhere. Marschall, 2016 No relevant outcomes. (569) Martel, 2014 Wrong study design: correlations only. (570) Martin, 2015 Wrong study design. Protocol for an RCT on an intervention for reducing (571) inappropriate prescribing (D-PRESCRIBE trial) Martin, 2017 No relevant outcomes. (572) Martin, 2017 No relevant outcomes. (573) Martin, 2018 Wrong population. Illicit use (methadone). (574) Martinez, 2014 No relevant outcomes. Survey of GPs ability to correctly diagnose neuropathic pain (575) and prescribe appropriate medication. Maslej, 2017 Order cancelled - Looks at side effects of antidepressants in people with CVD (576) compared to those without and shows harms lower in that group (mortality & CV events, not dependence etc.). Mastropietro Not relevant to review questions. Looks at tamper resistant formulations. 2013 (577) Matthias, 2018 No relevant outcomes. Patients' experiences of opioid use but no dependency or (578) withdrawal experiences reported. Mattick Wrong population. Not prescribed opioids for chronic pain (majority illicit use). 2014(579) Mattick Wrong population. Not prescribed opioids for chronic pain (majority illicit use).

Excluded studies

Study Exclusion reason 2009(580) Mattila, 1998 Published prior to 2008. (581) Matusow, 2018 No relevant outcomes. (582) Mauger, 2014 Systematic review with different protocol. Inadequate methodology. Includes non- (583) prescription opioid use. Maughan, 2016 Not relevant to review questions. Looks at effects of social prescribing. (584) Mayet, 2014 Wrong population. Mainly includes illicit drug dependence. (585) Mbaba, 2018 Wrong population. Illicit use (primarily heroin). (586) McAdam-Marx, Wrong population. Not limited to prescription opioid abuse/dependence. Includes 2010 (587) those under 18 years. Cannot distinguish between those with prescription opioid dependence and illicit dependence - as stated by study in limitations, McAuley, 2012 Wrong population/intervention. Intervention aimed to reduce overdose among (588) injecting drug users. McCabe, 2009 Wrong population. Not prescription use. (589) McCarthy, 2015 Wrong population. Opioid not included on pre-specified drug list. (590) McDermott, 2015 Not relevant to review questions. Looks at whether initial response is a predictor of (591) overall response. McDonald Wrong population. Not prescribed opioids for chronic pain (majority illicit use). 2016(592) McHugh, 2013 Wrong population - opioids but not currently prescribed for chronic pain (those with (593) chronic pain requiring ongoing management were excluded). McHugh, 2014 Wrong population - opioids but not currently prescribed for chronic pain (those with (594) chronic pain requiring ongoing management were excluded). McMahon, 2011 Not relevant to review questions. Primarily concerns the efficacy of various (595) medications in premature ejaculation treatment. McNicol Wrong intervention/comparison. Not approaches to prevent or treat dependence, but 2013(596) the effectiveness of opioids for chronic pain. Meade, 2010 Wrong population. Half sample used heroin and the rest opioids (not mention if (597) prescription). Mechcatie, 2018 Wrong study type. Letter. (598) Merchant, 2015 Wrong population. Prescription opioids 24%; rest where illicit. (599) Merrill 2012(600) Wrong study design. Survey. Michael, 2018 Not relevant to review questions. Looks at opioid prescribing intensity in the (601) emergency department. Michna 2004(602) Wrong study design: unadjusted data. Middleton, 2014 Wrong study design. Descriptive review of mechanisms of tolerance. Descriptive (603) review of biochemical mechanisms of a certain form of tolerance. No relevant outcomes. Midmer, Published prior to 2008. 2006(604)

Excluded studies

Study Exclusion reason Minegishi, 2018 Wrong study type. Protocol only. (605) Minozzi Wrong population. Not prescribed opioids for chronic pain (majority illicit use). 2013(606) Minozzi, 2013 Wrong population. Includes studies performed on those with acute pain or cancer (607) pain. Minozzi Wrong population. Not prescribed opioids for chronic pain (illicit use). 2011(608) Miotto, 2012 Wrong population. Illicit use - mainly heroin. (609) Miranda, 2018 Wrong population. 68% heroin dependent, 31% used prescription opiates, 33% (610) buprenorphine/naloxone, 9% methadone. Mitchell, 2002 Cancelled - conference abstract. (611) Mol, 2007(612) Published prior to 2008. Mogri, 2014 (613) Not relevant to review questions. Sleep apnoea in people with chronic pain. Mohlman, 2016 Wrong population. No mention of prescription drug use. (614) Moisset, 2016 Cancelled - not available in English. (615) Mokri, 2016 (616) Wrong population. Illicit opioids - no mention of prescription opioids and only all taking either heroin or opium as primary drug. Mol 2006(617) No relevant outcomes. Does not give outcomes for the different interventions - compares craving between those quitting, still using or using intermittently - not relevant Molfenter, 2017 Wrong study design. Protocol. (618) Moller, 1992 Cancelled - article not in English. (619) Momper Wrong population. Not solely people taking oxytocin and mostly illicit use. 2011(620) Montgomery, Wrong study design. Unclear if is a systematic review - methods described 2009 (621) insufficiently. No quality assessment mentioned. Moore, 2012 Wrong study design (higher quality evidence available). (622) Moore, 2014 Wrong population. Predominantly heroin users. (623) Moore, 2016 Wrong population. Patients dependent on prescription opioids but those taking (624) prescription opioids for a pain condition were excluded. Morasco No relevant outcomes. 2008(322) Morasco, 2013 Not relevant to review protocol. Stratifies into risk groups, then analyses factors in (625) those groups to determine association with risk. Making assumption the risk stratification is accurate… Morgan, 2018 Wrong population. Does not specify this was in a prescription opioid population - (626) likely to be illicit or at least a mixture. Morley, 2017 Not relevant question. Survey of predictors of misuse; use of illicit drugs associated (627) with greater odds of misuse. Morris 2014 (628) Wrong outcome and population.

Excluded studies

Study Exclusion reason Morton, Published prior to 2008. 1995(629) Moselhy 2010 Wrong population – heroin users. (630) Moselhy, 2012 Wrong population - illicit drug use. (631) Moy, 2011 (632) Review –different protocol. Search only up to 2007. Mugunthan, 2013 Cancelled - erratum/corrections. (633) Munshi Wrong study design. Abstract. 2013(634) Murphy, 2016 Wrong study design. Narrative review only. (635) Murphy, 2017 No relevant outcomes. Aim was to reduce unnecessary use of the emergency (636) department - not related to opioid dependence etc. directly – opioid misuse. Mutlu, 2015 (637) Wrong population. Illicit use – heroin. Mysels, 2011 Wrong population. Illicit use - heroin. (638) Nadpara, 2018 No relevant outcomes. Risk factors for overdose. (639) Nalamachu, 2011 Wrong study design. Combined analysis of 3 studies. Checked individually for (640) relevance. Naliboff No relevant outcomes. Study design useful but although it mentions the amount that 2011(641) were discontinued from the trial due to opiate misuse, this outcome is combined with those not complying with clinic protocol so can’t be analysed. Nardi 2010(642) Wrong study design - case series. Narayan 2011 Wrong study design. Systematic review that included case reports all published prior (643) to 2008. Nathan, Published prior to 2008. 1986(644) National No relevant outcomes. Audit of local services, no outcome data. Treatment Agency 2011 (645) Naumann, 2018 Wrong study design - PhD dissertation. (646) Nct, 1999 (647) Cancelled - clinical trial webpage. Nct, 1999 (648) Cancelled - clinical trial webpage. Nct, 1999 (649) Cancelled - clinical trial webpage. Nct, 1999 (650) Cancelled - clinical trial webpage. Nct, 1999 (651) Cancelled - clinical trial webpage. Nct, 1999 (652) Cancelled - clinical trial webpage. Nct, 1999 (653) Cancelled - clinical trial webpage. Nct, 2002 (654) Cancelled - clinical trial webpage. Nct, 2005 (655) Cancelled - clinical trial webpage. Nct, 2005 (656) Cancelled - clinical trial webpage. Nct, 2005 (657) Cancelled - clinical trial webpage.

Excluded studies

Study Exclusion reason Nct, 2005 (658) Cancelled - clinical trial webpage. Nct, 2005 (659) Cancelled - clinical trial webpage. Nct, 2005 (660) Cancelled - clinical trial webpage. Nct, 2005 (661) Cancelled - clinical trial webpage. Nct, 2005 (662) Cancelled - clinical trial webpage. Nct, 2005 (663) Cancelled - clinical trial webpage. Nct, 2005 (664) Cancelled - clinical trial webpage. Nct, 2006 (665) Cancelled - clinical trial webpage. Nct, 2006 (666) Cancelled - clinical trial webpage. Nct, 2006 (667) Cancelled - clinical trial webpage. Nct, 2006 (668) Cancelled - clinical trial webpage. Nct, 2006 (669) Cancelled - clinical trial webpage. Nct, 2007 (670) Cancelled - clinical trial webpage. Nct, 2007 (671) Cancelled - clinical trial webpage.. Nct, 2007 (672) Cancelled - clinical trial webpage. Nct, 2007 (673) Cancelled - clinical trial webpage. Nct, 2008 (674) Cancelled - clinical trial webpage. Nct, 2008 (675) Cancelled - clinical trial webpage. Nct, 2008 (676) Cancelled - clinical trial webpage. Nct, 2008 (677) Cancelled - clinical trial webpage. Nct, 2008 (678) Cancelled - clinical trial webpage. Nct, 2009 (679) Cancelled - clinical trial webpage. Nct, 2009 (680) Cancelled - clinical trial webpage. Nct, 2009 (681) Cancelled - clinical trial webpage. Nct, 2009 (682) Cancelled - clinical trial webpage. Nct, 2009 (683) Cancelled - clinical trial webpage. Nct, 2009 (684) Cancelled - clinical trial webpage. Nct, 2010 (685) Cancelled - clinical trial webpage. Nct, 2010 (686) Cancelled - clinical trial webpage. Nct, 2010 (687) Cancelled - clinical trial webpage. Nct, 2010 (688) Cancelled - clinical trial webpage. Nct, 2010 (689) Cancelled - clinical trial webpage. Nct, 2011 (690) Cancelled - clinical trial webpage. Nct, 2011 (691) Cancelled - clinical trial webpage. Nct, 2011 (692) Cancelled - clinical trial webpage. Nct, 2012 (693) Cancelled - clinical trial webpage. Nct, 2012 (694) Cancelled - clinical trial webpage. Nct, 2012 (695) Cancelled - clinical trial webpage. Nct, 2012 (696) Cancelled - clinical trial webpage. Nct, 2013 (697) Cancelled - clinical trial webpage. Nct, 2013 (698) Cancelled - clinical trial webpage. Nct, 2013 (699) Cancelled - clinical trial webpage. Nct, 2013 (700) Cancelled - clinical trial webpage.

Excluded studies

Study Exclusion reason Nct, 2014 (701) Cancelled - clinical trial webpage. Nct, 2014 (702) Cancelled - clinical trial webpage. Nct, 2017 (703) Cancelled - clinical trial webpage. Nelson, 2009 Not relevant to review questions. Literature review only. (704) Neumann Wrong population - majority hydrocodone abusers – not on pre-specified drug list. 2013(705) Neven, 2016 Wrong population. People visiting emergency department frequently with pain. (706) Nielsen, 2012 Not relevant to review questions. Analyses definitions of dependence and withdrawal. (707) Nielsen, 2012 Systematic review. Inadequate quality. No mention of quality assessment of included (708) studies. Outcomes reported insufficiently - Just mentions the symptoms doesn’t specify number of patients that experienced each. Nielsen, 2013 Not relevant to review questions. Comparing heroin to opioid prescription users. (709) Nielsen, 2014 Not relevant to review questions. Looks at characteristics associated with dependence (710) this is reporting on difficult induction of treatment for dependence. Nielsen, 2015 Not relevant to review questions. Comparing heroin to opioid prescription users. (711) Nielsen 2016(712) Wrong population. Does not specify dependence is of prescribed drugs. Background text implies it is illicit use. Nielsen, 2016 No relevant outcomes. (713) Nielsen, 2018 Wrong intervention (intraoperative ketamine). (714) Nielsen 2018(715) Systematic review. Inadequate methodology. No protocol, GRADE done but not by outcome, no data that can be analysed or reused. Nihr, 2016 (716) Wrong study type / wrong population. Report of opioid dependence treatment, looks to be mainly illicit opioids. Nikulina, 2016 Not relevant review question. Patient vs prescriber reported of aberrant medication (717) taking. Nilsen, 2010 (718) Wrong study design. Case series. Noble, 2008 (719) Systematic review. Inadequate methodology. All studies were pre2008. Noble 2010(720) No relevant outcomes. Does not report on harms of addiction just the risk of addiction in this group Noller, 2018 (721) Wrong population. Not prescription drug use. Nordfjærn, 2013 Wrong study design: risk of harms from benzodiazepine use (not dependence) (722) compared with those not using benzodiazepines. Norman, 2016 Wrong study type. Collates information to inform best practice, but concludes further (723) research is needed. Study aim to review best practices. Focus on codeine - not necessarily prescribed, also includes OTC use. Based in Canada. Northrup, 2015 Wrong population. Not prescription drug use. (724) Nosyk, 2011 (725) Wrong population. Assumed illicit drug uses as no mention of prescription. Nosyk, 2015 (726) Wrong population. Mixture of prescription opioid and heroin. Those abusing/dependent on prescription opioids not necessarily obtaining them originally for chronic pain from doctor. Majority heroin dependent. Those with prescription opioid-dependence not necessarily prescribed them - not mentioned. Could have

Excluded studies

Study Exclusion reason obtained illicitly without prescription. Novak, 2009 (727) Wrong population. Non-medical use of prescription analgesics (also aim of study not directly relevant - link to pain). Nuesch, 2009 Wrong study type. 'Cochrane highlight' - similar an abstract. Just one comment on (728) withdrawal symptoms after fentanyl treatment. Nunes, 2018 Wrong population. Opioid use disorder (not prescription) (729) Nurminen, 2014 Wrong study design (higher quality evidence available). (730) O'Connor Wrong study design / not relevant to review questions. Factor analysis to develop a 2008(731) tool. O'Connor, 2012 Wrong study design (higher quality evidence available). (732) Ogawa, 2010 Cancelled - article not in English. (733) Ogle, 2013 (734) Wrong study type. Guidance based on a literature review of guidelines. Oliva, 2017(735) Wrong population. Not limited to those receiving opioids for chronic pain - also includes cancer pain and may include acute pain. Also, outcomes are overdose and suicide attempts but not explicitly related to dependence in these patients. O'Mullan, 2015 No relevant outcomes. Qualitative study on women’s experiences of sexual side (736) effects of long-term SSRIs. Onyett Stephen, Published prior to 2008. 1988(737) Osmun, 2011 Wrong population. Not risk factors for dependence etc. just difference between a (738) group using oxycodone and a group not using it. Ostini, 2011 (739) Wrong population. People included on any drugs - ordered papers where used one of our drugs. Otto, 1993(740) Published prior to 2008. Otto 2010(741) Wrong population. Drugs not on our list (mainly clonazepam and some alprazolam) Oulis 2009(742) Wrong study design. Case report of one patient. Oulis, 2010 (743) Wrong study design / not relevant to review questions. Descriptive review of pregabalin. Pade, 2012 (744) Wrong population. Opioid abuse, prescription and illicit, no clear link with initial prescribed use. Page, 2018 (745) Wrong study design (higher quality evidence available). Page, 2018 (746) Wrong study type. Protocol only. Pani 2010(747) Wrong intervention. About treating depression not dependence. Pani 2013(748) Wrong population. Not prescription drug dependence. Paolucci, 2017 Wrong population. Doesn't state what the overused drug is - possible could be drugs (749) not included in our protocol. Papaleontiou, Systematic review – not relevant to this protocol. Focus on efficacy, safety and abuse 2010 (750) potential. Papazisis, 2010 Cancelled - conference abstract (751) Paquin 2014 (752) Wrong comparison. Unclear if comparative studies from review - treatment types were not compared. Pardo, 2017 (753) Wrong study design (higher quality evidence available). Park 2010 (754) Wrong population. Includes those with cancer pain.

Excluded studies

Study Exclusion reason Parker, 2018 Wrong population. Opioid misuse (heroin or illicit use of prescription drugs). (755) Parr, 2009 (756) Wrong population. Unclear if users in all studies obtained BZDs through prescription or whether they were using illicitly/recreationally. SR was also not a Cochrane SR. Parr, 2011 (757) Wrong study design (higher quality evidence available). Parran, 2010 Wrong population. Illicit drug use - 88% heroin users. (758) Passik, 2006(759) Published prior to 2008. Passik, 2008 (760) Wrong study type / not relevant to review question. Review of tools to assess opioid addiction. Passik, 2009 (761) Wrong study type. Narrative review only. Passik, 2011 (762) Wrong study design: Unadjusted data. Passik, 2014 (763) No relevant outcomes. Doesn't inform the harms. Pat- Published prior to 2008. Horenczyk,1998(7 64) Patrick 2014(765) Wrong population. Illicit prescription opioid abusers. Paulozzi Wrong population. Not specifically opioids for chronic pain - could include deaths 2006(766) from adverse short term events, misuse/abuse etc. Paulozzi, 2012 Wrong population: mixed licit and illicit use, wrong study design - not prognostic (767) factors. Pauly, 2018 (768) Wrong population: not all using prescription opioids and unclear if opioids were prescribed. Peckham, 2018 Exclude: wrong study design - overuse as risk factor rather than outcome. (769) Peirce, 2012 (770) No relevant outcomes. Study compared rates of doctor/pharmacy shopping in cases of drug related deaths and controls. Peles, 2008(771) Published prior to 2008. Penninga, 2016 Wrong study type / not relevant to review population. Looks at adverse events (772) reported from a treatment for intoxication, not efficacy of that treatment Perahia, 2008 Wrong study design. 2 RCTs combined but not systematically. Studies prior to 2008. (773) Pergolizzi, 2018 Wrong study type / not relevant to review population. Review of abuse-deterrent (774) opioids. Pergolizzi, Wrong study design - literature review. 2012(775) Pietrasanta 2011 Wrong population. Illicit use - mostly methadone maintenance. (776) Pikovsky, 2018 Cancelled - not available. Inspection of the abstract suggests illicit use. (777) Pilkonis, 2017 Not relevant to review questions. Development of an 'item bank' to measure (778) prescription medication abuse. Pilowsky, 2011 Wrong population. Alcohol use disorder. (779) Pimlott, Published prior to 2008. 2003(780) Pink 2012(781) Wrong study design. Assesses opinions/practicality of method for assessing problematic prescription drug use and discusses how this method was altered

Excluded studies

Study Exclusion reason following feedback. Pohjanoksa 2009 No relevant outcomes. No discussion of harms of withdrawal or dependency. (782) Pollack 2008(783) No relevant outcomes. Wrong intervention. Pollmann, 2015 Wrong study design. Review about the use of these drugs in community dwelling (784) adults. Polsky, 2010 Wrong population. Mainly illicit drug use (41% heroin, 8% mixed, 26% opiate (785) analgesics, but don't know if these were on prescription. Popovici, 2018 Wrong population. Population not limited to those with opioids for chronic pain. Also (786) talks about nonmedical use of prescription opioids being those obtaining without a prescription/using for euphoria, which is not relevant to the reviews. Posadzki Wrong population. Not prescribed opioids for chronic pain (illicit use); other groups 2016(787) not listed in the protocol (e.g. alcohol-dependence). Posadzki, 2016 Wrong study design. Overview of systematic reviews. (788) Potter, 2010 (789) Wrong population. Illicit drug use. Potter, 2015 (790) Wrong study design (higher quality evidence available). Pottie, 2018 (791) Systematic review using GRADE but results are reported narratively as recommendations. No extractable outcome data. Prakash, 2016 Wrong population. Opioid dependent but no mention of prescription. (792) Prendergast Wrong population. Illicit drug use - aim is to reduce HIV risk 2001(793) Priddy, 2018 Wrong study design: correlations only. (794) Prommer, 2017 Wrong study type. Guideline - not assessment of its use. (795) Puangkot 2011 Not relevant to review questions. Prevalence data only of use. (796) Public Health No relevant outcomes. Guidelines for prevention, no outcome data. England, 2013 (797). Puustinen, 2014 Wrong study design (higher quality evidence available). (798) Quagliato, 2018 Systematic review with different protocol. (799) Raheem, 2017 Not relevant to review question. Treatment of an opioid side effect. (800) Rahimi‐Movaghar Wrong population. Not prescribed opioids for chronic pain (illicit use). 2013(801) Raisch, 2012 Wrong population. Appears to be illicit opioid use - mentions possibility of extending (802) treatment to prescription opioids in the future. Raknes, 2017 Not relevant to review population / wrong population. Looks at effect of low dose (803) naltrexone on opioid consumption - but not in people dependent on opioids. Raoof, 2008 (804) Not relevant to review questions. Views and awareness of benzodiazepines. Raphael, 2013 No relevant outcomes. Aim isn't to reduce use of morphine, but to show that long (805) term use is effective and reducing dose increases pain and number dropping out. Raskin, 2014 No relevant outcomes. No harms associated with dependence/withdrawal just 6 (806) weeks AE from drug.

Excluded studies

Study Exclusion reason Rathlev, 2016 Wrong population. Not prescription drug use – OUD. (807) Reece, 2010 (808) Wrong population. Illicit opioid users – heroin. Reece, 2009 (809) Wrong population. Does not mention prescription addiction so presumed illicit. Reeve, 2017 (810) Systematic review – inadequate reporting of risk of bias. No quality assessment. Reid, 2002 (811) Published prior to 2008. Reifler 2012(812) Wrong population. Illicit abuse/misuse of prescription opioids - using for non-medical purposes to get high. Reimer, 2016 Wrong study type / not relevant to review question. Impact of misuse and diversion. (813) Reisman 2009 Wrong study design (higher quality evidence available). (814) Rickels, 1999(815) Published prior to 2008. Rigg, 2010 (816) Wrong population. Included participants predominantly obtaining and using initially in illicit ways - without prescription from doctor or using by improper methods Rikala, 2011 (817) Wrong population. Looks at factors associated with persistent use of psychotropic drugs. Not necessarily dependence. Ringwalt, 2014 Not relevant to review questions. Looks at which specialists prescribe opioids the (818) most - but cannot assume this as a risk factor. Ristanovic, 2009 No relevant outcomes. Treatment is for cataplexy, reports whether that increased (819) when initially withdrew. Comparison of number of cataplexy events between groups. Robinson, 2015 No relevant outcomes. Views on opioids, not experiences of harms of dependence (820) etc. Roehrs, 2012 Not relevant to review question. At months 1, 4 and 12 there was a 7 day withdrawal (821) period where placebo replaced the active intervention - study assesses this phase. Not assessing discontinuation or withdrawal, or demonstrating harms. Rojo-Mota, 2017 Wrong population / study type. ‘Substance addiction' not prescription. (822) Romach, Published prior to 2008. 1998(823) Romach, Published prior to 2008. 2000(824) Roman 2009 Wrong study design. Not a clinical study, no outcomes. (825) Rosenblum Wrong study design (higher quality evidence available). 2012(826) Rosenthal, 2013 Wrong population. Illicit use - excluded if taking opioids for chronic pain. (827) Rounsaville Wrong population. Illicit opiate use. 1983(828) Roussin, 2013 Wrong population. Non-prescription codeine use or antihistamines. (829) Roux, 2013 (830) Wrong population. Mix of prescription and illicit drug use - higher percentage of prescription though. Non-comparative, so only include if no better data. Rudolf, 2018 Wrong population. Illicit drug use. (831) Ruetsch, 2014 Wrong study design. Review of another included study. (832)

Excluded studies

Study Exclusion reason Ruetsch, 2012 Wrong population. No mention of prescription drugs so assume illicit use. (833) Ruetsch, 2010 Wrong population. Illicit drug use. (834) Sabzghabaee, Wrong population. Inappropriate use of naltrexone', no indication that abuse began 2012 (835) after dependence on prescription opioids. Safer, 2017 (836) No relevant outcomes. Only mentions relapse rate as main outcome - brief mentions of withdrawal symptoms but insufficient detail. Compares five different maintenance methodologies for antidepressant treatment - assessing rate of relapse, not dependence/withdrawal. Saleh, 2014 (837) Wrong population. Appears to be illicit use. Saleh, 2014 (838) Not relevant to review question. Predictors of successful withdrawal, not predictors of harms. Salehi, 2015 (839) Wrong population. Illicit use. People on a methadone maintenance programme. Salminen, 2009 No relevant outcomes. Falls only. Other papers reporting on this study with relevant (840) outcomes are included. Sanchez, 2017 Wrong population. Illicit drug use. (841) Saul, 1989(842) Published prior to 2008. Saulle 2017(843) Wrong population. Not prescribed opioids for chronic pain (illicit use). Saunders 2015 Wrong study design (higher quality evidence available). (844) Saxon, 2018 (845) Wrong population. Population not specifically prescription opioid dependence - likely to include those dependent on illicit drugs. Saxon 2013(846) Wrong population. Includes those using and dependent on illicit opioids - Not limited to prescription opioids. Saxon, 2010 (847) No relevant outcomes. States no reports of misuse or abuse; but not what harms these are individually. Schaffer 2010 No relevant outcomes. (848) Schepis, 2018 Wrong study design: t-stat and p-value only. (849) Scherbaum Wrong population. Illicit drug use. 2005(850) Schieffer, Published prior to 2008. 2005(851) Schifano Wrong population. Not prescribed medicines, abuse of prescribed medicines. 2018(852) Schroeder, 2018 Wrong population. Only mentions heroin in study - illicit users. (853) Schuman-Olivier, Not relevant to review question. Looks at factors linked to retention in substance use 2014 (854) disorder treatment programmes. Schuman-Olivier, Wrong population. Mixture of heroin & prescription opioid use 27% only used 2014 (855) prescription opioids. Schwarzer, 2015 No relevant outcomes. Apnoea adverse event of drug not harm of (856) withdrawal/dependency. Schweizer, Published prior to 2008. 1995(857) Schweizer Published prior to 2008.

Excluded studies

Study Exclusion reason 1990(858) Seay, 2017 (859) No relevant outcomes. Segal, 2017 (860) Not relevant to review question. Not looking at dependence etc. Sekhon, 2013 Not relevant review question. (861) Setnik, 2013 (862) Wrong intervention. Prevention as crushed pill releases naltrexone that stops morphine working effectively. Setnik, 2015 (863) Wrong study design (higher quality evidence available). Setnik, 2017 (864) Wrong study design. Comparison of methodology and results of 2 studies. Shigemura No relevant outcomes. Not specifically related to withdrawal or dependency. 2007(865) Shokrzadeh, 2017 Cancelled - article not in English. (866) Shorey, 2017 Wrong population. Mostly alcohol substance abuse. (867) Siassi, 1975 (868) No relevant outcomes. No patients showed any evidence of psychological or physical withdrawal symptoms 1 week post treatment. Sigmon, 2009 Wrong study design (higher quality evidence available). (869) Sigmon 2013(870) Wrong population. People using prescription opioids illicitly - those prescribed them for pain were excluded. Simon, 2013 (871) Wrong study design. 2 RCTs combined but not systematically. Simonsen, 2016 Systematic review with different protocol. Animal studies. (872) Sinnemaki, 2013 Not relevant to review protocol. Not specific to drug dependence etc. (873) Slofstra, 2017 Wrong study design. Protocol and methodology. (874) Sirdifield, 2017 Not relevant to review protocol. Review of qualitative studies - remit broader than (875) our protocol. Smith, 2010 (876) Wrong study design (higher quality evidence available). Socias, 2018 (877) Wrong study design. Design and rationale only. Soeffing, 2009 Wrong population. Illicit use. Mainly heroin. (878) Sohler, 2010 Wrong population. 15% prescribed opioids rest are illicit. (879) Sokol, 2018 (880) Wrong population. Opioid use disorder - not prescription opioids. Soldatos, Published prior to 2008. 1999(881) Soyka, 2011 (882) Cancelled - article not in English. Spiegel, Published prior to 2008. 1994(883) Spiller 2009 (884) Not relevant to review questions. Stannard Wrong intervention. About treating pain not dependence. 2016(885) Stanton Wrong population. Illicit drug abuse or addiction. 1997(886) Starrels, 2010 Wrong outcomes. Drug misuse including diversion etc.

Excluded studies

Study Exclusion reason (887) Steele, 2012 (888) Wrong population. Illicit drug use. Stein, 2015 (889) Wrong population. Does not mention opioids being prescribed to the patients in the prescription opioid group. Assume may have obtained illicitly as comparing with heroin. Compares life concerns of prescription opioid and heroin users. Stein, 2008 (890) No relevant outcomes. Wrong population. Illicit use. Stewart, 2007 Published prior to 2008. (891) Strain, 2011 (892) Wrong population. Heroin dependence only. Strang, 2017 Wrong population. Illicit drug use. (893) Strick, 2014 (894) No relevant outcomes. Effectiveness and tolerability outcomes only. Su 2018(895) Wrong population / study design (higher quality evidence available). Methadone being used in this scenario to treat opioid withdrawal syndrome. Subodh, 2017 Wrong population. (896) Subramaniam, Wrong population. Assumed to be illicit use due to mention of injection and no 2017 (897) mention of prescription throughout study. Sullivan No relevant outcomes. No qualitative outcomes (patient experience question). 2010(898) Sullivan, 2017 Wrong population. Illicit use. Excluded people with opioids prescribed for chronic (899) pain. Sun, 2016 (900) Wrong population. Opioid-naïve surgical patients. Syed, 2013 (901) Wrong population. Illicit use. Mainly heroin. Syed, 2018 (902) Wrong population. Surgical patients; opioids not prescribed for chronic pain. Sylvestre, 2012 No relevant outcomes. Harms of the drugs rather than discontinuation. (903) Szczypa Wrong study design. Abstract only. 2009(904) Tavakolian, 2016 Wrong population. No mention of prescription, appears to be illicit use ('opium' used (905) as a buzz word in the intervention). Tetrault, 2012 Wrong population. (906) Thakral, 2018 Wrong study design (higher quality evidence available). (907) Thorlund, 2016 Systematic review with different protocol. Not relevant to review question. (908) Tint, 2009 (909) Cancelled - erratum/corrections. Tkacz 2012 (910) Not relevant to review question. Comparison of demographic data between prescription and street opioid use. Tkacz, 2012 (911) Wrong population. Appears to be illicit use. Tofighi, 2017 Wrong population, intervention. Under half opioid addicts of whom 39% was heroin. (912) Tompkins, 2009 Wrong population. Illicit drug use. (913) Tompkins, 2014 Wrong population. Illicit use – heroin. (914) Tonks, 2008 (915) Wrong study design. Summary.

Excluded studies

Study Exclusion reason Tonks, 2010 (916) Wrong study design. Summary. Tonks, 2010 (917) Wrong study design. Summary. Tonks, 2011 (918) Wrong study design. Summary. Tourian 201(919) Wrong study design. Non-comparative, so only include if no better data. Tournebize Wrong intervention. Looks at adherence to risk reduction only, not what effect that 2016(920) has on opioid dependence etc. Townsend, 2008 Wrong study design (higher quality evidence available). (921) Tribl, 2013 (922) Not relevant to review questions. Systematic review does not directly address any of the review questions. Trivedi, 2010 No relevant outcomes. No discontinuation/withdrawal related outcomes reported. (923) Turk, 2008 (924) Systematic review. Inadequate methodology. Included studies pre-2008. Tyrer, 1981(925) Published prior to 2008. Tyrer, 1996(926) Published prior to 2008. Udelman, Published prior to 2008. 1990(927) Uebelacker, 2015 No relevant outcomes. (928) Urton 2017 (929) Wrong study design (higher quality evidence available). Uzun, 2010 (930) Study design. Review of literature- descriptive only. Vaapio, 2015 No relevant outcomes. Patients divided in to those who withdrew, those who reduced (931) and non-users and reported symptoms for each group. Vankova, 2014 Wrong population. Not on prescription drugs. (932) Veal, 2017 (933) Wrong study design. Letter. Veiraiah, 2012 Wrong population. Majority heroin users and others not specified as prescription (934) opioids. Velert Vila, 2012 Cancelled - article not in English. (935) Vicens 2006(936) Pre-2008 study. Vicens, 2011 Wrong study design. Study protocol. (937) Victorri 2011 Not relevant to review question. Over use due to abuse/misuse or lack of (938) effectiveness etc. Victorri-Vigneau Published prior to 2008. 2007(939) Vittengl, 2017 Weren't able to order, but abstract provided clarifying it's a letter to editor. (940) Vo, 2018 (941) Cancelled - not available. Inspection of the abstract suggests illicit use. Voderholzer Published pre-2008. Population isn’t a group of patients receiving the drugs for a 2001(942) condition - healthy males. Vogel, 2011 (943) Wrong outcome: prolonged BZD use and wrong population: not all using BZD on prescription. Vogel, 2017 (944) Wrong study design. Commentary. Von Korff, 2017 Wrong study design (higher quality evidence available). (945)

Excluded studies

Study Exclusion reason Voon, 2017 (946) Wrong study design / population / not relevant to review question. Include substance use generally as well as prescription opioids. Main focus is on prevalence. Voris John, 1989 Published prior to 2008. (947) Vorma, 2002(948) Published prior to 2008. Vorma 2010(949) Wrong population. Illicit users - those with medical illnesses were excluded. No mention of prescriptions. Vosburg Wrong population. Healthy individuals abusing prescription opioids via nasal route. 2012(950) Vosburg Wrong study design. No relevant outcomes. 2018(951) Voshaar Wrong outcomes. Identifies predictors of successful BZD discontinuation in patients - 2006(952) not factors associated with dependence or their harms. Vowles, 2015 Systematic review – different protocol. Only gives rates of behaviours and does not (953) mention factors. Gives rates of abuse/misuse across different studies and compares among different types of study design etc. Voyer, 2010 (954) No relevant outcomes. Number that self-diagnosed themselves as dependent. No harms. Wakeland, 2013 Wrong study design. Report of simulated model. (955) Walsh, 2017 (956) Wrong population. Illicit use - chronic pain requiring opioid therapy was an exclusion criterion. Wang, 2010 (957) Wrong population. Mix of illicit and prescription users - predominantly heroin. Wang, 2011 (958) No relevant outcomes - pain sensitivity. Three groups: low back pain with and without opioid therapy, and healthy controls with neither pain nor opioids. Wang, 2015 (959) Wrong study design. Case control study. Watson, 2010 Wrong outcomes. Mainly refers to efficacy/adverse events of taking opioids. (960) Mentions withdrawal but only whether or not they experienced withdrawal symptoms if missing doses, and not assessing factors contributing to this or the harms. Weber, 2010 Wrong study design. Literature review. (961) Webster Pre-2008 study. 2005(962) Webster, 2010 Intervention. Testing new formulation of drug to see if tamper proof. Assessing long (963) term safety of a formation that is less prone to abuse when crushed in the population of chronic pain. Webster, 2015 Wrong intervention. (964) Webster, 2016 Wrong population –illicit use. (965) Webster, 2016 Not relevant to review questions. Looking to see whether withdrawal reaction can be (966) initiated. Weiss 2011(967) Wrong population. Not all participants were prescribed for pain. May include some that obtained them illicitly. Only 40% had chronic pain diagnosis at time of entry. Weiss 2010(968) Wrong population. Includes some that were not prescribed them for chronic pain - obtaining them illicitly. Weiss, 2009 (969) Wrong population. Mix of drug/alcohol dependence (prescription not mentioned). Weiss, 2014 (970) Wrong outcomes. Although chronic pain population mentioned separately, only

Excluded studies

Study Exclusion reason mentions reasons for initiating opioid intake (not harms or factors contributing to dependence). Weiss, 2015 (971) Wrong population. Not limited to those being prescribed prescription opioids for chronic pain. Includes those may have obtained illicitly and not had them prescribed. Weiss, 2017 (972) Wrong population. Only 42% of the participants had chronic pain - reviews are only concerned with opioids for chronic pain. Welsh, 2018 (973) Wrong study design. Literature review rather than systematic review. Westanmo 2015 Wrong study design (higher quality evidence available). (974) Westbury, 2011 Wrong outcomes. Paper focuses on success of reducing antipsychotic/BZD due to risk (975) of falls etc. in elderly nursing homes. No mention of dependence throughout. Whitehead 2018 Summary of review. Cochrane systematic review it refers to (Eccleston) already (976) included. Whiteside, 2017 Wrong population. Includes various prescription drugs (e.g. Stimulants which are not (977) included in our review). Not limited to those initially prescribed the drugs legitimately - includes some illicitly obtained. Wightman Wrong population. Mostly heroin users or recreational opioid users. 2012(978) Willems, 2013 Wrong outcomes. Assessing whether dose of BZD increases over time in incident and (979) prevalent prescription users of BZD - not describing harms or factors contributing to harms of dependence. Wilsey 2009(980) Wrong outcomes. Assessing effects of short vs. long term opioids in patients at high risk of abusing prescription opioids for chronic pain - outcomes are effects of taking the drug rather than harms of dependence/withdrawal. Wilson, 2015 Wrong study design. Literature review. (981) Wittchen, 2011 Cancelled - article not in English. (982) Wobrock, 2008 Systematic review with different protocol. Looks at comorbid substance use disorder (983) and schizophrenia. Wobrock, 2009 Wrong intervention. Looks at comorbid substance use disorder and schizophrenia. (984) Wong 2014 (985) Wrong study design. Case cross-over study looking at adverse events. Woody 2008(986) Wrong population. Predominantly illicit users of opioids with no mention of prescription opioids. Also substantial proportion under 18 years. Woody 1983(987) Wrong population. Illicit opiate use. Worley, 2015 Not relevant to review questions. Looks at predictors of favourable or poor treatment (988) outcome (treatment for opioid dependence) in chronic pain subpopulation. Worley, 2015 Wrong population. Includes those obtaining illicitly (without prior prescription) and (989) legitimately from doctors etc. Worley, 2017 Not relevant to review questions. Predictors of response to treatment, not risk of (990) harms. Wright 2015(991) Systematic review. No quality assessment. References checked and relevant studies ordered. Wu, 2011 (992) Not relevant to review questions. Looks at how ceasing BZDs may lead to improvements in risk of dementia - not necessarily harms of BZD dependence. Xiang, 2007 (993) Wrong study design and population. Compares long term use with no use so not all are using the prescribed medicine. Xie, 2014 (994) Wrong population. Patients prescribed opioids (or not prescribed them) for various

Excluded studies

Study Exclusion reason conditions, not limited to chronic non-cancer pain (also includes cancer pain). Xue, 2012 (995) Wrong study design. Protocol only. Yang 2013(996) Wrong population. Opioid for premature ejaculation not chronic pain. Yang, 2006 (997) Cancelled - article not in English. Yarbrough, 2018 Wrong study design (higher quality evidence available). (998) Yasui-Furukori, Wrong study design. Non-comparative cohort. Separates participants into those with 2016 (999) and without discontinuation syndrome, but only reports results for those with the syndrome. Yen, 2015 (1000) Wrong study design: correlation only. Yeo, 2017 (1001) Systematic review with different protocol. Not relevant to review question. Yeung, 2017 Wrong study design. Study protocol. (1002) Yovell, 2016 No relevant outcomes. Effectiveness outcomes only. (1003) Zacny 2009(1004) Wrong population and no relevant outcomes. Zacny 2003(1005) Wrong population. Performed in people that were healthy and not receiving oxycodone for a medical reason. Measures effects of the drugs rather than harms of dependence/withdrawal. Zacny 2003(1006) Wrong population. Performed in people that were healthy and not receiving the drugs for a medical reason. Measures effects of the drugs rather than harms of dependence/withdrawal. Zacny 2008(1007) Wrong population. Performed in people that were healthy and not receiving the drugs for a medical reason. Measures effects of the drugs rather than harms of dependence/withdrawal. Zahradnik Wrong study design for Q2. Includes opioids, anxiolytics, hypnotics, sedatives and 2009(1008) caffeine -reports some separately. May include some abusing PDs rather than dependent on (see discussion section) Zajecka, 1998 Published pre2008. (1009) Zarghami, 2012 Wrong population. Not prescription use. (1010) Zedler, 2014 Wrong outcomes. Serious opioid-induced respiratory depression or overdose not (1011) related to opioid dependence, discontinuation or withdrawal. Zedler, 2015 Wrong outcomes. Serious opioid-induced respiratory depression or overdose not (1012) related to opioid dependence, discontinuation or withdrawal. Zgierska, 2016 Not relevant to review question. Although suggests could help reduce opioid use - (1013) opioids were allowed during the study. Zgierska, 2016 Not relevant to review question. Looking at effectiveness of an alternative (1014) intervention, but in people taking opioids Zhang, 2013 Cancelled - article not in English. (1015) Zheng 2008(1016) Wrong intervention. Aim to reduce opioid use by replacing with acupuncture in people who aren't addicted, rather than preventing addiction. Ziaaddini, 2012 Wrong population. Illicit use – opium (1017) Ziedonis, 2009 Cancelled - not available. Inspection of abstract suggests illicit drug use and alcohol (1018) dependence. Zippel-Schultz, Cancelled - not available. Inspection of abstract suggests illicit drug use.

Excluded studies

Study Exclusion reason 2016 (1019) Zisook, 2011 No relevant outcomes (1020) Zullich, Published prior to 2008. 1992(1021) Zullig, 2018 Wrong population. Illicit drug use – opioid use disorder. (1022)

List of medicines to be included

Appendix H: List of medicines to be included

This list refers to codes from BNF version 68. Drug class (for this analysis) BNF chapter Drugs included Opioids 4.7.2 Buprenorphine Codeine* Dextromoramide Diamorphine Dihydrocodeine** Dipipanone (including with cyclizine) Fentanyl Hydromorphone Meptazinol Methadone Morphine (including with cyclizine) Oxycodone (including with naloxone) Papaveretum Pentazocine Pentazocine Pethidine Tapentadol Tramadol (including with paracetamol) 4.7.1 Codeine with paracetamol = co-codamol* Dihydrocodeine with paracetamol = co-dydramol** Z-drugs 4.1.1 Zaleplon Zopiclone Zolpidem Benzodiazepines 4.1.1 (insomnia) Flurazepam Loprazolam Lormetazepam Nitrazepam Temazepam 4.1.2 (anxiety) Diazepam Chlordiazepoxide Lorazepam Oxazepam Gabapentinoids 4.7.3 Gabapentin 4.8.1 Pregabalin Antidepressants 4.3.1 (Tricyclics) Amitriptyline (including with perphenazine) Amoxapine Clomipramine

Dosulepin

List of medicines to be included

Drug class (for this analysis) BNF chapter Drugs included Doxepin Imipramine Lofepramine

Maprotiline

Mianserin

Nortriptyline

Protriptyline Trazodone Antidepressants Trimipramine (continued) 4.3.2 (MAOIs) Isocarboxazid Moclobemide Phenelzine Tranylcypromine 4.3.3 (SSRIs) Citalopram Escitalopram Fluoxetine Fluvoxamine Paroxetine Sertraline 4.3.4 (Other Agomelatine antidepressants) Duloxetine Flupentixol Mirtazapine Nefazodone Oxitriptan Reboxetine Tryptophan Venlafaxine Vortioxetine

* Although they are captured within different BNF chapters, codeine and co-codamol will be regarded as a single drug when considering co-prescribing within the opioid class.

** Although they are captured within different BNF chapters, dihydrocodeine and co-dydramol will be regarded as a single drug when considering co-prescribing within the opioid class.