For Peer Review Only

BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from

Pharmacologic interventions to improve sleep in hospitalized adults: A systematic review

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2016-012108

Article Type: Research

Date Submitted by the Author: 07-Apr-2016

Complete List of Authors: Kanji, Salmaan; The Ottawa Hospital, Pharmacy and Critical Care; The Ottawa Hospital Research Institute, Clinical Epidemiology Mera, Alexandru; The Ottawa Hospital, Pharmacy Hutton, Brian; University of Ottawa, Ottawa, Ontario, Canada, Clinical Epidemiology; The Ottawa Hospital Research Institute, Clinical Epidemiology Burry, Lisa; Mount Sinai Hospital, Pharmacy Rosenberg, Erin; The Ottawa Hospital, Critical Care MacDonald, Erika; The Ottawa Hospital, Pharmacy; The Ottawa Hospital Research Institute, Clinical Epidemiology Luks, Vanessa; Thunder Bay Regional Health Sciences Centre, Respirology

Primary Subject Pharmacology and therapeutics

Heading: http://bmjopen.bmj.com/

Secondary Subject Heading: Epidemiology, Evidence based practice, Respiratory medicine

SLEEP MEDICINE, CLINICAL PHARMACOLOGY, Sleep medicine < Keywords: ANAESTHETICS

on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 54 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 Pharmacologic interventions to improve sleep in hospitalized 5 6 adults: A systematic review 7 8 *Salmaan Kanji, Pharm D. 9 Departments of Pharmacy and Critical Care, The Ottawa Hospital 10 11 The Ottawa Hospital Research Institute 12 13 *Alexandru Mera, B.Sc. Pharm., ACPR 14 Department of Pharmacy, Hôpital Montfort - The Ottawa Hospital 15 For peer review only 16 Brian Hutton, MSc, PhD 17 18 Ottawa Hospital Research Institute 19 University of Ottawa School of Epidemiology, Public Health and Preventative Medicine 20 21 Lisa Burry, Pharm D. 22 Department of Pharmacy, Mount Sinai Hospital 23 24 University of Toronto 25 26 Erin Rosenberg, BSc., MD, MHA, FRCP(c) 27 Department of Critical Care, The Ottawa Hospital 28 29 Erika MacDonald, B.Sc. Pharm., ACPR, M.Sc, RPh 30 31 Department of Pharmacy, The Ottawa Hospital 32 The Ottawa Hospital Research Institute

33 http://bmjopen.bmj.com/ 34 Vanessa Luks 35 BASc, MD, FRCP(c), 36 37 Department of Respirology, Thunder Bay Regional Health Sciences Centre, 38 Northern Ontario School of Medicine 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 *Dr. Kanji and Mr. Mera contributed equally to this work and would like to be recognized as co- 43 44 primary authors. Dr. Kanji is the corresponding author. 45 46 47 Corresponding Author: Salmaan Kanji 48 The Ottawa Hospital 49 50 501 Smyth Rd 51 Ottawa, Ontario, Canada 52 K1H 8L6 53 [email protected] 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 54

2 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 Abstract 7 8 Objectives: Patients often suffer from disturbed sleep in hospital. Poor quality sleep in 9 10 11 hospitalized patients has been associated with significant morbidity and pharmacological sleep 12 13 aids are often prescribed. The objective of this systematic review is to evaluate the comparative 14 15 efficacy and safetyFor of pharmacologic peer interventions review used for sleep only in hospitalized patients. 16 17 18 19 Setting/Participants: We searched Medline, EMBASE, the Cochrane database and grey 20 21 literature for prospective studies that evaluated sleep in hospitalized adults after a pharmacologic 22 23 24 intervention. 25 26 27 28 Primary and Secondary Outcome Measures: Two reviewers assessed studies for inclusion and 29 30 31 extracted data for efficacy outcomes including sleep efficiency, sleep latency, sleep 32

33 fragmentation and objectively measured sleep stage distribution. Risk of bias was assessed and http://bmjopen.bmj.com/ 34 35 meta-analyses were planned contingent upon homogeneity of the included studies. 36 37 38 39 Results: After screening 1920 citations, 15 studies involving 861 patients were included. 40

41 Medications studied included benzodiazepines, nonbenzodiazepine sedatives, melatonin, on September 30, 2021 by guest. Protected copyright. 42 43 44 propofol, and dexmedetomidine. Five studies were deemed to be of high quality. Heterogeneity 45 46 and variable outcome reporting precluded meta-analysis in most cases. No consistent trends with 47 48 respect to sleep efficiency, quality or interruptions were observed identifying one drug or drug 49 50 51 class as superior to another or no treatment. Benzodiazepines appeared to be better than no 52 53 treatment with respect to sleep latency. Sleep stage distribution shows that sleep in hospital is 54 55 dominated by stage N1 and N2. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 54 BMJ Open

3 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Conclusion: There is insufficient evidence to suggest that pharmacotherapy improves the quality 4 5 6 or quantity of sleep in hospitalized patients suffering from poor sleep. No drug class or specific 7 8 drug was identified as superior even when compared to placebo or no treatment except for the 9 10 11 outcome of sleep latency, where benzodiazepines in general appear to outperform placebo or no 12 13 treatment. Although 15 studies were included, the quality of evidence was limited by both their 14 15 quality and size.For Larger, better-designed peer trials review in hospitalized adults only are needed. 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 54

4 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Article Summary: 4 5 6 • The purpose of this systematic review was to synthesize and summarize the available 7 8 evidence describing the efficacy and safety of drug therapy for sleep in acutely ill adults 9 10 11 in hospital. 12 13 • Our search of the literature identified 15 studies involving 861 patients and investigating 14 15 a varietyFor of sedatives peer and hypnotics review from different drug classes.only 16 17 18 • Studies were typically small and only a minority utilized an objective measure of sleep 19 20 such as polysomography. While it appears that drug therapy may reduce sleep latency it 21 22 23 does not improve the quality of restorative sleep in hospitalized patients. 24 25 26 27 Strengths and limitations of this study 28 29 30 • The strengths of this review included the use of a validated systematic search strategy of 31 32 multiple databases to identify relevant studies. Two investigators performed article

33 http://bmjopen.bmj.com/ 34 screening and data extraction and the research team consisted of physicians, pharmacists, 35 36 37 an epidemiologist/methodologist, a sleep expert and pharmacotherapy specialists. 38 39 • The main limitations of this review are related to the quality of the studies included. The 40

41 on September 30, 2021 by guest. Protected copyright. 42 studies included were typically small and outcomes were often measured subjectively. 43 44 Not all drugs currently being in clinical settings were identified in clinical trials and 45 46 comparative trials were identified without placebo controls for many drugs. Most studies 47 48 49 did not have a standard approach to the evaluation of safety and thus there was 50 51 insufficient data to adequately assess the safety of drug therapy for sleep in hospitalized 52 53 patients. 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 54 BMJ Open

5 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Key Messages: 4 5 6 • Sleep disruption and poor quality sleep is common among hospitalized adults with acute 7 8 illness 9 10 11 • Despite the fact that pharmacotherapy is commonly prescribed in hospital to improve 12 13 sleep there is no evidence that sleep quality or quantity is actually improved with drug 14 15 therapy For peer review only 16 17 18 • The available literature inadequately describes the safety of the drugs commonly 19 20 prescribed for sleep and thus the risk benefit ratio of this common practice is not clearly 21 22 23 in favor of benefit. 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 54

6 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Background 4 5 6 Sleep disturbances including difficulty initiating sleep, fragmented and non-restorative 7 8 sleep occur in up to 50% of acutely ill hospitalized patients.1 Sleep deprivation has been 9 10 11 associated with poor wound healing, memory disorders, delayed ventilator weaning, delirium 12 2-5 13 and mortality. Given the frequency of sleep disorders experienced by hospitalized patients and 14 15 their associated outcomes,For interventionspeer to reviewimprove sleep are frequently only attempted.1,6,7 Non- 16 17 18 pharmacologic interventions such as relaxation techniques, regulation of light/noise exposure are 19 20 often tried to counter new onset sleep disorders related to the environment.1,7 However, a 21 22 systematic review of non-pharmacologic interventions for sleep compared to no intervention did 23 24 8 25 not confirm a positive effect, citing a lack of high quality studies. Accordingly, sedative and 26 27 hypnotic pharmacotherapy is used in up to 88% of hospitalized patients.9 28 29 Currently there are no evidence-based recommendations or guidelines to direct the choice of 30 31 32 hypnotic drugs in the hospital setting, yet there are a plethora of agents that are used. The

33 http://bmjopen.bmj.com/ 34 objective of this systematic review was to identify, synthesize, and summarize the existing 35 36 37 evidence on the efficacy and safety of pharmacologic interventions used to improve sleep in 38 39 hospitalized adults. 40

41 on September 30, 2021 by guest. Protected copyright. 42 Methods 43 44 45 A protocol was developed prior to initiation of the review and was followed throughout the 46 47 review process. The protocol is available from the authors electronically upon request. 48 49 50 Selection of Studies 51 52 The Population-Intervention-Comparator-Outcome-Study design framework was used to 53 54 formulate the research question for this systematic review. The population of interest was 55 56 57 hospitalized adults; therefore we excluded non-hospitalized patients (i.e., healthy volunteers, 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 54 BMJ Open

7 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 outpatients, long-term care facility residents). We also excluded studies conducted solely in 4 5 6 patients diagnosed with a primary psychiatric illness or sleep apnea as these patient populations 7 8 exhibit different baseline sleep patterns irrespective of environment.1 Interventions of interest for 9 10 11 this systematic review included any pharmacologic intervention for sleep used alone or in 12 13 combination. This included: benzodiazepines, antidepressants, antihistamines, imidazopyridines 14 15 (e.g., zolpidem),For pyrazolopyrimidines peer (e.g., review zaleplon), cyclopyrrolones only (e.g., zopiclone ), 16 17 18 anticonvulsants (pregabalin/gabapentin), barbiturates, alpha-2-adrenergic agonists, 19 20 antipsychotics, melatonin or melatonin receptor agonists and natural health products. Studies 21 22 involving drugs that are no longer commercially available in any country were excluded. No 23 24 25 restrictions on dose for any agent were employed. Comparators of interest included placebo, no 26 27 treatment, or any other intervention. 28 29 Efficacy outcomes of interest were sleep efficiency, sleep latency, sleep interruptions and 30 31 32 sleep stage distribution measured by any means including polysomnography, bispectral index

33 http://bmjopen.bmj.com/ 34 monitoring (BIS), actigraphy, sleep questionnaires for patient self-reporting or direct observation 35 36 37 by a third party. However, studies solely relying on patient self-reporting for sleep outcomes of 38 10 39 interest were excluded. Sleep efficiency is defined as hours spent asleep divided by the total 40

41 hours of observation while sleep latency represents the duration of time taken to fall asleep. on September 30, 2021 by guest. Protected copyright. 42 43 44 Sleep fragmentation is defined as the number of arousals or awakenings in a designated sleep 45 46 period. Sleep stage distribution refers to the proportion of time a patient spends in each sleep 47 48 stage in a given period of observation. Secondary clinical outcomes included length of stay, 49 50 51 duration of mechanical ventilation, presence of delirium, occurrence of infection, cognitive 52 53 function, pain, glucose control, hypertension, myocardial infarction and mortality. Safety 54 55 outcomes of interest included overall occurrence of adverse events, serious adverse events 56 57 58 including those deemed life threatening, and withdrawal or discontinuation of therapy due to 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 54

8 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 adverse events. Study design comprised randomized controlled trials of pharmacological 4 5 6 interventions for sleep. Additionally, comparative cohort studies and prospective case series 7 8 were included only if sleep was evaluated using polysomnography. The language of publication 9 10 11 11 was restricted to English or French. 12 13 14 Literature Search Strategy 15 For peer review only 16 Relevant studies were identified using a sequential search approach. Initially, an informal 17 18 19 scoping exercise was conducted to identify potentially relevant trials in Ovid Medline with key 20 21 words such as “sleep”, “insomnia”, “drug therapy” and “in-patient”. Relevant studies identified 22 23 24 in the scoping exercise were provided to an information specialist to develop and validate a 25 26 search strategy for OVID Medline, Embase and Cochrane from inception to March 21, 2016. 27 28 The search strategy was peer-reviewed using PRESS (Peer Review of Electronic Search 29 30 31 Strategies) by a second information specialist and validated using the studies identified in the 32 12 33 scoping exercise. The detailed search strategy is provided in Appendix 1. In addition, the http://bmjopen.bmj.com/ 34 35 reference lists of included studies and reviews were screened. 36 37 38 Grey literature was searched via conference proceedings and abstracts (Canadian Sleep 39 40 Society, American Society of Sleep Medicine and European Sleep Research Society) for the last

41 on September 30, 2021 by guest. Protected copyright. 42 5 years, as well as clinical trial and systematic review registries, and Google Scholar®.13 43 44 45 46 Study selection 47 48 Titles and abstracts were independently reviewed for eligibility by two investigators (AM 49 50 51 and SK). Full-text articles identified as potentially eligible at first level screening were obtained 52 53 and reviewed for eligibility independently by the same two reviewers. Discrepancies were 54 55 resolved by discussion and consensus. The process of study selection was documented using a 56 57 14 58 PRISMA flow diagram. 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 54 BMJ Open

9 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 Data extraction and risk of bias assessment 8 9 Study level data was extracted using a pre-designed and piloted data collection form by 10 11 12 one author (AM) and checked by a second (SK). Means and standard deviations were estimated 13 14 using established methods when only medians and ranges were available from an included 15 For peer review only 16 study.15,16 In addition to the clinical and safety outcomes, extracted study data included the 17 18 19 following: study authors, year and journal of publication, country of study performance, funding 20 21 source, group sample sizes, study inclusion criteria, age distribution, gender distribution, patient 22 23 24 setting in hospital, and relevant concomitant medication use. Risk of bias was assessed using the 25 17 18 26 Cochrane Risk of Bias tool for randomized controlled trials and the Newcastle-Ottawa Scale 27 28 for cohort studies. Risk of bias was not assessed for prospective case series, as the risk of bias is 29 30 31 inherently high. 32

33 http://bmjopen.bmj.com/ 34 Synthesis of available studies 35 36 37 Clinical and methodological heterogeneity of the included studies was carefully 38 39 considered to determine whether meta-analysis would be conducted. Clinical heterogeneity was 40

41 assessed by comparing study populations, interventions and outcome measurements between on September 30, 2021 by guest. Protected copyright. 42 43 19 44 studies. Methodological heterogeneity was assessed by comparing methods for intervention 45 46 assignment, allocation concealment and blinding, as well as the extent of and reasons for loss to 47 48 follow-up. If heterogeneity was not identified between trials and reporting of clinical outcomes 49 50 51 was amenable to meta-analysis, a random-effects model was employed to calculate pooled risk 52 53 ratios for dichotomous outcomes (i.e., adverse events) and pooled mean differences for 54 55 continuous outcomes (i.e., sleep efficiency) using RevMan version 5.3 (the Nordic Cochrane 56 57 20,21 58 Center, the Cochrane Collaboration, Copenhagen, Denmark) . Studies were to be pooled by 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 54

10 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 drug and by drug class and compared with active controls or placebo separately depending on the 4 5 6 availability of data. In the event that 10 or more studies were pooled for any outcome, funnel 7 8 plots would be created and visually inspected to assess for publication bias.17 9 10 11 The PRISMA statement and its 27-item checklist were used in developing the summary 12 14 13 of findings for this review. A completed PRISMA checklist is provided in Appendix 2. 14 15 For peer review only 16 17 18 Results 19 20 Characteristics of included trials 21 22 Our search strategy identified a total of 1,920 citations, from which 11 randomized 23 24 25 controlled trials, two prospective cohort studies and two case series, totaling 861 patients, met 26 27 criteria for this review (Figure 1). Eight studies were conducted in the intensive care unit and 28 29 included patients requiring mechanical ventilation. Seven studies were conducted in acute care 30 31 32 hospital wards and included medical and surgical patients suffering from inadequate sleep. Six

33 http://bmjopen.bmj.com/ 34 studies measured sleep using polysomnography, two used BIS and the remaining measured sleep 35 36 37 using either third party observation or a combination of third party observation and patient self- 38 39 report (Tables 1 and 2). Medications studied included benzodiazepines, melatonin, propofol, 40

41 zolpidem, zopiclone and dexmedetomidine. In most studies the intervention was administered in on September 30, 2021 by guest. Protected copyright. 42 43 44 the evening and sleep was evaluated over night, however two of the ICU based studies evaluated 45 46 sleep while on continuous infusions of sedatives. Study durations ranged from 24 hours to 14 47 48 days. Seven of the randomized controlled trials had a high risk of bias in at least one criteria, 49 50 51 which impacted their overall quality (Figure 2). The prospective cohort studies included were of 52 53 high quality (Figure 3). 54 55 56 57 58 Sleep efficiency 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 54 BMJ Open

11 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Sleep efficiency was evaluated in seven studies (5 randomized controlled trials22-26 and 4 5 27,28 26,27 6 two cohort studies ), of which two identified statistically significant differences (Table 3, 7 8 Figure 4). Alexopoulou et al evaluated sleep with polysomnography in 13 of 16 enrolled patients 9 10 27 11 over the course of three nights in a prospective cross-over cohort study. On nights 1 and 3 12 13 patients received no treatment, while on night 2 they received dexmedetomidine. 14 15 PolysomnographyFor was not consistentlypeer viable, review and only ten patients only contributed to the reported 16 17 18 analysis, which showed that sleep efficiency was improved with dexmedetomidine (77.9% [SD: 19 20 65.6 - 80.2%]) as compared to no treatment (15.8% [6.4 – 51.6%], p=0.002). In a study by Oto et 21 22 al (2011), two different dosing strategies for midazolam were compared in a randomized 23 24 26 25 controlled trial of 22 critically ill patients. Both groups received sedation with a continuous 26 27 infusion of midazolam, but one group had their infusions interrupted during the day. Sleep 28 29 efficiency as measured by polysomnography was statistically significantly improved (p=0.047) 30 31 32 in the group that did not have their infusion interrupted during the day (97% [91.1 – 97.8%]

33 http://bmjopen.bmj.com/ 34 versus 81% [73.3 – 91.1%]). In a similar trial of 18 patients by Hardin et al, critically ill adults 35 36 37 were administered lorazepam via continuous infusion or an interrupted dosing strategy; unlike in 38 39 the previous study, differences between groups were not statistically significant (Table 3, Figure 40 28 41 4). on September 30, 2021 by guest. Protected copyright. 42 43 et al et al 44 Two randomized controlled trials (Bourne and Ibrahim ) that compared 45 46 melatonin with placebo failed to show a statistically significant difference in sleep efficiency as 47 48 measured by BIS or nurse observation.22,23 Similarly, two trials by Engelmann et al and Kondili 49 50 25 24 51 et al comparing propofol to flunitrazepam and no treatment did not show statistically 52 53 significant differences with respect to sleep efficiency as measured by BIS and 54 55 polysomnography, respectively. Only the two placebo controlled melatonin trials (total of 56 56 57 58 patients) were amenable to pooling of this outcome that revealed a non-statistically significant 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 54

12 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 trend in favor of melatonin as compared to placebo (mean difference of 8.0%, [95% confidence 4 5 6 interval: -1.5, 17.5]). 7 8 9 10 11 Sleep latency 12 13 Sleep latency was evaluated in five studies. Due to variability in assessment and 14 15 reporting metricsFor between studies,peer forest plotsreview were not created, onlyand the appropriateness of meta- 16 17 18 analysis was deemed low. Sleep latency data is reported for each study in Table 3. 19 20 Two randomized controlled trials, Feldmeier (n=50) and Gallais (n=51),29,30 compared 21 22 the effects of midazolam and oxazepam to placebo. One randomized controlled trial by 23 24 29 25 Lupolover et al (n=78) compared midazolam to oxazepam. Sleep latency was measured by 26 27 third party observation or patient questionnaire in all three studies. The studies by Feldmeier et al 28 29 and Gallais et al reported a statistically significant reduction in sleep latency with midazolam (15 30 31 32 mg orally) compared to placebo (median difference of 30-53%, p<0.05), however only

33 http://bmjopen.bmj.com/ 34 Feldmeier et al found oxazepam (15 mg orally) to be superior to placebo (median difference 35 36 29 et al 37 16%, p<0.05). Gallais found midazolam 15 mg orally to be superior to oxazepam 50 mg 38 30 39 orally (Table 3) as did Lupolover et al (15 mg doses for both midazolam and oxazepam), but 40 31 41 only after 3 days of treatment (table 3). on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 Two randomized control trials evaluated sleep latency with triazolam.32,33 Morgan et al 47 48 (n=357) compared triazolam to zolpidem and placebo, while Goetzke et al (n=79) compared it to 49 50 51 brotizolam and placebo. In both cases, drug therapy (any) appeared to be superior to placebo 52 53 with respect to sleep latency, however, no statistical analysis was provided by Goetzke et al.33 54 55 No statistically significant differences were identified between the active arms of either study 56 57 58 (Table 3). 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 54 BMJ Open

13 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 Sleep fragmentation 12 24-26 13 Sleep fragmentation was evaluated in five studies: three randomized controlled trials 14 15 and two cohort studiesFor27,28 (Tablepeer 3, Figure review 5). Variability in treatments only studied precluded the 16 17 18 opportunity for meta-analysis. Two randomized controlled trials studied the effect of propofol 19 20 on the number of awakenings compared to flunitrazepam (Engelmann et al)25 or no treatment 21 22 (Kondili et al)24. Engelmann et al (n=66), using BIS, found that propofol (median: 0.6 arousals 23 24 25 per hour of sleep) was associated with statistically significantly fewer sleep interruptions than 26 27 flunitrazepam (median: 1.1 arousals per hour of sleep in critically ill patients, p=0.041).25 28 29 However, Kondili et al (n=12), using polysomnography, found insufficient evidence of a 30 31 32 difference between patients receiving propofol (median [interquartile range]: 4.8 [1.3 - 14.6]

33 http://bmjopen.bmj.com/ 34 awakenings per hour of sleep) and those receiving no treatment (median [interquartile range]: 35 36 24 37 8.1 [2.9 – 16.2] awakenings per hour of sleep). 38 39 Dexmedetomidine compared to no treatment was found to statistically significantly 40

41 reduce the number of sleep interruptions in critically ill patients. Using polysomnography, on September 30, 2021 by guest. Protected copyright. 42 43 et al 44 Alexopoulou (n=23) found that patients receiving infusions of dexmedetomidine had a 45 46 median of 2.2 (interquartile range: 1.6-4.5) awakenings per hour while those receiving no 47 48 treatment had a median of 7.1 (interquartile range: 1.6-4.5) awakenings per hour (p=0.0023).27 49 50 51 Two trials comparing different dosing strategies for benzodiazepines in critically ill 52 53 patients yielded conflicting results. Oto et al (2011) found that continuous infusions of 54 55 midazolam were associated with fewer awakenings when compared to infusions incorporating 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 54

14 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 daytime interruptions, while Hardin et al found no difference between continuous infusions of 4 5 26,28 6 lorazepam (with and without a neuromuscular blocker) and an intermittent dosing strategy. 7 8 9 10 11 12 13 Sleep stage distribution 14 15 Two randomizedFor control peer trials,24,26 reviewtwo cohort studies27,28 only and two case series34,35 16 17 18 evaluated sleep architecture using polysomnography in a total of 74 patients. Treatment effects 19 20 from comparative studies are displayed in Figure 6, while data for all studies is found in Table 3. 21 22 Irrespective of treatment allocation, patients in hospital primarily experience light non- 23 24 25 restorative sleep, corresponding to stages N1 and N2. Only Hardin et al (continuous infusion of 26 27 lorazepam with or without neuromuscular blocker versus interrupted infusion) reported a 28 29 proportion of sleep as Stage N3 (restorative sleep), where Stage N3 represented 32-50% of total 30 31 32 sleep but insufficient evidence of a difference between groups. No other trial using

33 http://bmjopen.bmj.com/ 34 polysomnography was able to describe meaningful amounts of stage N3 sleep. REM sleep was 35 36 37 only described with midazolam infusions interrupted during the day (6.8%), continuous infusions 38 39 of midazolam (2%), interrupted lorazepam infusions (3.6%), and the no treatment arm of the 40 24,26,28,34 41 study by Kondili et al (1.4%). Given the variability in studied treatments, study designs on September 30, 2021 by guest. Protected copyright. 42 43 44 and the limited number of overall studies meta-analysis was not attempted. 45 46 47 48 Secondary outcomes 49 50 51 Delirium was not assessed in the included studies, however Ibrahim et al evaluated 52 53 agitation using the Riker Sedation-Agitation Scale and found no difference between melatonin 54 55 (n=14) and placebo groups (n=18).23 Cognitive function was evaluated in the study by Li Pi 56 57 58 Shan et al using the Mini Mental State Exam and no difference was found between lorazepam 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 54 BMJ Open

15 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 (n=9) and zopiclone (n=9).36 Narcotic requirements were evaluated in the study by Oto et al. 4 5 6 While more patients (9/11=81.8%) in the group assigned to midazolam infusions with daily 7 8 interruption required opioids than those assigned to the continuous infusion group (4/11=36.4%), 9 10 26 et al 11 there was no difference in median dose between groups. In the study by Hardin (n=18), 12 13 patients assigned to continuous infusions of lorazepam (n=6) required significantly more 14 15 morphine (0.48 ±For 0.52 mg/kg/day) peer than those review assigned to interrupted only infusions (0.11 ± 0.13 16 17 18 mg/kg/day) (n=6) or continuous infusions with neuromuscular blockade (0.14 ± 0.32 mg/kg/day) 19 20 (n=6) (p=0.031). No other differences were found between groups.28 21 22 23 Duration of mechanical ventilation, length of stay, occurrence of infection, glucose 24 25 control, hypertension, myocardial infarction and mortality were not described in any included 26 27 study. 28 29 30 31 32 Safety

33 http://bmjopen.bmj.com/ 34 No study reported occurrence of serious life threatening events. Seven of the 15 studies 35 36 22,25,29-33,36 37 did not report any safety information in the methods or results. Of the remaining eight, 38 39 no study reported any difference with respect to adverse events between the active treatment 40

41 on September 30, 2021 by guest. Protected copyright. 42 groups compared to the placebo or no treatment arms. 43 44 Engelmann et al reported that three patients developed respiratory depression after 45 46 flunitrazepam administration requiring intervention the insertion of a nasopharyngeal airway but 47 48 25 49 not discontinuation from study. One patient receiving dexmedetomidine in the study by 50 51 Alexopoulou et al had the study drug discontinued for bradycardia (heart rate less than 50 beats 52 53 per minute).27 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 54

16 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Patient conditioning, defined as somnolence or alertness the morning following 4 5 6 administration of hypnotics was reported in five studies. Four reported no differences between 7 8 drugs and/or placebo.29,32,33,36 The study by Lupolover et al described an improved state on 9 10 31 11 awakening by patient self-report in those taking midazolam compared to oxazepam. 12 13 14 15 HeadacheFor and nausea peer were the most review commonly reported onlyadverse events. Discontinuation 16 17 et al 32 18 of therapy due to nausea was only observed in the placebo group of Morgan . Allergic 19 20 type skin reactions were reported in two patients - one was associated with triazolam; the other, 21 22 described as pruritis, occurred in a patient with pancreatitis and whether the reaction was 23 24 31,33 25 attributed to benzodiazepines or placebo was not specified in the study. 26 27 28 29 30 31 32 Discussion

33 http://bmjopen.bmj.com/ 34 Hospitalized patients routinely suffer from sleep disorders attributed to both 35 36 37,38 37 pathophysiological and environmental factors. In a study of 280 hospitalized elderly patients, 38 39 21% reported new-onset insomnia, 38% reported moderate or severe insomnia, and 38% reported 40 39 41 sleep disturbances during admission. As such, pharmacologic agents are commonly on September 30, 2021 by guest. Protected copyright. 42 43 9 44 administered by health providers to improve sleep, in up to 88% of hospitalized patients. We set 45 46 out to summarize the existing evidence and compare studies looking at the efficacy and safety of 47 48 various pharmacologic agents for sleep. 49 50 51 52 Fifteen studies encompassing 861 patients met our inclusion criteria, however clinical 53 54 and methodological heterogeneity and variable outcome reporting were judged to preclude 55 56 57 performance of reliable meta-analyses. An apparent reduction in sleep latency with 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 54 BMJ Open

17 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 pharmacotherapy when compared to placebo or no treatment has been demonstrated, however 4 5 6 sleep efficiency and sleep fragmentation were not shown to be improved by pharmacologic 7 8 therapy when compared to placebo or no treatment. This is in keeping with expert opinion that 9 10 11 pharmacologic agents, while somewhat effective for sleep onset insomnia, are ineffective for 12 40 13 maintaining sleep. 14 15 Furthermore,For sleep architecturepeer in hospital,review as measured onlyby polysomnography, is 16 17 18 dominated by stage N1 and N2 sleep. Restorative N3 sleep or REM sleep was found to be rare 19 20 regardless of pharmacologic intervention. When sleep is highly fragmented, it is common to 21 22 observe a predominance of N1 sleep, as an arousal in an epoch of sleep on the EEG results in the 23 24 41 25 sleep stage being scored back to stage N1. Stage N3 sleep was observed only in a single study, 26 27 in patients receiving continuous or intermittently administered lorazepam.28 A paucity of N3 28 29 sleep is not unusual in middle-aged or elderly populations, where N3 typically constitutes only 5 30 31 42 32 to 10% of sleep. Furthermore, the sleep stages are influenced by the medications studied.

33 http://bmjopen.bmj.com/ 34 Benzodiazepines, for example, are potent suppressors of N3 sleep, and increase the presence of 35 36 43 37 sleep spindles. Spindles are the hallmark of N2 sleep, so it is common to see increased N2 38 39 sleep and decreased N3 sleep with benzodiazepines. Concurrent opioids will also decrease N3 40 44 41 sleep. The large percentage of N3 sleep seen in the study in which lorazepam was used is on September 30, 2021 by guest. Protected copyright. 42 43 28 44 surprising given the intervention is known to suppress N3 sleep. Furthermore, this finding is 45 46 not replicated in any other studies. A possible explanation is that N3 sleep is commonly over- 47 48 scored. REM sleep was only observed in four studies, independent of drug or placebo allocation. 49 50 51 REM sleep can also be influenced by current and prior medications. There is evidence to suggest 52 53 that benzodiazepines decrease REM sleep.45 Sleep stage distribution is also influenced by prior 54 55 sleep episodes and medications. For example, recent sleep deprivation will increase N3 sleep, 56 57 46 58 and REM sleep. Lastly, recent medication use prior to the polysomnography recording may 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 54

18 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 influence sleep stages (i.e., recent discontinuation of REM-suppressant such as a barbiturate, 4 5 6 anti-depressant, or alcohol, will result in REM rebound). In summary, the disappointing 7 8 performance of the sedatives in effecting greater amounts of N3 and REM is in part due to the 9 10 11 well-known effect of these drugs on sleep stages. In addition, the medications administered and 12 13 discontinued, as well as prior sleep during the run-in period can influence sleep architecture, and 14 15 although some ofFor the trials werepeer randomized, review it is possible that theseonly factors could be 16 17 18 confounding the results given the small sample sizes. 19 20 21 The lack of high-quality randomized controlled trials is one of the limitations of this 22 23 24 review. The diversity of outcome reporting and the high degree of methodologic heterogeneity 25 26 prevented statistical pooling of the results in almost all circumstances. Major sources of clinical 27 28 heterogeneity included large variability in drug dosing (i.e., oxazepam dosing ranged from 15 to 29 30 31 50 mg per dose) and the method by which sleep was evaluated. Only six of the 15 studies 32

33 included used polysomnography which is considered the gold standard for sleep evaluation. http://bmjopen.bmj.com/ 34 35 Only two of these were randomized controlled trials, which speaks to the practical difficulty in 36 37 38 using polysomonography in large trials and sick patients. In fact, 28 potentially eligible trials 39 40 were excluded at second level screening because the sole measure of sleep outcomes relied on

41 on September 30, 2021 by guest. Protected copyright. 42 patient self-reporting. We elected to exclude these studies as self-reporting has been 43 44 45 demonstrated to correlate poorly with objective assessments and is associated with systematic 46 47 bias.10 Furthermore, the general lack of safety outcome reporting does not allow a thorough 48 49 50 evaluation of risk. Given paucity of convincing efficacy data, the relatively high potential for 51 52 adverse events with these classes of drugs and the vulnerability of the population being studied 53 54 future studies must systematically evaluate risk in order for clinicians to adequately assess the 55 56 57 risk/benefit relationship of these interventions. Language restriction to English and French may 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 54 BMJ Open

19 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 also be considered a limitation although others have shown that this restriction may have little to 4 5 11 6 no impact on findings. 7 8 It is also important to recognize the paucity of placebo controlled trials in this setting. An 9 10 11 evidence based approach to clinical trial design would suggest that placebo controlled trials that 12 13 evaluate the efficacy of drug therapies be conducted in advance of comparative trials. While 14 15 placebo controlledFor trials are peeravailable in healthy review volunteers and otherwiseonly healthy community 16 17 18 based patients, the etiology of sleep disturbances is not the same as that of the acutely ill patient 19 20 sleeping in a hospital bed. Similarly the outcomes of such trials cannot be extrapolated to this 21 22 population either. Given that the majority of hospitalized patients receive pharmacologic sleep 23 24 25 aids in hospital in the absence of evidence to suggest that they are helpful and safe, well- 26 27 designed trials are needed. Future studies must be randomized, employ a placebo arm, use 28 29 standardized definitions of sleep related outcomes and evaluate sleep using objective measures 30 31 32 such as polysomnography.

33 http://bmjopen.bmj.com/ 34 35 Conclusions 36 37 38 While sleep is universally recognized as an essential body function to enable the 39 40 restoration of physiologic systems, sleep disturbances are common in hospitalized patients. A

41 on September 30, 2021 by guest. Protected copyright. 42 variety of pharmacologic agents are utilized in the majority of these patients with the 43 44 45 presumption that they improve the quality and quantity of sleep, however this review suggests 46 47 that the low to moderate quality evidence available does not support this notion. While drug 48 49 50 therapy may be associated with falling asleep faster (sleep latency) there is no evidence that drug 51 52 therapy improves the more important outcomes of sleep efficiency or quality. This may be due to 53 54 the fact that drug therapy is truly ineffective or that the quality of existing trials is inadequate to 55 56 57 show a difference. It is evident that larger, better quality trials are needed to definitively answer 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 54

20 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 this question. Future trials must use standardized definitions for sleep related outcomes, 4 5 6 objectives methods for measurement, a placebo arm and systematically evaluate risk. At this 7 8 time there is insufficient evidence to support the current usage of pharmacotherapy for the 9 10 11 treatment of sleep disturbances among hospitalized adults. 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

21 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Contributorship Statement: 4 5 6 All authors (SK, AM, BH, LB, ER, EM, VL) contributed to the study protocol. SK and AM 7 8 were responsible for abstract screening, study selection and data extraction. BH, SK and AM 9 10 11 performed the analysis and all authors contributed to the interpretation of results and writing of 12 13 the final manuscript. 14 15 For peer review only 16 17 18 Acknowledgements: 19 20 The authors would like to acknowledge the Risa Shorr MLS, information specialist, for her 21 22 assistance in developing and validating our search strategy. We would also like to acknowledge 23 24 25 Dr. Dean Fergusson for reviewing the manuscript. 26 27 28 29 Competing Interests: 30 31 32 All authors confirm that they have no competing interests to declare.

33 http://bmjopen.bmj.com/ 34 35 36 37 Funding: 38 39 None 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 Data Sharing Statement 45 46 All data was extracted from previously published papers which are referenced in the manuscript. 47 48 No additional or unpublished data is available. 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 54

33 sleep duration: how similar are they? Epidemiology 2008;19:838-45. http://bmjopen.bmj.com/ 34 11. Morrison A, Polisena J, Husereau D, et al. The effect of English-language restriction on 35 systematic review-based meta-analyses: a systematic review of empirical studies. International 36 37 journal of technology assessment in health care 2012;28:138-44. 38 12. Sampson M, McGowan J, Cogo E, Grimshaw J, Moher D, Lefebvre C. An evidence- 39 based practice guideline for the peer review of electronic search strategies. Journal of clinical 40 epidemiology 2009;62:944-52.

41 13. Balshem H, Stevens A, Ansari M, et al. Finding Grey Literature Evidence and Assessing on September 30, 2021 by guest. Protected copyright. 42 for Outcome and Analysis Reporting Biases When Comparing Medical Interventions: AHRQ 43 44 and the Effective Health Care Program. Methods Guide for Effectiveness and Comparative 45 Effectiveness Reviews. Rockville (MD)2008. 46 14. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for 47 systematic reviews and meta-analyses: the PRISMA statement. PLoS medicine 48 2009;6:e1000097. 49 50 15. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, 51 range, and the size of a sample. BMC medical research methodology 2005;5:13. 52 16. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation 53 from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 54 2014;14:135. 55 17. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for 56 57 assessing risk of bias in randomised trials. Bmj 2011;343:d5928. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 54 BMJ Open

23 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 18. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in 4 5 meta-analyses. 2013. (Accessed 28 Jan, 2015, at 6 http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp.) 7 19. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in 8 medicine 2002;21:1539-58. 9 20. SW R. Analyzing Effect Sizes: Random Effects Models: New York: Russell Sage 10 11 Foundation; 2009. 12 21. Viechtbauer W. Bias and Efficiency of Meta-Analytic Variance Estimators in the 13 Random-Effects Model. Journal of Educational and Behavioral Statistics 2005;30:261-93. 14 22. Bourne RS, Mills GH, Minelli C. Melatonin therapy to improve nocturnal sleep in 15 critically ill patients:For encouraging peer results fromreview a small randomised only controlled trial. Critical care 16 2008;12:R52. 17 18 23. Ibrahim MG, Bellomo R, Hart GK, et al. A double-blind placebo-controlled randomised 19 pilot study of nocturnal melatonin in tracheostomised patients. Critical care and resuscitation : 20 journal of the Australasian Academy of Critical Care Medicine 2006;8:187-91. 21 24. Kondili E, Alexopoulou C, Xirouchaki N, Georgopoulos D. Effects of propofol on sleep 22 quality in mechanically ventilated critically ill patients: a physiological study. Intensive care 23 24 medicine 2012;38:1640-6. 25 25. Engelmann C, Wallenborn J, Olthoff D, Kaisers UX, Ruffert H. Propofol versus 26 flunitrazepam for inducing and maintaining sleep in postoperative ICU patients. Indian journal of 27 critical care medicine : peer-reviewed, official publication of Indian Society of Critical Care 28 Medicine 2014;18:212-9. 29 26. Oto J, Yamamoto K, Koike S, Imanaka H, Nishimura M. Effect of daily sedative 30 31 interruption on sleep stages of mechanically ventilated patients receiving midazolam by infusion. 32 Anaesthesia and intensive care 2011;39:392-400.

33 27. Alexopoulou C, Kondili E, Diamantaki E, et al. Effects of dexmedetomidine on sleep http://bmjopen.bmj.com/ 34 quality in critically ill patients: a pilot study. Anesthesiology 2014;121:801-7. 35 28. Hardin KA, Seyal M, Stewart T, Bonekat HW. Sleep in critically ill chemically paralyzed 36 37 patients requiring mechanical ventilation. Chest 2006;129:1468-77. 38 29. Feldmeier C, Kapp W. Comparative clinical studies with midazolam, oxazepam and 39 placebo. British journal of clinical pharmacology 1983;16 Suppl 1:151S-5S. 40 30. Gallais H, Casanova P, Fabregat H. Midazolam and oxazepam in the treatment of

41 insomnia in hospitalized patients. British journal of clinical pharmacology 1983;16 Suppl on September 30, 2021 by guest. Protected copyright. 42 1:145S-9S. 43 44 31. Lupolover R, Ballmer U, Helcl J, Escher J, Pavletic B. Efficacy and safety of midazolam 45 and oxazepam in insomniacs. British journal of clinical pharmacology 1983;16 Suppl 1:139S- 46 43S. 47 32. Morgan PJ, Chapados R, Chung FF, Gauthier M, Knox JW, Le Lorier J. Evaluation of 48 zolpidem, triazolam, and placebo as hypnotic drugs the night before surgery. Journal of clinical 49 50 anesthesia 1997;9:97-102. 51 33. Goetzke E, Findeisen P, Welbers IB. Comparative study on the efficacy of and the 52 tolerance to the triazolodiazepines, triazolam and brotizolam. British journal of clinical 53 pharmacology 1983;16 Suppl 2:407S-12S. 54 34. Kim SJ, Park J, Lee YJ, et al. The Optimal Dose of Midazolam for Promoting Sleep in 55 Critically Ill Patients: A Pilot Study. Korean J Crit Care Med 2014;29:166-71. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 54

24 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 35. Oto J, Yamamoto K, Koike S, Onodera M, Imanaka H, Nishimura M. Sleep quality of 4 5 mechanically ventilated patients sedated with dexmedetomidine. Intensive care medicine 6 2012;38:1982-9. 7 36. Li Pi Shan RS, Ashworth NL. Comparison of lorazepam and zopiclone for insomnia in 8 patients with stroke and brain injury: a randomized, crossover, double-blinded trial. American 9 journal of physical medicine & rehabilitation / Association of Academic Physiatrists 10 11 2004;83:421-7. 12 37. Pisani MA, Friese RS, Gehlbach BK, Schwab RJ, Weinhouse GL, Jones SF. Sleep in the 13 Intensive Care Unit. American journal of respiratory and critical care medicine 2015. 14 38. Bano M, Chiaromanni F, Corrias M, et al. The influence of environmental factors on 15 sleep quality in hospitalizedFor peer medical patients. review Frontiers in neurology only 2014;5:267. 16 39. Isaia G, Corsinovi L, Bo M, et al. Insomnia among hospitalized elderly patients: 17 18 prevalence, clinical characteristics and risk factors. Archives of gerontology and geriatrics 19 2011;52:133-7. 20 40. Rosenberg RP. Sleep maintenance insomnia: strengths and weaknesses of current 21 pharmacologic therapies. Annals of clinical psychiatry : official journal of the American 22 Academy of Clinical Psychiatrists 2006;18:49-56. 23 24 41. Berry RB, Budhiraja R, Gottlieb DJ, et al. Rules for scoring respiratory events in sleep: 25 update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. Deliberations 26 of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. Journal 27 of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep 28 Medicine 2012;8:597-619. 29 42. Achermann P, Borbély AA. Chapter 37 - Sleep Homeostasis and Models of Sleep 30 31 Regulation. In: Dement MHKRC, ed. Principles and Practice of Sleep Medicine (Fifth Edition). 32 Philadelphia: W.B. Saunders; 2011:431-44.

33 43. Poyares D, Guilleminault C, Ohayon MM, Tufik S. Chronic benzodiazepine usage and http://bmjopen.bmj.com/ 34 withdrawal in insomnia patients. Journal of psychiatric research 2004;38:327-34. 35 44. Dimsdale JE, Norman D, DeJardin D, Wallace MS. The effect of opioids on sleep 36 37 architecture. Journal of clinical sleep medicine : JCSM : official publication of the American 38 Academy of Sleep Medicine 2007;3:33-6. 39 45. Borbely AA, Mattmann P, Loepfe M, Strauch I, Lehmann D. Effect of benzodiazepine 40 hypnotics on all-night sleep EEG spectra. Human neurobiology 1985;4:189-94.

41 46. Brunner DP, Dijk DJ, Tobler I, Borbely AA. Effect of partial sleep deprivation on sleep on September 30, 2021 by guest. Protected copyright. 42 stages and EEG power spectra: evidence for non-REM and REM sleep homeostasis. 43 44 Electroencephalography and clinical neurophysiology 1990;75:492-9. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from sedatives, analgesics Concomitant

Fentanyl, morphine for pain, additional midazolam for “as needed” sedation 4 patients 4receiving patients antidepressants None None reported None None reported None None reported None None reported

Patient Characteristics ICU patientsICU with tracheostomy, weaning from mechanical GSCventilation, >9, no sedatives for 12 hours ventilated ICU adult > 48hsedatives Adults Adults with diagnosis eitherof mechanical mechanical ventilation for 2 nights Post operative sedation or mechanical ventilation requiring requiring mechanical andventilation tracheostomy to assist weaning patients, receivingpatients, patients patients without

unit Setting care hospital Critical careCritical unit Critical careCritical unit Mechanically Critical careCritical unit patients ICU with Surgical Surgical step down Rehab unit inunit Rehab acute

hours Study Study reported BMJ Open http://bmjopen.bmj.com/ 14 nights / 14not nights for first 5 for hours)first 4 nights / 94 hours nights care Critical unit patients ICU 24 / 7 hours hours 48 / 9 hours hours (BIS only evaluated(BIS only observation period At least 48At / 8 hours Duration/Treatment hours hours for interrupted sedation group, 24 hours for continuous sedation group Not reported / 9

b c c b b a

b b 44% male 61% male 57% male 33% male 82% male 64% male 50% male 88% male 72±9 years 58.3% male 87.5% male 68±11 years Age and Sex 60.2±13 years on September 30, 2021 by guest. Protected copyright. 73(63-75) years 69.9± 12 years 57 (46-68) years 59.9±11.0 years 63 (54-72) years 58.7± 12.5 years 56.6(20-78) years

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml atients P For peer review only

(N) Placebo (n=18) Placebo (n=12) Midazolam continuous infusion (n=11) Flunitrazepam (n=32) Lorazepam orally (n=10) Melatonin (n=14) Melatonin (n=12) Midazolam infusion daytime with interruption (n=11) Propofol infusion (n=7) Propofol infusion (n=34) (n=6) No Treatment No Treatment Author, year Intervention and Li Pi etShan Li al., 2004 Ibrahim et Ibrahim al., 2006 Bourne et 2008 al., Otoal., et 2011 Kondili Kondili et al., 2012 (cross over) Engelmann et al., 2014 Trials RandomizedControlled Study Characteristics: Table1: Page 25 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 26 of 54 26 26 None None Reported None None Reported None None Reported None None Reported None None reported insomnia seeking hypnotic sleep aid hypnotic hypnotic due to sleep difficulties insomnia elective surgeryelective stroke/ acquiredstroke/ secondary insomnia patients patients with patients patients requiringpatients a patients patients with patients patients undergoing brain injury, brain injury, Hospital ward Hospitalized Acute careAcute unit Hospitalized Acute careAcute unit Hospitalized Acute careAcute unit Hospitalized adult reported reported reported reported BMJ Open 5 nights / 5 nights not 5 nights / 5 nights not 7 nights / 7 nights not http://bmjopen.bmj.com/ 14 nights / 14not nights 24 / 8 hours hours care Acute unit Hospitalized

b b b

58% male 49% male 32% male 39% male 50% male 27% male 37% male 52.5% male 27±5 years 34±13 years 33±13 years on September 30, 2021 by guest. Protected copyright. Mean 51.2Mean years Mean 50.8Mean years Mean 53.3Mean years 48.9±12.7 years Range 19-71Range years For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

For peer review only deviation standard ± opiclone orally Z Oxazepam orallyOxazepam (n=38) Placebo (n=20) Oxazepam orallyOxazepam (n=20) Triazolam orallyTriazolam (n=79) Placebo (n=16) Oxazepam Oxazepam (n=15) Placebo (n=118) Zolpidem (n=120) Midazolam orally (n=40) Midazolam orally (n=19) Brotizolam Brotizolam orally (n=79) Midazolam orally (n=19) Triazolam (n=119) (n=10) data presented as mean as presented data data presented as median as (range) presented data data presented as median (interquartile range) median (interquartile as presented data CI) (95% mean as presented data

(cross over) IV – intravenous intravenous – IV unit care – ICU Intensive index Bispectral – BIS a c d Lupolover et al., 1983 Gallais etGallais al., 1983 Goetzke et al., 1983 (cross over) Feldmeier Feldmeier et al., 1983 Morgan 1997 al., et b 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 27 27 sedatives, analgesics Concomitant None None reported None None reported None None reported None None reported

Patient patients patients patients sedation ventilated ICU ventilated ICU ventilated ICU ventilated ICU Characteristics patients requiringpatients continuous sedation patients requiringpatients IV

Setting Critical careCritical unit Mechanically

Study Study BMJ Open http://bmjopen.bmj.com/ 24 / 9 hours hours care Critical unit Mechanically 57 / 9 hours hours 24 / 24 hours hours care Critical unit Mechanically 24 / 24 hours hours care Critical unit Mechanically observation period Duration/Treatment

b b b

b 83% male 33% male 17% male 30% male 80% male 77% male 68±9 years Age and Sex 43.2±13 years on September 30, 2021 by guest. Protected copyright. 61.3±8.8 years 58.5±9.8 years 62 (55-69) years 67 (40.5-72) years

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

Patients For peer review only Intervention and (N) Lorazepam IV (continuous sedation with neuromuscular (n=6) Lorazepam IV (continuous sedation) (n=6) Lorazepam IV (intermittent sedation) (n=6) Dexmedetomidine IV infusion (n=10 ) Midazolam IV infusion (n=5) Dexmedetomidine IV infusion (n=13) blockade) No treatment (n=13)No treatment design) design) ( Author, year Series Studies andCohort Case Study Characteristics: Table2:

Hardin et al., 2006 (Cohort) Otoal., et 2012 (Case Series) Kim et al.,Kim 2014 (Case Series) Alexopoulou etAlexopoulou al., 2014 (cohortwith cross over) Page 27 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 28 of 54 28 28

nd nd and 2 st

night Measure of Effect rd Correlation between dose of andmidazolam Total Sleep (r=0.975,Time p=0.005) Correlation between dosemidazolam and stage II (r=0.975, p=0.005) night ; p=0.001night 2 between and and 3

Sleep EfficiencySleep FragmentationSleep Sleep EfficiencySleep StageSleep Distribution I : Stage p=0.006 IIStage : p=0.006 IIIStage : p=0.180 REM : p=0.173 p=0.002 1 between p = 0.0023

a a

a b a a a group a a Sleep Efficiency Sleep Sleep Efficiency Sleep 16.1(7.6-28.6) arousals/hr 7.1 (6.1-13.4) arousals/hr 15.8% (6.4-51.6%) StageSleep Distribution (%) Sleep Efficiency Sleep 34.3% (8-59.7) StageSleep Distribution (min) I =82.0Stage (60.5-372.5) IIStage =88.0 (19.0-621.0) IIIStage =0.0 (0-0) REM=10.0 (6.0-50.5) FragmentationSleep 77.9% (65.6-80.2%) StageSleep Distribution(%) Stage I =56.2Stage (24.7-79.3) IIStage =39.2 (20.7-66.4) IIIStage =0.0 (0-0) REM=0.0 (0-0.4) FragmentationSleep Stage I=16.1Stage (6.2-21.3) IIStage =78.7 (69.2-92.5) IIIStage =0.0 (0-0) REM=0.0 (0-0.4) FragmentationsSleep 2.2 (1.6-4.5) arousals/hr Sleep Efficiency Sleep ion ion by Outcome treatment

BMJ Open http://bmjopen.bmj.com/ reatment on day 1, dexmedetomidine on day 2) Polysomnography Polysomnography (no t Polysomnography

on September 30, 2021 by guest. Protected copyright. a

a Treatment group Method of evaluat sleep For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml (n=16 enrolled, 13 analyzed) Propofol 2mg/kg/hr BIS 31only in (analysis Midazolam IV 0.02(0.015- 0.04) mg/kg/hr (n=9 enrolled, 5 analyzed) Dexmedetomidine IV (0.6infusion [0.4-0.7] mcg/kg/hr) (n=16 enrolled, 10 analyzed) For peer cross Noover treatment review only

Author, year byOutcome of Study Findings, Summary Table3: Engelmann et al., 2014 Kim et al.,Kim 2014 Alexopoulou etAlexopoulou al., 2014 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 29 29 BIS: p = 0.777 Sleep p=0.623Diary: FragmentationSleep BIS: p = 0.041 StageSleep Distribution BIS p=0.016 values: p Stage = A: 0.044 B: Stage p = 0.004 Stage pC: = 0.69 QualitySleep Sleep pDiary: < 0.001 None None reported

b b

a b 188 (136-449) = a BIS: 94.7% (30.7-100) (median):Sleep diary 86% FragmentationSleep BIS (median): 0.6 arousals/hr Sleep Efficiency Sleep 52.3% (47-89.7) FragmentationSleep 9.3(3-19.5) arousals/hr StageSleep Distribution (min) I=76Stage (32-145) II Stage BIS: 92.7% (16.7-100) (median):Sleep diary 71% FragmentationSleep BIS (median): 1.1 StageSleep Distribution BIS values=78.7 (72.05-81) 30%Stage A= (4.0-81.3) B=Stage 55.4% (96-90) Stage 7.6%C= (0-83.7) QualitySleep Sleep (median):Diary 3/5 StageSleep Distribution BIS values=74.05 (65.68- 79.38) 7.4%Stage A= (1.1-98) B=Stage 36.6% (5.4-82.3) Stage 5.0%C= (0-50.7) QualitySleep Sleep (median)Diary = 2/5 Sleep Efficiency Sleep

BMJ Open http://bmjopen.bmj.com/ Polysomnography flunitrazepam patients) (PgShD)Sleep diary used to evaluate sleep quality, sleep andefficiency sleep fragmentation recordedSleep quality on 5 propofol and patients 25 point scale

on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml nalyzed) (n=34 34randomized, a Dexmedetomidine IV 0.2-0.7infusion mcg/kg/hr (n=10 enrolled, 10 analyzed) For peer Flunitrazepam Bolus 0.015mg/kg (32 randomized, 32 analyzed) review only

Otoal., et 2012 Page 29 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 30 of 54 30 30 Sleep EfficiencySleep FragmentationSleep StageSleep Distribution I: pStage = 1.0 II: Stage p = 0.66 III/IV:Stage p = 0.75 pREM: = 0.04 Sleep EfficiencySleep p = 0.047 FragmentationSleep StageSleep Distribution I: pStage = 0.46 II: Stage p = 0.55 III/IVStage : p = 0.03 pREM: = 0.01 p = 0.37 p = 0.33 p = 0.03

a a

a a a a

a a a a a EM=0.0 (0-0) Sleep Efficiency Sleep Stage III=0Stage (0.0-1.3) R 4.4 (2.5-8.3) arousals/hr StageSleep Distribution (min) 97% (91.1-97.8%) FragmentationSleep Sleep Efficiency Sleep I Stage = 96 (60-144) IIStage = 264 (222-360) III/IVStage = 1.8 (0.6-4.2) REM = 30 (12-48) 62.6% (13.1-85.9) FragmentationSleep 8.1 (2.9-16.2) arousals/hr StageSleep Distribution I =Stage 30.7% (4.6-66.7) IIStage = 46.1% (3.0-80.4) III/IVStage = 0% (0-00) REM = 1.4% (0-13) Sleep Efficiency Sleep 76.3% (28.4-96.9) FragmentationSleep 4.8 (1.3-14.6) arousals/hr StageSleep Distribution I =Stage 20.8% (5.6-80.6) IIStage = 48.9% (4.8-84) III/IVStage = 0% (0-5.8) REM = 0% (0-0) 81% (73.3-91.1%) FragmentationSleep Sleep Efficiency Sleep

BMJ Open http://bmjopen.bmj.com/ Polysomnography Polysomnography

on September 30, 2021 by guest. Protected copyright. continuous continuous with daytime with a a a a

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml infusion (n=11 11randomized, sedation sedation interruption (n=11 11randomized, analyzed) Midazolam infusion 2.7 (2.0-3.4)mg/kg Midazolam infusion 0.6 (0.5-0.9)mg/kg Propofol 0.86 infusion (0.67 - 1.25) mg/kg/h (n=7 randomized, 6 analyzed) (n=6 randomized, 6 analyzed) For peerNo propofol cross over review only

Otoal., et 2011 Kondili Kondili et al., 2012 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 31 31

= 0.98 Sleep EfficiencySleep FragmentationSleep StageSleep Distribution I: pStage = 0.40 II: Stage p = 0.20 III/IV:Stage p = 0.93 reportednot REM: Sleep Efficiency Sleep BIS: p = 0.09 pActigraphy: = 0.84 observation:Nurse p = 0.58 Patient self-report: p = 0.32

Sleep Efficiency Sleep p = 0.07 p = 0.90 p

d b b c c

d Sleep Efficiency Sleep 4226.7% ± FragmentationSleep 50% (0-72%) Stage I =5.8±6.6Stage II=57.7±Stage 23.8 III/IVStage =31.9±24.6 REM=3.6±5.7 50% (16-69) Sleep Efficiency Sleep BIS: 26% (17-36%) Actigraphy: 75% (67-83) 68) Patient self-report: 50% (43- 58) BIS: 39% (27-51) Actigraphy: 73% (53-93) 64) Patient self-report: 41% (24- 59) 4.24±2.4 arousals/hr StageSleep Distribution (%) 2.2 (0-2.6) arousals/hr StageSleep Distribution (min) I Stage = 180 (54-342) IIStage = 282(66-468) III/IVStage = 0(0-0) REM = 0 (0-9) Sleep Efficiency Sleep Sleep Efficiency Sleep Sleep Efficiency Sleep Nurse observation:Nurse 51% (35- Nurse observation:Nurse 45% (26- BMJ Open http://bmjopen.bmj.com/ ctigraphy Polysomnography Observation by Observation nurse Patient self report questionnaire

BIS A Nurse observationNurse d on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml analyzed) Intermittent Intermittent Lorazepam IV: 0.04± 0.04 mg/kg/day (n=6 randomized, 6 analyzed) Placebo orally (n=18 18randomized, analyzed) Melatonin 3mg orally (n=14 14randomized, analyzed) Placebo (n=12 12randomized, analyzed) For Melatonin orally10mg (n=13 12randomized, analyzed) peer review only

Hardin et al., 2006 Ibrahim et Ibrahim al., 2006 Bourne et 2008al., Page 31 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 32 of 54 32 32

QualitySleep Conditioning thein morning pAlertness: = 0.60 pTiredness: = 0.29 Total Time Sleep

p = 0.09 p = 0.17

d d

d d d

a a

a N/A =

d d Total Time sleep Sleep Efficiency Sleep Sleep Efficiency Sleep Total Time Sleep Stage I=7.4±5.9Stage II=48.8±Stage 22.3 III/IVStage =38.9±28.1 REM =78.8±19.6% FragmentationSleep 58.8±26.3% FragmentationSleep Alertness Alertness = 9 (8-10) =Tiredness 7.5 (5-10) 443±37.8 min QualitySleep 469±46.2 min QualitySleep 8.5 (7.5-10) Conditioning thein morning 4.79±3.7 arousals/hr StageSleep Distribution (%) I=20.8±35.5Stage II=28.34±Stage 35 III/IV=49.6±49Stage REM=N/A 8 (5-9) 4.4±3.7 arousals/hr StageSleep Distribution (%) BMJ Open http://bmjopen.bmj.com/ Patient questionnaire (10 insleep, thealertness morning) Nurse observationNurse point scale for quality of d d on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml .72±0.39 mg/kg/day Continuous Continuous Lorazepam IV 0 Lorazepam 0.5-1.0mg (n=10 9randomized, analyzed) Zopiclone Zopiclone 3.75-7.5mg (n=10 9randomized, analyzed) Continuous Continuous Lorazepam IV with neuromuscular blockade 0.62±0.20 mg/kg/day (n=6 randomized, 6 analyzed) (n=6 randomized, 6 analyzed) For peer review only

Li Pi etShan Li al., 2004 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 33 33 Sleep LatencySleep Both triazolam and zolpidem groups significantly were QualitySleep Both triazolam and zolpidem groups significantly were Morning Sleepiness differenceNo significant better than placebobetter than (p<0.001) significantlynot but different between other.each placebobetter than (p<0.001) significantlynot but different between other.each between (pgroups = 0.443)

(mean±SD)

(mean±SD) (mean±SD)

(mean±SD) (mean±SD)= a orning 2.0±0.1 Conditioning thein morning 62.8±2.5 Sleep Latency Sleep Median: Median: 60 minutes <30 min=22% <60 min=49% <90 min=64% <120 min=72% QualitySleep 2.7±0.1 Sleep Latency Sleep Median: Median: 25 min <30 min=42% <60 min=78% <90 min=92% <120 min=89% QualitySleep 2.0±0.1 Conditioning thein morning 58.4±2.7 Sleep LatencySleep Median: 30 minutes <30 min=45% <60 min=76% <90 min=87% <120 min=90% QualitySleep Conditioning thein m Alertness Alertness = 9 (6.5-10) =Tiredness 8 (5.5-8.5) BMJ Open http://bmjopen.bmj.com/ hours post hours dose only) Sleep quality: self-patient report 1=excellent 2=good 3=fair 4=poor Conditioning in the morning report)(Patient self Visual analog scale (0-100 sleepy0=very where and 100=not at all sleepy) Nurse observationNurse (first 2 on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Triazolam 0.25mgTriazolam (n=119 119randomized, analyzed) Placebo (n=118 118randomized, analyzed) For peer 10mgZolpidem (n=120 120randomized, reviewanalyzed) only

Morgan 1997 al., et Page 33 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 34 of 54 34 34 Total TimeSleep Reported no significant as difference between groups LatencySleep (<20min) p < 0.05 for both midazolam and oxazepam vs placebo

Not reported

430 min

: : 440 min : : : NR d d

d d

(mean±SD)= orning Total Time Sleep Sleep Latency Sleep Total Time Sleep Latency:Sleep 0% asleep20 within min 29% asleep40 min within Total Time Sleep Sleep Latency Sleep Conditioning thein m (<30 55.5 min): patients (70%) QualitySleep goodVery =18% Good =55% (<30 55.5 min): patients (70%) QualitySleep Very goodVery =17% Good =60%, Moderate =18% Poor =5% Conditioning thein morning Refreshed =47% Little tired =44% tiredVery =9% Latency:Sleep 16% asleep20 min within 73% asleep40 min within 59.0±2.9 Latency:Sleep 30% asleep20 min within 84% asleep40 min within BMJ Open http://bmjopen.bmj.com/ Observation by Observation nurse Patient questionnaire Frequencies (%) are approximated a barfrom graph Nurse observationNurse on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Brotizolam Brotizolam 0.25mg orally (n=86 79randomized, breakdownbut analyzed by group reported)not Midazolam orally15mg (n=19 19randomized, analyzed) Triazolam 0.25mgTriazolam orally Placebo (n=16 16randomized, analyzed) For 15mgOxazepam orally (n=15 15randomized, analyzed) peer review only

Goetzke et al., 1983 Feldmeier Feldmeier et al., 1983 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 35 35 Total Sleep Time Total Time Sleep compared to placebo no but difference between drugs LatencySleep Sleep onset theshorter was in midazolam group when compared placebowith (p<0.05) and oxazepam (p,0.05).There no was statistical difference between oxazepam and placebo groups. p < 0.05 drugs for when both

oor =6% oor =6% Total TimeSleep <360 min=10% 360-420 min=53% >420 min=37% LatencySleep <30 min=63% 30-50 min=21% >50 min=16% <360 min=12% 360-420 min=52% >420 min=36% LatencySleep <30 min=17% 30-50 min=44% Total Time Sleep Sleep Latency Sleep Moderate =21% P (<30 min): 36.5 (46%)patients QualitySleep Very goodVery =10% Good =32% Moderate =30% Poor =28% Conditioning thein morning Refreshed =31% Little tired =45% tiredVery =24% Conditioning thein morning Refreshed =49% Little tired =43% tiredVery =8% BMJ Open http://bmjopen.bmj.com/ Third party observationThird party (values are approximated a graph)from on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Midazolam orally15mg (n=19 18randomized, analyzed) Oxazepam 50mgOxazepam orally (n=20 17randomized, analyzed) ForPlacebo peer review only

Gallais etGallais al., 1983 Page 35 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 36 of 54 36 36

: : Total Sleep Time Total Time Sleep Latency Sleep FragmentationSleep

p > 0.05 p > 0.05 Not reported

Total TimeSleep 360 min=9.9% 420 min=41.6% 480 min=25.9% 540 min=10.9% 600 min=0.6% LatencySleep <20 min=15.2% 20-40 min=42.7% 40-60 min=22% >60 min=20.1% Total Sleep Time Total Time Sleep <360 min=30% 360-420 min=54% >420 min=16% Latency Sleep (arousals/night) 1-2=60.4% >2=23.7% Total TimeSleep 360 min=13.8% 420 min=22.5% 480 min=30.8% 540 min=14.3% 600 min=3.6% latencySleep <20 min=34.4% 20-40 min=48.7% 40-60 min=12.3% >60 min=4.7% FragmentationSleep <30 min=10% 30-50 min=50% >50 min=40% >50 min=39% None=15.9% BMJ Open http://bmjopen.bmj.com/

Nurse observationNurse on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Midazolam (4015mg randomized, 40 analyzed) Oxazepam 15mgOxazepam (38 randomized, 38 analyzed) Placebo (n=20 16randomized, analyzed) For peer review only

15. Lupolover et 1983 al., 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 37 37

leep leep Fragmentation S (arousals/night) 1-2=55.4% >2=27.1% None=17.5% BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

Fordeviation standard ± peer review only data presented as median (range) median (range) as presented data data presented as mean as presented data data presented as median (interquartile range) median (interquartile as presented data CI) (95% mean as presented data

a c d Sleep DiaryPittsburg – PgShD b NR = not reported reported not = NR Page 37 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 38 of 54 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 Figure 1: Process of Study Identification and Selection 37 38 39 40

39 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Study Random Allocation Blinding of Blinding of Incomplete Selective 4 (Author, year sequence concealment participants outcome outcome reporting 5 of publication) generation and assessment data 6 personnel 7 8 Engelman et Moderate Moderate Moderate Moderate Low Low 9 al., 2014 10 Kondili et al., Moderate High High Low Low Low 11 2012 12 Oto et al., Low Moderate High High Low Low 13 2011 14 Bourne et al., Low Moderate Low Moderate Low Low 15 2011 For peer review only 16 Ibrahim et al., Low Low Low High Moderate Moderate 17 2006 18 Li Pi Shan et al., Low Moderate Moderate Moderate High Low 19 2004 20 Morgan et al., Low Low Moderate Moderate Low Low 21 2004 22 Feldmeier et Moderate Low Moderate Low Low Low 23 24 al., 1986 25 Goetzke et al., Low Moderate Moderate Moderate High Low 26 1983 27 Gallais et al., Low Low Low Moderate High Low 28 1983 29 Lupolover et Moderate Moderate Moderate Moderate High Low 30 al., 1983 31 32 Figure 2: Risk of Bias for Randomized Controlled Trials using the Cochrane Risk of Bias Tool 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

40 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Study Selection of Studies Comparability of Ascertainment of 4 (Author, year of (max of 4 ) Groups exposure/outcome 5 publication) (max of 2 ) (max of 3 ) 6 Alexopolou et al., 2014 7 8 9 Hardin et al., 2006 10 11 12 Figure 3: Risk of Bias for Prospective Cohort Studies (NewCastle-Ottawa Scale) 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 Figure 4: Sleep Efficiency: Observed differences from randomized and non-randomized 35 studies and from active and inactive controlled studies (7 studies, 194 patients) 36 37 38 39 40

42 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 Figure 5: Sleep Fragmentation: Observed differences from randomized and non- 35 randomized studies and from active and inactive controlled studies (5 studies, 125 patients) 36 37 38 39 40

43 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Figure 6: Sleep Architecture: Observed differences from randomized and non randomized 32 studies and from active and inactive controlled studies (4 studies, 59 patients)

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

44 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Figure Legends: 4 5 6 Figure 1: Process of Study Identification and Selection 7 *Prospective cohort studies not using polysomnography to measure sleep 8 9 10 Figure 2: Risk of Bias for Randomized Controlled Trials using the Cochrane Risk of Bias 11 Tool 12 13 Risk of bias assessed as low, moderate or high for each category. 14 15 Figure 3: Risk ofFor Bias for Prospectivepeer Cohort review Studies (NewCastle-Ottawa only Scale) 16 17 Figure 4: Sleep Efficiency: Observed differences from randomized and non-randomized 18 19 studies and from active and inactive controlled studies (7 studies, 188 patients) 20 Sleep Efficiency was defined as the number of hours spent asleep divided by the total hours of 21 observation. In some studies, mean and SD values were estimated using an approached 22 described in the methods section. *=non-randomized study; IV=intravenous; 23 NMBA=neuromuscular blocking agent; SD=standard deviation 24 25 26 Figure 5: Sleep Fragmentation: Observed differences from randomized and non- 27 randomized studies and from active and inactive controlled studies (5 studies, 125 patients) 28 Sleep fragmentation was defined as the number of arousals or awakenings per hour of sleep. In 29 some studies, mean and SD values were estimated using an approached described in the methods 30 section. *=non-randomized study; IV=intravenous; NMBA=neuromuscular blocking agent; 31 SD=standard deviation; NR=not reported; NE=not estimable 32

33 http://bmjopen.bmj.com/ 34 Figure 6: Sleep Architecture: Observed differences from randomized and non randomized 35 studies and from active and inactive controlled studies (4 studies, 59 patients) 36 Sleep stage distribution defined as percentage of total sleep time patients spent in each stage. In 37 some studies, mean and SD values were estimated using an approached described in the methods 38 39 section.*=non-randomized study; IV=intravenous; cont=continuous dosing; int=intermittent 40 dosing; PRO=propafol; DEX=dexmedatomidine; LOR=lorazepam; MID=midazolam;

41 NMBA=neuromuscular blocking agent; No Trt=no treatment; MD=mean difference; on September 30, 2021 by guest. Protected copyright. 42 SD=standard deviation; NE=not estimable 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 45 of 54 BMJ Open

45 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Appendix 1 4 5 6 Database: Embase Classic+Embase <1947 to 2015 January 27>, Ovid MEDLINE(R) In-Process 7 & Other Non-Indexed Citations and Ovid MEDLINE(R) <1946 to Present> 8 Search Strategy: March 21, 2016 9 ------10 11 1 sleep/ or night sleep/ or sleep induction/ (119114) 12 2 exp *sleep disorder/ (48385) 13 3 (sleep or insomnia).tw. (281519) 14 4 1 or 2 or 3 (318752) 15 5 intensive careFor unit/ (128046) peer review only 16 6 *critical illness/ or *critically ill patient/ or *intensive care/ or *artificial ventilation/ 17 18 (107617) 19 7 *hospital patient/ (13433) 20 8 *postoperative care/ (25079) 21 9 *postoperative period/ (6176) 22 10 ((mechanical or artificial) adj ventilat$).tw. (76070) 23 24 11 (intensive care or icu).tw. (260630) 25 12 (critical$ adj2 (care or ill$)).tw. (116727) 26 13 (postoperativ$ or post operativ$ or inpatient$).tw. (1129711) 27 14 (hospital$ adj2 patient$).tw. (173950) 28 15 or/5-14 (1679818) 29 16 4 and 15 (15024) 30 31 17 exp *benzodiazepine derivative/ (64289) 32 18 *ramelteon/ (202)

33 19 ramelteon.tw. (515) http://bmjopen.bmj.com/ 34 20 *melatonin/ (27568) 35 21 Melatonin.tw. (39686) 36 37 22 *melatonin receptor agonist/ (31) 38 23 exp *antidepressant agent/ (161899) 39 24 exp *antihistaminic agent/ (79776) 40 25 *imidazopyridine derivative/ (193)

46 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 41 exp *neuroleptic agent/ (193754) 4 5 42 *olanzapine/ (5515) 6 43 olanzapine.tw. (15748) 7 44 *quetiapine/ (2936) 8 45 Quetiapine.tw. (8643) 9 46 *propofol/ (21766) 10 11 47 Propofol.tw. (36217) 12 48 *alpha 2 adrenergic receptor stimulating agent/ (1054) 13 49 *dexmedetomidine/ (3961) 14 50 Dexmedetomidine.tw. (5629) 15 51 exp *hypnoticFor sedative peer agent/ (137181) review only 16 52 estazolam/ (1148) 17 18 53 midazolam/ (41684) 19 54 temazepam/ (5932) 20 55 triazolam/ (6772) 21 56 flurazepam/ (5391) 22 57 flunitrazepam/ (10031) 23 24 58 (Estazolam or Midazolam or Temazepam or Triazolam or Flurazepam or 25 Flunitrazepam).tw. (36087) 26 59 (amitriptyline or nortriptyline or Mirtazapine or Trazodone or Doxepin).tw. (24214) 27 60 *amitriptyline/ or *nortriptyline/ or *mirtazapine/ or *trazodone/ or *doxepin/ (23976) 28 61 (Diphenhydramine or chlorpheniramine or hydroxyzine).tw. (13454) 29 62 Diphenhydramine.tw. (7480) 30 31 63 chlorpheniramine.tw. (4061) 32 64 hydroxyzine.tw. (2544)

33 65 or/17-64 (699111) http://bmjopen.bmj.com/ 34 66 16 and 65 (1685) 35 67 (animal experiment/ or animal/) not humans/ (6747169) 36 37 68 66 not 67 (1671) 38 69 68 use emczd (1285) 39 70 exp Sleep/ (231886) 40 71 exp Sleep Disorders/ (211280)

47 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 87 Receptors, Melatonin/ag (122) 4 5 88 ramelteon.tw. (515) 6 89 Melatonin/ (40427) 7 90 exp Antidepressive Agents/ (454850) 8 91 Histamine Antagonists/ (41545) 9 92 Imidazopyridine$.tw. (910) 10 11 93 Zolpidem.tw. (4150) 12 94 Pyridines/ (63930) 13 95 Pyrazolopyrimidine$.tw. (540) 14 96 Zaleplon.tw. (669) 15 97 Cyclopyrrolone$.tw.For (250)peer review only 16 98 eszopiclone.tw. (472) 17 18 99 zopiclone.tw. (1858) 19 100 exp Barbiturates/ (203683) 20 101 Chloral Hydrate/ (8173) 21 102 exp Antipsychotic Agents/ (348253) 22 103 olanzapine.tw. (15748) 23 24 104 Quetiapine.tw. (8643) 25 105 Propofol/ (49309) 26 106 Propofol.tw. (36217) 27 107 exp Adrenergic alpha-2 Receptor Agonists/ (24347) 28 108 Dexmedetomidine/ (6363) 29 109 Dexmedetomidine.tw. (5629) 30 31 110 exp "Hypnotics and Sedatives"/ (439023) 32 111 (Estazolam or Midazolam or Temazepam or Triazolam or Flurazepam or

33 Flunitrazepam).tw. (36087) http://bmjopen.bmj.com/ 34 112 Melatonin.tw. (39686) 35 113 (amitriptyline or nortriptyline or Mirtazapine or Trazodone or Doxepin).tw. (24214) 36 37 114 (Diphenhydramine or chlorpheniramine or hydroxyzine).tw. (13454) 38 115 or/86-114 (1336662) 39 116 85 and 115 (7050) 40 117 animals/ not humans/ (5087200)

48 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 6 exp Critical Care/ or exp Intensive Care/ or critical illness/ (2235) 4 5 7 Respiration, Artificial/ (2354) 6 8 *inpatients/ or *Hospitalization/ (0) 7 9 postoperative care/ or postoperative period/ (6109) 8 10 ((mechanical or artificial) adj ventilat$).tw,hw. (3634) 9 11 (intensive care or icu).tw,hw. (9709) 10 11 12 (critical$ adj2 (care or ill$)).tw,hw. (3550) 12 13 (postoperativ$ or post operativ$ or inpatient$).tw,hw. (63767) 13 14 (hospital$ adj2 patient$).tw,hw. (10443) 14 15 or/5-14 (82559) 15 16 4 and 15 (1240)For peer review only 16 17 exp Benzodiazepines/ (7485) 17 18 18 Receptors, Melatonin/ag (0) 19 19 ramelteon.tw,hw. (53) 20 20 Melatonin/ (667) 21 21 exp Antidepressive Agents/ (9517) 22 22 Histamine Antagonists/ (156) 23 24 23 Imidazopyridine$.tw,hw. (62) 25 24 Zolpidem.tw,hw. (442) 26 25 Pyridines/ (1773) 27 26 Pyrazolopyrimidine$.tw,hw. (14) 28 27 Zaleplon.tw,hw. (81) 29 28 Cyclopyrrolone$.tw,hw. (29) 30 31 29 eszopiclone.tw,hw. (78) 32 30 zopiclone.tw,hw. (272)

33 31 exp Barbiturates/ (1816) http://bmjopen.bmj.com/ 34 32 Chloral Hydrate/ (93) 35 33 exp Antipsychotic Agents/ (7162) 36 37 34 olanzapine.tw,hw. (1928) 38 35 Quetiapine.tw,hw. (846) 39 36 Propofol/ (3124) 40 37 Propofol.tw,hw. (6773)

41 38 exp Adrenergic alpha-2 Receptor Agonists/ (2227) on September 30, 2021 by guest. Protected copyright. 42 39 Dexmedetomidine/ (298) 43 44 40 Dexmedetomidine.tw,hw. (743) 45 41 exp "Hypnotics and Sedatives"/ (10693) 46 42 (Estazolam or Midazolam or Temazepam or Triazolam or Flurazepam or 47 Flunitrazepam).tw,hw. (6075) 48 43 Melatonin.tw,hw. (1099) 49 50 44 (amitriptyline or nortriptyline or Mirtazapine or Trazodone or Doxepin).tw,hw. (3497) 51 45 (Diphenhydramine or chlorpheniramine or hydroxyzine).tw,hw. (1352) 52 46 or/17-45 (41116) 53 47 16 and 46 (326) 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from

on page # # page on Reported Reported 7 7 6 6 4-5 N/A N/A 4-5 4 2 1 1 Appendix Appendix up) report up) characteristics and (e.g., analysis). - BMJ Open http://bmjopen.bmj.com/

on September 30, 2021 by guest. Protected copyright. List and define List variables all for data which sought were (e.g., PICOS,fundingsources) any and assumptionsand simplifications made. Describe method Describe of dataextraction from reports(e.g., forms, piloted inindependently, duplicate)and any processesfor and confirming obtaining from data investigators. State the State process for studies selecting (i.e., screening, eligibility, includedin systematic review, and, ifapplicable, included the meta in Present full electronicPresent search strategyfor least one at database, including limits any used, such could that repeated. it be Describe all Describe information sources databases (e.g., of with dates coverage, contact study with authorsidentify additionalto studies) in the search and searched. date last Specify study Specify characteristics PICOS, (e.g., length of follow - Indicate Indicate review if a protocolexists, if andcan it where Webaccessed (e.g.,be address), and, if available, provideregistration information registrationincluding number. Provide anProvide statement explicit of questions being addressedwithreference to participants, interventions,comparisons, outcomes, study and design(PICOS). Describe the Describe rationale for review inthe the context of what already is known. Provide a structured a Provide summary including, asapplicable: background; objectives;datasources; criteria, study eligibility participants, andinterventions; study appraisaland synthesis methods; limitations; results; conclusions and implications ofkey findings; systematic review registration number. Identify the Identify report systematic asa meta-analysis, review, both.or Checklist item item Checklist years considered, years language, publication status) used for ascriteria eligibility, giving rationale.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 1 8 2 4 7 9 3 6 # For peer review only 11

Data Data items Data collection Data process 10 Study selection Study Search Search Information sourcesInformation Eligibility criteria Eligibility Protocol and registrationProtocol 5 METHODS METHODS Objectives Objectives Rationale Rationale INTRODUCTION INTRODUCTION Structuredsummary ABSTRACT ABSTRACT Title Title TITLE TITLE Section/topic Section/topic Appendix 2: PRISMA Checklist PRISMA Appendix2: Page 49 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 50 of 54 50 50 7-8 16 16 9-13 9-13 N/A N/A N/A N/A N/A N/A 24-33 24-33 35 35 21-23 21-23 34 34 7 7 7 7 analysis. - specified. - foreach meta ) ) 2 BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. Discussstudy limitations at and outcome(e.g., level risk review-level ofand at bias), (e.g., retrieval incomplete of identifiedresearch, reporting bias). Summarizethe main findings the strengthincluding ofevidencefor main each outcome; their consider relevance to groups key (e.g., healthcare providers, users, and policy makers). Give results ofresults Give additional analyses, if (e.g.,done sensitivity subgroupor analyses, meta- [see regression Item 16]). Present results Present of any assessmentof risk across of (see Itembias studies 15). Present results Present of each meta-analysis done, including confidence intervalsand measures of consistency. For all outcomes Forall considered(benefits harms),or for study: each present, (a) simple summary for eachdata intervention group (b) effect estimates confidence intervals, and ideally with a forest plot. Present Present dataon risk of ofstudy and, each bias ifavailable, outcome any assessmentlevel (see item 12). For each study, Foreach present characteristics for which data extracted were (e.g., study PICOS,size, follow-upperiod) and providethe citations. Give numbersGive screened,of studies assessed for eligibility, and includedin the review, with reasonsfor exclusions each at stage, ideally flow with a diagram. Describe methods Describe of additional analyses (e.g., sensitivity subgroupor analyses, meta- regression), ifindicating done, prewhich were Specify Specify assessment any riskofthat may bias cumulative affectof evidence(e.g., the publicationbias, selective reporting withinstudies). Describe the Describe methods of handling dataand combining ofresults studies, if done, including measures(e.g., I consistency of State the State principal summarymeasures (e.g., risk ratio,means). difference in Describe methods Describe used for assessing risk studies of of bias individual (including specification of whetherthis done the at was studyoutcome or this level), and how information is be used to in datasynthesis. any

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 12 21 14 17 13 25 16 For peer review only23

tudies tudies

Riskindividual in of bias s Limitations Summaryof evidence 24 DISCUSSION DISCUSSION Riskacross of bias studies 22 Synthesisof results Results of Results individual studies 20 Riskstudies within of bias 19 Study characteristics Study 18 Study selection Study RESULTS RESULTS Riskacross of bias studies 15 Synthesisof results Summarymeasures Additional Additional analysis Additional Additional analyses 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 51 51 N/A N/A 16-17 16-17 BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. Describe sources Describe of funding for systematic the review support and (e.g., other supply of data); of funders role for systematic the review. Provide a a Provide general interpretationof the results thecontext in of other evidence, and implications forfuture research.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 27 26 For peer review only

Conclusions

Funding Funding FUNDING FUNDING Page 51 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from

Page 52 of 54

on page # # page on Reported Reported 7 7 6 6 4-5 N/A N/A 4-5 4 2 1 1 Appendix Appendix up) report up) characteristics and (e.g., analysis). - BMJ Open http://bmjopen.bmj.com/

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 1 8 2 4 7 9 3 6 # For peer review only 11

Data Data items Data collection Data process 10 Study selection Study Search Search Information sourcesInformation Eligibility criteria Eligibility Protocol and registrationProtocol 5 METHODS METHODS Objectives Objectives Rationale Rationale INTRODUCTION INTRODUCTION Structuredsummary ABSTRACT ABSTRACT Title Title TITLE TITLE Section/topic Section/topic Appendix 2: PRISMA Checklist PRISMA Appendix2: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 7-8 16 16 9-13 9-13 N/A N/A N/A N/A N/A N/A 24-33 24-33 35 35 21-23 21-23 34 34 7 7 7 7 analysis. - specified. - foreach meta ) ) 2 BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. Discussstudy limitations at and outcome(e.g., level risk review-level ofand at bias), (e.g., retrieval incomplete of identifiedresearch, reporting bias). Summarizethe main findings the strengthincluding ofevidencefor main each outcome; their consider relevance to groups key (e.g., healthcare providers, users, and policy makers). Give results ofresults Give additional analyses, if (e.g.,done sensitivity subgroupor analyses, meta- [see regression Item 16]). Present results Present of any assessmentof risk across of (see Itembias studies 15). Present results Present of each meta-analysis done, including confidence intervalsand measures of consistency. For all outcomes Forall considered(benefits harms),or for study: each present, (a) simple summary for eachdata intervention group (b) effect estimates confidence intervals, and ideally with a forest plot. Present Present dataon risk of ofstudy and, each bias ifavailable, outcome any assessmentlevel (see item 12). For each study, Foreach present characteristics for which data extracted were (e.g., study PICOS,size, follow-upperiod) and providethe citations. Give numbersGive screened,of studies assessed for eligibility, and includedin the review, with reasonsfor exclusions each at stage, ideally flow with a diagram. Describe methods Describe of additional analyses (e.g., sensitivity subgroupor analyses, meta- regression), ifindicating done, prewhich were Specify Specify assessment any riskofthat may bias cumulative affectof evidence(e.g., the publicationbias, selective reporting withinstudies). Describe the Describe methods of handling dataand combining ofresults studies, if done, including measures(e.g., I consistency of State the State principal summarymeasures (e.g., risk ratio,means). difference in Describe methods Describe used for assessing risk studies of of bias individual (including specification of whetherthis done the at was studyoutcome or this level), and how information is be used to in datasynthesis. any

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 12 21 14 17 13 25 16 For peer review only23

Limitations Summaryof evidence 24 DISCUSSION DISCUSSION Riskacross of bias studies 22 Synthesisof results Results of Results individual studies 20 Riskstudies within of bias 19 Study characteristics Study 18 Study selection Study RESULTS RESULTS Riskacross of bias studies 15 Synthesisof results Summarymeasures Riskindividual in of bias studies Additional Additional analysis Additional Additional analyses Page 53 of 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 54 of N/A N/A 16-17 16-17 BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. Describe sources Describe of funding for systematic the review support and (e.g., other supply of data); of funders role for systematic the review. Provide a a Provide general interpretationof the results thecontext in of other evidence, and implications forfuture research.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 27 26 For peer review only

Funding Funding FUNDING FUNDING Conclusions 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from

Pharmacologic interventions to improve sleep in hospitalized adults: A systematic review

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2016-012108.R1

Article Type: Research

Date Submitted by the Author: 15-Jun-2016

Primary Subject Pharmacology and therapeutics

Heading: http://bmjopen.bmj.com/

Secondary Subject Heading: Epidemiology, Evidence based practice, Respiratory medicine

SLEEP MEDICINE, CLINICAL PHARMACOLOGY, Sleep medicine < Keywords: ANAESTHETICS

on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 Pharmacologic interventions to improve sleep in hospitalized 5 6 adults: A systematic review 7 8 *Salmaan Kanji, Pharm D. 9 Departments of Pharmacy and Critical Care, The Ottawa Hospital 10 11 The Ottawa Hospital Research Institute 12 13 *Alexandru Mera, B.Sc. Pharm., ACPR 14 Department of Pharmacy, Hôpital Montfort - The Ottawa Hospital 15 For peer review only 16 Brian Hutton, MSc, PhD 17 18 Ottawa Hospital Research Institute 19 University of Ottawa School of Epidemiology, Public Health and Preventative Medicine 20 21 Lisa Burry, Pharm D. 22 Department of Pharmacy, Mount Sinai Hospital 23 24 University of Toronto 25 26 Erin Rosenberg, BSc., MD, MHA, FRCP(c) 27 Department of Critical Care, The Ottawa Hospital 28 29 Erika MacDonald, B.Sc. Pharm., ACPR, M.Sc, RPh 30 31 Department of Pharmacy, The Ottawa Hospital 32 The Ottawa Hospital Research Institute

33 http://bmjopen.bmj.com/ 34 Vanessa Luks 35 BASc, MD, FRCP(c), 36 37 Department of Respirology, Thunder Bay Regional Health Sciences Centre, 38 Northern Ontario School of Medicine 39 40

41 Medications studied included benzodiazepines, nonbenzodiazepine sedatives, melatonin, on September 30, 2021 by guest. Protected copyright. 42 43 44 propofol, and dexmedetomidine. Five studies were deemed to be of high quality. Heterogeneity 45 46 and variable outcome reporting precluded meta-analysis in most cases. No consistent trends with 47 48 respect to sleep efficiency, quality or interruptions were observed identifying one drug or drug 49 50 51 class as superior to another or no treatment. Benzodiazepines appeared to be better than no 52 53 treatment with respect to sleep latency but this was not consistently demonstrated across all 54 55 studies. Sleep stage distribution shows that sleep in hospital is dominated by stage N1 and N2. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 52 BMJ Open

3 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Conclusion: There is insufficient evidence to suggest that pharmacotherapy improves the quality 4 5 6 or quantity of sleep in hospitalized patients suffering from poor sleep. No drug class or specific 7 8 drug was identified as superior even when compared to placebo or no treatment. Although 15 9 10 11 studies were included, the quality of evidence was limited by both their quality and size. Larger, 12 13 better-designed trials in hospitalized adults are needed. 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

4 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Strengths and limitations of this study 4 5 6 • The strengths of this review include 7 8 o the use of a validated systematic search strategy of multiple databases to identify 9 10 11 relevant studies 12 13 o article screening and data extracted independently by two investigators 14 15 o theFor research teampeer consisted reviewof physicians, pharmacists, only an 16 17 18 epidemiologist/methodologist, a sleep expert and pharmacotherapy specialists. 19 20 • The main limitations of this review include 21 22 o The studies included were typically small, outcomes were often measured 23 24 25 subjectively and most studies did not have a standard approach to the evaluation 26 27 of safety 28 29 o 30 not all drugs currently being used in clinical settings were identified in clinical 31 32 trials and comparative trials were identified without placebo controls for many

33 http://bmjopen.bmj.com/ 34 drugs 35 36 37 38 39 40

5 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Background 4 5 6 Sleep disturbances including difficulty initiating sleep, fragmented and non-restorative 7 8 sleep occur in up to 50% of acutely ill hospitalized patients.1 Hospitalized patients routinely 9 10 2,3 11 suffer from sleep disorders attributed to both pathophysiological and environmental factors. In 12 13 a study of 280 hospitalized elderly patients, 21% reported new-onset insomnia, 38% reported 14 15 moderate or severeFor insomnia, peer and 38% reported review sleep disturbances only during admission.4 The 16 17 18 transient nature of acute illness coupled with the displacement and hospitalization of patients 19 20 represent unique features that differentiate insomnia associated with acute illness and 21 22 hospitalization from insomnia in otherwise healthy people. Sleep deprivation has been associated 23 24 25 with poor wound healing, memory disorders, delayed ventilator weaning, delirium and 26 27 mortality.5-8 Given the frequency of sleep disorders experienced by hospitalized patients and 28 29 their associated outcomes, interventions to improve sleep are frequently attempted.1,9,10 Non- 30 31 32 pharmacologic interventions such as relaxation techniques, regulation of light/noise exposure are

33 http://bmjopen.bmj.com/ 34 often tried to counter new onset sleep disorders related to the environment.1,10 However, a 35 36 37 systematic review of non-pharmacologic interventions for sleep compared to no intervention (in 38 39 hospitalized but non-critically ill patients) did not confirm a positive effect, citing a lack of high 40 11 41 quality studies. While environmental factors (related to light, noise, and activity) are common on September 30, 2021 by guest. Protected copyright. 42 43 44 in hospital and particularly in the ICU, limiting the potential impact of these on sleep quality is 45 46 not always easily accomplished. Accordingly, sedative and hypnotic pharmacotherapy is used in 47 48 up to 88% of hospitalized patients.12 49 50 51 Currently there are no evidence-based recommendations or guidelines to direct the choice of 52 53 hypnotic drugs in the hospital setting, yet there are a plethora of agents that are used. The 54 55 objective of this systematic review was to identify, synthesize, and summarize the existing 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 52

6 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 evidence on the efficacy and safety of pharmacologic interventions used to improve sleep in 4 5 6 hospitalized adults. 7 8 9 Methods 10 11 12 A protocol was developed prior to initiation of the review and was followed throughout the 13 14 review process. The protocol is available from the authors electronically upon request. 15 For peer review only 16 Selection of Studies 17 18 19 The Population-Intervention-Comparator-Outcome-Study design framework was used to 20 21 formulate the research question for this systematic review.13 The population of interest was 22 23 24 hospitalized adults in attempt to identify acutely ill patients. Studies of patients in any acute care 25 26 ward of a hospital (including medical, surgical and critical care units) were sought and non- 27 28 hospitalized patients (i.e., healthy volunteers, outpatients, long-term care facility residents) were 29 30 31 excluded. We also excluded studies conducted solely in patients diagnosed with a primary 32

33 psychiatric illness or sleep apnea as these patient populations exhibit different baseline sleep http://bmjopen.bmj.com/ 34 35 patterns irrespective of environment.1 Interventions of interest for this systematic review 36 37 38 included any pharmacologic intervention for sleep used alone or in combination. This included: 39 40 benzodiazepines, antidepressants, antihistamines, imidazopyridines (e.g., zolpidem),

41 on September 30, 2021 by guest. Protected copyright. 42 pyrazolopyrimidines (e.g., zaleplon), cyclopyrrolones (e.g., zopiclone ), anticonvulsants 43 44 45 (pregabalin/gabapentin), barbiturates, alpha-2-adrenergic agonists, antipsychotics, melatonin or 46 47 melatonin receptor agonists and natural health products. Studies involving drugs that are no 48 49 50 longer commercially available in any country were excluded. No restrictions on dose for any 51 52 agent were employed. Comparators of interest included placebo, no treatment, or any other 53 54 intervention. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 52 BMJ Open

7 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Efficacy outcomes of interest were sleep efficiency, sleep latency, sleep interruptions and 4 5 6 sleep stage distribution measured by any means including polysomnography, bispectral index 7 8 monitoring (BIS), actigraphy, sleep questionnaires for patient self-reporting or direct observation 9 10 11 by a third party. However, studies solely relying on patient self-reporting for sleep outcomes of 12 13 interest were excluded as subjective reports are poorly correlated with objective measures and 14 15 biased by systematicFor over-reporting. peer14,15 Sleep review efficiency is defined only as hours spent asleep divided 16 17 18 by period of observation (in hours) by study investigators while sleep latency represents the 19 20 duration of time taken to fall asleep. Sleep fragmentation is defined as the number of arousals or 21 22 awakenings in a designated sleep period. Sleep stage distribution refers to the proportion of time 23 24 25 a patient spends in each sleep stage in a given period of observation. Secondary clinical 26 27 outcomes included length of stay, duration of mechanical ventilation, presence of delirium, 28 29 occurrence of infection, cognitive function, pain, glucose control, hypertension, myocardial 30 31 32 infarction and mortality. Safety outcomes of interest included overall occurrence of adverse

33 http://bmjopen.bmj.com/ 34 events, serious adverse events including those deemed life threatening, and withdrawal or 35 36 37 discontinuation of therapy due to adverse events. Study design comprised randomized controlled 38 39 trials of pharmacological interventions for sleep. Additionally, comparative cohort studies and 40

41 prospective case series were included only if sleep was evaluated using polysomnography. The on September 30, 2021 by guest. Protected copyright. 42 43 16 44 language of publication was restricted to English or French. 45 46 47 Literature Search Strategy 48 49 50 Relevant studies were identified using a sequential search approach. Initially, an informal 51 52 scoping exercise was conducted to identify potentially relevant trials in Ovid Medline with key 53 54 words such as “sleep”, “insomnia”, “drug therapy” and “in-patient”. Relevant studies identified 55 56 57 in the scoping exercise were provided to an information specialist to develop and validate a 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 52

8 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 search strategy for OVID Medline, Embase and Cochrane from inception to March 21, 2016. 4 5 6 The search strategy was peer-reviewed using PRESS (Peer Review of Electronic Search 7 8 Strategies) by a second information specialist and validated using the studies identified in the 9 10 17 11 scoping exercise. The detailed search strategy is provided in Appendix 1. In addition, the 12 13 reference lists of included studies and reviews were screened. 14 15 Grey literatureFor was searchedpeer via conference review proceedings only and abstracts (Canadian Sleep 16 17 18 Society, American Academy of Sleep Medicine and European Sleep Research Society) for the 19 20 last 5 years, as well as clinical trial and systematic review registries, and Google Scholar®.18 21 22 23 24 Titles and abstracts were independently reviewed for eligibility by two investigators (AM 25 26 and SK). Full-text articles identified as potentially eligible at first level screening were obtained 27 28 and reviewed for eligibility independently by the same two reviewers. Discrepancies were 29 30 31 resolved by discussion and consensus. The process of study selection was documented using a 32 19 33 PRISMA flow diagram. http://bmjopen.bmj.com/ 34 35 36 37 38 39 Data extraction and risk of bias assessment 40

41 Study level data was extracted using a pre-designed and piloted data collection form by on September 30, 2021 by guest. Protected copyright. 42 43 44 one author (AM) and checked by a second (SK). Means and standard deviations were estimated 45 46 using established methods when only medians and ranges were available from an included 47 48 study.20,21 In addition to the clinical and safety outcomes, extracted study data included the 49 50 51 following: study authors, year and journal of publication, country of study performance, funding 52 53 source, group sample sizes, study inclusion criteria, age distribution, gender distribution, patient 54 55 setting in hospital, and relevant concomitant medication use. Risk of bias was assessed using the 56 57 22 23 58 Cochrane Risk of Bias tool for randomized controlled trials and the Newcastle-Ottawa Scale 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 52 BMJ Open

9 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 for cohort studies. Risk of bias was not assessed for prospective case series, as the risk of bias is 4 5 6 inherently high. 7 8 9 Synthesis of available studies 10 11 12 Clinical and methodological heterogeneity of the included studies was carefully 13 14 considered to determine whether meta-analysis would be conducted. Clinical heterogeneity was 15 For peer review only 16 assessed by comparing study populations, interventions and outcome measurements between 17 18 24 19 studies. Methodological heterogeneity was assessed by comparing methods for intervention 20 21 assignment, allocation concealment and blinding, as well as the extent of and reasons for loss to 22 23 24 follow-up. If heterogeneity was not identified between trials and reporting of clinical outcomes 25 26 was amenable to meta-analysis, a random-effects model was employed to calculate pooled risk 27 28 ratios for dichotomous outcomes (i.e., adverse events) and pooled mean differences for 29 30 31 continuous outcomes (i.e., sleep efficiency) using RevMan version 5.3 (the Nordic Cochrane 32 25,26 33 Center, the Cochrane Collaboration, Copenhagen, Denmark) . Studies were to be pooled by http://bmjopen.bmj.com/ 34 35 drug and by drug class and compared with active controls or placebo separately depending on the 36 37 38 availability of data. In the event that 10 or more studies were pooled for any outcome, funnel 39 40 plots would be created and visually inspected to assess for publication bias.22

41 on September 30, 2021 by guest. Protected copyright. 42 The PRISMA statement and its 27-item checklist were used in developing the summary 43 44 19 45 of findings for this review. A completed PRISMA checklist is provided in Appendix 2. 46 47 48 49 50 Results 51 52 Characteristics of included trials 53 54 Our search strategy identified a total of 1,920 citations, from which 11 randomized 55 56 57 controlled trials, two prospective cohort studies and two case series, totaling 861 patients, met 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 52

10 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 criteria for this review (Figure 1). Twenty-eight studies were excluded because the only method 4 5 6 of sleep assessment was via patient self-report. Eight studies were conducted in the intensive 7 8 care unit and included patients requiring mechanical ventilation. Seven studies were conducted 9 10 11 in acute care hospital wards and included medical and surgical patients suffering from 12 13 inadequate sleep. Six studies measured sleep using polysomnography, two used BIS and the 14 15 remaining measuredFor sleep usingpeer either third review party observation oronly a combination of third party 16 17 18 observation and patient self-report (Tables 1 and 2). Medications studied included 19 20 benzodiazepines, melatonin, propofol, zolpidem, zopiclone and dexmedetomidine. In most 21 22 studies the intervention was administered in the evening and sleep was evaluated over night, 23 24 25 however two of the ICU based studies evaluated sleep while on continuous infusions of 26 27 sedatives. Study durations ranged from 24 hours to 14 days. Seven of the randomized controlled 28 29 trials had a high risk of bias in at least one criteria, which impacted their overall quality (Figure 30 31 32 2). The prospective cohort studies included were of high quality (Figure 3).

33 http://bmjopen.bmj.com/ 34 35 36 37 Sleep efficiency 38 27-31 39 Sleep efficiency was evaluated in seven studies (5 randomized controlled trials and 40 32,33 31,32 41 two cohort studies ), of which two identified statistically significant differences (Table 3, on September 30, 2021 by guest. Protected copyright. 42 43 et al 44 Figure 4). Alexopoulou evaluated sleep with polysomnography in 13 of 16 enrolled patients 45 46 over the course of three nights in a prospective cross-over cohort study.32 On nights 1 and 3 47 48 patients received no treatment, while on night 2 they received dexmedetomidine. 49 50 51 Polysomnography was not consistently viable, and only ten patients contributed to the reported 52 53 analysis, which showed that sleep efficiency was improved with dexmedetomidine (77.9% [SD: 54 55 65.6 - 80.2%]) as compared to no treatment (15.8% [6.4 – 51.6%], p=0.002). In a study by Oto et 56 57 58 al (2011), two different dosing strategies for midazolam were compared in a randomized 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 52 BMJ Open

11 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 controlled trial of 22 critically ill patients.31 Both groups received sedation with a continuous 4 5 6 infusion of midazolam, but one group had their infusions interrupted during the day. Sleep 7 8 efficiency as measured by polysomnography was statistically significantly improved (p=0.047) 9 10 11 in the group that did not have their infusion interrupted during the day (97% [91.1 – 97.8%] 12 13 versus 81% [73.3 – 91.1%]). In a similar trial of 18 patients by Hardin et al, critically ill adults 14 15 were administeredFor lorazepam peer via continuous review infusion or an interrupted only dosing strategy; unlike in 16 17 18 the previous study, differences between groups were not statistically significant (Table 3, Figure 19 20 4).33 21 22 Two randomized controlled trials (Bourne et al and Ibrahim et al) that compared 23 24 25 melatonin with placebo failed to show a statistically significant difference in sleep efficiency as 26 27 measured by BIS or nurse observation.27,28 Similarly, two trials by Engelmann et al and Kondili 28 29 et al comparing propofol to flunitrazepam30 and no treatment29 did not show statistically 30 31 32 significant differences with respect to sleep efficiency as measured by BIS and

33 http://bmjopen.bmj.com/ 34 polysomnography, respectively. Only the two placebo controlled melatonin trials (total of 56 35 36 37 patients) were amenable to pooling of this outcome that revealed a non-statistically significant 38 39 trend in favor of melatonin as compared to placebo (mean difference of 8.0%, [95% confidence 40

41 interval: -1.5, 17.5]). on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 Sleep latency 47 48 Sleep latency was evaluated in five studies. Due to variability in assessment and 49 50 51 reporting metrics between studies, forest plots were not created, and the appropriateness of meta- 52 53 analysis was deemed low. Sleep latency data is reported for each study in Table 3. 54 55 Two randomized controlled trials, Feldmeier (n=50) and Gallais (n=51),34,35 compared 56 57 58 the effects of midazolam and oxazepam to placebo. One randomized controlled trial by 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 52

12 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Lupolover et al (n=78)34 compared midazolam to oxazepam. Sleep latency was measured by 4 5 6 third party observation or patient questionnaire in all three studies. The studies by Feldmeier et al 7 8 and Gallais et al reported a statistically significant reduction in sleep latency with midazolam (15 9 10 11 mg orally) compared to placebo (median difference of 30-53%, p<0.05), however only 12 13 Feldmeier et al found oxazepam (15 mg orally) to be superior to placebo (median difference 14 15 16%, p<0.05).34 ForGallais et alpeer found midazolam review 15 mg orally to beonly superior to oxazepam 50 mg 16 17 35 et al 18 orally (Table 3) as did Lupolover (15 mg doses for both midazolam and oxazepam), but 19 20 only after 3 days of treatment (table 3).36 21 22 23 24 37,38 25 Two randomized control trials evaluated sleep latency with triazolam. Morgan et al 26 27 (n=357) compared triazolam to zolpidem and placebo, while Goetzke et al (n=79) compared it to 28 29 brotizolam and placebo. In both cases, drug therapy (any) appeared to be superior to placebo 30 31 38 32 with respect to sleep latency, however, no statistical analysis was provided by Goetzke et al.

33 http://bmjopen.bmj.com/ 34 No statistically significant differences were identified between the active arms of either study 35 36 37 (Table 3). 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 Sleep fragmentation 47 48 Sleep fragmentation was evaluated in five studies: three randomized controlled trials29-31 49 50 32,33 51 and two cohort studies (Table 3, Figure 5). Variability in treatments studied precluded the 52 53 opportunity for meta-analysis. Two randomized controlled trials studied the effect of propofol 54 55 on the number of awakenings compared to flunitrazepam (Engelmann et al)30 or no treatment 56 57 29 58 (Kondili et al) . Engelmann et al (n=66), using BIS, found that propofol (median: 0.6 arousals 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 52 BMJ Open

13 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 per hour of sleep) was associated with statistically significantly fewer sleep interruptions than 4 5 30 6 flunitrazepam (median: 1.1 arousals per hour of sleep in critically ill patients, p=0.041). 7 8 However, Kondili et al (n=12), using polysomnography, found insufficient evidence of a 9 10 11 difference between patients receiving propofol (median [interquartile range]: 4.8 [1.3 - 14.6] 12 13 awakenings per hour of sleep) and those receiving no treatment (median [interquartile range]: 14 15 8.1 [2.9 – 16.2] awakeningsFor peerper hour of sleep). review29 only 16 17 18 Dexmedetomidine compared to no treatment was found to statistically significantly 19 20 reduce the number of sleep interruptions in critically ill patients. Using polysomnography, 21 22 Alexopoulou et al (n=23) found that patients receiving infusions of dexmedetomidine had a 23 24 25 median of 2.2 (interquartile range: 1.6-4.5) awakenings per hour while those receiving no 26 27 treatment had a median of 7.1 (interquartile range: 1.6-4.5) awakenings per hour (p=0.0023).32 28 29 Two trials comparing different dosing strategies for benzodiazepines in critically ill 30 31 32 patients yielded conflicting results. Oto et al (2011) found that continuous infusions of

33 http://bmjopen.bmj.com/ 34 midazolam were associated with fewer awakenings when compared to infusions incorporating 35 36 et al 37 daytime interruptions, while Hardin found no difference between continuous infusions of 38 31,33 39 lorazepam (with and without a neuromuscular blocker) and an intermittent dosing strategy. 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 Sleep stage distribution 47 48 Two randomized control trials,29,31 two cohort studies32,33 and two case series39,40 49 50 51 evaluated sleep architecture using polysomnography in a total of 74 patients. Treatment effects 52 53 from comparative studies are displayed in Figure 6, while data for all studies is found in Table 3. 54 55 Irrespective of treatment allocation, patients in hospital primarily experience light non- 56 57 58 restorative sleep, corresponding to stages N1 and N2. Only Hardin et al (continuous infusion of 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 52

14 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 lorazepam with or without neuromuscular blocker versus interrupted infusion) reported a 4 5 6 proportion of sleep as Stage N3 (restorative sleep), where Stage N3 represented 32-50% of total 7 8 sleep but insufficient evidence of a difference between groups. No other trial using 9 10 11 polysomnography was able to describe meaningful amounts of stage N3 sleep. REM sleep was 12 13 only described with midazolam infusions interrupted during the day (6.8%), continuous infusions 14 15 of midazolam (2%),For interrupted peer lorazepam reviewinfusions (3.6%), and only the no treatment arm of the 16 17 29,31,33,39 18 study by Kondili et al (1.4%). Given the variability in studied treatments, study designs 19 20 and the limited number of overall studies meta-analysis was not attempted. 21 22 23 24 25 Secondary outcomes 26 27 Delirium was not assessed in the included studies, however Ibrahim et al evaluated 28 29 agitation using the Riker Sedation-Agitation Scale and found no difference between melatonin 30 31 28 32 (n=14) and placebo groups (n=18). Cognitive function was evaluated in the study by Li Pi

33 http://bmjopen.bmj.com/ 34 Shan et al using the Mini Mental State Exam and no difference was found between lorazepam 35 36 41 et al 37 (n=9) and zopiclone (n=9). Narcotic requirements were evaluated in the study by Oto . 38 39 While more patients (9/11=81.8%) in the group assigned to midazolam infusions with daily 40

41 interruption required opioids than those assigned to the continuous infusion group (4/11=36.4%), on September 30, 2021 by guest. Protected copyright. 42 43 31 et al 44 there was no difference in median dose between groups. In the study by Hardin (n=18), 45 46 patients assigned to continuous infusions of lorazepam (n=6) required significantly more 47 48 morphine (0.48 ± 0.52 mg/kg/day) than those assigned to interrupted infusions (0.11 ± 0.13 49 50 51 mg/kg/day) (n=6) or continuous infusions with neuromuscular blockade (0.14 ± 0.32 mg/kg/day) 52 53 (n=6) (p=0.031). No other differences were found between groups.33 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 52 BMJ Open

15 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Duration of mechanical ventilation, length of stay, occurrence of infection, glucose 4 5 6 control, hypertension, myocardial infarction and mortality were not described in any included 7 8 study. 9 10 11 12 13 Safety 14 15 No studyFor reported occurrence peer of serious review life threatening onlyevents. Seven of the 15 studies 16 17 27,30,34-38,41 18 did not report any safety information in the methods or results. Of the remaining eight, 19 20 no study reported any difference with respect to adverse events between the active treatment 21 22 groups compared to the placebo or no treatment arms. 23 24 25 Engelmann et al reported that three patients developed respiratory depression after 26 27 flunitrazepam administration requiring intervention the insertion of a nasopharyngeal airway but 28 29 not discontinuation from study.30 One patient receiving dexmedetomidine in the study by 30 31 32 Alexopoulou et al had the study drug discontinued for bradycardia (heart rate less than 50 beats

33 http://bmjopen.bmj.com/ 34 per minute).32 35 36 37 38 39 Patient conditioning, defined as somnolence or alertness the morning following 40

41 administration of hypnotics was reported in five studies. Four reported no differences between on September 30, 2021 by guest. Protected copyright. 42 43 34,37,38,41 et al 44 drugs and/or placebo. The study by Lupolover described an improved state on 45 46 awakening by patient self-report in those taking midazolam compared to oxazepam.36 47 48 49 50 51 Headache and nausea were the most commonly reported adverse events. Discontinuation 52 53 of therapy due to nausea was only observed in the placebo group of Morgan et al.37 Allergic 54 55 type skin reactions were reported in two patients - one was associated with triazolam; the other, 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 52

16 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 described as pruritis, occurred in a patient with pancreatitis and whether the reaction was 4 5 36,38 6 attributed to benzodiazepines or placebo was not specified in the study. 7 8 9 10 11 12 13 Discussion 14 15 In this reviewFor we set peer out to summarize review the existing evidence only and compare studies looking 16 17 18 at the efficacy and safety of various pharmacologic agents for sleep. Fifteen studies 19 20 encompassing 861 patients met our inclusion criteria, however clinical and methodological 21 22 heterogeneity and variable outcome reporting were judged to preclude performance of reliable 23 24 25 meta-analyses in all cases except for two trials comparing melatonin to placebo with similar 26 27 methods for evaluating sleep efficiency. An apparent reduction in sleep latency with 28 29 pharmacotherapy when compared to placebo or no treatment has been demonstrated although not 30 31 32 consistently across all studies evaluating this metric. Sleep efficiency and sleep fragmentation

33 http://bmjopen.bmj.com/ 34 were not shown to be improved by pharmacologic therapy when compared to placebo or no 35 36 37 treatment. This is in keeping with expert opinion that pharmacologic agents, while somewhat 38 42 39 effective for sleep onset insomnia, are ineffective for maintaining sleep. 40

41 Furthermore, sleep architecture in hospital, as measured by polysomnography, is on September 30, 2021 by guest. Protected copyright. 42 43 44 dominated by stage N1 and N2 sleep. Restorative N3 sleep or REM sleep was found to be rare 45 46 regardless of pharmacologic intervention. When sleep is highly fragmented, it is common to 47 48 observe a predominance of N1 sleep, as an arousal in an epoch of sleep on the EEG results in the 49 50 43 51 sleep stage being scored back to stage N1. Stage N3 sleep was observed only in a single study, 52 53 in patients receiving continuous or intermittently administered lorazepam.33 A paucity of N3 54 55 sleep is not unusual in middle-aged or elderly populations, where N3 typically constitutes only 5 56 57 44 58 to 10% of sleep. Furthermore, the sleep stages are influenced by the medications studied. 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 52 BMJ Open

17 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Benzodiazepines, for example, are potent suppressors of N3 sleep, and increase the presence of 4 5 45 6 sleep spindles. Spindles are the hallmark of N2 sleep, so it is common to see increased N2 7 8 sleep and decreased N3 sleep with benzodiazepines. Concurrent opioids will also decrease N3 9 10 46 11 sleep. The large percentage of N3 sleep seen in the study in which lorazepam was used is 12 33 13 surprising given the intervention is known to suppress N3 sleep. Furthermore, this finding is 14 15 not replicated inFor any other studies.peer A possible review explanation is that only N3 sleep is commonly over- 16 17 18 scored. REM sleep was only observed in four studies, independent of drug or placebo allocation. 19 20 REM sleep can also be influenced by current and prior medications. There is evidence to suggest 21 22 that benzodiazepines decrease REM sleep.47 Sleep stage distribution is also influenced by prior 23 24 25 sleep episodes and medications. For example, recent sleep deprivation will increase N3 sleep, 26 27 and REM sleep.48 Lastly, recent medication use prior to the polysomnography recording may 28 29 influence sleep stages (i.e., recent discontinuation of REM-suppressant such as a barbiturate, 30 31 32 anti-depressant, or alcohol, will result in REM rebound). In summary, the disappointing

33 http://bmjopen.bmj.com/ 34 performance of the sedatives in effecting greater amounts of N3 and REM is in part due to the 35 36 37 well-known effect of these drugs on sleep stages. In addition, the medications administered and 38 39 discontinued, as well as prior sleep during the run-in period can influence sleep architecture, and 40

41 although some of the trials were randomized, it is possible that these factors could be on September 30, 2021 by guest. Protected copyright. 42 43 44 confounding the results given the small sample sizes. It is also important to recognize that while 45 46 medication may influence the quality of sleep in hospital, severity of illness and environmental 47 48 factors are important confounders. Severity of illness among patients included in this review 49 50 51 ranged from critical illness requiring mechanical ventilation to admissions for elective surgery. 52 53 Severity of illness was not described in some studies. Although our study selection strategy was 54 55 designed to identify acutely ill adults there variability in severity of illness within this group is 56 57 58 still large. Furthermore, most studies did not describe the presence or evaluate the impact of 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 52

18 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 environmental factors such as noise, light and interruptions to provide care that contribute to 4 5 6 insomnia in hospital. Attempts to minimize or standardize the influence of environmental factors 7 8 (i.e., use of earplugs or eye covers, “quiet times” in the unit) were not documented or, perhaps, 9 10 11 attempted. Given the results of our review and others of non-pharmacologic interventions a 12 13 multi-modal approach that involves minimization of environmental factors, pharmacotherapy 14 15 and treatment ofFor the underlying peer disease would review be worthy of future only study. 16 17 18 19 The lack of high-quality randomized controlled trials is one of the limitations of this 20 21 review. The diversity of outcome reporting and the high degree of methodologic heterogeneity 22 23 24 prevented statistical pooling of the results in almost all circumstances. Major sources of clinical 25 26 heterogeneity included large variability in drug dosing (i.e., oxazepam dosing ranged from 15 to 27 28 50 mg per dose), severity of illness in hospital and the method by which sleep was evaluated. 29 30 31 Only six of the 15 studies included used polysomnography which is considered the gold standard 32

33 for sleep evaluation. Only two of these were randomized controlled trials, which speaks to the http://bmjopen.bmj.com/ 34 35 practical difficulty in using polysomonography in large trials and sick patients. In fact, 28 36 37 38 potentially eligible trials were excluded at second level screening because the sole measure of 39 40 sleep outcomes relied on patient self-reporting. We elected to exclude these studies as self-

41 on September 30, 2021 by guest. Protected copyright. 42 reporting has been demonstrated to correlate poorly with objective assessments and is associated 43 44 14 45 with systematic bias. Furthermore, the general lack of safety outcome reporting does not allow 46 47 a thorough evaluation of risk. Given paucity of convincing efficacy data, the relatively high 48 49 50 potential for adverse events with these classes of drugs and the vulnerability of the population 51 52 being studied future studies must systematically evaluate risk in order for clinicians to 53 54 adequately assess the risk/benefit relationship of these interventions. Language restriction to 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 52 BMJ Open

19 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 English and French may also be considered a limitation although others have shown that this 4 5 16 6 restriction may have little to no impact on findings. 7 8 It is also important to recognize the paucity of placebo controlled trials in this setting. An 9 10 11 evidence based approach to clinical trial design would suggest that placebo controlled trials that 12 13 evaluate the efficacy of drug therapies be conducted in advance of comparative trials. While 14 15 placebo controlledFor trials are peeravailable in healthy review volunteers and otherwiseonly healthy community 16 17 18 based patients, the etiology of sleep disturbances is not the same as that of the acutely ill patient 19 20 sleeping in a hospital bed. Similarly the outcomes of such trials cannot be extrapolated to this 21 22 population either. Given that the majority of hospitalized patients receive pharmacologic sleep 23 24 25 aids in hospital in the absence of evidence to suggest that they are helpful and safe, well- 26 27 designed trials are needed. Future studies must be randomized, employ a placebo arm, use 28 29 standardized definitions of sleep related outcomes and evaluate sleep using objective measures 30 31 32 such as polysomnography.

41 on September 30, 2021 by guest. Protected copyright. 42 variety of pharmacologic agents are utilized in the majority of these patients with the 43 44 45 presumption that they improve the quality and quantity of sleep, however this review suggests 46 47 that the low to moderate quality evidence available does not support this notion. While drug 48 49 50 therapy may be associated with falling asleep faster (although not consistently demonstrated) 51 52 there is no evidence that drug therapy improves the more important outcomes of sleep efficiency 53 54 or quality. This may be due to the fact that drug therapy is truly ineffective or that the quality of 55 56 57 existing trials is inadequate to show a difference. It is evident that larger, better quality trials are 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 52

20 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 needed to definitively answer this question. Future trials must use standardized definitions for 4 5 6 sleep related outcomes, objectives methods for measurement, a placebo arm and systematically 7 8 evaluate risk. At this time there is insufficient evidence to support the current usage of 9 10 11 pharmacotherapy for the treatment of sleep disturbances among hospitalized adults. 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

33 http://bmjopen.bmj.com/ 34 35 36 37 Funding: 38 39 None 40

22 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 References 4 5 6 7 1. Young JS, Bourgeois JA, Hilty DM, Hardin KA. Sleep in hospitalized medical patients, part 8 1: factors affecting sleep. Journal of hospital medicine : an official publication of the Society of 9 Hospital Medicine 2008;3:473-82. 10 11 2. Pisani MA, Friese RS, Gehlbach BK, Schwab RJ, Weinhouse GL, Jones SF. Sleep in the 12 Intensive Care Unit. Am J Respir Crit Care Med 2015. 13 3. Bano M, Chiaromanni F, Corrias M, et al. The influence of environmental factors on 14 sleep quality in hospitalized medical patients. Frontiers in neurology 2014;5:267. 15 For peer review only 16 4. Isaia G, Corsinovi L, Bo M, et al. Insomnia among hospitalized elderly patients: 17 prevalence, clinical characteristics and risk factors. Archives of gerontology and geriatrics 18 2011;52:133-7. 19 20 5. Weinhouse GL, Schwab RJ, Watson PL, et al. Bench-to-bedside review: delirium in ICU 21 patients - importance of sleep deprivation. Crit Care 2009;13:234. 22 6. Kamdar BB, Niessen T, Colantuoni E, et al. Delirium Transitions in the Medical ICU: 23 Exploring the Role of Sleep Quality and Other Factors*. Crit Care Med 2015;43:135-41. 24 25 7. Raymond I, Ancoli-Israel S, Choiniere M. Sleep disturbances, pain and analgesia in adults 26 hospitalized for burn injuries. Sleep medicine 2004;5:551-9. 27 8. Manabe K, Matsui T, Yamaya M, et al. Sleep patterns and mortality among elderly 28 patients in a geriatric hospital. Gerontology 2000;46:318-22. 29 30 9. Morin CM, Jarrin DC. Epidemiology of Insomnia: Prevalence, Course, Risk Factors, and 31 Public Health Burden. Sleep Medicine Clinics 2013;8:281-97. 32 10. Young JS, Bourgeois JA, Hilty DM, Hardin KA. Sleep in hospitalized medical patients, part 33 2: behavioral and pharmacological management of sleep disturbances. Journal of hospital http://bmjopen.bmj.com/ 34 35 medicine : an official publication of the Society of Hospital Medicine 2009;4:50-9. 36 11. Tamrat R, Huynh-Le MP, Goyal M. Non-pharmacologic interventions to improve the 37 sleep of hospitalized patients: a systematic review. Journal of general internal medicine 38 2014;29:788-95. 39 40 12. Flaherty JH. Insomnia among hospitalized older persons. Clinics in geriatric medicine 2008;24:51-67, vi. 41 on September 30, 2021 by guest. Protected copyright. 42 13. Richardson WS, Wilson MC, Nishikawa J, Hayward RS. The well-built clinical question: a 43 key to evidence-based decisions. ACP J Club 1995;123:A12-3. 44 45 14. Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Self-reported and measured sleep 46 duration: how similar are they? Epidemiology 2008;19:838-45. 47 15. Cespedes EM, Hu FB, Redline S, et al. Comparison of Self-Reported Sleep Duration With 48 Actigraphy: Results From the Hispanic Community Health Study/Study of Latinos Sueno 49 50 Ancillary Study. Am J Epidemiol 2016;183:561-73. 51 16. Morrison A, Polisena J, Husereau D, et al. The effect of English-language restriction on 52 systematic review-based meta-analyses: a systematic review of empirical studies. International 53 journal of technology assessment in health care 2012;28:138-44. 54 55 17. Sampson M, McGowan J, Cogo E, Grimshaw J, Moher D, Lefebvre C. An evidence-based 56 practice guideline for the peer review of electronic search strategies. J Clin Epidemiol 57 2009;62:944-52. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 52 BMJ Open

23 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 18. Balshem H, Stevens A, Ansari M, et al. Finding Grey Literature Evidence and Assessing 4 5 for Outcome and Analysis Reporting Biases When Comparing Medical Interventions: AHRQ and 6 the Effective Health Care Program. Methods Guide for Effectiveness and Comparative 7 Effectiveness Reviews. Rockville (MD)2008. 8 19. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for 9 10 systematic reviews and meta-analyses: the PRISMA statement. PLoS medicine 11 2009;6:e1000097. 12 20. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, 13 range, and the size of a sample. BMC Med Res Methodol 2005;5:13. 14 15 21. Wan X, WangFor W, Liu peer J, Tong T. Estimating review the sample mean only and standard deviation from 16 the sample size, median, range and/or interquartile range. BMC Med Res Methodol 17 2014;14:135. 18 19 22. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for 20 assessing risk of bias in randomised trials. BMJ 2011;343:d5928. 21 23. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in 22 meta-analyses. 2013. (Accessed 28 Jan, 2015, at 23 24 http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp.) 25 24. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in 26 medicine 2002;21:1539-58. 27 25. SW R. Analyzing Effect Sizes: Random Effects Models: New York: Russell Sage 28 29 Foundation; 2009. 30 26. Viechtbauer W. Bias and Efficiency of Meta-Analytic Variance Estimators in the Random- 31 Effects Model. Journal of Educational and Behavioral Statistics 2005;30:261-93. 32 27. Bourne RS, Mills GH, Minelli C. Melatonin therapy to improve nocturnal sleep in critically

33 http://bmjopen.bmj.com/ 34 ill patients: encouraging results from a small randomised controlled trial. Crit Care 2008;12:R52. 35 28. Ibrahim MG, Bellomo R, Hart GK, et al. A double-blind placebo-controlled randomised 36 pilot study of nocturnal melatonin in tracheostomised patients. Critical care and resuscitation : 37 journal of the Australasian Academy of Critical Care Medicine 2006;8:187-91. 38 39 29. Kondili E, Alexopoulou C, Xirouchaki N, Georgopoulos D. Effects of propofol on sleep 40 quality in mechanically ventilated critically ill patients: a physiological study. Intensive Care Med

41 2012;38:1640-6. on September 30, 2021 by guest. Protected copyright. 42 30. Engelmann C, Wallenborn J, Olthoff D, Kaisers UX, Ruffert H. Propofol versus 43 44 flunitrazepam for inducing and maintaining sleep in postoperative ICU patients. Indian journal 45 of critical care medicine : peer-reviewed, official publication of Indian Society of Critical Care 46 Medicine 2014;18:212-9. 47 31. Oto J, Yamamoto K, Koike S, Imanaka H, Nishimura M. Effect of daily sedative 48 49 interruption on sleep stages of mechanically ventilated patients receiving midazolam by 50 infusion. Anaesthesia and intensive care 2011;39:392-400. 51 32. Alexopoulou C, Kondili E, Diamantaki E, et al. Effects of dexmedetomidine on sleep 52 quality in critically ill patients: a pilot study. Anesthesiology 2014;121:801-7. 53 54 33. Hardin KA, Seyal M, Stewart T, Bonekat HW. Sleep in critically ill chemically paralyzed 55 patients requiring mechanical ventilation. Chest 2006;129:1468-77. 56 34. Feldmeier C, Kapp W. Comparative clinical studies with midazolam, oxazepam and 57 placebo. Br J Clin Pharmacol 1983;16 Suppl 1:151S-5S. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 52

24 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 35. Gallais H, Casanova P, Fabregat H. Midazolam and oxazepam in the treatment of 4 5 insomnia in hospitalized patients. Br J Clin Pharmacol 1983;16 Suppl 1:145S-9S. 6 36. Lupolover R, Ballmer U, Helcl J, Escher J, Pavletic B. Efficacy and safety of midazolam and 7 oxazepam in insomniacs. Br J Clin Pharmacol 1983;16 Suppl 1:139S-43S. 8 37. Morgan PJ, Chapados R, Chung FF, Gauthier M, Knox JW, Le Lorier J. Evaluation of 9 10 zolpidem, triazolam, and placebo as hypnotic drugs the night before surgery. J Clin Anesth 11 1997;9:97-102. 12 38. Goetzke E, Findeisen P, Welbers IB. Comparative study on the efficacy of and the 13 tolerance to the triazolodiazepines, triazolam and brotizolam. Br J Clin Pharmacol 1983;16 14 15 Suppl 2:407S-12S.For peer review only 16 39. Kim SJ, Park J, Lee YJ, et al. The Optimal Dose of Midazolam for Promoting Sleep in 17 Critically Ill Patients: A Pilot Study. Korean J Crit Care Med 2014;29:166-71. 18 19 40. Oto J, Yamamoto K, Koike S, Onodera M, Imanaka H, Nishimura M. Sleep quality of 20 mechanically ventilated patients sedated with dexmedetomidine. Intensive Care Med 21 2012;38:1982-9. 22 41. Li Pi Shan RS, Ashworth NL. Comparison of lorazepam and zopiclone for insomnia in 23 24 patients with stroke and brain injury: a randomized, crossover, double-blinded trial. American 25 journal of physical medicine & rehabilitation / Association of Academic Physiatrists 26 2004;83:421-7. 27 42. Rosenberg RP. Sleep maintenance insomnia: strengths and weaknesses of current 28 29 pharmacologic therapies. Annals of clinical psychiatry : official journal of the American Academy 30 of Clinical Psychiatrists 2006;18:49-56. 31 43. Berry RB, Budhiraja R, Gottlieb DJ, et al. Rules for scoring respiratory events in sleep: 32 update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. Deliberations

33 http://bmjopen.bmj.com/ 34 of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. Journal 35 of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep 36 Medicine 2012;8:597-619. 37 44. Achermann P, Borbély AA. Chapter 37 - Sleep Homeostasis and Models of Sleep 38 39 Regulation. In: Dement MHKRC, ed. Principles and Practice of Sleep Medicine (Fifth Edition). 40 Philadelphia: W.B. Saunders; 2011:431-44.

41 45. Poyares D, Guilleminault C, Ohayon MM, Tufik S. Chronic benzodiazepine usage and on September 30, 2021 by guest. Protected copyright. 42 withdrawal in insomnia patients. Journal of psychiatric research 2004;38:327-34. 43 44 46. Dimsdale JE, Norman D, DeJardin D, Wallace MS. The effect of opioids on sleep 45 architecture. Journal of clinical sleep medicine : JCSM : official publication of the American 46 Academy of Sleep Medicine 2007;3:33-6. 47 47. Borbely AA, Mattmann P, Loepfe M, Strauch I, Lehmann D. Effect of benzodiazepine 48 49 hypnotics on all-night sleep EEG spectra. Human neurobiology 1985;4:189-94. 50 48. Brunner DP, Dijk DJ, Tobler I, Borbely AA. Effect of partial sleep deprivation on sleep 51 stages and EEG power spectra: evidence for non-REM and REM sleep homeostasis. 52 Electroencephalography and clinical neurophysiology 1990;75:492-9. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from sedatives, analgesics Concomitant

unit Setting care hospital Critical careCritical unit Critical careCritical unit Mechanically Critical careCritical unit patients ICU with Surgical Surgical step down Rehab unit inunit Rehab acute

b c c b b a

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml atients P For peer review only

(N) Placebo (n=18) Placebo (n=12) Midazolam continuous infusion (n=11) Flunitrazepam (n=32) Lorazepam orally (n=10) Melatonin (n=14) Melatonin (n=12) Midazolam infusion daytime with interruption (n=11) Propofol infusion (n=7) Propofol infusion (n=34) (n=6) No Treatment No Treatment Author, year Intervention and Li Pi etShan Li al., 2004 Ibrahim et Ibrahim al., 2006 Bourne et 2008 al., Otoal., et 2011 Kondili Kondili et al., 2012 (cross over) Engelmann et al., 2014 Trials RandomizedControlled Study Characteristics: Table1: Page 25 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 26 of 52 26 26 None None Reported None None Reported None None Reported None None Reported None None reported insomnia seeking hypnotic sleep aid hypnotic hypnotic due to sleep difficulties insomnia elective surgeryelective stroke/ acquiredstroke/ secondary insomnia patients patients with patients patients requiringpatients a patients patients with patients patients undergoing brain injury, brain injury, Hospital ward Hospitalized Acute careAcute unit Hospitalized Acute careAcute unit Hospitalized Acute careAcute unit Hospitalized adult reported reported reported reported BMJ Open 5 nights / 5 nights not 5 nights / 5 nights not 7 nights / 7 nights not http://bmjopen.bmj.com/ 14 nights / 14not nights 24 / 8 hours hours care Acute unit Hospitalized

b b b

Patient patients patients patients sedation ventilated ICU ventilated ICU ventilated ICU ventilated ICU Characteristics patients requiringpatients continuous sedation patients requiringpatients IV

Setting Critical careCritical unit Mechanically

b b b

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

Hardin et al., 2006 (Cohort) Otoal., et 2012 (Case Series) Kim et al.,Kim 2014 (Case Series) Alexopoulou etAlexopoulou al., 2014 (cohortwith cross over) Page 27 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 28 of 52 28 28

nd nd and 2 st

night Measure of Effect rd night ; p=0.001night 2 between and and 3

Sleep EfficiencySleep Correlation between dose of andmidazolam Total Sleep (r=0.975,Time p=0.005) Correlation between dosemidazolam and stage II (r=0.975, p=0.005) FragmentationSleep StageSleep Distribution I : Stage p=0.006 IIStage : p=0.006 IIIStage : p=0.180 REM : p=0.173 p=0.002 1 between p = 0.0023 a

a a

a a a a group a a Sleep Efficiency Sleep 7.1 (6.1-13.4) arousals/hr 15.8% (6.4-51.6%) StageSleep Distribution (%) Sleep Efficiency Sleep 34.3% (8-59.7) StageSleep Distribution (min) I =82.0Stage (60.5-372.5) IIStage =88.0 (19.0-621.0) IIIStage =0.0 (0-0) REM=10.0 (6.0-50.5) FragmentationSleep 77.9% (65.6-80.2%) StageSleep Distribution(%) Stage I =56.2Stage (24.7-79.3) IIStage =39.2 (20.7-66.4) IIIStage =0.0 (0-0) REM=0.0 (0-0.4) FragmentationSleep Stage I=16.1Stage (6.2-21.3) IIStage =78.7 (69.2-92.5) IIIStage =0.0 (0-0) REM=0.0 (0-0.4) FragmentationsSleep 2.2 (1.6-4.5) arousals/hr Sleep Efficiency Sleep

BMJ Open http://bmjopen.bmj.com/ reatment on day 1, dexmedetomidine on day 2) Polysomnography Polysomnography (no t Polysomnography

a Mechanically Ventilated Patients in an Intensive Care Unit Care in an Intensive Patients Ventilated Mechanically Treatment group Method of evaluation sleep by Outcome treatment For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml (n=16 enrolled, 13 analyzed) Midazolam IV 0.02(0.015- 0.04) mg/kg/hr (n=9 enrolled, 5 analyzed) Dexmedetomidine IV (0.6infusion [0.4-0.7] mcg/kg/hr) (n=16 enrolled, 10 analyzed) For peer cross Noover treatment review only

Author, year byOutcome of Study Findings, Summary Table3: Kim et al.,Kim 2014 Alexopoulou etAlexopoulou al., 2014 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 29 29 Sleep EfficiencySleep p = 0.047 FragmentationSleep Sleep EfficiencySleep FragmentationSleep StageSleep Distribution I: pStage = 1.0 II: Stage p = 0.66 III/IV:Stage p = 0.75 pREM: = 0.04 p = 0.03 p = 0.37 p = 0.33 None None reported

a a

a a a a

a a a a 188 (136-449) = a 16.1(7.6-28.6) arousals/hr 4.4 (2.5-8.3) arousals/hr Sleep Efficiency Sleep 62.6% (13.1-85.9) FragmentationSleep 8.1 (2.9-16.2) arousals/hr StageSleep Distribution I =Stage 30.7% (4.6-66.7) IIStage = 46.1% (3.0-80.4) III/IVStage = 0% (0-00) REM = 1.4% (0-13) Sleep Efficiency Sleep 76.3% (28.4-96.9) FragmentationSleep 4.8 (1.3-14.6) arousals/hr StageSleep Distribution I =Stage 20.8% (5.6-80.6) IIStage = 48.9% (4.8-84) III/IVStage = 0% (0-5.8) REM = 0% (0-0) Sleep Efficiency Sleep 52.3% (47-89.7) FragmentationSleep 9.3(3-19.5) arousals/hr StageSleep Distribution (min) I=76Stage (32-145) II Stage 81% (73.3-91.1%) FragmentationSleep Stage III=0Stage (0.0-1.3) REM=0.0 (0-0) Sleep Efficiency Sleep

BMJ Open http://bmjopen.bmj.com/ Polysomnography Polysomnography Polysomnography

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml sedation sedation interruption (n=11 11randomized, analyzed) Midazolam infusion 0.6 (0.5-0.9)mg/kg

Propofol 0.86 infusion (0.67 - 1.25) mg/kg/h (n=7 randomized, 6 analyzed) Dexmedetomidine IV 0.2-0.7infusion mcg/kg/hr (n=10 enrolled, 10 analyzed) (n=6 randomized, 6 analyzed) For peer reviewNo propofol cross over only

Otoal., et 2011 Kondili Kondili et al., 2012 Otoal., et 2012 Page 29 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 30 of 52 30 30

= 0.98 Sleep Efficiency Sleep BIS: p = 0.09 pActigraphy: = 0.84 Patient self-report: p = 0.32

Sleep Efficiency Sleep Sleep EfficiencySleep StageSleep Distribution I: pStage = 0.46 II: Stage p = 0.55 III/IVStage : p = 0.03 pREM: = 0.01 Nurse observation:Nurse p = 0.58 p a

d b b c c a

a a leep leep Stage Distribution Sleep Efficiency Sleep S Sleep Efficiency Sleep 50% (0-72%) 50% (16-69) Sleep Efficiency Sleep BIS: 26% (17-36%) Actigraphy: 75% (67-83) 68) Patient self-report: 50% (43- 58) BIS: 39% (27-51) Actigraphy: 73% (53-93) 64) Patient self-report: 41% (24- 59) 97% (91.1-97.8%) FragmentationSleep 2.2 (0-2.6) arousals/hr StageSleep Distribution (min) I Stage = 180 (54-342) IIStage = 282(66-468) III/IVStage = 0(0-0) REM = 0 (0-9) (min) I Stage = 96 (60-144) IIStage = 264 (222-360) III/IVStage = 1.8 (0.6-4.2) REM = 30 (12-48) Sleep Efficiency Sleep Sleep Efficiency Sleep Sleep Efficiency Sleep Nurse observation:Nurse 51% (35- Nurse observation:Nurse 45% (26- BMJ Open http://bmjopen.bmj.com/ ctigraphy Observation by Observation nurse Patient self report questionnaire

BIS A Nurse observationNurse on September 30, 2021 by guest. Protected copyright. continuous continuous a a

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml infusion (n=11 11randomized, analyzed) Intermittent Intermittent Lorazepam IV: Polysomnography Placebo orally (n=18 18randomized, analyzed) Melatonin 3mg orally (n=14 14randomized, analyzed) Placebo (n=12 12randomized, analyzed) Melatonin 10mg Melatonin orally10mg (n=13 12randomized, analyzed) Midazolam infusion 2.7 For (2.0-3.4)mg/kg peer review only

Hardin et al., 2006 Ibrahim et Ibrahim al., 2006 Bourne et 2008al., 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 31 31 Sleep EfficiencySleep BIS: p = 0.777 Sleep p=0.623Diary: FragmentationSleep StageSleep Distribution I: pStage = 0.40 II: Stage p = 0.20 III/IV:Stage p = 0.93 reportednot REM:

p = 0.07 p = 0.90

d d d

b d d N/A =

d d d leep leep Fragmentation Sleep Efficiency Sleep Sleep Efficiency Sleep Sleep Efficiency Sleep Stage I=7.4±5.9Stage II=48.8±Stage 22.3 III/IVStage =38.9±28.1 REM =78.8±19.6% FragmentationSleep 58.8±26.3% FragmentationSleep BIS: 94.7% (30.7-100) (median):Sleep diary 86% 4.79±3.7 arousals/hr StageSleep Distribution (%) I=20.8±35.5Stage II=28.34±Stage 35 III/IV=49.6±49Stage REM=N/A Stage I =5.8±6.6Stage II=57.7±Stage 23.8 III/IVStage =31.9±24.6 REM=3.6±5.7 4.4±3.7 arousals/hr StageSleep Distribution (%) 4.24±2.4 arousals/hr StageSleep Distribution (%) 4226.7% ± S

BMJ Open http://bmjopen.bmj.com/ BIS 31only in (analysis flunitrazepam patients)

propofol and patients 25

d d d on September 30, 2021 by guest. Protected copyright. Hospitalized Patients Not in an Intensive Care Unit Unit Care Not in an Intensive Patients Hospitalized For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 0.04± 0.04 mg/kg/day Propofol 2mg/kg/hr (n=34 34randomized, analyzed) Continuous Continuous Lorazepam IV with neuromuscular blockade 0.62±0.20 mg/kg/day (n=6 randomized, 6 analyzed) Continuous Continuous Lorazepam IV 0.72±0.39 mg/kg/day (n=6 randomized, 6 analyzed) (n=6 randomized, 6 analyzed) For peer review only

Engelmann et al., 2014 Page 31 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 32 of 52 32 32

QualitySleep Conditioning thein morning pAlertness: = 0.60 pTiredness: = 0.29 Total Time Sleep FragmentationSleep BIS: p = 0.041 StageSleep Distribution BIS p=0.016 values: p Stage = A: 0.044 B: Stage p = 0.004 Stage pC: = 0.69 QualitySleep Sleep pDiary: < 0.001 p = 0.09 p = 0.17

b b

d b a

a Total Time sleep FragmentationSleep BIS (median): 0.6 arousals/hr Total Time Sleep Alertness Alertness = 9 (8-10) =Tiredness 7.5 (5-10) 443±37.8 min 469±46.2 min QualitySleep 8.5 (7.5-10) Conditioning thein morning BIS: 92.7% (16.7-100) (median):Sleep diary 71% FragmentationSleep BIS (median): 1.1 StageSleep Distribution BIS values=78.7 (72.05-81) 30%Stage A= (4.0-81.3) B=Stage 55.4% (96-90) Stage 7.6%C= (0-83.7) QualitySleep Sleep (median):Diary 3/5 StageSleep Distribution BIS values=74.05 (65.68- 79.38) 7.4%Stage A= (1.1-98) B=Stage 36.6% (5.4-82.3) Stage 5.0%C= (0-50.7) QualitySleep Sleep (median)Diary = 2/5 Sleep Efficiency Sleep BMJ Open http://bmjopen.bmj.com/ leep diary leep(PgShD) diary used to S Patient questionnaire (10 insleep, thealertness morning) evaluate sleep quality, sleep andefficiency sleep fragmentation recordedSleep quality on 5 Nurse observationNurse point scale for quality of point scale

on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Lorazepam 0.5-1.0mg (n=10 9randomized, analyzed) Zopiclone Zopiclone 3.75-7.5mg (n=10 9randomized, For peer Flunitrazepam Bolus 0.015mg/kg (32 randomized, 32 analyzed) review only

Li Pi etShan Li al., 2004 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 33 33 Sleep LatencySleep Both triazolam and zolpidem groups significantly were QualitySleep Both triazolam and zolpidem groups significantly were Morning Sleepiness better than placebobetter than (p<0.001) significantlynot but different between other.each placebobetter than (p<0.001) significantlynot but different between other.each differenceNo significant between (pgroups = 0.443)

(mean±SD)

(mean±SD) a

(mean±SD) (mean±SD)= a leep leep Quality S 2.0±0.1 Conditioning thein morning 62.8±2.5 Sleep Latency Sleep Median: Median: 60 minutes <30 min=22% <60 min=49% <90 min=64% <120 min=72% Sleep Latency Sleep Median: Median: 25 min <30 min=42% <60 min=78% <90 min=92% <120 min=89% QualitySleep 2.0±0.1 Conditioning thein morning 58.4±2.7 Sleep LatencySleep Median: 30 minutes <30 min=45% <60 min=76% <90 min=87% <120 min=90% QualitySleep Alertness Alertness = 9 (6.5-10) =Tiredness 8 (5.5-8.5) 8 (5-9) Conditioning thein morning BMJ Open http://bmjopen.bmj.com/ hours post hours dose only) Sleep quality: self-patient report 1=excellent 2=good 3=fair 4=poor Conditioning in the morning report)(Patient self Visual analog scale (0-100 sleepy0=very where and 100=not at all sleepy) Nurse observationNurse (first 2 on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml analyzed) Triazolam 0.25mgTriazolam (n=119 119randomized, analyzed) Placebo (n=118 118randomized, analyzed) For peer review 10mgZolpidem (n=120 120randomized, analyzed) only

Morgan 1997 al., et Page 33 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 34 of 52 34 34 Total TimeSleep Reported no significant as difference between groups LatencySleep (<20min) p < 0.05 for both midazolam and oxazepam vs placebo

Not reported

430 min

: : 440 min : : : NR d d

d d

(mean±SD)

(mean±SD)= leep leep Quality Total Time Sleep Sleep Latency Sleep Total Time Sleep Latency:Sleep 0% asleep20 within min 29% asleep40 min within Total Time Sleep Sleep Latency Sleep S (<30 55.5 min): patients (70%) (<30 55.5 min): patients (70%) QualitySleep Very goodVery =17% Good =60%, Moderate =18% Poor =5% Conditioning thein morning Refreshed =47% Little tired =44% tiredVery =9% Latency:Sleep 16% asleep20 min within 73% asleep40 min within 2.7±0.1 Conditioning thein morning 59.0±2.9 Latency:Sleep 30% asleep20 min within 84% asleep40 min within BMJ Open http://bmjopen.bmj.com/ Observation by Observation nurse Patient questionnaire Frequencies (%) are approximated a barfrom graph Nurse observationNurse on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Brotizolam Brotizolam 0.25mg orally (n=86 79randomized, breakdownbut analyzed by group reported)not Midazolam orally15mg (n=19 19randomized, analyzed) Triazolam 0.25mgTriazolam orally Placebo (n=16 16randomized, analyzed) For 15mgOxazepam orally peer(n=15 15randomized, analyzed) review only

leep leep Quality Total TimeSleep <360 min=10% 360-420 min=53% >420 min=37% LatencySleep <30 min=63% 30-50 min=21% >50 min=16% <360 min=12% 360-420 min=52% >420 min=36% Total Time Sleep Sleep Latency Sleep S (<30 min): 36.5 (46%)patients QualitySleep Very goodVery =10% Good =32% Moderate =30% Poor =28% Conditioning thein morning Refreshed =31% Little tired =45% tiredVery =24% Very goodVery =18% Good =55% Moderate =21% Poor =6% Conditioning thein morning Refreshed =49% Little tired =43% tiredVery =8% BMJ Open http://bmjopen.bmj.com/ Third party observationThird party (values are approximated a graph)from on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Midazolam orally15mg (n=19 18randomized, analyzed) Oxazepam 50mgOxazepam orally (n=20 17randomized, analyzed) For Placebo peer review only

Gallais etGallais al., 1983 Page 35 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 36 of 52 36 36

: : Total Sleep Time Total Time Sleep Latency Sleep FragmentationSleep

oxazepam oxazepam and placebo groups. p > 0.05 p > 0.05 Not reported

leep leep Latency Total TimeSleep 360 min=9.9% 420 min=41.6% 480 min=25.9% 540 min=10.9% 600 min=0.6% LatencySleep Total Sleep Time Total Time Sleep <360 min=30% 360-420 min=54% >420 min=16% Latency Sleep (arousals/night) 1-2=60.4% >2=23.7% Total TimeSleep 360 min=13.8% 420 min=22.5% 480 min=30.8% 540 min=14.3% 600 min=3.6% latencySleep <20 min=34.4% 20-40 min=48.7% 40-60 min=12.3% >60 min=4.7% FragmentationSleep <30 min=10% 30-50 min=50% >50 min=40% S <30 min=17% 30-50 min=44% >50 min=39% None=15.9% BMJ Open http://bmjopen.bmj.com/

0-40 min=42.7% <20 min=15.2% 2 40-60 min=22% >60 min=20.1% FragmentationSleep (arousals/night) 1-2=55.4% >2=27.1% None=17.5% BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

For peerdeviation standard ± review only data presented as median (range) median (range) as presented data data presented as mean as presented data data presented as median (interquartile range) median (interquartile as presented data CI) (95% mean as presented data

a c d Sleep DiaryPittsburg – PgShD b NR = not reported reported not = NR Page 37 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 38 of 52 *=non-randomized study; IV=intravenous; study; IV=intravenous; *=non-randomized

BMJ Open http://bmjopen.bmj.com/ als using the Cochrane Risk of Bias Tool Tool Risk of Bias Cochrane the alsusing on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only active and from and studies randomized non-randomized from and differences Observed 5: Fragmentation: Sleep Figure NMBA=neuromuscular IV=intravenous; study; *=non-randomized methods section. inthe described an approached using estimated and active studies from and randomized and randomized non from Observed differences 6: Architecture: Sleep Figure patients) 59 studies, (4 studies controlled inactive SD and values mean some studies, stage. In ineach spent patients time of sleep total percentage as defined distribution stage Sleep cont=continuous study; IV=intravenous; section.*=non-randomized in methods described the an using approached estimated were Selection and of 1: Study Identification Process Figure sleep tomeasure polysomnography studies using not cohort *Prospective ControlledTri Randomized 2: for Risk of Bias Figure Risk of bias assessed as low, moderate or high for each category. each category. for high or moderate as low, of assessed Risk bias Scale) Studies(NewCastle-Ottawa Cohort Prospective 3: for Risk of Bias Figure and inactive active and studies from and non-randomized randomized from differences Observed 4: Efficiency: Sleep Figure patients) 188 studies, (7 studies controlled and mean studies, some observation. of hours In total the divided by asleep hours spent of number definedthe as was Efficiency Sleep section. methods inthe described an approached estimatedusing were values SD Figure Legends: Figure inactive controlled studies (5 studies, 125 patients) patients) 125 studies, (5 studies controlled inactive valueswere SD and mean some sleep.studies, of In perhour awakenings or of number arousals the as was defined fragmentation Sleep NMBA=neuromuscular MID=midazolam; LOR=lorazepam; DEX=dexmedatomidine; PRO=propafol; dosing; int=intermittent dosing;

NMBA=neuromuscular blocking agent; SD=standard deviation agent; SD=standard blocking NMBA=neuromuscular estimable NE=not reported; NR=not deviation; agent; SD=standard blocking blocking agent; No Trt=no treatment; MD=mean difference; SD=standard deviation; NE=not estimable estimable deviation; NE=not SD=standard difference; MD=mean treatment; Trt=no agent; No blocking 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 39 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 Figure 1: Process of Study Identification and Selection *Prospective cohort studies not using polysomnography to measure sleep 48 49 50 215x279mm (300 x 300 DPI) 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 40 of 52 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 Figure 2: Risk of Bias for Randomized Controlled Trials using the Cochrane Risk of Bias Tool Risk of bias assessed as low, moderate or high for each category. 48 49 50 215x279mm (300 x 300 DPI) 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 41 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 Figure 3: Risk of Bias for Prospective Cohort Studies (NewCastle-Ottawa Scale)

48 215x279mm (300 x 300 DPI) 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 42 of 52 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 Figure 4: Sleep Efficiency: Observed differences from randomized and non-randomized studies and from active and inactive controlled studies (7 studies, 188 patients) 48 Sleep Efficiency was defined as the number of hours spent asleep divided by the total hours of 49 observation. In some studies, mean and SD values were estimated using an approached described in the 50 methods section. *=non-randomized study; IV=intravenous; NMBA=neuromuscular blocking agent; 51 SD=standard deviation 52 53 54 215x279mm (300 x 300 DPI) 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 43 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 Figure 5: Sleep Fragmentation: Observed differences from randomized and non-randomized studies and from active and inactive controlled studies (5 studies, 125 patients) 48 Sleep fragmentation was defined as the number of arousals or awakenings per hour of sleep. In some 49 studies, mean and SD values were estimated using an approached described in the methods section. 50 *=non-randomized study; IV=intravenous; NMBA=neuromuscular blocking agent; SD=standard deviation; 51 NR=not reported; NE=not estimable 52 53 54 215x279mm (300 x 300 DPI) 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 44 of 52 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 30, 2021 by guest. Protected copyright. 42 43 44 45 46 47 Figure 6: Sleep Architecture: Observed differences from randomized and non randomized studies and from active and inactive controlled studies (4 studies, 59 patients) 48 Sleep stage distribution defined as percentage of total sleep time patients spent in each stage. In some 49 studies, mean and SD values were estimated using an approached described in the methods section.*=non- 50 randomized study; IV=intravenous; cont=continuous dosing; int=intermittent dosing; PRO=propafol; 51 DEX=dexmedatomidine; LOR=lorazepam; MID=midazolam; NMBA=neuromuscular blocking agent; No 52 Trt=no treatment; MD=mean difference; SD=standard deviation; NE=not estimable 53 54 55 215x279mm (300 x 300 DPI) 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 45 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 Appendix 1 4 5 6 Database: Embase Classic+Embase <1947 to 2015 January 27>, Ovid MEDLINE(R) In-Process 7 & Other Non-Indexed Citations and Ovid MEDLINE(R) <1946 to Present> 8 Search Strategy: March 21, 2016 9 ------10 11 1 sleep/ or night sleep/ or sleep induction/ (119114) 12 2 exp *sleep disorder/ (48385) 13 3 (sleep or insomnia).tw. (281519) 14 4 1 or 2 or 3 (318752) 15 5 intensive careFor unit/ (128046) peer review only 16 6 *critical illness/ or *critically ill patient/ or *intensive care/ or *artificial ventilation/ 17 18 (107617) 19 7 *hospital patient/ (13433) 20 8 *postoperative care/ (25079) 21 9 *postoperative period/ (6176) 22 10 ((mechanical or artificial) adj ventilat$).tw. (76070) 23 24 11 (intensive care or icu).tw. (260630) 25 12 (critical$ adj2 (care or ill$)).tw. (116727) 26 13 (postoperativ$ or post operativ$ or inpatient$).tw. (1129711) 27 14 (hospital$ adj2 patient$).tw. (173950) 28 15 or/5-14 (1679818) 29 16 4 and 15 (15024) 30 31 17 exp *benzodiazepine derivative/ (64289) 32 18 *ramelteon/ (202)

41 26 Imidazopyridine$.tw. (910) on September 30, 2021 by guest. Protected copyright. 42 27 *zolpidem/ (1381) 43 44 28 Zolpidem.tw. (4150) 45 29 *pyridine derivative/ (12737) 46 30 *pyrazolopyrimidine derivative/ (163) 47 31 Pyrazolopyrimidine$.tw. (540) 48 32 *zaleplon/ (238) 49 50 33 Zaleplon.tw. (669) 51 34 Cyclopyrrolone$.tw. (250) 52 35 *eszopiclone/ (164) 53 36 eszopiclone.tw. (472) 54 37 *zopiclone/ (826) 55 38 zopiclone.tw. (1858) 56 57 39 exp *barbituric acid derivative/ (69313) 58 40 *chloral hydrate/ (3654) 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 46 of 52 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 41 exp *neuroleptic agent/ (193754) 4 5 42 *olanzapine/ (5515) 6 43 olanzapine.tw. (15748) 7 44 *quetiapine/ (2936) 8 45 Quetiapine.tw. (8643) 9 46 *propofol/ (21766) 10 11 47 Propofol.tw. (36217) 12 48 *alpha 2 adrenergic receptor stimulating agent/ (1054) 13 49 *dexmedetomidine/ (3961) 14 50 Dexmedetomidine.tw. (5629) 15 51 exp *hypnoticFor sedative peer agent/ (137181) review only 16 52 estazolam/ (1148) 17 18 53 midazolam/ (41684) 19 54 temazepam/ (5932) 20 55 triazolam/ (6772) 21 56 flurazepam/ (5391) 22 57 flunitrazepam/ (10031) 23 24 58 (Estazolam or Midazolam or Temazepam or Triazolam or Flurazepam or 25 Flunitrazepam).tw. (36087) 26 59 (amitriptyline or nortriptyline or Mirtazapine or Trazodone or Doxepin).tw. (24214) 27 60 *amitriptyline/ or *nortriptyline/ or *mirtazapine/ or *trazodone/ or *doxepin/ (23976) 28 61 (Diphenhydramine or chlorpheniramine or hydroxyzine).tw. (13454) 29 62 Diphenhydramine.tw. (7480) 30 31 63 chlorpheniramine.tw. (4061) 32 64 hydroxyzine.tw. (2544)

41 72 (sleep or insomnia).tw. (281519) on September 30, 2021 by guest. Protected copyright. 42 73 or/70-72 (419131) 43 44 74 exp Intensive Care Units/ (149137) 45 75 exp Critical Care/ or exp Intensive Care/ or critical illness/ (556202) 46 76 Respiration, Artificial/ (129971) 47 77 *inpatients/ or *Hospitalization/ (69809) 48 78 postoperative care/ or postoperative period/ (293863) 49 50 79 ((mechanical or artificial) adj ventilat$).tw. (76070) 51 80 (intensive care or icu).tw. (260630) 52 81 (critical$ adj2 (care or ill$)).tw. (116727) 53 82 (postoperativ$ or post operativ$ or inpatient$).tw. (1129711) 54 83 (hospital$ adj2 patient$).tw. (173950) 55 84 or/74-83 (2184689) 56 57 85 73 and 84 (38589) 58 86 exp Benzodiazepines/ (216012) 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 47 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 87 Receptors, Melatonin/ag (122) 4 5 88 ramelteon.tw. (515) 6 89 Melatonin/ (40427) 7 90 exp Antidepressive Agents/ (454850) 8 91 Histamine Antagonists/ (41545) 9 92 Imidazopyridine$.tw. (910) 10 11 93 Zolpidem.tw. (4150) 12 94 Pyridines/ (63930) 13 95 Pyrazolopyrimidine$.tw. (540) 14 96 Zaleplon.tw. (669) 15 97 Cyclopyrrolone$.tw.For (250)peer review only 16 98 eszopiclone.tw. (472) 17 18 99 zopiclone.tw. (1858) 19 100 exp Barbiturates/ (203683) 20 101 Chloral Hydrate/ (8173) 21 102 exp Antipsychotic Agents/ (348253) 22 103 olanzapine.tw. (15748) 23 24 104 Quetiapine.tw. (8643) 25 105 Propofol/ (49309) 26 106 Propofol.tw. (36217) 27 107 exp Adrenergic alpha-2 Receptor Agonists/ (24347) 28 108 Dexmedetomidine/ (6363) 29 109 Dexmedetomidine.tw. (5629) 30 31 110 exp "Hypnotics and Sedatives"/ (439023) 32 111 (Estazolam or Midazolam or Temazepam or Triazolam or Flurazepam or

41 118 116 not 117 (7036) on September 30, 2021 by guest. Protected copyright. 42 119 118 use prmz (811) 43 44 120 69 or 119 (2096) 45 121 remove duplicates from 120 (1623) 46 122 121 use prmz (805) Medline 47 123 121 use emczd (818) Embase 48 49 50 Database: EBM Reviews - Cochrane Central Register of Controlled Trials 51 Search Strategy: March 21, 2016 52 ------53 1 exp Sleep/ (4021) 54 2 exp Sleep Disorders/ (3997) 55 3 (sleep or insomnia).tw,hw. (15412) 56 57 4 or/1-3 (15599) 58 5 exp Intensive Care Units/ (2098) 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 48 of 52 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 6 exp Critical Care/ or exp Intensive Care/ or critical illness/ (2235) 4 5 7 Respiration, Artificial/ (2354) 6 8 *inpatients/ or *Hospitalization/ (0) 7 9 postoperative care/ or postoperative period/ (6109) 8 10 ((mechanical or artificial) adj ventilat$).tw,hw. (3634) 9 11 (intensive care or icu).tw,hw. (9709) 10 11 12 (critical$ adj2 (care or ill$)).tw,hw. (3550) 12 13 (postoperativ$ or post operativ$ or inpatient$).tw,hw. (63767) 13 14 (hospital$ adj2 patient$).tw,hw. (10443) 14 15 or/5-14 (82559) 15 16 4 and 15 (1240)For peer review only 16 17 exp Benzodiazepines/ (7485) 17 18 18 Receptors, Melatonin/ag (0) 19 19 ramelteon.tw,hw. (53) 20 20 Melatonin/ (667) 21 21 exp Antidepressive Agents/ (9517) 22 22 Histamine Antagonists/ (156) 23 24 23 Imidazopyridine$.tw,hw. (62) 25 24 Zolpidem.tw,hw. (442) 26 25 Pyridines/ (1773) 27 26 Pyrazolopyrimidine$.tw,hw. (14) 28 27 Zaleplon.tw,hw. (81) 29 28 Cyclopyrrolone$.tw,hw. (29) 30 31 29 eszopiclone.tw,hw. (78) 32 30 zopiclone.tw,hw. (272)

41 38 exp Adrenergic alpha-2 Receptor Agonists/ (2227) on September 30, 2021 by guest. Protected copyright. 42 39 Dexmedetomidine/ (298) 43 44 40 Dexmedetomidine.tw,hw. (743) 45 41 exp "Hypnotics and Sedatives"/ (10693) 46 42 (Estazolam or Midazolam or Temazepam or Triazolam or Flurazepam or 47 Flunitrazepam).tw,hw. (6075) 48 43 Melatonin.tw,hw. (1099) 49 50 44 (amitriptyline or nortriptyline or Mirtazapine or Trazodone or Doxepin).tw,hw. (3497) 51 45 (Diphenhydramine or chlorpheniramine or hydroxyzine).tw,hw. (1352) 52 46 or/17-45 (41116) 53 47 16 and 46 (326) 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 49 of 52 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

Page 50 of 52

on page # # page on Reported Reported 7 7 6 6 4-5 N/A N/A 4-5 4 2 1 1 Appendix Appendix up) report up) characteristics and (e.g., analysis). - BMJ Open http://bmjopen.bmj.com/

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 1 8 2 4 7 9 3 6 # For peer review only 11

Data Data items Data collection Data process 10 Study selection Study Search Search Information sourcesInformation Eligibility criteria Eligibility Protocol and registrationProtocol 5 METHODS METHODS Objectives Objectives Rationale Rationale INTRODUCTION INTRODUCTION Structuredsummary ABSTRACT ABSTRACT Title Title TITLE TITLE Section/topic Section/topic Appendix 2: PRISMA Checklist PRISMA Appendix2: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from 7-8 16 16 9-13 9-13 N/A N/A N/A N/A N/A N/A 24-33 24-33 35 35 21-23 21-23 34 34 7 7 7 7 analysis. - specified. - foreach meta ) ) 2 BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. Discussstudy limitations at and outcome(e.g., level risk review-level ofand at bias), (e.g., retrieval incomplete of identifiedresearch, reporting bias). Summarizethe main findings the strengthincluding ofevidencefor main each outcome; their consider relevance to groups key (e.g., healthcare providers, users, and policy makers). Give results ofresults Give additional analyses, if (e.g.,done sensitivity subgroupor analyses, meta- [see regression Item 16]). Present results Present of any assessmentof risk across of (see Itembias studies 15). Present results Present of each meta-analysis done, including confidence intervalsand measures of consistency. For all outcomes Forall considered(benefits harms),or for study: each present, (a) simple summary for eachdata intervention group (b) effect estimates confidence intervals, and ideally with a forest plot. Present Present dataon risk of ofstudy and, each bias ifavailable, outcome any assessmentlevel (see item 12). For each study, Foreach present characteristics for which data extracted were (e.g., study PICOS,size, follow-upperiod) and providethe citations. Give numbersGive screened,of studies assessed for eligibility, and includedin the review, with reasonsfor exclusions each at stage, ideally flow with a diagram. Describe methods Describe of additional analyses (e.g., sensitivity subgroupor analyses, meta- regression), ifindicating done, prewhich were Specify Specify assessment any riskofthat may bias cumulative affectof evidence(e.g., the publicationbias, selective reporting withinstudies). Describe the Describe methods of handling dataand combining ofresults studies, if done, including measures(e.g., I consistency of State the State principal summarymeasures (e.g., risk ratio,means). difference in Describe methods Describe used for assessing risk studies of of bias individual (including specification of whetherthis done the at was studyoutcome or this level), and how information is be used to in datasynthesis. any

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 12 21 14 17 13 25 16 For peer review only23

Limitations Summaryof evidence 24 DISCUSSION DISCUSSION Riskacross of bias studies 22 Synthesisof results Results of Results individual studies 20 Riskstudies within of bias 19 Study characteristics Study 18 Study selection Study RESULTS RESULTS Riskacross of bias studies 15 Synthesisof results Summarymeasures Riskindividual in of bias studies Additional Additional analysis Additional Additional analyses Page 51 of 52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-012108 on 29 July 2016. Downloaded from Page 52 of N/A N/A 16-17 16-17 BMJ Open http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. Describe sources Describe of funding for systematic the review support and (e.g., other supply of data); of funders role for systematic the review. Provide a a Provide general interpretationof the results thecontext in of other evidence, and implications forfuture research.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 27 26 For peer review only