BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
An Investigation into Drug Products Withdrawn from the EU Market between 2002 and 2011 for Safety Reasons and the Evidence Used to Support the Decision Making
For peer review only Journal: BMJ Open
Manuscript ID: bmjopen-2013-004221
Article Type: Research
Date Submitted by the Author: 10-Oct-2013
Complete List of Authors: McNaughton, Rhian; Drug Safety Research Unit, ; University of Portsmouth, Huet, Gwenaël; Drug Safety Research Unit, Shakir, Saad; Drug Safety Research Unit, ; University of Portsmouth,
Primary Subject Public health Heading:
Secondary Subject Heading: Epidemiology, Evidence based practice, Health services research
Risk management < HEALTH SERVICES ADMINISTRATION & Keywords: MANAGEMENT, Adverse events < THERAPEUTICS, Clinical trials < THERAPEUTICS http://bmjopen.bmj.com/
on September 29, 2021 by guest. Protected copyright.
For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 An Investigation into Drug Products Withdrawn from the EU Market between 2002 4 5 and 2011 for Safety Reasons and the Evidence Used to Support the Decision Making 6 7 8 9 Rhian McNaughton, Gwenaël Huet, Saad Shakir 10 11 Drug Safety Research Unit, Southampton, SO31 1AA UK and University of Portsmouth, 12 13 Portsmouth PO1 2UP, UK Rhian McNaughton Research Fellow Drug Safety Research 14 15 Unit, Southampton,For SO31 peer 1AA UK review Gwenaël Huet Intern only Drug Safety Research Unit, 16 17 Southampton, SO31 1AA UK and University of Portsmouth, Portsmouth PO1 2UP, UK 18 19 Saad Shakir Director 20
21 22 Correspondence to: Rhian McNaughton, Drug Safety Research Unit, Southampton, SO31 1AA UK email: 23 24 [email protected], Tel: (023) 80408623, Fax: (023) 80408609 25 26 27 Keywords: Withdrawal – Drug Product – Pharmacovigilance – Evidence – Safety 28 29 30 31 Word Count - 2710 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 2 of 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 ABSTRACT 4 5 Background The objective of this study was to determine the nature of evidence used to support the 6 7 withdrawal of marketing authorisations of drug products for safety reasons throughout the EU between 8 2002 and 2011. 9 10 Methods Products withdrawn, either by a medicines agency or a marketing authorisation holder, during 11 12 the period 2002-2011 were identified by conducting detailed searches of the World Health Organization 13 (WHO), the EMA (European Medicines Agency) and national medicines agency websites throughout the 14 15 EU plus Norway,For Iceland andpeer Liechtenstein. review The scientific evidence only used to support the decision was 16 17 identified from a search within PubMed, the EMA and national medicines agencies websites. Information 18 about spontaneous case reports entered into EudraVigilance and unavailable on the EMA website, was 19 20 received by email from the EMA. 21 22 Results 19 drugs were withdrawn from the market, throughout the EU, for safety reasons from 2002 to 23 2011. Case reports were cited in 95% of withdrawals (18/19) and case-control studies, cohort studies, 24 25 randomized controlled trials (RCTs) or meta-analysis were cited in 63% of withdrawals (12/19). 26 27 Cardiovascular events or disorders were the main reason for withdrawal (9/19), followed by hepatic 28 disorders (4/19) and neurological or psychiatric disorders (4/19). 29 30 Conclusions This study has shown that the level of evidence used to support drug withdrawal has 31 32 improved during the last ten years, with an increased use of case-control studies, cohort studies, RCTs 33 and meta-analyses. This research has demonstrated that such studies have contributed to decision-making 34 http://bmjopen.bmj.com/ 35 in almost two-thirds of cases. 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 3 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Strengths and limitations of this study 4 5 • This study adds to the knowledge base regarding the level of evidence used to support 6 7 decisions to remove drug products from the market. 8 • Results of this research are in-line with the findings of other study groups who concluded that 9 10 there has been a shift in the level of evidence used e.g. results of randomized controlled trials 11 12 and cohort studies are used to justify a product withdrawal 13 • English and French websites and publications were only searched 14 15 • The INNFor of each drugpeer was included review in the PubMed search only strategy but as various drug names 16 17 exist for the same compound, a potential exists for the exclusion of some publications from 18 19 contributing evidence. 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 INTRODUCTION 4 5 The decision to remove a drug product from the market is not taken lightly either by the regulatory 6 7 authority or the marketing authorisation holder. A number of reasons contribute to the withdrawal such as 8 the emergence of new safety evidence and significant drug interactions. Ultimately, the risk-benefit 9 10 balance is paramount when making such decisions. The regulatory authorities review the evidence, such 11 12 as spontaneous case reports or results from a randomized controlled trial (RCT) and prepare an opinion 13 on whether the marketing authorisations of medicinal products containing that particular drug should be 14 15 maintained, changed,For suspended peer or withdrawn. review The interpretation only of the available safety data by the 16 17 numerous competent authorities across the European Union (EU) is wide-ranging as is the accessibility of 18 information regarding how they reach their conclusion. 19 20 21 22 In 2005, legislation was introduced to ensure marketing authorisation holders complete a risk 23 management plan (RMP) for their newly licensed product incorporating, where appropriate, post 24 25 marketing safety surveillance studies in the general population with regular assessment and appropriate 26 27 reporting to the CHMP [1] . The aim of the RMP is to ensure a more proactive approach to 28 pharmacovigilance by putting in place measures that allow for the early detection and minimization of 29 30 risks throughout a medicines lifecycle. This should, in theory, result in a reduction in the length of time 31 32 for drug withdrawal should a safety issue arise. Risk management planning in the EU has been further 33 enhanced by the introduction of new EU legislation for good pharmacovigilance practices (GVP) which 34 http://bmjopen.bmj.com/ 35 became applicable in July 2012 [2]. 36 37 38 Previous studies looked at the type of evidence used to support the withdrawal of products, for safety 39 40 reasons, in different individual countries and over varying time periods. Arnaiz et al. [3] reported on 41 et al 42 products withdrawn in Spain between 1990 to 1999, Clarke . [4] for products in USA and UK from on September 29, 2021 by guest. Protected copyright. 43 1999 to 2001 and Olivier et al. [5] for medicines in France from 1998 to 2004. They all concluded that 44 45 case reports were the mainstay of evidence used to support withdrawal decisions for pharmacovigilance 46 47 reasons; case-control studies, cohort studies and randomised controlled trials were used as evidence in 48 less than one-third of decisions. Olivier and colleagues revisited the issue six years later, between 2005 49 50 and 2011, and reported a shift in the level of evidence used [6]. They determined that of the 22 active 51 52 substances withdrawn from the French market within this 6 year time frame, 68% of the decisions used 53 multiple sources of evidence including clinical trials and pharmacoepidemiological studies. The objective 54 55 of this study is to explore the level of evidence used to support the withdrawal of marketing authorisation 56 57 of drug products in the time period 2002 to 2011 in the EU as a whole. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 METHODS 4 5 Drug products withdrawn from the EU market for safety or pharmacovigilance reasons, either by a 6 7 medicines agency or a marketing authorisation holder, during the period 2002-2011 were identified. This 8 was established by carrying out detailed searches of the websites of the World Health Organisation 9 10 (WHO), the EMA and those of national medicines agencies throughout the EU plus Norway, Iceland and 11 12 Liechtenstein. The WHO Pharmaceutical Newsletter and the EMA press releases provided the most 13 useable information. In addition, pharmaceutical companies were contacted to obtain more information 14 15 concerning datesFor of withdrawal peer in specific countries.review only 16 17 18 Prescription and over-the-counter products were included if their marketing authorisation was withdrawn 19 20 or suspended due to any safety reason such as an adverse drug reaction, abuse or misuse. Active 21 22 ingredients contained in a combination product were also included. Veterinary products, herbal drugs, 23 vaccines, excipients, biological drugs, radiopharmaceuticals products and diagnostic agents were 24 25 excluded. Drugs were not included if only one particular dose, dosage form or indication was withdrawn. 26 27 28 The scientific evidence leading to the withdrawal decision was identified from a search within PubMed, 29 30 the EMA (“Press releases”, “Questions and Answers” and “Scientific conclusions”) and national 31 32 medicines agencies websites. A request was also submitted to the EMA for information on the number of 33 cases entered into EudraVigilance related to the reason of withdrawal for each drug. 34 http://bmjopen.bmj.com/ 35 36 37 A search strategy was developed in PubMed (Table 1) by firstly searching under the International 38 Nonproprietary Name (INN) of the drug. Secondly, if over 200 publications were found, the INN 39 40 combined with the MeSH term describing the pharmacovigilance problem was entered. Thirdly, the INN 41
42 combined with the terms "adverse drug reaction", "adverse event", "toxicity" and "poisoning" was used. on September 29, 2021 by guest. Protected copyright. 43 Furthermore, papers published up to one year after the drug withdrawals were included to account for the 44 45 potential delay between the end of a study and its publication. 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 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 Data was collated from these sources and a comprehensive literature review was conducted. However, a 6 fully systematic review was not carried out due to time and resource constraints. Any evidence cited in 7 8 these sources was categorised according to study design including animal studies, case reports (including 9 10 spontaneous case reports, published case reports and case series), case-control studies, cohort studies, 11 RCTs, meta-analyses and other studies (including clinical trials not randomized and/or not controlled and 12 13 incidence studies). Only studies where statistically significant results were found were included. 14 15 For peer review only 16 17 RESULTS 18 Nineteen drugs were withdrawn for pharmacovigilance reasons in the EU from 2002 to 2011 (Table 2). 19 20 The anatomical therapeutic chemical (ATC) classes of drugs most represented in the nineteen drugs 21 22 withdrawn are "nervous system" (5/19 analgesics, antidepressants, antipsychotics and hypnotics), 23 "musculo-skeletal system" (5/19 non-steroidal anti-inflammatory drugs and muscle relaxants) and 24 25 "alimentary tract and metabolism" (4/19 antidiabetics and antiobesity drugs). 26 27 28 The average time to withdrawal was approximately 24 years (mean = 23.8) and with 95% confidence, that 29 30 estimate lies between 15 years and 33 years (95% CI 14.6 - 33.0) (Table 2). 31 32 33 Case reports were cited in 95% of withdrawals (18/19) and case-control studies, cohorts studies, RCTs or 34 http://bmjopen.bmj.com/ 35 meta-analysis were cited in 63% of withdrawals (12/19) (Table 3). 36 37 38 Cardiovascular events or disorders were the main reason for withdrawal (9/19), followed by hepatic 39 40 disorders (4/19) and neurological or psychiatric disorders (4/19). 41
42 on September 29, 2021 by guest. Protected copyright. 43 A list of the adverse drug reactions reported in EudraVigilance, correlating to the safety issue given as the 44 45 reason for withdrawal for each drug product is shown in Table 4. No results were returned for 46 47 ximelagatran / megalatran on querying the database for hepatotoxicity reports which is in line with the 48 EMA assessment which states that an RCT (EXTEND) study [7] ) elluicidated this safety concern [8] . A 49 50 high number of reports were recorded for rosiglitazone; 10,834 cases of cardiac disorders associated with 51 52 its use. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 DISCUSSION 4 5 This study has shown that the strength of evidence used to support the withdrawal of products relied on 6 7 information recorded from spontaneous reporting together with other pharmacoepidemiological and 8 clinical research studies. All decisions to revoke marketing authorisations described in this study utilised 9 10 spontaneous case reports, except for the withdrawal of ximegalatran [8]. Spontaneous case reporting 11 12 remains central to pharmacovigilance as it allows for the detection of suspected adverse drug reactions 13 and significant safety signals. However, spontaneous reporting should be used in conjunction with other 14 15 pharmacovigilanceFor methodologies peer such as observationalreview studies andonly clinical trials to further evaluate any 16 17 such safety signals due to the potential for over and under-reporting, confounding, bias and 18 misclassification. The randomised controlled trial is commonly assumed to be the gold standard for 19 20 clinical research [9] but adverse drug reactions may not become evident until a medicine is used in the 21 22 larger population, post-authorisation, where patients are more likely to have co-morbidities and other risk 23 factors. The outcome of pharmacoepidemiology studies and the conduct of systematic reviews, even with 24 25 their limitations particularly bias, confounding and misclassification, can provide useful additional 26 27 evidence on the nature and frequency of adverse drug reactions. As a direct country comparison, Clarke 28 et. al [4] stated that six products were withdrawn in the UK in between 1999 and 2001, namely 29 30 grepafloxacin, cisapride, pumactant, cerivastatin, droperidol and levacetylmethadol. The mean length of 31 32 time these drugs remained on the market is approximately 10 years. Droperidol was available for 33 approximately 30 years as it was marketed before the licencing system began in 1971 (personal 34 http://bmjopen.bmj.com/ 35 communication from MHRA (2013). In contrast, only 3 products were withdrawn from the UK market 36 37 between 2009 and 2011; orciprenaline, sibutramine and rosiglitazone, marking a 50% reduction in the 38 number of drug products withdrawn over the same time period. The mean length of time these products 39 40 remained on the market was also 10 years. Although the time taken to withdraw a drug product appears 41
42 unchanged over the ten year period within the UK, the evidence used to aid the decision making has been on September 29, 2021 by guest. Protected copyright. 43 heightened. Spontaneous case reporting was used to support the case for withdrawal in 5 of the 6 44 45 products between 1999 and 2001, while only one randomized controlled trial was included as evidence 46 47 [4]. In this study, both spontaneous case reporting and randomized controlled trials were shown to have 48 contributed to all three withdrawn products, which suggests an increase in the level of evidence used to 49 50 support decisions drug product withdrawal from the market. 51 52 53 The findings of this study are in-line with those of the study conducted by Paludetta and colleagues [6] in 54 55 2012 who observed a shift in the nature of the safety data used as justification for the withdrawal of 56 57 products in France in the time periods 1998 to 2004 and 2005 to 2011. One explanation the authors gave 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 for this change was the implementation of the European Clinical Trial Directive in 2004 and the 4 5 subsequent analysis of clinical trial safety data by pharmacovigilance experts. The evolution of 6 pharmacovigilance legislation over the past decade is likely to have contributed to the availability of 7 8 additional data through the conduct of post-marketing safety surveillance allowing regulators and 9 10 companies to include such information in the drug products safety portfolio. It is difficult to ascertain 11 whether the introduction of risk management plans for newly licensed medicinal products has influenced 12 13 the length of time a drug with safety concerns remains on the market as only two drug products included 14 15 in this study wereFor marketed peer after 2005. Rimonabant review and sitaxentan only were licensed in 2006 and removed 16 from the market in 2008 and 2010 respectively. At the time of authorisation, information was known 17 18 about the psychiatric side effects of rimonabant [10] and liver toxicity associated with sitaxentan [11] and 19 20 the benefit-risk balance deemed to be favourable. However, in the case of rimonabant, as data became 21 available when used in the real-life clinical setting, the risk of obese and overweight patients developing 22 23 psychiatric side effects was double that of patients taking placebo [10]. The marketing authorisation 24 25 holder voluntarily removed sitaxentan from the market and cancelled on-going clinical trials as a result of 26 two cases of fatal liver injury [11]. During the time period studied, an additional twelve drugs were 27 28 withdrawn in some but not all EU countries. In some cases, it was not feasible to establish the reasons for 29 30 the withdrawal e.g. astemizole has been withdrawn from the UK, France and Spain, but information about 31 its availability in other EU countries is difficult to establish. Another example, the NSAID nimesulide, 32 33 was withdrawn in Spain, Finland, Belgium and Ireland but is still available in fifteen EU member states. 34 http://bmjopen.bmj.com/ 35 The EMA had a favourable opinion of the risk-benefit balance after a review of spontaneous reports 36 epipdemiological studies and other published studies in 2010 [12]. This was in spite of evidence of an 37 38 increased risk of hepatotoxicity compared with other NSAIDs, although restrictions are in place to limit 39 40 the number of liver related side effects. Ultimately, it would seem that there is a disparity in opinion of 41 risk-benefit within the EU between decision makers at the EMA and various regulatory authorities. 42 on September 29, 2021 by guest. Protected copyright. 43 44 45 The main limitation of this study was the impact of various language barriers across the EU. Websites of 46 competent authorities are written in their local language and resource constraints did not allow for 47 48 translations. Only information from the websites of those competent authorities where English and French 49 50 are spoken locally or an English or French version is available was utilised. This may also have had a 51 bearing on the studies included as evidence in this research as again only English and French publications 52 53 were included. After a website search, competent authorities and pharmaceutical companies were 54 55 contacted in order to obtain further information on drug product withdrawals. The response rate from 56 competent authorities was 70% and from pharmaceutical companies 55%. Again, this may have been due 57 58 to a language barrier problem, but it may also be due to a lack of available information on drugs 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 withdrawn for several years or an unwillingness to share information. Competent authorities or 4 5 companies may also not have a complete list of products removed from the market. Over time, the way in 6 which drug products have been licensed in the EU has changed considerably. It is relatively easy to 7 8 determine how products licensed centrally were withdrawn from the market. However, for drugs that 9 10 have been licensed for a considerable number of years, in countries that have recently joined the EU, and 11 for those that have multiple drug names; it can be very difficult to obtain conclusive, consistent and clear 12 13 information. The effect of different drug names may have led to potential exclusions of study evidence 14 15 from our PubMedFor search strategy peer as the INN review of each drug was used only rather than an exhaustive drug name 16 list. Due to time constraints, it was not possible to conduct a full investigation into why drug products 17 18 had been removed from the market in some but not all countries and to find detailed information about 19 20 the decision-making process. 21 22 23 It would be useful to investigate the impact of the introduction of RMPs in a future study. This could be 24 25 done as an investigation of how results from post-marketing studies, conducted as part of RMPs, aid in 26 the assessment of any safety issues. In theory, the introduction of RMPs should expedite postmarketing 27 28 drug safety decisions. 29 30 31 This study has shown that the strength of evidence has improved during the last ten years. Indeed, the 32 33 results show an increased use of case-control studies, cohort studies, RCTs and meta-analyses as 34 http://bmjopen.bmj.com/ 35 justification for the withdrawal of a marketed product. Previously only one third of decisions used 36 evidence from observational studies or clinical trials [4] but this study showed that they had contributed 37 38 to decision making in almost two-thirds of withdrawals. Spontaneous case reports remain the most 39 40 significant method of pharmacovigilance. 41
42 on September 29, 2021 by guest. Protected copyright. 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 10 of 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 1: Search strategy on PubMed 4 5 6 1 [International Nonproprietary Name (INN)] 7 8 2 [INN] AND [MeSH term describing the pharmacovigilance problem] 9 10 11 3 [INN] AND [adverse drug reaction OR adverse event OR toxicity OR 12 13 poisoning] 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 11 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 2: Alphabetical list of drugs withdrawn for safety reasons in all EU member states between 2002 and 2011. 4 Drug name Drug class or use Year first Year of Length of time on market Adverse reaction or safety 5 marketed withdrawal (years) concern 6 7 Aceprometazine + Hypnotic 1988 2011 23 Cumulative adverse effects - 8 Acepromazine + Misuse – Fatal side effect 9 Clorazepate 10 11 Benfluorex Anorectic and 1974 2009 35 Heart valve disease – 12 hypolipidemic Pulmonary hypertension 13 Bufexamac NSAID ~1970 2010 40 Contact allergic reactions 14 Buflomedil Vasodilator (α and α 1974 2011 37 Neurological and cardiac 15 For peer1 2 review only 16 receptor antagonist) disorders (sometimes fatal) 17 Carisoprodol Muscle relaxant 1959 2007 48 Intoxication - Psychomotor 18 impairment - Addiction – 19 Misuse 20 Clobutinol Cough suppressant 1961 2007 46 QT prolongation 21 (centrally acting) 22 Dextropropoxyphene Opioid painkiller ~1960 2009 49 Fatal overdose 23
24 25 Lumiracoxib NSAID (COX-2 inhibitor) 2003 2007 4 Hepatotoxicity 26 27 Nefazodone Antidepressant 1994 2003 9 Hepatotoxicity 28 29 Orciprenaline Sympathomimetic (non- 1961 2010 49 Cardiac disorders 30 specific β-agonist) 31 Rimonabant Treatment of obesity 2006 2008 2 Psychiatric disorders 32 33 (cannabinoid receptor 34 antagonist) http://bmjopen.bmj.com/ 35 Rofecoxib NSAID (COX-2 inhibitor) 1999 2004 5 Thrombotic events 36 37 Rosiglitazone Anti-diabetic treatment 2000 2010 10 Cardiovascular disorders 38 (PPAR agonist) 39 Sibutramine Treatment of obesity 1999 2010 11 Cardiovascular disorders 40 (serotonin-noradrenaline 41 re-uptake inhibitor) 42 on September 29, 2021 by guest. Protected copyright. 43 Sitaxentan Antihypertensive 2006 2010 4 Hepatotoxicity 44 (endothelin receptor 45 antagonist) 46 Thioridazine Neuroleptic (α-adrenergic 1958 2005 47 Cardiac disorders 47 and dopaminergic receptor 48 antagonist) 49 Valdecoxib NSAID (COX-2 inhibitor) 2003 2005 2 Cardiovascular and cutaneous 50 51 disorders 52 Veralipride Neuroleptic (and 1979 2007 28 Neurologic and psychiatric 53 dopaminergic receptor disorders 54 antagonist) 55 Ximelagatran / Anticoagulant (thrombin 2003 2006 3 Hepatotoxicity 56 Megalatran inhibitor) 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 3: List of evidence used to support medicinal product withdrawals in all EU member states 4 5 between 2002 and 2011 derived from EMA reports, PubMed literature search and websites of competent 6 7 authorities. 8 9 Drug name Case reports Animal Case- Cohort RCTs Meta- †Others control analysis 10 studies 11 12 13 Aceprometazine + x x 14 Acepromazine + 15 ClorazepateFor peer review only 16 17 Benfluorex x x x x 18 19 Bufexamac x x x 20 Buflomedil x x 21 22 Carisoprodol x x x x x 23 24 Clobutinol x x x 25 Dextropropoxyphene x x 26 27 Lumiracoxib x x 28 29 Nefazodone x x 30 Orciprenaline x x 31 32 Rimonabant x x x 33 34 Rofecoxib x x x x x http://bmjopen.bmj.com/ 35 Rosiglitazone x x x x x 36 37 Sibutramine x x x 38 Sitaxentan x x 39 40 Thioridazine x x x x x 41
42 Valdecoxib x x x on September 29, 2021 by guest. Protected copyright. 43 Veralipride x 44 45 Ximelagatran / Megalatran 46 x
47 48 † other studies include non-randomized and/or not controlled clinical trials and incidence studies. 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 13 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 4: Count of case reports of selected withdrawn drug products submitted to Eudravigilance up to 30 4 5 June 2012. 6 Drug name MedDRA Level Term Total 7 8 Aceprometazine + HLGT Fatal Outcomes 3 9 Acepromazine + Clorazepate 10 11 Benfluorex HLGT Cardiac valve disorders 3439 12 PT Pulmonary hypertension 181 13 Bufexamac SOC Immune systems disorders 10 14 Buflomedil HLGT Cardiac arrhythmias 36 15 For peer review only HLGT Fatal outcomes^ 14 16 17 HLGT Seizures (inc. subtypes) 35 18 Carisoprodol HLGT Mental impairment disorders 18 19 HLGT Movement disorders (inc. parkinsonism) 15 20 Clobutinol HLGT Cardiac arrhythmias 14 21 22 Dextropropoxyphene HLGT Fatal outcomes^ 55 23
24 HLGT Overdoses 91 25 Lumiracoxib HLGT Hepatic and hepatobiliary disorders 92 26 27 Nefazodone HLGT Hepatic and hepatobiliary disorders 16 28 29 Orciprenaline HLGT Cardiac arrhythmias 15 30 31 Rimonabant* SOC Psychiatric disorders 918 32 33 PT Depression 545 34 Rofecoxib PT Myocardial infarction 6711 http://bmjopen.bmj.com/ 35 36 Rosiglitazone* SOC Cardiac disorders 10834 37 38 Sibutramine PT Blood pressure increased 69 39 PT Heart rate increased 23 40 41 Sitaxentan* PT Hepatitis 8
42 PT Hepatic Failure 6 on September 29, 2021 by guest. Protected copyright. 43 Thioridazine HLGT Cardiac arrhythmias 179 44 PT Sudden death 28 45 Valdecoxib* SOC Skin and subcutaneous tissue disorders 317 46 SOC Cardiac disorders 313 47 48 Veralipride HLGT Anxiety disorders and symptoms 142 49 HLT Depressive disorders 221 50 Ximelagatran / Megalatran No results were returned 51 52 Key: HLGT = Higher Level Group Term; PT = Preferred Term; SOC = System Organ Class 53 54 *- centrally authorised drug products. Information available on: www.adrreports.eu. Other information received 55 by email from EMA. 56 ^- the original condition, preceding the fatal outcome is unknown therefore there is a possibility of duplicate reporting. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 ACKNOWLEDGEMENTS 6 The Drug Safety Research Unit (DSRU) is a registered independent charity (No.327206) associated with 7 8 the University of Portsmouth. The DSRU receives unconditional donations from pharmaceutical 9 10 companies. The companies have no control on the conduct or the publication of its studies. The DSRU 11 has received such funds from the manufacturers of products included in this study. 12 13 14 15 All authors have Forcompleted thepeer Unified Competing review Interest form atwww.icmje.org/coi_disclosure.pdf only and declare: 16 no support from any organisation for the submitted work; Prof. Shakir reports personal fees from ICON (Contract 17 research organization), personal fees from Shire Pharmaceuticals, personal fees from ONO Pharmaceuticals, 18 19 personal fees from Intermune Pharma, personal fees from IPSEN, outside the submitted work; Mr Huet reports 20 personal fees from Axpharma, personal fees from Thea laboratories, outside the submitted work; Dr McNaughton 21 reports no financial relationships with any organisations that might have an interest in the submitted work in the 22 23 previous three years, no other relationships or activities that could appear to have influenced the submitted work. 24 25 26 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide 27 licence to the Publishers and its licensees in perpetuity, in all forms, formats and media (whether known now or created in the 28 future), to i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, 29 30 create adaptations, reprints, include within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) 31 create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the 32 inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) licence any third 33 party to do any or all of the above. 34 http://bmjopen.bmj.com/ 35 36 The lead author (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent 37 account of the study being reported; that no important aspects of the study have been omitted; and that any 38 39 discrepancies from the study as planned have been explained. 40 41
42 REFERENCE LIST on September 29, 2021 by guest. Protected copyright. 43 44 (1) European Medicines Agency (EMA) online. Available from URL: http://www ema europa eu 45 46 2013; [Accessed on 28 March 2013]. 47 48 (2) European Medicines Agency (EMA). Guideline on good pharmacovigilance practices (GVP) 49 50 51 Module V – Risk management systems. http://www ema europa 52 53 eu/docs/en_GB/document_library/Scientific_guideline/2012/06/WC500129134 pdf 2013; 54 55 [Accessed 16/05/2013]. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 16 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 (3) Arnaiz JA, Carné X, Riba N, Codina C, Ribas J, Trilla A. The use of evidence in 4 5 pharmacovigilance. Case reports as the reference source for drug withdrawals. Eur J Clin 6 7 Pharmacol 2001;57(1):89-91 8 9 10 (4) Clarke A, Deeks JJ, Shakir SA. An assessment of the publicly disseminated evidence of safety 11 12 used in decisions to withdraw medicinal products from the UK and US markets. Drug Saf 13 14 2006;29(2):175-81. 15 For peer review only 16 (5) Olivier P,Montastruc J L. The nature of the scientific evidence leading to drug withdrawals for 17 18 19 pharmacovigilance reasons in France. Pharmacoepidemiol Drug Saf 2006;15(11):808-12 20 21 (6) Paludetto M N, Olivier-Abbal P, Montastruc J L. Is spontaneous reporting always the most 22 23 important information supporting drug withdrawals for pharmacovigilance reasons in France? 24 25 Pharmacoepidemiol Drug Saf 2012;21(12):1289-94. 26 27 (7) Agnelli G, Eriksson B I, Cohen A T, Bergqvist D, Dahl O E, Lassen M R, Mouret P, Rosencher 28 29 30 N, Andersson M, Bylock A, Jensen E, Boberg B; EXTEND Study Group. Safety assessment of 31 32 new antithrombotic agents: lessons from the EXTEND study on ximelagatran. Thromb Res 33 34 2009;123(3):488-97. http://bmjopen.bmj.com/ 35 36 (8) European Medicines Agency (EMA). Press Release: AstraZeneca Withdraws its Application for 37 38 39 Ximelagatran 36-mg film-coated tablets. http://www ema europa 40 41 eu/docs/en_GB/document_library/Press_release/2010/02/WC500074073 pdf 2013; [Accessed
42 on September 29, 2021 by guest. Protected copyright. 43 15 Mar 2013] 44 45 (9) Carne X, Arnaiz J A. Methodological and political issues in clinical pharmacology research by 46 47 the year 2000. Eur J Clin Pharmacol 2000;55:781-5. 48 49
50 (10) European Medicines Agency (EMA). Assessment Report for Acomplia (Rimonabant). Available 51 52 from URL: http://www ema europa eu/docs/en_GB/document_library/EPAR_- 53 54 _Assessment_Report_-_Variation/human/000666/WC500021280 pdf 2009; [Accessed 28 55 56 March 2013]. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 16 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 (11) European Medicines Agency (EMA). Press release: Thelin (sitaxentan) to be withdrawn due to 4 5 cases of unpredictable serious liver injury. Available from URL: http://www ema europa 6 7 eu/ema/index jsp?curl=/pages/news_and_events/news/2010/12/news_detail_001161 8 9 10 jsp&murl=menus/news_and_events/news_and_events jsp&mid=WC0b01ac058004d5c1 11 12 2010; [Accessed 28 March 2013]. 13 14 (12) European Medicines Agency (EMA). Questions and Answers on the Review of Systemic 15 For peer review only 16 Medicines Containing Nimesulide. Available from URL: http://www ema europa 17 18 19 eu/docs/en_GB/document_library/Referrals_document/Nimesulide_31/WC500107957 pdf 20 21 2012; [Accessed on 28 March 2013]. 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
An Investigation into Drug Products Withdrawn from the EU Market between 2002 and 2011 for Safety Reasons and the Evidence Used to Support the Decision Making
For peer review only Journal: BMJ Open
Manuscript ID: bmjopen-2013-004221.R1
Article Type: Research
Date Submitted by the Author: 21-Nov-2013
Complete List of Authors: McNaughton, Rhian; Drug Safety Research Unit, ; University of Portsmouth, Huet, Gwenaël; Drug Safety Research Unit, Shakir, Saad; Drug Safety Research Unit, ; University of Portsmouth,
Primary Subject Public health Heading:
Epidemiology, Evidence based practice, Health services research, Secondary Subject Heading: Pharmacology and therapeutics
Risk management < HEALTH SERVICES ADMINISTRATION & Keywords: MANAGEMENT, Adverse events < THERAPEUTICS, Clinical trials <
THERAPEUTICS http://bmjopen.bmj.com/
on September 29, 2021 by guest. Protected copyright.
For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 An Investigation into Drug Products Withdrawn from the EU Market between 2002 4 5 and 2011 for Safety Reasons and the Evidence Used to Support the Decision Making 6 7 8 9 Rhian McNaughton, Gwenaël Huet, Saad Shakir 10 11 Drug Safety Research Unit, Southampton, SO31 1AA UK and University of Portsmouth, 12 13 Portsmouth PO1 2UP, UK Rhian McNaughton Research Fellow Drug Safety Research 14 15 Unit, Southampton,For SO31 peer 1AA UK review Gwenaël Huet Intern only Drug Safety Research Unit, 16 17 Southampton, SO31 1AA UK and University of Portsmouth, Portsmouth PO1 2UP, UK 18 19 Saad Shakir Director 20
21 22 Correspondence to: Rhian McNaughton, Drug Safety Research Unit, Southampton, SO31 1AA UK email: 23 24 [email protected], Tel: (023) 80408623, Fax: (023) 80408609 25 26 27 Keywords: Withdrawal – Drug Product – Pharmacovigilance – Evidence – Safety 28 29 30 31 Word Count 2430 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 2 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 ABSTRACT 4 5 Background The objective of this study was to determine the nature of evidence used to support the 6 7 withdrawal of marketing authorisations of drug products for safety reasons throughout the EU between 8 2002 and 2011. 9 10 Methods Products withdrawn, either by a medicines agency or a marketing authorisation holder, during 11 12 the period 2002 2011 were identified by conducting detailed searches of the World Health Organization 13 (WHO), the EMA (European Medicines Agency) and national medicines agency websites throughout the 14 15 EU plus Norway,For Iceland andpeer Liechtenstein. review The scientific evidence only used to support the decision was 16 17 identified from a search within PubMed, the EMA and national medicines agencies websites. Information 18 about spontaneous case reports entered into EudraVigilance and unavailable on the EMA website, was 19 20 received by email from the EMA. 21 22 Results 19 drugs were withdrawn from the market, throughout the EU, for safety reasons from 2002 to 23 2011. Case reports were cited in 95% of withdrawals (18/19) and case control studies (4/19), cohort 24 25 studies (4/19), randomized controlled trials (RCTs) (12/19) or meta analysis (5/19) were cited in 63% of 26 27 withdrawals (12/19). Cardiovascular events or disorders were the main reason for withdrawal (9/19), 28 followed by hepatic disorders (4/19) and neurological or psychiatric disorders (4/19). 29 30 Conclusions This study has shown that the level of evidence used to support drug withdrawal has 31 32 improved during the last ten years, with an increased use of case control studies, cohort studies, RCTs 33 and meta analyses. This research has demonstrated that such studies have contributed to decision making 34 http://bmjopen.bmj.com/ 35 in almost two thirds of cases. 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 3 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Strengths and limitations of this study 4 5 • This study adds to the knowledge base regarding the level of evidence used to support 6 7 decisions to remove drug products from the market. 8 • Results of this research are in line with the findings of other study groups who concluded that 9 10 there has been a shift in the level of evidence used e.g. results of randomized controlled trials 11 12 and cohort studies are used to justify a product withdrawal. 13 • English and French websites and publications were only searched. 14 15 • The INNFor of each drugpeer was included review in the PubMed search only strategy but as various drug names 16 17 exist for the same compound, a potential exists for the exclusion of some publications from 18 19 contributing evidence. 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 INTRODUCTION 4 5 The decision to remove a drug product from the market is not taken lightly either by the regulatory 6 7 authority or the marketing authorisation holder. A number of reasons contribute to the withdrawal such as 8 the emergence of new safety evidence and significant drug interactions. Ultimately, the risk benefit 9 10 balance is paramount when making such decisions. The regulatory authorities review the evidence, such 11 12 as spontaneous case reports or results from a randomized controlled trial (RCT) and prepare an opinion 13 on whether the marketing authorisations of medicinal products containing that particular drug should be 14 15 maintained, changed,For suspended peer or withdrawn. review The interpretation only of the available safety data by the 16 17 numerous competent authorities across the European Union (EU) is wide ranging as is the accessibility of 18 information regarding how they reach their conclusion. 19 20 21 22 In 2005, legislation was introduced to ensure marketing authorisation holders complete a risk 23 management plan (RMP) for their newly licensed product incorporating, where appropriate, post 24 25 marketing safety surveillance studies in the general population with regular assessment and appropriate 26 27 reporting to the CHMP [1]. The aim of the RMP is to ensure a more proactive approach to 28 pharmacovigilance by putting in place measures that allow for the early detection and minimization of 29 30 risks throughout a medicines lifecycle. This should, in theory, result in a reduction in the length of time 31 32 for drug withdrawal should a safety issue arise. Risk management planning in the EU has been further 33 enhanced by the introduction of new EU legislation for good pharmacovigilance practices (GVP) which 34 http://bmjopen.bmj.com/ 35 became applicable in July 2012 [2]. 36 37 38 Previous studies looked at the type of evidence used to support the withdrawal of products, for safety 39 40 reasons, in different individual countries and over varying time periods. Arnaiz et al. [3] reported on 41 et al 42 products withdrawn in Spain between 1990 to 1999, Clarke . [4] for products in USA and UK from on September 29, 2021 by guest. Protected copyright. 43 1999 to 2001 and Olivier et al. [5] for medicines in France from 1998 to 2004. They all concluded that 44 45 case reports were the mainstay of evidence used to support withdrawal decisions for pharmacovigilance 46 47 reasons; case control studies, cohort studies and randomised controlled trials were used as evidence in 48 less than one third of decisions. Olivier and colleagues revisited the issue six years later, between 2005 49 50 and 2011, and reported a shift in the level of evidence used [6]. They determined that of the 22 active 51 52 substances withdrawn from the French market within this 6 year time frame, 68% of the decisions used 53 multiple sources of evidence including clinical trials and pharmacoepidemiological studies. The objective 54 55 of this study is to explore the level of evidence used to support the withdrawal of marketing authorisation 56 57 of drug products in the time period 2002 to 2011 in the EU as a whole. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 METHODS 4 5 Drug products withdrawn from the EU market for safety or pharmacovigilance reasons, either by a 6 7 medicines agency or a marketing authorisation holder, during the period 2002 2011 were identified. This 8 was established by carrying out detailed searches of the websites of the World Health Organisation 9 10 (WHO), the EMA and those of national medicines agencies throughout the EU plus Norway, Iceland and 11 12 Liechtenstein. The WHO Pharmaceutical Newsletter and the EMA press releases provided the most 13 useable information. In addition, pharmaceutical companies were contacted to obtain more information 14 15 concerning datesFor of withdrawal peer in specific countries.review only 16 17 18 Prescription and over the counter products were included if their marketing authorisation was withdrawn 19 20 or suspended due to any safety reason such as an adverse drug reaction, abuse or misuse. Active 21 22 ingredients contained in a combination product were also included. Veterinary products, herbal drugs, 23 vaccines, excipients, biological drugs, radiopharmaceuticals products and diagnostic agents were 24 25 excluded. Drugs were not included if only one particular dose, dosage form or indication was withdrawn. 26 27 28 The scientific evidence leading to the withdrawal decision was identified from a search within PubMed, 29 30 the EMA website (published documents searched included: “Press releases”, “Questions and Answers” 31 32 and “Scientific conclusions”) and national medicines agencies websites. Within the “Questions and 33 Answers” reports and “Press releases”, prepared by the EMA, a summary of the reviewed evidence with 34 http://bmjopen.bmj.com/ 35 reference to study type (e.g. RCT, animal study) is often included. A request was also submitted to the 36 37 EMA for information on the number of cases entered into EudraVigilance related to the reason of 38 withdrawal for each drug. 39 40 41
42 A search strategy was developed in PubMed by firstly searching under the International Nonproprietary on September 29, 2021 by guest. Protected copyright. 43 Name (INN) of the drug. Secondly, if over 200 publications were found, the INN combined with the 44 45 MeSH term describing the pharmacovigilance problem was entered. Thirdly, the INN combined with the 46 47 terms "adverse drug reaction", "adverse event", "toxicity" and "poisoning" was used. Furthermore, papers 48 published up to one year after the drug withdrawals were included to account for the potential delay 49 50 between the end of a study and its publication. 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 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 Data was collated from these sources and a comprehensive literature review was conducted. However, a 6 fully systematic review was not carried out due to time and resource constraints. Any evidence cited in 7 8 these sources was categorised according to study design including animal studies, case reports (including 9 10 spontaneous case reports, published case reports and case series), case control studies, cohort studies, 11 RCTs, meta analyses and other studies (including clinical trials not randomized and/or not controlled and 12 13 incidence studies). Only studies where statistically significant results were found were included. 14 15 For peer review only 16 17 RESULTS 18 Nineteen drugs were withdrawn for pharmacovigilance reasons in the EU from 2002 to 2011 (Table 1). 19 20 The anatomical therapeutic chemical (ATC) classes of drugs most represented in the nineteen drugs 21 22 withdrawn are "nervous system" (5/19 analgesics, antidepressants, antipsychotics and hypnotics), 23 "musculo skeletal system" (5/19 non steroidal anti inflammatory drugs and muscle relaxants) and 24 25 "alimentary tract and metabolism" (4/19 antidiabetics and antiobesity drugs). 26 27 28 The median time to withdrawal was 23 years with an interquartile range (IQR) of 4 and 46 (Table 1). 29 30 31 32 Case reports were cited in 95% of withdrawals (18/19) and case control studies (4/19), cohorts studies 33 (4/19), RCTs (12/19) and meta analysis (5/19) were cited in 63% of withdrawals (12/19) (Table 2). 34 http://bmjopen.bmj.com/ 35 36 37 Cardiovascular events or disorders were the main reason for withdrawal (9/19), followed by hepatic 38 disorders (4/19) and neurological or psychiatric disorders (4/19). 39 40 41
42 A list of the adverse drug reactions reported in EudraVigilance, correlating to the safety issue given as the on September 29, 2021 by guest. Protected copyright. 43 reason for withdrawal for each drug product is shown in Table 3. No results were returned for 44 45 ximelagatran / melagatran on querying the database for hepatotoxicity reports which is in line with the 46 47 EMA assessment which states that an RCT (EXTEND) study [7] ) elucidated this safety concern [8] . A 48 high number of reports were recorded for rosiglitazone; 10,834 cases of cardiac disorders associated with 49 50 its use. 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 DISCUSSION 4 5 This study has shown that the strength of evidence used to support the withdrawal of products relied on 6 7 information recorded from spontaneous reporting together with other pharmacoepidemiological and 8 clinical research studies. All decisions to revoke marketing authorisations described in this study utilised 9 10 spontaneous case reports, except for the withdrawal of ximelagatran [8]. Spontaneous case reporting 11 12 remains central to pharmacovigilance as it allows for the detection of suspected adverse drug reactions 13 and significant safety signals. However, spontaneous reporting should be used in conjunction with other 14 15 pharmacovigilanceFor methodologies peer such as observationalreview studies andonly clinical trials to further evaluate any 16 17 such safety signals due to the potential for over and under reporting, confounding, bias and 18 misclassification. The randomised controlled trial is commonly assumed to be the gold standard for 19 20 clinical research [9] but adverse drug reactions may not become evident until a medicine is used in the 21 22 larger population, post authorisation, where patients are more likely to have co morbidities and other risk 23 factors. The outcome of pharmacoepidemiology studies and the conduct of systematic reviews, even with 24 25 their limitations particularly bias, confounding and misclassification, can provide useful additional 26 27 evidence on the nature and frequency of adverse drug reactions. As a direct country comparison, Clarke 28 et. al [4] stated that six products were withdrawn in the UK in between 1999 and 2001, namely 29 30 grepafloxacin (marketed for 2 years), cisapride (12 years), pumactant (6 years), cerivastatin (4 years), 31 32 droperidol (30 years) and levacetylmethadol (4 years). The median length of time these drugs remained 33 on the UK market was 5 years (IQR 4 10.5). Droperidol was available for approximately 30 years as it 34 http://bmjopen.bmj.com/ 35 was marketed before the licencing system began in 1971 (personal communication from MHRA (2013)). 36 37 In contrast, only 3 products were withdrawn from the UK market between 2009 and 2011; orciprenaline 38 sibutramine, and rosiglitazone marking a 50% reduction in the number of drug products withdrawn over 39 40 the same time period. The median length of time these products remained on the market was 11 years 41
42 (IQR 10.5 30). Although the length of time taken to withdraw a drug product appears to have doubled on September 29, 2021 by guest. Protected copyright. 43 (mainly due to the length of time orciprenaline was available, 49 years) over the ten year period, the 44 45 evidence used to aid the decision making has been heightened. Spontaneous case reporting was used to 46 47 support the case for withdrawal in 5 of the 6 products between 1999 and 2001, while only one 48 randomized controlled trial was included as evidence [4]. In this study, both spontaneous case reporting 49 50 and randomized controlled trials were shown to have contributed to all three withdrawn products, which 51 52 suggests an increase in the level of evidence used to support decisions drug product withdrawal from the 53 market. 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 The findings of this study are in line with those of the study conducted by Paludetta and colleagues [6] in 4 5 2012 who observed a shift in the nature of the safety data used as justification for the withdrawal of 6 products in France in the time periods 1998 to 2004 and 2005 to 2011. One explanation the authors gave 7 8 for this change was the implementation of the European Clinical Trial Directive in 2004 and the 9 10 subsequent analysis of clinical trial safety data by pharmacovigilance experts. The evolution of 11 pharmacovigilance legislation over the past decade is likely to have contributed to the availability of 12 13 additional data through the conduct of post marketing safety surveillance allowing regulators and 14 15 companies to includeFor such peer information in review the drug products safety only portfolio. It is difficult to ascertain 16 whether the introduction of risk management plans for newly licensed medicinal products has influenced 17 18 the length of time a drug with safety concerns remains on the market as only two drug products included 19 20 in this study were marketed after 2005. Rimonabant and sitaxentan were licensed in 2006 and removed 21 from the market in 2008 and 2010 respectively. At the time of authorisation, information was known 22 23 about the psychiatric side effects of rimonabant [10] and liver toxicity associated with sitaxentan [11] and 24 25 the benefit risk balance deemed to be favourable. However, in the case of rimonabant, as data became 26 available when used in the real life clinical setting, the risk of obese and overweight patients developing 27 28 psychiatric side effects was double that of patients taking placebo [10]. The marketing authorisation 29 30 holder voluntarily removed sitaxentan from the market and cancelled on going clinical trials as a result of 31 two cases of fatal liver injury [11]. During the time period studied, an additional twelve drugs were 32 33 withdrawn in some but not all EU countries. In some cases, it was not feasible to establish the reasons for 34 http://bmjopen.bmj.com/ 35 the withdrawal e.g. astemizole has been withdrawn from the UK, France and Spain, but information about 36 its availability in other EU countries is difficult to establish. Another example, the NSAID nimesulide, 37 38 was withdrawn in Spain, Finland, Belgium and Ireland but is still available in fifteen EU member states. 39 40 The EMA had a favourable opinion of the risk benefit balance after a review of spontaneous reports 41 epipdemiological studies and other published studies in 2010 [12]. This was in spite of evidence of an 42 on September 29, 2021 by guest. Protected copyright. 43 increased risk of hepatotoxicity compared with other NSAIDs, although restrictions are in place to limit 44 45 the number of liver related side effects. Ultimately, it would seem that there is a disparity in opinion of 46 risk benefit within the EU between decision makers at the EMA and various regulatory authorities. 47 48 49 50 The main limitation of this study was the impact of various language barriers across the EU. Websites of 51 competent authorities are written in their local language and resource constraints did not allow for 52 53 translations. Only information from the websites of those competent authorities where English and French 54 55 are spoken locally or an English or French version is available was utilised. This may also have had a 56 bearing on the studies included as evidence in this research as again only English and French publications 57 58 were included. After a website search, competent authorities and pharmaceutical companies were 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 contacted in order to obtain further information on drug product withdrawals. The response rate from 4 5 competent authorities was 70% and from pharmaceutical companies 55%. Again, this may have been due 6 to a language barrier problem, but it may also be due to a lack of available information on drugs 7 8 withdrawn for several years or an unwillingness to share information. Competent authorities or 9 10 companies may also not have a complete list of products removed from the market. Over time, the way in 11 which drug products have been licensed in the EU has changed considerably. It is relatively easy to 12 13 determine how products licensed centrally were withdrawn from the market. However, for drugs that 14 15 have been licensedFor for a considerable peer number review of years, in countries only that have recently joined the EU, and 16 for those that have multiple drug names; it can be very difficult to obtain conclusive, consistent and clear 17 18 information. The effect of different drug names may have led to potential exclusions of study evidence 19 20 from our PubMed search strategy as the INN of each drug was used rather than an exhaustive drug name 21 list. Due to time constraints, it was not possible to conduct a full investigation into why drug products 22 23 had been removed from the market in some but not all countries and to find detailed information about 24 25 the decision making process. 26 27 28 It would be useful to investigate the impact of the introduction of RMPs in a future study. This could be 29 30 done as an investigation of how results from post marketing studies, conducted as part of RMPs, aid in 31 the assessment of any safety issues. In theory, the introduction of RMPs should expedite postmarketing 32 33 drug safety decisions. 34 http://bmjopen.bmj.com/ 35 36 This study has shown that the strength of evidence has improved during the last ten years. Indeed, the 37 38 results show an increased use of case control studies, cohort studies, RCTs and meta analyses as 39 40 justification for the withdrawal of a marketed product. Previously only one third of decisions used 41 evidence from observational studies or clinical trials [4] but this study showed that they had contributed 42 on September 29, 2021 by guest. Protected copyright. 43 to decision making in almost two thirds of withdrawals. Spontaneous case reports remain the most 44 45 significant method of pharmacovigilance. 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 10 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 1: List of drugs withdrawn for safety reasons in all EU member states between 2002 and 2011 grouped by 4 adverse drug reaction or safety concern.. 5 6 Drug name Drug class or use Year first Year of Length of time on market Adverse reaction or safety 7 marketed withdrawal (years) concern 8 Rofecoxib NSAID (COX 2 inhibitor) 1999 2004 5 Thrombotic events 9 10 Thioridazine Neuroleptic (α adrenergic 1958 2005 47 Cardiac disorders 11 and dopaminergic receptor 12 13 antagonist) 14 Valdecoxib NSAID (COX 2 inhibitor) 2003 2005 2 Cardiovascular and cutaneous 15 For peer review only disorders 16 Rosiglitazone Anti diabetic treatment 2000 2010 10 Cardiovascular disorders 17 (PPAR agonist) 18 Sibutramine Treatment of obesity 1999 2010 11 Cardiovascular disorders 19 (serotonin noradrenaline 20 re uptake inhibitor) 21 22 Orciprenaline Sympathomimetic (non 1961 2010 49 Cardiac disorders 23 specific β agonist) 24 Benfluorex Anorectic and 1974 2009 35 Heart valve disease – 25 hypolipidemic Pulmonary hypertension 26 Clobutinol Cough suppressant 1961 2007 46 QT prolongation 27 (centrally acting) 28 Buflomedil Vasodilator (α1 and α2 1974 2011 37 Neurological and cardiac 29 30 receptor antagonist) disorders (sometimes fatal) 31 Veralipride Neuroleptic (and 1979 2007 28 Neurologic and psychiatric 32 dopaminergic receptor disorders 33 antagonist) 34 Rimonabant Treatment of obesity 2006 2008 2 Psychiatric disorders http://bmjopen.bmj.com/ 35 (cannabinoid receptor 36 antagonist) 37 Carisoprodol Muscle relaxant 1959 2007 48 Intoxication Psychomotor 38 39 impairment Addiction – 40 Misuse 41 Aceprometazine + Hypnotic 1988 2011 23 Cumulative adverse effects
42 Acepromazine + Misuse – Fatal side effect on September 29, 2021 by guest. Protected copyright. 43 Clorazepate 44 45 Dextropropoxyphene Opioid painkiller ~1960 2009 49 Fatal overdose 46
47 48 Nefazodone Antidepressant 1994 2003 9 Hepatotoxicity 49 50 Ximelagatran / Anticoagulant (thrombin 2003 2006 3 Hepatotoxicity 51 Melagatran inhibitor) 52 Lumiracoxib NSAID (COX 2 inhibitor) 2003 2007 4 Hepatotoxicity 53 54 Sitaxentan Antihypertensive 2006 2010 4 Hepatotoxicity 55 (endothelin receptor 56 57 antagonist) 58 Bufexamac NSAID ~1970 2010 40 Contact allergic reactions 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 2: List of evidence used to support medicinal product withdrawals in all EU member states 4 5 between 2002 and 2011 derived from EMA reports, PubMed literature search and websites of competent 6 7 authorities. 8 Drug name Case reports Animal Case- Cohort RCTs Meta- †Others 9 studies control analysis 10 Rofecoxib x x x x x 11 12 Thioridazine x x x x x 13 Valdecoxib x x x 14 Rosiglitazone x x x x x 15 SibutramineFor peerx review onlyx x 16 Orciprenaline x x 17 18 Benfluorex x x x x 19 Clobutinol x x x 20 Buflomedil x x 21 Veralipride x 22 Rimonabant x x x 23 Carisoprodol x x x x x 24 25 Aceprometazine + x x 26 Acepromazine + Clorazepate 27 Dextropropoxyphene x x 28 Nefazodone x x 29 Ximelagatran / Melagatran x 30 31 Lumiracoxib x x 32 Sitaxentan x x 33 Bufexamac x x x http://bmjopen.bmj.com/ 34 35 † other studies include non randomized and/or not controlled clinical trials and incidence studies. 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 12 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 Table 3: Count of case reports of selected withdrawn drug products submitted to Eudravigilance up to 30 4 5 June 2012. 6 Drug name MedDRA Level Term Total 7 8 Rofecoxib PT Myocardial infarction 6711 9 10 Thioridazine HLGT Cardiac arrhythmias 179 11 PT Sudden death 28 12 Valdecoxib* SOC Skin and subcutaneous tissue disorders 317 13 SOC Cardiac disorders 313 14 Rosiglitazone* SOC Cardiac disorders 10834 15 For peer review only
16 17 Sibutramine PT Blood pressure increased 69 18 PT Heart rate increased 23 19 Orciprenaline HLGT Cardiac arrhythmias 15 20 21 Benfluorex HLGT Cardiac valve disorders 3439 22 PT Pulmonary hypertension 181 23 24 Clobutinol HLGT Cardiac arrhythmias 14 25 26 Buflomedil HLGT Cardiac arrhythmias 36 27 HLGT Fatal outcomes^ 14 28 HLGT Seizures (inc. subtypes) 35 29 Veralipride HLGT Anxiety disorders and symptoms 142 30 HLT Depressive disorders 221 31 32 Rimonabant* SOC Psychiatric disorders 918 33 PT Depression 545 34 Carisoprodol HLGT Mental impairment disorders 18 http://bmjopen.bmj.com/ 35 HLGT Movement disorders (inc. parkinsonism) 15 36 Aceprometazine + HLGT Fatal Outcomes 3 37 Acepromazine + Clorazepate 38
39 40 Dextropropoxyphene HLGT Fatal outcomes^ 55 41 HLGT Overdoses 91
42 Nefazodone HLGT Hepatic and hepatobiliary disorders 16 on September 29, 2021 by guest. Protected copyright. 43 44 Ximelagatran / Melagatran No results were returned 45 46 Lumiracoxib HLGT Hepatic and hepatobiliary disorders 92 47 48 49 Sitaxentan* PT Hepatitis 8 50 PT Hepatic Failure 6 51 Bufexamac SOC Immune systems disorders 10 52
53 54 Key: HLGT = Higher Level Group Term; PT = Preferred Term; SOC = System Organ Class 55 *- centrally authorised drug products. Information available on: www.adrreports.eu. Other information received 56 57 by email from EMA. 58 ^- the original condition, preceding the fatal outcome is unknown therefore there is a possibility of duplicate reporting. 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 ACKNOWLEDGEMENTS 6 The Drug Safety Research Unit (DSRU) is a registered independent charity (No.327206) associated with 7 8 the University of Portsmouth. The DSRU receives unconditional donations from pharmaceutical 9 10 companies. The companies have no control on the conduct or the publication of its studies. The DSRU 11 has received such funds from the manufacturers of products included in this study. 12 13 14 15 All authors have Forcompleted thepeer Unified Competing review Interest form atwww.icmje.org/coi_disclosure.pdf only and declare: 16 no support from any organisation for the submitted work; Prof. Shakir reports personal fees from ICON (Contract 17 research organization), personal fees from Shire Pharmaceuticals, personal fees from ONO Pharmaceuticals, 18 19 personal fees from Intermune Pharma, personal fees from IPSEN, outside the submitted work; Mr Huet reports 20 personal fees from Axpharma, personal fees from Thea laboratories, outside the submitted work; Dr McNaughton 21 reports no financial relationships with any organisations that might have an interest in the submitted work in the 22 23 previous three years, no other relationships or activities that could appear to have influenced the submitted work. 24 25 26 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide 27 licence to the Publishers and its licensees in perpetuity, in all forms, formats and media (whether known now or created in the 28 future), to i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, 29 30 create adaptations, reprints, include within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) 31 create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the 32 inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) licence any third 33 party to do any or all of the above. 34 http://bmjopen.bmj.com/ 35 36 The lead author (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent 37 account of the study being reported; that no important aspects of the study have been omitted; and that any 38 39 discrepancies from the study as planned have been explained. 40 41
42 CONTRIBUTORSHIP on September 29, 2021 by guest. Protected copyright. 43 Rhian McNaughton, Saad Shakir and Gwenael Huet were all involved in the conception and design of the 44 45 study, interpretation of data, drafting and revision of the article and final approval of the version for 46 47 publication. Gwenael Huet and Rhian McNaughton were responsible for data acquisition. 48 DATA SHARING 49 50 No additional data available. 51 52 COMPETING INTERESTS 53 All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf and 54 55 declare: no support from any organisation for the submitted work; Prof. Shakir reports personal fees from 56 57 ICON (Contract research organization), personal fees from Shire Pharmaceuticals, personal fees from 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 ONO Pharmaceuticals, personal fees from Intermune Pharma, personal fees from IPSEN, outside the 4 5 submitted work; Mr Huet reports personal fees from Axpharma, personal fees from Thea laboratories, 6 outside the submitted work; Dr McNaughton reports no financial relationships with any organisations that 7 8 might have an interest in the submitted work in the previous three years, no other relationships or 9 10 activities that could appear to have influenced the submitted work. 11 12 13 14 REFERENCE LIST 15 For peer review only 16 (1) European Medicines Agency (EMA) online. Available from URL: http://www ema europa eu 17 18 19 2013; [Accessed on 28 March 2013]. 20 21 (2) European Medicines Agency (EMA). Guideline on good pharmacovigilance practices (GVP) 22 23 Module V – Risk management systems. http://www ema europa 24 25 eu/docs/en_GB/document_library/Scientific_guideline/2012/06/WC500129134 pdf 2013; 26 27 [Accessed 16/05/2013]. 28 29 30 (3) Arnaiz JA, Carné X, Riba N, et al. The use of evidence in pharmacovigilance. Case reports as the 31 32 reference source for drug withdrawals. Eur J Clin Pharmacol 2001;57(1):89 91 33 34 (4) Clarke A, Deeks JJ, Shakir SA. An assessment of the publicly disseminated evidence of safety http://bmjopen.bmj.com/ 35 36 used in decisions to withdraw medicinal products from the UK and US markets. Drug Saf 37 38 39 2006;29(2):175 81. 40 41 (5) Olivier P,Montastruc J L. The nature of the scientific evidence leading to drug withdrawals for
42 on September 29, 2021 by guest. Protected copyright. 43 pharmacovigilance reasons in France. Pharmacoepidemiol Drug Saf 2006;15(11):808 12 44 45 (6) Paludetto M N, Olivier Abbal P, Montastruc J L. Is spontaneous reporting always the most 46 47 important information supporting drug withdrawals for pharmacovigilance reasons in France? 48 49 50 Pharmacoepidemiol Drug Saf 2012;21(12):1289 94. 51 52 (7) Agnelli G, Eriksson B I, Cohen A T, et al. EXTEND Study Group. Safety assessment of new 53 54 antithrombotic agents: lessons from the EXTEND study on ximelagatran. Thromb Res 55 56 2009;123(3):488 97. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 (8) European Medicines Agency (EMA). Press Release: AstraZeneca Withdraws its Application for 4 5 Ximelagatran 36 mg film coated tablets. http://www ema europa 6 7 eu/docs/en_GB/document_library/Press_release/2010/02/WC500074073 pdf 2013; [Accessed 8 9 10 15 Mar 2013] 11 12 (9) Carne X, Arnaiz J A. Methodological and political issues in clinical pharmacology research by 13 14 the year 2000. Eur J Clin Pharmacol 2000;55:781 5. 15 For peer review only 16 (10) European Medicines Agency (EMA). Assessment Report for Acomplia (Rimonabant). Available 17 18 19 from URL: http://www ema europa eu/docs/en_GB/document_library/EPAR_- 20 21 _Assessment_Report_-_Variation/human/000666/WC500021280 pdf 2009; [Accessed 28 22 23 March 2013]. 24 25 (11) European Medicines Agency (EMA). Press release: Thelin (sitaxentan) to be withdrawn due to 26 27 cases of unpredictable serious liver injury. Available from URL: http://www ema europa 28 29 30 eu/ema/index jsp?curl=/pages/news_and_events/news/2010/12/news_detail_001161 31 32 jsp&murl=menus/news_and_events/news_and_events jsp&mid=WC0b01ac058004d5c1 33 34 2010; [Accessed 28 March 2013]. http://bmjopen.bmj.com/ 35 36 (12) European Medicines Agency (EMA). Questions and Answers on the Review of Systemic 37 38 39 Medicines Containing Nimesulide. Available from URL: http://www ema europa 40 41 eu/docs/en_GB/document_library/Referrals_document/Nimesulide_31/WC500107957 pdf
42 on September 29, 2021 by guest. Protected copyright. 43 2012; [Accessed on 28 March 2013]. 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 16 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 An Investigation into Drug Products Withdrawn from the EU Market between 2002 9 10 and 2011 for Safety Reasons and the Evidence Used to Support the Decision Making 11 12 13 Rhian McNaughton, Gwenaël Huet, Saad Shakir 14 Drug Safety Research Unit, Southampton, SO31 1AA UK and University of Portsmouth, 15 For peer review only 16 Portsmouth PO1 2UP, UK Rhian McNaughton Research Fellow Drug Safety Research 17 18 Unit, Southampton, SO31 1AA UK Gwenaël Huet Intern Drug Safety Research Unit, 19 Southampton, SO31 1AA UK and University of Portsmouth, Portsmouth PO1 2UP, UK 20 21 Saad Shakir Director 22 23 24 Correspondence to: Rhian McNaughton, Drug Safety Research Unit, Southampton, SO31 1AA UK email: 25 [email protected], Tel: (023) 80408623, Fax: (023) 80408609 26
27 28 Keywords: Withdrawal – Drug Product – Pharmacovigilance – Evidence – Safety 29 30 Word Count 24302710 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 17 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 ABSTRACT 9 10 Background The objective of this study was to determine the nature of evidence used to support the 11 withdrawal of marketing authorisations of drug products for safety reasons throughout the EU between 12 2002 and 2011. 13 14 Methods Products withdrawn, either by a medicines agency or a marketing authorisation holder, during 15 the period 2002 2011 were identifiedFor by conductingpeer detailed searchesreview of the World Health only Organization 16 (WHO), the EMA (European Medicines Agency) and national medicines agency websites throughout the 17 18 EU plus Norway, Iceland and Liechtenstein. The scientific evidence used to support the decision was 19 identified from a search within PubMed, the EMA and national medicines agencies websites. Information 20 about spontaneous case reports entered into EudraVigilance and unavailable on the EMA website, was 21 22 received by email from the EMA. 23 Results 19 drugs were withdrawn from the market, throughout the EU, for safety reasons from 2002 to 24 25 2011. Case reports were cited in 95% of withdrawals (18/19) and case control studies (4/19), cohort 26 studies (4/19), randomized controlled trials (RCTs) (12/19) or meta analysis (5/19) were cited in 63% of 27 withdrawals (12/19). Cardiovascular events or disorders were the main reason for withdrawal (9/19), 28 29 followed by hepatic disorders (4/19) and neurological or psychiatric disorders (4/19). 30 Conclusions This study has shown that the level of evidence used to support drug withdrawal has 31 improved during the last ten years, with an increased use of case control studies, cohort studies, RCTs 32 33 and meta analyses. This research has demonstrated that such studies have contributed to decision making 34 in almost two thirds of cases. http://bmjopen.bmj.com/ 35
36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 18 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Strengths and limitations of this study 9 10 • This study adds to the knowledge base regarding the level of evidence used to support 11 decisions to remove drug products from the market. 12 • Results of this research are in line with the findings of other study groups who concluded that 13 14 there has been a shift in the level of evidence used e.g. results of randomized controlled trials 15 and cohort studies areFor used to justify peer a product withdrawal review. only 16 17 • English and French websites and publications were only searched. 18 • The INN of each drug was included in the PubMed search strategy but as various drug names 19 exist for the same compound, a potential exists for the exclusion of some publications from 20 21 contributing evidence. 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 19 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 INTRODUCTION 9 10 The decision to remove a drug product from the market is not taken lightly either by the regulatory 11 authority or the marketing authorisation holder. A number of reasons contribute to the withdrawal such as 12 the emergence of new safety evidence and significant drug interactions. Ultimately, the risk benefit 13 14 balance is paramount when making such decisions. The regulatory authorities review the evidence, such 15 as spontaneous case reports Foror results from peer a randomized controlled review trial (RCT) and prepare only an opinion 16 on whether the marketing authorisations of medicinal products containing that particular drug should be 17 18 maintained, changed, suspended or withdrawn. The interpretation of the available safety data by the 19 numerous competent authorities across the European Union (EU) is wide ranging as is the accessibility of 20 information regarding how they reach their conclusion. 21 22 23 In 2005, legislation was introduced to ensure marketing authorisation holders complete a risk 24 25 management plan (RMP) for their newly licensed product incorporating, where appropriate, post 26 marketing safety surveillance studies in the general population with regular assessment and appropriate 27 reporting to the CHMP [1] . The aim of the RMP is to ensure a more proactive approach to 28 29 pharmacovigilance by putting in place measures that allow for the early detection and minimization of 30 risks throughout a medicines lifecycle. This should, in theory, result in a reduction in the length of time 31 for drug withdrawal should a safety issue arise. Risk management planning in the EU has been further 32 33 enhanced by the introduction of new EU legislation for good pharmacovigilance practices (GVP) which 34 became applicable in July 2012 [2]. http://bmjopen.bmj.com/ 35
36 37 Previous studies looked at the type of evidence used to support the withdrawal of products, for safety 38 reasons, in different individual countries and over varying time periods. Arnaiz et al. [3] reported on 39 et al 40 products withdrawn in Spain between 1990 to 1999, Clarke . [4] for products in USA and UK from 41 1999 to 2001 and Olivier et al. [5] for medicines in France from 1998 to 2004. They all concluded that 42 case reports were the mainstay of evidence used to support withdrawal decisions for pharmacovigilance on September 29, 2021 by guest. Protected copyright. 43 44 reasons; case control studies, cohort studies and randomised controlled trials were used as evidence in 45 less than one third of decisions. Olivier and colleagues revisited the issue six years later, between 2005 46 and 2011, and reported a shift in the level of evidence used [6]. They determined that of the 22 active 47 48 substances withdrawn from the French market within this 6 year time frame, 68% of the decisions used 49 multiple sources of evidence including clinical trials and pharmacoepidemiological studies. The objective 50 51 of this study is to explore the level of evidence used to support the withdrawal of marketing authorisation 52 of drug products in the time period 2002 to 2011 in the EU as a whole. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 METHODS 9 10 Drug products withdrawn from the EU market for safety or pharmacovigilance reasons, either by a 11 medicines agency or a marketing authorisation holder, during the period 2002 2011 were identified. This 12 was established by carrying out detailed searches of the websites of the World Health Organisation 13 14 (WHO), the EMA and those of national medicines agencies throughout the EU plus Norway, Iceland and 15 Liechtenstein. The WHO PharmaceuticalFor peer Newsletter and thereview EMA press releases provided only the most 16 useable information. In addition, pharmaceutical companies were contacted to obtain more information 17 18 concerning dates of withdrawal in specific countries. 19 20 Prescription and over the counter products were included if their marketing authorisation was withdrawn 21 22 or suspended due to any safety reason such as an adverse drug reaction, abuse or misuse. Active 23 ingredients contained in a combination product were also included. Veterinary products, herbal drugs, 24 25 vaccines, excipients, biological drugs, radiopharmaceuticals products and diagnostic agents were 26 excluded. Drugs were not included if only one particular dose, dosage form or indication was withdrawn. 27 28 29 The scientific evidence leading to the withdrawal decision was identified from a search within PubMed, 30 the EMA website (published documents searched included: “Press releases”, “Questions and Answers” 31 and “Scientific conclusions”) and national medicines agencies websites. Within the “Questions and 32 33 Answers” reports and “Press releases”, prepared by the EMA, a summary of the reviewed evidence with 34 reference to study type (e.g. RCT, animal study) is often included. A request was also submitted to the http://bmjopen.bmj.com/ 35 EMA for information on the number of cases entered into EudraVigilance related to the reason of 36 37 withdrawal for each drug. 38 39 40 A search strategy was developed in PubMed (Table 1) by firstly searching under the International 41 Nonproprietary Name (INN) of the drug. Secondly, if over 200 publications were found, the INN 42 combined with the MeSH term describing the pharmacovigilance problem was entered. Thirdly, the INN on September 29, 2021 by guest. Protected copyright. 43 44 combined with the terms "adverse drug reaction", "adverse event", "toxicity" and "poisoning" was used. 45 Furthermore, papers published up to one year after the drug withdrawals were included to account for the 46 potential delay between the end of a study and its publication. 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 21 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 9 Data was collated from these sources and a comprehensive literature review was conducted. However, a 10 11 fully systematic review was not carried out due to time and resource constraints. Any evidence cited in 12 these sources was categorised according to study design including animal studies, case reports (including 13 spontaneous case reports, published case reports and case series), case control studies, cohort studies, 14 15 RCTs, meta analyses and otherFor studies (including peer clinical trials review not randomized and/or not only controlled and 16 incidence studies). Only studies where statistically significant results were found were included. 17
18 19 RESULTS 20 Nineteen drugs were withdrawn for pharmacovigilance reasons in the EU from 2002 to 2011 (Table 12). 21 22 The anatomical therapeutic chemical (ATC) classes of drugs most represented in the nineteen drugs 23 withdrawn are "nervous system" (5/19 analgesics, antidepressants, antipsychotics and hypnotics), 24 25 "musculo skeletal system" (5/19 non steroidal anti inflammatory drugs and muscle relaxants) and 26 "alimentary tract and metabolism" (4/19 antidiabetics and antiobesity drugs). 27 28 29 The average median time to withdrawal was approximately 234 years (mean = 23.8) and with 95% 30 confidence, that estimate lies between 15 years and 33 years (95% CI 14.6 33.0) (Table 2). with an 31 interquartile range (IQR) of 4 and 46 (Table 1). 32 33 34 Case reports were cited in 95% of withdrawals (18/19) and case control studies (4/19), cohorts studies http://bmjopen.bmj.com/ 35 (4/19), RCTs (12/19) andor meta analysis (5/19) were cited in 63% of withdrawals (12/19) (Table 23). 36 37 38 Cardiovascular events or disorders were the main reason for withdrawal (9/19), followed by hepatic 39 40 disorders (4/19) and neurological or psychiatric disorders (4/19). 41 42 A list of the adverse drug reactions reported in EudraVigilance, correlating to the safety issue given as the on September 29, 2021 by guest. Protected copyright. 43 44 reason for withdrawal for each drug product is shown in Table 34. No results were returned for 45 ximelagatran / megalatran melagatran on querying the database for hepatotoxicity reports which is in line 46 with the EMA assessment which states that an RCT (EXTEND) study [7] ) elluicidatedelucidated this 47 48 safety concern [8] . A high number of reports were recorded for rosiglitazone; 10,834 cases of cardiac 49 disorders associated with its use. 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 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 DISCUSSION 9 10 This study has shown that the strength of evidence used to support the withdrawal of products relied on 11 information recorded from spontaneous reporting together with other pharmacoepidemiological and 12 clinical research studies. All decisions to revoke marketing authorisations described in this study utilised 13 14 spontaneous case reports, except for the withdrawal of ximegalatran ximelagatran [8]. Spontaneous case 15 reporting remains central toFor pharmacovigilance peer as it allows forreview the detection of suspected only adverse drug 16 reactions and significant safety signals. However, spontaneous reporting should be used in conjunction 17 18 with other pharmacovigilance methodologies such as observational studies and clinical trials to further 19 evaluate any such safety signals due to the potential for over and under reporting, confounding, bias and 20 misclassification. The randomised controlled trial is commonly assumed to be the gold standard for 21 22 clinical research [9] but adverse drug reactions may not become evident until a medicine is used in the 23 larger population, post authorisation, where patients are more likely to have co morbidities and other risk 24 25 factors. The outcome of pharmacoepidemiology studies and the conduct of systematic reviews, even with 26 their limitations particularly bias, confounding and misclassification, can provide useful additional 27 evidence on the nature and frequency of adverse drug reactions. As a direct country comparison, Clarke 28 et. al 29 [4] stated that six products were withdrawn in the UK in between 1999 and 2001, namely 30 grepafloxacin (marketed for 2 years), cisapride (12 years), pumactant (6 years), cerivastatin (4 years), 31 droperidol (30 years) and levacetylmethadol (4 years). The mean median length of time these drugs 32 33 remained on the UK market iswas approximately 10 5 years (IQR 4 10.5). Droperidol was available for 34 approximately 30 years as it was marketed before the licencing system began in 1971 (personal http://bmjopen.bmj.com/ 35 communication from MHRA (2013)). In contrast, only 3 products were withdrawn from the UK UK 36 37 market between 2009 and 2011; orciprenaline , sibutramine, and rosiglitazone , marking a 50% reduction 38 in the number of drug products withdrawn over the same time period. The mean median length of time 39 40 these products remained on the market was also 110 years (IQR 10.5 30). Although the length of time 41 taken to withdraw a drug product appears to have doubled (mainly due to the length of time orciprenaline 42 was available, 49 years) Although the time taken to withdraw a drug product appears unchanged over the on September 29, 2021 by guest. Protected copyright. 43 44 ten year period within the UK, the evidence used to aid the decision making has been heightened. 45 Spontaneous case reporting was used to support the case for withdrawal in 5 of the 6 products between 46 1999 and 2001, while only one randomized controlled trial was included as evidence [4]. In this study, 47 48 both spontaneous case reporting and randomized controlled trials were shown to have contributed to all 49 three withdrawn products, which suggests an increase in the level of evidence used to support decisions 50 51 drug product withdrawal from the market. 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 The findings of this study are in line with those of the study conducted by Paludetta and colleagues [6] in 9 2012 who observed a shift in the nature of the safety data used as justification for the withdrawal of 10 11 products in France in the time periods 1998 to 2004 and 2005 to 2011. One explanation the authors gave 12 for this change was the implementation of the European Clinical Trial Directive in 2004 and the 13 subsequent analysis of clinical trial safety data by pharmacovigilance experts. The evolution of 14 15 pharmacovigilance legislationFor over the past peer decade is likely review to have contributed to the only availability of 16 additional data through the conduct of post marketing safety surveillance allowing regulators and 17 companies to include such information in the drug products safety portfolio. It is difficult to ascertain 18 19 whether the introduction of risk management plans for newly licensed medicinal products has influenced 20 the length of time a drug with safety concerns remains on the market as only two drug products included 21 22 in this study were marketed after 2005. Rimonabant and sitaxentan were licensed in 2006 and removed 23 from the market in 2008 and 2010 respectively. At the time of authorisation, information was known 24 about the psychiatric side effects of rimonabant [10] and liver toxicity associated with sitaxentan [11] and 25 26 the benefit risk balance deemed to be favourable. However, in the case of rimonabant, as data became 27 available when used in the real life clinical setting, the risk of obese and overweight patients developing 28 psychiatric side effects was double that of patients taking placebo [10]. The marketing authorisation 29 30 holder voluntarily removed sitaxentan from the market and cancelled on going clinical trials as a result of 31 two cases of fatal liver injury [11]. During the time period studied, an additional twelve drugs were 32 33 withdrawn in some but not all EU countries. In some cases, it was not feasible to establish the reasons for 34 the withdrawal e.g. astemizole has been withdrawn from the UK, France and Spain, but information about http://bmjopen.bmj.com/ 35 its availability in other EU countries is difficult to establish. Another example, the NSAID nimesulide, 36 37 was withdrawn in Spain, Finland, Belgium and Ireland but is still available in fifteen EU member states. 38 The EMA had a favourable opinion of the risk benefit balance after a review of spontaneous reports 39 epipdemiological studies and other published studies in 2010 [12]. This was in spite of evidence of an 40 41 increased risk of hepatotoxicity compared with other NSAIDs, although restrictions are in place to limit
42 the number of liver related side effects. Ultimately, it would seem that there is a disparity in opinion of on September 29, 2021 by guest. Protected copyright. 43 risk benefit within the EU between decision makers at the EMA and various regulatory authorities. 44 45 46 The main limitation of this study was the impact of various language barriers across the EU. Websites of 47 48 competent authorities are written in their local language and resource constraints did not allow for 49 translations. Only information from the websites of those competent authorities where English and French 50 are spoken locally or an English or French version is available was utilised. This may also have had a 51 52 bearing on the studies included as evidence in this research as again only English and French publications 53 were included. After a website search, competent authorities and pharmaceutical companies were 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 contacted in order to obtain further information on drug product withdrawals. The response rate from 9 competent authorities was 70% and from pharmaceutical companies 55%. Again, this may have been due 10 11 to a language barrier problem, but it may also be due to a lack of available information on drugs 12 withdrawn for several years or an unwillingness to share information. Competent authorities or 13 companies may also not have a complete list of products removed from the market. Over time, the way in 14 15 which drug products have beenFor licensed inpeer the EU has changed review considerably. It is relativelyonly easy to 16 determine how products licensed centrally were withdrawn from the market. However, for drugs that 17 have been licensed for a considerable number of years, in countries that have recently joined the EU, and 18 19 for those that have multiple drug names; it can be very difficult to obtain conclusive, consistent and clear 20 information. The effect of different drug names may have led to potential exclusions of study evidence 21 22 from our PubMed search strategy as the INN of each drug was used rather than an exhaustive drug name 23 list. Due to time constraints, it was not possible to conduct a full investigation into why drug products 24 had been removed from the market in some but not all countries and to find detailed information about 25 26 the decision making process. 27 28 It would be useful to investigate the impact of the introduction of RMPs in a future study. This could be 29 30 done as an investigation of how results from post marketing studies, conducted as part of RMPs, aid in 31 the assessment of any safety issues. In theory, the introduction of RMPs should expedite postmarketing 32 33 drug safety decisions. 34 http://bmjopen.bmj.com/ 35 This study has shown that the strength of evidence has improved during the last ten years. Indeed, the 36 37 results show an increased use of case control studies, cohort studies, RCTs and meta analyses as 38 justification for the withdrawal of a marketed product. Previously only one third of decisions used 39 evidence from observational studies or clinical trials [4] but this study showed that they had contributed 40 41 to decision making in almost two thirds of withdrawals. Spontaneous case reports remain the most
42 significant method of pharmacovigilance. on September 29, 2021 by guest. Protected copyright. 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 25 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Table 1: Search strategy on PubMed 9 10 1 [International Nonproprietary Name (INN)] 11 12 2 [INN] AND [MeSH term describing the pharmacovigilance problem] 13 14 15 3 [INN] AND [adverseFor drug reaction peer OR adverse eventreview OR toxicity OR only 16 poisoning] 17
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 26 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. 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 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 27 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Table 1: List of drugs withdrawn for safety reasons in all EU member states between 2002 and 2011 grouped by 9 adverse drug reaction or safety concern.Table 2: Alphabetical list of drugs withdrawn for safety reasons in all EU 10 member states between 2002 and 2011. 11 12 Drug name Drug class or use Year first Year of Length of time on market Adverse reaction or safety marketed withdrawal (years) concern 13 Aceprometazine + Hypnotic 1988 2011 23 Cumulative adverse effects 14 Acepromazine + Misuse – Fatal side effect 15 Clorazepate For peer review only 16 17 Benfluorex Anorectic and 1974 2009 35 Heart valve disease – 18 hypolipidemic Pulmonary hypertension 19 Bufexamac NSAID ~1970 2010 40 Contact allergic reactions 20 Buflomedil Vasodilator (α1 and α2 1974 2011 37 Neurological and cardiac 21 receptor antagonist) disorders (sometimes fatal) 22 Carisoprodol Muscle relaxant 1959 2007 48 Intoxication Psychomotor 23 impairment Addiction – 24 Misuse Clobutinol Cough suppressant 1961 2007 46 QT prolongation 25 (centrally acting) 26 Dextropropoxyphene Opioid painkiller ~1960 2009 49 Fatal overdose 27 28 Lumiracoxib NSAID (COX 2 inhibitor) 2003 2007 4 Hepatotoxicity 29 30 Nefazodone Antidepressant 1994 2003 9 Hepatotoxicity 31 32 Orciprenaline Sympathomimetic (non 1961 2010 49 Cardiac disorders 33 specific β agonist) 34 Rimonabant Treatment of obesity 2006 2008 2 Psychiatric disorders http://bmjopen.bmj.com/ 35 (cannabinoid receptor 36 antagonist) 37 Rofecoxib NSAID (COX 2 inhibitor) 1999 2004 5 Thrombotic events
38 Rosiglitazone Anti diabetic treatment 2000 2010 10 Cardiovascular disorders 39 (PPAR agonist) 40 Sibutramine Treatment of obesity 1999 2010 11 Cardiovascular disorders 41 (serotonin noradrenaline
42 re uptake inhibitor) on September 29, 2021 by guest. Protected copyright. 43 Sitaxentan Antihypertensive 2006 2010 4 Hepatotoxicity 44 (endothelin receptor 45 antagonist) 46 Thioridazine Neuroleptic (α adrenergic 1958 2005 47 Cardiac disorders 47 and dopaminergic receptor 48 antagonist) 49 Valdecoxib NSAID (COX 2 inhibitor) 2003 2005 2 Cardiovascular and cutaneous disorders 50 Veralipride Neuroleptic (and 1979 2007 28 Neurologic and psychiatric 51 dopaminergic receptor disorders 52 antagonist) 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Ximelagatran / Anticoagulant (thrombin 2003 2006 3 Hepatotoxicity 9 Megalatran inhibitor) 10 Drug name Drug class or use Year first Year of Length of time on market Adverse reaction or safety 11 marketed withdrawal (years) concern Rofecoxib NSAID (COX 2 inhibitor) 1999 2004 5 Thrombotic events 12
13 Thioridazine Neuroleptic (α adrenergic 1958 2005 47 Cardiac disorders 14 and dopaminergic receptor 15 antagonist)For peer review only 16 Valdecoxib NSAID (COX 2 inhibitor) 2003 2005 2 Cardiovascular and cutaneous 17 disorders 18 Rosiglitazone Anti diabetic treatment 2000 2010 10 Cardiovascular disorders 19 (PPAR agonist) 20 Sibutramine Treatment of obesity 1999 2010 11 Cardiovascular disorders 21 (serotonin noradrenaline 22 re uptake inhibitor) 23 Orciprenaline Sympathomimetic (non 1961 2010 49 Cardiac disorders specific β agonist) 24 Benfluorex Anorectic and 1974 2009 35 Heart valve disease – 25 hypolipidemic Pulmonary hypertension 26 Clobutinol Cough suppressant 1961 2007 46 QT prolongation 27 (centrally acting) 28 Buflomedil Vasodilator (α1 and α2 1974 2011 37 Neurological and cardiac 29 receptor antagonist) disorders (sometimes fatal) 30 Veralipride Neuroleptic (and 1979 2007 28 Neurologic and psychiatric 31 dopaminergic receptor disorders 32 antagonist) 33 Rimonabant Treatment of obesity 2006 2008 2 Psychiatric disorders 34 (cannabinoid receptor http://bmjopen.bmj.com/ 35 antagonist) 36 Carisoprodol Muscle relaxant 1959 2007 48 Intoxication Psychomotor impairment Addiction – 37 Misuse 38 Aceprometazine + Hypnotic 1988 2011 23 Cumulative adverse effects 39 Acepromazine + Misuse – Fatal side effect 40 Clorazepate 41
42 Dextropropoxyphene Opioid painkiller ~1960 2009 49 Fatal overdose on September 29, 2021 by guest. Protected copyright. 43 44 Nefazodone Antidepressant 1994 2003 9 Hepatotoxicity 45 46 Ximelagatran / Anticoagulant (thrombin 2003 2006 3 Hepatotoxicity 47 Melagatran inhibitor) 48 Lumiracoxib NSAID (COX 2 inhibitor) 2003 2007 4 Hepatotoxicity
49 Sitaxentan Antihypertensive 2006 2010 4 Hepatotoxicity 50 (endothelin receptor 51 antagonist) 52 Bufexamac NSAID ~1970 2010 40 Contact allergic reactions 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Table 23: List of evidence used to support medicinal product withdrawals in all EU member states 9 between 2002 and 2011 derived from EMA reports, PubMed literature search and websites of competent 10 11 authorities.
12 Drug name Case reports Animal Case- Cohort RCTs Meta- †Others Formatted: Centered 13 studies control analysis Formatted Table Rofecoxib x x x x x 14 Formatted: Centered 15 Thioridazine x x x x x For peer review only Formatted: Centered Valdecoxib x x x 16 Formatted: Centered Rosiglitazone x x x x x 17 Formatted: Centered 18 Sibutramine x x x Formatted: Centered 19 Orciprenaline x x Formatted: Centered 20 Benfluorex x x x x Formatted: Centered 21 Clobutinol x x x Formatted: Centered 22 Buflomedil x x Formatted: Centered 23 Veralipride x Formatted: Centered 24 Rimonabant x x x Formatted: Centered 25 Carisoprodol x x x x x Formatted: Centered 26 Aceprometazine + x x Formatted: Centered 27 Acepromazine + Clorazepate 28 Dextropropoxyphene x x Formatted: Centered 29 Nefazodone x x Formatted: Centered 30 Ximelagatran / Melagatran x Formatted: Centered Lumiracoxib x x 31 Formatted: Centered Sitaxentan x x 32 Formatted: Centered Bufexamac x x x 33 Formatted: Centered 34 Drug name Case reports Animal Case- Cohort RCTs Meta- †Others http://bmjopen.bmj.com/ control analysis 35 studies 36 37 Aceprometazine + 38 x x Acepromazine + 39 Clorazepate 40 41 Benfluorex x x x x
42 Bufexamac x x x on September 29, 2021 by guest. Protected copyright. 43 44 Buflomedil x x 45 Carisoprodol x x x x x
46 Clobutinol x x x 47 48 Dextropropoxyphene x x 49 Lumiracoxib x x
50 Nefazodone x x 51 52 Orciprenaline x x 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Rimonabant x x x 9 Rofecoxib x x x x x 10 11 Rosiglitazone x x x x x 12 Sibutramine x x x 13 Sitaxentan x x 14 15 Thioridazine Forx x peerx reviewx x only 16 Valdecoxib x x x 17 Veralipride x 18 Ximelagatran / Megalatran 19 x 20 21 Formatted: Centered 22 † other studies include non randomized and/or not controlled clinical trials and incidence studies. 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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 31 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Table 34: Count of case reports of selected withdrawn drug products submitted to Eudravigilance up to 9 30 June 2012. 10 11 Drug name MedDRA Level Term Total 12 Rofecoxib PT Myocardial infarction 6711 13 14 Thioridazine HLGT Cardiac arrhythmias 179 PT Sudden death 28 15 Valdecoxib* ForSOC peerSkin and subcutaneous tissuereview disorders only317 16 SOC Cardiac disorders 313 17 Rosiglitazone* SOC Cardiac disorders 10834 18 19 Sibutramine PT Blood pressure increased 69 20 PT Heart rate increased 23 21 Orciprenaline HLGT Cardiac arrhythmias 15 22 23 Benfluorex HLGT Cardiac valve disorders 3439 24 PT Pulmonary hypertension 181 25 Clobutinol HLGT Cardiac arrhythmias 14 26 27 Buflomedil HLGT Cardiac arrhythmias 36 28 HLGT Fatal outcomes^ 14 29 HLGT Seizures (inc. subtypes) 35 Veralipride HLGT Anxiety disorders and symptoms 142 30 HLT Depressive disorders 221 31 Rimonabant* SOC Psychiatric disorders 918 32 PT Depression 545 33 Carisoprodol HLGT Mental impairment disorders 18 http://bmjopen.bmj.com/ 34 HLGT Movement disorders (inc. parkinsonism) 15 35 Aceprometazine + HLGT Fatal Outcomes 3 36 Acepromazine + Clorazepate 37 38 Dextropropoxyphene HLGT Fatal outcomes^ 55 39 HLGT Overdoses 91 40 Nefazodone HLGT Hepatic and hepatobiliary disorders 16 41 Ximelagatran / Melagatran 42 No results were returned on September 29, 2021 by guest. Protected copyright. 43 44 Lumiracoxib HLGT Hepatic and hepatobiliary disorders 92
45 Sitaxentan* PT Hepatitis 8 46 PT Hepatic Failure 6 47 Bufexamac SOC Immune systems disorders 10 48 49 50 Drug name MedDRA Level Term Total 51 Aceprometazine + HLGT Fatal Outcomes 3 Acepromazine + Clorazepate 52
53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 Benfluorex HLGT Cardiac valve disorders 3439 9 PT Pulmonary hypertension 181 10 Bufexamac SOC Immune systems disorders 10 11 Buflomedil HLGT Cardiac arrhythmias 36 12 HLGT Fatal outcomes^ 14 13 HLGT Seizures (inc. subtypes) 35 Carisoprodol HLGT Mental impairment disorders 18 14 HLGT Movement disorders (inc. parkinsonism) 15 15 Clobutinol ForHLGT peerCardiac arrhythmias review only14 16 17 Dextropropoxyphene HLGT Fatal outcomes^ 55 18 HLGT Overdoses 91 19 Lumiracoxib HLGT Hepatic and hepatobiliary disorders 92 20 21 Nefazodone HLGT Hepatic and hepatobiliary disorders 16 22 23 Orciprenaline HLGT Cardiac arrhythmias 15 24 25 Rimonabant* SOC Psychiatric disorders 918 26 PT Depression 545 27 Rofecoxib PT Myocardial infarction 6711
28 Rosiglitazone* SOC Cardiac disorders 10834 29 30 Sibutramine PT Blood pressure increased 69 31 PT Heart rate increased 23 32 Sitaxentan* PT Hepatitis 8 33 PT Hepatic Failure 6 http://bmjopen.bmj.com/ 34 Thioridazine HLGT Cardiac arrhythmias 179 35 PT Sudden death 28 36 Valdecoxib* SOC Skin and subcutaneous tissue disorders 317 37 SOC Cardiac disorders 313 38 Veralipride HLGT Anxiety disorders and symptoms 142 39 HLT Depressive disorders 221 Ximelagatran / Megalatran 40 No results were returned 41
42 Key: HLGT = Higher Level Group Term; PT = Preferred Term; SOC = System Organ Class on September 29, 2021 by guest. Protected copyright. 43 *- centrally authorised drug products. Information available on: www.adrreports.eu. Other information received 44 by email from EMA. 45 ^- the original condition, preceding the fatal outcome is unknown therefore there is a possibility of duplicate reporting. 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 33 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 9 ACKNOWLEDGEMENTS 10 11 The Drug Safety Research Unit (DSRU) is a registered independent charity (No.327206) associated with 12 the University of Portsmouth. The DSRU receives unconditional donations from pharmaceutical 13 companies. The companies have no control on the conduct or the publication of its studies. The DSRU 14 15 has received such funds fromFor the manufacturers peer of products included review in this study. only 16 17 All authors have completed the Unified Competing Interest form atwww.icmje.org/coi_disclosure.pdf and declare: 18 19 no support from any organisation for the submitted work; Prof. Shakir reports personal fees from ICON (Contract 20 research organization), personal fees from Shire Pharmaceuticals, personal fees from ONO Pharmaceuticals, 21 personal fees from Intermune Pharma, personal fees from IPSEN, outside the submitted work; Mr Huet reports 22 personal fees from Axpharma, personal fees from Thea laboratories, outside the submitted work; Dr McNaughton 23 reports no financial relationships with any organisations that might have an interest in the submitted work in the 24 previous three years, no other relationships or activities that could appear to have influenced the submitted work. 25
26 27 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide 28 licence to the Publishers and its licensees in perpetuity, in all forms, formats and media (whether known now or created in the 29 future), to i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, 30 create adaptations, reprints, include within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) 31 create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the 32 inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) licence any third 33 party to do any or all of the above. 34 http://bmjopen.bmj.com/ 35 The lead author (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent 36 account of the study being reported; that no important aspects of the study have been omitted; and that any 37 discrepancies from the study as planned have been explained. 38 39 40 REFERENCE LIST 41 (1) European Medicines Agency (EMA) online. Available from URL: http://www ema europa eu 42 on September 29, 2021 by guest. Protected copyright. 43 2013; [Accessed on 28 March 2013]. 44 45 (2) European Medicines Agency (EMA). Guideline on good pharmacovigilance practices (GVP) 46 47 Module V – Risk management systems. http://www ema europa 48 49 eu/docs/en_GB/document_library/Scientific_guideline/2012/06/WC500129134 pdf 2013; 50 51 [Accessed 16/05/2013]. 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 35 BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 (3) Arnaiz JA, Carné X, Riba N, Codina C, Ribas J, Trilla A. The use of evidence in 9 10 pharmacovigilance. Case reports as the reference source for drug withdrawals. Eur J Clin 11 12 Pharmacol 2001;57(1):89 91 13 (4) Clarke A, Deeks JJ, Shakir SA. An assessment of the publicly disseminated evidence of safety 14 15 used in decisions toFor withdraw medicinal peer products fromreview the UK and US markets. only Drug Saf 16 17 2006;29(2):175 81. 18 19 (5) Olivier P,Montastruc J L. The nature of the scientific evidence leading to drug withdrawals for 20 21 pharmacovigilance reasons in France. Pharmacoepidemiol Drug Saf 2006;15(11):808 12 22 (6) Paludetto M N, Olivier Abbal P, Montastruc J L. Is spontaneous reporting always the most 23 24 important information supporting drug withdrawals for pharmacovigilance reasons in France? 25 26 Pharmacoepidemiol Drug Saf 2012;21(12):1289 94. 27 28 (7) Agnelli G, Eriksson B I, Cohen A T, Bergqvist D, Dahl O E, Lassen M R, Mouret P, Rosencher 29 30 N, Andersson M, Bylock A, Jensen E, Boberg B; EXTEND Study Group. Safety assessment of 31 32 new antithrombotic agents: lessons from the EXTEND study on ximelagatran. Thromb Res 33
2009;123(3):488 97. http://bmjopen.bmj.com/ 34 35 (8) European Medicines Agency (EMA). Press Release: AstraZeneca Withdraws its Application for 36 37 Ximelagatran 36 mg film coated tablets. http://www ema europa 38 39 eu/docs/en_GB/document_library/Press_release/2010/02/WC500074073 pdf 2013; [Accessed 40 41 15 Mar 2013]
42 on September 29, 2021 by guest. Protected copyright. 43 (9) Carne X, Arnaiz J A. Methodological and political issues in clinical pharmacology research by 44 the year 2000. Eur J Clin Pharmacol 2000;55:781 5. 45 46 (10) European Medicines Agency (EMA). Assessment Report for Acomplia (Rimonabant). Available 47 48 from URL: http://www ema europa eu/docs/en_GB/document_library/EPAR_- 49 50 _Assessment_Report_-_Variation/human/000666/WC500021280 pdf 2009; [Accessed 28 51 52 March 2013]. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from
1 2 3 4 5 6 7 8 (11) European Medicines Agency (EMA). Press release: Thelin (sitaxentan) to be withdrawn due to 9 10 cases of unpredictable serious liver injury. Available from URL: http://www ema europa 11 12 eu/ema/index jsp?curl=/pages/news_and_events/news/2010/12/news_detail_001161 13 jsp&murl=menus/news_and_events/news_and_events jsp&mid=WC0b01ac058004d5c1 14 15 2010; [Accessed 28 MarchFor 2013]. peer review only 16 17 (12) European Medicines Agency (EMA). Questions and Answers on the Review of Systemic 18 19 Medicines Containing Nimesulide. Available from URL: http://www ema europa 20 21 eu/docs/en_GB/document_library/Referrals_document/Nimesulide_31/WC500107957 pdf 22 2012; [Accessed on 28 March 2013]. 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41
42 on September 29, 2021 by guest. Protected copyright. 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