An Investigation Into Drug Products Withdrawn from the EU Market Between 2002 and 2011 for Safety Reasons and the Evidence Used to Support the Decision- Making

Open Access Research 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 Downloaded from An investigation into drug products withdrawn from the EU market between 2002 and 2011 for safety reasons and the evidence used to support the decision- making Rhian McNaughton,1,2 Gwenaël Huet,1 Saad Shakir1,2 To cite: McNaughton R, ABSTRACT Strengths and limitations of this study Huet G, Shakir S. An Objectives: The objective of this study was to investigation into drug determine the nature of evidence used to support the ▪ products withdrawn from the This study adds to the knowledge base regarding withdrawal of marketing authorisations of drug EU market between 2002 and the level of evidence used to support decisions 2011 for safety reasons and products for safety reasons throughout the European to remove drug products from the market. the evidence used to support Union (EU) between 2002 and 2011. ▪ Results of this research are in line with the find- the decision-making. BMJ Setting: Products withdrawn, either by a medicines ings of other study groups which concluded that Open 2014;4:e004221. agency or a marketing authorisation holder, during the there has been a shift in the level of evidence doi:10.1136/bmjopen-2013- period 2002–2011 were identified by conducting used, for example, results of randomised con- 004221 detailed searches of the WHO, the European Medicines trolled trials and cohort studies are used to Agency (EMA) and national medicines agency websites justify a product withdrawal. ▸ Prepublication history for throughout the EU plus Norway, Iceland and ▪ English and French websites and publications this paper is available online. Liechtenstein. The scientific evidence used to support were only searched. To view these files please the decision was identified from a search within ▪ The international non-proprietary name of each visit the journal online PubMed, the EMA and national medicines agencies drug was included in the PubMed search strat- (http://dx.doi.org/10.1136/ websites. Information about spontaneous case reports egy but as various drug names exist for the bmjopen-2013-004221). http://bmjopen.bmj.com/ entered into EudraVigilance and unavailable on the same compound, a potential exists for the exclu- Received 10 October 2013 EMA website was received by email from the EMA. sion of some publications from contributing Revised 21 November 2013 Results: 19 drugs were withdrawn from the market, evidence. Accepted 29 November 2013 throughout the EU, for safety reasons from 2002 to 2011. Case reports were cited in 95% of withdrawals (18/19) and case–control studies (4/19), cohort studies contribute to the withdrawal such as the emer- (4/19), randomised controlled trials (RCTs) (12/19) or gence of new safety evidence and significant meta-analysis (5/19) were cited in 63% of withdrawals drug interactions. Ultimately, the risk-benefit (12/19). Cardiovascular events or disorders were the balance is paramount when making such deci- on September 24, 2021 by guest. Protected copyright. main reason for withdrawal (9/19), followed by hepatic sions. The regulatory authorities review the evi- disorders (4/19) and neurological or psychiatric dence, such as spontaneous case reports or disorders (4/19). results from a randomised controlled trial Conclusions: This study has shown that the level of (RCT) and prepare an opinion on whether evidence used to support drug withdrawal has the marketing authorisations of medicinal improved during the past 10 years, with an increased products containing that particular drug use of case–control studies, cohort studies, RCTs and meta-analyses. This research has demonstrated that should be maintained, changed, suspended such studies have contributed to decision-making in or withdrawn. The interpretation of the avail- almost two-thirds of cases. able safety data by the numerous competent authorities across the European Union (EU) is wide ranging as is the accessibility of infor- 1Drug Safety Research Unit, mation regarding how they reach their Southampton, UK conclusion. 2University of Portsmouth, INTRODUCTION In 2005, legislation was introduced to Portsmouth, UK The decision to remove a drug product from ensure marketing authorisation holders com- Correspondence to the market is not taken lightly either by the plete a risk management plan (RMP) for Dr Rhian McNaughton; regulatory authority or the marketing author- their newly licensed product incorporating, [email protected] isation holder. A number of reasons where appropriate, post-marketing safety McNaughton R, Huet G, Shakir S. BMJ Open 2014;4:e004221. doi:10.1136/bmjopen-2013-004221 1 Open Access BMJ Open: first published as 10.1136/bmjopen-2013-004221 on 15 January 2014. Downloaded from surveillance studies in the general population with not included if only one particular dose, dosage form or regular assessment and appropriate reporting to the indication was withdrawn. CHMP.1 The aim of the RMP is to ensure a more pro- The scientific evidence leading to the withdrawal deci- active approach to pharmacovigilance by putting in sion was identified from a search within PubMed, the place measures that allow for the early detection and EMA website (published documents searched included: minimisation of risks throughout a medicine’s lifecycle. ‘Press releases’, ‘Questions and Answers’ and ‘Scientific This should, in theory, result in a reduction in the conclusions’) and national medicines agencies websites. length of time for drug withdrawal should a safety issue Within the ‘Questions and Answers’ reports and ‘Press arise. Risk management planning in the EU has been releases’, prepared by the EMA, a summary of the further enhanced by the introduction of new EU legisla- reviewed evidence with reference to study type (eg, RCT, tion for good pharmacovigilance practices which animal study) is often included. A request was also sub- became applicable in July 2012.2 mitted to the EMA for information on the number of Previous studies looked at the type of evidence used to cases entered into EudraVigilance related to the reason support the withdrawal of products, for safety reasons, in of withdrawal for each drug. different individual countries and over varying time A search strategy was developed in PubMed by first, periods. Arnaiz et al3 reported on products withdrawn in searching under the International Non-proprietary Spain between 1990 and 1999, Clarke et al4 for products Name (INN) of the drug. Second, if over 200 publica- in the USA and the UK from 1999 to 2001 and Olivier tions were found, the INN combined with the MeSH and Montastruc5 for medicines in France from 1998 to term describing the pharmacovigilance problem was 2004. They all concluded that case reports were the entered. Third, the INN combined with the terms mainstay of evidence used to support withdrawal deci- ‘adverse drug reaction’, ‘adverse event’, ‘toxicity’ and sions for pharmacovigilance reasons; case–control ‘poisoning’ was used. Furthermore, articles published up studies, cohort studies and RCTs were used as evidence to 1 year after the drug withdrawals were included to in less than one-third of decisions. Paludetto et al6 revis- account for the potential delay between the end of a ited the issue 6 years later, between 2005 and 2011, and study and its publication. reported a shift in the level of evidence used. They Data were collated from these sources and a compre- determined that, of the 22 active substances withdrawn hensive literature review was conducted. However, a fully from the French market within this 6-year time frame, systematic review was not carried out due to time and 68% of the decisions used multiple sources of evidence resource constraints. Any evidence cited in these sources including clinical trials and pharmacoepidemiological was categorised according to study design including studies. The objective of this study was to explore the animal studies, case reports (including spontaneous case – level of evidence used to support the withdrawal of mar- reports, published case reports and case series), case http://bmjopen.bmj.com/ keting authorisation of drug products in the time period control studies, cohort studies, RCTs, meta-analyses and 2002–2011 in the EU as a whole. other studies (including clinical trials not randomised and/or not controlled and incidence studies). Only studies where statistically significant results were found METHODS were included. Drug products withdrawn from the EU market for safety or pharmacovigilance reasons, either by a medicines agency or a marketing authorisation holder, during the on September 24, 2021 by guest. Protected copyright. period 2002–2011 were identified. This was established RESULTS by carrying out detailed searches on the websites of the Nineteen drugs were withdrawn for pharmacovigilance WHO, the European Medicines Agency (EMA) and reasons in the EU from 2002 to 2011 (table 1). The ana- those of national medicines agencies throughout the EU tomical therapeutic chemical classes of drugs most plus Norway, Iceland and Liechtenstein. The WHO represented in the nineteen drugs withdrawn are Pharmaceutical Newsletter and the EMA press releases ‘nervous system’ (5/19 analgesics, antidepressants, anti- provided the most useable information. In addition, psychotics and hypnotics), ‘musculoskeletal system’ pharmaceutical companies were contacted to obtain (5/19 non-steroidal anti-inflammatory drugs (NSAIDs) more information concerning dates of withdrawal in spe- and muscle relaxants) and ‘alimentary tract and metab- cific countries. olism’ (4/19 antidiabetics and antiobesity drugs). Prescription and over-the-counter products were The median time to withdrawal was 23 years with an included if their marketing authorisation was withdrawn IQR of 4 and 46 (table 1). or suspended due to any safety reason such as an Case reports were cited in 95% of withdrawals (18/19) adverse drug reaction, abuse or misuse. Active ingredi- and case–control studies (4/19), cohorts studies (4/19), ents contained in a combination product were also RCTs (12/19) and meta-analysis (5/19) were cited in included.

Load more