Open Access Research BMJ Open: first published as 10.1136/bmjopen-2016-011666 on 30 June 2016. Downloaded from Regulatory approval of pharmaceuticals without a randomised controlled study: analysis of EMA and FDA approvals 1999–2014 Anthony J Hatswell,1 Gianluca Baio,1 Jesse A Berlin,2 Alar Irs,3 Nick Freemantle4 To cite: Hatswell AJ, Baio G, ABSTRACT et al Strengths and limitations of this study Berlin JA, . Regulatory Introduction: The efficacy of pharmaceuticals is most approval of pharmaceuticals often demonstrated by randomised controlled trials ▪ without a randomised This work is the first systematic attempt to iden- (RCTs); however, in some cases, regulatory controlled study: analysis of tify drugs approved without randomised EMA and FDA approvals applications lack RCT evidence. evidence. 1999–2014. BMJ Open Objective: To investigate the number and type of ▪ The in-depth review identifies all evidence for 2016;6:e011666. these approvals over the past 15 years by the European treatments throughout the clinical development doi:10.1136/bmjopen-2016- Medicines Agency (EMA) and the US Food and Drug programme, not just the pivotal study. 011666 Administration (FDA). ▪ Because of the different remits and processes of Methods: Drug approval data were downloaded from the European Medicines Agency and the Food ▸ Prepublication history for the EMA website and the ‘Drugs@FDA’ database for all and Drug Administration, it is not possible to this paper is available online. decisions on pharmaceuticals published from 1 compare across all disease areas. To view these files please January 1999 to 8 May 2014. The details of eligible ▪ Despite identifying treatments without rando- visit the journal online applications were extracted, including the therapeutic mised evidence, it is not in the scope of this (http://dx.doi.org/10.1136/ area, type of approval and review period. study to conclude on the appropriateness of bmjopen-2016-011666). Results: Over the period of the study, 76 unique approval on the basis of non-randomised data. ▪ Received 25 February 2016 indications were granted without RCT results (44 by Due to the lack of follow-up studies, it is also Revised 27 May 2016 the EMA and 60 by the FDA), demonstrating that a not possible to reach a conclusion on the effi- Accepted 1 June 2016 substantial number of treatments reach the market cacy of these products. http://bmjopen.bmj.com/ without undergoing an RCT. The majority was for haematological malignancies (34), with the next most common areas being oncology (15) and metabolic properly designed and conducted, provide un- conditions (15). Of the applications made to both biased estimates of treatment effect.1 However, agencies with a comparable data package, the FDA there are occasions when a therapy is adminis- granted more approvals (43/44 vs 35/44) and took less tered to all patients within a trial; this is a fre- time to review products (8.7 vs 15.5 months). quent step in the development process of Products reached the market first in the USA in 30 of pharmaceuticals. While it is not necessarily a on September 28, 2021 by guest. Protected copyright. 34 cases (mean 13.1 months) due to companies preferred approach, the product can be sub- making FDA submission before EMA submissions and mitted to regulatory agencies for approval fol- faster FDA review time. lowing these studies. This can occur with Discussion: Despite the frequency with which approvals are granted without RCT results, there is no treatments that are new to the market, or for the extension of a licensed treatment to a dif- 1Department of Statistical systematic monitoring of such treatments to confirm Science, University College their effectiveness or consistency regarding when this ferent patient population, with the argument London, London, UK form of evidence is appropriate. We recommend a that clinical equipoise no longer applies. As 2Johnson & Johnson, more open debate on the role of marketing the treatment is obviously effective, there is no Titusville, New Jersey, USA authorisations granted without RCT results, and the clinical uncertainty, and as such, to withhold 3 Department of development of guidelines on what constitutes an the treatment from one group in a trial would Pharmacology, University of acceptable data package for regulators. 2–4 Tartu, Tartu, Estonia be unethical or impractical. 4Department of Primary Care Uncontrolled clinical studies may take & Population Health, several forms that include aspects of random- University College London, isation (figure 1). For example, patients may London, UK INTRODUCTION be randomised to different dosages (or regi- fi Correspondence to The ef cacy of pharmaceuticals is usually mens) of the experimental treatment. Other Dr Nick Freemantle; nicholas. evaluated for regulatory purposes using ran- data packages may include a clinical study [email protected] domised controlled trials (RCTs) that, when compared to a historical control, such as a Hatswell AJ, et al. BMJ Open 2016;6:e011666. doi:10.1136/bmjopen-2016-011666 1 Open Access BMJ Open: first published as 10.1136/bmjopen-2016-011666 on 30 June 2016. Downloaded from Figure 1 Types of uncontrolled studies used to support regulatory applications compared to randomised controlled trials. previously published study of similar patients treated available regulatory documents. In the case of the EMA, with the relevant comparator.5 Finally, other treatments the main sources of information were the European may have no control arm or control data, but instead Public Assessment Reports and the Committee of rely on an ‘obvious’ difference, or avoidance of an Human Medicinal Products Assessment Reports, while expected outcome—for example, death from poisoning. for the FDA, these were the approved label and the The common factor in each of these designs is that the Medical Officer’s report. study in isolation is insufficient to demonstrate the All sources were taken directly from the websites of effectiveness of the treatment (ie, external data are the EMA and the FDA. The EMA approval database was needed) as it does not contain an internal control arm downloaded on 8 May 2014 (http://www.ema.europa. (either active or placebo) against which outcomes can eu/ema/), and the FDA ‘Drugs@FDA’ database was be compared. downloaded on 15 May 2014 (http://www.fda.gov/Drugs/). Following the clinical study programme, the data These databases were used to explore drug approvals package for a treatment is submitted to the regulator, based on uncontrolled trial results from 1 January 1999 who must give their approval before a drug is allowed on to 8 May 2014. This date range was chosen as it started themarket.IntheEUandtheUSA,thereareslightlydif- several years after the beginning of the EMA Centralised ferent paths to approval. Historically, pharmaceutical Procedure (1995), giving time for the process to become http://bmjopen.bmj.com/ licensing in the EU was the responsibility of individual established. Each approval was reviewed in turn to iden- countries; however, with the development of the cen- tify applications made without supportive RCT results. tralised authorisation procedure by the European Following deduplication, all products not approved by Medicines Agency (EMA) in 1995, companies submit a both agencies were searched for individually on the dossier to the EMA. The EMA reviews the submission, website of the other agency with no date restriction to asks questions to the company as a part of the process, understand whether they were approved before the date and then issues a decision that is binding throughout the period. Searches were also conducted on the website of 6 EU. In the USA, manufacturers submit a New Drug the manufacturer, and in PubMed and Google News for on September 28, 2021 by guest. Protected copyright. Application (NDA) to the Food and Drug Administration information regarding submissions, in order to identify (FDA) which will review the data submitted before any products that did not appear in the databases use making a decision on the request.7 for the main analysis. The objective of this study was to review new pharma- The aim of our study was to look at newly approved ceutical approvals by the EMA and the FDA, from 1999 indications for pharmaceuticals, which led to the exclu- to 2014, to understand how frequently applications (and sion of several types of product, listed in table 1. approvals) are made solely on the basis of uncontrolled Applications were compared against the exclusion cri- studies, without either the pivotal study or supportive teria, and if excluded, the reason for exclusion was studies being RCTs in the indication (previous work has noted in a hierarchical fashion. If a regulatory applica- focused solely whether the pivotal study contained a tion for a product was made to one agency but did not control arm8). A secondary analysis compared the appear in the results for the other agency, a search with approval rates of the EMA and the FDA for evidence of no date restriction on was performed to identify any differences in the approach taken to comparable appli- applications made outside the review window (to ensure cations in approval rate and timing. all comparable approvals have been included). The included data were for each approved indication, METHODS other approved indications for the treatment, submission Details of the trials conducted for the licensing applica- and decision dates, disease area and supporting evidence tion of treatments were identified through publicly for the approval. 2 Hatswell AJ, et al. BMJ Open 2016;6:e011666. doi:10.1136/bmjopen-2016-011666 Open Access BMJ Open: first published as 10.1136/bmjopen-2016-011666 on 30 June 2016. Downloaded from Table 1 Exclusion criteria and rationale Exclusion criteria Rationale Generic drugs Licensed on the basis of similarity to existing drugs and would result in duplication Biosimilar drugs The interpretation of data from trials of biosimilars is likely to be informed by data available regarding the original drug (in mechanism and effect on a condition).