Evaluating Sex Differences in Medical Device Clinical Trials
16 Mar 2012
Proponiamo un articolo pubblicato online su JAMA il 29 febbraio 2012, che illustra i contenuti del documento redatto nel dicembre scorso dalla US Food and Drug Administration sull’importanza dell’arruolamento delle donne negli studi clinici.
The safety and effectiveness of medical treatments can differ in men and women for many reasons related to different epidemiologic characteristics, physiology, and body size. In general, women have higher bleeding rates and procedural morbidity and mortality than men, which means that their risk/benefit ratios for many implanted medical devices can differ from men. Therefore, sex-specific safety and effectiveness data are necessary for informed patient decision making. In 2008, the US Food and Drug Administration (FDA) convened 2 workshops on this topic that included multiple stakeholders—patient groups, medical device industry representatives, academia, and government officials. In December 2011, the FDA released a draft guidance informed by these workshops, entitled “Evaluation of Sex Differences in Medical Device Clinical Studies.”1 This document discusses the underrepresentation of women in medical device clinical trials and provides recommendations for increasing enrollment of women and performing sex-specific analyses. A 90-day public comment period began on December 19, 2011.
The FDA draft guidance document provides a number of helpful guidelines to support adequate enrollment of women in clinical trials. For studies that have not yet been completed, the draft guidance recommends keeping enrollment open until a prespecified number of women are enrolled. Following this recommendation so that women are enrolled in proportion to disease prevalence would do the most to ensure sufficient data about the effect of a medical device on women. The draft guidance document also recommends that outcomes data be analyzed by sex. By not limiting its guidance to new studies, the FDA is indicating that it hopes changes will occur more immediately. This approach is laudable; the difference between men and women in terms of effectiveness and safety of medical devices has long been understudied. The recommendations in this draft guidance may help ensure safe and effective use of medical devices in both sexes.
However, the notation on this guidance stating that it “Contains Nonbinding Recommendations” raises the fundamental question of whether these important guidelines will be followed. This is not a new question. In 1994, FDA issued a “Gender Bias” directive stating that FDA Summaries of Safety and Effectiveness Data, which are released on premarket approval for all high-risk medical devices, should examine clinical trial data for differences in safety and efficacy by sex. However, an examination of compliance with this directive found that it is routinely ignored. Among 78 high-risk cardiovascular devices approved during the years 2000 through 2007, the overall patient population in these studies was 67% men2 with no increase in the enrollment of women over time,3 although women make up 46% of all patients with cardiovascular disease.
The analyses by sex specified in the 1994 FDA directive occurred in only 41% of studies (51/123).3 When these analyses were completed, 26% (12/47) identified some difference in device safety or effectiveness by sex. In 28% of device studies (34/123), no sex distribution of the patient population was stated. As a result, clinically important sex-specific differences related to use of many medical devices are unknown.
For example, the implantable cardioverter-defibrillator (ICD), a device shown to be lifesaving in selected populations, is commonly implanted in men and women. At the time of device approval, there were gaps in the data because of very low enrollment of women in the pivotal studies. In 2003 the Medicare Coverage and Advisory Committee was convened pursuant to a manufacturer’s request for expanded Medicare coverage of ICDs for primary prevention (which was granted). During the meeting, when asked why there were so few women in the pivotal Multicenter Automatic Defibrillator Implantation Trial II, the trial’s principal investigator stated, “we are contemplating a trial in the future to almost exclusively focus on women, because we don’t think they have been adequately represented.”4 Nearly a decade later, after hundreds of thousands of ICD implantations, that trial has not started.
The consequences of this failure are now emerging. A meta-analysis of randomized controlled trials showed that ICDs are not associated with reductions in all-cause mortality when the device is used for primary prevention in women with heart failure.5 An exploration of sex-specific data for ICD usage prior to approval and coverage of these devices would have better informed patient and physician use and recommendations. However, after FDA approval of this device for use in both sexes, device sponsors have no incentive to complete studies that may identify risk and subsequently limit sales.Another illustrative example is a left ventricular assist device that was approved for patients with advanced heart failure based on data from just 44 women (23% of study population).6 Although women were noted to have a 3-fold increase in stroke risk compared with men as well as trends toward increased bleeding and infection,7 the device manufacturer specifically advertises that this device’s “small size and quiet operation make the HeartMate II suitable for a wider range of patients, including women and those of smaller stature.”8 Such marketing, while perhaps well meaning, is misleading in the absence of sufficient numbers of women to ensure safety and effectiveness.
Although the FDA required a postapproval study on “gender-specific outcomes,”7 the initial publication on the postmarketing study did not include these data.The experience of the National Institutes of Health (NIH) with sex bias guidances also suggests that enforcement, tracking, and consequences for noncompliance are necessary. The NIH adopted a policy statement in 1986 encouraging enrollment of women in all federally funded research but noted a few years later that little progress had been made. Thus, in the early 1990s, inclusion of women became a requirement for NIH funding with annual progress reports. These changes have had a substantial effect.The NIH policy illustrates 2 lessons for the FDA—active assessment of compliance and consequences for noncompliance. Real-time transparency, with reporting through FDA-TRACK9 at monthly intervals for clinical trials that are actively recruiting, would increase the likelihood that adequate numbers of women are enrolled. Tracking should involve specific information about the percentage of women enrolled in studies. If the FDA monitors progress in real time and notes the preferential exclusion of women, it can prospectively signal to device sponsors that the deficiencies in enrollment of women need to be addressed for the device to be considered for approval.The FDA should return device applications that have not enrolled sufficient numbers of women (or justified the reason for insufficient enrollment of women) or performed sex-specific analyses per the guidance document. Specifically, they should be performed for primary effectiveness end points, primary safety end points, and key secondary end points.
This step would ensure that before approval, medical devices have been sufficiently evaluated in women as well as men. The FDA should also make sex-specific data available for devices that are not approved. Additionally, the FDA should have a postmarketing requirement that the manufacturers of all implanted medical devices provide additional outcomes data after 1 year, again with consequences (fines) for noncompliance.Transparency is critical in this process and should not just apply at the trial planning and recruitment phase; sex-specific data from FDA submissions should also be made publicly available. The agency should be required to disclose if companies do not comply with the sex bias guidance. The practice of keeping sex-specific analyses “confidential” must end.For the excellent provisions in this new draft FDA guidance to have a positive effect on patient health, they must be implemented.1 It is time that FDA approval of medical devices is based on data that they are safe and effective in both women and men.
