Underrepresentation of Women in Late-Breaking Cardiovascular Clinical Trials

Introduction

Cardiovascular disease (CVD) is the leading cause of death among both women and men, yet the inclusion of women in clinical trials remains suboptimal. ^1,2 Meaningful sex-based differences in symptomatology, pathophysiology, and clinical outcomes of CVD have been demonstrated. ³ A dearth of sex-specific data as it pertains to women and CVD persists, making risk–benefit decisions with regard to drug, device, and behavioral interventional therapies uncertain and limit the application of novel therapies. ^4,5 As such, the federal organizations, including the National Institutes of Health and the U.S. Food and Drug Administration, have mandated demographic analyses and proposed a coordinated approach to the inclusion of women via the Health of Women Program Strategic Plan, resulting in an increase in the inclusion of women and minority populations in clinical trials, as well as more sex-specific analyses of results. ⁶

However, prior research has suggested that women remain underrepresented in clinical trials, particularly in trials that address device-based therapies, coronary artery disease, and heart failure. ^{7 –9} A recent analysis suggested that on average, individuals who identify as women make up about 25% of major clinical trials evaluating drug-eluting stents, ventricular assist devices, implantable cardioverter defibrillators, and cardiac resynchronization defibrillators. ¹⁰ Other studies have found that relative to the prevalence in the disease population, women have been more adequately enrolled in trials of hypertension and pulmonary artery hypertension, as compared with arrhythmia and heart failure trials. ¹ Similar trends have been identified in other medical subspecialties, including neurology, as well as in gastroenterology. ^{11 –13}

Among clinical trials, inadequate enrollment, trial retention, and sex-specific reporting are particularly prominent in interventional and device-based trials. ^14,15 Late-Breaking Clinical Trials (LBCTs) at national conferences garner wide publicity and dissemination, often being advertised prominently to attendees as the most cutting-edge, highly-anticipated science of the year and being presented on the main stage at flagship conferences. LBCT are frequently highlighted in subsequent publications, including review articles summarizing high-impact trials from major conferences, reaping additional notice. ^16,17 However, the representation of women in high-impact trials presented at major cardiovascular conferences remains unknown. This analysis seeks to characterize the inclusion of women in the LBCT at three major cardiovascular meetings to better understand current barriers to enrollment, reporting, and retention.

Methods

Results

Of 68 trials, inclusion of women ranged from 0% to 71%, with a mean of 35.2% ± 13.6%. In total, 29.4% of trials had <25% female enrollment and 85.3% of trials had <50% female enrollment. Only 47.1% of trials reported sex-specific analyses. Of trials reporting racial demographics, on average 72.9% of participants were white and mean age of participation 65.1 years (Table 2). Of trials included, 32.4% addressed general or preventative interventions, 22.1% heart failure, 14.7% interventional, 19.1% electrophysiology, and 11.8% other interventions. Medication-based therapies were assessed in 53.0% of trials, while interventional procedures were assessed in 27.9% of trials. The remaining 19.1% of trials addressed a variety of interventions, including cardiac rehabilitation, nutritional interventions (e.g., alcohol use, coffee consumption), educational curricula (e.g., clinician counseling as a quality improvement initiative), telemedicine programs (e.g., home blood pressure monitoring, remote access to lipid counseling), and electronic medical record based interventions (e.g., calculation of 1 year mortality).

The average IPR was 0.76 ± 0.27 for all trials and did not vary significantly based on conference, trial center, geographic region, or funding source (data not shown), with only 15 trials exceeding an IPR of 1.0 (Fig. 1). The average IPR varied by subspecialty, with a statistical difference noted between interventional cardiology and heart failure (0.65 vs. 0.88, p = 0.02). In addition, the average IPR was significantly lower for procedure-based trials compared with medication trials (0.61 vs. 0.78, p = 0.008), mean age <65 years of age versus ≥65 years (0.67 vs. 0.81, p = 0.03), and those with <1500 participants versus ≥1500 participants (0.69 vs. 0.84, p = 0.03). Only 21.2% of the authorship block was women, with 4/68 women presenting authors and 8/55 women senior authors, although without statistical difference in IPR by presenting or senior author gender.

OPEN IN VIEWER

FIG. 1.

Inclusion of women in late-breaking clinical trials by prevalence of disease. (A) IPR by subspecialty, (B) IPR by intervention, (C) IPR by funding source, (D) IPR by conference. Thick solid line indicates median value, thin dashed line indicates mean value. IPR, inclusion to prevalence ratio.

Discussion

LBCTs present novel, practice-changing evidence to the medical community on a prominent stage and as such, should represent the gold standard for diversity, equity, and inclusion in clinical trial enrollment. This analysis of three major cardiovascular conferences in 2021 suggests that previously documented low inclusion of women in clinical trials, particularly in interventional and procedural spheres, persists for LBCTs despite ongoing efforts to achieve gender parity. ⁴

Equitable trial enrollment has been a growing focus for both academic and professional organizations over the past decade. Despite well-documented disparities, coordinated strategic initiatives to incentivize the inclusion of female participants are lacking. Investigators have postulated that contributors to limited trial enrollment of women may fall into several spheres of influence, including the patient, clinical care team, local research team, clinical trial infrastructure, and broader societal expectations. ¹⁵ Strategies to augment the enrollment of women and underrepresented minorities in clinical trials may include reducing barriers to access interventional care and referral to subspeciality care, as well as extending age-based inclusion criteria. The COVID-19 pandemic has shed light upon the utility of pragmatic trial design in the acute care setting, allowing for expanded outreach capabilities and exploration of virtual interventions and follow-up. ^23,24

Outside of drug- and device-based clinical trials, the remaining trials of nutritional- and cardiac rehabilitation-based interventions demonstrate a signal toward improved enrollment of women (IPR 0.94), although a specific subgroup analysis was not performed due to small subgroup size. Similarly, trials of telemedicine- and electronic medical record-based therapies demonstrate a signal toward improved enrollment of women, highlighting an opportunity to further explore the role for virtual, telemedicine-based clinical trial platforms to augment the enrollment of women. Telemedicine may reduce barriers to trial participation by decreasing physical, logistical, and transportation barriers for participants.

CVD presents often at a later age in women as compared with men; as such, previous studies have demonstrated that greater female enrollment in trials may be achieved through the recruitment of older population. ²⁵ This finding was reflected in the LBCTs assessed, with trials that had a mean age <65 years, including fewer women, while trials that had a mean age >65 enrolling a higher proportion of women. Future analyses may consider calculating age-stratified IPRs by disease state to identify whether and the extent to which liberalizing the upper age limit for inclusion might improve equitable gender enrollment. Special attention should also be paid to diversification of clinical trial leadership, as a diverse team may be more likely to enroll a more diverse population. In fact, a recent study found that cardiovascular trials led by women enrolled more female participants than trials enrolled by men (44.9% vs. 37.9%, p < 0.001). ²⁶

In this sample, only 4/68 reported first authors were female, 8/55 reported last authors were female, and 21% of the total authorship block were female, demonstrating significant room for growth. Specific benchmarks for diversity of gender should be set a priori in trial design with predetermined checkpoints and benchmarks to assess for goal attainment.

It should be noted that this analysis has several limitations. Although LBCTs garner additional attention on a prominent stage at national flagship conferences, presentation under the designation of “late-breaking” may also reflect the timing of data collection and analysis; future studies may consider assessment of trials designated as “Key Sessions” or those presented on the main stage at large conferences. Previous analyses of Altimetric scores to denote overall trial impact have noted a discrepancy between media attention and number of citations, suggesting that digital attention may stand in contrast to academic importance, limiting the use of such metrics. ²⁷ Second, a single year of LBCT data from three major conferences may not be generalizable across years. In particular, the COVID-19 pandemic may influence both trial enrollment and study publication, given that women were disproportionately affected by the pandemic in terms of care-taking burden, employment, grant funding, and publication. ²⁸ Third, the sample size of 68 studies may preclude detection of statistically significant differences in many study characteristics, given limited statistical power.

Despite gradual improvement in the inclusion of women in clinical trials, we find that enrollment of women continues to lag in LBCT at the three largest cardiology conferences. We encourage cardiology societies to select—and stakeholders to design—high-impact trials that enroll patients that are representative of the population.