Stroke clinical trials under enroll women, due in part to upper age limits

Acute stroke trials underenrolled women, and trials with an upper age threshold of 80 years further limited enrollment, according to findings from a database search of more than 100 trials published in JAMA Neurology.

“Under enrollment of women (relative to their frequency in the underlying stroke population) raises questions about the generalizability and representativeness of the data,” Mathew J. Reeves, BVSc, PhD, FAHA, professor in the department of epidemiology and biostatistics at Michigan State University, told Healio Neurology. “It threatens the validity of the evidence base if the trials that we base much of our clinical guidelines on include fewer women. More directly, it means that any sex-specific estimates of treatment effects are going to be more imprecise because they are based on fewer women.”

Reeves and colleagues performed a database search for articles between January 1, 2010, and June 11, 2020, on acute stroke therapeutic trials to determine the extent of underenrollment while accounting for the underlying prevalence of stroke in men and in women. Eligible articles included final primary results from phase 2 or 3 randomized clinical trials that enrolled patients with an acute stroke of any type and applied a therapeutic intervention within 1 month. Two independent authors extracted data from the trials and matched it to the proportion of women in underlying stroke populations as estimated by the Global Burden of Disease database. Researchers quantified the enrollment of women using a modified version of the enrollment disparity difference (EDD).

The search resulted in 1,529 results. After applying eligibility criteria, the analysis included 115 trials with 121,105 randomized patients. Female patients accounted for 52,522 (43.4%) of the trial participants.

The random-effects pooled EDD was -0.053 (95% CI, -0.065 to -0.040), indicating an underenrollment of women by an absolute difference of 5.3% in relation to their representation the underlying stroke population. Enrollment in the 115 trials varied; nine studies had an overenrollment of women with an EDD of 0.117 (95% CI, 0.084-0.15) with underenrollment of 6.7% in the remaining 106 trials (summary EDD, 0.067; 95% CI, 0.078 to 0.057). Trials of secondary prevention therapies exhibited the greatest enrollment disparities regarding sex.

“We expected to see a larger deficit of women in the clinical trials that examined thrombolysis (tPA) and endovascular therapy (EVT), but we did not see that. Overall, the deficit of women in these 115 trials was 5% lower, but this was the same for the tPA and EVT trials,” Reeves said. “The other surprise was that trials that had women leadership (a woman was listed as the first author or senior author) did no better in terms of the equitable enrollment of women compared with trials led by men.”

During a metaregression analysis, the researchers also found that trials with an upper age limit of 80 years or younger produced a 6% decrease in women enrollment despite it being well-known that first stroke occurs more often in women at an older age.

“Women with stroke tend to be older than men, so more of them will be excluded. This approach ends up being both ageist as well as sexist,” Reeves said. “We recommend that trial organizers replace age criteria with whatever age-associated characteristics they are really worried about (eg, frailty, fall risk, bleeding risk). The exclusion criteria should be specific to these features and not old age itself. Removing the age-based exclusion criteria could have a big effect on enrollment.”

In a related editorial, Safi U. Khan, MD, MS, of the department of medicine at West Virginia University, Morgantown, and Erin D. Michos, MD, MHS, of the division of cardiology at John Hopkins School of Medicine, discussed the historical precedent of sex disparities in scientific research.

“In a study of 60 lipid-lowering therapy trials, only 53% [of] trials reported outcomes according to sex,” they wrote. “This is important, since failure to disaggregate data between sexes generates inequitable treatments and health outcomes.”

Citing stroke as the second leading cause of death worldwide in 2016, including the deaths of 2.6 million women, Khan and Michos called on researchers to maintain proportionate enrollment in cardiovascular trials, as lesser representation risks “further compounding the existing cardiovascular health crisis.”

Reference:

Khan SU, et al. JAMA Neurol. 2021;doi:10.1001/jamaneurol.2021.0624.