Sex differences in early study discontinuation and study drug withdrawal: insights from REDUCE-IT
M F Van Der Bijl, D L Bhatt, H M den Ruijter, I C D Westendorp, S B Ketchum, A Lira Peneda, A Schut, Y Appelman, E Boersma, J E Roeters van LennepAbstract
Background
Efforts to improve enrollment of women in cardiovascular disease (CVD) trials have not addressed retention after enrollment. Analyses of the LoDoCo2 and pooled TIMI trials suggest that women may be more likely than men to prematurely discontinue study drug or withdraw consent. Whether this represents a broader phenomenon and the factors underlying these differences remain uncertain.
Purpose
To assess whether sex differences in retention observed in coronary artery disease trials apply to the REDUCE-IT trial, which includes patients with both established and high-risk CVD.
Methods
Post hoc analysis of the randomised, double-blind, placebo-controlled REDUCE-IT trial of statin-treated patients with established CVD or diabetes plus additional risk factors, controlled LDL cholesterol, and moderately elevated triglycerides, randomised to icosapent ethyl 4 g daily or placebo. Early study discontinuation and study drug withdrawal were defined as permanent cessation of the trial (medication) for reasons other than a primary nonfatal end point, death, or administrative censoring. Cox proportional hazards models were used to assess sex differences in early study discontinuation and study drug withdrawal. The multivariable model adjusted for covariates identified as significant in univariable analyses (including treatment allocation, eGFR, and smoking status) and was stratified by geographic region, cardiovascular risk, and ezetimibe use.
Results
Among 8,179 participants, 28.8% were women. During a median follow-up of 4.9 years, early study discontinuation occurred in 14.2% of women and 9.1% of men (HRadj, 1.40; 95% CI, 1.22–1.62, P < 0.0001). Women discontinued earlier than men (median (Q1, Q3), 697 (340, 1145) vs 728.5 (324, 1227) days), resulting in reduced potential follow-up time. Consistent patterns were observed for women and men in both placebo (HRadj, 1.49; 95% CI, 1.22–1.81, P <0.0001) and active therapy arms (HRadj, 1.30; 95% CI, 1.05–1.61, P = 0.01). Study drug withdrawal occurred in 41.4% of women and 32.2% of men, in both groups due to patient preference (47.1%), adverse event (27.6%), other (20.4%), physician decision (4.0%) and lack of efficacy (0.9%), (HRadj, 1.31; 95% CI, 1.21–1.42, P < 0.0001). Women withdrew their study drug earlier than men (median (Q1, Q3), 645 (266, 1105) vs 704 (276, 1233) days), resulting in reduced potential time on trial medication. Consistent patterns were observed for women and men in both placebo (HRadj, 1.36; 95% CI, 1.22–1.52, P < 0.0001) and active therapy arms (HRadj, 1.25; 95% CI, 1.11–1.41, P = 0.0003).
Conclusions
Women in the REDUCE-IT trial had higher rates of early study discontinuation and drug withdrawal than men, regardless of treatment. This reduced their representation and exposure over time, potentially affecting generalisability. Strategies to improve retention of women in CVD trials are needed, and further research is warranted to understand the underlying causes.