The Importance of Subgroup Analysis in Drug-Eluting Stent Trials.

SEE PAGE 28 C ardiovascular drugs and devices may have different safety and efficacy profiles in certain populations. Performance of separate trials or subgroup analysis is important if there is a biological reason, either intrinsic and/or extrinsic, for a varied response to a therapy. For example, compared with men, women have different body size, hormonal milieu, sex-specific physiology, and sociocultural influences. Randomized controlled trials are generally powered for the primary outcome in all enrolled individuals, and the results are generalized to subgroups. Pre-specified or post-hoc analyses of subgroups within a study are often exploratory due to lack of randomization and small numbers. Caution must be exercised in the number of subsets examined due to the potential for chance findings. In addition, interaction terms in statistical models should only be pursued if there is an a priori hypothesis being examined. In trials of percutaneous coronary intervention (PCI), the low enrollment of women has led to less robust data and the inability to study the interaction between sex and other important clinical characteristics, for example, diabetes mellitus and chronic kidney disease (CKD), which are associated with high rates of major adverse cardiovascular events in PCI trials. In an effort to examine the performance of drug-eluting stents (DES) in women, the Society for Cardiovascular Angiography and Interventions’ Women in Innovation Initiative held a Gender Data Forum. Through collaborative efforts of cardiology societies, study investigators, and industry, an individual patient-level