Trends in characteristics of cardiovascular clinical trials 2001-2012

Background Efficient conduct of clinical trials is essential for the timely generation of critical medical knowledge. Methods We systematically assessed size, duration, enrollment rates, and geographic distribution of randomized cardiovascular trials published between 2001 and 2012 in the 8 highest-impact journals in general medicine and cardiology. Results Of the 1,224 trials, 27.0% were conducted in North America, 36.5% in Western Europe, and 7.7% in other countries, and 28.8% were multiregional. Trials enrolled a median of 452 patients (interquartile range 167-1,530) in 20 sites (2-76). Median duration was 2.1 (1.3-3.3) years, with an estimated enrollment rate of 1.1 (0.5-3.5) patients/site per month. Between 2001-2003 and 2009-2012, the proportion of North American trials decreased from 34.5% to 25.7% (P =.006), whereas that of multiregional trials (from 26.0% to 30.3%; P =.046) and trials conducted in other countries (from 4.6% to 10.3%; P =.012) increased. Over time, trials involved more patients (from 400 to 500 [median]; P =.032) and sites (from 20 to 22; P =.049), multiregional trials involved more countries (from 12 to 18; P =.031), and enrollment rate declined from 1.2 to 0.9 patients/site per month (P =.017). The proportion of trials meeting their primary end point ("positive") decreased from 69% to 57% (P <.001). Trials with higher enrollment rates were more likely to be positive (odds ratio 1.20 per doubling, 95% CI 1.12-1.29), as were industry-sponsored compared with government-sponsored trials (odds ratio 2.62, 95% CI 1.67-4.12). Conclusions From 2001 to 2012, cardiovascular clinical trials have become larger, more global, and less likely to meet their primary end point. Enrollment rates have declined, requiring more sites and regions.