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.