Cholesterol trials and mortality.

Abstract

An overview of clinical trials can reveal a class effect on mortality that is not apparent from individual trials. Most large trials of lipid pharmacotherapy are not powered to detect differences in mortality and instead assess efficacy with composite cardiovascular endpoints. We illustrate the importance of all-cause mortality data by comparing survival in three different sets of the larger controlled lipid trials that underpin meta-analyses. These trials are for fibrates and statins. Fibrate treatment in five of the six main trials was associated with a decrease in survival, one fibrate trial showed a non-significant reduction in mortality that can be explained by a different target population. In secondary prevention, statin treatment increased survival in all five of the main trials, absolute mean increase ranged from 0.43% to 3.33%, the median change was 1.75%, which occurred in the largest trial. In primary prevention, statin treatment increased survival in six of the seven main trials, absolute mean change in survival ranged from -0.09% to 0.89%, median 0.49%. Composite safety endpoints are rare in these trials. The failure to address composite safety endpoints in most lipid trials precludes a balanced summary of risk-benefit when a composite has been used for efficacy. Class effects on survival provide informative summaries of the risk-benefit of lipid pharmacotherapy. We consider that the presentation of key mortality/survival data adds to existing meta-analyses to aid personal treatment decisions.