Meta-analysis, Clinical Trials, and Transferability of Research Results Into Practice

Objective: To evaluate, in the comprehensive scenario of "evidence-based" medicine, the transferability of the results of published randomized clinical trials and meta-analyses on cholesterol-lowering interventions to clinical practice. Method: Overview of randomized clinical trials on cholesterol-lowering interventions in the secondary prevention of coronary heart disease. Results: The present overview on secondary prevention of coronary heart disease included 34 trials with cholesterol-lowering interventions in 24 968 individuals. There was a 12.5% mortality in the group that was allocated active intervention and a 17.2% mortality in the control group (risk reduction, 13%; 95% confidence interval, —19% to —6%). Coronary and cardiovascular odds of deaths were significantly reduced. No clear association was found between noncoronary mortality and cholesterol-lowering interventions. Baseline total cholesterol levels had no clear influence on total mortality. Intermediate (10%-20%) and high (>20%) total cholesterol reductions were associated with similar reductions in the odds of death (—23% and —30%, respectively). No conclusion could be reached for patients who were less represented in the studies (ie, women and elderly persons). Patients with more complicated baseline clinical conditions (eg, congestive heart failure) had little nonsignificant benefit from cholesterol-lowering interventions. Conclusions: The effect of cholesterol-lowering interventions at least in the secondary prevention of coronary heart disease can be considered as established, but the transferability of such results to real-life patients remains the critical, unanswered question. (Arch Intern Med. 1996;156:1158-1172)