Surrogate end-points vs. all-cause mortality as end-points in mega trials on post-myocardial infarction.

To provide secondary prevention after MI, a dozen strategies (aspirin, beta blocker, streptokinase, verapamil, ACE inhibitor, nitrate, HMG coA reductase inhibitor, heparin, estrogen, CABG, exercise, coumadin) reduce morbidity and mortality. Prerequisites in mega trials are overall deaths vs. surrogate endpoints; reinfarction, coronary deaths; patent arteries, atherosclerosis regression, improved ejection fraction, PVC suppression, etc. For event rates of 10-15%, an effective therapy requires 10-50% reduction of risk, an N of 1×103 to 3.6 × 104, a sufficient population from which to choose (i.e., 6-60% of screened or 6×102 to 6.4×105) and 100ʹs of 106 dollars. Post myocardial infarction event rate in patients with ejection fraction <40% is high - 10-40% per year. In comparison, pristine hypertensives, i.e., SHEP (event rate 1% per year) with 36 and 27% reductions in stroke/ischemic HD events and LV mass, yielded no overall mortality benefit. Ventricular ectopic suppression (class IC agents) in CAST led to increased overall mortality. Surrogates must be juxtaposed reliably to mortality overall. Single trials with inferior “N” may be combined to metaanalyses with significant mortality reduction. Sometimes, meta analysis is to analysis as metaphysics is to physics. Surrogate vs. mortality endpoints is not a difficult choice; in either case, the result should be relevant to the appropriate patient, other trials, and pathogenic mechanisms.