On the design and analysis of multicenter trials in acute renal failure

Improving outcomes for patients with acute renal failure (ARF) is one of the most urgent tasks for 21st century nephrologists and other critical care physicians. The incidence of hospital-acquired ARF is rising (to 5% or more), partly because of changes in the demographic and clinical characteristics of the hospitalized patient population. When ARF is severe enough to require dialysis, in-hospital mortality rates are distressingly high, in excess of 50% in most published studies. In the past several years, a number of attempts have been made to influence the course of ARF, either relatively early (eg, atrial natriuretic peptide [ANP], insulin growth factor-1 [IGF-1]) or coincident with the dialysis procedure in severe cases (eg, variations in dialysis membrane and modality). Several lessons can be learned from these efforts. This article will briefly address issues in the design and analysis of clinical trials, focusing on elements of special interest to practitioners of renal medicine.