BACKGROUND: Some studies have suggested that early institution of renal replacement therapy (RRT) might be associated with improved outcomes in patients with acute renal failure (ARF). STUDY DESIGN: A systematic review and meta-analysis of randomized controlled trials and cohort comparative studies to assess the effect of early RRT on mortality in patients with ARF. SETTING & POPULATION: Hospitalized adult patients with ARF. SELECTION CRITERIA FOR STUDIES: We searched several databases for studies that compared the effect of "early" and "late" RRT initiation on mortality in patients with ARF. We included studies of various designs. INTERVENTION: Early RRT as defined in the individual studies. OUTCOMES: The primary outcome measure was the effect of early RRT on mortality stratified by study design. The pooled risk ratio (RR) for mortality was compiled using a random-effects model. Heterogeneity was evaluated by means of subgroup analysis and meta-regression. RESULTS: We identified 23 studies (5 randomized or quasi-randomized controlled trials, 1 prospective and 16 retrospective comparative cohort studies, and 1 single-arm study with a historic control group). By using meta-analysis of randomized trials, early RRT was associated with a nonsignificant 36% mortality risk reduction (RR, 0.64; 95% confidence interval, 0.40 to 1.05; P = 0.08). Conversely, in cohort studies, early RRT was associated with a statistically significant 28% mortality risk reduction (RR, 0.72; 95% confidence interval, 0.64 to 0.82; P < 0.001). The overall test for heterogeneity among cohort studies was significant (P = 0.005). Meta-regression yielded no significant associations; however, early dialysis therapy was associated more strongly with lower mortality in smaller studies (n < 100) by means of subgroup analysis. LIMITATIONS: Paucity of randomized controlled trials, use of variable definitions of early RRT, and publication bias preclude definitive conclusions. CONCLUSION: This hypothesis-generating meta-analysis suggests that early initiation of RRT in patients with ARF might be associated with improved survival, calling for an adequately powered randomized controlled trial to address this question.