Fifteen studies were included in the review (2955 patients): three RCTs (143 patients), two prospective (1,480 patients) and 10 retrospective studies (1,332 patients). Both randomised and observational studies were of relatively low quality.
Early renal replacement therapy was associated with a 29% reduction in risk of mortality compared with late renal replacement therapy but with high heterogeneity (RR 0.71; 95% CI 0.59 to 0.86; 15 studies; Ι²=79%).
Subgroup analyses found a similar reduction in mortality for early renal replacement therapy compared with late renal replacement therapy when using continuous renal replacement therapy but with less heterogeneity (RR 0.69; 95% CI 0.56 to 0.84; seven studies, Ι² =33%). There was a 74% reduction in risk of mortality for early renal replacement therapy compared with late renal replacement therapy when using intermittent haemodialysis (RR 0.26; 95% CI 0.15 to 0.45; two studies; Ι²=0%). The pooled effect for the mixed modality suggested no benefit (RR 0.94; 95% CI 0.88 to 1.01; six studies; Ι²=81%) but there was very high heterogeneity.
Sensitivity analyses found no impact when removing one study at a time on continuous renal replacement therapy or mixed modality. However, one study on intermittent haemodialysis found no statistically significant difference.
The authors stated that according to their funnel plot analysis there was no significant evidence of publication bias. But acknowledged it wasn't possible to rule out that such bias was present due to the small number of included studies.