Continuous renal replacement therapy (CRRT) is used to treat renal failure in children. Despite widespread use of the technique, little research has evaluated how variations in dose of replacement fluid or degree of fluid overload at initiation relate to outcomes. We conducted a retrospective review of patients treated with convective CRRT at our institution, using a multivariable Cox regression model. Children with high fluid overload (>10%) at CRRT initiation were at 3.02 times greater risk of mortality than those with low or no fluid overload [95% confidence interval (CI) 1.50-6.10, P=0.002]. The hazard ratio for death in children treated with high-dose convective clearance was not statistically significant. Our data support previous findings that volume overload in excess of 10% is strongly correlated with poor outcome. We favor early institution of CRRT, before excessive fluid overload occurs. In contrast to findings in adults, we find no advantage to higher rates of convective clearance. Given the risks and increased complexity associated with high-volume hemofiltration, we recommend further study prior to widespread adoption of high-level convection in children treated with continuous veno-venous hemofiltration.