Children on continuous renal replacement therapy: prognostic factors.

OBJECTIVES To identify prognostic factors in children receiving continuous renal replacement therapy. DESIGN Historical cohort study. SETTING Neonatal and paediatric intensive care unit of a Hong Kong hospital. PATIENTS Neonatal or paediatric patients who received continuous renal replacement therapy from January 1998 to December 2008. RESULTS In all, 37 patients who received 39 episodes of continuous renal replacement therapy were identified. The male-to-female ratio was 1.5:1. Among the 39 episodes, 15 (39%) were performed on neonates with a mean birth weight of 2.6 (standard deviation, 0.7; range, 0.9-3.7) kg, and 24 (62%) were performed on paediatric patients with a mean age of 7.9 years (standard deviation, 6.4 years; range, 6 months to 18 years). The overall mortality was 41%; in the neonatal and paediatric groups it was 60% and 29%, respectively. There was no significant difference in the mean and maximal ultrafiltration rate in survivors and non-survivors. Multivariate analysis identified the PRISM III score and fluid overload as independent predictors of mortality. Kaplan-Meier survival analysis showed that patients with pre-continuous renal replacement therapy fluid overload of 5.5% or more was associated with reduced survival in the intensive care unit as compared to those having less severe fluid overload (P=0.011). In neonatal patients, there was a higher proportion with multi-organ failure and severe fluid overload. CONCLUSION High PRISM III scores and the degree of pre-continuous renal replacement therapy fluid overload were independent predictors of mortality.