Estimating creatinine clearance in children: comparison of three methods

Three methods of estimating creatinine clearance (Clcr) in children were compared in pediatric patients ranging in age from 1 to 16 years. Sixty-eight measured Clcr determinations, normalized to ml/min/1.73 sq m of body surface area, were performed on 58 patients. Estimated Clcr values were calculated using the methods of Schwartz et al. (Method I), Shull et al. (Method II), and Rudd et al. (Method III), and each method was correlated with the measured creatinine clearance. Mean prediction error (estimated Clcr minus measured Clcr) also was compared for the methods to assess the relative prediction accuracy. Measured and estimated Clcr values correlated well (Method I, r = 0.90; Method II, r = 0.85; Method III, r = 0.80; p less than 0.0001). Mean prediction errors for Methods I and III (4.1 and -2.8 ml/min/1.73 sq m, respectively) were not significantly different; however, both methods were significantly more accurate (p less than 0.01) than Method II (mean prediction error: -17.4 ml/min/1.73 sq m). In this study, Method I, which was based on height and serum creatinine, was found to have the greatest accuracy and simplicity and to be the most clinically useful method of the three methods. The applicability of the three methods to pediatric patients with severe renal disease is uncertain.