Predicting GFR in children and adults: a comparison of the Cockcroft-Gault, Schwartz, and modification of diet in renal disease formulas.

BACKGROUND: A rapid prediction of glomerular filtration rate (GFR) is often needed in clinics. Formulas based on plasma creatinine level are being increasingly used, Schwartz for children, supposed to give GFR; Cockcroft-Gault for adults, supposed to indicate the creatinine clearance; and a recent formula introduced by the Modification of Diet in Renal Disease (MDRD) group. Our objective was to test whether one single formula could suffice and which one gives the best estimation of GFR.

METHODS: In 198 children (with two kidneys, single kidney, or renal transplant) and 116 adults (single kidney and transplanted), we measured inulin clearance and creatinine clearance and calculated Cockcroft-Gault, MDRD and, in children only, Schwartz. Data were compared with analysis of variance (ANOVA), regression statistics, and concordance studies.

RESULTS: In patients over 12 years of age, Cockcroft-Gault was almost similar to GFR corrected for body surface and creatinine clearance exceeded GFR by more than 20%; Schwartz was above creatinine clearance excepted for transplanted children. In younger children, no prediction approached GFR. Predictions were well correlated with GFR, but concordance studies showed Schwartz with dispersed results and GFR overestimated (20 mL/min/1.73 m2); Cockcroft-Gault was close to GFR and results were dispersed; MDRD in children gave a large overestimation and badly dispersed results; in transplanted adults its prediction was good.

CONCLUSION: Cockcroft-Gault prediction could be used for children over 12 years of age and adults; it should not be considered as creatinine clearance but as GFR corrected for body surface, it is merely a prediction, 95% of the results are between +/-40 mL/min/1.73 m(2) in children and +/-30 mL/min/1.72 m(2) in adults. In younger children no formula is satisfying.