Performance of creatinine clearance equations on the original Cockcroft-Gault population

Abstract

Prediction of endogenous creatinine clearance by mathematical equations such as the Cockcroft-Gault formula is used in clinical practice in spite of the reported concern for their limited predictability. The aim of this study is to determine whether the measured creatinine clearance can be predicted accurately by a number of published equations including the recently modified Cockcroft-Gault formula = Cockcroft-Gault formula x 1.73 m2/body surface area from the original Cockcroft-Gault population. The performance of the mathematical equations in patients with different creatinine clearance and body mass indices was assessed by computing accuracy at different percentiles, bias and precision from the original Cockcroft-Gault data. Refitting the modified formula to the Cockcroft-Gault data gave superior results compared to the original Cockcroft-Gault formula with an overall accuracy in the general and subgroup analysis above 70% agreement within 30% estimate of the measured creatinine clearance. On the other hand, analysis of the other equations, including the original Cockcroft-Gault, demonstrated a limited accuracy to predict creatinine clearance particularly in patients with creatinine clearance below 50 ml/min with an overall accuracy in less than 1/3 of the calculated creatinine clearance within 30% range from the measured creatinine clearance. The current creatinine clearance equations and even the original Cockcroft-Gault formula did not accurately predict the measured creatinine clearance. Normalization for body surface area in the original Cockcroft-Gault formula demonstrated more accuracy to estimate creatinine clearance, particularly in patients with diminished renal function and is recommended to physicians who wish to use the Cockcroft-Gault formula in their practice until more credible formulas are developed.