Poor correlation between published methods to predict creatinine clearance and measured creatinine clearance in...

The purpose of this study was to evaluate the predictive ability of six published creatinine clearance (CCr) equations in healthy human immunodeficiency virus (HIV) infected individuals. A 24-h urine collection to determine CCr was done on an out-patient basis in 18 subjects. Predicted CCr was compared with the measured values, and the predictive performance was assessed with percentage mean error (bias) and percentage root mean error (precision). Mean +/- standard deviation measured CCr was 107 +/- 35 mL/min/1.73 M2. CCr determined using each of the published equations correlated poorly with measured values. CCr determined using Hull methods was significantly different from the measured values. Though Cockcroft and Gault and Jelliffe methods had the lowest mean bias and greater precision, a significant range of difference from measured CCr was observed (-12 to +28%). All methods over estimated the measured CCr in HIV-infected individuals. Until other approaches are developed, a 24 h urine collection may be the best approach for assessing renal function in HIV-infected individuals, especially in those receiving medications with narrow therapeutic indices that are cleared by the kidney.