The accuracy of different methods of calculating 24-hour creatinine clearance in patients with unstable renal function was compared using simulated data (based on a one-compartment pharmacokinetic model), as well as data from postrenal transplant patients. When creatinine clearance was calculated from the urinary creatinine excretion and a serum creatinine concentration, the use of the midpoint serum creatinine concentration produced the lowest degree of error. Therefore, this method is recommended for routine clinical determination of creatinine clearance in such patients. When the urinary creatinine excretion was unknown, an iteration method produced the lowest degree of error among four methods, and therefore is recommended to estimate creatinine clearance in such patients.