Existing methods for predicting creatinine clearance provide accurate estimates for normal-weight patients but not for patients who are obese. Studies into this problem began with an animal model of obesity, the obese overfed rat. Mean creatinine clearance was found to vary in direct proportion to fat-free body mass, determined in both obese and normal animals. The relevance of this observation to renal function in humans was evaluated by analyzing published studies reporting creatinine clearance and creatinine excretion rates in obese and normal persons. Measured creatinine clearance correlated well with estimated fat-free body mass (r = 0.772, p less than 0.02), and urinary excretion of creatinine normalized to fat-free mass correlated impressively with age (r = 0.960). Formulas derived from these observations allow for the prediction of creatinine clearance at steady state: (formula; see text) In initial tests of these formulas, their predictions appeared to be as accurate as existing methods for the normal-weight population and far superior to these methods when applied to the obese population. Therefore, when creatinine clearance is not measured in obese patients, the estimation of this parameter with the proposed formulas should improve the ability to select the appropriate dose for drugs that are cleared principally by renal filtration.