The logarithmic relationship of serum creatinine and creatinine clearance was analyzed in 100 adult patient studies using a geometric regression technique. Each sex was independently analyzed, and the subsequently derived regression formulae were age corrected resulting in mathematical models useful in estimating creatinine clearance from serum creatinine concentrations. These formulae were tested prospectively in another group of 100 patient studies in which creatinine clearance had been determined, and the results compared to values derived by use of two other published formulae. This newer method resulted in a closer distribution of data around a line of identity compared to other formulae and allowed for a good "bedside" estimation of creatinine clearance from serum creatinine concentration.