MEASURED VERSUS ESTIMATED CREATININE CLEARANCE IN A HIGH-FUNCTIONING ELDERLY SAMPLE - MACARTHUR-FOUNDATION STUDY OF SUCCESSFUL AGING

Objective: To assess the validity of several equations for estimating creatinine clearance in a large sample of high-functioning, community- dwelling elderly. Design: Serum and 12-hour urine samples were collect ed and assayed for creatinine using the Jaffe total chromagen method. Fifteen clearance-estimating equations were evaluated for bias, accura cy, correlation with measured clearance values, and frequency of erron eous placement into renal function categories. Stepwise regression mod eling and reliability testing were performed on a split sample to cons truct and assess a novel creatinine-clearance-estimating equation.Sett ing: New Haven, Connecticut, East Boston, Massachusetts, and a five-co unty region in and around Durham, North Carolina. Participants: A subs ample of community-dwelling men and women (age range 70-79 years) from the Established Populations for Epidemiological Studies of the Elderl y was screened for physical and cognitive functioning and placed into high-, medium-, and low-functioning groups (n = 1354). High-functionin g respondents who provided blood and complete urine samples (n = 762) were included in the present study. Results: In general, estimated cre atinine clearance was more closely correlated to measured values in ma les than in females. Most equations underestimated creatinine clearanc e, with average bias ranging from -33.1 mL/min to +19.6 mL/min. Predic tive accuracy ranged from 18.2 mL/min to 38.0 mL/min. Equations were v ariable in their erroneous placement of individuals into renal functio n categories. Regression modeling yielded an equation which contained novel components but failed to provide better estimates of creatinine clearance than those already available. Conclusions: The equations eva luated here provide unacceptable predictions of creatinine clearance i n normally aging individuals. We advocate the use of serum drug concen tration measurements when available and encourage investigation into t imed urine collections of short duration as alternatives to clearance- estimating equations in the elderly.