The reliability of different formulae to predict creatinine clearance

Verhave JC, Balje-Volkers CP, Hillege HL, de Zeeuw D, de Jong PE (University Medical Center Groningen, Groningen University, Institute of Drug Exploration, Groningen, the Netherlands). The reliability of different formulae to predict creatinine clearance. J Intern Med 2003; 253: 563–573. Objectives. Creatinine clearance (CCR) is a commonly used tool to measure glomerular filtration rate (GFR) in clinical practice. This tool requires collection of 24-h urine, which is quite bothersome. Several different formulae have been used to estimate GFR using plasma creatinine and other easy formulae to obtain biometrical data. We examined 10 formulae and compared them with actually measured CCR in a large sample of the general population. Design. Cross-sectional cohort study. Setting. University hospital outpatient clinic, a population based study. Subjects. A total of 8592 inhabitants of the city of Groningen, 28–75 years of age. The cohort is enriched for microalbuminuria. Results. In general, the formulae did not give an accurate estimation of CCR, particularly not in male and in obese subjects. Six formulae, including the Cockcroft–Gault gave a fairly good estimation of CCR in the overall population and in subgroups of specific gender, body mass index and age. All formulae however, gave an underestimation of the measured CCR in higher ranges of CCR and an overestimation in the lower ranges. Moreover, the age-related decline of CCR is hard to approximate with a formula. Conclusions. We conclude that formulae to estimate CCR in the general population, although giving a fairly good estimate of mean CCR, do not offer reliable data on CCR in the upper and lower ranges and do not adequately estimate the age–related decline in CCR.