Creatinine clearance estimation from serum creatinine values: evaluation and comparison of five prediction formulae...

Predictability of creatinine clearance (CrCl) from serum creatinine would reduce the cost of renal care and obviate the need for 24-hours urine collection. Correlations have been established between serum creatinine (Scr) and 24-hour urinary creatinine clearance with derivation of various formulae. We have tested the applicability of these formulae in 34 Nigerian patients (22 males, 12 females) aged 18 to 58 years, (mean age 34.97 +/- 11.20 years) with established chronic renal failure (CRF) mean Scr level 742.26 +/- 388.15 mol/L. 32 age and sex matched healthy adults with serum creatinine values below 120 umol/l., served as controls. Serum creatinine and 24 hour CrCl levels were determined on two consecutive occasions. Creatinine clearance values were also derived from Scr using each of the established prediction formulae: Cockcrof and Gault [1] Gates [2] Hull et al [3]; Jelliffe [4]; and Mawer et al [5]. A relationship was sought between measured CrCl and the predicted values (derived) using the stated formulae. Regression equation were generated and correlation coefficient r, coefficient of determination r2, F-ration, prediction error, all defining the nature and strength of relationship were determined. We observed that good and statistically significant correlations exist between measured CrCl values and those predicated from the formulae (r ranging from 0.908 to 0.968 and r20.82 to 0.93 P = 0.000) and that a linear relationship exists in all cases. Cockcroft and Gault formula gave the highest coefficient of determination r2 = 0.94. It is concluded that the existing formulae are adequate for determining CrCl from Scr and should be frequently used in the long term follow-up of patients with Chronic Renal Failure (CRF) in our setting.