A Formula for the Estimation of 24-Hour Urinary Creatinine Excretion: A Derivation and Validation Study

Twenty-four-hour urinary creatinine (Cr) excretion (24h-uCr) is the basis of Cr clearance and urinary protein-Cr ratio, and it is related to frailty, worsening kidney function, and mortality in patients with chronic kidney disease. Although subjects with lower estimated glomerular filtration rate (eGFR) tend to have lower 24h-uCr, previous formulae for the estimation of 24h-uCr did not include Cr as a predictor.

This retrospective study included patients admitted to the Department of Nephrology at our hospital (derivation cohort and validation cohort: patients admitted between April 2016 and March 2020). The prediction formula of 24h-uCr was calculated using a multivariate linear regression model with the bootstrap method. Age, height, weight, sex, Cr, and cystatin C were used as predictors.

The derivation and validation cohorts included 187 and 63 patients, respectively. The characteristics of the derivation and validation cohorts were as follows: age 73 (61-79.5) years and 70 (58.5-79) years; males, 61.5% and 60.3%; eGFR_Cr 27.0 (13.7-48.6) mL/min/1.73 m² and 26.3 (14.0-51.5) mL/min/1.73 m²; and 24-hour urinary protein excretion 0.79 (0.17-2.12) g/day and 1.08 (0.26-2.55) g/day, respectively. Seven prediction formulae were derived. In all models, the Pearson's correlation coefficient was relatively high and statistically significant. However, previous models tended to overestimate the 24h-uCr. Furthermore, the predicted 24h-uCr calculated by the models that do not include Cr as a predictor fluctuates depending on the eGFR_Cr.

The best formula for predicting 24h-uCr (mg/day) in a wide range of eGFR populations is a Cr-containing formula: [-9.04 × age (years) + 8.03 × weight (kg) + 0.66 × height (cm) + 188.59 (if male) - 32.11 × Cr (mg/dL) + 779.14].