Variability in creatinine excretion in adult diabetic, overweight men and women: consequences on creatinine-based classification of renal disease.

BACKGROUND: It is crucial to estimate renal function in diabetic patients. However, formulas are inadequate in this population whereas creatinine clearance (C(cr)) on a 24-h urine collection may be valuable only if we can improve its reproducibility.
OBJECTIVE: To evaluate in diabetic patients whether standardised procedures of 24-h urine collection improve the day-to-day variability in creatinine urinary excretion and the subsequent precision of the measured C(cr).
METHODS: The C(cr) from two consecutive 24-h urine collections was measured in 201 consecutive diabetic inpatients. Procedures of 24-h urine collection were standardised, and implementation was supervised at a diabetes clinic.
RESULTS: Pearson's correlation coefficients of the two 24-h creatinine urinary excretion were significant (r(2)=0.64 in women and r(2)=0.65 in men, p<0.0001) but the daily variability in creatinine urinary excretion was high (14.9% in women and 17.4% in men). As a consequence, the agreement between the two consecutive measurements of C(cr) was poor. First, Bland-Altman plots showed large 95% limits of agreement (-34.3 to 34.6 mL/min/1.73 m(2) in women and -39.0 to 52.0 mL/min/1.73 m(2) in men). Secondly, there was a poor agreement for classifying patients according to the National Kidney Foundation classification >90, 60-89.9, 30-59.9, and <30 mL/min/1.73 m(2) (Kappa coefficients=0.61, 0.42, 0.65, and 0.74, respectively).
CONCLUSIONS: Despite standardised procedures of 24-h urine collection, day-to-day variability in creatinine urinary excretion in adult diabetic men and women remains important, and may lead to misclassification of renal disease.