Variability of albumin excretion in diabetes

The variability of urinary albumin excretion was determined in 3 consecutive overnight urine collections from 152 adult type I diabetic patients (study A) and in two 1-h urine collections from 57 adult type I and type II diabetic patients, obtained at a 10-wk interval (= median, range = 2-30 wk) (study B). In both studies the coefficient of variation (CV) of urinary albumin excretion recorded as urinary albumin concentration (mg/l) was highest (study A: CV 51.1%; study B: CV 61.8%), whereas the CV of the urinary albumin/urinary creatinine ratio (mg/mmol) was lowest (study A: CV 41.9%; study B: CV 50.8%). In study A the CV of urinary albumin excretion rate (microgram/min) (CV 43.5%) did not differ from that of urinary albumin/urinary creatinine ratio. In study B, however, the CV of urinary albumin excretion rate (CV 57.2%) was higher than that of the urinary albumin/urinary creatinine ratio and of fractional albumin clearance (CV 51.0%). The variability of urinary albumin excretion was not related to its magnitude and could not be attributed to variations in plasma glucose, glycosylated haemoglobin or renal tubular glucose reabsorption. This high variability stresses the need to obtain several urine samples before the amount of albumin excreted can be reliably estimated. It is recommended that one does not use urinary albumin concentration but urinary albumin/urinary creatinine ratio or fractional albumin clearance in the assessment of urinary albumin excretion in diabetes management.