Biologic variation of urinary albumin: consequences for analysis, specimen collection, interpretation of results, and screening programs.

Studies on the analytic and biologic variability of albumin concentration, albumin/creatinine ratio, and albumin excretion rate in first morning, random, and 24-hour urine specimens from healthy subjects suggest that (1) first morning specimens are preferred, (2) results should be expressed as albumin concentration, (3) assay of creatinine confers little advantage, (4) an analytical precision of coefficient of variation (CV) less than 18% is satisfactory, and (5) semiquantitative or qualitative analyses are suitable for screening programs. The intraindividual variation of albumin concentration in first morning specimens from diabetics is such that no threshold value gives the desired 100% nosological sensitivity. However, a threshold value of 30 mg/L confers 100% specificity, and a single abnormal result therefore requires initiation of therapy. Patients with negative results should continue to be monitored regularly.