In 86 patients with non-insulin-dependent diabetes mellitus who were negative for proteinuria by a test paper method, the albumin excretion rate was compared to the albumin/creatinine ratio, and a baseline albumin/creatinine ratio value was determined at which therapy to arrest the progression of nephropathy should be initiated. The albumin excretion rate, albumin/creatinine ratio, and urinary creatinine excretion were determined from urine samples obtained at outpatient visits. The reproducibility of time-restricted urine sampling was investigated using the creatinine excretion rate. The mean coefficient of variation was found to be 42%, and inaccurate urine sampling appeared to cause variation in the albumin excretion rate. A significant difference was evident between males and females in creatinine excretion (0.823 +/- 0.152 mg/min for males and 0.577 +/- 0.182 mg/min for females, p < 0.001). The baseline albumin/creatinine ratio value for initiating therapy was established to be 36 mg/g creatinine for males and 51 mg/g creatinine for females. Use of the albumin/creatinine ratio appears to provide a simple and valuable index for determining when to initiate therapy for diabetic nephropathy.