Correlation between plasma and urine glucose in diabetes.

To determine whether semiquantitative glucose measurements of spot urine specimens accurately reflect prevailing plasma glucose levels, we compared reported levels from 400 second-voided urines to simultaneous plasma determinations from 246 adult diabetics. Quantitative urine levels and plasma glucose levels correlated. However, when semiquantitative urinary determinations were compared to plasma glucose stratified into 0 to 149, 150 to 199, and greater than 200 mg/dL, 75% of the urine samples associated with plasma levels from 150 to 199 mg/dL were negative by Diastix, and 16.5% of samples negative by Diastix were in the 200+ mg/dL plasma range. Only 9% of samples from 0 to 149 mg/dL showed any positive Diastix readings. Because of the low sensitivity of semiquantitative methods, we fell that, except for detection of marked hyperglycemia, spot urine glucose determinations are inadequate as the sole means of clinical assessment for management of diabetic patients. Home glucose monitoring may be a better alternative for follow-up of these patients.