Use of protein: creatinine ratio in a random spot urine sample for predicting significant proteinuria in diabetes mellitus.

Present study was undertaken during a period of 6 months (September 2008-February 2009) to see an correlation of 24 hours urine protein estimation with random spot protein-creatinine (P:C) ratio among a diabetic patients. The study comprised of 144 patients aged 30-70 years, recruited from Kantipur hospital, Kathmandu. The 24-hr urine sample was collected, followed by spot random urine sample. Both samples were analyzed for protein and creatinine excretion. An informed consent was taken from all participants. Sixteen inadequately collected urine samples as defined by (predicted creatinine--measured creatinine)/predicted creatinine > 0.2 were excluded from analysis. The Spearman's rank correlation between the spot urine P:C ratio and 24-hr total protein were performed by the Statistical Package for Social Service. At the P:C ratio cutoff of 0.15 and reference method (24-hr urine protein) cutoff of 150 mg/day, the correlation coefficient was found to be 0.892 (p < 0.001). The area under ROC curve at different cutoffs was 0.88 at 95.0% CI. The sensitivity and specificity of the P:C ratio to detect significant proteinuria at the cutoff of 0.15 are 96.6% and 74.4%. So the P:C ratio can predict significant proteinuria in diabetic subjects, avoiding the inconvenient 24-hr urine collection but the cutoff should be carefully selected for different patients group under different laboratory procedures and settings.