RESUMO
BACKGROUND: Many reports have described a correlation between the morning-urine protein /creatinine ratio (morning urine P/Cr) and the quantity of 24-h urine protein (Up), as well as regression formulas for Up with morning-urine P/Cr. However, there is no universal regression formula that can be used at all facilities. It is still controversial whether a qualitative calculation is required at outpatient clinics. To develop a practical and universal method, we used receiver operating characteristic (ROC) analysis to estimate Up from morning-urine P/Cr. METHODS: The subjects were 34 children (309 specimens) with kidney disease who had been admitted to Miyazaki Prefectural Hospital. We examined the correlations of P/Cr with Up and Up/body surface area (Up/BSA) using morning and daytime urine. We determined the cutoff values to estimate Up/BSA from morning-urine P/Cr with an ROC analysis. Next, we applied the values to specimens obtained from other facilities to show the universality of this approach. RESULTS: Up/BSA for samples in one hospital was significantly correlated with morning-urine P/C. When the morning-urine P/Cr ratio is >or=1.0 or >or=2.0, the Up/BSA ratio will exceed 0.5 or 1.0 (g/m(2) per day), respectively, and the efficiency was sufficiently high (efficiency for Up/BSA of >or=0.5: 88.0%, efficiency for Up/BSA of >or=1.0: 90.9%). When we analyzed samples from two other facilities with these cutoff values, both the sensitivity and specificity were greater than 80% for both facilities. CONCLUSIONS: The use of cutoff values of 1.0 and 2.0 for morning-urine P/Cr determined by ROC analysis could be a universal method for quantitatively estimating Up/BSA >or=0.5 and 1.0, respectively.
