Error in Estimation of 24-Hour Urine Protein Using Protein to Creatinine Ratio in Chronic Kidney Disease Patients

Abstract

   Urine protein to creatinine ratio (UPCR) is commonly used to estimate 24-hour urine protein (24hrUP). However, in our practice, we observe the discrepancy between both indices. This study aims to find an error in estimation of 24hrUP using UPCR. A cross-sectional study was conducted in Thammasat University Hospital during December 2016 and November 2017. Chronic kidney disease (CKD) patients were enrolled. All patients were measured proteinuria by 24-hour urine collection and UPCR using a standard method. Inadequate samples were excluded. We compared both indices using Pearson correlation. Bland and Altman plot was used to determine an agreement with pre-defined acceptable different range of ± 0.5. Furthermore, UPCR was stratified into subgroups to find the most reliable range to estimate proteinuria using linear regression models. Sixty-six of total 147 CKD patients were included for analysis. The main reason for exclusion was inadequate 24-hour urine collection (44% of complete urine examination). Mean age (SD) was 60.4 (15.0) years, 63.6% were male, and one-third had diabetes kidney disease. Mean estimated GFR (SD) was 58.7 (30.3) ml/min/1.73m². We found a high correlation between 24-hour urine protein and UPCR (r = 0.86; p < 0.001). Linear regression showed that the range of UPCR with the highest agreement with 24hrUP was between 0-2, with equation: 24hrUP = 0.99 * UPCR - 0.01. Systematic uptrend of error was observed when UPCR was more than 2. UPCR is highly correlated with 24hrUP. We suggest the most reliable range of UPCR used to estimate 24hrUP is ≤ 2. UPCR of more than 2 overestimates 24hrUP.