Estimating urine albumin to creatinine ratio from protein to creatinine ratio using same day measurement: validation of equations.

Abstract

Severity of chronic kidney disease is defined by glomerular filtration rate (GFR) and albuminuria (ACR) by the KDIGO and are related to cardiovascular outcomes and end-stage-kidney-failure. However, proteinuria (PCR) is more often available than ACR in records. Recently, equations were developed to estimate ACR from PCR. We investigated their performances in our population. In the academic medical hospital of Liège, we retrospectively analysed same day measurement of ACR and PCR and staged them according to the KDIGO A1-A2-A3 categories. Analyser Roche Cobas (R)gathered 2,633 urinalysis (May 2018-May 2019) and analyser Abbott Alinity (A) 2,386 urinalysis (May 2019-March 2020). We compared the KDIGO staging of mACR and eACR obtained from Weaver's and Sumida's equations. Median age was 63 [52;71]/64 [53;72] years old, 43/42% were female; 78/74% had diabetes; proportion of mACR-A1 was 65.6%/64.2%, A2 was 25.5%/25.5% and A3 was 8.8%/10.3% (Method R/A, respectively). Both equations gave similar distribution of KDIGO staging of eACR. Overall agreements were higher than 88% regardless of the analyser or of the equation. Performances in between equations were equivalent according to the multi-level AUC (multinomial logistic regression model). Good concordance was observed between mACR and eACR regardless of the equation or of the analyser. No patient with an A3-measured ACR was estimated within the KDIGO A1 category. Though ACR should be measured when clinically needed, it may be reasonably estimated from the PCR through these equations, for epidemiologic retrospective studies or research purposes.