Diagnostic Accuracy of Point-of-Care Tests for Detecting Albuminuria

Abstract

Background: Experts recommend screening for albuminuria in patients at risk for kidney disease. Purpose: To systematically review evidence about the diagnostic accuracy of point-of-care (POC) tests for detecting albuminuria in individuals for whom guidelines recommend such detection. Data Sources: Cochrane Library, EMBASE, Medion database, MEDLINE, and Science Citation Index from 1963 through 5 December 2013; hand searches of other relevant journals; and reference lists. Study Selection: Cross-sectional studies, published in any language, that compared the accuracy of machine-read POC tests of urinary albumin-creatinine ratio with that of laboratory measurement. Data Extraction: Two independent reviewers extracted study data and assessed study quality using the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies 2) tool. Data Synthesis: Sixteen studies (n = 3356 patients) that evaluated semiquantitative or quantitative POC tests and used random urine samples collected in primary or secondary ambulatory care settings met inclusion criteria. Pooling results from a bivariate randomeffects model gave sensitivity and specificity estimates of 76% (95% CI, 63% to 86%) and 93% (CI, 84% to 97%), respectively, for the semiquantitative test. Sensitivity and specificity estimates for the quantitative test were 96% (CI, 78% to 99%) and 98% (CI, 93% to 99%), respectively. The negative likelihood ratios for the semiquantitative and quantitative tests were 0.26 (CI, 0.16 to 0.40) and 0.04 (CI, 0.01 to 0.25), respectively. Limitation: Accuracy estimates were based on data from single-sample urine measurement, but guidelines require that diagnosis of albuminuria be based on at least 2 of 3 samples collected in a 6-month period. Conclusion: A negative semiquantitative POC test result does not rule out albuminuria, whereas quantitative POC testing meets required performance standards and can be used to rule out albuminuria.