Diagnostic accuracy of urine dipstick testing in screening for microalbuminuria in type 2 diabetes: a cohort study in primary care

• Published in: Family practice Vol. 30; no. 2; pp. 142 - 152
• Main Authors: Nagrebetsky, Alexander, Jin, Jing, Stevens, Richard, James, Timothy, Adler, Amanda, Park, Paul, Craven, Anthea, Shine, Brian, Farmer, Andrew
• Format: Journal Article
• Language: English
• Published: England 01.04.2013
• Subjects: Accuracy; Aged; Aged, 80 and over; Albuminuria - diagnosis; Albuminuria - epidemiology; Albuminuria - etiology; Albuminuria - urine; Clinical guidelines; Cohort analysis; Cohort Studies; Diabetes Mellitus, Type 2 - complications; Diabetes Mellitus, Type 2 - urine; England; Female; Humans; Laboratories; Male; Middle Aged; Point-of-Care Systems - economics; Prevalence; Primary Health Care; Prospective Studies; Reagent Strips - economics; Sensitivity and Specificity; Type 2 diabetes mellitus; Urine tests; England
• ISSN: 0263-2136; 1460-2229
• DOI: 10.1093/fampra/cms057

Clinical guidelines recommend annual screening for microalbuminuria in diabetes. Detection of microalbuminuria is important because it is associated with increased morbidity and mortality. Dipstick tests for microalbuminuria may be convenient, but their accuracy is uncertain. To assess the utility of urine dipstick testing for microalbuminuria in type 2 diabetes. In a 6-week cohort study in four general practices in Oxfordshire, UK, first-pass urine samples were obtained at two weekly intervals from patients with type 2 diabetes and tested in the practice using Micral-Test and Microalbustix urine dipsticks. Parallel samples were sent for laboratory albumin-creatinine ratio (ACR) assay. Results of single dipstick tests and sequences of dipstick and laboratory tests were compared with a clinical testing strategy based on current guidelines to assess the accuracy and estimate costs of testing. The prevalence of microalbuminuria was 12.5% (n = 88). Mean (standard deviation) age was 68 (10) years, 56 (57%) were men. Median (interquartile range) diabetes duration was 6.2 (2.0-10.0) years. The sensitivity and specificity, respectively, of a single Micral-Test were 91.7% and 44.0% and of a Microalbustix test 33.3% and 92.0%. Testing strategies involving dipstick and laboratory ACR measurements or dipstick tests had similar accuracy. The costs of using dipstick tests were overall lower than laboratory ACR-based testing. Dipstick testing in this study did not reliably identify diabetes patients with microalbuminuria. Although dipstick testing would decrease testing costs, it could either fail to diagnose most patients with microalbuminuria or increase the numbers of patients retested depending on the dipstick used.