PLoS ONE (Jan 2017)
Comparison of urine dipstick and albumin:creatinine ratio for chronic kidney disease screening: A population-based study.
Abstract
Chronic kidney disease (CKD) is usually diagnosed using the estimated glomerular filtration rate (eGFR) or kidney damage markers. The urine dipstick test is a widely used screening tool for albuminuria, a CKD marker. Although the urine albumin:creatinine ratio (ACR) has advantages over the dipstick test in sensitivity and quantification of levels, the two methods have not been compared in the general population. A total of 20,759 adults with urinalysis data in the Korea National Health and Nutrition Examination Survey 2011-2014 were examined. CKD risk categories were created using a combination of eGFR and albuminuria. Albuminuria was defined using an ACR cutoff of 30 mg/g or 300 mg/g and a urine dipstick cutoff of trace or 1+. The EQ-5D index was used for the health outcome. Prevalence estimates of ACR ≥30 mg/g and >300 mg/g vs dipstick ≥trace and ≥1+ in adults aged ≥20 years were 7.2% and 0.9% vs 9.1% and 1.2%, respectively. For ACR ≥30 mg/g detection, the sensitivity, specificity, and positive/negative predictive values of dipstick ≥trace were 43.6%, 93.6%, 34.6%, and 95.5%, respectively. When risk categories created based on dipstick cutoffs were compared with those based on ACR cutoffs, 10.4% of the total population was reclassified to different risk categories, with only 3.9% reclassified to the same CKD category. Akaike information criterion values were lower, and non-fatal disease burdens of CKD were larger, in models predicting EQ-5D index using ACR-based categories compared to those using dipstick-based categories, even after adjusting for confounders. In conclusion, the urine dipstick test had poor sensitivity and high false-discovery rates for ACR ≥30 mg/g detection, and classified a large number of individuals into different CKD risk categories compared with ACR-based categories. Therefore, ACR assessments in CKD screening appear beneficial for a more accurate prediction of worse quality of life.