Urine erythrocyte (Ery) and leukocyte (Leu) dipstick tests are essential for detecting microhematuria and urinary tract infection. Currently, there is no suggestion for establishing the cut-off limits in an ordinal scale test. This study aimed to establish the cut-off limits for urine Ery and Leu dipstick tests via probit regression. From 1 January 2016 to 30 June 2016, laboratory data were collected from patients at one teaching hospital whose specimen had analytical results from both a urine dipstick test and an automated urine particle analyzer. Probit regression was used to estimate the probability of positive urine dipstick results as a function of log-transformed urine Ery and Leu concentrations. Based on the analysis of 22,122 specimens, the estimated concentration that yields 50% positive results (C50) of the Ery weak+, 1+, 2+, 3+, and 4 + dipstick results were 14.6, 40.4, 51.6, 136.3 and 219.0 × 106/L, respectively. The estimated C50 of the Leu 1+, 2+ and 3 + dipstick results were 22.7, 67.9 and 283.9 × 106/L, respectively. The estimated values were different from arbitrary concentrations provided for each dipstick category by the manufacturer. If a quantitative comparison method/procedure is available, the cut-off limits of an ordinal scale test can be established using probit regression. Each laboratory should investigate the transferability of the arbitrary concentrations provided by the manufacturer, and if necessary, determine its own cut-off limits of urine Ery and Leu dipstick tests.
Read full abstract