A Dipstick Protein and Specific Gravity Algorithm Accurately Predicts Pathological Proteinuria

Background: The proper strategy to screen for early chronic kidney disease is debatable, but protein-creatinine ratio from a random urine sample (UPC) commonly is used. The purpose is to determine whether dipstick data effectively identify patients with increased UPC ratios. Because urine concentration affects proteinuria interpretation, we hypothesized that dipstick protein (DSP) and specific gravity (SG) are sufficient for screening. Methods: A hospital laboratory database was searched for urine samples simultaneously assayed for UPC ratio, DSP, and SG (n = 2,098). A random 70% of the cohort was used to generate a development model, which was validated in the remaining 30%. Samples were stratified according to DSP and SG values. A DSP versus SG matrix was created, and each sample was allocated to a discrete DSP-SG category. Proportions of samples with overt (UPC ratio ≥ 500 mg/g) and nephrotic-range proteinuria (UPC ratio ≥ 3,000 mg/g) were calculated for all 40 cells. Results: Optimum correlations between DSP-SG cells and UPC ratios were determined for the development model, yielding 97.0% negative predictive value (NPV) for a UPC ratio of 500 mg/g or less and 97.5% positive predictive value (PPV) for a UPC ratio of 500 mg/g or greater. NPV for a UPC ratio of 3,000 mg/g or less was 99.7%. Application of the model to the validation sample showed a 96.5% NPV for a UPC ratio of 500 mg/g or less, 99.4% PPV for a UPC ratio of 500 mg/g or greater, and 99.0% NPV for a UPC ratio of 3,000 mg/g or less. Conclusion: DSP and SG values effectively identify patients requiring proteinuria quantification by means of UPC ratio. A Web-based tool was developed that allows DSP and SG value entry and provides a recommendation regarding the need for proteinuria quantification.