Validity Of Scoring Urine Color At Different Light Intensities To Assess Urine Concentration

Abstract

PURPOSE: Light conditions may influence the accuracy and cut-off value of urine color (Uc) charts. The aim was to compare the accuracy of Uc scoring at different light conditions. METHODS: 178 previously collected samples with a median of 1.018 (IQR 1.012-1.023) urine specific gravity (USG) were stored at -20 °C. 30 mL samples were thawed and scored by 3 technicians for Uc in 16 different ways: four light conditions (fluorescent (224 lux), halogen (402 lux), LED panel light (1666 lux), and LED flashlight (1848 lux)), using two different color charts (8-color and 7-color Uc chart), and two different techniques (a sample over chart approach and a Uc scoring box). As freezing urine samples may affect Uc, the results' generalizability was analyzed using a subsample of frozen samples (n = 78) comparing to its originally measured fresh Uc based on LED panel scoring box condition. RESULTS: Aggregated Uc results were significantly different between light conditions (P < 0.01). Lower light intensity scored 1.5-2 shades darker than bright conditions, with 3.8 (2-5.5) for fluorescent light, 3.5 (2-5.8) for halogen light vs. 2.4 (1-5.2) for LED light panel, and 2.0 (1-3.5) the LED flashlight. The best diagnostic capability for identifying urine with a USG ≥1.020 using receiver operating characteristics (ROC) was found for the brightest light conditions (LED flashlight) based on an average area under the curve (AUC) of 0.86. Likewise, the optimal Uc cut-off value, obtained from ROC analysis, decreased from ≤4 (fluorescent and halogen) to ≤3 (LED conditions). The median Uc for both charts comparing fresh vs. frozen samples was 2, withthe 75th percentile for Uc +2 shades darker for frozen samples (P ≤ 0.004). Still, fresh vs. frozen samples showed a moderate (r = 0.64) to fair (r = 0.71) Spearman correlation for 8- and 7-color Uc charts, P < 0.001. The Bland-Altman analysis showed that misreporting bias (r ≤ -0.45, P < 0.001) was predominantly based on a limited number of darker scored Uc samples. CONCLUSIONS: Urine color scoring is affected by light conditions, resulting in higher accuracy and a one-shade lower Uc cut-off value at the brightest light conditions. While freezing affects Uc, these results are likely generalizable to samples measured in fresh conditions suggesting that Uc cut-off value should be adjusted to light conditions.