Abstract
CHD patients often present region-specific symptom clusters, such as "upper-body heat-related" (e.g., bitter taste) or "lower-body cold-related" (e.g., cold extremities), occurring independently or concurrently. Phenotype classification relies on subjective assessment and lacks quantitative indicators. This study aimed to establish a urine color-based quantitative model for objective CHD phenotype classification. From April 2023 to January 2024, a multicenter cross-sectional study involved 200 CHD patients and 240 healthy controls. Morning urine chromaticity was quantified using CIE Lab parameters (L, a, b values). The study included correlation analysis, two-way ANOVA, and hierarchical multinomial logistic regression. CHD patients had higher rates of upper-heat (44.50% vs. 25.42%, P = 0.033) and lower-cold (60.50% vs. 20.42%, P = 0.005) clusters than controls. Upper-heat clusters negatively correlated with L (r=-0.73) and positively with a (r = 0.79)/b (r = 0.74); lower-cold clusters showed opposite correlations (L: r = 0.81; a: r=-0.77; b: r=-0.73). Two-way ANOVA confirmed independent effects (η²=0.08-0.13, P < 0.01) with no interaction. Urine color parameters explained 86.3% of phenotypic variation, with model accuracy of 85.2%. This is the first study to validate urine color quantification via CIE Lab as an objective tool for CHD phenotypic classification, offering a novel auxiliary method for precise syndrome identification and targeted interventions.