Abstract
OBJECTIVE: To explore the effects of temperature and relative humidity on the tongue color of diabetic patients with different severity levels and their interactions. METHODS: A total of 1,711 participants were included, divided into five groups based on disease status (no diabetes, simple diabetes, diabetes with hypertension, diabetes with coronary heart disease, diabetes with both hypertension and coronary heart disease). RGB values of the tongue body and coating at five sites (tip, middle, root, left, and right) were collected (30 indicators in total). The month of visit (six categories) was used as a composite temperature-humidity proxy variable, with age, gender, BMI, smoking, and alcohol consumption as covariates. Multiple linear regression was conducted, including the interaction term of month × group, to perform Type III tests and FDR corrections for the 30 RGB indicators and principal component scores. RESULTS: Significant interaction effects were found between the month of visit and disease group on tongue color (PC1: F(20,1670)=12.87, P<0.0001; PC2: F(20,1670)=18.64, P<0.0001). In the diabetes with hypertension and coronary heart disease group, tongue redness and brightness increased by 2.74-3.19 standard deviations during the hot and humid months of July and August compared to colder months (P<0.0001), while the no diabetes group remained stable throughout the year. Forty-seven RGB interaction terms had FDR <0.05, with the ten most significant located at the tongue tip; the R-value of the tongue body was significantly associated with temperature increase, showing a clear dose-response relationship with the severity of complications (trend P<10(-)¹²). CONCLUSION: High temperature and humidity can significantly amplify tongue redness in diabetic patients, potentially involving a microcirculation expansion-endothelial dysfunction-oxidative stress cascade pathway, while tongue color remains highly stable in healthy individuals. This interaction effect suggests that RGB tongue indicators may serve as an objective dynamic biomarker reflecting cardiovascular metabolic risk under heat stress.