Abstract
Hypertension (HTN) is a major global health burden influenced by multiple factors. Dietary flavonoids have shown health benefits, yet limited epidemiological studies focus on individual flavonoid intake and HTN risk. The baseline characteristics and disease data of the population in this study are derived from National Health and Nutrition Examination Survey, and the flavonoid intake data come from food and nutrient database for dietary studies. After feature selection, we employed multivariable logistic regression to calculate odds ratio and quantify the protective effect of flavonoids. We then conducted sensitivity analyses using a generalized additive model to assess the nonlinear relationship between flavonoid intake and HTN, and finally performed Shapley additive explanations analysis with 3 machine learning models to explore the ranking of variable importance. This study found that higher intakes of individual flavonoids, particularly quercetin, epicatechin, and naringenin, and of flavonoid subclasses, such as flavanones and flavonols, were associated with a lower risk of HTN. Among these, quercetin and naringenin emerged as the most important individual flavonoids in the machine learning models. Furthermore, the associations between flavanones and naringenin with HTN were linear and negative, whereas quercetin, epicatechin, and flavones exhibited nonlinear relationships with HTN. This study suggests that consuming foods rich in individual flavonoids, such as quercetin, epicatechin, and naringenin, may aid in the control and prevention of HTN, thereby reducing the overall cardiovascular disease burden in the population.