Abstract
BACKGROUND: Diabetes mellitus (DM) is a prominent health concern worldwide, leading to the high incidence of disability and mortality and bringing in heavy healthcare and social burden. Plant-based diets are reported associated with a reduction of DM risk. Plant-based diets are rich in flavonoids, which possess properties such as scavenging free radicals and exerting both anti-inflammatory and antioxidant effects. PURPOSE: However, whether dietary flavonoids are associated with the prevalence of DM remains controversial. The potential reasons for contradictory epidemiological outcomes on the association between dietary flavonoids and DM prevalence have not been determined. METHODS: To address these limitations, we employed data from 22,481 participants in the National Health and Nutrition Examination Survey to explore the association between the intake of flavonoids and DM prevalence by weighted Logistic regression and weighted restricted cubic splines. RESULTS: We found that the prevalence of DM was inversely associated with the intake of total flavonoids in the second quartile [Odds Ratio (OR) 0.78 95% confidence interval (CI) (0.63, 0.97), p = 0.028], in the third quartile [0.76 (0.60, 0.97), p = 0.031], and in the fourth quartile [0.80 (0.65, 0.97), p = 0.027]. However, the p for trend was not significant [0.94 (0.88, 1.01), p = 0.096]. Moreover, the association between DM prevalence and the intake of total flavonoids was significantly influenced by race (p for interaction = 0.006). In Mexican Americans, there was a significant positive association between DM prevalence and total flavonoid intake within the third quartile [1.04 (1.02, 1.07), p = 0.003]. Total flavan-3-ol and subtotal catechin intake exhibited a non-linear U-shaped association with DM prevalence (p for non-linearity < 0.0001 and p for non-linearity < 0.0001, respectively). Compared to the first quartile of corresponding intakes, consumption within the third quartile of subtotal catechins [0.70 (0.55, 0.89), p = 0.005] and total flavan-3-ols [0.65 (0.50, 0.84), p = 0.002] was associated with a lower prevalence of DM. CONCLUSION: Taken together, our study may provide preliminary research evidence for personalized improvement of dietary habits to reduce the prevalence of diabetes.