Abstract
Flavonoids have been shown to be beneficial in a variety of inflammatory and metabolic diseases because of their anti-inflammatory and antioxidant properties. However, previous epidemiological studies have only demonstrated a negative correlation between flavonoid intake on inflammatory markers, and the optimal intake of dietary flavonoids and subclasses in terms of dietary anti-inflammatory efficacy remains undetermined. This study was based on 3 cycles (2007-2010, 2017-2018) of the National Health and Nutrition Examination Survey and the corresponding expanded flavonoid database. Weighted multiple linear regression was used to assess linear relationships between flavonoid intake and Dietary inflammation index (DII). Smoothed curve fit and a generalized additive model were used to investigate the nonlinear relationships and threshold effects, the 2-tailed linear regression model was used to find potential inflection points. A total of 12,724 adults were included in the study. After adjusting for potential confounders, flavonoid intake was significantly associated with DII, with the strongest negative association effect for flavonols (-0.40 [-0.45, -0.35]). In subgroup analyses stratified by sex, race, age, body mass index, education levels, and diabetes, flavonol intake maintained a significant negative linear correlation with DII. In addition, we found significant nonlinear relationships (L-shaped relationships) and threshold effects between total flavonoids, flavan-3-ols, and flavanols and DII, with inflection points of 437.65 mg/days, 157.79 mg/days, and 46.36 mg/days, respectively. Our results suggest a threshold for the dietary anti-inflammatory capacity of flavonoid intake in U.S. adults.