Intake and Sources of Dietary Fiber, Inflammation, and Cardiovascular Disease in Older US Adults.

Importance: Higher intake of dietary fiber has been associated with lower inflammation, but whether there are differences in this association by source of dietary fiber (ie, cereal, vegetable, or fruit) has not been studied to date. Objectives: To evaluate the associations of total fiber intake and source (ie, cereal, vegetable, and fruit fiber intake) with inflammation and to evaluate whether inflammation mediates the inverse association between dietary fiber intake and cardiovascular disease (CVD). Design, Setting, and Participants: At the baseline visit (1989-1990) of 4125 adults aged 65 years or older in an ongoing US cohort study, dietary intake was assessed by a food frequency questionnaire among study participants without prevalent CVD (stroke and myocardial infarction) at enrollment. Inflammation was assessed from blood samples collected at baseline with immunoassays for markers of inflammation. Multivariable linear regression models tested the association of dietary fiber intake with inflammation. Also assessed was whether each inflammatory marker and its composite derived from principal component analysis mediated the association of baseline cereal fiber intake with development of CVD (stroke, myocardial infarction, and atherosclerotic cardiovascular death) through June 2015. Data from June 1, 1989, through June 30, 2015, were analyzed. Exposures: Total fiber intake and sources of fiber (cereal, vegetable, and fruit). Main Outcomes and Measures: Systemic markers of inflammation. Cardiovascular disease was the outcome in the mediation analysis. Results: Of 4125 individuals, 0.1% (n = 3) were Asian or Pacific Islander, 4.4% (n = 183) were Black, 0.3% (n = 12) were Native American, 95.0% (n = 3918) were White, and 0.2% (n = 9) were classified as other. Among these 4125 individuals (2473 women [60%]; mean [SD] age, 72.6 [5.5] years; 183 Black individuals [4.4%]; and 3942 individuals of other races and ethnicitites [95.6%] [ie, race and ethnicity other than Black, self-classified by participant]), an increase in total fiber intake of 5 g/d was associated with significantly lower concentrations of C-reactive protein (adjusted mean difference, -0.05 SD; 95% CI, -0.08 to -0.01 SD; P = .007) and interleukin 1 receptor antagonist (adjusted mean difference, -0.04 SD; 95% CI, -0.07 to -0.01 SD; P < .02) but with higher concentrations of soluble CD163 (adjusted mean difference, 0.05 SD; 95% CI, 0.02-0.09 SD; P = .005). Among fiber sources, only cereal fiber was consistently associated with lower inflammation. Similarly, cereal fiber intake was associated with lower CVD incidence (adjusted hazard ratio, 0.90; 95% CI, 0.81-1.00; 1941 incident cases). The proportion of the observed association of cereal fiber with CVD mediated by inflammatory markers ranged from 1.5% for interleukin 18 to 14.2% for C-reactive protein, and 16.1% for their primary principal component. Conclusions and Relevance: Results of this study suggest that cereal fiber intake was associated with lower levels of various inflammatory markers and lower risk of CVD and that inflammation mediated approximately one-sixth of the association between cereal fiber intake and CVD.