Diet patterns and risk of sepsis in community-dwelling adults: a cohort study

BACKGROUND: Sepsis is the syndrome of body-wide inflammation triggered by infection and is a major public health problem. Diet plays a vital role in immune health but its association with sepsis in humans is unclear. METHODS: We examined 21,404 participants with available dietary data from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study, a national cohort of 30,239 black and white adults ≥45 years of age living in the US. The primary exposures of interest were five empirically derived diet patterns identified via factor analysis within REGARDS participants: “Convenience” (Chinese and Mexican foods, pasta, pizza, other mixed dishes), “Plant-based” (fruits, vegetables), “Southern” (added fats, fried foods, organ meats, sugar-sweetened beverages), “Sweets/Fats” (sugary foods) and “Alcohol/Salads” (alcohol, green-leafy vegetables, salad dressing). The main outcome of interest was investigator-adjudicated first hospitalized sepsis events. RESULTS: A total of 970 first sepsis events were observed over ~6 years of follow-up. In unadjusted analyses, greater adherence to Sweets/Fats and Southern patterns was associated with higher cumulative incidence of sepsis, whereas greater adherence to the Plant-based pattern was associated with lower incidence. After adjustment for sociodemographic, lifestyle and clinical factors, greater adherence to the Southern pattern remained associated with higher risk of sepsis (hazard ratio [HR] comparing the fourth to first quartile, HR 1.39, 95 % CI 1.11,1.73). Race modified the association of the Southern diet pattern with sepsis (P(interaction) = 0.01), with the Southern pattern being associated with modestly higher adjusted risk of sepsis in black as compared to white participants (HR comparing fourth vs. first quartile HR 1.42, 95 % CI 0.75,2.67 vs. 1.21, 95 % CI 0.93,1.57, respectively). CONCLUSION: A Southern pattern of eating was associated with higher risk of sepsis, particularly among black participants. Determining reasons for these findings may help to devise strategies to reduce sepsis risk. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1186/s12879-015-0981-1) contains supplementary material, which is available to authorized users.