Evaluating associations of oxylipin metabolites and fatty acids with incident and prevalent type 2 diabetes in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL)
Carolina G. Downie, Heather M. Highland, Annie Green Howard, Mona Alotaibi, Tali Elfassy, Christina Cordero, Jennifer E. Below, Mercedes Carnethon, Susan Cheng, Martha Daviglus, Olga Garcia-Bedoya, Robert Kaplan, Tasneem Khambaty, Yilun Li, Adam G. Lilly, Dan-Yu Lin, Tao Long, Frank J. Penedo, Lauren E. Petty, Bharat Thyagarajan, Hsiao-Chuan Tien, Mohit Jain, Barrett M. Welch, Kari E. North, Christy L. AveryBackground
Inflammation is a common pathological feature underlying type 2 diabetes (T2D) risk and progression. Determining the roles of fatty acids and bioactive lipid oxylipin inflammatory mediators could inform T2D preventative efforts and therapeutic targets.
Methods
Leveraging the Hispanic Community Health Study/Study of Latinos (HCHS/SOL), we evaluated cross-sectional and prospective associations of 81 oxylipins and polyunsaturated, monounsaturated, and saturated fatty acids with T2D. We assessed whether fatty acid levels differed at visit 2 by visit 1-prevalent T2D status and prioritized likely causal metabolites using Bayesian model averaging Mendelian randomization (MR-BMA).
Results
Fifty-four fatty acids and oxylipins were associated with prevalent T2D at visit 1 (15,138 participants; 2,962 T2D cases), four of which (adrenic, dihomo-gamma-linolenic, tricosanoic, and lignoceric acids) were also associated with incident T2D by visit 2 (8,548 participants; 1,376 T2D events), though effects were attenuated for adrenic and dihomo-gamma-linolenic acids upon further adjustment for impaired glucose tolerance. Sixteen out of 28 fatty acids were significantly different at visit 2 between those with and without prevalent T2D at visit 1 (9,115 participants; 2,151 T2D cases), holding visit 1-measured fatty acid levels constant. MR-BMA prioritized one diHETrE oxylipin as a likely causal risk factor for T2D.
Conclusions
Fatty acids and oxylipins showed complex patterns of association with T2D, including more associations likely reflecting overt disease processes in prevalent T2D, than incident risk-enhancing associations.