We aimed to quantify and generate comprehensive evidence on the associations of different fatty acids (FAs) with the risk of type 2 diabetes mellitus (T2DM). PubMed, MEDLINE, and Embase were searched for prospective cohort studies that examined the association between FA biomarkers and the risk of T2DM published before 18 May 2020. Random-effects meta-analyses of the effects of FA concentration on T2DM were performed. Thirty-three studies involving 95,810 adults (19,904 cases) were included. Divergent associations of different types of FAs with type 2 diabetes were observed. The pooled relative risk (RRs) of T2DM comparing the top versus the bottom tertile of saturated FAs (C14:0, C16:0, C18:0, total saturated FAs), monounsaturated FA (C16:1 n-7), polyunsaturated FA (C20:3 n-6, C22:4 n-6), and Δ-6-desaturase activities ranged from 1.19 to 1.80. Interestingly, unlike previous studies, we found a negative correlation between odd-chain saturated FAs (C15:0, C17:0), trans-FAs (trans-C16:1 n-7), total n-6, Δ-5-desaturase activities and risk of T2DM. The pooled RRs of T2DM comparing the top versus the bottom tertile of these FAs ranged from 0.62 to 0.78. No associations with T2DM were observed for the other FAs. Considerable heterogeneity was observed in our study, and no definitive conclusions can be made until further investigation has been carried out.
Keywords: Biomarkers; fatty acids; meta-analysis; systematic review; type 2 diabetes mellitus.