Introduction: Diabetes mellitus (DM) is one of the common problems in the global public health and it’s incidence is increasing rapidly worldwide each year. It is estimated that the prevalence of DM will dramatically increase from 425 million people in 2017 to 629 million in 2045. Studies have demonstrated that Non Alcoholic Fatty Liver Disease (NAFLD) is associated with metabolic syndrome, T2DM, atherosclerosis and cardio-vascular disease. NAFLD is a strong risk factor for T2DM and insulin resistance. NAFLD is the hepatic manifestation of metabolic syndrome. Fatty Liver Index (FLI) is an assessment index to diagnose Fatty Liver Disease (FLD). Furthermore, FLI may be used as a surrogate marker of FLD that may predict the risk of diabetes mellitus and coronary heart disease independent from other metabolic syndrome traits. We performed a dose-response meta-analysis to investigate the relationship between FLI and diabetes incidence in prospective cohort studies.

Methods: We conducted a systematic search of articles up to November 2018 in PubMed, SCOPUS, Cochrane library, and Embase. Hazard Ratios (HRs) with corresponding 95% confidence intervals (CIs) of studies were evaluated using meta-analysis with DerSimonian and Laird random-effects models to find combined HRs. Dose-response effect of this relationship was also assessed.

Results: Twenty-one studies providing 62,445 participants were included in the meta-analysis. Pooled results showed a significant association between FLI and risk of diabetes incidence [HR: 3.21, 95% CI: 2.53-4.8; P for heterogeneity: 0.001]. Subgroup analysis based on a gender, continent, and the quality of study could not identify the source of heterogeneity. The pooled HR from the random-effects dose-response model indicated a significant association between FLI level and risk of diabetes incidence (Coef = 0.024184, p= 0.001).

Conclusion: Our dose-response meta-analysis revealed that a direct relationship between FLI and HR of diabetes incidence.