AIM

To investigate the potential causal relationship between non-alcoholic fatty liver disease (NAFLD) and complications in type 1 diabetes (T1D) and type 2 diabetes (T2D).

MATERIALS AND METHODS

Two-sample Mendelian randomization (MR) analysis was conducted to appraise after controlling for the confounding factors. Genetic instrument variables for NAFLD surrogated by chronically elevated serum alanine transferase were derived from a recent genome-wide association study. Diabetes-related complications, including diabetic ketoacidosis, nephropathy and retinopathy, were included as outcomes. Four complementary MR methods were used to test reliability.

RESULTS

Genetically instrumented NAFLD showed a suggestive causal association with ketoacidosis in T1D (odds ratio [OR]: 1.574; 95% confidence interval [CI]: 1.076, 2.302; P = .019; false discovery rate [FDR] = 0.096) and a significant causal association with early-stage kidney disease in T1D (OR: 1.249; 95% CI: 1.089, 1.432; P = 1.457 × 10 , FDR = 0.015). Sensitivity analysis indicated low heterogeneity, low pleiotropy and high reliability of the causal estimates. However, the MR analyses failed to show a causal association between NAFLD and T1D retinopathy, T2D ketoacidosis, nephropathy and retinopathy.

CONCLUSIONS

This study supports a causal effect of genetically driven chronic serum alanine aminotransferase-associated NAFLD on early-stage kidney disease in T1D and a suggestive causal effect on ketoacidosis in T1D. However, MR studies did not provide enough evidence to suggest that NAFLD independently increases the risk of retinopathy in T1D and of ketoacidosis, nephropathy and retinopathy in T2D.