ABSTRACT
BACKGROUND AND AIMS: Epidemiological evidence on the associations between female reproductive features and nonalcoholic fatty liver disease (NAFLD) is conflicting. To explore their causalities, we conducted a Mendelian randomization (MR) study. METHODS: Summary-level data were obtained, and univariable MR was performed to explore the causalities between female reproductive features and NAFLD. And we performed multivariable MR and MR mediation analysis to explore the mediation effects of educational attainment (EA) and body mass index (BMI) for these associations. Sensitivity analyses were performed to evaluate pleiotropy and heterogeneity. RESULTS: There were causal effects of age at menarche (AAMA) (odds ratio [OR]: 0.817, 95% confidence interval [CI]: 0.736-0.907, per year-increase), age at first birth (AFB) (OR: 0.851, 95%CI: 0.791-0.926, per year-increase) and age at first sexual intercourse (AFS) (OR: 0.676, 95%CI: 0.511-0.896, per standard deviation-increase) on NAFLD risk. Besides, the causal effects were also observed on NAFLD phenotypes including liver fat content (LFC) and alanine aminotransferase (ALT). Further mediation analysis showed that BMI mediated partial proportion of effects of AAMA and AFS on NAFLD/ALT, AFB on NAFLD/LFC/ALT, while EA mediated partial proportion of effects of AFB on NAFLD/LFC/ALT, and AFS on NAFLD/ALT. CONCLUSIONS: This study provided convincing evidence that early AAMA, AFB, and AFS were risk factors for NAFLD. Reproductive health education, obesity management, and education spread might be the beneficial strategies for NAFLD prevention.