INTRODUCTION AND OBJECTIVES: Non-alcoholic fatty liver disease (NAFLD) is the primary contributor to persistent chronic liver disease which derives cardiovascular disease, malignancies, and related mortality. There is an association between red blood cell (RBC) indices and the incidence of NAFLD, but the causal relationship has not been determined. We aimed to investigate the association through prospective and Mendelian randomization (MR) analyses. MATERIALS AND METHODS: The prospective study involved 237,016 participants from the UK Biobank. We employed Cox proportional hazard models and restricted cubic spline models to assess the association between RBC index and NAFLD, and used two-sample MR analysis to identify any causality. RESULTS: Over a mean follow-up of 8.64 years, 2,894 participants from UK Biobank developed NAFLD. The prospective study showed significant associations between high levels of hemoglobin (HGB) (hazard ratio [HR], 1.41
95 % confidence intervals [CI] 1.24-1.60
P <
0.002), RBC count (HR, 1.20
95 % CI, 1.07-1.36
P = 0.003) and an increased risk of NAFLD. MR analysis indicated a causal relationship between high HGB levels and NAFLD risk (Odds ratio [OR], 1.55
95 % CI, 1.11-2.18
P = 0.010). However, there was no observed causal relationship between RBC count and NAFLD. CONCLUSIONS: This prospective and MR analysis demonstrated a positive causal relationship between HGB levels and NAFLD. HGB can predict the risk of NAFLD, which can potentially be used as a large-scale non-invasive tool to dynamically monitor the occurrence and development of NAFLD.