Abstract
The associations between anthropometric measures and nonalcoholic fatty liver disease (NAFLD) risk have been investigated in observational studies, but results were inconsistent. This study conducted the first large-scale Mendelian randomization (MR) analyses to explore causal associations of single anthropometric measures and body shape derived from principal component analysis with the risk of NAFLD. This study utilized two-sample MR with genome-wide association study data from Europeans to investigate the causal relationship between anthropometric measures, body shape, and NAFLD. Body shape is defined by principal component analysis taken from 6 anthropometric measures (body mass index [BMI], weight, height, waist circumference [WC], hip circumference [HC], and waist-to-hip ratio [WHR]). Instrumental variables were single-nucleotide polymorphisms with P < 5 × 10-8, and data harmonization was performed. To assess result robustness, sensitivity analyses (MR-Egger, weighted median, leave-one-out analysis), and outlier detection (Mendelian Randomization Pleiotropy Residual Sum and Outlier) were conducted, alongside heterogeneity evaluation (Cochran Q test). Scatter and funnel plots were utilized to assess the exposure-outcome relationship and causal estimate consistency. The results of inverse-variance weighted analyses demonstrated that BMI (OR = 1.547, 95 CI% = 1.215-1.972, P < .001), weight (OR = 1.461, 95 CI% = 1.247-1.711, P < .001), WC (OR = 1.365, 95 CI% = 1.012-1.842, P = .042), HC (OR = 1.453, 95 CI% = 1.122-1.882, P = .005), WHR (OR = 2.765, 95 CI% = 1.931-3.957, P < .001), and WHR adjusted for BMI (OR = 2.226, 95 CI% = 1.639-3.024, P < .001) were positively related to NAFLD risk. In terms of body shape, the results of inverse-variance weighted (OR = 0.934, 95 CI% = 0.900-0.970, P < .001) and MR-Egger (OR = 0.890, 95 CI% = 0.810-0.978, P = .016) showed that PC3 (a tall, centrally obese body shape) was negatively associated with NAFLD. This study indicated that BMI, weight, WC, and HC were positively associated with NAFLD risk. WHR adjusted for BMI may be a more reliable predictor of NAFLD risk than WHR. A body shape related to tall individuals with centrally obese was inversely associated with NAFLD risk.