(P19-027-23) Evaluating the Role of Urinary Sodium in Non-Alcoholic Fatty Liver Disease: A Mendelian Randomization Study

Objectives: Observational studies suggested dietary sodium intake, proxied by urinary sodium, associated with higher non-alcoholic fatty liver disease (NAFLD) risk although these studies could be confounded by unhealthy lifestyle and socioeconomic position. Mendelian randomization (MR) can clarify whether these associations are causal. We assessed the association of urinary sodium with NAFLD and its markers (alanine transaminase (ALT) & aspartate transaminase (AST)) using 2 sample MR design.

Methods: We identified 43 single nucleotide polymorphism (SNPs) strongly (p < 5x10^-8) and independently (r² < 0.001) associated with urinary sodium from a large genome wide association study (GWAS) of urinary sodium (n=446,237) and applied to the summary statistics of NAFLD GWAS (cases: 8,434; controls: 770,180). We derived the causal effect of urinary sodium in NAFLD and its markers using inverse variance weighted method. To assess robustness of findings, we used additional sensitivity analyses which relied on other assumptions (e.g. weighted median and MR-Egger). MR-Egger intercept and F statistics were used to assess potential horizontal pleiotropy and weak instrument bias respectively.

Results: All instruments had F statistics >10, indicating weak instrument bias was unlikely. Urinary sodium was associated with higher NAFLD risk (Odds ratios (OR) 2.77 per log unit, 95% confidence interval (CI) 1.22 to 6.26, p=0.01) and ALT (β = 0.24, 95% CI 0.04 to 0.43, p=0.02). Results of sensitivity analyses were directionally consistent. However, urinary sodium was not associated with AST (β = -0.03, 95% CI -0.24 to 0.19, p=0.81). MR-Egger intercept did not indicate strong evidence for overall horizontal pleiotropy (p > 0.42).

Conclusions: Our study provides genetic evidence that increase in dietary salt intake, proxied by urinary sodium, is associated with higher risk of NAFLD. Reducing salt intake may be a possible approach to reduce NAFLD. Given this study focused only amongst European populations, whether these associations are present in other populations, such as East Asians, require further investigation.

Funding Sources: This study was partly funded by the Health and Medical Research Fund, Health Bureau, Hong Kong SAR, China (CFS-HKU1). The funder has no role in the design, analyses, interpretation of results or writing of the paper.