Causal relationship of hepatic fat with liver damage and insulin resistance in nonalcoholic fatty liver

Abstract

Background and aims: Nonalcoholic fatty liver disease is epidemiologically associated with hepatic and metabolic disorders. The aim of this study was to examine whether hepatic fat accumulation has a causal role in determining liver damage and insulin resistance.

Methods: We performed a Mendelian randomization analysis using risk alleles in PNPLA3, TM6SF2, GCKR and MBOAT7, and a polygenic risk score for hepatic fat, as instruments. We evaluated complementary cohorts of at-risk individuals and individuals from the general population: 1515 from the liver biopsy cohort (LBC), 3329 from the Swedish Obese Subjects Study (SOS) and 4570 from the population-based Dallas Heart Study (DHS).

Results: Hepatic fat was epidemiologically associated with liver damage, insulin resistance, dyslipidemia and hypertension. The impact of genetic variants on liver damage was proportional to their effect on hepatic fat accumulation. Genetically determined hepatic fat was associated with aminotransferases, and with inflammation, ballooning and fibrosis in the LBC. Furthermore, in the LBC, the causal association between hepatic fat and fibrosis was independent of disease activity, suggesting that a causal effect of long-term liver fat accumulation on liver disease is independent of inflammation. Genetically determined hepatic steatosis was associated with insulin resistance in the LBC and SOS. However, this association was dependent on liver damage severity. Genetically determined hepatic steatosis was associated with liver fibrosis/cirrhosis and with a small increase in risk of type 2 diabetes in publicly available databases.

Conclusion: These data suggest that long-term hepatic fat accumulation plays a causal role in the development of chronic liver disease.

Keywords: fibrosis; genetics; insulin resistance; mendelian randomization; nonalcoholic fatty liver disease; type 2 diabetes.

Figure 1

Study cohorts and outcomes. NASH, nonalcoholic steatohepatitis.

Figure 2

Comparison of the impact of risk variants: PNPLA3 I148M (rs738409), TM6SF2 E167K (rs58542926), GCKR P446L (rs1260326) and MBOAT7 rs641738 on hepatic fat vs. liver damage. Panel a: histological steatosis vs. ballooning in the LBC (R ² = 0.70). b: histological steatosis vs. necroinflammation in the LBC (R ² = 0.76). c: histological steatosis vs. fibrosis in the LBC (R ² = 0.97). d: hepatic fat content vs. serum ALT levels in the DHS (R ² = 1). Beta coefficients and 95% confidence intervals are shown for each variant. Correction added on 16 January 2018 after first online publication: In Figure 2B and D, the color labels for PNPLA3 and TM6SF2 have been interchanged for clarity]

Figure 3

Comparison of the epidemiological association of observed hepatic fat with NAFLD features (open circles) with the causal association of hepatic fat with NAFLD‐related features (filled circles) in the LBC (n = 1515, panel a), the SOS (n = 3329, panel b) and the DHS (n = 2736, panel c). T2D, type 2 diabetes; HOMA‐IR, homoeostasis model assessment‐insulin resistance index; HDL, high‐density lipoprotein; ALT, alanine aminotransferases; AST, aspartate aminotransferases. Results were adjusted for age, sex, BMI, modality of recruitment in the LBC and ethnicity in the DHS, use of statins and (for the metabolic parameters) the severity of liver fibrosis. Estimates are beta coefficients, and error bars are 95% confidence intervals.

Figure 4

Schematic representation of main study findings. Hepatic fat accumulation is causally associated with increased liver enzymes, hepatocellular damage, necroinflammation and fibrosis. TAG, triglycerides, AST, aspartate aminotransferases, ALT, alanine aminotransferases, HSCs, hepatic stellate cells.