Abstract
Observational studies suggest an association between nonalcoholic fatty liver disease (NAFLD) and liver cancer, but its causal nature remains unclear. A 2-sample Mendelian randomization (MR) analysis was performed using NAFLD and liver cancer summary statistics from genome-wide association study databases. Instrumental variables satisfying the 3 core MR assumptions were selected. Causal effects were estimated using inverse-variance weighted, MR-Egger, weighted median, and other methods, followed by sensitivity and power analyses. All 4 MR analyses demonstrated a positive causal association between NAFLD and liver cancer risk [odds ratio > 1, inverse-variance weighted P < .001]. Sensitivity analysis indicated no significant level of multiplicity or heterogeneity in the instrumental variables, and individual single nucleotide polymorphisms had no significant impact on the results. However, statistical power was insufficient. This study provides the first MR evidence demonstrating a genetically predicted causal relationship between NAFLD and liver cancer that is consistent across subtypes. Sensitivity analyses confirmed the absence of horizontal pleiotropy or heterogeneity, strengthening the robustness of the findings. These results offer genetic support for early NAFLD intervention to reduce the risk of liver cancer. However, the limited statistical power highlights the need for larger-scale genome-wide association study to identify more and stronger genetic instruments for a more precise quantification of the causal effect of NAFLD on liver cancer risk.
Keywords: liver cancer, Mendelian randomization, nonalcoholic fatty liver disease
1. Introduction
NAFLD affects approximately 25% of adults globally, with prevalence reaching 32% in Western countries, establishing it as a leading cause of chronic liver disease.[1,2] Driven by the epidemics of obesity and diabetes, NAFLD prevalence continues to rise, with its burden projected to increase 2- to 3-fold in Western nations and several Asian regions by 2030.[3] As a major cause of global cancer-related mortality, liver cancer pathogenesis is complex. Traditional epidemiological studies have linked metabolic factors like obesity and type 2 diabetes to increased liver cancer risk.[4,5] However, these associations are often confounded by factors such as smoking, alcohol consumption, and diet, making definitive causal inference challenging.
Observational studies indicate that NAFLD patients have a significantly higher risk of progressing to liver cancer compared to the general population.[6] However, whether this association represents a causal effect is debatable. Shared risk factors like obesity and insulin resistance between NAFLD and liver cancer may introduce reverse causation or confounding bias in conventional studies, potentially overestimating their direct link.[7] Therefore, rigorous causal inference methods are urgently needed to elucidate the biological relationship between NAFLD and liver cancer, providing a theoretical basis for early intervention strategies.
Mendelian randomization (MR) employs genetic variants associated with an exposure as instrumental variables (IVs). This approach mimics the randomization principle of controlled trials, effectively circumventing the confounding and reverse causation biases that plague traditional epidemiological studies.[8,9] Since genetic variants are randomly allocated at conception and generally independent of postnatal environmental factors, MR provides more robust genetic evidence for exposure-disease causality. While MR has been successfully applied to study associations between metabolic diseases and cancer, a gap remains in exploring the NAFLD–liver cancer causal relationship.
This study employed a MR approach, treating NAFLD as the exposure and liver cancer as the outcome. We selected genetic variants strongly associated with NAFLD as IVs to evaluate their causal effect. The results demonstrated a significant positive causal association between NAFLD and liver cancer risk across all 4 MR analyses (odds ratio [OR] > 1, inverse-variance weighted (IVW) method P < .001). This association was consistent for both hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma, with highly concordant effect estimates across different genetic instruments. Sensitivity analyses further confirmed the absence of horizontal pleiotropy or heterogeneity among the IVs, and no single single nucleotide polymorphism (SNP) unduly influenced the overall effect estimates, supporting the robustness of the findings.These results provide genetic evidence supporting the role of NAFLD in the pathogenesis of liver cancer and offer novel insights for primary prevention strategies and the development of therapeutic targets for liver cancer.
2. Materials and methods
2.1. Mendelian randomization study design
This study is based on 3 core MR assumptions: Relevance assumption: The selected IVs are strongly associated with the exposure (NAFLD); Independence assumption: The selected IVs are independent of confounders affecting both the exposure and outcome; Exclusion restriction assumption: The IVs influence the outcome (liver cancer) solely through the exposure (NAFLD). A 2-sample MR framework was used to assess the NAFLD–liver cancer causal relationship.
SNPs significantly associated with NAFLD (exposure) and present in the outcome datasets were selected as IVs to satisfy the relevance assumption. Liver cancer was the outcome. The LDlink database was queried to exclude SNPs associated with the outcome itself, aiming to satisfy the independence assumption. Heterogeneity was assessed using Cochran Q test, while horizontal pleiotropy was evaluated via MR-Egger regression and leave-one-out analysis to evaluate the exclusion restriction assumption. Two-sample MR analysis was performed using R software (version 4.4.1; R Foundation for Statistical Computing [R Core Team], Vienna, Austria).
2.2. Data sources
NAFLD and liver cancer genome-wide association study (GWAS) summary statistics were obtained from the IEU Open GWAS database (https://gwas.mrcieu.ac.uk/). All data were derived from individuals of European ancestry. Exposure datasets: ebi-a-GCST90091033, ebi-a-GCST90054782. Outcome datasets: ieu-b-4953 (liver cancer), ieu-b-4915 (liver/bile duct cancer) (Table 1). As this study involved secondary analysis of publicly available data, additional ethical approval was not required.
Table 1.
Summary of the GWAS included in this 2-sample MR study.
| Dataset | SNPs | ncase | ncontrol | Year | Population |
|---|---|---|---|---|---|
| ebi-a-GCST90091033 | 6,784,388 | 8434 | 770,180 | 2021 | European |
| ebi-a-GCST90054782 | 9,097,254 | 4761 | 373,227 | 2021 | European |
| ieu-b-4953 | 6,304,034 | 168 | 372,016 | 2021 | European |
| ieu-b-4915 | 7,687,713 | 350 | 372,016 | 2021 | European |
MR = Mendelian randomization, NAFLD = nonalcoholic fatty liver disease, SNP = single nucleotide polymorphisms.
2.3. Instrumental variables selection
SNPs strongly associated with NAFLD (P < 5 × 10‐8) were extracted from the exposure GWAS datasets. The clump_data function in R (parameters: R2 = 0.001, kb = 10,000) was applied for linkage disequilibrium pruning to ensure independence of selected SNPs. Weak instrument bias was mitigated by calculating the F-statistic [F = (N ‐ 2) × R2/(1 ‐ R2)]; SNPs with F > 10 were retained. The LDlink database (https://ldlink.nih.gov) was consulted to exclude SNPs directly associated with the outcome, satisfying the independence assumption. Allele and effect data for exposure and outcome were harmonized to determine the final set of SNPs for MR analysis. All procedures were performed using R software (version 4.4.1).
2.4. Mendelian randomization analysis
The primary causal effect estimate was obtained using the IVW method due to its high precision and statistical power. Supplementary analyses employed MR-Egger regression, weighted median (WME), weighted mode, and simple mode methods. Statistical significance was set at P < .05. Analyses were conducted using the R package “TwoSampleMR.”
2.5. Sensitivity and power analysis
Horizontal pleiotropy: Assessed via the intercept term in MR-Egger regression (P < .05 indicates significant pleiotropy). Heterogeneity: Evaluated using Cochran Q statistic (P < .05 indicates significant heterogeneity). Influence of individual SNPs: Evaluated using the leave-one-out method, which sequentially removes each SNP and recalculates the pooled effect estimate to assess if any single SNP disproportionately drives the result. To examine the directionality of the causal relationship and avoid reverse causation, Steiger filtering was applied; a Steiger P > .05 for an IV was considered indicative of reverse causality. All the above analyses were conducted using the “TwoSampleMR” package in R.
Statistical power for the MR analyses was calculated using an online tool (https://sb452.shinyapps.io/power/). The effect size derived from the IVW method was used as the input for power calculation. A statistical power level ≥80% was considered adequate.
3. Results
3.1. Selection of instrumental variables
Based on the predefined significance threshold, 246 and 436 NAFLD-associated SNPs were initially identified from datasets ebi-a-GCST90091033 and ebi-a-GCST90054782, respectively. Linkage disequilibrium clumping excluded 242 and 429 SNPs exhibiting nonrandom association, leaving 4 and 7 independent SNPs, respectively. The F-statistic for each retained SNP exceeded 10, indicating strong instruments and mitigating weak instrument bias. LDlink database queries confirmed no significant association between the retained SNPs and the outcome, satisfying the independence assumption (Table 2). Harmonization of exposure and outcome data finalized the SNP sets for MR analysis.
Table 2.
Information on the final screening of NAFLD SNPs from GWAS data.
| ID | SNP | Beta | SE | Alleles | P | F |
|---|---|---|---|---|---|---|
| ebi-a-GCST90091033 | ||||||
| 1 | rs28601761 | ‐0.110033 | 0.0164425 | G/C | 2.20191e‐11 | 44.782 |
| 2 | rs3747207 | 0.288601 | 0.019822 | A/G | 5.07341e‐48 | 211.981 |
| 3 | rs429358 | ‐0.136615 | 0.0239288 | C/T | 1.13501e‐08 | 32.595 |
| 4 | rs73001065 | 0.281009 | 0.0326406 | C/G | 7.35699e‐18 | 74.117 |
| ebi-a-GCST90054782 | ||||||
| 1 | rs1260326 | ‐0.136025 | 0.020866 | C/T | 2.5363e‐11 | 42.497 |
| 2 | rs17321515 | ‐0.154093 | 0.0207828 | G/A | 1.8134e‐13 | 54.974 |
| 3 | rs182611493 | 0.453817 | 0.0746899 | G/A | 2.3286e‐11 | 36.918 |
| 4 | rs2642442 | 0.13769 | 0.0227583 | T/C | 7.6713e‐10 | 36.603 |
| 5 | rs3747207 | 0.369714 | 0.0229366 | A/G | 6.7406e‐60 | 259.818 |
| 6 | rs429358 | ‐0.199223 | 0.0304411 | C/T | 2.1692e‐11 | 42.831 |
| 7 | rs73001065 | 0.345021 | 0.0348771 | C/G | 1.0814e‐24 | 97.860 |
GWAS = genome-wide association study, NAFLD = nonalcoholic fatty liver disease, SNP = single nucleotide polymorphisms.
3.2. Causal effect of NAFLD on liver cancer
All 4 MR analyses consistently demonstrated a significant positive causal association between NAFLD and the risk of liver cancer (all OR > 1).
The primary causal effect estimates, obtained using the IVW method as the main analytical approach, reached extreme statistical significance across all 4 exposure–outcome pairings (all P < .001; specific P-values: 9.01e‐09, 4.90e‐04, 1.25e‐10, and 3.14e‐07). Results from the WME method were also highly significant (specific P-values: 1.53e‐06, 2.00e‐05, 8.09e‐08, and 1.60e‐04), thereby robustly corroborating the findings from the IVW method. The forest plot in Figure 1 clearly illustrates these consistent positive causal effect estimates (OR > 1) along with their respective 95% confidence intervals (CIs).
Figure 1.
MR analysis results showing causal association between NAFLD and liver cancer. Forest plots demonstrate consistent positive causal effects (OR > 1) across 4 exposure–outcome combinations. IVW estimates were significant (all P < .001). (A) Exp: GCST90091033|Out: ieu-b-4953 (HCC). (B) Exp: GCST90091033|Out: ieu-b-4915 (liver/bile duct). (C) Exp: GCST90054782|Out: ieu-b-4953 (HCC). (D) Exp: GCST90054782|Out: ieu-b-4915 (liver/bile duct). Exp = NAFLD exposure dataset, HCC = hepatocellular carcinoma, IVW = inverse-variance weighted, MR = Mendelian randomization, NAFLD = nonalcoholic fatty liver disease, OR = odds ratio, Out = liver cancer outcome dataset.
Consistency in the direction of causal effects across different methods was visually confirmed by scatter plots, which depict the association between the genetic instrument effects on NAFLD and the corresponding effects on liver cancer. The regression slopes for all analytical methods (IVW, MR-Egger, weighted median, weighted mode, and simple mode) were positive, indicating unanimous agreement among methods that the causal direction points towards an increased risk of liver cancer due to NAFLD (Fig. 2).
Figure 2.
Scatter plots of causal relationship between NAFLD and liver cancer. Regression slopes from all methods (IVW, MR-Egger, weighted median, weighted mode, and simple mode) were positive, indicating consistent causal direction. (A)–(D) as defined in Figure 1. IVW = inverse-variance weighted, MR = Mendelian randomization, NAFLD = nonalcoholic fatty liver disease.
The contribution of individual SNPs was assessed via forest plots displaying the causal estimate derived from each SNP individually. The results revealed that the effect direction for the vast majority of individual SNPs was consistent with the overall IVW estimate (OR > 1). No SNPs demonstrated an effect in the opposite direction or exhibited outlier values, indicating that the overall causal estimate was not driven by any single influential SNP (Fig. 3).
Figure 3.
Forest plot of the causal relationship between NAFLD and liver cancer.The causal effect of NAFLD on liver cancer is estimated using each SNP singly. The MR estimate using all SNPs derived from the IVW and MR-Egger method is shown for comparison. (A)–(D) as defined in Figure 1. IVW = inverse-variance weighted, MR = Mendelian randomization, NAFLD = nonalcoholic fatty liver disease, SNP = single nucleotide polymorphisms.
Collectively, the results from the IVW and WME methods, supported by multiple supplementary MR approaches, provide strong evidence supporting NAFLD as an independent causal risk factor for liver cancer. The individual SNP analysis further confirms the robustness of the overall findings.
3.3. Sensitivity and power analysis
MR-Egger regression intercepts were close to 0 and nonsignificant (P > .05), indicating no directional horizontal pleiotropy. Cochran Q test results (P > .05) suggested no significant heterogeneity among SNPs (Table 3). Leave-one-out analyses showed that causal estimates remained stable and significant upon sequential removal of each SNP, confirming that no single SNP unduly influenced the results (Fig. 4). All SNPs passed Steiger testing (P < .05), indicating no reverse causality (Table 4). However, power analysis revealed insufficient statistical power (<80%) across all analyses (Table 3).
Table 3.
Sensitivity and power analysis results.
| Exposure | Outcome | Egger intercept test | Cochran Q test | Power | ||
|---|---|---|---|---|---|---|
| Egger intercept | P-value | Q-IVW | Q-value | |||
| ebi-a-GCST90091033 | ieu-b-4953 | 1.43e‐05 | .909 | 1.554 | 0.460 | 2.5% |
| ebi-a-GCST90091033 | ieu-b-4915 | ‐9.47e‐05 | .654 | 5.902 | 0.116 | 2.5% |
| ebi-a-GCST90054782 | ieu-b-4953 | ‐9.80e‐06 | .881 | 1.613 | 0.806 | 2.5% |
| ebi-a-GCST90054782 | ieu-b-4915 | ‐6.49e‐05 | .518 | 5.445 | 0.364 | 2.5% |
IVW = inverse-variance weighted.
Figure 4.
Leave-one-out analysis of NAFLD and liver cancer. Each black point represents the IVW estimate after excluding 1 SNP; red points show estimates using all SNPs. Results remained stable regardless of excluded SNP. (A)–(D) as defined in Figure 1. IVW = inverse-variance weighted, NAFLD = nonalcoholic fatty liver disease, SNP = single nucleotide polymorphisms.
Table 4.
Steiger filtering analysis results.
| Exposure:ebi-a-GCST90091033/outcome: ieu-b-4953/ieu-b-4915 | |||
|---|---|---|---|
| ID | SNP | Steiger filtering (P) | |
| 1 | rs28601761 | True (<.001) | True (<.001) |
| 2 | rs3747207 | True (<.001) | True (<.001) |
| 3 | rs429358 | True (<.001) | True (<.001) |
| 4 | rs73001065 | / | True (<.001) |
| Exposure:ebi-a-GCST90054782/outcome: ieu-b-4953/ieu-b-4915 | |||
| 1 | rs1260326 | True (<.001) | True (<.001) |
| 2 | rs17321515 | True (<.001) | True (<.001) |
| 3 | rs2642442 | True (<.001) | True (<.001) |
| 4 | rs3747207 | True(<.001) | True(<.001) |
| 5 | rs429358 | True (<.001) | True (<.001) |
| 6 | rs73001065 | / | True (<.001) |
4. Discussion
This MR study investigated the causal relationship between NAFLD and liver cancer. The results indicate a potential causal effect of NAFLD on both HCC and liver/bile duct cancer. The consistency of the positive association across different genetic instruments and exposure datasets strengthens the findings, further supported by the robustness demonstrated in sensitivity analyses. These results align with previous observational studies. A large cohort study involving 130 healthcare institutions reported that NAFLD patients had a 7.62-fold higher HCC incidence compared to non-NAFLD individuals.[10] This risk escalates markedly with cirrhosis, where annual HCC incidence reaches 10.6 per 1000 person-years versus only 0.08 per 1000 person-years in non-cirrhotic NAFLD.[11] Another study indicated a significantly elevated risk of cholangiocarcinoma in NAFLD patients (OR = 1.88, 95% CI: 1.25–2.83).[12] We attempted reverse MR analysis, but it yielded effect estimates that were logically implausible with exceedingly wide CIs, indicating model invalidity and a lack of biological plausibility for this causal direction.
From a mechanistic perspective, the progression from NAFLD to liver cancer involves multiple synergistic pathways: Chronic inflammation and immune microenvironment dysregulation: Hepatic lipid accumulation triggers the release of pro-inflammatory cytokines such as TNF-α and IL-6, persistently activating Kupffer cells and hepatic stellate cells, thereby promoting carcinogenesis through IKK/c-Jun, STAT, and NF-κB pathways.[13] Direct oncogenic effects of disrupted lipid metabolism: Excessive free fatty acid deposition induces endoplasmic reticulum stress and mitochondrial dysfunction, disrupts cellular signaling mechanisms, modulates gene transcription, and activates various oncogenic pathways.[13] Pro-tumorigenic microenvironment of progressive fibrosis: Activated HSCs secrete TGF-β and extracellular matrix components, forming fibrous septa that create a microenvironment conducive to the proliferation and invasion of liver cancer cells.[14,15] Driver role of insulin resistance: Hyperinsulinemia activates the IGF-1/PI3K/AKT and MAPK pathways, inhibiting apoptosis and promoting abnormal hepatocyte proliferation.[16,17]
Notably, NAFLD disease subtypes significantly influence liver cancer risk: nonalcoholic steatohepatitis, the progressive form of NAFLD characterized by hepatocyte ballooning and lobular inflammation, carries a substantially higher risk of liver cancer compared to simple steatosis.[7] The fibrosis stage represents a critical turning point: patients with advanced fibrosis (F3–F4) have an 8.5-fold increased risk of liver cancer compared to those with mild fibrosis (F0–F2), and cirrhosis constitutes the strongest predictor of liver cancer development.[18,19]
Two SNPs, rs3747207 and rs429358, were included in all 4 MR analyses. rs3747207 is located within the PNPLA3 gene, encoding Patatin-like phospholipase domain-containing protein 3, a member of the patatin-like phospholipase family.[4] While most studies have focused on the strong association between the PNPLA3 rs738409 variant (encoding the I148M mutation) and liver cancer risk,[11,20] research on rs3747207 remains limited. A study in a Chinese Han population found no overall association between rs3747207 and HCC susceptibility; however, age-stratified analysis revealed a significant association in subjects older than 55 years, suggesting an age-dependent effect of rs3747207 on liver cancer susceptibility.[21] rs429358 is located in the apolipoprotein E (APOE) gene, which plays a critical role in lipid metabolism and cholesterol transport. Our finding of a negative association between rs429358 and NAFLD aligns with previous reports indicating that the C allele reduces NAFLD risk, thereby potentially diminishing an upstream driver of liver cancer development.[22] Furthermore, rs429358 (APOE ε4) exhibits pleiotropic effects independent of NAFLD: it promotes amyloid-β oligomerization and pathological accumulation, representing the strongest genetic risk factor for late-onset Alzheimer disease[23]; in cardiovascular research, this locus is associated with higher low-density lipoprotein cholesterol levels and increased cardiovascular disease risk, partially independent of lipid metabolism[24]; it modulates the intensity of inflammatory responses in immune regulation[25]; and recent studies have linked APOE variants to sleep patterns.[26] These pleiotropic pathways suggest that rs429358 may influence health outcomes through neurodegenerative, cardiovascular, immune, and other non-NAFLD mechanisms.
This study provides genetic evidence supporting NAFLD as a modifiable risk factor for distinct pathological types of liver cancer, implying that early NAFLD management could reduce liver cancer incidence. Limitations should be acknowledged. First, minor inconsistencies among the supplementary MR methods (WME, weighted mode, and simple mode) may arise from methodological characteristics or data features, warranting further investigation. Second, the exclusive reliance on European ancestry GWAS data necessitates validation of the results’ generalizability across diverse ethnic populations.
5. Conclusion
This study provides the first MR evidence demonstrating a genetically predicted causal relationship between NAFLD and liver cancer that is consistent across subtypes. Sensitivity analyses confirmed the absence of horizontal pleiotropy or heterogeneity, strengthening the robustness of the findings. These results offer genetic support for early NAFLD intervention to reduce the risk of liver cancer. However, the limited statistical power highlights the need for larger-scale GWAS to identify more and stronger genetic instruments for a more precise quantification of the causal effect of NAFLD on liver cancer risk.
Author contributions
Conceptualization: Feihua Chen, Xifeng Zhang.
Data curation: Feihua Chen, Xifeng Zhang.
Formal analysis: Xifeng Zhang.
Funding acquisition: Feihua Chen.
Investigation: Xifeng Zhang.
Methodology: Feihua Chen, Xifeng Zhang.
Project administration: Feihua Chen, Xifeng Zhang.
Resources: Feihua Chen.
Supervision: Feihua Chen.
Software: Xifeng Zhang.
Validation: Feihua Chen.
Visualization: Xifeng Zhang.
Writing – original draft: Feihua Chen, Xifeng Zhang.
Abbreviations:
- APOE
- apolipoprotein E
- CI
- confidence interval
- GWAS
- genome-wide association study
- HCC
- hepatocellular carcinoma
- IV
- instrumental variable
- IVW
- inverse-variance weighted
- MR
- Mendelian randomization
- NAFLD
- nonalcoholic fatty liver disease
- OR
- odds ratio
- SNP
- single nucleotide polymorphisms
- WME
- weighted median estimator
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are publicly available.
How to cite this article: Chen F, Zhang X. Association between NAFLD and liver cancer: A two-sample Mendelian randomization study. Medicine 2025;104:42(e45125).
References
- [1].Chandrasekaran P, Weiskirchen R. The pivotal role of the membrane-bound O-acyltransferase domain containing 7 in non-alcoholic fatty liver disease. Liver. 2023;4:1–14. [Google Scholar]
- [2].Marchisello S, Di Pino A, Scicali R, et al. Pathophysiological, molecular and therapeutic issues of nonalcoholic fatty liver disease: an overview. Int J Mol Sci. 2019;20:1948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Yuan S, Chen J, Li X, et al. Lifestyle and metabolic factors for nonalcoholic fatty liver disease: Mendelian randomization study. Eur J Epidemiol. 2022;37:723–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Chen M, Liu J, Xia X, Wang Y, Zheng H. Causal relationship between non-alcoholic fatty liver disease and sarcopenia: a bidirectional Mendelian randomization study. Front Med (Lausanne). 2024;11:1422499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Mcglynn KA, Petrick JL, Serag HBE. Epidemiology of hepatocellular carcinoma. Hepatology. 2021;73(Suppl 1):4–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Bjorkstrom K, Widman L, Hagstrom H. Risk of hepatic and extrahepatic cancer in NAFLD: a population-based cohort study. Liver Int. 2022;42:820–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Shah PA, Patil R, Harrison SA. NAFLD-related hepatocellular carcinoma: the growing challenge. Hepatology. 2023;77:323–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Sekula P, Del Greco M F, Pattaro C, Köttgen A. Mendelian randomization as an approach to assess causality using observational data. J Am Soc Nephrol. 2016;27:3253–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23:R89–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Kanwal F, Kramer JR, Mapakshi S, et al. Risk of hepatocellular cancer in patients with non-alcoholic fatty liver disease. Gastroenterology. 2018;155:1828–37.e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Motta BM, Masarone M, Torre P, Persico M. From non-alcoholic steatohepatitis (NASH) to hepatocellular carcinoma (HCC): epidemiology, incidence, predictions, risk factors, and prevention. Cancers (Basel). 2023;15:5458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Xiao J, Ng CH, Chan KE, et al. Hepatic, extra-hepatic outcomes and causes of mortality in NAFLD – an umbrella overview of systematic review of meta-analysis. J Clin Exp Hepatol. 2023;13:656–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Chen Y, Wang W, Morgan MP, Robson T, Annett S. Obesity, non-alcoholic fatty liver disease and hepatocellular carcinoma: current status and therapeutic targets. Front Endocrinol (Lausanne). 2023;14:1148934. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Llovet JM, Willoughby CE, Singal AG, et al. Nonalcoholic steatohepatitis-related hepatocellular carcinoma: pathogenesis and treatment. Nat Rev Gastroenterol Hepatol. 2023;20:487–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Song Y, Wei J, Li R, et al. Tyrosine kinase receptor B attenuates liver fibrosis by inhibiting TGF-beta/SMAD signaling. Hepatology. 2023;78:1433–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Dhamija E, Paul SB, Kedia S. Non-alcoholic fatty liver disease associated with hepatocellular carcinoma: an increasing concern. Indian J Med Res. 2019;149:9–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Rodriguez-Lara A, Rueda-Robles A, Sáez-Lara MJ, Plaza-Diaz J, Álvarez-Mercado AI. From non-alcoholic fatty liver disease to liver cancer: microbiota and inflammation as key players. Pathogens. 2023;12:940. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Phoolchund AGS, Khakoo SI. MASLD and the development of HCC: pathogenesis and therapeutic challenges. Cancers (Basel). 2024;16:259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Thomas JA, Kendall BJ, El-Serag HB, Thrift AP, Macdonald GA. Hepatocellular and extrahepatic cancer risk in people with non-alcoholic fatty liver disease. Lancet Gastroenterol Hepatol. 2024;9:159–69. [DOI] [PubMed] [Google Scholar]
- [20].Huang DQ, El-Serag HB, Loomba R. Global epidemiology of NAFLD-related HCC: trends, predictions, risk factors and prevention. Nat Rev Gastroenterol Hepatol. 2021;18:223–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Gong D, Li S, Yu Z, Wang K, Qiao X, Wu C. Contribution of PNPLA3 gene polymorphisms to hepatocellular carcinoma susceptibility in the Chinese Han population. BMC Med Genomics. 2022;15:248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Innes H, Nischalke HD, Guha IN, et al. The rs429358 locus in apolipoprotein E is associated with hepatocellular carcinoma in patients with cirrhosis. Hepatol Commun. 2022;6:1213–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Wu LY, Real R, Martinez-Carrasco A, et al. Investigation of the genetic aetiology of Lewy body diseases with and without dementia. Brain Commun. 2024;6:fcae190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Civeira-Marin M, Cenarro A, Marco-Benedí V, et al. APOE genotypes modulate inflammation independently of their effect on lipid metabolism. Int J Mol Sci. 2022;23:12947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Feringa FM, Elmagd AA, Kruijff ID, et al. AD risk variants in APOE differentially affect lipid metabolism in iPSC derived neurons, astrocytes and microglia. Alzheimers Demen. 2023;19:1. [Google Scholar]
- [26].Moyses-Oliveira M, Zamariolli M, Tempaku PF, et al. Pleiotropic effects of APOE variants on a sleep-based adult epidemiological cohort. Sleep Med. 2025;131:106490. [DOI] [PubMed] [Google Scholar]