In this systematic review and meta-analysis of 109 studies and 118,302 individuals, we found an overall MAF(G) of 0.45 (adjusted to 0.39 due to publication bias) in patients with MASLD. The MAF(G) varied geographically, with the highest rates in Latin America and the lowest in Europe. Notably, no African study was included. MASLD patients carrying the PNPLA3 variant had reduced adiposity, altered fat metabolism, and worse liver damage/histology than those with the PNPLA3 CC genotype. In addition, only the PNPLA3 GG genotype was associated with worse clinical outcomes, including mortality and liver-related events. These findings have important implications for understanding the influence of genetic factors on how and in which MASLD patients under PNPLA3 lens. By quantifying the genetic contribution to phenotypic differences in MASLD, we can move toward individualized management of the MASLD patient.
To our knowledge, this study is the first to assess the global prevalence of the PNPLA3 variant in MASLD using a meta-analytic approach. Previous meta-analyses have consistently shown an association between the PNPLA3 G allele and the onset/severity of MASLD.22,23 However, these meta-analyses mainly evaluated PNPLA3 variant-related associations without adopting formal adequate inclusion criteria to minimize the risk of selection bias. Compared with these meta-analyses, we used stringent inclusion criteria for MASLD patients and analyzed more studies. Thus, we examined various aspects of the PNPLA3 variant influence, including genetic epidemiology, clinical characteristics, histological features, and long-term outcomes. We also formally assessed in depth the presence and impact of publication bias. Additionally, we tested the effect of steatosis-related factors as potential moderators of the expression of the G-risk allele at the PNPLA3 gene.
Genetic and sociodemographic factors influence MASLD severity.24 Understanding their complex interplay is crucial for identifying high-risk individuals. Our study revealed a higher MAF(G) of 0.39 in MASLD patients (after adjustment for publication bias) compared to the general population (0.26).25 This enrichment was highest in Latin America and lowest in Europe, reflecting known geographic differences in MASLD genetics.26 This pattern underscores the susceptibility of certain populations, especially Hispanics, to more severe disease. Beyond genetics, it is also important to address health disparities due to limited healthcare access and information.27 Educational interventions aimed at raising awareness of genetic risk, particularly in Hispanic populations, should be considered.28 Similar attention should be given to East Asian patients, especially in Japan, due to their high MAF(G) in the context of the rising burden of MASLD in Asian.29 Conversely, our screening process identified an absence of African studies. Our publication bias analyses suggested that studies with a lower MAF(G) may remain unpublished. This may be due to a lack of research in Africa, where this variant is less common.26 Furthermore, it highlights the importance of addressing the gap through further research in African regions. Like Hispanics, Black populations experience health disparities27, so it is critical to explore the interplay between the genetic epidemiology of MASLD and the environment in African populations, considering the diversity within Africa.
From a pathophysiological perspective, the PNPLA3 variant causes a loss of function of the PNPLA3 protein in lipid droplets, reducing triglyceride and retinyl ester hydrolysis in hepatocytes 30, ultimately leading to liver steatosis.31 Luukkonen et al. recently linked this steatosis to hepatic mitochondrial dysfunction, decreased de novo lipogenesis and increased ketogenesis.32 These mechanisms drive MASLD progression through liver injury and fibrosis.31 Our study aligns with this, demonstrating the clinical impact of the PNPLA3 variant on the reduction of TG levels and the increase in liver injury enzymes. In addition, we observed that patients with PNPLA3 GG had lower FBG and were less likely to have diabetes mellitus than those with the PNPLA3 CC genotype, showing the interplay between the PNPLA3 variant and glucose metabolism.33,34 Interestingly, our meta-regressions showed that diabetes mellitus and FBG levels influenced AST levels, potentially modulating the response of the PNPLA3 gene and amplifying the effect of the G-risk allele. The biological plausibility of this interaction involves, at least in part, pathways related to insulin resistance, inflammation, and dyslipidemia. These findings highlight the potential benefits of anti-diabetics not only for managing MASLD complications, but also for mitigating the high genetic risk profile associated with the PNPLA3 variant. Consistent with previous studies, our results showed a significant additive effect of the G-risk allele on histological features.35 This was evident in increased histological scores and higher odds/proportions of MASH and fibrosis across different PNPLA3 genotypes. Therefore, the PNPLA3 gene is a promising target to consider in MASLD treatment, whether at the level of RNA, proteins or metabolic pathways.36 Whilst no current drugs modulate it, promising clinical trials are underway, and personalized medicine approaches targeting PNPLA3 offer significant prospects for MASLD prevention and treatment.36
Emerging evidence suggest PNPLA3 genotype influence MASLD treatment response. A systematic review found that carriers of PNPLA3 G allele showed varied response to therapies, with lower response to omega-3 fatty acids and dapagliflozin, but higher response to lifestyle interventions.37 Similarly, a retrospective study observed increased ALT reduction with semaglutide in PNPLA3 variant carriers.38 Our findings on the clinical and histological impact of the PNPLA3 variant in patients with MASLD support this rationale, raising an important question: how well do MASLD/MASH therapeutic studies reflect the genetic risk profile of participants? Incorporating PNPLA3 genotyping in clinical trials may ensure balanced representation of individuals at high genetic risk in both study arms, followed by sub-analyses to assess treatment response and adverse effects on genetic profiles.37,39 A personalized approach would identify MASLD patients most likely to benefit from specific treatments. Our data can inform the design of such trials by providing insights into the genetic epidemiology of MASLD patients.
Recent studies have linked the PNPLA3 G allele to increased risk of LREs,40,41 particularly in PNPLA3 GG genotype.17,42 Our study confirms this association and further demonstrates higher risks of cirrhosis, liver-related mortality, and overall mortality in these patients. This suggests the potential of using PNPLA3 genotyping for prognostic information and risk stratification. On the other hand, no link was found between PNPLA3 variant and cardiovascular outcomes. Interestingly, combining non-invasive tests or clinical information (e.g., presence of diabetes) with the PNPLA3 genotyping might improve risk stratification for liver outcomes.40,43 However, current guidelines do not recommend routine genotyping in MASLD management.44,45 While incorporating PNPLA3 genotyping into clinical practice seems promising, challenges including ethical considerations, cost-effectiveness, unclear improvement over existing methods (e.g., histological staging and non-invasive scores), and patient communication strategies need careful evaluation.10 Further research is needed to address these challenges and provide a framework for potential clinical implementation.
PNPLA3-environmental interactions contribute significantly to shaping the MASLD phenotype. The well-known PNPLA3-adiposity interaction9 was evident in our meta-regressions, with higher WC and BMI amplifying the impact of PNPLA3 variant on liver injury enzymes. This suggests potential benefits of genetic screening for patients with high adiposity and targeted weight loss interventions for MASLD patients at high genetic risk, as previously observed.46 However, optimal treatment strategies for MASLD patients at low genetic risk require further investigation. Additionally, age emerged as a moderator for PNPLA3 variant effect on AST, implying a potential age-related influence on the activity of the PNPLA3 gene. While interactions like PNPLA3-female sex and diet were not identified, they might act as confounders and warrant further investigation.47,48 Unraveling these complex PNPLA3-environment interactions is crucial to identify distinct MASLD phenotypes and promote personalized medicine approaches.48
This study has several limitations, some of which are intrinsic to the included studies. First, the absence of African studies potentially overestimates and limits generalizability of our overall findings. We addressed this through publication bias adjustments, but further research in Africa is crucial. Second, limited ethnic reporting hindered subgroup analyses by ethnicity, though we considered ethnic distributions in interpreting results (e.g., the predominance of Hispanics in Latin America). Third, high heterogeneity (i.e., I² ≥ 75%) was observed in most analyses. Subsequent subgroup and sensitivity analyses were conducted, but results should be interpreted cautiously in the light of CIs. Fourth, we included studies using non-standard MASLD diagnoses, although biopsy (i.e., the gold standard) remained the most frequent method in the eligible studies (42.2%). Fifth, we acknowledge the high exclusion rate due to unsatisfactory outcomes (n = 66) or overlapping population (n = 50). The main reasons are the lack of complete PNPLA3 genotype distribution and the large number of collaborative studies without cohort-stratified data, respectively. Most of our efforts to obtain data were unsuccessful. Nevertheless, our large sample size provides a valuable representation. Finally, interactions between PNPLA3 and other loci (e.g., TM6SF2 and MBOAT7) were beyond the scope of this study.
Despite these limitations, this study has important strengths. First, it provides the most comprehensive analysis to date of the global frequency, clinical/histological presentation, and long-term outcomes of the PNPLA3 variant in MASLD. Second, we aggregated a large sample from different regions of the world (the Americas, Asia, and Europe), providing valuable worldwide data to understand the variation in the burden of MASLD complications; this approach allows the implementation of interventions to target populations at high risk of worsening MASLD. Third, our review identified important gaps, including the need for better representation of African regions, integrating genetics into clinical trials, unraveling PNPLA3-environment interactions, and addressing challenges in routine genotyping. Fourth, the overall quality of the eligible studies was acceptable, suggesting an overall moderate risk of bias, according to the NOS/HWE and Q-Genie tools. Lastly, we followed the guidelines for genetic association meta-analysis by reporting the genetic model and conducting sensitivity analyses for studies in HWE.
In conclusion, our systematic review and meta-analysis demonstrate a substantial global frequency of the PNPLA3 variant in MASLD patients, with significant geographic variations. MASLD patients with PNPLA3 variant exhibit lower adiposity, altered fat metabolism and more severe liver damage than those with the PNPLA3 CC genotype. Importantly, only PNPLA3 GG genotype carriers had increased risk of liver outcomes. Our findings suggest an amplifying effect of adiposity, diabetes mellitus, glucose, and age on PNPLA3 expression. These results highlight the need for a personalized medicine approaches applied to clinical management of MASLD. Integrating PNPLA3 genotyping into MASLD treatment trials is welcome. Further research is needed to delve the genetic epidemiology of underrepresented regions, explore challenges of routine genotyping implementation, and investigate the influence of gene-environment interactions on treatment and prognostic.