Association between metabolic dysfunction-associated steatotic liver disease and risk of early-onset advanced colorectal neoplasia: a systematic review and meta-analysis

Introduction

According to the Global Burden of Disease study, the global number of incident cases of early-onset colorectal cancer (EOCRC) increased from approximately 105,300 in 1990 to 210,100 in 2021. Over the same period, the age-standardized incidence rate rose from 5.43 to 6.13 per 100,000 population, reflecting an average annual growth rate of approximately 0.39%. In 2021, East Asia recorded the highest burden, with approximately 81,700 new cases, and China accounted for the largest share, approximately 78,700 cases (1). In high-income countries, including the United States, Australia, South Korea, and Singapore, as well as in selected Asian regions, incidence rates among adults aged 20–49 years have accelerated since the mid-1990s, with annual increases ranging from 1% to 3%. This trend has elevated the proportion of EOCRC within the overall colorectal cancer (CRC) caseload (2). As the principal precancerous lesion of CRC, adenoma detection has mirrored this shift toward younger-onset disease. Among average-risk individuals aged 45–49 years, the adenoma detection rate (ADR) ranges from 27% to 34%, while the advanced ADR is approximately 5–9%, nearing rates observed in older age groups (3). Traditional studies indicate that unresected advanced adenomas carry a 5–10% or higher risk of progression to malignancy within 10 years. In early-onset cases, disease progression is typically more rapid and biologically aggressive, resulting in approximately 75% of young patients presenting with advanced-stage EOCRC (stages III/IV) at the time of diagnosis (4). This pattern is partly attributable to diagnostic delays, or to the under-recognition of the malignant potential of high-risk adenomas.

In response to this emerging epidemic, both the United States and Australia have lowered the recommended starting age for CRC screening to 45 years. Similarly, several Asian countries have enhanced their early detection strategies through colonoscopy and fecal immunochemical testing, emphasizing the urgent need for lifestyle modification (5). To ensure terminological consistency throughout this review, we adopt the unified concept of “early-onset advanced colorectal neoplasia” (eoACRN), defined as advanced colorectal neoplasia diagnosed in individuals under 50 years of age. This classification primarily includes high-risk colorectal adenomas (CRA) and CRC, thereby reinforcing a public health framework focused on earlier identification and prevention.

Non-alcoholic fatty liver disease (NAFLD) is a chronic liver disease characterized by a fat content exceeding 5% in hepatocytes, with the exclusion of significant alcohol consumption and other secondary liver injury factors such as viral hepatitis or drug-induced liver damage. The disease spectrum of NAFLD can progress from simple hepatic steatosis to non-alcoholic steatohepatitis (NASH), liver fibrosis, cirrhosis, and ultimately hepatocellular carcinoma (6). In 2021, the estimated global prevalence of NAFLD was approximately 1.27 billion, with incidence and disability-adjusted life years showing an upward trend, primarily affecting young and middle-aged populations in regions with low to middle social population indices (7). In 2023, a multi-society Delphi consensus officially renamed NAFLD to metabolic dysfunction-associated steatotic liver disease (MASLD), replacing NAFLD and the transitional term metabolic associated fatty liver disease (MAFLD), to more accurately reflect its core pathophysiological mechanisms driven by metabolic syndrome (7,8). Given that many early studies adopted the definitions of NAFLD or MAFLD, this review uniformly uses the term “MASLD” to encompass clinically relevant studies utilizing the definitions of MASLD, MAFLD, or historical NAFLD.

MASLD and eoACRN share numerous epidemiological risk factors, including obesity, diabetes mellitus, high-fat diets, and unhealthy lifestyle behaviors (9). Recent high-quality cohort studies have demonstrated that MASLD is not merely a concurrent manifestation of metabolic derangements, but may also function as an independent risk factor for eoACRN. In 2024, Zhao et al. conducted a meta-analysis of 13 cohort studies, reporting that MASLD was significantly associated with an increased risk of CRC [pooled hazard ratio (HR) 1.25; 95% confidence interval (CI): 1.15–1.36] and colorectal adenoma (HR 1.38; 95% CI 1.17–1.64). Subgroup analyses confirmed the robustness of this association across geographic regions, diagnostic modalities, and study quality strata, underscoring the clinical utility of non-invasive markers (10). In 2025, a nationwide population-based cohort study by Kim et al. in Korea further established that MASLD/NAFLD, as defined by the fatty liver index (FLI), was strongly linked to the risk of EOCRC, particularly in tumors located in the left colon and rectum (11).

In light of the persistently rising global incidence of EOCRC and its increasingly aggressive clinical behavior, improving risk stratification for eoACRN and optimizing early screening strategies have become clinical priorities. we conducted a systematic review and meta-analytic methods to synthesize the latest evidence from large cohort studies, quantifying the association between MASLD and eoACRN risk, with inclusion of high-risk adenomas to better capture early colorectal tumorigenesis. As early-onset advanced adenomas and EOCRC represent distinct stages of colorectal carcinogenesis, eoACRN is considered a composite early neoplastic outcome. We present this article in accordance with the PRISMA reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0401/rc) (12).

Methods

Methodology of searching

We conducted a comprehensive search across multiple electronic databases, including PubMed, Embase, Web of Science, Cochrane Library, and the China National Knowledge Infrastructure (CNKI). The search strategy integrated Medical Subject Headings (MeSH) terms and free-text terms related to MASLD, MAFLD, and NAFLD, alongside relevant historical definitions. It also included terms associated with colorectal tumors, colorectal adenomas, advanced adenomas, and eoCRC. There were no language or geographical restrictions during the search process. To ensure the comprehensiveness of the literature inclusion, the reference lists of all included studies and relevant reviews were manually screened.

Due to the fact that many studies were conducted prior to the introduction of the MASLD nomenclature, we included studies defined by NAFLD, MAFLD, or MASLD, as long as they assessed hepatic steatosis associated with metabolic dysfunction, which qualifies them for inclusion. Due to the unavailability of individual participant-level data, we were unable to uniformly reclassify all participants according to the latest MASLD criteria. “ eoACRN “ refers to advanced adenomas and/or CRC in the early-onset population, excluding late-stage or metastatic cancers. We carefully reviewed studies that employed terms such as high-risk adenomas, advanced adenomas, or advanced colorectal tumors to ensure that the outcome definitions were consistent with the eoACRN. The detailed search strategy can be found in Appendix 1.

Study selection

Inclusion criteria studies were included if they met all of the following criteria:

• These studies are observational (cross-sectional, case-control, or cohort studies) that investigated the association between MASLD and the risk of developing eoACRN.
• They reported odds ratios (OR) and their 95% CIs, or provided sufficient raw data to facilitate the calculation of OR and 95% CIs;
• The diagnostic criteria for high-risk CRA and advanced colorectal tumors were clearly defined, and the outcome determination repository was described in detail;
• It was clarified that MASLD is diagnosed based on specific criteria, and the outcome determination repository was clearly described;
• There were no restrictions regarding the participants’ race, gender, ethnicity, or comorbidity status.
• They involved early-onset populations or provided data that allowed for the extraction of outcomes for early-onset colorectal tumors based on the age thresholds used in the original studies.

Studies that meet any of the following criteria are excluded:

• They are laboratory investigations, letters, abstracts, reviews, conference reports, meta-analyses, commentary articles, or case reports;
• The study population is limited to highly selected groups unrelated to the review question, including: Patients with chronic liver disease or cirrhosis primarily caused by competing etiologies (e.g., viral hepatitis, alcohol consumption, drug-induced liver injury); Patients who are candidates for liver transplantation due to cirrhosis;
• Duplicate publications or overlapping populations, in which case the most comprehensive or recent study is retained.

Two reviewers (Y.Z. and X.H.) independently evaluated the eligibility of each study. Any disagreements were resolved through discussion between the two reviewers and a third author of the paper. The PRISMA flow diagram documents and summarizes the reasons for exclusion at the full-text stage. Tables S1,S2 provide detailed definitions and study characteristics.

Data extraction

Data were extracted independently by the same two reviewers (Y.Z. and X.H.) using a pre-piloted Excel form. Data were extracted using a standardized data extraction form, summarizing the following key information: the number of patients with comorbid MASLD and eoACRN (or the corresponding OR); the first author; the year of publication; the total sample size of the included studies; the country or institution where the study was conducted; the type of study design; the diagnostic method for MASLD; the diagnostic method for high-risk CRA or advanced colorectal tumors; the diagnostic method for diabetes mellitus; the number (or OR) of MASLD-eoACRN patients with comorbid diabetes; the BMI diagnostic criteria; and the number (or OR) of obese or overweight MASLD-eoACRN patients, as well as the number (or OR) of male versus female MASLD-eoACRN patients.

In the preliminary analysis, the OR was used as a common summary effect measure. The adjusted OR was prioritized for extraction when available. In cases where the adjusted OR was not reported but sufficient raw data was available, the crude OR and the corresponding 95% CI were calculated. Studies reporting other relevant metrics (such as relative risk or hazard ratio) were not considered interchangeable with the OR unless sufficient information was available for appropriate coordination or sensitivity analysis. Missing or unavailable information was recorded as “not reported.” The extracted features are summarized in Appendix 3 to provide a standardized overview of the included studies.

Quality assessment

The Newcastle-Ottawa Scale (NOS) was employed to assess the methodological quality of the included observational studies. Cohort studies were evaluated using the standard NOS for cohort studies, while cross-sectional studies were assessed using the modified NOS (13,14). Two reviewers (Y.Z. and X.H.) independently conducted the quality assessments and resolved discrepancies through discussion or consultation with a third reviewer. An additional evaluation of the certainty of evidence for each outcome was conducted using the GRADE framework (15). Since all included studies were observational, the initial certainty of evidence was rated as low. When appropriate, the certainty of evidence was downgraded due to risks of bias, inconsistency, indirectness, imprecision, and publication bias. Potential upgrading factors, including large effect sizes, evidence of a dose-response gradient, and plausible residual confounding factors that might reduce the observed associations, were considered, but upgrades were not applied unless explicitly supported by available data. The GRADE criteria profile is presented in Figure S1.

Statistical analysis

The primary endpoint of this study was to assess the association between eoACRN and MASLD. All analyses were conducted using Stata 18.0 SE software (StataCorp, College Station, TX, USA). The OR is commonly used as a summary effect measure. Due to variations in definitions and measurement methods across different studies, a random-effects model was employed for data pooling. The Cochran’s Q test and I² statistic were used to assess statistical heterogeneity. An I² value greater than 50% is considered indicative of substantial heterogeneity. Cumulative meta-analysis was conducted based on publication year to explore temporal changes in the combined estimates. Subgroup analyses were performed based on study-level characteristics when sufficient data were available, in order to investigate potential sources of heterogeneity. Further sensitivity analyses were conducted to evaluate the robustness of the combined estimates. Galbraith plots were utilized to identify potential outlier or influential studies contributing to heterogeneity. Formal tests for funnel plot asymmetry were only conducted when there were at least 10 studies available for a given analysis. For analyses containing fewer than 10 studies, formal assessments of publication bias were not performed due to limited statistical power and the risk of misleading results.

Results

Features of selected studies

The screening process commenced with a preliminary review of all abstracts to exclude articles that did not align with the research topic and to ensure relevance to the study question. Following this, the primary reviewer (S.Q.) conducted full-text screening to verify that each study met the predefined inclusion criteria. The initial search yielded 2,661 records (410 from PubMed, 415 from Web of Science, 1,646 from Embase, five from the Cochrane Library, and 185 from CNKI), which excluded 689 duplicates. After a thorough examination of titles and abstracts, 1,958 records were excluded, leaving 14 articles that qualified for full-text review. After a thorough review of the entire text, four studies were excluded due to predefined eligibility criteria, including incompatible study populations, definitions of outcomes, or insufficient data. Ultimately, 10 studies were included in the meta-analysis (the detailed selection process is shown in Figure 1).

Figure 1 PRISMA flow diagram for the study selection process. The diagram illustrates the identification, screening, eligibility assessment, and inclusion of studies in the systematic review and meta-analysis.

Figure S1 summarizes the main characteristics of the ten studies included in the final analysis, which collectively involved 76,688 participants. The included studies were observational, comprising cohort and cross-sectional designs. All studies were conducted in Asia, including five from South Korea, three from Japan, and two from China. Most studies adjusted for key covariates such as age and sex, although adjustments for other potential confounders varied across studies. Data on study design, exposure and outcome definitions, age thresholds, effect measures, and covariate adjustments were extracted using a standardized framework, with missing information recorded as “not reported” in Figure S1.

We employed the NOS and its modified version to assess the methodological quality of the studies: among the 7 cohort studies, 1 study scored 9 points, 4 studies scored 8 points, and 2 studies scored 7 points, indicating predominantly good quality. Among the 3 cross-sectional studies, 2 scored 6 points and 1 scored 5 points, reflecting an overall moderate to low quality. Furthermore, we applied the GRADE evidence grading system for a comprehensive evaluation, which revealed that 4 studies were rated as ‘moderate’ quality and 6 studies as ‘low’ quality. Given the limited number of included studies (n=10), the evidence rated as ‘low’ or higher still holds certain clinical relevance and can provide a reference for clinical decision-making. For detailed GRADE Evidence Profile, see Table S3. The NOS assessment can be found in Tables S4,S5.

Main outcomes of standard and cumulative meta-analysis

A total of 10 studies were included to evaluate the association between MASLD and eoACRN risk. Of these, two focused on high-risk colorectal adenomas (16,17), and eight on advanced colorectal neoplasia (11,18-24). Compared with non-MASLD individuals, MASLD was associated with a significantly increased risk of eoACRN (OR =1.39, 95% CI: 1.28–1.51), although substantial heterogeneity was observed (I²=92.1%, P<0.001), potentially due to differences in population characteristics, MASLD diagnostic criteria, and follow-up duration (Figure 2). In region-based subgroup analyses, similar effect estimates were observed in Korean (OR =1.38, 95% CI: 1.24–1.54) and Japanese studies (OR =1.36, 95% CI: 1.18–1.56), although heterogeneity remained high within both subgroups (I²=94.5% and 88.6%, respectively). In contrast, Chinese studies showed a higher pooled risk (OR =1.74, 95% CI: 1.30–2.32) with no significant heterogeneity (I²=0%). The difference between subgroups was statistically significant (P<0.05), suggesting that geographic region may contribute to heterogeneity.

Figure 2 Forest plot showing the association between MASLD and eoACRN, including regional subgroup analyses. CI, confidence interval; eoACRN, early-onset advanced colorectal neoplasia; MASLD, metabolic dysfunction-associated steatotic liver disease; OR, odds ratio.

Further analyses showed that MASLD was associated with an increased risk of EOCRC (OR =1.29, 95% CI: 1.21–1.38), although heterogeneity remained substantial (I²=86.2%, P<0.001). The association was stronger for high-risk/advanced adenomas (OR =2.03, 95% CI: 1.81–2.28), with low heterogeneity (I²=0.0%). These findings suggest that tumor subtype may also contribute to between-study heterogeneity (Figure 3).

Figure 3 Forest plot of the association between MASLD and EOCRC as well as high-risk adenomas. CI, confidence interval; EOCRC, early-onset colorectal cancer; MASLD, metabolic dysfunction-associated steatotic liver disease; OR, odds ratio.

Cumulative meta-analysis indicated that evidence for an association between MASLD and eoACRN has accumulated since 2019. Although later studies refined the precision of the pooled estimates, the direction and magnitude of the association remained stable. Given the substantial heterogeneity and variability in study design and outcome definitions, these findings should be interpreted as demonstrating temporal robustness rather than providing confirmatory evidence (Figure 4).

Figure 4 Cumulative meta-analysis of the 10 included studies assessing the association between MASLD and eoACRN risk. CI, confidence interval; eoACRN, early-onset advanced colorectal neoplasia; MASLD, metabolic dysfunction-associated steatotic liver disease; OR, odds ratio; REML, restricted maximum likelihood.

Subgroup analyses

Four studies evaluated the risk of eoACRN in MASLD patients based on obesity status, using a BMI >25 kg/m² as the exposed group and BMI ≤25 kg/m² as the control group (11,16,17,24). The combined OR was 2.31 (95% CI: 0.97–5.49), which did not reach statistical significance and exhibited extremely high heterogeneity (I²=97.85%), with a wide CI indicating considerable uncertainty in the estimate (Figure 5A). Furthermore, three additional studies defined the exposed group as BMI >23 kg/m² and the control group as BMI ≤23 kg/m² (16-18), yielding a combined OR of 3.19 (95% CI: 0.57–17.87), again demonstrating very high heterogeneity (I²=97.66%) and significant imprecision (Figure 5B).

Figure 5 Subgroup analysis of eoACRN risk among individuals with MASLD. (All analyses were restricted to participants diagnosed with MASLD.) (A) Compares individuals with BMI >25 vs. ≤25 kg/m². (B) Compares individuals with BMI >23 vs. ≤23 kg/m². (C) Compares patients with diabetes vs. those without diabetes. (D) Compares males vs. females. In each comparison, the latter group serves as the reference category. These subgroup analyses should be interpreted with caution, as some are based on a limited number of studies and demonstrate significant heterogeneity. BMI, body mass index; CI, confidence interval; eoACRN, early-onset advanced colorectal neoplasia; MASLD, metabolic dysfunction-associated steatotic liver disease; OR, odds ratio; REML, restricted maximum likelihood.

Four studies assessed the association between diabetes status and risk of eoACRN in MASLD patients, with non-diabetic patients serving as the comparator group (11,16-18). The pooled analysis showed a positive association between diabetes and eoACRN risk, which did not reach statistical significance (OR =2.16, 95% CI: 0.98–4.74, P=0.06), accompanied by substantial heterogeneity (I²=96.49%) (Figure 5C).

Three studies assessed sex differences in the risk of eoACRN among MASLD patients, with females serving as the reference group (16,17,22). The pooled OR for males was 3.93 (95% CI: 1.56–9.90), indicating a statistically significant association. However, this finding was based on only three studies and was accompanied by very high heterogeneity (I²=96.49%, P<0.001) and a wide CI, suggesting substantial uncertainty in the effect estimate. Therefore, these sex-specific results should be considered exploratory and require validation in large, multicenter, prospective cohort studies (Figure 5D).

Evidence regarding tumor site specificity remains limited and is largely derived from individual studies rather than pooled analyses. A subgroup analysis from a single study showed that the incidence of EOCRC in patients with MASLD was higher for colon tumors than for rectal tumors (RR =1.68, 95% CI: 1.55–1.82) (22). Similarly, Kim et al. reported a dose-response association between FLI and EOCRC risk that appeared restricted to the left colon and rectum, with no significant association observed in the right colon (11). However, given the limited number of studies and inconsistent reporting of tumor location across the literature, site-specific associations remain inconclusive and should be interpreted with caution.

In summary, the existing literature suggests a significant association between MASLD and the risk of eoACRN. Subgroup analyses indicate that the magnitude of this association may vary across sex, diabetes status, BMI categories, and tumor characteristics, suggesting potential effect modification. However, most subgroup analyses were based on only three to four studies and showed substantial heterogeneity. Therefore, these findings remain exploratory and are insufficient to confirm effect modification, warranting further validation in larger, high-quality studies.

Sensitivity analyses

Sensitivity analysis was conducted using a leave-one-out approach, in which each study was sequentially excluded and the pooled effect estimate recalculated. The original pooled OR was 1.39 (95% CI: 1.28–1.51). After sequential exclusion of individual studies, the pooled OR remained stable, ranging from 1.22 to 1.68. Notably, exclusion of the study by Chang et al. slightly reduced the pooled OR to 1.31 (95% CI: 1.22–1.40), which remained within the original CI and did not alter statistical significance. These results suggest that the overall association was not driven by any single study; however, interpretation should remain cautious given the substantial heterogeneity observed in the primary analysis (Figure 6A).

Figure 6 Leave-one-out sensitivity analysis and Galbraith plot for the association between MASLD and eoACRN. CI, confidence interval; eoACRN, early-onset advanced colorectal neoplasia; MASLD, metabolic dysfunction-associated steatotic liver disease; OR, odds ratio; REML, restricted maximum likelihood.

In addition, a Galbraith plot was used to assess heterogeneity and potential outliers. Most studies clustered around the regression line within the 95% confidence limits, with only one study showing slight deviation, and no clear outliers were identified. Nevertheless, this does not indicate low heterogeneity, as substantial heterogeneity persisted in the primary analysis (I²=92.1%). These findings suggest that the observed heterogeneity is more likely attributable to underlying clinical and methodological differences across studies rather than the influence of a single outlying study (Figure 6B).

In the sensitivity analysis excluding cross-sectional studies, all included populations were restricted to patients with EOCRC, and the pooled effect estimate remained statistically significant (OR =1.28, 95% CI: 1.20–1.37), although attenuated compared with the primary analysis. Country-specific analyses showed a marked reduction in heterogeneity in the South Korean subgroup (from 94.5% to 9.3%), suggesting that cross-sectional studies may contribute substantially to heterogeneity in this subgroup. However, persistently high heterogeneity in other countries indicates that differences in population characteristics, MASLD definitions, and outcome ascertainment may continue to influence the pooled estimates (Figure S2).

In subgroup analyses related to diabetes, preliminary results suggested a positive association between diabetes and eoACRN risk in patients with MASLD, although this did not reach statistical significance (OR =2.16, 95% CI: 0.98–4.74, P=0.06). After excluding the study by Kim et al., sensitivity analysis yielded a larger pooled estimate that became statistically significant (OR =2.93, 95% CI: 1.34–6.43, P=0.01), although substantial heterogeneity persisted (I²=88.14%) (Figure S3). These findings suggest that the pooled effect estimate for diabetes is sensitive to the inclusion of individual studies and should be interpreted as exploratory evidence.

Furthermore, sensitivity analyses combining studies using different BMI thresholds to define overweight/obesity showed a significant association (OR =2.67, 95% CI: 1.19–6.00); however, substantial heterogeneity was observed across studies (I²=98.32%), which may partly reflect differences in BMI definitions, population characteristics, and study designs (Figure S4).

Discussion

In recent years, metabolic dysfunction—including obesity, T2DM, and MASLD—has been established as a major driver of eoACRN. The gut-liver axis provides a biologically plausible framework linking hepatic steatosis and colorectal tumors (25). Hepatic lipid accumulation not only induces local inflammation but also promotes the transformation of colonic adenomas into carcinomas through systemic chronic low-grade inflammation and hyperinsulinemia, with these effects manifesting even in the early stages of metabolic injury (26). Consequently, the present study simultaneously enrolled patients with high-risk adenomas and advanced CRC to underscore the importance of early screening in high-risk populations.

Current evidence suggests that MASLD is associated with an increased risk of eoACRN. The underlying mechanisms have been proposed in both experimental and clinical studies and may involve insulin resistance, chronic low-grade systemic inflammation, gut microbiota dysbiosis, and metabolic reprogramming (27,28). A prospective cohort study by Li et al. demonstrated that a higher burden of metabolic abnormalities is associated with a progressively increased risk of CRC (29). Consistently, Yoon et al. reported an elevated CRC risk in MASLD patients with normal lipid levels and lower waist circumference, supporting a role for metabolic reprogramming even in non-obese phenotypes (30). Moreover, gut microbiota dysbiosis in MASLD is characterized by reduced microbial diversity, depletion of beneficial taxa, and enrichment of opportunistic pathogens, leading to impaired intestinal barrier integrity and increased permeability (31). This facilitates translocation of lipopolysaccharide via the portal circulation, triggering systemic and hepatic inflammation. These processes promote colonic epithelial proliferation, DNA damage, and adenoma–carcinoma progression, forming the core of the gut–liver–colon axis (32). Studies also indicate that adipokine-mediated inflammation and proliferative signaling may contribute to the increased risk of colorectal adenomas in MASLD (33). Collectively, these metabolic, inflammatory, and microbial pathways interact synergistically and may explain the tumor-promoting effects of MASLD even in non-cirrhotic stages.

This study synthesizes outcomes related to early colorectal tumorigenesis using a composite endpoint encompassing advanced adenomas and EOCRC. However, advanced adenomas and EOCRC represent distinct stages along the spectrum of colorectal carcinogenesis. Therefore, pooled estimates for eoACRN should be interpreted with caution, as separate analyses of EOCRC and advanced adenoma outcomes remain clinically important. Furthermore, in the included studies, MASLD-related hepatic steatosis was associated with an increased risk of eoACRN in Asian populations. However, substantial heterogeneity was observed across studies; thus, these findings should be interpreted as evidence of associations derived from heterogeneous observational studies rather than as evidence sufficient to infer causality.

In the stratified analysis of diabetes, patients with both MASLD and diabetes showed a trend toward an increased risk of eoACRN, although this association did not reach statistical significance. In the post-hoc sensitivity analysis excluding the study by Kim et al., a larger pooled effect estimate that became statistically significant was observed; however, substantial heterogeneity persisted, and the results remained sensitive to the inclusion of individual studies. Therefore, whether diabetes modifies the association between MASLD and eoACRN remains inconclusive based on current evidence. However, from a biological perspective, insulin resistance and compensatory hyperinsulinemia represent plausible mechanistic links between MASLD, diabetes, and colorectal tumorigenesis (34). In addition, a meta-analysis including 19 studies has identified type 2 diabetes as an independent risk factor for EOCRC (OR =1.43, 95% CI: 1.08–1.80), with a further increased risk observed in individuals with poor glycemic control (OR range: 1.37–1.59) (35). These findings suggest that type 2 diabetes may remain an important metabolic comorbidity contributing to colorectal tumor risk, even in the context of MASLD.

Previous studies have suggested that the association between MASLD and colorectal tumorigenesis is not entirely dependent on obesity status. Chang et al. reported that young patients with MASLD, regardless of obesity status, had significantly higher prevalences of both low-risk and high-risk adenomas, with adjusted ORs of 1.30 and 1.40, respectively. This association persisted even among lean individuals with a BMI <23 kg/m² (16). Consistently, Xiong et al. demonstrated a 2.3-fold increased risk of colorectal adenoma in lean MASLD patients (95% CI: 1.5–3.6) (36). However, a 2026 health screening cohort study from Taiwan did not observe an additional association between lean MASLD and adenoma risk, highlighting inconsistencies in the current evidence and suggesting that BMI alone may be an insufficient surrogate for the metabolic risk profile of patients with MASLD (37). In our BMI-stratified analyses, pooled estimates indicated a trend toward increased risk of eoACRN in MASLD patients, regardless of whether BMI cut-offs of 23 or 25 kg/m² were used to define overweight/obesity; however, these associations did not reach statistical significance and were accompanied by wide CIs and substantial between-study heterogeneity. Notably, in sensitivity analyses pooling studies with different BMI thresholds, higher BMI was significantly associated with an increased risk of eoACRN in MASLD patients. Nevertheless, this result remained highly heterogeneous (I²>95%), indicating substantial variability across studies. Overall, the current evidence suggests that elevated BMI may act as a potential risk enhancer rather than an independent determinant of eoACRN in MASLD. Given the emerging concepts of “metabolically obese normal weight” and “lean MASLD”, future studies should incorporate more sensitive measures of adiposity and metabolic dysfunction, such as waist circumference, visceral fat area, body fat distribution, and metabolic biomarkers, to better characterize obesity phenotypes and clarify their role in the development of eoACRN among patients with MASLD.

This study showed that among patients with eoACRN and comorbid MASLD, males exhibited a higher risk than females; however, the CIs were wide and substantial heterogeneity was observed across studies. Notably, a health screening cohort of 4,436 individuals undergoing colonoscopy and abdominal ultrasound found that although the overall detection rate of colorectal adenomas was slightly higher in males with MASLD, the association between MASLD and advanced colorectal adenomas appeared more pronounced in females (OR =3.93, 95% CI: 1.02–15.11, P=0.046) (38). Similarly, Yoon et al. reported a non-significant association between MASLD and CRC in women, potentially due to limited statistical power or a predominantly premenopausal population, in whom estrogen deficiency may further potentiate the pro-tumorigenic effects of MASLD. In addition, MASLD has been shown to independently increase CRC risk in women (adjusted HR=1.10, 95% CI: 1.00–1.20), with a significantly higher 7-year cumulative incidence (0.47% vs. 0.43%, P=0.006) (27). These sex-specific patterns highlight the complex interplay between sex hormones and metabolic dysfunction, suggesting that pooled effect estimates may not fully capture underlying biological heterogeneity. Future studies incorporating larger real-world datasets are needed to generate more robust and clinically generalizable evidence.

Current CRC screening recommendations are primarily age-based; however, a proportion of eoACRN/EOCRC occurs before the initiation of routine screening. Our findings suggest that MASLD-related metabolic dysfunction may provide a potential basis for future risk stratification models. Nevertheless, the current evidence is insufficient to support MASLD-based recommendations for earlier screening, shortened surveillance intervals, or disease-specific monitoring strategies.

The review protocol was not registered in PROSPERO, which is considered a limitation. In addition, this study has several limitations. First, eoACRN was defined as a composite endpoint including advanced adenomas and/or EOCRC. Although this captures early neoplastic lesions relevant to prevention, advanced adenomas and EOCRC represent distinct clinical and biological entities, limiting the interpretability of pooled estimates; future studies should report these outcomes separately whenever possible. Second, substantial heterogeneity was observed, likely due to differences in study design, MASLD definitions (NAFLD, MAFLD, and MASLD), diagnostic criteria, and outcome ascertainment. The lack of individual participant data precluded harmonization under unified MASLD criteria, and variability in alcohol intake thresholds further contributed to inconsistency. Third, all studies were conducted in Asian populations, mainly from Korea, Japan, and China, limiting generalizability to non-Asian populations due to differences in genetic, metabolic, and lifestyle factors. Minor overlap among studies may exist due to the use of multiple national South Korean datasets, potentially causing duplicate sample sizes, bias in pooled effect estimates, and underestimated CIs. Fourth, subgroup analyses were based on only three to four studies and showed substantial heterogeneity; thus, these results should be considered exploratory. Sensitivity analyses should likewise be interpreted as post hoc findings rather than primary evidence. Fifth, most studies were observational, including cross-sectional designs; therefore, results reflect associations rather than causality. Residual confounding and detection bias are likely due to inconsistent adjustment for lifestyle, dietary, and metabolic factors, and increased surveillance in metabolic disease populations. Sixth, publication bias could not be reliably assessed due to the small number of studies, particularly in subgroup analyses; thus, funnel plot and Egger’s test results should be interpreted cautiously. Finally, the methodological quality of included studies was variable. Study quality was assessed using the NOS, and certainty of evidence was evaluated using the GRADE framework. Overall, the certainty of evidence was rated as low for most outcomes due to reliance on observational data, inconsistency, imprecision, indirectness, and potential publication bias. These limitations should be considered when interpreting the pooled estimates.

Future research should prioritize large, multicenter prospective cohort studies with standardized definitions of MASLD, harmonized alcohol intake thresholds, consistent age cutoffs, and separate reporting of EOCRC and advanced adenomas. Detailed adjustment for metabolic, lifestyle, and healthcare utilization factors will be essential to clarify whether MASLD independently contributes to early colorectal tumorigenesis and whether specific metabolic phenotypes can improve clinically useful risk stratification.

Conclusions

In conclusion, in Asian populations, MASLD-associated hepatic steatosis may be associated with an increased risk of eoACRN. However, due to substantial between-study heterogeneity and limited subgroup evidence, no definitive conclusions can be drawn regarding risk modification by BMI, sex, or diabetes status. Nevertheless, the current evidence suggests that male MASLD patients with higher BMI may represent a potentially high-risk subgroup, for whom enhanced CRC screening could be considered.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0401/rc

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0401/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0401/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Meng Y, Tan Z, Zhen J, et al. Global, regional, and national burden of early-onset colorectal cancer from 1990 to 2021: a systematic analysis based on the global burden of disease study 2021. BMC Med 2025;23:34. [Crossref] [PubMed]
2. Akimoto N, Ugai T, Zhong R, et al. Rising incidence of early-onset colorectal cancer - a call to action. Nat Rev Clin Oncol 2021;18:230-43. [Crossref] [PubMed]
3. Kim DH. Epidemiology of colonic adenoma and cancer. Clin Endosc 2026;59:333-44. [Crossref] [PubMed]
4. Kyrochristou ID, Lianos GD, Kyrochristou GD, et al. Early-onset colorectal cancer in patients younger than 50 years: a systematic review of the literature. Ann Gastroenterol 2025;38:364-79. [Crossref] [PubMed]
5. Jayakrishnan T, Ng K. Early-Onset Gastrointestinal Cancers: A Review. JAMA 2025;334:1373-85. [Crossref] [PubMed]
6. Canivet CM, Boursier J, Loomba R. New Nomenclature for Nonalcoholic Fatty Liver Disease: Understanding Metabolic Dysfunction-Associated Steatotic Liver Disease, Metabolic Dysfunction- and Alcohol-Associated Liver Disease, and Their Implications in Clinical Practice. Semin Liver Dis 2024;44:35-42. [Crossref] [PubMed]
7. Zhong Z, Xu M, Luo X, et al. Global, regional, and national burden of NAFLD in adults aged 20-49 years (1990-2021): a comprehensive analysis of trends, decomposition, and health inequality. BMC Gastroenterol 2025;26:23. [Crossref] [PubMed]
8. Ali SMJ, Lai M. Metabolic Dysfunction-Associated Steatotic Liver Disease. Ann Intern Med 2025;178:ITC1-ITC16. [Crossref] [PubMed]
9. Parizadeh SM, Parizadeh SA, Alizade-Noghani M, et al. Association between non-alcoholic fatty liver disease and colorectal cancer. Expert Rev Gastroenterol Hepatol 2019;13:633-41. [Crossref] [PubMed]
10. Zhao JF, Zhou BG, Lv Y, et al. Association between metabolic dysfunction-associated steatotic liver disease and risk of colorectal cancer or colorectal adenoma: an updated meta-analysis of cohort studies. Front Oncol 2024;14:1368965. [Crossref] [PubMed]
11. Kim S, Jung JH, Han K, et al. Association Between Nonalcoholic Fatty Liver Disease and Risk of Early-onset Colorectal Cancer. Clin Gastroenterol Hepatol 2025;23:2550-2558.e2. [Crossref] [PubMed]
12. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Int J Surg 2021;88:105906. [Crossref] [PubMed]
13. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Hospital Research Institute; 2000. Available online:
14. Gualdi-Russo E, Zaccagni L. The Newcastle–Ottawa Scale for Assessing the Quality of Studies in Systematic Reviews. Publications 2026;14:4.
15. Guyatt G, Oxman A, Akl E, et al. GRADE guidelines: 1. Introduction—GRADE evidence profiles and summary of findings tables. J Clin Epidemiol 2011;64:383-94.
16. Chang J, Chang Y, Cho Y, et al. Metabolic-associated fatty liver disease is associated with colorectal adenomas in young and older Korean adults. Liver Int 2023;43:2548-59. [Crossref] [PubMed]
17. Peng S, Song X, Ma X, et al. Metabolic-associated fatty liver disease and Colorectal Adenoma: A Cross-Sectional Study in a Metabolically Characterized Xinjiang Population, China. Dig Liver Dis 2025;57:2384-9. [Crossref] [PubMed]
18. Bae JK, Park JJ, Kim SU, et al. Metabolic dysfunction-associated fatty liver disease is an independent risk factor for advanced colorectal adenoma or cancer. Eur J Gastroenterol Hepatol 2026;38:632-8. [Crossref] [PubMed]
19. Choi YJ, Lee DH, Han KD. Association between high fatty liver index and development of colorectal cancer: a nationwide cohort study with 21,592,374 Korean. Korean J Intern Med 2020;35:1354-63. [Crossref] [PubMed]
20. Hamaguchi M, Hashimoto Y, Obora A, et al. Non-alcoholic fatty liver disease with obesity as an independent predictor for incident gastric and colorectal cancer: a population-based longitudinal study. BMJ Open Gastroenterol 2019;6:e000295. [Crossref] [PubMed]
21. Kimura T, Tamaki N, Wakabayashi SI, et al. Colorectal Cancer Incidence in Steatotic Liver Disease (MASLD, MetALD, and ALD). Clin Gastroenterol Hepatol 2025;23:2197-2204.e2. [Crossref] [PubMed]
22. Park JH, Hong JY, Shen JJ, et al. Increased Risk of Young-Onset Digestive Tract Cancers Among Young Adults Age 20-39 Years With Nonalcoholic Fatty Liver Disease: A Nationwide Cohort Study. J Clin Oncol 2023;41:3363-73. [Crossref] [PubMed]
23. Tamaki N, Kimura T, Wakabayashi SI, et al. Gastrointestinal Cancer Risk in Steatotic Liver Diseases: MASLD, MetALD and ALD. Aliment Pharmacol Ther 2026;63:979-91. [Crossref] [PubMed]
24. Wu PH, Chung CH, Wang YH, et al. Association between nonalcoholic fatty liver disease and colorectal cancer: A population-based study. Medicine (Baltimore) 2023;102:e33867. [Crossref] [PubMed]
25. Joon A, Sharma A, Jalandra R, et al. Nonalcoholic Fatty Liver Disease and Gut-liver Axis: Role of Intestinal Microbiota and Therapeutic Mechanisms. J Transl Gastroenterol 2024;2:38-51.
26. Chakraborty D, Wang J. Nonalcoholic fatty liver disease and colorectal cancer: Correlation and missing links. Life Sci 2020;262:118507. [Crossref] [PubMed]
27. Dua A, Kumari R, Singh M, et al. Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD): the interplay of gut microbiome, insulin resistance, and diabetes. Front Med (Lausanne) 2025;12:1618275. [Crossref] [PubMed]
28. Targher G, Byrne CD, Tilg H. MASLD: a systemic metabolic disorder with cardiovascular and malignant complications. Gut 2024;73:691-702. [Crossref] [PubMed]
29. Li Y, Ma XM, Jia JG, et al. Association between metabolic dysfunction-associated steatotic liver disease, metabolic dysfunction subtypes and risk of colorectal cancer: A prospective cohort study. Clin Res Hepatol Gastroenterol 2025;49:102573. [Crossref] [PubMed]
30. Yoon CI, Lee HS, Jeon S, et al. Heterogeneous Colorectal Cancer Risk in Women with Metabolic Dysfunction-Associated Steatotic Liver Disease by Age, Lipid, and Waist-Circumference: A Nationwide Cohort Study. Cancers (Basel) 2025;18:125. [Crossref] [PubMed]
31. Luo J, Wei Y, Fan Z, et al. Gut microbiota as a metabolic mediator to prevent or attenuate MASLD. Critical Reviews in Food Science and Nutrition 2026;66:1446-62.
32. Jayachandran M, Qu S. Non-alcoholic fatty liver disease and gut microbial dysbiosis- underlying mechanisms and gut microbiota mediated treatment strategies. Rev Endocr Metab Disord 2023;24:1189-204. [Crossref] [PubMed]
33. Mitsala A, Tsalikidis C, Romanidis K, et al. Non-Alcoholic Fatty Liver Disease and Extrahepatic Cancers: A Wolf in Sheep’s Clothing? Curr Oncol 2022;29:4478-510. [Crossref] [PubMed]
34. Talamantes S, Lisjak M, Gilglioni EH, et al. Non-alcoholic fatty liver disease and diabetes mellitus as growing aetiologies of hepatocellular carcinoma. JHEP Rep 2023;5:100811. [Crossref] [PubMed]
35. Khoa Ta HD, Nguyen NN, Ho DKN, et al. Association of diabetes mellitus with early-onset colorectal cancer: A systematic review and meta-analysis of 19 studies including 10 million individuals and 30,000 events. Diabetes Metab Syndr 2023;17:102828. [Crossref] [PubMed]
36. Xiong J, Wu Y, Chen D, et al. MAFLD with central obesity is associated with increased risk of colorectal adenoma and high-risk adenoma. BMC Gastroenterol 2024;24:138. [Crossref] [PubMed]
37. Wu KW, Liang KS, Hsu YC, et al. Association between metabolic dysfunction-associated steatotic liver disease, lean metabolic dysfunction-associated steatotic liver disease, and colorectal neoplasms: a cross-sectional study. Eur J Gastroenterol Hepatol 2026;38:216-21. [Crossref] [PubMed]
38. Gong Y, Kang J, Wang X, et al. Increased detection rates of advanced colorectal adenoma in women with metabolic dysfunction-associated fatty liver disease. Heliyon 2023;9:e22391. [Crossref] [PubMed]