Association between vitamin D and metabolic-associated steatotic liver disease in type 2 diabetes mellitus patients: a systematic review and meta-analysis Authors: Bruno Basil, Winifred Adiri, Ejike Lawrence Ugwu, Ngozi Ijeoma Okoro Published: 2025 ABSTRACT AbstractBackgroundMetabolic-associated steatotic liver disease (MASLD) is a common liver disorder among adults with type 2 diabetes mellitus (T2DM) and is associated with increased risk of liver-related complications and cardiovascular disease. Vitamin D deficiency has been proposed as a potential modifiable factor influencing liver fat accumulation and metabolic dysfunction, but evidence for its role in T2DM remains inconsistent.Main bodyThis systematic review and meta-analysis was conducted to evaluate the association between circulating vitamin D levels and MASLD in adults with T2DM. Comprehensive searches of multiple databases and other sources from January 2000 to January 2025 identified 158 records, of which ten studies including 4,151 participants met inclusion criteria. Most studies were conducted in Asia and the Middle East, with limited data from other regions. Studies measured vitamin D using validated laboratory methods and assessed liver disease using imaging, elastography, or biomarker-based algorithms. Pooled analysis using a random-effects model showed that individuals with liver disease had significantly lower vitamin D concentrations compared to those without the disease, with a moderate effect size (approximately half a standard deviation lower vitamin D levels on average). Considerable heterogeneity was observed across studies, largely explained by differences in study design and geographic region, while sex did not significantly influence the association. Sensitivity analyses and tests for publication bias indicated the findings were robust. Most included studies were cross-sectional, limiting causal inference, and the geographic distribution was concentrated in specific regions, highlighting gaps in evidence from other populations.ConclusionLower vitamin D levels are moderately associated with liver fat accumulation in adults with type 2 diabetes. Although the findings support the potential role of vitamin D as a marker of metabolic liver dysfunction, the predominance of cross-sectional studies and regional concentration of research limit definitive conclusions. Future longitudinal and interventional studies across diverse populations are needed to determine whether optimizing vitamin D status can influence the development or progression of liver disease in this high-risk population.Keywords: Hepatic steatosis, Insulin resistance, 25-hydroxycholecalciferol, 25-hydroxyvitamin D, Vitamin D insufficiency, Non-alcoholic fatty liver disease, Vitamin D supplementation, Hepatic fibrosis, Systematic synthesis, Heterogeneity Introduction Metabolic-associated steatotic liver disease (MASLD), previously termed non-alcoholic fatty liver disease (NAFLD), has emerged as the most common chronic liver disorder worldwide, with recent estimates placing its global prevalence at around 38% in adults, with projections indicating an increase to over 55% by 2040 if current trends persist [1, 2]. This widespread condition is not uniformly distributed; prevalence rates are higher in regions with rising obesity and type 2 diabetes mellitus (T2DM), reflecting the close interplay between metabolic dysfunction and hepatic steatosis [3–5]. T2DM in particular represents the strongest metabolic risk factor for MASLD. More than 65% of individuals with T2DM are estimated to have MASLD, and these patients face a substantially higher risk of progression to metabolic-associated steatohepatitis (MASH), advanced fibrosis, cirrhosis, and hepatocellular carcinoma compared with their non-diabetic counterparts [4, 6]. The coexistence of MASLD and T2DM not only increases liver-related morbidity and mortality but also exacerbates cardiovascular risk, which remains the leading cause of death in this population [7, 8]. Global projections suggest that the number of adults living with T2DM will surpass 850 million by 2050 [9], as such, the parallel rise in MASLD emphasizes the urgency of identifying modifiable risk factors that could influence its development and progression. Among such factors, vitamin D deficiency has generated growing scientific interest. Vitamin D insufficiency is highly prevalent in patients with MASLD [10], a condition that disproportionately affects individuals with T2DM, raising the possibility that it may represent a shared and modifiable determinant of disease severity. Recent large syntheses and primary studies indicate that vitamin D deficiency is highly prevalent among people with T2DM, affecting roughly half to two-thirds of patients in pooled analyses and contemporary cohorts, which underscores the public-health relevance of examining vitamin D’s relationship with MASLD in this population [11, 12]. Although it has been long recognized for its role in calcium and bone metabolism, vitamin D is now established as a pleiotropic hormone with wide-ranging metabolic, endocrine, and immunomodulatory actions [13]. Several biological mechanisms support its potential role in MASLD pathogenesis including enhancement of insulin sensitivity through effects on pancreatic β-cell function and peripheral glucose uptake, regulation of lipid metabolism by modulating hepatic de novo lipogenesis and fatty acid oxidation, and exhibiting anti-inflammatory and anti-fibrotic properties by suppressing pro-inflammatory cytokine signaling, thereby reducing hepatic stellate cell activation, and modulating the renin–angiotensin system [10]. Recent mechanistic studies indicate that vitamin D may reduce hepatic fibrosis by modulating autophagy and inhibiting pyroptosis in liver sinusoidal endothelial cells [14]. These mechanisms provide a strong rationale for hypothesizing that low vitamin D level is associated with, and may worsen hepatic steatosis, necroinflammation, and fibrosis in individuals with T2DM, whereas supplementation might attenuate both metabolic and hepatic injury. Despite this compelling biological plausibility, the clinical evidence to date remains inconclusive. Observational studies often report inverse associations between serum vitamin D concentrations and the presence or severity of MASLD in individuals with various metabolic disorders including T2DM, but the results vary, and many fail to retain statistical significance after adjusting for potential confounders such as adiposity, glycaemic control, and insulin resistance [15–18]. Similarly, interventional studies assessing vitamin D supplementation have yielded mixed outcomes. Some randomized controlled trials have demonstrated improvements in hepatic steatosis, insulin resistance, or systemic inflammatory markers, while others have reported negligible or no benefits [19, 20]. The heterogeneity in study design, baseline vitamin D status, dosing regimens, patient characteristics, and methods of assessing MASLD contributes to these discrepancies and complicates interpretation of the evidence. The uncertainty surrounding the role of vitamin D in MASLD among patients with T2DM reinforces the need for a systematic evaluation of the existing evidence. Clarifying whether serum vitamin D deficiency is associated with the presence and severity of MASLD in this high-risk population could have important implications for risk stratification, early detection, and clinical management. This systematic review and meta-analysis therefore aim to quantitatively synthesize observational and post-hoc trial data to evaluate the relationship between circulating serum 25-hydroxyvitamin D [25(OH)D] levels and MASLD in adults with T2DM, and to examine potential sources of heterogeneity across studies, including study design and participant characteristics. By integrating epidemiological findings with mechanistic insights, this work seeks to resolve existing uncertainties and advance understanding of vitamin D’s potential role in the natural history of MASLD in individuals with T2DM. Methods Protocol and reporting This systematic review and meta-analysis followed the PRISMA 2020 guideline. A protocol specifying the research question, eligibility criteria, outcomes, and analysis plan was prospectively registered on PROSPERO (CRD42025634176). Eligibility criteria Eligible studies enrolled adults (≥ 18 years) with type 2 diabetes mellitus (T2DM) and reported circulating 25-hydroxyvitamin D [25(OH)D] measured using a validated laboratory method (e.g., ELISA/CLIA/ECLIA or LC–MS/MS). We included observational designs (cross-sectional, case–control, cohort) and post-hoc analyses of randomized trials when they provided extractable quantitative data bearing on metabolic-associated steatotic liver disease (MASLD/NAFLD). The primary outcome was the presence/diagnosis of MASLD/NAFLD, defined by imaging (e.g., abdominal ultrasound or elastography), biomarker/algorithmic criteria (e.g., NAFLD-LFS), or biopsy. Secondary outcomes comprised indices of disease severity (e.g., steatosis grade, fibrosis stage, liver enzymes). We excluded studies of non-T2DM populations (e.g., type 1 or gestational diabetes), liver disease of alternative aetiologies (alcoholic, viral, autoimmune), non-human or paediatric populations, non-English publications, and reports lacking sufficient information to compute or transform an effect size for quantitative synthesis. Information sources and search strategy We searched EBSCOhost, Web of Science, ScienceDirect, and Scopus for human studies published from January 2000 through January 2025, restricted to English. Search strategies combined controlled vocabulary and free-text terms for vitamin D (e.g., “25-hydroxyvitamin D”, “vitamin D”), fatty liver (e.g., “MASLD”, “NAFLD”), and T2DM, with Boolean operators and database-specific filters. We also screened reference lists of included studies and relevant reviews to identify additional records. Study selection Two reviewers independently screened titles/abstracts and assessed full texts against the eligibility criteria. Disagreements were resolved by discussion or, if needed, adjudication by a third reviewer. Reasons for full-text exclusion were recorded and the study flow is depicted in the PRISMA diagram. Data extraction Using a piloted form, two reviewers independently extracted: author/year, country/region, design, sample size, participant characteristics (age, sex), vitamin D assay (and units), group means/SDs and sample sizes where available, MASLD definition/diagnostic modality, thresholds for vitamin D deficiency/insufficiency, severity measures, adjusted covariates, and effect estimates (with precision). When multiple eligible groups or time points were reported, we followed prespecified rules to avoid unit-of-analysis errors (e.g., combining arms or selecting one contrast per study). Discrepancies were reconciled by consensus, with unresolved disagreements adjudicated by a third reviewer. Risk of bias assessment Non-randomized evidence was appraised using ROBINS-I (confounding, selection, exposure classification, deviations, missing data, outcome measurement, selective reporting). Randomized evidence considered in sensitivity analyses was evaluated with RoB 2. Assessments were conducted independently in duplicate with consensus agreement. Effect size computation and data handling For the primary comparison of vitamin D levels in MASLD versus non-MASLD groups, we computed standardized mean differences (SMD; Hedges g) from group means, SDs, and sample sizes. Where units differed, vitamin D values were harmonized using standard conversions (e.g., 1 ng/mL = 2.5 nmol/L). To maintain independence when studies reported multiple eligible subgroups, we combined arms using fixed-effects formulas or selected a single prespecified comparison. Synthesis methods Primary meta-analyses pooled SMDs with inverse-variance random-effects models; between-study variance (τ2) was estimated by the DerSimonian–Laird method of moments. We report pooled effects with 95% confidence intervals (CI) and summarize heterogeneity with Q, I², and τ2. A 95% prediction interval was calculated for random-effects models. Fixed-effect models were performed as sensitivity analyses. Statistical significance was set at two-sided α = 0.05. Moderator and subgroup analyses To explain heterogeneity a priori, we evaluated study design and region as categorical moderators using mixed-effects models (random effects within levels; fixed contrasts across levels), and gender as a continuous meta-regression covariate. In CMA, categorical moderators were encoded as dummy contrasts versus a reference category, yielding regression coefficients interpreted as differences in SMD relative to the reference; the overall moderator was tested with the omnibus QMQ_MQM​ statistic. For meta-regression, we estimated the slope (β), SE, 95% CI, and z test and, when available, reported residual τ2 and R2 (proportion of between-study variance explained). These specifications match the moderation outputs presented in the Results (design contrasts “Design: 1/2”, region contrasts, and the continuous “Gender (F)” term). Small-study effects and sensitivity analyses We assessed publication/small-study effects visually with funnel plots and analytically with Egger’s regression and Begg–Mazumdar rank correlation. These assessments were pre-specified in the study protocol as part of the planned bias and sensitivity analyses. We applied Duval & Tweedie trim-and-fill and computed the classic fail-safe N as robustness checks. Additional sensitivity analyses included leave-one-out/influence diagnostics and fixed-effect re-estimation. Statistical software All analyses were conducted in Comprehensive Meta-Analysis (CMA) v4, following standard meta-analytic procedures. Result Study selection The search identified 158 records (153 via databases/registries and 5 via other sources). After screening and full-text review, 10 studies met inclusion criteria (PRISMA, Fig. 1). PRISMA diagram Characteristics of included studies Across k = 10 studies (N = 4,151), most were cross-sectional; single studies were case–control, cross-sectional case–control, retrospective observational, and a post-hoc RCT. Studies were conducted in China (n = 5), Egypt (n = 2), Iran (n = 1), United States (n = 1), and Russia (n = 1). Vitamin D assays included ELISA, CLIA/ECLIA, and LC–MS/HPLC; MASLD/NAFLD was diagnosed predominantly by ultrasound, with single studies using NAFLD-LFS or elastography. Full characteristics appear in Table 1. Table 1Characteristics of included studiesStudy (Author, Year)Country/RegionStudy Design N Mean age NAFLD (± SD)Mean age non-NAFLD (± SD)Gender NAFLD (M/F or %)Gender non-NAFLD (M/F or %)BMI NAFLD (kg/m²)BMI non-NAFLD (kg/m²)Vitamin D assay MethodMASLD diagnosis methodAdjusted variables/methodNadia A. Elshamy et al. (2015) [21]EgyptCross-sectional case–control8047.044.340.025.0--ELISAAbdominal ultrasonographyInsulin resistance (not clearly reported in regression format)Hanaa Taha Kandeel et al. (2019) [22]EgyptCase-Control Study5048.646.734.925.5--ELISAHepatic Steatosis Index (HSI) and Abdominal Ultrasound-Lei Xiu et al. (2021) [23]ChinaCross-sectional30051.540.545.738.526.526.2Automated Roche® 601 analyzerB-mode ultrasound + transaminase + alcohol historyAge, waist circumference, TGs, albumin, insulin, FBG, HOMA-IR, ALT, AST, BMI, durationYuanyuan Zhang et al. (2021). [24]ChinaRetrospective Observational Study33957.263.2--26.924.7High-Performance Liquid Chromatography–Tandem Mass Spectrometry (HPLC-MS/MS)Abdominal UltrasoundAge, BMI, GGT, Cystatin C, HDL, HOMA-IR (CP), 25-(OH)D; Multivariate Logistic RegressionVahid Sheikhi et al. (2022) [25]IranCross-Sectional Study111054.149.4--27.426.1Enzyme immunoassayAbdominal ultrasoundAge, BMI, HbA1c, insulin use, diabetes duration, ALTYuling Xing et al. (2022) [18]ChinaCross-Sectional43458.7---25.7-Electrochemiluminescence immunoassay (ECLIA, COBAS e 601, Roche, Germany)Abdominal ultrasoundAge, BMI, hypertension, T2DM duration, albumin, ALT, AST, GGT, FPG, lipid profile, HbA1cCorbin et al. (2023) [26]United StatesRandomized controlled trial (RCT)125------Mostly liquid chromatography-tandem mass spectrometry (LC-MS/MS)NAFLD-Liver Fat Score (NAFLD-LFS)Age, BMI, metabolic syndrome, insulin levels, AST/ALT ratio; Multivariate regressionYagoda A. V. et al. (2023) [27]RussiaCross-sectional9057.960.683.370.033.132.2Enzyme-linked immunosorbent assay (ELISA)Abdominal ultrasonographyage, HOMA-IR, BMI, GOT (AST), HDL cholesterolNan Huang et al. (2024) [28]ChinaCross-Sectional Study1513----26.424.4High-Performance Liquid Chromatography–Mass Spectrometry (HPLC-MS/MS)Elastography (Liver Stiffness Measurement)Age, sex, BMI, HbA1c, insulin, and HOMA-IRWei Xiang et al. (2024) [29]ChinaCross-Sectional Study11048.357.056.950.026.622.5Serum 25(OH)D3 chemiluminescence assayAbdominal ultrasoundAge, BMI, HbA1c, ALT, FBG, Gender; Multivariate logistic regression. Data not reported in the original publication are denoted by (-) Characteristics of included studies Data not reported in the original publication are denoted by (-) Overall Meta-Analytic effect Using standardized mean differences (SMD; d), the random-effects model showed a moderate negative association: d = − 0.47 (SE = 0.13), 95% CI [− 0.72, − 0.21], z = − 3.59, p < 0.001. For reference, the fixed-effect estimate was d = − 0.27 (SE = 0.03), 95% CI [− 0.33, − 0.20], z = − 8.30, p < 0.001 (Forest plot, Fig. 2). Forest plot Heterogeneity and prediction interval Between-study heterogeneity was substantial: Q(9) = 118.22, p < 0.001; I² = 92.39% with τ² = 0.14 (τ = 0.374). The 95% prediction interval spanned − 1.380 to 0.447, indicating that true effects across comparable settings may range from large negative to small positive. Moderator analyses Study design significantly moderated the association between vitamin D levels and MASLD (QM(2) = 9.84, p = 0.007; Table 2). Controlled case studies showed a stronger inverse association than randomized controlled trials, suggesting that observational designs may detect larger effect sizes, potentially due to differences in study methodology or residual confounding. Table 2Moderator analyses (random-effects models; coefficients are differences in SMD vs. reference)ModeratorLevel/Term (vs. ref)β (slope)SE95% CIz P Study design RCT0.230.36−0.47 to 0.930.640.52Controlled Case study−0.760.30−1.36 to − 0.17−2.510.012Omnibus (design)QM = 9.84df = 2; 0.007Q partition (model/residual/total)10.21/17.92/28.13df 3/6/9p = 0.009/0.019/< 0.001 Region America1.880.451.005 to 2.7604.20< 0.001Asia1.540.330.901 to 2.1854.70< 0.001Europe1.210.480.270 to 2.1402.530.011Omnibus (region)QM = 25.93df = 3; < 0.001Q partition (model/residual/total)25.93/8.67/34.60df 3/6/9p = < 0.001/0.196/< 0.001 Gender Gender (F)*−0.0120.008−0.029 to 0.005−1.490.136SMD = standardized mean difference; SE = standard error; CI = confidence interval; QM = model chi-square statistic for moderator test; Q partition = decomposition of heterogeneity into model, residual, and total components; df = degrees of freedom; p = two-tailed significance level. * Export label is “Gender (F)” (slope per 1% increase in female participants). If reporting % male, reverse the sign and rescale (e.g., per 10% male). Moderator analyses (random-effects models; coefficients are differences in SMD vs. reference) SMD = standardized mean difference; SE = standard error; CI = confidence interval; QM = model chi-square statistic for moderator test; Q partition = decomposition of heterogeneity into model, residual, and total components; df = degrees of freedom; p = two-tailed significance level. * Export label is “Gender (F)” (slope per 1% increase in female participants). If reporting % male, reverse the sign and rescale (e.g., per 10% male). Region also emerged as a significant moderator (QM(3) = 25.93, p < 0.001), with studies from America, Asia, and Europe showing higher effect estimates compared with the reference region. This indicates that geographic, environmental, or population-specific factors may influence the vitamin D–MASLD relationship. In contrast, gender did not significantly moderate the association (β = −0.012 per 1% increase in female participants, p = 0.136). Overall, these findings suggest that study design and geographic region contribute meaningfully to heterogeneity in effect sizes, whereas participant sex appears to have minimal impact. Small-Study Effects/Publication bias Funnel plot inspection did not suggest marked asymmetry (Fig. 3). Egger’s regression intercept = − 3.29 (SE = 2.24), 95% CI [− 8.45, 1.88], p = 0.180 (two-tailed). Begg–Mazumdar rank correlation test yielded τ = −0.296, p = 0.213, indicating no significant small-study effects. Duval and Tweedie’s trim-and-fill procedure imputed 0 studies under both fixed- and random-effects models. The classic fail-safe N indicated that approximately 235 “null” studies would be required to raise the overall meta-analytic p-value > 0.05. Funnel plot Discussion This systematic review and meta-analysis demonstrates that individuals with type 2 diabetes mellitus (T2DM) who have metabolic-associated steatotic liver disease (MASLD) exhibit significantly lower circulating 25-hydroxyvitamin D [25(OH)D] concentrations than those without MASLD. The pooled standardized mean difference indicates a moderate, inverse association, suggesting that vitamin D insufficiency may be linked with hepatic steatosis in diabetic populations. Although the overall pattern of association was consistent across studies, there was substantial heterogeneity, with differences in study design and geographical region contributing significantly to the variation in effect sizes. This result highlights the complexity of vitamin D’s role in metabolic liver disease, supports the hypothesis that lower vitamin D status is associated with hepatic steatosis among individuals with T2DM, and reinforces the importance of considering study design and regional context when interpreting its association with MASLD in T2DM. Our findings are broadly consistent with prior observational studies reporting inverse associations between vitamin D status and MASLD in the study population. Several studies have documented lower serum 25(OH)D concentrations in patients with hepatic steatosis, independent of glycaemic control, body mass index, or insulin resistance [15, 30, 31]. However, the magnitude of the association has varied across studies, which may reflect differences in population characteristics, vitamin D assay methodology, MASLD diagnostic criteria, and regional variations in sun exposure, diet, and genetic predisposition. By quantitatively synthesizing these data, this meta-analysis provides a more precise estimate of the association and highlights the influence of study design and region on effect size. In contrast to observational findings, other studies investigating vitamin D supplementation in NAFLD/MASLD have generally produced null or inconsistent results. Notably, a randomized controlled trials (RCTs) reported no significant improvements in hepatic fat content, transaminases, CK-18 fragments, or metabolic endpoints in T2DM patients with MASLD [32]. Other meta-analyses of RCTs in MASLD populations have likewise found benefits for metabolic indices (e.g. ALT, insulin resistance) but not consistently for core hepatic outcomes [33]. Such inconsistencies suggest that observed cross-sectional associations may reflect correlation rather than causation, or that supplementation effects are contingent on baseline deficiency, dosage, duration, or other modifying factors. Substantial heterogeneity was observed across studies in the vitamin D–MASLD association, with study design and geographic region emerging as significant moderators. Cross-sectional and case–control studies often report stronger inverse associations than post-hoc analyses of trials, possibly as a result of residual confounding, reverse causation or differences in exposure and outcome ascertainment [34, 35]. Regional variation likewise appears to influence effect size, likely because of differences in sunlight exposure, dietary vitamin D intake, baseline vitamin D deficiency prevalence, and genetic or ethnic differences in vitamin D metabolism [15, 35]. Other plausible sources of variability include differences in MASLD/NAFLD diagnostic modality, vitamin D assay method, and season and latitude of sampling, which influence sun exposure and thus endogenous vitamin D production. Incomplete adjustment for lifestyle factors also likely contributes to heterogeneity; key confounders such as physical activity levels, detailed dietary intake, and the degree of sun exposure were not uniformly reported or controlled for across the included studies [15]. Notably, most included studies originated from East Asia (particularly China) and the Middle East, with only one from Europe (Russia) and none from sub-Saharan Africa, Western Europe, or South America. This regional concentration may partly explain the observed heterogeneity and highlights a significant evidence gap in underrepresented populations with differing sunlight exposure, dietary patterns, and genetic backgrounds. Importantly, some studies have failed to find any association between vitamin D and fibrosis stage which suggests that vitamin D relationships may be stronger for steatosis than for more advanced histopathology [36]. The wide prediction interval for future settings emphasizes that the inverse link between vitamin D and MASLD, while generally robust, is highly context-dependent and must be interpreted in light of study characteristics and local environmental or population-specific factors [35]. Clinically, the findings from this study suggest that in T2DM patients, lower vitamin D levels may serve as a marker of increased hepatic steatosis risk. Given the high prevalence of MASLD in diabetic populations and the relatively low cost and risk of vitamin D testing and supplementation, these results raise the question of whether routine vitamin D screening should be integrated into clinical practice for this high-risk group. Monitoring vitamin D status might be considered as part of comprehensive metabolic risk evaluation. Specifically, in settings with a high, documented prevalence of both MASLD and vitamin D insufficiency, screening could help identify a subgroup of patients who may benefit most from correcting deficiency, even if the primary benefit remains skeletal health. Such screening could also serve as an additional indicator, alongside metabolic-panel abnormalities, to raise clinical suspicion for underlying MASLD, potentially prompting earlier non-invasive liver assessment. However, current evidence does not yet support prescribing vitamin D solely to prevent or treat MASLD. From a public health perspective, strategies to improve vitamin D sufficiency through supplementation, food fortification, and safe sunlight exposure promotion remain plausible, low-risk adjuncts, especially in populations with documented deficiency. If future trials confirm hepatic benefits, such measures could offer synergistic gains in metabolic and liver health. Nonetheless, until more convincing interventional data emerge, established approaches like weight loss, glycaemic control, and lipid therapy should remain the mainstay of MASLD prevention and management in T2DM. This study is the first comprehensive meta-analysis focused specifically on adults with T2DM, a group particularly vulnerable to MASLD. The inclusion of studies with standardized vitamin D assays and validated diagnostic methods for MASLD strengthened internal validity, while rigorous procedures such as duplicate data extraction, ROBINS-I risk assessment, and multiple sensitivity and moderator analyses, enhanced reliability and interpretability of the findings. Nonetheless, some limitations should be acknowledged. Most included studies were observational and cross-sectional, limiting causal inference and allowing potential residual confounding. Significant heterogeneity persisted due to variations in study populations, diagnostic definitions, and assay methods. Data on vitamin D supplementation were scarce, precluding conclusions about therapeutic implications. Furthermore, the evidence base was geographically skewed, with most studies from Asia and the Middle East and very limited data from Africa and Latin America. This underrepresentation restricts the generalizability of findings and reinforces the need for further research in diverse populations and settings. Conclusions This systematic review and meta-analysis provides quantitative evidence that lower circulating 25-hydroxyvitamin D levels are significantly associated with the presence of MASLD among individuals with T2DM, supporting the hypothesis that vitamin D deficiency may contribute to hepatic steatosis in this high-risk population. However, the observed heterogeneity across studies, largely influenced by differences in study design, geographic region, and diagnostic approaches, emphasizes the complexity of this relationship. Although the findings highlight vitamin D status as a potential marker of metabolic and hepatic dysfunction, current evidence remains insufficient to establish a causal role or to recommend vitamin D supplementation as a therapeutic strategy for MASLD in T2DM. Further large-scale, well-designed longitudinal and interventional studies are warranted, particularly those using standardized vitamin D assays and validated MASLD diagnostic criteria, to clarify causal pathways, determine dose–response effects, and assess whether improving vitamin D status can meaningfully alter the natural history of MASLD in diabetic populations. Such future research should include diverse geographic and ethnic populations to address the current gaps in evidence and improve generalizability. Acknowledgements The authors thank the respective study teams whose data were included in this meta-analysis for making their work publicly available. No additional acknowledgments apply. Abbreviations Metabolic–associated steatotic liver disease Non–alcoholic fatty liver disease Metabolic–associated steatohepatitis Type 2 diabetes mellitus 25–hydroxyvitamin D Randomized controlled trial Enzyme–linked immunosorbent assay Chemiluminescence immunoassay Electrochemiluminescence immunoassay MS/MS–Liquid chromatography–tandem mass spectrometry Hepatic Steatosis Index Body mass index IR–Homeostatic Model Assessment of Insulin Resistance Alanine aminotransferase Aspartate aminotransferase Gamma–glutamyl transferase Triglycerides High–density lipoprotein Fasting blood glucose LFS–NAFLD–Liver Fat Score Preferred Reporting Items for Systematic Reviews and Meta–Analyses International Prospective Register of Systematic Reviews I–Risk Of Bias In Non–randomized Studies of Interventions Author contributions BB conceptualized the study and led the overall design. BB and ELU developed the protocol, performed the literature search, and conducted study selection. ELU performed the statistical analyses. WA and NIO contributed to study planning, interpretation of results, and critical revision of the manuscript. All authors read, reviewed, and approved the final manuscript. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Data availability All data generated or analyzed during this study are included in this published article. Additional data and extraction sheets are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This study is a systematic review and meta-analysis of previously published studies; no primary human or animal data were collected. Ethical approval was therefore not required. Informed consent Not applicable, as all data analyzed were derived from previously published studies. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data generated or analyzed during this study are included in this published article. Additional data and extraction sheets are available from the corresponding author on reasonable request.