Hepatic metabolic reprogramming after metabolic and bariatric surgery: Gene-nutrition interactions and precision nutrition

Abstract

Background

Metabolic dysfunction-associated steatotic liver disease (MASLD) is highly prevalent among individuals with obesity evaluated for metabolic and bariatric surgery (MBS) and may persist or recur despite substantial postoperative weight loss. This review examines the postoperative state as a dynamic model of hepatic metabolic reprogramming and gene–nutrition interaction, integrating surgical anatomy, dietary intake, nutrient absorption, metabolic response, weight trajectory, and inherited susceptibility.

Methods

This was a narrative clinical review informed by a structured literature search of MEDLINE (via PubMed), the Cochrane Library, and Epistemonikos (final search, August 3, 2026). Human evidence on hepatic outcomes after MBS, genetic determinants of liver disease, and postoperative nutritional exposures was prioritized; mechanistic and non-surgical human studies were considered only where direct postoperative evidence was limited.

Results

MBS rapidly reduces hepatic fat and improves insulin sensitivity, whereas resolution of metabolic dysfunction-associated steatohepatitis (MASH) and regression of fibrosis follow slower and partly dissociated trajectories. Genetic variants affecting hepatic lipid storage, lipid export, glucose-driven lipogenesis, and fibrogenesis may contribute to residual hepatic risk, but no study has jointly assessed genotype, repeated postoperative dietary exposure, and direct liver outcomes. Current evidence supports adequate protein intake, procedure-specific micronutrient supplementation, sustained dietary quality, metabolic risk control, alcohol assessment, and phenotype-guided liver surveillance regardless of genotype. Pharmacotherapy with demonstrated histological efficacy in MASH may be considered for residual MASH or weight recurrence, although it has not been evaluated specifically after MBS.

Conclusions

Genetic information may refine the interpretation of residual hepatic risk, but genotype-guided procedure selection, dietary patterns, macronutrient targets, or supplement doses are not supported. Prospective, adequately powered, multi-ancestry studies integrating longitudinal nutrition, genetic susceptibility, and direct hepatic outcomes are needed to determine whether genotype-informed nutrition improves outcomes beyond high-quality phenotype-guided care.

Graphical abstract

Conceptual framework of hepatic metabolic reprogramming after metabolic and bariatric surgery.Postoperative liver outcomes reflect the interplay between surgery-induced metabolic changes, evolving nutritional exposures, and genetic susceptibility. Although surgery generally improves hepatic steatosis, inflammation, insulin resistance, and other metabolic abnormalities, fibrosis improvement may be slower and variable, supporting phenotype-guided surveillance and individualized nutritional care. Abbreviations: GCKR, glucokinase regulator; HSD17B13, hydroxysteroid 17-beta dehydrogenase 13; MBOAT7, membrane bound O-acyltransferase domain containing 7; MTARC1, mitochondrial amidoxime reducing component 1; PNPLA3, patatin like phospholipase domain containing 3; TM6SF2, transmembrane 6 superfamily member 2.

1. Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly termed nonalcoholic fatty liver disease (NAFLD), spans a continuum from simple steatosis to metabolic dysfunction-associated steatohepatitis (MASH), progressive fibrosis, cirrhosis, and hepatocellular carcinoma [1], [2]. For consistency, MASLD and MASH are used throughout this review, including when describing studies that applied the earlier terminology [2]. Its close association with obesity, insulin resistance, type 2 diabetes, dyslipidemia, and hypertension makes MASLD highly relevant in candidates for metabolic and bariatric surgery (MBS) [1]. Because fibrosis severity is a major determinant of liver-related prognosis, the benefit of obesity treatment cannot be judged by weight loss alone [3], [4]. Current guidance recognizes MBS as a therapeutic option for appropriately selected patients with obesity and non-cirrhotic MASLD, while emphasizing fibrosis assessment and multidisciplinary follow-up [3], [5]. Beyond sustained weight reduction, MBS alters glucose homeostasis, adipose-tissue lipolysis, incretin signaling, bile-acid metabolism, appetite regulation, and the gut–liver axis [6], [7]. These changes are accompanied by marked improvements in steatosis and MASH, with fibrosis regression in a substantial proportion of patients [8], [9], [10], [11]. Yet hepatic recovery is not uniform: MASH may resolve relatively early, whereas fibrosis often improves more slowly and can remain incomplete years after surgery [4], [12], [13]. Observational studies further suggest that MBS is associated with lower risks of major adverse liver and cardiovascular outcomes in selected patients with fibrotic MASH and in carefully selected patients with compensated MASH-related cirrhosis [14], [15].

Substantial weight loss does not necessarily imply complete resolution of liver disease. Fibrosis may persist, remain stable, or occasionally progress despite favorable changes in body weight and metabolic status [16], [17], [18]. This heterogeneity may reflect baseline fibrosis, age, diabetes, durability of weight loss, procedure type, alcohol exposure, medication use, and postoperative dietary behavior [18], [19], [20], while the method and timing of liver assessment can influence the apparent trajectory of recovery [21]. Nutrition adds a further source of variation: gastric restriction, altered food tolerance, malabsorption, supplementation, and weight regain modify protein intake, carbohydrate and fat quality, and micronutrient status over time [22], [23]. Patients with similar weight loss may therefore follow different nutritional and hepatic trajectories.

Inherited susceptibility may explain part of this residual variability. MASLD is a polygenic disorder, and variants in patatin-like phospholipase domain-containing 3 (PNPLA3), transmembrane 6 superfamily member 2 (TM6SF2), glucokinase regulator (GCKR), membrane-bound O-acyltransferase domain-containing 7 (MBOAT7), hydroxysteroid 17-beta dehydrogenase 13 (HSD17B13), and mitochondrial amidoxime reducing component 1 (MTARC1) influence hepatic lipid storage, lipid export, de novo lipogenesis, inflammation, and fibrogenesis [24], [25], [26]. Their effects vary with adiposity, diabetes, alcohol exposure, and diet [27], [28], [29], [30]. MBS therefore provides an unusual setting in which inherited susceptibility is superimposed on rapid and sustained changes in energy intake, nutrient exposure, adipose-derived substrate flux, and gastrointestinal physiology [31], [32]. This creates a biological rationale for gene–nutrition interactions, particularly during the transition from rapid weight loss to weight stabilization and, in some patients, weight regain.

Despite this rationale, the evidence needed for clinical translation remains incomplete. Existing nutritional-genomics reviews have addressed genetic susceptibility to MASLD and nutrient–gene interactions across broad populations rather than the metabolic and nutritional environment created by MBS [33], [34]. Post-bariatric genetic studies have focused mainly on weight-loss outcomes, whereas nutritional studies rarely incorporate genotyping or direct hepatic endpoints [35], [36]. This review critically synthesizes evidence on how genetic susceptibility, surgically induced metabolic changes, postoperative nutrition, and weight trajectory may interact to influence liver outcomes after MBS, and evaluates their potential role in hepatic risk stratification, nutritional monitoring, and precision nutrition. One principle frames the review: genetic information may refine the interpretation of residual hepatic risk, but it does not currently justify genotype-specific procedure selection, dietary prescriptions, supplementation, or surveillance [3], [37]. This principle is stated here once and applied throughout.

2. Methods

This was a narrative clinical review informed by a structured literature search conducted in MEDLINE via PubMed, the Cochrane Library, and Epistemonikos, with the final search performed on August 3, 2026. No restrictions were applied by publication date or language. Search strategies combined controlled vocabulary, including Medical Subject Headings such as “Bariatric Surgery,” “Nutrigenomics,” “Genetic Predisposition to Disease,” and “Non-alcoholic Fatty Liver Disease,” with free-text terms for metabolic and bariatric surgery, MASLD, hepatic outcomes, postoperative nutrition, genetic variants, gene–nutrition interactions, and precision nutrition, adapted to each database. Reference lists of relevant articles were also screened. Two authors independently screened titles and abstracts and reviewed potentially relevant full texts; disagreements were resolved by consensus or by a third reviewer. Evidence was prioritized in the following order: human studies after MBS with direct hepatic endpoints; non-bariatric human studies; and mechanistic or experimental studies, which were used only to contextualize gaps in postoperative evidence and not as a basis for recommendations. Case reports, guidance superseded by updated versions, and published corrections without the original article were not used to support statements. Findings were synthesized narratively. The variant-level evidence summarized in Table 3 was rated by author consensus using prespecified criteria (study design, consistency, directness to the post-MBS setting, and availability of hepatic outcomes); these ratings reflect author judgment and do not constitute a formal certainty-of-evidence assessment such as GRADE. Where reported in the original studies, effect estimates are presented with 95% confidence intervals (CIs).

Table 3.

Variant-level grading of evidence relevant to precision nutrition and liver health after MBS.

Variant	Biological mechanism	MASLD evidence	Post-MBS hepatic evidence	Genotype × diet evidence	Current clinical relevance
PNPLA3 rs738409 C > G (p.I148 M; G risk allele)	Hepatic triglyceride retention through impaired hydrolysis, altered PUFA mobilization, and possible stellate-cell–mediated fibrogenesis [38], [39].	●●● Steatosis, MASH, fibrosis, cirrhosis, and liver-related events; effects amplified by metabolic and clinical risk factors [28].	● One prospective cohort (n = 84) showed greater 1-year liver-fat reduction in carriers; no genotype-specific data on postoperative diet, MASH, fibrosis, or liver-related events [40].	●● Observational interactions reported for diet quality, carbohydrate, fat quality, and selected nutrients, but findings are inconsistent and one RCT was null [27], [41], [42], [43], [44].	May refine liver-risk assessment; does not support a genotype-specific postoperative diet, procedure, or supplementation.
TM6SF2 rs58542926 C > T (p.E167K; T risk allele)	Reduced ApoB lipidation and VLDL secretion, promoting hepatic lipid retention [45], [46].	●●● Steatosis, progressive liver injury, and liver-related events [37].	● Evaluated in one small MBS cohort; did not independently predict 1-year liver-fat improvement; postoperative diet not assessed by genotype [40].	● Non-bariatric imaging data suggest interactions with Mediterranean diet adherence and selected food groups [27].	May refine hepatic-risk assessment; should not guide procedure choice, dietary-fat targets, or supplementation.
GCKR rs1260326 C > T (p.P446L; T risk allele)	Increased hepatic glycolytic flux and de novo lipogenesis, with lower fasting glucose but higher triglycerides and liver fat [47].	●● Associated with steatosis and glucose–lipid phenotypes; associations with advanced disease are heterogeneous [37].	○ No longitudinal bariatric evidence for postoperative steatosis, MASH, or fibrosis outcomes.	● Biological plausibility for carbohydrate/fructose interactions, but no validated bariatric or interventional evidence.	Research hypothesis only; no genotype-specific postoperative carbohydrate or fructose target.
MBOAT7 rs641738 C > T (T risk allele; noncoding)	Reduced expression alters phosphatidylinositol remodeling and lipid signaling, potentially promoting steatosis and fibrosis [25].	●● Associated with steatosis, inflammation, and fibrosis, with population and ancestry heterogeneity.	● Included in one small MBS cohort; did not independently predict 1-year liver-fat improvement [40].	● Non-bariatric treatment-response heterogeneity reported; no replicated diet-specific interaction.	Research hypothesis only; should not guide diet, supplementation, or surveillance.
HSD17B13 rs72613567 T > TA (protective loss-of-function variant)	Reduced hepatic HSD17B13 activity may attenuate inflammatory and fibrogenic pathways [25].	●● Associated with lower MASH and fibrosis risk, with variable effect magnitude.	○ No postoperative evidence for MASH resolution, fibrosis regression, or liver-related events.	○ No validated human genotype–diet interaction.	Risk-refinement/research relevance only; should not alter standard monitoring.
MTARC1 rs2642438 G > A (p.A165T; protective A allele)	Reduced hepatic MTARC1 abundance may favor lipid handling, triglyceride export, and fatty-acid oxidation [25].	●● Protective associations with steatosis, cirrhosis, and liver-related mortality; predictive value remains under study.	○ No longitudinal bariatric evidence.	○ No validated human genotype–diet interaction.	Research relevance only; no basis for modifying postoperative nutrition or surveillance.

Note: Evidence ratings were assigned by author consensus (author judgment) according to study design, consistency, directness to the post-MBS setting, and availability of hepatic outcomes; they do not constitute a formal certainty-of-evidence assessment. ●●●, consistent evidence from multiple human cohorts and/or longitudinal outcome studies; ●●, replicated human evidence with limitations related to heterogeneity, sample size, or outcomes; ●, isolated, indirect, or non-bariatric evidence; ○, no qualifying evidence identified among the studies reviewed. Post-MBS hepatic evidence refers to studies evaluating the genetic variant in relation to a direct hepatic outcome after MBS. No study identified simultaneously assessed genotype, postoperative nutritional exposure, and direct hepatic outcomes. Abbreviations: ApoB, apolipoprotein B; GCKR, glucokinase regulator; HSD17B13, hydroxysteroid 17-beta dehydrogenase 13; MASH, metabolic dysfunction-associated steatohepatitis; MASLD, metabolic dysfunction-associated steatotic liver disease; MBOAT7, membrane-bound O-acyltransferase domain-containing 7; MBS, metabolic and bariatric surgery; MTARC1, mitochondrial amidoxime reducing component 1; PNPLA3, patatin-like phospholipase domain-containing 3; PUFA, polyunsaturated fatty acid; RCT, randomized controlled trial; TM6SF2, transmembrane 6 superfamily member 2; VLDL, very-low-density lipoprotein.

3. Hepatic metabolic adaptations after MBS

3.1. Rapid changes in liver fat content, insulin sensitivity, and lipid flux

Hepatic adaptation begins before maximal weight loss has occurred [48]. In the Oseberg randomized trial of 100 patients with obesity and type 2 diabetes, liver fat fraction declined similarly after sleeve gastrectomy (SG; −19.7%; 95% CI, −22.5% to −16.9%) and Roux-en-Y gastric bypass (RYGB; −21.5%; 95% CI, −24.3% to −18.6%) at 1 year, and 94% and 100% of patients, respectively, had no or low-grade steatosis. In contrast, the enhanced liver fibrosis score category remained stable in 77% of patients and worsened in 18%, illustrating that hepatic triglyceride depletion precedes extracellular-matrix remodeling [49]. Similarly, in nine patients undergoing SG, the magnetic resonance imaging (MRI)-derived proton density fat fraction fell from 14.1% to 8.9% within 1 month, whereas liver stiffness did not change during 6 months [50].

Much of this early improvement reflects the energy deficit and weight loss shared by different operations. Two months after SG or RYGB (n = 16), two-step hyperinsulinemic–euglycemic clamps showed comparable improvements in hepatic, muscle, and adipose-tissue insulin sensitivity [51]. In a matched prospective cohort of 22 patients with type 2 diabetes who achieved approximately 18% weight loss by RYGB or diet alone, improvements in intrahepatic triglyceride content, 24-h glucose, free-fatty-acid and insulin profiles, β-cell function, and multiorgan insulin sensitivity did not differ between groups [52]. Surgery therefore remains clinically superior mainly because it achieves larger and more durable weight loss.

Anatomy-dependent signals may add to these weight-loss-mediated effects. One month after SG, an increase in circulating neuregulin-4 was associated with reduced hepatic steatosis before substantial weight loss; the supporting mechanistic data derive largely from rodent and cell models [53]. In 21 patients with paired liver biopsies, hepatic oxidative-phosphorylation capacity and mitochondrial DNA content increased 12 months after surgery, indicating that hepatic adaptation also involves slower mitochondrial remodeling [54].

3.2. Alterations in bile-acid signaling and the gut-liver axis

After MBS, bile acids act as endocrine signals through the farnesoid X receptor, G protein-coupled bile acid receptor 1, and fibroblast growth factor 19 (FGF19), influencing hepatic lipogenesis, gluconeogenesis, inflammation, and glucagon-like peptide-1 (GLP-1) secretion [55]. This response is procedure-specific. RYGB increased basal bile flow, bile-acid retention, and fasting plasma bile acids, with a brief postprandial bile-acid rise and lower FGF19 release, whereas after SG bile flow and plasma bile acids were largely comparable to those of non-operated controls despite accelerated gastric emptying [56]. The timing of intestinal bile-acid exposure may therefore matter as much as the size of the circulating pool.

In a subanalysis of the Oseberg trial (RYGB, n = 39; SG, n = 38), both operations altered the fecal microbiome, with larger changes after RYGB; type 2 diabetes remission was associated with greater microbial gene richness and fermentation and butyrate-production potential, independently of weight loss [57]. Because diet, medication withdrawal, weight loss, and altered transit simultaneously reshape the microbiome, such associations should not be interpreted as causal for hepatic improvement [58], [59].

3.3. Differences between restrictive and malabsorptive procedures

The conventional distinction between “restrictive” and “malabsorptive” operations is biologically incomplete. SG preserves intestinal continuity but alters gastric emptying, nutrient sensing, and incretin release, whereas RYGB excludes the duodenum and proximal jejunum, producing faster distal nutrient delivery and larger postprandial incretin responses; biliopancreatic diversion with duodenal switch and single-anastomosis duodeno-ileal procedures generate a stronger hypoabsorptive state [60]. Hepatic effects therefore lie along a continuum determined by anatomy, nutrient flow, energy balance, and endocrine signaling.

Current comparative evidence does not establish hepatic superiority of RYGB over SG. In paired biopsies from 40 patients, both procedures reduced histological activity to a similar extent at 12 months, RYGB produced a greater reduction in steatosis, and fibrosis declined non-significantly [61]. A meta-analysis of three randomized trials (322 patients) found that both operations reduced advanced fibrosis, with a non-significant advantage for RYGB in fibrosis improvement (odds ratio [OR], 1.63; 95% CI, 0.94–2.81) [62]. In the BRAVES randomized trial, MASH resolution without fibrosis worsening at 1 year occurred in 56% of patients after RYGB and 57% after SG, compared with 16% after lifestyle modification plus best medical care (relative probability versus lifestyle modification, 3.60 [95% CI, 2.19–5.92] and 3.67 [95% CI, 2.23–6.02], respectively) [9].

Greater malabsorption may intensify weight loss but introduces distinct liver-related risks. Protein–energy malnutrition, bacterial overgrowth, increased intestinal permeability, and endotoxemia may promote liver injury, and the extent of intestinal bypass shows a dose-dependent relationship with progressive liver injury, particularly after biliopancreatic diversion or procedures with very short common channels [63], [64]. These uncommon but severe outcomes show why hepatic benefit cannot be inferred solely from the magnitude of postoperative weight loss.

3.4. MBS as a physiological model of hepatic metabolic reprogramming

MBS acts as a controlled physiological perturbation in which caloric intake, gastrointestinal anatomy, enteroendocrine signaling, bile flow, microbial ecology, and hepatic substrate availability change over different timescales [6]. The discordant kinetics of steatosis clearance, insulin-sensitivity restoration, mitochondrial adaptation, and fibrosis regression indicate that hepatic recovery comprises several biologically distinct processes rather than a single weight-dependent pathway [13], [49], [54]. Human MBS cohorts combining intraoperative liver, adipose, and jejunal sampling with longitudinal multi-omic phenotyping offer a platform for stratifying these responses by genotype and postoperative nutrition [65].

As summarized in Fig. 1, rapid energy restriction and metabolic improvement initially favor hepatic recovery, whereas long-term outcomes increasingly reflect dietary intake, nutritional adequacy, weight trajectory, and recurrence of metabolic disease. Steatosis, MASH activity, and fibrosis follow partly distinct temporal trajectories, and genotype-dependent modification of these phases remains a hypothesis requiring prospective post-MBS validation.

Fig. 1.

Time-dependent gene–nutrition–liver interactions after MBS.

MBS produces a dynamic sequence of metabolic, nutritional, and hepatic changes. During rapid weight loss, marked energy restriction, improved insulin sensitivity, reduced hepatic lipid accumulation, and surgery-related changes in bile acids, gut hormones, and the microbiome favor early improvement in steatosis and metabolic dysfunction-associated steatohepatitis activity. As weight and food intake stabilize, dietary quality, protein and micronutrient adequacy, supplementation, alcohol exposure, and metabolic trajectory become increasingly relevant. During long-term maintenance or weight regain, refined carbohydrate and saturated-fat intake, alcohol use, diabetes recurrence, and deteriorating diet quality may restore hepatic metabolic pressure. Genetic susceptibility related to hepatic lipid handling and fibrogenesis may influence residual or recurrent liver disease across these phases; however, direct postoperative gene–nutrition interactions remain unproven. Hepatic outcomes follow different temporal trajectories: steatosis may improve rapidly but recur, steatohepatitis activity may resolve earlier, whereas fibrosis regression is generally slower and may remain incomplete. Solid arrows represent relationships supported by post-MBS human evidence; dashed arrows represent proposed interactions requiring prospective validation. Phase durations are approximate and may vary according to procedure and patient characteristics [9], [13], [18], [23], [27], [41]. Abbreviations: GCKR, glucokinase regulator; MASH, metabolic dysfunction-associated steatohepatitis; MBOAT7, membrane-bound O-acyltransferase domain-containing 7; MBS, metabolic and bariatric surgery; PNPLA3, patatin-like phospholipase domain-containing 3; TM6SF2, transmembrane 6 superfamily member 2. Created with BioRender.com.

4. Genetic modifiers of liver outcomes following MBS

4.1. Evidence for genotype-dependent variability in MASLD resolution and recurrence

The hepatic response to MBS reflects the interaction between a major metabolic perturbation and a liver whose capacity for lipid export, triglyceride remodeling, inflammation, and fibrogenesis is partly inherited [25]. Genetic variation is unlikely to determine a binary response; instead, it may modify baseline severity, the magnitude of steatosis clearance, residual fibrosis, and susceptibility to recurrence. Postoperative heterogeneity should therefore be interpreted as overlapping steatotic, inflammatory, and fibrotic phenotypes rather than a single category of “resolution” [26].

In a prospective cohort of 84 patients, carriers of the PNPLA3 rs738409-G allele (p.I148 M) lost more weight and more liver fat 1 year after MBS than noncarriers, and PNPLA3 genotype independently predicted steatosis improvement, whereas TM6SF2 rs58542926-T (p.E167K) and MBOAT7 rs641738-T did not [40]; this cohort was also identified in a systematic review of genetic variants and MBS outcomes [35]. This apparently favorable response is not protection: carriers often begin with more hepatic fat, and residual risk depends on postoperative fat, fibrosis, weight trajectory, and metabolic control. Small samples, heterogeneous operations, variable endpoints, and limited replication preclude a reliable genotype-specific estimate of surgical benefit.

Genotype may also dissociate biochemical from histological recovery. Outside MBS, a systematic review found moderate-certainty evidence that PNPLA3 148 M homozygotes benefit less from omega-3 carboxylic acids and low-certainty evidence that they respond more to lifestyle interventions for liver-fat reduction [66]. Whether genotype predicts MASH resolution, fibrosis regression, or liver-related events after MBS independently of weight loss has not been established [28], [66]. A genotype-associated change in MRI-measured steatosis therefore cannot be equated with modification of long-term hepatic risk.

Recurrence may be particularly sensitive to gene–environment interaction because surgery does not remove inherited susceptibility. In a natural experiment involving 357 patients with MASLD during the coronavirus disease 2019 (COVID-19) lockdown, the PNPLA3 GG genotype was independently associated with weight gain, which in turn was associated with worsening glycemia and aminotransferase levels [67]. In the UK Biobank, genetic risk amplified the associations between diet and liver fat [27]. Neither study was conducted after MBS, and no genotype-specific postoperative surveillance schedule has been validated.

Inherited effects on weight regulation may indirectly modify recurrence. Variants affecting appetite, incretin signaling, adipocyte biology, and energy expenditure have been associated with weight or metabolic outcomes after MBS, although findings for FTO and MC4R are inconsistent [36], [68]. In a cohort of 375 patients followed for 6 years, a multivariant score was associated with lower weight loss and greater weight regain [69]. Such variants do not establish recurrent MASLD directly, but predisposition to lesser weight loss or to regain may recreate conditions in which liver-specific alleles become relevant.

4.2. Key genetic variants influencing postoperative hepatic response

The PNPLA3 rs738409 C > G variant (p.I148 M; G risk allele) alters lipid-droplet metabolism. PNPLA3 is nutritionally regulated and accumulates on hepatocyte lipid droplets under lipogenic conditions; the wild-type protein mobilizes polyunsaturated fatty acids from triglycerides and facilitates the secretion of large very-low-density lipoprotein (VLDL) particles, functions that are impaired by the 148 M variant [38]. Experimental data further indicate that the 148 M protein inhibits adipose triglyceride lipase-mediated lipolysis [39]. Postoperative caloric restriction may suppress the conditions that expose this defect, plausibly explaining early liver-fat loss among carriers, whereas refeeding, weight regain, or high-sugar intake may re-expose it; these postoperative effects remain hypotheses [38], [39].

The TM6SF2 rs58542926 C > T variant (p.E167K; T risk allele) impairs the lipidation of apolipoprotein B (ApoB)-containing particles and limits hepatic triglyceride export; carriers have lower circulating triglycerides and lower cardiovascular risk but greater liver risk [45], [46], [70]. Low serum triglycerides should therefore not be interpreted as low postoperative liver risk. Bariatric studies have not demonstrated TM6SF2-dependent liver-fat resolution [40].

The GCKR rs1260326 C > T variant (p.P446L; T risk allele) weakens glucokinase regulatory protein inhibition, increasing glycolytic flux and substrate supply for de novo lipogenesis, so that lower fasting glucose may coexist with higher triglycerides and hepatic fat [47]. Reduced postoperative carbohydrate delivery may mute this effect, whereas later refined-carbohydrate or fructose exposure could reactivate it. No adequately powered MBS cohort has tested whether GCKR predicts postoperative steatosis recurrence [71].

Additional loci show that fat accumulation and tissue injury are genetically separable: MBOAT7 rs641738 C > T is associated with liver fat, inflammation, and fibrosis, whereas the HSD17B13 rs72613567 T > TA loss-of-function variant and the MTARC1 rs2642438 G > A (p.A165T) variant are protective [25]. In 298 patients with MASLD, including 150 from a bariatric surgery unit, a weighted polygenic score combining PNPLA3, TM6SF2, MBOAT7, and GCKR correlated with aminotransferases and the fibrosis-4 index (FIB-4) [72]. This cross-sectional evidence favors polygenic over single-allele interpretation but does not address the postoperative response.

Risk alleles should not be combined into an unweighted count because their effects on hepatic triglyceride retention, circulating lipoproteins, glucose homeostasis, and cardiovascular risk differ. Jamialahmadi et al. partitioned variants according to the concordance between liver fat and circulating triglycerides, identifying liver-specific and systemic genetic profiles [73]. Independently, Raverdy et al. identified comparable clinical subtypes using six routinely available variables and observed distinct long-term trajectories despite similar baseline liver severity [74] (Table 1). After MBS, cardiometabolic remission could therefore coexist with persistent liver-specific susceptibility, a hypothesis that requires evaluation with direct hepatic endpoints.

Table 1.

Mechanistic and clinical characteristics of liver-specific and cardiometabolic MASLD phenotypes and their potential implications following MBS.

Dimension	Genetic partitioning	Clinical clustering
Study design and population	An adiposity-adjusted genome-wide analysis using liver imaging data from 36,394 participants identified 27 previously unknown loci, six of which were replicated in 3903 participants from four independent cohorts [73]	Clustering based on age, BMI, ALT, HbA1c, LDL cholesterol, and TG was performed in 1389 individuals with obesity. The clusters were reproduced in three biopsy cohorts comprising 1099 participants and evaluated longitudinally in the UK Biobank [74]
Identification of subtypes	A discordant pPRS comprising 10 variants associated increased hepatic fat with lower or nonparallel circulating TG and was interpreted as liver-specific. A concordant pPRS comprising 13 variants associated parallel increases in hepatic and circulating TG and represented systemic MASLD [73]	Among six clusters, cluster 5 was classified as liver-specific because of higher ALT at a younger age, whereas cluster 2 was classified as cardiometabolic because of marked dysglycemia, hypertriglyceridemia, hypertension, and dyslipidemia [74]
Biological basis	The discordant score included PNPLA3 and TM6SF2 variants and was enriched in lipid and TG homeostasis pathways, consistent with impaired VLDL secretion and hepatic lipid retention. The concordant score was enriched in insulin-receptor signaling and glucose-homeostasis pathways, consistent with increased substrate uptake or synthesis and reduced β-oxidation [73]	The liver-specific cluster was enriched in PNPLA3 rs738409-G (p.I148 M) and the PRS-HFC and showed transcriptomic activation of lipid-droplet homeostasis, intrahepatic lipid transport, inflammatory, and wound-healing processes. The cardiometabolic cluster showed increased cholesterol synthesis and glycolysis and a metabolomic profile reflecting dysglycemia and systemic metabolic dysfunction [74]
Hepatic outcomes	Both pPRSs were associated with MASLD and HCC, but hepatic associations were stronger for the discordant, liver-specific score [73]	The subtypes had similarly severe liver phenotypes at baseline. Compared with the control cluster, adjusted HRs for incident chronic liver disease were 4.52 (95% CI, 3.88–5.26) for the liver-specific cluster and 4.04 (95% CI, 3.50–4.66) for the cardiometabolic cluster [74]
Extrahepatic outcomes	The discordant pPRS was associated with reduced cardiovascular risk, whereas the concordant pPRS increased the risks of CVD, heart failure, hypertension, and diabetes. Both scores increased diabetes susceptibility, although the association was stronger for the concordant score [73]	Compared with the control cluster, adjusted HRs for CVD were 1.18 (95% CI, 1.07–1.31) for liver-specific and 1.80 (95% CI, 1.68–1.93) for cardiometabolic MASLD; corresponding HRs for T2D were 2.91 (95% CI, 2.62–3.23) and 6.82 (95% CI, 6.01–7.73), respectively [74]

Abbreviations: ALT, alanine aminotransferase; BMI, body mass index; CI, confidence interval; CVD, cardiovascular disease; HbA1c, glycated hemoglobin; HCC, hepatocellular carcinoma; HR, hazard ratio; LDL, low-density lipoprotein; MASLD, metabolic dysfunction-associated steatotic liver disease; MBS, metabolic and bariatric surgery; PNPLA3, patatin-like phospholipase domain-containing 3; pPRS, partitioned polygenic risk score; PRS-HFC, polygenic risk score for hepatic fat content; T2D, type 2 diabetes; TG, triglycerides; TM6SF2, transmembrane 6 superfamily member 2; VLDL, very-low-density lipoprotein.

Note: Jamialahmadi et al. provided the genetic and mechanistic classification, whereas Raverdy et al. provided the clinical, histological, molecular, and longitudinal characterization. Neither study directly assessed differential outcomes after MBS.

4.3. Clinical relevance of genetic heterogeneity in post-MBS care

The immediate clinical value of genetic heterogeneity lies in risk refinement, not surgical candidacy. A high-risk genotype should not lead to withholding an intervention that can markedly reduce hepatic and cardiometabolic stress. Surgical candidacy and follow-up should continue to be based on liver phenotype, fibrosis stage, metabolic disease, and weight trajectory [3]. In longitudinal MASLD cohorts, PNPLA3, TM6SF2, and GCKR genotypes contributed to liver-related events [28], [37]. Among 2075 adults with biopsy-confirmed MASLD, the PNPLA3 G allele was associated with major adverse liver outcomes (adjusted subdistribution hazard ratio, 1.4; 95% CI, 1.07–1.8), but advanced fibrosis carried a much larger risk (7.8; 95% CI, 4.4–13.8) [28]. Genotype is therefore most informative when combined with age, diabetes, sex, and fibrosis.

Sequential assessment with FIB-4 and elastography remains the recommended approach [3], [5]. In 1304 patients with biopsy-proven MASLD, a two-step strategy combining FIB-4 with PNPLA3 and TM6SF2 genotypes separated groups with 10-year cumulative liver-event rates of 2.1%, 28.9%, and 53.5% [37]. This study was not conducted after MBS. Rapid weight loss changes aminotransferases, platelets, and body composition and therefore alters the performance of non-invasive tests [21]. Fixed genetic information should complement, not replace, these dynamic markers.

The 2024 European MASLD guideline allows the consideration of genetic risk profiles in specialized centers, while sequential non-invasive fibrosis assessment, metabolic risk control, dietary quality, and individualized surveillance remain the foundation of management [3]. A systematic review of genetic variants and MBS outcomes (52 of 1572 screened studies) found that most studies assessed weight loss, few examined metabolic or hepatic outcomes, and findings for individual variants were often inconsistent [35].

Clinical use would also require ancestry-aware interpretation because allele frequencies, linkage disequilibrium, and score performance vary across populations [75], [76].

5. Post-MBS nutrition as a determinant of genotype-dependent liver health

5.1. Protein adequacy and micronutrient status in hepatic recovery

Post-MBS hepatic recovery occurs during profound energy restriction and loss of fat-free mass [77]. Adequate protein intake helps preserve lean mass during major weight loss [78], [79]. However, intake is frequently insufficient. In a prospective RYGB cohort (n = 28), protein intake fell from 80 g/day before surgery to 45 g/day at 1 month, and only 25% of participants reached 60 g/day by 6 months [80]. Among 133 patients in China, mean intake remained below recommendations at 6 and 12 months (53.5 and 56.4 g/day), the proportion meeting targets declined from 59.4% to 53.4%, and deficiencies of 25-hydroxyvitamin D (91.7%), ferritin (27.1%), albumin (14.3%), and hemoglobin (12.8%) were common at 12 months [81]. Neither cohort assessed liver outcomes or genotype.

In a 3-week preoperative hypocaloric trial (n = 19), a high-protein diet reduced intrahepatic lipid by 42.6%, whereas a low-protein diet produced no significant change despite similar weight loss [82]. After surgery, a meta-analysis of eight randomized trials found that protein intake above recommendations increased weight loss (mean difference, −4.95 kg; 95% CI, −9.41 to −0.49) and fat-mass loss (−7.64 kg; 95% CI, −14.01 to −1.28) but preserved fat-free mass only after SG (6.52 kg; 95% CI, 0.99–12.02) [83]. Evidence therefore supports meeting established requirements rather than indiscriminate high-dose supplementation; no genotype-specific protein target has been validated.

Severe protein–energy malnutrition can turn a hepatoprotective intervention into a cause of liver injury. Acute liver injury or failure has been described after bypass procedures when inadequate intake, malabsorption, rapid weight loss, and intestinal-barrier dysfunction coexist [64], [84]. Falling albumin, edema, muscle wasting, diarrhea, or unexplained aminotransferase elevation therefore warrant prompt nutritional and surgical evaluation [84], [85].

Micronutrient vulnerability varies by procedure and time. In an Australian cohort (gastric bypass, n = 66; SG, n = 144), deficiencies of thiamine (21% versus 6% preoperatively), vitamin A (21% versus 3%), and selenium (59% versus 39%) increased within 1–3 months after gastric bypass, folate deficiency emerged 6 months after SG (OR, 13; 95% CI, 2–84), and inflammation was independently associated with several deficiencies, complicating biomarker interpretation [22]. In a 12-week randomized trial after one-anastomosis gastric bypass (n = 60), a multistrain probiotic and micronutrient mixture did not change the primary outcome, alanine aminotransferase, although aspartate aminotransferase and a non-invasive fibrosis score improved [86].

Genotype-specific micronutrient requirements remain unproven. PNPLA3 participates in hepatic stellate-cell retinoid metabolism, which makes vitamin A status biologically relevant but does not justify genotype-guided retinol supplementation [87]. Among 452 adults with biopsy-confirmed MASLD, intakes of choline, methionine, and n-3 polyunsaturated fatty acids were inversely associated with significant fibrosis mainly in PNPLA3 G-allele carriers [42]; this cross-sectional, non-bariatric evidence does not support a postoperative supplementation strategy.

5.2. Carbohydrate quality, fructose exposure, and liver fat reaccumulation

Early carbohydrate restriction does not guarantee sustained dietary quality. Six months after surgery (n = 107), absolute energy intake decreased, but the energy contribution of mono- and disaccharides increased (+4.2 percentage points) and vegetable and whole-grain intakes declined [23]. At 12 months, marked reductions in energy and sucrose intake coexisted with lower intakes of fiber and of eicosapentaenoic plus docosahexaenoic acid [88]. Between 6 and 12 months, the carbohydrate contribution to energy rose and the protein contribution fell [81]. Food tolerance also changes after surgery and restricts the food groups that patients can consume [89]. None of these studies assessed genotype or liver outcomes.

Fructose is relevant to recurrent steatosis because hepatic fructolysis bypasses a major glycolytic control step and, particularly when it adds energy, activates carbohydrate-responsive element-binding protein and de novo lipogenesis [90]. Surgery initially suppresses these pathways through energy restriction and improved insulin sensitivity, but frequent sugar-sweetened beverages or sucrose-rich foods may restore lipogenic signaling during weight stabilization [90]. This supports established advice to limit added sugars [3] but not genotype-specific carbohydrate limits.

PNPLA3 and GCKR may modify the hepatic response to renewed carbohydrate exposure [25]. However, a study of 288 Mexican-origin adults found no significant interaction between PNPLA3 and dietary intake on steatosis [43], and no postoperative study has validated genotype-specific carbohydrate thresholds.

Liver-fat reaccumulation after MBS may reflect weight regain, recurrent diabetes or insulin resistance, alcohol intake, and physical inactivity. In a paired-biopsy cohort, 15 of 18 patients (83.3%) with preoperative MASH achieved resolution without fibrosis worsening, together with improvements in insulin resistance and leptin, but persistence after weight regain was not assessed [91]. A systematic review found no consistent association between postoperative macronutrient composition and long-term weight outcomes [92]. Reassessing carbohydrate sources, liquid sugars, fiber, energy intake, alcohol, and weight trajectory is therefore more informative than relying on broad macronutrient categories.

5.3. Dietary fat composition and genotype-specific lipid handling

Fat remains relevant after bypass procedures. Under a controlled diet, fat absorption after RYGB was only moderately lower than in matched controls (88% versus 93%), although postprandial chylomicron and triglyceride kinetics were markedly altered [93]. Lower circulating triglycerides thus do not indicate limited fat delivery. Within an energy-appropriate diet, the emphasis should be on fat quality, favoring mono- and polyunsaturated over saturated fats [94], [95].

In a 12-week weight-maintaining trial in 98 older men homozygous for PNPLA3 rs738409, a diet following fat-quality recommendations reduced MRI-measured liver fat in both CC and GG groups, whereas an average Finnish diet increased it; hepatic triglyceride concentration measured by spectroscopy rose only in GG participants receiving the average diet (genotype × time interaction, p = 0.027) [41]. The short, non-bariatric design precludes a genotype-specific postoperative fat prescription.

TM6SF2 carriers may show favorable plasma triglycerides despite impaired hepatic lipid export, whereas GCKR effects appear more closely related to carbohydrate availability than to fat intake [46], [47]. Bariatric cohorts have not tested whether fat recommendations should differ by either variant.

Long-chain n-3 polyunsaturated fatty acids suppress de novo lipogenesis and favor fatty-acid oxidation: in 38 healthy men, 4 g/day of eicosapentaenoic plus docosahexaenoic acid for 8 weeks reduced intrahepatic triglycerides and hepatic lipogenesis, although plasma glucose increased [96]. Including n-3-rich foods within a balanced postoperative diet is appropriate, but genotype-guided supplementation is not supported [42], [66].

5.4. Interaction between diet composition and genetic susceptibility post-surgery

Surgery changes both sides of a gene–diet interaction, reducing energy intake, adipose lipolysis, insulin, and hepatic substrate flux while altering absorption, bile acids, transit, tolerance, and the microbiota [97]. Genetic liability may therefore be least visible during rapid weight loss and re-emerge after stabilization or regain, a hypothesis not yet tested. In the UK Biobank (n = 21,619), healthier dietary patterns were associated with lower MRI-measured liver fat and inflammation/fibrosis, and dietary effects on liver fat were up to 3.8-fold larger in PNPLA3 GG than in CC individuals [27]. High genetic risk may therefore increase the potential benefit of sustained diet quality rather than make dietary treatment futile.

As summarized in Table 2, human evidence does not support PNPLA3-specific postoperative diets. A cross-sectional study of Mexican-origin adults [43] and a 12-week randomized trial comparing hypocaloric Mediterranean and low-fat diets in 250 adults with MASLD [44] found no significant PNPLA3–diet interaction.

Table 2.

Evidence hierarchy for genotype–nutrition interactions relevant to liver health after MBS.

Evidence setting	Study	Design and population	Genetic factor; nutritional exposure	Main finding (interaction tested)	Relevance after MBS
Post-MBS, indirect	Zafirovska et al. [35]	Systematic review; 52 studies; adults undergoing MBS	Multiple variants, including PNPLA3, TM6SF2, and MBOAT7; diet not analyzed by genotype	Mostly weight-loss outcomes; few hepatic outcomes (not evaluated)	Genetic variation studied after MBS, but no genotype–nutrition–liver evidence
Post-MBS, indirect	Heusschen et al. [23]	Cohort; n = 107 (RYGB 87, SG 20); baseline and 6 months	Not assessed; energy, nutrients, food groups, guideline adherence	Diet quality worsened in several domains despite lower energy intake; no hepatic endpoint (not evaluated)	Caloric restriction does not ensure diet quality
Post-MBS, indirect	Li et al. [81]	Cohort; n = 133 (SG or gastric bypass); 6 and 12 months	Not assessed; nutrients, protein-powder use, biomarkers	Protein inadequacy and deficiencies persisted; no hepatic endpoint (not evaluated)	Supports repeated nutritional assessment
Non-bariatric, observational	Vilar-Gomez et al. [42]	Cross-sectional; n = 452; biopsy-confirmed MASLD	PNPLA3 rs738409; carbohydrate, n-3 PUFA, isoflavones, methionine, choline	Diet–fibrosis associations significant mainly in G-allele carriers (yes)	Plausible effect modification; not postoperative
Non-bariatric, observational	Morrill et al. [43]	Cross-sectional; n = 288 Mexican-origin adults; CAP	PNPLA3 rs738409; sugars, fructose, PUFA, fiber	No diet–steatosis association or interaction (yes, null)	Inconsistent replication
Non-bariatric, observational	Oh et al. [98]	Case–control; 2950 cases and 12,907 controls; KoGES	PNPLA3 rs738409; macro- and micronutrients, sodium, kimchi	Sodium and kimchi interactions (yes)	Population-specific; hypothesis-generating
Non-bariatric, observational	Chen et al. [27]	UK Biobank; n = 21,619; liver MRI	PNPLA3, TM6SF2, 16-variant steatosis PRS; Mediterranean diet, food groups	Dietary effects on liver fat up to 3.8-fold larger in PNPLA3 GG than CC (yes)	Genetic modification of diet–liver associations
Dietary RCT	Dogay Us et al. [44]	RCT; n = 250 adults with MASLD; 12 weeks	PNPLA3 rs738409; hypocaloric Mediterranean vs low-fat diet	Both diets reduced CAP and LSM; no interaction (yes, null)	Does not support diet selection by PNPLA3
Dietary RCT	Tauriainen et al. [41]	RCT; n = 98 older men (CC vs GG); 12 weeks	PNPLA3 rs738409; recommended vs average dietary-fat quality	Liver TG increased only in GG on the average diet (yes)	Biological plausibility of genotype-dependent response to fat quality

Abbreviations: CAP, controlled attenuation parameter; CC and GG, homozygous genotypes at PNPLA3 rs738409; KoGES, Korean Genome and Epidemiology Study; LSM, liver stiffness measurement; MASLD, metabolic dysfunction-associated steatotic liver disease; MBOAT7, membrane-bound O-acyltransferase domain-containing 7; MBS, metabolic and bariatric surgery; MRI, magnetic resonance imaging; n-3 PUFA, omega-3 polyunsaturated fatty acids; PNPLA3, patatin-like phospholipase domain-containing 3; PRS, polygenic risk score; PUFA, polyunsaturated fatty acids; RCT, randomized controlled trial; RYGB, Roux-en-Y gastric bypass; SG, sleeve gastrectomy; TG, triglycerides; TM6SF2, transmembrane 6 superfamily member 2.

Note: “Interaction tested” indicates whether a formal interaction or effect-modification analysis between genetic variation and a nutritional exposure or intervention was performed; “yes, null” indicates that the interaction was tested but was not statistically significant. No study was identified that simultaneously evaluated genotype, postoperative nutritional exposure, and longitudinal hepatic outcomes after MBS; all studies therefore provide indirect evidence.

6. Precision nutrition considerations for liver health after MBS

Recommendations are classified as evidence-based standard care, potential genotype-informed considerations, or research hypotheses. Table 3 summarizes, by variant, the strength and directness of the evidence and its current clinical applicability; the ratings reflect author judgment based on the prespecified criteria described in Section 2.

6.1. Tailoring dietary recommendations based on genetic risk profiles

Precision nutrition after MBS should integrate inherited susceptibility with the surgical procedure, liver phenotype, dietary tolerance, metabolic response, and weight trajectory [99]. Adequate protein intake, lifelong procedure-specific micronutrient supplementation, limited added sugars, high-quality dietary fats, and nutrient-dense foods remain standard care for all patients, regardless of genotype [79], [100], [101], [102].

For PNPLA3 148 M carriers, sustained energy balance, prevention of weight regain, limited sugar-sweetened beverages and refined carbohydrates, and fiber-rich plant foods are appropriate, but these recommendations are not specific to carriers [103]. Studies of overall diet and fat quality have produced inconsistent and population-specific findings [27], [41], [98]; greater emphasis on established fat-quality advice may be reasonable, whereas a genotype-specific fat target remains unvalidated.

For TM6SF2 and GCKR, the distinct vulnerabilities described in Section 4.2—impaired lipid export and glucose-driven lipogenesis, respectively—support cautious interpretation of favorable plasma triglycerides in TM6SF2 carriers and emphasis on glycemic control and carbohydrate quality in GCKR carriers, but they do not establish separate macronutrient prescriptions [26], [71].

Because dietary trials have not demonstrated clinically meaningful genotype–diet interactions, genotype should not determine the choice between Mediterranean, low-fat, or other dietary patterns [44]. Recommendations should instead be adapted to postoperative capacity, gastrointestinal tolerance, cultural preferences, and cardiometabolic needs [104], [105].

6.2. Identifying patients requiring intensified nutritional monitoring

Preoperative MASH, significant fibrosis, type 2 diabetes, and older age should remain the primary determinants of long-term liver surveillance after MBS [3], [18], [28], [106]. Successful weight loss should not end surveillance in these patients: in a biopsy cohort with advanced MASH, advanced fibrosis persisted in 47% of patients with baseline advanced fibrosis at a mean of 6 years after surgery, despite MASH resolution [18]. Known high-risk PNPLA3 or TM6SF2 genotypes may be considered additional risk modifiers, not independent indications for intensified monitoring [37].

Procedure type is an established determinant of nutritional risk. RYGB, biliopancreatic diversion, duodenal switch, and other hypoabsorptive procedures require closer assessment [85], [102]. Among patients with a body mass index (BMI) ≥50 kg/m², malnutrition occurred in 8.3% after duodenal switch versus 1.2% after RYGB (OR, 5.53; 95% CI, 1.35–22.44) [107]. Pre-existing deficiencies should be corrected before surgery, and monitoring should be intensified in patients with longer bypass limbs, chronic diarrhea, poor supplement adherence, or previous deficiencies [22], [102].

Inadequate protein intake is another indication for closer follow-up. Assessment should include the grams of protein consumed, food tolerance, muscle strength, and body composition rather than serum albumin alone [80], [83]. Persistent intake below requirements, rapid loss of lean mass, edema, or hypoalbuminemia warrants dietitian-led intervention and investigation for malabsorption or surgical complications [84].

Clinical warning signs—recurrent vomiting, dysphagia, diarrhea, edema, muscle wasting, neuropathic symptoms, jaundice, easy bruising, or unexplained aminotransferase elevation—should override routine follow-up intervals. They should prompt coordinated bariatric, hepatology, and clinical-nutrition assessment rather than empirical escalation of supplementation [64], [84], [108].

Weight regain and metabolic recurrence are later indications for intensified monitoring. Increasing waist circumference, recurrent diabetes, rising triglycerides, liquid calories, alcohol exposure, and declining physical activity can restore hepatic fatty-acid and carbohydrate delivery [20], [109]. After clinically meaningful weight regain, liver enzymes, dietary intake, alcohol exposure, glycemic control, and non-invasive fibrosis markers should be reassessed [3].

6.3. Integrating liver-focused nutrition into standard MBS follow-up protocols

Contemporary guidelines recommend structured perioperative and long-term nutritional and metabolic management after MBS, including assessment for deficiencies and procedure-specific supplementation [79], [101], [102], [110]. Integration should begin before surgery with a combined nutritional and hepatic risk assessment. Among potential MBS candidates in the United States, 76.7% (95% CI, 71.7–81.0) had steatosis and 23.1% (95% CI, 17.8–29.3) had elevated liver stiffness [111]. Preoperative assessment should therefore include fibrosis stage, diabetes, alcohol intake, dietary quality, muscle status, existing deficiencies, and the planned procedure [3], [112]. A written handover can then help postoperative teams distinguish routine nutritional prevention from liver-directed surveillance.

During the first postoperative year, liver-focused nutrition should be integrated into the standard progression from liquids to solid foods. Each phase should maintain hydration and protein intake while introducing vegetables, legumes, whole grains, fruit, and unsaturated fats according to tolerance [79], [102]. Because diet quality and food tolerance may deteriorate despite reduced energy intake, they should be assessed independently of weight loss [23], [89].

Biochemical monitoring is recommended at 3, 6, and 12 months in the first postoperative year and at least annually thereafter, with core measures including blood count, ferritin, folate, vitamin B12, vitamin D, calcium, parathyroid hormone, and renal and liver function [101], [102]. Vitamins A, E, and K, zinc, copper, and selenium should be assessed after hypoabsorptive procedures or when deficiency is suspected [102]. Results should be interpreted together with inflammation, supplementation, and dietary intake [22].

Liver assessment should accompany nutritional monitoring. Weight trajectory, waist circumference, diabetes status, alcohol intake, aminotransferases, platelets, albumin, and FIB-4 can be reviewed at scheduled visits, and elastography or MRI-based assessment can be considered in patients with baseline fibrosis, persistent abnormalities, discordant results, or meaningful weight regain [3]. Non-invasive tests behave differently after surgery: in a large biopsy cohort, the mean values of most tests decreased after surgery but FIB-4 did not, and their performance for detecting fibrosis resolution was better at 5 years than at 1 year [21]. Normalization of aminotransferases should therefore not be interpreted as histological resolution.

Follow-up intensity should be organized primarily by surgical anatomy, dietary intake, deficiencies, fibrosis risk, symptoms, and metabolic trajectory. Patients with uncomplicated surgery, adequate intake, corrected deficiencies, low fibrosis risk, and stable weight may follow a standard annual pathway. Previous MASH, intermediate fibrosis tests, recurrent diabetes, inadequate protein intake, or repeated deficiencies justify more frequent review [3], [113]. Advanced fibrosis, synthetic dysfunction, severe malnutrition, or unexplained hepatic deterioration should prompt hepatology referral and multidisciplinary review [108].

Structured records should capture procedure type, dietary exposures, supplement adherence, deficiencies, weight change, diabetes status, and longitudinal liver measurements [114], [115].

6.4. Pharmacotherapy for persistent or recurrent obesity and residual MASH after MBS

Weight recurrence, insufficient weight loss, and residual MASH are increasingly managed with obesity pharmacotherapy, but evidence specific to the liver after MBS is limited [116]. In a randomized trial of 48 patients with suboptimal weight loss 12–36 months after MBS, liraglutide produced greater 12-month weight loss than placebo (between-group difference, 7.1 kg; 95% CI, 3.9–10.3) [117]. A meta-analysis of six studies (300 participants), including non-randomized designs, reported significant additional weight loss with semaglutide after MBS [118]. Neither study reported histological liver outcomes.

For residual MASH, efficacy data come from non-surgical populations. In the phase 3 ESSENCE trial in patients with MASH and stage 2–3 fibrosis, semaglutide 2.4 mg weekly achieved MASH resolution without fibrosis worsening in 62.9% versus 34.3% of patients receiving placebo (difference, 28.7 percentage points; 95% CI, 21.1–36.2) and fibrosis improvement without MASH worsening in 36.8% versus 22.4% (difference, 14.4 percentage points; 95% CI, 7.5–21.3) at 72 weeks [119]. On this basis, semaglutide received accelerated approval in the United States in 2025 for MASH with moderate-to-advanced fibrosis [120]. The liver-directed thyroid hormone receptor-β agonist resmetirom was superior to placebo for both histological endpoints in a phase 3 trial [121], and in a phase 2 trial the dual glucose-dependent insulinotropic polypeptide and GLP-1 receptor agonist tirzepatide 15 mg achieved MASH resolution in 62% versus 10% with placebo (difference, 53 percentage points; 95% CI, 37–69) [122]. These trials did not specifically evaluate patients after MBS, and how these agents interact with altered postoperative anatomy and nutrient absorption is uncertain.

Pharmacotherapy after MBS also has nutritional implications. In a small post-SG cohort (n = 29), semaglutide produced additional weight and glycemic improvement but was followed by decreases in serum albumin, vitamin B12, and zinc [123]. Further reductions in intake may compound postoperative protein and micronutrient deficits and lean-mass loss. Pharmacotherapy for weight recurrence or residual MASH after MBS should therefore be selected according to fibrosis stage and metabolic indication, accompanied by protein and micronutrient monitoring, and evaluated prospectively with direct hepatic endpoints [3], [120].

6.5. Considerations for adolescents

MBS is increasingly used in adolescents with severe obesity. Among eligible adolescents in the United States, utilization rose 7.7-fold between 2016 and 2025, although at its peak only 0.24% of eligible adolescents underwent surgery, predominantly SG [124]. Pediatric guidelines recommend considering MBS for adolescents with severe obesity and comorbidities including MASH [125], [126]. Liver disease is common in this population: in 296 children and adolescents undergoing SG, liver biopsy showed MASLD in 76%, MASH in 39.9%, and stage ≥2 fibrosis in 37.2% [127].

Hepatic outcomes in adolescents appear favorable. In the Teen Longitudinal Assessment of Bariatric Surgery (Teen-LABS) cohort (n = 219), the proportion with elevated alanine aminotransferase fell from 71% at baseline to 36% at 2 years, and the decline was associated with metabolic improvement independently of weight loss [128]. In a prospective controlled cohort of 42 adolescents with biopsy-confirmed MASLD, SG was associated with greater 52-week histological improvement than comprehensive lifestyle intervention, including fibrosis improvement (risk difference, 0.36; 95% CI, 0.12–0.59) [129]. No study has evaluated gene–nutrition interactions on liver outcomes after adolescent MBS. Because surgery occurs during growth and is followed by decades of exposure, lifelong nutritional follow-up, monitoring of supplement adherence, and structured transition to adult care are particularly important [125]; genetic information should not influence surgical eligibility in this group.

7. Translational challenges and research priorities

7.1. Practical limitations to the clinical translation of genotype-guided nutrition

A useful genetic test must improve prediction beyond procedure type, baseline fibrosis, diabetes, and weight trajectory, and it must lead to an intervention that improves hepatic outcomes. Neither condition has been met after MBS. In a multi-ancestry analysis of 10,960 individuals, a higher BMI polygenic score was only modestly associated with lower weight loss after MBS (+0.7% per standard deviation), and the effect attenuated in sensitivity analyses [130]. Although weight loss is not a liver endpoint, this finding illustrates the limited incremental value of current genetic predictors when a powerful anatomical intervention dominates the response.

Most studies evaluate one variant, one nutrient, and a short-term surrogate, and replication is inconsistent [66], [92]. Each proposed genotype-informed recommendation would require prospective comparison with standard nutrition, including assessment of hepatic benefit, adverse effects, adherence, and durability.

A genetic report is static, whereas postoperative phenotype and exposures change. BMI, diabetes, and alcohol exposure modify the expression of genetic susceptibility [29], [30], and diet, supplementation, and metabolic trajectory evolve after surgery [131]. A single genotype cannot replace serial reassessment of these exposures.

Technical standardization is a further obstacle. Polygenic scores depend on variant selection, weighting, reference populations, platforms, imputation, and thresholds, and they are not calibrated for post-MBS decisions [73], [132]. Implementation would require transparent reporting of score construction, ancestry, analytical performance, and intended use, with external validation against liver outcomes after specific operations.

Workforce readiness is also limited: in a survey of 169 nutrition professionals, 94% reported insufficient knowledge of personalized nutrition and only 9.5% had received nutrigenetics training [133].

7.2. Equity, accessibility, and ethical considerations

Ancestry is the most immediate equity challenge. Genome-wide studies and polygenic scores remain disproportionately derived from European-ancestry populations, and score accuracy varies across the genetic-ancestry continuum and even within broad ancestry groups [76], [134]. Applying unvalidated scores to Latin American, African, Asian, Indigenous, or admixed patients could create false reassurance or unnecessary surveillance [132].

Financial feasibility extends beyond the cost of genotyping. Residence in areas with higher food-insecurity rates was independently associated with lower percentage excess weight loss 1 year after MBS [135]. Genetic stratification without access to nutritional counseling, appropriate foods, supplements, and monitoring could widen disparities; access to these established services should take priority over genetic testing [136].

When genetic testing is undertaken, informed consent should cover the purpose and limitations of the test, incidental findings, data storage and reuse, familial implications, and the right not to know. Particular caution is needed with direct-to-consumer services, which vary in validity and regulatory oversight [137], [138].

Risk communication should avoid both weight stigma [139] and genetic determinism. Patients and clinicians frequently misinterpret polygenic risk estimates, and below-average scores may generate false reassurance [140], [141]. Probabilistic communication within genetic counseling is therefore essential [142], and genetic findings should never be used to withhold indicated MBS, nutritional support, or liver surveillance.

7.3. Need for interdisciplinary collaboration among surgeons, dietitians, and genetic specialists

If genetic testing is undertaken in research or specialized care, interpretation requires a shared pathway. The surgeon contributes knowledge of anatomy, absorption, and complications; the dietitian translates hepatic and nutritional phenotype into feasible food and supplement plans; a genetic counselor addresses test validity, ancestry limitations, consent, and risk communication; and hepatology input is needed when fibrosis is advanced or findings are discordant [142], [143]. Before testing, the team should define how the result could change management.

Multidisciplinary care should be matched to need. In a meta-analysis including 1533 adults undergoing MBS, intensive postoperative multidisciplinary interventions increased weight loss (standardized mean difference, −0.94; 95% CI, −1.27 to −0.61) [144]. A continuum-of-care model integrating obesity medicine and metabolic surgery is consistent with this approach [145].

Shared records should make data provenance, timing, and completeness explicit, because differential information capture in electronic health records can distort estimates of long-term benefits and harms after MBS [146].

7.4. Research priorities

Future research should prioritize adequately powered, multicenter, multi-ancestry prospective cohorts that begin before MBS and extend beyond the weight-loss phase. These cohorts should integrate baseline genotyping with repeated assessment of diet, body composition, alcohol use, medications, physical activity, and metabolic status, and should analyze steatosis, MASH activity, fibrosis, and clinical events as distinct outcomes [136], [147], [148]. Genetic information should be tested for incremental predictive value beyond baseline fibrosis, diabetes, weight trajectory, FIB-4, and elastography [37]. The central interventional question is whether genotype-informed nutrition improves outcomes beyond high-quality phenotype-guided care; trials should prespecify variants and interaction analyses, ensure adequate power, and assess fibrosis, nutritional deficiencies, lean mass, adherence, quality of life, costs, and harms [149], [150].

8. Limitations

This review has several limitations. First, it is a narrative review; although it was informed by a structured search with independent screening, it did not include formal risk-of-bias assessment or quantitative synthesis, and study selection may have been influenced by author judgment. Second, the variant-level ratings in Table 3 reflect author judgment based on prespecified criteria rather than a formal certainty-of-evidence framework. Third, much of the genetic and gene–diet evidence derives from non-bariatric populations, cross-sectional designs, or small cohorts, and no study has jointly assessed genotype, repeated postoperative dietary exposure, and direct hepatic outcomes; conclusions about postoperative gene–nutrition interactions are therefore inferential. Fourth, the available genetic studies are dominated by European-ancestry populations, which limits generalizability. Fifth, the search was restricted to three databases, and evidence on pharmacotherapy for MASH is evolving rapidly, so relevant recent studies may have been missed. Finally, mechanistic and animal data were used only to contextualize gaps and may not translate to humans.

9. Conclusions

MBS provides a dynamic model of gene–nutrition–liver interaction because it simultaneously modifies energy intake, nutrient absorption, weight trajectory, glycemic control, and bile-acid signaling. Steatosis improves rapidly and MASH often resolves, but fibrosis may regress slowly or incompletely, so metabolic recovery does not eliminate residual liver risk. Postoperative nutrition is a time-varying determinant of hepatic recovery: diet quality, protein adequacy, micronutrient status, alcohol use, and weight trajectory continue to evolve after the rapid weight-loss phase. Variants in PNPLA3, TM6SF2, GCKR, and related genes provide biological plausibility for gene–nutrition interactions, but whether they modify postoperative hepatic outcomes remains unproven, and current genetic predictors add little to clinical measures. Several clinically relevant implications emerge from these findings:

• •Base postoperative liver care on phenotype: assess fibrosis before surgery and continue surveillance after weight loss in patients with preoperative MASH, significant fibrosis, type 2 diabetes, older age, or weight regain, using serial non-invasive tests rather than aminotransferases alone.
• •Treat postoperative nutrition as part of hepatic care: ensure adequate protein intake, procedure-specific micronutrient supplementation and monitoring, limited added sugars and alcohol, and high-quality dietary fat for all patients, regardless of genotype.
• •Do not use genotype to select the procedure, diet, supplements, or monitoring intensity; for residual MASH or weight recurrence, pharmacotherapy with proven efficacy may be considered alongside nutritional monitoring.

Author contributions

Conceptualization: C.R.-G., N.M.R.-G., D.S.-R., L.S.; Methodology: C.R.-G., N.M.R.-G., D.S.-R.; Investigation: C.R.-G., N.M.R.-G., D.J.G., G.R., L.C., V.P.; Writing – original draft: C.R.-G., N.M.R.-G., D.J.G., G.R., L.C., V.P., D.S.-R., L.S.; Writing – review & editing: C.R.-G., N.M.R.-G., D.J.G., G.R., L.C., V.P., D.S.-R., L.S.; Validation: G.R., L.C., V.P., L.S.; Supervision: D.S.-R., L.S.

Data availability

No new datasets were generated or analyzed for this narrative clinical review; all evidence discussed derives from published sources cited in the manuscript.

Ethics statement

Not applicable. This narrative clinical review did not involve new research with human participants or animals.

Declaration of generative AI and AI-assisted technologies

During the preparation of this work, the author(s) used artificial intelligence (AI)–assisted tools solely for language editing and grammar checking to improve clarity and readability of the manuscript. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the content of the publication.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.