Skip to content

Resmetirom and thyromimetics

TL;DR — Resmetirom is an oral, liver-directed, thyroid hormone receptor-β-selective agonist and the first drug approved for MASH. In MAESTRO-NASH (n=966 in the primary analysis population; biopsy-confirmed NASH, F1B–F3), NASH resolution with no worsening of fibrosis at week 52 occurred in 25.9% (80 mg) and 29.9% (100 mg) versus 9.7% placebo, and ≥1-stage fibrosis improvement with no worsening of NAS in 24.2% and 25.9% versus 14.2% (all p<0.001) (Harrison 2024, PMID 38324483, NCT03900429). LDL-cholesterol fell 13.6% and 16.3% versus +0.1% on placebo at week 24 — a genuine secondary benefit given that cardiovascular disease is the leading cause of death in this population. The FDA granted accelerated (conditional) approval in March 2024 for non-cirrhotic MASH with moderate-to-advanced fibrosis (F2–F3), in conjunction with diet and exercise (Keam 2024, PMID 38771485). Two things follow from "accelerated": the approval rests on a histological surrogate, and the 54-month outcome phase of MAESTRO-NASH is still running. The absolute treatment effects are modest — number needed to treat for fibrosis improvement is about 9–10 against placebo — and the drug is not approved in cirrhosis, where the fibrosis-stage evidence stops. In network meta-analysis of 29 randomised trials (n=9,324), resmetirom is significantly better than placebo for both endpoints but ranks mid-field: pegozafermin, cilofexor+firsocostat and cilofexor+selonsertib rank highest for fibrosis regression, and pegozafermin, survodutide and tirzepatide for MASH resolution (Souza 2025, PMID 39903735).

Mechanism

MASLD is associated with suboptimal thyroid status — subclinical hypothyroidism or low-normal thyroid function — and within the steatotic liver intracellular thyroid hormone concentrations are low and THR activation is reduced. The association is real but smaller than the drug's premise implies, and worth quantifying rather than asserting:

Exposure Outcome Effect Source
Primary hypothyroidism (levothyroxine, subclinical or overt), 24 cross-sectional studies, ~76.5M individuals prevalent MASLD OR 1.43 (1.23–1.66), I²=89% Mantovani 2024, PMID 38782564
Same MASH or advanced fibrosis (5 studies) OR 2.84 (2.07–3.90), I²=0% Mantovani 2024, PMID 38782564
Same, 4 longitudinal cohorts, median 4.5 y incident MASLD HR 1.39 (0.98–1.97) — not significant Mantovani 2024, PMID 38782564
Subclinical hypothyroidism, 10 studies, n=71,332 prevalent MASLD OR 1.46 (1.23–1.73), I²=36% Amdetsion 2025, PMID 41357781
Subclinical hypothyroidism, 4 cohorts, n=31,518 incident MASLD HR 1.59 (1.05–2.40); prospective-only HR 1.90 (1.50–2.39), I²=0 Amdetsion 2025, PMID 41357781
Hypothyroidism, 12,172 biopsy-confirmed MASLD vs 56,831 matched controls (Sweden, ESPRESSO) MASLD OR 1.68 (1.36–2.06); stable against 10,682 sibling controls Yuan 2025, PMID 40342634
Hyperthyroidism, same cohort MASLD OR 0.17 (0.05–0.56) Yuan 2025, PMID 40342634

Two qualifications matter. First, the cross-sectional and longitudinal estimates disagree in the larger meta-analysis (OR 1.43 prevalent, HR 1.39 non-significant incident) — the association is better established as co-occurrence than as antecedence, and the two syntheses reach opposite conclusions about the prospective signal (PMIDs: 38782564, 41357781). Second, the absolute numbers are small: hypothyroidism was present in 2.5% of biopsy-confirmed MASLD cases and 1.4% of controls, and Mendelian randomisation attributed ~41% of the association to metabolic disorders rather than to thyroid signalling directly (Yuan 2025, PMID 40342634). Resmetirom therefore does not work by correcting a prevalent deficiency; it exploits a pathway that is intact in most patients. Thyroid hormone reduces liver triglyceride by stimulating fatty-acid disposal through lipophagy and β-oxidation, and lowers LDL-cholesterol. Because thyroid hormone acts in heart and bone as well as liver, the drug class was engineered for selectivity: reduced affinity for THR-α and a tissue-distribution profile differing from endogenous thyroid hormone, to avoid cardiovascular and skeletal effects (Ratziu 2025, PMID 39428045).

This is a hepatic-disposal mechanism rather than a supply-side one — contrast with the incretins, which act mainly by reducing energy intake and weight (GLP-1 and incretin therapy), and with the fat-source apportionment in pathogenesis.

The MAESTRO programme

Trial Design Population Primary endpoint(s) Result
MGL-3196-05 phase 2 (Harrison 2019, PMID 31727409, NCT02912260) 36-week RCT, 2:1, 25 US centres; 84 resmetirom 80 mg vs 41 placebo biopsy NASH F1–F3, hepatic fat ≥10% by MRI-PDFF relative change in MRI-PDFF hepatic fat at week 12 −32.9% vs −10.4%; LS mean difference −22.5% (95% CI −32.9 to −12.2), p<0.0001. Week 36: −37.3% vs −8.5%; difference −28.8% (−42.0 to −15.7), p<0.0001
MAESTRO-NAFLD-1 (Harrison 2023, PMID 37845512, NCT04197479) 52-week phase 3 safety trial; 100 mg (n=325), 80 mg (n=327), placebo (n=320), plus open-label 100 mg (n=171) NAFLD with presumed NASH incidence of treatment-emergent adverse events TEAEs 86.1% (100 mg), 88.4% (80 mg), 81.8% (placebo); excess diarrhoea and nausea at initiation. Secondary (80 mg, 100 mg vs placebo): LDL-C −11.1%, −12.6%; apoB −15.6%, −18.0%; triglycerides −15.4%, −20.4%; 16-week hepatic fat −34.9%, −38.6% (p<0.0001); liver stiffness −1.02, −1.70; 52-week hepatic fat −28.8, −33.9
MAESTRO-NASH (Harrison 2024, PMID 38324483, NCT03900429) ongoing 54-month phase 3; 1:1:1; primary analysis population 966 (80 mg 322, 100 mg 323, placebo 321) biopsy-confirmed NASH, fibrosis F1B, F2 or F3 at week 52: (1) NASH resolution incl. NAS reduction ≥2 with no worsening of fibrosis; (2) fibrosis improvement ≥1 stage with no worsening of NAS see below

MAESTRO-NASH week-52 results:

Endpoint Placebo Resmetirom 80 mg Resmetirom 100 mg
NASH resolution, no worsening of fibrosis 9.7% 25.9% 29.9%
Fibrosis improvement ≥1 stage, no worsening of NAS 14.2% 24.2% 25.9%
LDL-C change to week 24 +0.1% −13.6% −16.3%
Serious adverse events 11.5% 10.9% 12.7%

All histological and LDL comparisons p<0.001. Diarrhoea and nausea were more frequent with resmetirom. Converting to absolute terms: the fibrosis-improvement difference is 10.0 and 11.7 percentage points, giving NNT ≈ 10 and ≈ 9; the NASH-resolution difference is 16.2 and 20.2 points, NNT ≈ 6 and ≈ 5. These are respectable for a chronic liver disease and unremarkable set against ≥10% weight loss by lifestyle (lifestyle and weight loss).

The non-invasive signal, and where it came from. The phase 2 programme included a 36-week open-label active-treatment extension in 31 consenting patients (14 of them former placebo recipients) with persistently elevated liver enzymes. At OLE week 36, MRI-PDFF fell −11.1% absolute (SE 1.5) and −52.3% relative (SE 4.4), both p<0.0001; LDL-C −26.1% (4.5), apoB −23.8% (3.0), triglycerides −19.6% (5.4) / −46.1 mg/dL; transient-elastography liver stiffness −2.1 kPa (0.8), p=0.015; and PRO-C3 −9.8 ng/mL (2.3), p=0.0004, in those with baseline PRO-C3 ≥10 (Harrison 2021, PMID 33860116). PRO-C3/C3M — the ratio interpreted as net fibrogenesis — fell −0.76 (−1.27 to −0.24), p=0.0044 in the main study and −0.68, p<0.0001 in the OLE. This small, unblinded, enzyme-selected extension is the origin of the idea that resmetirom response can be tracked non-invasively in an individual patient; it is hypothesis-generating and has never been validated against the histological endpoint that licensed the drug.

Digital pathology re-reads the same biopsies. A prespecified secondary analysis applied AI-based second-harmonic-generation quantification (qFibrosis) to the paired MAESTRO-NASH biopsies (n=966). Categorical qFibrosis stage improved by ≥1 in 24.4% (80 mg) and 22.3% (100 mg) more patients than placebo (nominal p≤0.001), and placebo-corrected continuous qFibrosis fell −0.95 (95% CI −1.22 to −0.69) and −1.08 (−1.34 to −0.81) (Schattenberg 2026, PMID 41895606). Of 30 outcome-associated collagen features, the six correlating most strongly with pathologist stage and with liver stiffness were in the portal tract and Zone 2 — and portal-region and chicken-wire features showed the largest treatment reductions. That regional pattern is hypothesis-generating; this analysis did not test a mechanism. The categorical qFibrosis treatment difference (24.4 percentage points at 80 mg) was wider than the conventional pathologist-staged difference (10.0 points), but the continuous qFibrosis result is reported in score units and cannot be compared as a percentage-point contrast.

Approval, and what "accelerated" constrains

Resmetirom (Rezdiffra) received accelerated FDA approval in March 2024, in conjunction with diet and exercise, for adults with non-cirrhotic MASH with moderate-to-advanced fibrosis consistent with F2–F3; it was simultaneously under EU regulatory review (Keam 2024, PMID 38771485). Accelerated approval is granted on a surrogate reasonably likely to predict clinical benefit, with confirmation required — here, from the ongoing outcome phase of MAESTRO-NASH.

The approval created an immediate practical problem, since the licensed population is defined histologically but is expected to be identified non-invasively. An expert panel addressed exactly this: how to identify F2–F3 patients without biopsy, how to exclude patients with more advanced disease who should not be treated until further data emerge, and when to stop treatment (Noureddin 2024, PMID 39038768). The exclusion problem is the sharper one — cirrhosis is precisely the population in whom the drug is untested and the population non-invasive tests most often flag ambiguously. See noninvasive assessment.

Interaction with weight loss and diabetes therapy

A prespecified secondary analysis of MAESTRO-NASH asked whether background metabolic therapy or weight change modified the effect. At baseline 13–17% of patients (all with T2D) were on stable GLP-1 receptor agonist or SGLT2-inhibitor therapy; no weight loss above baseline occurred on those agents during the trial (Noureddin 2025, PMID 41127972):

Comparison MASH resolution Fibrosis improvement MRI-PDFF Liver stiffness
Resmetirom 100 mg with ≥5% weight loss 56.6% 40.6% −69% −4.6 kPa
Resmetirom 100 mg with <5% weight loss 33.8% 31.5% −46% −2.3 kPa

Patients on SGLT2 inhibitors or GLP-1 receptor agonists had similar rates of MASH resolution and fibrosis improvement with resmetirom as patients not on those therapies — i.e. no interference — while ≥5% weight loss enhanced resmetirom's efficacy. Since this compares outcome-defined subgroups rather than randomised arms, the weight-loss contrast is subject to the same confounding as the lifestyle dose–response literature, and should not be read as evidence that inducing weight loss would produce that increment.

Where resmetirom sits among MASH drugs

Network meta-analysis of 29 randomised controlled trials in biopsy-proven MASH published January 2020–December 2024 (n=9,324), with SUCRA rankings (Souza 2025, PMID 39903735):

Endpoint Agents significantly better than placebo Top-ranked (SUCRA)
Fibrosis improvement ≥1 stage without worsening MASH pegozafermin, cilofexor+firsocostat, denifanstat, survodutide, obeticholic acid, tirzepatide, resmetirom, semaglutide pegozafermin 79.92; cilofexor+firsocostat 71.38; cilofexor+selonsertib 69.11
MASH resolution without worsening fibrosis pegozafermin, survodutide, tirzepatide, efruxifermin, liraglutide, vitamin E + pioglitazone, resmetirom, semaglutide, pioglitazone, denifanstat, lanifibranor pegozafermin 91.75; survodutide 90.87; tirzepatide 84.70

Resmetirom is the only one of these that is approved. Ranking above it are agents at earlier phases, in smaller trials, with wider confidence intervals and — for the combination arms — no phase 3 programme. Network meta-analysis across trials with different populations, placebo responses and biopsy-reading conventions is a weak basis for head-to-head inference; the ranking is a hypothesis about relative efficacy, not a demonstration of it.

And the rankings do not replicate. A second network meta-analysis, of 48 RCTs and 10,119 participants, restricted to fibrosis improvement, found only pegozafermin, obeticholic acid and resmetirom superior to placebo in the F1–F3 analysis, and ranked pegbelfermin first (SUCRA 77.11%) with pegozafermin second (74.91%) — pegbelfermin does not appear at all in the Souza ranking, and obeticholic acid, an agent the FDA declined twice, topped the 1.5-year analysis (SUCRA 81.64%) (Han 2026, PMID 41811770). For liver stiffness by VCTE at 0.5 years, pegozafermin ranked first (SUCRA 96.98%). Two contemporaneous network meta-analyses of overlapping trial sets thus produce different podiums (PMIDs: 39903735, 41811770). The disagreement is informative about the method, not about the drugs: SUCRA rankings are unstable to inclusion criteria, endpoint definition and time window, and neither analysis should be read as evidence that any unapproved agent outperforms resmetirom. See other pharmacotherapy and clinical trials landscape.

Other THR-β agonists

The class is broader than resmetirom, and the design logic — hepatic targeting plus THR-β selectivity to spare cardiac and skeletal THR-α effects — is shared (Ratziu 2025, PMID 39428045). Resmetirom is described as the most advanced compound in the class. Exact PubMed re-run on 2026-09-02: ((MASLD[Title/Abstract] OR NAFLD[Title/Abstract] OR MASH[Title/Abstract] OR NASH[Title/Abstract] OR "steatotic liver disease"[Title/Abstract]) AND (VK2809[Title/Abstract] OR VK-2809[Title/Abstract] OR VOYAGE[Title/Abstract])) — 8 records; no peer-reviewed phase 2b VOYAGE efficacy report for VK2809 was retrieved.

The most informative non-resmetirom compound is HSK31679, a resmetirom derivative studied in a randomised, double-blind, placebo-controlled multiple-ascending-dose cohort (n=32 active, n=8 placebo) alongside germ-free mouse work (Zhang 2025, PMID 39181210). Two findings bear on the whole class. First, HSK31679 out-performed resmetirom in specific-pathogen-free mice but not in germ-free mice — the increment is microbiome-dependent, mediated by steric hindrance of bacterial glucosylceramide synthase (GCS) at residues R123 and Y401 and consequent loss of monoglucosylation of Cer(d18:1/16:0) and Cer(d18:1/24:1). Second, in humans, 160 mg HSK31679 shifted the peripheral immune compartment (fewer CD8α⁺ dendritic cells and MINCLE⁺ macrophages) only in participants with high faecal GCS activity. If replicated, faecal GCS activity would be the first mechanistic predictor of THR-β agonist response, and would supply an explanation for the "remarkable heterogeneity of individual clinical efficacy" that the authors take as their starting point. It is a phase 1-scale cohort in a mouse-anchored paper, so it is a hypothesis, not a biomarker.

Cost, value and access — the unresolved economics

Two lifetime microsimulations of the same drug reach opposite verdicts, and the reason is instructive.

Analysis Timing Model ICER vs standard of care Verdict at $100,000/QALY
Javanbakht 2023, PMID 36104546 pre-approval, informed by the phase 2 trial, US commercial payer Markov, lifetime $53,929/QALY (cost +$66,764; +1.24 QALY) cost-effective, up to a daily price of $72.00
Le 2025, PMID 40577015 post-approval, agent-based state-transition microsimulation, 14 histological health states, 200,000 simulated F2/F3 patients (mean age 57.1) lifetime, 2023 USD $140,134/QALY (cost +$36,255; +0.26 QALY) not cost-effective

The gap is almost entirely in the QALY gain — 1.24 versus 0.26, a factor of nearly five — not in cost. In the later model, assuming no discontinuation raised the ICER to $318,740/QALY, versus $140,134/QALY in the base case. However, the same abstract also states that lower discontinuation increased cost-effectiveness; that verbal statement conflicts with the reported no-discontinuation ICER and is retained here as a source-internal inconsistency rather than reconciled. The reported supportable annual prices were $10,914, $15,406 and $19,879 at $50,000, $100,000 and $150,000/QALY in the base case, versus $5,645–$10,619/year without discontinuation (Le 2025, PMID 40577015). Treatment offset $18,499/patient in modelled MASLD-related medical costs. Neither model had outcome data; both extrapolated from histological surrogates.

What access actually looks like. A single-centre retrospective series of the first cohort prescribed resmetirom after approval (137 prescribed, 113 treated ≥3 months, April–September 2024) found nausea/vomiting in 12.4% and diarrhoea in 12.4%, discontinuation in 8.8%, and mean ALT falling 49.96 → 45.38 IU/L and AST 40.15 → 36.15 IU/L over three months (Shuaibi 2025, PMID 40688390). Insurance appeals were required for 24.1% of patients; 93.4% were ultimately approved; and the mean delay between approval and dispensing was 12.5 days. The reported obstacle is specific and points back at noninvasive assessment: payers required invasive fibrosis testing for approval — reinstating the biopsy that the non-invasive pathway was designed to avoid. The three-month enzyme change is uncontrolled and small; it is evidence about tolerability and administrative friction, not efficacy.

Open questions

  • Does resmetirom reduce hepatic clinical outcomes? This is the question accelerated approval deferred. MAESTRO-NASH runs to 54 months with a clinical-outcome component; the published week-52 report is histological (PMID 38324483). Until that reports, the drug is licensed on a surrogate whose validity as a treatment-effect transfer measure is itself unestablished (natural history).
  • Does it work in cirrhosis? The trial enrolled F1B–F3 and the label excludes cirrhosis (PMID 38771485). The expert panel's principal practical instruction is to exclude cirrhotic patients (PMID 39038768) — meaning the population with the highest event rate is untreated by the only approved drug. See cirrhosis and decompensation.
  • How should F2–F3 be identified without biopsy? The label is histological, the practice is non-invasive, and non-invasive tests have a large indeterminate zone and region-varying accuracy (PMID 41100867). No prospective study has validated a non-invasive selection rule against the trial's histological entry criteria.
  • When should treatment stop? The expert panel proposed stopping criteria (PMID 39038768) but there is no trial evidence on discontinuation, on duration of benefit, or on rebound.
  • Is the LDL reduction clinically meaningful here? A 13.6–16.3% LDL fall in a population whose leading cause of death is cardiovascular (see cardiovascular and extrahepatic outcomes) is potentially the drug's largest population-level benefit, and no trial has been designed to measure it.
  • Is resmetirom worth its price? Two lifetime models disagree by a factor of nearly five on QALYs gained (1.24 vs 0.26) and land on opposite sides of the $100,000/QALY threshold (PMIDs: 36104546, 40577015). The disagreement cannot be resolved from the trial data, because both extrapolate a 52-week histological surrogate; only the MAESTRO-NASH outcome phase can settle it.
  • Does faecal glucosylceramide-synthase activity predict THR-β agonist response? A microbiome-dependent mechanism and a candidate human predictor were reported for the resmetirom derivative HSK31679 in a 40-participant MAD cohort (PMID 39181210). It has not been tested against resmetirom response in MAESTRO-NASH, and would be the class's first response biomarker if it held.
  • Does digital qFibrosis add information beyond pathologist staging? Categorical qFibrosis separated 80-mg treatment from placebo by 24.4 percentage points where conventional pathologist staging separated it by 10.0, on the same biopsies (PMIDs: 41895606, 38324483). The continuous qFibrosis result is in score units, so it should not be described as a percentage-point treatment contrast.
  • Why does the antifibrotic effect concentrate in the portal tract? The largest resmetirom-associated reductions were in portal-region and chicken-wire collagen features (PMID 41895606), whereas MASH fibrosis is conventionally described as beginning pericentrally (histology and biopsy). No mechanistic account of this regional preference has been published.
  • Does combination with an incretin add? The MAESTRO-NASH secondary analysis found no interference and a weight-loss-associated increment (PMID 41127972), but was observational within the trial. No randomised combination trial has reported.

References

  1. Harrison SA, Bedossa P, Guy CD, et al. A Phase 3, Randomized, Controlled Trial of Resmetirom in NASH with Liver Fibrosis. N Engl J Med. 2024;390(6):497-509. PMID 38324483
  2. Harrison SA, Taub R, Neff GW, et al. Resmetirom for nonalcoholic fatty liver disease: a randomized, double-blind, placebo-controlled phase 3 trial. Nat Med. 2023;29(11):2919-2928. PMID 37845512
  3. Harrison SA, Bashir MR, Guy CD, et al. Resmetirom (MGL-3196) for the treatment of non-alcoholic steatohepatitis: a multicentre, randomised, double-blind, placebo-controlled, phase 2 trial. Lancet. 2019;394(10213):2012-2024. PMID 31727409
  4. Keam SJ. Resmetirom: First Approval. Drugs. 2024;84(6):729-735. PMID 38771485
  5. Noureddin M, Charlton MR, Harrison SA, et al. Expert Panel Recommendations: Practical Clinical Applications for Initiating and Monitoring Resmetirom in Patients With MASH/NASH and Moderate to Noncirrhotic Advanced Fibrosis. Clin Gastroenterol Hepatol. 2024;22(12):2367-2377. PMID 39038768
  6. Ratziu V, Scanlan TS, Bruinstroop E. Thyroid hormone receptor-β analogues for the treatment of metabolic dysfunction-associated steatohepatitis (MASH). J Hepatol. 2025;82(2):375-387. PMID 39428045
  7. Noureddin M, Rinella M, Taub R, et al. Effects of Resmetirom on Metabolic-Dysfunction Associated Steatohepatitis in Patients With Weight Loss and/or Diabetes Taking Glucagon-Like Peptide-1 Receptor Agonists and Other Diabetes Therapies: A Secondary Analysis of the MAESTRO-NASH Trial. Aliment Pharmacol Ther. 2025;62(11-12):1089-1099. PMID 41127972
  8. Souza M, Al-Sharif L, Antunes VLJ, et al. Comparison of pharmacological therapies in metabolic dysfunction-associated steatohepatitis for fibrosis regression and MASH resolution: Systematic review and network meta-analysis. Hepatology. 2025;82(6):1523-1533. PMID 39903735
  9. Younossi ZM, de Avila L, Petta S, et al. Global performance of non-invasive tests in MASLD: Insights from the G-MASLD study. Hepatology. 2026;84(1):161-174. PMID 41100867
  10. Harrison SA, Bashir M, Moussa SE, et al. Effects of Resmetirom on Noninvasive Endpoints in a 36-Week Phase 2 Active Treatment Extension Study in Patients With NASH. Hepatol Commun. 2021;5(4):573-588. PMID 33860116
  11. Schattenberg JM, Bedossa P, Guy CD, et al. Quantitative regression of qFibrosis with resmetirom: Exploratory histologic endpoints from the MAESTRO-NASH phase III clinical trial. J Hepatol. 2026;85(2):202-211. PMID 41895606
  12. Le P, Dasarathy S, Herman WH, et al. Value-Based Pricing of Resmetirom for Metabolic Dysfunction-Associated Steatotic Liver Disease. JAMA Netw Open. 2025;8(6):e2517122. PMID 40577015
  13. Javanbakht M, Fishman J, Moloney E, et al. Early Cost-Effectiveness and Price Threshold Analyses of Resmetirom: An Investigational Treatment for Management of Nonalcoholic Steatohepatitis. Pharmacoecon Open. 2023;7(1):93-110. PMID 36104546
  14. Shuaibi S, Tobal I, Gore J, et al. A Real-World Experience With Resmetirom: Tolerability and Access. Gastro Hep Adv. 2025;4(9):100709. PMID 40688390
  15. Han X, Ai X, Sun Y, et al. Comparative efficacy of pharmacologic therapies for MASLD in improving fibrosis: systematic review and network meta-analysis. Eur J Gastroenterol Hepatol. 2026;38(4):407-415. PMID 41811770
  16. Mantovani A, Csermely A, Bilson J, et al. Association between primary hypothyroidism and metabolic dysfunction-associated steatotic liver disease: an updated meta-analysis. Gut. 2024;73(9):1554-1561. PMID 38782564
  17. Amdetsion GY, Pan CW, Tebeje H, et al. Association Between Subclinical Hypothyroidism and MASLD: A Systematic Review and Meta-Analysis. Int J Hepatol. 2025;2025:8133686. PMID 41357781
  18. Yuan S, Ebrahimi F, Bergman D, et al. Thyroid dysfunction in MASLD: Results of a nationwide study. JHEP Rep. 2025;7(5):101369. PMID 40342634
  19. Zhang YH, Xie R, Dai CS, et al. Thyroid hormone receptor-beta agonist HSK31679 alleviates MASLD by modulating gut microbial sphingolipids. J Hepatol. 2025;82(2):189-202. PMID 39181210