FGF21
The hepatic peptide hormone of fasting and ketogenic states — and the target behind the most exciting current NASH/MASH pipeline. Native FGF21 has a half-life measured in hours; the clinical drugs are engineered Fc-fusions and PEGylated variants dosed weekly.
What is FGF21?
FGF21 is a 181-amino-acid member of the fibroblast growth factor superfamily, produced predominantly by the liver under conditions of metabolic stress — prolonged fasting, ketogenic diets, protein restriction, mitochondrial dysfunction, and chronic alcohol exposure. Unlike canonical FGFs, it lacks the heparin-binding domain and functions as an endocrine (rather than paracrine) hormone, circulating and acting on distant tissues. It signals through FGFR1c paired with the co-receptor β-Klotho, expressed in adipose tissue, brain, pancreas, and liver. Metabolic effects include increased insulin sensitivity, hepatic fat reduction, increased adiponectin, and — uniquely among metabolic hormones — central effects on macronutrient preference (reduced sweet-food preference and reduced alcohol intake). Native FGF21 has a circulating half-life under two hours, so therapeutic development has focused on engineered long-acting analogs: efruxifermin (Fc-fusion), pegozafermin (PEGylated), aldafermin (FGF19 analog, mechanistically adjacent), BOS-580, and PF-05231023. The NASH/MASH pipeline is the leading clinical application, and in 2025 it consolidated into large pharma: Novo Nordisk acquired Akero Therapeutics (efruxifermin), Roche acquired 89bio (pegozafermin), and GSK took over efimosfermin alfa, formerly BOS-580, from Boston Pharmaceuticals. All three are now in Phase 3.
What FGF21 Is Investigated For
FGF21 has one of the strongest pipelines in contemporary metabolic medicine — the HARMONY Phase 2b trial of efruxifermin in NASH/MASH reported positive fibrosis-improvement data in 2023, with a 96-week follow-up in 2025 in which the 50 mg dose held its advantage and the 28 mg dose did not, and pegozafermin has shown similar results in MASH plus a completed Phase 3 in severe hypertriglyceridemia. The record is not uniform: the dedicated SYMMETRY trial of efruxifermin in compensated cirrhosis missed its 36-week primary endpoint, leaving the class's reach into F4 disease unsettled. FGF21's uniqueness among metabolic hormones is its central action on macronutrient preference: the rodent data showing reduced sweet-food and reduced ethanol intake has translated, in early clinical work, to intriguing signals in alcohol-use-disorder populations. None of this translates to a legitimate self-administered use pattern — the clinically relevant agents are engineered biologics with weekly parenteral dosing, not available on the compounded-peptide market. Research-grade native FGF21 sold online is not the engineered therapeutic variant. The page is positioned as a biology and pipeline reference, not a dosing guide.
History & Discovery
FGF21 was identified and cloned around the turn of the 2000s as a novel member of the fibroblast growth factor superfamily with unusual properties — it lacked the heparin-binding domain that restricts canonical FGFs to paracrine signaling, it was produced mainly by the liver, and it had no obvious developmental or mitogenic role. Its biological function was unclear until Kharitonenkov and colleagues at Eli Lilly published the 2005 JCI paper demonstrating potent insulin-sensitizing and glucose-lowering effects in diabetic rodent models and in human adipocytes. Non-human primate data came later, with Pfizer's long-acting PF-05231023 in obese cynomolgus monkeys. That 2005 paper launched the therapeutic pipeline. The mechanistic biology was consolidated over the next few years. Inagaki and colleagues (2007 Cell Metabolism) showed that FGF21 is induced in the liver during fasting and ketogenic states through PPARα-driven transcription, positioning it as a key endocrine mediator of the adaptive starvation response. The β-Klotho co-receptor was identified as obligatory for FGF21 signaling, clarifying the tissue-specificity pattern. CNS effects on sweet-food and alcohol preference were mapped to hypothalamic β-Klotho-expressing neurons. The first-in-human trial of LY2405319 (an early Eli Lilly engineered variant) by Gaich and colleagues in 2013 confirmed metabolic effects in humans. Subsequent development focused on engineered long-acting variants with weekly dosing pharmacokinetics — efruxifermin (Akero's bivalent Fc-fusion), pegozafermin (89bio's PEGylated variant), aldafermin (NGM's FGF19 analog, mechanistically adjacent), BOS-580, and PF-05231023. The NASH indication emerged as the leading clinical target because of the combination of hepatic effects, metabolic benefit, and a regulatory context (FDA-endorsed surrogate endpoints) favorable to fibrosis-focused trials. The HARMONY Phase 2b trial of efruxifermin, read out in 2023, was the most consequential FGF21 clinical data to date, with the 96-week extension in 2025 supporting durability. Severe hypertriglyceridemia has emerged as a parallel indication for pegozafermin. Alcohol-use-disorder trials are an emerging area of interest.
How It Works
FGF21 is the hormone your liver makes when it thinks food is scarce — it rises in fasting, on keto diets, and in protein restriction. It tells adipose tissue to burn fat, tells the liver to make less fat, raises adiponectin (a marker of metabolic health), and — unusually — acts on the brain to reduce cravings for sweets and, in animal studies, for alcohol. The drugs in development (efruxifermin, pegozafermin) are engineered long-acting versions that exploit this biology for fatty liver disease. Native FGF21 is not used therapeutically because it breaks down too quickly.
FGF21 signals through FGFR1c (with lower-affinity activity at FGFR2c and FGFR3c) paired with the obligatory co-receptor β-Klotho, which is expressed in adipose tissue, pancreas, liver, and discrete CNS regions — most relevantly the hypothalamic paraventricular nucleus and suprachiasmatic nucleus. The tissue expression pattern of β-Klotho determines which tissues respond to circulating FGF21, distinguishing it from the canonical paracrine FGF signaling mode. Hepatic FGF21 transcription is driven by PPARα under fasting / ketogenic conditions and by ATF4 under mitochondrial stress and amino-acid imbalance. Peripheral effects include direct action on adipocytes to increase glucose uptake and adiponectin secretion (FGF21 elevates adiponectin robustly — this is one of its most reliable pharmacodynamic biomarkers), lipolysis regulation, and white-to-beige adipocyte browning (rodent data is clearer than human). Hepatic effects include reduced de novo lipogenesis and increased fatty-acid oxidation. CNS effects mediate suppression of sweet-food preference (via β-Klotho-expressing neurons, with reduced dopamine in the nucleus accumbens) and reduced alcohol consumption. The alcohol circuit has been localised twice: to a projection-specific subpopulation of β-Klotho neurons in the basolateral amygdala, and to a paraventricular-nucleus oxytocin to ventral-tegmental-area dopamine loop that lengthens the interval between drinks. In alcohol-preferring non-human primates an FGF21 analogue halved alcohol intake. In obesity and NAFLD, serum FGF21 is paradoxically elevated — interpreted by some as a state of FGF21 resistance analogous to leptin resistance. Whether elevated baseline FGF21 in obese patients blunts response to therapeutic FGF21 analogs is partially supported by pharmacodynamic data but does not prevent therapeutic effect, because therapeutic exposures are far above physiologic levels and override the apparent resistance.
Evidence Snapshot
Human Clinical Evidence
Moderate-to-strong for NASH/MASH: efruxifermin's HARMONY Phase 2b met its 24-week fibrosis endpoint at both doses (RR 2.3 and 2.2, p=0.025 and p=0.036) and at 96 weeks only the 50 mg dose separated from placebo (49% vs 19%, p=0.0030; the 28 mg arm reached 30%, p=0.19). The same agent's SYMMETRY trial in compensated (F4) cirrhosis missed its 36-week primary endpoint outright, with only a nominal 96-week signal (29% vs 11% fibrosis reduction, difference 16 percentage points, 95% CI 2 to 30). Pegozafermin showed comparable Phase 2b results in pre-cirrhotic MASH. The honest summary is that this is the field's most convincing antifibrotic signal, not an unbroken run of wins. A 2026 meta-analysis of 10 randomized trials (1,113 patients) put the class-level effect at roughly a doubling of ≥1-stage fibrosis improvement versus placebo (RR ~2.25), and a 2026 pooled analysis found efruxifermin's benefit was consistent across PNPLA3 genotypes — an important point given PNPLA3 is the strongest common genetic risk factor for MASH progression. Moderate for severe hypertriglyceridemia. Emerging for alcohol-use disorder. Weak for longevity or healthy-adult metabolic intervention — no trial in those populations has read out meaningfully.
Animal / Preclinical
Strong. FGF21 biology is extensively characterized across rodent and non-human-primate models. The discovery papers (Kharitonenkov 2005) and mechanistic work on fasting biology (Inagaki, Reitman) anchor a well-understood pathway.
Mechanistic Rationale
Strong. FGFR1c / β-Klotho signaling, PPARα-driven hepatic induction, and the adipose/hepatic/central action triad are thoroughly characterized.
Research Gaps & Open Questions
What the current literature has not yet settled about FGF21:
- 01Long-term (multi-year) safety of chronic FGF21-pathway stimulation, particularly bone-density effects.
- 02Whether fibrosis improvements on FGF21 therapy translate to reduced cirrhosis and hepatocellular carcinoma incidence over years — the definitive endpoint still awaits Phase 3 readouts.
- 03Whether CNS effects on alcohol and sweet-food preference translate to clinically meaningful use cases at doses tolerable for chronic administration.
- 04Whether combination with GLP-1 agonists produces synergistic rather than additive metabolic benefit in NASH and metabolic syndrome — early preclinical work on unimolecular GLP-1/FGF21 dual agonists shows complementary effects on weight, glycemia, and hepatic fibrosis in diet-induced steatohepatitis models, but no human combination-therapy outcome data exist yet.
- 05Whether the FGF21 benefit extends to compensated (F4) cirrhosis at all — the one dedicated trial (SYMMETRY) missed its 36-week primary endpoint, and the Phase 3 cirrhosis programmes do not reach primary completion until 2028 (pegozafermin) and 2030 (efruxifermin).
- 06Why obesity produces elevated baseline FGF21 — is this compensatory, dysregulatory, or incidental, and does it predict therapeutic response?
- 07Whether engineering of FGF21 analogs can separate the liver-beneficial actions from the bone-turnover signals seen in some programs.
- 08Role in longevity pathways — endogenous FGF21 rises in caloric restriction and ketogenic states, but whether chronic pharmacologic FGF21 elevation in healthy adults extends lifespan (as rodent CR models would predict) has not been tested.
Forms & Administration
Native FGF21 is not administered therapeutically — its half-life is too short (under 2 hours) for practical use. The engineered clinical analogs are subcutaneous injectables: efruxifermin is a bivalent Fc-fusion dosed weekly in NASH trials; pegozafermin is a PEGylated variant dosed weekly or every two weeks; aldafermin is mechanistically adjacent (FGF19 analog). Efimosfermin alfa (formerly BOS-580) is a long-acting Fc-fusion dosed monthly rather than weekly, now owned by GSK and in Phase 3 MASH trials. PF-05231023 was discontinued, and aldafermin's only registered Phase 2/3 was withdrawn in 2024, so neither is an active programme. None are FDA-approved as of 2026. Research-chemical 'FGF21' sold online is native protein, not equivalent to the engineered therapeutic variants, and should not be presumed to reproduce the clinical-drug pharmacology.
Dosing & Protocols
The ranges below reflect protocols commonly discussed in the literature and by clinicians — not a prescription. Actual dosing for any individual should be determined by a qualified healthcare provider who knows the patient.
Typical Range
No FDA-approved FGF21 therapy exists. Clinical-trial dosing: efruxifermin 28 mg or 50 mg once weekly SC in NASH trials; pegozafermin 15 mg or 30 mg weekly or 44 mg every two weeks in NASH and severe-hypertriglyceridemia trials; efimosfermin alfa monthly SC in Phase 3 MASH trials; aldafermin 1 mg daily SC in NASH (FGF19 analog, not FGF21 proper — and a dormant programme).
Frequency
Once weekly SC injection for efruxifermin-class Fc-fusion analogs; weekly or every-two-weeks for pegozafermin-class PEGylated variants. The engineered half-lives (multiple days to weeks) support these regimens. Native FGF21 at 1–2 hour half-life would require multiple-times-daily dosing.
Timing Considerations
Time of day
Not time-of-day specific — the engineered variants are weekly or biweekly and pharmacokinetically span multiple days.
Relative to meals
Not meal-timed in the clinical protocols; SC absorption is not meaningfully affected by food intake.
Relative to exercise
Not tied to exercise timing.
Weekly dosing schedules are typically fixed to a day of the week for convenience; no pharmacodynamic benefit to specific times has been established.
Protocol Notes
There is no legitimate at-home or compounded-peptide pathway for FGF21-class agents. The engineered therapeutics are investigational biologics administered in clinical trials. Research-grade native FGF21 sold online is not equivalent — the pharmacokinetic profile is fundamentally different, and its identity and purity are unverified. For readers interested in the biology as a guide to dietary intervention: endogenous FGF21 rises robustly during prolonged fasting and strict ketogenic-diet adherence, driven by PPARα. Whether dietary-induced FGF21 elevation produces metabolic benefits independent of the weight-loss effect of those interventions is unresolved. The engineered therapeutics produce sustained exposures far above anything the fasting response generates — a far more intense pharmacology than dietary manipulation can reproduce.
FGF21 analogs are investigational biologics administered under clinical trial protocols. Nothing on this page constitutes a suggested self-use protocol, and no compounded-peptide source legitimately reproduces the engineered long-acting therapeutics.
Timeline of Effects
Onset
Metabolic biomarker responses (adiponectin elevation, triglyceride reduction) emerge within days to the first week of dosing in trial participants. Hepatic fat reduction on MRI-PDFF is typically detectable by 4–8 weeks. Fibrosis improvement on paired biopsy requires 24+ weeks of continuous dosing.
Peak Effect
Hepatic fat reduction accrues over the first few months of continuous dosing and is the basis for the 12–16 week MRI-PDFF assessments used in trials. Fibrosis improvement continues to accrue through week 96 in the HARMONY extension — though only at the 50 mg dose, which was the only arm to separate from placebo at that timepoint. Adiponectin elevation plateaus within the first weeks and is the cleanest early signal that the drug is engaging its target.
After Discontinuation
Hepatic fat partially rebounds within weeks of discontinuation — the effect is pharmacologic and reversible, not disease-modifying in a durable sense without continued dosing. Fibrosis benefit accrual stops with discontinuation; whether established fibrosis improvement reverses on stopping is still being characterized in trial extensions.
Monitoring & Measurement
Bloodwork & Labs
- •Serum adiponectin — reliably rises on FGF21 therapy and is a well-validated pharmacodynamic marker
- •Serum FGF21 (ELISA) — elevated in NASH, fasting, and mitochondrial disease; rises further on engineered-analog therapy
- •Liver enzymes (ALT, AST) — decrease on FGF21-pathway therapy in NASH
- •Fasting triglycerides, HDL — respond favorably to FGF21-class agents
- •Bone-turnover markers (CTX, P1NP) — monitored in longer trials because of bone-density concerns
Functional & Performance Tests
- •MRI-PDFF (proton-density fat fraction) — primary imaging biomarker for hepatic steatosis
- •MR elastography or FibroScan — non-invasive fibrosis assessment
- •Liver biopsy with paired pre/post specimens — the regulatory endpoint for NASH fibrosis claims
When to Test
Baseline before initiation, with adiponectin and liver enzymes reassessed at 4–8 weeks, MRI-PDFF at 12–16 weeks, paired biopsy at 24+ weeks in NASH trials. Bone-turnover markers at baseline and every 6 months in longer-duration trials.
Interpretation & Notes
Serum FGF21 measurement is not clinically informative for most patients — absolute levels vary widely with fasting state, metabolic health, mitochondrial dysfunction, and alcohol intake. Adiponectin, by contrast, is the cleanest pharmacodynamic marker of FGF21-analog engagement: it rises reliably and dose-dependently, is easy to measure, and has well-characterized interpretation. MRI-PDFF is the gold-standard imaging biomarker for hepatic steatosis response, though a 2026 secondary analysis of the pegozafermin Phase 2b set the useful bar higher than the conventional 30% decline: only a 'super response' of at least a 50% relative fall in liver fat independently predicted fibrosis regression without worsening MASH (adjusted OR 3.08, 95% CI 1.32–7.19), while the 30% threshold did not. Liver biopsy remains the regulatory gold standard for fibrosis claims and is required for pivotal NASH endpoints.
Common Questions
What is FGF21's claim to fame in NASH / MASH?
FGF21 analogs produce reductions in hepatic fat content (MRI-PDFF), liver enzymes (ALT, AST), and — importantly — fibrosis on paired biopsy, across multiple Phase 2 and Phase 2b trials. The HARMONY trial of efruxifermin showed fibrosis improvement at 24 weeks for both doses and at 96 weeks for the 50 mg dose only. A parallel trial in compensated cirrhosis (SYMMETRY) missed its 36-week endpoint, so how far the class reaches into F4 disease is still unsettled. The mechanism combines direct hepatic lipogenesis suppression, increased fatty-acid oxidation, improved insulin sensitivity, and adiponectin elevation — a multi-pronged metabolic profile that may explain why FGF21 has succeeded where single-target NASH candidates have disappointed.
Why does FGF21 reduce alcohol and sweet-food preference?
FGF21 signals through β-Klotho on hypothalamic paraventricular-nucleus neurons and related central circuits to suppress preference for sweet macronutrients and, in rodent and early human work, ethanol consumption. The central actions are mechanistically distinct from the peripheral metabolic effects on liver and adipose tissue. Whether this translates to a therapeutic use case in alcohol-use disorder is under investigation — the preclinical signal is strong but human trial data remain early.
Does endogenous FGF21 rise with fasting or keto diets?
Yes, robustly. Serum FGF21 rises during prolonged fasting, ketogenic-diet adherence, and protein restriction, with the hepatic induction driven principally through PPARα activation. The rise is a marker of the hepatic adaptive-starvation response. Whether any dietary strategy meaningfully 'boosts' therapeutic FGF21 signaling is unresolved — engineered therapeutics drive exposures far above anything dietary manipulation produces, which is also the reason the apparent FGF21 resistance of obesity does not prevent them working.
Is there human evidence that FGF21 affects drinking, or is it all rodent data?
There is human evidence, but it is genetic rather than interventional. A 2026 Mendelian-randomisation study in the Journal of Hepatology found that genetically higher FGF21 protein levels were associated with lower problematic alcohol use and a lower risk of alcohol-related liver disease, with receptor analyses implicating hippocampal FGFR3 and basal-ganglia KLB expression — a liver-brain axis consistent with the rodent circuit work. That is a strong argument for running the trial, not a substitute for one: Mendelian randomisation tells you about lifelong small differences in a protein, not about what a weekly injection does to a person already drinking heavily. No FGF21 analogue has completed a randomised trial with an alcohol-consumption endpoint.
Can I buy FGF21 as a peptide for personal use?
No legitimate path. Research-grade FGF21 sold online is native 181-AA protein with a half-life of 1–2 hours, not the engineered Fc-fusion or PEGylated analogs that provide weekly-dosing pharmacokinetics in clinical trials. Even setting aside identity and purity concerns, the pharmacology of research-chemical FGF21 would not reproduce the clinical-drug profile. The engineered analogs are not on the compounded-peptide market.
Who FGF21 Is NOT For
- •No approved indication — investigational only. Clinical trial enrollment is the only legitimate pathway.
- •Active malignancy — FGF signaling has complex roles in cancer biology and is typically an exclusion criterion in trial protocols.
- •Pregnancy and lactation — insufficient data.
- •Patients with significant osteoporosis or recent fracture — bone-turnover signals in some programs prompt caution.
- •Known hypersensitivity to the specific agent's excipients or carrier proteins.
Drug & Supplement Interactions
Formal drug interaction studies with FGF21 analogs are limited because all agents are investigational. Theoretical and trial-protocol considerations: combination with GLP-1 receptor agonists is an active research question, but the evidence that the two are complementary rather than antagonistic is preclinical — unimolecular GLP-1/FGF21 dual agonists improve weight, glycaemia and hepatic fibrosis in diet-induced steatohepatitis models — not clinical. Any effect of FGF21 on thyroid hormone metabolism is theoretical at present and has not been characterized at trial doses. Glucose-lowering medications (insulin, sulfonylureas) may plausibly require dose reduction given FGF21's insulin-sensitizing effect, by analogy with GLP-1 agonists, although the Phase 1 multiple-ascending-dose study of PF-05231023 in type 2 diabetes found no significant effect on glycaemic control. No clinically characterized pharmacokinetic interaction has been reported for any FGF21 analogue.
Safety Profile
Common Side Effects
Cautions
- • Bone-density monitoring is a consideration in long-term trials. The concern is specific, not vague: FGF21 gain of function decreases bone mass in mice by inhibiting osteoblastogenesis and driving marrow stem cells toward adipogenesis through potentiation of PPAR-γ, and in humans the long-acting analogue PF-05231023 produced dose-dependent changes in multiple markers of bone formation and resorption
- • Not appropriate in patients with active malignancy without oncology input
- • Interactions with other metabolic agents (GLP-1 agonists, thyroid hormones) not fully characterized
- • Pregnancy — insufficient human data across the pipeline
What We Don't Know
The long-term (multi-year) consequences of chronic FGF21-pathway stimulation are not yet characterized — the longest trial exposures are under two years. Whether the bone-turnover signals observed in some Phase 2 programs translate to clinically meaningful bone-density loss over years is unresolved. Whether the central alcohol- and sweet-preference effects observed in rodents translate robustly to humans at therapeutic doses awaits definitive trials.
Legal Status
United States
No FGF21-pathway agent is FDA-approved as of 2026. Efruxifermin, pegozafermin, aldafermin, BOS-580, and PF-05231023 are investigational. The FDA granted pegozafermin Breakthrough Therapy Designation in NASH in September 2023, on the strength of the ENLIVEN Phase 2b data. Ownership of the leading programmes consolidated into large pharma over 2025: Novo Nordisk completed its acquisition of Akero Therapeutics (efruxifermin) on 9 December 2025, Roche acquired 89bio (pegozafermin) and is now the sponsor of record on new pegozafermin trials, and GSK is the sponsor of the Phase 3 programme for efimosfermin alfa, acquired from Boston Pharmaceuticals. Research-chemical 'FGF21' is not an FDA-regulated therapy.
International
Status parallels the US — investigational across the EU, UK, Japan, and other major markets. NASH / MASH regulatory frameworks differ slightly across regions but no approval has emerged as of 2026.
Sports & Competition
WADA's 2027 Prohibited List names fibroblast growth factors (FGFs) explicitly under S2.3, Growth Factors and Growth Factor Modulators, alongside hepatocyte growth factor and IGF-1 — a class prohibited at all times, in and out of competition. An engineered FGF21 analogue sits squarely inside that entry; it is not a borderline case requiring the catch-all clause. The practical reality is that no FGF21-class agent is commercially available, and access is trial-restricted.
Regulatory status changes over time. Verify current local rules with a qualified professional.
Myths & Misconceptions
Myth
FGF21 is a fasting-mimetic peptide you can inject.
Reality
Native FGF21 has a half-life under two hours and would require impractical dosing frequency. The 'fasting-mimetic' framing oversimplifies — FGF21 is one component of a complex fasting adaptive response that also includes shifts in insulin, glucagon, cortisol, ketone bodies, autophagy, and more. Pharmacologic FGF21 monotherapy does not reproduce the full fasting state.
Myth
FGF21 is already approved for fatty liver disease.
Reality
As of 2026, no FGF21-pathway agent has FDA approval for NASH/MASH or any other indication. Efruxifermin, pegozafermin and efimosfermin alfa are all in Phase 3 development, but no regulatory approval has occurred. Readers sometimes conflate this with resmetirom, which is approved for noncirrhotic MASH with F2–F3 fibrosis — a thyroid-hormone-receptor-beta agonist on an entirely different pathway, not an FGF21 analogue.
Myth
FGF21 causes weight loss on the scale of GLP-1 agonists.
Reality
FGF21 analogues produce only modest weight loss compared with the 10–20% seen with semaglutide and tirzepatide — and the interesting part is that they induce fibrosis regression anyway, which is why combination strategies with incretins are being discussed rather than head-to-head substitution. The FGF21 weight-loss effect is secondary to its metabolic-improvement and hepatic-fat-reduction profile, not a primary indication. The NASH and hypertriglyceridemia indications do not depend on weight loss.
Myth
FGF21 reverses cirrhosis.
Reality
The one dedicated trial says otherwise. SYMMETRY tested efruxifermin in compensated (F4) cirrhosis caused by MASH and missed its 36-week primary endpoint — 18% and 19% of treated patients had fibrosis reduction without MASH worsening versus 13% on placebo, with confidence intervals spanning zero. A separation did appear at 96 weeks in the 50 mg arm (29% vs 11%), but it was nominal, in a Phase 2b trial that had already failed its primary. Phase 3 cirrhosis programmes for both efruxifermin and pegozafermin are running; neither reads out before 2028.
Myth
FGF21 blocks alcohol cravings in humans.
Reality
The animal data is strong and unusually well mapped. FGF21 reduces sweet and alcohol preference in mice through central β-Klotho with corresponding falls in nucleus accumbens dopamine; the circuit was later localised to a projection-specific population of β-Klotho neurons in the basolateral amygdala, and separately to an FGF21–oxytocin–dopamine loop from the hypothalamic paraventricular nucleus to the ventral tegmental area. An FGF21 analogue cut alcohol intake by 50% in alcohol-preferring non-human primates. What does not exist is a single randomised human trial with an alcohol-consumption endpoint. The human evidence is one 2026 Mendelian-randomisation analysis — genetic association, not treatment. The scientific interest is real; the clinical claim is ahead of the evidence.
Myth
FGF21 is safer than GLP-1 agonists for metabolic indications.
Reality
Head-to-head comparative safety has not been established. FGF21 analogs have distinct adverse-event profiles — bone-turnover signals, adiponectin elevation, mild GI effects — that differ from GLP-1 side effects (nausea, gastroparesis, gallbladder issues). Each class has trade-offs; the 'safer' framing is marketing language that outruns the evidence.
Published Research
20 studiesThe use of FGF21 analogues in MASH and MASH cirrhosis — metabolic master regulators or liver-specific mechanisms?
Holleboom et al. 2026 JHEP Reports — the best current synthesis of why FGF21 analogues work in MASH, separating the liver-directed mechanisms (suppression of de novo lipogenesis, attenuation of hepatocellular stress, inhibition of stellate-cell activation) from the systemic ones (the FGF21–adiponectin axis, insulin sensitisation, central control of macronutrient preference). Also the clearest statement of the class's defining oddity: potent fibrosis regression despite only modest weight loss.
A unimolecular GLP-1 and FGF21 dual agonist for treatment of metabolic dysfunction-associated steatohepatitis
Fibroblast Growth Factor 21 Analogues Improve Fibrosis in Metabolic Dysfunction-Associated Steatohepatitis: An Updated Systematic Review and Meta-Analysis
Consistent efficacy of efruxifermin across PNPLA3 genotypes in phase 2b trials in MASH
Genetic evidence that FGF21 signaling reduces problematic alcohol use and alcohol-related liver disease
Clinical Utility of a 50% and 30% Decline in MRI-PDFF in Predicting Fibrosis Improvement in MASH
Negative feedback regulation of alcohol ingestion through the FGF21-PVH oxytocin-VTA dopamine system
Safety and efficacy of once-weekly efruxifermin versus placebo in metabolic dysfunction-associated steatohepatitis (HARMONY): 96-week results from a multicentre, randomised, double-blind, placebo-controlled, phase 2b trial
Noureddin et al. 2025 Lancet — the 96-week follow-up from HARMONY. Read it for the dose split rather than the headline: 49% of the 50 mg group achieved ≥1-stage fibrosis improvement without MASH worsening versus 19% on placebo (p=0.0030), while the 28 mg group reached 30% and did not separate (p=0.19). The data behind the ongoing Phase 3 programme, and the reason 50 mg is the dose carried forward.
Efruxifermin in Compensated Liver Cirrhosis Caused by MASH
Noureddin et al. 2025 NEJM — the SYMMETRY trial, and the most important negative result in the FGF21 literature. In compensated (F4) cirrhosis caused by MASH, efruxifermin did not significantly reduce fibrosis at 36 weeks (18% and 19% versus 13% on placebo; p=0.62 and p=0.52). A nominal separation appeared at 96 weeks for the 50 mg dose (29% vs 11%), but the trial had already missed. Anyone citing FGF21 as a cirrhosis treatment should read this one first.
Safety and efficacy of once-weekly efruxifermin versus placebo in non-alcoholic steatohepatitis (HARMONY): a multicentre, randomised, double-blind, placebo-controlled, phase 2b trial
Randomized, Controlled Trial of the FGF21 Analogue Pegozafermin in NASH
The FGF21 analog pegozafermin in severe hypertriglyceridemia: a randomized phase 2 trial
FGF21 suppresses alcohol consumption through an amygdalo-striatal circuit
A Long-Acting FGF21 Molecule, PF-05231023, Decreases Body Weight and Improves Lipid Profile in Non-human Primates and Type 2 Diabetic Subjects
FGF21 Regulates Sweet and Alcohol Preference
The effects of LY2405319, an FGF21 analog, in obese human subjects with type 2 diabetes
Fibroblast growth factor 21 promotes bone loss by potentiating the effects of peroxisome proliferator-activated receptor γ
Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast growth factor 21
FGF21: a missing link in the biology of fasting
FGF-21 as a novel metabolic regulator
Kharitonenkov et al. 2005 Journal of Clinical Investigation — the foundational paper establishing FGF21 as a therapeutic metabolic-regulator candidate, showing that FGF-21 regulates glucose uptake in mouse 3T3-L1 and primary human adipocytes and that therapeutic administration normalised plasma glucose and triglycerides in ob/ob and db/db mice, without mitogenicity, hypoglycaemia or weight gain at any dose tested. The paper that launched the clinical pipeline.
Quick Facts
- Class
- Hepatokine
- Tier
- B
- Evidence
- Moderate
- Safety
- Moderate Data
- Updated
- Oct 2026
- Citations
- 20PubMed
Also known as
Tags
Related Goals
Conditions Discussed
Evidence Score
Clinical Trials
View Clinical TrialsLinks to ClinicalTrials.gov for reference. Listing does not imply endorsement.