Most of the excitement in fatty liver disease this year has belonged to GLP-1 drugs. A different hormone is quietly building its own case.
At The Liver Meeting 2026 in Denver, researchers presented one-year results for efimosfermin, an experimental once-monthly injection owned by GSK that mimics the metabolic hormone FGF21.
The study involved 39 people with metabolic dysfunction-associated steatohepatitis, or MASH, whose livers already had moderate to advanced scarring. Fifteen had compensated cirrhosis, the most advanced stage of fibrosis.
After 52 weeks of treatment, investigators reported substantial reductions in liver fat, lower liver stiffness measurements and favorable changes in blood markers associated with fibrosis.
In one group, average liver stiffness fell by 42.6%. In another, the reported reduction reached 48.5%.
Those numbers are intriguing. They also require an unusually large asterisk.
The 48.5% result came from just two people. More importantly, none of the 39 participants received placebo. Everyone received the experimental drug, and investigators and participants knew which treatment was being administered.
The evidence is human clinical research, but it is not controlled evidence that the drug reversed liver scarring.
That distinction is the central story behind GSK’s latest liver results.
What FGF21 Is and Why GSK Paid $1.2 Billion for Efimosfermin
Fibroblast growth factor 21, abbreviated FGF21, is a metabolic signaling hormone involved in regulating energy and nutrient metabolism.
Its biology differs from GLP-1, the gut-hormone pathway targeted by semaglutide and other incretin medicines.
Efimosfermin alfa, previously known as BOS-580, is an engineered, long-acting FGF21 analog. It is designed for subcutaneous administration once every four weeks.
GSK acquired rights to the drug through its purchase of Boston Pharmaceuticals subsidiary BP Asset IX. The companies announced the agreement in May 2025, and GSK completed the acquisition on July 7, 2025.
The financial commitment was substantial: $1.2 billion upfront, with up to another $800 million in success-based milestone payments.
That transaction placed a potential value of $2 billion on an investigational liver medicine that had not yet completed pivotal Phase 3 testing.
The target disease is MASH, the inflammatory form of metabolic dysfunction-associated steatotic liver disease. It develops when excess liver fat is accompanied by cellular injury and inflammation, sometimes progressing to fibrosis.
Fibrosis is categorized from F0 to F4. F2 represents significant fibrosis, F3 advanced fibrosis, and F4 cirrhosis.
Once cirrhosis develops, the clinical questions become more serious. Researchers want to know not only whether liver fat or laboratory markers improve, but whether treatment can delay liver failure, decompensation, transplantation or death.
That is why GSK’s interest in a hormone acting through a different metabolic pathway matters.
Existing incretin-based therapies are already changing the MASH treatment landscape. Grey Pepper has examined how semaglutide and Rezdiffra compare in real-world fatty liver treatment.
But compensated cirrhosis remains a major unmet need.
FGF21 is not another GLP-1 drug. GSK is paying to test an entirely different hormonal approach to advanced liver disease.
What the New Efimosfermin Phase 2 Study Found
The latest findings were presented at the American Association for the Study of Liver Diseases annual meeting, held in Denver from November 5 through November 9, 2026.
The presentation, abstract 2299, was led by Mazen Noureddin and colleagues and described a separate open-label Phase 2 cohort followed for 52 weeks.
A total of 39 adults participated. Twenty-four had F2 or F3 fibrosis, while 15 had F4 fibrosis, meaning compensated cirrhosis.
The participants’ mean age was 60.6 years, and 51.3% were women.
Researchers administered efimosfermin at 225 mg or 300 mg every four weeks.
Among the F2/F3 participants, 13 were assigned to 225 mg and 11 to 300 mg. Among those with F4 cirrhosis, seven received 225 mg and eight received 300 mg.
The investigators evaluated several measurements, including liver fat on magnetic resonance imaging, liver stiffness using vibration-controlled transient elastography, the Enhanced Liver Fibrosis score, and a blood marker called PRO-C3.
Liver fat was measured using MRI-proton density fat fraction, or MRI-PDFF.
At 52 weeks, the mean relative reduction in MRI-PDFF was 58.5% among seven evaluable F2/F3 participants receiving 300 mg.
Among six evaluable participants with cirrhosis receiving 300 mg, the average relative reduction was 71.1%.
Those are substantial changes in a liver-fat measurement. But the denominators are important: seven and six participants, not the full cohort.
The reported abstract did not provide corresponding liver-fat results for the 225 mg groups.
The liver stiffness results also moved in a favorable direction.
Among participants with F2/F3 fibrosis, mean stiffness decreased by 33.6% in the 225 mg group, based on 11 evaluable participants, and 42.6% in the 300 mg group, based on seven.
Among participants with F4 cirrhosis, stiffness fell by 48.5% in the 225 mg group, based on only two evaluable participants, and by 20.7% in the 300 mg group, based on seven.
That last comparison demonstrates why subgroup sizes matter so much. A result from two people is extremely sensitive to individual variation and cannot provide a stable estimate of treatment efficacy.
It is also not evidence that the lower dose works better in cirrhosis. The groups were too small, and the analysis was not designed to establish superiority between doses.
Other measurements offered supporting signals.
The Enhanced Liver Fibrosis score, which incorporates blood markers associated with extracellular matrix remodeling, decreased across the reported groups.
PRO-C3, a marker associated with type III collagen formation and active fibrogenesis, also declined.
Those changes are consistent with potentially favorable biological effects. However, changes in biomarkers are not interchangeable with direct proof that established scar tissue has disappeared.
Investigators also reported broadly similar drug exposure in participants with and without cirrhosis at the same dose.
That pharmacokinetic finding matters because advanced liver disease can alter drug handling. It does not, by itself, demonstrate comparable efficacy or safety in both populations.
A lower liver stiffness reading is an encouraging measurement. It is not the same as proving that cirrhosis has been reversed.
What the Safety Results Showed, and What They Left Out
Safety is particularly important when an experimental metabolic medicine is administered to people with advanced liver disease.
In the newly reported cohort, no serious adverse events occurred among the 24 participants with F2 or F3 fibrosis.
Among the 15 participants with F4 cirrhosis, three experienced a serious adverse event, representing 20% of that group.
The investigators judged none of those events to be related to efimosfermin.
That is reassuring as far as the reported safety assessment goes. But a determination that an event is unrelated to treatment is an investigator’s assessment, not a guarantee that every possible causal relationship has been excluded.
The abstract summary did not provide a complete breakdown of gastrointestinal adverse events, treatment discontinuations or the specific serious events observed.
Those omissions prevent a comprehensive evaluation of tolerability in the cirrhosis population.
There is more safety information from a separate, earlier randomized Phase 2 study.
That trial enrolled 84 adults with biopsy-confirmed F2 or F3 MASH and randomized them to efimosfermin 300 mg or placebo for 24 weeks.
Its results were published in The Lancet in February 2026.
Adverse events occurred in 67% of participants receiving efimosfermin, compared with 55% receiving placebo.
Most were mild or moderate, and gastrointestinal complaints were among the most frequent events, generally appearing during the early weeks of treatment.
The earlier study did not identify clinically meaningful differences in vital signs or clinically significant severe laboratory abnormalities between the treatment groups.
But the earlier participants had F2 or F3 fibrosis, not established F4 cirrhosis. Its safety results therefore cannot be assumed to apply unchanged to people with more advanced disease.
A 39-person cohort can identify some safety signals. It cannot reliably exclude uncommon or delayed adverse effects.
Three serious adverse events occurred in the cirrhosis group. None was judged treatment-related, but the population was far too small to settle the long-term safety question.
Why the Missing Placebo Group Matters So Much
The study’s most important limitation is not a particular percentage. It is the design.
This was an open-label, uncontrolled cohort. Every participant received efimosfermin, and there was no group receiving a matching placebo injection.
Researchers can therefore describe changes from baseline, but they cannot confidently determine how much of those changes was caused by the experimental drug.
That distinction is especially important for noninvasive measurements of liver disease.
Liver stiffness is influenced by fibrosis, but it can also change because of inflammation, congestion and other physiological factors.
A reduction in stiffness may be consistent with reduced scarring. It does not, by itself, establish histological regression of scar tissue.
Similarly, MRI-PDFF measures liver fat rather than the amount of collagen deposited in fibrotic tissue.
A liver can lose a substantial amount of fat without necessarily reversing established cirrhosis.
Changes in ELF and PRO-C3 provide additional biological context, but they remain indirect markers.
Without a placebo group, variation in disease activity, changes in other care, regression toward average measurements and chance can all complicate interpretation.
The very small subgroups add another problem. The reported 48.5% reduction in stiffness among cirrhosis patients on 225 mg came from two evaluable individuals.
The finding deserves to be reported, but not treated as an established treatment effect.
The conference report also did not provide confidence intervals or formal comparative significance tests for the 52-week stiffness outcomes.
Nor did it report follow-up liver biopsy results demonstrating fibrosis-stage regression in this cohort.
As of this report, the findings have been presented as a conference abstract rather than as a complete peer-reviewed publication detailing the 52-week cohort.
Several authors were affiliated with GSK or Boston Pharmaceuticals. Those relationships are relevant to the transparency and interpretation of company-supported research, although they do not invalidate the measurements.
The most important distinction is that efimosfermin has previously undergone placebo-controlled testing. The absence of a placebo group applies specifically to this new 39-person cohort, not to the drug’s entire clinical development program.
In the earlier randomized Phase 2 study, 45.2% of patients in the evaluable biopsy analysis receiving efimosfermin achieved at least one stage of fibrosis improvement without worsening MASH, compared with 20.6% receiving placebo.
MASH resolution without worsening fibrosis occurred in 67.7% versus 29.4%, respectively.
Those findings offer more directly controlled evidence of histological benefit in F2/F3 disease.
They do not automatically prove the same benefit in compensated F4 cirrhosis.
Before a new treatment can claim to reverse cirrhosis, researchers need more than an encouraging imaging signal. They need rigorous comparisons and evidence of outcomes that matter to patients.
What GSK’s NEBULA Phase 3 Trials Need to Prove
GSK has already moved efimosfermin into a larger late-stage development program.
The company lists ZENITH-1 and ZENITH-2 as Phase 3 trials in MASH with F2/F3 fibrosis, alongside NEBULA-1 and NEBULA-2 for compensated MASH cirrhosis.
NEBULA-1, registered as NCT07701993, is a randomized, placebo-controlled, double-blind Phase 3 study in adults with compensated cirrhosis.
The study began in July 2026 and is recruiting participants. Its registry lists an estimated primary completion date of June 23, 2033.
The trial is designed to assess time to an adjudicated composite liver-related clinical outcome, moving beyond changes in liver fat and fibrosis biomarkers alone.
That distinction is crucial.
For people with compensated cirrhosis, the most consequential outcomes include progression to decompensation, serious liver complications, transplantation and survival.
A medicine that improves laboratory markers but fails to prevent clinical deterioration would have a very different value from one that meaningfully changes those outcomes.
The long timeline is not necessarily a sign of trouble. Serious liver complications can take years to develop, and trials designed to measure them require extended observation.
Nevertheless, the 2033 estimate means the definitive answer for NEBULA-1 is still years away, and trial timelines can change.
Efimosfermin is also being investigated in additional liver-disease settings, reflecting GSK’s broader interest in FGF21 signaling.
But no amount of pipeline expansion substitutes for demonstrating that a treatment improves meaningful outcomes in randomized human trials.
As of November 2026, efimosfermin remains investigational and is not an approved treatment for MASH or cirrhosis.
NEBULA must answer the question that a 39-person uncontrolled study cannot: does efimosfermin prevent serious liver complications, rather than merely improve the measurements associated with them?
Why Efimosfermin Matters Beyond the GLP-1 Obesity Race
The broader scientific interest in efimosfermin comes from the possibility of targeting liver disease through pathways other than the familiar incretin receptors.
GLP-1-based medicines have established their ability to reduce body weight, and some have demonstrated benefits in patients with MASH.
Other investigational medicines, including the GLP-1/glucagon dual agonist survodutide, are being evaluated for their effects on liver disease alongside metabolic outcomes.
FGF21 analogs offer a different approach, acting through a metabolic signaling system with potential effects on hepatic fat accumulation and pathways associated with inflammation and fibrosis.
If future randomized trials demonstrate meaningful benefits, FGF21-directed therapies could expand treatment options rather than simply compete on weight-loss percentages.
Researchers may also explore combinations of therapies acting through different biological mechanisms.
However, combining mechanisms remains a research strategy, not proof of superior outcomes or acceptable safety.
There is a commercial story here too.
GSK committed $1.2 billion upfront because the potential market for treatments that prevent advanced liver disease is substantial. Yet the company is still funding studies whose decisive outcomes may not arrive for years.
The latest efimosfermin study adds a useful signal, particularly because it includes patients with compensated cirrhosis who have historically had limited disease-directed treatment options.
It does not establish a commercial product, prove clinical benefit in F4 disease or resolve the question of long-term safety.
The most interesting result may ultimately prove to be the change in liver stiffness. Or it may prove to have been influenced by changes in inflammation and other nonfibrotic processes.
Only better-controlled research can separate those explanations.
For now, efimosfermin’s new liver data are a promising sketch drawn with a thin pencil: 39 people, no placebo, small evaluable subgroups and favorable noninvasive measurements.
The drug’s earlier placebo-controlled results provide a stronger scientific foundation in noncirrhotic MASH, but its most difficult clinical question remains unanswered.
GSK has shown that liver stiffness measurements can fall during monthly efimosfermin treatment. Whether that means scar tissue has meaningfully regressed, and whether patients ultimately avoid liver failure or other complications, is still a question for Phase 3.
Sources and Further Reading
- HCPLive — Once-Monthly Efimosfermin Improves Fibrosis Markers in Phase 2 MASH Study. November 6, 2026. Reports The Liver Meeting 2026 abstract 2299.
- GSK — GSK Showcases Progress in Liver Portfolio at AASLD 2026. November 5, 2026. Official confirmation of abstract 2299 and the Phase 3 program.
- Noureddin M, et al. Efimosfermin Alfa (BOS-580) Once per Month in People With MASH and F2 or F3 Fibrosis: A Randomized, Double-Blind, Placebo-Controlled Phase 2 Trial. The Lancet, 2026;407:794–804.
- The Lancet — Original Efimosfermin Phase 2 Research Publication. DOI: 10.1016/S0140-6736(25)02276-7.
- The Lancet — Department of Error, May 9, 2026. Correction to participant counts shown in figures from the randomized Phase 2 report.
- GSK Clinical Study Register — NEBULA-1 Phase 3 Trial, Study 306300, NCT07701993. Study design, outcomes and planned completion dates.
- GSK — Completion of Efimosfermin Acquisition From Boston Pharmaceuticals. July 2025. Acquisition terms and development strategy.
- ClinicalTrials.gov — Efimosfermin Phase 2 Study Registration, NCT04880031.
