A peer-reviewed Phase 2 trial explains why the oral GLP-1 candidate TERN-601 was discontinued: modest weight loss, high dropout, and three delayed liver-enzyme cases.
60-Second Clinical Brief
Research · September 24, 2026A peer-reviewed Phase 2 trial explains why the oral GLP-1 candidate TERN-601 was discontinued: modest weight loss, high dropout, and three delayed liver-enzyme cases.
TERN-601 did what an oral GLP-1 pill is supposed to do: it reduced appetite, slowed gastric emptying, and moved body weight down. Then three participants developed a liver signal serious enough to stop the program.
That is the uncomfortable lesson in the newly published Phase 1 and Phase 2 data. The drug was not simply beaten by a stronger competitor. Its benefit-risk equation failed. In the best-performing Phase 2 group, the placebo-adjusted weight change was 4.6% after 12 weeks. At the same time, two cases were judged consistent with drug-induced liver injury, and a third could not be cleared of that possibility.
This matters beyond one discontinued asset. TERN-601 belongs to the fast-growing field of oral, non-peptide GLP-1 receptor agonists. These are small molecules designed to activate the same receptor pathway as injectable peptide drugs while avoiding injections, cold-chain logistics, and the absorption limits of oral peptides. They are convenient in theory. Their chemistry also creates risks that cannot be inferred from injectable semaglutide or tirzepatide.
FALCON was a 12-week, randomized, double-blind, placebo-controlled Phase 2 study conducted at 17 outpatient sites in the United States. It enrolled 167 adults with obesity or overweight without diabetes. Of those, 134 received TERN-601 across four dosing schedules and 33 received placebo. Participants started at 100 mg daily and escalated toward 250 mg, 500 mg, a slower 500 mg schedule, or 750 mg. Unlike oral semaglutide, dosing did not have to be coordinated with food.
The trial population was mostly female, at roughly 79% in both the active and placebo groups. Participants receiving TERN-601 had mean baseline body weights around 99 to 102 kilograms and mean body mass indexes around 36 to 37 kg/m². The primary endpoint was percentage change in body weight at week 12.
The results were dose-related, but not neatly so. Mean weight changes from baseline were 2.1% at 250 mg, 3.9% with slow titration to 500 mg, 4.9% at 500 mg, and 3.3% at 750 mg. Placebo participants lost 0.3%. That produced placebo-adjusted differences of 1.8%, 3.6%, 4.6%, and 3.0%, respectively. The 250 mg comparison narrowly missed the study’s stated statistical threshold; the higher-dose schedules met it.
The 750 mg group did not outperform 500 mg. That matters because increasing exposure did not buy a cleaner efficacy story. At week 12, 46.4% of the standard 500 mg group achieved at least 5% weight loss, compared with 39.3% in the slow-titration 500 mg group, 25.0% at 750 mg, 14.8% at 250 mg, and 15.4% with placebo. These are short-term results, not a projection of one-year weight loss.
Most adverse events looked familiar for a GLP-1 study. Nausea, vomiting, and other gastrointestinal effects were common and generally mild to moderate. But treatment persistence was weak. Only 77.2% of randomized participants completed treatment. Early discontinuation occurred in 24.6% of those assigned to TERN-601 and 15.2% of those assigned to placebo. The paper reports that about 12% of active-treatment participants stopped because of adverse events, mainly gastrointestinal ones.
The decisive issue appeared in liver blood tests. Three participants receiving TERN-601 developed elevations in alanine aminotransferase or aspartate aminotransferase beginning at week 8 or later. The peaks occurred several weeks after the last dose, which is one reason short exposure studies can miss a delayed signal. Two cases occurred in the 500 mg group and one in the 750 mg group.
An independent expert from the Drug-Induced Liver Injury Network reviewed all three. Two profiles were considered consistent with potential drug-induced liver injury; one was supported by liver biopsy. The third participant had gallstones as a possible alternative explanation, but drug-induced injury could not be ruled out. One participant met the pattern of a potential Hy’s Law case, a regulatory warning sign involving liver-cell injury together with elevated bilirubin when another cause is not evident.
There is an important piece of context: all transaminase and bilirubin elevations resolved without treatment. No participant had an alkaline phosphatase value above 1.5 times the upper limit of normal, and no active-treatment participant had ALT or AST above 10 times the upper limit. Resolution is reassuring for those individuals. It does not make the signal acceptable for a chronic obesity medicine that must compete against several effective alternatives.
The affected participants did not simply have extreme drug exposure. In the two cases with pharmacokinetic substudy data, exposure was close to the group mean. Preclinical testing and the 28-day Phase 1 study had not predicted the toxicity. That removes two easy explanations and makes patient selection harder.
TERN-601 is a small molecule, not a peptide. It activates the GLP-1 receptor, but its structure, metabolism, and off-target interactions differ from peptide-based GLP-1 medicines. A safety problem tied to one chemical scaffold should not be casually transferred to semaglutide, tirzepatide, or every oral candidate.
Still, TERN-601 is not an isolated warning. Lotiglipron, another oral small-molecule GLP-1 agonist, was stopped after liver-enzyme elevations emerged in Phase 2. Its published trial included 901 treated participants. Transaminase elevations occurred in 6.6% of lotiglipron-treated participants in the diabetes cohort and 6.0% in the obesity cohort, and investigators could not identify a reliable at-risk subgroup. Danuglipron development also ended after efficacy, tolerability, and liver-safety concerns.
The TERN-601 authors point out that these three molecules share a similar core scaffold or pharmacophore. That supports a plausible chemistry-specific problem, but it does not prove a mechanism. The trial was too small and too short to map exactly which structural feature caused injury. The correct conclusion is narrower: repeated hepatotoxicity signals in related molecules demand stronger screening and longer observation before a new oral GLP-1 is treated as interchangeable with established medicines.
The wider field is not uniformly failing. In a 2026 Phase 2b trial, aleniglipron produced placebo-adjusted mean weight reductions of 8.2% to 11.3% at week 36 across three doses. Investigators reported no drug-induced liver injury events, although treatment-related discontinuation across active groups was 10.4%. That does not guarantee long-term liver safety, but it separates aleniglipron’s published profile from TERN-601’s.
HRS-7535 also produced weight loss in a 26-week Chinese Phase 2 obesity trial. At 180 mg, mean weight change was 9.36%, compared with 2.50% for placebo. A separate 16-week diabetes trial reported no ALT or AST elevations above three times the upper limit of normal. Again, that is not a clean bill of health for every population or exposure duration. It shows why “oral GLP-1” is a delivery and target description, not a shared safety identity.
The commercial bar is also rising. A once-daily pill must offer enough weight loss, manageable gastrointestinal effects, and a convincing safety profile to compete with established injections and approved oral options. TERN-601’s strongest placebo-adjusted result was meaningful for 12 weeks, but modest relative to that bar. The 24.6% active-treatment discontinuation rate and delayed liver signal made further development hard to justify.
TERN-601 is not an available treatment, and the published data do not support using it through research-chemical sellers or underground peptide channels. The development program was discontinued. There is no approved dose, no clinical monitoring protocol, and no legitimate risk-benefit case for self-experimentation.
For researchers, the case argues for longer liver monitoring, including follow-up after dosing ends. The first enzyme elevations began at week 8 or later, and peaks appeared weeks after the final dose. A short first-in-human study can therefore look clean while missing the clinically relevant window. Future programs built on related chemistry need to explain why their molecules should behave differently, rather than relying only on receptor selectivity or animal toxicology.
For clinicians, the lesson is not to order extra liver tests for every person taking an approved GLP-1 solely because TERN-601 failed. Monitoring should follow the product label, the patient’s medical history, symptoms, and clinical judgment. But unexplained jaundice, dark urine, severe fatigue, persistent right-upper-abdominal pain, or abnormal liver tests always deserve medical assessment. Patients should not stop prescribed medication without speaking with their clinician.
The most useful drug-development stories are not always the winners. TERN-601 confirms that oral small molecules can activate the GLP-1 pathway and produce short-term weight loss without food restrictions. It also shows why convenience is not enough. A chronic obesity drug has to survive prolonged exposure in a broad population, where a low-frequency toxicity can outweigh a moderate efficacy gain.
The signal is strong enough to explain why the program stopped, but not broad enough to condemn oral GLP-1 medicines as a class. The next question is structural: can developers preserve oral receptor activation while moving away from the chemistry shared by TERN-601, lotiglipron, and danuglipron? Candidates that answer that question with longer randomized data, not just clean Phase 1 results, will deserve attention.
Full open-access TERN-601 study at PubMed Central.
ClinicalTrials.gov record NCT06854952.
Amin NB et al. Lotiglipron Phase 2 study. PMID 39415344.
Rosenstock J et al. Aleniglipron Phase 2b trial. PMID 42249138.
Gu W et al. HRS-7535 obesity Phase 2 trial. PMID 42331800.
Guo L et al. HRS-7535 diabetes randomized trial. PMID 42234428.
Evidence grade: B-. Randomized, placebo-controlled human evidence supports both the weight-loss effect and the liver-safety signal. Confidence is reduced by the 12-week duration, modest sample size, sponsor involvement, available-case efficacy analysis, and only three potential liver-injury cases.
Medical disclaimer: This article is for education and research context only. It is not medical advice and does not recommend TERN-601 or any unapproved drug. TERN-601 development was discontinued. Decisions about approved GLP-1 medicines, liver testing, symptoms, or treatment changes should be made with a licensed clinician. Oria exists to help readers separate early research signals from established clinical evidence.
