The Endpoint Changes the Story

This week’s research shared a useful theme: the endpoint you choose can change the story you tell. A dual-incretin candidate produced large glucose and weight signals, but in a trial too small to settle comparisons with semaglutide. Oral semaglutide reduced heavy drinking days, but missed its primary craving endpoint. A topical antimicrobial peptide improved the clinical appearance of infected diabetic foot ulcers, while broad bacterial eradication remained uncertain.

That is not a week of failed studies. It is a week of studies that demand better questions. Is an exploratory p-value enough to call one incretin superior? Does fewer severe drinking episodes matter even when average daily drinking does not move? Can a wound treatment be clinically useful before cultures turn negative? The answers may be yes, but only if we keep the design, comparator, population, and follow-up attached to every claim.

We are covering BGM0504, semaglutide, tirzepatide, and peceleganan (PL-5). Together they span systemic metabolic signaling, reward-related behavior, and local antimicrobial action. They also show why protocol thinking should begin with mechanism and evidence quality—not with the largest percentage in an abstract.

BGM0504: a strong metabolic signal, not a superiority verdict

BGM0504 is a once-weekly peptide that activates both GLP-1 and GIP receptors, putting it in the same broad mechanistic class as tirzepatide rather than a single-pathway GLP-1 drug. In a randomized Phase 2 study of 64 Chinese adults with type 2 diabetes, weekly doses of 5, 10, and 15 mg reduced HbA1c by 1.72, 1.94, and 2.48 percentage points after participants reached target dose; placebo rose by 0.28 points, while open-label semaglutide 1 mg reduced HbA1c by 1.43 points. Fasting and two-hour post-meal glucose also improved in a dose-responsive pattern, which makes the efficacy signal more credible than one isolated endpoint. Weight fell by 4.83 kg at 10 mg and 8.25 kg at 15 mg, with placebo-adjusted differences of 6.58% and 9.55%. Those numbers justify larger studies. They do not establish that BGM0504 has beaten semaglutide or tirzepatide. Each dose group contained only about a dozen people, baseline HbA1c was higher in the 15 mg group than in the semaglutide group, the active-comparator arm was open-label, and semaglutide was tested at 1 mg rather than the 2.4 mg obesity dose. Tolerability is also part of the dose-response: diarrhea affected 75% of the 15 mg group, and every participant at that dose reported a treatment-emergent adverse event. The practical read is that dual-agonist design continues to produce potent metabolic effects, while the usable dose will be defined by long-term benefit, gastrointestinal burden, and active-comparator trials—not by the best 12-week number. BGM0504 remains investigational; online products using its name do not inherit the identity or controls of the trial drug.

Read the full analysis →

Oral semaglutide and alcohol: the secondary outcomes carry the signal

The newest semaglutide study enrolled 50 treatment-seeking adults with moderate to severe alcohol use disorder and randomized them to eight weeks of oral semaglutide or placebo. Participants received 3 mg daily for four weeks, then 7 mg daily for four weeks. Semaglutide did not significantly improve the primary endpoint, a laboratory measure of cue-elicited craving at week six, and it did not significantly reduce drinks per day. It did, however, reduce heavy drinking days (model estimate b=-0.580; 95% CI -1.012 to -0.148), drinks per drinking day (b=-1.177; 95% CI -2.307 to -0.047), and craving reported in ordinary life (b=-2.195; 95% CI -4.174 to -0.216). More participants also moved down at least one World Health Organization risk-drinking level. That uneven pattern is biologically and clinically plausible: GLP-1 signaling affects appetite, gastric emptying, and reward-related behavior, so it may change the intensity of some drinking episodes without uniformly suppressing every dimension of alcohol use. But a plausible mechanism does not rescue a missed primary endpoint. Small samples and multiple secondary analyses can produce estimates that shrink on replication. Context helps: two earlier randomized semaglutide trials also found signals for heavy drinking or drinking intensity, including a 26-week study in 108 adults with alcohol use disorder and obesity in which heavy drinking days fell 41.1 percentage points with semaglutide plus cognitive behavioral therapy versus 26.4 points with placebo plus therapy. Repetition across different formulations is encouraging, but semaglutide is not established as an alcohol-use-disorder treatment. The next trial needs a clinically direct primary endpoint, longer follow-up, detailed adverse-event reporting, and a design that can separate weight change, nausea, altered reward, and standard addiction care. Alcohol withdrawal can be dangerous; this is not a self-directed use case.

Read the full analysis →

Peceleganan: clinical improvement and bacterial clearance are different outcomes

Peceleganan, also called PL-5, is a topical antimicrobial peptide designed to disrupt microbial membranes. In a multicenter, randomized, double-blind trial of mild-to-moderate diabetic foot ulcer infections, PL-5 spray plus standardized debridement produced a significantly better short-term clinical response than placebo plus debridement one day after treatment ended. The accessible abstract does not report the absolute response rates or total randomized population, so the size and precision of that benefit are hard to judge. The culture results were more informative—and less tidy. Overall microbiological eradication was 57.89% with PL-5 versus 33.33% with placebo at the end-of-treatment assessment, and 64.71% versus 40.00% seven days later; neither comparison reached statistical significance. In a resistant-bacteria subgroup, clearance was 71.43% versus 50%, a significant difference, but the subgroup denominators and prespecification are not clear from the abstract. This distinction matters because less redness, drainage, pain, or swelling is valuable, yet it is not the same as microbiological cure, ulcer closure, avoidance of amputation, or freedom from recurrence. Earlier human trials point in the same complicated direction: a 570-person Phase 3 wound-infection study reported day-eight clinical efficacy of 90.4% with 2% peceleganan versus 78.7% with silver sulfadiazine, while bacterial clearance was lower with peceleganan, 24.0% versus 46.0%. Topical delivery may create high local exposure with limited systemic absorption, and membrane disruption may be useful against resistant organisms. Still, infected diabetic feet can deteriorate quickly, and PL-5 has not shown that debridement, vascular assessment, off-loading, systemic antibiotics, surgery, or urgent medical care can be skipped. The compound is most interesting as a potential complement to modern wound care, not as a home-treatment substitute.

Read the full analysis →

What to watch next

The next useful evidence will be harder to compress into a headline. For BGM0504, look for complete Phase 3 tables, longer safety follow-up, and adequately powered comparisons with tirzepatide or semaglutide at relevant doses. For semaglutide in alcohol use disorder, watch whether larger trials preselect heavy drinking or relapse as the primary endpoint and whether benefit persists beyond a few months. For peceleganan, demand absolute response rates, pathogen-specific clearance, resistant-subgroup denominators, ulcer healing, recurrence, antibiotic exposure, and limb outcomes. A good protocol is not built from positive endpoints alone. It is built by deciding which endpoint matters before looking at the result.

Model the interactions, not just the headlines

This week connected BGM0504’s GLP-1/GIP co-agonism with semaglutide’s GLP-1 signaling, put both beside tirzepatide as the established dual-agonist comparator, and contrasted those systemic drugs with peceleganan’s local membrane-disrupting action. Those mechanisms are not interchangeable, and neither are their evidence levels. The BioStack Generator helps you map semaglutide and tirzepatide against your goals, current medications, timing, and overlapping effects, while keeping investigational compounds such as BGM0504 and peceleganan in the evidence context they deserve. Use it to expose redundancies and conflicts before they become a schedule. It takes about two minutes.

Build your evidence-aware BioStack →

The Weekly Dose is educational and is not medical advice. Do not start, stop, combine, or change prescription or investigational treatments without a qualified clinician who knows your history.