Six days ago the news section explained why Amylyx's avexitide ↗ blocks the GLP-1 receptor to fix post-bariatric hypoglycemia. On May 15, Obesity Reviews published an Eisa and Barood meta-analysis ↗ of seven studies and 337 patients showing that semaglutide ↗ (which does the opposite, it activates the same receptor) fixes the same complication. Both numbers are robust. The receptor is not the mechanism. The meal is.
The result
The meta-analysis pooled three randomized controlled trials, three cohort studies, and a case series, all evaluating subcutaneous semaglutide in patients who developed dumping syndrome or reactive hypoglycemia after bariatric surgery. The search ran through September 2025 and was registered prospectively. Effect sizes were standardized mean differences because the included studies used different symptom-score instruments, but the direction was consistent across every endpoint the authors looked at.
Dumping symptom scores improved by 1.18 standard deviations (roughly: patients went from severe to mild on the typical scoring tools used, 95% CI 0.72 to 1.64, heterogeneity I² 39%). Hypoglycemic episode frequency dropped by 0.85 standard deviations (about a halving of episode counts in the trials that reported them, 95% CI 0.48 to 1.22, I² 35%). Time spent in hypoglycemia fell by 0.62 standard deviations. Glycemic variability fell by 0.71. Quality of life improved by 0.73. Gastrointestinal adverse events were common, as they are everywhere semaglutide is studied, and were rarely the reason patients stopped the drug.
The heterogeneity numbers (35-39%) sit at the low end of "moderate." That is the cleanest signal a meta-analysis of seven mixed-design studies usually produces. The result is not driven by a single outlier trial.
The mechanism puzzle
Six days ago the avexitide piece ↗ laid out the standard PBH model. Roux-en-Y gastric bypass moves the stomach pouch upstream of most of the upper GI tract, so a meal arrives in the distal small intestine faster than the digestive physiology was set up to handle. Nutrient-stimulated L-cells in the lower small intestine release GLP-1 ↗ in oversized pulses. The pulses drive pancreatic beta-cells to over-secrete insulin. Blood glucose crashes two to three hours after a meal. Avexitide, which is exendin (9-39) ↗, occupies enough GLP-1 receptors on islet beta-cells to clip the inappropriate insulin peak, and the hypoglycemia stops.
Semaglutide does the opposite at the same receptor. It is an agonist with a multi-day half-life that holds the receptor in a continuously activated state. The model from the avexitide piece predicts that more GLP-1 signaling, not less, should make the runaway insulin response worse, not better.
The reconciliation is that the avexitide model is downstream of the actual cause. The runaway L-cell pulse in PBH is triggered by a runaway carbohydrate surge in the small intestine. Semaglutide slows gastric emptying enough (a well-documented effect at every dose where the drug is studied) that the surge stops happening. No surge, no L-cell pulse, no beta-cell over-secretion, no hypoglycemia. The receptor activation at the islet still occurs, but it occurs against a flattened post-prandial glucose curve where it has no over-secretion to cause. Chronic supra-physiological occupancy probably also desensitizes the beta-cell to the endogenous post-meal GLP-1 spike, but the gastric-emptying effect is sufficient to explain the result on its own.
Two drugs at the same receptor in opposite directions, both producing the same therapeutic outcome by acting on different parts of the chain. Avexitide intercepts the receptor amplification step. Semaglutide intercepts the trigger that drives the receptor amplification step.
What the meta-analysis does not settle
The studies it pooled mostly used standard obesity-dose semaglutide (titrated to 2.4 mg weekly or the diabetes-dose equivalent). None of them were designed against avexitide. Whether the two drugs are interchangeable, complementary, or staged by severity is an unanswered question that the LUCIDITY Phase 3 readout ↗ in the third quarter of 2026 will not directly answer either, because LUCIDITY is placebo-controlled. The closest the field has to a head-to-head is the implicit comparison this meta-analysis now allows against the historical avexitide Phase 2 numbers.
What this changes for peptide design. The GLP-1R has had two decades of agonist optimization (longer half-life, more biased signaling, dual and triple co-agonism). It has six days of clinical-stage antagonist optimization since the avexitide piece. The mechanistic point that both directions of intervention at this receptor produce real symptomatic benefit in the same indication makes a stronger case than either drug alone for the proposition that GLP-1 receptor pharmacology has a richer functional space than the "more agonism, more weight loss" framing has explored.
Two doors into the same room. Different keys. Same lock.