The biomarker that made the GLP-1 alcohol use disorder case structural

The GLP-1 drug class just crossed a line addiction medicine has been arguing about for two years: a blood biomarker moved in the right direction. Patient surveys have always been the soft underbelly of the GLP-1 → alcohol use disorder (AUD) hypothesis, because the people signing the daily drinking logs are the people trying to cut back. When phosphatidylethanol (PEth), a direct residue of recent drinking, fell 38% on Altimmune's pemvidutide while climbing 5.9% on placebo in the Reclaim phase 2 readout, structural evidence caught up to the headline. The primary endpoint was a survey-reported reduction in heavy drinking days; the 38% PEth drop was a secondary/exploratory biomarker measure.

Objective endpoints buy you something surveys cannot in a stigmatized indication: a number that cannot be politely answered. The 4.2 versus 2.75 fewer heavy drinking days the survey picked up is the same modest gap critics waved off since semaglutide's first AUD signals; the PEth divergence is what makes that gap harder to dismiss. William Blair's note called the biomarker drop the strongest validation of the class effect, and the framing is right for a specific reason: the placebo arm getting worse is the control the field has been missing.

The repeatable mechanism: gut-hormone agonists suppressing the rewarding loop behind compulsive intake, tested in obesity, then diabetes, now addiction, with the cleanest evidence coming from biomarkers the patient cannot game. Eli Lilly's next-gen dual-acting GLP-1 candidate is in phase 3 AUD; Novo Nordisk's semaglutide has early AUD data, though William Blair noted a more robust profile for pemvidutide at the doses tested. Pemvidutide does not have to win the race to have made the class legible.