Why First-Gen GLP-1s Failed for Addiction (And Semaglutide Might Succeed)
Four clinical trials of exenatide came up empty. Researchers think semaglutide's different pharmacology could change the equation—and early data supports that hope.
Exenatide (Byetta/Bydureon) showed no significant effect on alcohol, cocaine, or binge eating in clinical trials. But semaglutide binds more tightly to receptors, lasts longer, and penetrates the brain better—and the first RCT for alcohol showed positive results.
The First-Generation Failures
When animal studies showed that GLP-1 receptor agonists reduced alcohol and drug self-administration in rodents, researchers were excited to test them in humans. But the first wave of clinical trials—mostly using exenatide (the oldest GLP-1 drug, approved in 2005)—came up largely empty.
Exenatide for Alcohol Use Disorder (Denmark, 2022)
Primary Endpoint Failed127 patients with alcohol dependence received weekly exenatide or placebo for 26 weeks. All participants also received behavioral therapy.
Result: No significant difference in heavy drinking days between groups. Both groups reduced drinking by about the same amount.
However: Brain imaging showed exenatide reduced activity in the ventral striatum when participants viewed alcohol images. And an exploratory analysis found benefit in a subgroup with BMI ≥30.
Exenatide for Cocaine Use (Yale, 2021)
No Effect30 individuals with cocaine use disorder received a single acute dose of exenatide or placebo before cocaine self-administration in a laboratory setting.
Result: No effect on cocaine self-administration or subjective responses to cocaine.
Limitation: Single dose, acute treatment. May not reflect what happens with chronic dosing.
Dulaglutide for Smoking Cessation (Switzerland, 2024)
Primary Endpoint FailedDouble-blind trial testing dulaglutide (another older GLP-1) for smoking cessation.
Result: No significant effect on smoking cessation rates at 12-month follow-up.
Exenatide for Binge Eating
No EffectTrial testing exenatide for binge eating disorder.
Result: No significant benefit over placebo.
"That was a surprise for us." — Anders Fink-Jensen, MD, on the exenatide alcohol trial results
Why Researchers Think Semaglutide Is Different
Despite the negative trials, researchers didn't abandon the GLP-1 hypothesis. Instead, they reasoned that the newer, more potent GLP-1 drugs might succeed where older ones failed. Here's why:
| Property | Exenatide (Byetta) | Semaglutide (Ozempic/Wegovy) |
|---|---|---|
| Half-life | 2-4 hours | ~7 days |
| Receptor binding affinity | Moderate | High |
| Dosing frequency | Twice daily (or weekly for ER) | Once weekly |
| Weight loss efficacy | ~5% body weight | ~15% body weight |
| Brain penetration | Limited (uncertain) | Appears better (uncertain) |
Higher Receptor Occupancy
Semaglutide binds more tightly to GLP-1 receptors and maintains binding for longer. This could translate to stronger effects on reward pathways.
Consistent Drug Levels
Once-weekly dosing maintains steadier blood levels. With twice-daily exenatide, levels fluctuate dramatically—possibly insufficient for sustained CNS effects.
Greater Weight Loss Signal
Semaglutide produces ~3x more weight loss than exenatide. If CNS effects parallel metabolic effects, semaglutide's brain activity should also be stronger.
Better Brain Penetration?
Structural differences may allow semaglutide to cross the blood-brain barrier more effectively, though this isn't definitively proven.
"Based on the potency of exenatide and semaglutide, we expect that semaglutide will cause a stronger reduction in alcohol intake."
— Anders Fink-Jensen, MD, Psychiatric Centre Copenhagen (before the first semaglutide trial)
The Semaglutide Alcohol Trial: First Positive Result
In February 2025, researchers at UNC Chapel Hill published the first randomized controlled trial testing semaglutide specifically for alcohol use disorder. Unlike the exenatide trial, this one showed positive results.
Semaglutide for Alcohol Use Disorder (UNC, 2025)
Positive Results48 adults with AUD received either low-dose semaglutide (up to 1.0 mg weekly) or placebo for 9 weeks.
Primary finding: Semaglutide reduced alcohol consumption in a laboratory self-administration test.
Secondary findings: Reduced drinks per drinking day, significantly reduced weekly alcohol craving.
Bonus finding: In participants who smoked, semaglutide also reduced cigarettes per day.
The exenatide alcohol trial (127 patients, 26 weeks) failed to show benefit. The semaglutide trial (48 patients, 9 weeks) succeeded. While the semaglutide trial was smaller and shorter, the fact that it showed signal where exenatide didn't supports the hypothesis that drug potency matters.
The "Simplest Interpretation"
In their JAMA Psychiatry paper, the UNC researchers directly addressed why semaglutide might work when exenatide didn't:
"Although differences in sample characteristics or study design could contribute to differential evidence for efficacy across these 2 trials, the superior efficacy of semaglutide vs exenatide is likely the simplest interpretation." — Hendershot et al., JAMA Psychiatry, 2025
In other words: it's probably not that the populations or study designs differed that much. Semaglutide is simply a more potent drug.
The Observational Data Supports This
The JAMA Network Open study comparing semaglutide to other diabetes medications found something striking: semaglutide outperformed not just metformin and insulin, but also other GLP-1s (liraglutide, dulaglutide, exenatide) in reducing opioid overdose risk.
Hazard ratio for semaglutide vs. other GLP-1s: 0.32 (68% lower overdose risk). This wasn't just GLP-1s being better than other diabetes drugs—it was semaglutide specifically being better than its class cousins.
What About Tirzepatide?
Tirzepatide (Mounjaro/Zepbound) is even newer and even more potent than semaglutide for weight loss. It targets both GLP-1 and GIP receptors. Early observational data has linked it to reduced alcohol consumption, but no controlled trials have been completed yet.
If the pattern holds—more potent GLP-1s = better addiction effects—tirzepatide could theoretically be even more effective than semaglutide. But this remains entirely speculative.
The exenatide trials had limitations. The alcohol trial showed some positive signals (brain imaging, obese subgroup). The cocaine trial used only a single acute dose. It's possible that with different designs, older GLP-1s could show efficacy. We shouldn't assume only semaglutide works.
Trials Now Underway
Based on the hypothesis that newer = better, almost all ongoing addiction trials are using semaglutide (or occasionally tirzepatide):
NCT05895643: Semaglutide for alcohol in patients with AUD + obesity
NCT05891587: Semaglutide for alcohol with brain imaging
NCT06015893: Semaglutide for opioid abstinence
UTHealth Houston: Semaglutide for cocaine ($2.7M NIH-funded)
No major trials are testing exenatide or liraglutide for new addiction indications. The field has moved on.
First-generation GLP-1s (exenatide, dulaglutide, liraglutide) failed in clinical trials for addiction—but showed some intriguing signals. Semaglutide's first controlled trial succeeded. The most likely explanation is pharmacological: semaglutide's superior potency translates to stronger effects on the brain's reward system. Larger trials will determine if this pattern holds.