Why GLP-1 Plus Therapy Beat Therapy Alone for Drinking

The 2026 trial did not test a GLP-1 against therapy. It tested therapy plus a GLP-1 against therapy plus placebo, and that distinction is the whole story.

What You'll Discover:

• What the National Institutes of Health actually reported.

• Why the placebo group's improvement is the most useful number in the study.

• How this mirrors twenty years of naltrexone research.

• What medication plus support looks like in an ordinary week.

• Where a GLP-1 fits, and where naltrexone does.

Most coverage of the 2026 semaglutide trial ran with a simple headline. GLP-1 reduces heavy drinking.

The actual finding was narrower and far more useful. Every person in that trial received cognitive behavioral therapy. The medication was added on top of it.

That framing changes what the result means for anyone deciding what to do about their own drinking.

What the Trial Actually Found

The National Institutes of Health reported in April 2026 that adding a weekly GLP-1 to cognitive behavioral therapy further reduced heavy drinking compared with therapy plus placebo.

The study enrolled 108 treatment-seeking patients with alcohol use disorder and comorbid obesity, and ran for 26 weeks.

Everyone received standard cognitive behavioral therapy. On top of that, participants received either weekly semaglutide or weekly placebo injections.

Participants on semaglutide saw a 41.1% reduction in heavy drinking days. That was 13.7 percentage points better than the placebo group managed.

Blood-alcohol biomarkers supported the self-reported drinking data, which matters more than it sounds. People consistently underreport how much they drink, and the biomarkers moved in the same direction as the diaries.

The researchers also reported a number needed to treat of 4.3, meaning roughly four to five people would need the treatment for one additional person to benefit.

The Detail Most Coverage Skipped

Here is the number almost nobody reported. The placebo group reduced their heavy drinking days by 26.4 percentage points.

Those people received a saline injection and cognitive behavioral therapy. Nothing else.

The therapy alone carried them roughly two-thirds of the way to where the medication group ended up.

That is not a knock on the medication. It is direct evidence that behavioral support does real, measurable work, and that a trial without it would have looked very different.

Heavy drinking days, the outcome both arms were measured on, is defined in the National Institute on Alcohol Abuse and Alcoholism's guidance on alcohol drinking patterns. It is a hard behavioral count rather than a self-assessment.

Both halves worked. The combination worked better. That is the finding.

It also has a practical implication most readers will care about more than the statistics. If you are only offered one half of this, you are being offered less than the evidence supports.

Plenty of people get handed a prescription with no follow-up. Plenty of others get sent to a support group with no discussion of medication at all.

The trials keep saying the same thing about that split. Each half works. Neither half is the whole treatment.

Therapy Alone Versus Therapy Plus Medication‍

‍Why the Combination Beats Either Half

Drinking is held in place by two separate forces, and each half of the treatment handles one of them.

The biological half is craving and reward. Alcohol drives a dopamine surge through the brain's reward circuits, and after enough repetitions the system starts anticipating it on schedule.

That anticipation is not a decision. It arrives before you have consciously chosen anything, which is exactly why advice to try harder tends to fail.

Medication works on that layer. It lowers the pull, so the moment of choice starts from a quieter baseline. We cover how GLP-1 medications appear to do this in our piece on whether GLP-1 medications reduce alcohol cravings.

The behavioral half is everything a medication cannot reach. The specific bar you walk past on the way home, the 6pm slump, the friend who orders for the table, the quiet belief that you earned it today.

Therapy works on that layer. It replaces a hope about Thursday night with a plan for Thursday night.

Neither half solves the other's problem. That is why trials combining them keep outperforming trials that do not.

This Is Not a New Pattern

Twenty years before anyone tested a GLP-1 on drinking, the same result showed up with naltrexone.

The COMBINE study, published in JAMA in 2006, randomized patients into nine different treatment groups. It is still one of the largest alcohol treatment trials ever conducted.

Participants who received naltrexone plus medical management reached 80.6% days abstinent, compared with 75.1% for those on placebo plus medical management.

Notice that the control group still did well. Structured clinical support by itself moved people substantially, exactly as therapy alone did in the 2026 trial.

Medication added to that support did better than the support alone. The pattern has now repeated across two decades and two completely unrelated drug classes.

That is about as close to a settled finding as alcohol research produces, and it is the part of the 2026 study least likely to be overturned by a bigger trial. We go deeper into the comparison in our article on naltrexone versus therapy for alcohol.

What an Integrated Solution Looks Like Week to Week

Descriptions of combined treatment tend to stay abstract. Here is what it actually involves.

A physician reviews your history, your drinking pattern, and your goals, then prescribes if it is appropriate. That first conversation sets both the dose and the plan.

You take the medication on a schedule you can realistically keep. For naltrexone that is one 50mg tablet, usually daily, taken at least an hour before you expect to drink.

You track your drinking honestly. That sounds trivial and it is not, because most people underestimate their intake by a wide margin until they start writing it down.

You have someone to check in with. A coach, a clinician, a group, or all three, depending on what the week looks like.

And then you adjust. Dose, timing, and strategy all get revised based on what the tracking shows.

That last step is what separates real care from a prescription arriving in the mail. Our overview of medication-assisted treatment for alcohol at home walks through how it works when the whole arrangement happens online.

What "Support" Actually Means Here

The word therapy makes a lot of people picture a couch and an hour of talking about childhood. Cognitive behavioral therapy for alcohol use disorder is not that.

It is closer to skills training. You map the situations that reliably produce a heavy night, then build specific responses to them in advance.

A typical session might cover what to do when a colleague hands you a drink you did not ask for. Or how to get through the first twenty minutes at home after work, which is when most people's resolve gets tested.

Some of it is coping strategy. Some of it is scheduling. A lot of it is the simple accountability of telling another person what actually happened last week.

That is why the placebo arm improved as much as it did. Structured attention on the problem changes the problem, even before a medication enters the picture.

The version delivered online tends to compress this into shorter, more frequent contact. Daily check-ins, drink logging, and a coach who reads the log, rather than one 50-minute block a week.

Neither format is inherently better. What matters is that the contact is regular and that someone is actually looking at the numbers with you.

Where a GLP-1 Fits and Where Naltrexone Does

Naltrexone is FDA-approved for alcohol use disorder and has been since 1994. It carries decades of trial data and a mechanism that is well understood.

GLP-1 medications are not approved for alcohol use disorder. The evidence is emerging and genuinely promising, and right now it rests on two small randomized trials.

They also appear to act at different points in the sequence. Naltrexone blunts the reward that follows a drink, which is why people taking it often stop after one or two.

GLP-1 medications seem to work on the wanting that arrives before the first drink is poured.

Some people use both together under physician supervision. Our article on taking a GLP-1 and naltrexone together covers how physicians think through that combination.

Choose Your Horizon supports both paths. Naltrexone-assisted care sits at the center of what we do, and physician-guided GLP-1 options are available for people who are appropriate candidates.

What to Ask For When You Set This Up

If the medication-plus-support finding is the real lesson, then the useful question is what to actually request from whoever is treating you.

Ask what the medication is expected to do and on what timeline. Naltrexone typically starts producing noticeable change in two to four weeks, and knowing that prevents an early quit.

Ask who reviews your drinking data and how often. A plan nobody looks at is not a plan.

Ask what happens if the first approach does not work. Dose adjustments, timing changes, and added support are all normal, and a good setup expects them rather than treating them as failure.

Ask whether your goal is allowed to be reduction rather than abstinence. Both trials discussed here measured heavy drinking days, not total sobriety, because reduction is a legitimate clinical target.

Those four questions will tell you very quickly whether you are being offered real care or just a refill.

The Limits Worth Naming

The 2026 trial had 108 participants at a single center. That is a modest study by any standard, and treatment effects frequently shrink when trials scale up.

Every participant also had obesity. An earlier GLP-1 trial found no benefit across its full sample but a strong response in the subgroup with obesity, which means the population studied matters a great deal here.

Twenty-six weeks is short relative to how long drinking patterns take to build. No data has been published on what happens after the injections stop.

Cross-trial comparisons of number needed to treat also deserve caution. Different populations, different endpoints, and different lengths make those figures much less comparable than they look.

None of this weakens the combination finding. The medication-plus-support result is the sturdiest thing in the study, precisely because it replicates something already well established.

Single trials get overturned all the time. Findings that show up repeatedly, in different populations, with different medications, tend to hold.

The combination effect is in the second category. That is the reason it deserves more weight than the specific percentages attached to it.

The Takeaway

The 2026 trial is being read as a story about a medication. It is really a story about a structure.

Medication lowers the biological pull. Support handles the situations, the habits, and the reasons underneath them.

Together they beat either one alone, and that has now been demonstrated across two drug classes and twenty years of clinical, neurological, and behavioral science.

If you are working out what to do about your own drinking, that is the practical lesson. Ask for both halves, not one.

You also do not need to reach a crisis point first. Wanting fewer heavy nights and clearer mornings is a complete reason to start.

And the numbers from both arms of that trial are worth holding onto. People improved with support alone, and improved more with support plus medication.

Neither group had to hit bottom first. They just had to start.

Frequently Asked Questions

Did the GLP-1 trial test medication without therapy?

No. Every participant received standard cognitive behavioral therapy, and the comparison was therapy plus semaglutide versus therapy plus placebo. The trial cannot tell us how the medication performs on its own.

How much did therapy alone help in that trial?

The placebo group reduced heavy drinking days by 26.4 percentage points over 26 weeks. That is a substantial result from behavioral support plus regular clinical contact.

Is combining medication and therapy better than either alone?

That is what the evidence keeps showing. The 2006 COMBINE study found the pattern with naltrexone, and the 2026 trial found it again with a GLP-1.

Can I take naltrexone and a GLP-1 at the same time?

Some people do, under physician supervision, because the two act on different parts of the reward pathway. It is a decision for a prescribing doctor who knows your full medical history.

Do I need in-person therapy for this to work?

Not necessarily. Structured support delivered online, including coaching, drink tracking, and regular clinical check-ins, is the model most modern alcohol care now uses.

Is a GLP-1 approved for alcohol use disorder?

No. Semaglutide is approved for type 2 diabetes and weight management, and any use aimed at drinking is off-label. Naltrexone is the medication with FDA approval for alcohol use disorder.

Find out whether naltrexone and real support fit your situation. Take the online Alcohol Use Assessment and a physician will review your history and your goals with you.