Here is a number worth sitting with before your next pharmacology block: 1.6%. That is the share of American adults with alcohol use disorder who took any medication for it in 2019, according to Han and colleagues in JAMA Psychiatry. Roughly 14 million people met criteria that year. About 223,000 of them got a prescription.
Now picture yourself six months into practice. A patient sits across from you who has tried naltrexone and gotten nothing from it, white-knuckled through acamprosate’s three-times-daily schedule until they quit taking it, and is still drinking a fifth of vodka a day. What’s your next move? A new paper trail suggests the field is starting to have an answer, and it involves two cheap generic drugs most of us already know from other contexts.
The Study Carlat Flagged
In its May 2026 issue, The Carlat Psychiatry Report ran a Research Update by Christine Tran-Boynes, DO, titled “Prazosin and Cyproheptadine for AUD: A Promising Combination.” The trial under review came from Henri-Jean Aubin and colleagues at Université Paris-Saclay, published in Addiction in 2024.
Here is the design. They enrolled 154 adults with severe alcohol use disorder who were drinking at high-risk levels, meaning more than 60 grams of alcohol a day for men and 40 grams for women. For twelve weeks, patients were randomized to one of three arms: a high-dose combination of prazosin 10 mg and cyproheptadine 12 mg daily, a low-dose combination of prazosin 5 mg and cyproheptadine 8 mg daily, or placebo. Everyone also got brief psychosocial support focused on adherence and cutting back.
The primary outcome was alcohol consumption, not abstinence. That choice matters, and we will come back to it.
According to the summary in Psychiatric News, the high-dose group cut daily drinking by 23.6 grams more than placebo. The low-dose group cut it by 18.4 grams more. A standard US drink holds 14 grams of alcohol, so that works out to roughly 1.7 and 1.3 drinks per day beyond what placebo accomplished. In the heaviest drinkers (over 100 grams a day for men, over 60 for women), the high dose pulled an extra 29.8 grams off the daily total, a bit more than two drinks.
There was also a dose-response pattern, which is the kind of detail that makes a signal harder to dismiss. More drug, bigger effect.
Why These Two Drugs?
You already know prazosin. It is an alpha-1 blocker, and many of you will have learned it as the PTSD nightmare drug or the old-school BPH and hypertension agent. Cyproheptadine is a first-generation antihistamine with serotonin 5-HT2A antagonism, the one that shows up in serotonin syndrome discussions and appetite stimulation in kids.
The authors’ reasoning, reported by Psychiatric News, is that the two drugs hit different parts of the problem. Prazosin works on noradrenergic pathways tied to craving and stress reactivity. Cyproheptadine works on serotonin receptors tied to impulsive behavior. Put them together and you are covering both the “I need a drink because I’m wound up” pathway and the “I had one before I decided to” pathway.
Whether that mechanistic story holds up is still a hypothesis. But it gives you a framework for thinking about why combining agents might work when single agents fall short, and that kind of reasoning shows up on exams.
Read the Fine Print Like a Clinician
This is where I want you to slow down, because the way Carlat handles research is a model for how you should handle it too.
First, it is one phase 2 trial. One. The authors themselves said the results “warrant” phase 3 work, which is researcher-speak for “we are not done.” Second, twelve weeks is short for a disorder that behaves like a chronic relapsing condition. Third, and Dr. Tran-Boynes flags this directly, the study was funded by Kinnov Therapeutics, the company developing a fixed-dose prazosin-cyproheptadine combination pill. The trial used the separate generic drugs, but the sponsor has a stake in the answer.
That does not make the data wrong. It does mean you hold it more loosely. At this site we do not lean on pharmaceutical marketing, so I am telling you plainly where the money came from and why the Carlat reviewers called the evidence preliminary. Your future patients deserve that same discipline from you.
On safety, the Carlat summary says no serious adverse events were attributed to the combination, and more than 90 percent of adverse events were mild or moderate. Sedation and orthostatic hypotension, the two things you would expect from an alpha blocker plus a sedating antihistamine, showed up at similar rates across all three groups, including placebo. I would still not take that as permission to be casual. Your sample here was people who agreed to join a trial. Your real-world patient might be older, on an antihypertensive, drinking heavily enough to be volume-depleted, or all three.
Why Drinking Reduction Is a Legitimate Goal
Notice what the investigators chose to measure. They tracked how much people drank, not whether they stayed completely dry. Aubin’s group said they picked that endpoint because many people with AUD want to cut down rather than stop.
If your training so far has been abstinence-or-bust, this may feel uncomfortable. It shouldn’t. Reduced drinking lowers risk. Patients who would never walk into a program that demands total sobriety will often show up for one that meets them where they are, and then some of them go on to stop altogether. Harm reduction is not a consolation prize in addiction medicine. It is often the front door.
What the Guidelines Still Say
Do not walk out of this post thinking prazosin plus cyproheptadine has displaced anything. It hasn’t. The American Psychiatric Association’s 2018 practice guideline on pharmacological treatment of AUD still recommends naltrexone or acamprosate for moderate to severe AUD when patients have a goal of reducing or stopping drinking, with disulfiram as an alternative, and topiramate and gabapentin as second-line options in specific situations. If a board question asks for the first-line choice for an AUD patient without opioid use, the answer is not today’s study.
The Carlat takeaway is deliberately modest: this combination could be considered off-label for patients with severe AUD who have not responded to naltrexone or acamprosate, with the understanding that the evidence is thin. Think of it as a possible third or fourth step on the ladder, not the first rung.
If you do try it down the road, the monitoring plan writes itself. Check orthostatic blood pressure before and after starting. Titrate slowly. Warn patients about sedation, particularly at the higher dose. Ask about other antihypertensives, PDE5 inhibitors, and other sedating medications. Screen for pregnancy where relevant and be aware that both drugs have their own interaction profiles. And document why you chose off-label treatment, what you told the patient about the limits of the evidence, and what the plan is if it does not work.
The Bigger Lesson: Medication Is Underused, Not Overused
I keep coming back to that 1.6 percent. Han and colleagues wrote that medications are “rarely prescribed to and used by adults with AUD.” For comparison, we do not treat hypertension, diabetes, or depression with a rate that low and call it acceptable.
Part of the gap is stigma. Part of it is that many prescribers never received training in addiction medicine and quietly hand AUD off to someone else. Part of it is a system where patients land in detox, then rehab, then a meeting list, and nobody ever mentions that a pill exists which could help with craving. You are going to be one of the people who can change that. PMHNPs are positioned in outpatient clinics, telehealth practices, community mental health centers, and primary care settings where AUD sits right next to the depression and anxiety you are already treating.
That is a real career opportunity, not just a clinical one. The PMHNP who is comfortable starting naltrexone on a Tuesday afternoon, who can explain why acamprosate needs renal dosing, and who knows what to try when both fail, is the PMHNP who fills a gap most practices cannot fill.
What to Do With This for Boards
You will not see prazosin-cyproheptadine as a keyed answer on the ANCC or AANP exams. But the thinking behind this paper is exactly the sort of reasoning those exams reward. Here is how I would make it stick.
Know the first-line AUD medications cold: naltrexone (oral or injectable, contraindicated with opioid use or acute hepatitis), acamprosate (renally cleared, three-times-daily dosing, no hepatic metabolism), and disulfiram (needs a motivated, supervised patient). Know when you would reach for topiramate or gabapentin. Know why prazosin’s alpha-1 blockade means orthostasis and first-dose syncope. Know that cyproheptadine is a 5-HT2A and H1 antagonist and what that means for sedation and appetite. If you can explain those mechanisms in your own words, you can answer almost any variation a question writer throws at you.
Then practice the other skill: reading a study like a prescriber. Ask who paid for it. Ask how long it ran. Ask whether the outcome was one your patient cares about. Ask how big the effect is in grams or drinks per day, not just whether a p-value crossed a line. Those habits are the difference between a clinician who gets pushed around by every new paper and one who can say, “Interesting, let me see the replication.”
Where This Leaves You
Here is what I want you to carry out of this. A promising phase 2 trial, funded by a company with a product on the line, reviewed critically by an independent publication, is not a reason to rewrite your prescribing habits. It is a reason to stay curious, stay skeptical, and notice that psychiatry is still working out how to help the patients who fail the first two things we try.
Those patients are coming to you. Being the clinician who has an answer, and who is honest about how strong that answer is, is exactly what this profession needs.
You are putting in the late nights now so you can be that person later. If you want structured practice on pharmacology, addiction medicine, and the board-style reasoning that ties it all together, head over to pmhnpreview.org and keep building. The patients who haven’t had their first prescription yet are counting on someone like you.