Naltrexone for Alcohol Use Disorder

Naltrexone is an opioid antagonist. Drinking releases endogenous opioids that bind receptors in the brain's reward circuit, and naltrexone sits on those receptors so the surge lands softer. That is the whole mechanism. Nothing in it makes you ill, and nothing in it forces abstinence. If you drink while taking it, you will not flush or vomit the way disulfiram would make you. The FDA label describes 50 mg once daily as an adjunct to social and psychotherapeutic methods, tested for up to twelve weeks.

That distinction matters more than it sounds. Disulfiram works by threat, so it only helps while you keep swallowing it and know you have swallowed it. Naltrexone works by dulling the payoff, so the pull toward the fourth and fifth drink can weaken on an evening you were not planning to stop at all. The label is unusually blunt about the ceiling here, describing a modest improvement in the outcome of conventional treatment rather than a stand-alone answer. Read that as a floor to build on, not a disappointment.

The 2023 JAMA systematic review is the cleanest available answer to what this drug does. Pooling 118 randomized trials and 20,976 participants, it found that oral naltrexone at 50 mg a day carried a number needed to treat of 11 to prevent one return to heavy drinking, with a 95% confidence interval of 5 to 41. For preventing a return to any drinking at all, the number needed to treat was 18.

Eleven means that for every eleven prescriptions written, one drinker who would otherwise have gone back to heavy drinking does not.

An eleven is not a miracle and it is not nothing. Most drugs you already trust sit in that neighborhood. The same review put acamprosate at a number needed to treat of 11 for return to any drinking, which is better than naltrexone on that particular outcome. So the two are not interchangeable. Naltrexone earns its place on heavy drinking days, acamprosate on total abstinence, and the review names both as first-line options alongside psychosocial treatment rather than instead of it.

The older COMBINE trial, 1,383 volunteers across eleven US academic sites, found the same shape. Naltrexone with medical management raised percent days abstinent to 80.6 against 75.1 on placebo and cut the risk of a heavy drinking day with a hazard ratio of 0.72. Acamprosate showed no effect in that trial at all, in either direction. Read the two together and the honest summary is a real but moderate benefit that surfaces most reliably when somebody is checking in with you on a schedule.

Four medications get used for drinking in the United States and only two of them are FDA approved for it. Acamprosate is the head-to-head rival, and the trade is schedule against outcome: it dosed as two 333 mg tablets three times daily in the trials, which is six tablets a day against one, and it aims at staying stopped rather than at drinking less. Disulfiram is a different category entirely, a deterrent that depends on you fearing the reaction, and the naltrexone label warns that combining two potentially hepatotoxic medications is not ordinarily recommended.

Topiramate is the wildcard. A network meta-analysis of 32 trials in non-abstinent drinkers found topiramate superior to nalmefene, naltrexone and acamprosate on consumption outcomes, while noting its poor safety profile. That same analysis found no benefit for naltrexone on total alcohol consumption in that population, which is a real result and worth sitting with. The claims-data picture backs the tolerability angle: topiramate initiations lasted meaningfully longer before discontinuation than naltrexone initiations across 51,915 starts.

Disulfiram deserves a fairer hearing than it usually gets. It is the oldest of the four and it still does its job when somebody else watches you swallow it, which is why supervised dosing shows up in the trials that favor it. It asks nothing of your craving and everything of your fear. The practical point is that these four are not ranked in one column: your liver history, your drinking pattern, whether you are aiming at abstinence or at reduction, and how many tablets a day you will realistically take all move the answer.

Oral naltrexone comes as one strength: a 50 mg film-coated tablet. The alcohol dosing instruction is 50 mg once daily, the regimen the approval trials ran for up to twelve weeks, and the label notes plainly that other regimens and durations were not evaluated in those trials. There is no titration schedule for drinking, no starting dose to work up from.

You take one tablet. That simplicity is part of why a clinician can start it in a fifteen minute visit.

The extended-release injectable is a different proposition. It is 380 mg delivered deep intramuscularly as a gluteal injection every four weeks, alternating sides, using the customized needles that ship with the carton. Nobody self-administers this. You go somewhere every month and someone qualified assesses your body habitus and picks the needle length. In exchange you stop making a decision every single morning, which matters because the median time to stopping any alcohol medication was 30 days across 51,915 US initiations.

The pooled evidence rates the injection more cautiously than the marketing does. The 2023 review found injectable naltrexone associated with about five fewer drinking days over a 30-day treatment period and did not calculate a number needed to treat for it, because the strength of evidence did not reach the threshold that required. Oral at 50 mg is where the firm numbers live. Choose the injection for the adherence problem it solves rather than for a bigger pharmacological effect.

For the tablet, the abstinence-first rule is folklore rather than label. A Finnish factorial trial randomized 121 nonabstinent outpatients with alcohol dependence and concluded that detoxification is not required and targeted dosing taken when craving occurs maintained the reduction in heavy drinking. That trial is why a competent prescriber will not send you away for being a current drinker. Waiting until you have already quit means waiting for the thing the medication is supposed to help with. The injection is a genuine exception, and it is the correction most sources skip. In the 24-week pivotal alcohol trial, the effect concentrated in the 53 subjects, 8% of the study, who abstained completely in the week before their first dose, and the same treatment effects were not evident in the 571 who were actively drinking at initiation. Different route, different starting requirement.

If you are still drinking daily and want to begin now, the tablet is the form with evidence behind that choice.

The popular version of the Sinclair method is a pill an hour before you drink, forever, with no therapy attached. The trial it descends from, co-authored by Sinclair himself, is not that. All 121 participants took naltrexone or placebo daily for the first 12 weeks and only then moved to targeted dosing when craving for the following 20 weeks. Targeted dosing was the maintenance phase after a daily loading period, not the opening move.

The second correction is bigger. In that trial naltrexone was no better than placebo among those given supportive therapy. It separated only in the cognitive coping skills arm, where 27% had no relapse to heavy drinking across the full 32 weeks against 3% on coping plus placebo. The pill was not doing the work alone. Any version of the method sold as medication without structured behavioral work is quoting a trial that did not test it.

Targeted dosing on its own does have modern support. A 2022 double-blind randomized trial of 120 sexual and gender minority men who binge drink and have mild to moderate alcohol use disorder gave targeted 50 mg or placebo with weekly counseling for 12 weeks and reported fewer binge drinking days, an incidence rate ratio of 0.74, with effects sustained at six months after treatment. Weekly counseling again. Meanwhile a network meta-analysis in non-abstinent drinkers found no benefit for naltrexone on total alcohol consumption, so the evidence here is genuinely mixed.

This is the part where the answer is no, not maybe. Naltrexone is contraindicated if you are receiving opioid analgesics, currently dependent on opioids including methadone or buprenorphine maintenance, in acute opioid withdrawal, or carrying a positive urine screen for opioids. Taking it on board opioids precipitates withdrawal that the injectable label warns can be severe enough to require hospitalization, with a minimum opioid-free interval of 7 to 10 days recommended before the first injection. Two weeks if you are coming off buprenorphine or methadone.

The risk running the other direction is the one that kills. Blocking the receptors lowers your tolerance, so the label warns that after treatment stops a previously tolerated opioid dose may cause life-threatening intoxication, and fatal overdoses have been reported after discontinuation. The current labeling goes further and tells prescribers to strongly consider recommending or prescribing naloxone or nalmefene when starting and renewing. If you take opioids at all, including tramadol or a cough syrup you forgot about, say so before the prescription is written.

Naltrexone spent two decades under a boxed liver warning that it no longer carries, and the reason matters. The whole signal traces to a study in which naltrexone was given at 300 mg a day to obese subjects. In that study 19% of the naltrexone group, 5 of 26, developed elevated serum transaminases against 0 of 24 on placebo, and the levels returned toward baseline within weeks. Three hundred milligrams is six times the 50 mg used for drinking. The dose that generated the fear is not the dose you would be prescribed.

FDA agreed. The archived 2013 label revision records, in its own Recent Major Changes table, Boxed Warning Removed 7/2013 and Contraindications for Acute Hepatitis or Liver Failure Removed 7/2013. Current labeling still carries hepatotoxicity as a warning, still tells you to report symptoms of liver trouble immediately, and still flags a higher risk of hepatocellular injury with single doses above 50 mg. What it no longer does is treat 50 mg daily as a liver hazard, or bar the drug outright from anyone with hepatitis.

Everyday tolerability is a different conversation and mostly means nausea. In an open-label safety study of roughly 570 drinkers, nausea ran 10%, headache 7%, and dizziness, nervousness and fatigue 4% each. The 2023 pooled review put the nausea risk ratio at 1.73 against placebo. Most of it fades. None of this is a substitute for a clinician reviewing your liver history, your other medications and your drinking pattern before deciding whether the drug suits you.

Low-dose naltrexone shares a molecule with the alcohol treatment and almost nothing else. A 2026 narrative review defines it as naltrexone prescribed off-label at 0.5 to 6.0 mg, with most studies using 4.5 mg daily, for chronic pain, autoimmune and neuroimmune conditions, and post-infectious syndromes. Reviewing 105 studies including 15 randomized controlled trials, the authors found that early positive findings from uncontrolled studies were rarely replicated in placebo-controlled trials and that current evidence does not support routine clinical use.

There is a supply-side consequence worth knowing before you go looking. The only oral naltrexone marketed in the United States is the 50 mg film-coated tablet. Everything below that strength is compounded from bulk powder, which means no FDA approval for the finished preparation and no bioequivalence testing behind it. Compounded medications are not reviewed by FDA for safety or efficacy. Whatever you conclude about low-dose naltrexone for other conditions, it is not the treatment the alcohol evidence in this page describes.

If you are already taking low-dose naltrexone for something else and drinking has now become the question, raise it rather than assume the doses stack. Four and a half milligrams is roughly a tenth of the alcohol dose, and none of the trial evidence on this page was generated at that strength. The opioid contraindication still applies at any dose. A clinician can tell you whether moving to 50 mg makes sense in your case, and what that would mean for whatever you began low-dose treatment for.

Start from what pharmacies pay, because it anchors everything else. On the CMS National Average Drug Acquisition Cost file effective 2026-08-19, naltrexone 50 mg tablets carried an acquisition cost of $1.13919 each, classified as generic. Thirty tablets is about $34 of actual drug. That is not a retail price and it is not what you will be quoted: dispensing fees, pharmacy margin, insurance design and the cost of the visit itself all sit on top, and cash prices vary widely between pharmacies in the same city.

Two comparisons from the same file put that in proportion. Acamprosate was $0.61033 per 333 mg tablet, cheaper per tablet but six tablets a day. The extended-release injectable was $1,716.48430 per 380 mg vial and carried the brand classification, roughly fifty times the monthly acquisition cost of tablets, plus an in-person administration visit every four weeks. If cost is the binding constraint, oral naltrexone is the least expensive of the FDA-approved options for drinking, and the one with the firmest effect numbers behind it.

Access is the harder problem. A national survey of 9,921 outpatient programs found only 45% provided any medication for alcohol use disorder. Remote prescribing matters for exactly that reason. Naltrexone is not a controlled substance, so a licensed clinician can assess and prescribe it by video in most states. None of this is medical advice, and whether the drug suits you is a licensed clinician's call. When you compare services, ask four things: will they prescribe while you are still drinking, do they screen you properly for opioid use, do they schedule follow-up past the first fill, and can they discuss the injection as well as the tablet.