The Sinclair Method: Naltrexone, Extinction, and the Honest Evidence
Quick answer: The Sinclair method means taking naltrexone about an hour before you drink, every time you drink, and continuing to drink while the medication blocks the reward. The idea is pharmacological extinction: a habit that stops paying off gradually weakens. The underlying drug is FDA-approved for alcohol dependence and the targeted-dosing evidence is real but thin — one 121-person trial and one 163-person trial, not a guideline-grade base. It is prescription-only, and it fails immediately if you skip the pill.
Most approaches to drinking ask you to stop first. The Sinclair method asks the opposite: keep drinking, but take a pill beforehand that removes the payoff, and let the habit fade on its own terms.
That inversion is why it gets attention, and why it gets oversold. The pharmacology is sound and the FDA-approved drug at its center is one of the best-studied options in the field. The specific protocol built on top of it has a much smaller evidence base than its popular reputation suggests. Both of those things are true at once, and you deserve both.
What the Sinclair Method Is
Naltrexone is an opioid antagonist. It occupies the opioid receptors that alcohol indirectly stimulates to produce the warm, rewarding rush of the first couple of drinks. With those receptors blocked, the alcohol still reaches your bloodstream and still impairs you — but the surge that trains the habit does not arrive.
The FDA label describes naltrexone as indicated in the treatment of alcohol dependence and for the blockade of the effects of exogenously administered opioids. That much is settled and unremarkable — it is one of the three FDA-approved medications NIAAA lists for alcohol use disorder, alongside acamprosate and disulfiram, and NIAAA notes that naltrexone can be started while a patient is still drinking.
What the Finnish researcher John David Sinclair added was a claim about timing. Take the dose before drinking rather than every morning, and you are not just dampening cravings — you are running an extinction procedure on the learned behavior itself.
The Extinction Idea, and Where It Came From
Extinction is a basic learning principle: a behavior that stops producing its reward weakens over repeated trials. Sinclair's argument was that drinking is such a behavior, reinforced thousands of times by an opioid-mediated reward, and that blocking the reward while the behavior continues should unwind the learning.
His 2001 paper in Alcohol and Alcoholism laid out the case. Reviewing eight double-blind placebo-controlled trials across five countries, he reported a pattern in which naltrexone helped when paired with therapy that accepted drinking and taught coping skills, while none of them found any significant benefit of naltrexone over placebo when combined with support for abstinence. Animal work pointed the same way: opioid antagonists worked when paired with drinking and did nothing during periods of abstinence.
From that he drew three recommendations, which are essentially the protocol as it is practiced today: give naltrexone to people who are still drinking, instruct them to take it only when drinking is anticipated, and continue treatment indefinitely.
That last word does a lot of work, and it is usually the part that gets left out of the summary.
What the Protocol Asks of You
Stated without the marketing:
- A prescription. Naltrexone is prescription-only. There is no version of this you do on your own.
- A dose roughly an hour before every drinking occasion. Not most occasions. The extinction logic depends on the reward being blocked each time, so an unmedicated night is a session of the old learning going back in.
- Continued drinking, at first. This is the point and also the problem: the method needs drinking episodes to work on. If your situation calls for stopping now, this is the wrong tool.
- Months, not weeks. Extinction curves are gradual. Expect the change to be measured in months of consistent use.
- Indefinite continuation. Sinclair's own recommendation. People who stop taking it before drinking generally see the old pattern return, because the reinforcement returns with it.
The FDA label puts the compliance issue in language worth quoting: naltrexone is expected to have a therapeutic effect only when given under external conditions that support continued use of the medication, and has not been shown to provide benefit except as part of an appropriate plan of management. A pill taken sometimes is not this method; it is a different, weaker thing.
Does the Sinclair Method Work? What the Trials Show
Two trials carry most of the weight, and both are small.
Heinälä 2001 is the pivotal one. It enrolled 121 nonabstinent outpatients with alcohol dependence, randomizing them to coping-skills therapy or supportive therapy, and to naltrexone 50 mg daily or placebo. The design ran 12 weeks of fixed daily dosing followed by 20 weeks of targeted dosing — medication taken only when craving struck. The headline result: 27% of the coping-plus-naltrexone group had no return to heavy drinking across the full 32 weeks, against 3% of the coping-plus-placebo group. The authors concluded that prior detoxification was not necessary and that targeted dosing maintained the reductions.
Read that carefully. It is a large relative difference on a small absolute base — 27% is roughly one person in four, and the trial ran under eight months. Dropout was 16.5% in the first 12 weeks and roughly doubled after that.
Kranzler 2009 tested targeted dosing directly, in 163 problem drinkers whose goal was moderation, comparing daily and targeted schedules against placebo over 12 weeks with daily interactive-voice reporting. Targeted naltrexone reduced drinks per drinking day at week 12 across both sexes, with the clearest effect in men. The authors supported the targeted approach while noting that additional strategies or more potent medications may be needed to improve on it.
For the wider drug, the evidence base is far larger. A 2023 systematic review in JAMA pooled 118 trials and 20,976 participants and found oral naltrexone at 50 mg/day had a number needed to treat of 11 to prevent one return to heavy drinking, and 18 to prevent a return to any drinking. It recommended oral naltrexone and acamprosate as first-line pharmacotherapies alongside psychosocial treatment.
Those NNTs are the honest scale of the effect. Naltrexone helps a meaningful minority of the people who take it. It is not a switch.
Where the 78% Figure Comes From
Search for this method and you will meet a number: a 78% success rate. It is worth knowing exactly what it is attached to.
A treatment provider site that promotes the protocol states it directly — the officially reported success rate of the Sinclair Method in the The Cure for Alcoholism book is 78%. That is the provenance: a popular book, not a randomized trial reporting that endpoint. No trial in the peer-reviewed record above produced a 78% figure, and the definition of "success" behind it is not one you can check.
This is not an accusation of bad faith. It is the ordinary difference between an advocacy number and a measured one. When you see the figure quoted, treat it as a claim from a book, and compare it against 27% versus 3% over 32 weeks in Heinälä, and an NNT of 11 in the pooled JAMA analysis. Those you can trace.
The Criticisms Worth Taking Seriously
It lives or dies on compliance. Every unmedicated drinking session works against the extinction. That is a heavy ask over months, in exactly the domain where consistency is hardest, and no trial has demonstrated the protocol working in people who take it inconsistently.
The targeted-dosing evidence is thin. Two modest trials, one of which mixed 12 weeks of daily dosing into the design before the targeted phase began. Compare that with more than a hundred trials behind daily naltrexone generally. Clinical guidelines list naltrexone; they do not list this specific protocol as a distinct recommended regimen.
It requires you to keep drinking. For someone with a damaged liver, a pregnancy, a legal requirement, or heavy daily dependence, "carry on drinking for a few months" is not a neutral instruction.
It is not a standalone. Sinclair's own 2001 analysis found the benefit appeared with coping-skills therapy and not with abstinence-oriented support. The medication was never the whole intervention, in his data or anyone else's. If therapy is the other half, our guide to therapy approaches for stopping drinking covers what the options look like.
The drug has real cautions. The label documents cases of hepatitis and clinically significant liver dysfunction with naltrexone exposure, and advises stopping if signs of acute hepatitis appear. Nausea is common early on.
Who the Sinclair Method Does Not Suit
Naltrexone is contraindicated in people receiving opioid analgesics, people currently dependent on opioids including those maintained on methadone or buprenorphine, and people in acute opioid withdrawal. That is not a soft warning: taking it while opioids are on board can precipitate severe withdrawal, and if you take opioids for pain, this method removes that pain relief.
Beyond the label, the protocol is a poor fit if you drink heavily every day and stopping abruptly would be risky, because the honest first question there is about withdrawal safety rather than about extinction. The withdrawal risk checker is the place to start, and heavy daily drinkers should be having a medical conversation before changing anything. It is also a poor fit if your goal is to stop now, tonight — the method's mechanism needs drinking to act on.
How to Raise It With a Prescriber
Nothing here is medical advice, and none of it substitutes for someone who knows your history. What you can do is arrive prepared.
Bring an honest count. Most people underestimate, not from dishonesty but because a large glass of wine reads as one drink; the alcohol units calculator converts what you actually pour into standard drinks. A screening score helps too — the alcohol addiction quiz runs a standard questionnaire in your browser with nothing saved or sent.
Then ask the questions that matter: is naltrexone appropriate for me, given my liver results and my other prescriptions; would you consider targeted dosing before drinking, or would you rather start daily; and what would we measure to know whether it is working. NIAAA notes these medications are prescribable in primary care and reach only 1.6% of adults with past-year alcohol use disorder — you may well be raising the subject before your doctor does. Our overview of medications that help you stop drinking is a reasonable thing to read first.
Choosing a Goal Without Choosing a Side
The method belongs to the moderation end of the field, and moderation and abstinence are often argued about as if one were a betrayal of the other. In the data they are simply different goals with different evidence behind them, and plenty of people move between them.
If you are somewhere in the middle — not sure whether you want less or none — you do not have to resolve that before doing anything, though our guide to moderation vs abstinence lays out what predicts success on each side. Cutting back has its own practical toolkit, zebra striping lowers the peak of a given night without requiring a decision about the next one, and the craving timer gives an urge somewhere to go while it passes. Understanding what a craving actually is tends to make all of it easier.
Whatever you choose, a night that did not go the way you planned is information about a trigger, not a verdict on you. The extinction curve in Sinclair's own model is not a straight line either.
Frequently Asked Questions
How does the Sinclair method actually work?
Naltrexone blocks the opioid receptors that alcohol stimulates to produce its rewarding rush. Taken about an hour before drinking, it removes the payoff while the behavior continues, which is the setup for extinction — a learned habit weakening because it no longer delivers. Sinclair's recommendation was to take it only when drinking is anticipated and to continue indefinitely.
Does the Sinclair method have real evidence behind it?
Partly. Naltrexone itself is strongly evidenced: a 2023 JAMA review of 118 trials found a number needed to treat of 11 to prevent a return to heavy drinking. The targeted before-drinking protocol rests on much less — mainly a 121-person Finnish trial and a 163-person US trial, both positive but small, and no major guideline lists the protocol as a distinct regimen.
Is the 78% success rate real?
That figure traces to a popular book rather than to a published trial, and the provider sites quoting it say so. No peer-reviewed trial in the record reports a 78% endpoint for this protocol. The traceable numbers are smaller: 27% versus 3% with no return to heavy drinking over 32 weeks in the Heinälä trial, and an NNT of 11 in the pooled analysis.
Can I try the Sinclair method without a doctor?
No. Naltrexone is prescription-only, and it is contraindicated for anyone taking opioid painkillers or maintained on methadone or buprenorphine — taking it in those situations can trigger severe withdrawal. The label also documents liver cautions. This needs a prescriber who has seen your history and your bloodwork.
What happens if I skip the pill and drink anyway?
That drinking session reinforces the old learning, because the reward is not blocked. The method depends on the block being in place every time, which is why compliance is its main weakness. Skipping does not undo everything, but the extinction curve is built from consistent trials, and a missed dose is a trial in the other direction.