Photo credits
Why Disulfiram Still Matters
Disulfiram is the oddball of the addiction pharmacopeia. Naltrexone, acamprosate, topiramate: every other medication used for alcohol use disorder (AUD) works by turning down the wanting, blunting the reward, quieting the craving, rebalancing glutamate. Disulfiram does none of that. It leaves the craving fully intact and installs a hard, chemical consequence for drinking instead. Take it, drink alcohol, and within minutes the patient is flushed, nauseated, pounding, and miserable. The patient still wants to drink. Drinking now costs something immediate and physical.
That mechanism is exactly why disulfiram fell out of favor, and exactly why it deserves a place in your toolkit. It fell out of favor because it fails the randomized-controlled-trial test almost by construction: you cannot blind a drug whose entire point is a reaction the patient has been warned about, and unsupervised patients simply stop taking it the day they decide to drink. So the tidy efficacy data look weak. The supervised trials tell a different story. When someone actually watches the patient take it, the effect size is roughly 0.8 versus supervised control, one of the largest signals in all of addiction medicine, though part of that figure is inflated by an apples-to-oranges comparison across study designs and shouldn't be read as proof disulfiram beats other treatments head-to-head (see Part 10). The accountability structure built around disulfiram makes or breaks it; the pharmacology doesn't.
Disulfiram is a deterrent, not a treatment for craving. It works when, and only when, it is given to a motivated, abstinence-committed patient inside a system of supervision and support. Confirm abstinence before the first dose, screen out the cardiac and hepatic patients who can't tolerate a reaction, educate obsessively about hidden alcohol, and pair the pill with a witness. Do those things and you hand a subset of your most serious AUD patients a powerful tool.
The origins explain the drug. Disulfiram wasn't designed as a medication: in the 1930s it was an industrial chemical used in rubber-tire manufacturing, and workers who handled it found they became violently ill after drinking. Someone made the connection, and by 1951 the FDA had approved it for alcoholism. Its therapeutic effect is, quite literally, a repurposed industrial accident.
Part 1: Indications
FDA-Approved Use
- Chronic alcohol use disorder: as an adjunct in motivated patients pursuing total abstinence
That is the entire on-label story. Disulfiram is not a first-line AUD medication in most guidelines (naltrexone holds that spot), but it occupies a specific and defensible niche.
The Right Patient
Disulfiram is a precision instrument, not a broad-spectrum one. The ideal candidate has a recognizable profile:
- Committed to total abstinence, not moderation: this is non-negotiable. Disulfiram offers nothing to the patient who wants to cut back. It is for the patient who resonates with "one drink is too many, and a thousand is not enough."
- Highly motivated, typically someone who has tried and failed to control drinking despite real, sometimes catastrophic, consequences: lost jobs, lost relationships, medical damage.
- Has external support: a spouse, partner, friend, sober-living staff, or clinic that can supervise or witness the daily dose. This single feature separates disulfiram that works from disulfiram that doesn't.
- Engaged in the rest of recovery: 12-step programs, psychotherapy, or both. Disulfiram is a scaffold around behavioral change, not a substitute for it.
Disulfiram's real mechanism of action is psychological, not pharmacologic. The pill instills a healthy fear of drinking that lasts the whole day: the patient knows that today, drinking means a reaction, so today the decision is already made. The acetaldehyde reaction is almost beside the point; most patients on disulfiram never actually experience it. What changes behavior is the deterrent, made real and daily by the act of taking the pill in front of someone.
Where It Works Best
- Severe AUD with an available supervisor. The classic win: the patient who cannot trust their own in-the-moment judgment outsources that judgment to a once-a-day decision, witnessed by someone they trust.
- Episodic, high-risk situations. Even a patient not on daily maintenance can use disulfiram tactically (starting it before a wedding, a holiday, a work event with an open bar) to take drinking off the table for a defined stretch.
- The patient who wants a commitment device. Some patients explicitly want the external constraint. Disulfiram gives it to them in a form they can see and feel.
Off-Label: Cocaine Use Disorder
Disulfiram has a secondary, dopaminergic mechanism (inhibition of dopamine β-hydroxylase) that blunts the rewarding effects of cocaine, and it has been studied, particularly in patients with combined alcohol-and-cocaine use, in several hundred patients across trials. The signal is real but modest, and the doses needed appear higher: evidence suggests >350 mg/day is required for a meaningful cocaine-blunting effect, versus the standard 250 mg for alcohol. Treat this as an off-label option for the polysubstance patient, not an established indication.
Who Is a Poor Candidate
- Anyone unwilling or unable to commit to total abstinence.
- Patients with no supervision structure: unsupervised disulfiram performs no better than placebo.
- Cognitive impairment severe enough that the patient can't reliably understand and follow the "avoid all alcohol, including hidden sources" instruction (this is a safety issue, not just an efficacy one).
- Patients with the specific medical contraindications below: cardiac disease, significant hepatic impairment, active psychosis, pregnancy.
The single best predictor of disulfiram success is not severity of illness or motivation on paper, it's whether there is a real person who will watch the patient take the pill. If you can't identify that person or structure, you are prescribing the weak, unsupervised version of the drug. Solve the supervision problem before you write the prescription.
Part 2: Workup and Candidacy
Disulfiram has a short pre-start checklist, but it matters. Two of its steps, confirming abstinence and screening the heart, are safety-critical.
Confirm abstinence: the hard rule. The patient must have been alcohol-free for at least 12 hours before the first dose. Give disulfiram to someone with alcohol still on board and you provoke the very reaction you're trying to hold in reserve. Confirm sobriety clinically (and, if there's any doubt, with a breathalyzer or blood alcohol level) before that first tablet. Many clinicians build in a longer cushion, a full 24 hours, to be safe.
Baseline labs. LFTs (AST, ALT, bilirubin) are the one lab you need: disulfiram is hepatotoxic, so establish a baseline before starting and to interpret later changes. Many AUD patients already have elevated LFTs from alcohol itself. There is no routine level to monitor (disulfiram has no therapeutic serum target) and no mandatory ECG for everyone, but see the cardiac screening step below.
Screen the heart and vessels: this drug is a stress test in a pill. The disulfiram-ethanol reaction produces tachycardia, hypotension, and cardiovascular strain. In a patient with meaningful cardiac or cerebrovascular disease, that strain can be dangerous. Take a careful cardiovascular history, and pursue further testing (ECG, and cardiology input where indicated) in anyone with cardiac risk. You need to be reasonably confident this patient could tolerate a significant physiologic challenge, because that is exactly what a reaction is.
Screen for psychosis. Disulfiram's dopaminergic activity can worsen or precipitate psychosis. Active psychosis (or a recent history of it) is a contraindication. Ask.
Assess hepatic function. Disulfiram requires at least fair liver function. Significant hepatic impairment or cirrhosis is a contraindication (see Special Populations). Mild-to-moderate transaminase elevations (common in this population and often driven by the alcohol rather than any drug) do not by themselves bar treatment.
Build the supervision plan. Not a lab, but it belongs in the workup. Identify the witness. Explain the two ground rules to the supervising partner (below). Confirm the patient understands and consents to total abstinence. Covert dosing is off the table: the boxed warning forbids giving disulfiram to an intoxicated patient or without the patient's full knowledge. Establish the call protocol. If you skip this, you've prescribed the version of disulfiram that doesn't work.
Give the supervisor two rules to keep the daily ritual from poisoning the relationship. First, no unpleasant conversations during the dose: the monitoring moment is not the time to relitigate old grievances. Second, thank each other afterward. Framed this way, the witnessed dose becomes a small daily act of trust-rebuilding rather than surveillance. Also set the call protocol: ask the supervisor to phone you if the patient misses two consecutive days, so you can address non-adherence early instead of discovering a relapse weeks later.
Part 3: How to Start and Dose
Older adultselderly patientsFDA labelIn general, dose selection should be cautious, usually starting at the low end of the dosing range.In practiceNo fixed dose adjustment; start 125–250 mg once daily. Cognitive changes may threaten comprehension of abstinence instruction and reliable adherence.
Duration and stoppingmaintenance and discontinuationFDA labelMust be continued daily and uninterrupted until the patient is fully recovered socially and a basis for permanent self-control is established.In practiceNo taper; the alcohol reaction lasts up to 2 weeks after the last dose. The dangerous moment is precisely the 2–3-month mark.
Formulation
Oral tablet, 250 mg standard. No extended-release, no injectable, no liquid to fuss with. The simplicity is a feature.
Starting Dose and Titration
- Confirm 12+ hours of abstinence first (see Part 2). This is the gate.
- Start: 125–250 mg once daily. Many clinicians simply start at 250 mg.
- Timing: dose at bedtime if it makes the patient drowsy (the label says it is usually taken in the morning, and on retiring by patients who feel a sedative effect). The main early side effect is fatigue/drowsiness, so dosing at night lets the patient sleep through it.
- Typical maintenance target: 250 mg/day. Range 125–500 mg/day. Titrate upward only if there's a specific reason.
The Dose Ceiling: Lower Is Better
The historical practice of dosing 1,000–3,000 mg/day is abandoned, and for good reason. Those high doses drove the drug's dangerous reputation: the rare deaths from disulfiram-ethanol reactions occurred largely in that high-dose era and with deliberate provocation testing. Modern low-dose disulfiram (250 mg) is both safer and just as effective as the old high doses. There is no efficacy payoff for going high in AUD.
"Two-fifty and a witness." The dose that matters is about 250 mg (the label's average maintenance dose); the intervention that matters is the person watching it go down.
Cocaine use disorder (off-label): the exception to "stay low." Evidence suggests >350 mg/day is needed for the cocaine-blunting effect. For AUD alone, there is no reason to climb.
There Is No Serum Level to Chase
Unlike lithium, disulfiram has no trough level, no therapeutic window to titrate into, and no steady-state blood test. Dose to a fixed milligram target and monitor adherence and liver, not a concentration. Day-to-day management is simpler for it: the entire safety system rests on abstinence confirmation, LFTs, and the supervision structure.
Set the Timeline Expectation
Tell patients what recovery feels like on a clock. The pill's protection is immediate (it's active the day they take it), but the ease comes later. Around 2–3 months of sobriety, the brain has healed enough that cravings lessen and confidence returns; paradoxically, that's often when the temptation to stop the disulfiram rises ("I feel fine now, I don't need it"). Warn patients that the good stretch is not the finish line. It's the payoff for continuing.
Part 4: Monitoring
Monitoring disulfiram is refreshingly light compared with lithium, but the two things you do monitor matter.
Liver Function
| Timepoint | Draw |
|---|---|
| Baseline | LFTs (before the first dose) |
| 10 to 14 days after starting | LFTs (the follow-up draw the label suggests) |
| Every ~6 months during treatment | LFTs |
| Any time hepatotoxicity is suspected | LFTs promptly (jaundice, dark urine, RUQ pain, malaise, anorexia) |
- Interpretation: most mildly elevated LFTs in this population reflect alcohol, not disulfiram. Mild-to-moderate elevations generally tolerate continued treatment with closer follow-up.
- Act on a clear signal: a marked or rising transaminase pattern, or clinical signs of hepatitis, warrants holding the drug and evaluating. Disulfiram-induced hepatitis is rare but can be serious.
Adherence: the Real Monitoring Target
For disulfiram, adherence is efficacy, so monitor it as deliberately as you'd monitor a lab:
- Use the supervisor. The witnessed daily dose is your primary adherence measure.
- The two-day call rule. Ask the supervisor to call you if the patient misses two consecutive doses: early warning of drift.
- Track the outcomes that matter at each visit: drinks per day, days abstinent, relapse events, and the life-domain markers that reflect real recovery (marriage/relationship stability, job retention, paying bills, stable housing). These tell you whether the whole structure is holding, not just whether the pill is being swallowed.
Cardiovascular Check-In
No routine repeat ECG is required for a stable patient, but revisit cardiac status if the patient develops new cardiac symptoms or if their risk profile changes.
Part 5: Side Effects and How to Manage Them
Disulfiram has two entirely separate side-effect stories. One is the ordinary tolerability profile of the drug taken without alcohol, mostly minor. The other is the disulfiram-ethanol reaction, which is the drug's whole reason for existing and gets its own section (Part 6). Don't conflate them.
The Everyday Side Effects (No Alcohol Involved)
These are the effects of simply taking disulfiram, and most are mild.
- Fatigue / drowsiness. Common early, usually improves. Management: dose at bedtime.
- Metallic or garlic-like taste in the mouth. Expected, harmless. Reassure.
- Bad breath / body odor. Warn the patient in advance; reinforce dental hygiene. Usually tolerable.
- Maculopapular rash. Monitor. Usually manageable, but a significant or spreading rash may require discontinuation.
- Mild nausea / GI upset. Usually mild. Be sure it isn't a low-grade reaction from an unrecognized alcohol exposure; ask about mouthwash, cough syrup, sauces.
- Headache. Generally mild.
Peripheral Neuropathy
Disulfiram can cause a dose- and duration-related peripheral neuropathy. Ask about and monitor for paresthesias: numbness, tingling, burning in the feet or hands. It is usually reversible if caught and the drug stopped, so take new sensory complaints seriously and examine.
Hepatotoxicity
Rare at 250 mg but real, and the reason for the LFT schedule. It is idiosyncratic, can range from asymptomatic enzyme elevation to overt hepatitis, and at its worst can progress to fulminant, sometimes fatal hepatic failure requiring transplant, occasionally even after many months of uneventful therapy. Counsel the patient to report jaundice, dark urine, right-upper-quadrant pain, unusual fatigue, or loss of appetite immediately. Baseline and periodic LFTs plus symptom vigilance are the safety net (Part 4).
Psychiatric Effects
- Psychosis: disulfiram can worsen or precipitate psychosis via its dopaminergic effects, hence the contraindication in psychotic patients. If new psychotic symptoms emerge, stop the drug.
- Depression / mood changes: exercise caution in severe depression with suicidal ideation; the mechanism of any mood effect is unclear, but the caution is warranted.
The everyday side effects are, with the exception of neuropathy and hepatotoxicity, nuisances rather than hazards, and they're worth naming up front so the patient isn't alarmed. The metallic taste and the bad breath in particular surprise people; a five-second warning at prescribing prevents a worried phone call and a premature discontinuation.
Part 6: The Disulfiram-Ethanol Reaction
The reaction is disulfiram. Understanding it cold, and teaching it to your patient cold, is the core of prescribing this drug safely.
The Mechanism
Normally, the body metabolizes alcohol in two steps: alcohol to acetaldehyde (via alcohol dehydrogenase), then acetaldehyde to harmless acetate (via aldehyde dehydrogenase). Disulfiram irreversibly inhibits aldehyde dehydrogenase. The second step stalls. Acetaldehyde, the toxic intermediate responsible for much of a hangover's misery, accumulates to high levels the moment alcohol enters the system. The reaction is an instant, amplified, chemically enforced hangover.
Because the enzyme inhibition is irreversible, the effect does not switch off when the last pill wears off. The body has to synthesize new enzyme, so the disulfiram-ethanol reaction can persist for up to two weeks after the last dose. The two-week tail has real clinical value: a patient who impulsively decides to drink two days after quitting the drug is still protected, which builds in a window for reconsideration.
The Reaction, in Order
- Onset: 5–15 minutes after alcohol ingestion.
- Typical symptoms: flushing (face, neck), throbbing headache, a pounding sensation in the head and neck, sweating, thirst, nausea and often copious vomiting, chest pain, palpitations, tachycardia, shortness of breath and hyperventilation, hypotension, dizziness/vertigo, blurred vision, weakness, marked uneasiness, and confusion.
- Severe reactions (uncommon): respiratory depression, cardiovascular collapse, arrhythmias, myocardial infarction, acute heart failure, seizures, unconsciousness, and, historically and rarely, death.
How Dangerous Is It, Really?
The drug's reputation outruns the modern data. The historical fatalities clustered in the era of 1,000–3,000 mg dosing and deliberate provocation ("challenge") testing, which the label says has been largely abandoned. Lowering the dose to 250 mg is what tamed the safety profile. That is precisely why you screen out cardiac and cerebrovascular patients up front: the danger is concentrated in the patient whose heart or vessels can't absorb the stress, not in the reaction per se at modern doses.
Management of a Reaction
- Most reactions are self-limited and supportive care (rest, reassurance, fluids, antiemetics) suffices.
- A severe reaction is an emergency: significant chest pain, syncope, hypotension, arrhythmia, respiratory distress warrants ED evaluation and supportive/resuscitative care. There is no specific antidote; treatment is supportive (IV fluids for hypotension, oxygen, cardiac monitoring).
Hidden Alcohol: the Education That Prevents Accidental Reactions
Most disulfiram reactions in practice are accidental: the patient never meant to drink. The job at prescribing is to inventory the hidden sources with the patient, explicitly:
- Mouthwash (many contain alcohol; switch to an alcohol-free brand today)
- OTC cough, cold, and flu preparations (many liquid formulations are alcohol-based; read labels)
- Cooking wine and dishes cooked with wine, beer, or spirits (sauces, marinades)
- Vanilla extract and other alcohol-based flavor extracts
- Wine used for religious/sacramental purposes
- Certain foods and fermented products that carry trace alcohol content
- Aftershaves and colognes applied to skin
- Other topical products containing alcohol
Don't hand the patient a vague "avoid alcohol." Walk the list, item by item, at the prescribing visit. And specifically tell them to buy alcohol-free mouthwash today, before they leave the office.
Part 7: Drug Interactions
Disulfiram's interactions come from two directions: it inhibits several hepatic metabolizing enzymes (raising levels of other drugs), and one antibiotic must be actively avoided.
Disulfiram Raises Levels of These Drugs (via Enzyme Inhibition)
Disulfiram slows the rate at which certain other drugs are metabolized, so their levels and the risk of toxicity can rise:
- Warfarin: reduced clearance can raise the INR and bleeding risk. Monitor INR closely and adjust the warfarin dose when starting or stopping disulfiram.
- Phenytoin: levels can rise into the toxic range. Monitor phenytoin levels and watch for toxicity.
- Benzodiazepines metabolized by oxidation, especially chlordiazepoxide and diazepam: reduced clearance, increased sedation and accumulation. Prefer benzodiazepines cleared by glucuronidation (lorazepam, oxazepam, temazepam), which are largely spared, if a benzodiazepine is needed. Monitor either way.
- Tricyclic antidepressants: reduced clearance; monitor for elevated TCA effects/levels.
- Other anticonvulsants (e.g., valproate): levels can shift; monitor.
The combination has been associated with confusional states and psychotic reactions, and both drugs independently produce disulfiram-like reactions with alcohol. Avoid the pairing. Be alert that metronidazole itself, and other agents like certain cephalosporins, can independently cause a disulfiram-like reaction with alcohol even in patients not on disulfiram.
Alcohol: the Intended "Interaction"
The disulfiram-ethanol reaction (Part 6) is, mechanistically, a drug interaction, the one deliberately being set up. Everything in the hidden-alcohol education section is interaction management.
Management Strategy
- Before starting disulfiram, review the full medication list for warfarin, phenytoin, oxidatively metabolized benzodiazepines, TCAs, and anticonvulsants, and plan to monitor levels/effects.
- Time INR and phenytoin checks around both starting and stopping disulfiram; the enzyme inhibition unwinds gradually given the drug's long duration of action.
- Screen specifically for metronidazole (including short courses for dental or GI infections) and avoid.
Part 8: Special Populations
Pregnancy
Avoid. Safe use of disulfiram in pregnancy has not been established, and the disulfiram-ethanol reaction poses added risk. For AUD in pregnancy, favor non-pharmacologic support and safer strategies; do not use disulfiram.
Lactation
Avoid. Disulfiram is expected to pass into breast milk, and safety data are lacking.
Hepatic Impairment
A relative-to-absolute contraindication depending on severity. Disulfiram is hepatotoxic and requires at least fair liver function. Cirrhosis or significant hepatic impairment is a contraindication. Mild-to-moderate transaminase elevation (common from alcohol itself) does not automatically bar treatment, but demands baseline LFTs, closer monitoring, and a low threshold to stop.
Cardiac and Cerebrovascular Disease
Contraindicated in significant cardiovascular or cerebrovascular disease. The disulfiram-ethanol reaction is a physiologic stress test; a heart or cerebral circulation that can't tolerate tachycardia and hypotension turns a manageable reaction into a catastrophe. This is the screening step you cannot skip (Part 2). Heart failure is at minimum a strong relative contraindication.
Elderly
No fixed dose adjustment, but two age-linked concerns dominate: cardiovascular disease is more prevalent (making the cardiac screen more important), and cognitive changes may threaten both comprehension of the abstinence instruction and reliable adherence. Assess both carefully before prescribing.
Renal Impairment
Disulfiram is hepatically metabolized, and no specific renal dose adjustment is standard. Renal impairment is not a primary concern for this drug the way hepatic and cardiac status are.
Cognitive Impairment
A contraindication when significant. The patient must understand the mechanism, commit to abstinence, and reliably avoid hidden alcohol sources, all of which require intact cognition. This is a safety line, not just an efficacy one.
Active Psychosis
Contraindicated. Disulfiram's dopaminergic activity can worsen or precipitate psychosis.
Part 9: Discontinuation
Stopping disulfiram is mechanically simple, and this is one of the few places it's easier than the alternatives.
- No withdrawal syndrome, no taper required. Disulfiram can be discontinued abruptly. There is nothing analogous to lithium's rebound or baclofen's seizure risk.
- But the drug's effect lingers. Because aldehyde dehydrogenase inhibition is irreversible, the alcohol reaction persists for up to two weeks after the last dose. Patients must understand this: quitting the pill does not make it safe to drink that evening. This lingering window is a safety feature, it protects the impulsive patient in the days after they stop, but it must be taught, not assumed.
- The relapse vulnerability outlives the drug. Disulfiram was never treating the craving; it was containing it. When the pill stops, the underlying AUD is still there. Discontinuation should be a planned step within an intact recovery structure, ongoing therapy, support groups, and other AUD pharmacotherapy if appropriate, not a solo decision made during a good stretch. The dangerous moment is precisely the 2–3-month mark when the patient feels well and reasons themselves off the drug.
The message to give at discontinuation mirrors the one for maintenance: "Feeling good is the medicine working, not proof you're done. If we stop, we stop on purpose, with the rest of your support still in place, and remember the reaction still works for about two weeks, so drinking isn't safe just because you stopped the pill."
Part 10: Disulfiram vs the Alternatives
Choosing among AUD medications is mostly about matching mechanism to patient goal. Disulfiram sits at one pole (deterrence) while the others cluster around craving and reward.
The Mechanistic Map
| Drug | Mechanism | Requires Abstinence? | Needs Supervision? | Core Use Case |
|---|---|---|---|---|
| Disulfiram | Aldehyde dehydrogenase inhibition, aversive reaction (deterrent) | Yes, total | Yes, critical | Motivated, abstinence-committed patient with a supervisor |
| Naltrexone | Opioid antagonist, blunts alcohol reward/euphoria | No (harm reduction OK) | No (improves adherence) | First-line; reduces heavy drinking, works even without full abstinence |
| Acamprosate | GABA-glutamate normalization | No | No | Post-detox maintenance of abstinence; safe in hepatic disease |
vs Naltrexone
Naltrexone is first-line for AUD and has the strongest evidence base (it was the winner in the large COMBINE trial). It works by blunting the euphoric rush after the first drink, so it reduces heavy drinking and supports harm reduction: it does not require the patient to commit to total abstinence, and it needs no supervisor. It's also gentler on the cardiovascular system. Choose naltrexone as the default; reach for disulfiram when the patient specifically wants (or needs) a hard external constraint, is committed to abstinence, and has a supervision structure, or when opioid use precludes naltrexone. Naltrexone dosing: 50 mg PO daily, or 380 mg IM monthly (Vivitrol).
vs Acamprosate
Acamprosate works via GABA-glutamate rebalancing and is the safest of the three (no cardiac risk, no hepatotoxicity, usable in liver disease) but it is also the weakest, and it flatly failed to beat placebo in COMBINE. Its high TID pill burden hurts adherence. A meta-analysis pooling three head-to-head trials put supervised disulfiram ahead of acamprosate, though all three were open-label, so read that margin with the blinding caveat in mind. Acamprosate's niche is the post-detox patient who needs a safe maintenance agent, especially with hepatic disease that rules disulfiram out.
vs Topiramate
Topiramate (off-label) cut the percentage of heavy drinking days by about 16 points more than placebo in the 371-patient Johnson trial. It works via GABA-up/glutamate-down mechanisms and, like naltrexone, permits moderation rather than demanding abstinence. Its cognitive and sensory side effects ("Dopamax": word-finding trouble, paresthesias) are the trade-off. A different tool for a different patient, one who wants to cut down rather than commit to a deterrent.
An Honest Word on the Efficacy Comparison
Disulfiram's headline effect sizes (roughly 0.8 supervised) are partly inflated by an apples-to-oranges artifact: disulfiram is scored against a total-abstinence standard, while naltrexone and topiramate "succeed" by reducing drinking without full abstinence. So don't read a bigger number as unambiguous superiority. The truthful summary: supervised disulfiram is effective in the right, committed patient; unsupervised disulfiram is not; and naltrexone remains the more broadly useful first-line agent.
Why Disulfiram Fell Out of Use, and Why That's Partly a Mistake
Disulfiram declined for reasons that are mostly not clinical: it fails conventional double-blind RCTs (you can't blind the reaction), unsupervised use looks like placebo, and its 1,000–3,000 mg-era reputation for danger stuck long after low-dose practice made it safe. But recent meta-analyses and the alcohol-mortality crisis have revived interest: pooled open-label trials put disulfiram ahead of naltrexone and acamprosate in head-to-head comparisons. For the committed, supervised patient, it is one of the more powerful things you can offer.
Mechanism
Disulfiram's primary action is irreversible inhibition of aldehyde dehydrogenase. Blocking this enzyme stalls alcohol metabolism at acetaldehyde, which accumulates and produces the aversive disulfiram-ethanol reaction. Because inhibition is irreversible, the effect persists until new enzyme is synthesized, hence the roughly two-week duration after stopping; that duration reflects how long aldehyde dehydrogenase takes to regenerate, not disulfiram's own elimination half-life (which is only several hours, via an active metabolite, diethyldithiocarbamate). A secondary mechanism (inhibition of dopamine β-hydroxylase, raising CNS dopamine) underlies its blunting of cocaine reward and also its capacity to worsen psychosis. The therapeutic effect in AUD is not pharmacologic craving-reduction at all: it is behavioral and psychological, a daily, witnessed act that installs a healthy fear of drinking. The pharmacology is the deterrent's enforcement mechanism; the accountability structure is the treatment.
Bedside Cheat Sheet
Before starting
- Confirm ≥12 hours (ideally 24) alcohol-free before the first dose, non-negotiable
- Baseline LFTs. Screen cardiac/cerebrovascular status; screen for psychosis and cognitive capacity
- Build the supervision plan: identify the witness; unsupervised disulfiram ≈ placebo
Starting & dosing
- 250 mg PO once daily (range 125–500 mg); the label says usually morning, with bedtime for patients who feel sedated. Fatigue is the reason for night dosing
- Low dose is the standard: 250 mg is as effective as, and far safer than, the abandoned 1,000–3,000 mg era
- No serum level to monitor. Cocaine off-label: evidence suggests >350 mg/day is needed (label maximum 500 mg/day)
Monitoring & the reaction
- LFTs at baseline and at 10 to 14 days (label), then periodically, about every 6 months in practice (and on any hepatotoxicity symptom)
- Monitor adherence via the supervisor; two-day miss triggers a supervisor call
- Onset 5–15 min after alcohol: flushing, headache, nausea/vomiting, tachycardia, hypotension. Severe (rare) means ED
- Effect lasts up to 2 weeks after stopping
Interactions & contraindications
- Disulfiram raises: warfarin (watch INR), phenytoin (watch levels), diazepam/chlordiazepoxide, TCAs, anticonvulsants
- Avoid metronidazole.
- Contraindicated: cardiac/cerebrovascular disease, significant hepatic impairment/cirrhosis, active psychosis, pregnancy, significant cognitive impairment, and the patient who won't commit to total abstinence
- Discontinuation: no taper, no withdrawal, but the reaction persists ~2 weeks
Disulfiram asks something unusual of the prescriber. It doesn't need a monitoring apparatus of blood levels and titration. It needs an accountability structure: a committed patient, a trusted witness, a walked-through list of hidden alcohol, and a plan for the day the patient feels well enough to want to quit. In exchange it offers what none of the craving-focused drugs can, an immediate, tangible, daily consequence that takes the in-the-moment decision to drink off the table entirely. It is not the right tool for most AUD patients, and it is not a first-line agent. But for the motivated patient who wants a hard line drawn, and has someone to help hold it, supervised low-dose disulfiram is no crude relic of psychiatry's past. It is one of the most powerful deterrents we have, and its power lies in the structure built around the pill.