# Alcohol use disorder

Alcohol use disorder in adults: reward, learning and physical adaptation, safe withdrawal, medicines, psychological treatment and continuing support.

Information, not medical advice. Discuss personal concerns with a qualified health professional.

[Watch the film](https://www.youtube.com/watch?v=hH4AJ-1yKQU) · [All films and essays](https://evidencepress.org/why-minds-break/films/)

Alcohol use disorder develops when drinking becomes persistently difficult to control and continues despite significant problems. It can involve craving, drinking more than intended, repeated unsuccessful attempts to cut down, and alcohol taking priority over other parts of life. People reach this point by different routes. Inherited susceptibility, childhood and adult circumstances, opportunities to drink, and alcohol's effects on learning and the nervous system can all contribute.

Three things need to be distinguished: how much someone drinks, physical dependence, and alcohol use disorder. Physical dependence means the body has adapted so that reducing or stopping alcohol can produce withdrawal. It often accompanies severe alcohol problems, but withdrawal is not required for the disorder. Someone can have serious difficulty controlling drinking without being physically dependent. Conversely, neither a drinking total nor a brain scan explains an individual's whole situation. [NIAAA's guide to risk, diagnosis and recovery](https://www.niaaa.nih.gov/health-professionals-communities/core-resource-on-alcohol/alcohol-use-disorder-risk-diagnosis-recovery)

## Why some people are more vulnerable

Inherited differences contribute substantially, but they do not determine who will develop a problem. A review of twin and adoption studies estimated that inherited differences accounted for about 49% of the variation in susceptibility in the populations studied, with a plausible range of 43–53%. This does not mean that half of an individual's disorder is genetic, or that half of cases are unavoidable. It describes differences across a population living in particular circumstances. Family history also reflects shared environments. [Verhulst and colleagues' review of twin and adoption studies](https://pubmed.ncbi.nlm.nih.gov/25171596/)

Most genetic contributions are individually small. A large study involving more than a million people, including about 102,000 with alcohol use disorder, identified 37 associated regions of the genome. Finding a region associated with a diagnosis does not establish which gene matters or how it acts. Genetic influences on drinking quantity also overlap only partly with influences on impaired control and disorder. Some susceptibility is shared with broader difficulties regulating behaviour or with other mental health conditions; shared genetic associations do not show that one condition causes another. [The large genetic study of alcohol use disorder](https://www.nature.com/articles/s44220-024-00353-8)

One unusually clear biological example concerns alcohol breakdown. The body converts alcohol first into acetaldehyde and then into acetate. Variants in the genes involved can change this process. Reduced activity of the enzyme ALDH2 allows acetaldehyde to accumulate, producing flushing and other unpleasant effects that often discourage drinking. Some variants affecting ADH1B, involved in the first step, also reduce risk, although their effects are not identical to the pronounced flushing associated with reduced ALDH2 activity. These variants occur at different frequencies in different populations. They illustrate a specific route from a genetic difference to a bodily response and behaviour, rather than a general genetic explanation of every case. [Review of alcohol-metabolising genes](https://pmc.ncbi.nlm.nih.gov/articles/PMC6286250/)

Development matters too. Long-term studies find that early problems with impulsivity, conduct or regulating behaviour predict later alcohol use disorder more consistently than anxiety or depression alone. In a review of 12 studies involving about 19,400 people, the association was an odds ratio of 1.62 for these outward behavioural difficulties, with an estimated range of 1.39–1.90, and 1.21 for inward distress such as anxiety and depression, with a range of 1.05–1.39. These are comparisons of odds, not proof that the difficulties caused corresponding increases in risk. Studies differed, and selective publication could have exaggerated some associations. [Review of childhood and adolescent pathways to alcohol use disorder](https://pubmed.ncbi.nlm.nih.gov/31035781/)

For one person, drinking may begin with excitement, social rewards and impulsive decisions. For another, relief from worry, painful memories or low mood may become important. Anxiety and depression can precede alcohol problems, result from them, or share underlying causes with them. The idea that everyone with alcohol problems is simply treating an earlier emotional disorder is therefore too narrow.

Childhood abuse, neglect, violence and instability are associated with later substance problems. Possible pathways include changes in stress responses, difficulty managing emotions, disrupted relationships, and greater exposure to harmful situations. But adversity often occurs alongside poverty, family substance problems and other disadvantages. Retrospective accounts can be imperfect, and many reviews combine alcohol with other substances. The evidence supports taking adversity seriously without assuming that a particular experience independently caused a particular person's disorder. [Review of reviews on childhood adversity and substance misuse](https://pubmed.ncbi.nlm.nih.gov/36411791/)

Family circumstances can make a difference even when inherited risk is high. A Swedish study compared high-risk siblings brought up in their original families with siblings adopted into screened homes. Among 1,316 full-sibling and 4,623 half-sibling groups, the adopted groups had lower rates of recorded alcohol use disorder: the adjusted rates over time were about 24% and 23% lower respectively. Adoption was not random, and the records mainly captured more serious cases, so this is not a precise estimate of what changing any child's home would achieve. It nevertheless supports an environmental contribution beyond a simple resemblance between parents and children. [The Swedish adoption study](https://pmc.ncbi.nlm.nih.gov/articles/PMC7870033/)

Starting to drink young predicts later problems, but the age of the first drink is also a marker of access to alcohol, family circumstances and pre-existing vulnerability. In a review of prospective studies, accounting more carefully for these factors weakened or removed several associations. That does not make heavy drinking in adolescence harmless; it means that postponing one first drink is not, by itself, a complete explanation of prevention. At a population level, price, availability and marketing shape exposure. Evidence supports policies that increase prices or restrict availability and marketing to reduce consumption, without allowing an exact prediction of how many individual cases of disorder each policy will prevent. [Review of drinking initiation and later problems](https://pmc.ncbi.nlm.nih.gov/articles/PMC4158030/), [IARC's evidence summary on alcohol policies](https://www.iarc.who.int/wp-content/uploads/2025/10/IARC_Evidence_Summary_Brief_6.pdf)

## What repeated drinking changes

Alcohol acts on several kinds of chemical communication between nerve cells. Broadly, it increases some of the effects of GABA, a chemical involved in slowing activity, and reduces some effects of glutamate, involved in activating cells and learning. This is not a simple whole-brain switch: effects depend on dose, cell type, previous drinking and the connections between cells. Slowing a cell that normally restrains another cell can even release that second cell from restraint. [Review of alcohol's effects on brain signalling](https://pmc.ncbi.nlm.nih.gov/articles/PMC6566861/)

With repeated exposure, the nervous system compensates. This helps explain tolerance, when the same amount has less effect, and withdrawal, when alcohol levels fall but the compensating changes remain. The resulting overactivity can cause shaking, sweating, anxiety, poor sleep and a racing pulse. Severe withdrawal can cause seizures or delirium, involving profound confusion and disturbed awareness. Repeated withdrawal episodes can increase susceptibility to severe symptoms, sometimes called kindling, although this does not mean every episode inevitably becomes worse. Physical adaptation is important, but it does not explain all impaired control or why difficulties can continue long after acute withdrawal ends. [Review of the addiction cycle and adaptation](https://pmc.ncbi.nlm.nih.gov/articles/PMC6135092/)

Alcohol also teaches. Enjoyment, stimulation or relief makes a behaviour more likely to be repeated. Places, people, times of day, feelings and routines become associated with drinking. Later, these cues can attract attention and trigger wanting before a person has consciously decided to drink. Dopamine helps with motivation and learning about rewards; it is not simply a measure of pleasure. [NIAAA's explanation of addiction and recovery in the brain](https://www.niaaa.nih.gov/health-professionals-communities/core-resource-on-alcohol/neuroscience-brain-addiction-and-recovery)

Human studies show that alcohol can release dopamine in reward-related brain regions. Some studies of recently detoxified people also find reduced availability of certain dopamine receptors or a smaller response to a stimulant challenge. These findings concern particular groups and testing conditions. They do not establish that everyone with alcohol problems began with a dopamine shortage. A response to a laboratory drug challenge is also different from a response to a personally meaningful drinking cue. [Review of dopamine findings in alcohol use disorder](https://pubmed.ncbi.nlm.nih.gov/37963449/)

Nor does the pleasure of drinking necessarily disappear as problems develop. A study following 190 young drinkers for ten years repeatedly compared their responses to alcohol and a placebo. Stronger initial stimulation, liking and wanting predicted later disorder, and stimulation and wanting increased among those with the greatest subsequent difficulties. For some people, rewarding effects remain important; for others, particularly with substantial dependence, avoiding withdrawal or distress becomes increasingly influential. Both patterns can occur together. [The ten-year alcohol-response study](https://pubmed.ncbi.nlm.nih.gov/33397141/)

Researchers often organise these processes into three connected phases: drinking and reward; withdrawal and distress; and anticipation, craving and difficulty regulating behaviour. Brain systems involved in reward learning, stress, bodily awareness and planning all contribute. This is a useful framework rather than a compulsory sequence that everyone follows. Much of the detailed evidence about particular stress chemicals, including corticotropin-releasing factor and dynorphin, comes from animal experiments. Such experiments can show a causal mechanism in the model studied without proving that it dominates every human case. [Review of the three-stage framework](https://pubmed.ncbi.nlm.nih.gov/27475769/)

Heavy sustained drinking can make periods without alcohol more anxious, irritable, sleepless or low in mood. Drinking then provides temporary relief, strengthening the pattern. These feelings may also arise from an independent mental health condition or difficult circumstances. The body's stress system is not simply permanently turned up: a review of 46 studies found mixed cortisol findings during ongoing disorder, increased activity during withdrawal, and often a muted response to stress in early abstinence. What happens later is less clear. [Review of cortisol and alcohol use disorder](https://www.tara.tcd.ie/items/7033e856-fee5-4ca1-ab6e-8c416dd81be5)

Problems with attention, planning, stopping a response and weighing delayed consequences can also matter. They may predate heavy drinking, develop through exposure, or become worse when stress and craving compete for attention. A synthesis of 144 reviews identified reward, regulation and negative emotion as recurring areas of difficulty. This does not provide a test that identifies the cause in an individual. [Review of psychological and brain mechanisms](https://pubmed.ncbi.nlm.nih.gov/35295672/)

Habit is another part of the picture, but habitual behaviour and compulsive behaviour are not identical. A familiar routine can become automatic, while continued drinking can also remain directed towards a strongly valued immediate outcome, such as relief. Research does not establish a universal switch from thoughtful choice to an entirely automatic habit. Understanding what drinking currently does for a person is therefore still relevant. [Review of habit and goal-directed behaviour in addiction](https://pubmed.ncbi.nlm.nih.gov/39030169/)

## Recovery, uncertainty and newer biological ideas

Many difficulties can improve. A review of 16 studies following people over time, including 783 people with alcohol use disorder and 390 comparison participants, found that several aspects of attention, memory, perception and planning improved towards usual levels over roughly six to twelve months of abstinence; some improved earlier. Recovery varies with health, brain injury, other conditions and circumstances. Repeating cognitive tests can itself improve scores, people lost to follow-up may differ from those retained, and abstinence was not always verified equally well. These findings offer grounds for hope, not a promise of complete recovery on a fixed timetable. [Review of cognitive recovery during abstinence](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0296043)

Other possible mechanisms include inflammation, changes in gut bacteria and the gut lining, and changes in how genes are switched on or off. These ideas are biologically plausible but less settled. A review of 50 clinical studies of the gut and alcohol problems found that most were snapshots taken at one time, often in predominantly male groups with alcohol problems or liver disease. Diet, medicines, liver damage and drinking itself can affect the findings. A marker accompanying liver disease is not automatically a cause of impaired control over drinking. Studies of changes in gene regulation also face small samples and the difficulty of inferring what happens in the brain from other tissues. [Review of the gut–alcohol relationship](https://pubmed.ncbi.nlm.nih.gov/38719790/), [Review of emerging mechanisms](https://www.jci.org/articles/view/172885)

More broadly, comparing brains after years of drinking cannot cleanly separate earlier vulnerability from alcohol exposure, withdrawal, smoking, nutrition, medicines and other illnesses. Much research concerns severe, treatment-seeking cases and disproportionately represents men and people of European ancestry. Findings cannot simply be extended to every milder community case. Different studies also measure different things: consumption, dependence, craving, withdrawal and diagnosed disorder overlap without being interchangeable.

The next useful studies will follow diverse people before problems begin, test whether changing a proposed mechanism actually improves behaviour and wellbeing, and establish whether biological or psychological profiles improve treatment choice beyond a careful ordinary assessment. There is currently no routinely validated brain scan, molecular test or biological subtype that tells clinicians the best treatment for an individual. [Review of alcohol use disorder and the research agenda](https://pmc.ncbi.nlm.nih.gov/articles/PMC11747793/)

## Choosing a goal and stopping safely

Treatment should begin with an assessment of drinking, physical dependence, health, other medicines, mental health, practical circumstances and the person's priorities. Abstinence is often the safest goal with moderate or severe dependence or significant physical or mental illness. Supported reduction may be appropriate in other circumstances. If a person initially wants to reduce rather than stop, care can still address risk and review the goal as things change. Most treatment evidence concerns adults seeking help; pregnancy, adolescence and complicated withdrawal require more specific assessment. [UK guidance on assessment and recovery planning](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/4-assessment-and-treatment-and-recovery-planning)

**A person who may be physically dependent should seek medical advice before abruptly stopping or sharply reducing alcohol.** Withdrawal can be dangerous. Seizures or severe confusion require emergency help. Clinicians assess whether withdrawal can be managed safely in the community or needs inpatient care, considering previous withdrawal, health, other substance use and available support. [UK guidance on medically assisted withdrawal](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/11-community-based-medically-assisted-withdrawal)

Short-term, medically supervised benzodiazepine treatment is the usual first-line medication approach to significant alcohol withdrawal. Thiamine, vitamin B1, is used to prevent or treat a serious deficiency-related brain disorder called Wernicke's encephalopathy; the preparation and route depend on the clinical situation. Withdrawal medicines need monitoring and are normally stopped after the episode. Completing withdrawal is an important beginning, but it does not by itself address the learned patterns and circumstances that can sustain alcohol problems. [UK guidance on pharmacological treatment](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/10-pharmacological-interventions)

## Medicines that help beyond withdrawal

Naltrexone and acamprosate have the broadest established evidence among medicines used to reduce return to drinking. They serve somewhat different purposes, and suitability matters as much as a league table. Naltrexone can help prevent a return to heavy drinking and can also support abstinence. Current opioid use can make it unsuitable because it blocks opioid effects and can precipitate withdrawal; liver health and other medicines require assessment. Nausea and difficulty taking it consistently can limit usefulness. Some clinical guidance allows naltrexone to start while someone is still drinking after assessment, rather than requiring a period of abstinence first. Any withdrawal risk from reducing drinking still needs a plan. [New York State clinical guidance on alcohol use disorder](https://www.suguidelinesnys.org/guideline/alcohol-use-disorder/)

Acamprosate is particularly used to help maintain abstinence after withdrawal. Kidney problems can limit its use, diarrhoea is a common adverse effect, and taking it several times daily can be inconvenient. Disulfiram is an alternative for selected people aiming for abstinence, especially when taking it can be supervised or supported. It creates an aversive reaction to alcohol rather than simply removing craving. That reaction can be dangerous, so selection, understanding and monitoring are essential. [UK medication guidance](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/10-pharmacological-interventions)

A 2023 review of 118 trials involving 20,976 people estimated that treating about eleven people with acamprosate rather than placebo prevented one additional return to any drinking. The corresponding estimate for oral naltrexone, studied at 50 mg a day, was about eighteen for any drinking and eleven for heavy drinking. These figures concern trial follow-up, generally at least twelve weeks, and usually compare medication with placebo when both groups also receive support. Eleven treated for one additional benefit means roughly nine extra people benefit per hundred treated; it does not mean only one in eleven improves. The estimates were uncertain: the ranges were 1–32 for acamprosate, 4–32 for naltrexone preventing any drinking, and 5–41 for preventing heavy drinking. Direct comparisons did not establish one universal winner. [The 2023 review of alcohol use disorder medicines](https://jamanetwork.com/journals/jama/fullarticle/2811435)

An injectable form of naltrexone can help when daily tablets are difficult. However, a 2025 study of 248 hospitalised patients did not find a significant difference in the primary drinking outcome between monthly injection and oral treatment at three months. Convenience, preferences, cost and access remain important; the injection is not automatically more effective for everyone. [Trial comparing oral and injectable naltrexone](https://pubmed.ncbi.nlm.nih.gov/40257810/)

Several other medicines have narrower or less consistent roles. Topiramate can reduce drinking but may cause tingling and difficulties with memory or thinking. It is used outside its specific alcohol-treatment licence in some settings; that status does not itself mean it is ineffective. Gabapentin may help selected patients, but sedation, kidney function and misuse risk matter. Baclofen is sometimes considered by specialists, including in some liver-related circumstances. A review found that it probably reduced return to drinking and increased abstinent days, especially after withdrawal, without showing the same clear benefit for heavy-drinking days or drinks per drinking day. Sedation, overdose risk and abrupt withdrawal from baclofen require care. Nalmefene has a restricted role in reducing consumption for selected people who do not need immediate withdrawal treatment, alongside support. [British Association for Psychopharmacology guidance](https://journals.sagepub.com/doi/10.1177/02698811251399593), [Cochrane review of baclofen](https://www.cochrane.org/evidence/CD012557_baclofen-alcohol-use-disorder)

## Psychological treatment, mutual aid and continuing care

Psychological treatment can help people recognise high-risk situations, manage urges, practise different responses and rebuild routines. Cognitive behavioural therapy links thoughts, feelings and actions; relapse prevention prepares for situations in which drinking is more likely. Motivational approaches help clarify goals and resolve mixed feelings about change. Work with social networks can make daily life more supportive of recovery. Couples work can help when a partner is willing and supportive, but current UK guidance does not recommend it where there is domestic abuse or the partner is also drinking problematically. Brief advice can help risky drinking, but is not a substitute for comprehensive care for dependence. [UK guidance on psychosocial interventions](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/5-psychosocial-interventions), [Guidance on brief interventions](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/3-identification-and-brief-interventions)

No single credible psychological therapy is best for every person and every outcome. Comparisons against no help usually answer an easier question than comparisons against another active treatment. The quality and consistency of the evidence also vary between approaches and outcomes. That is compatible with psychological help being a central part of care. [WHO guidance on psychosocial treatment](https://www.who.int/teams/mental-health-and-substance-use/treatment-care/mental-health-gap-action-programme/evidence-centre/alcohol-use-disorders/psychosocial-interventions-for-management-of-alcohol-dependence)

Alcoholics Anonymous and professionally delivered Twelve-Step Facilitation have substantial evidence for supporting abstinence. A Cochrane review covering 27 studies and 10,565 participants found that structured, manual-guided Twelve-Step Facilitation improved continuous abstinence at twelve months compared with other established approaches in the relevant trials. The relative increase was about 21%, with an estimated range of 3–42%, based on two studies with 1,936 participants. That is not a 21-percentage-point improvement, and it does not show superiority for every measure of drinking intensity or alcohol-related harm. Nor is evidence for a structured intervention identical to evidence for every form of informal meeting attendance. [Cochrane review of AA and Twelve-Step Facilitation](https://www.cochrane.org/evidence/CD012880_alcoholics-anonymous-aa-and-other-12-step-programs-alcohol-use-disorder)

People should have meaningful choices. Other mutual-aid options, including SMART Recovery, may fit different preferences, although their comparative trial evidence is less extensive. Peer relationships can provide practical knowledge, belonging and support beyond appointments. [UK guidance on recovery support and lived-experience initiatives](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/6-recovery-support-services-and-lived-experience-initiatives)

Combining medication with psychological and practical support is generally recommended. A 2025 review of 21 trials involving 4,746 people found that adding medicine to psychosocial treatment reduced return to heavy drinking over eight to twenty-four weeks. An illustrative comparison was about 74 people per hundred returning to heavy drinking with psychosocial treatment alone versus 64 with the combination. This does not mean that everyone's alcohol intake falls by fourteen per cent. The reverse question—how much specialist psychotherapy adds when medication already comes with structured clinical management—has a less certain answer. [Cochrane review of combined medication and psychological treatment](https://pmc.ncbi.nlm.nih.gov/articles/PMC11924338/), [WHO recommendation on combined treatment](https://www.who.int/teams/mental-health-and-substance-use/treatment-care/mental-health-gap-action-programme/evidence-centre/alcohol-use-disorders/combined-psychosocial-and-pharmacological-interventions-for-adults-with-alcohol-dependence)

The large COMBINE trial illustrates the distinction. Among 1,383 participants, naltrexone with supportive medical management and a substantial behavioural intervention each helped, but combining them did not produce a further clear improvement. Medical management included genuine follow-up and support, not merely handing over a prescription. Acamprosate did not help in that particular study, despite benefits across the wider evidence. Differences in participants, treatment context or adherence may contribute to such variation, but the explanation is not settled. [The COMBINE trial](https://jamanetwork.com/journals/jama/fullarticle/202789)

Access to medication should not depend on waiting for scarce specialist psychotherapy. Current UK guidance commonly recommends reviewing treatment around six months and extending it when beneficial, including to twelve months; treatment decisions should not be reduced to an inflexible maximum. Mental health treatment and help with housing, relationships or employment can proceed alongside alcohol care. There is no general reason to defer necessary mental health care for months while waiting for abstinence. Residential care is selected according to needs, and a lapse should prompt reassessment and support rather than the end of care. [UK medication guidance](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/10-pharmacological-interventions), [Guidance on co-occurring mental health conditions](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/18-people-with-co-occurring-mental-health-conditions)

## Promising treatments that still need clearer answers

New findings deserve attention without displacing established care prematurely. In a 2026 trial of 108 adults with alcohol use disorder and obesity, weekly semaglutide or placebo was added to cognitive behavioural therapy for twenty-six weeks. Semaglutide produced an additional reduction of about 13.7 percentage points in heavy-drinking days; the plausible range was 5.4–22 points. Eighty-eight participants completed treatment, and gastrointestinal side effects were more frequent. This single-centre study does not establish benefits for people across all body weights, long-term durability, or superiority to established alcohol medicines. Current UK guidelines predate these newer trials, so their omission should not be read as a considered rejection of the findings. [The 2026 semaglutide trial](https://researchprofiles.ku.dk/en/publications/once-weekly-semaglutide-versus-placebo-in-patients-with-alcohol-u/)

Another 2026 study, involving fifty adults given oral semaglutide or placebo, did not improve its main laboratory craving outcome, although a secondary measure of heavy-drinking days improved. A positive secondary finding alongside a negative main outcome calls for replication, rather than being presented as an unqualified success. [The oral semaglutide trial](https://pubmed.ncbi.nlm.nih.gov/42522065/)

Psilocybin with psychotherapy reduced heavy-drinking days in a 2022 trial that randomised 95 people. However, participants often recognised which treatment they had received, allowing expectations to influence results. A smaller 2025 study, with 37 people analysed after withdrawal, did not find improved abstinence or consumption after a different single-dose treatment programme. The studies used different settings and amounts of therapy, so the second is not an exact repetition of the first. [The 2022 psilocybin trial](https://jamanetwork.com/journals/jamapsychiatry/fullarticle/2795625), [The 2025 post-withdrawal psilocybin trial](https://pubmed.ncbi.nlm.nih.gov/40144690/)

A trial involving 96 people found that ketamine treatment increased abstinent days at six months, without a significant improvement in relapse rate. A 2026 trial of electrical brain stimulation through electrodes on the scalp, involving 337 people, found fewer heavy-drinking days, but no significant improvement in its other main outcome, total alcohol consumption. Missing drinking diaries and follow-up complicate interpretation. These approaches remain promising research areas rather than established replacements for routine treatment. [The ketamine trial](https://pubmed.ncbi.nlm.nih.gov/35012326/), [The 2026 brain-stimulation trial](https://pubmed.ncbi.nlm.nih.gov/42091052/)

Across treatment research, drinking less, avoiding heavy-drinking days, remaining completely abstinent, reducing craving and improving health are different achievements. A treatment can help one without clearly helping another. Self-reported drinking, missing follow-up, small studies, multiple outcomes and difficulty concealing treatment allocation all affect confidence. Evidence is generally stronger for drinking outcomes than for long-term effects on mortality, quality of life and daily functioning; limited direct evidence for those outcomes does not mean reducing harmful drinking has no health value.

The most useful next steps include larger comparisons of new and established treatments, longer follow-up, better representation of people with other illnesses or unstable living conditions, and clearer evidence about which combinations are practical and effective. For current care, the strongest approach remains individual assessment, safe withdrawal when needed, appropriate medication, psychological and social support, and sustained follow-up that can adapt when circumstances change.

---

AI-assisted adaptation of supplied research; not independently clinically reviewed. Original essay text: CC0 1.0. Published 7 November 2026.