Psychiatrists often take a “don’t ask, don’t treat” approach to alcohol use disorders, often because of the seeming futility of treatment. Why address addiction when many patients will continue to drink? Addiction, not surprisingly, behaves like other mental disorders. No one is too shocked when a patient with a history of major depression develops a new episode. Although we shoot for full remission, many patients are left with residual symptoms. The same is true of alcoholism: it often follows a relapsing-remitting course characterized by partial remission. In longitudinal studies stretching decades, only about 40% to 50% of patients achieve stable recovery with the remainder using alcohol to varying degrees (Vaillant GE, Arch Gen Psychiatry 1996;53(3):243–249; Öjesjö L et al, J Stud Alcohol 2000;61(2):320–322).

In this article, I will focus on medication management of alcoholism, with the understanding that the best treatment combines medications with some form of psychotherapy or other therapeutic support, such as AA meetings. See table 1 for a summary of the medications covered in this article.

Disulfiram
Disulfiram (Antabuse) inhibits a liver enzyme needed to metabolize alcohol. When alcohol is consumed, this leads to a build-up of acetaldehyde, which causes the characteristic disulfiram reaction: flushing, nausea, palpitations, and other miserable symptoms. In theory, these symptoms should serve as an extreme disincentive to drink.

Although disulfiram has been around forever (it was initially approved by the FDA in 1951), only a few decent clinical trials exist and their results are mixed. The most recent meta-analysis, which acknowledges issues with the underlying data, demonstrates that disulfiram may slightly increase abstinence in the short-term (Jørgensen CH et al, Alcohol Clin Exp Res 2011;May 25, online ahead of print). Studies involving supervised administration—medication compliance is a major problem—have shown more convincing benefit.

This plus-minus efficacy must be weighed against possible harms to patients. I often describe the disulfiram reaction, which involves tachycardia and hypotension, as a stress test in pill form. You need to be reasonably assured, either by history or diagnostic testing, that the patient can survive a major cardiovascular challenge before issuing a prescription. Although hepatotoxicity is rare, patients should also have at least fair liver function.

Disulfiram can pose special challenges for patients with other mental disorders. It rarely causes psychosis by increasing dopamine through its effects on catecholamine production. Disulfiram can also reduce the clearance of some benzodiazepines, tricyclic antidepressants, and anticonvulsants (Fishman MJ et al, ASAM Patient Placement Criteria: Supplement on Pharmacotherapies for Alcohol Use Disorders, 2010:62).

In light of this, I struggle with patient selection. Patients who have previously succeeded with disulfiram can usually convince me to prescribe it. Ditto for healthy patients with the goal of abstinence who struggle with impulsive drinking. I start at 125 mg per day and eventually titrate to 250 mg to 500 mg, if tolerated. You need to counsel patients to avoid all alcohol, which is often present in things we don’t think about—mouthwash, over-the-counter cough-and-cold preparations, wine used for religious purposes, and certain foods (eg, those con taining vanilla extract).

Naltrexone
Naltrexone (ReVia) hit the U.S. market in 1984. As an opioid blocker (antagonist), it blunts the rewarding aspects of alcohol use. On average, it reduces heavy drinking by 17% and drinking days by 4% (Rösner S et al, Opioid antagonists for alcohol dependence. Cochrane Database Syst Rev 2010, Issue 12). In the United States, heavy drinking for men is generally defined as more than four drinks per day or more than 14 drinks per week; for women, the corresponding thresholds are three and seven. Although naltrexone also reduces the amount of alcohol consumed per drinking day, it does not increase abstinence.

The latest buzz with naltrexone—although it’s getting to be old news by now—is the long-acting, injectable formulation (Vivitrol) that received FDA approval in 2006. Administered once a month, it reduces heavy drinking by about 25% (Garbutt JC et al, JAMA 2005;293(13):1617–1625). Patients who are abstinent from alcohol for seven days prior to injection (“lead-in abstinence”) appear to realize the greatest benefit (80% versus 21% reduction in heavy drinking).

Although compliance with oral naltrexone can be a problem, the substantial cost of Vivitrol is part of my clinical calculus. At about $1,200 per injection versus $85 for a month of tablets, third party payers are loathe to authorize Vivitrol without a compelling rationale. As there are no clinical trials comparing the two formulations, decisions need to occur on a patient-by-patient basis. I tend to reserve Vivitrol for chronic alcohol abusers with profound psychosocial impairment and a documented history of poor compliance with the oral formulation.

I typically start oral naltrexone at 25 mg per day and titrate to a target of 50 mg to 100 mg over several weeks. The injectable dose is 380 mg per month. Chronic use of opioid analgesic medications is an absolute contraindication and patients should be off of all opioids for at least a week before starting therapy. Common side effects include nausea, anorexia, and abdominal discomfort, although these can typically be avoided with gentle introduction. Side effects unique to Vivitrol include injection site pain and sterile abscesses.

Patients with mild-to-moderate liver enzyme elevations can usually take naltrexone safely. You should consult with an addiction specialist or gastroenterologist before prescribing if clear hepatic insufficiency (decreased albumin, increased bilirubin) or cirrhosis is present. After taking baseline liver function tests, ongoing monitoring is only necessary if there are symptoms to warrant it (Rie RK et al, Principles of Addiction Medicine, 4th ed. Philadelphia; LW&W, 2009).

Acamprosate
Acamprosate (Campral) was being used in Europe for a long time before receiving U.S. approval in 2004. Its mechanism of action is a bit of a mystery; however, the prevailing theory is that acamprosate normalizes GABA-glutamate balance, which is especially important in early recovery when there is theoretically too much glutamate and not enough GABA.

The acamprosate literature in recent years has been inconsistent and therefore hard to interpret. The most rigorous meta-analyses still suggest that acamprosate reduces lapse/relapse by 14% compared to placebo but does not affect heavy drinking. Acamprosate has been shown to be either inferior or superior to naltrexone, depending on the trial, with no difference between the two when data are pooled (Rösner S et al, Acamprosate for alcohol dependence. Cochrane Database Sys Rev 2010, Issue 9).

There has been a lot of hand-waving about who is most appropriate for acamprosate and when to start it. Acamprosate is likely more effective if a patient’s stated goal is abstinence and there is a period of lead-in abstinence prior to initiating therapy (Chick J et al, Alcohol 2000;35(2):176–187). Moreover, a small open-labeled trial demonstrated that patients given acamprosate during detoxification did worse in the long run compared those who began therapy later (Kampman KM et al, Addict Behav 2009;34(6–7):581–586).

Although acamprosate’s efficacy is modest, it scores points for safety and tolerability. Diarrhea is the most common side effect but this rarely leads to discontinuation of therapy. Acamprosate is renally excreted, which makes it a good choice for patients with hepatic dysfunction unless they also have chronic kidney disease (consult with an addiction specialist or nephrologist for recommendations about dose reduction). The standard regimen, 666 mg three times a day, poses some compliance problems due to dose frequency.

Topiramate
Topiramate (Topamax) was approved by the FDA in 1996 for seizure disorder, and although it does not have formal approval for alcohol use disorders, it is generally considered to be part of the standard addiction pharmacopeia. Topiramate has a complex mechanism of action but, at minimum, appears to fiddle with the GABA-glutamate teetertotter. It acts as an anticonvulsant by increasing GABA, the brain’s primary inhibitory neurotransmitter, and decreasing glutamate, the brain’s primary excitatory neurotransmitter. This is also its presumed mechanism of action for alcohol dependence.

Like naltrexone, topiramate primarily affects heavy drinking. A recent meta-analysis demonstrated that it reduced heavy drinking days by about 23% with small improvements in abstinence (Arbaizar B et al, Actas Esp Psiquiatr 2010;38(1):8–12). Although the trials used in the meta-analysis were small and more data are needed, topiramate appears to be at least equivalent to naltrexone in head-to-head trials.

Topiramate has a number of clinically relevant side effects including cognitive slowing (“Dopamax”), anorexia and sensory symptoms (numbness, tingling, and taste alteration). It is renally excreted and can cause acid-base issues and kidney stones in some patients. Despite this, I’ve found it to be well tolerated with slow titration. I start with 25 mg per day and increase by 25 mg per week to an initial target in the 150 mg to 200 mg per day range. Doses as high as 300 mg are a viable option (Johnson BA et al, JAMA 2007;298(14):1641–1651), although I typically reserve this for partial responders who aren’t experiencing side effects.

Baclofen
Baclofen (Lioresal) is a very old drug recently made new. Originally approved by the FDA in 1977 for neuromuscular spasticity, it acts as an agonist at a subtype of GABA receptors. Like acamprosate and topiramate, this is felt to modulate GABA-glutamate balance. Most clinical trials in addiction have narrow inclusion criteria that limit external generalizability. Patients with major mental disorders, multiple substance use disorders, and general medical conditions are usually tossed out. The remaining patients are great for research purposes but don’t resemble any of the patients that we typically treat. It was thus refreshing when a trial was conducted on alcohol dependent patients with a history of cirrhosis (Addolorato G et al, Lancet 2007;370(9603);1915–1922). Thirty of the 42 patients (71%) allocated to baclofen achieved abstinence compared to only 12 of the 42 (29%) who received placebo. As would be expected, there were no adverse events related to the liver (baclofen is renally excreted). Dosing was initially 5 mg three times per day, increasing to 10 mg three times per day beginning on the fourth day.

I generally initiate baclofen in this range but often titrate much higher (for example, I presently have a patient on 30 mg three times per day). The autobiographical case report that stimulated interest in baclofen involved a physician with alcohol dependence who titrated as high as 270 mg per day before backing off to a maintenance dose of 120 mg (Ameisen O, Alcohol Alcoholism 2005;40(2):147–150).

Baclofen is generally well tolerated. Cognitive slowing, somnolence, and neuromuscular
weakness, which are dose related, are possible side effects.

Mix and Match
Combining medications with complementary mechanisms of action has intuitive appeal, but clinical trials have yielded disappointing results. One of the more famous trials, The COMBINE Study, found that naltrexone plus acamprosate was no better than naltrexone alone (Anton RF et al, JAMA 2006;295(17):2003–2017). Despite this, some of us cross our fingers and mix and match. If you go this route, I’d suggest sequential introduction to pinpoint possible benefits and harms.