There are now four medications that are approved by the FDA for the treatment of alcohol dependence, so it is high time we take a fresh look at what our options are.

Antabuse. The scientific evidence that Antabuse (disulfiram) actually works is old and not terribly impressive; nonetheless, with over 40 years of experience prescribing it, most addiction psychiatrists are convinced that can be highly effective when given to real patients.

The largest study was published in 1986 (JAMA 1986;256:1449). In this study, 600 alcoholic males in the VA system were randomly assigned to three treatment arms: 200 patients were given Antabuse 250 mg QD, 200 were given Antabuse at only 1 mg QD, and 200 were given riboflavin. After one year of treatment, there was good news and bad news. The bad news was that there was no difference in the percentage of men remaining abstinent. The good news was that in the subpopulation of men who attended all the scheduled follow-up research appointments, Antabuse 250 significantly decreased drinking days vs. Antabuse 1 mg and vitamin. The bottom line: Those alcoholics who are highly motivated will do well on Antabuse.

Antabuse works by preventing the complete metabolism of alcohol, causing a build up of acetaldehyde and a consequent array of extremely unpleasant symptoms, such as nausea, vomiting, headache, dizziness, and flushing. But its real mechanism of action is instilling a healthy fear of drinking in alcoholics who are motivated enough to actually take the pill every day.

Most Antabuse-savvy clinicians start patients on 250 mg QD. If a patient reports drinking on 250 with no reaction, bump it up to 500. The common side effects are initial fatigue (so start it at bedtime), a metallic taste in the mouth, and GI side effects such as nausea or diarrhea. Since Antabuse can rarely cause liver toxicity, it’s good to check LFTs (liver function tests) at baseline, and then periodically thereafter.

Naltrexone. Naltrexone is an opioid antagonist that quickly received FDA approval on the strength of impressive data showing that it helped prevent relapse to alcoholism (Arch Gen Psychiatry 1992;49:876-880). Over time, a number of other studies have emerged, some positive, but others negative. The overall characterization of its effect is only “modest,” according to one meta-analysis (Alcohol Clin Exp Res 2001 Sep;25(9):1335-41).

Naltrexone is thought to work by attenuating that euphoric rush that alcoholics get after the first drink, and therefore making it less likely that patients who peer over the edge of the wagon will actually fall off it. Indeed, the studies show that naltrexone is better at preventing heavy drinking than it is at ensuring complete abstinence.

Just before we went to press with this issue, the largest-ever comparative study of alcoholism treatment was published and sheds more light on naltrexone’s role. The study in question is the COMBINE study, which was funded by the National Institute on Alcohol and Alcoholism (NIAAA) and had no pharmaceutical industry support (JAMA 2006;295:2003-2017). A total of 1,383 patients with alcohol dependence were assigned to nine different treatment arms, involving various combinations of naltrexone, acamprosate, and cognitive behavioral therapy (CBT).

As befits a large and complex study, attempting to read the labyrinthine results section is enough to make you reach for a drink. But the bottom-line conclusions are:

• Naltrexone (dosed at an average of 88 mg QD) was the big winner, beating placebo in percent of days abstinent and relapse to heavy drinking.


• Acamprosate (average 2537 mg QD) was the surprise loser, performing no better than placebo on any measure.


• “Medical management,” meaning nine 20 minute sessions (over 16 weeks) oriented around taking medication and maintaining abstinence, was quite a potent treatment alone, and actually outperformed twenty sessions of CBI. In fact, patients assigned to CBI plus placebo did better than those having CBI alone, implying that simply taking a pill, any pill, is important for alcoholics, possibly because it reinforces the “medical” nature of the problem and therefore is less stigmatizing.

To prescribe naltrexone, start at 25 mg QD for three days to minimize initial nausea, then increase to the final dose of 50 mg QD. Check LFTs at baseline and then every few months thereafter, since the medication carries a black box warning about the risk of drug-induced hepatitis.

Vivitrol. The results of the COMBINE study were welcome news for a company called Alkermes, which is about to release Vivitrol (naltrexone extended-release injectable), a long-acting depot version of naltrexone. (The medication will be marketed by Cephalon.) Recently approved by the FDA for the treatment of alcoholism, one gluteal injection of Vivitrol lasts 30 days, and gradually delivers 380 mg of naltrexone directly into the bloodstream. If you do the math, this is an average of only 12 mg of naltrexone per day. The reason such a mini-dose can still be effective is that the injectable version avoids the dreaded “first-pass effect,” in which oral medications are absorbed by the intestine, enter the portal circulation and go directly to the liver, where a significant amount of drug is metabolized and therefore inactivated before it can enter the brain.

Alkermes/Cephalon will be arguing that Vivitrol is crucial for alcoholics because compliance with oral medication is so poor. While it’s hard to argue with this logic, only studies explicitly comparing the two formulations will tell us whether Vivitrol really is worth recommending.

Vivitrol is not yet available in pharmacies, but should be sometime this June. Company sources tell us that they will set up an 800 phone number to help patients get the medication and to locate a place where the injection can be given. Interestingly, in the company studies, patients were started on Vivitrol right away, without first taking oral naltrexone, and tolerated this well (see www.vivitrol.com for prescribing information).

Campral. The COMBINE study was unkind to Campral (acamprosate), but Forest Laboratories is playing it cool, referring to the spectacularly unimpressive results as no more than a “single data point” (quoted in the Los Angeles Times, May 3, 2006). That may be true, but Campral has been plagued with efficacy-challenged data points ever since it applied for FDA approval.

At first, in 2002, FDA rejected Campral because the only positive data Forest could muster were three studies that had been done in Europe (the one large study conducted in the U.S. showed no separation between Campral and placebo). The agency looked at the European data, determined that the methodology was unreliable, and rejected the application, requiring that all the data be “audited” before they would reconsider it. This painstaking process took Forest two years, but Campral was eventually rewarded with FDA approval in July of 2004.

The new data are discouraging for psychiatrists, particularly since a study published in 2003 showed that the combination of naltrexone and Campral was more effective at preventing relapse than placebo or either drug alone (Arch Gen Psychiatry 2003;60:92-99). The COMBINE study, on the other hand, showed no benefit of combined treatment.

Campral presumably works (if it does work) by toning down the glutamate system, which remains hyperexcitable after alcohol abusers stop drinking (J Clin Psychiatry 2001;62[suppl 20]:42-48). It is dosed at two 333 mg tablets three times a day, making compliance a struggle for many patients. The only common side effect is diarrhea (about 10% of patients) and otherwise it appears to be perfectly safe, with no black box warnings in the package insert.

TCPR Verdict:
Naltrexone’s hot; Campral, cold as ice.