Chances are good that we under-medicate alcoholics. According to one estimate, only 10% of alcoholics receive medications as part of their treatment (Jonas DE et al, JAMA 2014;311(18):1889–1900). That’s too bad, because these medications work.
In the interview in this issue, we learn from Amy R. Krentzman, MSW, PhD, about techniques from positive psychology that may help treat addiction (see “Using Positive Psychology to Help People with Addictions”), and in prior issues of CATR we have covered Alcoholics Anonymous, motivational interviewing, and other non-drug approaches. In this article, we focus strictly on medications that appear to be effective, both those approved by the US Food and Drug Administration (FDA) and those that are commonly used off-label.
How do we define an effective medication for alcohol dependence? While complete abstinence is often the goal, many studies are more realistic and settle for decreasing the frequency of heavy drinking (defined as more than five drinks per day for men and four for women) (Dawson DA, J Subst Abuse 2000;12(1–2):79–91).
Every doctor has his or her own approach to using these meds. Some swear by Antabuse—others find it nearly useless. Some prescribe Vivitrol, while others prefer the oral version of naltrexone. And there are many acamprosate lovers out there as well. We’ll go through the evidence so that you can fine-tune your own personal prescribing algorithm.
Naltrexone (ReVia, Vivitrol) is an opioid blocker, but oddly it is much more effective as a treatment for alcoholism than for narcotic abuse. In fact, you should consider it a first-line option for most patients. It’s safe to take if the patient is still drinking (unlike Antabuse) but it should be avoided in people with liver disease or those taking opioids. Several meta-analyses have reported that naltrexone decreases any drinking, but also heavy drinking, more than placebo (Jonas DE et al, op.cit; Rosner S et al, Cochrane Database Syst Rev 2010;(12):CD001867). The Cochrane meta-analysis included 50 randomized studies with almost 7,800 patients with alcohol dependence and found the risk of heavy drinking was decreased by 83% compared to the placebo patients. The dose most often used is 50 mg/day but some trials have used 100 mg/day.
Patients taking naltrexone will most often complain about nausea, headache, and dizziness but these tend to go away after continued treatment. Liver enzymes were elevated by five-fold in about 1.8% of naltrexone patients in one study, but returned to normal after it was discontinued; so checking liver enzymes should be part of the treatment with naltrexone.
For patients who have trouble sticking with their meds, a monthly injectable form of naltrexone (Vivitrol) can be an option. It’s not clear whether the injectable form is as effective as the oral since they haven’t been compared head-to-head. A large study with the intramuscular (IM) form showed a 25% decrease in heavy drinking compared to placebo, a less robust difference than seen with the oral formation (Garbutt JC et al, JAMA 2005;293(13):1617–1625). It’s usually given as a 380 mg IM to the gluteal muscles every four weeks. Patients may still have nausea with the IM version and some patients may also have injection site reactions, some of them serious (eg, induration, cellulitis, hematoma, abscess, necrosis).
Acamprosate (Campral) is a good option for patients with liver failure, acute hepatitis, or liver enzymes that are elevated more than three to five times normal levels. Its efficacy evidence is a bit mixed and the research suggests that it may be a better choice for patients who are already abstinent, in order to maintain sobriety by reducing craving.
One meta-analysis of acamprosate treatment combined nearly 7,000 patients from 24 trials and found that the drug reduced relapse rates vs. placebo. In a population of alcohol-dependent patients, acamprosate would be expected to prevent drinking after detox in one out of nine patients who would otherwise have relapsed (Rosner S et al, Cochrane Database Syst Rev 2010;(9):CD004332). Another meta-analysis of more than 4,000 patients from 17 European studies found that six-month abstinence rates were significantly higher with acamprosate than with placebo—36% vs. 23% (Mann K et al, Alcohol Clin Exp Res 2004;28(1):51–63). On the other hand, three studies from the US and Australia found no benefit of acamprosate over placebo in abstinence rates or time to relapse. One of these, the COMBINE study, compared both acamprosate and naltrexone to placebo and found acamprosate was no better than placebo while naltrexone was significantly better (Anton RF et al, JAMA 2006;295(17):2003–2017).
It’s likely that differences in studies and patients contributed at least in part to these conflicting findings but, for now, acamprosate can be thought of as an effective option for some patients, particularly those who can’t take naltrexone because of liver issues. It’s a bit less convenient than the once daily naltrexone because it’s taken three times a day in doses of 666 mg each. It’s pretty well tolerated, with diarrhea and fatigue, as the most common complaints.
Disulfiram (Antabuse) is really only for those who are highly motivated to maintain abstinence, have good medication adherence, or will be taking it in a supervised way. It inhibits aldehyde dehydrogenase, which is the enzyme needed to break down alcohol’s main metabolite, acetaldehyde. Drinking while taking disulfiram will cause unpleasant effects such as sweating, headache, flushing, shortness of breath, a drop in blood pressure, racing heart, nausea, and vomiting. The patient knows there may be severe and potentially serious effects so this hopefully motivates them to not drink. How well does it work?
There are fewer studies, but a recent meta-analysis of two trials with almost 500 patients found that disulfiram was no better than placebo on any measure of drinking or relapse (Jonas DE et al, op. cit). Another large, year-long study of 605 US veterans also found it to be no better than placebo for abstinence or time to relapse, but it did reduce drinking days in those who did drink (Fuller RK et al, JAMA 1986;256(11):1449–1455).
These poor results may have to do with subjects’ poor compliance with disulfiram. When patients’ use of the drug is supervised, they do better. For example, one such study compared naltrexone, acamprosate, and disulfiram, combined with psychological interventions, over 52 weeks in 243 patients in Finland. Unlike many studies, the disulfiram patients in this study were required to give the name of a person responsible for supervising them when they take their medicine (Laaksonen E et al, Alcohol Alcohol 2008;43(1):53–61). Disulfiram fared better than both naltrexone and acamprosate in the first 12 weeks on all outcome measures—time to first drink, number of heavy drinking days, average weekly alcohol consumed, and number of abstinent days. For the rest of the yearlong study, the other two drugs caught up with disulfiram on all outcome measures except number of abstinent days, for which disulfiram was the ultimate winner.
When starting disulfiram, patients must have not had a drink for at least 12 hours. The usual dose is 250 mg/day. Some patients may tell you they had a drink on this dose without feeling any part of the “Antabuse reaction." In these patients, you should titrate the dose up to 500 mg/day. Side effects are generally mild with headache, drowsiness, or fatigue most common. An important thing to tell patients is that they could have the “Antabuse reaction” with hidden forms of ethanol (eg, mouthwash, cold medicines, aftershave, or perfume, etc.) and that the reaction may still occur up to 14 days after stopping disulfiram.
Non-FDA Approved Options:
Topiramate (Topamax), an anticonvulsant, is commonly used off-label for alcoholism, and with good reason. All four placebo-controlled studies reported a significant decrease in alcohol use with topiramate compared to placebo. In one 14-week study, for example, the percent of heavy drinking days was 44% in topiramate patients compared to 52% in placebo patients (Johnson BA et al, JAMA 2007;298(14):1641–1651). The topiramate group also did better on abstinent days, number of drinks per day, and plasma gamma glutamyl transferase (GGT), a marker for alcohol intake. There have also been three studies comparing naltrexone to topiramate suggesting topiramate is as effective as naltrexone—unfortunately, the studies were small, only 12 weeks long and two were not controlled.
The dose of topiramate used in most studies was about 200 mg/day (divided twice a day) and side effects were considerable, especially at higher doses. The most common were cognitive impairment (eg, word-finding difficulties), numbness or tingling, weight loss, headache, fatigue, dizziness, and depression. Titrating the dose slowly and using the lowest effective dose might help patients tolerate this drug better.
Baclofen, Gabapentin and SSRIs
Baclofen (Lioresal), a muscle relaxant, at 30 mg/day has shown mixed efficacy compared to placebo. In two studies with a total of 123 patients, it showed higher rates of abstinence but in a third study of 80 patients, it was no better than placebo on any measure (Muzyk AJ et al, CNS Drugs 2012;26(1):69–78). Some have suggested a higher dose of 60 mg/day may be more effective but this needs to be studied and such high doses might lead to baclofen abuse.
There have been a few studies suggesting decreased drinking with 900 mg–1800 mg/day of gabapentin (Neurontin), an anticonvulsant, but the studies weren’t rigorously designed and had small numbers of patients, so it’s hard to make any comparisons with other options (Furieri FA & Nakamura-Palacios EM, J Clin Psychiatry 2007;68(11):1691–1700).
A meta-analysis of seven studies using selective serotonin reuptake inhibitors (SSRIs) didn’t find them to be effective in alcohol dependence unless patients also had a co-occurring depression, in which case patients did drink less overall (Nunes EV & Levin FR, JAMA 2004;291(15):1887–1896).
Dr. Carlat’s Verdict: Naltrexone is likely the most effective medication for most patients, but there are several other options.