Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Alcohol dependence treatments

Systematic review
Results
Acamprosate
Naltrexone
Comment

Alcohol dependence is a bad thing. It has high impacts on individuals, their health, their family, and the economy. There are lots of suggestions for dealing with alcohol dependence, including drug treatments. A new systematic review [1] gives us a good idea of the efficacy of two of those treatments, acamprosate and naltrexone.

Systematic review

The review sought randomised studies of full, published papers comparing acamprosate or naltrexone with placebo or other control without medication. Studies had to be longer than two weeks. The period covered in searching four electronic databases was 1990-end 2002.

Results

Acamprosate

For acamprosate there were 13 placebo-controlled trials with about 4,000 subjects. Twelve were conducted in an ambulatory setting and were included in the review. All used standard definitions of alcohol dependence and all subjects had undergone a previous detoxification process. Most were double blind, and all had adequate quality scores. All but one reported an intention-to-treat result. Duration was three to 24 months, with most being six to 12 months. Most were definitely funded by the manufacturer of acamprosate; for four no information on funding source was available. The dose of acamprosate was most often set by body weight, being about 2,000 mg daily in those over 60 kg, and 1,300 mg a day in those below 60 kg.

Abstinence rates (percentage of patients completing the study without ingesting alcohol) varied (Figure 1), but with acamprosate were consistently higher than placebo. The overall abstinence rate was 23% with acamprosate and 15% with placebo, giving a number needed to treat to generate one more abstinent patient of 12 (95% CI 9 to 17; Table 1).



Figure 1: Abstinence rates with acamprosate and placebo







Table 1: Outcomes for benefit and adverse events with acamprosate and naltrexone compared with placebo in randomised trials



Number of
Event rate (%) with
Outcome
Trials
Patients
Treatment
Placebo
Relative benefit/risk
(95% CI)
NNT/NNH
(95% CI)
Acamprosate
Abstinence
11
3322
23
15
1.6 (1.4 to 1.9)
12 (9 to 17)
Compliance
12
3959
53
47
1.1 (1.05 to 1.2)
18 (12 to 43)
Gastrointestinal adverse events
10
3425
17
11
1.5 (1.3 to 1.8)
17 (12 to 27)
Adverse event discontinuation
9
2697
2.8
2.2
1.3 (0.8 to 2.1)
Naltrexone
Abstinence
10
1077
35
30
1.2 (0.98 to 1.4)
Relapse
14
2071
37
48
0.78 (0.71 to 0.86)
10 (7 to 16)
Gastrointestinal adverse events
17
2564
24
16
1.5 (1.3 to 1.8)
13 (9 to 21)
Neuropsychiatric
17
2564
43
37
1.14 (1.03 to 1.26)
19 (11 to 73)
Adverse event discontinuation
11
1892
10
3.6
2.9 (1.8 to 4.2)
15 (11 to 22)

NNH is bold



Acamprosate also had a small but significant increase in treatment compliance (Table 1). Gastrointestinal adverse events were higher with acamprosate (17% versus 11% with placebo, producing an NNH of 17; 12 to 27). Adverse event withdrawals were low, and not significantly higher with acamprosate than with placebo.

Naltrexone

For naltrexone there were 19 trials, of which 12 were randomised and double blind, and from which information was taken; all had adequate quality scores. Information was available from over 2,000 subjects. One other (large) study was double-blind, but did not specify that it was randomised.

Studies used standard definitions of alcohol dependence, with subjects having undergone a previous detoxification. All were in an ambulatory setting. Duration was three to 18 months, with most studies of three or six months. Most were funded from charitable or government sources, with six funded in whole or part by a manufacturer. The dose of naltrexone was generally 50 mg daily.

There was no consistent reduction in abstinence rate with naltrexone. The abstinence rate with naltrexone of 35% was not significantly higher than the 30% with placebo (Table 1). There was a significant reduction in relapse rate to 37% with naltrexone compared with 48% with placebo (Figure 2), with an NNT of 10 (7 to 16).

Naltrexone produced a higher rate of gastrointestinal adverse events than placebo, with an NNH of 13 (9 to 21; Table 1). It also produced a higher rate of neuropsychiatric adverse events, with an NNH of 19 (11 to 73), but not serious neuropsychiatric adverse events. Adverse event discontinuation was significantly higher with naltrexone, with an NNH of 15 (11 to 22).



Figure 2: Relapse rate with naltrexone and placebo





Comment

Here we have two therapies, assessed in good trials which were of adequate quality, were valid, and with a reasonable number of patients in the combined trials. The message is reasonably clear, with acamprosate clearly better than placebo, with few adverse events, and naltrexone of questionable benefit and with more adverse events.

In randomised trials of alcohol treatment, some untreated participants (waiting list, no treatment, placebo) will give up or reduce their drinking by themselves. A review of untreated individuals [2] showed an average abstinence rate of 21% in 17 studies, with a mean level of consumption of 31 drinks a week, about six per week lower than the baseline of 37 drinks per week in 19 studies. This gives an indication of what can be achieved by doing nothing, and provides a benchmark for active treatments.

Which is the best choice? There may be differences in opinion about outcomes and their utility, but it would be difficult to argue that naltrexone was better than acamprosate.

In the past some would quibble. They might argue that these are indirect comparisons, and we could trust only a direct head-to-head comparison. There is one such. It had 157 patients, and concluded that naltrexone was better than acamprosate, though not by much. Bandolier 110 informed us that indirect comparison fulfilling quality, validity, and size criteria were at least as good, if not better, than direct comparisons lacking size.

The main question, though, may be what outcome you want. If it is abstinence, acamprosate clearly wins. If it is controlling consumption, rather than achieving abstinence, then naltrexone may have a role.

References:

  1. B Carmen et al. Efficacy and safety of naltrexone and acamprosate in the treatment of alcohol dependence: a systematic review. Addiction 2004 99: 811-828.
  2. A Moyer, JW Finney. Outcomes for untreated individuals involved in randomized trials of alcohol treatment. Journal of Substance Abuse Treatment 2002 23: 247-252.

previous or next story