|Acamprosate for alcohol dependence: a sex-specific meta-analysis based on individual patient data
|Mason BJ, Lehert P
This high quality individual patient data review of 22 trials concluded that acamprosate had a significant effect, compared with placebo, in improving the rates of abstinence for women and men with alcohol dependence. There were no gender differences in tolerability or efficacy. The conclusions of the review are likely to be reliable.
To assess gender differences in the efficacy, safety and tolerability of acamprosate for those dependent on alcohol.
PubMed, PsycINFO and The Cochrane Library were searched, using specified terms, with no language restrictions (inception to February 2011). Bibliographies were searched and the proceedings of major scientific meetings were checked.
Prospective, placebo-controlled, double-blind parallel-group trials were eligible for inclusion. They had to have random assignment of patients with alcohol dependence to acamprosate or placebo and to report at least one quantitative measure of drinking behaviour, defined as the percentage of abstinent days, the proportion of patients maintaining abstinence, or the percentage of days without heavy drinking.
The included trials were reasonably similar for admission criteria, outcome variables and provision of concomitant alcohol counselling. They typically included adults meeting the American Psychiatric Association 1987 or 1994 criteria for alcohol dependence. They excluded participants with psychiatric disorders or using medication that might confound the results, such as naltrexone. Participants had a mean age of 43 years and more than 40% had severe alcohol dependence.
The number of reviewers assessing trial eligibility was not reported.
Assessment of study quality
Methodological quality was assessed using a Chalmers score for the sample size, randomisation methods, methods of blinding, selection and withdrawal criteria, and statistical analysis. Only trials scoring more than 50 out of 100 points on the Chalmers scale were included. Individual patient data were used to compare the baseline characteristics of men and women, but any differences in baseline characteristics between trial arms were not reported.
Two reviewers independently assessed trial quality.
Trial authors and the manufacturer of acamprosate were contacted for individual patient data. These were collated to allow regression analyses of efficacy outcomes and analysis of variance of tolerability outcomes. Tolerability was assessed using the rates of treatment completion, discontinuation due to adverse events, the occurrence of one or more adverse events of moderate or severe intensity, digestive system adverse events, and medication compliance.
Individual patient data were collected for outcomes based on the Timeline Followback Interview or similar trial-specific interviews that recorded the frequency of drinking and were converted to standard definitions. Missing tolerability data were imputed, assuming random omission and using Anderson's method (1957, see Other Publications of Related Interest). Loss to follow-up was conservatively attributed to drinking.
Methods of synthesis
The individual patient data were synthesised, using a one-stage regression, with a random treatment effect and a fixed treatment covariate interaction, assuming a common baseline risk across trials. Within- and across-trial interaction coefficients were not distinguished. Tolerability outcomes were analysed, using analysis of variance, with treatment and gender as the main interaction effects. Similar two-stage approaches, pooling using DerSimonian and Laird, were used with individual patient data to generate summary data.
Forest plots, summary effects and statistical tests of heterogeneity (Cochran Q) were presented.
Results of the review
Individual patient data were available from 22 out of 23 eligible trials, with 1,317 women and 4,794 men. Fourteen trials lasted six months or more, and five of these lasted for one year.
There were statistically significant beneficial effects of acamprosate relative to placebo across all measured endpoints. The incremental gain in percentage of abstinent days was 10% (95% CI 7 to 14). The incremental gain in percentage of days with no heavy drinking was 11% (95% CI 8 to 15). The odds of complete abstinence were 1.9 (95% CI 1.6 to 2.2). The odds of no heavy drinking were 1.9 (95% CI 1.6 to 2.3).
There was no significant difference between men and women for any efficacy or tolerability outcome. Heterogeneity was not statistically significant across trials for women, but it was for men.
Acamprosate had a significant effect, compared with placebo, in improving the rates of abstinence for both women and men with alcohol dependence. There were no gender differences in the efficacy or tolerability of acamprosate.
The review addressed a clearly defined question, using appropriate inclusion criteria to ensure replicability. The individual patient data were acquired, using an appropriate search strategy, but it was unclear if trial authors were asked to identify further trials and unpublished information. The authors retrieved data from 22 out of 23 trials, and availability bias is unlikely. Trial quality was assessed at a total level; sensitivity to trial quality was not reported. The authors used the individual patient data to compare the baseline characteristics of women and men, but they did not report any baseline differences between trial arms.
The interaction terms were always small and non-significant indicating that there was no evidence of a different treatment effect by gender. Heterogeneity within the female subgroup was not statistically significant, but it was not explored, in either subgroup, using additional individual patient data covariates. Missing data were imputed, but the extent of missing data and the reasons for them were not fully described, so the potential for bias is unclear. Tolerability data were synthesised using analysis of variance rather than meta-analysis, but the magnitude of differences due to gender was always small and not statistically significant.
The conclusion that there was no difference between genders in treatment efficacy and in tolerability is likely to be reliable.
Implications of the review for practice and research
Practice: The authors stated that treatment providers could routinely consider acamprosate for women and men with alcohol dependence, taking into account treatment goals and preferences.
Research: The authors stated that further investigation of the efficacy of acamprosate for women with alcohol dependence and comorbid psychiatric disorders might be needed.
Supported by grants from the National Institute on Alcohol Abuse and Alcoholism, USA.
Mason BJ, Lehert P. Acamprosate for alcohol dependence: a sex-specific meta-analysis based on individual patient data. Alcoholism: Clinical and Experimental Research 2012; 36(3): 497-508
Other publications of related interest
Anderson TW. Maximum likelihood estimates for a multivariate normal distribution when some observations are missing. Journal of the American Statistical Association 1957; 52: 200-203.
Subject indexing assigned by NLM
Adult; Alcohol Deterrents /adverse effects /therapeutic use; Alcoholism /drug therapy; Female; Humans; Male; Randomized Controlled Trials as Topic /psychology /statistics & numerical data; Sex Characteristics; Taurine /adverse effects /analogs & derivatives /therapeutic use
Database entry date
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.