Table. Summary of Findings and Strength of Evidence From Trials Assessing Efficacy of Medications With at Least Low Strength of Evidence for Benefit for Alcohol Use Disordera.
| Acamprosate | Baclofen | Disulfiram | Gabapentin | Naltrexone | Topiramate | ||||
|---|---|---|---|---|---|---|---|---|---|
| 50 mg/d, oral | 100 mg/d, oral | Injection | Any dose | ||||||
| Return to any drinking | |||||||||
| No. of studies | 20 | 8 | 3 | 3 | 16 | 3 | 2 | 25 | 1 |
| No. of participants | 6380 | 995 | 622 | 522 | 2347 | 946 | 939 | 4604 | 106 |
| Results effect size (95% CI) | RR, 0.88 (0.83-0.93) | RR, 0.83 (0.70-0.98) | RR, 1.03 (0.90-1.17) | RR, 0.92 (0.83-1.02) | RR, 0.93 (0.87-0.99) | RR, 0.97 (0.91-1.03) | RR, 0.96 (0.90-1.03) | RR, 0.95 (0.92-0.99) | Topiramate, 53.8%; placebo, 72.2% |
| Number needed to treat (95% CI)c | 11 (1-32) | 18 (4-32) | |||||||
| Strength of evidence | Moderate | Low | Low (no effect) | Low | Moderate | Low (no effect) | Low (no effect) | Moderate | Insufficient |
| Return to heavy drinking | |||||||||
| No. of studies | 7 | 4 | 0 | 3 | 23 | 2 | 2 | 27 | 1 |
| No. of participants | 2496 | 483 | 0 | 522 | 3170 | 858 | 615 | 4645 | 170 |
| Results effect size (95% CI) | RR, 0.99 (0.94-1.05) | RR, 0.92 (0.80-1.06) | RR, 0.90 (0.82-0.98) | RR, 0.81 (0.72-0.90) | RR, 0.93 (0.84-1.01) | RR, 1.00 (0.82-1.21) | RR, 0.86 (0.80-0.93) | Topiramate, 10%; placebo, 14% | |
| Number needed to treat (95% CI)c | 11 (5-41) | ||||||||
| Strength of evidence | Moderate (no effect) | Low (no effect) | Insufficient | Low | Moderate | Low (no effect) | Low (no effect) | Moderate | Insufficient |
| Percentage of drinking days | |||||||||
| No. of studies | 14 | 5 | 2 | 1 | 15 | 3 | 2 | 24d | 8 |
| No. of participants | 4916 | 714 | 290 | 112 | 1992 | 1023 | 467 | 4021 | 1080 |
| Results effect size (95% CI)b | WMD, −8.3 (−12.2 to −4.4) | WMD, −5.55 (−18.79 to 7.69) | No significant difference | No significant difference | WMD, −5.1 (−7.16 to −3.04) | WMD, −2.3 (−5.60 to 0.99) | WMD, −4.99 (−9.49 to 0.49) | WMD, −4.51 (−6.26 to −2.77) | WMD, −7.2 (−14.3 to −0.1) |
| Strength of evidence | Moderate | Low (no effect) | Insufficient | Insufficient | Moderate | Low | Low | Moderate | Moderate |
| Percentage of heavy drinking days | |||||||||
| No. of studies | 2 | 9 | 0 | 3 | 7 | 2 | 3 | 13 | 9 |
| No. of participants | 123 | 1112 | 0 | 600 | 624 | 423 | 956 | 2167 | 1210 |
| Results effect size (95% CI)b | WMD, −3.4 (−6.45 to 5.86) | WMD, −2.16 (−7.34 to 3.02) | No significant difference | WMD, −4.3 (−7.60 to −0.91) | WMD, −3.1 (−5.8 to −0.3) | WMD, −4.68 (−8.63 to −0.73) | WMD, −3.92 (−5.86 to −1.97) | WMD, −6.2 (−10.9 to −1.4) | |
| Strength of evidence | Insufficient | Low (no effect) | Insufficient | Low (no effect) | Moderate | Low | Low | Moderate | Moderate |
| Drinks per drinking day | |||||||||
| No. of studies | 2 | 2 | 0 | 2 | 9 | 1 | 0 | 16 | 7 |
| No. of participants | 139 | 146 | 0 | 428 | 1018 | 240 | 0 | 2011 | 922 |
| Results effect size (95% CI)b | WMD, 0.6 (−1.43 to 2.64) | WMD, 0.85 (−2.23 to 3.93) | No significant difference | WMD, −0.49 (−0.92 to −0.06) | WMD, 1.9 (−1.5 to 5.2) | WMD, −0.85 (−1.44 to −0.26) | WMD, −2.0 (−3.1 to −1.0) | ||
| Strength of evidence | Insufficient | Low (no effect) | Insufficient | Low (no effect) | Low | Insufficient | Insufficient | Low | Moderate |
| Motor vehicle crashes or injuries | |||||||||
| No. of studies | 0e | 0 | 0 | 0 | 0 | 2 | |||
| No. of participants | 0 | 0 | 0 | 0 | 0 | 541 | |||
| Results effect size (95% CI)b | Reduced risk | ||||||||
| Strength of evidence | Insufficient | Insufficient | Insufficient | Insufficient | Insufficient | Low | |||
| Quality of life or function | |||||||||
| No. of studies | 1 | 2 | 0 | 0 | 5 | 2 | |||
| No. of participants | 612f | 384g | 0 | 0 | 1844h | 118i | |||
| Results effect size (95% CI)b | No significant differencej | No significant difference | Some conflicting resultsk | No significant difference | |||||
| Strength of evidence | Insufficient | Low (no effect) | Insufficient | Insufficient | Insufficient | Low (no effect) | |||
| Mortality | |||||||||
| No. of studies | 8 | 4 | 0 | 0 | 6 | 3 | |||
| No. of participants | 2677 | 660 | 0 | 0 | 1738 | 507 | |||
| Results effect size (95% CI)b | 7 events (acamprosate) vs 6 events (placebo) | 8 baclofen vs 3 placebo | 1 event (naltrexone) vs 2 events (placebo) | Not reported | |||||
| Strength of evidence | Insufficient | Insufficient | Insufficient | Insufficient | Insufficient | Insufficient | |||
Abbreviations: RR, risk ratio; WMD, weighted mean difference.
Blank cells indicate data not applicable. Strength of evidence was not rated for naltrexone by dose. Heavy drinking days was defined as ≥4 drinks/d for women and ≥5 drinks/d for men.
Negative effect sizes favor intervention over placebo/control.
Lack of entry for number needed to treat indicates that the relative risk (95% CI) was not statistically significant, so the investigators did not calculate a number needed to treat or the effect measure was not one that allows direct calculation of number needed to treat (eg, WMD).
One study contained 2 treatment groups included in the meta-analysis.79
Results were not reported for each treatment group separately, but there were no clinically significant differences across treatment groups.
Quality of life and functioning were assessed with the World Health Organization Quality of Life (WHOQOL) and 12-item Short-Form Health Survey (SF-12) version 2 physical and mental health scores.
Quality of life and functioning were assessed with the Quality of Life Enjoyment and Satisfaction Questionnaire and the 36-item Short Form Health Survey (SF-36).
Each trial used a different measure to assess quality of life and functioning, including the Short Inventory of Problems, SF-36, WHOQOL, SF-12 version 2 physical and mental health scores, Drinker Inventory of Consequences, and SF-12.
Quality of life was assessed with the SF-36.
Results were not reported for each treatment group separately, but there were no clinically significant differences across treatment groups.
One study rated as having unclear risk of bias reported that 1 patient in the placebo group died by “accident.” No other details on the cause or nature of the accident were provided.44 That study also reported 1 injury in the acamprosate group and 2 in the placebo group. Another study, rated as having high risk of bias, reported a “traffic accident” in the acamprosate group.80