| Medication | Outcome | N Studies; N Subjectsa | Results Effect Size (95% CI)b | NNTc | Strength of Evidence |
|---|---|---|---|---|---|
| ACA | Return to any drinking | 16; 4,847 | RD: -0.09 (-0.14 to -0.04) | 12 | Moderate |
| Return to heavy drinking | 7; 2,496 | RD: -0.01 (-0.04 to 0.03) | NA | Moderate | |
| % DDs | 13; 4,485 | WMD: -8.8 (-12.8 to -4.8) | NA | Moderate | |
| % HDDs | 1; 100 | WMD: -2.6 (-11.4 to 6.2) | NA | Insufficient | |
| Drinks per DD | 1; 116 | WMD: 0.4 (-1.8 to 2.6) | NA | Insufficient | |
| Accidents or injuries | 0;d 0 | NA | NA | Insufficient | |
| QoL or function | 1; 612 | NSD | NA | Insufficient | |
| Mortality | 8; 2,677 | 7 events (ACA) vs. 6 events (placebo) | NA | Insufficient | |
| DIS | Return to any drinking | 2; 492 | RD: -0.04 (-0.11 to 0.03)e | NA | Low |
| Return to heavy drinking | 0; 0 | NA | NA | Insufficient | |
| % DDs | 2; 290 | NSDf | NA | Insufficient | |
| % HDDs | 0; 0 | NA | NA | Insufficient | |
| Drinks per DD | 0; 0 | NA | NA | Insufficient | |
| Accidents or injuries | 0; 0 | NA | NA | Insufficient | |
| QoL or function | 0; 0 | NA | NA | Insufficient | |
| Mortality | 0; 0 | NA | NA | Insufficient | |
| NTX 50 mg oral | Return to any drinking | 16; 2,347 | RD: -0.05 (-0.10 to -0.00) | 20 | Moderate |
| Return to heavy drinking | 19; 2,875 | RD: -0.09 (-0.13 to -0.04) | 12 | Moderate | |
| % DDs | 15; 1,992 | WMD: -5.4 (-7.5 to -3.2) | NA | Moderate | |
| % HDDs | 6; 521 | WMD: -4.1 (-7.6 to -0.61) | NA | Moderate | |
| Drinks per DD | 9; 1,018 | WMD: -0.49 (-0.92 to -0.06) | NA | Low | |
| NTX 100 mg oral | Return to any drinking | 3; 946 | RD: -0.03 (-0.08 to 0.02) | NA | Low |
| Return to heavy drinking | 2; 858 | RD: -0.05 (-0.11 to 0.01) | NA | Low | |
| % DDs | 2; 858 | WMD: -0.9 (-4.2 to 2.5) | NA | Low | |
| % HDDs | 2; 423 | WMD: -3.1 (-5.8 to -0.3) | NA | Low | |
| Drinks per DD | 1; 240 | WMD: 1.9 (-1.5 to 5.2) | NA | Insufficient | |
| NTX injection | Return to any drinking | 2; 939 | RD: -0.04 (-0.10 to 0.03) | NA | Low |
| Return to heavy drinking | 2; 615 | RD: -0.01 (-0.14 to 0.13) | NA | Low | |
| % DDs | 1; 315 | WMD: -8.6 (-16.0 to -1.2) | NA | Insufficient | |
| % HDDs | 2g; 926 | WMD: -4.6 (-8.5 to -0.56) | NA | Low | |
| Drinks per DD | 0; 0 | NA | NA | Insufficient | |
| NTX (any dose) | Accidents or injuries | 0; 0 | NA | NA | Insufficient |
| QoL or function | 4; 1,513 | Some conflicting resultsh | NA | Insufficient | |
| Mortality | 6; 1,738 | 1 event (NTX) vs. 2 events (placebo) | NA | Insufficient |
ACA = acamprosate; CI = confidence interval; DD, drinking day; DIS = disulfiram; FDA = U.S. Food and Drug Administration; HDD, heavy drinking day; N = number; NA = not applicable; NNT = number needed to treat; NSD = no statistically significant difference; NTX = naltrexone; QoL, quality of life; RD = risk difference; vs. = versus; WMD = weighted mean difference.
Includes only studies rated as low or medium risk of bias included in the main analyses; these numbers do not include studies rated as high or unclear risk of bias that were included in sensitivity analyses.
Negative effect sizes favor intervention over placebo/control.
NA entry for numbers needed to treat (NNT) indicates that the risk difference (95% CI) was not statistically significant, so we did not calculate a NNT, or that the effect measure was not one that allows direct calculation of NNT (e.g., WMD).
One study rated as unclear risk of bias reported that one patient in the placebo group died by “accident.” No other details on the cause or nature of the accident were provided.78 That study also reported 1 injury in the acamprosate group and 2 in the placebo group. Another study, rated high risk of bias, reported a traffic accident in the acamprosate group.184
From meta-analysis of disulfiram 250 mg versus control (disulfiram 1 mg).92,93 Meta-analysis including studies rated as high risk of bias also found no significant difference (RD, -0.00; 95% CI, -0.10 to 0.09). Similarly, our meta-analysis found no statistically significant difference between disulfiram 250 mg per day and riboflavin (i.e., no disulfiram) (RD, -0.04; 95% CI, -0.11 to 0.03).
One study (N=128) reported similar percentages and no significant difference;93 the other reported that disulfiram was favored among the subset of subjects (N=162 of 605 subjects) who drank and had a complete set of assessment interviews, but it did not report this outcome for the full randomized sample.92 Overall, evidence was insufficient due to imprecision, inconsistency, and indirectness.
Contains data from personal communication (B. Silverman, November 14, 2013).
Unable to pool data. Two studies found no significant difference between naltrexone- and placebo-treated subjects.102,193 One study reported that patients receiving injectable naltrexone 380 mg per month had greater improvement on the mental health summary score than those receiving placebo at 24 weeks (8.2 versus 6.2, p=0.044).131 One study measured alcohol-related consequences (with the DrInC) and reported that more subjects who received placebo (N=34) had at least 1 alcohol-related consequence than those who received naltrexone (N=34): 76 percent versus 45 percent, p=0.02.42
From: Discussion
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.