Twenty-nine RCTs, providing 34 data sets, were included in the review (n=2,398 in total): 3 minimal intervention (n=526), 1 systematic discontinuation alone (n=84), 5 systematic discontinuation with psychotherapy (n=342) and 25 systematic discontinuation with medication (n=1,261).
The quality of the included studies was deemed to be moderate to excellent: the total scores ranged from 8 to 17.
Benzodiazepine discontinuation strategies.
A significantly higher discontinuation success rate was found in those receiving minimal intervention compared with usual care (OR 2.8, 95% CI: 1.6, 5.1), based on 3 studies (n=526). No statistical heterogeneity was found.
One study, which compared systematic discontinuation alone with usual care (n=84), found a significantly higher discontinuation success rate in those receiving systematic discontinuation (OR 6.1, 95% CI: 2.0, 18.6).
Benzodiazepine discontinuation augmentation strategies.
A higher success rate was found for systematic discontinuation with psychotherapy compared with systematic discontinuation alone (OR 1.8, 95% CI: 1.1, 2.9), based on 5 studies (n=342). Statistically significant between-study heterogeneity was present.
Systematic discontinuation with medication (21 studies, 25 data sets, n=1,261).
Significantly higher discontinuation success rates were found in patients receiving systematic discontinuation plus imipramine (OR 3.1, 95% CI: 1.1, 9.4, p=0.03), while higher success rates of borderline significance were found in patients receiving systematic discontinuation plus carbamazopine (OR 3.5, 95% CI: 0.9, 16.7, p=0.06), compared with those receiving systematic discontinuation alone; based on 3 studies (n=94) and 2 studies (n=75), respectively. No statistical heterogeneity was found. No significant difference in the discontinuation rates was found between groups for systematic discontinuation augmented with trazodone, propranolol or buspirone compared with systematic discontinuation alone. Statistical heterogeneity was found in the studies evaluating augmentation with propranolol and buspirone. The authors reported that this heterogeneity was explained by differences in the tapering procedure (propranolol only), transfer to long-acting agent, baseline benzodiazepine dosage, type of benzodiazepine, and the diagnosis of included patients (propranolol and buspirone).
Eleven studies evaluated systematic discontinuation with a different single pharmacological agent (aldpidem, progesterone, dothiepin, hydroxyzine, melatonin, Asp Mg, valproate, homeogene 46, sedative PC, paroxetine and flumazenil). Melatonin, valproate and flumazenil were all found to demonstrate higher benzodiazepine discontinuation success rates in comparison with the control groups.