Sixty four studies, 4309 participants, met the inclusion criteria.
- Comparing benzodiazepines versus placebo, benzodiazepines performed better for seizures, 3 studies, 324 participants, RR 0.16 (0.04 to 0.69), no statistically significant difference for the other outcomes considered.
- Comparing benzodiazepines versus other drugs, there is a trend in favour of benzodiazepines for seizure and delirium control, severe life threatening side effect, dropouts, dropouts due to side effects and patient's global assessment score. A trend in favour of control group was observed for CIWA-Ar scores at 48 hours and at the end of treatment. The results reach statistical significance only in one study, with 61 participants, results on Hamilton anxiety rating scale favour control MD -1.60 (-2.59 to -0.61)
- Comparing different benzodiazepines among themselves,results never reached statistical significance but chlordiazepoxide performed better
- Comparing benzodiazepine plus other drug versus other drug, results never reached statistical significance.
- In the comparison of fixed-schedule versus symptom-triggered regimens, results from a single study, with 159 participants, favour symptom-triggered regimens MD -1.10 [-3.27, 1.07] for CIWA-Ar scores at the end of treatment. Differences in isolated trials should be interpreted very cautiously.
