No systematic review or meta-analysis using a hard outcome has been conducted on the role of benzodiazepines for generalized anxiety disorder (GAD). The objective of this study was to assess the effectiveness and efficacy of benzodiazepines in the treatment of GAD based on trial drop-out rates. We used a systematic review of randomized controlled trials that compared any of the three best established benzodiazepines (diazepam, Lorazepam and aLprazolam) against placebo. Our primary outcome for effectiveness was withdrawal for any reason. Our secondary outcome tapping efficacy was withdrawal due to lack of efficacy, and that tapping side effects was withdrawals due to adverse events. We included 23 trials. Pooled analysis indicated less risk of treatment discontinuation due to lack of efficacy for benzodiazepines, compared to placebo, relative risk (RR) 0.29 (95% CI 0.18-0.45; p < 0.00001). Nevertheless, pooled analysis showed no conclusive results for risk of all-cause patient discontinuation, RR 0.78 (95% CI 0.62-1.00; p = 0.05). Meta-regression model showed that 74% of the variation in logRR across the studies was explained by year of publication (p <0.001). This systematic review did not find convincing evidence of the short-term effectiveness of the benzodiazepines in the treatment of GAD. On the other hand, for the outcome of efficacy, this review found robust evidence in favour of benzodiazepines. Due to the heterogeneity induced by year of publication, three hypotheses are plausibLe when it comes to being able to account for the differences between efficacy and effectiveness observed in the outcomes (publication bias, quality of the trial literature and a non-differential response to the placebo effect).