We tested the hypothesis that stroke outcomes in patients with preadmission use of benzodiazepine are worse. In a prospective cohort study, we recruited patients with acute ischaemic stroke. Mortality, functional outcomes and cognition were evaluated at 8 and 90 days after stroke. 370 patients were included. 62 (18.5%) of the 336 remaining patients were treated with benzodiazepines when stroke occurred, and they did not receive any other psychotropic drug. The mortality rate was higher in benzodiazepines users than non-users at day 8 (2.2% vs 8.1%, p=0.034) and day 90 (8.1% vs 25.9%, p=0.0001). After controlling for baseline differences using propensity-score matching, only the difference in mortality rate at day 90 was of borderline of significance, with a matched OR of 3.93 (95% CI, 0.91 to 16.98). In propensity-score-adjusted cohort, this difference remained significant with a similar treatment effect size (adjusted OR, 3.50; 95% CI, 1.57 to 7.76). A higher rate of poor functional outcome at day 8 and day 90 defined bymodified Rankin scale (mRS) ≥2 or by theBarthel index (BI) <95 was found in benzodiazepines users. In propensity-score-adjusted cohort, only the difference in mRS≥2 at day 90 remained significant (adjusted OR, 1.89; 95% CI, 1.02 to 3.48). In survivors at day 8 and at day 90, there was no significant difference in cognitive evaluation. Our study has shown that preadmission use of benzodiazepines could be associated with increased post-stroke mortality at 90 days. These findings do not support a putative neuroprotective effect of γ-aminobutyric acidA NCT00763217.