Background: Benzodiazepines have been proposed both as a neuroprotectant and risk factor for pneumonia in acute stroke. Aims: We assessed the impact of benzodiazepine exposure on the modified Rankin scale score distribution at 90 days as well as pneumonia rates among patients registered in a trials archive. Method: We used an age, baseline National Institutes of Health Stroke Score, and thrombolysis-rate adjusted Cochran-Mantel-Haenszel test to test significance (P) followed by proportional odds logistic regression analysis to estimate the odds ratios for improved modified Rankin scale score, and binary logistic regression to estimate the odds ratio for developing pneumonia. Results: Data were available for 5938 patients, of whom 1800 received benzodiazepines. No association of benzodiazepine use and overall stroke outcome could be found (odds ratio 0·90, 95% confidence interval 0·82-1·00, P=0·121). Pneumonia occurred in 12·8% of patients treated with benzodiazepines and in 13·6% of the controls (odds ratio 0·99, 95% confidence interval 0·83-1·18, P=0·904). Conclusion: In this nonrandomized comparison, treatment with benzodiazepines as a concomitant medication had no independent impact on stroke outcome.