Background Little is known about the longer-term effects of adjunctive benzodiazepines on symptom response during treatment in patients with bipolar disorders. Methods The study sample consisted of 482 patients with bipolar I or II disorder enrolled in a 6-month, randomized, multi-site comparison of lithium- and quetiapine-based treatment. Changes in clinical measures (BISS total and subscales, CGI-BP, and CGI-Efficacy Index) were compared between participants who did and did not receive benzodiazepine treatment at baseline or during follow-up. Selected outcomes were also compared between patients who did and did not initiate benzodiazepines during follow-up using stabilized inverse probability weighted analyses. Results Significant improvement in all outcome measures occurred within each benzodiazepine exposure group. Benzodiazepine users (at baseline or during follow-up) experienced significantly less improvement in BISS total, BISS irritability, and CGI-BP scores than did benzodiazepine non-users. There were no significant differences in these measures between patients who did and did not initiate benzodiazepines during follow-up in the weighted analyses. There was no significant effect of benzodiazepine use on any outcome measure in patients with comorbid anxiety or substance use disorders. Limitations This is a secondary analysis of data from a randomized effectiveness trial that was not designed to address differential treatment response according to benzodiazepine use. Conclusions Adjunctive benzodiazepines may not significantly affect clinical outcome in lithium- or quetiapine-treated patients with bipolar I or II disorder over 6 months, after controlling for potential confounding factors. © 2014 Elsevier B.V.