Background: Breast cancer central nervous system (CNS) metastases are an increasingly important problem because of high CNS relapse rates in patients treated with trastuzumab and/or taxanes. Patients and methods: We evaluated data from 2887 node-positive breast cancer patients randomised in the BIG 02-98 trial comparing anthracycline-based adjuvant chemotherapy (control arms) to anthracycline-docetaxel-based sequential or concurrent chemotherapy (experimental arms). After a median follow-up of 5 years, 403 patients had died and detailed information on CNS relapse was collected for these patients. Results: CNS relapse occurred in 4.0% of control patients and3.7% of docetaxel-treated patients. CNS relapse occurred in 27% of deceased patients in both treatment groups. CNS relapse was usually accompanied by neurologic symptoms (90%), and 25% of patients with CNS relapse died without evidence of extra-CNS relapse. Only 20% of patients survived 1 year from the diagnosis of CNS relapse. Prognosis of CNS relapse was worse for patients with meningeal carcinomatosis when compared with brain metastases. Unexpected findings included a higher rate of positive cerebrospinal fluid cytology (8% versus 3%) and more frequent use of magnetic resonance imaging for diagnosis (47% versus 30%) in the docetaxel-treated patients. Conclusion: There is no evidence that adjuvant docetaxel treatment is associated with an increased frequency of CNS relapse. © The Author 2008. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved.