BACKGROUND: Several randomised trials have confirmed the benefit of adjuvant trastuzumab for patients with HER2-positive early breast cancer. However, concern has been expressed that adjuvant trastuzumab might be associated with an increased frequency of CNS relapses. We assessed the frequency and course of CNS relapses, either as first event or at any time, using data from the HERA trial. METHODS: We estimated the cumulative incidence of first disease-free survival (DFS) events in the CNS versus other sites by competing risks analysis in patients with HER2-positive early breast cancer who had been randomly assigned to receive 1 year of trastuzumab or to observation in the HERA trial after a median follow-up of 4 years (IQR 3·5-4·8). To obtain further information about CNS relapse at any time before death, we circulated a data collection form to investigators to obtain standardised information about CNS events that occurred in all patients who had died before July, 2009. We estimated the cumulative incidence of CNS relapse at any time with a competing risks analysis. RESULTS: Of 3401 patients who had been assigned to receive 1 year of trastuzumab or to observation, 69 (2%) had a CNS relapse as first DFS event and 747 (22%) had a first DFS event not in the CNS. The frequency of CNS relapses as first DFS event did not differ between the group given 1 year of trastuzumab (37 [2%] of 1703 patients) and the observation group (32 [2%] of 1698; p=0·55 [Gray's test]). 481 data collection forms were distributed, of which 413 (86%) were returned. The proportion of patients who had died and experienced a CNS relapse was numerically higher in the observation group (129 [57%] of 227) than in the group given trastuzumab for 1 year (88 [47%] of 186; p=0·06 [Gray's test]). Most CNS relapses were symptomatic (189 [87%] of 217). CONCLUSION: Adjuvant trastuzumab does not increase the risk of CNS relapse in patients with HER2-positive early breast cancer. FUNDING: None.