Three perioperative factors impair host defence against recurrence during cancer surgery: the surgical stress response, use of volatile anaesthetic, and opioids for analgesia. All factors are ameliorated by regional anaesthesia-analgesia. We tested the primary hypothesis that breast cancer recurrence after potentially curative surgery is lower with regional anaesthesia-analgesia using paravertebral blocks and the anaesthetic propofol than with general anaesthesia with the volatile anaesthetic sevoflurane and opioid analgesia. A second hypothesis was that regional anaesthesia-analgesia reduces persistent incisional pain. We did a randomised controlled trial at 13 hospitals in Argentina, Austria, China, Germany, Ireland, New Zealand, Singapore, and the USA. Women (age <85 years) having potentially curative primary breast cancer resections were randomised by computer to either regional anaesthesia-analgesia (paravertebral blocks and propofol) or general anaesthesia (sevoflurane) and opioid analgesia. The primary outcome was local or metastatic breast cancer recurrence. The secondary outcome was incisional pain at 6 months and 12 months. Primary analyses were done under intention-to-treat principles. This trial is registered with ClinicalTrials.gov, NCT00418457. The study was stopped after a preplanned futility boundary was crossed. Between Jan 30, 2007, and Jan 18, 2018, 2132 women were enrolled to the study, of whom 24 were excluded before surgery. 1043 were assigned to regional anaesthesia-analgesia and 1065 were allocated to general anaesthesia. Baseline characteristics were well balanced between study groups. Median follow-up was 36 (IQR 24-49) months. Among women assigned regional anaesthesia-analgesia, 102 (10%) recurrences were reported, compared with 111 (10%) recurrences among those allocated general anaesthesia (hazard ratio 0·97, 95% CI 0·74-1·28; p=0·84). Incisional pain was reported by 442 (52%) of 856 patients assigned to regional anaesthesia-analgesia and 456 (52%) of 872 patients allocated to general anaesthesia at 6 months, and by 239 (28%) of 854 patients and 232 (27%) of 852 patients, respectively, at 12 months (overall interim-adjusted odds ratio 1·00, 95% CI 0·85-1·17; p=0·99). Neuropathic breast pain did not differ by anaesthetic technique and was reported by 87 (10%) of 859 patients assigned to regional anaesthesia-analgesia and 89 (10%) of 870 patients allocated to general anaesthesia at 6 months, and by 57 (7%) of 857 patients and 57 (7%) of 854 patients, respectively, at 12 months. In our study population, regional anaesthesia-analgesia (paravertebral block and propofol) did not reduce breast cancer recurrence after potentially curative surgery compared with volatile anaesthesia (sevoflurane) and opioids. The frequency and severity of persistent incisional breast pain was unaffected by anaesthetic technique. Clinicians can use regional or general anaesthesia with respect to breast cancer recurrence and persistent incisional pain. Sisk Healthcare Foundation (Ireland), Eccles Breast Cancer Research Fund, British Journal of Anaesthesia International, College of Anaesthetists of Ireland, Peking Union Medical College Hospital, Science Fund for Junior Faculty 2016, Central Bank of Austria, and National Healthcare Group.