Abstract Background The implementation of screening programmes in Sweden during the mid-1990s increased the number of small node-negative breast cancers. In this era before staging by sentinel node biopsy, routine axillary dissection for staging of early breast cancer was questioned owing to the increased morbidity and lack of perceived benefit. The long-term risk of axillary recurrence when axillary staging is omitted remains unclear. Methods This prospective observational multicentre cohort study included Swedish women diagnosed with breast cancer between 1997 and 2002. The patients had clinically node-negative, pT1a–b, grade I–II tumours. No axillary staging or dissection was performed. The primary outcome was ipsilateral axillary recurrence and survival. Results A total of 1543 patients were included. Breast-conserving surgery (BCS) was performed in 94·0 per cent and the rest underwent mastectomy. After surgery, 58·1 per cent of the women received adjuvant radiotherapy, 11·9 per cent adjuvant endocrine therapy and 31·5 per cent did not receive any adjuvant treatment. After a median follow-up of 15·5 years, 6·4 per cent developed contralateral breast cancer and 16·5 per cent experienced a recurrence. The first recurrence was local in 116, regional in 47 and distant in 59 patients. The breast cancer-specific survival rate was 93·7 per cent after 15 years. There were no differences in overall or breast cancer-specific survival between patients who received adjuvant radiotherapy and those who did not. Only 3·0 per cent of patients had an axillary recurrence, which was isolated in only 1·0 per cent. Conclusion Axillary surgery can safely be omitted in patients with low-grade, T1a–b, cN0 breast cancers. This large prospective cohort with 15-year follow-up had a very low incidence of axillary recurrences and high breast cancer-specific survival rate.