Background: 5-year results of the UK Standardisation of Breast Radiotherapy (START) trials suggested that lower total doses of radiotherapy delivered in fewer, larger doses (fractions) are at least as safe and effective as the historical standard regimen (50 Gy in 25 fractions) for women after primary surgery for early breast cancer. In this prespecified analysis, we report the 10-year follow-up of the START trials testing 13 fraction and 15 fraction regimens. Methods: From 1999 to 2002, women with completely excised invasive breast cancer (pT1-3a, pN0-1, M0) were enrolled from 35 UK radiotherapy centres. Patients were randomly assigned to a treatment regimen after primary surgery followed by chemotherapy and endocrine treatment (where prescribed). Randomisation was computer-generated and stratified by centre, type of primary surgery (breast-conservation surgery or mastectomy), and tumour bed boost radiotherapy. In START-A, a regimen of 50 Gy in 25 fractions over 5 weeks was compared with 41·6 Gy or 39 Gy in 13 fractions over 5 weeks. In START-B, a regimen of 50 Gy in 25 fractions over 5 weeks was compared with 40 Gy in 15 fractions over 3 weeks. Eligibility criteria included age older than 18 years and no immediate surgical reconstruction. Primary endpoints were local-regional tumour relapse and late normal tissue effects. Analysis was by intention to treat. Follow-up data are still being collected. This study is registered as an International Standard Randomised Controlled Trial, number ISRCTN59368779. Findings: START-A enrolled 2236 women. Median follow-up was 9·3 years (IQR 8·0-10·0), after which 139 local-regional relapses had occurred. 10-year rates of local-regional relapse did not differ significantly between the 41·6 Gy and 50 Gy regimen groups (6·3%, 95% CI 4·7-8·5 vs 7·4%, 5·5-10·0; hazard ratio [HR] 0·91, 95% CI 0·59-1·38; p=0·65) or the 39 Gy (8·8%, 95% CI 6·7-11·4) and 50 Gy regimen groups (HR 1·18, 95% CI 0·79-1·76; p=0·41). In START-A, moderate or marked breast induration, telangiectasia, and breast oedema were significantly less common normal tissue effects in the 39 Gy group than in the 50 Gy group. Normal tissue effects did not differ significantly between 41·6 Gy and 50 Gy groups. START-B enrolled 2215 women. Median follow-up was 9·9 years (IQR 7·5-10·1), after which 95 local-regional relapses had occurred. The proportion of patients with local-regional relapse at 10 years did not differ significantly between the 40 Gy group (4·3%, 95% CI 3·2-5·9) and the 50 Gy group (5·5%, 95% CI 4·2-7·2; HR 0·77, 95% CI 0·51-1·16; p=0·21). In START-B, breast shrinkage, telangiectasia, and breast oedema were significantly less common normal tissue effects in the 40 Gy group than in the 50 Gy group. Interpretation: Long-term follow-up confirms that appropriately dosed hypofractionated radiotherapy is safe and effective for patients with early breast cancer. The results support the continued use of 40 Gy in 15 fractions, which has already been adopted by most UK centres as the standard of care for women requiring adjuvant radiotherapy for invasive early breast cancer. Funding: Cancer Research UK, UK Medical Research Council, UK Department of Health.