Background: Population-based studies of treatment of locally recurrent rectal cancer (LRRC) are lacking. The aim was to investigate the surgical treatment of patients with LRRC at a national population-based level.Methods: All patients undergoing abdominal resection for primary rectal cancer between 1995 and 2002 in Sweden with LRRC as a first event were included. Detailed information about treatment, complications and outcomes was collected from the medical records. The patients were analysed in three groups: patients who had resection of the LRRC, those treated without tumour resection and patients who received best supportive care only.Results: In all, 426 patients were included in the study. Of these, 149 (35⋅0 per cent) underwent tumour resection, 193 (45⋅3 per cent) had treatment without tumour resection and 84 (19⋅7 per cent) received best supportive care. Abdominoperineal resection was the most frequent surgical procedure, performed in 65 patients (43⋅6 per cent of those who had tumour resection). Thirteen patients had total pelvic exenteration. In total, 63⋅8 per cent of those whose tumour was resected had potentially curative surgery. After tumour resection, 62 patients (41⋅6 per cent) had a complication within 30 days. Patients who received surgical treatment without tumour resection had a lower complication rate but a significantly higher 30-day mortality rate than those who underwent tumour resection (10 versus 1⋅3 per cent respectively; P = 0⋅002). Of all patients included in the study, 22⋅3 per cent had potentially curative treatment and the 3-year survival rate for these patients was 56 per cent.Conclusion: LRRC is a serious condition with overall poor outcome. Patients undergoing curative surgery have an acceptable survival rate but substantial morbidity. There is room for improvement in the management of patients with LRRC.