Background: The choice of operation for complicated diverticular disease is contentious. The aim of this study was to investigate adverse events following restorative (primary resection and anastomosis, PRA) and non-restorative (Hartmann's procedure, HP) surgery for complicated diverticular disease. Methods: Five hundred and thirty-nine patients who presented with complicated diverticular disease in 42 centres over a 12-month period from January 2003 were considered for the study. Data were collected prospectively from 248 patients (46.0 per cent) who underwent PRA and 167 (31.0 per cent) who had HP. A propensity score was developed for case-mix adjustment. Multifactorial logistic regression was used to evaluate differences in operative outcomes. Results: Mortality, surgical and medical complication rates were 4.0, 31.0 and 13.7 per cent respectively after PRA, and 23.4, 53.3 and 40.7 per cent for HP (all P < 0.001). After adjusting for the propensity score, the HP group had a 2.1- and 1.9-fold increase in medical and surgical complications respectively compared with those who had PRA, whereas the operative mortality rate was not significantly different. Non-colorectal surgeons performed a significantly higher proportion of HPs in the non-elective setting than colorectal surgeons (80.6 versus 60.4 per cent; χ2 = 8.31, 1 d.f., P = 0.004). Conclusion: PRA with or without a proximal diversion is more often performed non-electively by specialist colorectal surgeons. It may be a safe procedure for complicated diverticular disease in selected patients as it may be associated with fewer postoperative adverse events.