BACKGROUND: Evidence that thiopurines impact on the risk of surgery in elderly onset inflammatory bowel disease (EO-IBD) is lacking. We aimed to compare the rates of surgery in EO-IBD (>60 years at diagnosis) with adult-onset IBD (18-59 yrs), and examine the impact of thiopurines on surgical risk in EO-IBD. METHODS: Using a U.K. database between 1990 and 2010, we compared rates of surgery between adult-onset IBD and EO-IBD using survival analysis. Ulcerative colitis (UC) and Crohn's disease (CD) were analyzed separately. Cox proportional hazard modeling was used to determine the adjusted relative risk of surgery. We further assessed the impact of duration of thiopurine treatment on risk of surgery. RESULTS: We identified 2758 of 9515 patients with UC and 1349 of 6490 patients with CD, with EO-IBD. Cumulative 1, 5, and 10 years risk of colectomy was similar in EO-UC (2.2, 4.5, and 5.8%, respectively) and AO-UC (2.2, 5.0, and 7.3%, respectively; P = 0.15). Cumulative 1, 5, and 10 years risk of first intestinal surgery was lower in EO-CD (9.5, 14.6, and 17.9%, respectively) than AO-CD (12.2, 19.0, and 24.4%, respectively; P < 0.001). Early steroid use, steroid dependency, and thiopurine use was associated with higher risk of colectomy in EO-UC. Among EO-UC receiving thiopurines for >12 months, there was a 70% reduction in risk of colectomy (hazard ratio. 0.30; 95% confidence interval, 0.15-0.58). Thiopurines were not associated with a reduced risk of surgery in EO-CD. CONCLUSIONS: Risk of colectomy in EO-UC does not differ from AO-UC, but the risk of surgery in EO-CD is significantly lower than in AO-CD. Sustained thiopurine use of 12 months or more duration in EO-UC reduces the risk colectomy, but does not impact on the risk of surgery in EO-CD. These findings are important given the greater risk of thiopurine-associated lymphoma in the elderly.