INTRODUCTION: Elderly onset (>60 yrs at diagnosis) Crohn's disease (CD) seems to be associated with a better outcome than when diagnosed earlier in life. The aim of this study was to compare the natural history of patients with CD older than 70 years at diagnosis with that of elderly patients diagnosed between 60 and 70 years in the EPIMAD population-based registry. METHODS: Three hundred seventy patients with elderly onset CD diagnosed between January 1988 and December 2006 were identified. Among them, 188 (63%) were older than 70 years at diagnosis (≥70 yrs). Clinical presentation, disease location, and behavior at diagnosis and also natural history, surgery needs, and drug exposure were recorded, with a median follow-up of 4.5 years (1.1; 8.3) in CD diagnosed after 70 years and of 7.8 years (3.3; 12.1) in CD diagnosed between 60 and 70 years, respectively. RESULTS: CD incidence in elderly patients diagnosed ≥70 years was 2.3/100,000 inhabitants, compared with 2.6/100,000 in elderly patients diagnosed below the age of 70 (60-69 yrs). The proportion of males was lower in patients ≥70 years than in patients aged 60 to 69 (31% versus 45%, P = 0.006). Clinical presentation at diagnosis was similar in both groups. Pure colonic location (L2) was more frequent among patients >70 years both at diagnosis (73% versus 57%, P = 0.004) and maximal follow-up (70% versus 47%, P < 0.0001). Disease extension (from L1 or L2 to L3) was not significantly different among patients >70 years and patients aged 60 to 69 years (hazard ratio [HR] = 2.0 [0.9; 4.5] for 60 to 69 yrs, P = 0.09). The most frequent behavior in the 2 groups was inflammatory, both at diagnosis (75% versus 80%, P = 0.43) and at maximal follow-up (69% versus 70%, P = 0.55). There was no significant difference in patients >70 years compared with patients aged 60 to 69 years regarding treatment with 5-aminosalicylic acid (P = 0.72), oral corticosteroids (P = 0.83), and anti-tumor necrosis factor therapies (P = 0.37). However, the use of immunosuppressants was significantly less frequent in patients >70 years (HR = 2.1 [1.3; 3.5] for 60 to 69 yrs, P = 0.003). Risk of surgery was similar in both groups (P = 0.72). Extraintestinal manifestations at diagnosis were significantly associated with an evolution to complicated behavior (HR = 2.7 [1.0; 7.0], P = 0.045), immunosuppressant treatment (HR = 2.9 [1.4; 6.0], P = 0.006), and corticosteroid use (HR = 3.3 [1.8; 6.1], P < 0.0001). CONCLUSIONS: The natural history of CD in elderly patients diagnosed over the age of 70 is mild with low disease extension and complicated behavior. This needs to be taken into account when establishing therapeutic strategies.