We confirmed the genome-wide significance of four regions previously associated with Crohn’s disease prognosis, but did not identify any further associations with prognosis. In their index study, Lee et al.6 adjusted for disease location before inspecting associations between disease prognosis and 170 susceptibility SNPs. This was done because some of the criteria used to define severe disease (e.g., need for recurrent surgery) could lead to an over-representation of patients with ileal disease, for whom surgery is more commonly used because the operation carries lower morbidity than colonic surgery and does not leave a permanent stoma. Considering location as a disease subtype, adjustment for location might modify the prognosis associations for SNPs with effects on particular disease locations (Fig. 4). We did not adjust for location here, as our aims were to identify associations with prognosis independently of possible mechanism, and to demonstrate the utility of our approach on published summary statistics.
