Does cigarette smoking influence the phenotype of Crohn's disease? Analysis using the Montreal classification

OBJECTIVES: The clinical subclassification of Crohn's disease by phenotype has recently been reevaluated. We have investigated the relationships between smoking habit, age at diagnosis, disease location, and progression to stricturing or penetrating complications using the Montreal classification. METHODS: 408 patients (157 male, median age 29.4 yr) were assessed. Data were collected on smoking habit, age at diagnosis, anatomical distribution, and disease behavior. Follow-up data were available on all patients (median 10 yr). RESULTS: At diagnosis, ex-smokers (N = 53) were older than nonsmokers (N = 177) or current smokers (N = 178, medians 43.2 vs 28.3 or 28.9 yr, respectively, P < 0.001). Disease location differed according to smoking habit at diagnosis (chi(2) = 24.1, P = 0.02) as current smokers had less colonic (L2) disease than nonsmokers or ex-smokers (30% vs 45%, 50%, respectively). In univariate Kaplan-Meier survival analysis, smoking habit at diagnosis was not associated with time to development of stricturing disease, internal penetrating disease, perianal penetrating disease, or time to first surgery. Patients with isolated colonic (L2) disease were slower to develop strictures (P < 0.001) or internal penetrating disease (P = 0.001) and to require surgery (P < 0.001). Cox models with smoking habit as time-dependent covariates showed that, relative to ileal (L1) location of disease, progression to stricturing disease was less rapid for patients with colonic (L2) disease (HR 0.140, P < 0.001), but not independently affected by smoking habit. Progression to surgery was also slower for colonic (L2) than ileal (L1) disease location (HR 0.273, P < 0.001), but was independent of smoking habit. CONCLUSIONS: Smoking habit was associated with age at diagnosis and disease location in Crohn's disease, while disease location was associated with the rate of development of stricturing complications and requirement for surgery. The pathogenic basis of these observations needs to be explained.