Background: Cigarette smoking affects susceptibility to ulcerative colitis (UC), but its effects on age at diagnosis, disease extent, and need for surgery are less well defined. We examined these parameters in a detailed retrospective analysis of a large cohort of well-characterized UC patients.

Methods: 499 UC patients (254 male, median age 34.3 yr) were studied. Data were collected on smoking habits, smoking load (pack-years), age at recruitment, age at diagnosis, surgery, and disease extent. Colonoscopic and histological data at both diagnosis and follow-up (median follow-up time 4.6 yr) were available on 349 patients.

Results: Ex-smokers were older at diagnosis than current or nonsmokers, (46.5 yr vs 31.1 or 29.4 yr, respectively, P < 0.001). Before diagnosis, ex-smokers had a higher smoking load than current smokers (13.0 vs 6.94 pack-years, P < 0.001). A Cox model for age at diagnosis, with smoking as a time-dependent covariate, showed that at any age, ex-smokers were significantly more likely to develop UC than current smokers (hazard ratio 1.8, 95% CI 1.41-2.44, P < 0.001). For current smokers at latest colonoscopy, those with extensive disease were the lightest smokers (median 0.320 pack-years), whereas those with healthy colons were the heaviest smokers (median 9.18 pack-years, P= 0.006). At 5 yr, regression of extensive disease was more frequent in current than ex-smokers or nonsmokers (30% current smokers vs 8% nonsmokers and 5% ex-smokers, chi(2)= 30.4, P < 0.001) but these differences were not maintained over a longer time period.

Conclusions: Smoking habit influences the age at diagnosis and changes in disease extent in UC. Mechanisms are likely to be complex and require further investigation.