Aspects of the etiology and survival of lower gastrointestinal cancers

Abstract

In the first two papers, the Inpatient Register was used to identify patients hospitalized with (i) condylomata acuminata, or with (ii) benign anal lesions. In both papers, the cancer incidence in the cohort was compared to the general population generating standardized incidence ratios (SIRs). Cancer cases occurring during the (i) first year, or (ii) first three years of follow-up and accrued person-years were excluded.

(i) Between 1965 and 1999, 10,971 patients were hospitalized with condylomata acuminata. During a median follow-up of 13 years, 473 cancer cases occurred, and the corresponding SIR for all cancers was 1.5. Although based on few cases, the risks of all anogenital cancers, except cervical cancer, were significantly elevated. The pattern of relative risks for different types of cancers resembled the patterns described for patients with a suppressed immune defense.

(ii) Between 1965 and 2002, 45,186 patients were hospitalized with benign inflammatory anal lesions (anal fissure, fistula, and perianal abscess), and 79,808 patients were hospitalized with hemorrhoids. Patients with inflammatory anal lesions had a 3-fold increased risk of anal cancer that persisted over time. There was no persistent risk elevation in patients hospitalized with hemorrhoids.

The third and fourth papers were based on the Construction Workers Cohort consisting of 336,381 males. Detailed information on tobacco use, smoking and Swedish moist snuff (snus), was collected at cohort entry in 1971-1992. Never-users of any tobacco served as reference group.

(iii) The third paper studied tobacco use and the incidence of colorectal and anal cancer. After up to 37 years of follow-up, pure smoking was marginally associated with colon cancer risk, modestly associated with rectal cancer risk, and there was a substantially increased risk of anal cancer. Snus use was not significantly associated with neither colorectal, nor anal cancer risk. However, the point estimates of colon cancer risk were similar in snus users and smokers.

(iiii) The fourth paper studied the impact of tobacco use on cancer survival, with a specific interest in colorectal cancer. There were 40,230 incident cancer cases in the cohort. Both smoking and snus use were associated with an increased risk of death, even though the estimates tended to be slightly higher for smokers. Smokers had a borderline 25% increased risk of rectal cancer-specific death, whereas there was no excess risk of colon cancer death. Snus use was not significantly associated with colorectal cancer death. Using data on comorbidity, a stratified analysis revealed no substantial differences between those with and without comorbid conditions. This suggests that an excess fatality among tobacco users might be a biological effect, possibly exerted by nicotine.