SEARCH

SEARCH BY CITATION

Keywords:

  • smoking;
  • colorectal neoplasms;
  • cohort studies

Abstract

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Epidemiological studies have consistently found a positive association between cigarette smoking and risk of colorectal adenomas, so the absence of a clear association between smoking and colorectal cancer risk may seem paradoxical. However, if colorectal cancer develops only after an induction period of about 35 years, as has been proposed recently, then studies in which all subjects have fewer than about 35 years between smoking commencement and assessment of outcome would be unlikely to detect this association. Few studies have examined smoking of several decades' duration among women. Therefore, in the cohort study reported here, we used proportional hazards models to estimate hazard ratios relating cigarette smoking to colorectal cancer risk among 89,835 women aged 40–59 years at recruitment into the Canadian National Breast Screening Study, a randomized controlled trial of mammography screening for breast cancer. During an average 10.6 years of follow-up (936,433 person-years), a total of 527 women were diagnosed with incident colorectal cancer (363 colon and 164 rectal). We found that smoking was associated with increased risk of rectal cancer 30 years or more after commencement, and especially with smoking of 40 years' duration or longer (hazard ratio=3.14, 95% CI=1.33–7.42). There was little evidence for altered risk of colon cancer. These results, along with those of other recent studies, support the hypothesis that tobacco smoking is an initiator, rather than a promoter, of rectal cancer. However, the results do not support an association with colon cancer risk, even with smoking of very long duration and high intensity. © 2002 Wiley-Liss, Inc.

A positive association between cigarette smoking and risk of colorectal adenomatous polyps has been observed consistently.1 This association may be causal given the identification of numerous carcinogenic compounds produced by burning tobacco,2, 3, 4 the possibility of both direct effects of those compounds on digestive tract tissues and of systemic effects, and the tobacco-induced genetic damage to colorectal epithelium observed experimentally.5 Since most colorectal cancers develop from adenomas,6, 7, 8 the fact that many epidemiological studies have shown no clear association between cigarette smoking and colorectal cancer risk may seem surprising.

Several hypotheses have been proposed to explain the discrepant findings for colorectal cancers and adenomas. Since colorectal polyps may remain undetected throughout life, spuriously positive associations with adenomas (ascertainment bias) may result if smokers are more likely to undergo endoscopic examination than non-smokers. However, there exists no clear evidence for such bias. Furthermore, since some of the adenoma studies in which positive associations were observed limited their analyses to subjects who had undergone endoscopy,9, 10 it has also been proposed that nicotine-induced colon motility would lead to more effective bowel clearance before colonoscopy with consequent greater sensitivity for the detection of small polyps in smokers. On the other hand, if colorectal cancer emerges only after a 35-year induction period, as has been proposed recently,1 studies with an insufficiently long interval between the commencement of smoking and the end of follow-up would be unlikely to detect an association. A long induction period is suggested by the 3–4 decades that separate germline mutations in the adenomatosis polyposis coli (APC) gene and subsequent colorectal cancer diagnosis.1, 11 In keeping with this, studies in which smoking of considerable duration has been examined have been more likely to show positive associations with colorectal cancer risk.1 However, most of these studies have been conducted in men, since women did not smoke in substantial numbers prior to the 1950s, at least in North America.12 Therefore, in the cohort study reported here, we examined the association between cigarette smoking and the risk of colorectal cancer in a cohort of women with up to 4 decades of smoking duration at recruitment, who were subsequently followed for an average of 10.6 years.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Study Population

Our investigation was conducted using data from the Canadian National Breast Screening Study (NBSS). The NBSS is a multi-center randomized controlled trial of mammography screening for breast cancer in 89,835 women aged 40–59 years at recruitment.13 Participants were recruited between 1980 and 1985 by various means, including personal invitation by letter, group mailings to employees of large institutions and to members of professional associations, advertisements in newspapers and public service announcements on radio and television.

Questionnaires

On enrollment in the NBSS, all participants completed a questionnaire that sought data on demographic characteristics, lifestyle (including cigarette smoking), menstrual and reproductive history, use of oral contraceptives and replacement estrogens. Regarding smoking history, participants were first asked whether or not they had ever smoked. Women who had ever smoked were then asked how many cigarettes they smoked per day, for how many years they had smoked and the year they had ceased smoking (former smokers only). Starting in 1982, a questionnaire regarding diet and physical activity was distributed to all new attendees at all screening centers, and to women returning to the screening centers for re-screening. By the time that the dietary questionnaire was introduced, some women had already been enrolled in the study and were not seen again at the screening centers. A total of 56,837 women returned completed dietary questionnaires. Therefore, analyses were performed both on the entire cohort and among women for whom information on physical activity was available (see below).

Case Definition and Ascertainment

Outcome (incident colorectal cancer or death) was ascertained by means of computerized record linkage to the National Mortality Database and to the Canadian Cancer Database, both of which are maintained by Statistics Canada. There is good evidence from the NBSS and from other sources that the use of record linkage to ascertain incident cancer cases and deaths in Canada is both accurate and complete.14, 15 After excluding 66 women with prevalent colorectal cancer at baseline, we identified 527 incident colorectal cancers (363 colon and 164 rectal) in total.

Colon cancers were defined as those occurring above the peritoneal delineation of the abdominal cavity and rectal cancers were defined as those occurring below this delineation. For subanalyses by colon cancer site, proximal colon cancers were defined as those occurring from the cecum through the splenic flexure (n=172), and distal colon cancers were defined as those occurring from the descending colon through the sigmoid colon (n=148).

Statistical Analysis

Follow-up of the cohort was continued until the date of diagnosis, the date of death or the end of the follow-up period (December 31, 1993), whichever was the earliest. Cox proportional hazards models were used to estimate hazard ratios (RR) and 95% confidence intervals (CI) for the association between cigarette smoking and colorectal cancer risk. Age at smoking commencement was calculated for each smoker by subtracting her total years of smoking (and the time since quitting for ex-smokers) from her age at recruitment. Multivariate models included the following potential confounders: age in 5-year age groups, Quetelet's Index, educational level (less than high school, high school and university), menopausal status (pre-, peri- and post-menopausal), hormone replacement therapy (never + tertiles of duration), vigorous physical activity and alcohol consumption. Alcohol consumption and physical activity were categorized by tertiles or as “missing,” where the latter group was comprised of women who did not complete the questionnaire regarding physical activity and diet. The lowest tertile of alcohol consumption was comprised of non-drinkers. To assess the possibility of residual confounding among women with missing information on physical activity, we conducted additional analyses limited to the 56,837 women who completed questionnaires regarding physical activity and diet. For tests of trend in risk across successive levels of categorical variables, median values of each category were fitted in the risk models as successive integers.16

RESULTS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

On average, participants were followed for 10.6 years, yielding a total of 936,433 person-years of follow-up for the cohort. The average ages at diagnosis of colon and rectal cancer were 57.8 and 57.5 years, respectively. Since the results were very similar in age-adjusted and multivariate-adjusted models, only the multivariate-adjusted rate ratios for cigarette smoking are shown in Table I.

Table I. Adjusted Rate Ratios for Colorectal Cancer in Relation to Cigarette Smoking1
Smoking measureCases/person-yearsColorectal cancer RR (95% CI)Colon cancer RR (95% CI)Rectal cancer RR (95% CI)
  • 1

    Multivariate models included age in 5-year age groups, Quetelet's index (quartiles), education level (less than high school, high school, and university), vigorous physical activity (hr per day in tertiles, and “missing”), hormone replacement therapy (never + 4 levels of duration), menopausal status (pre, peri and post) and alcohol consumption (tertiles, and “missing”).–*All p-values are from 2-sided tests.

Smoking status
Never-smokers274/498,5161.0 (referent)1.0 (referent)1.0 (referent)
Ever-smokers253/222,8701.07 (0.90–1.28)0.98 (0.79–1.22)1.30 (0.95–1.78)
Ex-smokers145/102,1821.15 (0.94–1.41)1.03 (0.80–1.33)1.44 (1.00–2.06)
Current smokers108/120,6881.00 (0.80–1.26)0.93 (0.71–1.24)1.17 (0.78–1.75)
Cigarettes/day
Never-smokers274/498,5161.0 (referent)1.0 (referent)1.0 (referent)
1–956/102,1821.01 (0.75–1.35)0.89 (0.61–1.28)1.31 (0.80–2.14)
10–1978/120,6881.24 (0.96–1.60)0.94 (0.67–1.32)1.98 (1.32–2.96)
20–2993/166,8461.11 (0.87–1.41)1.16 (0.87–1.53)0.97 (0.61–1.56)
30–3912/29,4140.83 (0.46–1.48)0.87 (0.44–1.69)0.72 (0.23–2.29)
40+8/23,1940.71 (0.35–1.43)0.63 (0.26–1.52)0.90 (0.28–2.85)
p for trend*0.880.990.82
Years Smoked
Never-smokers274/498,5161.0 (referent)1.0 (referent)1.0 (referent)
1–942/84,3981.04 (0.75–1.44)0.93 (0.61–1.40)1.31 (0.75–2.28)
10–1953/113,2761.00 (0.74–1.35)0.90 (0.62–1.30)1.24 (0.75–2.05)
20–2983/156,6211.07 (0.89–1.47)1.04 (0.77–1.42)1.37 (0.89–2.11)
30–3961/79,9401.14 (0.87–1.54)1.16 (0.83–1.63)1.12 (0.65–1.94)
40+12/9,0841.30 (0.68–2.47)0.68 (0.25–1.86)3.14 (1.33–7.42)
p for trend*0.170.660.07
Years since smoking commenced
Never-smokers274/498,5161.0 (referent)1.0 (referent)1.0 (referent)
1–912/16,459158 (0.89–2.83)1.50 (0.74–3.05)1.76 (0.64–4.82)
10–1924/63,9710.88 (0.58–1.33)0.84 (0.50–1.40)0.97 (0.47–2.02)
20–2985/207,3430.97 (0.75–1.25)0.91 (0.67–1.24)1.11 (0.72–1.73)
30–39105/136,8161.18 (0.93–1.49)1.05 (0.79–1.39)1.52 (1.01–1.26)
40+22/16,7361.42 (0.89–2.27)1.12 (0.62–2.04)2.27 (1.06–4.87)
p for trend*0.210.980.03

Smoking status (ever, former or current) was not clearly associated with the risk of colorectal cancer as a whole or with the risk of colon or rectal cancer when these endpoints were examined separately (Table I). Similarly, we observed no clear association with smoking 40 cigarettes per day or more, smoking for 40 pack-years or more (data not shown) or the commencement of smoking at age 15 years or younger (data not shown).

For colon cancer, there was also no clear association with smoking of long duration or having commenced smoking many years earlier (latency). For proximal tumors, the multivariate-adjusted rate ratio for smoking 40 years or more compared to never smoking was 1.07 (95% CI 0.33–3.46, p for trend = 0.79). For distal tumors, the multivariate-adjusted rate ratio for smoking 30 years or more compared to never smoking was 1.16 (95% CI 0.67–2.03, p for trend = 0.94). There were not enough cases of distal colon cancer in the highest smoking category (40 or more years' duration) to estimate the relative risk. Among ex-smokers, years since smoking ceased was not associated with colon cancer risk (data not shown).

With respect to rectal cancer, a non-significant increased risk was observed among both current and former smokers compared to never smokers (Table I), although there was no clear association with the number of cigarettes smoked per day, the number of pack-years smoked or the age at which smoking commenced (data not shown). In contrast, smoking of 40 years' duration or longer was associated with a 3-fold increased risk of rectal cancer. However, there were only 8 cases in the highest category of smoking duration. Women who had commenced smoking at least 30 years earlier were also at increased risk of rectal cancer, a finding that was based on a larger number of cases in that smoking category (n=44), and the dose-risk trend was statistically significant. Among ex-smokers, years since smoking ceased was not clearly associated with rectal cancer risk. The relative risks for rectal cancer and 95% confidence intervals for increasing categories of time since quit (1–9 years, 10–19 years, 20+ years) compared with never smokers was 1.57 (0.97–2.54), 1.28 (0.73–2.25) and 1.23 (0.63–2.39), respectively, p for trend = 0.35. The addition of smoking duration to the model appeared to nullify that result, however, while the association with very long duration remained essentially unchanged.

The results for smoking were not altered by further adjustment for use of oral contraceptives, parity or dietary intake of folic acid (data not shown). In addition, the results were also similar when we restricted the analyses to the 56,837 women who completed questionnaires regarding physical activity (and diet), though statistical precision was reduced; these women were similar to the remainder of the cohort with respect to the variables included in the multivariate models. Regarding the covariates, we observed a positive association between obesity and colorectal cancer, but only among premenopausal women, a statistically non-significant 30% decrease in colorectal cancer risk with long duration of oral contraceptive use, and an 83% increased risk of rectal cancer (but not colon cancer) among women with a long duration of hormone replacement therapy.

DISCUSSION

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In our large prospective cohort study, we found that smoking was associated with increased risk of rectal cancer 30 years or more after commencement and especially with smoking of 40 years' duration or longer. In contrast, smoking of shorter durations was not associated with the risk of colorectal cancer at any cancer site. These results are consistent with the hypothesis that tobacco smoking is an initiator, rather than a promoter, of rectal cancer. However, the results do not support an association with colon cancer risk, even with smoking of very long duration or high intensity.

Among the strengths of our study was the large sample size of our cohort of women and the relatively long-term follow-up. The completeness of follow-up of the cohort reduces the likelihood that our results reflect bias due to differential follow-up of long-term smokers compared to non-smokers. On the other hand, although we did adjust our estimates for a wide range of potentially confounding variables, we cannot exclude the possibility of residual confounding by other factors. Furthermore, the low number of rectal cancer cases in the highest category of smoking duration raises the possibility that the findings for that smoking measure were due to chance. However, the positive association between smoking latency and rectal cancer risk was based on a larger number of cases and is, therefore, less likely to reflect random variations in risk. Finally, it is possible that smoking habits changed during follow-up, in which case non-differential misclassification of smoking would have tended to attenuate the rate ratios.

The 2- to 3-fold increased risk of rectal cancer with long-term smoking in our data is similar in magnitude to the 2- to 3-fold increased risks observed with long-term smoking in several prospective cohort studies conducted over the past few years, mostly in men.1 In addition, while some of these previous studies have found similar positive associations for colon and rectal cancers with long-term smoking, or even stronger associations with colon cancer,1 most studies that have observed differences in the association at the 2 cancer sites have found a positive association that was stronger with, or limited to, rectal cancer risk.1 In the Nurses' Health Study,9 for example, a stronger association was observed for the risk of rectal cancer than colon cancer. In an update of the prospective cohort study among male British doctors,17 no clear association was observed between long-term smoking and colon cancer mortality, but a statistically significant positive association was observed with rectal cancer mortality. Although a large prospective cohort study of male Swedish construction workers18 failed to find an association between smoking of 40 years' duration or longer and colorectal cancer risk, the results were suggestive of an increased risk of rectal cancer. Smoking of long duration was also associated more strongly with increased rectal cancer risk in a cohort of male American veterans19 and most recently, in a cohort of Swedish twins.20 In the latter study, which included both males and females, the association was observed only among males, likely due to the relative scarcity of females who had smoked for a great number of years. As with the Nurses' Health Study,9 the results of the present study suggest that smoking of long duration may also be associated with increased risk in females.

The results of the present study support the hypothesis that tobacco smoking is an initiator, rather than a promoter, of rectal cancer. As with several previous studies, we found that long-term smoking is associated more strongly with increased risk of rectal cancer than colon cancer. Although the reason for this is unclear, it is possible that, with the progressive absorption of water from the stool during its passage through the colon,21 the ingested carcinogens from tobacco smoke become progressively more concentrated and deleterious to the colorectal epithelium. However, given the low numbers of rectal cancer cases in the highest categories of smoking duration and latency in most studies that have examined these measures, the interpretation of this finding should be made cautiously. The association between cigarette smoking and colorectal cancer risk might be clarified further through the analysis of data from several pooled prospective cohort studies, as has been done recently22 in the study of risk factors for breast cancer.

REFERENCES

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Giovannucci E. An updated review of the epidemiological evidence that cigarette smoking increases risk of colorectal cancer. Cancer Epidemiol Biomarkers Prev 2001;10: 72531.
  • 2
    IARC. Tobacco smoking. IARC Monographs on the evaluation of the carcinogenic risk of chemicals to humans. IARC Sci Publ 1986;38: 1397.
  • 3
    Hoffmann D, Hoffmann I. The changing cigarette, 1950–1995. J Toxicol Environ Health 1997;50: 30764.
  • 4
    Alexandrov K, Rojas M, Kadlubar FF, Lang NP, Bartsch H. Evidence of anti-benzo[a]pyrene diolepoxide-DNA adduct formation in human colon mucosa. Carcinogenesis 1996;17: 20813.
  • 5
    Manabe S, Tohyama K, Wada O, Aramaki T. Detection of a carcinogen, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), in cigarette smoke condensate. Carcinogenesis 1991;12: 19457.
  • 6
    Shimoda T, Ikegami M, Fujisaki J, Matsui T, Aizawa S, Ishikawa E. Early colorectal carcinoma with special reference to its development de novo. Cancer 1989;64: 113846.
  • 7
    Jass JR. Do all colorectal carcinomas arise in preexisting adenomas? World J Surg 1989;13: 4551.
  • 8
    Kune GA. Causes and control of colorectal cancer: a model for cancer prevention. In: Norwell, MA: Kluwer Academic Publishers, 1996. 13954.
  • 9
    Giovannucci E, Colditz GA, Stampfer MJ, Hunter D, Rosner BA, Willett WC, Speizer FE. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. women. J Natl Cancer Inst 1994;86: 1929.
  • 10
    Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Ascherio A, Kearney J, Willett WC. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. men. J Natl Cancer Inst 1994;86: 18391.
  • 11
    Vasen HF, Griffioen G, Offerhaus GJ, Den Hartog Jager FC, Van Leeuwen-Cornelisse IS, Meera Khan P, Lamers CB, Van Slooten EA. The value of screening and central registration of families with familial adenomatous polyposis: a study of 82 families in The Netherlands. Dis Colon Rectum 1990;33: 22730.
  • 12
    Pierce JP, Fiore MC, Novotny TE, Hatziandreu EJ, Davis RM. Trends in cigarette smoking in the United States. Projections to the year 2000. J Am Med Assoc 1989;261: 615.
  • 13
    Miller AB, Baines CJ, To T, Wall C. 1992. Canadian National Breast Screening Study. I. Breast cancer detection and death rates among women aged 4–49 years. II. Breast cancer detection and death rates among women aged 5–59 years. CMAJ 1992;147: 145988.
  • 14
    Shannon HS, Jamieson E, Walsh C, Julian JA, Fair ME, Buffet A. Comparison of individual follow-up and computerized record linkage using the Canadian Mortality Data Base. Can J Public Health 1989;80: 547.
  • 15
    Robles SC, Marrett LD, Clarke EA, Risch HA. An application of capture-recapture methods to the estimation of completeness of cancer registration. J Clin Epidemiol 1988;41: 495501.
  • 16
    Rothman KJ, Greenland, S. Modern Epidemiology, 2nd ed. Philadelphia: Lippincott-Raven, 1998. 319.
  • 17
    Doll R, Peto R, Wheatley K, Gray R, Sutherland I. Mortality in relation to smoking: 40 years' observations on male British doctors. Br Med J 1994;309: 90111.
  • 18
    Nyren O, Bergstrom R, Nystrom L, Engholm G, Ekbom A, Adami HO, Knutsson A, Stjernberg N. Smoking and colorectal cancer: a 20-year follow-up study of Swedish construction workers. J Natl Cancer Inst 1996;88: 13027.
  • 19
    Heineman EF, Zahm SH, McLaughlin JK, Vaught JB. Increased risk of colorectal cancer among smokers: results of a 26-year follow-up of US veterans and a review. Int J Cancer 1994;59: 72838.
  • 20
    Terry P, Ekbom A, Lichtenstein P, Feychting M, Wolk A. Long-term tobacco smoking and colorectal cancer in a prospective cohort study. Int J Cancer 2001;91: 5857.
  • 21
    Johnson LR. Fluid and electrolyte absorption. In: JohnsonLR, ed. Gastrointestinal physiology, 6th ed. St. Louis, MO: Mosby, 2001. 14353.
  • 22
    Smith-Warner SA, Spiegelman D, Yaun SS, van den Brandt PA, Folsom AR, Goldbohm RA, Graham S, Holmberg L, Howe GR, Marshall JR, Miller AB, Potter JD, et al. Alcohol and breast cancer in women: a pooled analysis of cohort studies. J Am Med Assoc 1998;279: 53540.