Cigarette Smoking and Ulcerative Colitis: A Case-Control Study

mayo Cigarette Smoking and Ulcerative Colitis: A Case-Control Study MARC

D.

SILVERSTEIN,

M.D.,

BRET

A.

LASHNER,

M.D.,*

AND STEPHEN

* Objective: To determine whether the previously reported decreased risk of ulcerative colitis in current smokers and increased risk in former smokers are explained by age, sex, race, ethnicity, or socioeconomic status. * Design: We conducted a case-control study at a university hospital gastroenterology clinic. • Material and Methods: One hundred patients with ulcerative colitis and 100 age- and sex-matched community control subjects were randomly selected for a telephone interview to collect information on smoking habits, race, religion, income, education, and occupation. Smoking habits at the onset of symptoms were analyzed with use of conditional logistic regression for matched data to obtain adjusted odds ratios and 95% confidence intervals for current or former smokers. • Results: In comparison with those who had never smoked, current smokers were less likely to have ulInitial observations of the low rate of cigarette smoking among patients with ulcerative colitis were identified in investigations of risk factors to explain the low rate of cardiovascular-related mortality among patients with ulcerative colitis1 and the potentialrole of risk factors for pulmonary disease among such patients.2 Striking case reports of the possible beneficial effects of cigarette smoking3 and nicotine chewing gum4 on the severity of symptoms in patients with ulcerative colitis led to further observational studies. Casecontrol studies showed decreased smoking among patients with ulcerative colitis in comparison with control subjects among clinic patients,5 7 hospital patients,89 and community residents.10'4 Examination of potential confounding variables, including ingestion of alcohol, consumption of coffee, and socioeconomic status, did not explain the observed association. A possible explanation may be cessation of smoking as a result of subtle, early prodromal symptoms of the disFrom the Department of Medicine, University of Chicago Medical Center, Chicago, Illinois. ♦Current address: Cleveland Clinic Foundation, Cleveland, Ohio. Address reprint requests to Dr. M. D. Silverstein at his current address: Division of Area General Internal Medicine, Mayo Clinic Rochester, 200 First Street SW, Rochester, MN 55905. Mayo Clin Proc 1994;69:425-429

B.

HANAUER,

M.D.

cerative colitis: odds ratio = 0.13; 95% confidence interval = 0.05 to 0.38. Former smokers had no increased risk for ulcerative colitis: odds ratio = 1.24; 95% confidence interval = 0.52 to 2.95. No doseresponse effect was noted on the basis of pack-years of cigarette smoking, and among former smokers, the interval since quitting smoking was not significantly associated with the relative risk of ulcerative colitis. No confounding effect was detected from race, religion, income, education, or occupation. • Conclusion: An association seems to exist between ulcerative colitis and nonsmoking; perhaps patients with ulcerative colitis who smoke are less likely to experience symptoms than are nonsmokers because of the effects of nicotine. (Mayo Clin Proc 1994; 69:425-429) CI = confidence interval

ease. Some studies, however, directly ascertained smoking at the time of onset of symptoms rather than at the time of diagnosis and found that lifetime nonsmokers have an increased risk of ulcerative colitis. Some,7·'0·'2·'5 but not all,5·68·9 published studies showed that former smokers may have an increased risk of ulcerative colitis for several years after cessation of smoking. We undertook the current casecontrol study to assess whether nonsmoking is associated with ulcerative colitis by using age- and sex-matched community control subjects and by adjusting for race, ethnicity, and socioeconomic status as potential confounding factors. A secondary goal of the study was to assess the risk, if any, associated with former smoking. METHODS Cases.—The University of Chicago Gastroenterology Clinic is an urban, teaching hospital-based clinic that has served as a national referral center for patients with inflammatory bowel disease. All new patients referred to the Gastroenterology Clinic for assessment and found to have ulcerative colitis were eligible "cases" for this study. Ulcerative colitis was diagnosed on the basis of a clinical syndrome of diarrhea or hematochezia; radiologie, colonoscopic, or pathologic confirmation; and negative results of stool cultures.

425

© / 994 Mayo Foundation for Medical Education and Research

426

CIGARETTE SMOKING AND ULCERATIVE COLITIS

Controls.—Age-matched (±5 years) and sex-matched community residents with the same area code and telephone exchange as the cases were selected as "controls" by using a modification of the method initially developed by Waksberg.16,17 A maximum of six telephone calls during morning, afternoon, and early evening hours of weekdays was made to identify residences. When a residence was identified, the purpose of the call was explained to the person who answered the telephone. If an age- and sex-matched respondent was available, verbal informed consent was sought for completion of the telephone interview. If no match was available or if the matched person declined to participate, a new telephone number was selected. The process was repeated until one control was matched to each case. Usually, data from controls were collected concurrently with data from the matched cases; however, a delay of several weeks occasionally occurred because of the time needed for the study assistant to make the necessary telephone calls to identify a residence with an available matched control who consented to participate. Study Subjects.—On the basis of an estimated 30% prevalence of smoking among controls'8 and a reported odds ratio of 0.3, we calculated that, in a matched study, 93 cases and controls would be necessary to provide 90% power with a two-sided a level of 0.05.19 Initially, we had 102 pairs of cases and matched controls. A careful review of the medical records that summarized the diagnostic evaluation indicated that 2 patients did not fulfill strict study criteria for ulcerative colitis, and they were excluded from further study; thus, 100 case and control pairs were available for analysis. Collection of Data.—The telephone interview of cases and each age- and sex-matched control was conducted by one of two trained interviewers, who used a standard questionnaire. Respondents were asked whether they were smokers and, if so, whether they currently smoked cigarettes. Current cigarette smokers were asked to report the age at which they had begun smoking regularly and the amount smoked. Former cigarette smokers were asked the age at which they had begun smoking, how much they smoked most recently before stopping, and the age at which they had stopped smoking. Potential controls who had ever had inflammatory bowel disease, colitis (unspecified), ulcerative colitis, Crohn's disease, or regional enteritis, but not spastic colon or irritable bowel syndrome, were excluded from analysis. All respondents were asked to classify themselves by race, religion, education, income, and occupation. Analysis ofData.—The smoking habit of cases before the onset of symptoms was the exposure of interest. For the controls, the smoking habit at the same age (or the current age for controls who were younger than the matched case when disease developed) as the onset of symptoms of the matched case was the exposure of interest. Conditional

Mayo Clin Proc, May 1994, Vol 69

logistic regression for matched case-control studies was performed; cigarette smoking was the exposure of interest, and race, religion, education, income, and occupation were potential confounding variables of interest. Because each control was matched to a 5-year interval around the corresponding case's age, age was included in the analyses. In some of the conditional logistic regression models, the covariate of religion (highly significant with use of McNemar's test) failed to come into the model because no discordant pairs of one type could be found, and the standard error of the estimate became infinite. Therefore, for analyses with the covariate of religion, unconditional logistic regression was performed, in which age and sex (the matching variables) were included with other covariates of interest. Odds ratios and 95% confidence intervals (CI) for the odds ratios were calculated.20 The odds ratio was used to estimate the relative risk of ulcerative colitis in cigarette smokers in comparison with those who had never smoked (never smokers). Analyses were performed for the number of pack-years of cigarette smoking to test for a dose-response effect among current or former cigarette smokers in comparison with never smokers. The risk of ulcerative colitis among former smokers in comparison with never smokers was analyzed by interval of time since quitting smoking among the former smokers. RESULTS The 54 female and 46 male cases were matched to controls of the same gender. The overall mean age was 36.7 years for the cases and 36.6 years for the controls, and the standard deviation was 13.4 years for both groups. The age of each matched control was within 1 year of the age of the corresponding case in 22%, within 2 years in 41%, within 3 years in 63%, and within 4 years in 81 %. Only 8% of respondents at residences called for identification of controls refused to participate. Among the cases, 69 had never smoked, 22 were former smokers, and 9 were current smokers. Among the controls, 51 had never smoked, 12 were former smokers, and 37 were current smokers. In comparison with never smokers, the odds ratio for ulcerative colitis among current smokers (0.13) was significantly different from 1 (95% CI, 0.05 to 0.38), and the odds ratio for former smokers ( 1.24) was not significantly different from 1 (95% CI, 0.52 to 2.95). Race, religion, education, income, and occupation were significantly associated with ulcerative colitis (Table 1). In multivariate analyses of the relationship of ulcerative colitis to cigarette smoking, religion and any one of the three variables related to socioeconomic status (education, income, or occupation) were significant; the remaining two socioeconomic status variables were not significant. The two best models were religion and occupation or religion and income. The relationship of smoking habits to ulcerative colitis, adjusted

Mayo Clin Proc, May 1994, Vol 69

CIGARETTE SMOKING AND ULCERATIVE COLITIS

Table 1.—Odds Ratios and 95% Confidence Intervals for Ulcerative Colitis, Stratified by Certain Variables of Study Subjects* Variable Current smoker versus never smoked Former smoker versus never smoked White versus other Jewish versus other:]: Catholic versus other Annual income >$40,000 versus other Education: college degree versus other Occupation§ Professional Sales or clerical Retired or homemaker

No. of subjects!

Odds ratio

95% CI

200

0.13

0.05-0.38

200 196 198 196

1.24 3.75 8.78 0.53

0.52-2.95 1.24-11.30 3.69-20.92 0.29-0.98

142

3.67

1.49-9.04

162 168

2.89

1.35-6.17

6.43 5.15 4.25

1.79-23.09 1.27-20.87 1.26-17.97

*CI = confidence interval. tSome respondents declined to provide some information on race, religion, income, education, and occupation. ^Determined by unconditional logistic regression. §In comparison with blue-collar workers.

for these covariates, is shown in Table 2. The odds ratio for ulcerative colitis among current smokers was significantly less than 1 in all analyses, and the odds ratio for former smokers was never significantly different from 1. Collectively, these analyses indicate little or no confounding of the significant relationship of nonsmoking to ulcerative colitis by ethnicity and socioeconomic status. Former smokers did not have a significantly increased risk of ulcerative colitis. Among current smokers, no dose-response relationship was noted in the risk of ulcerative colitis in comparison with never smokers when smokers were classified by the number of pack-years of cigarette smoking (Table 3). Among former smokers, in comparison with never smokers, the risk of ulcerative colitis was not uniformly increased. The relative risk of ulcerative colitis was increased during the first 2 years after quitting cigarette smoking but was not significant (P = 0.057; 95% CI, 0.94 to 60.34) (Table 4). Similar results were obtained when religion, income, and occupation were included in the analyses of interval after quitting cigarette smoking. DISCUSSION In this study, patients with ulcerative colitis who were referred to the University of Chicago Gastroenterology Clinic were less likely to be smokers at the time of onset of their symptoms than were age- and sex-matched residents from their community. The association between nonsmoking and

427

ulcerative colitis was not explained by age or sex; by measures of socioeconomic status such as education, income, and occupation; or by ethnicity as characterized by race and religion. Although patients with ulcerative colitis were slightly more likely to be former smokers than were age- and sex-matched community residents, this trend was observed only in those who had stopped smoking during the 2 years before the onset of symptoms of ulcerative colitis, and the change in smoking behavior may have been the result of unrecognized prodromal symptoms or a chance occurrence. The strengths of our current study include the use of community controls selected from the same telephone exchange as cases; the control of potential confounding by demographic characteristics, socioeconomic factors, and ethnicity; and the analysis of risk factors at the time of onset of symptoms of ulcerative colitis, rather than at the time of diagnosis. Our study was limited by the necessity of using patient telephone reports of smoking habits and of other potential confounding variables and the reluctance of some respondents to provide information on income, education, or occupation, which was used to control for confounding by socioeconomic status. We believe that our finding that patients with ulcerative colitis are less likely to smoke than are community residents is a true association, not attributable to demographic factors, socioeconomic factors, ethnicity, or chance. Our finding substantiates results of previous studies that also showed an association between ulcerative colitis and nonsmoking 5.8-10,12,15 j n con trast, previous studies from our institution and elsewhere demonstrated a strong association between Crohn's disease and smoking.,7·9·12·15·21 A meta-analysis of

Table 2.—Adjusted Odds Ratios and 95% Confidence Intervals for Ulcerative Colitis, Stratified by Cigarette Smoking Habits of Study Subjects* Current smoker Odds 95% CI ratio

Former smoker Odds ratio 95% CI

0.13 0.13 0.13 0.22

0.05-0.38 0.04-0.50 0.03-0.48 0.08-0.58t

1.24 1.18 1.11 1.60

0.52-2.95 0.45-3.14 0.37-3.33 0.56-4.56

0.23

0.09-0.6 I t

1.37

0.49-3.83

Covariates Occupation Income Religion, occupation:): Religion, incomet

*CI = confidence interval. tDetermined by unconditional logistic regression. tModels that included religion and occupation were not improved by inclusion of race, education, or income. Similarly, models that included religion and income were not improved by inclusion of race, education, or occupation.

428

Mayo Clin Proc, May 1994, Vol 69

CIGARETTE SMOKING AND ULCERATIVE COLITIS

Table 3.—Odds Ratios and 95% Confidence Intervals for Ulcerative Colitis Among Cigarette Smokers, Stratified by Cumulative Amount of Cigarettes Smoked Before Onset of Symptoms* Subjects (no.)

Odds ratio

95% CI

7 1 4

0.74 3.83 1.96

0.25-2.20 0.75-19.46 0.49-7.90

3 3 3

18 9 10

0.05 0.27 0.14

0.01-0.26 0.07-1.02 0.03-0.67

69

51

1.0

Referent group

Smoking status

Pack-years (no.)

Cases

Former

<10 11-20 >20

9 5 8

Current

<10 11-20 >20

Never

Controls

*CI = confidence interval. prior studies of the role of smoking in inflammatory bowel disease also concluded that these opposite effects of smoking and the development of ulcerative colitis and Crohn's disease are true relationships.22 Our study did not confirm a previously reported association between former smoking and ulcerative colitis.7'012 Nevertheless, the results of our study are not inconsistent with this hypothesis. The positive association between ulcerative colitis and a previous smoking habit was not significant and was observed only in those who had stopped smoking within 2 years of the onset of symptoms of ulcerative colitis. Thus, smoking cessation may have resulted from early (subclinical) prodromal symptoms, or this finding may possibly be ascribed to bias from an increased tendency of patients to report smoking cessation in comparison with community residents who were not patients.

mechanisms that can explain the association. Studies should examine the relationship between nicotine or other components of cigarettes and catecholamines, the effects of endogenous mediators such as prostaglandins,23 the relationship of smoking to colonie and rectal blood flow,24 and the characteristics of colonie mucus.25 At present, the biologic basis for these observations remains obscure. One study26 showed a possible beneficial effect of nicotine-replacement therapy, originally used to moderate symptoms of nicotine withdrawal in smoking-cessation programs, in modulating symptoms of ulcerative colitis. Further clinical trials of nicotine therapy for ulcerative colitis are warranted. We speculate that patients with ulcerative colitis who smoke are less likely to experience active symptoms or signs of ulcerative colitis than are nonsmokers and that this finding is due to nicotine.

CONCLUSION If an actual association exists between ulcerative colitis and nonsmoking, further research is needed to identify biologic

ACKNOWLEDGMENT We thank Tanya M. Petterson, M.S., for assistance with data analysis.

Table 4.—Odds Ratios and 95% Confidence Intervals for Ulcerative Colitis Among Former Cigarette Smokers, Stratified by Time Elapsed Since Last Use* Smoking status Former

Time since last use (mo) <24 25-48 49-72 >72

Current Never

*CI = confidence interval.

Subjects (no.) Cases Controls

Odds ratio

95% CI 0.94-60.34 0.05-5.51 0.06-10.10 0.11-1.69

11 1 3 7

1 2 1 8

7.54 0.50 0.77 0.44

9

37

0.12

0.04-0.37

69

51

1.0

Referent group

Mayo Clin Proc, May 1994, Vol 69

REFERENCES 1. Gyde S, Prior P, Dew MJ, Saunders V, Waterhouse JAH, Allan RN. Mortality in ulcerative colitis. Gastroenterology 1982;83:36-43 2. Heatley RV, Thomas P, Prokipchuk EJ, Gauldie J, Sieniewicz DJ, Bienenstock J. Pulmonary function abnormalities in patients with inflammatory bowel disease. Q J Med 1982; 51:241-250 3. de Castella H. Non-smoking: a feature of ulcerative colitis [letter]. BMJ 1982; 284:1706 4. Roberts CJ, Diggle R. Non-smoking: a feature of ulcerative colitis [letter]. BMJ 1982:285:440 5. Harries AD, Baird A, Rhodes J. Non-smoking: a feature of ulcerative colitis. BMJ 1982; 284:706 6. Logan RF, Edmond M, Somerville KW, Langman MJS. Smoking and ulcerative colitis. BMJ 1984;288:751-753 7. Thornton JR, Emmett PM, Heaton KW. Smoking, sugar, and inflammatory bowel disease. BMJ 1985;290:1786-1787 8. Jick H, Walker AM. Cigarette smoking and ulcerative colitis. NEnglJMed 1983;308:261-263 9. Franceschi S, Panza E, La Vecchia C, Parazzini F, Decarli A, Bianchi Porro G. Nonspecific inflammatory bowel disease and smoking. Am J Epidemiol 1987;125:445-452 10. Boyko EJ, Koepsell TD, Perera DR, Inui TS. Risk of ulcerative colitis among former and current cigarette smokers. N EnglJMed 1987;316:707-710 11. Calkins B, Lilienfeld A, Mendeloff A, Garland C, Monk M, Garland F. Smoking factors in ulcerative colitis and Crohn's disease in Baltimore [abstract]. Am J Epidemiol 1984; 120:498-499 12. Lindberg E, Tysk C, Andersson K, Järnerot G. Smoking and inflammatory bowel disease: a case control study. Gut 1988; 29:352-357 13. Persson P-G, Ahlbom A. Hellers G. Inflammatory bowel disease and tobacco smoke—a case-control study. Gut 1990; 31:1377-1381

CIGARETTE SMOKING AND ULCERATIVE COLITIS

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

429

Lorusso D, Leo S, Misciagna G, Guerra V. Cigarette smoking and ulcerative colitis: a case control study. Hepatogastroenterology 1989; 36:202-204 Tobin MV, Logan RFA, Langman MJS, McConnell RB, Gilmore IT. Cigarette smoking and inflammatory bowel disease. Gastroenterology 1987;93:316-321 Waksberg J. Sampling methods for random digit dialing. J AmStatAssoc 1978;73:40-46 Groves RM, Kahn RL. Surveys by Telephone. New York: Academic Press, 1979 Silverstein MD. Smoking habits of ambulatory patients at a university teaching hospital. J Smoking Rel Dis 1992; 3(No. 3):231-239 Schlesselman JJ. Case Control Studies: Design, Conduct, Analysis. New York: Oxford University Press, 1982 Breslow NE, Day NE. Statistical Methods in Cancer Research. Lyon (France): International Agency for Research on Cancer, 1980 Silverstein MD, Lashner BA, Hanauer SB, Evans AA, Kirsner JB. Cigarette smoking in Crohn's disease. Am J Gastroenterol 1989;84:31-33 Calkins BM. A meta-analysis of the role of smoking in inflammatory bowel disease. Dig Dis Sei 1989; 34:18411854 Motley RJ, Rhodes J, Williams G, Tavares IA, Bennett A. Smoking, eicosanoids and ulcerative colitis. J Pharm Pharmacol 1990;42:288-289 Srivastava ED, Russell MAH, Feyerabend C, Rhodes J. Effect of ulcerative colitis and smoking on rectal blood flow. Gut 1990;31:1021-1024 Cope GF, Heatley RV, Kelleher J. Smoking and colonie mucus in ulcerative colitis. BMJ 1986; 293:481 Lashner BA, Hanauer SB, Silverstein MD. Testing nicotine gum for ulcerative colitis patients: experience with singlepatient trials. Dig Dis Sei 1990;35:827-832