Trentham-Dietz, Amy*†; Nichols, Hazel B.*; Egan, Kathleen M.‡; Titus-Ernstoff, Linda§; Hampton, John M.*; Newcomb, Polly A.¶
Cigarette smoking is consistently associated with most malignancies except breast cancer.1,2 Smoking behavior has generally been explored in relation to invasive breast cancer, with modest effects observed according to age at initiation, duration, and cessation of smoking. These subtle changes in risk may be due, in part, to the joint antiestrogenic and carcinogenic properties of cigarette smoke and underscore the importance of accurately capturing the window of exposure.3–6
Since the advent of population mammographic screening in the 1980s, the incidence of breast carcinoma in situ has increased more than 7-fold. This increase has been observed predominately among women age 50 years and older who follow the recommended screening guidelines.7 An estimated 62,000 cases of breast carcinoma in situ will be diagnosed in the United States in 2007.8 Ductal carcinoma in situ, in particular, is often identified through the use of mammograms,7 whereas lobular carcinoma in situ may more likely be identified as an incidental finding.
Although survival rates exceed 95%,9 breast carcinoma in situ has clear links to invasive disease. Women with a history of breast carcinoma in situ have a 2- to 9-fold excess risk of developing invasive breast cancer.10,11 Women with breast carcinoma in situ may suffer substantial worry and concern after their diagnosis, and many experience a second breast cancer diagnosis despite aggressive treatments.12,13 Identification of factors related to incidence is therefore of significant public health interest.
Although the associations with cigarette smoking have been explored extensively for invasive breast cancer,4–6 the relation to in situ cancer has less frequently been studied. To further investigate these associations, we examined data from a population-based case-control study.
This analysis was performed using data from the Three State Study, a population-based case-control study conducted at the University of Wisconsin, Madison, Wisconsin; the Harvard School of Public Health, Boston, Massachusetts; and Dartmouth University, Lebanon, New Hampshire. The study was conducted according to institutionally approved protocols at each study site.
Selection of Cases
Women ages 20–74 years residing in Wisconsin, Massachusetts (excluding metropolitan Boston), and New Hampshire with a new diagnosis of breast carcinoma in situ (ICD-O version 2 C50.0-C50.9) reported to each state's cancer registry during 1997–2001 were eligible for this study. The physician of record for each eligible case was contacted by mail and provided an opportunity to actively refuse participation on the case patient's behalf. Interviews were conducted from February 1997 to May 2001. Eligibility was limited to case subjects with listed telephone numbers, driver's licenses verified by self-report (if less than 65 years of age), and known dates of diagnosis. A total of 2269 cases were eligible for the study. Of these cases, physicians refused contact with 58 (2.6%), 17 (0.7%) were deceased, 63 (2.8%) could not be located, and 244 (10.8%) refused to participate. Overall, 1887 (83%) women were interviewed. Participation was similar between states. Nine interviewed cases were considered unreliable by the interviewers, leaving 1878 cases available for analysis. Of the interviewed cases, 99% were confirmed by histology, cytology, or other means according to the registry reports. For this study, in situ cases were distinguished according to the fifth digit behavior code (in situ = 2) of the morphology code.14 Subtypes of breast carcinoma in situ were defined as lobular morphology (8520) and ductal/nonlobular (8500, 8501, 8503, 8504, 8010, and 8140).14 Histologic classification (ductal or lobular) was not available for 10% (n = 184) of cases. Of the remaining 1694 cases, 87% (n = 1471) were ductal/nonlobular and 13% (n = 223) were lobular.
Selection of Controls
Community controls were randomly selected during 1997–2001 in each state using 2 sampling frames: those under 65 years of age were selected from lists of licensed drivers, and those 65 to 74 years of age were selected from a roster of Medicare beneficiaries compiled by the Centers for Medicare & Medicaid Services. Controls were selected at random within 5-year age strata to yield an age distribution similar to the cases enrolled in each state. Controls were required to have no personal history of breast cancer, and a listed telephone number. Of the 10,690 potential controls approached for participation, 86 (0.8%) were deceased, 475 (4.4%) could not be located, and 2074 (19.4%) refused to participate. Interviews were obtained for 8055 (75%) of these women. Fourteen interviewed controls were considered unreliable by the interviewers; hence, 8041 controls were available for analysis.
Case subjects and controls were sent letters briefly describing the study before they were contacted by trained telephone interviewers. Interviews for case and control participants were conducted contemporaneously. The 40-minute interview elicited detailed information on smoking, screening mammography, reproductive/menstrual history, use of exogenous hormones (oral contraceptives and postmenopausal therapy), recreational physical activity, occupational history, alcohol consumption, height and weight, personal and family medical history, and demographic factors. Information about the woman's personal and family history of cancer was obtained at the end of the interview to maintain blinding. For 95% of cases and 93% of controls, the interviewers reported being unaware of the woman's case-control status until the end of the interview.
Participants who indicated during the telephone interview that they had smoked more than 100 cigarettes before the referent date were classified as smokers. Those who had not smoked more than 100 cigarettes were classified as never-smokers. We categorized study participants who reported smoking during the year preceding the referent date as “current smokers.” Duration of smoking was defined as the number of years participants reported smoking regularly. Pack-years were calculated as the average number of packs of cigarettes smoked per day times the smoking duration in years. We created categories of years of smoking, pack-years and age of smoking initiation based on approximate quartiles among the controls. Categories of cigarettes per day and years since smoking cessation were created in 10-unit intervals.
We defined a woman as postmenopausal if she reported a natural menopause (no menstrual periods for at least 6 months) before the referent date. Women who reported taking postmenopausal hormones and still having periods were classified as (1) premenopausal if their referent ages were in the first decile of age at natural menopause among the controls (<41 years of age for current smokers and <43 years of age for nonsmokers), (2) postmenopausal if their referent ages were in the highest decile for age at natural menopause in the control group (≥54 years of age for current smokers and ≥56 years of age for nonsmokers) with age at menopause defined as unknown, and (3) unknown menopausal status if at intermediate ages (second to ninth decile). Menopausal status was unknown for 527 controls (7%) and 132 cases (7%); reasons included report of a hysterectomy but not bilateral oophorectomy (n = 361) or current use of postmenopausal hormones (n = 198) among women age 43–55 (age 41–53 for current smokers), reports of being unsure whether a menstrual period had occurred within 6 months (n = 68), and reports of unspecified female surgery (n = 28) or never having periods (n = 4). We evaluated alcohol use as frequency of beer, red wine, white wine, or liquor consumption during the year previous to the referent date. Weight gain (or loss) was determined by subtracting each participant's weight at age 18 from her weight one year before the referent date (“recent weight”). Body mass index (BMI) was calculated as recent weight (kg)/tallest adult height (m).2
For each case, a referent date was defined as the registry-supplied date of breast carcinoma in situ diagnosis. For comparability, the control subjects were assigned a referent date that corresponded to the average diagnosis date for cases of similar ages (within 5-year strata by state). Referent age was defined as the woman's age at the referent date. Only exposures that occurred before the assigned referent date were included in analyses.
We considered the following potential confounders: age at menarche, age at first full-term pregnancy, parity, menopausal status, age at menopause, oral contraceptive and postmenopausal hormone use, family history of breast cancer, education, body mass index, weight change since age 18, alcohol consumption, personal history of benign breast disease, and mammography use. Variables that were associated with risk of breast carcinoma in situ (P ≤ 0.05 in age and state-adjusted models) were included in multivariate models. Our study population was 94.8% white (1799 cases and 7605 controls); thus we did not evaluate breast carcinoma in situ risk associations by race or ethnicity.
Multivariate models were adjusted for age (<40, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, ≥70 years), state (MA, NH, WI), age at menarche (<12, 12, 13, ≥14, unknown), age at first birth (<20, 20–24, 25–29, ≥30, unknown), parity (≤1, 2, ≥3, unknown), menopausal status (premenopausal, postmenopausal, unknown), age at menopause (<45, 45–49, 50–54, ≥55, unknown), oral contraceptive use (never, ever), postmenopausal hormone use (never, former, current), family history of breast cancer (yes, no, unknown), education (less than high school diploma, high school diploma, some college, college diploma, unknown), weight at age 18 (continuous), height (continuous), weight change since age 18 (weight loss, weight gains of 0–15 lbs, 16–30 lbs, 31–50 lbs, >50 lbs, unknown), personal history of benign breast disease (yes, no, unknown), alcohol consumption (0 drinks per week, <7, 7–13, ≥14, unknown), and number of mammograms within 5 years before the referent date (none, less than five, 5 or more, unknown). Although age at menarche and alcohol consumption were not associated with breast carcinoma in situ in preliminary models, these 2 variables were included in multivariate models as decided a priori.
Odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer were produced using multivariate logistic regression models15 adjusted for the covariates described above. To obtain P-values for trend, we included select variables as continuous linear terms in regression models. Effect modification was evaluated by inclusion of cross-product interaction terms in logistic models and measuring the change in the log-likelihood using χ2 tests. The analyses described above were performed using SAS version 9.1 software (SAS Institute, Cary, NC).
Table 1 presents the distribution of age at diagnosis, histologic type, and the method of detection for breast carcinoma in situ cases. As expected, the majority (77.5%; n = 1456) of breast carcinoma in situ cases were discovered by screening mammogram.
In multivariate models, ORs for breast carcinoma in situ increased with older age at first birth and menopause, low parity, oral contraceptive and postmenopausal hormone use, a personal history of benign breast disease, greater frequency of mammography screening, a family history of breast cancer, and (among postmenopausal women) weight gain since age 18 (Table 2). Compared with premenopausal women, postmenopausal women had a lower odds of breast carcinoma in situ. Odds did not appear to be strongly associated with age at menarche, education, or alcohol consumption.
Among women in our study, approximately half reported ever smoking (49% of cases; 52% of controls). Of these, 68% of cases and 59% of controls were categorized as former smokers and 32% of cases and 41% of controls as current smokers. Smoking history was strongly related to frequency of mammography among control women: annual screening examinations were reported by 47% of former smokers, 42% of never-smokers, and 32.2% of current smokers. Current smokers (22%) were twice as likely as former smokers (11%) to report no screening mammograms in this period, with intermediate rates in never-smokers (14%; P for trend <0.001).
In the adjusted models, we observed a lower odds ratio for breast carcinoma in situ (OR = 0.82; 95% CI = 0.70−0.96) among current, but not former (1.00; 0.88–1.13) smokers, compared with never-smokers (Table 3). We did not observe a clear pattern of risk associated with number of cigarettes per day, duration of smoking in years, or pack-years of smoking. Odds ratios were less than 1.0 among women who initiated smoking at younger ages. Women who started smoking at age 18 or younger had 0.81 times the risk compared with never-smokers. Risk increased towards the null with greater time since smoking cessation; women who smoked more recently appeared to be at decreased risk compared with never-smokers. In models restricted to parous women, we also observed lower odds ratios among women who initiated smoking habits after the first birth, compared with never-smokers (0.74; 0.57–0.96). The direction of the association was the same for both premenopausal and postmenopausal women (OR = 0.73, 95% CI = 0.53–1.01 and OR = 0.84, 95% CI = 0.67−1.07, respectively).
Heterogeneity of the relation between cigarette smoking and breast carcinoma in situ was explored according to age as well as the factors shown in Table 2 (data not shown). Among these, interaction was suggested only with frequency of mammogram screening in the 5 years preceding the referent date (χ2 = 16.092, 2 df, P < 0.001 for never/ever smoking versus any mammography and χ2 = 19.673, 4 df, P < 0.001 for never/former/current smoking versus 0/<5/5 mammograms).
Most women reported annual screening mammography prior to the referent date. The association between smoking history for women who were frequently screened was similar to the results overall. In multivariate models restricted to women who reported having annual mammograms in the 5-year period, ORs for breast carcinoma in situ were less than 1.0 among current smokers (OR = 0.70, 95% CI = 0.56−0.89) and former smokers (0.87; 0.74−1.03) compared with never-smokers (Table 4). Among women at least 50 years of age who reported annual mammograms, the odds ratio associated with current smoking was similar in multivariate models, both adjusted (61; 0.46−0.81) and unadjusted (0.72; 0.58−0.89) for pack years of smoking. In the models restricted to women who reported annual mammography, odds ratios were lowest among women who started smoking at younger ages, and who started smoking after their first full-term pregnancy. However, these trends were not stable nor was there a clear dose-response pattern. No consistent associations were observed among women undergoing less frequent mammogram screening during this period (OR = 0.82, 95% 0.62−1.09 for current smoking; OR = 1.20, 95% CI 0.96−1.51 for past smoking).
For the relatively few women (n = 196 cases; 1191 controls) reporting no mammograms in the preceding 5 years, we observed ORs greater than 1.0 for former (1.69 [1.13–2.53]) and current (1.29 [0.85–1.96]) smoking (Table 4). In this group, smoking more than a pack per day and smoking for longer periods were associated with increased risk.
Cigarette smoking has rarely been considered in studies that have sought to determine whether risk factor associations are similar for invasive and in situ breast cancer.16–18 We are aware of only 3 studies that examined breast carcinoma in situ risk according to cigarette smoking. In a 1995 combined analysis of invasive and in situ breast cancer, Brinton et al reported a null association for smoking and breast cancer risk (OR = 0.96; 95% CI = 0.8−1.1); however, only 14% of the 1,648 breast cancer cases were in situ disease.19 In their 1998 report, Gammon et al20 observed a reduced risk of breast carcinoma in situ among current smokers age 45 and younger (0.83; 95% CI = 0.56−1.37); the number of in situ cases was not reported. A 2001 study by Claus et al21 did not detect an association between smoking (ever/never) and risk of ductal cancer in situ (n = 838 cases, 986 controls; OR = 1.01; 95% CI = 0.82−1.26) or lobular cancer in situ (n= 123 cases, 986 controls; OR = 1.02; CI = 0.58−1.80). The studies by Brinton and Gammon et al did not evaluate mammographic screening; Claus’ findings adjusted for history of at least one screening mammogram in the 5-year period 1 year before the referent date.
The relation of cigarette smoking with invasive breast cancer is complex; changes in risk associated with initiation, duration, and cessation of smoking have been reported.4–6 Cigarette smoking is thought to exert conflicting forces on breast cancer risk by applying both antiestrogenic and carcinogenic effects. Within the invasive breast cancer literature, studies have reported an increase in breast cancer risk associated with initiating smoking habits at an early age,22,23 before the first birth,6,23 and among premenopausal women in particular.22 In a 2000 report, Manjer et al24 observed an inverse relationship between invasive breast cancer risk and greater time since smoking cessation. In our study, risk estimates for breast carcinoma in situ increased toward the null with greater time since smoking cessation; however, this relation was less apparent after restricting models to women with recent annual mammograms. Additionally, women who started smoking at younger ages or after a first birth had the lowest odds ratios for breast carcinoma in situ, compared with never-smokers.
Differences in the observed relations suggest that selection bias may have influenced the association between smoking and risk of breast carcinoma in situ. Smoking may be associated with both the frequency of health screening behaviors and the willingness to participate in health surveys. National surveys have demonstrated that current smokers are half as likely to take advantage of cancer screening such as mammography.25,26 In the current study, control women who smoked cigarettes were substantially less likely to receive routine mammography than former and never-smokers. The diagnosis of breast carcinoma in situ is more common in women with higher socioeconomic status.27 Differential response rates according to smoking history and breast screening may have distorted associations of smoking and breast carcinoma in situ risk in the current study. Although participation rates in both cases and controls were relatively high, and our risk estimates were adjusted for several factors that are often correlated with health-seeking behavior (such as educational level and postmenopausal hormone therapy), the results may still be influenced by residual bias.
Previous research has suggested that studies of breast carcinoma in situ should most appropriately use data from women who have undergone mammographic screening to address this concern.28,29 The presence of undiagnosed breast carcinoma in situ within the control group potentially could bias associations towards the null. Because odds ratio estimates among unscreened women were greater than 1.0, the influence of selection bias due under-participation from control women who smoke is likely greater than any bias due to undiagnosed disease among controls. The lack of a detectable dose-response according to the amount, duration, or recency of smoking among women who reported annual mammograms also reinforces concerns that the data were affected by bias.
Chance or uncontrolled confounding as an explanation for our findings cannot be ruled out. Alternatively, it may be possible that the long-term carcinogenic effects of smoking early in life dominate in relation to invasive breast cancer risk, while the more immediate antiestrogenic effects of current or recent smoking dominate the relation with risk of in situ breast cancer. The window of exposure is likely important; full differentiation of breast tissue occurs following the first full-term pregnancy and lactation, and may result in a reduced susceptibility to tumor initiating effects.30 Researchers have theorized that the carcinogenic effects of cigarette smoke are maximized in premenopausal women who initiate smoking around menarche and before the first birth (a period of high endogenous estrogen when antiestrogenic effects associated with cigarette smoking would be relatively weak), while antiestrogenic effects would be most pronounced among postmenopausal women who commenced smoking after the first full-term birth, especially those who have gained weight in adulthood. In postmenopausal women, endogenous estrogen levels are in part derived from androgens in adipose tissue, and may be more sensitive to antiestrogenic effects of cigarette smoke.31 In our data, we observed lower odds ratios for breast carcinoma in situ among women who started smoking before age 19, or during a period of potential high susceptibility for undifferentiated breast tissue. However, among both postmenopausal and premenopausal women, odds ratios for breast cancer were also lower among women who initiated smoking after their first full-term birth, relative to parous never-smokers.
The antiestrogenic effect attributed to cigarette smoking may vary according to exogenous hormone use, menopausal status, or body mass index.6,22 We did not observe effect modification by these factors within our study population. However, our results did suggest that the association between risk of breast carcinoma in situ and smoking status depended on mammography screening behavior. Reductions in breast carcinoma in situ risk associated with smoking status were limited to women who received a mammogram in the 5-year period 1 year before breast cancer diagnosis in the cases or comparable period in controls. In contrast, former smoking was associated with elevated odds ratios among women reporting no mammograms during this period. As noted above, selection bias may have influenced the results, especially among women who did not report any recent mammograms. In our study, case women were more likely than controls to report annual mammograms in the 5 years before diagnosis. This is consistent with the evidence that breast carcinoma in situ is usually diagnosed through mammography (78% of cases in the current study), and women undergoing regular breast screening would be more likely to have a prevalent cancer diagnosed.
Women ages 20–74 years participated in our study; during this period screening was advised only for women ages 50 and older. In some circumstances, women may have included diagnostic mammography when reporting frequency of screening mammograms, which would contribute to an elevated risk associated with the procedure. To ensure that screening mammograms reported by younger women were not a reflection of subclinical disease symptoms, detection of a lump, or a family history, we restricted smoking analyses to women ages 50 and older who had had at least one mammogram in the 5-year period of interest. In this group, the relation of current smoking and decreased risk of breast carcinoma in situ remained unchanged. We also performed analyses of age at smoking initiation, time since cessation, and smoking in relation to the first full term birth both adjusted and unadjusted for pack-years. In the full study population, the pattern of association was unaltered.
Associations observed in this study with factors other than smoking are similar to previously published reports. Our results are consistent regarding increased risk of breast carcinoma in situ associated with older ages at first birth18,21 and at menopause,21 low parity,17,18,21 a personal history of benign breast disease,16–18 a family history of breast cancer,16–18,21 and higher education.17 Similar associations are observed for invasive breast cancer risk. Two recent case-control studies reported that women with breast cancer in situ were less likely to be postmenopausal than similarly-aged controls.29,32 A previous study reported an inverse association between body mass index and risk of breast carcinoma in situ;17 while our findings did not indicate an association with body mass index, we did observe an increase in situ cancer risk in relation to weight gain. We observed a null association for risk of breast cancer in situ in relation to alcohol consumption as have other investigators.17,21,33
Our study findings are strengthened by the large study population and the use of standardized instruments, including self-reported mammographic screening information. In addition, we were able to categorize a wide range of smoking measures6 and adjust for known breast cancer risk factors. However, some studies have suggested that not accounting for passive smoking could weaken the assumed positive association between (invasive) breast cancer and smoking,22 and we were unable to account for exposure to passive cigarette smoke. Recall bias is an inherent concern with the use of self-reported, retrospective data. Reassuringly, self-reported lifetime smoking has been highly reproducible in other studies and we have previously established the strong reliability of covariate data from our survey instrument.34–37
In conclusion, our findings suggest that current smoking is inversely associated with risk of breast carcinoma in situ. This association was restricted to women who reported frequent screening mammography. The interaction with mammography frequency reaffirms the importance of capturing screening information in studies of breast carcinoma in situ to adequately address the potential for selection bias and the necessity of an appropriate comparison group.
We thank Henry Anderson, Patrick L. Remington, Meir J. Stampfer, Walter C. Willett, John A. Baron, and E. Robert Greenberg; Laura Stephenson and the staff of the Wisconsin Cancer Reporting System; Susan T. Gershman and the staff of the Massachusetts Tumor Registry; Marguerite Stevens and the staff of the New Hampshire Cancer Registry; and Mary Pankratz, Linda Haskins, Jerry Phipps, Heidi Judge, Laura Mignone, and Shafika Abrahams-Gessel, along with the study interviewers in all 3 states for assistance with data collection. We are especially grateful to the study participants, whose generosity made this research possible.
1. Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst. 1981;66:1191–1308.
2. Women and smoking: a report of the Surgeon General. Executive summary. MMWR Recomm Rep. 2002;51(RR-12):i–iv;1–13.
3. Baron JA, La Vecchia C, Levi F. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol. 1990;162:502–514.
4. Hamajima N, Hirose K, Tajima K, et al. Alcohol, tobacco and breast cancer–collaborative reanalysis of individual data from 53 epidemiological studies, including 58,515 women with breast cancer and 95,067 women without the disease. Br J Cancer. 2002;87:1234–1245.
5. Morabia A. Smoking (active and passive) and breast cancer: epidemiologic evidence up to June 2001. Environ Mol Mutagen. 2002;39:89–95.
6. Terry PD, Rohan TE. Cigarette smoking and the risk of breast cancer in women: a review of the literature. Cancer Epidemiol Biomarkers Prev. 2002;11:953–971.
7. Breast Cancer Facts & Figures 2005–2006. Atlanta: American Cancer Society, Inc.; 2006.
8. Cancer Facts & Figures 2007. Atlanta: American Cancer Society; 2007.
9. Ernster VL, Barclay J, Kerlikowske K, et al. Mortality among women with ductal carcinoma in situ of the breast in the population-based surveillance, epidemiology and end results program. Arch Intern Med. 2000;160:953–958.
10. Warnberg F, Yuen J, Holmberg L. Risk of subsequent invasive breast cancer after breast carcinoma in situ. Lancet. 2000;355:724–725.
11. Li CI, Malone KE, Saltzman BS, Daling JR. Risk of invasive breast carcinoma among women diagnosed with ductal carcinoma in situ and lobular carcinoma in situ, 1988–2001. Cancer. 2006;106:2104–2112.
12. Ernster VL, Barclay J. Increases in ductal carcinoma in situ (DCIS) of the breast in relation to mammography: a dilemma. J Natl Cancer Inst Monogr. 1997;151–156.
13. Burstein HJ, Polyak K, Wong JS, Lester SC, Kaelin CM. Ductal carcinoma in situ of the breast. N Engl J Med. 2004;350:1430–1441.
14. Percy C, Van Holten V, Muir C. International Classification of Diseases for Oncology. 2 ed. Geneva, Switzerland: The World Health Organization; 1990.
15. Breslow NE, Day NE. Statistical methods in cancer research. Vol. 1. The analysis of case-control studies. Lyon: IARC Scientific Publication; 1980;32.
16. Longnecker MP, Bernstein L, Paganini-Hill A, et al. Risk factors for in situ breast cancer. Cancer Epidemiol Biomarkers Prev. 1996;5:961–965.
17. Weiss HA, Brinton LA, Brogan D, et al. Epidemiology of in situ and invasive breast cancer in women aged under 45. Br J Cancer. 1996;73:1298–1305.
18. Trentham-Dietz A, Newcomb PA, Storer BE, et al. Risk factors for carcinoma in situ of the breast. Cancer Epidemiol Biomarkers Prev. 2000;9:697–703.
19. Brinton LA, Daling JR, Liff JM, et al. Oral contraceptives and breast cancer risk among younger women. J Natl Cancer Inst. 1995;87:827–835.
20. Gammon MD, Schoenberg JB, Teitelbaum SL, et al. Cigarette smoking and breast cancer risk among young women (United States). Cancer Causes Control. 1998;9:583–590.
21. Claus EB, Stowe M, Carter D. Breast carcinoma in situ: risk factors and screening patterns. J Natl Cancer Inst. 2001;93:1811–1817.
22. Khuder SA, Mutgi AB, Nugent S. Smoking and breast cancer: a meta-analysis. Rev Environ Health. 2001;16:253–261.
23. Gram IT, Braaten T, Terry PD, et al. Breast cancer risk among women who start smoking as teenagers. Cancer Epidemiol Biomarkers Prev. 2005;14:61–66.
24. Manjer J, Berglund G, Bondesson L, et al. Breast cancer incidence in relation to smoking cessation. Breast Cancer Res Treat. 2000;61:121–129.
25. Rakowski W, Clark MA, Truchil R, et al. Smoking status and mammography among women aged 50–75 in the 2002 behavioral risk factor surveillance system. Women Health. 2005;41:1–21.
26. Rakowski W, Meissner H, Vernon SW, et al. Correlates of repeat and recent mammography for women ages 45 to 75 in the 2002 to 2003 Health Information National Trends Survey (HINTS 2003). Cancer Epidemiol Biomarkers Prev. 2006;15:2093–2101.
27. Dershaw DD, Loring L, True SJ, et al. A comparison of screening mammography results from programs for women of different socioeconomic status. Cancer. 1998;82:1692–1697.
28. Patel AV, Press MF, Meeske K, et al. Lifetime recreational exercise activity and risk of breast carcinoma in situ. Cancer. 2003;98:2161–2169.
29. Meeske K, Press M, Patel A, et al. Impact of reproductive factors and lactation on breast carcinoma in situ risk. Int J Cancer. 2004;110:102–109.
30. Russo J, Russo IH. Toward a physiological approach to breast cancer prevention. Cancer Epidemiol Biomarkers Prev. 1994;3:353–364.
31. Band PR, Le ND, Fang R, et al. Carcinogenic and endocrine disrupting effects of cigarette smoke and risk of breast cancer. Lancet. 2002;360:1044–1049.
32. Gill JK, Press MF, Patel AV, et al. Oral contraceptive use and risk of breast carcinoma in situ (United States). Cancer Causes Control. 2006;17:1155–1162.
33. Terry MB, Zhang FF, Kabat G, et al. Lifetime alcohol intake and breast cancer risk. Ann Epidemiol. 2006;16:230–240.
34. Krall EA, Valadian I, Dwyer JT, Gardner J. Accuracy of recalled smoking data. Am J Public Health. 1989;79:200–202.
35. Slattery ML, Hunt SC, French TK, et al. Validity of cigarette smoking habits in three epidemiologic studies in Utah. Prev Med. 1989;18:11–19.
36. Longnecker MP, Newcomb PA, Mittendorf R, et al. Risk of breast cancer in relation to lifetime alcohol consumption. J Natl Cancer Inst. 1995;87:923–929.
37. Czermak C, Lehofer M, Gasser-Steiner P, et al. Test-retest reliability of a lifetime drug use questionnaire. Addict Behav. 2005;30:361–368.
© 2007 Lippincott Williams & Wilkins, Inc.