PMC:13911 JSONTXT

{"project":"Colil","denotations":[{"id":"T1","span":{"begin":9453,"end":9455},"obj":"9605646"},{"id":"T2","span":{"begin":8772,"end":8773},"obj":"9883788"},{"id":"T3","span":{"begin":8774,"end":8775},"obj":"8610705"},{"id":"T4","span":{"begin":8776,"end":8778},"obj":"9596479"},{"id":"T5","span":{"begin":8779,"end":8781},"obj":"9596478"},{"id":"T6","span":{"begin":8670,"end":8671},"obj":"9883788"},{"id":"T7","span":{"begin":8672,"end":8673},"obj":"8610705"},{"id":"T8","span":{"begin":8674,"end":8675},"obj":"7613275"},{"id":"T9","span":{"begin":8543,"end":8544},"obj":"2178432"},{"id":"T10","span":{"begin":8545,"end":8546},"obj":"8405197"},{"id":"T11","span":{"begin":8547,"end":8548},"obj":"9611614"},{"id":"T12","span":{"begin":8549,"end":8550},"obj":"8877046"},{"id":"T13","span":{"begin":9584,"end":9586},"obj":"9596479"},{"id":"T14","span":{"begin":10625,"end":10627},"obj":"10466767"},{"id":"T15","span":{"begin":15162,"end":15163},"obj":"10189043"},{"id":"T16","span":{"begin":15164,"end":15166},"obj":"8818604"}],"namespaces":[{"prefix":"_base","uri":"http://pubannotation.org/docs/sourcedb/PubMed/sourceid/"}],"text":"Smoking and high-risk mammographic parenchymal patterns: a \t\t case-control study \n\nCurrent smoking was strongly and inversely associated with high-risk \t\t\t patterns, after adjustment for concomitant risk factors. Relative to never \t\t\t smokers, current smokers were significantly less likely to have a high-risk \t\t\t pattern. Similar results were obtained when the analysis was confined to \t\t\t postmenopausal women. Past smoking was not related to the mammographic \t\t\t parenchymal patterns. The overall effect in postmenopausal women lost its \t\t\t significance when adjusted for other risk factors for P2/DY patterns that were \t\t\t found to be significant in the present study, although the results are still \t\t\t strongly suggestive. The present data indicate that adjustment for current \t\t\t smoking status is important when evaluating the relationship between \t\t\t mammographic parenchymal pattern and breast cancer risk. They also indicate \t\t\t that smoking is a prominent potential confounder when analyzing effects of \t\t\t other risk factors such as obesity-related variables. It appears that \t\t\t parenchymal patterns may act as an informative biomarker of the effect of \t\t\t cigarette smoking on breast cancer risk.\n\nAbstract\nIntroduction:\nOverall, epidemiological studies [1,2,3,4] have \t\t\t\treported no substantial association between cigarette smoking and the risk of \t\t\t\tbreast cancer. Some studies [5,6,7] reported a significant increase of \t\t\t\tbreast cancer risk among smokers. In recent studies that addressed the \t\t\t\tassociation between breast cancer and cigarette smoking, however, there was \t\t\t\tsome suggestion of a decreased risk [8,9,10], especially among current smokers, \t\t\t\tranging from approximately 10 to 30% [9,10]. Brunet et al [11] \t\t\t\treported that smoking might reduce the risk of breast cancer by 44% in carriers \t\t\t\tof BRCA1 or BRCA2 gene mutations. Wolfe [12] described four different mammographic patterns created by \t\t\t\tvariations in the relative amounts of fat, epithelial and connective tissue in \t\t\t\tthe breast, designated N1, P1, P2 and DY. Women with either P2 or DY pattern \t\t\t\tare considered at greater risk for breast cancer than those with N1 or P1 \t\t\t\tpattern [12,13,14,15]. There are no published studies \t\t\t\tthat assessed the relationship between smoking and mammographic parenchymal \t\t\t\tpatterns.\n\nAims:\nTo evaluate whether mammographic parenchymal patterns as \t\t\t\tclassified by Wolfe, which have been positively associated with breast cancer \t\t\t\trisk, are affected by smoking. In this case-control study, nested within the \t\t\t\tEuropean Prospective Investigation on Cancer in Norfolk (EPIC-Norfolk) cohort \t\t\t\t[16], the association between smoking habits and \t\t\t\tmammographic parenchymal patterns are examined. The full results will be \t\t\t\tpublished elsewhere.\n\nMethods:\nStudy subjects were members of the EPIC cohort in Norwich who also \t\t\t\tattended the prevalence screening round at the Norwich Breast Screening Centre \t\t\t\tbetween November 1989 and December 1997, and were free of breast cancer at that \t\t\t\tscreening. Cases were defined as women with a P2/DY Wolfe's mammographic \t\t\t\tparenchymal pattern on the prevalence screen mammograms. A total of 203 women \t\t\t\twith P2/DY patterns were identified as cases and were individually matched by \t\t\t\tdate of birth (within 1 year) and date of prevalence screening (within 3 \t\t\t\tmonths) with 203 women with N1/P1 patterns who served as control \t\t\t\tindividuals.\nTwo views, the mediolateral and craniocaudal mammograms, of both \t\t\t\tbreasts were independently reviewed by two of the authors (ES and RW) to \t\t\t\tdetermine the Wolfe mammographic parenchymal pattern.\nConsiderable information on health and lifestyle factors was \t\t\t\tavailable from the EPIC Health and Lifestyle Questionnaire [16]. In the present study we examined the subjects' personal \t\t\t\thistory of benign breast diseases, menstrual and reproductive factors, oral \t\t\t\tcontraception and hormone replacement therapy, smoking, and anthropometric \t\t\t\tinformation such as body mass index and waist:hip ratio.\nOdds ratios (ORs) and their 95% confidence intervals (CIs) were \t\t\t\tcalculated by conditional logistic regression [17], and \t\t\t\twere adjusted for possible confounding factors.\n\nResults:\nThe characteristics of the cases and controls are presented in \t\t\t\tTable 1. Cases were leaner than controls. A larger \t\t\t\tpercentage of cases were nulliparous, premenopausal, current hormone \t\t\t\treplacement therapy users, had a personal history of benign breast diseases, \t\t\t\tand had had a hysterectomy. A larger proportion of controls had more than three \t\t\t\tbirths and were current smokers.\nTable 2 shows the unadjusted and adjusted OR \t\t\t\testimates for Wolfe's high-risk mammographic parenchymal patterns and smoking \t\t\t\tin the total study population and in postmenopausal women separately. Current \t\t\t\tsmoking was strongly and inversely associated with high-risk patterns, after \t\t\t\tadjustment for concomitant risk factors. Relative to never smokers, current \t\t\t\tsmokers were significantly less likely to have a high-risk pattern (OR 0.37, \t\t\t\t95% CI 0.14-0.94). Similar results were obtained when the analysis was confined \t\t\t\tto postmenopausal women. Past smoking was not related to mammographic \t\t\t\tparenchymal patterns. The overall effect in postmenopausal women lost its \t\t\t\tsignificance when adjusted for other risk factors for P2/DY patterns that were \t\t\t\tfound to be significant in the present study, although the results were still \t\t\t\tstrongly suggestive. There was no interaction between cigarette smoking and \t\t\t\tbody mass index.\n\nDiscussion:\nIn the present study we found a strong inverse relationship \t\t\t\tbetween current smoking and high-risk mammographic parenchymal patterns of \t\t\t\tbreast tissue as classified by Wolfe [12]. These \t\t\t\tfindings are not completely unprecedented; Greendale et al [18] found a reduced risk of breast density in association with \t\t\t\tsmoking, although the magnitude of the reduction was unclear. The present \t\t\t\tfindings suggest that this reduction is large.\nRecent studies [9,10] \t\t\t\thave suggested that breast cancer risk may be reduced among current smokers. In \t\t\t\ta multicentre Italian case-control study, Braga et al [10] found that, relative to nonsmokers, current smokers had a \t\t\t\treduced risk of breast cancer (OR 0.84, 95% CI 0.7-1.0). These findings were \t\t\t\trecently supported by Gammon et al [9], who \t\t\t\treported that breast cancer risk in younger women (younger than 45 years) may \t\t\t\tbe reduced among current smokers who began smoking at an early age (OR 0.59, \t\t\t\t95% CI 0.41-0.85 for age 15 years or younger) and among long-term smokers (OR \t\t\t\t0.70, 95% CI 0.52-0.94 for those who had smoked for 21 years or more).\nThe possible protective effect of smoking might be due to its \t\t\t\tanti-oestrogenic effect [1,2,19]. Recently there has been renewed interest in the potential \t\t\t\teffect of smoking on breast cancer risk, and whether individuals may respond \t\t\t\tdifferently on the basis of differences in metabolism of bioproducts of smoking \t\t\t\t[20,21]. Different relationships \t\t\t\tbetween smoking and breast cancer risk have been suggested that are dependent \t\t\t\ton the rapid or slow status of acetylators of aromatic amines [20,21]. More recent studies [22,23], however, do not support these \t\t\t\tfindings.\nThe present study design minimized the opportunity for bias to \t\t\t\tinfluence the findings. Because subjects were unaware of their own case-control \t\t\t\tstatus, the possibility of recall bias in reporting smoking status was \t\t\t\tminimized. Systematic error in the assessment of mammograms was avoided because \t\t\t\treading was done without knowledge of the risk factor data. Furthermore, the \t\t\t\tassociations observed are unlikely to be explained by the confounding effect of \t\t\t\tother known breast cancer risk factors, because we adjusted for these in the \t\t\t\tanalysis. We did not have information on passive smoking status, however, which \t\t\t\thas recently been reported to be a possible confounder [5,6,21,24].\nThe present data indicate that adjustment for current smoking \t\t\t\tstatus is important when evaluating the relationship between mammographic \t\t\t\tparenchymal pattern and breast cancer risk. They also indicate smoking as a \t\t\t\tprominent potential confounder when analyzing effects of other risk factors \t\t\t\tsuch as obesity-related variables. It seems that parenchymal patterns may act \t\t\t\tas an informative biomarker of the effect of cigarette smoking on breast cancer \t\t\t\trisk. \n\nIntroduction\nOverall, epidemiological studies [1,2,3,4] have \t\t reported no substantial association between cigarette smoking and the risk of \t\t breast cancer. Some studies [5,6,7] reported a significant increase of \t\t breast cancer risk among smokers. It has been suggested [5,6,21,24] that passive exposure to cigarette smoking may alter prior \t\t associations seen when only active smoking was assessed, with increased risk \t\t being observed for passive smoking exposure. Furthermore, there is a \t\t possibility of heterogeneity in the response to the carcinogenic effect of \t\t smoking, which might explain inconsistent findings for cigarette smoking as a \t\t risk factor for breast cancer [20].\nIn recent studies that addressed the association between breast cancer \t\t and cigarette smoking, however, there was some suggestion of a decreased risk \t\t [8,9,10], \t\t especially among current smokers, ranging from approximately 10 to 30% [9,10]. Brunet et al [11] reported that smoking might reduce the risk of breast \t\t cancer by 44% in carriers of BRCA1 or BRCA2 gene \t\t mutations.\nWolfe [21] described four different \t\t mammographic patterns that are created by variations in the relative amounts of \t\t fat, epithelial and connective tissue in the breast, designated N1, P1, P2 and \t\t DY. Women with either P2 or DY patterns are considered to be at greater risk \t\t for breast cancer than those with N1 or P1 pattern [12,13,14,15].\nThere are no published studies that assessed the relationship between \t\t smoking and mammographic parenchymal patterns.\nThe aim of the present study was to evaluate whether mammographic \t\t parenchymal patterns as classified by Wolfe [12], which \t\t have been positively associated with breast cancer risk, are affected by \t\t smoking. In the present case-control study, nested within the European \t\t Prospective Investigation on Cancer in Norfolk (EPIC-Norfolk) cohort [16], \t\t the association between smoking habits and mammographic parenchymal patterns \t\t are examined. The full results will be published elsewhere.\n\nMaterials and methods\nStudy subjects were members of the EPIC cohort in Norwich [16], who also attended the prevalence screening round at the \t\t Norwich Breast Screening Centre between November 1989 and December 1997 and \t\t were free of breast cancer at that screening. A case-control study was designed \t\t within this cohort.\nCases were defined as women with a P2/DY Wolfe's mammographic \t\t parenchymal pattern on the prevalence screen mammogram. Assuming a 2.5-fold \t\t increase in risk of P2/DY mammographic patterns from the lowest quintile of a \t\t quantitative factor to the highest, 200 cases and 200 controls will yield a \t\t power of approximately 90%. A total of 203 women with P2/DY patterns were \t\t identified as cases and were individually matched by date of birth (within 1 \t\t year) and date of prevalence screening (within 3 months) to 203 women with \t\t N1/P1 patterns who served as controls. Additional information regarding case \t\t selection is presented elsewhere [25].\nWe examined the screening records of each woman. Mammograms of both \t\t breasts were collected. Two views, the mediolateral and craniocaudal \t\t mammograms, of both breasts were independently reviewed by two of the authors \t\t (ES and RW) to determine the Wolfe mammographic parenchymal pattern. The \t\t inter-reader agreement in the classification of mammographic parenchymal \t\t patterns was 95% on the four pattern categories, and 99% when the P2 and DY \t\t categories were combined, but for the purposes of the present study we used \t\t only the films in which we agreed on the patterns.\nConsiderable information on health and lifestyle factors was available \t\t from the EPIC Health and Lifestyle Questionnaire [16]. \t\t In the present study we examined the subjects' personal and family history of \t\t benign breast diseases and cancer, menstrual and reproductive factors, oral \t\t contraception and hormone replacement therapy, physical activity, smoking, and \t\t anthropometric information such as body mass index and waist:hip ratio.\n\nStatistical methods\nOdds ratios (ORs) and their 95% confidence intervals (CIs) were \t\t\t calculated by conditional logistic regression, which takes into account the \t\t\t matching of controls to cases [17]. Adjustment was \t\t\t performed for those variables that were previously found to be associated with \t\t\t high-risk mammographic parenchymal patterns [25].\n\nResults\nThe characteristics of the cases and controls are presented in Table \t\t 1. The mean age of cases and controls was similar \t\t (because they were matched on date of birth). Cases were leaner than controls. \t\t A larger percentage of cases were nulliparous, similar proportions of cases and \t\t controls had between one and three births, and a larger proportion of controls \t\t had more than three births. A larger proportion of cases were pre-menopausal, \t\t current hormone replacement therapy users, had a personal history of benign \t\t breast diseases, and had had a hysterectomy, whereas a larger proportion of \t\t controls were current smokers. The cases and controls were similar with respect \t\t to age at menarche and age at menopause.\nTable 2 shows the unadjusted and adjusted OR \t\t estimates for Wolfe's high-risk mammographic parenchymal patterns and smoking \t\t in the total study population and in post-menopausal women separately. Current \t\t smoking was strongly and inversely associated with high-risk patterns, after \t\t adjustment for concomitant risk factors. Relative to never smokers, current \t\t smokers were significantly less likely to have a high-risk pattern (OR 0.37, \t\t 95% CI 0.14-0.94). Similar results were obtained when the analysis was confined \t\t to postmenopausal women. Past smoking was not related to the mammographic \t\t parenchymal patterns. The overall effect in postmenopausal women lost its \t\t statistical significance when adjusted for other risk factors for P2/DY \t\t patterns that were found to be significant in this study, although the results \t\t are still strongly suggestive. There was no interaction between cigarette \t\t smoking and body mass index (P =0.73 and 0.72 in the whole study \t\t population and in postmenopausal women, respectively).\n\nDiscussion\nIn the present study, we found a strong inverse relationship between \t\t current smoking and mammographic parenchymal patterns of breast tissue as \t\t classified by Wolfe [12]. These findings are not \t\t completely unprecedented; Greendale et al [18] \t\t found a reduced risk of breast density in association with smoking, although \t\t the magnitude of the reduction was unclear. Our findings suggest that this \t\t reduction is large.\nRecent studies [9,10] \t\t suggest that breast cancer risk may be reduced among current smokers. In a \t\t multicentre Italian case-control study, Braga et al [10] found that, relative to nonsmokers, current smokers had a \t\t reduced risk of breast cancer (OR 0.84, 95% CI 0.7-1.0). These findings were \t\t recently supported by Gammon et al [9], who \t\t reported that breast cancer risk in younger women (younger than 45 years) may \t\t be reduced among current smokers who began smoking at an early age (OR 0.59, \t\t 95% CI 0.41-0.85 for age 15 years or younger) and among long-term smokers (OR \t\t 0.70, 95% CI 0.52-0.94 for those who had smoked for 21 years or longer).\nThe possible protective effect might be due to the anti-oestrogenic \t\t effect of smoking [1,2,19]. Exposure to cigarette smoking causes an earlier menopause \t\t [1,26]. Smoking appears to alter \t\t the metabolism of oestradiol leading to enhanced formation of the inactive \t\t catechol estrogens [1]. Furthermore, smoking increases \t\t circulating androgens through adrenal cortical stimulation [2], but the conversion rates of androgens to oestrogens are \t\t lower in those who smoke [27]. There has been a recent \t\t resurgence of interest in the potential effect of smoking on breast cancer \t\t risk, and whether individuals may respond differently on the basis of \t\t differences in metabolism of bioproducts of smoking [20,21]. Different relationships between \t\t smoking and breast cancer risk have been suggested that are dependent on the \t\t rapid or slow status of acetylators of aromatic amines [20,21], rapid acetylators being better \t\t able to inactivate the potential carcinogenic tobacco compounds. More recent \t\t studies [22,23] do not support \t\t these findings, however.\nThe present study design minimized the opportunity for bias to \t\t influence the findings. Systematic error in the assessment of mammograms was \t\t avoided because reading was done without knowledge of the risk factor data. \t\t Because subjects were unaware of their own case-control status, the possibility \t\t of recall bias in reporting smoking status was minimized. Furthermore, the \t\t associations observed are unlikely to be explained by the confounding effect of \t\t other known breast cancer risk factors, because we adjusted for these in the \t\t analysis. We did not have information on passive smoking status, however, which \t\t has recently been reported as a possible confounder [5,6,21,24].\nAlthough, ideally we would have liked to evaluate the relationship \t\t between intensity and duration of smoking and mammographic parenchymal patterns \t\t among current smokers, the numbers were too small to perform the analysis. \t\t Trends for intensity and duration of smoking were not monotonic, and P \t\t values were inconclusive (between 0.05 and 0.1). Age at menopause and time \t\t since menopause were not related to mammographic patterns in the present study \t\t (data not shown). Although current smokers were likely to have an early \t\t menopause (70% of current smokers were postmenopausal before age 50 years), \t\t there was no difference among mean age at menopause in the three smoking \t\t categories (P = 0.15). There was no difference in time since menopause \t\t among current smokers.\nThese data indicate that adjustment for current smoking status is \t\t important when evaluating the relationship between mammographic parenchymal \t\t patterns and breast cancer risk. They also indicate smoking to be a prominent \t\t potential confounder when analyzing effects of other risk factors, such as \t\t obesity-related variables. It appears that parenchymal patterns may act as an \t\t informative biomarker of the effect of cigarette smoking on breast cancer \t\t risk. "}