Cigarette Smoking and Increased Risk of Mucinous Epithelial Ovarian Cancer

Transcription

1 American Journal of Epidemiology Copyright 2004 by the Johns Hopkins Bloomberg School of Public Health All rights reserved Vol. 159, No. 2 Printed in U.S.A. DOI: /aje/kwh015 Cigarette Smoking and Increased Risk of Mucinous Epithelial Ovarian Cancer Yuqing Zhang 1, Patricia F. Coogan 2, Julie R. Palmer 2, Brian L. Strom 3, and Lynn Rosenberg 2 1 Clinical Epidemiology Research and Training Unit, Boston University School of Medicine, Boston, MA. 2 Slone Epidemiology Center, Boston University, Boston, MA. 3 Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, and Division of General Internal Medicine, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA. Received for publication May 7, 2003; accepted for publication July 21, Several studies have reported that cigarette smoking is associated with an increased risk of mucinous ovarian cancer, but other studies have failed to find such a relation. Using data from the Case-Control Surveillance Study, begun in four US cities in 1976, the authors conducted a case-control study ( ) to examine the association between cigarette smoking and the risk of ovarian cancer of different cell types. Among 709 incident cases of epithelial ovarian cancer, 402 were serous, 74 were mucinous, 106 were endometrioid, and 127 were of other cell types. For mucinous ovarian cancer, the odds ratios were 1.5 (95% confidence interval (CI): 0.7, 3.4) among women who smoked less than one pack of cigarettes per day, 1.4 (95% CI: 0.6, 3.5) among women who smoked one pack per day, and 2.9 (95% CI: 1.2, 7.5) among women who smoked more than one pack per day, relative to never smokers. The odds ratios were 2.5 (95% CI: 1.1, 5.4) for ex-smokers and 1.4 (95% CI: 0.7, 2.9) for current smokers. While women with up to 15 pack-years of smoking had an almost 2.5 times increased risk of mucinous ovarian cancer, such an increased risk was not found among those with more than 15 pack-years of smoking. There was no association between cigarette smoking and epithelial ovarian cancer of other cell types. Despite inconsistencies in the data, these results strengthen the evidence that cigarette smoking may play a role in the development of mucinous ovarian cancer but not ovarian cancer of other cell types. case-control studies; neoplasms by histologic type; ovarian neoplasms; smoking Abbreviation: CI, confidence interval. Ovarian cancer is one of the most common malignancies among women: It accounts for nearly 4 percent of all cancers and ranks second among gynecologic cancers (1). The American Cancer Society estimates that approximately 25,000 new cases of ovarian cancer are diagnosed every year in the United States. To date, few risk factors have been identified. With the exception of oral contraceptive use, risk factors (such as age, parity, and family history of ovarian cancer) are difficult to modify or are nonmodifiable (2). Ten to twenty percent of epithelial ovarian cancers are of the mucinous type, a histologic type that often indicates a poor prognosis if the disease is advanced. In the late 1980s, Koch (3) reported that the percentage of smokers was significantly higher among women with mucinous ovarian cancer than among those with ovarian cancer of other histologic types. Recently, three other studies provided additional evidence that cigarette smoking is associated with increased risk of mucinous ovarian cancer (4 6), but a populationbased case-control study found no association (7). Therefore, we examined cigarette smoking in relation to different histologic types of epithelial ovarian cancer using data from the Case-Control Surveillance Study. MATERIALS AND METHODS The Case-Control Surveillance Study has been conducted in hospitals in four US cities (Boston, Massachusetts; New York, New York; Philadelphia, Pennsylvania; and Baltimore, Maryland) since 1976 and is still in progress. Briefly, subjects were interviewed in the hospital by trained nurse interviewers who used a structured questionnaire to collect information on demographic factors, medical and reproduc- Correspondence to Dr. Yuqing Zhang, Clinical Epidemiology Research and Training Unit, Room A203, Boston University School of Medicine, 715 Albany Street, Boston, MA ( yuqing@bu.edu). 133

2 134 Zhang et al. tive history, family history of cancer, and several lifestyle factors. Until 1997, the study included adult patients up to 69 years of age; after 1997, the upper age limit was raised to 79 years. Over the course of the study, different diagnoses were given priority for interview. Each nurse interviewer generally covered several hospitals and would spend specific days at each one. The nurse interviewers identified potentially eligible patients through examination of admission lists and ward logs. The proportion of hospitalized patients with a particular diagnosis who were enrolled in the Case-Control Surveillance Study varied, depending on whether the diagnosis was being given priority, the level of staffing, and whether the patient s physician permitted his/her patients to be included in the study. The major reason for failure to enroll eligible patients was that they were unavailable because they were undergoing tests or treatments or had visitors. To guard against potential selection bias from referrals to the hospital, study investigators enrolled only patients who lived in areas that were within an hour s drive of the hospital; to ensure that this criterion was met, the nurse interviewers were supplied with lists of acceptable zip codes. The participation rate among patients targeted for interview was 95 percent. Seven previous papers on risk factors for ovarian cancer have been based on Case-Control Surveillance Study data (8 14). A detailed description of the study s methods is available elsewhere (15). For cigarette smoking, subjects were first asked about their current smoking status that is, whether they were a never, past, or current smoker (had smoked within the previous year). Current and past smokers were then asked about number of years of cigarette smoking, number of cigarettes smoked per day, and age at which they had started smoking. Past smokers were asked about the number of years since stopping smoking. The question on age at starting smoking was not asked until After discharge, each patient s diagnosis, including the primary diagnosis that led to hospital admission, was abstracted from the hospital record. The present study included women interviewed through the end of Selection of cases Eligible cases were 731 women aged years who met the following criteria: 1) a diagnosis of epithelial ovarian cancer recorded in the discharge summary or pathology report that had been made within the year before the current admission and 2) no other primary cancer or history of cancer. Women whose tumors were classified as borderline malignancies were not eligible. Case women with epithelial ovarian cancer were classified into four groups based on pathology reports (9): serous cell type, mucinous cell type, endometrioid cell type, and other cell types. The latter group included cases with undifferentiated carcinoma, cases with mixed epithelial adenocarcinoma, and cases with an unknown histologic type. Of the 731 case women, information on smoking was available for 709 (97 percent), and they comprised the final case series. Of the latter cases, 577 (81 percent) had been diagnosed within the preceding 6 months; TABLE 1. Characteristics (%) of women with epithelial ovarian cancer and controls, Case-Control Surveillance Study, the median age at cancer diagnosis was 51 years, and 136 case women (19 percent) were under 40 years of age. Selection of controls Cases (n = 709) Controls (n = 951) Age (years) < Race White Black Other Education (years) < Age (years) at menarche < Menopausal status Premenopausal Postmenopausal Parity Oral contraceptive use No Yes Estrogen replacement therapy No Yes Controls were selected from a pool of 2,223 women aged years with no history of cancer or bilateral oophorectomy who had been admitted to the hospital for diseases that we considered unassociated with cigarette smoking. Eligible diagnoses included appendicitis, hernia of the abdominal cavity, and traumatic injury, other than osteoporotic fractures. Subjects with missing information on smoking status were not eligible as potential controls. The median age was 41 years, and 48 percent of subjects were under age 40.

3 Cigarette Smoking and Ovarian Cancer 135 TABLE 2. Relation of smoking status to the risk of ovarian cancer, by histologic type, Case-Control Surveillance Study, Histologic type and smoking status Serous cases controls * OR, odds ratio; CI, confidence interval. Adjusted for race, education, body mass index (weight (kg)/height (m) 2 ), age at menarche, menopausal status, parity, oral contraceptive use, and postmenopausal hormone use. Reference category. Crude OR* 95% CI* Multivariable adjusted OR Nonsmoker % CI Ex-smoker , , 1.3 Current smoker , , 1.1 Mucinous Nonsmoker Ex-smoker , , 5.4 Current smoker , , 2.9 Endometrioid Nonsmoker Ex-smoker , , 2.1 Current smoker , , 1.7 Other Nonsmoker Ex-smoker , , 1.4 Current smoker , , 2.0 From this pool of controls, we randomly selected for each case up to three controls matched by age (within the same 5- year age category), year of interview (within the same 5-year category), and geographic region (New York City, Boston, Philadelphia, or Baltimore). The final control series comprised 951 women, who were eligible to serve as controls for more than one cell type: 684 for cases of serous ovarian cancer, 195 for cases of mucinous ovarian cancer, 259 for cases of endometrioid ovarian cancer, and 297 for cases of other types of epithelial ovarian cancer. Statistical analysis A woman was considered a nonsmoker if she had never smoked in her life, a current smoker if she had smoked within the year before interview, and an ex-smoker if she had stopped smoking at least 1 year prior to interview. We divided the number of cigarettes smoked per day into three categories: less than one pack per day, one pack per day, and more than one pack per day. We calculated cumulative exposure to cigarettes in pack-years by multiplying the average number of cigarettes smoked per day by the number of years of smoking and dividing by 20. We then divided the packyears into three categories: pack-years, packyears, and more than 15 pack-years. For the smokers, the age at starting smoking was divided into two categories: less than 20 years of age and age 20 years or more. We examined the relation of cigarette smoking to the estimated relative risk (odds ratio) of each histologic type of ovarian cancer using a conditional logistic regression model. In the multivariable regression model, we adjusted for race, education, body mass index (weight (kg)/height (m) 2 ), age at menarche, menopausal status, parity, oral contraceptive use, and postmenopausal hormone use. RESULTS The age-standardized proportions of subjects by smoking status were similar among controls with different disease diagnoses. Among the controls with a diagnosis of appendicitis or hernia of the abdominal cavity, the proportions of never smokers, past smokers, and current smokers were 47 percent, 20 percent, and 33 percent, respectively. The corresponding figures among controls with a diagnosis of traumatic injury were 47 percent, 17 percent, and 36 percent. The two subgroups of controls were also similar in terms of categories of number of cigarettes smoked per day and packyears of cigarette smoking. Most characteristics, including age at interview, educational level, age at menarche, and estrogen use, were similarly distributed between cases and controls (table 1). The proportion of White women was significantly higher among ovarian cancer cases than among controls. While case women were more likely to be postmenopausal, had a lower prevalence of high parity ( 4), and were less likely to have used oral contraceptives than their matched controls, these differences were not statistically significant. The relation of smoking status to the risk of ovarian cancer by histologic type is shown in table 2. The multivariable adjusted odds ratios for mucinous ovarian cancer were 1.4

4 136 Zhang et al. TABLE 3. Relation of average number of cigarettes smoked per day to the risk of ovarian cancer, by histologic type, Case-Control Surveillance Study, Histologic type and number of cigarettes smoked per day Serous cases controls * OR, odds ratio; CI, confidence interval. Adjusted for race, education, body mass index (weight (kg)/height (m) 2 ), age at menarche, menopausal status, parity, oral contraceptive use, and postmenopausal hormone use. Reference category. Crude OR* 95% CI* Multivariable adjusted OR % CI , , , , 1.3 > , , 1.2 Mucinous , , , , 3.5 > , , 7.5 Endometrioid , , , , 1.7 > , , 2.9 Other , , , , 1.7 > , , 1.5 (95 percent confidence interval (CI): 0.7, 2.9) for current smokers and 2.5 (95 percent CI: 1.1, 5.4) for ex-smokers, as compared with never smokers. The relation of smoking status to the risk of mucinous ovarian cancer was little changed when current smoking was redefined as smoking within the previous 5 years. There was no association of current or past smoking with the risk of serous, endometrioid, or other types of ovarian cancer. Data on average number of cigarettes smoked per day in relation to each ovarian cancer cell type are presented in table 3. Compared with never smokers, women who smoked more than one pack per day had almost three times the risk of mucinous ovarian cancer (odds ratio = 2.9, 95 percent CI: 1.2, 7.5). In contrast, average number of cigarettes smoked per day was not associated with the risk of serous, endometrioid, or other types of ovarian cancer. Women with up to 15 pack-years of smoking had an almost 2.5 times increased risk of mucinous ovarian cancer. The odds ratio was close to 1 for those with more than 15 pack-years of smoking (table 4). Data on age at starting smoking were available for approximately 80 percent of the cases. The odds ratio was higher for women who started smoking after age 20 years (odds ratio = 1.9, 95 percent CI: 0.8, 4.7) than for those who started smoking at a younger age (odds ratio = 1.3, 95 percent CI: 0.5, 3.1), but the estimates were compatible with a uniform value. There was no association between age at commencement of smoking and risk of serous, endometrioid, or other types of ovarian cancer. DISCUSSION In this hospital-based case-control study, we confirmed previous findings (4 6) that cigarette smoking is associated with an increased risk of mucinous ovarian cancer. Compared with never smokers, women who smoked more than one pack of cigarettes per day had an almost threefold increased risk of mucinous ovarian cancer. No such association was found for other types of epithelial ovarian cancer. However, our study showed a stronger association of mucinous ovarian cancer with past smoking than with current smoking status, even after we redefined current smokers as those who smoked during the 5 years prior to interview. Furthermore, there was no evidence of a dose-response relation with amount of smoking, as determined by pack-years. These inconsistencies in the data may have been due to chance, given the relatively small number of cases of mucinous ovarian cancer. To our knowledge, at least four studies (4 7) have examined the relation between cigarette smoking and different

5 Cigarette Smoking and Ovarian Cancer 137 TABLE 4. Relation of pack-years of cigarette smoking to the risk of ovarian cancer, by histologic type, Case-Control Surveillance Study, * Cell type and pack-years of cigarette smoking Serous cases controls * Subjects with missing data on pack-years of cigarette smoking were excluded from the analysis. OR, odds ratio; CI, confidence interval. Adjusted for race, education, body mass index (weight (kg)/height (m) 2 ), age at menarche, menopausal status, parity, oral contraceptive use, and postmenopausal hormone use. Reference category. Crude OR 95% CI Multivariable adjusted OR % CI , , , , 1.1 > , , 1.4 Mucinous , , , , 6.2 > , , 2.8 Endometrioid , , , , 1.5 > , , 1.8 Other , , , , 2.6 > , , 1.7 types of epithelial ovarian cancer. Using data from the Cancer and Steroid Hormone Study, Marchbanks et al. (5) reported an odds ratio of 2.9 for mucinous ovarian cancer among current smokers and an odds ratio of 2.3 among ever smokers, relative to never smokers. The risk of mucinous ovarian cancer among current smokers significantly increased as the number of cigarettes smoked increased, regardless of years since first smoking or age at first smoking. No association was found between cigarette smoking and risk of other types of epithelial ovarian cancer. Two subsequent population-based case-control studies showed similar findings (4, 6). However, such an association was not observed in another population-based case-control study (7). In that study, more than 40 percent of case women with ovarian cancer could not be interviewed because of death, an unknown address, or refusal by the patient or her physician, and approximately 30 percent of control women approached also declined to participate. It is possible that nonparticipants, especially those who died, were more likely to have been heavy smokers. If that were the case, the effect of smoking on risk of ovarian cancer may have been diluted. A biologic mechanism for a link between cigarette smoking and risk of mucinous ovarian cancer has not been well established. Scully (16, 17) suggested that mucinous ovarian tumors are histologically similar to cervical and colon epithelial cancer cells, both of which have been associated with cigarette smoking (18 20). In addition, Zenzes et al. (21, 22) found cotinine, a major metabolite of nicotine, and benzo(a)pyrene-dna adducts in the granulosa lutein cells of women exposed to cigarette smoking. Our study had some limitations. First, there will have been some degree of histologic misclassification, since the pathology reports were provided by different pathologists and there was no standardized review procedure. However, in one of our early studies of ovarian cancer (10), 79 percent of the original pathologic diagnoses of histologic type were confirmed by an independent pathologist with extensive experience in gynecologic pathology. This suggests that there was reasonable agreement with the original classification. Furthermore, of the cases for which histologic type was classified in the present study, 12.3 percent of the cancers were mucinous. This proportion is similar to those ( percent) found in population-based case-control studies that have assessed the relation of smoking to risk of epithelial ovarian cancer (4, 6, 7). Second, of 709 epithelial ovarian cancer cases, the cell type for 105 cases could not be classified because of a lack of such information in the pathology report. However, unless most case women with missing information on cell type had mucinous cancers and were also nonsmokers or light

6 138 Zhang et al. smokers, this should not have materially changed the present findings. Third, information on age at starting smoking was not collected from approximately 20 percent of the women. This limited the study s power to assess the relation between age at starting smoking and risk of ovarian cancer of different cell types. Among subjects whose data were available, we did not find any association between age at starting to smoke and risk of mucinous ovarian cancer. Results from the two previous studies did not provide evidence for an association between age at starting smoking and risk of mucinous ovarian cancer (4, 6). Selection of appropriate controls in a case-control study is a challenge, especially when the researchers interest is in assessing the effect of cigarette smoking using hospital controls. Many diseases for which patients are hospitalized are likely to be associated with cigarette smoking. It has been well documented that the prevalence of cigarette smoking is higher in hospitalized patients than in the general population (23 25). Such selection bias would have resulted in underestimation of an adverse effect of cigarette smoking on risk of mucinous ovarian cancer in the present study. We selected our controls from patients with diseases we regarded as unrelated to tobacco use (26). Indirect evidence for the validity of our control selection is the fact that the percentages of women who smoked were similar across subgroups of controls. In addition, there was a greater proportion of White women and a smaller proportion of oral contraceptive users among cases than among controls, which is consistent with the results of previous studies (27 31). It is unlikely that biased reporting of cigarette smoking accounts for the present results, because there is no reason to suspect that patients with mucinous epithelial ovarian cancer would have reported their smoking differently from patients with other types of ovarian cancer. In conclusion, despite some inconsistencies in our data, these results strengthen the evidence that cigarette smoking may play a role in the development of mucinous ovarian cancer. ACKNOWLEDGMENTS This work was supported by grant CA45762 from the National Cancer Institute. Additional support was provided by grant FD-U from the Food and Drug Administration. The authors thank the many physicians who allowed their patients to be interviewed; the nurse interviewers who collected the data; Marguerite Angeloni, who coordinated data collection; and Leonard Gaetano, who was responsible for data management. The following hospitals participated in this study: New York, New York Brookhaven Memorial Hospital, Lenox Hill Hospital, Memorial Sloan-Kettering Cancer Center, and New York Hospital; Philadelphia, Pennsylvania American Oncologic Hospital, Crozier Chester Medical Center, Hahnemann University Hospital, Hospital of the Medical College of Pennsylvania, Hospital of the University of Pennsylvania, Lankenau Hospital, Montgomery Hospital, Pennsylvania Hospital, Presbyterian Hospital, and Thomas Jefferson University Hospital; Boston, Massachusetts Sancta Maria Hospital, Beth Israel Hospital, Newton Wellesley Hospital, Mount Auburn Hospital, Massachusetts General Hospital, Brigham and Women s Hospital, University Hospital, and New England Medical Center; Baltimore, Maryland Johns Hopkins Hospital, University of Maryland Medical Center, Sinai Hospital, Greater Baltimore Medical Center, and Mercy Medical Center. REFERENCES 1. American Cancer Society. Cancer facts and figures New York, NY: American Cancer Society, Edmondson RJ, Monaghan JM. The epidemiology of ovarian cancer. Int J Gynecol Cancer 2001;11: Koch M. Risk factors in ovarian cancer of epithelial origin by pathological diagnosis. Br J Cancer 1989;59: Green A, Purdie D, Bain C, et al. Cigarette smoking and risk of epithelial ovarian cancer (Australia). Cancer Causes Control 2001;12: Marchbanks PA, Wilson H, Bastos E, et al. Cigarette smoking and epithelial ovarian cancer by histologic type. Obstet Gynecol 2000;95: Modugno F, Ness RB, Cottreau CM. Cigarette smoking and the risk of mucinous and nonmucinous epithelial ovarian cancer. Epidemiology 2002;13: Kuper H, Titus-Ernstoff L, Harlow BL, et al. Population based study of coffee, alcohol and tobacco use and risk of ovarian cancer. Int J Cancer 2000;88: Rosenberg L, Shapiro S, Slone D, et al. Epithelial ovarian cancer and combination oral contraceptives. JAMA 1982;247: Miller DR, Rosenberg L, Kaufman DW, et al. Epithelial ovarian cancer and coffee drinking. Int J Epidemiol 1987;16: Kaufman DW, Kelly JP, Welch WR, et al. Noncontraceptive estrogen use and epithelial ovarian cancer. Am J Epidemiol 1989;130: Rosenberg L, Palmer JR, Zauber AG, et al. A case-control study of oral contraceptive use and invasive epithelial ovarian cancer. Am J Epidemiol 1994;139: Rosenberg L, Palmer JR, Zauber AG, et al. Relation of benzodiazepine use to the risk of selected cancers: breast, large bowel, malignant melanoma, lung, endometrium, ovary, non- Hodgkin s lymphoma, testis, Hodgkin s disease, thyroid, and liver. Am J Epidemiol 1995;141: Rosenberg L, Palmer JR, Rao RS, et al. A case-control study of analgesic use and ovarian cancer. Cancer Epidemiol Biomarkers Prev 2000;9: Coogan PF, Rosenberg L, Palmer JR, et al. Risk of ovarian cancer according to use of antidepressants, phenothiazines, and benzodiazepines (United States). Cancer Causes Control 2000; 11: Shapiro S. Case-control surveillance. In: Strom BL, ed. Pharmacoepidemiology. 3rd ed. New York, NY: John Wiley and Sons, Inc, 2000: Scully RE. Tumors of the ovary and maldeveloped gonads. (Atlas of tumor pathology, second series, fascicle 16). Washington, DC: Armed Forces Institute of Pathology, Scully RE. Classification of human ovarian tumors. Environ Health Perspect 1987;73:15 24.

7 Cigarette Smoking and Ovarian Cancer Ngelangel C, Munoz N, Bosch FX, et al. Causes of cervical cancer in the Philippines: a case-control study. J Natl Cancer Inst 1998;90: Chao A, Thun MJ, Jacobs EJ, et al. Cigarette smoking and colorectal cancer mortality in the Cancer Prevention Study II. J Natl Cancer Inst 2000;92: Slattery ML, Potter JD, Friedman GD, et al. Tobacco use and colon cancer. Int J Cancer 1997;70: Zenzes MT, Puy LA, Bielecki R. Immunodetection of cotinine protein in granulosa-lutein cells of women exposed to cigarette smoke. Fertil Steril 1997;68: Zenzes MT, Puy LA, Bielecki R. Immunodetection of benzo[a]pyrene adducts in ovarian cells of women exposed to cigarette smoke. Mol Hum Reprod 1998;4: Mantel N, Haenszel W. Statistical aspects of the analysis of data from the retrospective studies of disease. J Natl Cancer Inst 1959;22: West DW, Schuman KL, Lyon JL, et al. Differences in risk estimations from a hospital and a population-based case-control study. Int J Epidemiol 1984;13: Morabia A, Stellman SD, Wynder EL. Smoking prevalence in neighborhood and hospital controls: implications for hospitalbased case-control studies. J Clin Epidemiol 1996;49: Slone D, Shapiro S, Rosenberg L, et al. Relation of cigarette smoking to myocardial infarction in young women. N Engl J Med 1978;298: Weiss NS, Peterson AS. Racial variation in the incidence of ovarian cancer in the United States. Am J Epidemiol 1978;107: John EM, Whittemore AS, Harris R, et al. Characteristics relating to ovarian cancer risk: collaborative analysis of seven U.S. case-control studies. Epithelial ovarian cancer in black women. Collaborative Ovarian Cancer Group. J Natl Cancer Inst 1993; 85: Jensen JT, Speroff L. Health benefits of oral contraceptives. Obstet Gynecol Clin North Am 2000;27: The reduction in risk of ovarian cancer associated with oralcontraceptive use. The Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. N Engl J Med 1987; 316: Modan B, Hartge P, Hirsh-Yechezkel G, et al. Parity, oral contraceptives, and the risk of ovarian cancer among carriers and noncarriers of a BRCA1 or BRCA2 mutation. N Engl J Med 2001;345: