Cigarette smoking and risk of borderline and invasive epithelial ovarian cancer Inger T. Gram 1,2 * , Tonje Braaten 1 , Hans-Olov Adami 3,4,5 , Eiliv Lund 1 and Elisabete Weiderpass 3,5,6 1 Institute of Community Medicine, University of Tromsø, Breivika N-9037 Tromsø, Norway 2 Norwegian Centre for Telemedicine, University Hospital of North Norway, N-9038 Tromsø, Norway 3 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, SE-17177 Stockholm, Sweden 4 Department of Epidemiology, Harvard University, 677 Huntignton Avenue, Boston, MA, 02115, USA 5 Cancer Registry of Norway, Montebello, N-0310, Oslo, Norway 6 Samfundet Folkh € alsan, Topeliusgatan 20, FI-00250 Helsingfors, Finland Studies regarding the association between smoking and risk of epi- thelial ovarian cancer (EOC) are inconsistent. The purpose of this study was to examine the association between smoking and EOC, overall and according to invasiveness and histological subtype in a cohort of women with a high proportion of smokers at enrolment. We followed 103,081 women, aged 30–50 years in 1991/1992, from the Norwegian-Swedish Women’s Lifestyle and Health cohort. The women completed a questionnaire on personal characteristics and exposures at enrolment and were subsequently followed with linkages to national registers through December 31, 2004. We used Cox proportional hazard regression models to estimate hazard ratio (RR) of EOC with 95% confidence intervals (CIs) associated with different measures of smoking exposures adjusting for confounding variables. Altogether 343 [241 (70%) invasive and 102(30%) borderline] incident EOC cases were identified. Former [HR 5 2.2(95% CI 1.0–4.7)] and current [HR 5 2.7(95% CI 1.2–5.7)] smokers had a more than doubling in risk for border- line tumors compared to never smokers. Women who had smoked for more than 20 years had 3 times [HR 5 3.1(95% CI 1.5–6.7)] the risk of borderline tumors compared to never smokers. A test for trend according to smoking status was almost significant for mucinous tumors (p-trend 5 0.05). A significant dose response relationship was found according to smoking intensity [pack- years; (0–9, 0–14, ≥ 15)] and duration [number of years; (0–10, 11–20, ≥ 20)] for borderline and serous tumors (p-trends < 0.05). In conclusion, smoking may increase the risk of borderline EOC. ' 2007 Wiley-Liss, Inc. Key words: epidemiology; women; smoking; prospective cohort; epithelial ovarian cancer; borderline tumors; Norway; Sweden Relatively little is known about the causes of ovarian cancer. Established factors that increase the risk are age, nulliparity, infer- tility, family history of ovarian cancer, whereas increasing parity, oral contraceptive use, tubal ligation and hysterectomy reduce the risk. 1,2 Several other factors such as a late age at menarche, an early menopause, a late age at first and last birth, breast feeding and physical activity have been linked with decreased risk, while obesity, postmenopausal hormone therapy use, and alcohol con- sumption have been linked with increased risk. However, many studies did not find these associations. Several etiologic hypothe- ses; the incessant ovulation, the retrograde transport, the inflam- mation, exogenous carcinogens, and some related to specific hor- mones, have been postulated. None of the suggested hypotheses so far can explain all the epidemiological data. 1–3 According to the latest Monograph on ‘‘Tobacco smoke and involuntary smoking’’ from the International Agency for Research on Cancer there is insufficient evidence to draw any conclusions regarding the possible effect of smoking on ovarian cancer risk. 4 Risch et al. was the first to suggest that mucinous and nonmuci- nous ovarian cancer may have different etiology. 5 A recent sys- tematic review, including a meta-analysis, about smoking and ovarian cancer lends support to this hypothesis. Jordan et al. found that there was a significant doubling in risk of mucinous, but not of serous, endometroid and clear cell ovarian cancers among cur- rent smokers compared to never smokers. 6 This meta-analyses included data from 1 cohort study, 7 8 population based case-con- trol studies 8–15 and 1 pooled analysis of 10 case-control studies from the US. 16 Because the use of tobacco is rising sharply globally, 17 smok- ing may generate large numbers of ovarian cancers if a causal association exists. The purpose of this study was to examine the association between smoking and epithelial ovarian cancer (EOC), overall, according to invasiveness and histological sub- type in a cohort of women with a high proportion of smokers at enrolment. Material and methods Study cohort The Norwegian-Swedish Women’s Lifestyle and Health Cohort Study was initiated in 1991/1992. In Norway, a nationwide ran- dom sample of 100,000 women, born between 1943 and 1957, was drawn from the National Population Register at Statistics Nor- way. In Sweden, a random sample of 96,000 women, born between 1942 and 1962 residing in the Uppsala Health Care Region (which comprises about one-sixth of the Swedish popula- tion), was drawn from the National Population Register at Statis- tics Sweden (Stockholm, Sweden). All women received a letter inviting them to participate in the study. The letter requested the women to provide written informed consent, and contained a comprehensive questionnaire to be com- pleted and returned in a prestamped envelope. The common set of questions included detailed assessment of smoking habits, alcohol consumption, contraceptive use, reproductive history, history of breast cancer in the mother and sister(s), height and current weight (allowing us to calculate BMI as weight in kilograms divided by the square of height in meters), and other aspects of lifestyle hab- its. In both countries, the national Data Inspection Boards and the regional Medical Ethical Committees approved the study. Smoking assessment The questionnaire elicited information on current and previous smoking history. Women who reported to be ever smokers were asked to fill in an 8 by 7 table with preset categories for number of cigarettes smoked daily at different age intervals. On the basis of this table, we further categorized ever smokers according to cur- rent and former smoking status, age of smoking initiation, smok- ing duration; average number of cigarettes smoked daily, and pack-years of smoking (i.e. number of cigarettes smoked per day, Grant sponsors: In Norway: The National Cancer Institute of the U.S; Grant number: CA 52449; the Norwegian Cancer Society; Grant number: DNK 90050; and the Aakre Foundation. In Sweden: The Swedish Council for Planning and Coordination of Research, the Swedish Cancer Society, STINT (The Swedish Foundation for International Cooperation in Research and Higher Education), Organon, Pharmacia, Medical Products Agency, Schering-Plough, AFA and the US Army Breast Cancer Research Program. *Correspondence to: Inger T. Gram, Institute of Community Medicine, University of Tromsø, N - 9037 Tromsø, Norway. Fax: 147-77-64-48-31. E-mail: inger.gram@ism.uit.no Received 27 February 2007; Accepted after revision 31 July 2007 DOI 10.1002/ijc.23108 Published online 4 October 2007 in Wiley InterScience (www.interscience. wiley.com). Int. J. Cancer: 122, 647–652 (2008) v ' 2007 Wiley-Liss, Inc. Publication of the International Union Against Cancer