1 Epidemiology and Health Epidemiology and Health Volume: 38, Article ID: e2016060, 6 pages https://doi.org/10.4178/epih.e2016060 PERSPECTIVE Open Access The effect of smoking on lung cancer: ethnic differences and the smoking paradox Keum Ji Jung1, Christina Jeon1,2, Sun Ha Jee1 1Institute for Health Promotion, Department of Epidemiology and Health Promotion, Graduate School of Public Health, Yonsei University, Seoul; 2Department of Public Health, Graduate School, Yonsei University, Seoul, Korea The objectives of this review were to determine whether the smoking paradox still exists and to summarize possible explanations for the smoking paradox. Based on published data, we compared the risk of cigarette smoking for lung cancer in Western and Asian countries. We extracted data from the relevant studies about an- nual tobacco consumption, lung cancer mortality rates according to smoking status from each country, and possible explanations for the smoking paradox. A significantly greater risk of lung cancer death was found among current smokers in Asian countries than among nonsmokers, with relative risks (RRs) of 4.0 to 4.6 for Koreans, 3.7 to 5.1 for Japanese, and 2.4 to 6.5 for Chinese. Although a significantly greater risk of lung can- cer was present among current smokers in Asian countries, the RRs in Asian countries were much lower than those reported in Western countries (range, 9.4 to 23.2). Possible explanations for the smoking paradox includ- ed epidemiologic characteristics, such as the smoking amount, age at smoking initiation, and the use of filtered or mild tobacco. The smoking paradox definitely exists, but may be explained by major epidemiologic charac- teristics. Therefore, the smoking paradox should not be interpreted as indicating that tobacco is safer or less harmful for Asians. KEY WORDS: Smoking, Lung neoplasms, Risk, Ethnic groups INTRODUCTION Tobacco has a long history of use starting in the early Ameri- cas. Following the discovery of tobacco in the New World by Christopher Columbus in 1492 [1,2], it was introduced to Eu- ropean countries. Cigarettes became popular after the Industri- al Revolution. Tobacco was popularized in two especially signif- icant periods. The first wave of popularization occurred in West- ern countries at the time of World War I, which started in 1914. The second wave was in Asian countries around the 1940s, an epidemic that started 40 years after the European countries. Popularization continued until scientific studies in the mid-20th century demonstrated the negative health effects of tobacco smoking, including lung cancer. Despite the lack of data for these periods, the prevalence of smoking is thought to have increased during World Wars I and II. In 1935, corresponding to the development of a machine by a British company, the prevalence of smoking increased. The tobacco industry requires mass production of cigarettes. The first epidemiological paper on this topic published in the first half of the 20th century stated that “smoking is connected with the shortening of life span,” and smokers began to reflect on their health because they observed some negative health effects of smoking [3]. In 1950, cigarette companies initiated a shift in cigarette design from a non-filter design to a filter design. Pro- duction of filter cigarettes skyrocketed from 0.5% in 1950 to 87.7% in 1975 [4]. In 1950, Doll & Hill [5] conducted the British Doctors’ Study to determine the association between smoking and health con- sequences. In 1964, the first Surgeon General’s Report was pub- lished in the US, which allowed governmental enforcement of Correspondence: Sun Ha Jee Institute for Health Promotion, Department of Epidemiology and Health Promotion, Graduate School of Public Health, Yonsei University, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 03722, Korea Tel: +82-2-2228-1523, Fax: +82-2-365-5118, E-mail: jsunha@yuhs.ac Received: Dec 12, 2016, Accepted: Dec 20, 2016, Published: Dec 20, 2016 This article is available from: http://e-epih.org/ 2016, Korean Society of Epidemiology This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. http://crossmark.crossref.org/dialog/?doi=10.4178/epih.e2016060&domain=pdf 2 Epidemiology and Health 2016;38:e2016060 smoking bans in public places and strongly recommended smok- ing cessation [6]. Forty years later, another comprehensive re- port entitled Tobacco Free * Japan was published in 2004 in Japan [7]. At that time, findings on smoking and cancer risk ob- tained using data from the Korean Cancer Prevention Study (KCPS) were also published [8]. Interestingly, however, the mag- nitude of the effect of smoking on lung cancer was different be- tween Western and Asian countries, with a smaller effect size observed in Asian countries. Investigators named this phenom- enon “the smoking paradox,” and have tried to explain it [9,10]. The objectives of this review were to investigate whether the smoking paradox still exists and to summarize possible expla- nations for the smoking paradox (Figure 1). DOES THE SMOKING PARADOX EXIST? To determine whether the risk of smoking for lung cancer is lower in Asian countries, which have been referred to as the smoking paradox, we reviewed selected well-established epide- miologic studies from Western and Asian countries. The results are presented below. United Kingdom The British Doctors’ Study is the oldest epidemiologic study on this topic from the United Kingdom (UK) and has become very well known. This UK prospective study of smoking and death began in 1950, included 34,439 British male doctors, and has now continued for 50 years. The researchers published a 20-year follow-up study in 1976, a 40-year follow-up in 1994, and a 50-year follow-up in 2004 [11-13]. Recently, results from a 60-year follow-up study were presented at a meeting in hon- or of Richard Doll’s centenary at Oxford University on October 28, 2012. In the latest three reports (the 20-year, 40-year, and 50-year follow-up) the relative risk (RR) of smoking for lung cancer was found to have increased over the duration of follow-up. The RR was approximately eight at the time of the 20-year follow-up study, but increased to above 14 when the participants were followed up for more than 40 years (Figure 2). US In the US, the most well-known prospective cohort that we reviewed was the US veterans study [14]. The tobacco use of almost 250,000 US veterans was surveyed via questionnaire in the 1950s. Together with six other large studies and several case- control studies, the early results of this investigation provided the basis for the first series of reports by the Surgeon General on the health effects of smoking, which identified lung cancer and a number of other smoking-related diseases [6]. In a graph showing the results from the US veterans study cohort published in 1995, the RRs of smoking for laryngeal cancer and lung can- cer were 17.8 and 11.7, respectively [14]. Additionally, the Cancer Prevention Study (CPS), the largest prospective cohort, was reviewed. In 1959, 1,078,894 subjects were recruited by the American Cancer Society to participate Figure 1. Landmarks of the history of tobacco, from discovery to popularization. Landmark tobacco history Popularization in Western Popularization in Asian World War I, II 1492 1900 1940 1950 1964 2004, 2005 2016 Figure 2. Relative risk (RR) of smoking for lung cancer, UK. Adapted from Doll et al. [11-13]. 16 14 12 10 8 6 4 2 0 20 40 50 Duration of follow-up (yr) R R 3 Jung KJ et al.: Smoking effect on lung cancer: smoking paradox in the 12-year study, CPS-I. CPS-II, a prospective mortality study, was initiated by the American Cancer Society in 1982 [15]. The American Cancer Society recruited subjects in 1982 among the friends, neighbors, and acquaintances of volunteers. The tobac- co use of almost more than 1.2 million US inhabitants was sur- veyed by the questionnaire in 1982. The key findings were re- markable. On average, the mortality among non-smokers was 8.6, while the mortality among smokers was remarkably high (101.3 per 100,000 persons). The estimated RR was 11.9. Smok- ing accounted for over 95% of lung cancer deaths in men and 92% in women. China China is by far the largest producer of cigarettes in the world. Between 1915 and 1945, annual cigarette consumption rose from approximately 0 to 100 billion. China is also the largest consumer market in the world, with over 300 million smokers consuming 1.7 trillion cigarettes in 1997 (unpublished data). Three prospective studies of smoking and the RR for lung can- cer have been conducted: the Shanghai cohort, whose sample size was 18,244, and two prospective cohorts studied by Lam et al. [16] and Yuan et al. [17]. In the Shanghai cohort, male res- idents of Shanghai, China were enrolled from January 1, 1986, through September 30, 1989 and were actively follow ed via annual visits during the study period [17]. However, the cohort size was small and the duration of follow-up was relatively short. In China, Liu and Peto published a paper [18] presenting a ret- rospective proportional mortality study of one million deaths in 1998. The goal of that study was to examine the hazards at an early phase of the growing epidemic of death from tobacco in China. From 1989 to 1991, they interviewed the surviving family members of one million people who died from 1986 to 1998 in 98 areas of China. For both sexes, the lung cancer rates for individuals 35 to 69 of age were approximately 2.6 times higher in smokers than in nonsmokers among men and 2.0 times greater in smokers than in non-smokers in women. Of all deaths attributed to tobacco, 45% were due to chronic obstructive pul- monary disease and 15% were due to lung cancer. Japan In 2004, Tobacco Free * Japan, a comprehensive and detailed report on smoking and health similar to the US Surgeon Gen- eral’s report was published in Japan. Although the report covered numerous epidemiologic studies, we selected the Japan Collab- orative Cohort (JACC) study for this review [19]. The JACC study was sponsored by the Ministry of Education, Culture, Sports, Science, and Technology of Japan. The cohort was established from 1988 to 1990, with 46,465 men and 64,327 women aged 40 to 79 years recruited from 45 study areas throughout Japan. The RRs for lung cancer in smokers were 4.46 in men and 3.58 in women. In men, the population-attributable risk (PAR) was 52.2% for current smokers and 14.8% for ex-smokers. Korea In Korea, despite the introduction of tobacco 400 years ago, its widespread use in the form of manufactured cigarettes only emerged after 1945. The first national brand, produced by a gov- ernment monopoly, was named Seung Ri, meaning “indepen- dence from Japan.” Cigarette consumption rose sharply after the end of the Korean War in 1953. The rise has been particu- larly pronounced since 1960, as Korea became economically prosperous, with a peak around 2000 (Figure 3) [20]. Korean tobacco consumption trends have shown an increase since 1945. In Figure 3, the tobacco consumption amount unit is one mil- lion packs per year, and the red line shows the number of deaths from lung cancer. Since 1981, the year of the first mortality re- port, the number of lung cancer deaths in Korea continuously increased until 2015, when a slight drop was observed for the first time. The KCPS is a prospective cohort study with a follow- up period of 24 years that was designed to assess the risk factors for mortality, incidence, and hospital admission due to cancer and other chronic diseases. The KCPS cohort included 1,329,525 Koreans from 30 to 88 years of age who underwent biennial medical evaluations from 1992 to 1995. The first report based on the KCPS was published in 2004. The RRs of smoking for lung cancer were 4.6 in men and 2.5 in women [20]. Summary of selected publications Two major observations were made based on our analysis of the data. First, a significantly greater risk of lung cancer mortal- ity was found among current smokers than among non-smok- ers in Asian countries (RRs, 4.0 to 4.6 for Koreans; 3.7 to 5.1 for Japanese; and 2.4 to 6.5 for Chinese). Second, although a significantly increased risk of lung cancer was present among current smokers in Asian countries, the RRs in Asian countries were much lower than those reported in Western countries (range, 6,000 5,000 4,000 3,000 2,000 1,000 0 19 45 19 50 19 55 19 60 19 65 19 70 19 75 19 80 19 85 19 90 19 95 20 00 20 05 20 10 20 15 Year M ill io n pa ck Tobacco consumption No. of death from lung cancer Figure 3. Tobacco consumption and lung cancer deaths in Korea. 4 Epidemiology and Health 2016;38:e2016060 9.4 to 23.2). The results were very similar among men, with the RRs in Asian countries being much lower than those reported in Western countries. Asian countries had a lower PAR for smok- ing and lung cancer than Western countries. In summary, numer- ous epidemiological studies have consistently reported smoking to be a risk factor for lung cancer. However, the magnitude of the risk for lung cancer mortality associated with cigarette smok- ing has been reported to be lower in Asian countries than in West- ern countries. POSSIBLE EXPLANATIONS FOR THE SMOKING PARADOX Different epidemics Cigarette consumption in Japan and Korea has followed a pat- tern similar to that observed among adults in the US, although the major increase took place 40 years later [7]. One of the possi- ble explanations for this discrepancy may be the differences in the exposure amount of cigarettes smoked in different eras in Western countries and Asian countries. The per capita cigarette consumption in the US reached 3,000 cigarettes per year in the 1940s, while Korea and Japan reached the same level in the 1970s. Therefore, we may hypothesize that Asian countries have not yet experienced the full effect of cigarette smoking on lung cancer. Age at smoking initiation and amount of smoking Another explanation might be the age of smoking initiation. It has been found that current smokers in Asian countries began smoking at an older age. A strong trend has been found for an earlier smoking initiation age in later birth cohorts. While the age at smoking initiation for Korean men and women has grad- ually decreased in recent birth cohorts, they still began smoking much later than men and women in the US [20]. The average smoking amount among Asian smokers was relatively small in the past. In China, the average daily cigarette consumption for men was one in 1952, increasing to 16 in 1996 [18]. In Korea, the weighted mean of smoking quantity among current smok- ers was only a half-pack in 1980, which dramatically increased to almost one pack in 2000 (unpublished data). Toxicity and filter The contents of tobacco itself have not significantly changed over time. Basic physiology and genetic factors are likewise not changeable over time. However, tobacco products may have changed over time. Since consumer demand has increased for mild tobacco due to health concerns, tobacco industries have produced products such as Mild Seven or Right, which are known to have low tar and low nicotine. Recently, such brands have become very popular in Asian countries. Previous studies have found that the use of low-nicotine products did not lead smok- ers to change the quantity of nicotine consumption per day. This is because smokers who use low-nicotine products tend to in- crease the number of cigarettes they smoke to meet their needs for nicotine. Therefore, the effects of low-tar or low-nicotine prod- uct use on various diseases are known to be similar to those of standard cigarettes. However, Philip Morris, British American Tobacco, and other tobacco companies considered the possibility of reducing ciga- rette toxicity through the use of filters. Although filtered brands constitute over 90% of the entire cigarette market [4], epidemi- ologic data support the conclusion that cigarette filters have done little to protect smokers. In 2011, investigators from the US and Japan reported that the shift from non-filter to filter cigarettes appears to have merely altered the most frequent type of lung cancer, from squamous cell carcinoma to adenocarcinoma [21]. Lung cancer risk of never smokers The data previously discussed in this review indicate that lung cancer mortality among nonsmokers across countries varies, and is likely to be lower in Western countries than in Asian coun- tries. For men, lung cancer mortality per 100,000 persons among non-smokers in Western countries ranged from 14 to 17, while in Asians it ranged from 19.2 to 36.0. Death rates of non-smok- ers in all Asian cohorts were higher than those of non-smokers in Western countries. However, the death rates for current smo- kers in all Asian cohorts were lower than those in Western coun- tries. Comparing the death rates and exposure levels of current smokers indicates that one reason for the lower RR in Asians is the presence of higher death rates among non-smokers combined with lower death rates among smokers (Figure 4). Another com- Figure 4. Hypothetical graph of relative risk (RR) in relation to death rate and smoking status. 100 10 Smoker SmokerNever- smoker Never- smoker Western Asian D ea th ra te /1 00 ,0 00 RR > 10 RR = 5 5 Jung KJ et al.: Smoking effect on lung cancer: smoking paradox prehensive study reported that the death rate from lung cancer among never smokers was higher in men than in women, and higher in Asians residing in Asia than in individuals of Europe- an descent [22]. This study used participants from the CPS to identify mortality rates among those of European descent and participants from the KCPS as the Asian sample. The morality rates per 100,000 persons among Europeans and Asians were 12.0 and 26.0 in men and 9.5 and 16.1 in women, respectively. One of the possible reasons for the increased lung cancer risk among never smokers in Asian countries is poor ventilation com- bined with the use of coal, which Asian women have traditional- ly been exposed to [23]. Genetic risk scores involving telomere length are also a possible factor explaining differences in lung cancer risk among never smokers [24]. Genetic differences The association of DNA methylation with smoking has been reported to vary according among ethnic groups [25]. Current smokers in Europe have been found to smoke more per day, to have started smoking earlier, and, therefore, to have accumulat- ed more pack-years than South Asians. Differences were found in DNA methylation between South Asians (57%) and Europe- ans (53%) among current smokers. However, the difference was not sufficiently large and may not explain the smoking paradox. Another issue is epidermal growth factor receptor (EGFR). The EGFR gene is a major oncogene in lung cancer adenocarcino- ma in Asians. The expression of EGFR in both never smokers and ever smokers was found to be much higher in Japan than in the US [26]. These differences may explain the smoking par- adox, at least for lung cancer adenocarcinoma in Asians. Others In summary, smoking is the most convincing causal risk fac- tor for lung cancer, with over 90% of attributable risk in West- ern countries. Alcohol consumption, cholesterol, total and satu- rated fat, and animal fat are other possible risk factors. Vegeta- bles and fruits are convincing protective factors, while physical activity, carotenoids, vitamin C, and selenium are potentially beneficial. In a report that compared the outcomes of tobacco smoking (e.g., lung cancer) between Japan and the US, no dif- ferences were shown regarding diet, degree of inhalation, or al- cohol consumption. However, more toxicity due to the carcino- gens in cigarettes and an earlier onset age of smoking were found in the population in the US [27]. CONCLUSION Based on previous epidemiological studies, the smoking para- dox definitely exists. However, while the smoking paradox may be explained by major epidemiologic differences between West- ern and Asian populations, such as smoking amount, age at smok- ing initiation, and the use of filtered or mild tobacco, we cannot conclude how much these factors may explain this phenome- non, as relevant genetic differences have not been sufficiently elucidated. The smoking paradox itself bears important implications for public health. Based on previous studies that examined differ- ences in epidemiological characteristics related to the smoking paradox, the Asian population may not yet have fully experi- enced the damage from smoking due to a comparably shorter tobacco epidemic history than Western countries. Second, the Asian population has been found to be characterized by a late age of smoking initiation, and since the smoking rate has incre- ased and the initiation age of adolescents has decreased recent- ly, the damage from smoking may increase in the future. In the public health field, practitioners should be especially cautious of misstating the smoking paradox by saying that tobacco is saf- er or less harmful for Asians. Although tobacco brands change over time, the contents of tobacco do not, meaning that no safe tobacco brand exists in the world. ACKNOWLEDGEMENTS This study was supported by a grant from the National R&D Program for Cancer Control, Ministry of Health and Welfare, Republic of Korea (1631020). CONFLICT OF INTEREST The authors have no conflicts of interest to declare for this study. ORCID Keum Ji Jung http://orcid.org/0000-0003-4993-0666 Christina Jeon http://orcid.org/0000-0001-7584-6228 Sun Ha Jee http://orcid.org/0000-0001-9519-3068 REFERENCES 1. Gately I. La diva nicotina: the story of how tobacco seduced the world. London: Scribner; 2002, p. 1-320. 2. Musk AW, de Klerk NH. History of tobacco and health. Respirology 2003;8:286-290. 3. Grzybowski A. The history of antitobacco actions in the last 500 years. part. 1. Non-medical actions. Przegl Lek 2006;63:1126-1130 (Polish). 6 Epidemiology and Health 2016;38:e2016060 4. Jarvik ME, Cullen JW, Gritz ER, Vogt TM, West LJ. Research on smoking behavior; 2007 [cited 2016 Dec 25]. Available from: https:// archives.drugabuse.gov/pdf/monographs/17.pdf. 5. Doll R, Hill AB. The mortality of doctors in relation to their smoking habits; a preliminary report. Br Med J 1954;1:1451-1455. 6. US Surgeon General’s Advisory Committee on Smoking and Health. Smoking and health; report of the advisory committee to the Surgeon General of the Public Health Service. Washington, DC: US Depart- ment of Health, Education, and Welfare; 1964, p. 23-40. 7. Mochizuki-Kobayashi Y, Samet JM, Yamaguchi N. Tobacco free * Japan: recommendations for tobacco control policy; 2004 [cited 2016 Dec 4]. Available from: http://escholarship.org/uc/item/9x27v23m. 8. Jee SH, Samet JM, Ohrr H, Kim JH, Kim IS. Smoking and cancer risk in Korean men and women. Cancer Causes Control 2004;15:341- 348. 9. Kawaguchi T, Matsumura A, Fukai S, Tamura A, Saito R, Zell JA, et al. Japanese ethnicity compared with Caucasian ethnicity and never- smoking status are independent favorable prognostic factors for over- all survival in non-small cell lung cancer: a collaborative epidemio- logic study of the National Hospital Organization Study Group for Lung Cancer (NHSGLC) in Japan and a Southern California Region- al Cancer Registry databases. J Thorac Oncol 2010;5:1001-1010. 10. Ahn MJ, Lee J, Park YH, Ahn JS, Ziogas A, Zell JA, et al. Korean ethnicity as compared with white ethnicity is an independent favor- able prognostic factor for overall survival in non-small cell lung can- cer before and after the oral epidermal growth factor receptor tyro- sine kinase inhibitor era. J Thorac Oncol 2010;5:1185-1196. 11. Doll R, Peto R. Mortality in relation to smoking: 20 years’ observa- tions on male British doctors. Br Med J 1976;2:1525-1536. 12. Doll R, Peto R, Wheatley K, Gray R, Sutherland I. Mortality in rela- tion to smoking: 40 years’ observations on male British doctors. BMJ 1994;309:901-911. 13. Doll R, Peto R, Boreham J, Sutherland I. Mortality in relation to smok- ing: 50 years’ observations on male British doctors. BMJ 2004;328: 1519. 14. McLaughlin JK, Hrubec Z, Blot WJ, Fraumeni JF Jr. Smoking and cancer mortality among U.S. veterans: a 26-year follow-up. Int J Can- cer 1995;60:190-193. 15. Thun MJ, Myers DG, Day-Lally C, Namboodiri MM, Calle EE, Flan- ders WD, et al. Age and the exposure-response relationships between cigarette smoking and premature death in cancer prevention study II [cited 2016 Dec 4]. Available from: https://cancercontrol.cancer.gov/ brp/tcrb/monographs/8/m8_5.pdf. 16. Lam TH, Xu L, Schooling CM, Chan WM, Lee SY, Leung GM. Smok- ing and mortality in a prospective cohort study of elderly Chinese in Hong Kong. Addiction 2015;110:502-510. 17. Yuan JM, Ross RK, Wang XL, Gao YT, Henderson BE, Yu MC. Mor- bidity and mortality in relation to cigarette smoking in Shanghai, Chi- na. A prospective male cohort study. JAMA 1996;275:1646-1650. 18. Liu BQ, Peto R, Chen ZM, Boreham J, Wu YP, Li JY, et al. Emerg- ing tobacco hazards in China: 1. Retrospective proportional mortality study of one million deaths. BMJ 1998;317:1411-1422. 19. Ando M, Wakai K, Seki N, Tamakoshi A, Suzuki K, Ito Y, et al. At- tributable and absolute risk of lung cancer death by smoking status: findings from the Japan Collaborative Cohort Study. Int J Cancer 2003; 105:249-254. 20. Jee SH, Ohrr H, Sull JW, Samet JM. Cigarette smoking, alcohol drink- ing, hepatitis B, and risk for hepatocellular carcinoma in Korea. J Natl Cancer Inst 2004;96:1851-1856. 21. Ito H, Matsuo K, Tanaka H, Koestler DC, Ombao H, Fulton J, et al. Nonfilter and filter cigarette consumption and the incidence of lung cancer by histological type in Japan and the United States: analysis of 30-year data from population-based cancer registries. Int J Cancer 2011;128:1918-1928. 22. Thun MJ, Hannan LM, Adams-Campbell LL, Boffetta P, Buring JE, Feskanich D, et al. Lung cancer occurrence in never-smokers: an anal- ysis of 13 cohorts and 22 cancer registry studies. PLoS Med 2008;5:e185. 23. Kim C, Gao YT, Xiang YB, Barone-Adesi F, Zhang Y, Hosgood HD, et al. Home kitchen ventilation, cooking fuels, and lung cancer risk in a prospective cohort of never smoking women in Shanghai, China. Int J Cancer 2015;136:632-638. 24. Machiela MJ, Hsiung CA, Shu XO, Seow WJ, Wang Z, Matsuo K, et al. Genetic variants associated with longer telomere length are as- sociated with increased lung cancer risk among never-smoking wom- en in Asia: a report from the female lung cancer consortium in Asia. Int J Cancer 2015;137:311-319. 25. Elliott HR, Tillin T, McArdle WL, Ho K, Duggirala A, Frayling TM, et al. Differences in smoking associated DNA methylation patterns in South Asians and Europeans. Clin Epigenetics 2014;6:4. 26. Saito M, Shiraishi K, Kunitoh H, Takenoshita S, Yokota J, Kohno T. Gene aberrations for precision medicine against lung adenocarcino- ma. Cancer Sci 2016;107:713-720. 27. Takahashi K, Bokura H, Kobayashi S, Iijima K, Nagai A, Yamaguchi S. Metabolic syndrome increases the risk of ischemic stroke in wom- en. Intern Med 2007;46:643-648.