Risk of Functional Ovarian Cyst: Effects of Smoking and Marijuana Use according to Body Mass Index

Transcription

1 American Journal of Epidemiology Copyright ª 2005 by the Johns Hopkins Bloomberg School of Public Health All rights reserved Vol. 161, No. 6 Printed in U.S.A. DOI: /aje/kwi080 Risk of Functional Ovarian Cyst: Effects of Smoking and Marijuana Use according to Body Mass Index Victoria L. Holt 1,2, Kara L. Cushing-Haugen 1, and Janet R. Daling 1,2 1 Program in Epidemiology, Fred Hutchinson Cancer Research Center, Seattle, WA. 2 Department of Epidemiology, School of Public Health and Community Medicine, University of Washington, Seattle, WA. Received for publication August 12, 2004; accepted for publication October 27, Smoking is one of the few risk factors that have been identified for functional ovarian cysts, and results of one epidemiologic study suggest that body mass index (BMI; weight (kg)/height (m) 2 ) may modify the effect of this exposure. The current study assessed the association of cigarette smoking and marijuana use with functional ovarian cyst risk by using data from a population-based case-control study of 586 incident functional ovarian cyst cases and 757 age-matched controls in a large health maintenance organization in Washington State. In multivariate analyses controlling for age, education, and reference year, the authors found an increase in risk associated with current cigarette smoking among women whose BMI was <20 (odds ratio (OR) ¼ 2.48, 95% confidence interval (CI): 1.32, 4.64) or (OR ¼ 1.60, : 1.04, 2.46) but not >25 (OR ¼ 0.85, : 0.53, 1.37). Corresponding risks associated with current marijuana use were BMI <20, OR ¼ 2.05 (: 0.89, 4.75); BMI 20 25, OR ¼ 1.78 (: 1.00, 3.17); and BMI >25, OR ¼ 0.72 (: 0.36, 1.42). Study results indicate that increased BMI may attenuate the adverse effect of smoking on the risk of functional ovarian cyst. body mass index; case-control studies; marijuana smoking; ovarian cysts; smoking Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio. Functional ovarian cysts are benign cystic structures of two types (follicular cysts, corpus luteum cysts) that occur in reproductive-age women. Although they can be self-limiting and regress without treatment, they may cause substantial pelvic pain and menstrual dysfunction. Persistent functional ovarian cysts may be removed surgically, and there are nearly 200,000 hospitalizations annually among US women with this gynecologic condition (1). Epidemiologic studies have identified few risk factors for the development of functional ovarian cyst; clinicians have posited that altered serum gonadotropin levels cause the condition (2). Cigarette smoking is known to affect gonadotropins, ovarian hormones, and ovarian function (3 5), yet the results of previous epidemiologic studies of this exposure in relation to cyst development are somewhat mixed. Smoking was found to significantly increase ovarian cyst risk in two US studies, a retrospective cohort study with self-reported disease ascertainment (6) and a small case-control study that used existing medical records for case finding and data collection (7). Conversely, an Italian hospital-based case-control study reported no significant associations with current or former smoking (8). Although marijuana also affects gonadotropin levels and ovarian function (9), to our knowledge there has been no epidemiologic research on the possible association between marijuana use and ovarian cyst development. We conducted this analysis to investigate further the relations between cigarette and marijuana smoking and the risk of functional ovarian cyst. Because the results of the cohort study noted above suggested that the effect of cigarette smoking on cyst risk may be stronger among leaner women, a primary goal was to determine whether heterogeneity of effect across body mass index (BMI) categories existed. MATERIALS AND METHODS Full details of the methods used to conduct this health maintenance organization-based case-control study have Correspondence to Dr. Victoria L. Holt, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North (M4 B874), PO Box 19024, Seattle, WA ( vholt@u.washington.edu). 520

2 Smoking and Ovarian Cyst Risk according to Body Mass Index 521 been reported elsewhere (10). Briefly, Group Health Cooperative inpatient, outpatient, and radiology databases were used to identify as potential cases all female enrollees aged years diagnosed by ultrasound or surgical examination with a functional ovarian cyst (International Classification of Diseases, Ninth Revision, codes ) 2 cm or larger between January 1, 1990, and June 30, After medical record review and telephone query of potential cases, we excluded women who did not speak English, were pregnant at the time of diagnosis, or had been enrolled in the Group Health Cooperative less than 6 months before the reference month (the month of the cyst diagnosis). Group Health Cooperative lists of all women aged years enrolled in the health maintenance organization between January 1, 1990, and June 30, 1994, were the basis for our control group selection. We chose potential controls randomly from these lists, stratified by 5-year age groups, with the number of controls in each age group proportional to the number of cases in that age group. Reference months were assigned to potential controls according to the distribution of cases reference months. After telephone query of potential controls, we excluded women who did not speak English, had had a bilateral oophorectomy, were pregnant during the reference month, or had been enrolled in the Group Health Cooperative less than 6 months before the reference month. After written informed consent was obtained, we were able to interview 586 of the 750 eligible potential cases (78.1 percent) and 757 of the 921 eligible potential controls (82.2 percent). All study protocols were approved by the Fred Hutchinson Cancer Research Center and the Group Health Cooperative human subjects review committees. For the analyses presented in this paper, we excluded three cases and two controls for whom interview data were unreliable. Additionally, we excluded 60 cases and 33 controls who had received an infertility diagnosis or had undergone infertility treatment, because ovarian cysts are associated with infertility treatment and these women s smoking habits may have changed as a result of an infertility diagnosis. A total of 523 cases and 722 controls remained for analysis. For 395 cases (75.5 percent), the ovarian cyst diagnosis was made during a clinic visit for cyst symptoms; for 128 cases (24.5 percent), the cyst was found incidentally. Study participants were interviewed in person by trained female interviewers using a structured questionnaire to elicit a lifetime history of a variety of demographic, medical, and lifestyle factors. All interview information collected pertained to events or characteristics up until the reference month. We asked participants whether they had smoked more than 100 cigarettes in their lifetime and, if so, at what age they began to smoke. If these women did not still smoke in the reference month, we obtained the age at which they had stopped. Additionally, we ascertained the average number of cigarettes smoked per day overall and in the 3 months before the reference month, and the total number of years of smoking before the reference month. For these analyses, we considered current cigarette smokers to be women who smoked in the 3 months before the reference month; former smokers were those who smoked at least 100 cigarettes in their lifetime but did not smoke in the 3 months before the reference month; and never smokers were those who smoked fewer than 100 cigarettes total in their lifetime. Similarly, we asked participants whether they had ever smoked marijuana and, if so, at what age they first did so. We then ascertained how often these women smoked marijuana when they first began to smoke, and how much they smoked on a day during which they smoked. If the participant s marijuana smoking habits had ever changed, we ascertained the details noted above for each change in smoking habits up until the reference month. For the current analyses, we classified as marijuana never users women who had smoked marijuana fewer than five times in their lifetime; current users were women who smoked marijuana at least five times in their lifetime and who considered themselves marijuana smokers as of the reference month; and former users were women who had smoked marijuana at least five times and considered themselves not to be marijuana smokers as of the reference month. Alcohol consumption in the 12 months prior to the reference month was ascertained as part of a food frequency questionnaire. Odds ratios and 95 percent confidence intervals for the associations of cigarette smoking and marijuana use with risk of a functional ovarian cyst were calculated by using unconditional logistic regression, with Stata statistical software (11). Two comparisons were made in the analyses of each exposure: 1) former versus never smoker and 2) current versus never smoker. We tested for effect modification by BMI (defined as weight (kg)/height (m 2 )) using the likelihood ratio test and categorizing BMI as recommended by the National Research Council s Committee on Diet and Health (12): underweight (<20.0), normal weight ( ), and overweight (>25.0). We found significant (p < 0.05) heterogeneity of effect by BMI for cigarette smoking but not for marijuana use; for consistency and to provide additional information, in this paper we present both sets of analyses separately by BMI category. We examined as potential confounders the following variables: race/ethnicity, marital status, educational level, income, marijuana use (for smoking analyses), cigarette smoking (for marijuana analyses), pregnancies, livebirths, spontaneous abortions, induced abortions, tubal sterilization, current oral contraceptive use, previous ovarian cyst, and alcohol consumption. We included potential confounding variables in the multivariate models if they were matching variables or changed the odds ratio for ovarian cyst risk associated with smoking or marijuana use by 10 percent or more (13). Age, educational level, and reference year met these criteria for the smoking analyses; and age, cigarette smoking, and reference year met the criteria for the marijuana analyses. RESULTS Women with cysts were similar to those without (controls) regarding age and race/ethnicity, but cases were more likely than controls to be single, and they had lower educational and income levels (table 1). Pregnancy histories of cases and controls were similar, with the exception that cases were more likely to have a history of spontaneous abortion. Additionally, cases were less likely than controls

3 522 Holt et al. TABLE 1. Demographic, reproductive, and other selected characteristics of ovarian cyst cases and controls, Group Health Cooperative, Washington State, (n ¼ 523) (n ¼ 722) No. % No. % Age (years) < Race/ethnicity White African American Asian Hispanic Native American Other Refused/unknown 1 Marital status Married Living as married Single Highest educational level High school Some college College graduate Refused/unknown 1 Household income ($) <15, ,000 29, ,000 45, >45, Refused/unknown 7 8 No. of pregnancies Table continues to be using oral contraceptives at the reference date and more likely to have had a tubal sterilization. also were more likely to have a prior ovarian cyst diagnosis, weigh over 70.8 kg, and have a BMI of >25.0. Smoking at least 100 cigarettes in their lifetime was reported by 46.5 percent of cases and 38.5 percent of controls; 28.7 percent of cases and 19.9 percent of controls smoked at the reference date. Among current smokers, the mean numbers of cigarettes smoked per day were 12.4 for TABLE 1. Continued (n ¼ 523) (n ¼ 722) No. % No. % No. of livebirths No. of spontaneous abortions No. of induced abortions Oral contraceptive use Never Former Current Tubal sterilization Previous ovarian cyst* Weight (kg) < < > Body mass indexy < > * Excludes two cases and one control whose ovarian cyst history was unknown. y Weight (kg)/height (m) 2. cases whose BMI was <20.0, 14.4 for cases whose BMI was , and 15.4 for cases whose BMI was >25.0. The corresponding mean numbers for controls were 12.7, 12.6, and 12.9 cigarettes per day. In multivariate analyses stratified on BMI and adjusted for age, educational level, and reference year, we found significant elevations in risk of a functional ovarian cyst diagnosis associated with current smoking among women whose BMI was <20.0 (odds ratio (OR) ¼ 2.48, 95 percent confidence interval (CI): 1.32, 4.64) or (OR ¼ 1.60, 95 percent CI: 1.04, 2.46). Among women whose BMI was >25.0, smoking was not associated with ovarian cyst risk (OR ¼ 0.85, 95 percent CI: 0.53, 1.37; table 2). Smoking marijuana at least five times was reported by 54.0 percent of cases and 48.5 percent of controls; 13.0 percent of cases and 9.0 percent of controls were current marijuana

4 Smoking and Ovarian Cyst Risk according to Body Mass Index 523 TABLE 2. Risk of functional ovarian cyst by cigarette smoking status and body mass index,* Group Health Cooperative, Washington State, Smoking status (n ¼ 104) Body mass index <20.0 Body mass index Body mass index >25.0 (n ¼ 134) ORy,z y (n ¼ 241) (n ¼ 394) (n ¼ 178) (n ¼ 193) Current , , , 1.37 Former , , , 1.26 Never Reference Reference Reference * Weight (kg)/height (m) 2. y OR, odds ratio; CI, confidence interval. z Adjusted for subject age, reference year, and educational level. Excludes one control whose educational level was unknown. smokers. In multivariate unconditional logistic regression analyses stratified on BMI and adjusted for age, cigarette smoking, and reference year, we found marijuana-associated ovarian cyst risk patterns that were similar to cigaretteassociated risks (table 3). Nonsignificantly elevated disease risks were found for current marijuana users whose BMI was <20.0 (OR ¼ 2.05, 95 percent CI: 0.89, 4.75) or (OR ¼ 1.78, 95 percent CI: 1.00, 3.17) but not for users whose BMI was >25.0 (OR ¼ 0.72, 95 percent CI: 0.36, 1.42). DISCUSSION In this study, associations between cigarette smoking and diagnosis of functional ovarian cyst varied significantly according to women s BMI. Cyst risk was elevated with current smoking for underweight and normal-weight women but not for overweight women. Two prior epidemiologic studies found significant positive associations between cigarette smoking and functional ovarian cysts among women of all BMIs combined. In a 1988 retrospective cohort study, Wyshak et al. (6) asked 5,398 female college alumnae (2,622 former college athletes and 2,776 nonathletes) aged years if they had ever had ovarian cysts. The authors reported a lower occurrence of cysts in every age decade among never smokers than among women who were former or current smokers at the time of the follow-up interview (OR ¼ 1.5 for ever vs. never smokers). A small 1994 case-control study conducted in the same health maintenance organization as the current study and based on medical record review found a somewhat stronger association: women with functional ovarian cysts were twice as likely as controls to be current smokers (7). Both studies had shortcomings. The cohort study relied on self-reported historical disease ascertainment and could not determine smoking status at the time of the reported cyst diagnosis; the case-control study could not differentiate between former and never smokers; and neither study evaluated the influence on the smoking ovarian cyst association of variables such as BMI, marijuana use, and alcohol consumption. In apparent contrast to the studies cited above, a hospitalbased case-control study conducted in Milan, Italy, did not find a significant association between current or former smoking and cyst risk (8). However, in that study, the adjusted odds ratio for current smoking was 1.4, and the small number of cases (n ¼ 68) and the use of hospitalized controls may have affected the results and their interpretation. Similarly, the authors of a cross-sectional clinical study of 335 randomly selected reproductive-age women undergoing transvaginal ultrasound examination reported no association between smoking and ovarian cysts in their population (14). The power of that study was quite limited as well, and the authors found that 40.9 percent of the 22 women with cystic ovarian lesions were current smokers, compared with 22.7 percent of the 309 women without cysts. TABLE 3. Risk of functional ovarian cyst by marijuana use and body mass index,* Group Health Cooperative, Washington State, Marijuana use (n ¼ 104) Body mass index <20.0 Body mass index Body mass index >25.0 (n ¼ 134) ORy,z (n ¼ 240) (n ¼ 394) (n ¼ 178) (n ¼ 194) Current , , , 1.42 Former , , , 1.12 Never Reference Reference Reference * Weight (kg)/height (m) 2. y OR, odds ratio; CI, confidence interval. z Adjusted for subject age, reference year, and cigarette smoking. Excludes one case whose marijuana use was unknown.

5 524 Holt et al. To our knowledge, no prior study has investigated marijuana use in relation to functional ovarian cyst risk, either as a primary exposure or as a potential confounder of other associations. Although marijuana use did not confound our smoking analyses, we did find an increased ovarian cyst risk associated with marijuana use for underweight and normal-weight women. However, current marijuana use in our population was much less common than cigarette smoking, and our risk estimates were neither precise nor significant. Additionally, it is possible that marijuana users in our population smoked marijuana too infrequently to enable us to evaluate this potential association conclusively. A major strength of the current health maintenance organization-based study was the large sample size, which enabled us to examine smoking ovarian cyst relations precisely and test for effect modification. Rather than rely on a highly selective group of women who underwent surgery or ultrasound examination and were found not to have cysts, we chose controls randomly from Group Health Cooperative enrollment lists, assuring that they were representative of the well-defined population from which the cases arose. However, because we did not conduct diagnostic surgical or radiologic examinations of controls, the potential for inclusion of some number of undiagnosed cases in the control group existed in our study. Two cross-sectional studies of transvaginal ultrasound examination of randomly selected reproductive-age women found undiagnosed ovarian cysts in 6 7 percent of participants (14, 15). The percentage of controls who were undiagnosed cases may be somewhat lower in our study population for two reasons. There are no financial barriers to obtaining medical care within the Group Health Cooperative, so any symptomatic enrollee is likely to obtain medical attention that would lead to diagnosis. Additionally, we required 6 months of Group Health Cooperative enrollment prior to the reference month to ensure that all participants had the opportunity to establish care. We found that 83 percent of our controls had had Group Health Cooperative medical visits within 12 months of the reference month, at which asymptomatic cysts could be diagnosed. Because the opportunity for diagnosis did not differ appreciably by smoking status or BMI, the presence of a small number of undiagnosed cysts in our control group would likely have resulted in only a slight bias of our results toward the null. In contrast to two prior epidemiologic studies, we were able in this interview study to assess the ovarian cyst risk associated with current and former smoking separately. Similar to Parazzini et al. (8), we found higher risks for current smoking than for former smoking. Recognizing that decisions about the time since smoking cessation used to define subjects as former smokers may affect risk estimates (16), we decided to classify women as former smokers if they had last smoked 3 or more months prior to the reference month, for two reasons. First, functional ovarian cysts are thought to arise fairly rapidly and may regress spontaneously over the course of a few months; therefore, the etiologically relevant period is likely to be the few months prior to cyst diagnosis. Second, the effect of smoking on serum levels of the gonadotropin follicle-stimulating hormone is seen in current smokers only (3). Similarly, we did not use pack-years as a measurement of smoking behavior because neither duration nor intensity of past smoking affects levels of follicle-stimulating hormone. To our knowledge, this study is the first to determine the existence of heterogeneity of the association between smoking and ovarian cyst by BMI. The potential for such effect modification was suggested by the results of Wyshak et al. s 1988 study (6) of female college alumnae. In that study, the relation between smoking and cyst risk was stronger for women who had been athletes in college than for nonathletic women (relative risk ¼ 1.90 vs. relative risk ¼ 1.25). Because the authors found that former athletes were somewhat leaner than nonathletes, their finding of differential risk by college athleticism accords with our finding of effect modification by BMI. Hormonal alterations associated with smoking, marijuana use, and BMI in reproductive-age women may help explain our findings. Smoking affects serum gonadotropin levels, alteration of which is the hypothesized mechanism of cyst development (2). Compared with nonsmokers, smokers have lower follicular-phase serum levels of the gonadotropin luteinizing hormone (4) and higher levels of folliclestimulating hormone (3, 17). Additionally, in reproductiveage women, smoking alters production and/or metabolism of estrogen (18 22) and progesterone (4, 23). Finally, smoking affects menstrual function: smokers have shorter, more irregular cycles than nonsmokers do (24), with increased likelihood of anovulation and short luteal phase (5). Marijuana also affects reproductive hormones. Studies using 9-tetrahydrocannabinol, the active ingredient in marijuana, have shown a dose-related suppression of pulsatile luteinizing hormone secretion in rats (25) and decreased progesterone levels during the luteal phase of the menstrual cycle, irregular cycle length, and anovulation in monkeys (26, 27). Similar effects have been seen in women who smoke marijuana: substantial suppression of plasma levels of luteinizing hormone during the luteal phase of the menstrual cycle (9), shortened luteal phase and overall cycle length, and anovulation (28). Increased BMI also has hormonal effects in reproductive-age women. Leptin, a hormone produced by adipose tissue with circulating levels closely correlated with degree of adiposity, counteracts the inhibitory effect of fasting on secretion of luteinizing hormone (29); thus, we can speculate that leptin may counteract the luteinizing hormone-suppressive effects of smoking as well. In summary, our results indicate that increased BMI may attenuate the adverse impact of smoking on the risk of functional ovarian cyst. Further investigation of the combined effect of smoking and BMI in relation to ovarian cysts as well as other hormonally related conditions in reproductive-age women is warranted. ACKNOWLEDGMENTS This research was supported by grant 1 R01 HD from the US National Institute of Child Health and Human Development.

6 Smoking and Ovarian Cyst Risk according to Body Mass Index 525 REFERENCES 1. Velebil P, Wingo PA, Xia Z, et al. Rate of hospitalization for gynecologic disorders among reproductive-age women in the United States. Obstet Gynecol 1995;86: Hernandez E, Atkinson BF. Clinical gynecologic pathology. Philadelphia, PA: W B Saunders Co, Cooper GS, Baird DD, Hulka BS, et al. Follicle-stimulating hormone concentrations in relation to active and passive smoking. Obstet Gynecol 1995;85: Zumoff B, Miller L, Levit CD, et al. The effect of smoking on serum progesterone, estradiol, and luteinizing hormone levels over a menstrual cycle in normal women. Steroids 1990;55: Windham GC, Elkin EP, Swan SH, et al. Cigarette smoking and effects on menstrual function. Obstet Gynecol 1999;93: Wyshak G, Frisch RE, Albright TE, et al. Smoking and cysts of the ovary. Int J Fertil 1988;33: Holt VL, Daling JR, McKnight B, et al. Cigarette smoking and functional ovarian cysts. Am J Epidemiol 1994;139: Parazzini F, Moroni S, Negri E, et al. Risk factors for functional ovarian cysts. Epidemiology 1996;7: Mendelson JH, Mello NK, Elllingboe J, et al. Marihuana smoking suppresses luteinizing hormone in women. J Pharmacol Exp Ther 1986;237: Holt VL, Cushing-Haugen KL, Daling JR. Oral contraceptives, tubal sterilization, and functional ovarian cyst risk. Obstet Gynecol 2003;102: StataCorp. Stata statistical software: release 6.0. College Station, TX: Stata Corporation, Committee on Diet and Health. Diet and health: implications for reducing chronic disease risk. Washington, DC: National Academy Press, Maldonado G, Greenland S. Simulation study of confounderselection strategies. Am J Epidemiol 1993;138: Borgfeldt C, Andolf E. Transvaginal sonographic ovarian findings in a random sample of women years old. Ultrasound Obstet Gynecol 1999;13: Christensen JT, Boldsen JL, Westergaard JG. Functional ovarian cysts in premenopausal and gynecologically healthy women. Contraception 2002;66: Leffondré K, Abrahamowicz M, Siemiatycki J, et al. Modeling smoking history: a comparison of different approaches. Am J Epidemiol 2002;156: Cramer DW, Barbieri RL, Fraer AR, et al. Determinants of early follicular phase gonadotrophin and estradiol concentrations in women of late reproductive age. Hum Reprod 2002;17: MacMahon B, Trichopoulos D, Cole P, et al. Cigarette smoking and urinary estrogens. N Engl J Med 1982;307: Barbieri RL, Kistner RW. Endometriosis. In: Kistner RW, ed. Gynecology principles and practice. Chicago, IL: Year Book Medical Publishers, Baron JA, La Vecchia C, Levi F. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol 1990; 162: Westhoff C, Gentile G, Lee J, et al. Predictors of ovarian steroid secretion in reproductive-age women. Am J Epidemiol 1996;144: Van Voorhis BJ, Dawson JD, Stovall DW, et al. The effects of smoking on ovarian function and fertility during assisted reproduction cycles. Obstet Gynecol 1996;88: Berta L, Frairia R, Fortunati N, et al. Smoking effects on the hormonal balance of fertile women. Horm Res 1992;37: Kato I, Toniolo P, Koenig KL, et al. Epidemiologic correlates with menstrual cycle length in middle aged women. Eur J Epidemiol 1999;15: Murphy LL, Munoz RM, Adrian BA, et al. Function of cannabinoid receptors in the neuroendocrine regulation of hormone secretion. Neurobiol Dis 1998;5: Asch RH, Smith CG, Siler-Khodr TM, et al. Effects of delta 9-tetrahydrocannabinol during the follicular phase of the rhesus monkey (Macaca mulatta). J Clin Endocrinol Metab 1981;52: Almirez RG, Smith CG, Asch RH. The effects of marijuana extract and delta 9-tetrahydrocannabinol on luteal function in the rhesus monkey. Fertil Steril 1983;39: Abel EL. Marihuana and sex: a critical survey. Drug Alcohol Depend 1981;8: Caprio M, Fabbrini E, Isidori AM, et al. Leptin in reproduction. Trends Endocrinol Metab 2001;12:65 72.