Obstetrics & Gynecology:
Cigarette Smoking, Androgen Levels, and Hot Flushes in Midlife Women
Cochran, Chrissy J. PhD1; Gallicchio, Lisa PhD2; Miller, Susan R. ScD3; Zacur, Howard MD, PhD1,3; Flaws, Jodi A. PhD4
From the 1Division of Reproductive Biology, Department of Biochemistry and Molecular Biology, Johns Hopkins School of Public Health, Baltimore, Maryland; the 2Prevention and Research Center, Weinberg Center for Women’s Health & Medicine, Mercy Medical Center, Baltimore, Maryland; the 3Department of Gynecology and Obstetrics, Johns Hopkins University School of Medicine, Baltimore, Maryland; and the 4Department of Veterinary Biosciences, University of Illinois, Urbana, Illinois.
Supported by NIH grant AG18400 and a grant from the Women’s Health Research Group at the University of Maryland, Baltimore.
Corresponding author: Jodi A. Flaws, PhD, University of Illinois, Department of Veterinary Biosciences, 3223 VMBS Building, 2001 S. Lincoln Ave., Urbana, IL 61802; e-mail: email@example.com.
Financial Disclosure The authors have no potential conflicts of interest to disclose.
OBJECTIVE: To test the hypothesis that cigarette smoking is associated with hot flushes through a mechanism involving androgen levels, progesterone levels, sex hormone–binding globulin levels, or the ratio of androgens to estrogens.
METHODS: Women with and without hot flushes were recruited from Baltimore, Maryland, and the surrounding counties. Women were between 45 and 54 years of age, with at least three menstrual periods in the previous 12 months, and were not postmenopausal. Study participants completed a questionnaire and gave a blood sample for hormone measurements.
RESULTS: Current smokers had significantly higher androstenedione levels and a higher androgen-to-estrogen ratio than never smokers. Current smokers had significantly lower progesterone levels compared with never smokers. Former and current cigarette smokers had increased odds of experiencing hot flushes compared with never smokers (former: odds ratio [OR] 1.41, 95% confidence interval [CI] 0.99–2.01; current: OR 2.43, 95% CI 1.28–4.62). This association, however, was not attenuated by the addition of hormones to the smoking and hot-flush model.
CONCLUSION: Cigarette smoking is associated with hot flushes through a mechanism that may not involve alterations in hormone levels or their ratios.
LEVEL OF EVIDENCE: II
Hot flushes are the most common complaint of women during the menopausal transition.1 Despite the decades of research that have been performed, few risk factors for hot flushes have been identified. One of the most common risk factors studied in association with hot flushes is cigarette smoking. Our previous work showed that women who had ever smoked cigarettes had 1.6-fold increased odds of experiencing hot flushes compared with women who had never smoked cigarettes.2 In addition, women who were heavy smokers had fourfold increased odds of experiencing hot flushes compared with women who were never smokers.2 We hypothesized that cigarette smoking is associated with hot flushes because the chemicals in cigarette smoke lower estrogen levels in women,3–5 and low estrogen levels have been associated with hot flushes.1,6–9 This hypothesis, however, was not supported because estrogen levels did not mediate the association between cigarette smoking and hot flushes.2
It may be that cigarette smoke is modulating other hormone levels, therefore altering the odds of experiencing hot flushes. For example, higher androgen levels have been observed in smokers compared with nonsmokers.10–17 Although estrogen levels are similar in smokers and nonsmokers, higher androgen levels raise the androgen-to-estrogen ratio, which has been shown to be associated with hot flushes.18 Further, progesterone levels have been reported to be associated with smoking status,13 and low progesterone levels have been associated with hot flushes.18
Little is known about the association between androgen levels, progesterone levels, and smoking status or whether androgen and progesterone levels associated with smoking status mediate the association between cigarette smoking and hot flushes. Thus, a cross-sectional study was employed to study the associations between cigarette smoking and androgen levels, the androgen-to-estrogen ratio, and progesterone levels. In addition, we also examined whether the hormones associated with cigarette smoking mediated the association between cigarette smoking and hot flushes.
MATERIALS AND METHODS
Sample methods have been described in detail elsewhere.2,19 Briefly, this study was conducted among residents of Baltimore, Maryland, and the surrounding counties from 2000–2004. The names and addresses of women residing in the selected area who were 45 to 54 years of age were obtained from AccuData America (Fort Meyers, FL). All women on the list were mailed an invitation to participate in a cross-sectional study of midlife health. Interested participants contacted the clinic, and, if the staff determined they were eligible to participate, a visit was scheduled. All clinic visits were scheduled in the morning (8:30–10:00 am), and the women were instructed to fast overnight before the visit. At the clinic visit, participants were weighed, measured, and had their blood drawn for hormone assays. Each woman was eligible for study participation if she was between 45 and 54 years of age and had intact ovaries and a uterus. To ensure that women enrolled in the study were not postmenopausal, women were eligible only if they reported having at least 3 menstrual periods in the previous 12 months. Women were excluded if they were pregnant, were taking hormone replacement therapy or hormonal contraception, or had a history of cancer. All participants in this study gave written informed consent according to procedures approved by the University of Illinois, University of Maryland School of Medicine, and Johns Hopkins University institutional review boards.
The participants also completed the study-specific questionnaire, which included questions regarding demographic information, reproductive history, menstrual cycle characteristics, hormonal contraceptive use, menopausal symptoms, hormone replacement therapy use, medical and family history, smoking history, and alcohol consumption. A woman was considered to be an ever smoker if she answered yes to the question, “Have you ever smoked cigarettes?” A woman was considered to be a current smoker if she answered yes to the question, “Do you still smoke cigarettes?” Women also were asked to estimate how many cigarettes they smoke(d) per day and how many years they have (had) been smoking cigarettes. Current alcohol users were defined as those women who had consumed at least 12 alcoholic beverages in the previous 12 months. Body mass index (BMI) was calculated using the National Institutes of Health online BMI calculator. Body mass index is calculated as weight (kg)/[height (m)]2. Participants were characterized as normal weight (BMI 24.9 or lower), overweight (BMI 25.0–29.9), or obese (BMI 30.0 or higher).
Information also was collected via the study questionnaire about hot flushes. Specifically, information was collected on whether the woman had experienced hot flushes, whether the woman had a hot flush in the previous 30 days, the number of hot flushes experienced within the previous 30 days, the severity and frequency of hot flushes, and the length of time each woman had experienced hot flushes. In terms of severity, each woman was asked to describe her hot flushes as mild (sensation of heat without sweating), moderate (sensation of heat with sweating), or severe (sensation of heat with sweating that disrupts usual activity). In terms of frequency of hot flushes, each woman was asked to describe her hot flushes as occurring every hour, every 2–5 hours, every 6–11 hours, every 12–23 hours, 1–2 days per week, 3–4 days per week, 5–6 days per week, 2–3 days per month, 1 day per month, less than 1 day per month, or never. Each woman also was asked to describe the duration of her hot flushes as occurring for less than 1 month, 1–5 months, 6–11 months, 1–2 years, 3–4 years, or 5 years or longer.
Serum concentrations of sex hormone–binding globulin, estradiol (E2), estrone, testosterone, androstenedione, progesterone, and dehydroepiandrosterone sulfate (DHEAS) were measured using enzyme-linked immunosorbent assays (ELISAs). Enzyme-linked immunosorbent assay kits for E2, testosterone, androstenedione, and DHEAS were obtained from Diagnostic Systems Laboratories, Inc. (Webster, TX). Enzyme-linked immunosorbent assay kits for estrone, progesterone, and sex hormone–binding globulin were obtained from American Laboratory Products Company (Windham, NH). The assays were run using the manufacturers’ instructions. All assays were conducted in the same laboratory by a single investigator. All samples were run in duplicate, and mean values for each participant were used in the analysis. The laboratory personnel were blinded with respect to any information concerning study participants. For quality-control purposes, samples from women with hot flushes and women without hot flushes were run within the same laboratory batches. In addition, positive controls containing known amounts of E2, estrone, testosterone, androstenedione, progesterone, DHEAS, or sex hormone–binding globulin were included in each batch. Further, some samples were run in multiple assays to ensure that the assay values did not dramatically shift over time.
The minimum detection limits and intra-assay coefficients of variation were as follows: E2 7 pg/mL, 3.3±0.17%; estrone 10 pg/mL, 4.8±0.25%; testosterone 0.04 ng/mL, 2.2±0.56%; androstenedione 0.03 ng/mL, 2.5±0.60%; DHEAS 15 ng/mL, 1.9±0.63%; progesterone 0.1 ng/mL, 2.1±0.65%; and sex hormone–binding globulin 0.1 nmol/L, 2.4±0.67%. No E2, estrone, testosterone, androstenedione, DHEAS, or sex hormone–binding globulin samples were below the limit of detection. For progesterone samples that were below the limit of detection (n=66), the value was set at the limit of detection (0.1 ng/mL). The average inter-assay coefficient of variation for all assays was less than 5%.
The free E2 index was estimated using a conversion factor to change pg/mL of E2 to nmol/L: 100×(total E2×0.003671)/sex hormone–binding globulin. The free testosterone index also was estimated using a conversion factor to change ng/mL of testosterone to nmol/L: 100×(total testosterone×3.467)/sex hormone–binding globulin. The other hormones measured in this study do not bind appreciably to sex hormone–binding globulin; therefore, free index calculations for these hormones were not performed.
Sample size calculations were performed before starting the main study on hot flushes. Based on our preliminary data, we estimated that smokers would have a twofold increase in the odds of experiencing hot flushes compared with nonsmokers (95% confidence interval [CI] 1.2–3.5). We also assumed that about 20% of the women in the target population would be current smokers or have a lifetime history of smoking. Under these assumptions, we calculated that 165 women who reported experiencing hot flushes and 165 who reported never experiencing hot flushes would be necessary to detect a statistically significant twofold difference with a power of 80% and a 2-sided alpha of 0.05. Sample size calculations based on detecting significant differences in mean E2 levels between those women with and without hot flushes also were calculated based on data from published literature8; these calculations showed that, for the E2 analyses, 50 women with hot flushes and 50 women without hot flushes would be needed.
Women were excluded from this analysis if they were missing data on smoking status (one woman with hot flushes). In all analyses, hormone levels and BMI were log transformed because none of these variables were normally distributed when graphed. Characteristics of women with and without hot flushes and of smokers and nonsmokers were compared using χ2 analyses. Associations between hormone levels and smoking status were examined using general linear models. To examine whether cigarette smoking was associated with the experiencing of hot flushes, the odds ratio (OR) of experiencing hot flushes was calculated using a logistic regression model. To examine whether any association between cigarette smoking and hot flushes was due in part to hormone levels, each hormone that was found to be associated with smoking status was added to the confounder-adjusted cigarette-smoking and hot-flush model. Factors were considered confounders if they were associated (P<.1) with cigarette-smoking and hot-flush status. Variables remaining in the final analyses were age, race, BMI, alcohol use, and number of days since last menstrual period. The number of pack-years smoked was calculated as follows: (number of cigarettes smoked per day×the number of years smoked)/20. All analyses were performed using SPSS 11.0 (Chicago, IL). A P value of less than .05 was considered to be statistically significant.
Characteristics of the study sample are shown in Table 1. Women with hot flushes were more likely to be older, of black race, obese (BMI 30.0 or higher), and not current alcohol users compared with women without hot flushes. Of the women who reported ever experiencing hot flushes, approximately two thirds reported having hot flushes in the previous 30 days, roughly two thirds reported experiencing moderate or severe hot flushes, and about half reported having hot flushes on a daily or weekly basis.
Women who currently smoked cigarettes were more likely to be younger than women who formerly smoked or never smoked cigarettes (Table 2). In addition, women who currently smoked were more likely to be of black race and obese compared with women who formerly smoked or never smoked. Former smokers were more likely to consume alcohol than never smokers. In regards to hot flushes, current and former smokers were more likely to have experienced recent (in the previous 30 days), moderate, or severe hot flushes compared with never smokers. There was no difference reported, however, in the frequency of hot flushes between the smoking groups.
Mean levels of androstenedione and progesterone and the ratio of total androgens to total estrogens differed significantly by smoking status (Table 3). Specifically, former and current smokers had higher androstenedione levels than did women who had never smoked cigarettes. Conversely, progesterone levels were lower in current smokers compared with former and never smokers. Further, there was a tendency for a higher ratio of total androgens to total estrogens among women who reported being smokers at the time of enrollment into the study (P=.04). Levels of sex hormone–binding globulin and the free estrogen index and the free testosterone index were similar between all categories of smokers (data not shown).
Table 4 shows the relation between cigarette smoking and hot flushes. In agreement with our previously published data,2 the confounder-adjusted (age, race, BMI, alcohol use, and days since last menstrual period) odds of experiencing any hot flushes were significantly increased in women who had ever smoked (OR 1.55, 95% CI 1.11–2.17) compared with women who had never smoked. When ever smokers were separated into former and current smokers, the former smokers were 1.4 times and the current smokers 2.4 times more likely to report experiencing any hot flushes compared with never smokers. For women who ever smoked cigarettes, the greater number of pack-years smoked was associated with greater odds of reporting hot flushes (1–25 pack-years OR 1.36, 95% CI 0.96–1.95; more than 25 pack-years OR 2.59, 95% CI 1.32–5.07) compared with women who never smoked cigarettes. In addition, although the numbers were small, current smokers who had smoked 1–25 pack-years were 6.2 times more likely to experience any hot flushes compared with never smokers. There were no women without hot flushes in the more than 25 pack-years category. In addition, current smokers who smoked up to a pack of cigarettes a day were twice as likely to experience hot flushes compared with never smokers.
The odds of experiencing any hot flushes remained significantly elevated, and relatively unchanged, for women who were ever smokers compared with never smokers after additional adjustment for those hormones shown to be associated with smoking: androstenedione (OR 1.57, 95% CI 1.12–2.19), progesterone (OR 1.52, 95% CI 1.08–2.13), and T+ASD/E2+E1 (OR 1.49, 95% CI 1.06–2.09). The odds of experiencing hot flushes were also relatively unchanged for the other categories of cigarette smoking examined when the hormone variables were added to the model.
In addition to examining the association between cigarette smoking and any hot flushes, the odds of experiencing hot flushes within the previous 30 days, moderate or severe hot flushes, and hot flushes that last for a period of a year or more were examined. Results similar to those seen for cigarette smoking and any hot flushes were observed (data not shown).
In this study, we observed that current cigarette smoking is associated with higher androstenedione levels, a higher total androgen–to–total estrogen ratio, and lower progesterone levels. In addition, we observed that hormone levels did not mediate the association between current cigarette smoking and the experiencing of hot flushes.
Our data indicating that androstenedione levels are higher in current smokers are consistent with several previous studies.10,12–17,20 Androstenedione is predominately derived from the adrenal gland; therefore, these data are also consistent with the observation that one mode of action of nicotine in cigarette smoke is to stimulate the adrenal gland.21 Our data suggest that smoking may permanently alter the amount of androstenedione produced by the adrenal gland because the level of androstenedione was similar among current and former smokers but was significantly lower in never smokers. In addition, androstenedione levels in former smokers were not significantly different by time since smoking cessation (data not shown). In addition, nicotine may stimulate the pituitary gland to produce adrenocorticotropic hormone, which also stimulates the adrenal gland to produce more androstenedione.15 Longcope and Johnston12 also have suggested that there is a decrease in the clearance of androstenedione in smokers compared with nonsmokers, leading to increased androstenedione levels. Luteinizing hormone levels also have been shown to be increased in smokers,12 suggesting, along with increased adrenocorticotropic hormone levels, that nicotine is stimulating the pituitary and hypothalamus. The hypothalamus contains nicotinic cholinergic receptors. When nicotine binds to these receptors, increases in heart rate, blood pressure, and respiration become apparent.22 These are some of the same “symptoms” of hot flushes. Because higher androstenedione levels are noted in smokers but do not mediate the association between smoking and hot flushes, androstenedione is not likely part of the causal pathway of hot flushes and is therefore an additional effect of cigarette smoking. Because testosterone, E2, and estrone levels were not affected by cigarette smoking, it makes sense to find a higher total androgen–to–total estrogen ratio in current smokers compared with nonsmokers because androstenedione levels are elevated.
Our results do not concur with previous studies that measured progesterone levels in smokers. We report significantly lower progesterone levels in smokers compared with nonsmokers. In previous studies of premenopausal women, there was no difference detected in progesterone levels between smokers and nonsmokers.11,23 In a study of postmenopausal women, however, progesterone levels were noted to be higher in smokers compared with nonsmokers.13 There may be several reasons for these differences. All of the previous studies had small sample sizes, and the ages of the women were different than in our study. Premenopausal smokers also seem to have a more robust hypothalamus-pituitary-ovarian axis that allows them to maintain hormone levels that are similar to those of nonsmokers. This is shown in numerous studies of premenopausal smokers and nonsmokers who have similar levels of all hormones studied.11–12,23–25
We previously showed that low progesterone levels are associated with experiencing hot flushes.18 Therefore, we were surprised to find that, although progesterone levels were significantly lower in smokers compared with nonsmokers, progesterone did not mediate the association between smoking and hot flushes. This finding indicates that progesterone is not within the causal pathway of cigarette smoking and its relation to hot flushes.
Hot flushes have been studied for many years, and a good, simple, reliable method of determining hot flushes (other than by interview or questionnaire) has not been developed for wide use in large studies. Freedman26 has performed studies measuring various physiological (core body temperature, respiratory exchange ratio, skin temperature, skin conductance, sweat rate) and biological (3-methoxy-4-hydroxyphenylglycol) parameters in association with hot flushes. Unfortunately, none of these measurements correlates with a woman’s perception of a hot flush 100% of the time. The work to develop more sensitive approaches for collecting hot flush data may allow the association between androgen levels, hot flushes, and cigarette smoking to be explored more fully in the future.
There are several limitations of this study that may affect the interpretation of results. First, smoking history was obtained by self-report; therefore, there may be some misclassification of exposure in this study. There is no reason to suspect, however, that this misclassification would differ based on hot flush status, and, hence, the estimates reported would be biased toward the null and may actually be larger than those observed. In addition, because this study was of a cross-sectional design, we were unable to determine the temporality of the association between smoking and hot flushes. Based on questionnaire data, it seemed that age of smoking preceded age of hot flushes among those who were both smokers and reported experiencing hot flushes. Finally, we collected only one blood sample per participant in the study. Variability was limited by collecting all samples during the same 2-hour window in the morning after an overnight fast and adjusting analyses involving hormone levels for number of days since last menstrual period.
From the data presented here and in previous studies, it seems that cigarette smoking stimulates the production of androstenedione from the adrenal gland. Although androstenedione levels are increased and progesterone levels are decreased in smokers compared with nonsmokers, hormone levels may not be involved in the association between cigarette smoking and hot flushes. Rather, the effect is probably a more direct stimulatory action of nicotine on the nicotinic receptors in the hypothalamus.
1. Overlie I, Moen MH, Holte A, Finset A. Androgens and estrogens in relation to hot flushes during the menopausal transition. Maturitas 2002;41:69–77.
2. Gallicchio L, Miller SR, Visvanathan K, Lewis LM, Babus J, Zacur H, et al. Cigarette smoking, estrogen levels, and hot flashes in midlife women. Maturitas 2006;53:133–43.
3. Whiteman MK, Staropoli CA, Benedict JC, Borgeest C, Flaws JA. Risk factors for hot flashes in midlife women. J Women’s Health 2003;12:459–72.
4. Whiteman MK, Staropoli CA, Langenberg PW, McCarter RJ, Kjerulff KH, Flaws JA. Smoking, body mass, and hot flashes in midlife women. Obstet Gynecol 2003;101:264–72.
5. Li C, Samsioe G, Borgfeldt C, Lidfeldt J, Agardh CD, Nerbrand C. Menopause-related symptoms: what are the background factors? A prospective population-based cohort study of Swedish women (The Women’s Health in Lund Area study). Am J Obstet Gynecol 2003;189:1646–53.
6. Woods NF, Smith-Dijulio K, Percival DB, Tao EY, Taylor HJ, Mitchell ES. Symptoms during the menopausal transition and early postmenopause and their relation to endocrine levels over time: observations from the Seattle midlife women’s health study [published erratum appears in J Womens Health (Larchmt) 2007;16:1379]. J Womens Health (Larchmt) 2007;16:667–77.
7. Randolph JF Jr, Sowers M, Bondarenko I, Gold EB, Greendale GA, Bromberger JT, et al. The relationship of longitudinal change in reproductive hormones and vasomotor symptoms during the menopausal transition. J Clin Endocrinol Metab 2005;90:6106–12.
8. Erlik Y, Meldrum DR, Judd HL. Estrogen levels in postmenopausal women with hot flashes. Obstet Gynecol 1982;59:403–7.
9. Visvanathan K, Gallicchio L, Schilling C, Babus JK, Lewis LM, Miller SR, et al. Cytochrome gene polymorphisms, serum estrogens, and hot flushes in midlife women. Obstet Gynecol 2005;106:1372–81.
10. Bancroft J, Cawood EH. Androgens and the menopause; a study of 40-60 year-old women. Clin Endocrinol 1996;45:577–87.
11. Key TJ, Pike MC, Baron JA, Moore JW, Wang DY, Thomas BS, et al. Cigarette smoking and steroid hormones in women. J Steroid Biochem Mol Biol 1991;39:529–34.
12. Longcope C, Johnston CC Jr. Androgen and estrogen dynamics in pre- and postmenopausal women: a comparison between smokers and nonsmokers. J Clin Endocrinol Metab 1988;67:379–83.
13. Friedman AJ, Ravnikar VA, Barbieri RL. Serum steroid hormone profiles in postmenopausal smokers and nonsmokers. Fertil Steril 1987;47:398–401.
14. Cauley JA, Gutai JP, Kuller LH, LeDonne D, Powell JG. The epidemiology of serum sex hormones in postmenopausal women. Am J Epidemiol 1989;129:1120–31.
15. Schlemmer A, Jensen J, Riis BJ, Christiansen C. Smoking induces increased androgen levels in early post-menopausal women. Maturitas 1990;12:99–104.
16. Khaw KT, Tazuke S, Barrett-Connor E. Cigarette smoking and levels of adrenal androgens in postmenopausal women. N Engl J Med 1988;318:1705–9.
17. Manjer J, Johansson R, Lenner P. Smoking as a determinant for plasma levels of testosterone, androstenedione, and DHEAs in postmenopausal women. Eur J Epidemiol 2005;20:331–7.
18. Schilling C, Gallicchio L, Miller SR, Langenberg P, Zacur H, Flaws JA. Genetic polymorphisms, hormone levels, and hot flashes in midlife women. Maturitas 2007;57:120–31.
19. Schilling C, Gallicchio L, Miller SR, Babus JK, Lewis LM, Zacur H, et al. Current alcohol use is associated with a reduced risk of hot flashes in midlife women. Alcohol Alcohol 2005;40:563–8.
20. Baron JA, Comi RJ, Cryns V, Brinck-Johnsen T, Mercer NG. The effect of cigarette smoking on adrenal cortical hormones. J Pharmacol Exp Ther 1995;272:151–5.
21. Ramos KS, Chacon E, Acosta D Jr. Toxic responses of the heart and vascular systems. In: Klaassen CD, editor. Casarett & Doull’s toxicology: the basic science of poisons. 5th ed. New York (NY): McGraw-Hill; 1996.;p. 487–527.
22. Anthony DC, Montine TJ, Graham DG. Toxic responses of the nervous system. In: Klaassen CD, editor. Casarett & Doull’s toxicology: the basic science of poisons. 5th ed. New York (NY): McGraw-Hill; 1996.;p. 463–86.
23. Thomas EJ, Edridge W, Weddell A, McGill A, McGarrigle HH. The impact of cigarette smoking on the plasma concentrations of gonadotrophins, ovarian steroids and androgens upon the metabolism of oestrogens in the human female. Hum Reprod 1993;8:1187–93.
24. Law MR, Cheng R, Hackshaw AK, Allaway S, Hale AK. Cigarette smoking, sex hormones and bone density in women. Eur J Epidemiol 1997;13:553–8.
25. Daniel M, Martin AD, Drinkwater DT. Cigarette smoking, steroid hormones, and bone mineral density in young women. Calcif Tissue Int 1992;50:300–5.
26. Freedman RR. Biochemical, metabolic, and vascular mechanisms in menopausal hot flashes. Fertil Steril 1998;70: 332–7.
© 2008 The American College of Obstetricians and Gynecologists