Approximately 40%–70% of perimenopausal women experience hot flashes, and many of these women seek medical attention for treatment of their hot flashes.1,2 Hot flashes negatively impact the quality of life for women by causing sleep disturbances that often result in fatigue, irritability, forgetfulness, acute physical discomfort, and negative effects on work.3 Additionally, hot flashes may be associated with the development of serious medical conditions, such as Alzheimer disease,4 osteoporosis,5 and depression.6
Despite the impact of hot flashes on women's lives, few studies have attempted to identify factors that may predispose women to hot flashes. Such information would be valuable for identifying women at high risk for hot flashes as they approach midlife and for developing appropriate prevention strategies that may include lifestyle modifications. The few published studies that have examined risk factors for hot flashes suggest that potentially modifiable factors associated with lifestyle, such as smoking7–10 or high body mass index (BMI)9,11,12 may influence risk for hot flashes. However, many of these studies have been limited by small sample size, lack of detail on hot flash history, and lack of information on potential confounding factors. Furthermore, few studies have examined risk factors in relation to more severe or frequent hot flashes, which cause the most disruption in quality of life for women. Additionally, most studies of hot flashes have been conducted in clinic populations that may be more likely to report menopausal symptoms or to have spoken with a physician about menopause.13 Thus, the purpose of this study was to investigate potentially modifiable risk factors for hot flashes. To do this, we assessed the associations between smoking and BMI with any, moderate to severe, and daily hot flashes in a population-based sample of women in the Baltimore metropolitan area.
MATERIALS AND METHODS
A cross-sectional study of midlife women was conducted in 2001 among residents of the Baltimore metropolitan region who reported their history of hot flashes and other information through a mailed survey. The University of Maryland Institutional Review Board approved all aspects of the study. The target population for this study included women aged 40–60 years residing in the Baltimore Metropolitan Statistical Area, as defined by the U.S. Bureau of the Census.
A listing of all postal zip codes in the Baltimore Metropolitan Statistical Area and the percent of nonwhite (non-Hispanic) individuals residing in each zip code was obtained from 1990 U.S. Census Data. It was estimated that an adequate sample size could be obtained from mailing surveys to six zip codes in the Baltimore Metropolitan Statistical Area. In an effort to recruit minority women, four of the six zip codes to which surveys were mailed were randomly selected from among 21 zip codes with at least 30% nonwhite residents. To help ensure an adequate response rate, the remaining two zip codes were selected for their high response rates to previous mailed women's health surveys.
Names and addresses of women in the selected age range residing in the six selected zip codes were obtained from a commercial mailing house that compiles names from public sources (Department of Motor Vehicles, voter registration). Surveys were mailed to a random sample of 3000 names selected from the 13,378 names on this list. All surveys were sent by first class metered mail along with an introductory letter and a postage-paid, self-addressed return envelope. As part of an accompanying trial, women were randomized to receive experimental incentives (introductory postcard, money, lottery ticket) to assess the incentives' effect on response rates. All persons who had not returned a survey within 21 days after the initial mailing were sent a reminder postcard.
Surveys that were not delivered because of incorrect addresses (n = 81) or those that were returned by the recipient indicating that they were ineligible because of age (n = 16) or gender (n = 1) were re-mailed to randomly chosen individuals remaining from the original list of names. One survey with an incorrect address was returned after the completion of the study and was therefore not replaced. Thus, the total number of mailed surveys was 2999.
Although the main outcome of this study was hot flashes, efforts were made to avoid reporting bias by presenting the survey as part of a general study called “Study of Women's Health in Midlife.” The 15-page, double-sided survey took 15–25 minutes to complete and asked questions regarding demographic information, pregnancy history, hormonal contraceptive use, menstrual history, menopausal symptoms, hormone replacement therapy (HRT) use, medical and family history, health behaviors (smoking, alcohol use, vitamin use), and attitudes toward menopause.
A detailed hot flash history was obtained though a series of questions that asked for information on the following: whether the woman had ever had hot flashes; whether the hot flashes may have occurred for reasons other than approaching menopause (pregnancy, medication, illness); the ages when hot flashes occurred; the severity and frequency of hot flashes; and whether medical attention was sought because of hot flashes. In all analyses, three outcomes were examined: any hot flashes, moderate to severe hot flashes, and daily hot flashes. A woman was considered as having any hot flashes if she indicated any hot flashes that were due to approaching menopause and not due to other factors. A woman was classified as having moderate to severe hot flashes if she had hot flashes not due to other factors that were described as a sensation of heat accompanied by sweating that may have interrupted usual activity. Daily hot flashes were hot flashes not due to other factors that occurred on an average of at least one per day.
A detailed smoking history was obtained that included information on the ages when the woman smoked and the average amount smoked per day during those times. For analysis, smoking was examined as current or former smoking, amount of cigarettes currently smoked, and pack-years of smoking. Body mass index was calculated in kg/m2 from self-reported height and weight at the time of the survey. Body mass index was categorized for analysis as 24.9 or less, 25.0–29.9, or 30 or more kg/m2, in accordance with the World Health Organization's definitions of normal, overweight, and obese.
Women were asked about their menstrual history, including whether they had experienced any changes in bleeding or regularity, as well as reasons for any cessation of menses. From this information, menopausal status was defined in five categories: Premenopausal: last menstrual period (LMP) within the past 3 months and no changes in bleeding or regularity in the past year and no hysterectomy or bilateral oophorectomy; Perimenopausal: 1) LMP within the past year but not within the past 3 months and no hysterectomy or bilateral oophorectomy, or 2) LMP within the past 3 months and changes in either bleeding or regularity in the past year and no hysterectomy or bilateral oophorectomy; Naturally postmenopausal: LMP more than 12 months ago and periods did not stop because of surgery, chemotherapy, or hormones; Artificially postmenopausal: periods stopped because of surgical removal of the uterus (with or without removal of ovaries) or chemotherapy/radiation; Hormones: this category in included women who stated that their periods stopped because of taking hormones (oral contraceptives or HRT), as well as premenopausal or perimenopausal women currently taking hormones.
For analysis, exclusions included those with missing information regarding ever having hot flashes (n = 1), menopausal status (n = 22), or current age (n = 5). Additionally, women reporting Turner syndrome (n = 2) or premature ovarian failure (n = 12) were excluded because they likely have different hormonal disturbances and thus different risk factors for hot flashes than other women. Respondent numbers differ for some analyses because of missing data for some questions.
The population characteristics of those with and without any, moderate to severe, and daily hot flashes were compared using χ2 tests. The unadjusted associations with the risk factor of interest (smoking and BMI) and each of the three hot flash outcome variables were assessed with unadjusted odds ratios (ORs). Interactions of age, menopausal status, race, smoking, and BMI with the respective independent variables were first examined by stratification to determine if the stratum-specific estimates of the OR differed from each other. If there was an indication of an interaction from these stratified analyses, the statistical significance of the interaction terms was examined in logistic regression models. Other population factors were examined for their associations with the risk factors of interest to assess which may act as confounders.
Unconditional logistic regression was used to estimate the association between each of the independent variables (measures of smoking history and BMI) and the three hot flash outcome variables, adjusting for potential confounders. Models were first fit adjusting for age and menopausal status. Factors were then included in the models if they were associated with both hot flashes and the independent variable of interest or if they were strongly associated with hot flashes. If significant interaction terms were detected (P < .05), models were fit separately in each strata.
The mean age of the study population was 49.9 years. The study population was predominately nonblack (80.4%), and the participants tended to be of higher socioeconomic status, with over half reporting at least a college degree and more than half reporting an annual income of at least $61,000 (Table 1). At the time of the survey, 26% of participants were taking HRT. Any hot flashes were reported by over half of the population. Almost 40% of the population reported moderate to severe hot flashes, and almost 20% reported daily hot flashes. Close to half of women reporting any hot flashes (43%) had sought medical treatment for their hot flashes.
More than half of the study population (51%) smoked during their life; 16% of the population were smokers at the time of the survey. Of these current smokers, 43% smoked more than 1 pack of cigarettes on average per day, and 41% smoked for more than 25 pack-years.
The results of analyses estimating the association between measures of smoking history and any, moderate to severe, and daily hot flashes are shown in Table 2. Current smokers were more likely to report any hot flashes compared with those who never smoked, although the increase in risk was no longer significant after adjustment for age, menopausal status, BMI, race, HRT use, herbal supplement use, nulliparity, and tubal ligation (OR = 1.47, 95% confidence interval [CI] 0.94, 2.29). Former smokers were not significantly more likely to report any hot flashes than those who never smoked. Among current smokers, the number of cigarettes smoked did not seem to affect risk for any hot flashes, though those who smoked for a greater number of pack-years (more than 25) had higher odds of any hot flashes compared with never-smokers than those smoking fewer pack-years.
Current smoking, as well as the amount of cigarettes smoked and pack-years of smoking, were associated with the occurrence of moderate to severe hot flashes (Table 2). Current smokers had 1.9 times the odds of never smokers for reporting moderate to severe hot flashes. Although a greater proportion of former smokers reported moderate to severe hot flashes than never smokers, this difference was not significant. Among current smokers, the amount smoked was associated with a trend of increasing odds: those currently smoking more than 1 pack of cigarettes per day were more than 2.5 times more likely to report moderate to severe hot flashes than never-smokers (adjusted OR = 2.68), whereas those smoking 1 pack of cigarettes or less per day were only somewhat more likely to report moderate to severe hot flashes than never-smokers (adjusted OR = 1.53). Similarly, current smokers who smoked the equivalent of more than 25 pack-years were at a higher risk for moderate to severe hot flashes (adjusted OR = 2.98) compared with never-smokers than were those smoking 1–25 pack-years (adjusted OR = 1.43). Results for daily hot flashes were similar to those for moderate to severe hot flashes (Table 2). The associations between smoking history and any, moderate to severe, and daily hot flashes were consistent across strata defined by menopausal status, age, HRT use, race, and BMI.
The results of analyses estimating the association between BMI and moderate to severe hot flashes are shown in Table 3. The association between BMI and any and daily hot flashes varied across subgroups defined by age; these results are shown in Table 4. Increasing BMI was associated with an increased risk for moderate to severe hot flashes (Table 3). The odds of reporting hot flashes was more than two times greater in those with BMI of 30 kg/m2 or more compared with those with BMI of 24.9 kg/m2 or less (adjusted OR = 2.11). Those in the middle BMI group (25.0–29.9 kg/m2) had a slightly higher odds for moderate to severe hot flashes compared with those in the lowest BMI group (adjusted OR = 1.22). This association between increased odds for moderate to severe hot flashes with increasing BMI was consistent across strata defined by age, menopausal status, smoking status, and race.
Similar to the results for moderate to severe hot flashes, there was an indication that those with higher BMI were more likely than those with lower BMI to report any hot flashes in the entire study population. However, the association differed significantly by both age (P = .02 for interaction) and menopausal status (P = .03). Increasing BMI was associated with an increased risk for any hot flashes in woman aged 40–50, but not among those aged 51–60 (Table 4). The association between increasing BMI and risk for any hot flashes was present among premenopausal or perimenopausal women, but not postmenopausal women (data not shown). As with any hot flashes, the association between BMI and daily hot flashes differed significantly by both age (P = .03 for interaction) and menopausal status (P = .04). When stratified by age, there was a suggestion of an association between BMI and daily hot flashes among women aged 40–50, but not among those aged 51–60 (Table 4). Results were similar from analyses stratified by menopausal status; there was a suggestion of an association between increasing BMI and daily hot flashes only among premenopausal or perimenopausal women.
The results of this cross-sectional, population-based survey suggest that potentially modifiable factors, such as current smoking and high BMI, may predispose a woman to more severe or frequent hot flashes. Furthermore, among current smokers there is a trend toward increasing risk for any, moderate to severe, and daily hot flashes with increasing amount smoked and pack-years of smoking. These findings are consistent with several other studies that also report any smoking to be associated with an increased risk for any hot flashes7–9 and for bothersome hot flashes.10 Additionally, some investigators have reported an increased risk for hot flashes with increasing pack-years of smoking8 or, to a lesser extent, with amount currently smoked.10
However, two studies reported results that are inconsistent with those found in this study. In a small study (n = 334) of naturally postmenopausal women, smoking at the time of menopause was not independently associated with hot flashes in the entire study population, but it was associated with an increased risk for hot flashes in women with lower BMI.1 Guthrie et al found no significant difference in smoking status (current versus not current) between women who experienced hot flashes and those who did not experience hot flashes in a population-based sample of women aged 48–59 in Australia.14 This lack of association may stem from the fact that women were asked whether they had been bothered by hot flashes in the preceding 2 weeks instead of obtaining a detailed hot flash history, as was done in this study.
Former smokers in this study did not have an increased risk for any or moderate to severe hot flashes, compared with never-smokers. Although former smokers were at a modestly increased risk for daily hot flashes compared with never-smokers, this increase was not as great as that for current smokers. Similar findings were reported by Gold et al, who found that former smokers in a large, community-based sample of women, although at an increased risk for hot flashes, were less likely than current smokers to report hot flashes.9
Although the precise role smoking may play in the etiology of hot flashes is unclear, it likely involves changes in endogenous estrogen levels. Several reports have noted lower active estrogen levels in smokers than nonsmokers, and collectively these reports suggest that smoking could alter estrogen metabolism by several pathways.15–27 Cigarette smoking may directly deplete estrogen levels by interacting with the CYP450 enzyme system,15–18,25–27 which is responsible for the metabolism of the chemicals in cigarette smoke as well as for the metabolism of estrogen. The chemicals in cigarette smoke may reduce the conversion of androgens to estrogens by inhibiting aromatase activity.28 Additionally, the chemicals in cigarette smoke may reduce estrogen levels by destroying ovarian follicles25,29–31 or may indirectly alter estrogen metabolism by reducing body weight.1,32
It is possible that the findings in this study regarding smoking and hot flashes are due to inadequate control for potential confounders, as smokers typically suffer more comorbid conditions and are generally less health conscious than nonsmokers. These factors, in turn, may also be associated with the occurrence of hot flashes. This possibility was explored using data available in this study by controlling for such factors as attitudes toward menopause, depressive symptoms, and other comorbid conditions, including thyroid disease and heart disease. Controlling for these factors did not alter the association between smoking and hot flashes. Furthermore, as current smoking is consistently associated with earlier age at natural menopause,33,34 the association was further explored both controlling for and stratifying by menopausal status, with no change in the results.
In this study population, BMI was positively associated with the risk for moderate to severe hot flashes in the entire study population and with the risk for any or daily hot flashes among younger women (aged 40–50) and premenopausal or perimenopausal women, but not among older women (aged 51–60) or postmenopausal women. Several other studies have also found the risk of hot flashes to be higher in women with higher BMI than in those with lower BMI.9,11,12 Schwingl et al found that the effects of BMI on hot flash risk varied by smoking status, in that higher BMI was only associated with hot flashes in smokers.1 In agreement with the hypothesis that heavier women have higher estrogen levels and in contrast to the present study, some investigators found that thinner women were more likely to report hot flashes than heavier women.35,36 However, neither of these studies used BMI as their measure of body size, instead using measures that may capture slightly different information about body composition, such as percent ideal weight36 or number of kilograms overweight.35
The finding in this study that the effect of higher BMI on the risk for any or daily hot flashes existed only in younger or premenopausal or perimenopausal women is consistent with the results of two other studies.9,37 In a sample of women from a population-based breast cancer screening project in the Netherlands, den Tonkelaar et al found the risk of hot flashes to be higher in heavier women only among those aged 40–44, but not in those aged 54–69.37 Additionally, Gold et al reported in a cross-sectional analysis of the Study of Women Across the Nation that, despite an overall association of high BMI with reporting of hot flashes, there was no such association in postmenopausal or late perimenopausal women.9
There is evidence to suggest that obesity can lead to earlier ovarian insufficiency, as Klinga et al found that obese women experienced increased levels of follicle-stimulating hormone and decreased levels of estradiol an average of 4 years earlier than nonobese women.38 Thus, the association between BMI and hot flashes may exist only for younger women because of earlier entry into the menopausal transition or a longer perimenopausal period, which itself is associated with an increased risk for hot flashes.39 As the majority of older and/or postmenopausal women have already experienced hot flashes, there may not be an effect for BMI on the risk of ever having hot flashes. Alternatively, high BMI may protect against hot flashes in older women, who are more likely than younger women to be estrogen deficient, through peripheral conversion of androgens to estrogens in fat tissue.1,9,10,36
There is a widely accepted clinical hypothesis that women with low body weight are more likely to experience hot flashes than heavier women. The prevailing wisdom is that obese women have higher levels of estrogen than thin women because of greater peripheral conversion of androgens to estrogens in fat tissue.1,9,10,36 The results of this and other studies that find increasing BMI to be associated with risk for hot flashes contradict the long-standing hypothesis that heavier women have higher endogenous estrogen levels and thus a lower risk for hot flashes. In addition to evidence suggesting that obese women may experience earlier ovarian insufficiency,38 other studies lend biological support to the findings of this study. While forming estrogens from circulating androgen precursors, adipose tissue also produces hormones (leptin, tumor necrosis factor α) that may suppress ovarian steroid production40,41 and may influence thermoregulation,42 thus impacting the risk for hot flashes.
The cross-sectional nature of this study limits the conclusions that can be drawn. As a cross-sectional study, we cannot be certain that the risk factors of interest temporally preceded and thus predisposed to the occurrence of hot flashes. However, for certain variables, such as smoking, hysterectomy, oophorectomy, and tubal ligation, the ages when these exposures occurred were recorded, as were the ages when hot flashes occurred. Although current BMI was used as a proxy for BMI at the time of hot flashes, it is unlikely that BMI changed substantially for most women. Thus, with reasonable certainty, it can be assumed that smoking and high BMI preceded the occurrence of hot flashes. However, as the underlying biological changes that lead to hot flashes are unknown, it is difficult to comment on whether the exposures of interest occurred prior to these biological changes. Thus, prospective studies are still needed to establish the temporal sequence of events leading to hot flashes.
This study, together with previous reports, suggests that modifiable factors, such as smoking or high BMI, may increase a woman's risk for hot flashes. The research in this area is sparse and marked by methodological limitations, so additional studies are needed to confirm these findings. Although HRT is an effective treatment for hot flashes in most women, others find no relief with this therapy, and there are many women for whom estrogen is contraindicated or for whom the potential risks associated with HRT are major concerns. Thus, it is imperative that additional research be conducted to further elucidate lifestyle factors that may be associated with risk for hot flashes so as to offer alternatives to HRT, such as lifestyle modifications. Furthermore, as factors that predispose women to hot flashes are clarified, preventive measures can be targeted to women at greatest risk for hot flashes. Such measures may help to maintain women's quality of life as they go through menopause as well as prevent some conditions associated with the hormonal changes that accompany menopause.
1. Schwingl PJ, Hulka BS, Harlow SD. Risk factors for menopausal hot flashes. Obstet Gynecol 1994;84:29–34.
2. Blumberg G, Kaplan B, Rabinerson D, Goldman GA, Kitai E, Neri A. Women's attitudes towards menopause and hormone replacement therapy. Int J Gynaecol Obstet 1996;54:271–7.
3. Oldenhave A, Jaszmann LJ, Haspels AA, Everaerd WT. Impact of climacteric on well-being. Am J Obstet Gynecol 1993;168:772–80.
4. Birge SJ. Is there a role for estrogen replacement therapy in the prevention and treatment of dementia? J Am Geriatr Soc 1996;44:865–70.
5. Lee SJ, Kanis JA. An association between osteoporosis and premenstrual symptoms and postmenopausal symptoms. Bone Miner 1994;24:127–34.
6. Avis NE, Crawford S, Stellato R, Longcope C. Longitudinal study of hormone levels and depression among women transitioning through menopause. Climacteric 2001;4:243–9.
7. Kuh DL, Wadsworth M, Hardy R. Women's health in midlife: The influence of the menopause, social factors and health in earlier life. Br J Obstet Gynaecol 1997;104:923–33.
8. Staropoli CA, Flaws JA, Bush TL, Moulton AW. Predictors of menopausal hot flashes. J Womens Health 1998;7:1149–55.
9. Gold EB, Sternfeld B, Kelsey JL, Brown C, Mouton C, Reame N, et al. Relation to demographic and lifestyle factors to symptoms in a multi-racial/ethnic population of women 40–55 years of age. Am J Epidemiol 2000;152:463–73.
10. Avis NE, Crawford SL, McKinlay SM. Psychosocial, behavioral, and health factors related to menopause symptomatology. Womens Health 1997;3:103–20.
11. Wilbur J, Miller AM, Montgomery A, Chandler P. Socio-demographic characteristics, biological factors, and symptom reporting in midlife women. Menopause 1998;5:43–51.
12. Freeman EW, Sammel MD, Grisso JA, Battistini M, Garcia-Espagna B, Hollander L. Hot flashes in the late reproductive years: Risk factors for African American and Caucasian women. J Womens Health Gend Based Med 2001; 10:67–76.
13. Leidy LE, Canali C, Callahan WE. The medicalization of menopause: Implications for recruitment of study participants. Menopause 2000;7:193–9.
14. Guthrie JR, Dennerstein L, Hopper JL, Burger HG. Hot flushes, menstrual status, and hormone levels in a population-based sample of midlife women. Obstet Gynecol 1996;88:437–42.
15. Cassidenti DL, Vijod AG, Vijod MA, Stanczyk FZ, Lobo RA. Short-term effects of smoking on the pharmacokinetic profiles of micronized estradiol in postmenopausal women. Am J Obstet Gynecol 1990;163:1953–60.
16. Westhoff C, Gentile G, Lee J, Zacur H, Helbig D. Predictors of ovarian steroid secretion in reproductive-age women. Am J Epidemiol 1996;144:381–8.
17. Michnovicz JJ, Hershcopf RJ, Naganuma H, Bradlow HL, Fishman J. Increased 2-hydroxylation of estradiol as a possible mechanism for the anti-estrogenic effect of cigarette smoking. N Engl J Med 1986;315:1305–9.
18. Baron JA, La Vecchia C, Levi F. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol 1990;162:502–14.
19. Jensen J, Christiansen C, Rodbro P. Cigarette smoking, serum estrogens, and bone loss during hormone-replacement therapy early after menopause. N Engl J Med 1985; 313:973–5.
20. Sharara FI, Beatse SN, Leonardi MR, Navot D, Scott RT Jr. Cigarette smoking accelerates the development of diminished ovarian reserve as evidenced by the clomiphene citrate challenge test. Fertil Steril 1994;62:257–62.
21. McKinlay SM, Bifano NL, McKinlay JB. Smoking and age at menopause in women. Ann Intern Med 1985;103:350–6.
22. Mattison DR, Shiromizu K, Nightingale MS. Oocyte destruction by polycyclic aromatic hydrocarbons. Am J Ind Med 1983;4:191–202.
23. Takizawa K, Yagi H, Jerina DM, Mattison DR. Murine strain differences in ovotoxicity following intraovarian injection with benzo(a)pyrene, (+)-(7R,8S)-oxide, (−)-(7R,8R)-dihydrodiol, or (+)-(7R,8S)-diol-(9S,10R)-epoxide-2. Cancer Res 1984;44:2571–6.
24. Jull JW, Phillips AJ. The effects of 7,12-dimethylbenz(a)anthracene on the ovarian response of mice and rats to gonadotrophins. Cancer Res 1969;29:1977–87.
25. Mattison DR. The mechanisms of action of reproductive toxins. Am J Ind Med 1983;4:65–79.
26. Bengtsson M, Rydstrom J. Regulation of carcinogen metabolism in the rat ovary by the estrous cycle and gonadotropin. Science 1983;219:1437–8.
27. Bengtsson M, Montelius J, Mankowitz L, Rydstrom J. Metabolism of polycyclic aromatic hydrocarbons in the rat ovary: Comparison with metabolism in adrenal and liver tissues. Biochem Pharmacol 1983;32:129–36.
28. Barbieri RL, McShane PM, Ryan KJ. Constituents of cigarette smoke inhibit human granulosa cell aromatase. Fertil Steril 1986;46:232–6.
29. Shamsuddin AK, Gan R. Immunocytochemical localizatin of benzo(a)pyrene-DNA adducts in human tissues. Hum Pathol 1988;19:309–15.
30. Basler A, Rohrborn G. Chromosome aberrations in oocytes of NMRI mice and bone marrow cells of Chinese hamsters induced with 3,4-benzpyrene. Mutat Res 1976; 38:327–32.
31. Davidson GE, Dawson GW. The induction of somatic mutations in mouse embryos by benzo(a)pyrene. Arch Toxicol 1977;38:99–103.
32. 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.
33. Brambilla DJ, McKinlay SM. A prospective study of factors affecting age at menopause. J Clin Epidemiol 1989;42:1031–1039.
34. Gold EB, Bromberger J, Crawford S, Samuels S, Greendale GA, Harlow SD, et al. Factors associated with age at natural menopause in a multiethnic sample of midlife women. Am J Epidemiol 2001;153:865–74.
35. Campagnoli C, Morra G, Belforte P, Belforte L, Prelato Tousijn L. Climacteric symptoms according to body weight in women of different socio-economic groups. Maturitas 1981;3:279–87.
36. Erlik Y, Meldrum DR, Judd HL. Estrogen levels in postmenopausal women with hot flashes. Obstet Gynecol 1982;59:403–7.
37. den Tonkelaar I, Seidell JC, van Noord PA. Obesity and fat distribution in relation to hot flashes in Dutch women from the DOM-project. Maturitas 1996;23:301–5.
38. Klinga K, von Holst T, Runnebaum B. Influence of severe obesity on peripheral hormone concentrations in pre- and postmenopausal women. Eur J Obstet Gynecol Reprod Biol 1983;15:103–12.
39. McKinlay SM, Brambilla DJ, Posner JG. The normal menopause transition. Maturitas 1992;14:103–15.
40. Zachow RJ, Magoffin DA. Direct intraovarian effects of leptin: Impairment of the synergistic action of insulin-like growth factor-I on follicle-stimulating hormone-dependent estradiol-17β production by rat ovarian granulosa cells. Endocrinology 1997;138:847–50.
41. Rice VM, Limback SD, Roby KF, Terranova PF. Tumor necrosis factor alpha inhibition of follicle-stimulating hormone-induced granulosa cell estradiol secretion in the human does not involve reduction of cAMP secretion but inhibition at post-cAMP site(s). Endocrine 1999;10:19–23.
42. Luheshi GN, Gardner JD, Rushforth DA, Loudon AS, Rothwell NJ. Leptin actions on food intake and body temperature are mediated by IL-1. Proc Natl Acad Sci 1999;96:7047–52.