| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Institute of Community Medicine (Å.B., B.S., M.M., V.F., G.K.R.B.), University of Tromsø, N-9037 Tromsø, Norway; Departments of Clinical Chemistry (J.Su.), Internal Medicine (J.Sv.), and Obstetrics and Gynecology (G.A., P.Ø.), University Hospital of North Norway, N-9038 Tromsø, Norway
Address all correspondence and requests for reprints to: Åshild Bjørnerem, Institute of Community Medicine, University of Tromsø, N-9037 Tromsø, Norway. E-mail: ashild.bjornerem{at}ism.uit.no.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Total estradiol increased with age among men (P < 0.001), with or without adjustment for BMI and lifestyle factors. FSH increased with age both in men (P < 0.001) as well as pre- (P < 0.001) and postmenopausal women (P = 0.01) after similar adjustment, and DHEAS decreased with age in both sexes (P < 0.001).
With increasing BMI, free estradiol increased in men (P = 0.004), total and free estradiol increased in postmenopausal women (P < 0.001), and FSH decreased in men (P = 0.03) and postmenopausal women (P < 0.001).
Men with chronic diseases had lower levels of DHEAS, compared with healthy men (P < 0.001). Smokers had higher DHEAS levels than nonsmokers. Further studies are needed to confirm these hormonal changes with age and disease.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
There are major differences in the levels of sex hormones with age among men and women. Hormonal changes through the menstrual cycle and the menopausal transition are well described, a greater lack of clarity is present among postmenopausal women. The menopausal transition in women is characterized by a relatively abrupt decrease in the level of total estradiol and an increase in the level of FSH (10). The effects of age and body mass index (BMI) on the levels of total estradiol and FSH are relatively small, compared with the large effect of the menopause (10). Decreased or unchanged FSH (11, 12, 13) and decreased or unchanged estradiol (12, 13) with age are reported among postmenopausal women. Most authors describe a fall in FSH whereas estradiol is unchanged with age in this group, although divergent results exist. Use of hormone replacement therapy (HRT) increases total serum estradiol and decreases the level of FSH (14). Dehydroepiandrosterone sulfate (DHEAS) levels gradually decrease with age independent of menopausal transition (15, 16).
In men, the influence of age on estradiol level is not clear. Decreased (17, 18) and unchanged levels are reported (1, 19, 20, 21). The FSH levels increase (19, 22), and the DHEAS levels gradually decrease with age in men (15, 19).
BMI is positively associated with total estradiol (12, 21) and negatively associated with FSH (12, 23) in both men and postmenopausal women. BMI is negatively associated with DHEAS among men (7, 21, 24) but not among women (7, 25).
Smokers have higher levels of DHEAS than nonsmokers of both sexes (7, 26); otherwise smoking and other lifestyle factors are reported to be either marginally or not at all associated with sex hormones (5, 21). Except for being a large source of sex hormones (27), the role of DHEAS in aging and disease is still insufficiently understood (7, 8, 9) and remains controversial.
The aim of this study was to describe the distribution of total and free estradiol, FSH, and DHEAS by age, sex, menopausal status, and use of HRT as well as the associations of total and free estradiol, FSH, and DHEAS with BMI, selected lifestyle factors, and chronic diseases in an unselected general population.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The Tromsø study was initiated in 1974 with surveys repeated in 1979–1980, 1986–1987, and 1994–1995. The study is a single-center, population-based prospective study in the municipality of Tromsø, Northern Norway. The fourth survey of the Tromsø population started in September 1994 and was completed in September 1995. The University of Tromsø, in cooperation with the National Health Screening Service, conducted the survey. The Regional Committee for Medical Research Ethics recommended the study. All the participants gave informed written consent.
All inhabitants over 24 yr of age were invited, and 27,159 subjects (77%) participated in the main survey (phase I). A protocol similar to that used during previous surveys in this population was followed (28, 29). From the main survey, all men aged 55–74 yr, all women aged 50–74 yr, and 5–10% random samples of the other age groups of both sexes were invited to an extended examination, including forearm bone densitometry and blood sampling (phase II). In addition, 328 male participants of the Family Intervention Study, selected on the basis of high total cholesterol or low high-density lipoprotein to total cholesterol ratio (30), were also invited. A total of 7948 individuals participated in this Tromsø Osteoporosis Study (31). Among these study participants, a subgroup of 3685 was selected for hormone assays, and blood samples from 3564 participants were available for analyses. We excluded 10 participants due to pregnancy (n = 4) or use of testosterone (n = 2), progestin (n = 1), selective estrogen receptor modulator (n = 1), or corticosteroids (n = 2). To make the analyses consistent, we also excluded 47 participants with missing values on one or more hormones, BMI, or lifestyle factors. Of the total 1952 women, all 380 women on combined HRT, estriol, or hormonal contraception are described separately but not included in the regression models. A total of 1555 men and 1572 women with hormone measurements are thus included in the main part of this study.
Questionnaires
Two extensive self-administered questionnaires were used as instruments to gain information on a broad set of variables. In the present study, we included information on previous and present diseases such as cancer; asthma; diabetes; stroke and ischemic heart disease (IHD); use of any medication; current smoking status; and consumption of cigarettes, alcoholic beverages, and coffee. An alcohol intake score was constructed by adding the number of glasses of beer, wine, and spirits, assuming an equal amount of alcohol in one unit of each type. A physical activity score was made by adding the hours per week of moderate and hard physical activity, giving the hours with hard activity double weight: score = moderate + 2 hard.
Menopause and use of hormonal therapy
Women were asked to report whether they were pregnant; the date of the last menstrual period if they still were menstruating; and if not, their age at menopause. They were also asked about the use of oral, transcutaneous, or vaginal estrogen therapy, contraceptive pills, and hormonal intrauterine device and the brand they currently used.
At blood sampling, menstruating women were not asked to report their date of the last menstrual period; therefore, the information on menstrual phase at blood sampling is lacking.
Women who reported use of hormonal therapy (n = 380) were on combined HRT (n = 215), oral or vaginal estriol (n = 111), contraceptive pills (n = 22), or levonorgestrel-releasing intrauterine system (LGS IUS) (n = 32). Only nonusers of these hormones were classified into pre-, peri-, or postmenopausal groups because the menopausal status could not be determined accurately among some of the hormone users. We have chosen to present results with the highest possible n, and therefore our definition of postmenopausal status was based on the self-reported menopause and age. Women who reported that they had stopped menstruating over a year ago (n = 900) or were 54 yr of age or older (n = 434) were defined as postmenopausal (n = 1334). Women who had stopped menstruating within the last 3–12 months and were 45 yr or older or with unreported last menstrual period and between 45 and 54 yr were considered perimenopausal (n = 33). Women who had their last menstrual period within the last 3 months (n = 176) or within the last 3–12 months and were younger than 45 yr of age (n = 6) or had missing menstruation data and were younger than 45 yr of age (n = 23) were defined as premenopausal (n = 205). This definition of menopausal status left none of the women undefined due to missing or conflicting values. Participants were not asked about previous hysterectomy or oophorectomy; accordingly, information on surgical cause of menopause was not available. Total number of postmenopausal years was calculated as current age minus age at menopause when present.
Measurements
Height and weight were measured in light clothing without shoes, and BMI was calculated as weight in kilograms divided by the square of height in meters.
Hormone assays
Nonfasting blood samples were taken between 0800 and 1600 h. Serum samples were stored at –70 C until they were first thawed for analyses of sex hormones in 2001, after a storage time of 6–7 yr. All hormones and SHBG were measured on Immulite 2000 (Diagnostic Products Corp., Los Angeles, CA). Total estradiol and DHEAS measurements were based on competitive immunoassays, whereas FSH and SHBG measurements were based on immunometric assays. The assays were run daily on randomly selected samples in batches of 100 per day.
The intra- and interassay coefficient of variation for the measurements of total estradiol, FSH, DHEAS, and SHBG were between 3.5 and 10%, depending on the level, and the levels of sensitivity were 10 pmol/liter, 0.5 IU/liter, 1.0 µmol/liter, and 1.0 nmol/liter, respectively. Samples with values below functional assay sensitivity were given a value midway between zero and assay sensitivity for the analyses: total estradiol (354 women and 61 men), FSH (eight women), and DHEAS (419 women and 85 men). Total estradiol values greater than 7340 were recoded to 7340 pmol/liter (two women), FSH greater than 170 was recoded to 170 IU/liter (one woman), and SHBG greater than 180 was recoded to 180 nmol/liter (36 women and seven men).
Free estradiol values were calculated from total estradiol and SHBG levels according to the equation:
 |
Statistical analysis
The SAS software package (version 8.2, SAS Institute Inc., Cary, NC), was used for both data management and analysis. For statistical analysis the values of the hormones were log transformed to correct for skewed distribution. Differences in means between groups were tested by ANOVA, analysis of covariance (ANCOVA), and adjusted t tests for multiple comparisons of pairs of means. Data are presented stratified by age. Linear trend by age was analyzed in linear regression models. Multiple linear regressions were used to model total estradiol, free estradiol, FSH, and DHEAS with the following independent variables: age, BMI, smoking, alcohol and coffee consumption, physical activity, sampling hour and season. Season and sampling hour were dropped from the final model because adjustment for them did not change any of the results. The same set of independent variables, which was significantly associated with at least one hormone, was used systematically in all hormone models among both women and men. The variables describing smoking and alcohol use had a substantial proportion of zero values. Therefore, we analyzed these variables dichotomized to check for any difference in hormone levels between smokers vs. nonsmokers and alcohol users vs. nonusers. Thereafter we checked for any dose-response effects among smokers and users of alcohol. Number of cigarettes, number of alcohol units, age at menarche, and number of menopausal years were not associated with any hormone when they were used as independent variables in multiple linear regression models. These variables are therefore not presented in the results.
The analyses were performed separately for pre- and postmenopausal women. The perimenopausal women were excluded from the multiple regression analyses due to small numbers. All P values are based on analyses of log-transformed values, and P values are two sided. However, means and confidence limits are transformed back to the original units when presented.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. The mean, median, and 95th percentile for age at menopause were 48, 49, and 55 yr, respectively, and 85% of women not using any hormone were postmenopausal by an average of 12 yr. The mean age was 38.1 (range 25–54) yr, 50.5 (range 39–57) yr, and 64.1 (range 37–83) yr among pre-, peri-, and postmenopausal women, respectively.
|
Age and sex
Age- and sex-stratified distributions of total and free estradiol, FSH, and DHEAS are shown in Table 2 and Fig. 1. Among men, total (P < 0.001) and free estradiol levels (P = 0.04) were positively associated with age and were, respectively, 33 and 12% higher at age older than70 yr, compared with men younger than 40 yr. Among postmenopausal women, total (P = 0.04) and free estradiol levels (P = 0.008) were negatively associated with age and were 90% lower at age older than 70 yr, compared with premenopausal women younger than 40 yr. In men and premenopausal women, FSH was positively associated with age (P < 0.001). Men and women older than 70 yr had, respectively, 2.4- and 13.2-fold higher levels of FSH than those younger than 40 yr. An abrupt change in the levels of estradiol and FSH was obvious among women at approximately 50 yr. DHEAS was negatively associated with age in both sexes (P < 0.001). The DHEAS levels at age older than 70 yr were 32 and 29% of the levels at age younger than 40 yr among men and women, respectively. The levels of DHEAS were higher among men than women (P < 0.001). Except for free estradiol among men (P = 0.09), the associations between age and all hormones studied were significant among men (P < 0.001) and postmenopausal women (P = 0.02, P < 0.001, P = 0.01, and P < 0.001) after adjusting for BMI and lifestyle factors. Among premenopausal women only FSH and DHEAS were associated with age after similar adjustment (P < 0.001).
|
|
Among men, only free estradiol was positively associated (P = 0.004), whereas FSH was negatively associated (P = 0.03) with BMI after adjustment for age and lifestyle factors (Fig. 2). Among postmenopausal women, BMI was positively associated with total and free estradiol and negatively associated with FSH after similar adjustment (P < 0.001) (Fig. 2). The levels of total and free estradiol were, respectively, 39 and 71% higher at BMI of 30 kg/m2 or greater, compared with postmenopausal women with BMI less than 25 kg/m2. The levels of DHEAS were not associated with BMI for either sex.
|
Chronic diseases
The DHEAS levels were lower among 243 men with self-reported IHD (P < 0.001), 43 men with stroke (P = 0.001), 50 men with diabetes (P < 0.001), 55 men with cancer (P = 0.007), and 110 men with asthma (P = 0.003), compared with 1128 healthy men after adjusting for age, BMI, and lifestyle factors (Table 3).
|
The mean levels of all hormones studied were different among pre-, peri-, and postmenopausal groups (P < 0.001) (Table 4). After adjustment for age, BMI, and lifestyle factors, the differences in hormone levels by menopausal status persisted regarding total and free estradiol and FSH but not DHEAS.
|
). The levels of DHEAS did not differ between users of HRT and nonusers. Users of contraceptive pills had a significantly lower mean level of free estradiol and FSH. The levels of estradiol did not differ between users of estriol and nonusers.
|
Smoking was positively associated with DHEAS among men (P < 0.001) (Fig. 3) and premenopausal (P = 0.02) and postmenopausal women (P = 0.03) (data not shown). Coffee was positively associated with FSH and DHEAS (P = 0.03) among men and DHEAS among premenopausal women (P = 0.01) and postmenopausal women (P = 0.03). This was the case when the significance of the independent variables smoking and coffee was analyzed in separate models among women. With both variables in the model, none were significant (Fig. 3), probably due to an intercorrelation (Pearson correlation coefficient r = 0.38). Estradiol levels were not associated with smoking for either sex (Fig. 3).
|
Total and free estradiol were significantly associated with season among men and postmenopausal women (P < 0.001 by partial F test). However, adjustment for season did not change any of the above reported associations.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
To our knowledge, this is the largest population-based study in which FSH, DHEAS, and estradiol of both sexes have been measured. The high response rate (77% of the eligible population) in the main survey assures generalizability of results to a majority of the source population. However, the nonresponding minority might have characteristics that differ substantially from those found in the study population. In a previous paper (34), participants of the main study who were nonresponders at the second examination (phase II) were compared with responders of phase II. They were not found to be healthier in any age or sex group, as judged by the self-reported prevalence of chronic diseases. We looked for selection bias by comparing characteristics of the subgroup that participated in the hormone studies (n = 3507) with the total phase II survey population (n = 7948). The subgroup of men did not have significantly different characteristics. Although there was a mean age difference of 1 yr (59.4 vs. 58.2 yr, P < 0.001), higher prevalence of IHD (8.6 vs. 7.1%, P < 0.05) and higher percentage of teetotalers (27.3 vs. 24.5%, P < 0.01) among the subgroup of women participating in hormone studies in comparison with the total phase II survey population, this is not expected to have influenced our results substantially.
Our study is cross-sectional; therefore, the direction of the associations cannot be determined. The cross-sectional estimates of associations between hormonal levels and age or BMI are not true measures of longitudinal changes, and individual change in hormonal levels cannot be indicated.
Self-reported age at menopause is the most straightforward way of classifying menopausal status according to the recommendation of the World Health Organization (35). Because 93% of women with missing menopause data were older than 53 yr, the classification based on age is not expected to have biased this study. Results did not differ substantially whether or not women with a missing response to the menopause question were included.
Lack of information on hysterectomy before natural menopause is a limitation of our study (36). Nevertheless, because the rate of hysterectomy in Norway is reported to be remarkably low (37), we believe that this does not introduce a serious misclassification problem.
The mean interval between filling in the questionnaire and blood sampling was 5 wk, but 78% of the women were examined within an 8-wk period. Menopausal status or use of medications may have changed during that interval. This may have weakened true associations among premenopausal women.
Assays with high sensitivity were used to measure the hormone levels. However, total estradiol and DHEAS values were below the assay sensitivity level in 18 and 21% of postmenopausal women, respectively. Therefore, it was not possible to obtain normal distribution of these hormones in this group. Blood samples taken between 0800 and 1600 h were used for the hormone measurements. Moderate diurnal variations of estradiol, FSH, and DHEAS are reported in previous mostly small studies (38, 39). In agreement with Verkasalo et al. (5), we did not find any systematic daytime variation in any hormone among either men or pre- and postmenopausal women. Adjustment for sampling hour did not change any result. Serum samples were frozen at –70 C for approximately 6.5 yr, and the hormone levels were measured when the samples were thawed for the first time. Levels of steroid hormones have been shown to be relatively stable in frozen plasma stored for 3–10 yr (15, 40). The delayed analyses are therefore not likely to represent a major problem. Free estradiol was calculated according to Vermeulen et al. (33), recently evaluated by Rinaldi et al. (41), and found to be a simple and reliable index of free estradiol.
No previous population-based study has reported increasing estradiol with increasing age among men. Recent large population-based studies describe levels of total estradiol unchanged (19, 21) or decreased with age (17, 18) among men. Although the reason for this discrepancy may be cohort effects and different selection criteria, our study contributes to continued uncertainty on the issue.
Among men, most of the circulating estradiol is derived from peripheral aromatization of the circulating precursor testosterone (20). The lower levels of testosterone with increasing age (19, 42) do not explain the increased levels of estradiol. A possible explanation might be increased levels of aromatase enzymes with age and the age-associated increase in fat tissue even without weight gain (20, 43, 44). Free estradiol increased to a lesser extent than total estradiol, which is explained by the well-known increase in SHBG with age (19, 42).
Whether FSH levels among postmenopausal women decrease or remain stable with increasing age remains controversial. Previous studies report a decrease in FSH (11, 12) or a stable high level of FSH (13) but no increase with age in this group. However, most of these studies are small (n = 32–60) except a population-based study by Kwekkeboom et al. (12), who did not account for the use of HRT that is known to lower the level of FSH (14). No previous study to our knowledge has reported increasing FSH with increasing age among postmenopausal women. Despite small changes in the absolute levels of FSH, we found that the FSH levels do not fall with age among postmenopausal women as thought previously (45). A possible explanation is decreased negative feedback on the pituitary gland due to lower levels of estradiol in this group of women. Among men, the levels of FSH at age older than 70 yr were higher than in those younger than 40 yr, but the age-related difference in the levels was less in comparison with women, as previously reported (10, 22).
The lower levels of DHEAS associated with increasing age of both sexes and higher DHEAS levels in men than in women confirm previously well-documented results (15, 26).
Our findings also confirm the well-known associations between BMI and the hormones as reported previously (12, 21). The higher levels of free estradiol but not total estradiol with increasing BMI in men is explained by the decrease of SHBG with increasing BMI (21, 42). In addition, the larger increase of free estradiol than total estradiol levels with increasing BMI among postmenopausal women is also explained by this same fall in SHBG (5, 25).
Lower DHEAS levels among men with IHD, compared with healthy men, have been reported in cross-sectional studies (9), although results from longitudinal studies are ambiguous (7, 8). The discrepancy in findings among studies may reflect the different end points used (9). DHEAS may be associated with IHD morbidity but not with cardiovascular mortality, or a more complex relationship between DHEAS and other biological processes may exist. However, among women DHEAS was not associated with IHD or other chronic diseases. This is in agreement with previous reports (7, 46). As previously described, smokers had higher levels of DHEAS (7, 26). This intriguing finding of higher levels of DHEAS among smokers and lower levels of DHEAS among men with IHD has so far not been explained by any biological mechanism. The clinical significance of DHEAS in cardiovascular disease remains uncertain. Further prospective studies are needed for better understanding of the biological effects of DHEAS both in men and women.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: ANCOVA, Analysis of covariance; BMI, body mass index; CI, confidence interval; DHEAS, dehydroepiandrosterone sulphate; HRT, hormone replacement therapy; IHD, ischemic heart disease; LGS IUS, levonorgestrel-releasing intrauterine system.
Received April 22, 2004.
Accepted September 7, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-subunit decline in aging normal postmenopausal women. J Clin Endocrinol Metab 70:944–950[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  G Tomei, M. Rosati, M Ciarrocca, A Capozzella, B Pimpinella, T Palitti, T Casale, M Di Famiani, C Filippelli, C Cangemi, et al. Urban stressors and plasmatic 17-{beta}-estradiol (E2) in male exposed workers Toxicology and Industrial Health, October 1, 2007; 23(9): 537 - 543. [Abstract] [PDF]
|
|
|
|
 |
|
 S. C Larsson and A. Wolk Obesity and colon and rectal cancer risk: a meta-analysis of prospective studies Am. J. Clinical Nutrition, September 1, 2007; 86(3): 556 - 565. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bjornerem, L. A. Ahmed, R. M. Joakimsen, G. K R. Berntsen, V. Fonnebo, L. Jorgensen, P. Oian, E. Seeman, and B. Straume A prospective study of sex steroids, sex hormone-binding globulin, and non-vertebral fractures in women and men: the Tromso Study Eur. J. Endocrinol., July 1, 2007; 157(1): 119 - 125. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-F. Chan, M. Dowsett, E. Folkerd, S. Bingham, N. Wareham, R. Luben, A. Welch, and K.-T. Khaw Usual Physical Activity and Endogenous Sex Hormones in Postmenopausal Women: The European Prospective Investigation into Cancer-Norfolk Population Study Cancer Epidemiol. Biomarkers Prev., May 1, 2007; 16(5): 900 - 905. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bjornerem, B. Straume, P. Oian, and G. K. R. Berntsen Seasonal Variation of Estradiol, Follicle Stimulating Hormone, and Dehydroepiandrosterone Sulfate in Women and Men J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 3798 - 3802. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Orwoll, L. C. Lambert, L. M. Marshall, K. Phipps, J. Blank, E. Barrett-Connor, J. Cauley, K. Ensrud, S. Cummings, and for the Osteoporotic Fractures in Men Study Group Testosterone and Estradiol among Older Men J. Clin. Endocrinol. Metab., April 1, 2006; 91(4): 1336 - 1344. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
