SEARCH

SEARCH BY CITATION

Keywords:

  • women;
  • lean;
  • metabolism;
  • hormones;
  • aromatase gene

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: In the current obesity epidemic, the ability to remain lean is beginning to be uncommon. Therefore, it was considered of interest to characterize such subjects.

Research Methods and Procedures: From a population of premenopausal women (n = 270), all 40 years of age, those with a similar body mass index (BMI) as women at the age of 21 years, born the same year (BMI = 21.1 kg/m2) were selected among nonsmokers and compared with the remaining nonsmoking women.

Results: Lean women showed, as expected, low waist-to-hip circumference ratio and abdominal sagittal diameter as well as absence of other disease risk factors. Compared with the remaining women, 17β-estradiol was high and androgens were low, whereas insulin-like growth factor I and thyroid hormones showed no differences. Dihydroepiandrosterone sulfate was lower, whereas cortisol, measured in saliva repeatedly over a day, and adrenocorticotropin hormone were not different. Results from questionnaires indicated higher education and socioeconomic status, frequent sports activities, and better psychosocial adaptation and psychological health. A tetranucleotide repeat polymorphism in the fifth intron of the aromatase P450 gene was longer among the lean (187 base pairs) than the rest of the women. Women with opposite phylogenetic characteristic have a short microsatellite (168 base pairs) in this gene locus.

Discussion: Lean, nonsmoking women enjoy an excellent health in not only anthropometric and metabolic factors, but also in neuroendocrine, endocrine, and psychological variables. The endocrine measurements suggest a well-functioning aromatase, which in turn might have a genetic background, contributing to health. The aromatase gene might be important for regulation of body fat mass.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The obesity pandemic has now reached such proportions that it is considered one of the major global health problems. This has aroused national and international administrative bodies to plan strategies for counteractions (1). Overweight and obesity are now found in a majority of subjects in several urbanized countries. The situation is frequently worse in women than in men (1).

No previous world epidemic has reached such proportions, were it plague or tuberculosis, or the current problems with acquired immunodeficiency syndrome. Such a dramatic prevalence may lead to the consideration that overweight and obesity are normal consequences of evolutionary characteristics for survival, which now are allowed to be fully developed. From this aspect, subjects remaining lean might even be considered abnormal. A powerful argument against this view, however, is the prevalent, serious diseases particularly associated with central obesity (2). It is of interest to try to characterize subjects who have managed to avoid becoming overweight or obese, because such information might be helpful in current efforts to counteract the obesity epidemic.

This problem was approached in this study. Nonsmoking middle-aged women were characterized, who had been able to keep their body weight at the average level of young women.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Population

The population studied has been described previously in detail (3, 4), and only a brief description is given here. All women born on uneven days in 1956, and living in Göteborg, Sweden (n = 1464), were selected from population registers and examined in two phases. The first phase consisted of questionnaires and self-measurements. The second phase was performed at the laboratory and comprised an invited subcohort of 450 women, of whom 270 (60.0%) volunteered to participate. Postmenopausal women, defined as no menstruations during the recent past 6 months, were excluded from the study. All examinations were performed in the follicular phase of the menstrual cycle (days 5 to 10).

Current smokers were first excluded from calculations to avoid the influence of smoking on body weight. This left 189 nonsmoking women. The population of lean women was selected by setting an upper limit of BMI at 23 kg/m2, which gave an average BMI of 21.1 (1.23) kg/m2 (mean and SD), similar to the reference group (BMI = 21.1 kg/m2), which consisted of Swedish women born from 1955 to 1958, examined at the age of 21 years (5). This left 77 nonsmoking, lean women for studies, comprising 28.5% of the total number of women examined. The remaining (BMI ≥ 23 kg/m2) nonsmoking women were 112.

Among the lean women, 12 (15.6%) used anti-conception pills and 21 (27.3%) intrauterine anti-conception devices, of which 8 (10.4%) contained hormones. Among the remaining nonsmoking women, corresponding figures were 8 (7.1%), 36 (32.1%), and 7 (6.3%). These differences between the two groups were not significantly different (Mann–Whitney U test).

All women gave written, informed consent, and the study was approved by the Ethical Committee of the University of Göteborg.

Examinations

Two physicians (F.B. and R.R.) performed a clinical examination, including history of current and previous diseases and a physical examination. Collection of questionnaires and anthropometric examinations were performed by the same technician, who also took a venous blood sample after the women had fasted over night. Before venipuncture, blood pressure was measured on the right arm using an automatic digital blood pressure measuring equipment (UA-751; A&D Company, Tokyo, Japan). With the woman sitting, a first blood pressure was measured after 5 minutes rest and a second after another 5 minutes rest, and the results averaged. Heart rate was recorded simultaneously.

Questionnaires

The questionnaires have been described in detail previously (6, 7) and contained questions about a number of psychosocial and socioeconomic details, including work conditions and education, spare time activities as well as sleeping problems. The Alcohol Use Disorders Identification Test was used for definition of alcohol habits. Symptoms of depression, anxiety, and compulsiveness were recorded according to the Comprehensive Psychopathology Rating Scale. Furthermore, a questionnaire was used defining behavior conventionally considered as masculine or feminine (8). Finally, the women were asked about their playing preferences as girls, whether they preferred to play with girls or boys, with toys conventionally considered as boys’ or girls’ toys, and whether they were often involved in physical fights. These observations were summarized as a tomboy index (4). The questions were graded 1 to 5 or 1 to 7 and treated quantitatively.

Anthropometry

These measurements were performed after the women had been fasting overnight. Body weight was measured to the nearest 0.1 kg with the woman in underwear and the height to the nearest 0.01 m. The waist circumference was measured half-way between the costal arc and the iliac crest, in a normal respiratory position, the hip circumference over the widest part of the gluteal region, and the waist-to-hip circumference ratio calculated. The abdominal sagittal diameter was recorded as the distance between the examination table and the highest point of the abdomen, providing an estimation of visceral fat mass (9).

Hormones, Serum Lipids, and Blood Glucose

Saliva cortisol was measured in seven samples collected during an ordinary working day from morning to bedtime, including after a standardized lunch as described previously. Lunch cortisol was defined as the added cortisol measurements 30, 45, and 60 minutes after finishing lunch. Perception of stress was reported during the day of measurements to provide a stress-related cortisol. Furthermore, a low-dose (0.5 mg) dexamethasone suppression test was also performed at home (10).

Total testosterone (T), 17β-estradiol, sex hormone-binding globulin (SHBG) and follicle-stimulating hormone (FSH) were analyzed using chemoluminescent enzyme immunoassays (Diagnostic Products, Los Angeles, CA). Luteinizing hormone (LH), adrenocorticotropic hormone (ACTH), free testosterone, androstenedione and dehydroepiandrosterone sulfate (DHEAS) were determined by radioimmunoassays, using assay kits from Serono Diagnostics (Braintree, MA) (LH) and Diagnostic Products (all other). Insulin was measured by radioimmunoassay (Phadebas, Pharmacia Insulin RIA 100; Kabi-Pharmacia Diagnostics, Uppsala, Sweden) and insulin-like growth factor I (IGF I) was determined by radioimmunoassay (Nichols Institute Diagnostics, San Juan Capistrano, CA).

Thyroid-stimulating hormone and thyroxin were measured by an enzyme immunoassay method (Ortho-Clinical Diagnostics, Stockholm, Sweden). Saliva cortisol was measured by radioimmunoassay (Orion Diagnostica Cortisol RIA; Orion Corporation, Espoo, Finland). Glucose was determined enzymatically and serum lipids as previously described (11).

Gene Analyses

The number of tetranucleotide repeats in the fourth intron of the aromatase P450 gene were determined in genomic DNA extracted from blood cells. After amplification by polymerase chain reaction, the size of the fragments obtained was determined with electrophoresis as previously described in detail. Alleles with 168, 171, 175, 179, 183, 187, and 191 base pairs (bp) were found (3, 12).

Statistical Methods

The SPSS software (SPSS for Windows, Release 9.0; SPSS, Chicago, IL) was used. Differences between groups were analyzed with Student'st test, preceded by Levene's test for homogenicity of variances. For observations with n < 30, we also applied the Mann–Whitney U test. Correlation analyses were performed according to Spearman. A two-tailed p value <0.05 was considered significant.

To analyze putative impacts of the stretches of the alleles of the aromatase microsatellite the following comparisons between the both groups were performed.

  • 1
    . The length (bp) of the two alleles averaged ([allele 1 + allele 2]/2, for example [168 + 187]/2, etc.).
  • 2
    . The number of separate microsatellite length (168 bp, 171 bp, etc.).
  • 3
    . The number of homozygotes (168/168 bp, 171/171 bp, etc.).
  • 4
    . Combinations including the 187 bp allele (168/187, 171/187, etc.).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The distribution of the BMI in the whole material (n = 270) is seen in Figure 1. The average and SD were 24.7 and 4.30. The borderlines of BMI > 30 kg/m2 (obesity) and > 25 kg/m2 (overweight) (1) separated out 36 (13.3%) and 103 (38.1%) women, respectively.

image

Figure 1. The distribution of values for body mass index (BMI; n = 270). Abscissa, BMI; ordinate, n.

Download figure to PowerPoint

The anthropometric measurements were considerably lower in the lean women, as expected (Table 1). The metabolic and hemodynamic values were also lower except total cholesterol and heart rate, whereas high-density lipoprotein cholesterol was elevated.

Table 1.  Descriptive statistics and comparisons between nonsmoking lean and the remaining women
 Lean women 77Remaining women 112 
nMeanSDMeanSDp value
  1. LDL, low-density lipoprotein; HDL, high-density lipoprotein; BP, blood pressure; NS, not significant.

Body mass index (kg/m2)21.101.2927.373.990.000
Waist/hip circumference ratio0.7630.0480.8340.0700.000
Sagittal diameter (cm)17.411.1220.932.680.000
Glucose (mM)4.220.344.490.890.011
Insulin (μU/mL)6.662.408.893.810.000
Triglycerides (mM)0.800.311.100.560.000
Total cholesterol (mM)4.990.735.170.84NS
LDL cholesterol (mM)2.900.683.130.720.026
HDL cholesterol (mM)1.750.331.560.390.001
Systolic BP (mm Hg)105.713.6113.913.20.000
Diastolic BP (mm Hg)62.99.666.69.60.011
Heart rate (beats/minute)65.410.165.69.2NS

Hormone values are presented in Table 2. 17β-estradiol and SHBG were higher, and all androgens, including androstenedione and DHEAS, were considerably lower in the lean women. Furthermore, the ratios of 17-β estradiol over total or free T or androstenedione were much higher in the lean women. There were no differences in FSH, LH, IGF-I, thyroid hormones, or ACTH. These values were recalculated for the lean women, excluding those who were using anti-conception pills or hormone-containing intrauterine anticonception devices. There was no difference in the statistical results with these women excluded (data not shown).

Table 2.  Hormone values in nonsmoking lean and the remaining women
 Lean women 77Remaining women 112 
nMeanSDMeanSDp value
  1. SHBG, sex hormone-binding globuline; DHEAS, dihydroepiandrosterone sulphate; FSH, follicle-stimulating hormone; LH, luteinizing hormone; IGF-I, insulin-like growth factor I; TSH, thyroid-stimulating hormone; ACTH, adrenocorticotropin hormone.

17-β estradiol (pM)2552321831350.015
SHBG (nM)58.925.239.320.90.000
Testosterone (nM)1.670.812.070.970.004
Free testosterone (pM)3.251.815.252.500.000
Androstenedione (nM)5.761.977.683.460.000
DHEAS (μM)3.501.524.091.720.017
17-β estradiol/testosterone1891891231500.012
17-β estradiol/free testosterone10413047620.001
17-β estradiol/androstenededione505729280.003
FSH (IU/liter)14.68.912.36.5NS
LH (IU/liter)3.92.113.733.2NS
IGF-I (ug/liter)1995420353NS
TSH (mU/liter)1.241.001.260.87NS
Thyroxin (pM)16.03.915.32.6NS
ACTH (ng/liter)24.015.326.514.3NS

Cortisol measurements, including morning cortisols, slope of diurnal cortisol curve, lunch or total cortisol, stress-related cortisol, or dexamethasone suppression of morning cortisols, revealed no differences between the nonsmoking groups (data not shown).

Table 3 shows that the lean women had higher scores in education, better work situation, economy, spare time, general health, physical conditioning, memory, and self-confidence and were more frequently engaged in sports than the remaining women. Housing conditions, appetite, and alcohol consumption revealed no significant differences (data not shown).

Table 3.  Socioeconomic and psychosocial factors in nonsmoking lean and the remaining women*
 Lean women 77Remaining women 112 
nMeanSDMeanSDp value
  • *

    For explanation of variables see references 6 and 7.

Education5.451.084.811.480.001
Work situation4.911.164.131.620.000
Economy4.911.534.351.540.018
Spare time5.291.164.701.550.003
General health5.871.035.261.220.000
Sports3.111.262.451.030.000
Physical conditioning4.771.123.771.350.000
Memory5.770.915.171.200.000
Self-confidence5.081.094.661.400.023

The lean women had less depressive and anxiety symptoms than the other nonsmoking women, whereas occurrence of compulsive symptoms were not different. The lean women were sleeping better and showed a trend to less frequently waking up during night (Table 4).

Table 4.  Indices of depression, anxiety and compulsion, and sleeping conditions in nonsmoking lean and remaining women
 Lean women 68 to 76Remaining women 98 to 112 
n*MeanSDMeanSDp value
  • NS indicates not significant.

  • *

    Differences in n due to missing values.

Depression index2.882.543.832.970.032
Anxiety index3.702.245.163.350.001
Compulsive index2.462.003.022.60NS
Quality of sleep5.631.475.151.760.045
Awakening during night2.451.002.761.140.057

The index of femininity was lower in the lean than in the rest of the women (means: 45.4 vs. 46.9, SD: 4.0 and 5.1, p = 0.04) but the masculinity index was not different. The examination of expression of tomboyish behavior revealed no differences (data not shown).

The distribution of the average length of tetranucleotide alleles of the microsatellite in the fifth intron of the aromatase gene is seen inFigure 2 both for the nonsmoking lean and the rest of the women. There was an apparent increase of the mean allele stretch with 187 bp in the lean women (21% vs. 9.8%), with a corresponding moderately lower frequency in the shorter alleles. In Pearson χ2 test the frequency of allele 187 was significantly higher in the lean than in the rest of the women (p = 0.024), whereas there were no significant differences in the other allele frequencies. Compared with the rest of the women, the lean women had a significantly longer average total microsatellite stretch (means: 178.8 and 176.1, with SDs of 6.02 and 5.89, p = 0.003). Furthermore, the frequency of 187/187 bp allele homozygotes was higher in the lean women (lean, n = 16, 20.8%, other, n = 10, 8.9%,p < 0.020), whereas no differences were found for homozygotes for other alleles. The frequency of heterozygotes in which the 187 bp allele was included (168/187, 171/187, and 175/187) showed no differences between groups.

image

Figure 2. The distribution of the average length of tetranucleotide (TTTA) repeats in the fifth intron of the aromatase gene in nonsmoking lean (n = 78) and the remaining (n = 112) women. Abscissa, base pairs; ordinate, n.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

In this work, 40-year-old women who had the same BMIs as women at 21 years of age were characterized. To avoid the influence of smoking, which usually is associated with relative leanness, smokers were excluded from the analyzes in an attempt to reveal other potential explanatory factors.

Among nonsmokers, the lean women in comparison with the remaining middle-aged women not only had a lower BMI, due to preselection, but also lower waist-to-hip circumference ratio and sagittal diameter. This was expected because these measurements of centralization of body fat often follow BMI measurements (13). Metabolic and hemodynamic risk factors were also lower in the lean women; this also was expected from a large number of previous similar comparisons.

Saliva instead of serum cortisol was measured during an ordinary working day to avoid confounding disturbance of a hospital or laboratory environment including venous puncture. No differences were seen in ACTH or cortisol measurements. DHEAS and androstenedione were lower in the lean women, suggesting a low, normal adrenal activity.

The 17β-estradiol values were higher and androgens lower in the lean women, indicating a well-functioning hypothalamo-pituitary-gonadal axis. The IGF-I and thyroid hormone values did not differ between the groups, suggesting no involvement of the central growth hormone and thyroid axes.

There were no changes in results when women using hormone-containing anticonception pills or devices were excluded in comparisons. This is probably due to the relatively small number of such women.

Blood pressure was lower in the lean women. Elevated blood pressure is considered to be caused mainly by an arousal of the central sympathetic nervous system (14), suggesting that also this autonomic pathway was normally regulated in the lean women.

Results of the questionnaires suggested that the lean women had higher levels of education, better economy, and were more satisfied with their work than the rest of the women. In addition, sports activities were more prevalent in their spare time. Symptoms of depression and anxiety were lower than in the rest of the women. There were no tendencies toward masculine, aggressive behavior at adult age or as reported from childhood.

One may ask whether there are potential cause–effect relationships in the correlations reported in this work. It has been suggested that central neuroendocrine and autonomic pathways are frequently aroused in obesity, particularly of the central subtype (15). Possibilities have also been suggested that such abnormalities, with the hypothalamic-pituitary-adrenal axis regulation in a dominant position, may be responsible for the generation of peripheral disease risk factors (13). The observations presented here seem to support such a contention because when neuroendocrine and autonomic regulations are normal, such as in the lean women examined, then also anthropometric, endocrine, metabolic, and hemodynamic risk factors for disease are absent.

Aromatase converts androgens to estrogens (16). A poorly functioning enzyme, therefore, would be expected to be associated with elevated circulating androgens and low estrogens in women. In the lean nonsmoking women, the estrogen/androgen ratios were clearly high, suggesting a well-functioning aromatase, although direct measurements of enzyme activity were not performed. The aromatase gene was characterized by a long microsatellite in the fifth intron. This is in sharp contrast to hyperandrogenic women in this cohort with abdominal obesity who show powerful, consistent associations to other risk factors, and who have recently been shown to have a short microsatellite in this domain with 168 bp (3). In fact, the short and long alleles in this gene locus seem to form totally separated genetic subgroups (Figure 2) with opposite phylogenetic characteristics.

A short microsatellite in this gene locus has previously also been observed in women with increased risk to develop mammary carcinoma and osteoporosis (17, 18). Hyperandrogenicity with such a short microsatellite in women is a powerful risk factor for type 2 diabetes, myocardial infarction, and female carcinomas in prospective studies (19, 20, 21). Thus, it is apparent that the length of this microsatellite in the fifth intron of the aromatase gene is closely associated with health and disease in women.

The lengths of microsatellite stretches have recently been found to be followed by different functional power of the protein produced, possibly by interference with the regulation of gene transcription (17, 18). For example, a long trinucleotide (CAG) repeat stretch in the transactivating domain of the androgen receptor gene is associated with a weak signaling androgen receptor and vice versa.

The studies reported here as well as in previous work in hyperandrogenic women (3) show opposite directions of repeat length and presumed aromatase function, short repeats associated with poor aromatase function and vice versa. Recently it has been reported that knock-out of the aromatase gene is followed by elevated androgens as well as central obesity, demonstrating the involvement of this enzyme in the regulation of body fat mass (22). Androgen administration to women is followed by increased accumulation of central body fat (23, 24), and elevated androgens are associated with severe health problems (3, 19, 20, 21). Therefore, we speculate that the short microsatellite in the fourth intron of the aromatase gene is associated with a poorly functioning enzyme, hyperandrogenicity, and abdominal obesity, whereas a long microsatellite is a sign of an efficient enzyme with low androgens and leanness.

In summary, lean women are characterized by health not only in endocrine, anthropometric, metabolic, and hemodynamic variables, but also in neuroendocrine and autonomic functions. These findings strengthen a contention that these factors are functionally associated. There are also genetic characteristics of the aromatase gene in such women, associated with a presumably well-functioning enzyme as judged indirectly by peripheral concentrations of estrogens and androgens. High estrogens and low androgens probably help to maintain health in women.

To what extent these hormonal changes are associated with a well-adapted psychosocial life and socioeconomic progress is, however, difficult to judge. Lean women seem to be devoid of aggressive male behavior, also as children, and apparently also have a robust psychological health. It cannot be judged from the data presented to which extent this is secondary to their physical health. The high education among the lean women might perhaps make them amenable to healthy food habits. There was probably not an overrepresentation of restrained eaters because the lean women did not report different appetite than the remaining women. Furthermore, physical activity was reported to be high, which would also decrease their tendency to become obese.

It seems possible to make the general conclusion that women who manage to keep their body weight stable since youth also have excellent physical and psychological health.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This study was supported by the Swedish Medical Research Council (Project K 2000-72x-00, 251-38B and 8868). Anna Nilsson is acknowledged for excellent technical assistance.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    World Health Organization (1998) Obesity: Preventing and Managing the Global Epidemic: Report of a WHO Consultation on Obesity. WHO/NUT/NCD/981. World Health Organization: Geneva, Switzerland.
  • 2
    Björntorp, P. (1999) Overweight is risking fate. Ballière's Clin Endocrinol Metab. 13: 4769.
  • 3
    Baghaei, F., Rosmond, R., Westberg, L., et al (2000) Polymorphism in CYP19 aromatase gene and associations with androgens and risk factors in premenopausal women. (in press).
  • 4
    Rosmond, R., Baghaei, F., Holm, G., Björntorp, P. (2000) Gender-related behavior during childhood and associations with adult abdominal obesity: a nested case-control study in women. J Wom Health Gend Based Med. 9: 413419.
  • 5
    Karlberg, P., Taranger, J., Engström, I., Karlberg, J., Landström, T., Lichtenstein, H. (1976) Physical growth from birth to 16 years and longitudinal outcome of the study during the same age period. Acta Pedriatr Scand Suppl. 258: 776.
  • 6
    Tibblin, G., Tibblin, B., Peciva, S., Kullman, F., Svärdsudd, K. (1990) The Göteborg Quality of Life Instrument—an assessment of well-being and symptoms among men born 1913 and 1923: methods and validity. Scand J Prim Health Care. 1: 3338.
  • 7
    Rosmond, R., Björntorp, P. (1999) Psychosocial and socioeconomic factors in women and their relationship to obesity and regional fat distribution. Int J Obes Relat Metab Disord. 23: 138145.
  • 8
    Bem, S. (1974) The measurement of psychological androgyny. J Consult Clin Psychol. 42: 155162.
  • 9
    Kvist, H., Chowdhury, B., Grangård, U., Tylén, U., Sjöström, L. (1988) Total and visceral adipose tissue volumes derived from measurements with computed tomography in adult men and women: predictive equations. Am J Clin Nutr. 48: 13511361.
  • 10
    Rosmond, R., Dallman, M. F., Björntorp, P. (1998) Stress-related cortisol secretion in men: relationships with abdominal obesity and endocrine, metabolic and hemodynamic abnormalities. J Clin Endocrinol Metab. 83: 18539.
  • 11
    Wiklund, O., Fager, G., Craigh, I. H. (1980) Alphalipoprotein cholesterol levels in relation to acute myocardial infarction and its risk factors. Scand J Clin Lab Invest. 40: 239247.
  • 12
    Westberg, L., Baghaei, F., Hellstrand, M., et al (2001) Relation between polymorphisms of the androgen receptor and the estrogen receptor β genes and serum levels of androgens in women. J Clin Endocrinol Metab. 86: 25622568.
  • 13
    Björntorp, P. (1992) Visceral obesity: a civilization syndrome. Obes Res. 1: 216222.
  • 14
    Folkow, B. (1982) Physiological aspects of primary hypertension. Physiol Rev. 62: 384504.
  • 15
    Björntorp, P., Holm, G., Rosmond, R. (1999) Hypothalamic arousal, insulin resistance and type 2 diabetes mellitus. Diabetes Med. 16: 111.
  • 16
    Ackerman, G. E., Smith, M. E., Mendelson, C. R., MacDonald, P. C., Simpson, E. R. (1981) Aromatization of androstenedione by human adipose tissue stromal cells in monolayer culture. J Clin Endocrinol Metab. 53: 412417.
  • 17
    Riggins, G. J., Lokey, L. K., Chastain, J. L., et al (1992) Human genes containing trinucleotide repeats. Nat Genet. 2: 186191.
  • 18
    Comings, D. E. (1998) Polygenic inheritance and micro/minisatellites. Mol Psychiatry. 3: 2131.
  • 19
    Lindstedt, G., Lundberg, P. A., Lapidus, L., Lundgren, H., Bengtsson, C., Björntorp, P. (1991) Low sex-hormone-binding globulin concentration as independent risk factor for development of NIDDM: 12-yr follow-up of population study of women in Gothenburg, Sweden. Diabetes. 40: 123128.
  • 20
    Haffner, S. M. (1996) Sex hormone-binding globulin, hyperinsulinemia, insulin resistance and noninsulin-dependent diabetes mellitus. Horm Res. 45: 233237.
  • 21
    Lapidus, L., Helgesson, Ö, Merck, C., Björntorp, P. (1988) Adipose tissue distribution and female carcinomas: a 12-year follow-up of the participants in the population study of women in Gothenburg, Sweden. Int J Obes Relat Metab Disord. 12: 361368.
  • 22
    Jones, M. E. E., Thorburn, A. W., Britt, K. L., et al (2000) Aromatase-deficient (ArKO) mice have a phenotype of increased adiposity. Proc Nat Acad Sci USA. 97: 1273512740.
  • 23
    Elbers, J. M. H., Asscheman, H., Seidell, J. C., Megens, J. A., Gooren, L. J. G. (1997) Long-term testosterone administration increases visceral fat in female to male transsexuals. J Clin Endocrinol Metab. 82: 20442047.
  • 24
    Lovejoy, J. C., Bray, G. A., Bourgeois, M. O., et al (1996) Exogen androgens influence on body composition and regional fat distribution in obese postmenopausal women: a clinical research center study. J Clin Endocrinol Metab. 81: 21892203.