Influence of body mass index and age on the grade of hair growth in hirsute women of reproductive ages*

Transcription

1 FERTILITY AND STERILITY Copyright 1988 The American Fertility Society Vol. 50, No. 2, August 1988 Printed in U.S.A. Influence of body mass index and age on the grade of hair growth in hirsute women of reproductive ages* Kristiina Ruutiainen, M.D. t+ Risto Erkkola, M.D.t Matti A. Gronroos, Ph.D. Kerttu Irjala, M.D. II Turku University Central Hospital and Turku University, Turku, Finland The relationships between the hair growth in different body regions, body mass index (BMI) and age were studied in 225 women of reproductive ages referred for hirsutism. The regularity of the cycles was registered, and 109 of the patients were interviewed for their maximum weight, teenage obesity, and age of menarche. The serum androgens were measured in the follicular phase. The results indicate that facial hirsutism is associated with BMI (p = 0.41, P < ) independently of age and the testosterone (T) to sex hormone-binding globulin (SHBG) ratio. Facial hirsutism is also correlated with age (p = 0.37, P < ) irrespective of BMI (p = 0.26, P < ) or the T /SHBG ratio (p = 0.43, P < ). In contrast, the hair growth on trunk area is related to the T/SHBG ratio (p = 0.35, P < ) but not to BMI or age when the correlations are adjusted for the grade of hyperandrogenemia. The women with severe facial hirsutism had a higher maximum weight (P < ) and more teenage obesity (P < 0.01) than other hirsute patients. They also had a slightly earlier menarche compared with their agemates than the women with mild or absent facial hair (P < 0.05). The data suggest differencies in the regulation of hair growth between the face and trunk areas. Fertil Steril 50:260, 1988 Hirsutism is associated with increased androgen production rates 1 and with high total and free androgen levels. 2 3 Androgen production and metabolism have been increased in obesity, 4 too, and the low serum concentrations of sex hormone-binding globulin (SHBG) in overweight women 5 result in high levels of biologically active testosterone (T). 6 7 Similarities in androgen metabolism have been considered to suggest an association between obesity and hirsutism. 8 Received August 13, 1987; revised and accepted April5, * Supported by a grant from the Emil Aaltonen Foundation, Tampere, and from the Duodecim Foundation, Helsinki, Finland. t Department of Obstetrics and Gynecology, Turku University Central Hospital. :j: Reprint requests: Kristiina Ruutiainen, M.D., Department of Obstetrics and Gynecology, Turku University Central Hospital, Turku SF-20520, Finland. Department of Statistics, Turku University. II Central Laboratory, Turku University Central Hospital. 260 Ruutiainen et al. Hirsutism, weight and age Based on answers to questionnaires obtained from more than 26,000 premenopausal women, Hartz et al. 9 were able to demonstrate a relationship between obesity and virile hair growth with facial hair. Facial hirsutism was increased with a longer duration of obesity. The androgen profiles of the patients were not studied. In hirsute women of reproductive ages, the serum androgen concentrations are significantly decreased by age. 7 However, more than 20 years ago, Ferriman and Gallwey 10 noticed that the hairiness tended to increase on the face and disappear from other sites with advancing years. Whether this was an aging phenomenon or associated with the increased frequency of obesity in older women was not studied. The present work was undertaken to explore the relationships of hirsutism in different body regions to body mass index (BMI) or age after eliminating the effect of changing androgen levels. Another purpose was to search after possible Fertility and Sterility

2 differences in previous overweight, menstrual pattern, age of menarche, and serum androgen par ameters between patient groups with different localization of virile hair growth. Patients MATERIALS AND METHODS The study group consisted of 229 consecutive premenopausal patients referred for hirsutism. All patients were clinically examined by one of us (KR). Four patients with an adrenogenital syndrome or an androgen-secreting tumor were excluded from consideration. The age of the patients ranged from 17 to 52 years (mean 30.8, standard deviation [SD] 5.2). A complete gynecologic examination was made. Pelvic ultrasound, laparoscopy, computer tomography of the sella and the adrenals, ( 131 l)iodocholesterol scanning of the adrenals, selective catheterization of the adrenal and ovarian veins, and adrenocorticotropin (ACTH) stimulation and dexamethasone suppression tests were made when needed to exclude a possibility of a virilizing tumor. All except one patient had normal cortisol diurnal variations, but she also had normal urine free cortisol secretion. None of the women had true hyperprolactinemia (prolactin [PRL] values< 40 ~g/1), and all were euthyreotic. Ninetyseven of the women had regular menstrual cycles (28 ± 4 days), 41 had slightly irregular cycles (28 ± 7 days), and 83 were oligomenorrheic (menstrual cycles > 35 days). Two young patients who had previously had regular cycles no longer menstruated at the time of the study as they had undergone hysterectomy. The data of menstrual cycles failed in two patients. The terminal hair growth in 11 body regions was graded according to a modification of the criteria of Ferriman and Gallwey. 10 Steps of 0.5 were used, and the scores ranged from 0.0 to 4.0. The body areas defined by Ferriman and Gallwey were grouped as follows: face (regions 1-2), trunk (3-8 and 10), and total (1-11). The quantitation of hirsutism preceded the determination of the laboratory values. The grade of hirsutism on the face area ranged from 0.0 to 8.0 (mean 3.0, SD 2.3); on the trunk area, from 0.0 to 18.0 (mean 6.1, SD 3.6); and on the total body area, from 0.0 to 31.0 (mean 12.6, SD 5.6). BMI, the ratio weight/height 2 was assessed by a nomograph method of Thomas et al. 11 The BMI of the patients ranged from 16 to 49 kg/m 2 (mean 24.6, SD 5.8) and the weight of the women from 43 to 137 kg. The historical data of possible teenage obesity was obtained from 109 women and of the maximum weight of the patient from 90 women. The teenage obesity was defined by the patient herself as follows: 1 = lean or normal weight, 2 = overweight, 3 = very obese. The maximum weight of the women ranged from 46 to 135 kg (mean 79.3, SD 19.9). The historical data of the age of menarche were obtained from 109 women. The patients' menarcheal age ranged from 10 to 20 years (mean 12.7, SD 1.5). The difference between the patient's menarche and the mean age of menarche in the respective age group(.:\ Mage) was calculated by regarding 14 years as the mean of normal menarche in women born 1940 or earlier and 13 years in women born 1941 or later. 12 The difference was expressed in steps of half years, -1 meaning the menarche 1 year earlier and year later than that of the agemates. The patients were divided into groups according to the location of the excessive hair growth. The patients in the group A had pure facial hirsutism with the facial score of terminal hair ::::: 4.0 but with the trunk score of <5.0. The group B consisted of women with male-pattern hair scattered both on the face (score::::: 4.0) and on the trunk (score;;:: 5.0) areas. The females in the group C had their excessive hair growth concentrated on the trunk area (trunk score ::::: 8.0, facial scores; 3.0). Hormone Assays The blood samples collected in the early follicular phase (days 3 to 7) of the menstrual cycle were included (N = 200). In the oligomenorrheic women, the specimens were taken during a progesteronechallenge withdrawal bleeding with dydrogesterone 10 mg/day for 10 daysp The sampling was carried out between 8 A.M. and 9 A.M. The patients had not used any hormonal medication for at least 2 months before the study. The serum values fort, androstendione (A), dehydroepiandrosterone sulfate (DHEAS), SHBG, estradiol (E 2), luteinizing hormone (LH), and follicle-stimulating hormone (FSH) were measured by radioimmunoassay (RIA) as described earlier. 2 The values for A missed in two women and those for SHBG in two other women. The serum T values ranged from 1.1 to 5.5 nmol/1 (mean 2.4, SD 0.7), the serum A values from 2.0 to 20.1 nmol/1 (mean 8.1, SD 3.3), DHEAS from Vol. 50, No.2, August 1988 Ruutiainen et al. Hirsutism, weight and age 261

3 1.6 to 16.8 umol/1 (mean 8.2, SD 3.2), SHBG from 12.0 to nmol/1 (mean 49.1, SD 22.3), and the T/SHBG ratio (10 3 ) from 12.0 to (mean 59.1, SD 33.6). In our laboratory, the normal values are for T 0.8 to 2.5 nmol/1, for A 4.0 to 9.0 nmol/1, for DHEAS 2.0 to 9.0 ~mol/1, and for SHBG 35.0 to nmol/l. Statistical Methods The main part of the statistical analysis was done using BMDP computer program, and the simultaneous statistical inference available in it was utilized. Kruskall-Wallis' analysis of variance, the two-sided Mann-Whitney U-test, and Spearman's rank correlation coefficient were used for the nonparametric statistical calculations. The calculation of partial Spearman's correlation is not straightforward. A rank analogue of partial product-moment correlation coefficient, which is sometimes used, is not even asymptotically distribution free. Instead of it, we used a partial Spearman's rank correlation coefficient proposed by Shirahata. 14 Shirahata 15 showed that the coefficient has good asymptotic properties. RESULTS There was a significant positive correlation between facial hirsutism and BMI (p = 0.41, P <, Table 1) independently of the effect of age or the T /SHBG ratio. The facial hair growth and age were also significantly related (p = 0.37, P <, Table 2) independently of the grade of hyperandrogenemia (p = 0.43, P < after adjusting for the T /SHBG ratio). The association of facial hirsutism to age was, however, to some de- Table 2 The Correlations (Simple and Partial Corrected for the Body Mass Index [BMI] and for the Testosterone/Sex Hormone-Binding Globulin [T/SHBG] Ratio) Between the Hair Growth and the Age Hirsutism score Face Trunk Total Simple (N = 223) b NSd p < b p < c p <. d NS, not significant. Age Corrected Corrected forbmi fort/shbg (N = 223) (N = 198) NSd -0.14" NSd gree a consequence of the high BMI in elderly women (the correlation between age and BMI was p = 0.40, P < ), but a correlation of p = 0.26, P < 0.01 (Table 2) was still found after eliminating the effect of BMI. No correlation was found between facial hirsutism and the T /SHBG ratio when the effect of the higher T/SHBG ratio in obese women 7 was eliminated (Table 3). There was only a weak relationship between the hirsutism score on trunk area and BMI (p = 0.16, P < 0.05, Table 1). It appeared to be a consequence of the higher androgenicity in obese women, 7 as no correlation could be found after adjusting for the T/SHBG ratio. Similarily, the negative correlation between the grade of hirsutism on trunk area and age (p = -0.21, P < 0.01, Table 2) was associated to the decrease in serum androgens 7 by age. The excessive hair growth on trunk area was correlated to the T/SHBG ratio (p = 0.35, P <, Table 3). Table 1 The Correlations (Simple and Partial Corrected for the Effect of the Age and the Testosterone/Sex Hormone Binding Globulin [T/SHBG] ratio) Between the Hair Growth and the Body Mass Index (BMI) Hirsutism score Face Trunk Total Simple (N = 223) 0.41" 0.16" 0.27c p < b p < c p <. d NS, not significant. BMI Corrected for age (N = 223) C 262 Ruutiainen et al. Hirsutism, weight and age Corrected fort/shbg (N = 198) 0.41" NSd 0.20b Table 3 The Correlations (Simple and Partial Corrected for the Body Mass Index [BMI] and for the Age) Between the Hair Growth and the Testosterone/Sex Hormone-Binding Globulin (T/SHBG) Ratio T/SHBG Corrected Corrected Hirsutism Simple forbmi for age score (N = 198) (N = 198) (N = 198) Face 0.21" NS 0.29b Trunk 0.35b 0.26b 0.29b Total 0.32b 0.21" 0.29b p < b p <. NS, not significant. Fertility and Sterility

4 Table 4 The Differences in the BMI, Age, and Menarche Between Two Hirsute Patient Groups ( <4)a Hirsutism score Face Trunk Total Age (yr) BMI Max. weight (kg) Teenage obesity Menarche ~Mage Oligomenorrhea(%) Facial score~ 4 (n = 89) 5.4 ± ± ± ± ± ± ± ± ± Facial score < 4 (n = 114) 1.4 ± ± ± ± ± ± ± ± ± P< NSb O.ol NS 0.05 NS a The differences in testosterone (T), androstendione, sex hromone-binding globulin (SHBG), the T/SHBG ratio, estradiol, and the LH/FSH ratio were not significant. The dehydroepiandrosterone sulfate values were slightly lower (P < 0.05) in the former group with a higher mean age. b NS, not significant. The effects of BMI and age on this correlation were weak. When the women with severe facial hair growth (hirsutism score on face ~ 4.0, the groups A+ B) were compared to the other hirsute patients (facial score < 4.0) there was a significant difference in BMI (P < ) and in the maximum weight (P <, Table 4) between the groups. Teenage obesity was also more common (P < 0.01) in women with severe versus mild or absent facial hirsutism. Moreover, there was a significant difference in the age (P < ) between these two patient groups (Table 4). No difference was found in the age of menarche. However, when taking into account the decline of the menarcheal age since the beginning of the century (~ M age), it appeared that the women with severe facial hirsutism had an earlier menarche compared with their agemates. The two groups with different localization of the excessive hair growth were similar in terms of their current hormonal values. When comparing the groups A, B, and C, the patients with facial hirsutism (groups A and B) had a higher BMI (P < 0.01) and maximum weight (P < 0.01) than the women with hirsutism concentrated only on the trunk area (Group C, Table 5). The groups B and C differed from each other in terms of their teenage obesity (P < 0.05, Table 5). The patients with pure facial hirsutism (group A) had an earlier menarche (P < 0.05) and~ Mage (P < 0.01) than the women in groups B and C (Table 5). The women in group B who had the highest total hirsutism score were more often oligomenorrheic than the women in group A (P < 0.05, Table 5). The hormone values were similar in all three patient groups. DISCUSSION The results of the present study indicate a close relationship between facial hirsutism and obesity. This is in agreement with the analysis of questionnaires obtained from more than 26,000 premenopausal females 9 : the women who considered themselves hirsute concerning the face and possibly Table 5 The Differences in the BMI, Age, and Menarche Between Three Patient Groups with Pure Facial Hirsutism (Group A, n = 35), with Hirsutism Both on the Face and on the Trunk (Group B, n = 54) and with Hirsutism Concentrated only on the Trunk (Group C, n = 26)a Group A Group B Group C A-B A-C B-C P< P< P< Hirsutism score Face 5.3 ± ± 1.4 Trunk 2.7 ± ± 3.6 Total 10.6 ± ± 5.1 Age (yr) 36.1 ± ± 7.7 BMI 27.0 ± ± 6.9 Max. weight (kg) 86.3 ± ± 18.7 Teenage obesity 1.8 ± ± 0.8 Menarche (yrs) 12.2 ± ± 1.3 ~Mage -0.9± ± 1.3 Oligomenorrhea(%) a The differences in testosterone (T), androstendione, sex hormone-binding globulin (SHBG), the T/SHBG ratio, and the LH/FSH ratio were insignificant. The dehydroepiandrosterone sulfate values were slightly lower in group A than in group C (P 1.4 ± 1.0 NSb 10.2 ± 1.9 NS 15.8 ± ± ± 4.4 NS ± 12.7 NS O.ol O.ol 1.3 ± 0.6 NS NS ± NS +0.3 ± NS NS NS < 0.05), and the estradiol values slightly higher in group B than in group C (P < 0.05). b NS, not significant. Vol. 50, No.2, August 1988 Ruutiainen et al. Hirsutism, weight and age 263

5 other areas were 4 kg heavier than the women who considered themselves normal. The association between facial hirsutism and body mass was also seen in the present work when comparing the groups with severe versus mild facial hair growth: the former had a higher mean of maximum weight and more teenage obesity. Hartz et al. 9 noticed that excess facial hair was associated with previous obesity even after adjusting for present obesity. We found that the significant positive correlation between facial hair and BMI was independent of the current grade of hyperandrogenemia. On the other hand, the positive correlation between facial hirsutism and the T /SHBG ratio appeared to be a consequence of a higher T /SHBG ratio in obese women, 7 and we could not find any relationship between facial hairiness and the grade of hyper androgenemia after adjusting the correlation for the effect of BMI. The correlation of the facial hair growth to current and previous overweight could be due to metabolic alterations associated primarily with obesity. On the other hand, obesity, heaviness before menarche, and hirsutism are features of PCOD, and the relationship between hirsutism and overweight does not necessarily imply a causalship. Facial hirsutism appears to be significantly associated with age, too, despite the decline in the serum androgen values during aging. 7 When one possible contributing factor, the effect of a high BMI in elderly women, is eliminated, facial hirsutism and age do still relate. Formerly, also Ferriman and Gallwey 10 noticed that hair tended to increase on the face and disappear from other sites with advancing years. They did not, however, calculate the statistical significance of the change in the hairiness. Neither did they study whether this was an independent aging phenomenon or secondarily associated with changes in serum androgens or incidence of obesity. Although most hirsute women overproduce a variety of active androgens, the expression of androgen action is modulated by peripheral tissue events. The enzyme Sa-reductase (5a-R) converts T to dihydrotestosterone (DHT), which mediates androgen action in the target tissues. The 5a-R activity is stimulated by androgens 16 in some skin areas. Once the enzyme becomes activated, it remains sensitive to low androgen levels for its life cycle. 17 It has been suggested that the increase in 5a-R in some normoandrogenemic hirsute patients might be due to an enzyme induction by hyperandrogenemia at puberty. 18 Also facial hair is dependent on the 5a-R activity since the male patients with 5a-R deficiency hermaphroditismus have absent facial hair in spite of partial signs of virilism in puberty. 19 Direct studies on the 5a-R activity on facial area and its dependency on androgens have not yet been published. Changes in enzyme activity and peripheral androgen metabolism might explain the correlation of facial hair growth with obesity and age in spite of the natural decline in serum androgen values. This could be one possible reason for the confusing clinical picture of virile hair growth on face in normoandrogenemic patients. It remains, however, to be studied whether aging phenomena independent of androgen action could increase the facial coarse terminal hair classically considered androgen dependent. The hirsutism on the trunk area was strongly associated with the T /SHBG ratio even after elimination of the effect of BMI. In contrast to facial hirsutism, the hair growth in trunk region does not correlate with age, and it has only a weak positive correlation with BMI. Moreover, this relationship was found to be associated with the higher levels of biologically active T in obese patients 7 because the correlation was not significant after adjusting it for the T/SHBG ratio. Formerly, Hartz et al. 9 were also unable to find evidence that women with virile hair growth other than on the face would be significantly heavier than normal women. The results might be explained by differences in the local androgen sensitivity or in the duration of the enzyme activation between face and trunk regions. The latter hypothesis is supported by a patient case of our own clinic: a 21-year-old woman was operated on in 1981 because of a rapidly virilizing neoplasm in her left adrenal. Postoperatively, the very high levels of serum androgens, especially T and DHEAS, were normalized, and the virile hair growth on trunk area disappeared in a few months. However, the severe facial hirsutism (Ferriman-Gallwey scale of 7) still persists 5 years after the operation. It appeared in our study that the women with severe facial hirsutism had a relatively earlier menarche than the other hirsute patients. The differences in the menarcheal age between the patient groups might reflect differences in body mass development in adolescence, as childhood obesity is associated with early menarche Childhood obesity is also associated with menstrual disturbancies, 21 and obese adolescent girls with irregular menses have been shown to have 264 Ruutiainen et al. Hirsutism, weight and age Fertility and Sterility

6 hormonal changes similar to those in PCOD. 22 In the present study, the only significant difference in the current menstrual patterns of the different patient groups was a high number of oligomenorrheic patients in the group with the highest total hirsutism score. Unfortunately, the design of our work does not allow us to answer the question if these patients had more menstrual abnormalities already in their adolescence. In summary, facial hirsutism is associated to obesity and aging irrespecively of the current T I SHBG ratio. On the contrary, the excessive hair growth in the trunk region correlates to the T I SHBG ratio but not with BMI or age when adjusted for the grade of hyperandrogenemia. The differences in the hair growth between the face and trunk regions need further evaluation. REFERENCES 1. Bardin CW, Lipsett MB: Testosterone and androstendione blood production rates in normal women, with idiopathic hirsutism and polycystic ovaries. J Clin Invest 46:891, Ruutiainen K, Erkkola R, Kaihola H-L, Santti R: The grade of hirsutism correlated to serum androgen levels and hormonal indices. Acta Obstet Gynecol Scand 64:629, Hosseinian AH, Kim MH, Rosenfield RL: Obesity and oligomenorrhea are associated with hyper androgenism independent of hirsutism. J Clin Endocrinol Metab 42:765, Samojlik E, Kirschner MA, Silber D, Schneider G, Ertel NH: Elevated production and metabolic clearance rates of androgens in morbidly obese women. J Clin Endocrinol Metab 59:949, Kopelman PG, Pilkington TRE, White N, Jeffcoate SL: Abnormal sex steroid secretion and binding in massively obese women. Clin Endocrinol12:363, Zhang Y, Stern B, Rebar RW: Endocrine comparison of obese menstruating and amenorrheic women. J Clin Endocrinol Metab 58:1077, Ruutiainen K, Erkkola R, Gronroos MA, Irjala K: Androgen parameters in hirsute women: correlations with body mass index and age. Fertil Steril 50:255, Kirschner MA, Samojlik E, Silber D: A comparison of androgen production and clearance in hirsute and obese women. J Steroid Biochem 19:607, Hartz AJ, Barboriak PN, Wong A, Katayama KP, Rimm AA: The association of obesity with infertility and related menstrual abnormalities in women. Int J Obes 3:57, Ferriman D, Gallwey JD: Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 24:1440, Thomas AE, McKay DA, Cutlip MB: A nomograph method for assessing body weight. Am J Clin Nutr 29:302, Widholm 0, Kantero R-L: A statistical analysis ofthe menstrual patterns of 8000 Finnish girls and their mothers. Acta Obstet Gynecol Scand 30(Suppl14):7, Heineman MJ, Thomas CMG, Doesburg WH, Rolland R: Hormonal characteristics of women with clinical features of the polycystic ovary syndrome. Eur J Obstet Gynec Reprod Bioi 17:263, Shirahata S: Tests of partial correlation in a linear model. Biometrika 64:162, Shirahata S: Rank tests of partial correlation. Bull Mathern Statistics 19:9, Mowszowicz I, Kirchhoffer MO, Kuttenn F, Mauvais-Jarvis P: Dihydrotestosterone stimulates 5a-reductase activity in pubic skin fibroblasts (PSF). J Clin Endocrinol Metab 56:320, Thomas JP, Oake RJ: Androgen metabolism in the skin of the hirsute women. J Clin Endocrinol Metab 38:19, Mauvais-Jarvis P: Regulation of androgen receptor and 5areductase in the skin of normal and hirsute women. Clin Endocrinol Metab 15:307, Imperato-McGinley J: The syndromes in male pseudohermaphroditism and 5a-reductase deficiency. In Regulation of androgen action, Edited by N Bruchovsky, A Chapdelaine, F Neumann. Berlin (West), R Brucner, 1985, p Frisch RE, Revelle R: Height and weight at menarche and a hypothesis of menarche. Arch Dis Child 46:695, Yen SSC: The polycystic ovary syndrome. Clin Endocrinol (OxO 12:117, Kaufman ED, Mosman J, Sutton M: Characterization of basal estrogen and androgen levels and gonadotropin release patterns in the obese adolescent females. J Pediatr 98:990, 1981 Vol. 50, No.2, August 1988 Ruutiainen et al. Hirsutism, weight and age 265