Maturitas, 12 (1990) 137-143 Elsevier Scientific Publishers Ireland Ltd.

137

MAT 00583

Age, steroids and bone mineral content S. Rozenberg,

H. Ham, D. Bosson, A. Peretz and C. Robyn

Interdisciplinary Group on Osteoporosis, Free University of Brussek (VUB-ULB), Hospital, Hoogstraat 322, loo0 Brussek (Belgium)

St Peter’s

(Received 1 June 1989; revision received 24 October 1989; accepted 10 March 1990)

The possible existence of correlations between bone mineral content (BMC), age and serum levels of steroid hormones was investigated. It was found that dehydroepiandrosterone sulphate (DHEA-S), oestradiol (E*) and delta4androstenedione (A) were correlated with BMC, whereas oestrone (E,) and testosterone (T) were not. Partial correlations after adjustment for age were significant (P < 0.05) only between E, and DZfEA-S and BMC at the L2-L4 lumbar site (BMC-1) and between DHEA-S (P < 0.01) and BMC at the midradius site (BMC-r). Stepwise multiple regression analysis showed that, apart from age, E, was the only factor to fit (P < 0.05) into the mathematical model with BMC-I as the dependent variable, while DHEA-S was the only factor to fit (P < 0.01) with BMC-r as the dependent variable. These data suggest that different hormonal inBuences are related to BMC at different sites, namely E, to lumbar trabecular bone (LZL4) and DHEA-S to cortical bone (midradius). (Key words: Osteoporosis, Oestrogens, Androgens, Steroid hormones, Bone parameters, Aging)

Introduction Osteoporosis is an age-related disorder characterized by reduced bone mass [l]. Women suffer from this disease more frequently than men [l] and the fact that bone loss increases after the menopause has been extensively documented [2-51. Oestrogen replacement therapy prevents accelerated bone loss as a consequence of natural or surgical menopause [3,4]. Other amenorrhoeic conditions with associated hypogonadism have been linked with bone loss [a]. It is widely accepted that oestrogen deficiency is responsible for the observed accelerated bone loss. There is controversy, however, concerning the patterns of agerelated bone loss at different sites in the skeleton and their relationship with the oestrogen deficiency that occurs at menopause and thereafter (71. The production of sex steroid hormones other than oestrogens also decreases with age and may influence bone mineral content (BMC) [8-lo]. The purpose of this cross sectional study, conducted in a large group of women, was to investigate the relationship between BMC at different sites in the skeleton and the serum levels of steroid hormones of ovarian and adrenal origin. 0378~5122/90/$03.50 0 1990 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland

138

Subjects and methods A review was carried out of recorded BMC and steroid hormone measurements in 617 women inpatients and outpatients aged 40 or over (mean * SEM 63.3 f 0.3 years) held in a databank compiled for the study of osteoporosis. The patients were classified into 5-year age groups (see Fig. 1) as follows: 40-45 (n = 59), 46-50 (n = 98), 51-55 (n = 96), 56-60 (n = 108), 61-65 (n = 72) , 66-70 (n = 42), 71-75 (n = 47), 76-80 (n = 44) and > 80 (n = 51). None of these women had received oestrogen-progestogen or oestrogen replacement therapy. The BMC measurements were made at the midradius site (BMC-r) in the nondominant arm and in the lumbar (L2-L4) spine (BMC-1). BMC-r was measured by single photon absorptiometry (Cameron Nordland 278 Bone Densitometer, U.S.A.) and BMC-1 was measured by dual photon absorptiometry (NOVO, BMC-lab 22a, Denmark). The reproducibility of the readings in clinical practice was to within 4% [l 11. Serum 17/I-oestradiol (EJ, oestrone (E,) androstenedione (A), testosterone (T) and dehydroepiandrosterone sulphate (DHEA-S) were measured by radioimmunoassay using polyclonal antibodies [ 12- 151. The intra-assay coefficients of variation were below 7% and the interassay coefficients of variation ranged between 7% and 13%. One-way analysis of variance was performed to test the statistical significance of the differences between the BMC values in the various age groups [16]. The significance of the differences (070)between BMC-1 and BMC-r was tested using the bivariate t-test [16]. Relationships between age, BMC and hormones were tested by correlation analyses (I?) and partial correlation analyses (I?$ after adjustment for age [16]. Prediction of BMC based on age and hormone values was tested by stepwise multiple regression analysis. The BMC variation fraction (R2) that could be explained by the mathematical model used in our investigation was calculated after the successive introduction of new variables (age and hormones). Each time a new variable was introduced into the model, the statistical significance of the increase in R2 was evaluated. However, the increase in R2 did not indicate what fraction of the variation was attributable to the newly introduced variable [ 161. Results The decreases in BMC-1 and BMC-r are shown in Fig. 1 for the various 5-year age groups. They are expressed as percentages of the mean value in the 40-45 group, in which the decreases in BMC-1 and BMC-r were, by definition, deemed to be 0%. Both BMC-1 (R = - 0.43, P < 0.001) and BMC-r (R = - 0.59, P < 0.001) decreased significantly with age. BMC-1 decreased more than BMC-r in the 56-60 group (P < O.OOl), while the latter decreased more in the 70-75 (P < 0.05), 76-80 (P < 0.01) and> 80 (P < 0.001) groups. When menopausal status was taken into consideration no significant differences were found between

1-3wa

(IO-0

pUa

>

W~I

d

‘60-o = *LIpm I I-O = “& s-vgHa > d ‘CI ‘0 = *H)S-VZHa Yl!M AIUO PV+JJO~

‘dpwadsa~

~~13 ‘ 3 qlp

@U=!JIuf&

paupxua~ ~-3~8 ~t3qlpunoj uaql S’IZM 31 .a% .xoj itxaurym[ptz r&e paleIn3pm alar sanph *a :IOO*O>d ‘II-O= 8 :V PUB 100’0 > d ‘WO - = 8 :‘a ‘tIOO’0 > d 'LE'O- = H :s-VgHa) a%a qph pa~qa1103 d~a~ya%au 0s~ alaM sauoulroq pi0 -JajS aXI$

‘(LlaAyXxiSaJ

‘65yj -

‘dlayqadsal

‘IE*O =

(~la+l~adsaJ ‘blaAy3adsar ~-3yya

‘I()‘()

> d ‘PI.0

= 8 pW !sO’O > d ‘I I()‘() = 8) V PUB

'10'0> d ‘PI.0= 2lPUe :IOO’O > d ‘61’0 = Xl ‘El ‘(100-O > d

put3 1-3~8

u) uawow

g pus zz*0 =

=

8 pule Et-0 -

8) S-VBHCI W!J+’4aW!sod Pm (100’0 > d = 8) a?& ql!~ Alahyla%au papIa3103 alah

my1 pa@aAal (y) sluap!JJao3

pmedouauwsod

-aid aql ui (~oyj

aq

> d pm

uy utql

10fyo

uo!yqa~Jo:, aq3 Jo uoyqn3~3 $CIahyadsa~ ‘89 pm EP =

(QaA!padSaJ

> d) lay%y

$2

Qiue3g@?~s

pm

ss

= u) pmedouam

alaM sari@@@I-3wa

aql

sdnoS3 0~1 asay UI *dluo ‘-3Ha 01 pM?aJ qqrn sdnoS3 Ss-Is put3 OS--9p aql u! am aql %yaq sum aql ‘1-3gya 10 I-~yva laqlja 01 pm%al ql!M dnol% S~-_OP aql u! uawow (~1 = U) ptmadoriaux-Isod pua (zp = U) pmmdouaux-aJd aql

moJ% 3% (100.0 > d) 08 < PUB (IO’0 > cf) 08-9L ‘(SOO’O > b SL-OL W U! J)WB?Jfi~llu=?@!s sm ~allef aql apqm ‘dno18 a% 09-95 aql u! ‘-3~~8 II! aql mqi ~alw~3 (rt)cro > d) ~~IUBXJ$~~ mn 1-3~8 u! aseamap aqL ‘o,+,oaq 01 plump s! SSOI3~18 ‘uo!lprJap bq ‘qcnqm u! ‘dnol% a%8 sp--op aql u! 3~8 jo aSwuamd e se paswdxa ale sari@@@ *dnorIl a% nza&s 01 %tqp1033~ sa~!s (J-D~@ sn!pelp!w pua (1-3~~) lequtnl p~-z? aql le (ssol 3yva oh) lualuo3 pzslaytu auoq u! as8ai3aa *I *8!~

01-99

08-91

09-99

09-9,

%GE-

’

1 %O&-

6Ef

Stepwise multiple regression analysis showed that age accounted for 17.2% ( =

R2 x 100%) and 34.9% of the decreases in BMC-1 and BMC-r, respectively. All the regression and correlation analyses relating to the post-menopausal women yielded identical results to those relating to the entire study population. E, fitted into the equation (multiple regression equation analysis) with BMC-1 as the dependent variable, increasing R2 (17.9%) slightly but significantly, but it did not fit the equation when BMC-r was included. DHEA-S significantly increased R2 (35.8%) with BMC-r as the dependent variable, but it was rejected from the equation that included BMC-1 once E, was introduced. Discussion In the immediately post-menopausal 56-60 age group, BMC loss was faster at the lumbar site (mainly trabecular bone) than at the midradial site (mainly cortical bone). Moreover, women who became post-menopausal before the age of 55 already had significantly decreased BMC at the lumbar (but not the midradius) site in comparison with pre-menopausal women of the same age. Excess bone loss during the immediate post-menopause corresponds to type I osteoporosis [ 171. During the 61-70 decade, there was no difference in bone loss at the trabecular and cortical sites. After age 70, BMC loss became progressively more pronounced at the midradial site and affected cortical bone more markedly than trabecular bone. Excess bone loss during this period of life corresponds to type II or senile osteoporosis [ 171. The developments in bone loss profiles after the menopause described here confirm and extend the observations reported by other investigators [17]. There is also a decline in the circulating levels of steroid hormones at the menopause [18]. The most dramatic is the fall in E,, which is produced almost exclusively by the ovary [ 181. Serum E, levels were already found to have reached their lowest values in the 51-60 decade of age [18]. Similar rapid falls in T and A have also been described [18], but they are less marked than the drop in E,. This is probably due to the significant adrenal contribution to androgen production and to the relatively smaller reduction in ovarian androgens after the menopause [ 191. The decline in DHEA-S after the menopause is more progressive than that in E,, T or A. The serum levels of DHEA-S, which is almost exclusively of adrenal origin, fall to their minimal values between 60 and 70 years of age, i.e. one decade later than those of serum E,; the mechanism is still not completely understood [20]. Oophorectomy in pre-menopausal women results in only a modest decrease in serum DHEA-S as compared with that which follows natural menopause [20]. It is thus apparent that non-ovarian factors control DHEA-S secretion, the declining influence of which appears to be responsible for post-menopausal decrease in circulating DHEA-S. When post-menopausal BMC and hormone profiles are compared, it is seen that the rather slow decrease in BMC at the radial site seems to be more closely

141

related to the slow decrease in DHEA-S, while the rapid fall in BMC at the lumbar site seems to be linked more strongly to the steep drop in E,. However, the BMC decline at both sites may be age-related rather than hormone-related. In order to determine whether the correlations between hormone profiles and bone loss rates were coincidental, statistical analyses were therefore conducted after adjustment for age. It was then found that only E, and DHEA-S remained correlated with BMC at the lumbar site and DHEA-S alone remained correlated with BMC at the midradial site. Moreover, in the stepwise multiple regression analyses that were performed E, was the only factor, apart from age, that fitted the equation with BMC-1 as the dependent variable. However, with BMC-r as the dependent variable, DHEA-S was the only factor to fit the equation besides age. The outcome of these age-adjusted data analyses confirmed that post-menopausal trabecular bone loss is related to the cessation of the ovarian E, secretion and that the later loss in cortical bone is related to the decline in DHEA-S. BMC loss at both the radial and the lumbar sites appears to be much more closely related to age than to hormone deficiency, age and hormonal changes together accounting for only some 35% of the variations. Thus, major factors still remain to be identified in the aetiology of osteoporosis. Available data on the influence of exogenous oestrogens on BMC at various sites in the skeleton serve to provide an indication of the preferential site of oestrogen action on bone, expecially since exogenous oestrogens do not alter circulating levels of DHEA-S [20]. There is a large body of data which indicates that exogenous oestrogens reduce bone loss at virtually all sites in the skeleton when administered at the time of the menopause or shortly thereafter [4,5,21,22]. Epidemiological studies have revealed that oestrogens not only prevent bone loss but also fractures, particularly vertebral crush fractures and fractures of the hip [231. Vertebrae contain essentially trabecular bone, while the hip comprises a mixture of trabecular and cortical bone [ 171. However, hip fractures occur much later in life than vertebral fractures and it has accordingly been suggested that hip fractures may be influenced not only by oestrogen deficiency but also other factors such as impairment of vitamin D metabolism [24]. Possible endocrine control of these factors should therefore be considered. Contradictory data have been reported on the relationship between androgens and bone loss. Lower-than-normal serum androgen levels were recorded in osteoporotic women by some authors [9,10] but not by others [25,26]. Moreover, significant relationships between endogenous E, (converted from androgens) and BMC have been described in patients [27]. In young women with supraphysiological levels of endogenous androgens, higher-than-normal trabecular bone density has been reported [28]. Furthermore, adrenal androgens have been shown to play an essential role in the maintenance of bone mass in post-menopausal women suffering from Addison’s disease [29]. In addition, it has been reported, on the basis of clinical trials, that androgenic medication reduces bone loss [301. However, the action of the drugs used is not clearly understood; the mechanism may involve indirect effects on bone due to modification of the muscular mass or direct effects on bone cells. In vitro studies have provided some evidence that

142

androgens exert proliferative effects on bone and bone cells [31,32] which are not mediated by estrogens. In the present study, significant relationships could not be demonstrated between BMC at either of the sites investigated and T, A, or even E, after the data were adjusted for age, although correlations were still found between BMC1 and in particular BMC-r (cortical bone) and the adrenal androgen DHEA-S. Nordin et al. also reported low serum DHEA levels in osteoporotic patients [lo]. The mechanism whereby patients with low DHEA and DHEA-S levels suffer reduced cortical bone mass remains to be clarified. In conclusion, on the basis of the findings of our analyses of the BMC and biological data, we can state that both E, and DHEA-S play a role in bone maintenance. E, would appear to exert a greater influence on trabecular bone (L2L4), while either DHEA(-S) or factors that control its secretion and/or metabolism (or even other factors that it controls itself) would seem to have the greater effect in the case of cortical bone (midradius). Acknowledgements This work was supported by Research Grant No 3452787 from the ‘Fonds de La Recherche Scientifique Medicale’ of Belgium. We are grateful to Dr S. Lynch (Birmingham and Midland Hospital for Women, Sparkhill, UK) and to the National Hormone and Pituitary Program (National Institute of Arthritis, Diabetes and Digestive and Kidney Diseases-NIADDK, Baltimore, Maryland, USA) for reagents used in the radioimmunoassays, as well as to Mrs J. Delogne-Desnoeck and Mrs C. Nandance for their valuable technical assistance. References NIH consensus conference. .I Am Med Assoc 1984; 252: 799-802. Albright F, Smith PH, Richardson AM. Postmenopausal osteoporosis. Its clinical features. J Am Med Assoc 1941; 116: 2465-2474. Lindsay R, Hart DM, Forrest C, Baird C. Prevention of spinal osteoporosis in oophorectomised women. Lancet 1980; II: 1151-1154. 4 Christiansen C, Christensen MS, McNair P, Hagen C, Stocklund KE, Transb$l I. Prevention of early postmenopausal bone loss: controlled 2-year study in 315 normal females. Eur J Clin Invest 1980; 10: 273-279. 5 Richelson LS, Wahner HW, Melton III LJ, Riggs BL. Relative contributions of aging and estrogen deficiency to postmenopausal bone loss. N Engl J Med 1984; 311: 1273-1275. 6 Klibanski J, Neer RM, Beitins EZ, Ridgaway EC, Zervas NT, McArthur JW. Decreased bone density in hyperprolactinemic women. N Engl J Med 1985; 303: 1511-1514. Riggs BL, Wahner HW, Melton III LJ, Richelson LS, Judd HL, Offord KP. Rates of bone loss in the appendicular and axial skeleton of women. J Clin Invest 1986; 77: 1487-1491. Crilly RG, Francis PM, Nordin BEC. Steroid hormones, ageing and bone. Clin Endocrinol Metab 1981; 10: 115-139. 9 Marshall DH, Crilly RG, Nordin BEC. Plasma androstenedione and oestrone levels in normal and osteoporotic postmenopausal women. Br Med J 1977; 2: 1177-l 179. 10 Nordin BEC, Robertson A, Seamark RF, Bridges A, Philcox JC, Need AG, Horowitz M, Morris HA, Deam S. The relation between calcium absorption, serum dehydroepiandrosterone and vertebral mineral density in postmenopausal women. J Clin Endocrinol Metab 1985; 60: 651657.

143 11 12

13 14 15 16 17 18 19

20 21 22 23

24 25 26 27

28 29 30

31 32

Rozenberg S, Peretz A, Robyn C, Ham H. La reproductibilite des mesures du contenu osseux et de la densite osseuse par absorptiometrie biphotonique. J Biophys Biomtc 1987; II: 167-170. Delvoye P, Demaegd M, Uwayitu-Nyampeta L, Robyn C. Serum prolactin, gonadotrophins and oestradiol in menstruating and amenorrheic mothers during two years of lactation. Am J Obstet Gynecol 1978; 130: 635-639. Lejeune-Lenain C, Walter R, Franckson JRM. A direct radioimmunoassay for plasma delta-4androstenedione. Application to children. Clin Chim Acta 1979; 94: 327-329. Delbeke D, Lejeune-Lenain C. Simultaneous radioimmunoassay measurement of testosterone and dihydrotestosterone without chromatography. Ann Biol Clin 1982; 40: 579-584. Buster JE, Abraham GE. Radioimmunoassay of plasma dehydroepiandrosterone sulfate. Anal Lett 1972; 5: 543-551. Norusis MJ. Statistical Program for Social Sciences (SPSS/PC software) 1984, SPSS Inc. Chicago, USA. Riggs BL, Melton III LJ. Involution osteoporosis. N Engl J Med 1986; 315: 1676-1685. Rozenberg S, Bosson D, Peretz A, Caufriez A, Robyn C. Serum levels of gonadotrophins and steroid hormones in the post-menopause and later life. Maturitas 1988; 10: 215-224. Judd HL, Judd GE, Lucas WE, Yen SSC. Endocrine function of postmenopausal ovary: concentrations of androgens and oestrogens in ovarian and peripheral vein blood. J Clin Endocrinol Metab 1982; 39: 1020-1024. Cumming DC, Rebar RW, Hopper BR, Yen SSC. Evidence for an influence of the ovary on circulating dehydroepiandrosterone sulfate levels. J Clin Endocrinol Metab 1982; 54: 1069-1071. Rhs BJ, Christiansen C. Measurement of spinal or peripheral bone mass to estimate early postmenopausal bone loss? Am J Med 1988; 84: 646-653. Slemenda CW, Johnston CC. Bone mass measurement: which site to measure? Am J Med 1988; 84: 645. Kiel DP, Felson DT, Anderson JJ, Wilson PWF, Moskowitz MA. Hip fracture and the use of oestrogens in postmenopausal women: the Framingham study. N Engl J Med 1987; 317: 11691174. Lindsay R. The menopause: sex steroids and osteoporosis. Clin Obstet Gynecol 1987; 30: 847859. Manolagas AC, Anderson DC. Adrenal steroids and the development of osteoporosis in oophorectomised women. Lancet 1979; II: 597-600. Davidson BJ, Riggs BL, Wahner HW, Judd HL. Endogenous cortisol and sex steroids in patients with osteoporotic spinal fractures. Obstet Gynecol 1983; 61: 275-278. Cauley JA, Gutal JP, Sandler RB, La Porte RE, Kuller LH, Sashin D. The relationship of endogenous oestrogen to bone density and bone area in normal postmenopausal women. Am J Epidemiol 1986; 125: 752-761. Buchanan JR, Hospodar P, Myers C, Leuenberger P, Demers LM. Effects of excess endogenous androgens on bone density in young women. J Clin Endocrinol Metab 1988; 67: 937-943. Devogelaer JP, Crabbe J, Nagant de Deuxchaisnes C. Bone density in Addison’s disease: evidence for an effect of adrenal androgens on bone mass. Br Med J 1987; 294: 798-800. Geusens P, Dequeker J, Verstraeten A, Nijs J, VanHolsbeeck M. Bone mineral content, cortical thickness and fracture rate in osteoporotic women after withdrawal of treatment with nandrolone decanoate, l-alpha hydroxyvitamin D3, or intermittent calcium infusions. Maturitas 1986; 8: 281-289. Kasperk CH, Wergedal JE, Farley JR, Linkhart TA, Turner RT, Baylink DJ. Androgens directly stimulate proliferation of bone cells in vitro. Endocrinology 1989; 124: 1576-1578. Turner R, Wakley G, Hannon K, Bell N. Evidence that androgens are necessary for normal development and maintenance of bone. J Bone Min Res 1988; 3: S222.