The adrenal cortex and virilization

10 The Adrenal Cortex and Virilization T. JOSEPH McKENNA SEAN K. CUNNINGHAM THERESE LOUGHLIN

PHYSIOLOGY OF ADRENAL ANDROGENS Source, biosynthesis and secretion Androgens are secreted from the mature inner zones of the adrenal cortex, i.e. zona fasciculata and zona reticularis (Symington, 1969; O'Hare et ai, 1980). While dehydroepiandrosterone sulphate (DHAS) is almost uniquely derived from the adrenal and the vast majority of circulating DHA has a similar source, androstenedione is also secreted in significant quantities by the ovaries (Vermeulen, 1983). Approximately 25% of circulating testosterone in women arises from the adrenal, another 25% from the ovary and the remaining 50% has an extraglandular origin as the testosterone precursors DHA and androstenedione, mainly of adrenal origin, are converted to testosterone in blood (Vermeulen, 1983). Adrenal androgens, particularly androstenedione, can be converted to oestrogens, principally oestrone, and to a lesser extent testosterone is converted to oestradiol, in the periphery, for example, adipose tissue and liver (Figure 1) (Siiteri and MacDonald, 1973). The adrenal secretes significant quantities of androgens during fetal life, particularly DHAS (Tulchinsky, 1980), but there are only very low androgen concentrations present in circulation during the first five years of extra-uterine life. About the middle of the first decade plasma androgen levels start to rise (de Peretti and Forest, 1976). This is termed the 'adrenarche' and is clinically manifest by the appearance of pubic and axillary hair. In normal girls androgen levels continue to rise throughout the second decade. Plasma androgen levels are maintained in a relatively steady state until about the time of menopause when plasma DHAS levels gradually fall but subsequently become an important source of oestrogen for the post-menopausal woman (Marshall et al, 1978). The mechanism underlying the control of adrenal androgen secretion has not been established and it is a controversial topic. Basically, the issue revolves about whether there is a specific factor, or factors, responsible for Clinics in Endocrinology and Metabolism-Vol. 14. No.4, November 1985

997

998

T. J. McKENNA ET AL

ZONAE FASCICULATA AND RETICULARIS

I--------------------------~---------I • I

I I I

L~.s' S ~

PREGNENOLONE

: t"

~'s-l

PROGESTERONE

-

-

~IA':

;

: ZONA GLO~'ERULOSA ~

•

1-CZl : Cl9 ..... : ! 17-0H-PREGNENOLONE

t"

17·QH-PROGESTERONE

~

11-l1:0XYCDRTlSDL

;

:

----r:---

ANDROSTENEDIONE I ~ OESTRONE

t

h

TESTOSTERONE

1 I ~ OESTRADIOL I I I

!c CORTICOSTERONE

C18 ...

~

l"

:!

~

-- ---!b·---- --------tb--------1-j l1-DEOXYCORlICOSTERON

PERIPHERAL FAT

CORTISOL

,

I I I

---------------~~---------~

ALDOSTERONE

Figure 1. Schematic representation of adrenal and extraglandular steroidogenesis. Pathways bounded by - - indicate activity in the zona glomerulosa, ----- indicate activity in the zonae fasciculata and reticularis. Within the latter confines the area limited by . indicates activity in the cortisol-secreting cell and the area limited by '-'-'-'- indicates activity in the androgen-secreting cell. 6.5'S indicates the steroids termed dclta 5S and 6.4'S indicates those steroids termed delta 4S. (a) 3B-ol·dehydrogenase, 6.4-5-isomerase system; (b) 2I-hydroxylase; (c) ll-hydroxylase; (d) corticosterone methyl oxidase reactions types I and II; (e) 17-hydroxylase; (f) I7,20-desmolase; (g) steroid sulphotransferase; (h) 17-ketosteroid reductase; (i) peripheral aromatase. The number of carbon atoms in the various classes of steroids is indicated by C and a number, i.e. CZI, CI9, CIS.

stimulating adrenal androgen secretion which is distinct from adrenocorticotrophic hormome (ACfH) (Grumbach et aI, 1978; Parker and Odell, 1980), or does adrenal androgen secretion depend on evolution of androgen-secreting cells through a phase of increased sensitivity to stimulation by ACfH (Anderson, 1980)? Several situations exist in adult patients where plasma androgen and glucocorticoid levels are dissociated, e.g. secondary adrenal failure (Cutler et al, 1979), diabetes mellitus (Cohen et al, 1984) and during replacement treatment with small doses of glucocorticoid (McKenna et al, 1980). A particularly informative area of study is the hormonal pattern in patients with congenital adrenal hyperplasia of the 21-hydroxylase deficiency type. Integrated 24 hour plasma concentrations were obtained by pooling hourly samples from well controlled, regularly ovulating women with this disorder while on treatment with dexamethasone, 0.5 mg at night. The profiles obtained demonstrated that when ACTH levels were similar to those in normal women, DHAS, DHA and androstenedione levels were significantly suppressed below values seen in normal women, while testosterone levels were similar to those in normal

999

THE ADRENAL CORTEX AND VIRILIZATION

women and 17-hydroxyprogesterone values were markedly elevated (Figure 2) (McKenna et aI, 1980). It is not surpnsmg that 17hydroxyprogesterone levels should accumulate when production is normal under the influence of normal ACfH levels since 21-hydroxylase, the enzyme which enhances the conversion of 17-hydroxyprogesterone to cortisol, is deficient in these patients. Therefore, although ACTH levels were normal, adrenal androgen blood levels were suppressed while the levels of the cortisol and androgen substrate, 17-hydroxyprogesterone, were significantly elevated. This is a particularly striking finding as it occurs in patients who, in the untreated state, are characterized by adrenal androgen excess. These findings suggest that a dexamethasone-suppressible factor other than ACTH is involved in the stimulation of adrenal androgens. Furthermore, when the relationship between 17-hydroxyprogesterone and androgen is examined in children and adults with 21-hydroxylase deficiency, it is clear that for any level of precursor, more androgen is produced in adults than in children across the whole spectrum of control, from excessive suppression to untreated (Figure 3). However,

ACTH ng/ml NORMAL CAH

17-0H-PROGESTERONE nmol/l TESTOSTERONE nmol/l NORMAL CAH NORMAL CAH

50

50

25

0

0

0 LNS.....J

Lp
L

DHA-S IJrnolll NORf1AL CAH

DHA nmolll CAH NORMAL

ANDROSTENED lONE nmo III NORMAL CAH 10

20

20

5

10

10

0

0

0

Lp
NS --'

Lp
Lp
Figure 2. Integrated 24 hour plasma hormonal concentrations in normal women and in women with congenital adrenal hyperplasia (CAG) on treatment with dexamethasone, 0.5 mg each night, obtained during the follicular phase of the menstrual cycle in all subjects. In the dexamethasone-treated patients with 21-hydroxylase deficiency, while ACTH levels were normal, 17-0H progesterone was elevated but androstenedione, DBA and DHAS levels were suppressed.

1000

T. J. McKENNA ET AL

when the relationship between various androgens (i.e. DHAS and testosterone) is examined in children and adults, a similar relationship pertains. It appears therefore, that the as yet unidentified factor governing adrenal androgen secretion activates enzymes necessary for the conversion of C 2 1 to C l 9 steroids, i.e. 17,20-desmolase (Figure 1). Parker and Odell (1980) have suggested that they have identified such a factor and have designated it cortical androgen stimulation hormone (CASH). Genazzani et al (1983) have raised the possibility that B-lipotrophin (LPH) or B-endorphin may be responsible for adrenal androgen secretion because of observations of coincidental changes in blood levels between these proopiomelanocortin fragments and adrenal androgen levels throughout life. This concept suggests that the metabolism of proopiomelanocortin, which also yields ACTH within the pituitary, is different at different phases throughout life and is altered by exposure to small doses of dexamethasone. How these intracellular alterations in polypeptide metabolism might be controlled has not been addressed. In addition, the findings reported in Figure 2 suggest that androgens are not produced in the same cell as glucocorticoids and that 17-hydroxyprogesterone accumulating. within, or 'leaking' from 21-hydroxylase deficient cells, does not gain ready access to androgen biosynthetic pathways within or outside the affected cells. Otherwise elevated androstenedione and testosterone values would occur as a consequence of the excessive availability of the precursor, 17hydroxyprogesterone (Figure 3). 25 o PlASIIA 15 DHEA-S llG/,..-L 10

176~P::

HG/I'&..

100

so 50

.:

2511

o

20

100 J5lI 2002511

PLASI'A 17-0H-PROGESTERONE ..I-e

o

25

~

.'.i.!'

..

~

o

t.

20 o

PLAS~A 15

DHEA-S

V'/~I: o.~ o.~;".·;;r··,

D.5 1.0

1.5 2.0 2.5'4':55:'5

TESTOSTERONE ..I"L

0.5

, ,

'{~

1.0 1.5 2.0 2.5

~.5

5.5

TESTOSTERONE .,I-e - - - CHILDREN _

ADuLTS

Figure 3. Correlations of morning plasma steroid levels in adults (0) and children (.) with congenital adrenal hyperplasia of the 21-hydroxylase type. Left panel, DHAS and 17-hydroxyprogesterone: adults, r = 0.61; children, r = 0.56. Middle panel, 17-hydroxyprogesterone and testosterone: adults, r = 0.87; children, r = 0.76. Right panel, DHAS and testosterone: adults, r = 0.87; r = 0.49. From McKcnna ct al (1980), with kind permission of the authors and the publishers, Raven Press.

The physiological roles of adrenal androgens The role of adrenal androgens has not been clearly defined. In the fetus the adrenals are relatively large and steroid production mainly involves the D,5 precursor steroids (Tulchinsky, 1980). DHA is produced in large amounts and following 16-hydroxylation in the fetal liver it is converted to oestriol in the placenta. Urinary and plasma oestriollcvels may be used as indices of feto-placental well-being. However, the precise part played by oestriol

THE ADRENAL CORTEX AND VIRILIZATION

1001

in the gestational process has not been defined. Adrenal androgen secretion increases during the early part of the second decade of life and is responsible for the development of the sexual hair growth in the pubic and axillary regions. These physiological developments are more pronounced in males in whom additional androgens are derived from the testes. Adrenal androgens probably also contribute significantly to the emergence of the libido in women. Although adrenal androgen secretion decreases in women from the fifth or sixth decade of life onwards, the decline is much more gradual than that of the ovary at menopause (Vermeulen, 1983). In the post-menopausal years, circulating adrenal androgens are a more important source of oestrogens than are the ovaries. An important role for these oestrogens in maintaining skeletal integrity has been suggested by Marshall et al (1978). Measurement of androgens Today the capability exists to measure both the urinary excretion of androgens and metabolites and to assess the concentrations of androgens in blood (Grant and Beastall, 1983). Traditionally, adrenal androgen secretion has been assessed by measuring urinary 17-ketosteroid excretion. 17-ketosteroids mainly comprise DHA, DHAS, androstenedione and breakdown products of testosterone but not testosterone itself. More precise measurements of most of these steroids in urine and blood are now available. The development of sensitive, specific radioimmunoassays makes the measurements of the low concentrations of these steroids in blood samples feasible. Testosterone is a potent androgen which circulates mainly bound to sex hormone binding globulin (SHBG). It is probably only the unbound fraction, approximately 2%, which is biologically significant. Androgens tend to lower SHBG levels. It is not infrequent that normal total testosterone levels are associated with a suppressed SHBG value, yielding an elevated free testosterone level. Methods of measuring free testosterone are tedious and cumbersome, while SHBG measurement is considerably easier. Expression of testosterone as a function of SHBG provides an index of free testosterone levels (Cunningham et aI, 1983a). However, there is still a role for the measurement of urinary 17-ketosteroid excretion, particularly in pathological situations where abnormal patterns of metabolism may result in the production of an androgen which will not be recognized by the usual battery of radioimmunoassays available. However, the relatively non-specific urinary 17-ketosteroid assay will detect excessive excretion; this is mainly of importance for patients with an adrenal carcinoma. STATES OF ADRENAL ANDROGEN EXCESS AND VIRILIZATION In this section we will discuss the role of adrenal androgens in the pathogenesis of idiopathic hirsutism, polycystic ovary syndrome (peaS), Cushing's disease and virilizing adrenal tumours, benign and malignant. Congenital adrenal hyperplasia will be discussed elsewhere in this volume.

1002

T. J. McKENNA ET AL

Idiopathic hirsutism and

peos

It is important that we clearly define what the term 'idiopathic hirsutism' is used to convey in this chapter. This disorder is characterized by the growth of coarse, typically male type hair, in the distribution of normal male secondary sex hair (i.e. in the area of the beard, on the chest wall and abdomen, over the thighs and perineum) while regular, ovulatory menstrual cycles are maintained. In contrast, patients with peas may be similarly affected by hirsutism but in addition demonstrate all or some of the following features: oligomenorrhoea or amenorrhoea, obesity, enlarged ovaries demonstrating thickened capsules, atretic follicles and theca cell hyperplasia, and rarely severe manifestations of virilization, e.g. thinning of scalp hair and temporal recession typical of the normal male hair line, deepening of the voice, increased muscle mass and c1itorimegaly (Yen, 1980; Goldzieher, 1981). Acne vulgaris may be associated with either idiopathic hirsutism or polycystic ovary syndrome. Idiopathic hirsutism Patients with idiopathic hirsutism demonstrate elevations of testosterone, while the SHBG level is suppressed and free testosterone levels are elevated (Table 1). In addition, Table 1 shows that plasma andros-

Table 1. Hormonal profiles (mean ± s.d.) in idiopathic hirsutism, PCOS and obesity.

Testosterone nmol/l :j:SlIBG nrnol/l Testosterone/SllBG Androstenedione (~~) nmolll Oestradiol pmol/l Oestrone (E,) pmol/l EI/~~

Basal LlI mlU/ml LHRH Stimulated LlI mlU/ml Basal FSH mlU/ml LHRH Stimulated FSH mlU/ml Basal LHIFSH

Normal women

Idiopathic hirsutism

PCOS

Obesity

1.2 ± 0.34

tl.6 ± 0.55

tl.9 ± 0.6

1.15 ± 0.7

42 ± 13.5 3.09 ± 1.25

t31.9 ± 11.8 t5.55 ± 2.9

t26.9 ± 12.8 t9.05 ± 6.75

t20.8 ± 12.1 t6.4 ± 4.2

6.0 ± 1.73

t7.68 ± 2.58

t9.76 ± 3.31

5.69 ± 3.02

200.9 ± 135

171 ± 85

182.5 ± 117

128.1 ± 78.6

178.1 ± 70.8 33.7 ± 15.7

175 ± 87 25.7 ± 15.2

t293.2 ± 136 31.0 ± 12.7

'250.6 ± 123.4 '49.7 ± 24.2

4.6 ± 2.5

4.4 ± 1.5

12.3 ± 11.3

30.0 ± 55

5.3 ± 3.1

5.4 ± 2.0

'50.9 ± 39.1

'36.6 ± 31.5

4.1 ± 1.5

3.7 ± 1.1

'2.6 ± 1.3

3.4 ± 1.3

3.0 ± 1.8 1.1 ± 0.5

2.6 ± 1.8 1.3 ± 0.7

4.5 ± 4.9 '2.7 ± 1.3

3.0 ± 1.9 '1.7 ± 0.7

'p<0.05-0.005. tp
1003

TIlE ADRENAL CORTEX AND VIRILIZATION

tenedione is elevated in this disorder. 17a-hydroxyprogesterone (Moore et al, 1983), 17a-hydroxypregnenolone and pregnenolone levels are also raised in patients with idiopathic hirsutism (McKenna et al, 1977). DHA and DHAS may be elevated in these patients (Abraham et aI, 1973; Pugeat et al, 1982). From a review of Figure 1 outlining adrenal steroid biosynthesis it can be appreciated that the steroid pattern occurring in idiopathic hirsutism cannot be accounted for by a simple enzyme deficiency. Rather, it appears that the entire androgen biosynthetic pathway has been stimulated, although not in a highly efficient manner, since there is considerable leakage of precursor into the circulation. The source of this steroid excess in idiopathic hirsutism has been a matter of considerable controversy for many years (McKenna, 1978; Biffignandi et aI, 1984). Using various procedures such as sequential adrenal and ovarian vein sampling, adrenal or ovarian suppression and stimulation studies (Greenblatt and Mahesh, 1983), a wide range of conclusions have been reached variously suggesting that the ovary is almost exclusively involved (Kirschner et al, 1976), that the condition is almost equally of adrenal or ovarian origin (Stahl et aI, 1973; Northrop et al, 1975), or that the disorder can be exclusively explained by adrenal androgen excess (Oake et al, 1974). We have demonstrated increased androgen responsiveness to pharmacological adrenal stimulation using exogenous ACTH (Figure 4). More importantly we have also used metyrapone, an inhibitor of l lji-hydroxylasc, to induce hypocortisolism in order to stimulate pituitary secretion of ACTH and related peptides, derivatives of proopiomelanocortin. Since metyrapone does not perturb adrenal androgen production, this provides a probe which allows examination of adrenal androgen secretion in response to endogenous stimulation (Figure 5). The testosterone response to metyrapone in patients with idiopathic hirsutism was greater than that seen in normal women when both were examined during the early part of the follicular phase of the menstrual cycle (Figure 4) (Moore et al, 1983). This finding indicates an abnormality of adrenal androgens in patients with idiopathic hirsutism. In response to metyrapone the hypothalamic-pituitary unit secretes a family

TESTOSTEROI,E RfSPO:1SE TO ~ETYRAPO:i' 2.0

1.0

~JR~;'l

HIRSUTE

lo':W[N

\iJ~'I}~

ANPROSTEN'PIONE RFSPONSE TO ACTH 5.0

2.5

N;JRl";'l

I1IRSUTE

W:;~.[N

W8~.[N

DHEA RESPONSE TO ACTH HIRSuTE 50

l.'Gto'CN

I 25

Figure 4. The incremental androstenedione and DHA responses to stimulation with exogenous ACTH. are significantly higher in patients with idiopathic hirsutism than in normal subjects when both are studied during the follicular phase of the menstrual cycle. In addition. the incremental testosterone response following the administration of metyrapone is also significantly greater in patients with idiopathic hirsutism.

1004

T. J. McKENNA ET AL

ACTH

ACTH

ll-DEOXYCORTISOL

I

CORTISOL

II-DEO~IlRJISOL

r1ETYRAPONd~' ,

,

jJCORTISOL

ANDROGENS

Figure 5. Schematic representation of the mechanism of action of metyrapone whereby cortisol production is impaired. The resulting hypocortisolism brings about increased ACfH secretion and that of related pcptides with resulting increased endogenous stimulation of adrenal androgens, the biosynthesis of which is not perturbed by metyrapone.

of related peptides, proopiomelanocortin fragments, one or more of which may be specifically involved with control of adrenal androgen secretion. The observations made using metyrapone demonstrate an abnormality in the hypothalamic-pituitary-adrenal axis in idiopathic hirsutism and substantiate the significance of the abnormalities noted in the exogenous ACTH studies. In response to treatment with a small dose of dexamethasone, 0.5 mg given each night, a dose known to be associated with normal ACTH secretion in patients with congenital adrenal hyperplasia (McKenna et ai, 1980), there was a lowering of basal androgen levels and subjective and objective improvements in idiopathic hirsutism in excess of 50% of women treated for at least three months (Moore et ai, 1983). While these findings do not establish a unique role for the adrenal in the pathogenesis of idiopathic hirsutism in all patients they support the presence of adrenal androgen excess in perhaps the majority of women with this disorder. In addition, they indicate that modest adrenal suppression (Cunningham et ai, 1983b) may be associated with a significant clinical improvement. When acne coexists with hirsutism, clearing of these skin lesions occurs in approximately 75% of patients. In idiopathic hirsutism, oestrogen, oestradiol and oestrone levels are normal and basal luteinizing hormone (LH) and follicle-stimulating hormone (FSH) values do not deviate from normal and the LH and FSH responses to luteinizing hormone-releasing hormone (LHRH) are not significantly different from those seen in normal subjects when both were examined during the follicular phase of the menstrual cycle (Table 1). These findings differ considerably with those seen in pcas.

Treatment See sections on treatment of hirsutism in polycystic ovary syndrome. Polycystic ovary syndrome The hormonal profiles in patients with idiopathic hirsutism and pcas have been contrasted in Table 1. In the latter disorder there is evidence of

THE ADRENAL CORTEX AND VIRILIZATION

1005

abnormalities of the hypothalamic-pituitary unit (suppressed basal FSH and elevated LH responsiveness to LHRH) and of the ovaries (menstrual disturbances). From the definition of pcas it will be recalled that there are anatomical and microstructural ovarian abnormalities which are characteristic of the disorder. However, whether these abnormalities are primary or secondary events has yet to be established. It is possible that a primary abnormality in the hypothalamic-pituitary unit could result in the typical pattern of gonadotrophin secretion seen in pcas, with ovulation occurring only infrequently. In the normal ovary, androgens are produced in theca cells in response to stimulation with LH. The androgens are then transferred to granulosa cells which, when the aromatase system is stimulated by FSH, convert the androgens to oestrogens - the two cell mechanism of ovary oestrogen biosynthesis (Ryan and Smith, 1965). Tonic LH stimulation of theca cells would result in increased androgen production which, in the absence of normal follicular stimulation because of suppressed FSH activity, would have only limited conversion to oestrogens; ovarian androgen secretion would then increase and follicular development fail to occur (Erickson et ai, 1979). Alternatively, the primary abnormality might reside within the ovary, where theca cell hyperplasia could respond excessively to normal stimulation by LH, resulting in excessive androgen secretion which may undergo conversion to oestrogen in arornatose-containing peripheral tissues, principally adipocytes and the liver. The oestrogen excess could then feedback on the hypothalamic-pituitary unit to bring about abnormal LH and FSH secretion. Use of an oestrogen antagonist, clomiphene citrate, frequently results in FSH secretion, follicular development and ovulation (Yen et ai, 1970; Kistner, 1973). This observation supports a role for abnormal oestrogen feedback in generating the gonadotrophin abnormalities in pcas. From Table 1, it will be seen that while plasma oestrone levels are elevated in this disorder, oestradiol levels are normal. This finding has been previously noted in the study of pcas (Baird, 1979). A pivotal role for oestrone in the genesis of peas is suggested by the following observations: (a) the menstrual pattern is normal in patients with raised androgen levels but normal oestrone and gonadotrophin levels (Table 1); (b) menstrual disturbances are present in patients with elevated androgen levels, normal oestradiol but elevated oestrone levels (Table 1); (c) ovulation can be induced by the oestrogen antagonist clomiphene citrate when oestrone levels are elevated but oestradiol levels are normal; and (d) the typical abnormal gonadotrophin pattern in patients with peas can be exaggerated by the administration of oestrone (Chang et al, 1982). In normal women, while approximately 80% of oestradiol is derived from the ovary only 40% of oestrone arises from this source. Approximately 60% of oestrone arises from peripheral conversion of androstenedione, mainly of adrenal origin (Siiteri and MacDonald, 1973). An indirect index of aromatase activity may be provided by expressing oestrone as a function of androstenedione (Loughlin et al, 1985). Elevated oestrone levels could arise therefore by excessive excretion of oestrone by the ovary, excessive androstenedione secretion by the adrenal, excessive conversion of normal

1006

T. J. McKENNA ET AL

amounts of androstenedione to oestrone in the presence of increased aromatase activity or because of impaired clearance of oestrone. The studies of Siiteri and MacDonald (1973) exclude the possibility of excess ovarian oestrone production. Oestrone and oestradiol have related metabolic pathways. It is unlikely, therefore, that delayed clearance of oestrogens could account for the dissociation of oestrone and oestradiol levels seen in PCOS. It would appear, therefore, that to gain an understanding of the generation of oestrone excess in PCOS we should concentrate on possible increased availability of its precursor, androstenedione, and on aromatase activity. A suppressed oestrone to androstenedione (E 1 to A-t) ratio is consistent with reduced aromatase activity, while an increase in the E 1 to A-t ratio is consistent with increased aromatase activity. In patients with PCOS, plasma testosterone, androstenedione and oestrone levels are elevated with a normal E 1 to A-t ratio (Table 1). This suggests that elevated oestrone levels occur as a consequence of a normal rate of conversion of the excessive amount of androstenedione available, to yield elevated oestrone levels. A question then arises as to the origin of the excess androgens in PCOS - adrenal or ovarian. In a manner similar to that outlined earlier in this chapter we have used metyrapone in an attempt to identify an abnormality in adrenal androgen secretion in PCOS. In response to administration of a single dose of metyrapone at midnight, the incremental testosterone, and androstenedione responses observed in blood drawn 8-9 hours later, were significantly higher than those seen in normal women during the follicular phase of the menstrual cycle (Loughlin et al, 1985). Givens et al (1975) and Lachelin et al (1979) have reported excessive adrenal androgen responsiveness to exogenous ACTH in PCOS. Following three months of treatment with dexamethasone, 0.5 mg given at midnight, basal testosterone and androstenedione levels and the testosterone and androstenedione response to metyrapone returned to normal and 10 of 15 patients established regular menstrual bleeding (McKenna and Loughlin, 1984). These observations indicate that an adrenal abnormality exists in peos, the correction of which is associated with secondary readjustments occurring in the hypothalamic-pituitary-ovarian axis with a return of cyclicity. These findings favour the thesis that an adrenal abnormality is the primary event in the pathogenesis of PCOS (McKenna et ai, 1985). Had pituitary or ovarian abnormalities been primary we could not explain (a) the abnormal testosterone and androstenedione responses to endogenous adrenal stimulation, and (b) the success of a replacement dose of dexamethasone in correcting the abnormalities. Glucocorticoid excess which might suppress ovarian or pituitary activity, causes amenorrhoea (Boccuzzi et ai, 1975; Sakakura et ai, 1975) which contrasts with the return of menstrual cycles seen in our patients. In addition, the dose of dexamethasone used is known to facilitate regular ovulation, while maintaining normal ACTH and androgen levels, in congenital adrenal hyperplasia (Figure 2) (McKenna et al, 1980). 21-Hydroxylase deficient patients provide a useful model as the underlying abnormalities of adrenal androgen excess are well defined.

THE ADRENAL CORTEX AND VIRILIZATION

1007

Suggestions have been made that the oestrogen abnormalities might bring about perturbed adrenal steroid synthesis at the 3~-01-dehydrogenase, ~ .l.s-isomerase step resulting in adrenal DHAS excess (Figure 1) (Lobo et aI, 1982). However, there is no evidence that oestrone exercises such an influence and the increased levels in virtually all of the circulating androgen precursors cannot be explained by a single enzymatic defect (Table 1; Figure 1). As a result of these observations we suggest that adrenal androgen excess plays an important role in the genesis of pcas in many affected patients (McKenna et al, 1985). That an adrenal abnormality, similar to that in idiopathic hirsutism, is probably present in pcas is suggested by the similar androgen responses to metyrapone. It is worth noting that the androstenedione responsiveness is only excessive in pcas in which oestrone levels are elevated; androstenedione is the immediate precursor to oestrone. In pcas, oestrone levels probably rise into a critical range giving rise to gonadotrophin and ovarian abnormalities. Androstenedione levels are significantly higher in patients with pcas than in idiopathic hirsutism (Table 1). This indicates additional available androstenedione for conversion to oestrone. Therefore, when androstenedione excess is severe, oestrone levels may rise with the consequent development of pcas. This concept is compatible with the observations that pcas develops in patients with unequivocal severe androgen excess, i.e. androgen-secreting adrenal tumours (Kase et aI, 1963) and congenital adrenal hyperplasia (Lucis et aI, 1966; Rosenfield, 1985). In addition, in vivo studies reveal that androgen-treated animals develop a PCaS-like state (Mahesh, 1983). However, since androstenedione levels are elevated in idiopathic hirsutism but oestrone levels are normal, it is possible that decreased aromatase activity may protect these patients from the development of pcas (Table 1) (McKenna et aI, 1984b).

Obesity If the role of oestrone excess is central to the gonadotrophin and ovarian abnormalities in pcas, then elevation of the oestrogen in the absence of androgen excess may be associated with similar gonadotrophin and ovarian abnormalities. To examine this possibility we have studied eight obese amenorrhoeic patients, two of whom had mild hirsutism (Loughlin et aI, 1985). It will be recalled that aromatase is found in adipose tissue. While androstenedione levels were normal, oestrone levels were elevated, the LH to FSH ratio was raised and the LH response to LHRH in the obese patients was excessive (Table 1). Therefore, in the absence of elevated androstenedione levels, elevated oestrone levels were associated with gonadotrophin abnormalities similar to those seen in patients with typical pcas. In the obese patients, the E 1 to A .l ratio was raised, suggesting that the raised oestrone levels occurred as a consequence of the excessive conversion of normal amounts of androstenedione to oestrone. Therefore, similarly abnormal gonadotrophin secretion may occur in association with a high oestrone level, whether this occurs as a consequence of the provision of excessive substrate, or due to increased aromatase activity (Loughlin et

1008

T. J. McKENNA ET AL

al, 1985). These findings add further support to the concept that increased oestrone levels are primary to the gonadotrophin and ovarian abnormalities seen in patients with pcas. Increased aromatase action is probably related to the expanded pool of adipose tissue in obese patients. If this is so, reduction in body weight with consequent reduction in the conversion of androgens to oestrogens should bring about a correction of the abnormalities in the hypothalamicpituitary-ovarian axis and the resumption of regular menses. This sequence has been observed by Kopelman et al (1981). Although androstenedione levels are normal and the total testosterone level is also normal, SHBG levels are suppressed in patients with obesity and the testosterone/SHBG ratio is elevated (Cunningham et al, 1985 a,b). However, the degree of elevation is similar to that seen in patients with idiopathic hirsutism (Table 1). It is unlikely therefore that the mild testosterone excess seen in obese subjects can be implicated in the development of amenorrhoea. Suppression of SHBG levels in obese subjects has been previously noted but the mechanism underlying this phenomenon has not been fully explained. Correlation of elevated androgens and oestrogens with hirsutism and menstrual disturbances Idiopathic hirsutism is associated with high androgen levels but normal oestrogen and gonadotrophin levels. Typical PCOS is associated with hirsutism, menstrual disturbances, high androgens and high oestrone levels and abnormal gonadotrophin levels. Obese subjects have near normal androgens, elevated oestrone levels and abnormal gonadotrophin secretion. Therefore, the presence of hirsutism correlates closely with androgen excess, while amenorrhoea is associated with elevated oestrone levels. An exception to this general rule occurs with a form of presentation we have termed 'amenorrhoea with cryptic hyperandrogenaemia', in which patients who present with amenorrhoea but without hirsutism are found to have elevated androgen and oestrone levels (McKenna et al, 1983, 1984a,b). These patients may have normal levels of SHBG; SHBG is usually suppressed in hyperandrogenaemic states. The absence of hirsutism in these patients is probably related to relative insensitivity to androgens, at least involving the hair follicle and the process involved in the regulation of SHBG levels. Correction of hyperandrogenaemia is associated with the resumption of regular ovulation. Interrelation of adrenal androgens, idiopathic hirsutism, PCOS and obesity The majority of patients with idiopathic hirsutism or PCOS secrete excessive adrenal androgens. The absence of abnormal gonadotrophin secretion and resulting polycystic ovaries in idiopathic hirsutism is best explained by failure of these patients to develop elevated oestrone levels. Indeed, patients with idiopathic hirsutism convert androgens to oestrogens

THE ADRENAL CORTEX AND VIRILIZATION

1009

less readily than the normal subjects. This may be because patients with idiopathic hirsutism are of lower than average body weight or have a lower percentage of their body weight contributed to by adipose tissue. We suggest, therefore, that if patients with idiopathic hirsutism gain significant weight they may increase the conversion of androstenedione to oestrone and subsequently develop pcas. Under these circumstances pcas should respond either to suppression of adrenal androgen secretion, or to weight loss or to treatment with the anti-oestrogen clomiphene citrate. These deductions are consistent with our experience and that of others gained during the treatment of pcas. Therefore, body weight may play the major role in determining whether a patient with mild-moderate adrenal androgen excess demonstrates idiopathic hirsutism or pcas. Patients with normal adrenal androgen secretion but with marked obesity, may also develop pcas by excessive conversion of normal concentrations of adrenal androgen to elevated oestrone values. As a consequence, LH secretion is stimulated while that of FSH is suppressed. This brings about excessive stimulation of ovarian androgen production, the conversion of which to oestrogen is not favoured by failure to stimulate follicular development. Under these circumstances the availability of normal' amounts of adrenal androgen is essential to the initiation of the abnormalities leading to the development of pcas and ovarian hyperandrogenaemia in obese subjects. In resume, therefore, while hyperandrogenaemia in patients with idiopathic hirsutism may originate entirely in the adrenal, in patients with pcas developing as a consequence of adrenal androgen excess the eventual hyperandrogenaemia will have two significant sources, adrenal and ovarian. In contrast, when pcas develops as a result of obesity, the hyperandrogenaemia is solely of ovarian origin and adrenal suppression has no role to play in its treatment. In some patients, perhaps the majority with pcas, primary adrenal androgen excess and unrelated obesity may coexist to bring about the development of pcas which would have failed to occur if the same degree of androgen excess or the same extent of obesity had been present in isolation. These observations explain the frequent association of obesity and pcas. Treatment

We have previously proposed classification of pcas into primary and secondary forms (McKenna et aI, 1985). In the primary disorder there is no identifiable contributory abnormality, e.g. congenital adrenal hyperplasia, androgen-secreting adrenal tumours, or obesity; in secondary pcas such a predisposing condition coexists. The treatment of pcas depends both on its aetiology and on the precise goal of treatment. For secondary forms of pcas, treatment should be directed towards correction of the underlying abnormality, e.g. resection of adrenal tumour, weight loss for obesity. Some patients are anxious to conceive, the main concern of others is the possible significance of the abnormal menstrual pattern without any wish to become pregnant, while other patients require guidance on how to manage hirsutism.

1010

T. J. McKENNA ET AL

Amenorrhoea and infertility Patients wishing to become pregnant may be treated with clomiphene citrate, which is probably best undertaken in close collaboration with a gynaecologist (Yen et aI, 1970; Kistner, 1973). Alternatively, the introduction of treatment with a glucocorticoid, e.g. prednisone 5-10 mg per day or dexamethasone, 0.5 mg each night, is particularly effective in non-obese subjects with pcas, in whom a return of regular menstruation can be anticipated in excess of 50% of patients within three months of the initiation of treatment (Steinberger et aI, 1981; Loughlin et al, 1985). Adrenal suppression and clomiphene citrate may be used in combination with benefit (Chang and Abraham, 1976). Failure of patients to respond to adrenal suppression may occur in a patient with a primary adrenal disorder when pcas becomes self-sustaining. While adrenal androgen excess may initially be required to provide elevated oestrogen levels which perturb gonadotrophin secretion, the resulting theca cell hyperplasia and increased ovarian androgen production will then contribute to the raised oestrone level, thereby amplifying the abnormal gonadotrophin secretion resulting in further theca cell hyperplasia. Eventually the population of theca cells may provide androgen excess so that hyperoestronaemia may persist even after adrenal androgen secretion has been suppressed and that fraction of oestrone derived from adrenal androgens eliminated. In this manner, while adrenal androgen excess, or alternatively excessive androstenedione conversion to oestrone in obese patients, may be necessary to initiate the development of pcas, neither of these events may be required to sustain the disorder. For unusual patients with pcas who prove to be resistant to glucocorticoids and clomiphene citrate, treatment with gonadotrophins may be useful (Raj et al, 1977). Bilateral wedge resection of the ovaries is seldom utilized today, as less disruptive and more definitive forms of treatment are usually successful. The experimental use of bromocriptine in the treatment of pcas has been recently advocated (Spruce et aI, 1984). Repeated ultrasonographic examination of the ovary provides an excellent method of monitoring follicular development in response to therapeutic manoeuvres (Gordon et al, 1984).

Hirsutism Approximately 50% of patients will notice a reduction in the number of hairs and in the rate of hair growth, three to six months after the introduction of dexamethasone, 0.5 mg daily. A response to treatment can be predicted by a reduction in plasma levels of testosterone, the testosterone/SHBG ratio and androstenedione levels. The only significant side-effects we have encountered with this treatment is weight gain in patients who were obese at the onset of therapy: we have not noted significant weight gain in non-obese subjects treated in this way. The hypothalamic-pituitary glucocorticoid axis is only mildly suppressed if suppressed at all, by even long-term treatment (Cunningham et al, 1983b).

TIlE ADRENAL CORTEX AND VIRILIZATION

1011

Dexamethasone has the advantage over most other hormonal treatments for hirsutism, that regular ovulation is preserved or re-established by this treatment, which cannot be sustained during ovarian suppression. However, when dexamethasone does not prove to be successful, or in obese subjects, use of a combination of oestrogen-progestagen anovulant will bring about suppression of ovarian androgen secretion, may allow regression of theca cells and may have pharmacological effects which bring about suppression of adrenal steroid production (Givens, 1983). Since some progestagens have intrinsic androgenic properties it is essential that a non-androgenic progestagen be utilized. Another approach is to utilize anti-androgen therapy. Cyproterone acetate is an anti-androgenic progestagen, and it is used in a 'reversed sequential manner'; cyproterone acetate, 50-100 mg is given daily for 10 days from the fifth day of menstrual bleeding and this is accompanied by oestrogen, which is continued for an additional 11 days, after which the cycle is interrupted. This form of treatment is utilized to obtain regular withdrawal bleeding. Cyproterone acetate has a prolonged half-life and were it continued with the oestrogen during the second half of the therapeutic course, shedding of the endometrium might not occur for several days after stopping treatment. This treatment has been widely used and achieves good results in 75% of patients treated for 6-12 months (Hammerstein, 1980; Biffignandi et aI, 1984). If cyproterone acetate is used in the absence of oestrogen the possibility of ovulation exists and if pregnancy occurs, continued use of the antiandrogen will result in feminization of any male conceptus. Experimental anti-androgens used in the treatment of hirsutism with variable success include spironolactone (Molinatti et ai, 1983) and cimetidine (Vigersky et aI, 1980).

CUSHING'S SYNDROME Cushing's disease (pituitary ACTH excess) In this disorder hyperstimulation of the adrenal by ACTH brings about both cortisol and androgen excess (Rabin and McKenna, 1982). The androgen excess is probably responsible for some of the typical features of Cushing's syndrome, such as hirsutism, acne, thinning of scalp hair and, infrequently, c1itorimegaly. The plethoric facies of Cushing's syndrome may be in part contributed to by androgen-induced erythrocytosis. Patients with Cushing's syndrome typically have amenorrhoea. The development of this feature appears to have a complex aetiology. In the foregoing section, the roles of adrenal androgen excess and obesity, both features of Cushing's syndrome, have been extensively outlined in the development of pcas. However, in addition, glucocorticoid excess suppresses gonadotrophin secretion (Sakakura et aI, 1975) and also impairs gonadal responsiveness to stimulation (Saez et aI, 1977).

1012

T. J. McKENNA ET AL

Adrenal tumours Patients with pure cortisol-secreting adrenal adenomas do not demonstrate evidence of virilization. However, they may have increased hair growth of a downy type. This hair may be distinguished from androgen-dependent hair by its type (fine and lying along the skin rather than growing away from it) and distribution which is generalized, including the dorsum, rather than in the typical male pattern. Patients in whom cortisol and androgen excess occur as a consequence of a benign adrenal tumour have features very similar to those described for Cushing's disease (Rabin and McKenna, 1982). Very marked virilization may occur, however, in patients with Cushing's syndrome occurring as a consequence of an adrenal carcinoma. These patients may have markedly elevated plasma levels of testosterone and/or androstenedione and/or dehydroepiandrosterone. In addition to hirsutism patients may have marked scalp balding, increased muscle mass of a male type, lowering of the pitch of their voice and very marked c1itorimegaly. ANDROGEN-SECRETING ADRENAL TUMOURS In this section we will discuss the presentation of adrenal tumours which secrete androgen exclusively. These tumours are brought to clinical attention by evidence of virilization and menstrual disturbances in women. Androgen-secreting tumours are much less frequently recognized in men. In men, these tumours may cause infertility by suppression of gonadotrophin secretion. In women, hirsutism, usually with a relatively short history, i.e. developing rapidly over the preceding weeks to months, may be associated with acne, temporal recession (so that a male pattern hair-line develops), thinning of scalp hair, deepening of the voice, increased muscle mass, increased or decreased libido and clitoral enlargement (Gabrilove et aI, 1981; Talbert and Wing, 1983). The disorder may develop at any time from childhood to old age. Levels of circulating androgens, e.g. testosterone, androstenedione, DHA, DHAS, are increased, and/or urinary ketosteroid excretion rates are elevated in affected patients. Traditionally it was believed that normal urinary 17-ketosteroid excretion rates effectively excluded the presence of an adrenal lesion as a cause of virilization. However, a number of virtually pure testosteronesecreting tumours associated with normal urinary 17-ket.osteroid excretion have been reported (Burr et aI, 1973; Werk et al, 1973; Givens et aI, 1974). In almost all patients with virilizing adrenal tumours, plasma testosterone is markedly elevated. The main diagnostic problem is the distinction between an adrenal tumour andan ovarian-androgen secreting tumour. In some instances a mass may be palpable when an adenocarcinoma is present. The anatomical localization of the tumour is greatly facilitated using ultrasonography (Yeh et aI, 1978a,b), computerized tomography (CT. scanning) (Genant et aI, 1983), arteriography, selective venous sampling of adrenal and ovarian veins (Gabrilove et aI, 1981) and radio-isotope scanning of the adrenals (Gross et aI, 1984). Since

:;j

rn

>c :;>;l

m

Z

> r o o ~

Table 2. Cli nical pre sentation of virilizing disorders. Time o f o nse t to medi cal pre sentat ion

Men stru al distu rbances Viri lization"

Di agnosis

Time o f onse t

Idiopathic hirsu tism

Years Years Ma y p rese nt at birth , in adolesce nce o r, rar ely. in adulthood

+ + +

+ +

Rare

Congenit al adre nal hyperpl asia

15-25 Year s 15-25 Year s Con geni tal

Cu shin g's syndro me

An ytim e

Months- ye ars

+

+

Unusual

pcas

Androgen -secr et ing tum ours: adre nal or ova ria n Surreptitiou s and roge n

ingestion

Hirsu tism

+

Anyt ime

Weeks-months

+

+

+

A dulthoo d

?

+

+

+/-

Comm ent Enl arged o vari es ma y be palpable Small stature, evide nce of intrauterine a ndrog en exce ss, c .g, labial fusion . Usuall y primary amenorrhoea Coexi stin g evide nce o f glucocorticoi d excess Pa lpable mass ma y be p resent

m X

>-Z o

:5 :;>;l t=

N

~

(3 Z

Access to androge ns

+, Pre sent; - , A bsen t. "Virilizatio n is used her e to em brace the occurrence of all o r so me of the followi ng-deep ening of vocal p itch, male type hair-line and scalp thinning, increased musculat ure . clito rimcga ly. '

-o

VJ

1014

T. J. McKENNA ET AL

androgen-secreting tumours are usually large there is a high yield using ultrasound examination and Cf scanning which are non-invasive procedures. In the case of smaller tumours arteriography may identify a lesion not picked up by the non-invasive procedures. Similarly, measurement of the androgen levels known to be elevated in peripheral blood, in blood samples obtained from both adrenal veins and both ovarian veins, allows identification of the gradient in androgen levels which will exist between the venous blood draining the tumour and peripheral blood. Radioiodinated-cholesterol should be given to patients previously treated with dexamethasone to suppress ACfH-dependent adrenal activity. Cholesterol, which is a precursor for androgens, will accumulate in the autonomously functioning tumour. Adrenal carcinomas may fail to accumulate the radio-iodinated-cholesterol. Not all of these investigations will be necessary in order to establish a diagnosis. However, it is probably prudent to have two of the foregoing procedures yielding concordant results. In the very unusual virilized patient where these investigations either fail to demonstrate a tumour or yield inconsistent results, an exploratory laparotomy to examine both adrenal glands and both ovaries is warranted. However, it should be borne in mind that on rare occasions, emotionally disturbed patients may present with features suggestive of an androgensecreting tumour, who are surreptitiously taking androgens. Members of the medical and allied professions who have access to drugs, are particularly suspect in this area. Patients with late clinical manifestations of congenital adrenal hyperplasia may also present some diagnostic difficulties in this area and for this reason measurement of cortisol precursors, particularly 17-hydroxyprogesterone, should be performed. However, 17-hydroxyprogesterone levels and those of other precursor steroids may be markedly elevated in patients with adrenal adenocarcinoma but in contrast to congenital adrenal hyperplasia, they arc resistant to the suppressive effects of glucocorticoids (Table 3). Adrenal and ovarian stimulation and suppression tests using such agents as dexamethasone, oestrogen and progestagen, ACfH and human chorionic gonadotrophin (HCG) yield unreliable results. It is of particular interest that several androgen-secreting adrenal tumours may respond to stimulation with HCG (Gabrilove et aI, 1981). A recently reported patient who had a mildly active androgen-secreting adrenal tumour, continued to ovulate. When the patient became pregnant, virilization advanced rapidly and when the baby was subsequently delivered there was evidence of masculinization (Fuller et aI, 1983). The mother subsequently had a HCGresponsive adrenal androgen-secreting tumour removed. Presumably, the elevated levels of HCG during pregnancy stimulated the adrenal tumour so that androgen levels higher than in those in the non-pregnant state were attained. The treatment of adrenal tumour consists of resection of the tumour where possible and otherwise medical treatment (Rabin and McKenna, 1982). Where metastatic disease is present the adrenolytic agent o,p'-DDD may be useful, as mayan inhibitor of androgen biosynthesis, e.g. aminoglutethimide, which inhibits the conversion of cholesterol to preg-

-l

:::: >-o

m

Table 3. Laboratory evaluation of virilizing disorders.

Diagnosis Idiopathic hirsutism PCOS Congenital adrenal

Urinary 17ketosteroid excretion

Plasma testosterone and/or androstenedione

t t

In approx 50% of patients In approx 75% of patients

t t·

t t

In approx 50% of patients

~

Plasma 17a-OHprogesterone

Cortisol following dexamethasone

LH

FSH

N

Suppressed

N

N

rn

o

~

rn

NIt

Suppressed

t

!

t t·

t t·

Suppressed

NI t

Nt!

t

NIp

Not suppressed

NI t

Nt!

In approx 75% of patients

Miscellaneous

Ultrasonography may show enlarged ovaries with cysts

Androgensecreting tumours: Adrenal Ovarian

t

tt

t t

tIN N

NIp NIt

Variable Suppressed

! !

! !

X

>Z o

s

hypcrplasiat Cushing's syndrome

z >r o

May have evidence of pituitary or adrenal tumour: measure ACTH

:::: r

N ~

sz

Adrenal and ovarian ultrasonography, CT scanning, arteriography, selective venous sampling, adrenal radiocholesterol scanning, will all yield localizing information

t Elevated; ! Suppressed. 'Suppresses on treatment with replacement doses of glucocorticoids taken at night. tI7-0H-Progesterone and other precursors e.g. l l-dcoxycortisol, 17-0H-pregnenolone may be elevated in patients with an adrenal adenocarcinoma. :j:Congenital adrenal hyperplasia represented here is of the 21-hydroxylase deficiency type - sec relevant chapter for characteristics of l l-hydroxylasc and 3B-ol-dehydrogenase, /),,4•. - isomerase deficiencies.

......

o ...... VI

1016

T. J. McKENNA ET AL

nenolone (Figure 1). However, metyrapone does not inhibit androgen biosynthesis and since trilostane inhibits the conversion of dehydroepiandrosterone to androstenedione and testosterone, dehydroepiandrosterone (a weak androgen) will accumulate in large amounts during its use. CLINICAL PRESENTATION AND INVESTIGATION OF VIRILIZING DISORDERS Tables 2 and 3 provide suggested diagnostic guidelines, clinical and laboratory, respectively. This material should be supplemented from other relevant chapters thoughout this text. SUMMARY The physiological control of adrenal androgen secretion has not been definitively established. However, there is evidence to suggest. that a dexamethasone-suppressible factor other than ACTH may have a specific role to play. The majority of patients with idiopathic hirsutism (hirsutism associated with regular menstruation) have findings suggestive of adrenal androgen excess, including enhanced androgen responsiveness following administration of metyrapone, and respond to treatment with dexamethasone, 0.5 mg given each night. Patients with idiopathic hirsutism have elevated androgens but normal oestrogen and gonadotrophin levels. In contrast, while patients with polycystic ovary syndrome (PCaS) also demonstrate evidence of adrenal androgen excess, these patients have elevated oestrone levels and gonadotrophin secretion is abnormal. Approximately 50% of patients with pcas treated with dexamethasone resume regular menstruation. Oestrone excess appears to be primary to the abnormal gonadotrophin secretion and to the development of pcas. In non-obese patients with pcas elevated oestrone appears to occur as a consequence of the availability of the excessive amounts of its immediate' precursor, androstenedione, an androgen mainly of adrenal origin. Androstenedione is converted to oestrone in fat. Obese amenorrhoeic subjects have normal androstenedione values but elevated oestrone levels with abnormal gonadotrophin secretion as seen in pcas. These findings indicate that abnormal gonadotrophin secretion is associated with elevated oestrone levels whether these occur as a consequence of excessive adrenal androgen secretion, or the excessive conversion of normal amounts of available androstenedione. Patients with idiopathic hirsutism and elevated androstenedione levels but normal oestrone values appeared to be protected against the development of pcas by relatively poor conversion of androstenedione to oestrone. It is likely, therefore, that if patients with idiopathic hirsutism gain additional adipose tissue, elevated oestrone levels will result and pcas will develop. These observations explain the frequent association of pcas and obesity. There is a close clinical association between elevated androgen levels and hirsutism and between elevated

THE ADRENAL CORTEX AND VIRILIZATION

1017

oestrone levels and menstrual disturbances. However, some patients with amenorrhoea but without hirsutism may demonstrate marked elevations of androgens and oestrone, the correction of which leads to the resumption of regular ovulation. This presentation, 'amenorrhoea with cryptic hyperandrogenaemia', is probably explained by diminished sensitivity of androgen receptors. While adrenal androgen excess may be associated with hirsutism and menstrual disturbances in patients with Cushing's syndrome, the features of glucocorticoid excess usually predominate. When frank virilization develops, the presence of an androgen-secreting adrenal or ovarian tumour should be suspected. Ultrasonography, CT scanning, adrenal radio-isotope scanning, arteriography and selective sampling from adrenal and ovarian veins for measurement of androgens, will usually identify the source of androgen excess. Normal urinary 17-ketosteroid excretion may be found in association with a pure testosterone-secreting adrenal tumour. The possibility of surreptitious androgen ingestion should be considered prior to undertaking a diagnostic laparotomy when virilized patients fail to disclose adrenal or ovarian pathology despite employing the currently available highly sensitive localizing procedures. REFERENCES Abraham GE, Chakmakjian ZH, Buster JE & Marshall JR (1975) Ovarian and adrenal contributions to peripheral androgens in hirsute women. Obstetrics and Gynecology 46: 169-173. Anderson DC (1980) The adrenal androgen-stimulating hormone does not exist. Lancet ii: 454-456. Baird DT (1979) Polycystic ovary syndrome. In Jacobs HS (ed) Adl'ances in Gynaecological Endocrinology, pp 289-299. London: The Royal College of Obstetricians and Gynaecologists. Biffignandi P, Massucchetti C & Molianatti GM (1984) Femal hirsutism: pathophysiological considerations and therapeutic implications. Endocrine Reviews 4: 498-513. Boccuzzi G, Angeli A, Bisbocci D et al (1975) Effect of synthetic luteinizing hormone releasing hormone on the release of gonadotrophins in Cushing's disease. Journal of Clinical Endocrinology and Metabolism 40: 892-895. Burr 1M, Sullivan J, Graham T, Hartman WH & O'Neill J (1973) A testosterone-secreting tumour of the adrenal producing virilization in a female infant. Lancet ii: 6..B -644. Chang RJ & Abraham GE (1976) Effect of dexamethasone and clomiphene citrate on peripheral steroid levels and ovarian function in a hirsute amenorrhoeic patient. Fertility and Sterility 27: 64(}-646. Chang RJ, Mandel FP, Lu JKH & Judd HL (1982) Enhanced disparity of gonadotrophin secretion by oestrone in women with polycystic ovarian disease. Journal of Clinical Endocrinology and Metabolism 54: 49G-49..t Cohen HN, Paterson KR, Wallace AM et al (1984) Dissociation of adrenarche and gonadarche in diabetes mellitus. Clinical Endocrinology 20: 717-724. Cunningham SK, Loughlin T, Culliton M & McKenna TJ (1983a) Plasma sex hormone binding globulin and androgen levels in the management of hirsute patients. Acta Endocrinologica 104: 365-371. Cunningham SK, Moore A & McKenna TJ (l983b) Normal cortisol response to corticotrophin in patients with secondary adrenal failure. Archives of Internal Medicine 143: 2276-2279. Cunningham SK, Loughlin T, Culliton M & McKenna TJ (1985a) The role of sex steroids and sex hormone binding globulin in hirsutism and/or oligomcnorrhoea in obese and normal weight women. Irish Medical Journal 78: 208-212.

1018

T. J. McKENNA ET AL

Cunningham SK, Loughlin T, Culliton M & McKcnna TJ (1985b) The relationship between sex steroids and sex hormone binding globulin in plasma in physiological and pathological conditions. Annals of Clinical Biochemistry 22: 489-497. Cutler Gn Jr, Davis SE, Johnsonbaugh RE & Loriaux DL (1979) Dissociation of cortisol and adrenal androgen secretion in patients with secondary adrenal insufficiency. Jot/mal of Clinical Endocrinology and Metabolism 49: 60-1-609. Erickson GH, Hsueh AJW, Quigley ME, Rebar RW & Yen SSC (1979) Functional studies of aromatase activity in human granulosa cells from normal and polycystic ovaries. Jot/mal of Clinical Endocrinology and Metabolism 49: 514-519. Fuller PJ, Pettigrew IG, Pike JW & Stockigt JR (1983) An adrenal adenoma causing virilization of a mother and infant. Clinical Endocrinology 18: 143-153. Gabrilove JL, Seman AT, Sabet R, Mitty HA & Nicolis GL (1981) Virilizing adrenal adenoma with studies on the steroid content of the adrenal venous effluent and a review of the literature. Endocrine Reviews 2: 462-470. Genant HK, Turski PA & Moss AA (1983) Advances in CT assessment of metabolic and endocrine disorders. Adl'Onces in Internal Medicine 28: 409-447. Genazzani AR, Facchinetti F, Petraglai F et al (1983) Correlations between plasma levels of opioid pep tides and adrenal androgens in pre-puberty and puberty. Jot/mal of Steroid Biochemistry 19: 891-895. Givens JR (1983) Role of oral contraceptives in the treatment of hyperandrogcnism of hirsute women. In Mahesh VB & Greenblatt RB (eds) Hirsutism and Virilism: Pathogenesis, Diagnosis and Management, pp 351-367. Bristol: John Wright. Givens JR, Andersen RN, Wiser WL, Coleman SA & Fish SA (1974) A gonadotrophinresponsive adrenocortical adenoma. Jot/mal of Clinical Endocrinology and Metabolism 38: 126-133. Givens JR, Andersen RN, Ragland JB, Wiser WL & Umstot ES (1975) Adrenal function in hirsutism I. Diurnal change and response of plasma androstenedione, testosterone, 17-hydroxyprogesterone, cortisol, LH and FSH to dexamethasone and V2 unit of ACTII. Jot/mal of Clinical Endocrinology and Metabolism 40: 988-1000. Goldzieher JW (1981) Polycystic ovary disease. Fertility and Sterility 35: 371-394. Gordon PAL, Ayers An, Lowy C, Sonksen PH & Wheeler MJ (1984) Ultrasound of the polycystic ovary syndrome. Programme and Abstracts of Papers, 3rd Ioint Meeting of British Endocrine Societies, p 107. Grant JK & Beastall GH (1983) Androgens. In Clinical Biochemistry of Steroid Hormones: Methods and Applications, pp 144-171. London: Croom Helm. Greenblatt RB & Mahesh VB (1983) The androgenic polycystic ovary. In Mahesh VB & Greenblatt RB (eds) Hirsutism and Virilism: Pathogenesis, Diagnosis and Management, pp 213-237. Bristol: John Wright. Gross MD, Shapiro B, Thrall JH, Freitas JE & Bcierwaltes WII (1984) The scintigraphic imaging of endocrine organs. Endocrine Reviews 5: 221-281. Grumbach MM, Richards GE, Conte FA & Kaplan SL (1978) Clinical disorders of adrenal function and puberty. An assessment of the role of the adrenal cortex in normal and abnormal puberty in man and evidence for an ACTH-like pituitary adrenal androgen stimulating hormone. In James VHT et al (eds) The Endocrine Function of the Human Adrenal Cortex, pp 583-612. London: Academic Press. Hammerstein J (1980) Possibilities and limits of endocrine therapy. In Hammerstein J et al (eds) Androgenization in Women: Acne, Seborrhoea, Androgcnetic Alopecia and Hirsutism, pp 221-238. Amsterdam: Excerpta Medica. Kase N, Kowal J, Perloff W & Soffer U (1963) In vitro production of androgens by a virilizing adrenal adenoma and associated polycystic ovaries. Acta Endocrinologica 44: 15-19. Kirschner MA, Zucker IR & Jespersen D (1976) Idiopathic hirsutism - an ovarian abnormality. New England Journal of Medicine 294: 637-6-tO. Kistner RW (1973) The use of clomiphene citrate in the treatment of anovulation. Seminars in Drug Treatment 3: 159-175. Kopelman PG, White N, Pilkington TRE & Jeffcoate SL (1981) The effect of weight loss on sex steroid secretion and binding in massively obese women. Clinical Endocrinology 15: 113-116.

TilE ADRENAL CORTEX AND VIRILIZATION

1019

Lachelin GCL, Barnett M, Hopper BR. Brink G & Yen SSC (1979) Adrenal function in normal women and Women with polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism 49: 892-898. Lobo RA, Marsh CM. Goebelsmann U & Michell DR (1982) The modulating role of obesity and 17~-oestradiol (E 2) on bound and unbound E 2 and adrenal androgens in oophorectomized women. Journal of Clinical Endocrinology and Metabolism 54: 320-324. Loughlin T. Cunningham SK, Culliton M et al (1985) Altered androstenedione and oestrone dynamics associated with abnormal hormonal profiles in amenorrhoeic subjects with weight loss or obesity. Fertility and Sterility 43: 720-725. Loughlin T. Cunningham S, Moore A et al (1985) Journal of Clinical Endocrinology and Metabolism (in press). Lucis OJ, Hobkirk R, Hollenberg CH, MacDonald SA & BIahey P (1966) Polycystic ovaries associated with congenital adrenal hyperplasia. Canadian Medical Association Journal 94: 1-7. Mahesh VB (1983) Various concepts of pathogenesis of polycystic ovarian disease. In Mahesh VB & Greenblatt RB (eds) Hirsutism and Virilism: Pathogenesis, Diagnosis and Management, pp 247-276. Bristol: John Wright. Marshall DH, Crilly RG & Nordin BEC (1978) The relationship between plasma androstenedione and oestrone in untreated and corticosteroid treated post-menopausal women. Clinical Endocrinology 9: 407-412. McKenna TJ (1978) The adrenal cortex and menstrual disorders. In Givens JR (ed) Endocrine Causes of Menstrual Disorders, pp 372-407. Chicago: Year Book Medical Publishers. McKenna TJ, Miller RB & Liddle GW (1977) Plasma pregnenolone and 17hydroxypregncnolonc in patients with adrenal tumors, ACTH excess, or idiopathic hirsutism. Journal of Clinical Endocrinology and Metabolism 44: 231-236. McKenna TJ, Moore G, Orth ON et al (1980) The biosynthesis of androgens in 21-hydroxylase deficiency. In Genazzani AR. Thijssen JHH & Siiteri PK (eds) Adrenal Androgens, pp 135-139. New York: Raven Press. McKenna TJ. Moore A. Magee F & Cunningham S (1983) Amenorrhoea with cryptic hyperandrogenaemia. Journal of Clinical Endocrinology and Metabolism 56: 893---896. I McKenna TJ, Loughlin T, Culliton M, Smyth PPA & Cunningham S (1984a) The evolving hormonal profile in amenorrhoeic patients with cryptic hyperandrogenaemia. Irish Medical Journal 77: 344-346. McKenna TJ, Loughlin T, Daly L et al (1984b) Variable clinical and hormonal manifestations of hypcrandrogenacmia. Metabolism 33: 714-717. McKenna TJ, Loughlin T, Smyth PPA & Cunningham S (1985) The pathogenesis and treatment of polycystic ovary syndrome. In Flamigni C, Venturoli S & Givens JR (eds) Adolescence in Females, pp 405-419. Chicago: Year Book Medical Publishers. Molinatti GM, Messina M, Manicri C, Massucchetti C & Biffignandi P (1983) Current approaches to treatment of virilizing syndrome. In Molanatti GM, Martini L & James VHT (eds) Androgenization in Women, p 179. New York: Raven Press. Moore A, Magee F, Cunningham SK, Culliton M & McKenna TJ (1983) Adrenal abnormalities in idiopathic hirsutism. Clinical Endocrinology 18: 391-399. Northrop G, Archie JT. Patel SK & Wilbanks GO (1975) Adrenal and ovarian vein androgen levels and laparoscopic findings in hirsute women. American Journal of Obstetrics and Gynecology 122: 192-198. Oake RJ. Davies SJ, Mcl.achlan MSF & Thomas JP (1974) Plasma testosterone in adrenal and ovarian vein blood of hirsute women. Quarterly Journal of Medicine (New Series) 43: 603---613. O'Hare MJ, Nice EC & Neville AM (1980) Regulation of androgen secretion and sulfoconjugation in the human adult adrenal cortex: studies with primary monolayer cell cultures. In Genazzani AR, Thijssen JHH & Siiteri PA (eds) Adrenal Androgens, pp 7-25. New York: Raven Press. Parker LN & Odell WD (1980) Control of adrenal androgen secretion. Endocrine Reviews I: 392-410. de Peretti E & Forest MG (1976) Unconjugated dehydroepiandrosterone plasma levels in

1020

T. J. McKENNA ET AL

normal subjects from birth to adolescence in human: the usc of a sensitive radioimmunoassay. Journal of Clinical Endocrinology and Metabolism 43: 982-991. Pugeat M, Forest MG, Nisula BC et al (1982) Evidence of excessive androgen secretion by both the ovary and the adrenal in patients with idiopathic hirsutism. Obstetrics and Gynecology 59: 46-51: Rabin D & McKenna TJ (1982) Cushing's syndrome. Clinical Endocrinology and Metabolism; Principles and Practice, pp 434-454. New York: Grune & Stratton. Raj SG, Berger MJ, Grimes EM & Taylor ML (1977) The use of gonadotrophins for the induction of ovulation in women with polycystic ovarian disease. Fertility and Sterility 28: 1280-1284. Rosenfield RL (1985) Congenital adrenal hyperplasia and reproductive function in females. In Flamigni C, Venturoli F & Givens JR (eds) Adolescence in Females, pp 373-387. Chicago: Year Book Medical Publishers. Ryan KJ & Smith OW (1965) Biogenesis of steroid hormones in the human ovary. Recent Progress in Hormone Research 21: 367-402. Saez JM, Morera AM, Haour F & Evain D (1977) The effect of in-vivo administration of dexamethasone, corticotrophin and human chorionic gonadotrophin on steroidogenesis and protein and DNA synthesis of testicular interstitial cells in pre-pubertal rats. Endocrinology 101: 1256. Sakakura M, Takebe K & Nakagawa S (1975) Inhibition of luteinizing hormone secretion induced by synthetic LRII by long-term treatment with glucocorticoids in human subjects. Journal of Clinical Endocrinology and Metabolism 40: 774--779. Siiteri PK & MacDonald PC (1973) Role of extraglandular oestrogen in human endocrinology. In Greep RO & Astwood EB (eds) Handbook of Physiology Section 7: Endocrinology (Vol. 2), pp 615-629. Washington DC: American Physiological Society. Spruce BA, Kendal-Taylor P, Dunlop W et al (1984) The effect of bromocriptine in the polycystic ovary syndrome. Clinical Endocrinology 20: 481-488. Stahl NL, Teeslink CR, Beauchamps G & Greenblatt RB (1973) Serum testosterone in hirsute women: a comparison of adrenal, ovarian and peripheral vein values. Obstetrics and Gynecology 41: 650-654. Steinberger E, Smith KD & Rodriquez-Rigau U (1981) Hyperandrogenism and female infertility. In Crosignani PG & Rubin BL (cds) Endocrinology of Human Infertility: New Aspects, pp 327-342. London: Academic Press. Symington T (1969) Functional Pathology of the Human Adrenal Gland. pp 13-159. Edinburgh: Livingstone. Talbert LM & Wing RL (1983) Virilizing adrenal tumours. In Mahesh VB & Greenblatt RB (eds) Hirsutism and Virilism: Pathogenesis, Diagnosis and Management, pp 121-128. Bristol: John Wright. Tulchinsky D (1980) Adrenal androgens in pregnancy. In Genazzani AR, Thijssen JHH & Siiteri PK (eds) Adrenal Androgens, pp 189-198. New York: Raven Press. Vermeulen A (1983) Androgen secretion by the adrenals and gonads. In Mahesh VB & Greenblatt RB (eds) Hirsutism and Virilism: Pathogenesis, Diagnosis and Management, pp 17-34. Bristol: John Wright. Vigersky RA, Mehlman I, Glass AR & Smith CE (1980) Treatment of hirsute women with cimetidine: A preliminary report. New England Journal of Medicine 303: 1042. Werk EE, Sholiton U & Kalejs L (1973) Testosterone-secreting adrenal adenoma under gonadotrophin control. New England Journal of Medicine 289: 767-770. Yeh HC, Mitty HA, Rose J, Wolf BS & Gabrilove JL (1978a) Ultrasonography of adrenal masses: usual features. Radiology 127: 467-474. Yeh HC, Mitty HA, Rose J, Wolf BS & Gabrilove JL (1978b) Ultrasonography of adrenal masses: unusual manifestations. Radiology 127: 475-483. Yen SSC (1980) The polycystic ovary syndrome. Clinical Endocrinology 12: 177-208. Yen SSC, Vela P & Ryan KJ (1970) Effect of clomiphene citrate in polycystic ovary syndrome: relationship between serum gonadotrophin and corpus luteum function. Journal of Clinical Endocrinology and Metabolism 49: 514--519.