|
|
| |||||||||||||||||||||||||||||||||||||||
Background Hyperandrogenemia is the hallmark of the polycystic ovary syndrome, yet the relative contributions of the adrenal cortex and ovary to the overproduction of androgen remain unclear. To identify possible causes of adrenocortical overactivity, we studied the metabolism of adrenal and ovarian steroid hormones in women with this disorder.
Methods We measured 24-hour urinary excretion of steroid hormone metabolites by high-resolution capillary gas chromatography in 65 women with the polycystic ovary syndrome and 45 normal women matched for body-mass index.
Results After adjustment for body-mass index, the urinary excretion of testosterone and androstenedione metabolites was 1.9 times higher in the women with the syndrome than in the normal women, and the excretion of dehydroepiandrosterone metabolites (C19 steroid sulfates) and cortisol metabolites was 1.5 and 1.3 times higher, respectively (P<0.01 for all comparisons). The affected women also had significantly higher ratios of 11-oxo (oxygenated) metabolites to 11-hydroxy metabolites of cortisol (1.4 times higher, P<0.001) and of 11-oxo to 11-hydroxy metabolites of corticosterone (1.8 times higher, P<0.001). In the group with the polycystic ovary syndrome, 55 percent of the nonobese women and 24 percent of the obese women had ratios above the upper limit of normal; the ratios in the obese women did not differ significantly from those in the nonobese women.
Conclusions Adrenal secretion of cortisol and androgens is increased in women with the polycystic ovary syndrome. The increases may be explained by dysregulation of 11
-hydroxysteroid dehydrogenase causing increased oxidation of cortisol to cortisone, which cannot be accounted for by obesity.
Androgen excess is central to the pathophysiologic changes and clinical expression of PCOS. According to one hypothesis for the pathogenesis of PCOS, increased peripheral aromatization of androstenedione to estrone leads to excessive secretion of luteinizing hormone (LH), which leads to a further increase in ovarian androgen production8. According to an alternative hypothesis, hyperandrogenism results from dysregulation of ovarian cytochrome P-450 17-hydroxylase, a key enzyme in androgen synthesis; such dysregulation may be either dependent or independent of hypersecretion of LH9. A mechanism for gonadotropin-independent hyperandrogenemia is suggested by increased cortisol clearance, which leads to a compensatory rise in corticotropin secretion to maintain normal cortisol secretion, with the inevitable consequence of increased adrenal production of androgens10. To maintain tissue-specific aldosterone action when cortisol concentrations are normal, cortisol is rapidly oxidized to inactive cortisone by the oxidoreductase 11-hydroxysteroid dehydrogenase, thus excluding cortisol from access to mineralocorticoid receptors11. This enzyme is widely distributed in tissue, and its most important action occurs in the kidney. If the oxidase activity of 11-hydroxysteroid dehydrogenase is reduced, cortisol metabolism decreases and hypertension develops, mimicking the effects of excess aldosterone production; this mechanism explains the hypertension induced by licorice and carbenoxalone11. Conversely, if the oxidation and clearance of cortisol are increased, cortisol production driven by corticotropin increases, with the inevitable byproduct of increased adrenal synthesis of androgens.
The following findings are evidence of adrenocortical hyperfunction in PCOS: the secretion of dehydroepiandrosterone sulfate, an exclusively adrenal steroid,12 is increased in women with the syndrome,13 both basally and in response to corticotropin14,15; the administration of metyrapone to women with PCOS causes an excessive increase in serum levels of testosterone and androstenedione, and partial adrenal suppression with dexamethasone may correct the hyperandrogenemia16,17; adrenal uptake of [131I]iodocholesterol is increased in women with the disorder18; and the prevalence of polycystic ovaries is higher among women with congenital adrenal hyperplasia and their relatives19.
Other evidence suggests that the ovaries are the main source of excess androgens in PCOS: testosterone and androstenedione are the principal androgens secreted by these organs, and the synthesis of both is increased in polycystic ovarian tissue in vitro20; treatment of women with PCOS with a gonadotropin-releasing hormone-agonist analogue suppresses serum androstenedione and testosterone concentrations to levels found in women whose ovaries have been removed, whereas serum dehydroepiandrosterone sulfate concentrations are unaltered21; ablation of androgen-producing interstitial tissue of the ovaries by electrocautery or laser surgery temporarily decreases serum testosterone and androstenedione concentrations22; and direct sampling of adrenal and ovarian venous blood suggests that the chief source of excessive androgen levels is the ovary23,24.
Because of continued debate about the contributions of the ovaries and adrenal glands to hyperandrogenemia, we sought to detect disorders of adrenal steroid metabolism in women with PCOS by measuring the urinary excretion of steroid hormone metabolites, thereby assessing their rates of production and metabolism.
Methods
Subjects
We studied 65 women with PCOS whose mean (±SD) age was 25 ±6 years. All presented with either hirsutism (Ferriman-Gallwey score,25 >7) or menstrual irregularity (cycle length of >35 days or oligomenorrhea [intermenstrual interval, >42 days]), or both. All the women had polycystic ovaries on pelvic ultrasonography, according to the criteria of Adams et al.26. Thirty-four women (52 percent) were obese (body-mass index [the weight in kilograms divided by the square of the height in meters], 
25), and 44 (68 percent) had hirsutism. Fifty-six women (86 percent) had elevated serum testosterone concentrations, hirsutism, or both. The nonobese women with PCOS were compared with 27 nonobese normal women whose mean age was 31 ±2 years, and the obese women with PCOS were compared with 18 obese normal women whose mean age was 34 ±5 years. All the normal women had normal ovaries and regular ovulatory cycles as determined by serial ovarian ultrasonography and measurements of serum progesterone during the second half of their cycles. No woman had hyperprolactinemia, and all had normal serum concentrations of free thyroxine and thyrotropin. No woman had elevated urinary cortisol excretion or diabetes mellitus (fasting plasma glucose level, <120 mg per deciliter [<6.7 mmol per liter]), thereby excluding the possibilities of Cushing's syndrome and severe insulin resistance, respectively.
Serum and 24-hour urine samples were obtained within five days after the onset of menstruation or at random in women whose intermenstrual intervals were longer than three months (in which case the possibility of recent ovulation was excluded if the serum progesterone concentration was less than 1.6 µg per liter [<5 nmol per liter]). Serum 17
-hydroxyprogesterone was measured before and 60 minutes after the intramuscular administration of 250 mg of corticotropin (Synacthen, Ciba, Basel, Switzerland). Ovarian ultrasonography was performed in all women while their bladders were full.
The study protocol was approved by the hospital ethics committee, and all the women gave informed consent.
Steroid Measurements
Urinary cortisol was extracted with methylene chloride and measured by radioimmunoassay (Diagnostic Products, Los Angeles). The range in 89 normal women was 11 to 79 µg (30 to 220 nmol) per 24 hours. Serum LH and follicle-stimulating hormone (FSH) were measured by immunoradiometric assay (Serono Diagnostics, Woking, Surrey, United Kingdom), and the results expressed in terms of International Reference Preparations 68/40 (for LH) and 79/549 (for FSH). Serum total testosterone was measured by direct radioimmunoassay. The intraassay and interassay coefficients of variation were 3 percent and 12 percent, respectively6.
Urinary creatinine was assayed by an automated method based on the Jaffe (alkaline picrate) reaction. Urinary steroids were analyzed as previously described27 except that before hydrolysis, free glucuronide and glucuronide conjugates and steroid sulfate fractions were detected by chromatography on Sephadex LH-20. The fraction containing free and conjugated glucuronides was hydrolyzed with Helix pomatia digestive juice27. The S fraction was hydrolyzed for 48 hours at 37 °C after the addition of 5 ml of ethyl acetate acidified with 1 M hydrochloric acid. The intraassay and interassay coefficients of variation ranged from 7 to 21 percent and from 11 to 22 percent, respectively, for individual metabolites. The results were analyzed with particular reference to the groups of urinary steroid metabolites listed in Table 1 and shown in Figure 1. The C19 steroid sulfates, consisting of dehydroepiandrosterone sulfate and its major metabolites, are markers of adrenal androgen secretion. The levels of the principal urinary metabolites of cortisol (tetrahydrocortisone, tetrahydrocortisol, allotetrahydrocortisol, - and -cortols, and - and -cortolones) reflect daily cortisol production.
|
|
-reduced and 5-reduced metabolites of the latter two steroids reflects the relative rates of 5-reductase and 5-reductase activity. The ratios of total 11-oxo (oxygenated) metabolites to total 11-hydroxy metabolites of corticosterone and of cortisol were used as indexes of relative oxidation-reduction activity of 11-hydroxysteroid dehydrogenase. The normal reference range for steroid metabolites was defined as the mean (±2 SD) of values in the normal women (i.e., in the nonobese and obese subgroups). Statistical Analysis
The urinary steroid levels were expressed per gram of urinary creatinine to eliminate the influence of body mass on cortisol production28. The values were not normally distributed and were therefore logarithmically transformed before statistical analysis. Regression of the logarithm of each variable was performed on body-mass index and PCOS status, which was assigned a score of 0 if absent or 1 if present. The antilogarithm of the coefficient for PCOS for each variable was the ratio of the values in the women with PCOS to the values in the normal women. This method adjusts for body-mass index and eliminates the effect of obesity. The coefficients for PCOS are given as geometric means with 95 percent confidence intervals. A P value below 0.05 was considered to indicate statistical significance.
Results
The clinical and hormonal characteristics of the women with PCOS and the normal women, adjusted for body-mass index, are shown in Table 2. The mean serum LH and testosterone concentrations and ovarian volumes of the women with PCOS were higher, and their serum FSH concentrations and urinary cortisol excretion lower, than those of the normal women. Among the women with PCOS, the serum hormone values of those with hirsutism were not significantly different from values of those without this feature (data not shown).
|
-reduced to 5-reduced metabolites was not significantly different (Table 3). Values for all subjects are shown in Figure 2 according to whether or not they were obese (obesity was indicated by a body-mass index 25).
|
|
-hydroxyprogesterone concentrations recorded before and after corticotropin administration did not reveal late-onset adrenal 21-hydroxylase deficiency29 in any of the subjects. For each hormonal measurement, reference ranges (mean ±2 SD) were also derived from the results in both the obese and nonobese subgroups of normal women (Figure 2). Among the women with PCOS, 55 percent of those who were not obese and 24 percent of those who were had ratios of 11-oxo to 11-hydroxy metabolites of cortisol that were above the upper limits of normal (>1.7 and >2.2, respectively).
Discussion
Polycystic ovaries can be found in association with both early-onset and late-onset congenital adrenal hyperplasia,18 as well as Cushing's syndrome,30 adrenal androgen-secreting tumors,31 hyperprolactinemia,32 severe insulin resistance,33 and acromegaly34 and are associated with clinical syndromes often indistinguishable from primary PCOS. Long-term administration of androgens can produce similar morphologic changes in the ovaries35,36. However, the majority of women whose ovaries were polycystic as defined anatomically are classified as having primary PCOS.
All forms of PCOS are accompanied by chronic mild-to-moderate androgen excess,5,6,7 which is central to the pathogenesis of the morphologic changes in the ovaries and the clinical syndrome. The chronic anovulation typical of PCOS results in an increased number of atretic follicles (which become cysts) and increased interstitial tissue in the stroma of the ovaries3,37. The expanded stroma produces more androgen even when serum LH concentrations are normal, as in 30 to 40 percent of women with PCOS,2,4,6 with resultant systemic hyperandrogenemia, the most common biochemical abnormality of the syndrome. The primary source of the hyperandrogenemia may be extraovarian or intraovarian, and it is against this background that the relative roles of adrenal and ovarian sources of excess androgen production have been addressed2,8,9.
Our detailed analysis of the pattern of steroid metabolite excretion in a large group of women with PCOS demonstrated, as expected, increased urinary excretion of testosterone and androstenedione metabolites, although this finding does not help to identify the origin of the excess androgens. However, the increased excretion of metabolites of C19 steroids and cortisol in these women indicates that the steroidogenic activity of the zona reticularis and zona fasciculata of the adrenal cortex is increased. These changes cannot be explained solely by the accompanying obesity.
Since adrenal androgen production is corticotropin-dependent, it will occur whenever cortisol synthesis is limited, as in persons with congenital adrenal hyperplasia, cortisol resistance, or enhanced metabolism or clearance of cortisol. Cortisol clearance may be enhanced by changes in the relative activity of catabolic enzymes. Thus, Stewart et al.10 reported evidence of increased activity of 5-reductase, which converts cortisol to inactive 5-dihydrocortisol, in women with PCOS. Although we were unable to confirm these observations, we found increased ratios of 11-oxo to 11-hydroxy metabolites of both cortisol and corticosterone in women with the syndrome. These increases reflect increased 11-oxidation of cortisol and corticosterone. Such a mechanism provides an explanation for the increased production of cortisol that is evident from the increased excretion of urinary cortisol metabolites. Thus, increased metabolic clearance of cortisol may be a primary defect accounting for chronic adrenal hyperandrogenemia and subsequent ovarian changes in a large proportion of women with PCOS.
The interconversion of cortisol, which is metabolically active, and cortisone, which is inactive, is determined by the level of 11-hydroxysteroid dehydrogenase in many tissues, including the liver, kidney, skin, and adipose tissue11. Little is known about the regulation of the relative oxidoreductase activity of this enzyme in vivo. We considered the possibility that dysregulation of 11-hydroxysteroid dehydrogenase activity may be due to hyperandrogenemia itself, although this seems unlikely since the ratio of 11-oxo to 11-hydroxy metabolites of cortisol is normal in men, whose rate of testosterone production is 10 times the rate in women (unpublished data). Two of the major sites of 11-hydroxysteroid dehydrogenase activity -- the liver and adipose tissue -- are also targets of insulin action. Chronic hyperinsulinemia is a well-recognized feature of PCOS and is independent of obesity,38,39 and it is possible that hyperinsulinemia could cause dysregulation of 11-hydroxysteroid dehydrogenase activity in these tissues.
If enhanced clearance of cortisol results in adrenal overstimulation, and consequently increased adrenal androgen synthesis, it should be possible to detect increased serum corticotropin concentrations in women with PCOS in whom urinalysis shows an increased ratio of 11-oxo to 11-hydroxy metabolites of cortisol. However, subtle resetting of the hypothalamo-corticotropic-adrenal axis cannot be demonstrated in patients with biosynthetic defects due to late-onset congenital adrenal hyperplasia40. Alternatively, the adrenal cortex may have an enhanced sensitivity to normal amounts of corticotropin, a possibility supported by the observation that in women with PCOS, the serum level of 11-hydroxyandrostenedione (an exclusively adrenal androgen) increases in response to the administration of corticotropin-releasing hormone41.
In summary, increased urinary excretion of cortisol metabolites and C19 steroid sulfates provides evidence of increased adrenocortical activity in many women with PCOS. The increased ratio of the 11-oxo metabolites of cortisol and corticosterone to their 11-hydroxy metabolites in urine indicates enhanced oxidation by 11-hydroxysteroid dehydrogenase. We propose that this mechanism, whether primary or secondary, is responsible for chronic adrenal hyperandrogenism in many women with PCOS.
Source Information
From the Division of Biochemical Medicine, St. George's Hospital Medical School (A.R.), the Department of Clinical Biochemistry, the Royal London Hospital (H.T.), and the Department of Clinical Biochemistry, King's College School of Medicine (N.T.), London; and the Department of Endocrinology, North Staffordshire Royal Infirmary, Stoke-on-Trent, United Kingdom (R.C.).
Address reprint requests to Professor Clayton at the Department of Endocrinology, North Staffordshire Royal Infirmary, Hartshill, Stoke-on-Trent ST4 7LN, United Kingdom.
References
-reductase activity in polycystic ovary syndrome. Lancet 1990;335:431-433. [CrossRef][Medline]
-Hydroxysteroid dehydrogenase and enzyme-mediated receptor protection: life after liquorice? Clin Endocrinol (Oxf) 1991;35:281-289. [Medline]
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Hyperandrogenism in Polycystic Ovary Syndrome
Stewart P. M., Edwards C. R.W., Rodin D. A., Clayton R. N.
Extract |
Full Text
N Engl J Med 1994;
331:131-132, Jul 14, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
