Utilisation of gonadotrophin-releasing hormone (GnRH) analogue to differentiate ovarian from adrenal hyperandrogenism in postmenopausal women.

A 67-year-old lady reported an increase in hair growth on her face and abdomen during a routine clinic review for follow-up of her type 2 diabetes mellitus. The symptoms had been present for approximately 2 years, with more recent male pattern hair loss. She had no postmenopausal bleeding or weight loss. Her diabetes was well controlled on a basal-bolus insulin regime and metformin with a HbA1c of 51 mmol/mol. She had background retinopathy and treated hypertension and dyslipidaemia. She was obese with a BMI of 31.7 kg/m 2 . On physical examination, there was excess hair growth visible on her chin. Her abdominal hair had been waxed a couple of days before clinic so this was not visibly evident. A mild degree of temporal recession of the hairline was noted. Clitoromegaly was absent and there were no abdominal masses palpable.

E Bahaeldein prepared the case report. M J Brassill reviewed and edited the manuscript and was the physician responsible for both patients care.

This work did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

Histology showed a benign Leydig cell tumour of the left ovary with no evidence of malignancy. Testosterone level normalised to 0.88 nmol/L postoperatively and at further follow-up the patient noted a complete resolution of her hirsutism and improvement in her hairline.

An informed consent has been obtained from the patients for publication of this case report

A single dose of GnRH analogue was administered as per the method outlined in case 1. Her testosterone level suppressed significantly to 0.94 nmol/L following suppression of gonadotrophins. This supported an ovarian source of androgen excess and the patient proceeded to bilateral oophorectomy.

Postmenopausal hyperandrogenism is a rare but important diagnosis, particularly due to the possibility of an underlying malignant cause. Causes of postmenopausal hyperandrogenism can be divided into tumorous (benign or malignant) and non-tumorous. Even when the cause is benign the symptoms of hyperandrogenism may be distressing for the patient, and there is a possibility of increased cardiovascular morbidity due to long-term increased androgen exposure. Appropriate diagnosis and treatment is therefore essential. With an ovarian source of androgen excess determination of a benign diagnosis can typically only be made histologically. We report two cases of postmenopausal hyperandrogenism where imaging did not reveal an adrenal or ovarian abnormality. Response to GnRH analogue was successfully used to determine an ovarian source of hyperandrogenism in both cases. This was subsequently confirmed histologically post bilateral oophorectomy.

Androgen production in premenopausal women occurs in the adrenals, ovaries, and via peripheral conversion of prohormones ( Table 2 ). Overall androgen production gradually declines with age with a marked decline in adrenal androgens from the third decade onward ( 1 ). Ovarian testosterone production does not decrease after menopause leading to an increase in the relative contribution of ovarian testosterone production in postmenopausal women. Hirsutism and alopecia are the most common clinical signs of androgen excess in postmenopausal women, with anabolic appearance, lowering of voice and clitoromegaly occurring with higher androgen levels. Table 2 Source of androgen production in premenopausal women. Adrenal (%) Ovarian (%) Peripheral conversion (prohormone converted) DHEA-S 95–100 0–5 DHEA 50 20 30% (DHEA-S) Androstenedione 50 50 Testosterone 25 25 50% (androstenedione) Source of androgen production in premenopausal women. One of the key steps in the evaluation of postmenopausal hyperandrogenism is to differentiate an adrenal from an ovarian source of excess androgens. While an ovarian source is more common in this age group, there are no reliable clinical features to help with this differentiation. Measurement of other adrenal androgens such as DHEA-S may be helpful if significantly elevated but are not diagnostic. Elevated DHEA-S levels have been reported in cases of ovarian androgen secreting tumours and non-tumorous hyperandrogenism, while normal levels can be seen in adrenal androgen-secreting tumours ( 2 ). DHEA-S levels may also be elevated in patients with diabetes due to dysregulation of the HPA axis. Recommended imaging modalities for the ovary are transvaginal ultrasound or magnetic resonance imaging (MRI) and for the adrenal are CT or MRI ( 3 ). As MRI was not readily available in our institution we performed transvaginal ultrasound and CT. In many cases, such as ours, imaging may not yield a definitive answer. This scenario may be more common in postmenopausal women where an ovarian tumour may be too small to be visualised. The possibility of a non-functioning adrenal adenoma being detected on imaging must also be considered and the presence of an adrenal mass on imaging does not guarantee that it is the source of androgen excess. In these cases ovarian and adrenal venous sampling has been recommended ( 3 , 4 ). Our use of GnRH analogue as a diagnostic test is based on the principle that the secretion of androgens by ovarian tumours has been shown to be gonadotrophin dependent ( 5 ). Hence, the suppression of gonadotrophins by GnRH analogue administration will reduce ovarian androgen production but will not have any effect on adrenal androgen production. Androgen-secreting ovarian tumours are most commonly sex-cord stromal tumours or steroid tumours ( 6 ). Sertoli-Leydig cell tumours are the most common androgen-secreting sex-cord stromal tumour accounting for 0.5% of all ovarian neoplasms. They are typically unilateral and rarely malignant. Pure Sertoli cell tumours are rare and 90% occur in the reproductive years. Hilus cell tumours are steroid tumours arising from Leydig cells in the hilus of the ovary. They are very rare, accounting for 0.02% of all ovarian neoplasms. They typically present with symptoms of hirsutism and virilisation and are most likely to occur in postmenopausal women. Malignancy is extremely rare in hilus cell tumours. The most common non-tumorous ovarian cause of postmenopausal hyperandrogenism is ovarian hyperthecosis ( 3 ). This is a benign condition characterised by elevated testosterone levels in absence of elevation of other androgens. Patients typically present with slowly progressive symptoms of hyperandrogenism and virilisation. The exact aetiology of this condition is unknown however postmenopausal elevated gonadotrophin levels and insulin resistance are thought to play a role. The diagnosis can only be confirmed histologically. Previous case reports of postmenopausal hyperan­drogenism have described a suppressive effect of short and long-acting GnRH analogue on testosterone secretion from ovarian tumours ( 7 ). A 50% reduction in testosterone levels was seen with short-acting GnRH analogue and a normalisation of testosterone levels with long-acting GnRH analogue. GnRH analogue use has also been described in the treatment of postmenopausal ovarian hyperandrogenism as a non-surgical alternative. Successful suppression of testosterone levels and normalisation of symptoms was achieved in three cases described by Vollard et al . in their case series ( 8 ). Consideration of its use was suggested by the authors in patients who may be poor surgical candidates due to comorbidities or in patients who are unwilling to undergo surgery. Although ovarian malignancies are rare in this setting the possibility needs to be borne in mind and ongoing follow-up would be required. Suppression of androgen production from an adrenal source with GnRH analogue use has been described in postmenopausal women with adrenal adenomas and bilateral adrenal macronodular hyperplasia ( 9 ). This raises the possibility of a false-positive response to a GnRH analogue test. Of note, many of these cases did have adrenal abnormalities evident on imaging. There are no current recommendations for the use of GnRH analogue as a diagnostic tool to help confirm an ovarian source of androgen excess. Available algorithms have recommended combined adrenal and ovarian venous sampling if imaging is inconclusive ( 3 , 4 ). This procedure is technically difficult however, and requires an experienced operator. Even in a specialist unit overall success rates for catheterisation of all four veins was only 27% in one study ( 10 ). Our cases add to the literature supporting the effectiveness of GnRH in suppressing ovarian hyperandrogenism ( 7 , 8 ). We suggest that GnRH analogue use should be considered as the diagnostic test of choice in the setting of postmenopausal hyperandrogenism where imaging is inconclusive. It is non-invasive, does not require technical expertise and can be performed in an outpatient setting.

The authors declare that there is no conflict interest that could be perceived as prejudicing the impartiality of this case report.

Her hormonal profile showed a markedly elevated testosterone at 14.45 nmol/L with elevated FSH and LH consistent with menopause. Repeat testosterone confirmed elevation at 11.14 nmol/L. Full blood count, renal and liver profiles, TSH, HbA1c, random cortisol and prolactin levels were all within normal limits. The patient was not formally screened for hypercortisolism. DHEA-S was mildly elevated at 3.89 µmol/L (0.9–2.1) with remaining androgen levels normal as seen in Table 1 . Transvaginal ultrasound and contrast CT scans did not reveal any adrenal or adnexal abnormalities and the ovaries were reported as atrophic.