Background. An androgen-secreting tumor needs to be excluded in any woman with severe hyperandrogenism. We sought to characterize patients with biochemical hyperandrogenism in respect of tumor versus non-tumor etiologies, explore possible links between non-tumor hyperandrogenism and metabolic syndrome, and ascertain whether metformin therapy can elicit diagnostic reductions in serum testosterone (T). Patients and methods. Seven-year retrospective study of all women referred to a university hospital endocrinology service with baseline T >4.0 nmol/l. Dataset comprised age, menopausal status, body mass index (BMI), presence/absence of hypertension, diabetes, acanthosis or dyslipidemia, along with changes in BMI and serum T following intervention with metformin, oophorectomy or dexamethasone. Non-tumor hyperandrogenism was defined by normalization of serum T or >40% reduction from baseline. Results. Four out of 18 cases had adrenal carcinoma that was clinically obvious at initial presentation (one virilized, three Cushingoid). The remaining 14 were characterized by metabolic syndrome (BMI: 39.9 ± 8.1 kg/m2), serum T of 6.14 ± 1.6 nmol/l, and nadir serum T following intervention of 2.2 ± 1.04 nmol/l. Diagnostic reductions in serum T occurred in 11/12 patients treated with metformin. Conclusions. Non-tumor hyperandrogenism with markedly elevated serum T and associated metabolic syndrome is a defined clinical entity in postmenopause as well as in premenopausal women with polycystic ovary syndrome. This has hitherto been only sparsely documented in the published literature. A fall in serum T level in response to insulin-sensitizing therapy with metformin and lifestyle change may be a reassuring indicator that such women are highly unlikely to harbor an androgen-secreting tumor.
- Androgen-secreting tumor
- Metabolic syndrome
- Polycystic ovary syndrome
- Postmenopausal ovarian hyperthecosis