Polycystic ovary syndrome (PCOS) is a diagnosis of exclusion. We present two cases of women with oligomenorrhoea and high concentration of androstendione, suggestive of possible androgen-secreting tumour, caused by assay interference. The first patient, investigated for oligomenorrhoea, had no significant hirsutism or acne. Androstendione concentration was above 10.0 ng/ml (rr: 0.3-3.3 ng/ml). In order to rule out possible androgen-secreting tumour or hypercortisolaemia we performed 48-hour low dose dexamethasone suppression test (LDDST). This failed to demonstrate adequate suppression of androstendione (6.05 ng/ml and 9.32 ng/ml after the first and the second day, respectively). Pelvic ultrasound examination showed polycystic ovaries, while abdominal CT scan failed to show any ovarian or adrenal lesion. Despite such high androstendione concentrations, urinary steroid profile (gas chromatography/mass spectrometry method) yielded normal results. Hence a possibility of androstendione assay interference was raised. The second patient was also admitted for investigations of oligomenorrhoea. Clinical examination was unremarkable. There was a high concentration of testosterone 0.78 ng/ml (rr: 0.084-0.481 ng/ml) and androstendione above 10.0 ng/ml (rr: 0.3-3.3 ng/ml). LDDST failed to demonstrate any suppression of androstendione, while recalculated concentrations of androstendione after serial dilutions were markedly lower in comparison to initial values. Therefore, such high androstendione concentrations (i.e. above the upper limit of the assay) must have resulted from assay interference. In both cases a final diagnosis of PCOS was established. Conclusions: In the absence of clinical features, contrasting with unusually high androgen levels, a possibility of androgen assay interference should be considered in differential diagnosis of hyperandrogenism or PCOS.