Lack of linear relationship between hyperinsulinaemia and hyperandrogenism

Because of continued debate about the role of insulin in the development of hirsutism and in the induction of the polycystic ovary syndrome, we have evaluated the hormonal pattern in a group of hirsute patients.

Fifty-four hirsute patients (age range 18-39 years) of whom 26 patients were obese (O) (BMI 28-53 kg/m2 and W/H greater than 0.85), 12 with ultrasonographic evidence of polycystic ovaries (O PCO) and 14 with normal ovaries. Twenty-eight patients were within normal weight range, and, of these, 14 presented ultrasonographic evidence of polycystic ovaries and 14 had normal ovaries. Two groups of age-matched subjects (obese and normal weight), normally menstruating, without hirsutism or history of endocrinopathies or ultrasonographic evidence of polycystic ovaries, served as controls.

Androstenedione and testosterone were evaluated in all patients by RIA, following ether extraction, DHEAS, LH, FSH and insulin were evaluated directly by RIA. SHBG was evaluated by the concanavalin method. Free testosterone (FT%) was calculated according to the formula FT = 4.038-1.607 log SHBG. Integrated areas under the response curve were calculated for LH and insulin respectively following i.v. administration of GnRH (100 micrograms) or oral administration of glucose (75 g).

Results (mean +/- standard deviation) showed comparable values of androstenedione in all groups of obese patients and in obese controls (7.3 +/- 2.6 in patients with polycystic ovaries, 7.1 +/- 2.9 in non-polycystic ovary patients and 7.4 +/- 2.6 nmol/l in obese controls, respectively), regardless of baseline and area insulin, the presence or absence of polycystic ovaries, or hirsutism. SHBG levels showed a similar pattern (24 +/- 10, 23.8 +/- 7.9 and 36 +/- 19 nmol/l) as did the percentage of free testosterone, regardless of the presence or absence of hirsutism. Regression analysis of the insulin and LH values (baseline and area) against the androgens and SHBG plasma levels showed that only LH area correlated positively with testosterone (r = 0.36, P less than 0.03), androstenedione (r = 0.44, P less than 0.02), % free testosterone (r = 0.53, P less than 0.001), testosterone/SHBG ratio (r = 0.39, P less than 0.03) and inversely with SHBG (r = -0.57, P less than 0.001).

These results showed (1) no linear relationship between high levels of insulin, ovarian androgen production or free hormone availability, and (2) make it very doubtful that insulin plays a primary role in polycystic ovarian syndrome or hirsutism.