ESPE Abstracts

Background: Hyperinsulinemic androgen excess is the most common cause of hirsutism, acne, seborrhea and menstrual irregularity in adolescent girls. The ovarian androgen excess originates most often from an absolute or relative excess of fat (in adipose tissue and in organs such as the liver) and from the ensuing elevations in insulinaemia and gonadotropin secretion. There is no approved therapy for androgen excess in adolescent girls. The prime recommendation is to reduce body adiposity with lifestyle measures. The addition of an oral estro-progestagen contraceptive (OC) is a standard approach. An alternative is to add an insulin-sensitizing medication for those hyperandrogenic girls who are not sexually active.

Objective and hypotheses: Here we report a first randomized study comparing the effects of a widely prescribed OC (20 μg ethinylestradiol plus 100 μg levonorgestrel) to those of a low-dose insulin-sensitizing combination with pioglitazone (Pio, only 7.5 mg/d), spironolactone (Spi, 50 mg/d) and metformin (Met, 850 mg/d) in adolescent girls with hyperinsulinemic androgen excess and without need for contraception (n=30).

Method: Endocrine-metabolic markers, carotid intima-media thickness (US), visceral and hepatic adiposity (MRI).

Results: Over 1 year, both treatments attenuated the clinical and endocrine measures of androgen excess. However, pioglitazone-spironolactone-metformin (PioSpiMet) had more normalizing effects than OC, particularly on fasting and post-oGTT insulinemia, and on circulating GGT and C-reactive protein (all P between 0.01 and 0.001). OC and PioSpiMet had opposing effects on hepatic adiposity so that, after 1 year, the OC girls had about twice as much liver fat as the PioSpiMet girls (P<0.0001). There were no noteworthy side-effects in either subgroup.

Conclusion: PioSpiMet conferred more broadly normalizing effects than OC over 1 year. Given that adolescence is a potential time window of developmental plasticity (including for the liver), it will be of interest to study whether on-treatment divergences between OC and PioSpiMet girls do persist in the post-treatment phase.