Human chorionic gonadotrophin stimulation test as a predictor of ovarian response and pregnancy in IVF cycles stimulated with GnRH agonist gonadotrophin treatment: a pilot study.

BACKGROUND Recent evidence supports a specific and broad role of androgen produced by theca cells in reproductive physiology. This pilot study evaluated the usefulness of hCG theca stimulation test in predicting ovarian response and pregnancy. METHODS Prospective cohort study including 80 infertile women treated with IVF/ICSI. On Day 3 of the menstrual cycle preceding, the first IVF/ICSI cycle a blood sample was drawn to evaluate baseline FSH, estradiol (E(2)), 17-hydroxy-progesterone, androstenedione and testosterone levels. All women then received 250 µg recombinant hCG s.c. and underwent a second blood sampling 24 h after hCG injection to measurement steroid serum levels. RESULTS Percentage increment of E(2) but not its precursors was significantly higher in normo-responders and pregnancy cycles than in poor responders and non-pregnancy cycles (P = 0.03 and P = 0.02, respectively) diagnostic accuracy being 67 and 75%, respectively. The percentage increase in E(2) thus still fails in as many as 33 and 25% of patients in predicting ovarian response and pregnancy, respectively. In addition, E(2) concentrations are poorly reproducible and a wide range of variation in all serum steroids investigated-including E(2)-after hCG injection was observed. CONCLUSIONS The predictive power of the hCG test is based on E(2) but not androgen response to hCG injection. This test cannot be recommended in routine clinical practice because it is too laborious for screening purposes, shows great variability in the response obtained and its overall accuracy is not better than that reported for other available markers of ovarian reserve. The use of the currently available markers, antral follicle count and anti-Müllerian hormone, is therefore recommended.