Dynamic serum testosterone concentration may predict in vitro fertilization yield during ovarian stimulation by a gonadotrophin‐releasing hormone agonist long protocol in women with normal ovarian reserve

To determine whether dynamic changes in serum total testosterone (TT) levels during controlled ovarian stimulation (COS) by a GnRH‐a (gonadotrophin‐releasing hormone agonist) long protocol may predict pregnancy in women with normal ovarian reserve in in vitro fertilization.The TT measurements were added to routine hormone tests during COS. The TT dynamic changes, clinical pregnancy rate, and quality of oocytes and embryos of 109 patients were analyzed.Compared with the non‐pregnancy group, in the pregnancy group the TT level on Gn initial day (TTinitial‐d) increased sharply when the dominant follicle reached a diameter 10–12 mm [TTfΦ (10–12)‐d] and on human chorionic gonadotrophin (HCG) day (TTHCG‐d), also the ratios of TTfΦ (10–12)‐d to TTinitial‐d (1.23 ± 0.37 vs. 1.10 ± 0.58, p = 0.040) and TTHCG‐d to TTinitial‐d (2.32 ± 1.26 vs. 2.00 ± 1.43, p = 0.019) increased notably. Of the 2 TT ratios, the first tertile limit was regarded as the threshold of high TT ratios (1.00, 1.45). High tertiles had higher pregnancy rates than low tertiles (82.86% vs. 42.11%, p = 0.006; 71.43% vs. 44.83%, p = 0.040). The 2 TT ratios were positively associated with the number of metaphase II oocytes and good‐quality embryos.Of COS in a long GnRH‐a protocol with optimized outcome, serum TT kinetics appears to be characterized by sharp rises at the early and late stages of follicle growth. The ratios of TTΦ (10–12)‐d to TTinitial‐d and TTHCG‐d to TTinitial‐d may be predictors for pregnancy and qualitative outcomes of oocytes and embryos.