Premature progesterone rise at hCG injection in assisted reproductive technology (ART) is related to the down regulation protocol and has different impact on ART outcome according to protocol

Abstract

OBJECTIVE: In ART, a premature progesterone (P4) rise can be observed before triggering ovulation. The objective of this study was to analyze the relationships between P4 levels, the down regulation protocol and ART outcome. DESIGN: Retrospective analysis of a prospective ART cohort. MATERIALS AND METHODS: All cycles performed from 2005 to 2008 in 2 clinics collaborating with a single endocrinology/ART lab, with a serum P4 measurement on trigger day and down regulation using GnRH agonists, (short (SP) or long (LP) protocol), or antagonists (Ant) were included (n=6697). P4 was measured by chemiluminescence immunoassay, and classified in 3 groups: <1.5, 1.5-1.9 and ≥2.0 ng/ml. We analyzed the relationship between protocols and P4 and between P4 and cycle outcome (pregnancy rate per OPU (PR) and delivery rate per pregnancy (DR)) according to protocol. RESULTS: There was no difference in P4 between the 3 protocols. However, high levels were significantly detrimental on PR and DR for LP only. These results were confirmed by multivariate analysis, including women's age, endometriosis, polycystic ovaries, and gonadotropin, which showed a significant impact of high P4 on PR among only long protocol (OR=0.67; 95%CI=0.45-0.99), independent of the gonadotropin used.Tabled 1Progesterone level at HCG day (ng/mL)<1.51.5-1.9≥ 2.0pcycles (%)SP(n=501)86.88.25.0NSLP(n=4219)85.68.65.8NSAntagonists (n=1977)87.77.94.4NSPR/OPU (%)SP16.312.220.0NSLP27.924.119.80.01Antagonists23.723.713.8NSDR/pregnancy (%)SP(n=69)68.960.0100NSLP(n=1042)81.173.568.20.04Antagonists (n=422)74.465.758.3NS Open table in a new tab CONCLUSION: Elevated serum P4 at the time of HCG on the cycle outcome influences negatively cycle and pregnancy outcome, only in long down regulated GnRH agonists. However, the possible impact of some patients characteristics cannot be totally discarded, even if a multivariate model took into account the main confounders.