Abstract

Objective: This retrospective cohort study investigated trigger day progesterone to total oocyte count (P/OTotal) and progesterone to MII oocyte count (P/OMII) as potential predictors of live birth rate (LBR).

Methods: First cycles of 1,258 infertile patients who attended to the İzmir Tepecik Training and Research Hospital in vitro fertilization unit (IVF) between March 1, 2010 and November 1, 2016 and underwent intracytoplasmic sperm injection treatment following controlled ovarian stimulation + GnRH antagonist cycles with fresh embryo transfer were investigated.

Results: LBR for cycles with P levels ≤0.5, 0.5-1.5 and >1.5 ng/mL were 16.3% (50/307 cycles), 23.0% (180/783), and 19.6% (33/168), respectively. LBR was significantly lower in ≤0.5 ng/mL P level group than the 0.5-1.5 ng/mL group. After the exclusion of cycles with ≤0.5 ng/mL P level, P/OTotal and P/OMII ratios were found to have significant area under the curve values (0.544, p=0.049; 0.552, p=0.022, respectively) for predicting LBR, whereas P levels (0.509, p=0.686) on its own was not a significant predictor. In ROC models adjusted for cycle characteristics, P/OTotal and P/OMII ratio values showed neither better nor weaker prediction performance (0.608, p<0.001; 0.610, p<0.001, respectively) than using P and 1/O or P and 1/MII as separate parameters (0.610, p<0.001; 0.611, p<0.001, respectively).

Conclusion: The predictive performance of P/OTotal and P/OMII ratios for LBRs are similar when O, P, and MII are used as separate parameters. Consequently, these ratios can be used as feasible clinical markers and have the advantage of simpler interpretation.