Cycle day, estrogen level, and lead follicle size: analysis of 27,790 in vitro fertilization cycles to determine optimal start criteria for gonadotropin-releasing hormone antagonist

Cycle day, E2 level, and follicle size at time of antagonist start are all independent predictors of clinical pregnancy after IVF.

Volume 109, Issue 4, Pages 633–637


Brianna M. Lyttle Schumacher, M.D., Jennifer E. Mersereau, M.D., M.S.C.I., Anne Z. Steiner, M.D., M.P.H.



To determine the optimal criteria at which to start GnRH antagonists during controlled ovarian hyperstimulation (COH) for in vitro fertilization (IVF).


Retrospective clinical cohort.


IVF clinics.


Women undergoing fresh autologous IVF using GnRH antagonist for ovulation suppression during COH.


Measurement of lead follicle size, E2 level, and cycle day of stimulation on day of antagonist initiation.

Main Outcome Measure(s)

Clinical pregnancy rate (PR).


The highest clinical PR was achieved when the antagonist was started when a lead follicle reached 14–15.9 mm in size (mean clinical PR 21.3; 95% confidence interval [CI] 19.3, 23.6) on cycle day 6 (mean clinical PR 22.2; 95% CI 17, 28.4), or when the E2 level was between 500 and 599 pg/mL (mean clinical PR 25.4; 95% CI 19.5, 32.4). Starting antagonists when the E2 level was <300 or >1,100 pg/mL reduced the odds of clinical pregnancy by 40% (odds ratio 0.60, 95% CI 0.5, 0.7).


Cycle day, E2 level, and follicle size at time of antagonist start are all independent predictors of a clinical pregnancy after IVF. Initiating antagonists when the E2 level is extremely low (<300 pg/mL) or extremely high (>1,100 pg/mL) significantly reduces the odds of pregnancy.

Read the full text here.


Go to the profile of Micah J Hill
over 4 years ago

Thank you so much for the interesting article.   When should we start GnRH antagonists is such an important question we frequently debate in IVF rounds with our fellows. 

This is one of our articles for this week in our DC area NIH journal club.   And Im curious to your thoughts on the main discussion I anticipate.

You discuss the potential weaknesses of the retrospective design.  Your analysis is very in depth and multivariate.  The challenge I see is separating ovarian response from day of antagonist start.  From what I could tell, you were unable to incorporate ovarian reserve testing or follicle number  into the multivariate model.  In Table 2A, the groups with highest pregnancy rates seem to represent the groups with most normal responses to ovarian stimulation: ie on average follicles 14-16mm with antagonist start with E2 400-700 on cycle day 6-8. This is before antagonist start, so does this represent the right time to start antagonist or a perfectly responding group of patients?

Conversely, the poorest performing groups in Table 2B seems to represent extremes of ovarian response.  For example first line: day 6 with an E2 up to 1099 and a lead of 14mm seems to be patients with rapid recruitment of a few follicles.  Conversely the next line is E2 up to 199 with the same follicle size but 2 days later seems to represent a slow recruiter of a few follicles.  The last line is patients with a lead of 15.9mm but an E2 under 200 on cycle day 4 or less.  Looks like a single follicle recruiter.  So the poor pregnancy people seem to represent lower ovarian reserve.

So the good result antagonist start patients seem to represent a normal response to stimulation.  While the poor result patients seem to represents the extremes of response, either slow or fast recruitment of a lower number of follicles. The variation between A and B in Table 2 seems to suggest this strongly.

So the question Im sure we will debate is: since neither ovarian reserve nor the number of follicles were modeled, do these results represent the affect of the day of antagonist start or do they represent the patients underlying ovarian reserve and response to stimulation?  Given that follicle size, cycle day start, and E2, in figure 2 are all inverted V shaped, Im inclined to believe they represent normal response to stimulation.  Would love your thoughts!