Antimüllerian hormone and antral follicle count thresholds for hyperresponse risk assessment in in vitro fertilization: a Hyperresponse Risk Assessment consensus study.

Objective: To determine the serum antimüllerian hormone (AMH) and antral follicle count (AFC) thresholds indicating an increased risk of hyperresponse to ovarian stimulation (OS) during in vitro fertilization, as defined by the Hyperresponse Risk Assessment (HERA) Delphi Consensus.

Design: A retrospective multicenter cohort study.

Subjects: Women with normal ovarian reserve markers according to the POSEIDON criteria (AMH level of ≥1.2 ng/mL and AFC of ≥5) undergoing their first in vitro fertilization/ intracytoplasmic sperm injection cycle with conventional OS (follicle-stimulating hormone [FSH] level of ≥150 IU/d) using the gonadotropin-releasing hormone antagonist protocol (2015-2017) were included.

Exposure: Hyperresponse was defined as ≥15 retrieved oocytes, on the basis of the HERA definition, compared with non-HERA hyperresponders, defined as patients with ovarian reserve markers within the normal range per the POSEIDON criteria and with <15 oocytes retrieved.

Main Outcome Measures: The primary outcome was the AMH and AFC threshold values, indicating an increased risk of a hyperresponse, using receiver operator characteristic curves. Outcomes were further stratified by patients' age (<35 and ≥35 years). Multivariable logistic regression explored factors associated with an HERA hyperresponse.

Results: A total of 4,220 patients were included, of whom 2,132 (50.5%) were hyperresponders. Receiver operator characteristic curves revealed the following thresholds for a hyperresponse: AMH level of ≥4.38 ng/mL (area under the curve [AUC], 0.71) and AFC of ≥16 (AUC, 0.80) for the entire cohort; AMH level of ≥4.95 ng/mL (AUC, 0.68) and AFC of ≥18 (AUC, 0.76) for women aged <35 years (N = 3,056); and AMH level of ≥4.33 ng/mL (AUC, 0.77) and AFC of ≥15 (AUC, 0.86) for women aged ≥35 years (N = 1,164). Older women received higher median daily and total FSH doses than younger women. The AMH, AFC, female age, daily/total gonadotropin dose, type of gonadotropin, and trigger strategy were significant predictors for hyperresponse.

Conclusion: The AMH and AFC values at and above these thresholds warrant increased caution when planning gonadotropin dosing, regimen, and trigger strategies before OS. These thresholds were lower in older women, potentially due to higher FSH dosing in this population.