Association of 'normal' early follicular FSH concentrations with unexpected poor or suboptimal response when ovarian reserve markers are reassuring: a retrospective cohort study.

RESEARCH QUESTION: Are basal FSH measurements, when elevated within its normal range, useful for assessing overall ovarian response and predicting unexpected poor or suboptimal ovarian response?

DESIGN: Retrospective cohort study of ovarian stimulation cycles.

RESULTS: A total of 1058 ovarian stimulation cycles (891 first, 167 repeated) were included. Anti-Müllerian hormone (AMH) values were categorized into four (0 to ≤0.6, >0.6 to ≤1.2, >1.2 to ≤3.0, >3.0 to ≤6.25 ng/ml) and basal FSH levels into four groups (<25th percentile: >3.5 to 6.1 IU/ml; 25-75th percentile: >6.1 to ≤8.5 IU/ml; >75-90th percentile: >8.5 to ≤9.9 IU/ml; >90th percentile: >9.9 to ≤12.5 IU/ml). Including only first cycles, a significant independent effect of basal FSH on retrieved cumulus-oocyte complex (COC) count was seen for all basal FSH categories (>90th, >75 to ≤90th, >25 to ≤75th compared with ≤25th percentile, P < 0.001, P = 0.001 and P = 0.007, respectively), when adjusted for age, body mass index (BMI), AMH, antral follicle count (AFC), starting dose and gonadotrophin type. Including only first cycles, patients aged 35 years or older with AFC of 5 or above and AMH 1.2 ng/ml or above, showed significantly higher odds of unexpected poor or suboptimal response if they had higher basal FSH values. Most prominently in the above 90th percentile group (OR 8.64, 95% CI 2.84 to 28.47 compared with <25th percentile) but lower categories (>25th to ≤75th percentile: OR 3.04, 95% CI 1.42 t 6.99; >75th to ≤90th percentile: OR 3.47, 95% CI 1.28 to 9.83 compared with ≤25th percentile) also showed a significant association after adjusting for age, AMH, BMI, AFC, dose, and gonadotrophin type. In patients with a second cycle, an increase in FSH levels in the second round compared with the first was associated with fewer retrieved COCs (estimate: -0.44, 95% CI -0.44 to -0.05, P = 0.027). This effect was adjusted for changes in age, FSH, AFC, starting dose, stimulation duration and change in medication type.

CONCLUSIONS: Basal FSH is independently associated with overall ovarian response. Moreover, it is associated with unexpected poor or suboptimal response in patients, who would fulfill POSEIDON group 2 criteria after oocyte retrieval.