Impact of repeated antral follicle counts on the prediction of poor ovarian response in women undergoing in vitro fertilization.

OBJECTIVE: To study the value of a single antral follicle count and the additional value of repeated counts in different cycles for the prediction of poor ovarian response in IVF.
DESIGN: Prospective.
SETTING: Tertiary fertility center. PATIENT(S): One hundred twenty women undergoing their first IVF cycle. INTERVENTION(S): Measurement of the number of antral follicles on cycle day 3 in two spontaneous cycles. Ovarian response. RESULT(S): A single antral follicle count is clearly predictive of poor ovarian response and there is good agreement between repeated measurements in subsequent cycles (area under the receiver operating characteristic curve [ROC(AUC)]; cycle 1: 0.87, cycle 2: 0.85). In a logistic regression analysis, information obtained after the second cycle contributed significantly to the prediction of poor response by the antral follicle count of the first cycle. The predictive accuracy of the highest of two counts (ROC(AUC) 0.89) was slightly better than that of each single count. The predictive model with the highest count yielded slightly higher values of specificity and positive predictive value. Sensitivity, negative predictive value, and error rates were slightly lower. CONCLUSION(S): A single antral follicle count is a good predictor of poor ovarian response in IVF. Although the impact of a second antral follicle count on ovarian response predictions in IVF is statistically significant, clinical relevance is very limited. Repeating an antral follicle count in a subsequent cycle is not recommended.