Variation in antral follicle counts at different times in the menstrual cycle: does it matter?

Antral follicle count (AFC) variation was examined across the menstural cycle and its effect on clinical management assessed. In 79 women, AFC was documented in early (iAFC) and late follicular phase (sAFC). Absolute agreement between iAFC and sAFC and agreement for classification into categories of risk of extremes of ovarian response were examined. Ovarian stimulation protocols designed with iAFC and sAFC, and the predictive value of iAFC and sAFC for extremes of ovarian response, were compared in women undergoing ovarian stimulation. Significant differences were found between iAFC and sAFC (16 [IQR 9-24] versus 13 [IQR 7- 21]; P = 0.001), with moderate agreement for the classification into at risk of extremes of response (k = 0.525). Agreement for protocol selection based on either AFC (k = 0.750) and starting gonadotrophin dose was good (concordance correlation coefficient 0.970 [95% CI 0.951 to 0.982]). Predictive value for iAFC and sAFC was maintained for poor ovarian response and risk of ovarian hyperstimulation syndrome (OR 0.634 [0.427 to 0.920], 0.467 [0.233 to 0.935]) and (OR 1.049 [0.974 to 1.131], 1.140 [1.011 to 1.285]). Across the cycle, AFC varies but does not significantly affect ovarian stimulation protocol design and prediction of extreme ovarian response.