Can the fall in serum FSH during coasting in IVF/ICSI predict clinical outcomes?

This retrospective cohort study determined whether the total falls in serum FSH and oestradiol concentrations from start to end of coasting in IVF/intracytoplasmic sperm injection could predict clinical outcomes. Ninety-nine cycles, with gonadotrophin-releasing hormone-agonist down-regulation where coasting with serial serum oestradiol and FSH monitoring was adopted due to risk of severe ovarian hyperstimulation syndrome, were consecutively included. The primary clinical outcome was live-birth rate (LBR); other outcomes measured were number of oocytes retrieved and fertilization, implantation and clinical pregnancy rates. LBR for FSH fall >10 IU/l compared with 5–10 and < 5 IU/l were 45.4% versus 22.0% and 25.0%, respectively. Mean serum FSH fall was similar with and without live birth (8.4 ± 6.2 versus 7.3 ± 5.0 IU/l) as were mean oestradiol and FSH concentrations on HCG administration, oestradiol fall, percentage fall in FSH/oestradiol and duration of coasting. None of the variables efficiently predicted live birth on regression analysis. The AUC of FSH fall was 0.53 at 11.0 IU/l. Basal FSH, starting and total gonadotrophin dose and duration of coasting were positively correlated with FSH fall. A potentially clinically important association between live birth and FSH fall during coasting was apparent, which requires further evaluation.The purpose of this retrospective cohort study was to determine whether the magnitude of fall in the serum FSH and oestradiol concentrations from start to end of coasting in IVF/intracytoplasmic sperm injection cycles could predict the clinical outcomes. Gonadotrophin-releasing hormone-agonist down-regulated cycles (n = 99), where coasting with serial serum oestradiol and FSH monitoring was adopted due to risk of ovarian hyperstimulation, were consecutively included. Live birth was the primary clinical outcome measured; number of oocytes retrieved and fertilization, implantation and clinical pregnancy rates were the other outcomes examined. Live-birth rate tended to be high when FSH fall was >10 IU/l, compared with 5–10 IU/l and <5 IU/l, although not statistically significantly. Mean serum FSH fall were similar in live-birth and no-live-birth cycles (8.4 ± 6.2 versus 7.3 ± 5.0) as were mean oestradiol and FSH concentrations on hCG administration, oestradiol fall, percentage fall in FSH and oestradiol and duration of coasting. None of the variables efficiently predicted live birth. The area under the curve of FSH fall was 0.53. FSH fall of <11.0 IU/l was found to be more likely to predict negative outcome (specificity 84.72%) than predicting positive outcome when FSH fall was >11 IU/l (sensitivity 34.48%). Women’s basal FSH, starting and total gonadotrophin dose of ovarian stimulation and duration of coasting had direct positive correlation with the magnitude of FSH fall. A potentially clinically important rise in live birth in association with greater FSH fall during coasting was apparent, which requires further evaluation.