RESUMO
BACKGROUND: Unexpectedly poor response leading to IVF cycle cancellation is a distressing treatment outcome. We have prospectively assessed several markers of ovarian reserve in a high risk IVF population to determine their utility in predicting IVF cycle cancellation. METHODS: Eighty-four women at high risk of cycle cancellation due to raised FSH, previous poor response and/or age > or =40 years attending for high-dose short protocol IVF treatment had baseline measures of FSH, inhibin B, anti-Müllerian hormone (AMH), antral follicle count (AFC) and ovarian volume. A GnRH agonist was then administered and, 24 h later, estradiol (E(2)) and inhibin B measures were repeated. RESULTS: Fifty-seven per cent of patients in this study had a poor response to stimulation, and 15% were cancelled. Using multivariate logistic regression, we found that day 3 inhibin B levels were the best predictor of cycle cancellation with an area under the receiver operating curve (ROC AUC) = 0.78 (P = 0.017). When only considering baseline variables, mean ovarian volume was the best predictor of cycle cancellation (ROC AUC = 0.78; P = 0.016). AMH concentrations were the best predictor of a poor response (P = 0.003), and AMH was also predictive of cycle cancellation (P = 0.007) with very little inter-cycle variability. None of the parameters studied were predictive of ongoing pregnancy. CONCLUSIONS: This group of at-risk patients had a high rate of poor response to simulation and cancellation. Although several measures of ovarian reserve were able to predict cycle cancellation, none were able to predict pregnancy. AMH was predictive of both cycle cancellation and poor response with little inter-cycle variability.