OBJECTIVE: This study was designed to compare the exogenous FSH ovarian reserve test (EFORT) versus the clomiphene citrate challenge test (CCCT), basal FSH, and basal inhibin B, with respect to their ability to predict poor and/or hyperresponders in an IVF population. DESIGN: Prospective randomized controlled trial. SETTING:Fertility center of a university hospital. PATIENT(S): One hundred ten patients undergoing their first IVF cycle, randomized into two groups. INTERVENTION(S): Fifty-six patients underwent a CCCT, and 54 patients underwent an EFORT. In all patients, the test was followed by an IVF treatment. MAIN OUTCOME MEASURE(S): Ovarian response, expressed by the total number of retrieved oocytes. RESULT(S): Univariate logistic regression showed that the best predictor for poor response is the CCCT (area under receiver operator characteristic curve [ROC-AUC] = 0.87), with maximal accuracy of 0.89. Multiple logistic regression analysis did not produce a better model in terms of improving the prediction of poor response. For hyper response, univariate logistic regression showed that the best predictor is the inhibin B increment in the EFORT (ROC-AUC = 0.92) but with a low maximal accuracy of 0.78. Again, multiple logistic regression analysis did not produce a better model in terms of predicting hyper response. CONCLUSION(S): Our study, the first which compares the CCCT with the EFORT for the prediction of poor and hyperresponders, shows that the CCCT is superior for identification of low responders. The EFORT (inhibin B increment) is superior for prediction of hyper response at the cost of a high rate of false positives. Neither of the two tests seems adequate to act alone for identification of both poor and hyperresponders.
RCT Entities:
OBJECTIVE: This study was designed to compare the exogenous FSH ovarian reserve test (EFORT) versus the clomiphene citrate challenge test (CCCT), basal FSH, and basal inhibin B, with respect to their ability to predict poor and/or hyperresponders in an IVF population. DESIGN: Prospective randomized controlled trial. SETTING: Fertility center of a university hospital. PATIENT(S): One hundred ten patients undergoing their first IVF cycle, randomized into two groups. INTERVENTION(S): Fifty-six patients underwent a CCCT, and 54 patients underwent an EFORT. In all patients, the test was followed by an IVF treatment. MAIN OUTCOME MEASURE(S): Ovarian response, expressed by the total number of retrieved oocytes. RESULT(S): Univariate logistic regression showed that the best predictor for poor response is the CCCT (area under receiver operator characteristic curve [ROC-AUC] = 0.87), with maximal accuracy of 0.89. Multiple logistic regression analysis did not produce a better model in terms of improving the prediction of poor response. For hyper response, univariate logistic regression showed that the best predictor is the inhibin B increment in the EFORT (ROC-AUC = 0.92) but with a low maximal accuracy of 0.78. Again, multiple logistic regression analysis did not produce a better model in terms of predicting hyper response. CONCLUSION(S): Our study, the first which compares the CCCT with the EFORT for the prediction of poor and hyperresponders, shows that the CCCT is superior for identification of low responders. The EFORT (inhibin B increment) is superior for prediction of hyper response at the cost of a high rate of false positives. Neither of the two tests seems adequate to act alone for identification of both poor and hyperresponders.
Authors: Bruno Ramalho de Carvalho; Ana Carolina Japur de Sá Rosa e Silva; Júlio César Rosa e Silva; Rosana Maria dos Reis; Rui Alberto Ferriani; Marcos Felipe Silva de Sá Journal: J Assist Reprod Genet Date: 2008-08-05 Impact factor: 3.412