Michael H Jacobs1, Laura M Reuter2, Valerie L Baker3, LaTasha B Craig4, Denny Sakkas5, Eric Surrey6, Kevin J Doody7, Emily S Jungheim8, Aykut B Bayrak9, Martin Hund10, Wilma D J Verhagen-Kamerbeek10, Deborah Pardue11, Katharina Buck12, Barbara Timm13. 1. Fertility and IVF Center of Miami, 8950 North Kendall Drive, Suite 103, Miami FL 33176, USA. Electronic address: fivf@aol.com. 2. Midwest Fertility Specialists, 12188-A North Meridian Street, Suite 250, Carmel IN 46032, USA. 3. Stanford Fertility Specialists, 1195 West Fremont Avenue, Sunnyvale CA 94087, USA. 4. University of Oklahoma Health Sciences Center, 1100 North Lindsay Avenue, Oklahoma City OK 73104, USA. 5. Boston IVF, 130 Second Avenue, Waltham MA 02451, USA. 6. Colorado Center for Reproductive Medicine, 10290 RidgeGate Circle, Lone Tree CO 80124, USA. 7. Center for Assisted Reproduction, 1701 Park PI Avenue, Bedford TX 76022, USA. 8. Washington University Fertility and Reproductive Medicine Center, 4444 Forest Park Avenue No. 3100, St. Louis MO 63110, USA. 9. LA IVF, 2080 Century Park, Los Angeles CA 90067, USA. 10. Roche Diagnostics International Ltd, Forrenstrasse 2, Rotkreuz 6343, Switzerland. 11. Roche Diagnostics Operations Inc., 9115 Hague Road, Indianapolis IN 46250-0457, USA. 12. Roche Diagnostics GmbH, Nonnenwald 2, Penzberg 82377, Germany. 13. Heartland Center for Reproductive Medicine, 7308 South 142nd Street, Omaha NE 68138, USA.
Abstract
RESEARCH QUESTION: What concentration of anti-Müllerian hormone (AMH) corresponds to an antral follicle count (AFC) >15 for determination of ovarian reserve? DESIGN: A prospective study conducted at 13 US fertility clinics in women aged 21-44 years who presented for AFC evaluation by transvaginal ultrasound. Serum samples were collected at the time of AFC evaluation (menstrual cycle day 2-4). AMH concentrations were measured by the Elecsys® AMH immunoassay; oestradiol and follicle-stimulating hormone (FSH) concentrations were also measured. The serum AMH cut-off able to detect AFC >15 with high sensitivity was determined (derivation cohort). Clinical performance of the AMH assay at the derived cut-off was evaluated (validation cohort). Receiver operating characteristic (ROC) analyses were also performed. RESULTS: In the derivation cohort (n = 306), an optimal serum AMH cut-off value of 1.77 ng/ml was determined to correspond to AFC >15 with 89.63% sensitivity and 69.01% specificity, using the Elecsys AMH assay. In the validation cohort (n = 856), this 1.77 ng/ml cut-off could identify women with an AFC >15 with a sensitivity of 88.34% and a specificity of 68.29%; corresponding positive predictive and negative predictive values were 75.19% and 84.34%, respectively. ROC analyses demonstrated that AMH performed better than oestradiol or FSH in predicting AFC, with area under the curves of 85.7%, 57.1% and 69.7%, respectively, in the validation cohort. CONCLUSION: The Elecsys AMH immunoassay provides a robust and fully automated method to measure serum AMH levels. Women with AMH values below the cut-off of 1.77 ng/ml are unlikely to have AFC >15.
RESEARCH QUESTION: What concentration of anti-Müllerian hormone (AMH) corresponds to an antral follicle count (AFC) >15 for determination of ovarian reserve? DESIGN: A prospective study conducted at 13 US fertility clinics in women aged 21-44 years who presented for AFC evaluation by transvaginal ultrasound. Serum samples were collected at the time of AFC evaluation (menstrual cycle day 2-4). AMH concentrations were measured by the Elecsys® AMH immunoassay; oestradiol and follicle-stimulating hormone (FSH) concentrations were also measured. The serum AMH cut-off able to detect AFC >15 with high sensitivity was determined (derivation cohort). Clinical performance of the AMH assay at the derived cut-off was evaluated (validation cohort). Receiver operating characteristic (ROC) analyses were also performed. RESULTS: In the derivation cohort (n = 306), an optimal serum AMH cut-off value of 1.77 ng/ml was determined to correspond to AFC >15 with 89.63% sensitivity and 69.01% specificity, using the Elecsys AMH assay. In the validation cohort (n = 856), this 1.77 ng/ml cut-off could identify women with an AFC >15 with a sensitivity of 88.34% and a specificity of 68.29%; corresponding positive predictive and negative predictive values were 75.19% and 84.34%, respectively. ROC analyses demonstrated that AMH performed better than oestradiol or FSH in predicting AFC, with area under the curves of 85.7%, 57.1% and 69.7%, respectively, in the validation cohort. CONCLUSION: The Elecsys AMH immunoassay provides a robust and fully automated method to measure serum AMH levels. Women with AMH values below the cut-off of 1.77 ng/ml are unlikely to have AFC >15.