The role of serum AMH and FF AMH in predicting pregnancy outcome in the fresh cycle of IVF/ICSI: a meta-analysis

Purpose: This meta-analysis tries to find and confirm the true prognostic value of serum AMH and the follicle fluid AMH (FF AMH) on the outcome of ART. Method: We identified all studies published by March 2014 with data related to in vitro fertilization", "intracytoplasmic sperm injection", "assisted reproductive technology" and "antimullerian hormone" in Pubmed database. Studies were included if 2 x 2 tables for outcomes of pregnancy in IVF patients in relation to AMH could be constructed or studies which used T-tests to compare clinical indexes including AMH in pregnant and non-pregnant women. And all the patients were less than 46 years old. Results: A total of 26 studies could be used for this meta-analysis. Of these articles 22 studies could be constructed 2 x 2 tables, with 15 for predicting pregnancy and 7 for non-pregnancy. 11 studies used the analysis of T-test, with 7 articles were duplicated. And of the 11 articles, 8 were for the analysis of serum AMH in prediction of pregnancy, 3 were for FF AMH. Because of heterogeneity among studies, calculation of a summary point estimate for sensitivity and specificity was not possible. For the analysis of serum AMH on non-pregnancy, the heterogeneity was moderate (I-squared of 65.9%), the curves indicated positive find (the AUROC is 0.73, 95% CI is 0.69-0.77.). In the T-test group of serum AMH, the DOR for women with pregnancy outcome was 0.232 (95% confidence interval (CI): 0.034-0.43), with less heterogeneity (I-squared of 45.1%). Unfortunately, the predictive value of FF AMH on pregnancy is still unclear because of large heterogeneity (I-squared of 90.5%). Conclusion: Serum AMH, as an independent parameter, can predict pregnancy outcome after assisted conception and the positive correlation with serum AMH and non-pregnancy should not be ignored either. The predictive value of FF AMH on pregnancy is still unclear.