Prediction of IVF cycle success with antimullerian hormone levels

Abstract

BACKGROUND: Current estimates indicate approximately 6.7% of women in the US are infertile. Infertility can arise from many different conditions and can be both attributed to male and female factors. In women, measures of ovarian reserve such as AMH level or FSH level on day three of the menstrual cycle can be used to predict reproductive success. In conjunction with additional tests and lifestyle screening, these values help to guide the course of treatment, with some patients moving eventually to in vitro fertilization. In vitro fertilization therapy is cost-prohibitive if the patient does not have insurance coverage, an issue for the majority of patients across the United States. AIMS: To evaluate the correlation between cycle number, age, and AMH level with achieving a successful live birth through a statistical analysis of over 50,000 cycles of in vitro fertilization. Methods: Data was obtained from Practice Highway, a company that owns the electronic medical record (EMR) program eIVF, used by over 140 reproductive endocrinology practices across the country. The database provided a total of 88 variables and 138,526 IVF cycle observations ranging from 01/01/2005 to 09/09/2015. A thorough review of available literature was performed to determine the need for the correlations analyzed in this paper. Utilizing this information in conjunction with the scope of this paper, the number of variables was reduced to four. Linear regression and multivariate regression was performed on the four main study variables to identify possible correlations. Results: The results of our study indicate that there is not a strong correlational relationship between the variables analyzed. There was not a significant correlation between age and whether a live birth was achieved. As expected, the strongest relationship identified through our study is the decrease in AMH as age increases. Conclusions: Our study results indicate that there is no correlation between AMH and live birth, suggesting that this biomarker alone is insufficient to be used by physicians as a means of advising patients on IVF success. Additionally, measures such as number of cycles completed and age at cycle start can be useful when counseling patients on their probability of success with assisted reproductive technologies.