Fertility Forecasting: Advances in the Role of Ani-Müllerian Hormone as a Reproductive Biomarker

Abstract

This thesis aimed to investigate the role of Anti-Müllerian hormone (AMH) as a biomarker of current and future fertility. First a critical assessment of AMH as a quantitative and qualitative marker of ovarian function is presented after which the results of 7 original studies are described. In chapters 2 and 3 an individual patient database (IPD) approach was used to assess the value of AMH in predicting outcome in excessive and poor responders to ovarian hyperstimulation during in vitro fertilization (IVF). AMH was shown to be an excellent predictor of ovarian response but not of pregnancy. Next, a critical assessment of reproductive and lifestyle determinants of AMH, in a prospective population based cohort study, revealed that oral contraceptive use, pregnancy and smoking were associated with significantly reduced age-specific AMH values while BMI, waist circumference, alcohol consumption, physical exercise and socioeconomic status were not. The remaining chapters focus on the relationship between AMH and female age at natural menopause (ANM). In chapter 5, we compared the distribution of AMH in 27,563 subfertile to the distribution of ANM in 2,249 women with subfertility and used these distributions to make predictions of age at menopause using AMH. The similarity between the observed ANM and the predicted ANM support the hypothesis that age-related AMH decline is associated with menopause, thus making AMH an excellent candidate biomarker for ANM prediction. In Chapter 6 we compared the predictive performance of AMH with the predictive performance of mother’s ANM. We found that AMH and mother’s ANM are both predictive of ANM, but AMH is a more accurate predictor. In chapter 7 two models of menopausal age prediction were compared and crossvalidated. The results showed that both models discriminate well between women that enter menopause early or late during follow up. However, in one model the proportion of women who were predicted to enter menopause during follow-up coincided well with the actual observed proportion of women who entered menopause during follow up, while for the other model this was poor, owing to a difference in the way in which AMH was modelled as an age and time dependent predictor. Lastly, in a prospective population based sample of 1,163 women it was shown that a model including female age, BMI, pack-years of smoking, and menstrual cycle status discriminated well between women who entered menopause early or late during follow up. Adding AMH to this model increased the accuracy of the prediction of menopause occurring within 10 years, but the added value varied by age and reproductive characteristics of women. In conclusion, AMH has definite potential as a predictor of current and future fertility. As ANM is related to age at the end of female fertility, in the future, AMH may be useful in predicting until which age women remain fertile. However, at present, the precision of the prediction of ANM is not precise enough to be generally applied and important issues must be resolved around the laboratory technicalities of AMH measurements.