Abstract
It has been postulated that the menopause is preceded by a fixed interval of natural sterility related to the number of follicles in the ovaries; the ovarian reserve. This paper hypothesized that an estimate of ovarian reserve might make prediction of age at menopause, which is highly variable, possible. This could then allow improved management of age-related infertility and timely initiation of preventive measures for menopause-related conditions associated with early or late onset of the menopause.
This study proposed anti-Müllerian hormone (AMH), antral follicle count (AFC) and early-follicular follicle-stimulating hormone (FSH) as candidate markers for estimation of ovarian reserve and prediction of age at menopause.
The cohort comprised 302 normoovulatory women drawn from three previous studies. At time 1 (T1), an AFC was performed and AMH/early-follicular FSH concentration measured. At time 2 (T2), an average of 11.2 y post-T1, participants completed detailed questionnaires. On the basis of their answers, individuals were placed into five subgroups: regular cycle, menopausal transition, menopause, use of exogenous oestrogens or surgical removal of uterus/ovaries. Individuals in the latter two groups were excluded from the study, 45 could not be followed up and 12 had missing data, leaving 173 women available to be included in the statistical analysis.
AMH, AFC and FSH at T1 were all significantly correlated with time to menopause (P < 0.001). With the exception of FSH, all markers had good predictive capacities, with that of AMH being further improved when adjusted for age (C-statistic, 0.9). A nomogram for prediction of age at menopause is presented in this paper and shows the normal distribution shifts considerably depending on age-specific AMH.
Limitations of the study included the use of a cohort comprised of women from three previous studies. Furthermore, AMH concentrations in different groups were determined using different assays and a correction factor applied to align the results from the two assays. The numbers used for statistical analysis were fairly small, and larger studies would be required to confirm the findings of this paper. Most significantly, the spread of predicted age at menopause was large (an average of 5.8 y either side of the median), questioning whether AMH is an improvement on current predictors of age at menopause.