Predicting fertility.

Various predictors of fertility have been described, suggesting that none are ideal. The literature on tests of ovarian reserve is largely limited to women undergoing in vitro fertilization, and is reliant on the use of surrogate markers, such as cycle cancellation and number of oocytes retrieved, as reference standards. Currently available prediction models are far from ideal; most are applicable only to subfertile women seeking assisted reproduction, and lack external validation. Systematic reviews and meta-analyses of predictors of fertility are limited by their heterogeneity in terms of the population sampled, predictors tested and reference standards used. There is an urgent need for consensus in the design of these studies, definition of abnormal tests, and, above all, a need to use robust outcomes such as live birth as the reference standard. There are no reliable predictors of fertility that can guide women as to how long childbearing can be deferred.