A mathematical model for predicting a couple's chances of a successful pregnancy by IVF after one failed attempt has been developed by US researchers. The model is 1000 times more accurate at predicting a positive outcome than standard methods, which are mainly based on a woman's age. The model may help clinics give more personalised and reliable advice to couples.
In the new study, Prajna Banerjee and colleagues at California's Stanford University looked at existing data collected on the outcomes of over 1600 first-time IVF cycles.
The researchers developed a predictive model for assessing whether a second IVF treatment would lead to a live birth based on 52 clinical characteristics for each couple. These included hormone levels, the number of embryos created from eggs and sperm and how fast they developed, as well as the woman's age and body mass index. They validated their model against a second dataset from more than 600 cycles of IVF.
The most important factor predicting success after the second cycle is the rate the embryos developed during the first cycle of IVF. The researchers found this was four times more important than the woman's age. Dr Mylene Yao, assistant professor of obstetrics and gynaecology at Stanford and lead author of the study, said: 'while age is a major driving force towards decreased fertility, it does not have the highest importance once you have the embryo data'.
The team hope their approach will be used in IVF clinics to provide a more accurate and personalised estimate of birth probabilities in second cycles of IVF. This will help each couple decide whether to continue with IVF after a failed first attempt. 'Understanding their probability of having a live birth, based on their very own clinical data, may serve to support that decision', says Dr Yao.
Since the model is based on data from first cycles, however, it could not be used to predict initial IVF success.
Dr Allan Pacey, senior lecturer at the University of Sheffield and fertility expert, told the BBC: 'to have universal appeal the model must work well in lots of different clinics. IVF is not an exact science, so we've got to be careful not to give people false hope'.
The study was published last week in the journal Proceedings of the National Academy of Sciences.