Skip to main content

01-08-2010 | Gynaecology | Article

IVF prediction model developed for first cycle failures


Free PDF

MedWire News: US researchers have developed a predictive model to help couples who have had a failed IVF cycle decide whether to continue with a second cycle of treatment.

Currently, prediction models for such patients are based on age-related data.

"When used alone, age may be misleading as a prognostic factor, due to the complex and as yet uncharacterized relationships among age, hormonal response, and embryo parameters," write Mylene Yao (Stanford University School of Medicine, California) and colleagues.

The personalized model developed by the team includes data on a woman's previous response to hormone treatment, as well as previous embryo development parameters.

The team generated their prediction model using data for 1,676 first-IVF cycles, including assessment of the relative influence of different first-cycle variables on a live birth outcome.

The top three influential factors were the rate of previous blastocyst development (26 percent relative influence), the total amount of gonadotropins administered (10 percent), and the number of eight-cell embryos observed (9 percent).

Testing the model on an independent cohort of 634 cycles revealed it was 1,000 times more likely to predict an accurate response than an age-based model alone, report Yao et al.

MedWire ( is an independent clinical news service provided by Current Medicine Group, a trading division of Springer Healthcare Limited. © Springer Healthcare Ltd; 2010

By Sarah Guy