News

Fertility Treatment Success Predicted With Medical Analysis
By: Nicole Ostrow, Bloomberg
Published: July 19, 2010

A computer-based analysis can estimate a woman's chance of success from fertility treatments more accurately than age-based guidelines, Stanford University scientists said.

An analytic model developed in research goes beyond age to take into account women's height, weight, hormonal factors and the health of embryos, said Mylene Yao, author of a study released today in the Proceedings of the National Academy of Sciences. The researchers looked at 634 IVF treatments to test their model. In 60 percent of patients, the scientists when using the computer analysis found a significant difference in the calculation of likelihood of having a baby.

Age alone would be a better predictor of the success of fertility treatment in fewer than 1 in 1,000 cases, Yao said, citing statistical methods used to test the model. Computer- based analyses could help women decide whether to buy repeat IVF treatments that can cost $10,000 a try, Yao said. Almost 75 percent of IVF procedures fail to result in live births, according to the study.

"When people have the information about their own specific chances, they are going to feel more confident about pursuing the treatment," said Yao, an assistant professor of obstetrics and gynecology at Stanford, near Palo Alto, California, in a telephone interview on July 16 "Each patient is truly unique," she said. "Even though no two patients are alike, we're still able to give very predictive information."

In the research, computer analysis found, for instance, that one 38-year-old woman's likelihood of success from a second round of in vitro fertilization was 44 percent, topping age- related estimates of 22 percent to 28 percent, Yao said.

Fertility Spending

Yao said she plans to develop the test and offer it commercially through closely held Univfy Inc. of Los Altos, California.

About 7.3 million American women have infertility, according to the U.S. Centers for Disease Control and Prevention, based in Atlanta. The U.S. fertility industry generates more than $3 billion in revenue a year, according to the Center for Genetics & Society, a nonprofit public-affairs organization based in Berkley, California.

In vitro fertilization, or IVF, is a procedure in which a woman receives ovulation-stimulating drugs, and egg cells are removed from her and joined with sperm in a laboratory dish. The fertilized egg is placed in the woman's uterus to grow into a pregnancy. On average, most women have one to three cycles of IVF, Yao said.

Success Rates

Success rates for IVF vary, and age is the most important factor when women use their own eggs, according to the Society for Assisted Reproductive Technology, a professional group based in Birmingham, Alabama. About 31 percent of IVF cycles result in a live birth for women ages 35 to 37 who use their own fresh embryos, according to 2008 data from the society. That number drops to 22 percent for women ages 38 to 40.

Fertility treatment costs on average about $10,000 to $15,000 for each cycle or attempt at pregnancy, and isn't always covered by insurance, Yao said.

To develop their computer model in the first place, the researchers analyzed data from 1,676 IVF treatments performed at Stanford Hospital & Clinics from 2003 to 2006. The scientists identified 52 criteria, such as patient age, hormone levels and quality of a woman's eggs, that may help determine the chance of pregnancy, and developed a model that predicts a woman's likelihood of having a baby.

Yao said the new model can predict the chance of IVF success using information on 10 to 20 medical factors obtained for each woman.

Second Round

The predictions can only be made after the first round of IVF is done because the model needs data from that initial cycle.

Yao founded Univfy with a Stanford colleague and co-author, Wing Wong. The company would have to win approval for a test service from the U.S. Food and Drug Administration. Yao said she is licensing the test from the patent holder, Stanford.

Eventually Univfy may develop computer models to determine a woman's chance of getting pregnant with her first IVF treatment and the probability of having twins or multiple births, Yao said.

Dr. Cohen Comments

Matthew Cohen, director of reproductive endocrinology at North Shore-Long Island Jewish Health System, in Manhasset, New York, said while the analytical tool is interesting, it needs to be tested at clinics outside Stanford.

"It's just not ready for prime time use yet," he said in a telephone interview today.

Future studies also need to show if the results of the tests change a woman's decision to use IVF or how a doctor counsels patients, Cohen said.

"The fact that they went to the effort in showing that there are many more variables than age is teaching us something," Cohen said. "They're right. We need to look at other factors in helping to guide our patients."