Gender test's accuracy is questioned

The baby shower invitations are blue. So are the three blankets she has crocheted. Her husband painted the closet blue, too.

But according to her latest ultrasound scan, Jennifer Beers, a Web developer in Greenland, N.H., is expecting a baby girl. And what began as a splurge -- a $275 blood test to determine the gender of her fetus at just 10 weeks, and that brought news of a boy -- has brought unexpected unpleasantness.

''I wanted a girl in the first place," Beers mused in an e-mail to the Globe. ''But I think you get so used to one idea, you plan and fantasize, just to have the rug taken out from you."

Judging by a small but rising chorus on the Internet, Beers has some confused and angry company among the thousands of pregnant women who have bought the Baby Gender Mentor, a new test that advertises that it is 99.9 percent accurate at determining the gender of a fetus as early as five weeks after conception.

Ultrasound scans, usually performed after 16 weeks of pregnancy to check the fetus medically, are notoriously inaccurate for determining sex, but some of the women posting their tales of gender woe say that the ultrasound seems to clearly contradict the Baby Gender Mentor test -- showing an obvious penis or the lack thereof.

Dr. Jeffrey Ecker, chair of the ethics committee of The American College of Obstetricians and Gynecologists, said that as far as he is aware, the college has not looked into the gender test. But in general, he said, ''Whether it's a manufacturer asserting a claim or a patient that's asserting a problem, people really need to have the data -- and as far as I can tell, no one has it yet."

All this will likely be resolved in a few months. The gender test began to be marketed in June, and by wintertime, the first big crop of babies whose gender it predicted will be squalling their way into the world.

For now, some of the women who sought early clarity are decrying the added anxiety they got instead and questioning the test's accuracy.

In response to a Globe query about problems, a dozen women e-mailed saying they were unhappy with the uncertainty created by the test. The query was posted on a website, In-Gender.com, that is gathering many of the complaints.

The women, from around the country, described the wrenching flip-flop from expectations of one gender to the other, made especially painful for some by the sudden dawning of fears that perhaps the gender confusion meant something was wrong with the baby.

The lab that performs the test, Acu-Gen Biolab of Lowell, offers a double-your-money-back guarantee if a birth certificate proves it wrong. But it has given only four refunds so far out of a total of more than 4,000 tests done since this summer, said its scientific director, Dr. C. N. Wang. That is precisely the number of false results one would expect with a 99.9 percent accuracy rate.

And the company that markets the Baby Gender Mentor test, Mommy's Thinkin', argues that the Internet hubbub is overblown.

''People should realize that statistically, ultrasounds are only 80 to 90 percent accurate," said its president, Sherry Bonelli. ''Thousands and thousands of Baby Gender Mentor tests are being done, so statistically that means there are going to be hundreds and hundreds of ultrasounds that are going to be conflicting, but that's just based on ultrasound being inaccurate at determining gender. Is that cause for panic? No, that's not cause for panic."

The gender test is based on fetal DNA research that has been published in peer-reviewed scientific journals, but the particular technique used by Acu-Gen has not been published and reviewed and did not need US Food and Drug Administration approval because it does not provide an actual diagnosis.

Wang said he expected to get approval for the technique's patent application soon and to seek publication and peer review of his work within months, but for now, the lab's accuracy claims remain unverified.

He noted that aside from the inaccuracy of ultrasound, there could be several other explanations for discrepancies: Genetic and hormonal defects can make a genetic boy appear to be a girl, or vice versa.

And sometimes, he said, a woman is initially carrying twins but one of them fails to develop, a phenomenon known as a ''vanishing twin." The gender test works by detecting fragments of fetal DNA, so it could pick up a boy twin's early presence, for example, but later appear to be wrong because the boy has since vanished, leaving only a girl.

Of the four refunds, Wang said, one involved a vanishing twin, one stemmed from a lack of sufficient blood to test, and two resulted from calibration errors on a lab instrument on a particular day.

The gender test, he emphasized, is merely a stepping-stone toward his plans for more serious, diagnostic uses for the lab's early detection technology, such as finding sex-linked genetic disorders and other genetic problems. And the results it has brought have highlighted the inevitable complexity of nature, he said.

''It's a learning and adaptation processes for all of us, including myself, including all the women out there," Wang said.

Carey Goldberg is reachable at goldberg@globe.com.