In some cases, even if chromosomally normal embryos are transferred, the woman still miscarries. This occurs because of two factors. One is that we cannot yet test for all of the twenty-four chromosomes in a single cell. But another is the potential for error in diagnosis. You do not want to discard a healthy embryo (incorrectly thinking that it’s abnormal), and you do not want to transfer back an abnormal embryo to the patient (incorrectly thinking that it is normal). Because of a potential error rate of as high as 5 percent in trying to determine the number of chromosomes from testing just a single cell with FISH, it is possible that an embryo that was classified as normal may in fact be abnormal. The risk in the other direction is that an embryo that is thought to be abnormal (and is therefore not transferred to the woman) may turn out to have been normal.

Possible causes of error are failure of the FISH probe to hybridize, or the possibility that two chromosomes could be so close together on the slide that they appear as just one signal. Thus, an abnormality diagnosed as monosomic (only one copy instead of two) could in truth be a normal embryo with the chromosomes too close together to differentiate. Whether a chromosome is considered one or two is measured almost arbitrarily by the distance between the two spots. If the distance is less than one chromosome in diameter, it is considered to be one chromosome even though it might be two, and if the distance is greater than one chromosome in diameter, then it is considered to be two.

Because of the potential for errors, one should not take the decision to do PGD lightly, nor should it be done routinely as a matter of course in <a href="https://glowing.com/glow-fertility-program">IVF</a>. Rather, it needs to be restricted to women who have a specific reason for avoiding miscarriage, or an intense fear or risk of having an aneuploid offspring such as a Down syndrome child. Women who simply fail to get pregnant, particularly older women, who have the poorest pregnancy rates with <a href="https://glowing.com/glow-fertility-program">IVF</a>, may choose to forego PGD. These women want to avoid even the smallest risk of not transferring a normal embryo that might have been misdiagnosed as abnormal.

The Europeans have tried to solve the error issue by testing two cells from the eight-cell embryo, rather than just one. Presumably this would reduce the risk of error because the second biopsied cell would be able to verify or refute the results on the first cell. But that approach increases the risk that a normal embryo that should be transferred will be misdiagnosed as abnormal. For example, if you had only normal embryos, and you assume a 5 percent error rate, 95 percent of those normal embryos would be diagnosed as normal. Five percent would be diagnosed as abnormal (even though, in truth, this 5 percent are normal). Then, if you biopsied a second cell, once again 5 percent of the 95 percent of normal embryos that were diagnosed correctly as normal with the first cell would now be diagnosed incorrectly as abnormal. Thus, only 90 percent of normal embryos would still have a diagnosis of normal based on evaluating two cells instead of just one. Therefore, 10 percent of normal embryos would be discarded unnecessarily because of conflicting results in the two biopsied cells.

However, the benefit of checking two cells is that your chance of transferring an abnormal embryo that was diagnosed as normal would be only 0.25 percent (5 percent of 5 percent). So by biopsying two cells, the chance of transferring an abnormal embryo becomes almost infinitesimal, but you also risk not transferring 10 percent of otherwise healthy embryos. Therefore, in the United States, most doctors would only biopsy one cell, and take the 5 percent risk of transferring a chromosomally abnormal embryo.

There is, however, another problem with screening for aneuploidy even if the FISH technique were error-free. Many normal-appearing embryos are mosaics, meaning some of the cells in the embryo are chromosomally normal and some of the cells are abnormal. Thus, you might remove a cell from the embryo that is chromosomally normal, even though the rest of the embryo is abnormal. Conversely, you might remove an abnormal cell, and the rest of the embryo might be normal. Nonetheless, despite this 5 to 10 percent potential for error, the miscarriage rate is reduced by PGD, even though the pregnancy rate may be reduced also.

Although PGD has been a real breakthrough for many couples with recurrent miscarriages, there are a smaller number of miscarriages that are not caused by chromosomal errors. I will discuss those in the next sections of this chapter.

--