Sir

Your News Feature “The first cut” (Nature 445, 479–480; 2007) highlights a controversy about preimplantation genetic diagnosis (PGD), a procedure to test for genetic disease that is now more widely used for screening chromosomes in embryos after in vitro fertilization (IVF). Randomized clinical trials are needed to determine whether aneuploidy screening improves pregnancy rates, or whether embryo biopsy has any impact on the health of future children and adults. There is less doubt, however, that PGD encourages the waste of healthy embryos. Your News Feature pointed out that this can potentially happen if mosaic embryos normalize in culture, but a significant fraction may occur from misdiagnosis.

Chromosomes are identified by spots of colour from fluorescent in situ hybridization (FISH), which has 90–92% accuracy, by some estimates. Is this good enough? At present, about half the karyotype is usually screened: nine autosomes plus two sex chromosomes. But, for every additional chromosome examined, the probability of a false result increases until, when all are screened, more than half of all embryos might be rejected, even if the accuracy is as high as 96% (0.9624). If this prospect is not discouraging enough, then consider the FISH data themselves, even from the best screening centres.

During cell division, chromosome pairs normally segregate equally at opposite poles. When separation is premature or fails to happen, we expect to find an equal number of cells with an extra or a missing chromosome (monosomy), which is borne out by studies of human oocytes (see F. Pellestor et al. Hum. Genet. 112, 195–203; 2003). However, FISH data consistently show a significant bias and a two- to threefold excess of monosomy compared to trisomy. A higher error rate is anticipated for monosomy as hybridization can fail for many reasons, whereas an artefactual third signal is less likely. More perplexingly, however, the incidence of aneuploidy for each chromosome is so high that one wonders whether any embryo will be judged normal when the entire karyotype can be screened by this method.

Until a more reliable technology emerges, it is likely that many embryos will be needlessly discarded. This is perverse, as the goal is to improve pregnancy rates by selecting the best-quality embryos for transfer, particularly for older patients. The high incidence of aneuploidy in our species has encouraged this application of PGD, but best interests may not be served by liberal application and literal interpretation of current FISH technology. Sometimes a patient has no diploid embryos, causing her IVF cycle to be abandoned. Perhaps she should have any embryos with monosomy transferred, unless they are XO. At worst they will be nonviable, but at best they may produce a healthy child.