Our gentle descent toward eugenics.

We've just taken another step down the slippery slope toward eugenics.

The step involves "preimplantation genetic diagnosis," in which clinics take sperm and eggs, make embryos in lab dishes, and screen them for genetic flaws. Embryos without flaws are implanted in the mother's womb. Those with flaws are frozen or discarded.

The Human Fertilization and Embryology Authority  regulates PGD in Britain. Previously, it had approved PGD only to weed out genes that were nearly certain to cause a grave childhood disease or were certain to cause a grave adult disease. Last week, the HFEA stepped across that line. You can now chuck embryos in Britain for diseases that are more treatable, less likely to strike early in life, and less likely ever to occur in the person whom the embryo would become.

It's hard to argue with this step. It will spare many families a lot of suffering. But so will the next step down the slope, and the step after that. And there's no sign of a foothold ahead that will brake our slide.

The slide can be measured three ways. The first is "penetrance"—the probability that a gene will lead to disease. For breast and colon cancer, the HFEA has lowered its standard just a bit, from a 90 to an 80 percent probability. But the agency also cites the roles of the same genes in uterine, ovarian, and other cancers, where the probabilities range from 60 percent down to 30. The HFEA's report on the PGD decision, released this week, quotes the opinion of a "professional society" that "a 30-80% risk of developing the condition is significant enough" to justify PGD.

How low can we go? According to an internal HFEA briefing paper, "there are likely to be susceptibility genes identified that increase the chance of developing a condition by a very small amount (e.g. 1 or 2%)." The paper says such low-probability genes should be distinguished from high-probability genes, but it offers no logical place to draw a line. A bioethicist and an IVF expert quoted in the HFEA's report argue that if a disease is bad enough, embryos should be discardable for genes with even minimal probabilities of causing that disease.

The next sliding standard is treatability. The old rule of thumb was that you could screen for untreatable diseases. The rule toward which we're evolving is that you can screen for treatable diseases if the treatments are awful or unreliable. A disease that "required regular invasive treatment" would be serious enough to warrant PGD, says the HFEA. According to the agency, "There are treatments available for inherited cancer susceptibility conditions, however, survival rates are variable and there is still a risk that treatment will fail." Risk of failure, not certainty of failure, is the emerging justification.

The third sliding standard is age of onset. As the agency points out, "The conditions that have been licensed for PGD previously are serious conditions that are usually present in the child when it is born." Now, however, PGD will be allowed for inherited colon cancer, which is diagnosed "usually in the forties," and for inherited breast cancer, which tends to appear in the "late thirties and forties." It's worth noting that the disease probabilities the HFEA attributes to these genes are lifetime probabilities, which means the risks of the genes causing disease before age 40 or 50 are lower than advertised.

By today's standards, cancer in your 40s is quite young. But 100 years ago, life expectancy for a child born in the United States was less than 50 years. It's easy to see how our definition of early-onset illness loosens as our expectations rise. Today, with life expectancy approaching 80, the HFEA asks whether PGD should be used to screen out embryos with genes for diseases that "will not develop until the person is in their seventies or eighties."

Put the latter two factors together, and you have an additional problem. Treatability is dynamic. As the HFEA's internal document points out, "it is likely that in thirty years," when the genes in question begin to cause disease, "there will be more drugs available to treat cancers including inherited breast, ovarian and bowel cancers." That means we're chucking embryos today to avoid the possibility—not the known probability—that tomorrow, the diseases they get, if they get them, will still be too hard to treat.