Why I considered IVF despite not having any fertility issues, and then decided against it

In The dangers of reproducing while old, I mentioned that pre-implantation genetic testing could be a way for older prospective parents to improve outcomes. This led my partner and I to seriously consider doing IVF despite not having any fertility issues. At a high level, my conclusion after writing that post was:

After investigating this more, I now think:

Polygenic screening

I think polygenic screening1 has a lot of potential, but currently it has serious limitations, including:

  1. The models for certain traits like intelligence just aren’t very good at predicting. This is called the “missing heritability” problem, and it’s quite controversial exactly what’s going on, but some of the issues are clear. One is that current models are based on data that measured other things, like educational attainment, which are correlated with intelligence, but not perfectly (e.g. Deary et. al find 0.81). And it’s partly because some of the variance in traits we care about is probably caused by very rare variants which would require huge sample sizes to detect.
  2. The models are much better for people of certain ancestry than others, because of the data that they were created with. My impression is that the models currently work best for people of northern European ancestry.
  3. The expected variance amongst embryos from the same parents is pretty low. To have good odds of finding an embryo that has polygenic scores much better than the average of its genetic siblings (i.e. the other embryos parents will be choosing amongst2), one needs a lot of embryos to choose from.

Point 1 should decrease the appeal of polygenic screening for everyone. Point 2 decreased it for me because my children would be mostly not-northern European. As for the number of embryos, my partner took a test that suggested a round of egg retrieval would yield relatively few eggs. This meant she’d probably have to undergo multiple rounds of egg retrieval, which is expensive and unpleasant.

It’s hard to find good information on this topic. It’s not hard to find people saying that the technology doesn’t work, but I get the distinct whiff of motivated reasoning from articles like this one by Lior Pachter. Basically I think the author thinks polygenic screening is morally wrong or disgusting, and therefore he’s finding reasons to say it won’t work.

I want polygenic screening to work. If we were doing IVF anyways, I would definitely have the embryos polygenically screened.

IVF health outcomes

I found the excellent paper The health of children conceived by ART: ‘the chicken or the egg?’, which looks at studies that try to control for the systematic differences between people who pursue IVF and those who don’t. That review, and newer studies such as Sutclifffe (2023) did not find large differences in IVF babies later in life. But the best-controlled studies do find a pretty large increased risk of preterm birth (relative risk somewhere in 1.5 - 2). Elsewhere I found the absolute risk of preterm birth for a 36 year old mother is about 6%, so IVF might take that up to 9-12%. Since preterm birth is associated with a lot of bad health conditions, it would be surprising if IVF children are more likely to be born preterm but are equally healthy later on.

I’m wondering how it’s possible to reconcile the preterm birth and generally good outcomes in adults. Some possibilities:

My current guess is there’s a small but real tendency for people conceived via IVF to be less healthy later in life, though I’m extremely uncertain about the exact aspects of health, the magnitude and the frequency involved.

  1. If you are not familiar with polygenic screening, I recommend Gene Smith’s post 

  2. Unless you’re using donor gametes. In which case you might use multiple donors and compare across non-siblings.