Coronary Artery Disease: Hereditary Risk and Embryo Screening

When a parent has a heart attack in their fifties, or a sibling gets put on a statin in their forties, pretty much everyone in the family starts doing the quiet math about their kids. The honest answer to what they're asking is yes: a meaningful share of that risk is inherited. Twin studies put the heritability of coronary artery disease somewhere between 40% and 60%, with one Swedish registry following 20,966 twins for 36 years and landing at 0.57 for men and 0.38 for women.

But the inherited part of CAD is one of the most tractable in medicine. High polygenic risk for heart disease responds to lifestyle and statins more than high polygenic risk for almost anything else we screen for. And the genetic load driving most of that inherited risk is already visible at Day 5, before any embryo is transferred.

So the question isn't can I avoid passing this on. It's which levers are available at which points in life, and what the embryo-stage lever actually shows you.

How genetics drives heart disease

CAD is the leading cause of death in the United States. About 1 in 20 adults over 20 have it, and the CDC counted 919,032 cardiovascular deaths in 2023. Framingham data going back decades puts a 40-year-old man's lifetime risk of coronary heart disease at roughly 1 in 2. For women at 40, it's about 1 in 3. This isn't a rare disease. It's the background condition of American adulthood, and a meaningful fraction of who gets it comes down to which genetic variants you happened to inherit.

The genetic architecture splits cleanly into two categories. A small number of families carry a single high-effect mutation in one of three genes (LDLR, APOB, or PCSK9) that causes familial hypercholesterolemia. About 1 in 250 people have the heterozygous form of familial hypercholesterolemia, and untreated, men face a 50% chance of a coronary event by age 50 and women about 30% by 60. Roughly 1 in 10 people with an early heart attack turn out to have FH. For these families, preimplantation genetic testing for monogenic disorders (PGT-M) is the right tool, because the risk is driven by one known variant that can be tested for directly.

But most inherited CAD risk isn't monogenic. It's the sum of hundreds of common variants across the genome, each nudging risk up or down by a small amount. The latest CAD GWAS meta-analysis pooled over a million participants and identified 241 risk loci. Those common variants together explain about 36% of CAD heritability. No single one would show up on a monogenic test. Added together, they carry real predictive weight.

How much weight? Khera and colleagues built a genome-wide polygenic score in 2018 and found that 8% of the population ends up at greater than threefold CAD risk based on polygenic load alone. That's about 20 times more common than rare monogenic mutations conferring comparable risk. Most people whose hearts are going to give them trouble early aren't carrying a dramatic single-gene variant. They're sitting on the wrong end of a polygenic distribution.

One more piece worth naming: lipoprotein(a), or Lp(a). Blood levels of Lp(a) are over 90% heritable, driven mostly by two common variants in the LPA gene, and Lp(a) is independently causal for CAD. It also varies sharply by ancestry. Median Lp(a) concentrations are higher in people of African descent than in people of European descent, which is one of the specific reasons that polygenic scores calibrated on European data alone can't be exported to non-European families without recalibration.

Why hereditary doesn't mean destiny

CAD is unusual among the polygenic diseases we screen for: lifestyle works really well, even at the top of the genetic distribution.

In a 2016 NEJM paper, Khera's group looked at 55,685 people across four cohorts and stratified them by polygenic CAD risk. Among those in the highest genetic risk tier, a favorable lifestyle (not smoking, not obese, at least moderate exercise, and a reasonable diet) cut the 10-year coronary event rate from 10.7% to 5.1% in the ARIC cohort alone. Across all cohorts, the hazard ratio was 0.54. People born with the heaviest polygenic load on their cardiovascular system experienced events at roughly half the rate when they lived well.

This genuinely matters. If you're reading this because someone in your family had premature heart disease, and you're worried your kids will inherit the same risk, the genetic load you're worried about is unusually responsive to intervention. Statins, blood pressure control, and basic healthy habits change outcomes even at the high end of genetic risk. That's not true for every polygenic condition we screen for. It's a feature of cardiovascular biology specifically, which offers a lot of modifiable handles: cholesterol, blood pressure, inflammation, metabolic health.

So why think about embryo-stage screening at all?

Because the lifestyle and statin levers only get pulled after birth, on top of whatever genetic baseline a child starts with. The polygenic-load lever is available exactly once, before transfer. It doesn't replace adulthood prevention. It sets the starting point that adulthood prevention has to work from. For a family with premature CAD in their history, that starting point is one of the few things you actually get to choose between sibling embryos. And for the 8% of people whose polygenic load sits at monogenic-equivalent risk, that starting point matters a lot, because the adult work of managing cholesterol and blood pressure is harder and longer the heavier the inherited load.

What embryo screening can actually see

When parents go through IVF, they typically end up with several embryos at the blastocyst stage, around Day 5 or 6 of development. A small number of cells can be biopsied from each one without harming the embryo, and those cells carry the same DNA the future child would have. That DNA can be sequenced. From the sequence, we can compute a polygenic score for CAD and a range of other conditions.

Each embryo is a different draw from the same two parents. Which specific variants each one inherited was decided during meiosis (the shuffling of chromosomes when eggs and sperm form), and that shuffling produces embryos with meaningfully different polygenic risk profiles. Polygenic embryo screening measures those differences so that parents, working with their fertility team, have the information before deciding which embryo to transfer.

Embryos from the same IVF cycle are biologically siblings, which is why the right test for whether screening works is within-family validation rather than the population-level validation most of the field has historically leaned on. We validated 17 polygenic disease scores on sibling pairs, and 16 of them retained their predictive performance inside families, with within-family to population effect size ratios above 0.9. You can read more about why that test matters, but the short version is that it answers the question parents actually care about: does this score distinguish between my embryos, not just between strangers?

Two honest limits. Common variants capture roughly 36% of CAD heritability, not all of it. Rare variants, gene-environment interactions, and stochastic biology account for the rest. Screening shifts the distribution of likely outcomes across a family's embryos. It doesn't eliminate risk, and we don't frame it that way. We report absolute risk estimates for each embryo with their stated assumptions, so a family can see what the score says and what it doesn't.

CAD also shares genetic architecture with type 2 diabetes and other cardiometabolic conditions through shared lipid, insulin, and inflammation pathways. An embryo with a heavy polygenic burden for CAD often carries elevated risk for related metabolic conditions too. Part of what our counselors do when walking a family through results is surface those correlations, rather than letting each score sit in its own silo. If a family has a known pathogenic FH variant, the conversation starts with PGT-M, not PGT-P. What embryo screening is and what monogenic testing does are genuinely different tools answering different questions.

The last piece worth flagging is cross-ancestry calibration. Because Lp(a) and several other lipid traits vary substantially by ancestry, a CAD polygenic score calibrated only on European-ancestry reference data won't perform equally for a family of different background. We calibrate across 8 ancestry groups and report performance for each, rather than exporting a single score and hoping it generalizes.

If you're still weighing whether this makes sense for your situation, the decision guide walks through it with more depth. Most families find that the value depends on how many embryos they have, how strong the family history is, and which conditions they're most worried about.

If heart disease runs in your family and you're going through IVF, it's worth a conversation with someone who works with polygenic screening every day. You can talk with one of our counselors about what CAD screening would actually show for your embryos.