A patient brings you a report showing a “95th-percentile” coronary artery disease polygenic risk score (PRS) and asks whether they should start a statin. The useful question is not how impressive the percentile sounds; it is whether this particular score meaningfully changes the patient’s risk discussion and next decision.
A high PRS is a risk-enhancing signal—not a diagnosis, an absolute-risk estimate, or a prescription by itself. The 2026 U.S. dyslipidemia guideline lists high polygenic risk, if measured, among risk enhancers. For adults without ASCVD and with a borderline 10-year PREVENT-ASCVD risk estimate, considering risk enhancers is reasonable to personalize risk and the potential benefit of lipid-lowering therapy. The score still has to be interpreted beside the patient’s clinical risk and preferences.
A percentile ranks risk; it does not give the odds
A PRS summarizes inherited susceptibility associated with many common genetic variants. Its percentile describes where a person falls within the score’s reference distribution; it does not mean that a person has a 95% chance of developing coronary disease. Nor does a percentile by itself specify risk over 10 years or across a lifetime.
Absolute risk depends on the outcome and time horizon being estimated, as well as the person’s age and other clinical features. A similar relative increase in risk can translate to a small absolute difference for someone with low baseline risk and a larger one for someone whose baseline risk is already elevated.
The score’s reference group matters, too. Percentiles and performance can vary with the population used to develop or calibrate the score, the genetic ancestry represented, and the genotyping method. A result labeled “high” by one test is not automatically equivalent to the same percentile from another score—or to the same absolute risk in every patient.
| Report finding | What it can support | What it cannot establish alone |
|---|---|---|
| High CAD PRS | A signal to consider when discussing risk | A specific 10-year event probability or automatic statin indication |
| 95th-percentile result | A high rank in that score’s reference distribution | A 95% probability of disease |
| Low CAD PRS | Lower inherited susceptibility according to that score | Low overall risk or protection from other risk factors |
Check whether the score fits the patient and the decision
Before using a result, clarify what disease outcome the score predicts, how the score was developed and validated, what population its reference distribution represents, and whether the report provides calibrated absolute risk. A coronary artery disease score and a broader ASCVD risk estimate do not necessarily predict identical outcomes, so do not treat them as interchangeable.
Ancestry is relevant because a score may perform differently in populations unlike the group used to develop it. Genetic ancestry and social categories such as race or ethnicity are not interchangeable; neither a demographic label nor a vendor’s percentile guarantees that a score is well calibrated for an individual.
This is especially important when a patient arrives with a direct-to-consumer result. If the report does not make the model, intended use, or validation population clear, treat the number cautiously. Do not convert a relative score into a personalized absolute-risk figure by multiplying it into a clinical calculator unless a validated model explicitly supports that combination.
Use the PRS to inform—not replace—the prevention assessment
For primary-prevention lipid decisions in U.S. adults, start with the current guideline’s clinical risk assessment, including the PREVENT-ASCVD equations when applicable. Then review the patient’s actual risk factors, lipid results, premature ASCVD family history, and other relevant risk enhancers. The PRS may add context, but some genetic and clinical risk information can overlap; stacking numbers without a validated method can exaggerate precision.
A high, appropriately validated CAD PRS can strengthen a clinician–patient discussion, particularly when estimated risk is near a decision boundary. It does not automatically mean “start a statin,” and a low score should not overrule a clear clinical indication. The 2026 dyslipidemia guideline includes high polygenic risk, if measured, among risk enhancers; this is not a universal treatment threshold.
If the treatment decision remains uncertain, the 2026 guideline recommends CAC for further risk stratification in adults at intermediate 10-year risk and selected adults at borderline risk, generally men aged 40 or older and women aged 45 or older. CAC is not a reflex test required by a high PRS.
A markedly elevated LDL cholesterol or a strong pattern of premature disease in close relatives raises a different question: could this be familial hypercholesterolemia (FH)? A high PRS is not the same as a pathogenic FH variant, and it should not replace clinical FH assessment or indicated rare-variant testing. Conversely, a low PRS does not rule out FH or erase the implications of severe hypercholesterolemia.
Keep gene–environment interaction in its proper lane
A PRS estimates inherited susceptibility; it does not predict exactly how an individual’s smoking, blood pressure, diet, activity, or social environment will interact with that susceptibility. Environmental exposures and modifiable risk factors still shape disease risk, even though the inherited score itself does not change.
That distinction matters in counseling. A high score is not destiny, and a low score is not immunity. Address tobacco use, blood pressure, lipids, diabetes, and other established risks according to the patient’s overall clinical picture—not according to the genetic result alone.
Practical takeaways
- A PRS percentile is a rank within a reference distribution, not a probability of disease.
- Confirm the score’s outcome, validation population, and calibration before interpreting it.
- Use a validated clinical risk assessment and the patient’s full history first; treat a high CAD PRS as a risk enhancer, not a stand-alone statin order.
- If a lipid-treatment decision remains uncertain, consider guideline-appropriate CAC assessment rather than mechanically converting the PRS into absolute risk.
- Do not let a low PRS obscure severe LDL elevation, premature family history, or possible familial hypercholesterolemia.
