Iribarren et al1 recently reported that the risk associated with low-density lipoprotein cholesterol (LDL-C) is amplified by a polygenic risk score (PRS) for coronary artery disease, although formal tests for LDL-C∗PRS interaction were not statistically significant. Their analysis used a 12-variant coronary artery disease PRS. Joint PRS and LDL-C effects were estimated from a Cox model using low PRS and LDL-C <100 mg/dL as the reference group, so all HRs represent risk relative to that lowest-risk combination. This appears to contrast with a previous study2 reporting a significant LDL-C∗PRS interaction.
In their discussion, Iribarren et al1 compared their findings with those of Bolli et al2, stating that “unlike Bolli et al, the adjusted risk of high polygenic risk and LDL 130 to 160 mg/dL (HR: 2.1) was commensurate with intermediate polygenic risk and LDL 160 to 190 mg/dL (HR: 2.1), whereas in their study, it was commensurate with intermediate polygenic risk and LDL ≥190 mg/dL.” This comparison implicitly assumes that both studies used the same analytical reference framework.
However, this comparison is not valid because the 2 studies rely on different reference groups, making the reported risk gradients not directly comparable. In Bolli et al, the HR for LDL-C 130 to 160 mg/dL was calculated relative to individuals with optimal LDL-C within the high-PRS group. It was not calculated relative to low PRS with low LDL-C, which is the reference category used by Iribarren et al.
When the estimates from Bolli et al2 are recalculated under the analytical framework used by Iribarren et al1 (low PRS and LDL-C <100 mg/dL as the reference), the underlying effect sizes become fully comparable. Under this unified framework, the HR for high PRS with LDL-C 130 to 160 mg/dL in Bolli et al rises to 4.84 (95% CI: 2.89-8.10), considerably higher than the HR of 2.1 reported by Iribarren et al for the same LDL-C strata. Likewise, in Iribarren et al, the combination of high PRS and LDL-C ≥190 mg/dL carries an adjusted HR of approximately 3.5, whereas with the genome-wide PRS used in Bolli et al, the analogous category reaches an HR of 9.17 (95% CI: 5.42-15.50) (Figure 1).
Figure 1.
Coronary Artery Disease HRs Across Polygenic Risk Score and LDL Cholesterol Levels
HRs for coronary artery disease by polygenic risk score group and LDL cholesterol level using a genome-wide polygenic risk score (Allelica), with low polygenic risk score and LDL cholesterol <100 mg/dL as the reference group. CAD = coronary artery disease; LDL-C = low-density lipoprotein cholesterol; PRS = polygenic risk score.
These differences do not indicate biological inconsistency; instead, they reflect the substantially greater predictive performance of a genome-wide PRS compared with a 12-variant score. When risk estimates are expressed using the same reference group, the genome-wide PRS yields a far steeper and more clinically meaningful gradient of risk across LDL-C categories, including among individuals with moderate LDL-C elevations.
Footnotes
Dr Bottà is the CEO and Co-Founder of Allelica Inc. Dr Di Domenico is the CTO and Co-Founder of Allelica Inc.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.
References
- 1.Iribarren C., Lu M., Elosua R., Gulati M., Wong N.D., Rana J.S. Joint consideration of LDL-C and polygenic risk for coronary heart disease risk assessment. JACC Adv. 2025;4(11_Part_1) doi: 10.1016/j.jacadv.2025.102228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bolli A., Di Domenico P., Pastorino R., Busby G.B., Bottà G. Risk of coronary artery disease conferred by low-density lipoprotein cholesterol depends on polygenic background. Circulation. 2021;143(14):1452–1454. doi: 10.1161/CIRCULATIONAHA.120.051843. [DOI] [PMC free article] [PubMed] [Google Scholar]