Polygenic risk score is associated with high-risk plaque and coronary plaque burden in patients with non-obstructive coronary artery disease
E L Gaillard, C Y Y Beverloo, M Corver, S A J Chamuleau, C R Bezzina, L F Reeskamp, M Winter, S M Boekholdt, S J Jurgens, E S G Stroes, N S NurmohamedAbstract
Background
Cardiovascular disease is the leading cause of death, yet conventional risk tools remain limited. Coronary artery disease (CAD) results from a combination of lifestyle exposures and genetic predisposition. Polygenic risk scores (PRS) summarise CAD risk across a vast number of genetic variants into a single measure that is independent of conventional risk factors. Although PRS predicts future coronary events in general populations, their relationship with early atherosclerotic disease burden and high-risk plaque (HRP) on coronary computed tomography angiography (CCTA) is not established among asymptomatic individuals with non-obstructive CAD.
Purpose
This study aimed to determine whether a validated genome-wide CAD PRS (GPSmult) is independently associated with quantitative coronary plaque burden and HRP in asymptomatic individuals with non-obstructive CAD.
Methods
In this prospective study, patients with non-obstructive coronary atherosclerosis on CCTA underwent CAD PRS measurement. Quantitative plaque analysis was performed using a validated artificial intelligence-based algorithm (atherosclerosis imaging quantitative computed tomography [AI-QCT]). The association between CAD PRS and HRP prevalence was assessed using a logistic regression model adjusted for age, sex, and cardiovascular risk factors. The incremental predictive value of CAD PRS was evaluated across three sequential models: SCORE2 (Model 1), SCORE2 with extended risk factors (Model 2), and SCORE2 with extended risk factors and PRS (Model 3). Discrimination was assessed by the area under the receiver operating characteristic curve (AUC).
Results
A total of 312 patients (mean age 61 ± 6 years; 57% male) were included. Cardiovascular risk factors were similarly distributed across PRS groups. Patients with high CAD PRS had significantly greater plaque burden than those with low CAD PRS, including higher total plaque volume (99 vs 50 mm³), non-calcified plaque volume (70 vs 39 mm³), and calcified plaque volume (25 vs 7 mm³; all p<0.05; Figure 1). HRP prevalence increased across PRS tertiles: 44% (low), 46% (intermediate), and 64% (high; p=0.036). High CAD PRS was independently associated with HRP (OR 2.49, 95% CI 1.12 to 5.63; p=0.026). Discrimination for HRP improved progressively: Model 1 AUC 0.552 (95% CI 0.487 to 0.617), Model 2 AUC 0.637 (95% CI 0.574 to 0.700; p=0.03 vs Model 1), and Model 3 AUC 0.716 (95% CI 0.658 to 0.774; p=0.006 vs Model 2; Table 1).
Conclusion
In asymptomatic patients with non-obstructive CAD, a high PRS was independently associated with greater coronary plaque burden and HRP on CCTA. The integration of CAD PRS markedly improved the identification of high-risk individuals beyond clinical risk scoring. These findings suggest that PRS may serve as a tool to enhance risk assessment and guide treatment optimalisation in patients with subclinical atherosclerosis.