Identifying Risk-De-Escalating Markers in PREVENT-Defined Intermediate-Risk Older Adults: Insights From ASPREE

BACKGROUND: Previous studies showed that the Pooled Cohort Equations substantially overestimate atherosclerotic cardiovascular disease (ASCVD) risk in older adults, and the recently published PREVENT equation offers better performance. Identifying markers that can further refine risk estimates is essential for personalized prevention in this age group. This study evaluates the clinical utility of 6 markers for reclassifying intermediate-risk older adults (10-year PREVENT-estimated ASCVD risk, 7.5% to 20%), with a focus on de-escalation. METHODS: This post hoc analysis evaluated intermediate-risk older adults aged ≥70 years using data from ASPREE (Aspirin in Reducing Events in the Elderly; URL: https://www.clinicaltrials.gov ; Unique identifier: NCT01038583). ASPREE was a randomized trial of low-dose aspirin versus placebo in healthy older adults who were free of prior cardiovascular events at enrollment. Six markers were evaluated: the lowest quartile (Q1) of NT-proBNP (N-terminal pro-B-type natriuretic peptide), hs-TnI (high-sensitivity troponin I), and hs-CRP (high-sensitivity C-reactive protein), polygenic risk for lipoprotein(a) and polygenic risk for coronary artery disease, and absence of family history of ASCVD. Incident ASCVD events were adjudicated by expert panels. Statistical performance of each marker was assessed using diagnostic likelihood ratios, change in Harrell C statistic (ΔC), and net reclassification index relative to a baseline model incorporating variables included in the PREVENT equation. RESULTS: The study included 7764 participants (48% female; median age, 74 years [interquartile range, 72–77]), with a median follow-up of 10.3 years. During follow-up, 725 participants (9.3%) experienced ASCVD events. Q1 polygenic risk for coronary artery disease was the most powerful marker for down-grading risk, providing 37% risk reduction (diagnostic likelihood ratio: 0.627) and improving discrimination (ΔC: +1.44%; P =0.004) when compared with the baseline model, and correctly reclassifying 25.9% of nonevents downward (net reclassification index: 0.101). Q1 hs-TnI showed second-best performance (diagnostic likelihood ratio: 0.727; ΔC: +0.71%, net reclassification index: 0.075). CONCLUSIONS: Adding polygenic risk for coronary artery disease to the PREVENT equations may further personalize risk estimates and support clinical decision-making for older adults at intermediate risk for ASCVD events.