AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Age-Dependent White Matter Injury in Patent Foramen Ovale: Clinical Validation of a Literature-Trained Predictive Framework
By Roytberg G, Wasserlauf J, Zhelyakov E et al.
Background: Patent foramen ovale (PFO) has been associated with ischemic stroke and silent cerebral injury, but the clinical relevance and transportability of reported associations with cerebral white matter lesions (WMLs) remain uncertain. Because age-related cerebral small-vessel disease increases substantially with aging, younger populations may provide a more sensitive model for detecting PFO-associated cerebral effects. Objective: To evaluate age-dependent associations between transthoracic echocardiography (TTE)-detected PFO and MRI-defined WMLs and to determine whether literature-derived estimates of PFO-associated cerebral injury can be reproduced in routine clinical practice. Methods: We retrospectively analyzed 149 adults undergoing both TTE and brain MRI, including 37 patients with TTE-detected PFO and 112 controls. WMLs were defined as Fazekas grade ≥1. Age-stratified analyses were performed (<45 years, 45-60 years, and >60 years). Findings were compared with PAMAP (Practice-Anchored, Literature-Trained, Mode-Aware Predictive Framework), a transport model derived from published PFO-MRI studies and applied to the study cohort without modification. Results: Patients with PFO had a significantly higher prevalence of WMLs than controls (54% vs. 32%; OR 2.5, 95% CI 1.1-5.6, p=0.020). The largest relative difference was observed in adults <45 years, with WML prevalence of 40.0% in PFO-positive patients versus 5.9% in controls (p=0.018). Among patients aged 45-60 years, WML prevalence was 46.7% versus 28.3% (p=0.216), and among those >60 years, 90.0% versus 70.4% (p=0.393). Any neuroimaging abnormality was also more frequent in PFO-positive patients (68% vs. 34%, p<0.001). Although formal age-by-PFO interaction testing was not significant, the magnitude of the association progressively attenuated with increasing age. Literature-derived PAMAP predicted WML prevalence of 0.63 versus the observed 0.54, supporting successful transport of published PFO-WML associations to a real-world TTE-defined population. Conclusions: PFO is associated with an increased burden of cerebral WMLs in routine TTE clinical practice. The strongest relative association was observed in younger adults, a population in whom competing age-related small-vessel disease is less prevalent. Successful validation of literature-derived predictions in an independent clinical cohort supports the reproducibility of the PFO-WML relationship and highlights the potential value of integrating structural heart and neuroimaging phenotypes for individualized cerebrovascular risk assessment.