Patent Foramen Ovale on Transthoracic Echocardiography and Brain White-Matter Hyperintensities: A Transportability Analysis and Practice-Anchored Risk Framework

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Background: Meta-analyses suggest an association between patent foramen ovale (PFO) and white-matter hyperintensities (WMH), but pooled effect sizes do not clarify applicability to routine transthoracic echocardiography (TTE) or provide patient-level risk estimates. Objective: The objective of this study was to evaluate the association between TTE-detected PFO and MRI-defined WMH in routine practice and to develop a practice-anchored framework (PAMAP) that translates literature-derived evidence into individualized risk. Methods: We performed a retrospective, single-center, propensity-matched analysis of 149 adults undergoing TTE and brain MRI (37 PFO-positive, 112 controls). The primary endpoint was WMH (Fazekas ≥ 1). PAMAP synthesized 12 studies; 4 eligible studies were pooled using random-effects meta-analysis to derive a locked shunt coefficient (OR 3.65). The locked model used age and shunt (H); embolic context (E), and atrial stress (A) were neutral until refit. Transportability was assessed at the cohort level (expected vs. observed prevalence) and patient level (calibration, discrimination), followed by a minimal prespecified refit. Results: WMH was more frequent in PFO-positive versus control participants (54% vs. 32%). Literature-based expected prevalence approximated observed prevalence, supporting transportability. The locked model showed acceptable performance (calibration intercept 0.106; slope 0.912; Brier 0.188; AUC 0.756). A parsimonious refit improved performance (Brier 0.176; AUC 0.783), with the shunt term remaining significant (OR 2.45, 95% CI 1.23–4.88). Conclusions: PAMAP translates meta-analytic associations into a transportable patient-level risk framework. In routine TTE-defined PFO populations, the WMH association is preserved, suggesting that incidental PFO may mark early subclinical cerebral injury and enabling calibrated, individualized risk assessment.

This publication has 19 references indexed in Scilit: