Ethnic heterogeneity in anatomical and functional characteristics of patent foramen ovale and its clinical associations in a multi-ethnic population in Yunnan, China

Background Patent foramen ovale (PFO) is associated with paradoxical embolism and headache. In multi-ethnic regions, it remains unclear whether ethnicity independently contributes to PFO-related stroke or headache. Methods In this hospital-based cross-sectional cohort, consecutive patients with confirmed PFO were enrolled at the our Hospital Medical University from January 2019 to December 2025 (n=488): Han (n=253), Yi (n=125), Bai (n=72), and Hui (n=38). Demographics, vascular risk factors, and laboratory markers including D-dimer (DD2) were collected. Transthoracic echocardiography and contrast TTE assessed anatomical phenotype and right-to-left shunt (RLS). The Valsalva manoeuvre was standardised (40 mmHg for >10 s), and RLS was defined within 3 cardiac cycles after right atrial opacification. Stepwise logistic regression evaluated associations between ethnicity and outcomes (stroke, headache), with sequential adjustment for demographic factors, vascular risk factors, and lipids. Results PFO phenotype differed by ethnicity. At rest, persistent left-to-right shunting was less frequent in Yi patients (32.80%) than in Han (47.83%) and Bai (55.56%) patients (P<0.05). Atrial septal aneurysm was more common in Hui patients (15.79%) but was not significant after multiple-comparison correction. Overall stroke prevalence was 26.84%. Ethnicity was associated with stroke in univariable analysis (P=0.018), but became non-significant after adjusting for hypertension (OR 3.099), smoking (OR 4.124) and DD2 (OR 8.068), suggesting mediation by unequal risk-factor burden. Headache cases were younger and predominantly female; the ethnic association disappeared after adjustment for age and sex (P=0.322). Headache was less often documented among stroke patients (OR 0.325), suggesting ascertainment bias. In a 1-year follow-up of 131 stroke patients, no recurrent stroke or major bleeding occurred in either closure or medical-therapy groups. Conclusions PFO anatomical and functional phenotypes vary across ethnic groups in Yunnan, but ethnicity is not an independent risk factor for PFO-related stroke or headache. Apparent ethnic differences in stroke largely reflect disparities in conventional risk factors, whereas headache differences are mainly confounded by demographics and possible ascertainment bias. Risk assessment should prioritise shunt severity, atrial septal aneurysm, and systemic risk factors rather than ethnic labels.