Table 5:

Logistic regression model of factors associated with patent foramen ovale*

CharacteristicNo. (%) of patients with no PFO
n = 1206
No. (%) of patients with PFO
n = 66
Unadjusted OR (95% CI)Adjusted OR (95% CI)
Atrial fibrillation298 (24.7)9 (13.6)0.48 (0.22–0.94)NA
Sex, male634 (52.6)36 (54.5)1.08 (0.66–1.79)0.91 (0.54–1.53)
Age > 60 yr977 (81.0)37 (56.1)0.30 (0.18–0.50)0.34 (0.20–0.57)
Hypertension790 (65.5)34 (51.5)0.56 (0.34–0.92)0.77 (0.45–1.30)
Dyslipidemia275 (22.8)6 (9.1)0.34 (0.13–0.73)0.39 (0.15–0.84)
Diabetes mellitus327 (27.1)13 (19.7)0.66 (0.34–1.19)0.77 (0.39–1.42)
Prior stroke or TIA239 (19.8)4 (6.1)0.26 (0.08–0.64)0.31 (0.09–0.76)
Heart failure71 (5.9)2 (3.0)0.50 (0.08–1.64)NA
Coronary artery disease124 (10.3)4 (6.1)0.56 (0.17–1.39)NA
  • Note: CI = confidence interval, NA = not applicable, OR = odds ratio, PFO = patent foramen ovale, TIA = transient ischemic attack.

  • * We did not include atrial fibrillation, heart failure or coronary artery disease in our final model to avoid model overfitting and based on prior literature of risk factors for PFO.