Transcranial Doppler as a Screening Tool for High‐Risk Patent Foramen Ovale in Cryptogenic Stroke

ABSTRACT BACKGROUND AND PURPOSE The identification of high‐risk patent foramen ovale (PFO) is important for selecting suitable patients for PFO closure to prevent recurrent stroke in those with cryptogenic stroke. We aimed to evaluate the predictability of transcranial Doppler (TCD) in diagnosing high‐risk PFO compared with that of transesophageal echocardiography (TEE), which is not feasible for some stroke patients. METHODS We retrospectively reviewed the data of 461 cryptogenic stroke patients who underwent TEE and TCD for PFO evaluation. High‐risk PFO on TEE was defined as PFO with atrial septal aneurysm (phasic septal excursion ≥10 mm) or large PFO (≥2 mm). Spencer grading of right‐to‐left shunt was used to classify the amount of shunt on TCD. RESULTS PFO on TEE was observed for 242 (52.5%) patients, and high‐risk PFO was detected for 123 (26.7%) patients. However, PFO on TCD was observed for 336 (72.9%) patients. Among patients with significant shunt (Spencer grade III or higher) who underwent TCD after Valsalva maneuver (VM), 60.0% of patients had high‐risk PFO. However, only 5.3% of patients had high‐risk PFO among those without significant shunt. Receiver operating characteristic curves showed that significant shunt after VM had higher predictability (AUC = .876, 95% CI: .843‐.905) for detecting the high‐risk PFO compared with the predictability based on significant shunt at rest (AUC = .718, 95% CI: .674‐.759). ( P <.0001 for the differences between two AUCs). CONCLUSIONS TCD is a good screening tool for evaluating high‐risk PFO. VM is important for the evaluation of PFO. Patients with minimal or no shunt on TCD after VM are unlikely to have high‐risk PFO.