Background: In the pursuit of stroke etiology, cardiac imaging with bubble study often reveals an intracardiac shunt. We sought to determine the yield of bilateral lower extremity ultrasound (BLE US) and magnetic resonance venography (MRV) of the pelvis in detection of deep vein thrombosis (DVT) as a source for paradoxical embolization in patients with ischemic stroke who have an intracardiac shunt.
Methods: Patients admitted with AIS from 2008 through 2015 were included if the presence of intracardiac shunt was confirmed via echocardiography with bubble study. We examined the frequency of DVT (DVT+) among patients who had BLE US only, MRV pelvis only, or both BLE US and MRV. We then compared patients with evidence of DVT (DVT+) to those without evidence of DVT (DVT-).
Results: Among 1,992 patients with AIS, 145 patients had confirmed evidence of an intracardiac shunt; of these, 40 patients (27.6%) had neither BLE US nor MRV studies, 29 patients (20.0%) had BLE US only, 4 patients (2.8%) had MRV only, and 72 patients (49.7%) had both BLE US and MRV. Among those with BLE US only, 2 patients (6.9%) were DVT+; of those who had MRV only, 2 patients (50.0%) were DVT+; of those who had both BLE US and MRV, 11 (15.3%) were DVT+. Overall, 14.3% of patients with an intracardiac shunt and work-up for DVT were found to have an occult DVT. No statistically significant differences were detected in demographic, baseline, in-hospital, or discharge characteristics when comparing DVT+ and DVT- patients.
Discussion: The findings of this study support the use of BLE US and MRV to detect occult DVT among ischemic stroke patients with intracardiac shunt. In our sample population, 14.3% of patients were positive for occult DVT though no patient characteristics were associated with predicting the presence of DVT. This supports the need for adequate DVT screening in patients with AIS and positive bubble study.