Computed tomography and echocardiography together reveal more high-risk findings than echocardiography alone in the diagnostics of stroke etiology.

BACKGROUND: Cardioembolic stroke carries a major risk of stroke recurrence, which can be markedly reduced by early initiation of appropriate secondary prevention. We investigate whether combined examination of the heart, aorta, and cervicocranial arteries with computed tomography (CACC-CT) may improve the diagnosis of stroke etiology.
METHODS: Patients with suspected cardiogenic ischemic stroke or transient ischemic attack (n = 140; mean age 60 ± 10 years; 95 males) underwent CACC-CT and standard diagnostics including transthoracic and transesophageal echocardiography (TTE/TEE). Patients with atrial fibrillation were excluded because cardiac imaging will not affect to anticoagulant treatment. Imaging findings with a potential cardioembolic source were analyzed. Aortic and cardiac risk findings were evaluated independently. Consensus reading of 2 experts using the findings of both approaches and complemented by cardiac MRI when needed served as the reference standard.
RESULTS: In 101 patients (72%) the clinical diagnosis was stroke, and transient ischemic attack was confirmed in the remaining patients. Imaging findings associated with highly increased cardioembolic risk were detected in 22 patients (16%). Nine high-risk findings in 140 patients were found by TTE/TEE and this number rose to 25 high after performing both echocardiography and CACC-CT. No difference was found between CACC-CT and TTE/TEE in detecting patients with of at least one high-risk findings (sensitivity 68 vs. 41%, p = 0.052; specificity 98 vs. 99%; overall accuracy 94 vs. 90%). Combined use of CACC-CT and TTE/TEE was more sensitive than TTE/TEE alone for detecting patients with at least one cardiac or aortic high-risk finding (sensitivity 91 vs. 41%, p < 0.001; specificity 98 vs. 99%; overall accuracy 97 vs. 90%). TTE/TEE was insufficient for diagnosing myocardial infarction with left ventricular aneurysm, whereas the accuracy of CACC-CT was high. In 9 patients (6%) with normal or mild hypokinesia in TTE/TEE, CACC-CT and MRI showed myocardial infarction large enough to indicate anticoagulant therapy. In contrast, CACC-CT was not suitable for diagnosing small left artrial thrombi, patent foramen ovale or to measure left ventricular ejection fraction.
CONCLUSION: CACC-CT and TTE/TEE alone show limited accuracy for the diagnostics of stroke etiology. Therefore, CACC-CT could be a valuable tool in patients with cryptogenic stroke despite standard stroke diagnostics.