Need for Performance Protocols in TEE and TCD for Detection of Right to Left Shunts

• Published in: Journal of neuroimaging 2014-03, Vol.24 (2), p.144-148
• Main Authors: Chang, Jason, Darbonne, Carol, Drumm, Denise A., Teleb, Mohamed S., Frey, James L.
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Aged; Echocardiography; Echocardiography, Transesophageal - methods; Echocardiography, Transesophageal - standards; Female; Foramen Ovale, Patent - complications; Foramen Ovale, Patent - diagnostic imaging; Humans; Ischemic Attack, Transient - diagnostic imaging; Ischemic Attack, Transient - etiology; Male; Middle Aged; Needs Assessment; Neuroimaging; Observer Variation; paradoxical stroke; patent foramen ovale; Practice Guidelines as Topic; Reproducibility of Results; Sensitivity and Specificity; Stroke; transcranial Doppler; transesophageal echocardiogram; transthoracic echocardiogram; Ultrasonic imaging; Ultrasonography, Doppler, Transcranial - methods; Ultrasonography, Doppler, Transcranial - standards; United States; Young Adult
• ISSN: 1051-2284; 1552-6569
• DOI: 10.1111/j.1552-6569.2012.00729.x

ABSTRACT BACKGROUND AND PURPOSE Although transesophageal echocardiography (TEE) is the gold standard for right to left shunt detection, we observed that transcranial Doppler (TCD) was more sensitive and sought an explanation. METHODS We retrospectively evaluated results of TCD and TEE in 118 patients with cryptogenic stroke and transient ischemic attacks. TCDs were done as per modification of a published performance protocol and interpreted by a neurologist and radiologist. TEEs were performed and interpreted by five cardiologists without standardized protocol. Statistical methodology included χ2 tests, Fisher exact tests, and ANOVA. RESULTS Overall agreement between TCD and TEE was found for 76 of 118 patients. Sensitivities of TCD and TEE were 93.8% and 53.1%. Sensitivities for TCD interpreters were 61.1% and 64.1%. Sensitivities for TEE operators varied from 46.7% to 75.7%. We were unable to find a performance protocol for TEE microbubble studies in either published literature or certification organizations. CONCLUSIONS TEEs and TCDs are operator dependent and thus subject to false negatives. The lower yield and interoperator variability in TEE results appear to reflect the lack of performance protocols and engender concern about false negatives in community use. Consensus performance protocols and certification criteria for both modalities should have an impact on accuracy of shunt detection.