We thank Mojadidi et al . for their interest in our recent work and for their thought-provoking comments. In their letter, Mojadidi et al . emphasize the importance of additional techniques to maximize the detection rate of patent foramen ovale (PFO). We agree that techniques such as nitroglycerin administration for left ventricular unloading, repeated contrast injections, and use of saline with blood as a contrast agent are useful for this purpose, as well as the inferior vena cava (IVC) compression maneuver that we proposed. However, we have two concerns regarding the letter.

First, the sensitivity (89%) and specificity (91%) of transesophageal echocardiography (TEE) for the detection of PFO, which are based on a meta-analysis by Mojadidi et al ., appear to be underestimated. Of the studies included in the analysis, crucial drawbacks exist in two studies in particular, which showed excessively low sensitivity or specificity. Although one of these studies showed very low specificity (33%) of TEE for PFO detection, patients enrolled in the study were deemed eligible for transcatheter PFO closure. The specificity calculated using the small number of referential PFO-negative patients ( n = 3) might be statistically unreliable in this study. The other report by Augoustides et al . revealed low sensitivity (67%) of TEE for the detection of PFO in adult cardiac surgery patients. Most of the enrolled patients (86.1%) had severe left-sided heart disease, such as coronary heart disease and valvular heart disease, or had received ventricular assist devices or undergone heart transplantation. Identifying a PFO depends on creating a situation in which pressure in the right atrium exceeds pressure in the left atrium; when severe left ventricular dysfunction is present, left atrial pressure is usually elevated, and this might be expected to make it more difficult to demonstrate a PFO even though one is present. If these two studies were excluded from the analysis, the sensitivity and specificity of TEE for PFO detection would be higher than reported. Moreover, the IVC compression maneuver we developed is feasible and effective for detecting PFO and is comparable with the Valsalva maneuver. Unlike the Valsalva maneuver, the IVC compression maneuver is essentially a subject-independent procedure and can be performed in patients with cognitive disorders (e.g., patients with paradoxical cerebral embolism or those under deep sedation or general anesthesia) who cannot effectively perform the Valsalva maneuver. In addition, IVC compression can be performed under deep sedation in patients who cannot tolerate transesophageal probe stimulation or adequately perform the Valsalva maneuver under conscious sedation. We showed that use of IVC compression resulted in the detection of 22 new cases of PFO (8.7%) in 252 patients with Valsalva-negative results. Thus, use of this technique is expected to markedly increase the diagnostic accuracy of TEE for PFO detection.

Our second concern is that Mojadidi et al . misinterpret transcranial Doppler (TCD) as superior to TEE for PFO detection. In fact, the cited source notes that the diagnostic capability of TCD for the detection of PFO was comparable, but not significantly superior, to that of TEE. Furthermore, TCD cannot be used to distinguish an extracardiac shunt from PFO or atrial septal defect or to provide anatomic information regarding the location of the shunt or associated abnormalities, as Mojadidi et al . mention in their letter. In our transesophageal echocardiographic study, three patients with atrial septal defects were excluded from the 293 enrolled patients. Therefore, we do emphasize that TEE is considered the recommended method for detecting PFO with high sensitivity and specificity, as previously mentioned, and that TCD is an alternative or complemental technique.