Comparison of transthoracic echocardiography with second harmonic imaging, transcranial Doppler and transesophageal echocardiography for the detection of patent foramen ovale

Purpose: To compare contrast transthoracic echocardiography (c-TTE) with second harmonic imaging (SHI), contrast transcranial Doppler (c-TCD) and contrast transesophageal echocardiography (c-TTE), for the detection of a patent foramen ovale (PFO). Materilas and Methods: We enrolled 56 patients (pts) (20 men), mean age 51±14 years, underwent to TEE with following indications: migraine (15/56, 27%), cryptogenic stroke (32/56, 58%), atrial fibrillation (AF) (9/56, 27%). Written informed consent was obtained before the examinations from all patients. Three cc of saline solution and one cc of air were rapidly mixed until a homogenous solution was obtained (contrast solution, CS). The solution was injected, into the antecubital vein, as bolus. The test was performed with Valsalva manoeuvre (MV) which lasted at least 10 s. VM was medially executed 5 sec after the injection of the CS. C-TCD was performed by a temporal window. The test was deemed positive if at least one "hit" was recorded on the TCD trace within 40 sec after the injection. The results were classified as follows: 0 hit: test negative, 1-10 hits: small shunt, >10 hits: medium shunt and >10 hits with "curtain" effect: large shunt. By c-TTE and c-TEE, a PFO was diagnosed if at least one micro-bubble (MB) was detected in the left atrium (LA) within three cycles following the appearance of the CS in the right atrium (RA). The shunt was defined small (<10 MB), medium (>10 MB) and large if all the LA was opaque. TEE was realized by an experienced echocardiographer who was blinded to the results of the two others examinations. Results: a PFO was detected in 29/56 pts (51,8%) with c-TEE, in 26/56 (46,4%) with c-TCD and only in 18/56 (32,1%) with c-TTE .With c-TTE, any small PFO was detected, and medium and large PFO were identified in the same number of patients 9/18 (50%). With c-TEE, PFO was small in 5/29 (17,24%), medium in 5/29 (17,24%) and large in 19/29 (65,5%). With c-TCD, PFO was small in 2/26 (7,7%), medium in 4/26 (15%) and large in 20/26 (77%). Agreement between c-TCD and c-TEE was very high (K=0.89) with good (89%) sensibility and negative predictive value (90%) and excellent specificity (100%) and positive predictive value (100%). On the contrary, agreement between c-TTE and c-TEE was lower (K=0.61), c-TTE being less sensible (62%), with negative predictive value 71%, even if very specific (specificity and predictive positive value: 100%). Conclusions: Our study has shown that c-TCD can accurately and reliably detect PFO, unless the shunt is very small.