A score of non-contrast transthoracic echocardiography to detect patent foramen ovale

Background: The aim of this study was to develop a screening score of non-contrast transthoracic echocardiography (TTE) for patent foramen ovale (PFO) in patients with embolic stroke of undetermined source (ESUS). Methods: We performed a retrospective analysis of 218 consecutive patients with a recent ESUS from 2015 to 2018, who received TTE and transcranial Doppler (TCD) as routine examinations. PFO was diagnosed by the bubble test of TCD. Signicant differences of the TTE ndings between PFO group and non-PFO group were selected into a score. Results: PFO was diagnosed in 35.8% (78/218) of the patients. Compared with non-PFO group, a larger median aortic root diameter (ARd) (34mm vs. 32mm, p=0.005), a lower median peak E wave velocity (Em) (61.5cm/s vs. 68cm/s, p=0.005) and a lower incidence rate of mitral regurgitation (33.8% vs. 50.7%, p=0.016) were seen in PFO group. A three-point score of TTE criteria (including aortic root diameter>33mm, Em<72cm/s and without mitral regurgitation, AEM) was tested to be independently predict PFO (odd risk 1.95, 95%CI 1.37~2.78, p<0.001). AEM score ≥ 2 detected PFO with a sensitivity of 0.71 (95% CI 0.67(cid:0)0.85) and a specicity of 0.53 (95% CI 0.44 -0.55). Conclusion: The AEM score measured with non-contrast TTE can select ESUS patients for bubble test of TCD to detect PFO.


Introduction
Embolic stroke of undetermined source (ESUS), which accounts for approximately 25% of cases and patent foramen ovale (PFO) is considered to be one of the major causes of ESUS 1 . Ultrasonographic assessment of PFO, using bubble test with transthoracic echocardiograph (TTE), transesophageal echocardiography (TEE) and transcranial Doppler (TCD) remains the diagnostic approach of choice 2 .
Although the prevalence of cerebrovascular complications after bubble test were reported very low, stroke risk from paradoxical microbubble embolization can be clinically signi cant and cannot be guaranteed. 3 A safe and useful PFO-screening tool before contrast ultrasonographic assessment is important for clinical practice. In this study, we aimed to develop an easily measured screening tool by using non- Bubble test PFO was diagnosed by the bubble test. 5 9 mL isotonic saline solution, 1 mL of air, and 1 drop of the patient's blood were mixed through two 10 mL syringes connected by a three-way stopcock. The mixture was rapidly injected into the forearm vein during normal respiration and during the maintenance stage of Valsalva maneuver. Right-to-left shunt (RLS) was quanti ed by counting the number of microbubbles (MBs) within the rst 3 cardiac cycles. According to the number of MBs, patients were divided into two groups: PFO group, 1 or more MBs; non-PFO group, no MBs. 6 TTE Parameters of TTE were obtained by retrospectively reviewing the reports. TTE were performed in the left lateral decubitus position using standard imaging planes according to the American Society of Echocardiography recommendations during hospitalization. The aorta root diameters (ARd) and left atria anteroposterior diameters (LAAPd), end-diastolic left ventricle diameters (EDLVd), end-systolic left ventricle diameters (ESLVd), left ventricle ejection fraction (LVEF), peak E-wave velocity (Em) and peak Awave velocity (Am) were collected.

Statistical Analysis
Statistical analyses were performed using SPSS, version 22 (SPSS Inc, Chicago, IL). P value less than 0.05 was considered to indicate statistical signi cance. Continuous variables were compared by Mann-Whitney U test and categorical variables were compared by Chi-square or Fisher's exact test between PFO and non-PFO groups. Multivariate regression (including variables with p < 0.1) was used to access the association of variables with PFO. Receiver operating characteristic (ROC) analysis was performed to determine the optimal threshold of independent continuous variables in predicting PFO. Base on the previous results, a 3-point score, including 3 independent criteria, was tested by ROC for predicting PFO.
This study has been approved by the ethics committee of our institution.    Table 3. AEM ≥ 2 predicted PFO with a sensitivity of 0.71 and a speci city of 0.53, which was tested to be the best threshold to predict PFO with Youden index.

Discussion
We were able to develop a simple score to predict the detection of PFO in ESUS patients. The AEM score measured with non-contrast TTE could be used to identify high-risk patients for PFO and reduce the use of bubble test. This allows the clinician to easily screen patients at highest risk of PFO for further con rming exams. Moreover, our data addressed the echocardiographic changes of heart structures and functions in PFO patients, which could be useful to explore the potential mechanism of ESUS caused by PFO.
Dilatation of the aortic root may increase the risk of RLS by changing the angulation of the heart in such a way that ow streaming from the inferior vena cava into the right atrium is directed more towards the ostium secundum; thrombotic material is therefore more likely to cross into the systemic circulation, possibly causing a cryptogenic stroke. It has been reported that ARd, marked at the level of the sinuses of Valsalva (34 ± 4 vs 31 ± 3 mm, p < 0.01), is larger in PFO patients with cryptogenic stroke than in healthy people. 7 In this study, we compared the ARd in a ESUS patient cohort. All subjects had a homologous pro le and our results consisted with previous study, demonstrated a larger median ARd of 34 mm in PFO patients than those without PFO.
E-wave velocity re ects the left atrial (LA)-left ventricle (LV) pressure gradient during early diastole and is affected by alterations in the rate of LV relaxation and LA. Elevated LA pressure is associated with the absence of RLS in AF stroke patients and may prevent opening of a PFO. 8 In our study, patients with PFO had lower Em suggest a decreased LA pressure, may associated with a RLS-related stroke.
Mitral regurgitation is the most common valvular heart disorder in high-income countries, and its prevalence increases with age. 9 In a large-scale cohort of UK adults with 10 years of follow-up, elevated blood presure was continuously associated with an increased risk of mitral regurgitation. 10 Considering that the association between high blood pressure and mitral regurgitation is similar with that of arteriosclerosis stroke. ESUS patients with no mitral regurgitation may have fewer atherosclerotic risk factors, however, higher likelihood of PFO.
Several limitations of the present study need to be underlined. First, it was a single-center retrospective cohort study with a small sample size. Second, PFO was only diagnosed by the bubble test of TCD, not TEE. Although TEE was considered to be the standard technique for identifying a PFO, some patients were intolerant of this method. Finally, all patients in our study had a stroke. Thus, the differences between those with and without stroke in PFO were not examined, which needs to be addressed in further studies.

Conclusions
AEM is a risk score based solely on non-contrast TTE parameters that can easily be used to select patients for bubble test to increase the diagnostic yield of PFO after ESUS and might improve the secondary preventive strategy in order to prevent recurrent ischemic strokes.