Does patent foramen ovale presence procure favourable outcomes in patients with severe rheumatic mitral stenosis?

Mitral stenosis (MS) is tolerated for an extended period in patients with atrial septal defect (ASD) known as Lutembacher syndrome due to depressurizing effect. In a similar way, patients with patent foramen ovale (PFO) may have clinical benefits in severe MS. We aim to evaluate the clinical effects of PFO in rheumatic MS. Transthoracic and transesophageal echocardiography records of the patients with symptomatic severe MS were screened for the period between 2008 and 2019 in a single center. 320 symptomatic patients with severe MS were included and presence of PFO recorded. Left atrial appendix (LAA) thrombotic status was defined as clear, spontaneous echo contrast, and thrombus. Two different statistical models were used to determine the predictors of either smallest (mitral valve area) MVA at symptomatic presentation or more thrombogenic LAA. 34 patients had PFO. Multivariable ordinary least square model demonstrated that increase in systolic pulmonary arterial pressure, ejection fraction and presence of PFO were associated with smaller MVA on presentation. Multivariable proportional odds logistic regression model demonstrated that advanced age, increased left atrial diameter, absence of PFO were associated with more thrombotic status whereas larger MVA was associated with decreased thrombotic status in LAA. Presence of PFO in severe MS results in two clinical benefits as (i) being asymptomatic with smaller MVA and (ii) having less LAA thrombosis probably caused by depressurizing effect on the left atrial pressure. Our study could serve as an example for patient groups with expected symptomatic benefits from left atrium pressure offloading interventions.


Introduction
Rheumatic mitral valve stenosis (MS) pertains its place as acquired form of valvular pathology and still results in morbidity and mortality. Increased left atrial pressure results in delay in diastolic emptying, decrease in cardiac output [1], increase in left atrial pressure, pulmonary venous hypertension and superimposed arterial hypertension [2] play a role in pathophysiological events which cause patients to be symptomatic. Lutembacher described the Lutembacher syndrome in 1916 referred to his own name, which consists of MS accompanied by an atrial septal defect (ASD) [3]. In general, it is known that the MS is more benign, and well tolerated for a longer time period when ASD is present concomitantly [3].
Reduced left atrial appendage (LAA) functions in rheumatic MS were documented both in spectral and tissue Doppler echocardiography studies [4,5]. Decreased contraction of LAA causes spontaneous echo contrast (SEC) and thrombus formation even in MS patients with sinus rhythm [5][6][7].
Patent foramen ovale (PFO) is an interatrial septum abnormality which is detected in 27.3% of the normal population in autopsy series [8]. In a transesophageal echocardiography (TEE) study, PFO with left to right shunt was detected in 27% in left-sided cardiac lesions and it was speculated that this served to offload left atrial pressure referring to Lutembacher syndrome-like physiology [9].
Based on these pathophysiological phenomenon, practice of percutaneous implantation of shunt devices to the interatrial septum have been introduced as a nonpharmacological treatment alternative for heart failure. It was reported that these devices reduce left atrium (LA) pressure, particularly during exercise, and increase patient's exercise capacity [10].
Based on our clinical observations we hypothesized that in the presence of PFO, patients with MS who had smaller valvular areas are symptomatic at presentation, also the patients with PFO and MS concomitantly, thrombus and SEC in the LAA were observed fewer. Our primary focus was to investigate whether patients with PFO had a smaller mitral valve area (MVA) when admitted with symptoms. Secondary aim was to assess the relationship between thrombus formation frequency in the LAA and presence of a PFO. This retrospective study was designed to investigate whether concomitant presence of PFO makes a positive contribution to clinical presentation in two different pathways in severe MS patients by screening the TEE images in our hospital imaging database.

Study population
It is a single-center, retrospectively designed study. Medical records of the patients who underwent TEE due to symptomatic severe MS between 2008 and 2019, were screened. The patients with MVA < 1.5 cm 2 and at most mild to moderate mitral regurgitation (MR) or aortic regurgitation (AR) were included. The patients who had MVA > 1.5 cm 2 , moderate or severe MR or AR, tricuspid stenosis, severe pulmonary stenosis and regurgitation, left ventricle ejection fraction (LVEF) < 50%, previous history of percutaneous mitral balloon valvuloplasty (PMBV), ASD, partial and total pulmonary venous return anomalies, moderate or severe aortic stenosis were excluded.
After evaluation of these criteria, 320 patients with severe MS who underwent transthoracic echocardiography (TTE) and TEE to evaluate suitability for PMBV were included in the study. Transesophageal and TTE reports and TEE images archived to Picture Archiving and Communication Systems (PACS) system were reevaluated. The study was approved by the local ethics committee. The study protocol is in line with the Declaration of Helsinki [11].

Electrocardiography
Electrocardiographies (ECG) found in the medical records were revisited. Atrial fibrillation was defined by the absence of a P wave and irregular QRS complexes on 12-lead surface ECG.

Echocardiographic measurements
Echocardiography reports were screened and PACS TEE images of patients included in the study were reassessed. A standardized TEE protocol (19) was used to define valvular disease and atrial septum morphology.
Rheumatic MS evaluation was performed according to the European Association of Echocardiography and American Society of Echocardiography recommendations for echocardiographic assessment of valve stenosis [12].
Anatomical evaluation of the valve includes presence of commissural fusion (assessed with 2D and 3D echocardiography), leaflet thickness and motion, chordal shortening and thickening, presence of fibrotic nodules and calcification and extension (Fig. 1a). Continuous Doppler wave was used in apical views for mitral flow. Through this measurement, mean gradient was calculated by tracing diastolic mitral flow contour; pressure half time (PHT) was measured from E wave descending slope and MVA was calculated. With 3D echocardiography, the narrowest orifice measurement was made with the multiplanar reconstruction (MPR) method in which the valve was evaluated from three orthogonal planes (Fig. 1b). Planimetric (PLN) valve area measurement was made by determining the narrowest orifice by scanning from the apex to the base, from the 2D parasternal short axis at mid-diastole containing the commissures (Fig. 1c). Average values obtained from 3 to 5 beats were used for patients in atrial fibrillation. Systolic pulmonary arterial pressure (SPAP) was calculated using the maximum velocity of the tricuspid regurgitant jet and estimated right atrial pressure. Wilkins score was calculated by evaluating the valve mobility, thickness, calcification extent and thickness of the subvalvular apparatus.
Quantification of valvular regurgitations was performed according to recommendations for the echocardiographic assessment of native valvular regurgitation [13]. Mild MR was defined as vena contracta (VC) < 3 mm, effective regurgitant orifice area (EROA) < 20 mm 2 , regurgitant volume (RV) < 30 ml; mild to moderate MR was defined as VC intermediate (3 mm < VC < 7 mm), EROA 20-29 mm 2 , RV 30-44 ml. The VC was measured by optimizing gain/scale and from two orthogonal planes where the three components of the regurgitation jet (VC, proximal isovelocity surface area (PISA), jet into LA) were identified. For EROA and RV calculation, PISA radius was measured at mid-systole using the first aliasing at the appropriate Nyquist limit. EROA and RV were calculated using peak MR velocity and tissue velocity imaging. Mild AR was defined as VC < 3 mm, EROA < 10 mm 2 , RV < 30 ml; mild to moderate AR was defined as VC intermediate (3 mm < VC < 6 mm), EROA 10-19 mm 2 , RV 30-44 ml. Mild tricuspid regurgitation (TR) was defined as PISA radius < 5 mm, small, central jet; moderate TR was defined as PISA radius 6-9 mm; severe TR was defined as VC > 7 mm, PISA radius > 9 mm, EROA ≥ 40 mm 2 , RV ≥ 45 ml. PFO ( Fig. 2) was defined as a potential space or separation between the septum primum and secundum at the anterosuperior portion of the atrial septum, and a "stretched" PFO was present when atrial hemodynamics have enlarged the foramen and resulted in left to right or right to left shunting seen on Doppler imaging, depending on the differences in the right and LA pressure [14].
The SEC was recorded when dynamic smoke like echoes with characteristic rotational movement at an appropriate gain setting. LAA thrombus was defined as the presence of echogenic mass, which is distinctly different from pectinate muscles.

Clinical data
The patients' hypertension, diabetes mellitus, cerebrovascular event (CVA) records were obtained from the institutional electronic medical records system (HBYS: hospital information management system) and archive files. Neurology clinic attendance and history of transient ischemic attack or cerebrovascular accident were recorded. Acetyl salicylic acid and/or warfarin use was screened in medical records of patients.

Statistical analyses
All statistical analyzes were performed using "MASS", "rms", "ggplot2" and "ggridges" packages with R-software v. 3.5.1 (R statistical software, Institute for statistics and mathematics, Vienna, Austria). The continuous variables were presented as mean ± SD and median [interquartile range (IQR)] and as percentage (number of cases) for categorical variables. One sample Kolmogorov Smirnov test, Skewness and Kurtosis were used to test normality. Unpaired t test and Mann-Whitney U test were used to compare two groups of independent continuous variables as appropriate.  To allow for nonlinear associations, continuous predictors were modeled by using restricted cubic spline transformations. Regression coefficient or Odds ratio were presented change from 25 to 75th percentile for continuous variables. The relative importance of each predictor in the models was estimated with partial R 2 /partial X 2 value for each predictor divided by the model's total partial R 2 /X 2 , which estimates the independent contribution of the predictor to the variance of the outcome (partial R 2 for least square and X 2 for ordinal regression).
Multivariable ordinary least square model demonstrated that increase in SPAP, LVEF and presence of PFO were associated with decrease in MVA at presentation (Table 2). We showed the relative importance of each predictor in the Multivariable proportional odds logistic regression model demonstrated that increase in age and LA diameter, absence of PFO were associated with more favorable thrombotic PFO is a congenital interatrial septum abnormality which is not uncommon to persist in adult life and usually has an asymptomatic course [15]. A prevalence study investigating PFO frequency in patients with left-sided cardiac lesions (majority of the patients had MS) emphasized that reaching the critical LA pressure caused a small shunt that would disappear spontaneously after the correction of index lesion [9]. PFO shunts can be seen in both directions and this is determined by pressure gradient between left and right atrium. PFO shunt usually enables the high pressure to be decompressed without increasing the ventricular workload [9].
Congenital or iatrogenic defects of the interatrial septum in patients with increased LA and RA pressures provide insight for this phenomenon. Interatrial balloon septostomy [16] or IAS fenestration [17] have been incorporated into practice in severe pulmonary hypertension (PH) as one of the non-pharmacological treatment options to offload RA pressure. The rationale of decompressing the LA was inspired by the more benign course of MS in Lutembacher syndrome, and outcomes are associated with the diameter of the ASD as well as the degree of pulmonary vascular resistance [3,18]. In a case series published by Steinbrunn et al., cases with atrial shunt (mostly ASD) accompanying MS were hemodynamically evaluated and isolated PFOs were suggested to be secondary to LA stretch. They also observed that, patients with ASD had no paroxysmal dyspnea or pulmonary edema attacks. When the worsening of pulmonary congestion after ASD closure were investigated in two patients, they concluded that interatrial shunt decompressed LA and pulmonary venous system protectively [18]. Also, in iatrogenic ASDs occurring after PMBV; absence of newly developing or worsening pulmonary hypertension [19] and follow-up Qp/Qs of < 1.5 [20] pioneered in determining the safety margin for this procedure. Increasing LA pressure with exercise constitutes the cornerstone in both the symptomatology and the functional capacity in the heart failure with prevented ejection fraction (HFPEF) patient group [21,22].In the REDUCE LAP-HF study, after implantation of intracardiac shunt device in HFPEF patients, the exercise capacity and quality of life increased and peak pulmonary capillary wedge pressure decreased [23]. Simulation studies reported that the 8 mm iatrogenic defect is sufficient to decrease LA pressure without significant right ventricular volume load by keeping the Qp/Qs value around 1.3-1.4 [10]. Additionally, this device was shown to achieve improvement in pulmonary functional tests, both at rest and exercise, without impairment of systemic perfusion in a patient group with HFPEF [24].
As discussed previously, although there are many pathological factors determining the symptomatology in severe MS, LA pressure increase constitutes the cornerstone. In the interventional or surgical treatment algorithm, pulmonary arterial hypertension secondary to increased pulmonary venous pressure also has an important role. We excluded patients with moderate-severe MR from our study population to eliminate the confounding effect of concomitant LA volume overload that can be seen in these patients.
In this patient group, PFO opens a gateway for LA volume emptying so the patient gains the area of PFO diameter in addition to the MVA to offload the LA. Therefore, we would also expect the LA diameter to be smaller patients with PFO, but there was no statistical difference in between our patient group. Both physiologically and based on our findings, it seems reasonable for MS patients to remain asymptomatic for a longer period even with smaller valvular area. Pediatric cardiologists know that, in a newborn with congenital MS when PFO coexists, the gradient will be relatively low underestimating the stenosis. With this study, we showed that described patophysiological phenomenon might be of clinical benefit for our adult patient group with PFO and severe MS.
Interestingly, while assessing the predictors of smaller MVA at symptomatic presentation, LVEF has made a significant contribution even within normal limits. Although patients with LVEF < 50% were excluded, 'more' normal EF and increase in EF from 60 to 65 have resulted in a statistical difference. This emphasizes the importance of contractility to provide the optimum afterload for the left ventricle already affected with low-preload due to MS. High EF and a good myocardial contractility has the potential to delay onset of symptoms. Finally, the high SPAP is the pathophysiological consequence of Group 2 PH caused by high LA pressure and high PCWP in MS. It is not surprising that a smaller valve area initiates this pathological process sooner and higher pulmonary arterial pressures are recorded.

LAA function
LAA is the usual suspect of origin when a thromboembolic event occurs in the setting of MS. Independent of mitral valve pathology, it has been reported that there is a negative correlation between LAA contraction velocity and mean pulmonary capillary wedge pressure [25]. MS is known to decrease LAA flow velocities by impairment of active and passive emptying functions of LAA both in sinus rhythm and in atrial fibrillation [4,5]. Although this functional deterioration is primarily due to hemodynamically elevated LA pressure, it is also possible that the chronic inflammatory progression of rheumatic disease directly affects LA and LAA [5]. An electron microscopic study showed atrial endocardial damage in mitral valve disease [26]. SEC and thrombus formation are some of the clinical consequences of functional impairment. LAA function was reported to recover after intervention or medical treatment. In patients with MS following PMBV, the LAA functions improved, and SEC or thrombus formation was less frequent in LAA [7,27]. Additionally, with changes in loading conditions after heart failure treatment LAA functions were reported to be improved [28].
In various studies symptom duration [29,30], AF [30], LA diameter [30], severity of MS [31], age [29], dilatation of LAA with atrial body [32] have been determined as predictors of SEC and thrombus in severe MS. It has been shown in previous studies that MR is protective for thrombus and SEC formation by decreasing stagnation in LA [33] and that accompanied by severe MR decreases systemic embolism [34]. In our study, advanced age, larger LA diameter and smaller MVA were identified as predictors of thrombotic status in LAA in accordance with the literature. Severe MR patients were excluded however, patients with moderate MR did not show difference in LAA thrombus status. In a retrospective TEE study (1288 individual studies screened) exploring the relationship between SEC in the LA and CVA, SEC in SR was associated with high CVA prevalence independent of LA size and LAA function. One of the explanations for similar CVA rates in sinus rhythm and atrial fibrillation could be undetected paroxysmal atrial fibrillation as only ECG during TEE was reported. It is highly likely for patients with dilated LA to have paroxysmal AF episodes [35].
Other predictors of the LAA thrombus can be listed as larger LA diameter, advanced age and smaller MVA in accordance with the literature [28][29][30]. Not only statistically but also the clinically significant variables were evaluated, and it was shown that PFO as an independent predictor of LAA thrombotic status has an important contribution to the result.

Conclusion
Combination of PFO with MS can be interpreted as a variation providing clinical benefit to the acquired valvular disease by alleviating the elevated LA pressure which is the cornerstone of pathophysiology. Remaining asymptomatic for a longer period and having less thrombosis constitute two separate entities of this benefit. We believe that our study will also contribute to symptomatic benefit search with interventions targeted at LA pressure modulation.

Limitations
Retrospective data is the main restrictive feature of the study. Some parameters could not be reached or reanalyzed (LA volume index, right atrium volume index, right ventricule ejection fraction) from the TEE images recorded in the PACS system and the report records. Invasive MVA confirmation was not performed. Since the treatment options for MS are currently determined mainly with echocardiography, invasive hemodynamic pressure measurements were not available. SPAP was only evaluated by echocardiography, therfore it was not possible to distinguish between isolated postcapillary or combined pre and postcapillary PH.