Right heart echocardiographic variables and prediction of clinical severity in dogs with pulmonary stenosis

Abstract Background Pulmonary stenosis (PS) usually is evaluated using echocardiography. A multiparametric approach, in addition to the maximum pressure gradient (PG), might be indicated to better characterize PS severity and address its management. Hypothesis/Objectives Our hypothesis was that right heart size and function are associated with echocardiographic and clinical severity of pulmonary stenosis in dogs. Animals Client‐owned dogs with PS. Methods Prospective, multicenter, observational study. Enrolled dogs underwent complete echocardiographic examination. Associations among right heart echocardiographic variables, PS transvalvular PG >80 mm Hg and presence of clinical signs (exercise intolerance, syncope, right‐sided congestive failure, or some combination of these) were assessed using logistic regression analysis. Results Eighty‐eight dogs with PS. Twenty‐eight dogs were symptomatic. Increased right ventricular end‐diastolic free wall thickness (odds ratio [OR] > 100; 95% confidence interval [95%CI], 50‐ > 100; P = .01) and decreased aorta‐to‐pulmonary artery velocity time integral ratio (OR, < 0.001; 95%CI, 0.0‐0.001; P = .005) were independently associated with PS PG >80 mm Hg. Decreased tricuspid annular plane systolic excursion (OR, 0.35; 95%CI, 0.15‐0.77; P = .01) and increased right ventricular end‐diastolic area (OR, 1.4; 95%CI, 1.08‐2.02; P = .01) were independently associated with clinical severity. Conclusion and Clinical Importance Structural and functional right heart echocardiographic variables are associated with echocardiographic and clinical severity in dogs with PS. A multiparametric approach is advised to better assess PS severity.


| INTRODUCTION
Pulmonary stenosis (PS) is a common congenital heart diseases in dogs. 1,2 Most dogs are asymptomatic at the time of diagnosis, but some can develop clinical signs such as exercise intolerance, syncope, and right-sided congestive heart failure (R-CHF). Previous studies indicated that the presence of clinical signs, young age at diagnosis, pulmonary annulus hypoplasia, pulmonary transvalvular pressure gradient (PG), and severity of tricuspid regurgitation (TR) adversely affect survival in dogs with PS. [3][4][5] In addition to clinical signs, echocardiography is the most widely used method to evaluate PS severity.
According to current guidelines, severity is based on the PG obtained by transpulmonary peak velocity (PV Max ) using the modified Bernoulli equation. 6 However, assessing PS severity based only on PG might be inaccurate, and a multiparametric approach is expected to add important information for assessment of PS severity, management, and possibly prognosis. 7,8 In fact, as demonstrated from Gorlin's formula, PG is an echocardiographic parameter that is highly influenced by cardiac output compared to other echocardiographic variables such as the pulmonary valve area index (iPVA), aortic (V MaxAo ) to pulmonary maximum velocity (V MaxPV ) ratio (V MaxAo :V MaxPV ), and the aortic to pulmonary artery velocity time integral ratio (VTI Ao :VTI PV ). 7 Therefore, in some cases, PG might underestimate or overestimate the severity of PS, and thus a multiparametric approach might be indicated. Moreover, an echocardiographic multiparametric approach is recommended in human medicine to assess the severity of valve stenosis. 9 In particular for aortic stenosis, it is suggested to use valve area and mean transvalvular PG in all patients, whereas in selected patients, when additional information is needed, the velocity ratio and planimetry of the anatomic valve area should be considered. 9 For PS, it is suggested to evaluate associated lesions such as right ventricular (RV) hypertrophy and RV enlargement. 9 Studies in human and veterinary medicine indicate that echocardiographic assessment of right heart size and function is important to predict the presence of clinical signs and define prognosis in patients with right heart disease, both acquired and congenital. [8][9][10][11][12] However, a recent study evaluated right heart size and function in dogs affected by PS, and suggested the importance of a multiparametric approach for echocardiographic evaluation. 8 Therefore, our aim was to verify the hypothesis that right heart size and function are associated with echocardiographic and clinical severity of PS in dogs.

| Animals
Ours was a prospective, multicenter, observational study that enrolled client-owned dogs with PS at the Anicura Istituto Veterinario Novara and at the Department of Veterinary Sciences of the University of Pisa, with the owner giving signed consent. Dogs were consecutively included in the study over a 2.5-year period if they had clinical and echocardiographic findings compatible with congenital PS. Exclusion criteria were presence of other cardiac (eg, patent foramen ovale, tricuspid dysplasia) or systemic diseases, presence of atrial fibrillation, and if sedation was needed to perform echocardiography. Presence or absence of patent foramen ovale was assessed based on color Doppler examination in all dogs, and in selected unclear cases a bubble study was performed. Treatment with atenolol was not considered an exclusion criterion. Dogs were divided into 2 groups of PS echocardiographic severity based on peak PG: PS with a maximum (max) PG ≤80 mm Hg, and PS with a max PG >80 mm Hg. Differences on right heart echocardiographic variables were evaluated between these 2 groups of PS severity. The association between right heart echocardiographic variables and echocardiographic severity (PS with max PG >80 mm Hg) was assessed using logistic regression analysis. To assess clinical severity dogs were divided into 2 groups: asymptomatic and symptomatic. Dogs were considered symptomatic if the owner reported exercise intolerance, syncope, R-CHF, or a combination of these. Right-sided congestive heart failure was defined as the presence of ascites associated with jugular venous distension and dilated caudal vena cava.

| Echocardiographic examination
All dogs underwent standard transthoracic echocardiography, restrained in right and left lateral recumbency, with a simultaneous ECG tracing. 13 Echocardiographic examinations were performed by a board-certified cardiologist (OD) or a resident (VP, TV, FM) under the direct supervision of the same board-certified cardiologist using Vivid iQ e (GE Healthcare, 20 126, Milan, Italy) and Aplio 300 (Canon Medical Systems Europe, Zoetemermeer, the Netherlands) ultrasound systems with multi-frequency phased-array transducers. All measurement were performed by a single investigator (VP) and 3 cardiac cycles in sinus rhythm for each variable were averaged and used for statistical analysis. All valves were evaluated morphologically using 2-dimensional (2D) imaging and interrogated using color Doppler echocardiography.
The tricuspid valve apparatus was carefully evaluated in each dog to exclude the presence of tricuspid valve dysplasia, as previously described. 14 Tricuspid dysplasia was considered when anomalies of the valve leaflets (eg, thickened, elongated, short, fused) or chordae tendineae (eg, short, long, absent) or papillary muscle (eg, long, short, fuse, direct attachment with the valve leaflet), or some combination of these was present. 14 Aortic valve (AV a ) and pulmonic valve (PV a ) diameters were measured in early systole between the hinge points of maximally opened leaflets from the right parasternal long axis (RPLA) view optimized with the left ventricular outflow tract and from the right parasternal short axis (RPSA) view at the level of the heart base, respectively, in early to mid-systole, and the AV a :PV a ratio was calculated. 4 The morphology of PS was evaluated and classified as type A, type B, and intermediate as previously described. 4 Furthermore, the AV a :PV a ratio was utilized to define PV a hypoplasia when the AV a :PV a > 1.2. 4,6 Briefly, PS type A was defined when there was fusion and thickening of the valve cusps associated with a normal PV a , whereas PS type B was considered when the dog had PV a hypoplasia associated with thickened and immobile valve cusps. 6 Pulmonary stenosis with intermediate morphology was defined when the valve had characteristics between type A and B PS. 4 Moreover, the presence of supravalvular and subvalvular stenosis as well as the presence of coronary artery anomaly were evaluated as previously described. 5,15 Subvalvular and supravalvular PS were distinguished based on 2D anatomy of the right ventricular outflow tract, pulmonic valve and main pulmonary artery structures and evaluating the site of the narrowing and increased velocity seen with spectral and color Doppler. In particular, subvalvular PS was considered when the narrowing, generally determined by the presence of fibrous ring or fibromuscular hypertrophy, was below the pulmonic valve, whereas supravalvular PS was diagnosed when the narrowing was above the pulmonary valve, also including narrowing of the sino-tubular junction. 5,6,15,16 Aortic flow was acquired from the subcostal view using pulsed wave Doppler, with the sample volume positioned between the hinge point of valve leaflets, and pulmonary flow was obtained from the right parasternal short axis view using continuous wave Doppler. 5 Maximum transvalvular aortic velocity (Ao Vmax ) and PV Vmax was obtained and Ao Vmax :PV Vmax ratio was calculated. Velocity time integrals of aortic flow (VTI Ao ) and pulmonic flow (VTI PA ) were obtained after the manual tracing of flow profiles and mean PG were generated by the machine's software. Velocity time integrals were used to calculate the VTI Ao :VTI PV ratio and pulmonary valve area (PVA) using the continuity equation as follows: (cross sectional area of the aortic valve x VTI Ao )/VTI PV . 6 Assuming that the aortic root is a circle, cross sectional area of the aortic valve was obtained using the formula: π x (AV a /2) 2 . The PVA then was indexed to each dog's body surface area (BSA) as previously described. 6,7 The left apical 4 chamber view optimized for the right heart was used to evaluate right heart size and function during echocardiographic measurements. In particular, right atrium area (RAA) was measured by planimetry at the end of ventricular systole from the lateral aspect of the tricuspid annulus to the septal aspect, excluding the area between the leaflets and annulus, following the right atrium endocardium and excluding the caudal vena cava. 16 The RAA index (iRAA) was calculated by dividing RAA by BSA. 17 Right ventricular end diastolic area (RVEDA) and end systolic area (RVESA) were measured by planimetry, tracing the lateral aspect of the tricuspid annulus to the septal aspect and excluding the area of the tricuspid annulus, trabecular structures and papillary muscles following the RV endocardium. 18 The RVEDA index (iRVEDA) was calculated as the ratio of RVEDA to BSA. 18 Right ventricular free wall thickness at end-diastole (RVFWd) was measured at the mid-ventricular level excluding the pericardium and including the endocardium, from the left apical 4-chamber view as previously described. 19 The RVFWd was indexed to body weight (BW; iRVFWd = cm/kg 0.250 ). 19 Right ventricular systolic function was assessed by measuring tricuspid annular plane systolic excursion (TAPSE) and fractional area change of the RV (FAC). The TAPSE was obtained from M-mode recordings of the lateral aspect of the tricuspid valve annulus seen from the left apical 4-chamber view optimized for the right heart. 20,21 The FAC was derived from the RVEDA and RVESA using the formula: FAC = [(RVEDA-RV end systolic area)/RVEDA] x 100. 20 Both TAPSE and FAC were indexed to BW using the previously published scaling exponent: iTAPSE in mm/kg 0.297 and iFAC in %/kg À0.097 . 20 Tricuspid regurgitation severity was qualitatively evaluated using color Doppler and continuous wave Doppler. 13,17,22 Mild TR was considered when a small (<25% of RAA), central jet signal on the color Doppler and a faint parabolic TR jet signal on the continuous wave Doppler were present.
Moderate TR was considered when an intermediate TR jet (>25% and <50% of RAA) on the color Doppler and a dense parabolic TR jet signal on the continuous wave Doppler were observed. Severe TR was considered when a very large central jet or an eccentric jet (>50% of RAA) impinging on the wall on the color Doppler and a dense triangular with early peaking TR jet signal were seen on the continuous wave Doppler. For statistical purposes, the following scores were assigned to TR severity: no TR = 0, mild TR = 1, moderate TR = 2, severe TR = 3. Receiver operating curve analysis and Youden index were used to identity the best echocardiographic cut-offs of the variables independently associated with clinical severity and the presence of R-CHF.
Results of univariate and multivariate analysis to identify the echocardiographic variables associated with max PG > 80 mm Hg are presented in Table 2. Decreased VTI Ao :VTI PV and increased iRVFWd were independently associated with PS max PG >80 mm Hg. In particular, for each increase of 1 of VTI AV /VTI PV the risk of having a PS PG >80 mm Hg is decreased by >99%, whereas for each increase of 1 cm/kg 0.250 of iRVFWd the risk of having a PS PG >80 mm Hg is increased by >100 times.
The results of univariate and multivariate logistic regression analysis to identify echocardiographic variables associated with clinical severity of PS are presented in Table 3. Decreased iTAPSE and increased iRVEDA were independently associated with the presence of clinical signs, whereas decreased iFAC and increased TR severity were independently associated with the presence of R-CHF (Table 4).
Lastly, the results of the receiver operating curve analysis to calculate the cut-offs for those variables that were significant in multivariate analysis are presented in

| DISCUSSION
We identified right heart echocardiographic variables associated with echocardiographic and clinical severity in dogs with PS and their cutoffs for the presence of clinical signs of R-CHF.
Regarding echocardiographic severity, increased iRVFWd and decreased VTI Ao :VTI PV were independently associated with PS max PG > 80 mm Hg. In our study, consistent with a previous study, 8 almost all dogs with PS PG >80 mm Hg had an increased iRVFWd (86% and 97%, respectively). In fact, in patients with PS, the RV responds to the increased afterload with myocyte hypertrophy leading to increased RV wall thickness. 8 Therefore, it is not unexpected that iRVFWd is increased in the majority of dogs with a PS PG >80 mm Hg.
Similarly, the previous study found a mean value for VTI Ao :VTI PV of 0.16 in a population of dogs with PS PG >80 mm Hg. 8 The VTI Ao : VTI PV ratio is a less flow dependent index and, as suggested by a previous study, 7 might be a useful index to evaluate the severity of PS in dogs, especially in cases with high or low cardiac output.
Increased iRVEDA and decreased iTAPSE were independently associated with the presence of clinical signs. This observation is consistent with a previous study in dogs with PS that identified iTAPSE as the echocardiographic variable associated with the presence of clinical signs 8 and, as previously suggested, should be included in the echocardiographic evaluation of dogs with PS to help predict clinical severity. However, in our study, the percentage of dogs with decreased iTAPSE was lower than in a previous study 8  Note: The bold values represent the variables resulted statistically significant in the multivariable analysis. Abbreviations: 95% CI, 95% confidence interval; AV a :PV a , aorta-to-pulmonary artery annulus ratio; iFAC, fractional area change indexed to body weight; iPVA, pulmonary valve area indexed to body surface area; iRAA, right atrial area indexed to body surface area; iRVEDA right ventricular area at end-diastole indexed to body surface area; iRVFWd, right ventricular free wall thickness at end-diastole indexed to body weight; iTAPSE, tricuspid annular plane systolic excursion indexed to body weight; Max PG, maximum transpulmonary pressure gradient; OR, odds ratio; V MaxAo :V MaxPV , aorta-to-pulmonary artery maximum transvalvular velocity ratio; VTI Ao :VTI PV , aorta-to-pulmonary artery velocity time integral ratio; TR severity, tricuspid regurgitation severity.
decreased iTAPSE in previous studies originated from a population of adult dogs and not puppies. 20 In human medicine, adults and children have different reference ranges of TAPSE and lower values are observed in children. 29 In both our study and the previous study, 8  In another study 19 evaluating reference intervals and repeatability of right heart echocardiographic measurements, TAPSE showed a high coefficient of variation, with inter-operator variability of 21%.
Notwithstanding, TAPSE might not be the best echocardiographic index to evaluate RV systolic function because it is highly influenced by loading conditions, by left ventricular systolic function (ventricular interdependence), and because it is an angle-dependent parameter. 24 In human medicine, it has been noted that strain and strain rate might be superior in evaluating RV systolic function, because they are angle-independent, less load-dependent and can detect systolic dysfunction in the early stage of disease both in congenital and acquired right heart diseases. 26,28,30,31 However, to our knowledge, no studies have assessed strain and strain rate in evaluating systolic function in dogs affected by PS and additional studies are needed to assess this aspect of the disease.
Ours is the first study that identified the echocardiographic vari- In our study based only on PG of PS, 1 dog with afterload mismatch and R-CHF was included in the group having a PS max PG ≤80 mm Hg, which mostly consisted of asymptomatic dogs. This finding reinforces the concept of using a multiparametric approach to stratify PS severity not only based on PG, but also by evaluation of right heart structure (ie, iRAA, iRVEDA), function (ie, FAC, TAPSE), and taking into consideration less-flow dependent echocardiographic parameters (ie, iPVA, VTI Ao :VTI PV , V MaxAo :V MaxPV ). Moreover, right atrial enlargement, ventricular enlargement, and systolic dysfunction were defined based on previous published cut-offs derived from a population of healthy adult dogs. 18,19 In our study, almost 40% of dogs were <1 year of age, and it cannot be excluded that puppies have a different reference range for these variables compared to adult dogs.
Moreover, 44% of the dogs in our study were brachycephalic. A previous study determined that, in particular, English Bulldogs have a different leftsided cardiac chamber reference range for dimension, function, and geometry compared to the general dog population. 33 It is possible that differences also are present for RV echocardiographic parameters.
In our study, the results of the association between echocardiographic variables and R-CHF must be considered as preliminary based on the small sample size (only 11 dogs in R-CHF). However, the association between R-CHF and severity of TR was very strong and it is interesting that TR severity seems to be a main determinant of the presence of R-CHF in both congenital and acquired right heart pressure-overload diseases in dogs. 16 Lastly, in our study TR severity was classified based only on a qualitative and semiquantitative assessment of the color and spectral Doppler signals. The echocardiographic evaluation of TR severity remains challenging both in human and veterinary medicine, and may present substantial limitations. In human medicine 2-D and 3-dimensional quantitative echocardiographic indices have been proposed in the evaluation of TR severity (eg, vena contracta, proximal isovelocity surface area, regurgitant volume calculation). 37 However, to our knowledge, no studies in veterinary medicine evaluated these quantitative methods to assess TR severity and additional studies are needed to assess the feasibility, repeatability, and accuracy of these measurements in dogs.

| CONCLUSION
Our study supports the use of different right heart echocardiographic variables to evaluate dogs affected by PS. In particular, decreased RV ventricular systolic function (ie, iTAPSE and FAC) and RV dilatation (ie., iRVEDA) are associated with the presence of clinical signs.