Evaluation of right ventricular diastolic function, systolic function, and circulating galectin‐3 concentrations in dogs with pulmonary stenosis

Abstract

Background

Cardiovascular diseases with increased right ventricular (RV) afterload induce RV diastolic and systolic dysfunction, and myocardial fibrosis in humans. Studies in dogs with pulmonary stenosis (PS) evaluating RV diastolic function and markers of myocardial fibrosis are lacking.

Hypothesis/Objectives

Dogs with PS have echocardiographic evidence of RV diastolic and systolic dysfunction and increased serum concentrations of galectin‐3 (Gal‐3), a surrogate biomarker for myocardial fibrosis.

Animals

Forty client‐owned dogs (10 controls, 30 with PS).

Methods

Prospective study. All dogs had systemic blood pressure measurement, serum biochemical analysis, echocardiography, and measurement of serum Gal‐3 concentration performed.

Results

Variables of RV diastolic function were obtained in 39/40 dogs. Trans‐tricuspid flow velocity in early diastole to trans‐tricuspid flow velocity in late diastole ratios (RV E/A) were lower (P < .001) in dogs with PS (median, 0.94; range, 0.62‐2.04) compared to controls (1.78; 1.17‐2.35). Trans‐tricuspid flow velocity in early diastole to tricuspid annular myocardial velocity in early diastole ratios (RV E/e′) were higher (P < .001) in dogs with PS (11.55; 4.69‐28) compared to control (6.21; 5.16‐7.21). Variables of RV systolic function were lower in dogs with PS (P = <.001). Serum Gal‐3 concentration was higher (P = .002) in dogs with PS (285.1 pg/mL; 94.71‐406.97) compared to control dogs (162.83 pg/mL; 52.3‐232.82).

Conclusions and Clinical Importance

Dogs with PS have RV diastolic and systolic dysfunction, and increased Gal‐3 concentrations. These findings suggest the presence of RV myocardial fibrosis in dogs with PS, which could impact clinical management.

Abbreviations

Gal‐3

galectin‐3

iRAd

right atrial diameter indexed to body weight

iRV FAC

fractional area change of the right ventricle indexed to body weight

iRVFWd

right ventricular free wall thickness at end‐diastole indexed to body weight

iRVIDd

right ventricular internal diameter at end‐diastole indexed to body weight

iRV s'

peak systolic right ventricular myocardial velocity at the lateral tricuspid annulus indexed to body weight

iTAPSE

tricuspid annular plane systolic excursion indexed to body weight PS, pulmonary stenosis

RV

right ventricle

RV A

trans‐tricuspid flow velocity in late diastole

RV a′

myocardial velocity measured at the lateral tricuspid annulus in late diastole

RV decel

deceleration time of the RV E wave (RV E decel)

RV E

trans‐tricuspid flow velocity in early diastole

RV e′

myocardial velocity measured at the lateral tricuspid annulus in early diastole

RV E/A

ratio of trans‐tricuspid flow velocity in early diastole to trans‐tricuspid flow velocity in late diastole

RV e′/a′

ratio of myocardial velocity measured at the lateral tricuspid annulus in early diastole to myocardial velocity measured at the lateral tricuspid annulus in late diastole

RV E/e′

ratio of trans‐tricuspid flow velocity in early diastole to the myocardial velocity measured at the lateral tricuspid annulus in early diastole

RV IVRTc

isovolumic relaxation time (IVRT) corrected for heart rate

TR

tricuspid regurgitation

V _maxAV

maximal systolic ejection velocity of the aortic valve as measured from a subcostal view

V _maxAV/V _maxPV

ratio of the maximal systolic ejection velocity of the aortic valve over the maximal systolic ejection velocity of the pulmonary valve

V _TIAV/V _TIPV

ratio of the velocity time integral of systolic flow across the aortic valve over the velocity time integral of systolic flow across the pulmonary valve

1. INTRODUCTION

Pulmonary stenosis (PS) is a common congenital defect in dogs, resulting in obstruction to blood flow through the pulmonary valve and increased afterload to the right ventricle (RV). ¹ , ² , ³ , ⁴ Dogs with severe PS commonly develop clinical signs such as weakness and collapse, with some developing right‐sided congestive heart failure. ⁵ , ⁶ , ⁷ Valvular PS can be treated by balloon valvuloplasty with or without stent implantation, and these procedures generally are considered safe and effective in dogs and humans. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ However, some dogs experience restenosis, RV dysfunction, arrhythmias, or return of clinical signs or heart failure. ¹⁰ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Because it is unclear which dogs will deteriorate over time, there is no consensus on long‐term monitoring.

In humans with PS and other cardiac defects that result in increased RV afterload, ¹⁰ , ¹⁷ , ¹⁸ , ¹⁹ development of myocardial dysfunction, systolic or diastolic, indicates poor long‐term prognosis. ¹⁸ , ²⁰ Echocardiographic variables of RV diastolic function include the ratio of trans‐tricuspid flow in early (RV E) to late (RV A) diastole (RV E/A), the ratio of RV E to tricuspid annular myocardial velocities in early (RV e′) diastole (RV E/e′), the ratio of RV e′ to tricuspid annular myocardial velocities in late (RV a′) diastole (RV e′/a′), deceleration time of the E wave (RV E decel), and the isovolumic relaxation time (IVRT) corrected for heart rate (RV IVRTc). These variables help identify patients that require future intervention or have increased mortality. ²⁰ , ²¹ , ²² Numerous studies have shown that humans with congenital and acquired cardiac diseases develop myocardial fibrosis, a finding prominent in patients with worsened clinical severity. ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ Myocardial histopathology can detect myocardial fibrosis, but this technique only can be utilized after death or with invasive endomyocardial biopsies. Cardiac magnetic resonance imaging also can be used to detect myocardial fibrosis. One circulating biomarker that is a surrogate for myocardial fibrosis is galectin‐3 (Gal‐3). ²⁸ Galectin‐3 is a protein that activates fibroblasts in the myocardium, leading to fibrosis. ²⁹ Increased Gal‐3 concentrations are present in diseases with myocardial fibrosis, ¹⁷ , ²⁹ and often are found in humans with poor outcomes. ¹⁷ , ³⁰ , ³¹ , ³² , ³³ Although Gal‐3 has been evaluated in dogs, ³⁴ , ³⁵ a thorough evaluation of Gal‐3 has not been performed in dogs with PS. Studies evaluating RV systolic function have been performed in dogs with PS, ¹⁵ , ³⁶ but evaluation of RV diastolic function is rare in dogs. The importance of evaluating RV diastolic function variables and Gal‐3 in dogs with PS, in addition to RV systolic function variables, is unknown.

Our objectives were (a) evaluate RV diastolic function variables including RV E/A, RV E/e′, RV e′/a′, RV E decel, and RV IVRTc, and RV systolic function variables in healthy dogs and dogs with PS; (b) compare circulating Gal‐3 concentrations in dogs with PS to healthy control dogs. We hypothesized that dogs with PS would have evidence of RV diastolic and systolic dysfunction identified by echocardiography and that Gal‐3 concentrations would be higher in dogs with PS compared to healthy dogs.

2. MATERIALS AND METHODS

2.1. Animals

Forty dogs were prospectively enrolled between April 2021 and February 2023 (10 at Ohio State University College of Veterinary Medicine, and 30 at the Auburn University College of Veterinary Medicine). The Institutional Animal Care and Use Committee at each University approved the study before enrollment (2020‐A00000083, 2022‐4054), and owner consent was obtained. All dogs had the following tests performed: complete physical examinations, indirect systemic blood pressure measurements using Doppler ultrasonography on dogs in right lateral recumbency, serum biochemical analysis, serum Gal‐3 analysis, and transthoracic echocardiography on dogs restrained in right and left lateral recumbency. Indirect blood pressure measurements were taken 3 to 5 times on each patient, with the mean of the measurements recorded. Blood was obtained from a peripheral vein for serum biochemical analysis and Gal‐3 analysis. A portion of the blood obtained was used immediately for serum biochemical analysis, whereas 1 to 2 mL of blood was allowed to sit undisturbed at room temperature for 15 minutes to allow for stable clot formation. These samples then were centrifuged for 10 minutes, and serum was harvested and placed in polypropylene tubes that were immediately frozen at −80°C until batch analysis at the end of the study. Samples collected at Ohio State University College of Veterinary Medicine were shipped frozen on dry ice to the Auburn University College of Veterinary Medicine to be stored again at −80°C. When study enrollment was complete, the frozen serum samples were thawed for analysis. Serum galectin‐3 concentrations were measured using an ELISA (canine galectin‐3 ELISA kit; RayBiotech) in accordance with the manufacturer's instructions as previously described. ³⁷ Medical histories were obtained from all owners, including a detailed medication history. Dogs enrolled in the control group were not permitted any medications with known effects on the cardiovascular system. Exclusion criteria included any cardiovascular drug use in control dogs and any cardiovascular drug use other than atenolol in dogs with PS. Additional exclusion criteria included body weight <5.0 kg, systemic systolic blood pressure measurements >160 mmHg, ³⁸ having a clinically relevant hepatic disease (defined as inducing clinical signs or requiring hepato‐protectant medication), ³⁹ having dermatological disease, having a diagnosis of neoplasia, or having a concurrent cardiovascular disease other than PS.

2.2. Echocardiographic examination

Before echocardiographic examination, all dogs received 0.3 mg/kg butorphanol either IM or SC. In order to achieve the full effects of the administered sedation, the echocardiographic examination began at least 15 minutes after an IM injection and at least 30 minutes after an SC injection. The time and route of sedation administration were recorded. Transthoracic echocardiography was performed with dogs gently restrained in right and left lateral recumbency with a simultaneous ECG tracing. ⁴⁰ Echocardiographic examinations were performed by a board‐certified cardiologist (RLW) using the Vivid E95 ultrasound system (GE Medical Systems, Waukesha, Wisconsin) with multifrequency phased‐array transducers and subsequently analyzed on a digital off‐cart workstation (EchoPacs; GE Medical Systems, Waukesha, Wisconsin). The time that the echocardiographic examination began was recorded, as was heart rate and respiratory rate at the initiation of the echocardiographic examination. All measurements were performed by a single investigator (RLW) using the following protocol: in order to account for the effect of respiration, 5 to 7 cardiac cycles in sinus rhythm were averaged and used for statistical analysis. ⁴¹ , ⁴² The same ultrasound system and digital off‐cart workstation were utilized at both enrollment centers.

Cardiac valves were evaluated by 2‐dimensional (2D) imaging, and transvalvular flow and valvular regurgitation were evaluated using color Doppler echocardiography. Dogs were diagnosed with PS, and were considered to have type A stenosis, type B stenosis, or an intermediate morphology stenosis based on previously published criteria. ¹ Severity of tricuspid regurgitation (TR) was qualitatively assessed using color Doppler based on the following criteria of the observed regurgitant jet: trivial TR was <10% of the right atrial area; mild TR was 10% to 25% of the right atrial area; moderate TR was 25% to 50% of the right atrial area; and severe TR was >50% of the right atrial area. ¹⁵ , ⁴⁰ , ⁴³ The tricuspid valve was assessed in detail to exclude the presence of concurrent tricuspid valve dysplasia. The tricuspid valve was considered dysplastic if there were abnormalities of the valve leaflets (eg, thickened, elongated, fused), chordae tendineae (eg, long, short, absent), or papillary muscles (eg, direct attachment to the leaflet). ⁴⁴ Variables of left heart size and function were measured including the ratio of the left atrium to aorta, ⁴⁵ left ventricular internal dimensions at the end of diastole and systole normalized to body weight, and fractional shortening derived from M‐mode. ⁴⁶ Aortic flow was interrogated from the subcostal view, and pulmonary flow was interrogated from the right parasternal short‐axis view. ¹ , ⁴⁰ The pulsed‐wave Doppler‐derived profile of the transaortic flow was obtained by manually tracing the outer edge of the modal velocities throughout systole. The software package recorded the maximal systolic ejection velocity of the aortic valve (V _maxAV) and then calculated the velocity time integral of systolic flow through the aortic valve (V _TIAV). Using the same technique on continuous‐wave Doppler profiles of the pulmonary valve (PV), the maximal systolic ejection velocity of the pulmonary valve (V _maxPV) and velocity time integral of systolic flow through the pulmonary valve (V _TIPV) were obtained. The pressure gradient across the pulmonary valve (PG_PV) was calculated as PG_PV (mmHg) = 4 (V _maxPV). ² Velocity time integral ratio (V _TIAV/V _TIPV) was calculated as V _TIAV (cm)/V _TIPV (cm). Velocity ratio was calculated as V _maxAV (m/sec)/V _maxPV (m/sec).

Indices of right heart size were measured as previously described ¹⁴ , ⁴⁷ and indexed to body weight (BW) using previously published scaling exponents: right atrial diameter indexed to BW (iRAd) = cm/kg^{0.303 47} ; RV internal dimension at end‐diastole indexed to BW (iRVIDd) = cm/kg^{0.33 47} ; RV free wall thickness at end‐diastole indexed to BW (iRVFWd) = cm/kg^0.25. ⁴⁷ Measured indices of RV systolic function were obtained as previously described ¹⁴ , ⁴⁷ , ⁴⁸ and included fractional area change (FAC) of the RV (RV FAC), tricuspid annular plane systolic excursion (TAPSE) measured with M‐mode recordings, and pulsed‐wave tissue Doppler‐derived peak systolic RV myocardial velocity at the lateral tricuspid annulus (RV s'). Right ventricular FAC was calculated as ([RV area at end‐diastole − RV area at end‐systole]/RV area at end‐diastole) × 100. Indices of RV systolic function were indexed to BW using previously published scaling exponents: iRV FAC = %/kg^−0.097; iTAPSE = mm/kg^0.297; iRV s' = cm/sec/kg^0.23. ¹⁴

Indices of RV diastolic function were measured from the left apical 4‐chamber view optimized for the right heart. ¹⁴ , ⁴⁹ Particular care was taken to obtain optimal alignment of trans‐tricuspid blood flow (using pulsed‐wave spectral Doppler) and myocardial motion of the lateral tricuspid annulus (using pulsed‐wave tissue Doppler) for diastolic variables. Pulsed‐wave spectral Doppler variables were measured with a sample volume of 3 to 5 mm at the tips of the tricuspid leaflets; variables measured included RV E, RV A, and RV E decel. Pulsed‐wave tissue Doppler variables were measured with a sample volume of 2 to 3 mm at the lateral tricuspid annulus; these variables included RV IVRTc, ⁵⁰ RV e′ and RV a′. ⁴⁹ Calculated ratios of RV diastolic function included RV E/A, RV E/e′, and RV e′/a′. ⁴⁹ Calculation of RV IVRT corrected for heart rate was performed using the contemporaneous R‐to‐R interval measured on the ECG tracing using the following formula: RV IVRTc = IVRT/√RR. ⁵⁰

2.3. Statistical analyses

Statistical analysis was performed using commercially available software (JMP, version 16.0.0, SAS Institute Inc., Cary, North Carolina; and Graphpad Prism, version 9.3.1, La Jolla, California). A sample size calculation was prospectively performed to determine sample size for the study. Although no data for Gal‐3 concentrations or RV diastolic variables in dogs with PS have been published, we used the results from 2 similar studies that measured Gal‐3 concentrations to calculate sample size in order to attain a power of 0.8 and an α of 0.05. ¹⁷ , ³⁴ These calculations indicated groups between 8 and 17 would be sufficient to detect a difference in Gal‐3 concentrations. Study enrollment was designed to obtain 10 dogs in each of 4 groups: control dogs, mild PS, moderate PS, and severe PS. However, owing to the clinical population, a higher number of dogs with moderate and severe PS were enrolled. Descriptive statistics were generated and normality of data was tested using histogram analysis and the Shapiro‐Wilk test. Continuous data were reported as mean and SD if normally distributed, or if not normally distributed, they were reported as median and minimum and maximum. Student's t tests or Wilcoxon rank sum tests were performed to compare clinical data, echocardiographic variables, serum biochemical variables, and serum Gal‐3 concentrations between dogs with PS and control dogs, depending on data distribution. Analysis of correlation of age, velocity ratio, velocity time integral ratio, and variables of right heart size to variables of RV diastolic function, systolic function, and Gal‐3 was performed using Spearman's rank correlation and simple linear regression. The level of significance was considered as P ≤ .05, with Bonferroni correction performed to protect against type I error during multiple comparisons.

3. RESULTS

Forty‐one dogs were evaluated for our study, with 1 dog excluded because of a concurrent patent ductus arteriosus. Of the 40 dogs enrolled, 30 had PS and 10 were healthy control dogs. Eleven were female, and 29 were male. The study population had a median age of 28.4 months (minimum‐maximum, 4.5‐118.1 months) and a median weight of 18 kg (minimum‐maximum, 5.2‐57 kg). A summary of signalment data by group is presented in Table 1. No significant difference was found in age or weight of dogs with PS compared to control dogs (Table 1). No dogs had evidence of clinically relevant hepatic or dermatologic disease, and none were receiving hepato‐protectant medications or had a diagnosis of neoplasia. Dogs that received an SC injection of butorphanol (n = 9) had echocardiographic examination a median of 38 minutes (minimum‐maximum, 30‐59 minutes) after administration of the sedation. Dogs that received an IM injection of butorphanol (n = 31) had echocardiographic examination a median of 24 minutes (minimum‐maximum, 17‐44 minutes) after administration of the sedation. The median heart rate of all dogs at the start of the echocardiographic exam was 94 min⁻¹ (minimum‐maximum, 60‐166 min⁻¹). The median respiratory rate of all dogs at the start of the echocardiographic exam was 30 min⁻¹ (minimum‐maximum, 12‐60 min⁻¹). No significant difference in heart rate or respiratory rate was found between groups (P = .25 and P = .49, respectively). Of the 30 dogs with PS, 17 were receiving atenolol (median dosage, 2 mg/kg/day; minimum‐maximum, 0.8‐3.0 mg/kg/day) at the time of examination. Breeds of dogs enrolled in the study included mixed breed (n = 7), French Bulldog (n = 6), Golden retriever (n = 4), Pitbull terrier (n = 2), Chihuahua (n = 2), German Shorthair Pointer (n = 2), Miniature Schnauzer (n = 2), English Bulldog (n = 2), Rottweiler (n = 2), and 1 dog each from the following breeds: American Cocker Spaniel, Miniature Pinscher, Great Pyrenees, Samoyed, Bull terrier, German shepherd, Cavalier King Charles Spaniel, Doberman Pinscher, Australian shepherd, Boxer, and Pug. Tricuspid regurgitation was observed in 28/40 dogs, and its severity was trivial (n = 13), mild (n = 10), moderate (n = 4), and severe (n = 1). Of the 30 dogs with PS, the median PG_PV was 137 mmHg (minimum‐maximum, 21‐289 mmHg). Sixteen dogs had PS type A, 7 had PS type B, and 7 had an intermediate PS form. Measures of left heart size and function as well as V _TIAV/V _TIPV and V _maxAV/V _maxPV are presented in Table 1. The mean systemic blood pressure for control dogs was 143 ± 10 mmHg, which was significantly higher (P = < .001) than the mean systemic blood pressure in dogs with PS (mean 124 ± 20 mmHg). However, the clinical relevance of this difference is unknown. Importantly, no dog in the study was observed to have systemic hypertension. Biochemical variables of kidney (BUN and creatinine concentrations) and liver (alanine transferase, alkaline phosphatase and aspartate aminotransferase activities, and total bilirubin concentrations) function were not significantly different between dogs with PS and control dogs (all P > .2).

TABLE 1.

Demographic data, echocardiographic measures of left heart size and function, and aortic to pulmonary flow variables from 40 dogs.

Variables	Healthy (n = 10)	PS (n = 30)	P
Age (months)	19.5 (8.93‐53.9)	36.4 (4.5‐118.1)	.27
Weight (kg)	25.3 (5.4‐43)	14.9 (5.2‐57)	.19
LA/Ao	1.29 ± 0.18	1.28 ± 0.16	.62
FS %	38.92 ± 5.6	43.81 ± 10.49	.96
iLVIDd	1.49 ± 0.16	1.22 ± 0.23	<.001
iLVIDs	0.85 ± 0.15	0.66 ± 0.21	.006
V maxAV (m/sec)	1.45 ± 0.32	1.33 ± 0.25	.33
V maxAV/V maxPV	1.46 (0.75‐1.72)	0.24 (0.12‐0.57)	<.001
V TIAV/V TIPV	1.17 (0.57‐1.36)	0.15 (0.02‐0.45)	<.001

Note: Data reported as mean ± SD or median (minimum‐maximum).

Abbreviations: FS %, fractional shortening percentage of the left ventricle; iLVIDd, left ventricular internal dimension at the end of diastole indexed to body weight ⁴⁶ ; iLVIDs, left ventricular internal dimensions at the end of systole indexed to body weight ⁴⁶ ; LA/Ao, ratio of the left atrial diameter to aortic root diameter as measured from a right parasternal short‐axis view ⁴⁵ ; PS, pulmonary stenosis; V _maxAV, maximal systolic ejection velocity of the aortic valve as measured from a subcostal view; V _maxAV/V _maxPV, ratio of the maximal systolic ejection velocity of the aortic valve over the maximal systolic ejection velocity of the pulmonary valve; V _TIAV/V _TIPV, ratio of the velocity time integral of systolic flow across the aortic valve over the velocity time integral of systolic flow across the pulmonary valve.

As expected, variables of right heart size were significantly different between groups. The iRAd, iRVIDd, and iRVPWd all were significantly higher in the PS group compared to the control group. The median iRAd was 13.2 mm (minimum‐maximum, 7.8‐22.1 mm) in dogs with PS and 10.2 mm (minimum‐maximum, 8.8‐11.8 mm) in control dogs (P = < .001). The mean iRVIDd was 0.91 cm ± 0.21 cm in dogs with PS and 0.74 cm ± 0.16 cm in control dogs (P = .02). The median iRVPWd was 5.8 mm (minimum‐maximum, 2.4‐8.2 mm) in dogs with PS and 2.2 mm (minimum‐maximum, 1.3‐3.1 mm) in control dogs (P = < .001). Variables of RV systolic function were obtained in all 40 dogs (Table 2). The iRV FAC was not significantly different between groups (P = .56). However, iTAPSE and iRV s' both were significantly lower in the PS group compared to the control group (P = < .001 in both). Variables of RV diastolic function were obtained in 39/40 dogs, because 1 control dog had a persistently increased heart rate throughout the examination, which prevented trans‐tricuspid RV E and RV A separation. Two measures of RV diastolic function were not significantly different between dogs with PS and control dogs (RV decel, P = .48; RV IVRTc, P = .14; Table 2). However, other measures of RV diastolic function (RV E/A, RV E/e′, and RV e′/a′) were significantly different between control dogs and dogs with PS (Figure 1, Table 2). A significant correlation was not found between patient age and RV E/A, RV E/e′, or RV e′/a′ (all P ≥ .1).

TABLE 2.

Echocardiographic measures of right ventricular systolic and diastolic function from 40 dogs.

Variables	Healthy (n = 10)	PS (n = 30)	P
iRV FAC (%)	64.1 (53.8‐97.1)	64.5 (46.6‐93.3)	.56
iRV s′ (cm/s)	9.8 (4.9‐10.6)	4.2 (2.6‐7.9)	<.001
iTAPSE (mm)	5.7 (4.1‐6.4)	4.1 (2.6‐5.7)	<.001
RV E (m/s)	0.63 (0.49‐0.78)	0.62 (0.42‐1.22)	.63
RV E decel (ms)	145.84 ± 25	138.04 ± 37.58	.48
RV A (m/s)	0.38 ± 0.09	0.63 ± 0.18	<.001
RV e′ (cm/s)	11 (8‐15)	5.5 (3‐13)	<.001
RV a′ (cm/s)	8 (6‐17)	8 (4‐15)	.28
RV E/A	1.78 (1.17‐2.35)	0.94 (0.62‐2.04)	<.001
RV E/e′	6.21 (5.16‐7.21)	11.55 (4.69–28)	<.001
RV e′/a′	1.32 (0.58‐1.52)	0.67 (0.36‐1.63)	.003
RV IVRTc (ms)	71.36 (43.71‐101.65)	75.21 (48.99‐135.78)	.14

Note: Data reported as mean ± SD or median (minimum‐maximum). Bolded values denote statistical significance.

Abbreviations: iRV FAC, fractional area change of the right ventricle indexed to body weight; iRV s′, peak systolic right ventricular myocardial velocity at the lateral tricuspid annulus indexed to body weight; iTAPSE, tricuspid annular plane systolic excursion indexed to body weight; RV A, trans‐tricuspid flow velocity in late diastole; RV a′, myocardial velocity measured at the lateral tricuspid annulus in late diastole; RV decel, deceleration time of the RV E wave (RV E decel); RV E, trans‐tricuspid flow velocity in early diastole; RV e′, myocardial velocity measured at the lateral tricuspid annulus in early diastole: RV E/A, ratio of trans‐tricuspid flow velocity in early diastole to trans‐tricuspid flow velocity in late diastole; RV e′/a′, ratio of myocardial velocity measured at the lateral tricuspid annulus in early diastole to myocardial velocity measured at the lateral tricuspid annulus in late diastole; RV E/e′, ratio of trans‐tricuspid flow velocity in early diastole to the myocardial velocity measured at the lateral tricuspid annulus in early diastole; RV IVRTc, isovolumic relaxation time (IVRT) corrected for heart rate.

FIGURE 1.

Box and whisker plots of (A) the ratio of trans‐tricuspid flow velocity in early diastole to trans‐tricuspid flow velocity in late diastole (RV E/A), (B) the ratio of trans‐tricuspid flow velocity in early diastole to the myocardial velocity measured at the lateral tricuspid annulus in early diastole (RV E/e′), and (C), the ratio of myocardial velocity measured at the lateral tricuspid annulus in early diastole to myocardial velocity measured at the lateral tricuspid annulus in late diastole (RV e′/a′) in 10 healthy dogs (control) and 30 dogs with pulmonary stenosis (PS). The box represents the first and third quartiles, the middle line represents the median, and the whiskers represent the minimum and maximum values.

Significant linear correlations of V _maxAV/V _maxPV to E/A (r = .59; P = < .001), RV E/e′ (r = −.51; P = < .001), and RV e′/a′ (r = .50; P = .001) were observed. Similar significant linear correlations of V _TIAV/V _TIPV to RV E/A (r = .62; P = < .001), RV E/e′ (r = −.49; P = .001), and RV e′/a′ (r = .54; P = < .001) were observed. No significant correlation was observed between V _maxAV/V _maxPV and RV IVRTc (P = .34) or RV E decel (P = .36). No significant correlation was observed between V _TIAV/V _TIPV and RV IVRTc (P = .37) or RV E decel (P = .32). A significant linear correlation of iRAd and RV E/e′ (r = .46; P = .003) was observed.

Serum Gal‐3 concentrations were significantly higher (P = .002) in dogs with PS (285.1 pg/mL; 94.71‐406.97) compared to control dogs (162.83 pg/mL; 52.3‐232.82; Figure 2). No significant linear correlation was found between patient age and serum Gal‐3 concentrations (P = .53). Significant, negative linear correlations of serum Gal‐3 to V _maxAV/V _maxPV (r = −.47; P = .002) and V _TIAV/V _TIPV (r = −.48; P = .002) were observed.

FIGURE 2.

Box and whisker plot of serum galectin‐3 concentrations (pg/mL) in 10 healthy dogs (control) and 30 dogs with pulmonary stenosis (PS). The box represents the first and third quartiles, the middle line represents the median, and the whiskers represent the minimum and maximum values.

A significant linear correlation of serum Gal‐3 concentrations to RV E/A (r = −.39; P = .01) was observed, but all other indices of RV diastolic function did not have significant linear correlation to serum Gal‐3 (all P > .1). No significant linear correlation of serum Gal‐3 concentration to any variable of RV systolic function was observed (all P > .1).

4. DISCUSSION

Diastolic dysfunction of the RV was observed in the dogs with PS in our study, as demonstrated by significantly lower RV E/A and RV e′/a′, and significantly higher RV E/e′ in dogs with PS compared to healthy dogs. Similar to variables used to evaluate the diastolic function of the left ventricle, low values of RV E/A and RV e′/a′ may indicate impaired ventricular relaxation. ⁴⁹ In humans, RV E/A < 0.8 and RV e′/a′ < 0.5 are reported reference values used to indicate impaired RV relaxation. ⁴⁹ Although the dogs with PS in our study had median values above these cut‐offs established in humans (RV E/A, median 0.94; RV e′/a′, median 0.67), it is unclear how accurate these cut‐off values are for dogs. The influence of RV filling pressures also may influence these variables, with RV E/A < 0.8 indicating less severe diastolic dysfunction compared to an RV E/A between 0.8 and 2.1, which has been identified in humans with normal diastolic function and also as a pseudonormal filling pattern in humans. ⁴⁹ Old age is known to cause decreased RV E/A and RV e′/a′ in dogs and humans, ⁴² , ⁵¹ , ⁵² but no difference in age was found between the groups of our study and no correlation of either RV E/A or RV e′/a′ was found to age. These findings suggest that the differences observed in the dogs of our study were caused by the presence of PS. Similar to studies of the left ventricle, ⁵³ RV E/e′ is considered an index for RV filling pressure in humans. ⁵⁴ , ⁵⁵ In humans with surgically repaired tetralogy of Fallot, RV E/e′ was associated with a need for further intervention. ²⁰ Increased right atrial pressure also has been associated with increased RV E/e′ in humans. ⁵⁴ , ⁵⁵ These previously reported findings are consistent with those of our study, specifically that the RV E/e′ had a significant positive correlation with iRAd.

Circulating Gal‐3 has been experimentally validated as a marker of myocardial fibrosis in a number of veterinary species, ⁵⁶ , ⁵⁷ and it also has been correlated with the presence of myocardial fibrosis in humans detected by either cardiac magnetic resonance imaging (cMR) or histopathology. ²⁹ , ⁵⁸ Right ventricular myocardial fibrosis has been identified by histopathology in a client‐owned dog with PS and in an experimental model with increased RV afterload. ⁵⁹ , ⁶⁰ Right ventricular myocardial fibrosis also has been identified in humans with PS using cMR, ¹⁶ and a similar study in dogs with PS identified increased right ventricular stiffness using cMR. ⁶¹ Interestingly, in the study evaluating RV stiffness using cMR, no significant correlations were observed between cMR‐derived stiffness variables and either PG_PV or measures of RV systolic function. ⁶¹ This observation is similar to the findings of our study, where serum Gal‐3 concentrations were not significantly correlated with RV systolic variables. The presence of increased Gal‐3 concentrations and RV diastolic dysfunction in the dogs with PS in our study is suggestive of the presence of RV myocardial fibrosis, but further evaluation using histopathology should be considered.

Clinical outcomes also have been associated with Gal‐3 concentrations in humans. In a study of adults with congenital heart disease, humans with increased Gal‐3 concentrations were 6 times more likely to die or develop heart failure compared to those with lower Gal‐3 concentrations. ¹⁷ In this study, some of the highest Gal‐3 concentrations were found in diseases with increased RV afterload. ¹⁷ Another study found that humans with increased Gal‐3 concentrations at admission for acute heart failure had an odds ratio of 7.67 for 30‐day mortality. ³⁰ Galectin‐3 concentrations also were found to predict short‐ and long‐term success after cardiac intervention, because adults with increased Gal‐3 concentrations had poor 30‐day safety measures, 1‐year survival, and all‐cause mortality rates after transcatheter aortic valve replacement. ³¹ Similarly, increased Gal‐3 concentrations aided prediction of readmission and mortality in pediatric patients after congenital heart surgery. ³³ Galectin‐3 concentrations are associated with new‐onset atrial fibrillation and ventricular tachycardia in humans, ⁶² , ⁶³ both of which may be relevant to long‐term management of dogs with PS. Veterinary reports of circulating Gal‐3 concentrations in dogs with heart disease are lacking, ³⁴ , ³⁷ and the clinical importance of this biomarker in veterinary cardiology is currently unknown. However, the findings that Gal‐3 was significantly correlated to RV diastolic function and severity of pulmonary obstruction in the dogs of our study suggest that Gal‐3 may be a clinically important variable to monitor.

All dogs in our study received butorphanol as a sedative before echocardiography. The influence of respiration on echocardiographic variables is known to be greater for the right heart than for the left. ⁶⁴ , ⁶⁵ In a study of healthy children, mean increases of 18% to 26% were observed in RV E and RV A velocities during inspiration. ⁶⁴ For that reason, it is recommended to obtain measurements of RV variables either using a respirometer simultaneously with echocardiography, to ensure all measurements are performed during the same phase of respiration, or to obtain measurements from 5 to 7 cardiac cycles in sinus rhythm and average these values for analysis. ⁴⁹ The use of sedation was intended to help minimize changes in heart or respiratory rate among patients. We administered the sedation to all patients, regardless of temperament, so as to standardize our protocol.

Our study had some limitations. Measurement of RV diastolic function variables is influenced by phase of respiration and heart rate. We attempted to minimize the influence of these variables by having all dogs sedated before echocardiography and by averaging 5 to 7 measurements for each variable. However, it is unknown if it would have been more accurate to utilize a respirometer concurrently so that measurements could be verified as all being made in the same phase of respiration. We did not do so, which could represent a limitation to our data. Estimates of right atrial volume may provide clinically relevant information regarding RV diastolic function, and measurement of right atrial area may have provided a more precise evaluation of the right atrial dimension compared to the linear variable used in our study. Additionally, circulating Gal‐3 has been found to be increased in dogs with noncardiac diseases such as severe dermatologic disease including neoplasia ⁶⁶ , ⁶⁷ and endocrine disease, ⁶⁶ as well as in cirrhosis of the liver in human patients. ⁶⁸ In experimental studies of animals with cardiac disease and in studies of humans without the above‐listed diseases, concentrations of Gal‐3 correlate well with the degree of myocardial fibrosis observed. We excluded patients with severe liver disease, those with any dermatologic abnormalities, and those with systemic hypertension or a diagnosis of neoplasia. Additionally, we believe it is unlikely that our patients had undiagnosed neoplasia, because the median age of all dogs in our study was 28 months. However, it is possible that some dogs had concurrent diseases that were unknown to us and that could have impacted the measured Gal‐3 concentrations. Lastly, we did not perform histopathology or cMR of the RV to verify the presence of myocardial fibrosis.

In our study, we observed RV diastolic dysfunction and increased concentrations of circulating Gal‐3 in dogs with PS compared to healthy dogs, findings that suggest the presence of RV myocardial fibrosis in dogs with PS. Studies evaluating the combination of echocardiographically‐derived RV diastolic function variables and Gal‐3 in dogs with PS have not been performed previously. Additional prospective studies on the impact that RV diastolic dysfunction and Gal‐3 concentrations may have on clinical outcomes such as survival are warranted.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by the IACUC at The Ohio State University (#2020‐A00000083) and the Auburn University College of Veterinary Medicine (#2022‐4054).

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Winter RL, Maneval KL, Ferrel CS, Clark WA, Herrold EJ, Rhinehart JD. Evaluation of right ventricular diastolic function, systolic function, and circulating galectin‐3 concentrations in dogs with pulmonary stenosis. J Vet Intern Med. 2023;37(6):2030‐2038. doi: 10.1111/jvim.16890