- Email: [email protected]
Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease
Journal of Veterinary Cardiology (2017)
-, -e-
www.elsevier.com/locate/jvc
Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease*,** David Dickson, BVetMed CertAVP(VC) a, Domenico Caivano, DVM, PhD b, Jose Novo Matos, DVM c, Nuala Summerfield, BVMS d, Mark Rishniw, BVSc, MS, PhD e,* a
HeartVets, Porthcawl, Wales, UK Department of Veterinary Medicine, University of Perugia, Perugia, Italy c Vetsuisse Faculty University of Zurich, Zurich, Switzerland d New Priory Vets Brighton, Brighton, UK e Veterinary Information Network, Davis, CA, USA b
Received 10 January 2017; received in revised form 28 August 2017; accepted 11 September 2017
KEYWORDS Canine; Cardiac; Congestive heart failure
Abstract Objectives: To provide reference intervals for 2-dimensional linear and area-based estimates of left atrial (LA) function in healthy dogs and to evaluate the ability of estimates of LA function to differentiate dogs with subclinical myxomatous mitral valve disease (MMVD) and similarly affected dogs with congestive heart failure (CHF). Animals: Fifty-two healthy adult dogs, 88 dogs with MMVD of varying severity. Methods: Linear and area measurements from 2-dimensional echocardiographs in both right parasternal long and short axis views optimized for the left atrium were used to derive estimates of LA active emptying fraction, passive emptying fraction,
*
Preliminary data from this study were presented at the 24th ECVIM-CA Congress in Mainz in 2014. A unique aspect of the Journal of Veterinary Cardiology is the emphasis of additional web-based materials permitting the detailing of procedures and diagnostics. These materials can be viewed (by those readers with subscription access) by going to http://www.sciencedirect.com/science/journal/17602734. The issue to be viewed is clicked and the available PDF and image downloading is available via the Summary Plus link. The supplementary material for a given article appears at the end of the page. To view the material is to go to http://www.doi.org and enter the doi number unique to this paper which is indicated at the end of the manuscript. * Corresponding author. E-mail addresses: [email protected] (D. Dickson), [email protected] (D. Caivano), [email protected] (J.N. Matos), [email protected] (N. Summerfield), [email protected] (M. Rishniw). **
https://doi.org/10.1016/j.jvc.2017.09.003 1760-2734/ª 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
2
D. Dickson et al. expansion index, and total fractional emptying. Differences for each estimate were compared between healthy and MMVD dogs (based on ACVIM classification), and between MMVD dogs with subclinical disease and CHF that had similar LA dimensions. Diagnostic utility at identifying CHF was examined for dogs with subclinical MMVD and CHF. Relationships with bodyweight were assessed. Results: All estimates of LA function decreased with increasing ACVIM stage of mitral valve disease (p<0.05) and showed negative relationships with increasing LA size (all r2 values < 0.2), except for LA passive emptying fraction, which did not differ or correlate with LA size (p¼0.4). However, no index of LA function identified CHF better than measurements of LA size. Total LA fractional emptying and expansion index showed modest negative correlations with bodyweight. Conclusions: Estimates of LA function worsen with worsening MMVD but fail to discriminate dogs with CHF from those with subclinical MMVD any better than simple estimates of LA size. ª 2017 Elsevier B.V. All rights reserved.
Abbreviations 2D CHF ECG LA LAACT
two-dimensional congestive heart failure electrocardiograph(y) left atrium (atrial) left atrial active emptying fraction (based on linear measurements) LA:Ao ratio of the left atrial dimension to the aortic annulus dimension LAA:AoA ratio of the left atrial area to the aortic annulus area MMVD myxomatous mitral valve disease ROC receiver operating characteristic RPSA right parasternal short axis RPLA right parasternal long axis
Left atrial (LA) function is increasingly evaluated in veterinary cardiac patients to help identify disease severity [1e6]. Left atrial function comprises 3 components: it acts as a reservoir during ventricular systole (atrial diastole), it acts as a conduit during early ventricular diastole and diastasis and as an active pump during late ventricular diastole (atrial systole) [1,4,7]. Although simple measures of LA size (e.g. left atrial-to-aortic ratio) correlate with disease severity and are predictive of clinical outcomes in dogs, clinicians often observe dogs with LA of similar size that have different clinical disease status, e.g. subclinical disease vs. congestive heart failure (CHF). Whether these dogs might have different LA function to explain their clinical status remains largely undetermined.
While 3-dimensional assessment of function might be more accurate than 2-dimensional (2D) assessment [4], most clinicians performing echocardiographic evaluations of patients do not possess technology capable of 3-dimensional imaging and do not have the time for substantial off-line analysis. Furthermore, volumetric accuracy is not required if simpler methods are precise enough to successfully differentiate groups of patients with different LA function or clinical status. Two-dimensional echocardiography is commonly available, and evaluation of function requires little, if any, off-line analysis, and is minimally time-consuming. To the authors’ knowledge, no studies have determined reference intervals for 2D echocardiographic non-volumetric indices of LA function in healthy dogs. Furthermore, few studies have evaluated indices of LA function in dogs with varying degrees of LA pathology secondary to myxomatous mitral valve disease (MMVD) or examined the ability of these indices to identify the clinical status of these dogs. These studies suggested that LA function decreases with worsening MMVD and might differentiate dogs with different clinical disease classifications [3,4]. However, these studies failed to examine the benefit of estimating LA function over simply measuring the size of the LA. In addition, these studies failed to examine the ability of functional indices to discriminate between subclinical MMVD and CHF in dogs that had similar degrees of cardiac remodeling (i.e. dogs with similarly marked LA and LV enlargement that did, or did not, have the evidence of CHF). Therefore, we sought to establish reference intervals for 2D echocardiographically derived
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs indices of LA function in healthy dogs and to examine the ability of changes in these indices to identify MMVD severity, based on the ACVIM functional classification of MMVD.
Materials and methods Subjects Echocardiograms were prospectively performed on 140 client-owned dogs: 52 healthy dogs (dogs with an unremarkable history, physical examination, and normal echocardiogram), 25 dogs with subclinical mitral valve disease and no evidence of left atrial enlargement (ACVIM stage B1), 34 dogs with subclinical mitral valve disease and LA and ventricular enlargement (ACVIM stage B2), and 29 dogs with CHF (ACVIM stage C) [8]. Dogs with CHF were classified on the basis of severe left-sided heart disease, clinical signs of tachypnea and/or dyspnea, radiographic evidence of a pulmonary interstitial pattern, and, in all-but five dogs (which had not yet been administered diuretics at the time of imaging), a clear and positive response to diuretic therapy, manifesting as a normalization of respiratory rate and character. Twenty-four dogs with CHF were receiving combinations of furosemide, pimobendan, and angiotensin-converting enzyme inhibitors. Six of the stage B2 dogs and one stage B1 dog were receiving angiotensin-converting enzyme inhibitors or pimobendan at the time of evaluation. No institutional approval was sought, as the procedure was considered part of the routine evaluation of all dogs with MMVD. Owners of healthy dogs provided informed consent for the echocardiographic imaging.
Echocardiography All dogs underwent complete routine 2D echocardiography examinations by one of 4 investigators, using a variety of ultrasound equipment,f,g,h,i and appropriate phased-array transducers (ranging from 3.0 to 8.5 MHz). All echocardiograms were performed with simultaneous electrocardiographic (ECG) recording. Dogs were unsedated, and gently restrained in right lateral recumbency for all left atrial imaging.
f
MyLab Class C, Esaote, Genova, Italy. Vivid Q, GE Medical Systems Ltd, Buckinghamshire, UK. h Vivid I, GE Medical Systems Ltd, Buckinghamshire, UK. i Vivid 7, GE Healthcare, Chicago, IL, USA. g
3 Investigators obtained and stored cineloops of the left atrium from the right parasternal long-axis 4chamber (RPLA) view, and the right parasternal short-axis (RPSA) view optimized for the aortic valve and left atrium [9]. Investigators obtained and averaged three sets of measurements from different cardiac cycles (not necessarily consecutive) of the left atrium from the stored cineloops at the following timepoints: 1. For the RPSA view: (A) Early ventricular diastole, at the first frame in which aortic valve closure was observed, corresponding to maximal left atrial diameter; (B) Late ventricular diastole, at the last frame before aortic valve opening was observed, corresponding to the minimal left atrial diameter; (C) Diastasis, at the onset of the p-wave observed on the ECG. 2. For the RPLA view: (A) Early ventricular diastole, at the last frame before opening of the mitral valve was observed, corresponding to maximal left atrial diameter; (B) Late ventricular diastole, at the first frame after mitral valve closure was observed, corresponding to the minimal left atrial diameter; (C) Diastasis, at the onset of the p-wave observed on the ECG. Two sets of measurements were made at each time point: (A) Linear LA dimension, as previously described [10]. (B) Left atrial area, as previously described [10]. The averaged linear and area measurements were then used to calculate the indices of LA function (Table 1). Three aortic measurements were obtained from the RPSA view in early diastole, at the first frame in which aortic valve closure was observed and averaged [10,11]. We used a previously described method to differentiate dogs with MMVD in ACVIM stage B1 and stage B2 [12]. Briefly, we considered a dog to have no evidence of cardiac remodeling (stage B1) if 2 of the following 3 variables were below reference limits: LA:Ao (threshold value of 1.6), weightadjusted LA linear dimension (wLA; threshold value of 1.56), weight-adjusted LV linear
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
4
D. Dickson et al. Table 1 Linear
Area
Indices of left atrial function. Function assessed
Index
Abbreviation
Formula
LA reservoir
LA expansion index
LAEXP
LA contractile
LA active emptying fraction
LAACT
LA conduit
LA passive emptying fraction
LAPAS
LA reservoir
LA total emptying fraction
LAFT
LAMAX LAMIN )100 LAMIN LAP LAMIN )100 LAP LAMAX LAP )100 LAMAX LAMAX LAMIN )100 LAMAX LAAMAX LAAMIN )100 LAAMIN LAAP LAAMIN )100 LAAP LAAMAX LAAP )100 LAAMAX LAAMAX LAAMIN )100 LAAMAX
LA reservoir
LA expansion index
LAAEXP
LA contractile
LA active emptying fraction
LAAACT
LA conduit
LA passive emptying fraction
LAAPAS
LA reservoir
LA total emptying fraction
LAAFT
LA, left atrium; LAmax, maximal left atrial dimension; LAmin, minimal left atrial dimension; LAP, left atrial dimension at the onset of the p-wave on the ECG; LAEXP, left atrial expansion index (based on linear measurements); LAACT, left atrial active emptying fraction (based on linear measurements); LAPAS, left atrial passive emptying fraction (based on linear measurements); LAFT, left total atrial emptying fraction (based on linear measurements); LAAmax, maximal left atrial area; LAAmin, minimal left atrial area; LAAP, left atrial area at the onset of the p-wave on the ECG; LAAEXP, left atrial expansion index (based on area measurements); LAAACT, left atrial active emptying fraction (based on area measurements); LAAPAS, left atrial passive emptying fraction (based on area measurements); LAAFT, left atrial total emptying fraction (based on area measurements).
dimension in diastole (wLV; threshold value of 1.95). We derived the reference thresholds for the weight-adjusted estimates of LA and LV size from a previous study by Hall et al. [13]. Dogs with any 2 of these 3 variables above the reference limits we classified as ACVIM stage B2. Therefore, rare dogs could have LA:Ao > 1.6 but still be considered ACVIM stage B1, if both weight-adjusted variables were within reference intervals.
Statistical analyses Descriptive statistics were calculated for each index of LA function and are reported as median and range. Reference intervals were generated for each index of LA function using a reference interval calculator (RefValAdvisor) from both the RPSA and RPLA views [14]. The optimal method suggested by RefValAdvisor was used. If multiple methods were considered suitable, reference intervals from either untransformed or BoxeCox transformed data were used to provide the most narrow reference interval. To compare each index of LA function, obtained in the RPSA view for both linear and area dimensions, across the ACVIM stages, we first selected only those healthy dogs with bodyweights similar to dogs with MMVD. We used a KruskaleWallis test with post hoc multiple comparisons for these comparisons. For each KruskaleWallis test, we set the nominal p-value at 0.001 to preserve experiment-wise error rates. We also compared LA:Ao, obtained in the RPSA view for linear and area dimensions, across the ACVIM stages using the same statistical methodology. We did not perform
statistical analyses of indices of LA function across the ACVIM stages from the RPLA views. To examine the diagnostic utility of indices of LA function in differentiating dogs with CHF (ACVIM stage C) from dogs with subclinical MMVD (ACVIM stage B2), we first performed receiver operating characteristic (ROC) analysis using the entire cohort of dogs from both groups. We also performed this analysis for LA:Ao data. We then examined the indices of LA function to identify the best discriminators by comparing the areas under the ROC curves for the two best discriminating variables. Again, we examined only the indices of LA function from RPSA views. However, since ACVIM stage B2 dogs have a range of disease, extending from ‘almost normal’ to ‘severely affected, but not yet in CHF’, simply comparing the 2 groups would not be useful. Instead, we were interested to determine if indices of LA function could discriminate ACVIM stage B2 and ACVIM stage C dogs with similar disease severitydi.e. where simply evaluating the LA:Ao would not be helpful. Therefore, we selected a subset of ACVIM stage B2 and ACVIM stage C dogs with overlapping LA:Ao measurements, by selecting those dogs in ACVIM stage B2 that had a LA:Ao equal to or greater than the smallest LA:Ao in the dogs in ACVIM stage C, and selecting those dogs in ACVIM stage C that had a LA:Ao equal to or smaller than the largest LA:Ao of dogs in ACVIM stage B2. We then performed ROC analysis on this subset of dogs, to see if indices of LA function could separate out these 2 clinical groups. Because most of the ACVIM stage C dogs were stable and controlled with medications, we
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
5 between indices of LA function and either LA:Ao or bodyweight by simple linear regression.
Table 2 Signalment characteristics of the study populations.
Age (months); median [range] Sex Female intact Female spayed Male intact Male neutered Weight (kg); median [range]
Healthy dogs (n ¼ 52)
Dogs with MMVD (n ¼ 88)
48 [12, 144]
124 [36, 204]
22 11 13 6 14 [5, 54]
Results Population characteristics and reference intervals for indices of left atrial function
16 18 37 17 11 [2.5, 30]
Signalment characteristics for the 52 healthy and 88 diseased dogs are presented in Table 2. The reference intervals for indices of LA function in the 52 healthy dogs are displayed in Table 3.
MMVD, myxomatous mitral valve disease.
Left atrial function changes with worsening mitral valve disease examined whether the indices of LA function differed between the subset of ACVIM stage C dogs receiving diuretics (and other drugs) and those who had yet to receive any medications, using a Rank Sum Test. We set the nominal p-value for these comparisons at 0.05. Finally, after confirming that assumptions were not violated, we examined the relationship
Table 3
Left atrial functional indices determined by four different methods in 52 healthy dogs.
Linear indices
Area indices
Linear Area
All indices of LA function, except for LA passive emptying fraction (p¼0.4), decreased with increasing ACVIM stage of mitral valve disease, most often when dogs showed CHF (p<0.0001) (Figs. 1e4). As expected LA:Ao increased with increasing ACVIM stage of mitral valve disease (Fig. 5). Consistent with these observations,
Method/functional index
Sample size
LA RPSA view LAFT LAPAS LAACT LAEXP LA RPLA view LAFT LAPAS LAACT LAEXP LAA RPSA view LAAFT LAAPAS LAAACT LAAEXP LAA RPLA view LAAFT LAAPAS LAAACT LAAEXP LA:Ao RPSA view LA:Ao RPLA view LAA:AoA RPSA view LAA:AoA RPLA view
52
Lower reference limit (%) (95% CI)
Upper reference limit (%) (95% CI)
5 (4e9) 5 (6 to 0) 0 (2 to 6) 5 (4e10)
40 35 27 68
(33e41) (16e42) (25e27) (50e70)
10 (9e13) 1 (1 to 3) 1 (2 to 4) 11 (10e15)
33 19 24 50
(30e34) (17e21) (22e26) (43e52)
22 (19e24) 0 (0e4) 10 (7e13) 23 (16e29)
51 (48e54) 39 (29e43) 39 (36e42) 94 (87e100)
28 (24e31) 7 (7 to 3) 13 (9e17) 34 (33e39) 1.1 (1e1.2) 1.4 (1.3e1.5) 2 (2e2.2) 1.6 (1.6e2.1)
56 (53e58) 42 (32e43) 43 (40e46) 126 (107e130) 1.8 (1.8e1.8) 2.1 (2.1e2.2) 5.2 (3.9e5.6) 4.6 (3.8e4.8)
51
52
51
52 51 52 51
Please refer to Table 1 for abbreviations describing indices of LA function. LA:Ao, left atrial-to-aortic ratio; LAA:AoA, left atrial area-to-aortic area ratio; RPSA, right parasternal short-axis; RPLA, right parasternal long-axis.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
6 indices of LA function, except for LA passive emptying fraction, showed weak negative relationships with increasing LA size (as determined by LA:Ao) (Supplemental Fig. I).
Indices of left atrial function are no better at discriminating dogs with CHF from subclinical dogs than LA:Ao Indices of left atrial function showed modest sensitivity and specificity for differentiating the subset of all ACVIM stage B2 dogs from ACVIM stage C (CHF) dogs (Table 4). LA:Ao performed as well or better than any index of LA function in identifying dogs with CHF, when comparing areas under the curve of the ROC analysis (p<0.01). We then selected the subsets of ACVIM stage B2 and ACVIM stage C dogs that had overlapping LA:Ao, between 1.9 and 2.7, and examined diagnostic utility of left atrial active emptying fraction (LAACT), left atrial expansion index and LA:Ao (or the area counterparts e LAA:AoA, left atrial active emptying fraction (based on area measurements; LAAACT) and left atrial passive emptying fraction
D. Dickson et al. (based on area measurements)dbetween LAA:AoA 4.4 and 9.9). With both linear and area indices of LA function or size, neither of the 2 indices differentiated the 2 groups any better than LA size (LA:Ao or LAA:AoA; comparison of AUCs resulted in p>0.7, Table 5).
Dogs with CHF with or without therapy have similar indices of LA function We could find no differences in any indices of LA function (using either linear or area dimensions) when comparing the 24 dogs with CHF that were receiving therapy and the 5 dogs with CHF that had not yet been administered diuretics (p>0.8 for all comparisons).
Left atrial total emptying fraction and left atrial expansion index are weakly associated with bodyweight Both left total atrial emptying fraction and left atrial expansion index (slopes of 0.25 and 0.4, respectively, r2 ¼ 0.1 and p<0.03 for both
Figure 1 Left atrial (LA) active emptying fraction in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05).
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
7
Figure 2 Left atrial (LA) expansion index in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05).
Figure 3 Left atrial (LA) passive emptying fraction in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05). Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
8
D. Dickson et al.
Figure 4 Left atrial (LA) total fractional emptying in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05).
analyses), but not LAACT (p¼0.09) or left atrial passive emptying fraction (p¼0.3), showed weak negative associations with bodyweight in healthy dogs (Supplemental Fig. II).
Discussion Our study provides reference intervals for linear and area-based indices of left atrial function in dogs obtained from 2D echocardiograms. In addition, we have examined the effect of worsening mitral valve disease on 2D echocardiographic indices of LA function and, finally, the diagnostic utility of these indices in differentiating dogs with and without CHF secondary to MMVD. Our data suggest that while LA function worsens with increasing disease severity, whether using linear or area estimates, none of the indices of LA function differ sufficiently (when using 2D echocardiographic methods) between subclinical dogs and dogs with CHF to provide additional discriminatory information over a simple estimate of LA size. Our data support those of other investigators who have evaluated LA function in healthy dogs and dogs with MMVD [3]. Similar to findings by
Hollmer et al. [3], indices of LA function worsened with progressively more severe disease (based on ACVIM stages), except for conduit function, which, as might be expected, did not change. Likewise, Tidholm et al. [4] found that LA ejection fraction (a volumetric estimate of LA function) decreased as LA size increased in dogs with MMVD. Our findings are also consistent with findings in humans with increasing left atrial size [15], although we could not identify studies of LA function in humans with worsening mitral valve disease (most likely because these patients undergo valve repair before noticeable decline in LA function). However, our findings contrast those of Hollmer et al. [3], who suggested a strong discriminating ability of indices of LA function in identifying dogs with CHF. We observed that simple indices of LA function fail to differentiate dogs with subclinical MMVD and MMVD with CHF any better than a simple measurement of LA:Ao. We found that most indices of LA function were more specific, but less sensitive for identifying CHF than the corresponding LA:Ao measurement, but overall diagnostic accuracy did not differ. We suspect that the diagnostic test performance observed by Hollmer et al. [3] reflects the study
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
Figure 5 Left atrial-to-aortic annulus ratios in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear [LA:Ao] and (B) area [LAA:AoA] measurements. Categories with different letters above the box plot are different (p<0.05).
populations being compareddHollmer et al. [3] grouped all ACVIM stage B2 dogs and compared them to dogs with CHF. To confirm our suspicions, we examined the data from Hollmer et al. [3] and observed that ACVIM stage B2 and ACVIM stage C dogs could be effectively discriminated by simply evaluating the indexed minimal or maximal LA volume (which is essentially estimated by LA:Ao); indeed, LA volumes provided a better discrimination of subclinical and clinical MMVD than any functional indices in their study. The real value of indices of LA function as diagnostic tests in identifying dogs with MMVD and CHF depends on their ability to discriminate dogs with otherwise similar presentations, i.e. dogs in which the LA sizes are similar, but clinical presentations differ. Our study illustrates that in such a population of dogs (Table 4), 2D echocardiographic indices of LA function offered no better a discriminatory performance than did LA:Ao or LAA:AoA. The fact that LA:Ao still differentiated these two populations indicates that although the range of
9 LA:Ao were completely overlapping, the populations still had different LA sizes. Complicating this analysis was the small number of dogs with LA:Ao in the range of 1.8e2.7 in both ACVIM stages (B2 and C), as evidenced by the large confidence intervals around the point estimates of sensitivity and specificity. It is possible that more complex methods of estimating LA function might perform better than simple 2D echocardiographic estimates. Indeed, some of the investigators in this study have found that LA speckle tracking might provide estimates of LA function that can discriminate between dogs with subclinical MMVD and dogs with CHF.j We could find no differences in indices of LA function in dogs with untreated and stable CHF. However, we only had a small subset of dogs with untreated CHF at the time of echocardiographydtherefore, our ability to detect a difference, unless dramatic, is limited. When we examined the distribution of the functional index values of the 5 untreated CHF dogs and compared them to the treated CHF population, we could detect no trend or grouping at the lower limit of each index (Supplemental Fig. III). Nevertheless, additional studies with larger sample populations of treated and untreated dogs with CHF (or before/after studies) would be warranted to determine the effect of CHF therapy on LA functional indices. Our finding that bodyweight correlated weakly and negatively with total LA emptying fraction and LA reservoir function contrasts findings by LeBlanc et al. [2] and Hollmer et al. [7] who failed to observe any relationship with indices of LA function and bodyweight in healthy dogs, despite having dogs that spanned a similar range of bodyweights. However, Hollmer et al. [7] found a slight negative association between bodyweight (loosely based on increasing size of different breeds) and LA expansion index, similar to our observations. We did not examine intra- or inter-observer variability, because the study was not designed in a way to allow all aspects of variability to be examined. However, other studies have suggested that this variability is modest (<20%) for most echocardiographic estimates of LA size [2,16]. Given that all the measurements of LA linear and area dimensions were performed by cardiologists or residents under the supervision of cardiologists,
j Caivano D, Patata V, Birettoni F, Giorgi ME, Rishniw M, Porciello F. Left atrial function determined by 2-dimensional feature tracking echocardiography identifies dogs with congestive heart failure secondary to mitral valve disease. J Vet Intern Med 2015;29:471.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
10
D. Dickson et al.
Table 4 Diagnostic performance of select left atrial functional indices and left atrial-to-aortic ratio in differentiating dogs with subclinical (ACVIM stage B2) and dogs with clinical (ACVIM stage C) mitral valve disease. LA functional index Linear RPSA LAACT RPSA LAFT RPSA LAEXP LA:Ao Area RPSA LAAACT RPSA LAAFT RPSA LAAEXP LAA:AoA
Sample size
Threshold value
Sensitivity (95% CI)
Specificity (95% CI)
AUC
64 10.5 18.6 22.9 >2.2
77 90 90 87
(58e90) (74e98) (74e98) (69e96)
68 38 38 76
(50e83) (22e56) (22e56) (59e89)
0.71 0.64 0.64 0.88
12.4 19.5 24.3 >5.6
69 69 69 92
(48e86) (48e86) (48e86) (75e99)
93 90 90 73
(78e99) (74e98) (74e98) (54e88)
0.87 0.85 0.85 0.86
56
Please refer to Tables 1 and 3 for abbreviations describing indices of left atrial function. AUC, area under the curve. Analyses of linear indices included 34 ACVIM B2 dogs and 30 ACVIM C dogs. Analyses of area indices included 30 ACVIM B2 dogs and 26 ACVIM C dogs.
Table 5 Diagnostic performance of select left atrial functional indices and atrial size in differentiating dogs with subclinical (ACVIM stage B2) and clinical (ACVIM stage C) mitral valve disease with similarly sized left atria. LA functional index Linear RPSA LAACT RPSA LAEXP LA:Ao Area RPSA LAAACT RPSA LAAEXP LAA:AoA
Sample size
Threshold value
Sensitivity (95% CI)
Specificity (95% CI)
AUC
8.4 14.1 >2.2
75 (48e93) 53 (25e81) 77 (46e95)
79 (59e92) 91 (72e99) 65 (43e84)
0.74 0.75 0.70
21.3 50.6 >5.6
100 (72e100) 100 (72e100) 91 (59e99)
57 (35e77) 43 (23e66) 61 (39e80)
0.79 0.77 0.74
36
34
Please refer to Tables 1 and 3 for abbreviations describing indices of left atrial function. AUC, area under the curve. Analyses of linear indices included 23 ACVIM B2 dogs and 13 ACVIM C dogs. Analyses of area indices included 23 ACVIM B2 dogs and 12 ACVIM C dogs.
we have no reason to believe that the variability in our study would differ from that of other investigators. Indeed, examination of studies by other investigators suggests a similar dispersion of data across the populations of dogs without disease, and those with MMVD [3]. To see if there was any bias between investigators’ measurements, we compared the ranges of values obtained for each of the indices of LA function in healthy dogs by the 2 investigators who submitted most of the data and found that these were almost identical (data not shown). Therefore, we would anticipate that these investigators were measuring LA function similarly.
sufficiently discriminatory to differentiate dogs with subclinical disease from those with CHF if they have similarly sized left atria.
Conflict of Interest Statement The authors do not have any conflicts of interest to disclose.
Funding No funding was obtained for this study.
Conclusions Our study provides reference intervals for indices of LA function derived from 2D echocardiographic imaging, for both linear and area dimensions in two different views. While indices of LA function worsen with worsening MMVD, no index is
Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.jvc. 2017.09.003.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
11
References [1] Johns SM, Nelson OL, Gay JM. Left atrial function in cats with left-sided cardiac disease and pleural effusion or pulmonary edema. J Vet Intern Med 2012;26:1134e9. [2] LeBlanc N, Scollan K, Sisson D. Quantitative evaluation of left atrial volume and function by one-dimensional, twodimensional, and three-dimensional echocardiography in a population of normal dogs. J Vet Cardiol 2016;18:336e49. [3] Ho ¨llmer M, Willesen JL, Tolver A, Koch J. Left atrial volume and function in dogs with naturally occurring myxomatous mitral valve disease. J Vet Cardiol 2017;19:24e34. [4] Tidholm A, Ho ¨glund K, Ha ¨ggstro ¨m J, Bodega ˚rd-Westling A, Ljungvall I. Left atrial ejection fraction assessed by realtime 3-dimensional echocardiography in normal dogs and dogs with myxomatous mitral valve disease. J Vet Intern Med 2013;27:884e9. [5] Riesen SC, Schober KE, Cervenec RM, Bonagura JD. Effects of treatment with ivabradine and atenolol on reproducibility of echocardiographic indices of left heart size and function in healthy cats. J Vet Cardiol 2012;14:323e32. [6] Caivano D, Rishniw M, Patata V, Giorgi ME, Birettoni F, Porciello F. Left atrial deformation and phasic function determined by 2-dimensional speckle tracking echocardiography in healthy dogs. J Vet Cardiol 2016;18:146e55. [7] Ho ¨llmer M, Willesen JL, Tolver A, Koch J. Left atrial volume and phasic function in clinically healthy dogs of 12 different breeds. Vet J 2013;197:639e45. [8] Atkins C, Bonagura J, Ettinger S, Fox P, Gordon S, Haggstrom J, Hamlin R, Keene B, Luis-Fuentes V, Stepien R. Guidelines for the diagnosis and treatment of canine chronic valvular heart disease. J Vet Intern Med 2009;23:1142e50.
[9] Thomas WP, Gaber CE, Jacobs GJ, Kaplan PM, Lombard CW, Moise NS, Moses BL. Recommendations for standards in transthoracic two-dimensional echocardiography in the dog and cat. Echocardiography Committee of the Specialty of Cardiology, American College of Veterinary Internal Medicine. J Vet Intern Med 1993;7:247e52. [10] Rishniw M, Erb HN. Evaluation of four 2-dimensional echocardiographic methods of assessing left atrial size in dogs. J Vet Intern Med 2000;14:429e35. [11] Dickson D, Caivano D, Patteson M, Rishniw M. The times they are a-changin’: two-dimensional aortic valve measurements differ throughout diastole. J Vet Cardiol 2016; 18:15e25. [12] Ljungvall I, Rishniw M, Porciello F, Ferasin L, Ohad DG. Murmur intensity in small-breed dogs with myxomatous mitral valve disease reflects disease severity. J Small Anim Pract 2014;55:545e50. [13] Hall DJ, Cornell CC, Crawford S, Brown DJ. Meta-analysis of normal canine echocardiographic dimensional data using ratio indices. J Vet Cardiol 2008;10:11e23. [14] Geffre ´ A, Concordet D, Braun J-P, Trumel C. Reference value advisor: a new freeware set of macroinstructions to calculate reference intervals with Microsoft Excel. Vet Clin Pathol 2011;40:107e12. [15] Blondheim DS, Osipov A, Meisel SR, Frimerman A, Shochat M, Shotan A. Relation of left atrial size to function as determined by transesophageal echocardiography. Am J Cardiol 2005;96:457e63. [16] Chetboul V, Tidholm A, Nicolle A, Sampedrano CC, Gouni V, Pouchelon JL, Lefebvre HP, Concordet D. Effects of animal position and number of repeated measurements on selected two-dimensional and M-mode echocardiographic variables in healthy dogs. J Am Vet Med Assoc 2005;227:743e7.
Available online at www.sciencedirect.com
ScienceDirect
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
-, -e-
www.elsevier.com/locate/jvc
Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease*,** David Dickson, BVetMed CertAVP(VC) a, Domenico Caivano, DVM, PhD b, Jose Novo Matos, DVM c, Nuala Summerfield, BVMS d, Mark Rishniw, BVSc, MS, PhD e,* a
HeartVets, Porthcawl, Wales, UK Department of Veterinary Medicine, University of Perugia, Perugia, Italy c Vetsuisse Faculty University of Zurich, Zurich, Switzerland d New Priory Vets Brighton, Brighton, UK e Veterinary Information Network, Davis, CA, USA b
Received 10 January 2017; received in revised form 28 August 2017; accepted 11 September 2017
KEYWORDS Canine; Cardiac; Congestive heart failure
Abstract Objectives: To provide reference intervals for 2-dimensional linear and area-based estimates of left atrial (LA) function in healthy dogs and to evaluate the ability of estimates of LA function to differentiate dogs with subclinical myxomatous mitral valve disease (MMVD) and similarly affected dogs with congestive heart failure (CHF). Animals: Fifty-two healthy adult dogs, 88 dogs with MMVD of varying severity. Methods: Linear and area measurements from 2-dimensional echocardiographs in both right parasternal long and short axis views optimized for the left atrium were used to derive estimates of LA active emptying fraction, passive emptying fraction,
*
Preliminary data from this study were presented at the 24th ECVIM-CA Congress in Mainz in 2014. A unique aspect of the Journal of Veterinary Cardiology is the emphasis of additional web-based materials permitting the detailing of procedures and diagnostics. These materials can be viewed (by those readers with subscription access) by going to http://www.sciencedirect.com/science/journal/17602734. The issue to be viewed is clicked and the available PDF and image downloading is available via the Summary Plus link. The supplementary material for a given article appears at the end of the page. To view the material is to go to http://www.doi.org and enter the doi number unique to this paper which is indicated at the end of the manuscript. * Corresponding author. E-mail addresses: [email protected] (D. Dickson), [email protected] (D. Caivano), [email protected] (J.N. Matos), [email protected] (N. Summerfield), [email protected] (M. Rishniw). **
https://doi.org/10.1016/j.jvc.2017.09.003 1760-2734/ª 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
2
D. Dickson et al. expansion index, and total fractional emptying. Differences for each estimate were compared between healthy and MMVD dogs (based on ACVIM classification), and between MMVD dogs with subclinical disease and CHF that had similar LA dimensions. Diagnostic utility at identifying CHF was examined for dogs with subclinical MMVD and CHF. Relationships with bodyweight were assessed. Results: All estimates of LA function decreased with increasing ACVIM stage of mitral valve disease (p<0.05) and showed negative relationships with increasing LA size (all r2 values < 0.2), except for LA passive emptying fraction, which did not differ or correlate with LA size (p¼0.4). However, no index of LA function identified CHF better than measurements of LA size. Total LA fractional emptying and expansion index showed modest negative correlations with bodyweight. Conclusions: Estimates of LA function worsen with worsening MMVD but fail to discriminate dogs with CHF from those with subclinical MMVD any better than simple estimates of LA size. ª 2017 Elsevier B.V. All rights reserved.
Abbreviations 2D CHF ECG LA LAACT
two-dimensional congestive heart failure electrocardiograph(y) left atrium (atrial) left atrial active emptying fraction (based on linear measurements) LA:Ao ratio of the left atrial dimension to the aortic annulus dimension LAA:AoA ratio of the left atrial area to the aortic annulus area MMVD myxomatous mitral valve disease ROC receiver operating characteristic RPSA right parasternal short axis RPLA right parasternal long axis
Left atrial (LA) function is increasingly evaluated in veterinary cardiac patients to help identify disease severity [1e6]. Left atrial function comprises 3 components: it acts as a reservoir during ventricular systole (atrial diastole), it acts as a conduit during early ventricular diastole and diastasis and as an active pump during late ventricular diastole (atrial systole) [1,4,7]. Although simple measures of LA size (e.g. left atrial-to-aortic ratio) correlate with disease severity and are predictive of clinical outcomes in dogs, clinicians often observe dogs with LA of similar size that have different clinical disease status, e.g. subclinical disease vs. congestive heart failure (CHF). Whether these dogs might have different LA function to explain their clinical status remains largely undetermined.
While 3-dimensional assessment of function might be more accurate than 2-dimensional (2D) assessment [4], most clinicians performing echocardiographic evaluations of patients do not possess technology capable of 3-dimensional imaging and do not have the time for substantial off-line analysis. Furthermore, volumetric accuracy is not required if simpler methods are precise enough to successfully differentiate groups of patients with different LA function or clinical status. Two-dimensional echocardiography is commonly available, and evaluation of function requires little, if any, off-line analysis, and is minimally time-consuming. To the authors’ knowledge, no studies have determined reference intervals for 2D echocardiographic non-volumetric indices of LA function in healthy dogs. Furthermore, few studies have evaluated indices of LA function in dogs with varying degrees of LA pathology secondary to myxomatous mitral valve disease (MMVD) or examined the ability of these indices to identify the clinical status of these dogs. These studies suggested that LA function decreases with worsening MMVD and might differentiate dogs with different clinical disease classifications [3,4]. However, these studies failed to examine the benefit of estimating LA function over simply measuring the size of the LA. In addition, these studies failed to examine the ability of functional indices to discriminate between subclinical MMVD and CHF in dogs that had similar degrees of cardiac remodeling (i.e. dogs with similarly marked LA and LV enlargement that did, or did not, have the evidence of CHF). Therefore, we sought to establish reference intervals for 2D echocardiographically derived
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs indices of LA function in healthy dogs and to examine the ability of changes in these indices to identify MMVD severity, based on the ACVIM functional classification of MMVD.
Materials and methods Subjects Echocardiograms were prospectively performed on 140 client-owned dogs: 52 healthy dogs (dogs with an unremarkable history, physical examination, and normal echocardiogram), 25 dogs with subclinical mitral valve disease and no evidence of left atrial enlargement (ACVIM stage B1), 34 dogs with subclinical mitral valve disease and LA and ventricular enlargement (ACVIM stage B2), and 29 dogs with CHF (ACVIM stage C) [8]. Dogs with CHF were classified on the basis of severe left-sided heart disease, clinical signs of tachypnea and/or dyspnea, radiographic evidence of a pulmonary interstitial pattern, and, in all-but five dogs (which had not yet been administered diuretics at the time of imaging), a clear and positive response to diuretic therapy, manifesting as a normalization of respiratory rate and character. Twenty-four dogs with CHF were receiving combinations of furosemide, pimobendan, and angiotensin-converting enzyme inhibitors. Six of the stage B2 dogs and one stage B1 dog were receiving angiotensin-converting enzyme inhibitors or pimobendan at the time of evaluation. No institutional approval was sought, as the procedure was considered part of the routine evaluation of all dogs with MMVD. Owners of healthy dogs provided informed consent for the echocardiographic imaging.
Echocardiography All dogs underwent complete routine 2D echocardiography examinations by one of 4 investigators, using a variety of ultrasound equipment,f,g,h,i and appropriate phased-array transducers (ranging from 3.0 to 8.5 MHz). All echocardiograms were performed with simultaneous electrocardiographic (ECG) recording. Dogs were unsedated, and gently restrained in right lateral recumbency for all left atrial imaging.
f
MyLab Class C, Esaote, Genova, Italy. Vivid Q, GE Medical Systems Ltd, Buckinghamshire, UK. h Vivid I, GE Medical Systems Ltd, Buckinghamshire, UK. i Vivid 7, GE Healthcare, Chicago, IL, USA. g
3 Investigators obtained and stored cineloops of the left atrium from the right parasternal long-axis 4chamber (RPLA) view, and the right parasternal short-axis (RPSA) view optimized for the aortic valve and left atrium [9]. Investigators obtained and averaged three sets of measurements from different cardiac cycles (not necessarily consecutive) of the left atrium from the stored cineloops at the following timepoints: 1. For the RPSA view: (A) Early ventricular diastole, at the first frame in which aortic valve closure was observed, corresponding to maximal left atrial diameter; (B) Late ventricular diastole, at the last frame before aortic valve opening was observed, corresponding to the minimal left atrial diameter; (C) Diastasis, at the onset of the p-wave observed on the ECG. 2. For the RPLA view: (A) Early ventricular diastole, at the last frame before opening of the mitral valve was observed, corresponding to maximal left atrial diameter; (B) Late ventricular diastole, at the first frame after mitral valve closure was observed, corresponding to the minimal left atrial diameter; (C) Diastasis, at the onset of the p-wave observed on the ECG. Two sets of measurements were made at each time point: (A) Linear LA dimension, as previously described [10]. (B) Left atrial area, as previously described [10]. The averaged linear and area measurements were then used to calculate the indices of LA function (Table 1). Three aortic measurements were obtained from the RPSA view in early diastole, at the first frame in which aortic valve closure was observed and averaged [10,11]. We used a previously described method to differentiate dogs with MMVD in ACVIM stage B1 and stage B2 [12]. Briefly, we considered a dog to have no evidence of cardiac remodeling (stage B1) if 2 of the following 3 variables were below reference limits: LA:Ao (threshold value of 1.6), weightadjusted LA linear dimension (wLA; threshold value of 1.56), weight-adjusted LV linear
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
4
D. Dickson et al. Table 1 Linear
Area
Indices of left atrial function. Function assessed
Index
Abbreviation
Formula
LA reservoir
LA expansion index
LAEXP
LA contractile
LA active emptying fraction
LAACT
LA conduit
LA passive emptying fraction
LAPAS
LA reservoir
LA total emptying fraction
LAFT
LAMAX LAMIN )100 LAMIN LAP LAMIN )100 LAP LAMAX LAP )100 LAMAX LAMAX LAMIN )100 LAMAX LAAMAX LAAMIN )100 LAAMIN LAAP LAAMIN )100 LAAP LAAMAX LAAP )100 LAAMAX LAAMAX LAAMIN )100 LAAMAX
LA reservoir
LA expansion index
LAAEXP
LA contractile
LA active emptying fraction
LAAACT
LA conduit
LA passive emptying fraction
LAAPAS
LA reservoir
LA total emptying fraction
LAAFT
LA, left atrium; LAmax, maximal left atrial dimension; LAmin, minimal left atrial dimension; LAP, left atrial dimension at the onset of the p-wave on the ECG; LAEXP, left atrial expansion index (based on linear measurements); LAACT, left atrial active emptying fraction (based on linear measurements); LAPAS, left atrial passive emptying fraction (based on linear measurements); LAFT, left total atrial emptying fraction (based on linear measurements); LAAmax, maximal left atrial area; LAAmin, minimal left atrial area; LAAP, left atrial area at the onset of the p-wave on the ECG; LAAEXP, left atrial expansion index (based on area measurements); LAAACT, left atrial active emptying fraction (based on area measurements); LAAPAS, left atrial passive emptying fraction (based on area measurements); LAAFT, left atrial total emptying fraction (based on area measurements).
dimension in diastole (wLV; threshold value of 1.95). We derived the reference thresholds for the weight-adjusted estimates of LA and LV size from a previous study by Hall et al. [13]. Dogs with any 2 of these 3 variables above the reference limits we classified as ACVIM stage B2. Therefore, rare dogs could have LA:Ao > 1.6 but still be considered ACVIM stage B1, if both weight-adjusted variables were within reference intervals.
Statistical analyses Descriptive statistics were calculated for each index of LA function and are reported as median and range. Reference intervals were generated for each index of LA function using a reference interval calculator (RefValAdvisor) from both the RPSA and RPLA views [14]. The optimal method suggested by RefValAdvisor was used. If multiple methods were considered suitable, reference intervals from either untransformed or BoxeCox transformed data were used to provide the most narrow reference interval. To compare each index of LA function, obtained in the RPSA view for both linear and area dimensions, across the ACVIM stages, we first selected only those healthy dogs with bodyweights similar to dogs with MMVD. We used a KruskaleWallis test with post hoc multiple comparisons for these comparisons. For each KruskaleWallis test, we set the nominal p-value at 0.001 to preserve experiment-wise error rates. We also compared LA:Ao, obtained in the RPSA view for linear and area dimensions, across the ACVIM stages using the same statistical methodology. We did not perform
statistical analyses of indices of LA function across the ACVIM stages from the RPLA views. To examine the diagnostic utility of indices of LA function in differentiating dogs with CHF (ACVIM stage C) from dogs with subclinical MMVD (ACVIM stage B2), we first performed receiver operating characteristic (ROC) analysis using the entire cohort of dogs from both groups. We also performed this analysis for LA:Ao data. We then examined the indices of LA function to identify the best discriminators by comparing the areas under the ROC curves for the two best discriminating variables. Again, we examined only the indices of LA function from RPSA views. However, since ACVIM stage B2 dogs have a range of disease, extending from ‘almost normal’ to ‘severely affected, but not yet in CHF’, simply comparing the 2 groups would not be useful. Instead, we were interested to determine if indices of LA function could discriminate ACVIM stage B2 and ACVIM stage C dogs with similar disease severitydi.e. where simply evaluating the LA:Ao would not be helpful. Therefore, we selected a subset of ACVIM stage B2 and ACVIM stage C dogs with overlapping LA:Ao measurements, by selecting those dogs in ACVIM stage B2 that had a LA:Ao equal to or greater than the smallest LA:Ao in the dogs in ACVIM stage C, and selecting those dogs in ACVIM stage C that had a LA:Ao equal to or smaller than the largest LA:Ao of dogs in ACVIM stage B2. We then performed ROC analysis on this subset of dogs, to see if indices of LA function could separate out these 2 clinical groups. Because most of the ACVIM stage C dogs were stable and controlled with medications, we
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
5 between indices of LA function and either LA:Ao or bodyweight by simple linear regression.
Table 2 Signalment characteristics of the study populations.
Age (months); median [range] Sex Female intact Female spayed Male intact Male neutered Weight (kg); median [range]
Healthy dogs (n ¼ 52)
Dogs with MMVD (n ¼ 88)
48 [12, 144]
124 [36, 204]
22 11 13 6 14 [5, 54]
Results Population characteristics and reference intervals for indices of left atrial function
16 18 37 17 11 [2.5, 30]
Signalment characteristics for the 52 healthy and 88 diseased dogs are presented in Table 2. The reference intervals for indices of LA function in the 52 healthy dogs are displayed in Table 3.
MMVD, myxomatous mitral valve disease.
Left atrial function changes with worsening mitral valve disease examined whether the indices of LA function differed between the subset of ACVIM stage C dogs receiving diuretics (and other drugs) and those who had yet to receive any medications, using a Rank Sum Test. We set the nominal p-value for these comparisons at 0.05. Finally, after confirming that assumptions were not violated, we examined the relationship
Table 3
Left atrial functional indices determined by four different methods in 52 healthy dogs.
Linear indices
Area indices
Linear Area
All indices of LA function, except for LA passive emptying fraction (p¼0.4), decreased with increasing ACVIM stage of mitral valve disease, most often when dogs showed CHF (p<0.0001) (Figs. 1e4). As expected LA:Ao increased with increasing ACVIM stage of mitral valve disease (Fig. 5). Consistent with these observations,
Method/functional index
Sample size
LA RPSA view LAFT LAPAS LAACT LAEXP LA RPLA view LAFT LAPAS LAACT LAEXP LAA RPSA view LAAFT LAAPAS LAAACT LAAEXP LAA RPLA view LAAFT LAAPAS LAAACT LAAEXP LA:Ao RPSA view LA:Ao RPLA view LAA:AoA RPSA view LAA:AoA RPLA view
52
Lower reference limit (%) (95% CI)
Upper reference limit (%) (95% CI)
5 (4e9) 5 (6 to 0) 0 (2 to 6) 5 (4e10)
40 35 27 68
(33e41) (16e42) (25e27) (50e70)
10 (9e13) 1 (1 to 3) 1 (2 to 4) 11 (10e15)
33 19 24 50
(30e34) (17e21) (22e26) (43e52)
22 (19e24) 0 (0e4) 10 (7e13) 23 (16e29)
51 (48e54) 39 (29e43) 39 (36e42) 94 (87e100)
28 (24e31) 7 (7 to 3) 13 (9e17) 34 (33e39) 1.1 (1e1.2) 1.4 (1.3e1.5) 2 (2e2.2) 1.6 (1.6e2.1)
56 (53e58) 42 (32e43) 43 (40e46) 126 (107e130) 1.8 (1.8e1.8) 2.1 (2.1e2.2) 5.2 (3.9e5.6) 4.6 (3.8e4.8)
51
52
51
52 51 52 51
Please refer to Table 1 for abbreviations describing indices of LA function. LA:Ao, left atrial-to-aortic ratio; LAA:AoA, left atrial area-to-aortic area ratio; RPSA, right parasternal short-axis; RPLA, right parasternal long-axis.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
6 indices of LA function, except for LA passive emptying fraction, showed weak negative relationships with increasing LA size (as determined by LA:Ao) (Supplemental Fig. I).
Indices of left atrial function are no better at discriminating dogs with CHF from subclinical dogs than LA:Ao Indices of left atrial function showed modest sensitivity and specificity for differentiating the subset of all ACVIM stage B2 dogs from ACVIM stage C (CHF) dogs (Table 4). LA:Ao performed as well or better than any index of LA function in identifying dogs with CHF, when comparing areas under the curve of the ROC analysis (p<0.01). We then selected the subsets of ACVIM stage B2 and ACVIM stage C dogs that had overlapping LA:Ao, between 1.9 and 2.7, and examined diagnostic utility of left atrial active emptying fraction (LAACT), left atrial expansion index and LA:Ao (or the area counterparts e LAA:AoA, left atrial active emptying fraction (based on area measurements; LAAACT) and left atrial passive emptying fraction
D. Dickson et al. (based on area measurements)dbetween LAA:AoA 4.4 and 9.9). With both linear and area indices of LA function or size, neither of the 2 indices differentiated the 2 groups any better than LA size (LA:Ao or LAA:AoA; comparison of AUCs resulted in p>0.7, Table 5).
Dogs with CHF with or without therapy have similar indices of LA function We could find no differences in any indices of LA function (using either linear or area dimensions) when comparing the 24 dogs with CHF that were receiving therapy and the 5 dogs with CHF that had not yet been administered diuretics (p>0.8 for all comparisons).
Left atrial total emptying fraction and left atrial expansion index are weakly associated with bodyweight Both left total atrial emptying fraction and left atrial expansion index (slopes of 0.25 and 0.4, respectively, r2 ¼ 0.1 and p<0.03 for both
Figure 1 Left atrial (LA) active emptying fraction in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05).
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
7
Figure 2 Left atrial (LA) expansion index in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05).
Figure 3 Left atrial (LA) passive emptying fraction in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05). Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
8
D. Dickson et al.
Figure 4 Left atrial (LA) total fractional emptying in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear and (B) area measurements; right parasternal long axis (C) linear, and (D) area measurements. Categories with different letters above the box plot are different (p<0.05).
analyses), but not LAACT (p¼0.09) or left atrial passive emptying fraction (p¼0.3), showed weak negative associations with bodyweight in healthy dogs (Supplemental Fig. II).
Discussion Our study provides reference intervals for linear and area-based indices of left atrial function in dogs obtained from 2D echocardiograms. In addition, we have examined the effect of worsening mitral valve disease on 2D echocardiographic indices of LA function and, finally, the diagnostic utility of these indices in differentiating dogs with and without CHF secondary to MMVD. Our data suggest that while LA function worsens with increasing disease severity, whether using linear or area estimates, none of the indices of LA function differ sufficiently (when using 2D echocardiographic methods) between subclinical dogs and dogs with CHF to provide additional discriminatory information over a simple estimate of LA size. Our data support those of other investigators who have evaluated LA function in healthy dogs and dogs with MMVD [3]. Similar to findings by
Hollmer et al. [3], indices of LA function worsened with progressively more severe disease (based on ACVIM stages), except for conduit function, which, as might be expected, did not change. Likewise, Tidholm et al. [4] found that LA ejection fraction (a volumetric estimate of LA function) decreased as LA size increased in dogs with MMVD. Our findings are also consistent with findings in humans with increasing left atrial size [15], although we could not identify studies of LA function in humans with worsening mitral valve disease (most likely because these patients undergo valve repair before noticeable decline in LA function). However, our findings contrast those of Hollmer et al. [3], who suggested a strong discriminating ability of indices of LA function in identifying dogs with CHF. We observed that simple indices of LA function fail to differentiate dogs with subclinical MMVD and MMVD with CHF any better than a simple measurement of LA:Ao. We found that most indices of LA function were more specific, but less sensitive for identifying CHF than the corresponding LA:Ao measurement, but overall diagnostic accuracy did not differ. We suspect that the diagnostic test performance observed by Hollmer et al. [3] reflects the study
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
Figure 5 Left atrial-to-aortic annulus ratios in healthy dogs and dogs with myxomatous mitral valve disease of increasing severity (ACVIM classification). Right parasternal short axis (A) linear [LA:Ao] and (B) area [LAA:AoA] measurements. Categories with different letters above the box plot are different (p<0.05).
populations being compareddHollmer et al. [3] grouped all ACVIM stage B2 dogs and compared them to dogs with CHF. To confirm our suspicions, we examined the data from Hollmer et al. [3] and observed that ACVIM stage B2 and ACVIM stage C dogs could be effectively discriminated by simply evaluating the indexed minimal or maximal LA volume (which is essentially estimated by LA:Ao); indeed, LA volumes provided a better discrimination of subclinical and clinical MMVD than any functional indices in their study. The real value of indices of LA function as diagnostic tests in identifying dogs with MMVD and CHF depends on their ability to discriminate dogs with otherwise similar presentations, i.e. dogs in which the LA sizes are similar, but clinical presentations differ. Our study illustrates that in such a population of dogs (Table 4), 2D echocardiographic indices of LA function offered no better a discriminatory performance than did LA:Ao or LAA:AoA. The fact that LA:Ao still differentiated these two populations indicates that although the range of
9 LA:Ao were completely overlapping, the populations still had different LA sizes. Complicating this analysis was the small number of dogs with LA:Ao in the range of 1.8e2.7 in both ACVIM stages (B2 and C), as evidenced by the large confidence intervals around the point estimates of sensitivity and specificity. It is possible that more complex methods of estimating LA function might perform better than simple 2D echocardiographic estimates. Indeed, some of the investigators in this study have found that LA speckle tracking might provide estimates of LA function that can discriminate between dogs with subclinical MMVD and dogs with CHF.j We could find no differences in indices of LA function in dogs with untreated and stable CHF. However, we only had a small subset of dogs with untreated CHF at the time of echocardiographydtherefore, our ability to detect a difference, unless dramatic, is limited. When we examined the distribution of the functional index values of the 5 untreated CHF dogs and compared them to the treated CHF population, we could detect no trend or grouping at the lower limit of each index (Supplemental Fig. III). Nevertheless, additional studies with larger sample populations of treated and untreated dogs with CHF (or before/after studies) would be warranted to determine the effect of CHF therapy on LA functional indices. Our finding that bodyweight correlated weakly and negatively with total LA emptying fraction and LA reservoir function contrasts findings by LeBlanc et al. [2] and Hollmer et al. [7] who failed to observe any relationship with indices of LA function and bodyweight in healthy dogs, despite having dogs that spanned a similar range of bodyweights. However, Hollmer et al. [7] found a slight negative association between bodyweight (loosely based on increasing size of different breeds) and LA expansion index, similar to our observations. We did not examine intra- or inter-observer variability, because the study was not designed in a way to allow all aspects of variability to be examined. However, other studies have suggested that this variability is modest (<20%) for most echocardiographic estimates of LA size [2,16]. Given that all the measurements of LA linear and area dimensions were performed by cardiologists or residents under the supervision of cardiologists,
j Caivano D, Patata V, Birettoni F, Giorgi ME, Rishniw M, Porciello F. Left atrial function determined by 2-dimensional feature tracking echocardiography identifies dogs with congestive heart failure secondary to mitral valve disease. J Vet Intern Med 2015;29:471.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
10
D. Dickson et al.
Table 4 Diagnostic performance of select left atrial functional indices and left atrial-to-aortic ratio in differentiating dogs with subclinical (ACVIM stage B2) and dogs with clinical (ACVIM stage C) mitral valve disease. LA functional index Linear RPSA LAACT RPSA LAFT RPSA LAEXP LA:Ao Area RPSA LAAACT RPSA LAAFT RPSA LAAEXP LAA:AoA
Sample size
Threshold value
Sensitivity (95% CI)
Specificity (95% CI)
AUC
64 10.5 18.6 22.9 >2.2
77 90 90 87
(58e90) (74e98) (74e98) (69e96)
68 38 38 76
(50e83) (22e56) (22e56) (59e89)
0.71 0.64 0.64 0.88
12.4 19.5 24.3 >5.6
69 69 69 92
(48e86) (48e86) (48e86) (75e99)
93 90 90 73
(78e99) (74e98) (74e98) (54e88)
0.87 0.85 0.85 0.86
56
Please refer to Tables 1 and 3 for abbreviations describing indices of left atrial function. AUC, area under the curve. Analyses of linear indices included 34 ACVIM B2 dogs and 30 ACVIM C dogs. Analyses of area indices included 30 ACVIM B2 dogs and 26 ACVIM C dogs.
Table 5 Diagnostic performance of select left atrial functional indices and atrial size in differentiating dogs with subclinical (ACVIM stage B2) and clinical (ACVIM stage C) mitral valve disease with similarly sized left atria. LA functional index Linear RPSA LAACT RPSA LAEXP LA:Ao Area RPSA LAAACT RPSA LAAEXP LAA:AoA
Sample size
Threshold value
Sensitivity (95% CI)
Specificity (95% CI)
AUC
8.4 14.1 >2.2
75 (48e93) 53 (25e81) 77 (46e95)
79 (59e92) 91 (72e99) 65 (43e84)
0.74 0.75 0.70
21.3 50.6 >5.6
100 (72e100) 100 (72e100) 91 (59e99)
57 (35e77) 43 (23e66) 61 (39e80)
0.79 0.77 0.74
36
34
Please refer to Tables 1 and 3 for abbreviations describing indices of left atrial function. AUC, area under the curve. Analyses of linear indices included 23 ACVIM B2 dogs and 13 ACVIM C dogs. Analyses of area indices included 23 ACVIM B2 dogs and 12 ACVIM C dogs.
we have no reason to believe that the variability in our study would differ from that of other investigators. Indeed, examination of studies by other investigators suggests a similar dispersion of data across the populations of dogs without disease, and those with MMVD [3]. To see if there was any bias between investigators’ measurements, we compared the ranges of values obtained for each of the indices of LA function in healthy dogs by the 2 investigators who submitted most of the data and found that these were almost identical (data not shown). Therefore, we would anticipate that these investigators were measuring LA function similarly.
sufficiently discriminatory to differentiate dogs with subclinical disease from those with CHF if they have similarly sized left atria.
Conflict of Interest Statement The authors do not have any conflicts of interest to disclose.
Funding No funding was obtained for this study.
Conclusions Our study provides reference intervals for indices of LA function derived from 2D echocardiographic imaging, for both linear and area dimensions in two different views. While indices of LA function worsen with worsening MMVD, no index is
Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.jvc. 2017.09.003.
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003
Left atrial function in dogs
11
References [1] Johns SM, Nelson OL, Gay JM. Left atrial function in cats with left-sided cardiac disease and pleural effusion or pulmonary edema. J Vet Intern Med 2012;26:1134e9. [2] LeBlanc N, Scollan K, Sisson D. Quantitative evaluation of left atrial volume and function by one-dimensional, twodimensional, and three-dimensional echocardiography in a population of normal dogs. J Vet Cardiol 2016;18:336e49. [3] Ho ¨llmer M, Willesen JL, Tolver A, Koch J. Left atrial volume and function in dogs with naturally occurring myxomatous mitral valve disease. J Vet Cardiol 2017;19:24e34. [4] Tidholm A, Ho ¨glund K, Ha ¨ggstro ¨m J, Bodega ˚rd-Westling A, Ljungvall I. Left atrial ejection fraction assessed by realtime 3-dimensional echocardiography in normal dogs and dogs with myxomatous mitral valve disease. J Vet Intern Med 2013;27:884e9. [5] Riesen SC, Schober KE, Cervenec RM, Bonagura JD. Effects of treatment with ivabradine and atenolol on reproducibility of echocardiographic indices of left heart size and function in healthy cats. J Vet Cardiol 2012;14:323e32. [6] Caivano D, Rishniw M, Patata V, Giorgi ME, Birettoni F, Porciello F. Left atrial deformation and phasic function determined by 2-dimensional speckle tracking echocardiography in healthy dogs. J Vet Cardiol 2016;18:146e55. [7] Ho ¨llmer M, Willesen JL, Tolver A, Koch J. Left atrial volume and phasic function in clinically healthy dogs of 12 different breeds. Vet J 2013;197:639e45. [8] Atkins C, Bonagura J, Ettinger S, Fox P, Gordon S, Haggstrom J, Hamlin R, Keene B, Luis-Fuentes V, Stepien R. Guidelines for the diagnosis and treatment of canine chronic valvular heart disease. J Vet Intern Med 2009;23:1142e50.
[9] Thomas WP, Gaber CE, Jacobs GJ, Kaplan PM, Lombard CW, Moise NS, Moses BL. Recommendations for standards in transthoracic two-dimensional echocardiography in the dog and cat. Echocardiography Committee of the Specialty of Cardiology, American College of Veterinary Internal Medicine. J Vet Intern Med 1993;7:247e52. [10] Rishniw M, Erb HN. Evaluation of four 2-dimensional echocardiographic methods of assessing left atrial size in dogs. J Vet Intern Med 2000;14:429e35. [11] Dickson D, Caivano D, Patteson M, Rishniw M. The times they are a-changin’: two-dimensional aortic valve measurements differ throughout diastole. J Vet Cardiol 2016; 18:15e25. [12] Ljungvall I, Rishniw M, Porciello F, Ferasin L, Ohad DG. Murmur intensity in small-breed dogs with myxomatous mitral valve disease reflects disease severity. J Small Anim Pract 2014;55:545e50. [13] Hall DJ, Cornell CC, Crawford S, Brown DJ. Meta-analysis of normal canine echocardiographic dimensional data using ratio indices. J Vet Cardiol 2008;10:11e23. [14] Geffre ´ A, Concordet D, Braun J-P, Trumel C. Reference value advisor: a new freeware set of macroinstructions to calculate reference intervals with Microsoft Excel. Vet Clin Pathol 2011;40:107e12. [15] Blondheim DS, Osipov A, Meisel SR, Frimerman A, Shochat M, Shotan A. Relation of left atrial size to function as determined by transesophageal echocardiography. Am J Cardiol 2005;96:457e63. [16] Chetboul V, Tidholm A, Nicolle A, Sampedrano CC, Gouni V, Pouchelon JL, Lefebvre HP, Concordet D. Effects of animal position and number of repeated measurements on selected two-dimensional and M-mode echocardiographic variables in healthy dogs. J Am Vet Med Assoc 2005;227:743e7.
Available online at www.sciencedirect.com
ScienceDirect
Please cite this article in press as: Dickson D, et al., Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease, Journal of Veterinary Cardiology (2017), https://doi.org/10.1016/ j.jvc.2017.09.003