ecocardio indices and severity of mitral valve regurgitation in dogs with preclinical DMVD

SAVE OFFLINE

Received: 26 November 2018

Accepted: 6 February 2019

DOI: 10.1111/jvim.15461

STANDARD ARTICLE

Echocardiographic indices and severity of mitral regurgitation in dogs with preclinical degenerative mitral valve disease Éva Larouche-Lebel1† | Kerry A. Loughran1† | Mark A. Oyama1,2 1 Department of Clinical Sciences and Advanced Medicine, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania

Abstract Background: Describing severity of mitral regurgitation (MR) in dogs with degenerative mitral valve disease (DMVD) is challenging.

2

Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania

Hypothesis/Objectives: Mitral regurgitant fraction (RF), effective regurgitant orifice area (EROA),

Correspondence Mark A. Oyama, 3900 Delancey St, Philadelphia, PA 19104. Email: [email protected]

Animals: Fifty-seven dogs with preclinical DMVD including 36 without and 21 with cardiomegaly.

and the ratio of mitral regurgitant to aortic flow (QMR:QAo) can be calculated from routine echocardiographic measurements and provide additional information regarding MR severity. Methods: Prospective observational study. The expected relationships among RF, EROA, and QMR:QAo and 1-dimensional measurements including left atrium to aortic root diameter ratio (LA: Ao) and normalized left ventricular internal dimension at end-diastole (LVIDdN) were mathematically derived and calculated using echocardiographic data from the study population. Nonlinear goodness of fit was determined by calculation of the root mean standard error. The correlations between 1-dimensional and multidimensional indices were analyzed using receiver operating characteristic curves. Results: The relationships among RF, EROA, QMR:QAo, and both LA:Ao and LVIDdN were curvilinear, and the multidimensional indices differentiated MR of variable severity. By contrast, 1-dimensional measurements were insensitive to MR severity until RF equaled or exceeded 50%. Regurgitant fraction ≥50%, EROA to body surface area ≥0.347 and QMR:QAo ≥0.79 were strongly associated with LA:Ao ≥1.6 and LVIDdN ≥1.7. Conclusions and Clinical Importance: Regurgitant fraction, EROA, and QMR:QAo quantify MR severity in dogs with preclinical DMVD in a manner that 1-dimensional measurements do not. KEYWORDS

degenerative mitral valve disease, heart disease, myxomatous degeneration, regurgitant fraction

Abbreviations: Amax, peak left ventricular late filling velocity; AoD, aortic diameter; AoDN, normalized diameter of the aorta; AUC, area under the curve; BSA, body surface area; BW, body weight; CHF, congestive heart failure; CSA, cross-sectional area; DAo-STJ, diameter of the aorta at the level the sinotubular junction; DMVD, degenerative mitral valve disease; E:A, ratio of the peak early to late left ventricular filling velocity; Emax, peak left ventricular early filling velocity; EROA, effective regurgitant orifice area; FS, fractional shortening; FSV, forward stroke volume; ICC, intraclass correlation coefficient; IVSdN, normalized thickness of the interventricular septum at end-diastole; LA:Ao, left atrium to aortic root diameter ratio; LAD, left atrial diameter; LADN, normalized diameter of the left atrium; LV, left ventricular; LVET, left ventricular ejection time; LVIDdN, normalized left ventricular internal diameter at end-diastole; LVIDsN, normalized left ventricular internal diameter at end-systole; LVPWdN, normalized thickness of the left ventricular posterior wall at end-diastole; LVVd, left ventricular volume at end-diastole; LVVs, left ventricular volume at end-systole; MR VOL, mitral regurgitation volume; MR, mitral regurgitation; OR, odds ratio; PISA, proximal isovelocity surface area; QAo, aortic flow rate; QMR, mitral regurgitation flow rate; QMR:QAo, ratio of mitral regurgitation to aortic flow rate; RF, regurgitant fraction; RMSE, root mean standard error; ROC, receiveroperating characteristic; SBP, systolic blood pressure; TOT SV, total stroke volume; VTIAo-STJ, velocity time integral of aortic flow signal at the level of the sinotubular junction; VTIMI, velocity time integral of the mitral inflow signal; VTIMR, velocity time integral of the continuous wave Doppler mitral regurgitation jet. †

Éva Larouche-Lebel and Kerry A. Loughran contributed equally to this study.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2019 The Authors. Journal of Veterinary Internal Medicine published by Wiley Periodicals, Inc. on behalf of the American College of Veterinary Internal Medicine. J Vet Intern Med. 2019;1–10.

wileyonlinelibrary.com/journal/jvim

1

2

LAROUCHE-LEBEL ET AL.

enzyme inhibitors, positive inotropes, or vasodilators, administration of

1 | I N T RO D UC T I O N

parenteral fluids within the past 72 hours, evidence of significant heart Degenerative mitral valve disease (DMVD) in the dog is characterized by

disease other than DMVD, such as moderate or severe pulmonary

age-related morphological and structural changes in the valve that lead

hypertension or aortic insufficiency, or presence of significant extra-

1

to mitral regurgitation (MR). The MR caused by DMVD has been called

cardiac disease that might influence fluid balance or hemodynamics (ie,

“unrelentingly although variably progressive” and as such, stratification

hyperadrenocorticism, diabetes mellitus, chronic kidney disease, etc.).

of MR severity in both humans and dogs with DMVD is a subject of con-

Body surface area (BSA) in m2 was calculated using the following for-

siderable clinical interest.3 Specifically, determination of which dogs with

mula: BSA = (body weight (kg)0.67 × 10.1)/100.20 Indirect systolic blood

preclinical or asymptomatic MR will develop cardiomegaly is an impor-

pressure (SBP) was measured using the Doppler method (Doppler Flow

tant clinical question, particularly in light of the fact that treatment in this

Detector Model 811-B, Parks Medical Electronics, Inc., Aloha, Oregon).

2

cohort improves outcome.4 A substantial proportion of both dogs5–8

Echocardiograms (iE33, Philips Healthcare, Cambridge, Massachu-

remain in the preclinical stage of DMVD for extended

setts) were performed by a board-certified cardiologist or a resident in

periods of time whereas others experience progressive disease and clini-

training. Normalized LV internal dimension at end-diastole (LVIDdN)

cal development of congestive heart failure (CHF). To date, little data

and at end-systole (LVIDsN) as well as interventricular septal (IVSdN)

examining risk of cardiomegaly in dogs with early preclinical DMVD

and LV posterior wall thickness (LVPWdN) at end-diastole were calcu-

exist.

lated from LV M-mode or 2-D images obtained from the right short or

9,10

and humans

In humans with DMVD, echocardiography is widely used to assess

long axis views.21 The specific imaging plane used for measurement

however, in contrast to common practice in veterinary

was based on what the examiner regarded as the most accurate and

species, relatively little weight is given to 1-dimensional chamber sizes

representative view of the LV. Left atrial diameter (LAD), aortic root

such as left atrial or left ventricular (LV) diameter. Rather, severity of MR

dimension (AoD), and their ratio (LA:Ao) were measured from the 2D

is routinely assessed by multidimensional indices such as the mitral effec-

right short axis view.22 Normalized values for LAD (LADN) and AoD

tive regurgitant orifice area (EROA) and regurgitant fraction (RF) that

(AoDN) were calculated using previously reported formulae.21 Peak

take into account not only diastolic heart size, but importantly forward

early (Emax) and late (Amax) mitral inflow velocity were measured at

11

severity of MR,

Mitral EROA corresponds to

the tips of the mitral valve leaflets from the left apical view and their

the cross-sectional area (CSA) of the narrowest region (ie, the vena con-

ratio (E:A) was calculated. The velocity-time integral of aortic flow at

11–13

and mitral flow and systolic function.

tracta) of MR flow (QMR) as it moves from left ventricle to atrium during

the level of the sinotubular junction (VTIAo-STJ) and the LV ejection

LV systole, whereas mitral RF is the percentage of total LV stroke vol-

time (LVET) was measured from the pulsed wave Doppler tracing

ume (TOT SV) that regurgitates back through the mitral valve. In humans

obtained from the left apical view. The diameter of the aortic sinotub-

with DMVD, EROA ≥0.40 cm , regurgitant volume ≥60 mL, and RF

ular junction (DAo-STJ) was measured using the 2D image of the aorta

≥50% are established criteria for severe preclinical disease and increased

obtained from either the left cranial or right parasternal long axis view

2

and measured from inner edge to inner edge (ie, at the blood-tissue

12

risk for future clinical signs.

We hypothesized that EROA, RF, and ratio of mitral regurgitant to

interface). The velocity time integral of the MR flow (VTIMR) and dura-

aortic flow (QMR:QAo) would provide more information regarding MR

tion of MR flow were measured from the continuous wave Doppler

severity than 1-dimensional measures in dogs with preclinical DMVD.

tracing obtained from the left apical view.

how-

Left ventricular volume at end diastole (LVVd) and end-systole

ever, adoption into routine clinical practice is hindered by the complexity

(LVVs) were calculated using the formula for the volume of a prolate

of the echocardiographic procedures used to perform the necessary

ellipse with a length that is twice as long as the diameter (ie, volume =

measurements and calculations. Thus, we specifically sought to evaluate

diameter3), which has been previously validated primarily in the normal

relatively simple echocardiographic methods to derive EROA, RF, and

dog.23–25 Specifically, LVVd was calculated as LVIDd3 and LVVs was cal-

14–19

Previous studies have examined mitral EROA and RF in dogs,

QMR:QAo and to describe their relationships with indices of left heart enlargement.

culated as LVIDs3. Both volume measurements were subsequently indexed to body weight. The TOT SV was calculated as LVVd-LVVs. Forward stroke volume (FSV) was derived by multiplying VTIAo-STJ by the CSA of the sinotubular junction (CSA = π × (DAo-STJ/2)2). Mitral regurgi-

2 | MATERIALS AND METHODS

tant volume (MR VOL) was calculated as TOT SV-FSV. Values for QMR and aortic flow (QAo) in mL/sec were calculated as MR VOL divided by

The study was a single-site prospective observational cohort study. The

MR flow duration and FSV divided by LVET, respectively. Volume and

study utilized a convenience sample and all dogs recruited into the study

flow parameters were indexed to body weight. A unitless ratio of MR to

were already scheduled to undergo 2D, M-mode, and Doppler echocar-

aortic flow was calculated as QMR:QAo. Mitral RF was calculated as MR

diography and indirect blood pressure measurement as part of their visit

VOL divided by TOT SV and then multiplied by 100. Mitral EROA in cm2

to the Veterinary Hospital of the University of Pennsylvania. Eligible

was calculated by dividing MR VOL by VTIMR11 and then indexed to

dogs included those with a left apical systolic murmur and color flow

BSA (EROA:BSA). Representative echocardiographic measurements as

Doppler evidence of MR in association with thickened or prolapsing

well as a spreadsheet to enter and calculate mitral RF are presented as

mitral leaflets. Dogs with mild to moderate concurrent tricuspid regurgi-

Supplemental Materials. In order to validate findings based on the pro-

tation were eligible for the study. Exclusion criteria included current or

late ellipse method, the indices of interest were calculated a second time

previous episodes of CHF, receipt of diuretics, angiotensin-converting

using an alternate echocardiographic technique that did not rely on LV

3

LAROUCHE-LEBEL ET AL.

dimensions. Specifically, the product of the mitral inflow velocity time

contained a zero value.31 Finally, the AUC values related to RF, EROA,

integral (VTIMI) and the CSA of the mitral annulus measured at the tips

and QMR:QAo that were derived from the prolate ellipse method were

of the mitral valve leaflets in early diastole as recorded from the left api-

compared to AUC values that were derived from the mitral inflow

cal view (CSA = π × length between tips of the mitral leaflets/2)2) was

method.32

used to derive a value for TOT SV based on the mitral inflow technique

Significance was defined as P < .05. Data are reported as mean

(TOT SVMI).11 The TOT SVMI was then used to calculate alternate MR

(SD) or median (interquartile range). Correction for multiple compari-

VOL, RF, EROA, and QMR:QAo values specific to the mitral inflow tech-

sons was not performed. Statistical analyses and graphing were per-

nique. All echocardiographic and blood pressure measurements were

formed using commercial software (STATA 14.0, Stata Corp, College

the result of averaging 3 separate beats or readings.

Station, Texas; Prism 7.0, GraphPad, La Jolla, California; Solver plug in

Dogs were classified into 2 groups based on specific LV and atrial

for Microsoft Excel 14.7.2, Microsoft Corp, Redmond, Washington).

sizes that have been previously identified as important in the clinical

The majority of the study protocol was deemed exempt from

evolution of DMVD.4 Dogs in Group 1 included dogs that failed to meet

IACUC oversight and without need for informed owner consent based

either or both of the following criteria: LVIDdN ≥1.7 or LA:Ao ≥1.6.

on the fact that the study (1) involved a convenience sample of dogs

Group 2 included dogs that met both of the criteria: LVIDdN ≥1.7 and

that were scheduled to receive an echocardiographic exam and blood

LA:Ao ≥1.6. Data were tested for normality using the D'Agostino &

pressure measurement as part of their regularly scheduled hospital

Pearson normality test. Differences between groups were evaluated

visit, (2) the technical performance of the echocardiographic exam did

using t tests, Mann-Whitney U tests, or chi-square test. Plots of left

not deviate from what is standard practice at our hospital, and (3) any

heart size as the dependent variable and indices of MR severity as the

analysis of echocardiographic data for purposes of the current study

independent variable were constructed. Using simplifying assumptions,

was done offline. IACUC approval and owner consent was obtained

equations were derived that related 1-dimensional variables such as

for the 10 dogs that received 2 separate echocardiograms as part of

LVIDdN and LA:Ao to multidimensional variables such as RF, EROA, and

the effort to define interobserver variability.

QMR:QAo. Specifically, LVIDdN and LA:Ao were predicted to vary proportionally to RF raised to the −1/3 power (ie, Y = X−1/3 or Y = 1/X1/3) and for EROA and QMR:QAo, LVIDdN and LA:Ao were predicted to vary

3 | RE SU LT S

proportionally to the cube root of EROA or QMR:QAo (ie, Y = X1/3). Goodness of the predicted model fit to the actual data was assessed by

Echocardiography and blood pressure measurement were performed

calculation of root mean standard error (RMSE) values.

on 59 dogs. In 1 dog, satisfactory Doppler signals of the aortic outflow

Interobserver variability of the RF measurement was examined

could not be obtained. Another dog, a Bull terrier, had an unusually

in a subset of 10 dogs by having 2 individuals (Eva Larouche-Lebel

small aortic diameter (AoDN = 0.61) that was below the 2.5% value of

and Mark A. Oyama) acquire separate echocardiograms on the same

the reference range. Both of these dogs were excluded from further

dog during a single outpatient hospital visit. Interobserver variability

analysis. Thus, the analysis set included 57 dogs, including 36 dogs in

was determined by calculating the intraclass correlation coefficient

Group 1 and 21 dogs in Group 2. Table 1 displays the signalment,

(ICC) and by calculating the mean difference among observers.26 The

physical examination findings, blood pressure, and echocardiographic

strength of agreement was described as none, slight, fair, moderate, and

data of the study groups. Dogs in Group 2 had significantly greater

substantial for ICC values of 0-0.10, 0.11-0.40, 0.41-0.60, 0.61-0.81,

LVIDdN, LVIDsN, LADN, LA:Ao, Emax, Amax, and VTIMI values than

and 0.81-1.0, respectively.27

dogs in Group 1. Dogs in Group 2 had significantly lesser VTIAo-STJ

The utility of indices of MR severity to predict presence of clini-

and LVET values compared to Dogs in Group 1. There were no signifi-

cally relevant heart enlargement (ie, inclusion in Group 2) was evalu-

cant differences in age, sex, body weight, BSA, heart rate, SBP, IVSdN,

ated by construction of receiver-operating characteristic (ROC) curves

LVPWdN, fractional shortening (FS), AoDN, E:A, DAo-STJ, MR duration,

and calculation of various diagnostic metrics as follows. First, the

or VTIMR between groups. Table 2 displays mean or median echocar-

overall ability of each variable to predict patient group was assessed

diographic indices of MR severity by group calculated using the prolate

by the area under the curve (AUC). The AUC represents the probabil-

ellipse method. The LVVd, TOT SV, EROA, MR VOL, QMR, and QMR:

ity that a randomly selected subject with the condition of interest has

QAo were significantly greater in Group 2 vs Group 1. Indices of for-

a test result indicating greater suspicion of disease than a randomly

ward flow, including QAo, FSV, and cardiac index were significantly

selected individual without the condition.28 We defined AUC values

lower in Group 2 vs Group 1. Severity of MR as measured by RF was

29

whereas an AUC value

significantly higher in Group 2 (69% [10];) vs Group 1 (33% [18];

of 0.5 indicates the test is no better than chance. Second, sensitivity,

P < .0001). Agreement among observers regarding measurement of RF

specificity, positive likelihood ratio (ie, ratio of a positive test in the

using the prolate ellipse method was substantial with an ICC of 0.96.

affected vs ratio of positive test in the unaffected) and negative likeli-

The mean difference in RF values was −1% (95% CI, −6 to 4%) and this

hood ratio (ie, ratio of a negative test in the affected vs the ratio of a

value was not significantly different than a value of 0 (P = .56).

>0.85 as indicative of a clinically useful test,

negative test in the unaffected) for various variable values were calcu-

The relationships between LA:Ao and LVIDdN and RF, EROA:

lated. The diagnostic odds ratio (OR), which is the ratio of the odds of

BSA, and QMR:QAo were nonlinear and well described by the pre-

a positive test in the affected over the odds of the positive test in the

dicted regression formulae (Figure 1). As RF, EROA:BSA, and QMR:QAo

30

The probabi-

increased, LA:Ao and LVIDdN increased in a nonlinear fashion. Of

listic method was used to calculate OR in instances where any cell

note, the relationship between RF and LVIDdN and LA:Ao was

unaffected, was calculated for various variable values.

4

LAROUCHE-LEBEL ET AL.

relatively flat in dogs with small RF values and became increasingly steep as RF increased. Thus, despite the fact that there were minimal

TABLE 2

Echocardiographically derived measures in 57 dogs with preclinical degenerative mitral valve disease

to no increases in the 1-dimensional measures of heart size of dogs in

Variable

Group 1 (n = 36) Group 2 (n = 21)

P

Group 1, multidimensional echocardiographic data could be used to

LVVd:BW (mLs/kg)

3.18 (2.55-3.68)

4.99 (4.26-6.44)