Transthoracic echocardiography has been used in veterinary medicine as a noninvasive method for detailed evaluation of cardiac anatomy, function, and hemodynamics since the late 1970s.1,2 Dachshunds are predisposed to MMVD, and a polygenic mode of in heritance for MVP has been suggested.3–6 Mitral valve regurgitation secondary to MMVD is the most common cause of left atrial enlargement in small-breed dogs.7–9 Some echocardiographers prefer to measure left atrial and aortic dimensions by 2-D echocardiographic methods via the RPSA window instead of the M-mode method because the M-mode method has inherent limitations for obtaining reliable measurements of the maximum diameter of the left atrium owing to the potential for measuring the left auricle instead of the left atrial body.10,11
In clinically normal CKCSs, the mean (SD) LA:Ao is 1.03 (0.09) as determined by the use of 2-D echocardiography via the RPSA view at the level of the aortic cusps,10 which varies substantially from the upper limit for the LA:Ao (1.6) described for other breeds.11 In a large prospective study12 that involved 558 dogs with MMVD from 36 breeds, a LA:Ao > 1.7 was significantly associated with cardiac-related death. Despite the fact that multiple studies3–6 have been performed on Dachshunds with MMVD, there is a paucity of information regarding the LA:Ao for clinically normal dogs of that breed.
To date, approximately 20 studies13–32 have been published that contain reference ranges for M-mode echocardiographic measurements for specific dog breeds. Critics have questioned the usefulness of some of those reference ranges because of the small number of dogs used to calculate them.33,34 In 2004, the allometric scaling method was adapted to establish M-mode prediction intervals for all clinically normal dogs ranging in body weight from 2.5 to 90 kg regardless of breed.34 The allometric scaling 95% prediction intervals were calculated on the basis of M-mode measurements obtained from 494 dogs of 8 breeds.34 The allometric scaling equation is defined as Y = aMb, where Y represents each particular M-mode measurement, M is body weight, a is the proportionality constant, and b is the scaling exponent.34,35 Because dogs with chronic volume overload secondary to MMVD often have changes in left ventricular geometry,a many investigators have used the allometric scaling M-mode 95% prediction intervals to assess the extent of left ventricular geometric changes.
Echocardiographic measurements vary significantly among breeds of dogs, even after they are corrected for body weight or body surface area. The investigators of multiple studies18,25,36,37 report that breed and body conformation also influence the echocardiographic measurements of dogs. The primary objectives for the study reported here were to determine the LA:Ao and to establish 95% prediction intervals for left ventricular M-mode transthoracic echocardiographic measurements for clinically normal adult Dachshunds and compare those measurements with the left ventricular M-mode transthoracic echocardiographic 95% prediction intervals established by use of a multibreed allometric scaling approach.34 Because Dachshunds are characterized as chondrodysplastic dwarfs with a slightly rounded thoracic conformation and very short limbs, we hypothesized that the LA:Ao would differ significantly from that established for CKCSs. We also hypothesized that the transthoracic left ventricular M-mode prediction intervals established for clinically normal adult Dachshunds would differ from those derived from the multibreed allometric scaling approach.
Materials and Methods
Animals
Fifty-five Dachshunds between 1 and 7 years old and weighing between 5 and 12 kg from local breeders and clients, faculty, and staff of the Onderstepoort Veterinary Academic Hospital were recruited for the study during the 8-month period from December 2011 to July 2012. Owner consent was obtained for all dogs enrolled in the study. All study procedures were approved by and carried out in accordance with requirements established by the Institutional Animal Care Committee of the University of Pretoria.
For each dog, sex (including neuter status), body condition score (as determined on a 5-point scale38), and thoracic girth (at the level just caudal to the scapula) were recorded. Each dog was initially screened by means of a physical examination, cardiac auscultation, and a cursory echocardiographic examination. The presence or absence of MVP and MR was determined on the basis of evaluation of the right parasternal long-axis 4-chamber view and color flow Doppler echocardiography. Dogs that had previously diagnosed cardiac disease, were receiving any type of cardiotoxic medication, had a MVP > 1 mm, and had any MR detected on color flow Doppler echocardiography were excluded from the study. Dogs that passed the initial screening underwent advanced screening that included a CBC, peripheral blood smear evaluation, thoracic radiography (right lateral and dorsoventral views), resting ECG examination for 5 minutes, and 3 repeated systolic blood pressure measurements obtained by Doppler ultrasound.
Echocardiographic measurements
All echocardiographic evaluations were performed by use of an ultrasound machineb with a phased-array cardiac transducer (frequency, 2.6 to 7.7 MHz) equipped with AMM software. A complete echocardiographic examination was performed on each dog by the same investigator (CKL), who was a second-year European College of Veterinary Diagnostic Imaging resident with 6 years of clinical echocardiography experience at the time the study was conducted.
Three recommended standard 2-D echocardiographic methods11 (diameter, circumference, and the cross-sectional area) were used to acquire measurements of the left atrium and aorta from images obtained via the standard RPSA window at the level of the aortic valve where the commissure of the valve cusps were visible. The internal short-axis diameter of the aorta along the commissure between the noncoronary and right coronary aortic valve cusps was measured on the first frame following aortic valve closure, whereas the left atrial chamber size was obtained by measuring the line that extended from and parallel to the commissure between the noncoronary and left coronary aortic valve cusps to the distant margin of the left atrium in the same frame. In images where a pulmonary vein entered the left atrium at the caudolateral aspect, the edge of the left atrium was approximated by joining the visible edges of the left atrium with an imaginary curved line. The internal short-axis circumferences and internal cross-sectional areas of the left atrium and aorta were obtained by tracing their outlines on the same images and frames that were used to measure the respective diameters and automatically calculated by built-in ultrasound machine software. Echocardiographic images of satisfactory quality were stored in cineloop, and each variable was measured 3 times in different, but not necessarily consecutive, heart cycles. Software within the ultrasound machine recorded all measurements for the left atrium and aorta and calculated the LA:Ao and the mean for each variable.
M-mode echocardiographic measurements of the left ventricle were acquired in accordance with standard recommendations.39 Left ventricular measurements were obtained with the dog in right lateral recumbency by use of AMM echocardiography and measured on images obtained via the RPSA window at the chordal level, immediately below the mitral valve (Figure 1). The mean of 3 repeated measurements was recorded for IVSd, LVIDd, LVFWd, IVSs, LVIDs, and LVFWs. Additionally, the mean for 3 fractional shortening percentages was automatically calculated by software within the ultrasound machine; a simultaneous lead II ECG tracing was recorded to ensure that the R-R interval for the 3 repeated measurements did not vary by > 20%. All M-mode measurements were acquired by use of the leading edge of myocardial borders as described.40 The mean was also calculated for 3 repeated measurements of EPSS, which was measured from the maximum opening of the septal mitral valve leaflet in early diastole to the interventricular septum.41
Representative left ventricular M-mode transthoracic echocardiographic image obtained via the RPSA window (A) and M-mode tracing of the left ventricle with R-R interval time measurements (B) for a clinically normal adult Dachshund. A—The accuracy of the left ventricular measurements was maximized by the use of an AMM-generated sampling line (solid line) against which the short axis of the left ventricle was aligned perpendicularly to ensure that the ventricle was consistently bisected into 2 equal halves. Perfect perpendicular alignment of the short axis with a conventional M-mode sampling line (dashed line) can be difficult and result in unequal bisection of the left ventricle.
Citation: American Journal of Veterinary Research 77, 4; 10.2460/ajvr.77.4.374
Statistical analysis
Descriptive data (mean, SD, median, range, and approximate 95% prediction interval [mean ± 2SD]) were generated for the left atrial and aortic measurements, LA:Ao, EPSS, body weight, and left ventricular M-mode measurements. For each variable, the distribution of the data was assessed for normality by the Anderson-Darling test and visual examination of histograms. Body weight and some M-mode measurements were transformed by use of a natural logarithmic transformation to help normalize their respective distributions and allow for implementation of the allometric scaling approach by means of linear regression methods. Intraobserver repeatability was assessed by calculation of the CV for each measurement that was performed in triplicate. For each respective M-mode measurement, the proportionality constant (a), scaling parameter (b), and 95% prediction interval were estimated by fitting a linear regression model in which the M-mode measurement (following a natural logarithmic transformation when necessary) was the dependent variable and the natural logarithmic–transformed body weight was the independent variable. The importance of factors such as age, sex, neuter status, body condition score, and thoracic girth on the 95% prediction intervals for M-mode measurements in clinically normal dogs was assessed by adding each variable 1 by 1 to the allometric scaling regression model. The predictive ability of each variable was assessed by comparing the regression coefficient, P value, and change in the coefficient of determination (r2) of the robust model with those for the simple (univariate) allometric scaling model. The r2 is an estimate of the amount of variability in the dependent variable (ie, M-mode measurement) that is accounted for by changes in the independent variable (ie, body weight); the predictive ability of the independent variable increases as r2 increases. The validity of the multibreed allometric method34 was assessed by calculating the proportion of observed M-mode measurements from the present study that fell within the 95% prediction intervals calculated for the allometric method, estimating the Pearson correlation coefficient (r) between observed values and the values predicted on the basis of the allometric method for the present study, comparing those observed and predicted values with a paired t test, and creating scatter plots of observed and predicted values. All analyses were performed by 1 investigator (GTF) with commercially available software.c,d Values of P ≤ 0.05 were considered significant for all analyses.
Results
Dogs
Of the 55 (26 male and 29 female) adult Dachshunds initially recruited for the study, 36 were smooth haired, 15 were long haired, and 4 were wired haired. Mitral valve prolapse was observed in 25 (45.5%) dogs, of which 15 were excluded from the study because the MVP was > 1 mm. Of the 15 dogs excluded from the study, 3 had MR and 1 had a concomitant aortic aneurysm. Thus, 40 dogs were enrolled in the study, including 11 sexually intact males, 5 neutered males, 16 sexually intact females, and 8 spayed females. The mean age of the study dogs was 3.7 years (range, 1 to 7 years), and the mean body weight was 8.5 kg (range, 5 to 12.6 kg; Table 1).
Descriptive statistics for 40 clinically normal adult Dachshunds in which 2-D echocardiographic left atrial and aortic measurements and the associated LA:Ao were determined by each of 3 standardized methods (diameter, circumference, and cross-sectional area) and left ventricular measurements were determined by M-mode transthoracic echocardiography via the RPSA window.
| Variable | Mean (SD) | 95% Prediction interval* | Median (range) | Median (range) CV (%) |
|---|---|---|---|---|
| Age (y) | 3.7 (1.6) | 0.4–7.0 | 3.3 (1–7) | — |
| Body weight (kg) | 8.5 (2.1) | 4.3–12.6 | 8.2 (5–12.6) | — |
| Body condition score (5-point scale)† | 3.1 (0.6) | 1.9–4.3 | 3.0 (2.0–4.5) | — |
| Left atrial measurements | ||||
| Diameter method (cm) | 2.04 (0.26) | 1.52–2.56 | 2.04 (1.61–2.57) | 2.8 (0.3–26.6) |
| Circumference method (cm) | 10.93 (1.10) | 8.73–13.13 | 10.84 (8.34–13.44) | 4.1 (0.7–15.0) |
| Cross-sectional area method (cm2) | 5.26 (1.18) | 2.90–7.62 | 5.30 (3.10–7.65) | 4.6 (0.6–32.4) |
| Aortic measurements | ||||
| Diameter method (cm) | 1.46 (0.14) | 1.18–1.74 | 1.48 (1.17–1.80) | 1.7 (0.4–10.0) |
| Circumference method (cm) | 5.26 (0.56) | 4.14–6.38 | 5.22 (4.34–6.77) | 3.1 (0.4–14.7) |
| Cross-sectional area method (cm2) | 1.87 (0.40) | 1.07–2.67 | 1.85 (1.30–3.05) | 4.9 (1.5–30.4) |
| LA:Ao | ||||
| Diameter method | 1.40 (0.13) | 1.14–1.66 | 1.41 (1.19–1.65) | 2.8 (0.4–7.0) |
| Circumference method | 2.09 (0.17) | 1.75–2.43 | 2.09 (1.75–2.42) | 3.9 (0.7–9.0) |
| Cross-sectional area method | 2.85 (0.48) | 1.89–3.81 | 2.86 (1.87–3.86) | 7.0 (2.5–18.1) |
| Left ventricular measurements | ||||
| IVSd (mm) | 6.4 (0.8) | 4.8–7.9 | 6.3 (4.6–7.8) | 5.9 (0.0–19.0) |
| IVSs (mm) | 8.7 (1) | 6.7–10.7 | 8.8 (6.7–10.8) | 4.4 (0.6–17.2) |
| LVIDd (mm) | 27.7 (3.4) | 20.8–34.5 | 28.2 (21.6–34.5) | 2.3 (0.6–6.3) |
| LVIDs (mm)† | 16.3 (3) | 10.4–22.2 | 17.0 (11.1–21.1) | 3.6 (0.0–12.6) |
| LVFWd (mm) | 6.7 (0.9) | 4.9–8.4 | 6.8 (5.2–8.6) | 4.8 (0.0–15.6) |
| LVFWs (mm) | 9.9 (1.2) | 7.5–12.3 | 10.0 (7.2–12) | 4.2 (0.0–18.5) |
| Fractional shortening (%) | 41.4 (5.1) | 31.2–51.6 | 41.0 (31.6–53.4) | 4.5 (0.6–9.2) |
| EPSS (mm)† | 1.9 (0.8) | 0.3–3.5 | 1.6 (0.5–3.5) | 2.6 (0.5–5.9) |
For each dog, each echocardiographic variable was measured 3 times in different, but not necessarily consecutive, heart cycles.
95% Prediction interval = mean ± 2SD.
Variable not normally distributed.
— = Not calculated.
Echocardiographic measurements
Descriptive data for the left atrium and aorta and associated LA:Aos as determined by 3 standard methods of 2-D echocardiography and left ventricular measurements as determined by M-mode echocardiography were summarized (Table 1). Intraobserver variability of the measurements was considered good with the median CV for all measurements ranging from 1.7% to 7.0%.
Left ventricular M-mode 95% prediction intervals
For each left ventricular variable, the proportionality constant (a′) and scaling exponent (b′) for the allometric scaling equation were calculated by use of data obtained from the 40 Dachshunds of the present study and were compared to the proportionality constant (a) and scaling exponent (b) calculated for the same variables in a study34 of 494 adult dogs from 8 breeds (Table 2) The proportionality constants (a′) derived from the present study were comparable with the proportionality constants (a) derived from the previous study.34 However, the range for the scaling exponent (b'; 0.129 to 0.397) in the present study was wider than that for the previous study34 (b; 0.222 to 0.315) and was less consistent with the presumed index for linear measurements in the allometric equation, which was equivalent to (natural logarithmic–transformed body weight)0.333.
Coefficients for the proportionality constant (a) and scaling exponent (b) for the allometric equation used to estimate various left ventricular M-mode transthoracic echocardiographic measurements for the 40 clinically normal adult Dachshunds of Table 1 and 494 adult dogs from 8 breeds of a previous study.34
| Present study | Previous study34 | |||
|---|---|---|---|---|
| Variable | a' | b' | a | b |
| IVSd | 0.461 | 0.148 | 0.41 | 0.241 |
| IVSs | 0.659 | 0.129 | 0.58 | 0.240 |
| LVIDd | 1.441 | 0.306 | 1.53 | 0.294 |
| LVIDs | 0.694 | 0.397 | 0.95 | 0.315 |
| LVFWd | 0.399 | 0.239 | 0.42 | 0.232 |
| LVFWs | 0.658 | 0.189 | 0.64 | 0.222 |
The coefficients for the proportionality constant (a′) and scaling exponent (b′) calculated on the basis of data obtained from the 40 dogs of the present study (Table 2) were used to calculate the mean and 95% prediction intervals for each left ventricular M-mode measurement for clinically normal Dachshunds with various body weights (Table 3). For example, the mean LVIDd for an 8-kg Dachshund = 1.441 × 80.306 = 2.72 cm or 27.2 mm.
Mean (95% prediction interval) values for various left ventricular M-mode transthoracic echocardiographic measurements determined by use of the allometric scaling equation described and proportionality constants (a′) and scaling exponents (b′) for the dogs of Table 1 calculated in Table 2.
| Left ventricular measurements | ||||||
|---|---|---|---|---|---|---|
| Weight (kg) | IVSd (mm) | IVSs (mm) | LVIDd (mm) | LVIDs (mm) | LVFWd (mm) | LVFWs (mm) |
| 3.0 | 5.4 (4.0–7.3) | 7.6 (5.8–10.0) | 20.2 (15.8–25.7) | 10.7 (7.2–15.9) | 5.2 (3.9–6.9) | 8.1 (6.1–10.8) |
| 4.0 | 5.7 (4.3–7.5) | 7.9 (6.1–10.2) | 22.0 (17.6–27.6) | 12.0 (8.4–17.3) | 5.6 (4.2–7.3) | 8.6 (6.6–11.1) |
| 5.0 | 5.9 (4.5–7.6) | 8.1 (6.4–10.4) | 23.6 (19.0–29.3) | 13.2 (9.3–18.6) | 5.9 (4.5–7.6) | 8.9 (6.9–11.5) |
| 6.0 | 6.0 (4.6–7.8) | 8.3 (6.6–10.5) | 24.9 (20.2–30.8) | 14.1 (10.1–19.8) | 6.1 (4.8–7.9) | 9.2 (7.2–11.8) |
| 7.0 | 6.2 (4.8–7.9) | 8.5 (6.7–10.7) | 26.1 (21.2–32.2) | 15.0 (10.8–21.0) | 6.4 (5.0–8.1) | 9.5 (7.5–12.1) |
| 8.0 | 6.3 (4.9–8.1) | 8.6 (6.8–10.9) | 27.2 (22.1–33.5) | 15.9 (11.4–22.1) | 6.6 (5.1–8.4) | 9.8 (7.7–12.4) |
| 9.0 | 6.4 (5.0–8.2) | 8.8 (6.9–11.1) | 28.2 (23.0–34.7) | 16.6 (11.9–23.2) | 6.8 (5.3–8.6) | 10.0 (7.8–12.7) |
| 10.0 | 6.5 (5.0–8.4) | 8.9 (7.0–11.2) | 29.2 (23.7–35.9) | 17.3 (12.4–24.2) | 6.9 (5.4–8.9) | 10.2 (8.0–13.0) |
| 11.0 | 6.6 (5.1–8.5) | 9.0 (7.1–11.4) | 30.0 (24.3–37.0) | 18.0 (12.8–25.2) | 7.1 (5.5–9.1) | 10.4 (8.1–13.2) |
| 12.0 | 6.7 (5.1–8.7) | 9.1 (7.2–11.6) | 30.8 (24.9–38.1) | 18.6 (13.2–26.2) | 7.2 (5.6–9.3) | 10.5 (8.2–13.5) |
| 13.0 | 6.7 (5.2–8.8) | 9.2 (7.2–11.7) | 31.6 (25.5–39.2) | 19.2 (13.6–27.2) | 7.4 (5.7–9.5) | 10.7 (8.3–13.7) |
| 14.0 | 6.8 (5.2–8.9) | 9.3 (7.3–11.9) | 32.3 (26.0–40.2) | 19.8 (13.9–28.1) | 7.5 (5.8–9.7) | 10.8 (8.4–14.0) |
| 15.0 | 6.9 (5.3–9.0) | 9.4 (7.3–12.0) | 33.0 (26.5–41.2) | 20.3 (14.3–29.0) | 7.6 (5.9–9.9) | 11.0 (8.5–14.2) |
Comparison of 95% prediction intervals for left ventricular M-mode echocardiographic measurements calculated on the basis of Dachshund-specific data with 95% prediction intervals for those same measurements calculated on the basis of data obtained from adult dogs of various breeds in a previous study34
Compared with the 95% prediction intervals calculated in a previous study,34 the lower and upper limits of the 95% prediction interval for LVIDs were consistently 2 to 3 mm less, and the width of the 95% prediction intervals for all left ventricular M-mode measurements except those for LVIDd and LVIDs were narrower on the basis of data obtained from the Dachshunds of the present study.
For all 6 left ventricular M-mode measurements evaluated, ≥ 93% of the observed measurements for the Dachshunds of the present study were within the 95% prediction intervals calculated by the allometric scaling equation, which used the proportionality constants and scaling exponents estimated from data obtained from those dogs (Table 4). However, ≥ 95% of the observed measurements for only 4 variables (IVSd, IVSs, LVFWd, and LVFWs) and only 83% of observed measurements for LVIDs were within the 95% prediction intervals calculated by the allometric scaling equation which used proportionality constants and scaling exponents estimated from data obtained from dogs of a previous study.34 All of the observed measurements that were not within the respective 95% prediction intervals (regardless of proportionality constants and scaling exponents used) were below the lower prediction limit. In fact, the upper limits for the prediction intervals calculated by use of data from dogs of the previous study34 were substantially higher than the measurements observed for the Dachshunds of the present study, especially as body weight increased.
Number (percentage) of the observed left ventricular M-mode echocardiographic measurements obtained for the dogs of Table 1 that fell within, below, or above the 95% prediction intervals calculated by means of an allometric scaling equation in which the respective proportionality constants and scaling exponents were estimated from data obtained from those dogs (present study) or adult dogs of various breeds from a previous study.34
| 95% prediction interval based on data from present study | 95% prediction interval based on data from previous study34 | Correlation* | Mean difference† | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Variable | Within | Below | Above | Within | Below | Above | Pearson r | P value | Difference | P value |
| IVSd | 39 (97.5) | 1 (2.5) | 0 (0) | 40 (100) | 0 (0) | 0 (0) | 0.264 | 0.099 | −0.470 | 0.001 |
| IVSs | 39 (97.5) | 1 (2.5) | 0 (0) | 38 (95) | 2 (5) | 0 (0) | 0.268 | 0.094 | −0.915 | < 0.001 |
| LVIDd | 38 (95) | 2 (5) | 0 (0) | 37 (92.5) | 3 (7.5) | 0 (0) | 0.617 | < 0.001 | −0.829 | 0.059 |
| LVIDs | 37 (92.5) | 3 (7.5) | 0 (0) | 33 (82.5) | 7 (17.5) | 0 (0) | 0.550 | < 0.001 | −2.206 | < 0.001 |
| LVFWd | 40 (100) | 0 (0) | 0 (0) | 40 (100) | 0 (0) | 0 (0) | 0.441 | 0.004 | −0.195 | 0.127 |
| LVFWs | 39 (97.5) | I (2.5) | 0 (0) | 40 (100) | 0 (0) | 0 (0) | 0.363 | 0.022 | −0.354 | 0.056 |
Correlation was estimated between the observed measurements and the predicted measurements calculated on the basis of data obtained from dogs of a previous study.34
The mean difference was calculated as the observed measurement – the predicted measurement calculated on the basis of data obtained from dogs of a previous study,34 and the P values for the mean difference are for paired t tests.
The observed measurements of LVIDd, LVIDs, LVFWd, and LVFWs for the Dachshunds of the present study were significantly and positively correlated with the 95% predicted values calculated by use of data obtained from the dogs of the previous study34 (Table 4). The strength of those correlations ranged from weak (LVFWs, r = 0.363) to strong (LVIDd, r = 0.617). The observed measurements for IVSd and IVSs were not significantly correlated with the 95% prediction intervals calculated by use of data obtained from the dogs of the previous study.34
When added to the allometric scaling equation, neuter status and body condition score had significant negative associations with both LVIDd and LVIDs (Table 5). Age, sex, and thoracic girth were not significantly associated with any of the left ventricular M-mode measurements assessed.
Linear regression results for select variables when they were added to the allometric scaling regression model for prediction of left ventricular M-mode echocardiographic measurements for the dogs of Table 1.
| Left ventricular measurements | ||||||
|---|---|---|---|---|---|---|
| Variable | IVSd | IVSs | LVIDd | LVIDs | LVPWd | LVPWs |
| Age | ||||||
| Coefficient | −0.005 | 0.010 | −0.004 | −0.003 | 0.022 | 0.017 |
| r2 change | 0.4 | 1.9 | 0.2 | 0 | 7.6 | 4.8 |
| P value | 0.704 | 0.391 | 0.713 | 0.865 | 0.055 | 0.148 |
| Sex (referent = female) | ||||||
| Coefficient | −0.009 | −0.031 | −0.028 | −0.074 | −0.011 | 0.012 |
| r2 change | 0.1 | 1.7 | 1.2 | 3.7 | 0.2 | 0.2 |
| P value | 0.826 | 0.414 | 0.400 | 0.164 | 0.790 | 0.758 |
| Neuter status (referent = neutered) | ||||||
| Coefficient | 0.072 | 0.066 | −0.089 | −0.122 | 0.034 | 0.047 |
| r2 change | 6.8 | 6.7 | 10.7 | 8.8 | 1.4 | 3.0 |
| P value | 0.093 | 0.096 | 0.008 | 0.028 | 0.421 | 0.251 |
| Body condition score | ||||||
| Coefficient | 0.058 | 0.063 | −0.060 | −0.093 | 0.030 | 0.065 |
| r2 change | 6.8 | 6.7 | 10.7 | 8.8 | 1.4 | 3.0 |
| P value | 0.110 | 0.055 | 0.039 | 0.048 | 0.397 | 0.057 |
| Thoracic girth | ||||||
| Coefficient | 0.017 | 0.011 | 0.001 | −0.006 | −0.008 | −0.009 |
| r2 change | 6.3 | 3.1 | 0 | 0.3 | 1.2 | 1.9 |
| P value | 0.107 | 0.265 | 0.869 | 0.678 | 0.454 | 0.363 |
Discussion
In the present study, we calculated values for LA:Ao as determined by 2-D echocardiography and select left ventricular variables as determined by M-mode echocardiography in 40 clinically normal adult Dachshunds. We used the allometric scaling equation described in a previous study34 to generate 95% prediction intervals for the left ventricular variables on the basis of data obtained from the Dachshunds of the present study and compared those 95% prediction intervals with those calculated from data obtained from 494 adult dogs of 8 breeds. Our results suggested that the 95% prediction intervals calculated on the basis of data obtained in the present study were more representative for clinically normal adult Dachshunds than those calculated on the basis of data obtained from adult dogs of multiple breeds. Thus, the LA:Ao and 95% prediction intervals for left ventricular M-mode echocardiographic measurements reported in the present study may be useful as preliminary guidelines for clinically normal adult Dachshunds.
The upper limits (mean + 2SD) for the LA:Ao determined by standard diameter (1.66), circumference (2.43), and cross-sectional area (3.81) 2-D echocardiographic methods for the dogs of the present study were similar to those (1.59, 2.45, and 3.85, respectively) determined for 36 dogs of various breeds that were ≥ 9 months old and weighed between 10 and 40 kg.11 The mean LA:Ao for the clinically normal Dachshunds of the present study was markedly different from the mean LA:Ao for clinically normal CKCSs in another study,10 even though the body weights for the Dachshunds (range, 5.0 to 12.6 kg) and CKCSs (5.5 to 11.9 kg) were similar. The LA:Ao as determined by the diameter method was < 1.67 for all the Dachshunds of the present study and < 1.28 for all the CKCSs of that other study.10 The difference in the LA:Ao between those 2 breeds may be attributable to sample population differences, observer variability, or breed differences. The present study did not include any Dachshunds with MMVD; therefore, the proportion of Dachshunds with MMVD that might have a LA:Ao less than the upper limit observed for the clinically normal Dachshunds of this study is unknown. We did not use M-mode echocardiography to calculate the LA:Ao in this study because that method has inherent limitations such as difficulty in accurately measuring the maximum diameter of the aorta and the potential for measuring the left auricle instead of the left atrium.11
The presumed index of length was 0.333 for the allometric equation used to predict left ventricular M-mode echocardiographic measurements described in another study.34 The scaling exponents (b′) calculated for the allometric equation on the basis of data obtained from the Dachshunds of the present study ranged from 0.129 to 0.397. The deviation of those scaling exponents from the presumed index of length for the other study34 was likely caused by the nonquantifiable effects of thoracic conformation and breed specificity. The 494 dogs of that other study34 represented 8 breeds, most of which were large-breed dogs, whereas the present study included only Dachshunds, which are chondrodysplastic dwarfs. Other possible explanations for those deviations include the small sample size (n = 40) of the present study and the fairly narrow range of body weights for the Dachshunds of this study.
The prediction intervals for 4 of the 6 left ventricular M-mode measurements assessed for the Dachshunds of the present study were narrower than those calculated for the 494 dogs of multiple breeds in that other study.34 Additionally, only 8 of 240 (3%) observed left ventricular M-mode measurements fell outside of the 95% prediction intervals calculated from data obtained from the dogs of the present study (new prediction intervals), whereas 12 of 240 (5%) observed measurements fell outside of the 95% prediction intervals calculated from data obtained for the dogs of that other study34 (previous prediction intervals). This was particularly noticeable for LVIDs; of the 40 observed LVIDs measurements, only 3 (7.5%) did not fall within the new prediction interval, whereas 7 (17.5%) did not fall within the previous prediction interval. This suggested that the new prediction intervals were more representative for clinically normal adult Dachshunds than were the previous prediction intervals.34 The upper limits of the new prediction intervals were lower than the upper limits of the previous prediction intervals34 for all left ventricular variables assessed except LVIDd. The upper limits of the previous prediction intervals were likely skewed because the dogs of that study34 had various body conformations and a fairly wide range of body weights (2 to 95 kg) and body weight was normalized for calculation of the 95% prediction intervals. Although the new prediction intervals were narrower than the previous prediction intervals, we have not compared the new prediction intervals with observed measures obtained from Dachshunds with left ventricular volume overload. Additional study is required to determine whether the LA:Aos and 95% prediction intervals for left ventricular M-mode echocardiographic measurements determined in the present study can be used to accurately distinguish between clinically normal Dachshunds and Dachshunds with left-sided heart failure.
The fact that > 95% of the observed measurements for 4 of the 6 left ventricular M-mode variables assessed in the present study were within the respective previous prediction intervals34 was likely a function of the fairly wide width of those intervals. A similar phenomenon was observed in a study14 that involved Whippets. As alluded to earlier, the width of the previous prediction intervals34 was likely a reflection of the diverse population of dogs in terms of body size and conformation from which the data used to calculate those intervals was obtained.
Alternatively, the reason the new prediction intervals were narrower than the previous prediction intervals34 might have been associated with the use of AMM in the present study. The accuracy of left ventricular M-mode echocardiographic measurements is maximized when the sample line is aligned perpendicular to the short or long axis of the left ventricle, but that can be difficult to achieve by the use of conventional echocardiographic techniques. In the present study, use of AMM made alignment of the sampling line perpendicular to the short or long axis of the left ventricle easier and more precise than alignment by use of conventional M-mode techniques, which reduced the variation of the measurements42,43 and increased the reproducibility and accuracy of those measurements.44 The data used to calculate the previous prediction intervals34 was acquired by conventional M-mode echocardiography. Additionally, all measurements in the present study were acquired by 1 investigator (CKL) by the use of standard recommended echocardiographic techniques, which eliminated the potential for interobserver variability. The study34 from which the previous prediction intervals were obtained was retrospective, and the M-mode echocardiographic techniques used were not standardized (ie, measurements were obtained by 9 echocardiographers from both sedated and unsedated dogs).
The veterinary literature contains conflicting information regarding the association between sex and left ventricular M-mode echocardiographic measurements. Results of some studies25,31,45 suggest that there is no significant association between sex and left ventricular M-mode measurements, whereas IVSd and LVIDd varied significantly between male and female Estrela Mountain Dogs22 and LVFWd varied significantly between male and female Dogues de Bordeaux.23 In the present study, the addition of sex to the allometric linear regression model did not significantly affect predicted left ventricular M-mode measurements.
Results of other studies15,46,47 indicate a significant positive correlation between age and normalized left ventricular diameter and wall thickness. In the present study, age did not significantly affect the prediction of left ventricular M-mode measurements for adult Dachshunds. Thoracic girth likewise did not significantly affect the prediction of left ventricular M-mode measurements in adult Dachshunds, and we are not aware of any other studies that assessed the effect of girth on left ventricular M-mode measurements.
An interesting and novel finding of the present study was that neuter status and body condition score were significantly associated with both LVIDd and LVIDs. Neutered dogs and dogs with a high body condition score tended to have a lower LVIDd and LVIDs than did sexually intact dogs and dogs with the same weight but a lower body condition score. This finding might be a function of pericardial fat; neutered dogs and dogs with high body condition scores have proportionately more fat than lean body mass, compared with sexually intact dogs and dogs with lower body condition scores. However, only 13 of the 40 (33%) dogs of the present study were neutered, so the effect of neuter status on left ventricular M-mode measurements should be interpreted cautiously and warrants further investigation.
The 95% prediction interval for fractional shortening was 31% to 52% for the clinically normal adult Dachshunds of the present study, which was comparable to similar intervals (23% to 47%34 and 27% to 51%45) reported for adult dogs of various breeds. The measurement of EPSS is often performed because it is a practical and easily reproducible clinical index of left ventricular function.41 It has been suggested that abnormally increased EPSS values are indicative of systolic dysfunction, left ventricular volume overload, or mitral valve stenosis.41 For the Dachshunds of the present study, the maximum EPSS was 3.5 mm, which was substantially lower than the maximum EPSS (6.0 mm) measured for 50 healthy Beagle and German Shepherd Dogs in another study.41 That finding might be attributable to the smaller body size of Dachshunds, compared with the body sizes of Beagles and German Shepherd Dogs. However, clinically ill dogs with volume overload or systolic dysfunction were not included in the present study or the study41 that involved Beagles and German Shepherd Dogs; therefore, we could not estimate the extent of overlap between EPSS values of clinically normal Dachshunds and Dachshunds with volume overload or systolic dysfunction.
For the Dachshunds initially screened for the present study, the prevalence of MVP was 45.5% (25/55), which was less than the prevalence of MVP (60% [36/60]) for Dachshunds in another study5 and substantially less than that (86% [164/190]) reported for adult Dachshunds in a different study.3 The study3 in which the MVP prevalence was 86% involved Dachshunds from 18 families (a family consisted of both parents and ≥ 4 offspring). Mitral valve prolapse is a heritable condition, so the prevalence in that study3 may have been biased because of the high proportion of dogs that were related. The Dachshunds of the present study were from diverse origins, and although the pedigrees for those dogs were not assessed, it is likely that few had close familial relationships. Additionally, results of that other study3 indicate that coat type is significantly associated with MVP; long-haired Dachshunds are more likely to have MVP than are smooth-haired Dachshunds, which are more likely to have MVP than are wire-haired Dachshunds. Of the 55 Dachshunds initially screened for the present study, 36 (65%) had smooth hair, 15 (27%) had long hair, and 4 (7%) had wire hair, whereas of the 190 dogs in that other study,3 21 (11%) had smooth hair, 91 (48%) had long hair, and 78 (41%) had wire hair. So the difference in the prevalence of MVP observed between the present study and that other study3 might also have been affected by differences in the distributions of Dachshunds with various coat types in the respective study populations.
We recognize that the application of a natural logarithmic transformation to the left ventricular M-mode echocardiographic measurements that were not normally distributed prior to linear regression might not have been the ideal approach to assess those variables. However, identification of the best possible prediction model for each variable assessed was beyond the scope of the present study. Because of the large number of statistical tests performed, it is possible that some of the significant associations observed were simply due to chance (ie, for each test performed, there was a 5% chance of making a type I error [rejecting the null hypothesis when it is true]; therefore, when the tests were considered collectively, the cumulative probability of making at least 1 type I error was substantially higher).
In the present study, the means (SDs), medians, ranges, and 95% prediction intervals for the LA:Ao as determined by three 2-D conventional echocardiographic methods and 6 left ventricular M-mode echocardiographic variables (IVSDd, IVSDs, LVIDd, LVIDs, LVFWd, and LVFWs) for 40 clinically normal adult Dachshunds were provided. Results of an allometric linear regression model suggested that the 95% prediction intervals calculated in the present study were narrower and may be more representative for clinically normal adult Dachshunds than those calculated from data obtained from adult dogs of various breeds in another study.34 Thus, the 95% prediction intervals calculated in this study can be used as preliminary guidelines against which echocardiographic measurements obtained from other adult Dachshunds can be compared.
Acknowledgments
This manuscript represents a portion of a thesis submitted by Dr. Lim to the Department of Companion Animal Clinical Studies, University of Pretoria, as partial fulfillment of the requirements for a Master of Veterinary Medicine degree.
Presented in part at the 16th International Veterinary Radiology Association Meeting and European Veterinary Diagnostic Imaging Annual Meeting, Bursa, Turkey, September 2012.
ABBREVIATIONS
| AMM | Anatomic M-mode |
| CKCS | Cavalier King Charles Spaniel |
| CV | Coefficient of variation |
| EPSS | E point to septal separation |
| IVSd | Diastolic interventricular septal thickness |
| IVSs | Systolic interventricular septal thickness |
| LA:Ao | Left atrium-to-aorta ratio |
| LVFWd | Left ventricular free wall thickness at end diastole |
| LVFWs | Left ventricular free wall thickness at end systole |
| LVIDd | Left ventricular internal diameter at end diastole |
| LVIDs | Left ventricular internal diameter at end systole |
| MMVD | Myxomatous mitral valve disease |
| MR | Mitral valve regurgitation |
| MVP | Mitral valve prolapse |
| RPSA | Right parasternal short axis |
Footnotes
Ljungvall I. The left ventricle in dogs with myxomatous mitral valve disease—remodelling and overall performance. Doctoral thesis, Swedish University of Agricultural Sciences, Uppsala, Sweden, 2011.
Aloka Prosound F75, Hitachi Aloka Medical, Tokyo, Japan.
Minitab statistical software, version 13.32, Minitab Inc, State College, Penn.
SPSS, version 22, IBM Corp, Armonk, NY.
References
1. Dennis MO, Nealeigh RC, Pyle RL, et al. Echocardiographic assessment of normal and abnormal valvular function in Beagle dogs. Am J Vet Res 1978; 39: 1591–1598.
2. Mashiro I, Nelson RR, Cohn JN, et al. Ventricular dimensions measured noninvasively by echocardiography in the awake dog. J Appl Physiol 1976; 41: 953–959.
3. Olsen LH, Fredholm M, Pedersen HD. Epidemiology and inheritance of mitral valve prolapse in Dachshunds. J Vet Intern Med 1999; 13: 448–456.
4. Olsen LH, Martinussen T, Pedersen HD. Early echocardiographic predictors of myxomatous mitral valve disease in Dachshunds. Vet Rec 2003;152: 293–297.
5. Pedersen HD, Kristensen BØ, Nørby B, et al. Echocardiographic study of mitral valve prolapse in Dachshunds. Zentralbl Veterinarmed A 1996; 43: 103–110.
6. Pedersen HD, Olsen LH, Mow T, et al. Neuroendocrine changes in Dachshunds with mitral valve prolapse examined under different study conditions. Res Vet Sci 1999; 66: 11–17.
7. Beardow AW, Buchanan JW. Chronic mitral valve disease in Cavalier King Charles Spaniels: 95 cases (1987–1991). J Am Vet Med Assoc 1993; 203: 1023–1029.
8. Darke PGG. Valvular incompetence in Cavalier King Charles Spaniels. Vet Rec 1987; 120: 365–366.
9. Pedersen HD, Lorentzen KA, Kristensen BO. Echocardiographic mitral valve prolapse in Cavalier King Charles Spaniels: epidemiology and prognostic significance for regurgitation. Vet Rec 1999; 144: 315–320.
10. Hansson K, Häggström J, Kvart C, et al. Left atrial to aortic root indices using two-dimensional and M-mode echocardiography in Cavalier King Charles Spaniels with and without left atrial enlargement. Vet Radiol Ultrasound 2002; 43: 568–575.
11. Rishniw M, Erb HN. Evaluation of four 2-dimensional echocardiographic methods of assessing left atrial size in dogs. J Vet Intern Med 2000; 14: 429–435.
12. Borgarelli M, Savarino P, Crosara S, et al. Survival characteristics and prognostic variables of dogs with mitral regurgitation attributable to myxomatous valve disease. J Vet Intern Med 2008; 22: 120–128.
13. Baade H, Schober K, Oechtering G. Echocardiographic reference values in West Highland White Terriers with special regard to right heart function. Tierarztl Prax 2002; 30: 172–179.
14. Bavegems V, Duchateau L, Sys SU, et al. Echocardiographic reference values in Whippets. Vet Radiol Ultrasound 2007; 48: 230–238.
15. Bayón A, del Palacio MJF, Montes AM, et al. M-mode echocardiography study in growing Spanish Mastiffs. J Small Anim Pract 1994; 35: 473–479.
16. Borgarelli M, Tarducci A, Bussadori C, et al. Echocardiographic and echo-Doppler parameters in normal Great Danes. Eur Soc Vet Cardiol Newsl 1996; 10: 27–28.
17. Crippa L, Ferro E, Melloni E, et al. Echocardiographic parameters and indices in the normal Beagle dog. Lab Anim 1992; 26: 190–195.
18. Della Torre PK, Kirby AC, Church DB, et al. Echocardiographic measurements in Greyhounds, Whippets and Italian Greyhounds—dogs with a similar conformation but different size. Aust Vet J 2000; 78: 49–55.
19. Herrtage ME. Echocardiographic measurements in the normal Boxer, in Proceedings. 4th Annu Cong Eur Soc Vet Intern Med 1994;172–173.
20. Kayar A, Gonul R, Or ME, et al. M-mode echocardiographic parameters and indices in the normal German Shepherd dog. Vet Radiol Ultrasound 2006; 47: 482–486.
21. Kayar A, Uysal A. Determination of cardiac reference parameters using M-mode and 2-D echocardiographic techniques in adult Karabash dogs. Turk J Vet Anim Sci 2004; 28: 39–46.
22. Lobo L, Canada N, Bussadori C, et al. Transthoracic echocardiography in Estrela Mountain dogs: reference values for the breed. Vet J 2008; 177: 250–259.
23. Locatelli C, Santini A, Bonometti GA, et al. Echocardiographic values in clinically healthy adult Dogue de Bordeaux dogs. J Small Anim Pract 2011; 52: 246–253.
24. Lonsdale RA, Labuc RH, Robertson ID. Echocardiographic parameters in training compared with non-training Greyhounds. Vet Radiol Ultrasound 1998; 39: 325–330.
25. Muzzi RA, Muzzi LA, de Araújo RB, et al. Echocardiographic indices in normal German Shepherd Dogs. J Vet Sci 2006; 7: 193–198.
26. O'Leary CA, MacKay BM, Taplin RH, et al. Echocardiographic parameters in 14 healthy English Bull Terriers. Aust Vet J 2003; 81: 535–542.
27. Page A, Edmunds G, Atwell RB. Echocardiographic values in the Greyhound. Aust Vet J 1993; 70: 361–364.
28. Pietra M, Guglielmini C, Cipone M. Normal M-mode and Doppler echocardiographic values in the English Setter. Eur Soc Vet Cardiol Newsl 1998; 16: 8–14.
29. Sottiaux J, Amberger C. Normal echocardiographic values in the Dobermann Pinscher (French breeding). Eur Soc Vet Cardiol Newsl 1997;12: 18.
30. Vollmar A. Echocardiographic examinations in Deerhounds, reference values for echodiagnostic. Kleintierpraxis 1998; 43: 497–508.
31. Vollmar AC. Echocardiographic measurements in the Irish Wolfhound: reference values for the breed. J Am Anim Hosp Assoc 1999; 35: 271–277.
32. Vörös K, Hetyey C, Reiczigel J, et al. M-mode and two-dimensional echocardiographic reference values for three Hungarian dog breeds: Hungarian Vizsla, Mudi and Hungarian Greyhound. Acta Vet Hung 2009; 57: 217–227.
33. Brown DJ, Rush JE, MacGregor J, et al. M-mode echocardiographic ratio indices in normal dogs, cats, and horses: a novel quantitative method. J Vet Intern Med 2003; 17: 653–662.
34. Cornell CC, Kittleson MD, Della Torre P, et al. Allometric scaling of M-mode cardiac measurements in normal adult dogs. J Vet Intern Med 2004; 18: 311–321.
35. Schmidt-Nielsen K. The use of allometry. In: Schmidt-Nielsen K, ed. Scaling: why is animal size so important? New York: Cambridge University Press, 1984;21–32
36. Snyder PS, Sato T, Atkins CE. A comparison of echocardiographic indices of the nonracing, healthy Greyhound to reference values from other breeds. Vet Radiol Ultrasound 1995; 36: 387–392.
37. Morrison SA, Moise NS, Scarlett J, et al. Effect of breed and body weight on echocardiographic values in four breeds of dogs of differing somatotype. J Vet Intern Med 1992; 6: 220–224.
38. Impellizeri JA, Tetrick MA, Muir P. Effect of weight reduction on clinical signs of lameness in dogs with hip osteoarthritis. J Am Vet Med Assoc 2000; 216: 1089–1091.
39. Thomas WP, Gaber CE, Jacobs GJ, et al. 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: 247–252.
40. Sahn DJ, DeMaria A, Kisslo J, et al. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978; 58: 1072–1083.
41. Kirberger RM. Mitral wave E point to ventricular septal separation in the dog. J S Afr Vet Assoc 1991; 62: 163–166.
42. Carerj S, Micari A, Trono A, et al. Anatomical M-mode: an old-new technique. Echocardiography 2003; 20: 357–361.
43. Mele D, Pedini I, Alboni P, et al. Anatomic M-mode: a new technique for quantitative assessment of left ventricular size and function. Am J Cardiol 1998;81(suppl 1):82G–85G.
44. Oyama MA, Sisson DD. Assessment of cardiac chamber size using anatomic M-mode. Vet Radiol Ultrasound 2005; 46: 331–336.
45. Lombard CW. Normal values of canine M-mode echocardiogram. Am J Vet Res 1984; 45: 2015–2018.
46. Sisson D, Schaeffer D. Changes in linear dimensions of the heart, relative to body weight, as measured by M-mode echocardiography in growing dogs. Am J Vet Res 1991; 52: 1591–1596.
47. Sleeper MM, Henthorn PM, Vijayasarathy HV, et al. Dilated cardiomyopathy in juvenile Portuguese Water Dogs. J Vet Intern Med 2002; 16: 52–62.
