Myxomatous mitral valve disease is the most common cardiac disease of dogs and affects up to 90% of dogs > 10 years old within some small breeds.1 The disease is typically characterized by a long subclinical period during which a left apical systolic murmur gradually becomes more evident as mitral valve regurgitation progresses. Some dogs with MMVD develop clinically detectible LA and left ventricular dilatation, and in such dogs, the magnitude of LA enlargement is considered a reliable indicator of hemodynamic burden or disease severity.1 Clinical signs of MMVD, such as cough and syncope, and CHF can eventually develop in affected dogs and are generally associated with LA enlargement as the disease progresses.2
Results of multiple large clinical studies involving dogs with MMVD indicate that echocardiographic assessment of LA size is beneficial for evaluating the risk for CHF,3 when to intervene with medications prior to the onset of CHF,4,5 and prognosis.2,5–7 Thus, assessment of LA size is considered an important part of the diagnostic evaluation for dogs suspected of having MMVD. Echocardiographic LA size assessment usually involves indexing the left atrium to the aorta to serve as an internal control because body size varies widely among dogs. However, echocardiographic assessment of LA size is not always a practical option owing to the limited availability and cost of echocardiography in many areas and the training and expertise necessary to acquire and interpret echocardiographic assessments.
Thoracic radiography is recommended as part of the diagnostic evaluation for all dogs with suspected MMVD regardless of whether they have clinical signs.8 Compared with echocardiography, thoracic radiography is widely available, cost-effective, and considered part of the gold-standard procedures for diagnosis of left-sided CHF. However, radiographic assessment of LA size is primarily subjective and prone to error. Consequently, quantitative methods for radiographic estimation of LA size would be of clinical value, particularly when echocardiography is not readily available. Studies in which radiographic estimates of LA size have been compared with echocardiographic measurements of LA size (gold standard for LA measurement) are limited. Investigators of 1 study9 concluded that measurement of the tracheal bifurcation angle on dorsoventral thoracic radiographs was not reliable for estimation of LA size.
The purpose of the study reported here was to describe VLAS, a quantitative method to estimate LA size radiographically, and to determine its diagnostic value for prediction of echocardiographic LA enlargement in dogs with MMVD of varying severity. We hypothesized that VLAS would accurately predict the presence of echocardiographic LA enlargement in a large and diverse sample of dogs with MMVD.
Materials and Methods
Animals
All study procedures were reviewed and approved by the University of California-Davis Institutional Animal Care and Use Committee. All dogs were examined at the University of California-Davis Veterinary Medical Teaching Hospital, and owner consent was obtained for all dogs prior to study enrollment. Dogs with a left-sided systolic murmur that underwent a complete echocardiographic examination in conventional imaging planes as described10 and had 3-view thoracic radiographs obtained within 24 hours before or after the echocardiographic examination as part of a routine diagnostic evaluation were prospectively enrolled in the study in a sequential manner during the 7-month period from December 2015 to June 2016. Dogs were excluded from the study if thoracic radiography revealed overt malpositioning of the patient or the presence of thoracic vertebral abnormalities (eg, hemivertebrae). Dogs that were < 1 year old (eg, skeletally immature) and those with acute CHF secondary to chordae tendinae rupture were also excluded from the study. Dogs with tricuspid valve regurgitation were not excluded from the study because clinically irrelevant tricuspid valve regurgitation is common in dogs11 and does not affect LA size.
Dogs enrolled in the study were sedated with butorphanol (0.1 to 0.3 mg/kg [0.045 to 0.136 mg/lb], IV or IM) for radiographic and echocardiographic examinations at the discretion of the attending clinician. Likewise, the medication regimens for individual dogs were adjusted between radiographic and echocardiographic examinations as deemed necessary by the attending clinician. Myxomatous mitral valve disease was diagnosed on the basis of echocardiographic findings and was defined as the presence of valvular thickening, irregularity, and prolapse of the mitral valve apparatus and evidence of mitral valve regurgitation during systole as determined by color Doppler ultrasonography.12 Decompensated left-sided CHF was diagnosed on the basis of radiographic evidence of pulmonary edema that was deemed cardiogenic in origin by a cardiologist and the presence of respiratory difficulty that was responsive to furosemide administration.
Thoracic radiography
For each dog, right lateral, left lateral, and dorsoventral radiographic images of the thorax were obtained in a routine manner by use of a commercially available digital radiography system.a A digital caliperb was used to determine the VLAS on both the right lateral and left lateral thoracic images. First, a line was drawn and measured (in arbitrary units) from the center of the most ventral aspect of the carina to the most caudal aspect of the left atrium where it intersected with the dorsal border of the caudal vena cava. For the purpose of this study, the carina was defined as the radiolucent circular or ovoid structure within the trachea that represented the bifurcation of the left and right mainstem bronchi. Similar to the vertebral heart size method,13 a second line that was equal in length to the first was drawn beginning at the cranial edge of T4 and extending caudally just ventral and parallel to the vertebral canal (Figure 1). The VLAS was defined as the length of the second line expressed in vertebral-body units to the nearest 0.1 vertebra.
Representative right lateral thoracic radiographic image of a dog with LA enlargement secondary to MMVD that depicts measurement of the VLAS. A digital caliper was used to draw and measure (in arbitrary units) a line from the center of the most ventral aspect of the carina (bifurcation of the left and right mainstem bronchi) to the most caudal aspect of the left atrium where it intersects with the dorsal border of the caudal vena cava (#). Then, a second line that was equal in length to the first was drawn beginning at the cranial edge of T4 (*) and extending caudally just ventral and parallel to the vertebral canal. The VLAS was defined as the length of the second line expressed in vertebral-body units to the nearest 0.1 vertebra. For this dog, the VLAS was 3 vertebrae.
Citation: Journal of the American Veterinary Medical Association 253, 8; 10.2460/javma.253.8.1038
Representative right lateral thoracic radiographic image of a dog with LA enlargement secondary to MMVD that depicts measurement of the VLAS. A digital caliper was used to draw and measure (in arbitrary units) a line from the center of the most ventral aspect of the carina (bifurcation of the left and right mainstem bronchi) to the most caudal aspect of the left atrium where it intersects with the dorsal border of the caudal vena cava (#). Then, a second line that was equal in length to the first was drawn beginning at the cranial edge of T4 (*) and extending caudally just ventral and parallel to the vertebral canal. The VLAS was defined as the length of the second line expressed in vertebral-body units to the nearest 0.1 vertebra. For this dog, the VLAS was 3 vertebrae.
Citation: Journal of the American Veterinary Medical Association 253, 8; 10.2460/javma.253.8.1038
Representative right lateral thoracic radiographic image of a dog with LA enlargement secondary to MMVD that depicts measurement of the VLAS. A digital caliper was used to draw and measure (in arbitrary units) a line from the center of the most ventral aspect of the carina (bifurcation of the left and right mainstem bronchi) to the most caudal aspect of the left atrium where it intersects with the dorsal border of the caudal vena cava (#). Then, a second line that was equal in length to the first was drawn beginning at the cranial edge of T4 (*) and extending caudally just ventral and parallel to the vertebral canal. The VLAS was defined as the length of the second line expressed in vertebral-body units to the nearest 0.1 vertebra. For this dog, the VLAS was 3 vertebrae.
Citation: Journal of the American Veterinary Medical Association 253, 8; 10.2460/javma.253.8.1038
The same investigator (ELM) measured the VLAS on all lateral images. That investigator was unaware of (blinded to) the clinical diagnosis of all dogs and was also responsible for performing all echocardiographic measurements. The investigator was not present when thoracic radiographs were obtained or echocardiographic examinations were performed. For each dog, radiographic measurements were obtained before the echocardiographic measurements.
Echocardiography
All echocardiographic examinationsc were performed by a board-certified veterinary cardiologist or a cardiology resident under the supervision of a board-certified veterinary cardiologist. All echocardiographic variables were measured from 2-D echocardiographic images or views.
A digital off-cart workstationd was used for all echocardiographic measurements and calculations. Measurements were determined by the same investigator (ELM) who performed all radiographic measurements. That investigator was trained to evaluate the echocardiographic imaging planes and calculate the measurements evaluated in the study by a board-certified veterinary cardiologist (LCV). That training consisted of a one-on-one 3-day training period during which echocardiographic studies of several dogs were reviewed and measurements were demonstrated. The investigator reviewed current literature on the quantification of LA size in dogs, then performed the measurements under the direct supervision of the cardiologist. The cardiologist assessed the investigator's progress, and once the investigator's ability to perform the measurements closely matched that of the cardiologist, the training period was considered complete and the measurements for the study were acquired. All echocardiographic measurements made by the investigator were reviewed for accuracy by the cardiologist. A given variable was remeasured when there was substantial disagreement between the investigator and cardiologist. Each echocardiographic variable was measured 3 times, and the mean of those measurements was calculated and used for analysis. The investigator remained blinded to the radiographic measurements while performing the echocardiographic measurements.
A standard right parasternal long-axis 4-chamber view was used to assess LA size, which was determined by indexing the maximum dimension of the left atrium during systole to the aortic valve annulus (LA:AoLx). The aortic valve annulus was measured at the hinge points of the valve leaflets while they were open during early systole on a separate long-axis view that was optimized for assessment of the aortic valve annulus. The reference cutoff for LA:AoLx was < 2.5.e The LA size was also determined from a right parasternal short-axis view (LA:AoSx) by a conventional method,14,15 and the reference cutoff for LA:AoSx was < 1.5. For the purpose of this study, LA enlargement was defined as an LA:AoLx ≥ 2.6 and LA:AoSx ≥ 1.6. Those cutoffs were chosen because results of other studies4,5 indicate that dogs with MMVD and that degree of LA enlargement (eg, LA:AoSx ≥ 1.6) benefit from therapeutic intervention (eg, enalapril and pimobendan administration) prior to the onset of CHF.
Study group allocation
Dogs with MMVD were allocated to 1 of 4 groups in accordance with American College of Veterinary Internal Medicine guidelines.8 Dogs allocated to the control group had a left-sided physiologic or flow murmur but were free of echocardiographic abnormalities. Dogs allocated to the stage B1 group had MMVD with hemodynamically irrelevant mitral valve regurgitation as determined on the basis of an LA:AoSx < 1.6 or LA:AoLx < 2.6. Dogs allocated to the stage B2 group had MMVD with hemodynamically relevant mitral valve regurgitation as determined on the basis of an LA:AoSx ≥ 1.6 and LA:AoLx ≥ 2.6 but without evidence of previous (compensated) or concurrent (decompensated) CHF. Dogs allocated to the stage C-D group had MMVD and evidence of compensated or decompensated CHF.
Determination of VLAS measurement agreement
To determine the intraobserver agreement of VLAS measurements, 1 investigator (ELM) measured the VLAS on right and left lateral thoracic radiographic images for 20 dogs on 3 separate occasions. To determine interobserver agreement of VLAS measurements, 3 other investigators (LCV, KLP, and LRJ) individually measured the VLAS on right and left lateral thoracic radiographic images for the same 20 dogs. All investigators were blinded to the clinical diagnosis for each dog and the VLAS measurements determined by the other investigators.
Statistical analysis
Statistical analyses were performed with commercial software packages.f,g Descriptive statistics were generated. The distribution of data for continuous variables was assessed for normality by means of the D'Agostino-Pearson test. None of the variables were normally distributed, and results were reported as the median and IQR unless otherwise specified. Continuous variables were compared among the 4 study groups by means of the Kruskal-Wallis test followed by the Dunn test for pairwise comparisons when necessary. χ2 Tests were used to compare proportions among groups. The Spearman rank-order correlation coefficient (rS) was calculated to determine the strength of the association between VLAS and echocardiographic indices of LA size (LA:AoSx and LA:AoLx). The VLAS as determined from the right lateral thoracic radiograph versus that determined from the left lateral thoracic radiograph was compared with a Wilcoxon signed rank test and assessed for bias and limits of agreement by means of the Bland-Altman method. The extent of intraobserver and interobserver agreement for VLAS measurements was assessed by means of the ICC, and agreement was considered high when the ICC was > 0.75. Receiver operating characteristic analysis was used to assess the optimal cutoff value for VLAS (as measured on a right lateral thoracic radiograph) to predict LA enlargement as determined on the basis of 3 sets of criteria (LA:AoLx ≥ 2.6, LA:AoSx ≥ 1.6, and LA:AoLx ≥ 2.6 and LA:AoSx ≥ 1.6). For each set of criteria, the optimal clinically relevant cutoff value for VLAS was defined as the measurement that yielded the highest Youden index (1 + [sensitivity – specificity]), which reflected the optimal combination of sensitivity and specificity and minimized the amount of overlap between affected and unaffected dogs. The area under the ROC curve was used to assess the diagnostic accuracy and quantify the predictive value of VLAS. Values of P < 0.05 were considered significant for all analyses.
Results
Dogs
The study population consisted of 103 dogs, and descriptive data for each of the 4 study groups were summarized (Table 1). The control group consisted of 3 Chihuahuas and 1 dog from each of 12 other breeds. The stage B1 group consisted of 8 mixed-breed dogs, 4 Shih Tzus, 3 Labrador Retrievers, 3 Chihuahuas, 3 Toy Poodles, 3 Maltese, 2 Boston Terriers, 2 Pit Bull Terriers, 2 Cavalier King Charles Spaniels, 2 American Cocker Spaniels, and 1 dog from each of 8 other breeds. The stage B2 group consisted of 10 mixed-breed dogs, 4 Chihuahuas, 3 Dachshunds, 2 Cavalier King Charles Spaniels, and 1 dog from each of 7 other breeds. The stage C-D group consisted of 4 Cavalier King Charles Spaniels, 3 American Cocker Spaniels, 2 Chihuahuas, 2 mixed-breed dogs, and 1 dog from each of 11 other breeds.
Descriptive data for 103 dogs with a left-sided systolic murmur that underwent a complete echocardiographic examination and had 3-view (right lateral, left lateral, and ventrodorsal) thoracic radiographs obtained as part of a routine diagnostic evaluation at a veterinary teaching hospital from December 2015 through June 2016.
| Study group | |||||
|---|---|---|---|---|---|
| Variable | Control | Stage B1 | Stage B2 | Stage C-D | P value |
| No. of dogs | 15 | 40 | 26 | 22 | — |
| Body weight (kg) | 9.3 (5.4–27.2) | 8.7 (5.7–27.3) | 7.5 (5.4–13.2) | 8.4 (4.8–10.9) | 0.49 |
| Age (y) | 5.0 (2.4–11.0) | 10.3 (8.5–13.3)* | 11.1 (8.9–13.3)* | 10.3 (7.7–11.8)* | < 0.001 |
| No. (%) of females | 9 (60) | 20 (50) | 14 (54) | 13 (59) | 0.87 |
| Murmur grade (scale, 1–6) | 2 (1–2) | 3 (2–4)* | 4 (3–4)* | 4 (4–5)*† | < 0.001 |
| LA:AoSx | 1.3 (1.2–1.4) | 1.4 (1.3–1.5) | 2.0 (1.8–3.3)*† | 2.6 (2.0–3.4)*† | < 0.001 |
| LA:AoLx | 2.3 (2.1–2.4) | 2.5 (2.2–2.6) | 3.3 (2.9–3.7)*† | 4.1 (3.6–4.8)*† | < 0.001 |
| VLAS (No. of vertebrae) | 2.1 (1.8–2.3) | 2.1 (2.0–2.4) | 2.6 (2.3–2.9)*† | 3.0 (2.7–3.6)*† | < 0.001 |
Values represent the median (IQR) unless otherwise indicated. Dogs allocated to the control group had a left-sided physiologic or flow murmur but were free of echocardiographic abnormalities. Dogs allocated to the stage B1 group had MMVD with hemodynamically irrelevant mitral valve regurgitation as determined on the basis of an LA:AoSx < 1.6 or LA:AoLx < 2.6. Dogs allocated to the stage B2 group had MMVD with hemodynamically relevant mitral valve regurgitation as determined on the basis of an LA:AoSx ≥ 1.6 and LA:AoLx ≥ 2.6 but without evidence of previous (compensated) or concurrent (decompensated) CHF. Dogs allocated to the stage C-D group had MMVD and evidence of compensated or decompensated CHF.
Within a row, value differs significantly (P < 0.05) from that for the control group.
Within a row, value differs significantly (P < 0.05) from that for the stage B1 group.
— = Not determined.
The median age and left-sided murmur grade for dogs with MMVD (ie, dogs in the stage B1, B2, and C-D groups) were significantly greater than the corresponding values for dogs in the control group (Table 1). The sex distribution did not differ significantly (P = 0.87) among the 4 study groups. The LA size for dogs in the stage B2 and C-D groups was significantly greater than that for dogs in the control and stage B1 groups regardless of the method (LA:AoSx, LA:AoLx, or VLAS) used to assess LA size.
Of the 22 dogs in the stage C-D group, 12 (55%) had concurrent (decompensated) CHF and 10 (45%) had compensated CHF owing to the administration of various medications. Ten dogs in the stage C-D group received medications for the treatment of cardiac disease between the radiographic and echocardiographic examinations. All 10 of those dogs received furosemide, IV, and 7 also received pimobendan, PO. The medication regimens for the remaining 93 study dogs were not modified between the radiographic and echocardiographic examinations.
Correlation analyses
The VLAS was positively correlated with both LA:AoSx (rS, 0.70; 95% CI, 0.59 to 0.80; P < 0.001) and LA:AoLx (rS, 0.73; 95% CI, 0.63 to 0.81; P < 0.001; Figure 2). When the 10 dogs in the stage C-D group that received furosemide with or without pimobendan were removed from the analyses, the correlation between VLAS and LA:AoSx (rS, 0.73; 95% CI, 0.61 to 0.81) and LA:AoLx (rS, 0.76; 95% CI, 0.66 to 0.84) improved slightly.
Scatterplot of LA:AoSx (A) and LA:AoLx (B) versus VLAS for 103 dogs with a left-sided systolic murmur that underwent radiographic and echocardiographic examinations as part of a routine diagnostic evaluation at a veterinary teaching hospital from December 2015 through June 2016. Notice there was a significant (P < 0.001) positive correlation between VLAS and both LA:AoSx and LA:AoLx.
Citation: Journal of the American Veterinary Medical Association 253, 8; 10.2460/javma.253.8.1038
Scatterplot of LA:AoSx (A) and LA:AoLx (B) versus VLAS for 103 dogs with a left-sided systolic murmur that underwent radiographic and echocardiographic examinations as part of a routine diagnostic evaluation at a veterinary teaching hospital from December 2015 through June 2016. Notice there was a significant (P < 0.001) positive correlation between VLAS and both LA:AoSx and LA:AoLx.
Citation: Journal of the American Veterinary Medical Association 253, 8; 10.2460/javma.253.8.1038
Scatterplot of LA:AoSx (A) and LA:AoLx (B) versus VLAS for 103 dogs with a left-sided systolic murmur that underwent radiographic and echocardiographic examinations as part of a routine diagnostic evaluation at a veterinary teaching hospital from December 2015 through June 2016. Notice there was a significant (P < 0.001) positive correlation between VLAS and both LA:AoSx and LA:AoLx.
Citation: Journal of the American Veterinary Medical Association 253, 8; 10.2460/javma.253.8.1038
Diagnostic accuracy of VLAS for predicting LA enlargement
The diagnostic accuracy of and optimal LVAS cutoffs for the radiographic prediction of LA enlargement as determined by 3 sets of criteria for echocardiographic LA size were summarized (Table 2). Three cutoff values for VLAS were determined for each set of echocardiographic criteria for LA enlargement. One cutoff value represented the most clinically relevant value for VLAS (ie, value with the highest Youden index). One cutoff value represented the VLAS that yielded maximum sensitivity (ie, fewest false-negative results), and the remaining cutoff value represented the VLAS that yielded maximum specificity (ie, fewest false-positive results).
Diagnostic accuracy of various LVAS cutoffs for radiographic prediction of LA enlargement when LA enlargement was defined by each of 3 sets of echocardiographic criteria.
| ROC analysis | |||||||
|---|---|---|---|---|---|---|---|
| Echocardiographic criteria for LA enlargement | AUC (95% CI) | P value | VLAS cutoff (No. of vertebrae) | Cutoff type | Sensitivity (%) | Specificity (%) | Youden index |
| LA:AoLx ≥ 2.6 | 0.84 (0.77–0.92) | < 0.001 | ≥ 1.9 | Maximum sensitivity | 100 | 20 | 20 |
| ≥ 2.3 | Clinically relevant | 87 | 67 | 53 | |||
| ≥ 3.0 | Maximum specificity | 36 | 100 | 36 | |||
| LA:AoSx ≥ 1.6 | 0.84 (0.77–0.92) | < 0.001 | ≥ 1.9 | Maximum sensitivity | 100 | 21 | 21 |
| ≥ 2.5 | Clinically relevant | 67 | 84 | 50 | |||
| ≥ 2.8 | Maximum specificity | 49 | 100 | 49 | |||
| LA:AoLx ≥ 2.6 and | 0.88 (0.82–0.94) | < 0.001 | ≥ 1.9 | Maximum sensitivity | 100 | 16 | 16 |
| LA:AoSx ≥ 1.6 | ≥ 2.4 | Clinically relevant | 81 | 77 | 58 | ||
| ≥ 2.9 | Maximum specificity | 51 | 100 | 51 | |||
Data used for the analyses were obtained from the dogs of Table 1. The area under the ROC curve (AUC) was used to assess the diagnostic accuracy and quantify the predictive value of VLAS. Three VLAS cutoff values were determined for each set of echocardiographic criteria for LA enlargement. One cutoff value represented the most clinically relevant value for VLAS (ie, value with the highest Youden index). One cutoff value represented the VLAS that yielded maximum sensitivity (ie, fewest false-negative results), and the remaining cutoff value represented the VLAS that yielded maximum specificity (ie, fewest false-positive results).
See Table 1 for remainder of key.
VLAS measurement agreement
The VLAS as determined from right lateral thoracic radiographs (median, 2.3 vertebrae; IQR, 2.1 to 2.8 vertebrae) did not differ significantly (P = 0.40) from the VLAS as determined from left lateral thoracic radiographs (median, 2.2 vertebrae; IQR, 2.0 to 2.8 vertebrae). Bland-Altman analysis of the VLAS acquired from right versus left lateral thoracic radiographs revealed a mean ± SD bias of 0.02 ± 0.3 vertebrae (95% limits of agreement, −0.5 to 0.6 vertebrae). The intraobserver and interobserver ICCs for VLAS were high regardless of the lateral thoracic image (right or left) used for measurement. The intraobserver ICC for VLAS was 0.93 and 0.89 when measured on right and left lateral thoracic radiographs, respectively. The interobserver ICC for VLAS was 0.87 and 0.91 when measured on right and left lateral thoracic radiographs, respectively.
Discussion
The purpose of the study reported here was to describe how to measure VLAS, a new technique for quantifying LA size on thoracic radiographs, and determine its diagnostic value for predicting LA enlargement in dogs with MMVD when compared with echocardiographic measurement of LA size (gold standard). Results indicated that VLAS is a highly reproducible measurement that has a moderate positive correlation with echocardiographic measurements of LA size. Thus, the VLAS provided a quantitative measurement of LA size that can be determined from right or lateral thoracic radiographs, which will be useful for veterinarians who do not have ready access to echocardiography.
We believe that VLAS can provide valuable diagnostic information for dogs with a left-sided apical systolic murmur in conjunction with clinical signs of coughing or respiratory difficulty. Those signs often create a clinical conundrum for veterinarians because they can be caused by airway disease or cardiogenic pulmonary edema. Measurement of the VLAS could be helpful for determining whether the LA is enlarged. In our clinical experience, dogs with MMVD generally do not develop signs of CHF without at least some degree (frequently moderate) of LA enlargement. The exception to that observation is the rare dog with MMVD that develops acute CHF secondary to rupture of a major chordae tendinae; however, such dogs were excluded from the present study.
The veterinary literature contains few studies that describe quantitative radiographic measurement of the LA size in dogs. In 1 study,9 the diagnostic value of the tracheal bifurcation angle as determined from a dorsoventral thoracic radiograph for predicting LA:AoSx was evaluated. Unfortunately, the results of that study9 indicate that there is extensive overlap between the tracheal bifurcation angles of clinically normal dogs and dogs with varying degrees of LA enlargement; therefore, the sensitivity of the tracheal bifurcation angle was insufficient to support its use in clinical settings. In the present study, the sensitivity and specificity of VLAS for detection of LA enlargement were calculated when 2 echocardiographic measurements of LA size (LA:AoSx and LA:AoLx) were used as the gold standard. We believed that the use of 2 echocardiographic measurements to determine LA size would increase the likelihood of correctly diagnosing LA enlargement when it was present, and that comparison of VLAS with multiple echocardiographic measurements of LA size would be more advantageous than comparison with only 1 echocardiographic measurement for assessing its diagnostic value to detect LA enlargement. Also, the echocardiographic criteria (LA:AoLx ≥ 2.6 and LA:AoSx ≥ 1.6) used to diagnose LA enlargement in the present study were purposely selected because results of other studies4,5 indicate that dogs that meet those criteria likely have hemodynamically relevant MMVD and require therapeutic intervention. Therefore, accurate identification of those dogs, whether by radiographic (VLAS) or echocardiographic means, is clinically important for assessment of risks for CHF and anesthesia, development and modification of treatment regimens, and determining the prognosis. Results of the present study indicated that there was a significant positive correlation between VLAS and both LA:AoSx and LA:AoLx, and VLAS cutoffs of 2.3 to 2.5 vertebrae were associated with an LA:AoSx ≥ 1.6 and LA:AoLx ≥ 2.6 (or both). Thus, a VLAS ≥ 2.3 vertebrae can be used as a radiographic indicator of LA enlargement, and dogs with a VLAS ≥ 2.3 vertebrae likely have hemodynamically important MMVD.
In the present study, we used linear echocardiographic dimensions of the left atrium indexed to the aorta (LA:AoSx and LA:AoLx) to measure LA size and serve as clinical gold standards. However, those measurements are simply surrogates for LA volume estimates and do not represent a true gold standard of LA size (ie, LA volume). An example of a true gold standard is LA volume determined by multidetector CT,16 which is commonly measured in human medicine. We chose to compare VLAS with LA:AoSx and LA:AoLx because we believe most veterinarians use those ratios to quantify LA size. Also, linear measurement ratios represent an efficient and practical method to estimate LA size. Alternate volumetrically based methods for quantification of LA size in dogs are less efficient and largely unavailable, lack reference intervals, or, in the case of CT, require anesthesia. Finally, most of the veterinary literature regarding dogs with MMVD use linear measurement ratios of LA size, particularly LA:AoSx, to help guide clinical decisions.2–7
The present study was not without limitations. The clinical value of the VLAS cutoffs determined in this study is applicable only to dogs with left-sided systolic murmurs and MMVD. Further research is necessary to determine the diagnostic value of VLAS as a radiographic indicator of LA size for dogs with other diseases or in other scenarios. Also, measurement of the VLAS for a large number of healthy dogs of various breeds is necessary to establish reference intervals, as was done for vertebral heart size, a measurement performed in a similar manner as VLAS.17–19 For dogs suspected of having MMVD, VLAS should be considered complementary rather than confirmatory information, and echocardiography should be performed whenever possible. In the present study, use of VLAS to diagnose LA enlargement resulted in false-negative and false-positive results when an LA:AoSx ≥ 1.6, LA:AoLx ≥ 2.6, or both LA:AoSx ≥ 1.6 and LA:AoLx ≥ 2.6 were used as the gold standard. Moreover, 10 of the most severely affected dogs (dogs in the stage C-D group) received furosemide with or without pimobendan between the radiographic and echocardiographic examinations, which could have affected LA size measurements, even after only short-term administration.20,21 Although administration of those 2 drugs appeared to have a disadvantageous effect on our results, that effect was small, and removal of those 10 dogs from the analyses improved the positive correlations between VLAS and LA:AoSx and LA:AoLx only slightly. It should also be noted that all VLAS measurements were acquired from digital radiographs. The spatial and contrast resolutions differ slightly between digital and film-screen radiography. It is unlikely that those differences would substantially alter measurement of the VLAS; however, the superior contrast resolution of digital radiographs relative to that of film-screen radiographs might facilitate identification of the junction between the caudal border of the LA and caudal vena cava.
Results of the present study indicated that VLAS was an accurate predictor of LA enlargement in a large and diverse population of dogs with MMVD of varying severity. There was a moderate positive correlation between VLAS and echocardiographic estimates of LA size by linear measurement methods in both the long (LA:AoLx) and short (LA:AoSx) axes. Results also indicated that VLAS was a readily repeatable measurement, with a high level of agreement among measurements obtained by the same individual on multiple occasions as well as by multiple individuals. The VLAS measured on the right lateral thoracic image did not differ significantly from that measured on the left lateral thoracic image from the same dog; however, the 95% limits of agreement between the VLAS obtained from the right and left lateral images were fairly wide (−0.5 to 0.6 vertebrae). Therefore, the VLAS obtained from right lateral images should not be interchanged with that obtained from left lateral images when comparing that measurement among dogs or over time within the same dog. The findings of this study supported the use of VLAS to estimate LA size and predict LA enlargement in dogs with left-sided systolic murmurs suspected of having MMVD. Further research is necessary to determine the diagnostic value of VLAS for monitoring disease progression in dogs with MMVD and other diseases, such as dilated cardiomyopathy, characterized by LA enlargement.
Acknowledgments
Ms. Malcolm was supported by University of California-Davis School of Veterinary Medicine endowment funds and the Students Training in Advanced Research program. This study was not supported by a grant.
The authors declare that there were no conflicts of interest.
Presented as a poster presentation at the at the Merial-NIH National Veterinary Scholars Symposium, Columbus, Ohio, July 2016.
This work was performed at the William R. Pritchard Veterinary Medical Teaching Hospital at the University of California-Davis.
ABBREVIATIONS
| CHF | Congestive heart failure |
| CI | Confidence interval |
| ICC | Intraclass correlation coefficient |
| IQR | Interquartile (25th to 75th percentile) range |
| LA | Left atrial |
| LA:AoLx | Left atrium-to-aortic root ratio acquired from the long axis via echocardiography |
| LA:AoSx | Left atrium-to-aortic root ratio acquired from the short axis via echocardiography |
| MMVD | Myxomatous mitral valve disease |
| ROC | Receiver operating characteristic |
| VLAS | Vertebral left atrial size |
Footnotes
Sound Eklin, Carlsbad, Calif.
eFilm Workstation 3.3, Merge Healthcare Inc, Milwaukee, Wis.
IE33, Philips Healthcare, Andover, Mass.
Syngo Dynamic Workplace, version 10.0.01_HF04_Rev5 (Build 2884), Siemens Medical Solutions Inc, Malvern, Pa.
Strohm LE, Visser LC, Drost WT, et al. Two-dimensional long-axis echocardiographic ratios for assessment of left atrial and ventricular size in healthy dogs and dogs with mitral regurgitation (abstr), in Proceedings. Am Coll Vet Radiol Sci Conf 2016;28.
Prism 6 for Mac OS X, version 6.0f, GraphPad Software Inc, La Jolla, Calif.
MedCalc Statistical Software for Windows 10, version 16.4.3, MedCalc Software bvba, Ostend, Belgium.
References
1. Borgarelli M, Haggstrom J. Canine degenerative myxomatous mitral valve disease: natural history, clinical presentation and therapy. Vet Clin North Am Small Anim Pract 2010;40:651–663.
2. Borgarelli M, Crosara S, Lamb K, et al. Survival characteristics and prognostic variables of dogs with preclinical chronic degenerative mitral valve disease attributable to myxomatous degeneration. J Vet Intern Med 2012;26:69–75.
3. Reynolds CA, Brown DC, Rush JE, et al. Prediction of first onset of congestive heart failure in dogs with degenerative mitral valve disease: the PREDICT cohort study. J Vet Cardiol 2012;14:193–202.
4. Atkins CE, Keene BW, Brown WA, et al. Results of the veterinary enalapril trial to prove reduction in onset of heart failure in dogs chronically treated with enalapril alone for compensated, naturally occurring mitral valve insufficiency. J Am Vet Med Assoc 2007;231:1061–1069.
5. Boswood A, Häggström J, Gordon SG, et al. Effect of pimobendan in dogs with preclinical myxomatous mitral valve disease and cardiomegaly: the EPIC study—a randomized clinical trial. J Vet Intern Med 2016;30:1765–1779.
6. 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.
7. Häggström J, Boswood A, O'Grady M, et al. Effect of pimobendan or benazepril hydrochloride on survival times in dogs with congestive heart failure caused by naturally occurring myxomatous mitral valve disease: the QUEST study. J Vet Intern Med 2008;22:1124–1135.
8. Atkins C, Bonagura J, Ettinger S, et al. Guidelines for the diagnosis and treatment of canine chronic valvular heart disease. J Vet Intern Med 2009;23:1142–1150.
9. Le Roux A, Rademacher N, Saelinger C, et al. Value of tracheal bifurcation angle measurement as a radiographic sign of left atrial enlargement in dogs. Vet Radiol Ultrasound 2012;53:28–33.
10. 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.
11. Bonagura JD, Miller MW. Doppler echocardiography. II. Color Doppler imaging. Vet Clin North Am Small Anim Pract 1998;28:1361–1389 (vii).
12. Chetboul V, Tissier R. Echocardiographic assessment of canine degenerative mitral valve disease. J Vet Cardiol 2012;14:127–148.
13. Buchanan JW, Bücheler J. Vertebral scale system to measure canine heart size in radiographs. J Am Vet Med Assoc 1995;206:194–199.
14. 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.
15. 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.
16. Miyasaka Y, Tsujimoto S, Maeba H, et al. Left atrial volume by real-time three-dimensional echocardiography: validation by 64-slice multidetector computed tomography. J Am Soc Echocardiogr 2011;24:680–686.
17. Jepsen-Grant K, Pollard RE, Johnson LR. Vertebral heart scores in eight dog breeds. Vet Radiol Ultrasound 2013;54:3–8.
18. Lamb CR, Wikeley H, Boswood A, et al. Use of breed-specific ranges for the vertebral heart scale as an aid to the radiographic diagnosis of cardiac disease in dogs. Vet Rec 2001;148:707–711.
19. Marin LM, Brown J, McBrien C, et al. Vertebral heart size in retired racing Greyhounds. Vet Radiol Ultrasound 2007;48:332–334.
20. Häggström J, Lord PF, Höglund K, et al. Short-term hemodynamic and neuroendocrine effects of pimobendan and benazapril in dogs with myxomatous mitral valve disease and congestive heart failure. J Vet Intern Med 2013;27:1452–1462.
21. Fine DM, Durham HE Jr, Rossi NF, et al. Echocardiographic assessment of hemodynamic changes produced by two methods of inducing fluid deficit in dogs. J Vet Intern Med 2010;24:348–353.
