THE VERTEBRAL HEART SCORE (VHS) measurement has

VERTEBRAL HEART SCORES IN EIGHT DOG BREEDS K. JEPSEN-GRANT, R.E.POLLARD, L.R.JOHNSON The vertebral heart score (VHS) measurement is commonly used to provide a more objective measurement of cardiomegaly in canines. However, several studies have shown significant breed variations from the value previously established by Buchanan and Bücheler (9.7 ± 0.5). This study describes VHS measurements in Pug, Pomeranian, Yorkshire Terrier, Dachshund, Bulldog, Shih Tzu, Lhasa Apso, and Boston Terrier dog breeds. Dogs with two or three view thoracic radiographs, no subjective radiographic evidence of cardiomegaly, and no physical examination findings of heart murmurs or gallop rhythms were included in the study. The Pug, Pomeranian, Bulldog, and Boston Terrier groups were found to have a VHS significantly greater than 9.7 ± 0.5 (P < 0.00001, P = 0.0014, P < 0.0001, P < 0.00001, respectively). Body condition score was found to have a significant effect on the VHS of Lhasa Apso group. Anomalous vertebrae in the thoracic column were associated with a significant increase in VHS of the Bulldog (P = 0.028) and Boston Terrier (P = 0.0004) groups. Thoracic depth to width ratio did not have a significant effect on VHS. C 2012 Veterinary Radiology & Ultrasound. Key words: brachycephalic, canine, chondrodystrophic, vertebral heart score, VHS. Introduction THE VERTEBRAL HEART SCORE (VHS) measurement has been described as a more objective method for assessing cardiomegaly in dogs. The measurement is based on cardiac height and width and is normalized to overall body size by comparison to vertebral body length. 1 The original report of VHS measurements made by Buchanan and Bücheler evaluated 100 normal dogs representing a variety of breeds and established a reference range of 9.7 ± 0.5 vertebral bodies. 1 Since the original report, other studies have described breed-specific differences in this VHS reference range for Beagles, Greyhounds, and Whippets. 2 6 Additional studies have shown that nonprimary cardiac disease or changes in diagnostic technique can alter VHS measurement. In one previous study, cats with severe anemia showed a trend toward increased VHS in comparison to the normal VHS established in cats. 7 It was suggested that this trend was due to the activation of neuroendocrine compensatory mechanisms (such as the renin-angiotensinaldosterone system) leading to sodium and water retention and volume overload. Two previous studies found that values from right lateral radiographs were significantly greater than VHS values from left lateral radiographs. 2, 8 The original paper by Buchanan and Bücheler identified no signif- From the Department of Surgical and Radiological Sciences (Pollard), the Department of Medicine and Epidemiology (Johnson), Veterinary Medical Teaching Hospital, University of California, Davis, CA 95616 (Jepsen-Grant). Address correspondence and reprint request to R.E. Pollard, at the above address. E-mail: repollard@ucdavis.edu Received March 23, 2012; accepted for publication July 13, 2012. doi: 10.1111/j.1740-8261.2012.01976.x icant differences in left vs. right lateral radiographs. 1 In a study evaluating VHS of dogs under anesthesia, the VHS of manually ventilated dogs was smaller than the VHS of dogs with spontaneous ventilation. 9 A previous study has shown that the VHS of male dogs was greater than that of females. 5 A recent study evaluating airway collapse in small breed dogs with mitral regurgitation revealed that some dogs had a calculated VHS exceeding the normal range despite having normal cardiac dimensions on echocardiography. 10 Findings suggested that published reference ranges for VHS may not be applicable to small breed dogs. We hypothesized that the reference ranges for VHS of small breed, chondrodystrophic, or brachycephalic dogs would be higher than the published reference range of 9.7 ± 0.5 thoracic vertebrae. 1 If true, results of this study could help reduce false positive radiographic diagnoses of cardiomegaly in these dog breeds. Materials and Methods Medical records at the University of California, Davis William R. Pritchard Veterinary Medical Teaching Hospital were searched for Pugs, Pomeranians, Yorkshire Terriers, Dachshunds, Bulldogs, Shih Tzus, Lhasa Apsos, or Boston Terriers that had 2 or 3-view thoracic radiographs performed using a digital radiography system (Sound- Eklin, Carlsbad, CA) between January 1, 2006 and January 1, 2011. Inclusion criteria for the study were: age > 1.5 years, no record of a heart murmur or gallop sound Vet Radiol Ultrasound, Vol. 54, No. 1, 2013, pp 3 8. 3

4 JEPSEN-GRANTETAL. 2013 breed, dogs were grouped as having a low (1 5) or high (6 9) body condition score based on previously published criteria. 11 Thoracic depth to width ratio was calculated for each dog to determine the shape of the thorax for each dog. 1 Depth was assessed on the right lateral radiographic view from the dorsal margin of the xiphoid process to the ventral margin of the vertebral body, with the electronic calipers aligned perpendicular to the vertebral column (Fig. 2A). Width of FIG. 1. Demonstration of VHS measurement in a dog with a VHS of 9.4. detected on thoracic auscultation, and a radiographic report describing subjectively normal cardiac size. A random number generator was used to select a subset of 30 dogs from each breed that fit the inclusion criteria. The Bulldog category included French and English Bulldogs, as well as dogs whose breed was noted only as Bulldog. If the cardiac silhouette margins were obscured (i.e., by pleural effusion or alveolar lung pattern), that dog was removed from the study and another was selected from the database using the random number generator. Medical records of selected dogs were reviewed to extract body weight, body condition score (from 1 9), age at radiographic examination, and gender when available. If echocardiography was performed, results were also recorded. The VHS for all dogs was calculated as previously described (Fig. 1). 1 Electronic calipers were used to obtain dimensions of the long and short axis of the heart for comparison to the number of vertebral bodies, starting at thoracic vertebra number 4 (T4). A single observer (KJG) performed all VHS measurements, and was aware of patient signalment, physical examination results, and prior radiographic assessment. Using a right lateral radiographic view (except where noted), the long axis of the heart was measured from the ventral border of the carina to the most distant ventral contour of the heart, excluding any distance that had the radiographic opacity of fat. The short axis of the heart was determined perpendicular to the long axis at the level of the caudal vena cava and excluding any distance that had the radiographic opacity of fat. Measurements of both long and short axis were positioned over the thoracic vertebra, starting at T4 and the number of vertebra spanned was estimated to the nearest 0.1 vertebral body length. Addition of the two measurements resulted in the calculated vertebral heart score. The presence of hemi-vertebrae or abnormally shaped vertebrae within the thoracic vertebral column was recorded during data collection. Within each FIG. 2. (A) Thoracic depth measurement, from the xiphoid process to the perpendicular of the vertebral column. (B) Thoracic width measurement, measured from the insides of the 8th ribs.

VOL. 54, NO. 1 VERTEBRALHEARTSCORESIN EIGHT DOG BREEDS 5 the thorax was measured on a dorsoventral radiograph as the distance between the medial borders of the eight ribs (Fig. 2B). Dogs with a depth to width ratio < 0.75 were determined to have a barrel chest, while those with a depth to width ratio of > 1.25 were determined to have a deep chest. Statistical Analysis All statistical tests were selected and performed by the third author (L.R.J.). For all analyses, statistical testing was performed using GraphPad Prism (Version 5 San Diego, CA) and P < 0.05 was considered significant. Data were assessed for normality using the D Agostino & Pearson omnibus test and expressed as mean ± standard deviation (SD). 12 Mean VHS measurements were compared among dog breeds using analysis of variance. This analysis was repeated following exclusion of dogs with hemivertebrae. To assess effects of obesity, mean VHS measurements were compared between dogs with low (1 5) vs. high (6 9) Body condition score (BCS) within each breed using the Student s t-test. Mean VHS for each breed was compared to the published mean reference value (9.7) by use of a one-sample t-test. The number and percentage of dogs in each group with mean VHS > 2 SDs above the established reference range of 9.7 ± 0.5 was also calculated. Linear regression was used to assess the correlation between thoracic depth to width ratio and mean VHS within each breed. Results The sample population consisted of the following dogs: 30 Pugs, 18 Pomeranians, 30 Yorkshire Terriers, 29 Dachshunds, 30 Bulldogs, 30 Shih Tzus, 18 Lhasa Apsos, and 19 Boston Terriers. Clinical data on age, sex, body weight, and body condition score are summarized in Table 1. There was no significant difference in mean age among the dog breeds analyzed. Mean BCS differed significantly between Pugs (6.0 ± 1.6) and Lhaso Apsos (4.6 ± 1.4), P < 0.0036. Left lateral radiographs were used for VHS calculation in 1 Pomeranian (VHS of 11.3) and 1 Bulldog (VHS of 12.6). Right lateral radiographs were used for all other calculations. Mean and SD values for BCS and VHS are summarized, by breed group, in Table 1; and demonstrated graphically in Fig. 3. Mean VHS for Pug, Pomeranian, Bulldog, and Boston Terrier breed groups were significantly greater than the published reference VHS of 9.7 (P < 0.00001, P = 0.0014, P < 0.0001, P < 0.00001, respectively). Numbers of dogs within each breed group that had a VHS more than two SD above the published reference VHS are summarized in Table 1. Two Pugs (VHS of 14.1 and 11.1) and 1 Boston Terrier (VHS of 11) had echocardiograms that were assessed as normal. Thoracic hemivertebrae were identified in 2 Pugs, 13 Bulldogs, and 7 Boston Terriers. When VHS was recalculated after removal of these dogs, the breed mean VHS remained statistically different from the established value of 9.7 ± 0.5 (all P < 0.0001). Mean VHS for Bulldogs and Boston Terriers with normal vertebrae were statistically different from those with abnormal vertebrae (P = 0.028 and 0.0004, respectively) (Table 2). Effects of body condition score on VHS are summarized in Table 3. A significant effect was identified in the Lhaso Apso breed, with a higher VHS in high BCS group dogs (10.5 ± 0.6) vs. low BCS group dogs (9.4 ± 0.6), P = 0.007. Body condition score had a significant (P = 0.015), but weak correlation (R 2 = 0.03) with VHS across all dog breeds (Fig. 4). TABLE 1. Signalment Descriptions within Each Breed Group, Mean BCS and VHS for Each Breed Group, and Comparisons Between Breed Mean VHS vs. the Published Reference VHS of 9.7 Age (years) P Value # Dogs % Dogs M/F (mean ± SD) Weight (kg) BCS VHS (Breed VHS vs.) with VHS>2SD with VHS>2SD (N) (range) (mean ± SD) (mean ± SD) (mean ± SD) reference) vs. reference vs. reference Pug 11/19 6.9 ± 3.8 9.1 ± 2.7 6.0 ± 1.6 10.7 ± 0.9 <0.00001 13 43% (1.5 14) Pomeranian 9/9 7.4 ± 3.5 4.5 ± 2.4 5.5 ± 1.7 10.5 ± 0.9 0.0014 9 50% (2 13) Yorkshire Terrier 14/16 7.6 ± 4.3 3.2 ± 1.5 4.9 ± 1.4 9.9 ± 0.6 0.104 4 13% (2 15.5) Dachshund 11/18 6.8 ± 3.6 6.9 ± 2.8 5.8 ± 1.3 9.7 ± 0.5 0.788 2 7% (2.5 13.5) Bulldog 21/9 5.7 ± 3.2 26.7 ± 13.2 5.8 ± 1.2 12.7 ± 1.7 <0.0001 28 93% (2 14) Shih Tzu 9/21 6.5 ± 2.0 6.5 ± 2.0 5.1 ± 1.4 9.5 ± 0.6 0.059 0 0% (1.5 13) Lhasa Apso 9/9 7.4 ± 4.3 8.2 ± 3.2 4.6 ± 1.4 9.6 ± 0.8 0.57 1 6% (2 15) Boston Terrier 8/11 6.2 ± 3.2 8.9 ± 2.6 5.0 ± 0.9 11.7 ± 1.4 <0.00001 14 74% (2 13) M, male; F, female; N, number of dogs in each gender.

6 JEPSEN-GRANTETAL. 2013 FIG. 4. A significant (P = 0.015) but weak (R 2 = 0.03) correlation was found between BCS and VHS in all dogs. FIG. 3. Graphical representation of VHS measurements, with mean of each breed represented by the individual black lines. The highlighted box represents the reference value of 9.7 ± 0.5. All dog breed groups had individuals with a barrelshaped chest conformation (ratio <0.75), including 53% of Pugs and 80% of Yorkshire Terriers (Table 4). No significant correlation between VHS and thoracic depth to width ratio was identified for any breed group. Dogs within each breed were then divided into male and female groups and mean VHS between gender groups were compared. Values for each breed and gender group are sum- marized in Table 5. In the Yorkshire Terrier breed, mean VHS for females was greater than for males (P = 0.01). No other differences were identified. Discussion Findings from this study indicated that Pugs, Pomeranians, Bulldogs, and Boston Terriers without clinically apparent cardiac disease had significantly greater VHS values than the reference range of 9.7 ± 0.5 established by Buchanan and Bücheler. 2 6 In particular, the majority of Bulldogs had a VHS exceeding 12.0. One possible explanation for an increased VHS in chondrodystrophic dogs could be an artifactual increase in the number of TABLE 2. Comparisons Between Mean VHS for Normal vs. Abnormal Vertebrae Groups, within Bulldog and Boston Terrier Breeds Normal VHS 1 Abnormal VHS 2 P value vertebrae (N) (mean ± SD) vertebrae (N) (mean ± SD) (VHS1vs.2) Bulldog 17 12.1 ± 1.5 13 13.4 ± 1.6 0.028 Boston Terrier 12 11.4 ± 1.2 7 14.2 ± 1.6 0.0004 N = Number of Dogs in Each Vertebral Classification Group for Each Breed; VHS 1, VHS for Dogs with Normal Vertebrae; VHS 2, VHS for Dogs with Abnormal Vertebrae. TABLE 3. Comparisons Between Mean VHS for Low vs. High BCS Groups, Within Breeds BCS 1 5 VHS 1 BCS 6 9 VHS 2 P value (N) (mean ± SD) (N) (mean ± SD) (VHS 1 vs. 2) Pug 12 10.43 ± 0.58 19 10.95 ± 1.08 0.1307 Pomeranian 7 10.16 ± 0.95 10 10.82 ± 0.68 0.1139 Yorkshire Terrier 17 9.77 ± 0.59 9 9.86 ± 0.53 0.705 Dachshund 14 9.74 ± 0.61 11 9.69 ± 0.45 0.789 Bulldog 11 12.59 ± 1.64 17 12.73 ± 1.81 0.838 Shih Tzu 17 9.51 ± 0.59 11 9.4 ± 0.5 0.62 Lhasa Apso 13 9.37 ± 0.63 4 10.5 ± 0.634 0.007 Boston Terrier 13 11.98 ± 1.57 3 10.67 ± 0.49 0.1851 N, Number of Dogs in Each BCS Classification Group for Each Breed; VHS 1, VHS for Dogs with BCS 1 5; VHS 2, VHS for Dogs with BCS 6 9.

VOL. 54, NO. 1 VERTEBRALHEARTSCORESIN EIGHT DOG BREEDS 7 TABLE 4. Thoracic Depth to Width Ratios for Each Breed Group and Percentage of Dogs Within Each Group Meeting Published Criteria for Barrel-Shaped Chest Conformation (Ratio <0.75) Thoracic depth: N width ratio SD % < 0.75 Pug 30 0.75 0.8 53% Pomeranian 18 0.75 0.08 44% Yorkshire Terrier 30 0.7 0.07 80% Dachshund 29 0.83 0.07 10% Bulldog 30 0.77 0.1 33% Shih Tzu 30 0.79 0.06 23% Lhasa Apso 18 0.8 0.08 17% Boston Terrier 19 0.78 0.07 37% N, Number of Dogs in each Breed Group; SD, Standard Deviation for Thoracic Depth: Width Ratio Within Each Breed Group. TABLE 5. Comparisons Between Mean VHS for Males and Females Within Each Breed Group P value M/F Male Female (Male VHS vs. (N) VHS VHS Femal VHS) Pug 11/19 10.5 ± 0.7 10.8 ± 1.0 0.51 Pomeranian 9/9 10.3 ± 0.95 10.6 ± 0.8 0.45 Yorkshire Terrier 14/16 9.6 ± 0.4 10.2 ± 0.7 0.01 Dachshund 11/18 9.7 ± 0.36 9.7 ± 0.6 0.81 Bulldog 21/9 12.6 ± 1.7 12.9 ± 1.8 0.62 Shih Tzu 9/21 9.3 ± 0.6 9.6 ± 0.5 0.22 Lhasa Apso 9/9 9.9 ± 0.9 9.3 ± 0.6 0.12 Boston Terrier 8/11 11.6 ± 1.5 11.6 ± 1.5 1.0 M, Male; F, Female; N, Number in Each Gender. vertebrae spanned by cardiac measurements due to the presence of hemivertebrae. When dogs with anomalous vertebrae were removed from consideration, we found that the remaining dogs in some breeds persisted in having greater VHS measurements compared with the original VHS value of 9.7 ± 0.5. However, within the Bulldog and BostonTerrier breeds, VHS values were significantly greater in dogs with abnormal vertebrae compared to those with normal vertebrae. These findings indicate that vertebral abnormalities may artifactually increase VHS in Bulldogs and Boston Terrier breeds but may not alter VHS in other breeds. It is also possible that severe lordosis or kyphosis of the vertebral column could skew alignment of cardiac measurements and alter the measurement of VHS, however this factor was not investigated in our study. Another factor affecting VHS in small breeds could be breed-specific variations in vertebral body length relative to overall body size. A study directed at assessing vertebral body length compared to overall body size in multiple breeds would be necessary to further evaluate this effect. In this study, we attempted to exclude fat opacity from cardiac measurements during original scoring, although inclusion of some pericardial fat in the measurements was likely inevitable. To assess the effect of pericardial fat on VHS, we compared VHS in dogs with low/normal BCS vs. those of dogs with higher BCS. With the exception of the Lhasa Apso, the breed with the lowest BCS overall in this study, no significant differences were detected in VHS between dogs with low vs. high BCS. Although VHS was correlated with BCS, it was a very weak correlation and considered unlikely to be of clinical significance. We performed a correlation analysis between VHS and the thoracic depth to width ratio to determine whether a barrel chested vs. deep chested conformation was responsible for a significant portion of the variation in VHS among breeds. No significant correlation was identified. The highest proportion of barrel-chested dogs (thoracic depth to width ratios of <0.75) were in the Yorkshire Terriers (80%) and Pugs (53%). No dog in this study was identified as having a deep chest (ratio >1.25), which may have impacted our ability to establish a correlation of VHS to thoracic shape. One limitation of our study design was that we did not analyze effects of noncardiac-related diseases, anesthesia, or medical interventions, or whether right lateral radiographs were different from left lateral radiographs. Another limitation of our study was small sample size. Due to the presence of heart disease in many of the Lhasa Apso and Boston Terriers that were examined during our study s time frame, only 18 and 19 dogs in these breed groups, respectively, could be included. A larger sample size could possibly have yielded different results. However, even with low numbers, we were able to demonstrate that Boston Terriers had a significantly higher VHS than the reported reference value. In the case of Bulldogs, both English and French bulldogs were included in one group and evaluation of individual VHS of those specific bulldogs might have yielded different results. It is also possible that some of our dogs with a subjectively normal heart size on radiographs actually had occult heart disease, as only three of the dogs had echocardiography performed to assess the structure and function of the heart. Underlying heart disease could therefore have skewed the VHS values higher than the reported normal. However, none of the included dogs had clinical evidence of a heart murmur or gallop rhythm. Findings from our study indicate that application of a single, generic VHS measurement to small breed dogs, especially dogs with thoracic vertebral anomalies, could yield a false positive identification of cardiomegaly. A high body condition score could also yield a false positive diagnosis of cardiomegaly, however overall correlations between VHS and BCS were weak in our dogs. Findings support the use of breed-specific VHS reference ranges for objective measurements of cardiac size, however comparisons to subjective radiographic assessments of cardiac size and echocardiographic data are still recommended.

8 JEPSEN-GRANTETAL. 2013 1. Buchanan JW, Bücheler J. Vertebral scale system to measure heart size in radiographs. J Am Vet Med Assoc 1995;206:194 199. 2. Bavegems V, Van Caelenberg A, Duchateau L, Sys SU, Van Bree H, De Rick A. Vertebral heart size ranges specific for whippets. Vet Radiol Ultrasound 2005;46:400 403. 3. Kraetschmer S, Ludwig K, Meneses F, Nolte I, Simon D. Vertebral heart scale in the beagle dog. J Small Anim Pract 2008;49:240 243. 4. Lamb CR, Tyler M, Boswood A, Skelly BJ, Cain M. Assessment of the value of the vertebral heart scale in the radiographic diagnosis of cardiac disease in dogs. Vet Rec 2000;146:687 690. 5. Lamb CR, Wikeley H, Boswood A, Pfeiffer DU. 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. 6. Marin LM, Brown J, McBrien C, Baumwart R, Samii VF, Couto CG. Vertebral heart size in retired racing Greyhounds. Vet Radiol Ultrasound 2007;48:332 334. REFERENCES 7. Wilson HE, Jasani S, Wagner TB, Benigni L, Milne JR, Stokes AL, Luis-Fuentes V. Signs of left heart volume overload in severely anaemic cats. J Feline Med Surg 2010;12:904 909. 8. Greco A, Meomartino L, Raiano V, Fatone G, Brunetti A. Effect of left vs. right recumbency on the vertebral heart score in normal dogs. Vet Radiol Ultrasound 2008;49:454 455. 9. Webster N, Adams V, Dennis R. The effect of manual lung inflation vs. spontaneous inspiration on the cardiac silhouette in anesthetized dogs. Vet Radiol Ultrasound 2009;50(2):172 177. 10. Singh MK, Johnson LR, Kittleson KD, Pollard RE. Bronchomalacia in Dogs with Myxomatous Mitral Valve Disease. J Vet Intern Med 2012;26:312 319. 11. Laflamme D. Development and validation of a body condition score system for dogs. Canine Pract 1997;22:10 15. 12. D Agostino RB, Belanger A, D Agostino RB Jr. A suggestion for using powerful and informative tests of normality. Am Stat 1990;44:316 321.