Preliminary results presented at the American College Veterinary Internal Medicine Louisville, USA 2006.
Corresponding author: Michele Borgarelli, DVM, PhD, Dipl. ECVIM-CA (Cardiology), Department of Clinical Science, College of Veterinary Medicine, 111B Mosier Hall, KS 66506; e-mail: firstname.lastname@example.org.
Background: There are few studies evaluating the natural history and prognostic variables in chronic mitral valve disease (CMVI) in a heterogeneous population of dogs.
Objectives: To estimate survival and prognostic value of clinical and echocardiographic variables in dogs with CMVI of varying severity. Five hundred and fifty-eight dogs belonging to 36 breeds were studied.
Methods: Dogs were included after clinical examination and echocardiography. Long-term outcome was assessed by telephone interview with the owner.
Results: The mean follow-up time was 22.7 ± 13.6 months, and the median survival time was 19.5 ± 13.2 months. In univariate analysis, age>8 years, syncope, HR>140 bpm, dyspnea, arrhythmias, class of heart failure (International Small Animal Cardiac Health Council), furosemide therapy, end-systolic volume-index (ESV-I)>30 mL/m2, left atrial to aortic root ratio (LA/Ao)>1.7, E wave transmitral peak velocity (Emax)>1.2 m/s, and bilateral mitral valve leaflet engagement were associated with survival time when all causes of death were included. For the cardiac-related deaths, all the previous variables except dyspnea and EDV-I>100 mL/m2 were significantly associated with survival time. Significant variables in multivariate analysis (all causes of death) were syncope, LA/Ao>1.7 m/s, and Emax>1.2 m/s. For cardiac-related death, the only significant variable was LA/Ao>1.7.
Conclusions and Clinical Importance: Mild CMVI is a relatively benign condition in dogs. However, some clinical variables can identify dogs at a higher risk of death; these variables might be useful to identify individuals that need more frequent monitoring or therapeutic intervention.
Mitral regurgitation attributable to myxomatous valve disease, also known as chronic mitral valve insufficiency (CMVI), is the most common heart disease in dogs. It has been estimated that CMVI accounts for 75–80% of cardiac diseases in dogs.1 The disease is caused by progressive myxomatous degeneration of either the mitral valve alone or together with the tricuspid valve,1,2 leading to incomplete coaptation of the leaflets and valvular regurgitation. Myxomatous mitral valve disease has been shown to be associated with valve thickening and abnormal motion of the mitral leaflets upon echocardiogram analysis in both humans and dogs. The condition has been shown to represent the most common cause of mitral regurgitation in dogs.2–4 Gross and histological findings are comparable to the changes found in mitral valve prolapse (MVP) in humans,5 which is usually benign, although serious complications may develop. An age older than 50 years, depressed left ventricular function, moderate to severe mitral regurgitation, atrial enlargement, mitral valve thickness >5 mm, and atrial fibrillation have been reported to represent significant risk factors for cardiovascular events, including sudden death.6 In dogs, the disease is characterized by a slow progression over years, with many affected dogs that, because of the age of onset, never progress to reveal clinical signs of overt heart failure (HF) before death. Surprisingly, although CMVI is the most common cause of HF, there are very few studies evaluating the natural history and potential risk factors for progression of the disease in this species. Previously reported risk factors in dogs for progression of disease include age, gender, intensity of heart murmur, degree of valve prolapse, severity of valve lesions, and the degree of mitral valve regurgitation.7–10 However, these studies have generally been conducted on specific breeds such as Dachsund or Cavalier King Charles; prognostic data and survival analysis of a more heterogeneous population including large breed dogs are thus not available.
The aims of the present study were to estimate the survival times in dogs with varying severities of MR attributable to CMVI and to assess the prognostic value of clinical and echocardiographical variables on survival.
Material and Methods
The medical records of dogs examined at 4 referral centers in northern Italy between January 2000 and June 2005 were reviewed. From these records, 722 cases diagnosed with CMVI were identified. All dogs had been presented for a cardiology consultation because of previous identification of a heart murmur, because of the presence of clinical signs possibly indicating cardiovascular disorder including cough and exercise intolerance, or both. Out of the initial 722 cases, 157 were eliminated from further evaluation as it was impossible to contact the owner to obtain follow-up information or because echocardiographic examination had not been performed.
All dogs included in the study had to have undergone physical workup and echocardiographic examination. Echocardiographic inclusion criteria were the combination of presence of MVP, any degree of mitral valve leaflet thickening by 2-D echocardiography, and the identification of any degree of mitral valve regurgitation by color Doppler examination, with or without mitral valve thickening. The left ventricular fractional shortening (FS) had to be more than 20%. Finally, the owners had to be available for a telephone interview.
Dogs that presented with congenital heart disease or acquired cardiovascular disorders that directly or indirectly affected the mitral valve or its function, such as bacterial endocarditis or dilated cardiomyopathy, were excluded from the study. Mitral endocarditis was excluded based on clinical findings, and the lack of obvious large vegetative lesions with heterogeneous appearance, as detected by echocardiography.11 Dilated cardiomyopathy was excluded based on the presence of valve changes consistent with myxomatous mitral valve disease and MVP and the absence of echocardiographic criteria such as an FS<20%.12
The following data were obtained from case records: gender, age, body weight, heart rate (HR), presence and intensity of heart murmur, systolic blood pressure (SBP), presence of dyspnea, case history of syncope, presence of arrhythmias, and baseline treatment. Dyspnea was defined as labored or difficult breathing. Blood pressure was measured noninvasively with Doppler sphygmomanometry.a The following echocardiographic data were retrieved: end-diastolic and end-systolic volume indexes (EDV-I and ESV-I), left atrial to aortic root ratio (LA/Ao), description of mitral valve leaflet morphology (anterior, posterior, or both), and transmitral flow data. The latter included peak E-wave (Emax) velocity (early filling) and E-wave deceleration time (Edt). In each dog, based on clinical signs and thoracic radiographs, the severity of HF was classified according to the International Small Animal Cardiac Health Council (ISACHC) recommendations.13 All clinical datasets were reviewed by a single experienced investigator (MB).
All dogs had previously undergone a complete echocardiographic examination, which included transthoracic 2-D, M-mode, spectral, and color flow Doppler. Transducer arrays of 5.0–7.5 and 2.5–3.5 MHz were used.b Examinations were performed in conscious, unsedated dogs. Right parasternal M-mode recordings were obtained from short-axis views with the dogs positioned in right lateral recumbency, and the 2-D echocardiograms were obtained in accordance with techniques described elsewhere.14,15
The presence of MVP and mitral valve thickening was evaluated from the right parasternal long-axis, the right parasternal 4-chamber view, and left apical 4-chamber view. Mitral valve prolapse was defined as any systolic displacement of one or both mitral valve leaflets basal to the mitral annulus observed in at least two of these views.3 The presence of mitral valve regurgitation was evaluated by color Doppler in the right parasternal long-axis and left apical views.
All echocardiographic measurements were made by 4 investigators (MB,RAS,MP,DC), and were reviewed by 1 experienced investigator (MB) with videotape recordings. M-mode measurements were obtained according to the leading-edge-to-leading-edge method. The EDV and ESV were calculated by the Teicholz method: EDV=[7 × (EDD)3]/(2.4 + EDD) and ESV=[7 × (ESD)3]/(2.4 + ESD)16 and values were successively indexed for body surface area to obtain the EDV-I and the ESV-I. The LA/Ao was obtained from the 2-D short-axis view.17
Clinical Progress and Survival
The clinical progress of each dog was ascertained by telephone interview with the owner. The interviews were conducted by specifically trained senior students, and the results were recorded in an electronic questionnaire. The questionnaire consisted of questions with a definite number of possible answers, most commonly yes/no. The interviewer was not blinded to the clinical status of the dog at the initial examination. The owner was asked if the dog was dead or alive. If the dog was dead, the owner was asked if the dog had been euthanized or died spontaneously, reasons for euthanasia, and, in case of spontaneous death, the possible causes, including cardiac-related sudden death, presence of syncope, or progression of HF were probed. Cardiac-related death was defined as death occurring because of progression of clinical signs of HF. Dogs that were euthanized because of refractory HF were scored as cardiac-related deaths. In this study, sudden death was defined as death occurring during sleep or activity such as running, or within 2 hours after the dog showed sudden signs of HF (dyspnea). Sudden death was regarded as cardiac-related if no other cause of death was obvious. A survival analysis was performed on all causes of deaths and on cardiac-related deaths separately.
Statistical analysis was performed by a freeware statistical software package (R 2.3.0).c Normal distribution of data was assessed by the Shapiro Wilk normality test. Descriptive statistics were used for gender, age, body weight, heart rate, the presence and intensity of heart murmur, presence of dyspnea, syncope, class of HF and presence of arrhythmia, and all 2-D, M-mode, and Doppler-derived variables. Numerical variables were reported as the mean ± standard deviation (SD). Effects on survival of the 16 clinical, ECG, echocardiographic, and Doppler variables were evaluated, which included gender, dyspnea, syncope, age>8 years, weight>20 kg, HR>140 bpm, murmur>II/VI, SBP>140 mmHg, class of HF according to ISACHC classification, furosemide treatment (yes/no), affected leaflet (anterior, posterior, or both), EDV-I>100 mL/m2, ESV-I>30 mL/m2, LA/Ao>1.7, Emax>1.2 m/s, and Edt<80 ms.
Univariate Cox survival analysis was used in survival analysis to evaluate the hazard ratio of an adverse event. Survival curves, median survival times, and 95% CIs were obtained by the Kaplan-Meier method. Survival time was counted from the day of diagnosis of CMVI at the referral center to either the day of death or closing time of the study. End-point of the study was death (all causes). A subanalysis was performed including only deaths that were considered cardiac-related. Dogs available to follow up but for which the time point of death was not available or were still alive were censored. Univariate Cox survival model was used to define significant variables (P<.05) for subsequent analysis. Multivariate analysis was performed by Cox proportional hazard regression. Results of multivariate analysis were considered significant when P values were <.05 with a hazard ratio confidence interval of 95%, excluding 1.
Five hundred and fifty-eight dogs from 36 breeds were included in the analysis; 157 dogs had been excluded from the analysis, and of these, 143 (91.1%) were not included because it was not possible to contact the owner (ie change of address) and 14 (8.9%) because echocardiographic data were missing. The dogs not available for followup had different classes of HF: 76 were in class 1 (48.4%), 48 in class 2 (30.6%), and 19 (12.1%) were in class 3 ISACHC, whereas the ISACHC class in dogs with echocardiographic data missing could not be established. The majority of dogs included in the study were small (< 20 kg) mixed breed dogs (n=176; 31%) followed by Yorkshire Terrier (n=58; 10.4%) and Miniature Poodle (n=50; 9.0%). There were 180 females and 378 males; the mean age was 10.6 ± 2.62 years (range 3–18 years), and the mean weight was 11.6 ± 9.3 kg (range 2–67 kg). Four hundred and seventy-six dogs weighed ≤20 kg and 77>20 kg. At baseline examination, 67 dogs (12%) had dyspnea, 115 (20.6%) had a case history of syncope, and 17 dogs (3.0%) presente an arrhythmia on the ECG recording. The mean HR (n=171) was 121 ± 28 bpm. All dogs presented with a systolic murmur that was characterized as a soft midsystolic murmur (1–2/6) in 46 dogs (8.2%) and as a moderate to loud holosystolic murmur (3–5/6) in 347 dogs (62.2%). The intensity of the murmur was not characterized in 165 dogs (29.6%). Three hundred and two dogs (54.1%) had asymptomatic disease (class I ISACHC), 157 (28.1%) were in class II ISACHC, and 99 (17.8%) were in class III ISACHC. The mean EDV-I and ESV-I (n=555) were 107.1 ± 67.04 and 58.75 ± 59.3 mL/m2, respectively. The mean LA/Ao (n=526) was 1.8 ± 0.7. The mean Emax (n = 377) was 0.94 ± 0.45 and the mean Edt (n = 298) was 93 ± 28 ms. Tables 1a and 1b report clinical and echocardiographic variables for each ISACHC HF class.
Table 1a. Report on mean ± SD variable for each ISACHC heart failure class.
Mean ± SD
Mean ± SD
Mean ± SD
ESV-I, end systolic volume index; EDV-I, end diastolic volume index; LA/Ao, left atrium aortic root ratio; Edt, E wave deceleration time; E peak, peak velocity of E wave; HR, heart rate; SD, standard deviation; SBP, systolic blood pressure; ISACHC, International Small Animal Cardiac Health Council.
10.24 ± 2.78
11.16 ± 2.49
10.91 ± 2.13
11.89 ± 10.01
11.25 ± 8.46
11.12 ± 8.51
34.29 ± 29.67
65.59 ± 53.5
40.18 ± 26.91
93.94 ± 46.65
131.64 ± 75.9
188.39 ± 57.77
1.45 ± 0.47
2.04 ± 0.52
2.49 ± 0.79
95.03 ± 29.23
90.85 ± 31.04
73.76 ± 20.55
E peak (m/s)
0.78 ± 0.34
1.08 ± 0.43
1.41 ± 0.54
121.47 ± 26.65
122.63 ± 29.36
141.72 ± 17.15
161.30 ± 27.34
150.89 ± 32.08
142.50 ± 23.88
Table 1b. ISACHC class distribution for the discrete studied variables.
MVL, mitral valve leaflet; ISACHC, International Small Animal Cardiac Health Council.
By 2-D-echocardiogram, all dogs presented MVP and thickening of mitral valve leaflets. Description of which leaflet was affected was available for 189 dogs. In this subgroup bilateral mitral valve leaflet involvement was identified in 105 animals (55.5%), 80 (42.3%) had only the anterior leaflet affected, and 4 dogs (2.2%) had only the posterior leaflet affected. At baseline, 339 (60.7%) dogs were receiving medical therapy for HF. Of these dogs, 272 dogs received furosemide alone or in combination with other drugs. The baseline treatments are summarized in Table 2.
Table 2. Distribution of medical treatments at inclusion.
The mean followup time of this study was 22.7 ± 13.6 months. Altogether 252 dogs (45.2%) had died or were euthanized during the observation period; 174 deaths were considered as cardiac related. Ninety-one dogs (36.1%) had been euthanized, 71 (28.2%) of which had been requested because of refractory HF. Twenty-two dogs (8.7%) had sudden death. Of these, 5 were in class I, 8 were in class II, and 9 in class 3 ISACHC class at initial examination. The median survival time, regardless cause of death, was 19.5 ± 13.2 months (range 0.03–72 months). Of the 16 variables that were used as predictors in Kaplan-Meier analysis, age, syncope, HR >140 bpm, dyspnea, arrhythmias, class of HF, treatment with furosemide, ESV-I>30 mL/m2, LA/Ao>1.7, Emax>1.2 m/s, and bilateral mitral valve leaflet involvement were significantly associated with overall survival time (Fig 1; Table 3). For the cardiac-related deaths, all the previous significant variables, except dyspnea, had a significant relationship with survival. Moreover, EDV-I>100 mL/m2 had a significant effect on survival (Fig 2, Table 4). Syncope was not associated with an increased risk of sudden death.
Table 3. Probability of adverse effect of each of 15 studied variables in the univariate analysis when all causes of death were included in the dataset.
HR, heart rate; LA/Ao, left atrium aortic root ratio; ant MVL, anterior mitral valve leaflet; post-MVL, posterior mitral valve leaflet; EDV-I, end-diastolic volume index; ESV-I, end-systolic volume index; E peak, peak velocity of E wave; Edt, E wave deceleration time; CI, confidence interval; HRa, hazard ratio; NS, not significant.
n = 58
n = 553
n = 549
Age < 8 years
n = 172
n = 393
n = 510
n = 106
n = 394
n = 117
n = 526
n = 185
n = 555
n = 555
n = 377
E peak<1.2 m/s
E peak>1.2 m/s
n = 298
Table 4. Probability of adverse effects of each of 15 studied variables in the univariate analysis of the survival time when only cardiac related death was included in the dataset.
HR, heart rate; LA/Ao, left atrium aortic root ratio; ant MVL, anterior mitral valve leaflet; post-MVL, posterior mitral valve leaflet; EDV-I, end-diastolic volume index; ESV-I, end-systolic volume index; E peak, peak velocity of E wave; Edt, E wave deceleration time; HRa, hazard ratio; CI, confidence interval; NS, not significant.
n = 463
n = 460
n = 455
n = 141
n = 301
n = 408
n = 90
n = 330
n = 96
n = 436
n = 158
n = 459
n = 459
n = 158
E peak<1.2 m/s
E peak>1.2 m/s
n = 246
Multivariate Survival Analysis
In multivariate analysis, syncope (hazard ratio 3.3, 95% CI = 1.30–7.74, P<.05), LA/Ao>1.7 (hazard ratio = 3.70, 95% CI = 1.79–7.63, P<.01), Emax>1.2 (hazard ratio = 3.10, 95% CI = 1.48–6.25, P<.01) were independent predictors of all-cause mortality. When only dogs with cardiac-related death were included in the analysis, only LA/Ao>1.7 (hazard ratio = 2.1, 95% CI=1.31–3.91, P<.0001) had a significant effect on survival time, although Emax>1.2 m/s (P=.054) was very close to the chosen 5% level of significance. Mitral valve leaflet engagement and HR were excluded from multivariate analysis because of the relatively low number of dogs for which these data were available. Furthermore, furosemide treatment (yes/no) was also excluded from this analysis because it was highly covariate with class of HF.
Many studies have reported survival time and prognostic indicators in dogs with CMVI. However, these studies were focused mainly on specific breeds and did not include large breed dogs,8–11,18–20 or focused on specific aspects of the disease, such as influence on survival after chordal rupture21 or effect of therapy on survival time.22–24 Accordingly, the importance of the present study is that it documents the long-term outcome and influence of certain clinical and echocardiographic variables on survival in a large series of dogs of different breeds and weight affected with CMVI. In this large study, 384 of 558 dogs survived or died for causes unrelated to CMVI during the observation period. This finding may indicate that CMVI is a comparably benign condition as previously reported in both humans and dogs.1,22,25–28 This is particularly apparent for dogs in ISACHC class 1 HF. More than 60% of the dogs in this group were still alive after 70 months after initial diagnosis. This finding is in agreement with previous reports.21,22 The mean interval time to develop HF in 113 asymptomatic Cavalier King Charles Spaniels without any medical treatment was reported to be 27.2 ± 13.5 months.22 A recent study reporting survival data in a group of dogs with CMVI and chordae tendineae rupture reported a survival rate of 75% at 24 months for 28 dogs in class I ISACHC.21 However, in dogs CMVI has been linked to serious complications, including progressive HF, acute exacerbation of pulmonary congestion and edema, severe arrhythmias, right-side HF caused by pulmonary hypertension, and left atrial rupture and cardiac tamponade. Identification of animals that are at high risk for complications might allow targeting individuals in need of therapy for progression of HF. For low-risk patients, unwarranted owner concerns of HF might also be avoided.
The most significant clinical and echocardiographic independent predictors in our study for survival when all causes of deaths were included were syncope, LA/Ao ratio, and Emax. However, when only cardiac-related deaths were included only the LA/Ao ratio maintained statistical significance. Left atrial enlargement reflects the degree of severity and chronic nature of mitral valve regurgitation, whereas peak velocity of early transmitral filling is dependent on atrial volume and pressure.29 Consequently, increased LA/Ao ratio and Emax can be considered as indicators of left atrium volume overload and increased left atrial pressure. Therefore, the data from this study indirectly agree with reports from other studies in both humans and dogs, which suggest that severity of mitral valve regurgitation is a predictor of poor prognosis.9,28,30
Syncope was related to mortality when all causes of death were considered. Tussive fainting that may occur in conjunction with paroxysm of coughing is common in dogs with chronic pulmonary, bronchial disease, or both. In the latter case, the frequency and severity of episodes are often associated with severity of mitral regurgitation because the left atrium compresses the left mainstem bronchus. In dogs with severe CMVI, syncope may be induced by strenuous exercise, by cardiac tamponade caused by left atrial rupture, or by pulmonary hypertension.1,31 However, syncope can also be the consequence of tachyarrhythmias, which are associated with disease severity, the most common being atrial fibrillation or intermittent supraventricular tachycardia.1 Except in cases with primary lung disease, syncope is associated with the severity of mitral regurgitation, and it is therefore likely that the cases with syncope in our study had more severe mitral regurgitation. Consequently, it is not surprising it was associated with an increased risk of death.1
Chronic mitral valve disease has been reported to more commonly affect male dogs,1 which was confirmed in this study. However, gender did not influence survival. This study found that age more than 8 years was associated with a reduced survival time, both when all and only cardiac causes of death were included. This is in agreement with previous studies.1,32 However, age was not an independent prognostic variable in multivariate analysis, presumably because this variable was highly covariate with severity of MR.
In contrast to previous studies,8,22,32,33 we found no significant effect of heart murmur intensity on survival time. This finding can partly be explained by the fact that the present study included a greater diversity of breeds with different body sizes and chest conformations. Furthermore, the present study grouped only intensity of murmurs in 2 groups: those with a low intensity and those with moderate to high intensity murmurs. The outcome of this grouping was that a relatively low number of dogs were included in the low-intensity group, whereas the majority of dogs were included in the moderate- to high-intensity group. Presumably, the imbalance in number and the greater diversity of severity of MR in the moderate to high intensity group may contribute to the lack of significance of this variable in survival analysis. Although other studies included dogs in HF, the present study also included dogs with atrial fibrillation. This condition may cause a more pronounced reduction in the pressure gradient between the left atrium and ventricle and thereby a reduced murmur intensity.
Although the presence of dyspnea was significantly associated with an increased risk of death when all causes of death were included in the analysis, it was not significant when only cardiac deaths were included. This finding is surprising, but it could be caused by the high prevalence of primary respiratory disease in geriatric dogs. In these cases it is more likely that dyspnea is caused by the primary respiratory disease, and that this could determine shorter survival. On the other hand, the lack of statistical significance for dyspnea in dogs that died of HF might also be caused by the relatively low numbers of dogs in which the presence of this sign was reported in the clinical records.
The finding of sudden death in 8.9% is noteworthy. To the authors' knowledge, this is the first time sudden death has been reported in asymptomatic dogs with CMVI. Humans with MVP have twice the risk of the general population for sudden death.34,35 Ventricular arrhythmias, substantial mitral regurgitation, redundant tendinous chords, and depressed left ventricular function have been associated with an increased risk of sudden death in humans.5 In dogs, sudden death in animals with CMVI is a comparably uncommon event and has been associated with nonatherogenic dysplasia of the intramural coronary arteries.d Preliminary data from Holter recording in dogs with CMVI indicate that ventricular arrhythmias may be more common than previously thought.
In humans, systolic dysfunction has been associated with an increased risk of adverse events.35,36 In this regard, we recently reported that dogs with moderate HF caused by CMVI have reduced systolic function.37 However, increased ESV-I was associated with a reduction in survival time in univariate analysis, but was not an independent risk factor as it was associated with class of HF. Although not included in the multivariate analysis, the finding that bilateral mitral valve involvement represents a risk factor is noteworthy, and is in agreement with previous reports in both humans and dogs.8,36,38 It is also interesting to observe that only 4 of the 189 dogs for which this information was available presented isolated lesions of the posterior mitral valve leaflet. Similarly, the recent paper by Serres et al 21 also reported that only 4 of 114 dogs had a rupture of posterior mitral valve leaflet. This could reflect a different anatomy of the mitral valve leaflets, where the posterior leaflet appears to be smaller than the anterior.39 However, it is also possible that isolated posterior MVP is more difficult to identify using a right parastenal view because of its smaller extension, and this finding could represent an echocardiographic artifact for less severe cases.
A heart rate >140 bpm was associated with decreased survival time in univariate analysis. We did not include this data in multivariate analysis because of the low number of dogs for which this parameter was available. However, this is not an unexpected finding as increased HR is the consequence of increased sympathetic drive and neurohormonal activation, and thus reflects more severe disease.
Although this study was not designed to evaluate the effects of therapy on survival, it should be pointed that only 5.6% of the population received pimobendan. As pimobendan has become a more common treatment in Europe, it could be possible that our study could be used in the future as a historical control when more data on this drug are available.
Limitations of the Study
This study has some limitations. First, our population was not intentionally selected and consisted of a series of dogs visiting 4 animal clinics. This means that it may not represent the general canine population. In fact, the referral hospital population is more likely to represent a more advanced disease state compared with the general one, because some of these dogs were referred for the presence of clinical signs of heart disease. It is also possible that some dogs may have been referred when conventional treatment failed and this may have influenced the estimates of survival times. However, most of the dogs were referred for cardiological consultation at the time they were asymptomatic or mildly symptomatic; this is reflected by the fact that 82.2% of dogs were in class I or II ISACHC. Second, although the questionnaire was carefully designed, it is not possible to exclude that some deaths considered as cardiac related were because of other causes. However, there were small differences in the estimates of the studied risk factors between the 2 data sets used in univariate and multivariate analyses. Third, it is possible that therapy influenced outcome. We analyzed only differences between dogs treated with and without furosemide. However, because ongoing furosemide treatment was strictly a covariate with class of HF, we decided not to include it in multivariate analysis. Because this was a retrospective study and many combinations of drugs had been used, it was not possible to perform an appropriate analysis of effect of therapy on survival. Finally, the present study comprised a retrospective case series of dogs, which potentially makes it vulnerable for confounding factors. Although obvious confounding factors such as age and sex were controlled for in the present study, it is possible that the results could have been systematically influenced by other unknown factors.
This study shows that although a low degree CMVI can be considered a relatively benign condition in dogs, there are some echocardiographic variables that can identify dogs at higher risk of death. Specifically, syncope, a left-atrial aortic root ratio>1.7 and an Emax>1.2 m/s were shown to have the greatest potential to predict outcome in multivariate analysis. Univariate analysis also suggests that other clinical and echocardiographic variables, such as age, heart rate, and class of HF also have the potential to identify dogs at higher risk of death.
aUltrasonic Doppler Flow detector model 811-B, Parks Medical Electronics, Alhoa, OR USA
bMEGAS, ESAOTE Biomedica, Florence, Italy
cThe R Foundation for statistical computing version 2.3.0 (2006-04-24) ISBN 2-900051-07-0
dFalk T, Jonsson L, Olsen LH, Pedersen HD. Arteriosclerotic changes in myocardium, lung and kidney in dogs with chronic congestive heart failure and myxomatous mitral valve disease. J Vet Int Med 2005;19:932 (abstract)
eCrosara S, Perego E, Santilli RAS, et al. Holter monitoring in dogs with mitral regurgitation and different classes of heart failure. J Vet Int Med 2006;20:1535 (abstract)
This study has been granted by “Fondo Regione Piemonte Ricerca Scientifica Applicata 2004.”