This work was previously presented as an abstract at the 19th Congress of the European College of Veterinary Internal Medicine—Companion Animals, Porto, Portugal, September 2009.
Corresponding author: S. Eriksson, DVM, Specialist Degree in Small Animal Diseases, Minerva Foundation for Medical Research, Biomedicum 2U, Tukholmankatu 8, 00290 Helsinki, Finland; e-mail: firstname.lastname@example.org.
Background: Pulmonary edema and venous congestion are well-recognized signs of congestive heart failure (CHF) in advanced canine chronic mitral regurgitation (MR). However, little is known about pulmonary blood volume (PBV), blood pulmonary transit time (PTT), and the regulation of these.
Objectives: To measure and evaluate the relationships of PBV, forward stroke volume (FSV), and heart rate normalized blood pulmonary transit time (nPTT) in healthy dogs and dogs with MR.
Animals: Thirty-three Cavalier King Charles Spaniels; 11 healthy, 4 in modified New York Heart Association (NYHA) class I, 11 in class II, and 7 in CHF.
Methods: Heart rate normalized PTTs were measured by radionuclide angiocardiography. Left ventricular end diastolic and systolic diameter, left atrial/aortic root ratio, and FSV were measured by echocardiography. PBV and pulmonary blood volume index (PBVI) were calculated by established formulas.
Results: PBVI was 308 ± 56 (mean ± SD) mL/m2 for healthy dogs, 287 ± 51 mL/m2 in NYHA class I, 360 ± 66 mL/m2 in Class II, and 623 ± 232 mL/m2 in CHF (P= .0008). Heart rate normalized PTT, not FSV, was a predictor of PBV (r= 0.92 and 0.02, respectively).
Conclusions and Clinical Importance: Increased PBV, not decreased FSV, is the main cause of increased nPTT in MR. Increased nPTT can be used as an indicator of abnormal cardiopulmonary function in dogs with MR.
Mitral regurgitation (MR) caused by myxomatous mitral valve disease (MMVD) is the most common cause for congestive heart failure (CHF) and cardiac-related mortality in dogs. CHF can be seen as pulmonary edema and venous congestion on thoracic radiographs.1–3 This suggests that pulmonary blood volume (PBV) is increased. PBV can be estimated from forward stroke volume (FSV) and blood pulmonary transit time (PTT) normalized for heart rate (HR).4
The transit time of blood (t) through an organ is determined by the flow (F) and the blood volume (V) of the organ: t=V/F. PTT can be expressed standardized to HR by means of dividing PTT by mean R-R interval time (mRR). Then heart rate normalized PTT (nPTT) is the number of stroke volumes that the pulmonary vascular beds holds at a given moment, described by the relationship4–7:
where PBV is the pulmonary blood volume, HR is the heart rate, and FSV by definition refers to forward SV, not total SV. Cardiac output (CO) in turn equals FSV times HR. Therefore, PBV equals PTT times CO. Pulmonary blood volume index (PBVI) is calculated by dividing PBV by body surface area (BSA).
PBVI is 120–173 mL/m2 in anesthetized dogs.8 Normal values for PBVI in humans range from 200 to 400 mL/m2, with athletes having values in the upper range.5,7,9,10 PBV is higher and PTT longer in trained than untrained humans. This reflects an increased capacity of the pulmonary circulation, because PBV as percent of total plasma volume (PBV ratio) is equal.9 However, when PTT is normalized for HR (nPTT), changes between rest and exercise are minimal,6,11 making nPTT a useful index for cardiopulmonary function. Indeed, humans with heart failure have higher PBVs and nPTTs than healthy subjects.9 To date there are no studies on the dynamics of PBV in healthy dogs or in dogs with naturally occurring MMVD.
Dogs with enlarged hearts due to MR, but not yet in CHF, have increased nPTT and dogs with CHF had substantially increased nPTT.4 However, because of the study design, it was not possible to conclude whether this was because of an increase in PBV, decrease in FSV, or both. Therefore the aim of the present study was to evaluate the interrelationships of PBV, FSV, and nPTT in different clinical stages of MR as assessed by radiography, echocardiography, and clinical signs.
Materials and Methods
Fourteen male and 19 female client-owned Cavalier King Charles Spaniels with a mean age of 9.6 years (range 5–12.6) and mean body weight of 9.5 kg (5.2–13.6). The study population comprised dogs from the SVEP study12 as well as dogs recruited specifically for this study. All dogs were studied according to a separate protocol for this investigation. The protocol was approved by the ethical committees at the Universities in Helsinki (Finland) and Uppsala (Sweden). Informed consent was received from owners. A complete history was taken and thorough clinical examination was performed. Hemoglobin concentration, PCV, plasma sodium and potassium concentrations, urea, and creatine were measured. Only dogs with no signs of concurrent disease were included. Dogs were classified according to the (Scandinavian) modified New York Heart Association (NYHA) criteria as described previously.12 Briefly, dogs in class I had a mitral murmur without radiographic evidence of enlarged heart size, dogs in class II had increased global heart size and evident left atrial (LA) enlargement on radiography. The criteria for CHF were clinical signs of CHF and radiographic signs of pulmonary congestion and/or edema.2 Eleven dogs were healthy (N), 4 dogs had a heart murmur of mitral origin with no signs of heart enlargement (Class I), 11 had varying degrees of heart enlargement without clinical signs (Class II), and 7 had clinical and radiographic signs of CHF (CHF group). Three dogs were on enalapril prevention treatment, none of them in CHF. Only 1 dog had received treatment for CHF (digoxin, enalapril, and frusemide for 1 year). The dog relapsed into failure at time of evaluation.
Each dog was examined by first-pass radionuclide angiocardiography (FPRNA).4 Briefly, dogs were placed with their left side down on a gamma camera,a controlled by software.b An FPRNA was made with a bolus injection of 50–100 MBq of technetium 99 m-DTPA.c An ECG recording was started simultaneously. The time from the 1st notch in the right ventricular upslope, which indicated an ejection into the pulmonary trunk, to the 1st upslope of the LA curve, indicating activity had entered the LA, was measured as PTT. The mR-R during the peak-to-peak transit time was measured from the ECG strip. A single radiologist (P.L.) made all measurements and the mean of 2 consecutive FPRNA results was used. The coefficient of variation (CV%) for the consecutive measurements was 7.9%.
All dogs were examined by echocardiography by either A.E. (Helsinki) or K.H. (Uppsala) in conjunction with the FPRNA. Left ventricular end diastolic (LVIDd) and systolic (LVIDs) M-mode diameter was measured from the right parasternal view13 and values were normalized for body size.14 Dogs were grouped based on normalized left ventricular diastolic dimensions (nLVIDd) as normals (group N; nLVIDd ≤ 1.73, representing the 95% percentile upper limit), having moderate increase in left ventricular (LV) size (group 1; nLVIDd between 1.73 and 1.85), or having definite increase in LV size (group 2; nLVID over 1.85, the 97.5% percentile upper limit). LA and aortic root (Ao) diameters were measured from 2-dimensional images.15 Aortic flow was measured in a subcostal projection with continuous Doppler.16,17 The velocity time integral (VTI) and the HR for the preceding heart cycle were measured. Forward SV was estimated by the relation:
where CSA is the cross-sectional area of Ao measured in middiastole after closure of the aortic valves in a short-axis view according to the formula CSA =π×r2, and r (radius) is Ao diameter divided by 2. All measurements were done from at least 2 high quality images, the results averaged and FSV calculated from means. The intraobserver CV% for the consecutive measurements of CSA was 10.5 and for VTI 4.4. CO was calculated as HR × FSV. Forward stroke volume index (FSVI) and cardiac index were calculated as CO and FSV over BSA, respectively. End-diastolic volume (EDV) was assessed from M-mode recordings by the Teichholz method18 and FSV/EDV was used as a measure of effective ejection fraction.19
Thoracic radiographs were made in left lateral and ventrodorsal projections. The heart size was measured by a single radiologist (P.L.) in left lateral view by a slight modification of the vertebral heart scale (VHS) method,20 as described earlier.21 A VHS up to 10.9 was considered normal.21,22
Statistical calculations were performed by JMP software.d Normal distribution was evaluated by the Shapiro-Wilk test. Values were transformed (eg, logarithmic) as appropriate. Linear regression was used to assess associations between 2 variables and multivariate linear regression to find independent predictors of PBV. Multivariate regression was performed by entering all variables proceeded with a backward elimination of nonsignificant variables. A nonparametric test (Wilcoxon or Kruskall-Wallis) was used to evaluate differences between groups with unequal variances with Dunn's correction for several groups and analysis of variance (ANOVA) between groups with equal variances. Significant differences between ANOVA groups were evaluated with Tukey-Kramer HSD. PBV was indexed (PBVI) to BSA before comparing groups. The effects of treatment, sex, and age were investigated with both parametric and/or nonparametric tests, and also included in the multivariate analysis. As there was no effect on the outcomes, they were excluded from models.
Heart rate normalized PTT, not FSV, is the major predictor of PBV, with the regression formula PBV (mL) = 25.9 + 22.9 × nPTT (Fig 1). Though significant, the correlations between PBV and LVIDd, LVIDs and regurgitant volume, were only 0.51 (P= .0004), 0.45 (P= .0049), and 0.52 (P= .0015), respectively. The associations did not improve by indexing for BSA (PBVI) or normalizing for bodyweight (nLVIDd and nLVIDs).14 PBV increased with CO (r= 0.45, P= .005).
Multiple regression analysis using nPTT and FSV as predictor variables gave the best estimate for PBV (R2= 0.96, P < .0001). Heart rate normalized PTT alone explained 85% of the model. All other factors had no significant correlation.
Comparisons between Functional Heart Classes
Dogs in CHF had a significantly higher PBVI than dogs in other groups (Fig 2A). Likewise, only dogs with an nLVIDd exceeding the 97.5% normal upper limit (1.85) had increased PBVIs (Fig 2B). Forward SVI declined despite increase in total SVI (Table 1), but the decrease was not statistically significant. However, the effective ejection fraction, expressed as FSV/EDV (Fig 1), was decreased indicating an increased regurgitant fraction.
Table 1. Heart rate and echocardiographic variables for healthy dogs and dogs with mitral insufficiency grouped by the modified New York Heart Association scale.12
Normal (N= 11)
NYHA I (N= 4)
NYHA II (N= 11)
CHF (N= 7)
FSV/EDV, forward stroke volume over end-diastolic volume; CI, cardiac index (forward); HR, heart rate; FSVI, FSV index; Total SVI, SVI calculated from M-mode measurements (including the regurgitant volume); nPTT, heart-rate normalized blood pulmonary transit time; CHF, congestive heart failure; NYHA, New York Heart Association. Values are means ± standard deviation (SD). P-values refer to whole model. Symbols connect groups with significant difference , *P < .05 , #P < .01 , †P < .0001.
Results of the present study indicate that PBV increases in dogs with CHF attributable to MMVD. This is in accordance with previous reports in humans with MMVD5,7 and ischemic heart disease or left sided heart failure.7,10,23 PBVI in healthy humans5,7,9 was in the same range as we found in dogs. The good correlation between PBV and nPTT, with no correlation to FSV, is consistent with previous studies.5,9
Multivariate analysis revealed that nPTT describes 85% of the change in PBV, which therefore would be expected to increase with nPTT. nPTT is increased in class II dogs compared with dogs in CHF. nPTT is greater in dogs with enlarged hearts due to MR, but which are not yet in heart failure compared with healthy4 but, unlike the present study, dogs in class I were not included. We measured nonsignificantly lower nPTT values for dogs in class I (N= 4) than for dogs in class II (Table 1). According to power estimates, a minimum of 9 dogs in class I was required to demonstrate significance. Further studies are needed to refine the time course in nPTT increase during progression of disease. One would expect that treatment should affect nPTT and PBV but we found no effect in the multivariate analysis. A reasonable explanation is that only successful treatment should normalize values. Indeed, the only dog treated for CHF had been without clinical signs but relapsed into CHF (PBVI = 600 mL/m2 and nPTT = 10.9). It is therefore in concordance with the hypothesis that the increase in this dog was because treatment was ineffective.
Except for the correlation of PBV to LA (r= 0.73), the relationships to VHS and echocardiographic measurements of left heart size (r= 0.68, 0.51, and 0.45 for VHS, LVIDd, and LVIDs, respectively) as well as to regurgitant volume (r=0.53) were modest. Increase in heart size should logically precede increase in PBV. One reason for the limited correlation to VHS could be that VHS comprise total heart size, including the right ventricle. However, it is recently shown that the right ventricle is not dilated in dogs with mild-to-moderate MR without CHF.24 Though there is little doubt that LV size increases with progression of MR, another possible explanation for the modest correlation to LV measures is that increases in PBV are attenuated by multiple regulatory individual adaptive mechanisms.25,26 Further support for an individual compensatory capacitance is provided by the loglinear regression curve for FSV/EDV and PBV (Fig 1F). In fact, the regression curve could be split into one almost planar linear relationship for dogs not in failure, with an FSV/EDV over approximately 0.3, and a heterogeneous group consisting of dogs in CHF (FSV/EDV under 0.3) lying separately above this line. This split relationship indicates that dogs with MR are capable of maintaining a stable PBV despite an increased regurgitant volume until compensatory mechanisms fail resulting in a rise in PBV (Fig 2). However, we measured Ao diameter in a short-axis view in diastole and EDV by the Teichholz' method, which does not account for remodeling. The preferred methods, according to ASE recommendations,17 are in systole in long-axis and by the modified Simpson's rule, respectively. Further studies are therefore warranted to confirm a cut-off value and the potential clinical value of FSV/EDV as a predictor of increased PBV.
We found a negative trend in FSVI with progression of MR and a decrease in FSV/EDV despite an increase in total SVI (Table 1). Whether this reflects true myocardial dysfunction or is due to compensatory mechanisms remains unclear.27–29 It is well recognized that while the sympathetic nervous system is activated in heart failure, cardiac β1- and β2-receptors are downregulated.30 At the same time β3-receptors inhibiting cardiac contraction are upregulated.31 The net effect of noradrenergic stimulation in chronic MR may therefore be an increase in HR with preserved or decreased FSV. A decrease in FSV could in fact protect dogs from detrimental increase in PBV (Formula ) while the increased HR maintains CO (Table 1). Nevertheless, in spite of a wealth of data, we still do not know whether and to what extent β-receptor alterations are adaptive/protective or detrimental or both.32
Although increase in PBV is a prerequisite for pulmonary edema, a threshold for edema is not simple to estimate from PBV, because both are regulated by a myriad of mechanisms.25,26,33 Under normal conditions, the pulmonary arterial bed is a high-flow, low-pressure, low-resistance circuit. Pulmonary capillaries are capable of distending to accommodate increased flow and previously closed capillary beds are recruited in situations associated with increased CO, such as exercise. As a result, pulmonary vascular resistance (PVR) decreases and pulmonary arterial pressure (PAP) rises only minimally despite large increases in flow.26 The amount of pulmonary extravascular fluid is a function of the balance between capillary fluid filtration and reabsorption.34 Human HF patients have higher exercise induced increase in pulmonary vascular pressure with concurrent decrease in alveolar-capillary membrane conductance compared with normals.33 The increase in PBV in dogs with MR is initiated by the increase in LA pressure, which in turn increases pulmonary venous pressure (PVP). Therefore, increases in CO, PVR, or PVP will lead to increases in PAP.35,36 In chronic MR, LA compliance serves to buffer rises in LV and LA pressures.37 As MR progresses, the LA remains compliant and can receive the regurgitant blood volume, while increased lymphatic drainage of the pulmonary interstitium protects from pulmonary edema.38 Only in a very late stage of chronic MR does the stiffness and pressure volume curve shift toward reduced compliance,23,39–41 causing dramatic increase in PBV (Fig 2). As this is related to concurrent exercise sensitivity to rise in pulmonary blood pressures 33,34 and decrease in alveolar-capillary membrane conductance 23,41 dogs with MR become at risk of pulmonary edema.23
Several limitations of this study must be acknowledged. We used the modified NYHA classification12 to group dogs and the group sizes are small. Therefore possible existing differences in age, sex, treatment, and so on, between groups were not likely to be detected. Moreover, radiographic signs of congestion or edema were mandatory for CHF but this determination is subjective and affected by factors such as stage and depth of breathing, obesity, and exposure. As a consequence, some dogs with an elevation in PBV, with no radiographic signs of CHF, could erroneously have been classified in Class II. Moreover, because pulmonary edema is a dynamic state and fluid may resorb with rest,33 dogs in need of treatment (CHF) could in turn have been classified in Class II on the day of examination. This may explain part of the overlap between classes (Fig 2). Erroneous classification is one of the limitations of classifying study material into groups when measuring continuous variables. In this context it rather emphasizes the value of nPTT, with its close relation to PBV, as a useful indicator of cardiopulmonary hemodynamics.
Another source of variation is that PBV varies with work load,9 venous return and body position,23 and total blood volume,7,9 while PTT may vary with variations in regional blood flow in lung lobes caused by gravity and posture,42–44 the recruitment level of bronchial anastomoses affecting resistance, as well as equal lobe distribution of the tracer 5 and inaccuracy in timing and tightness of the tracer bolus.4,5 However, we made all FPRNAs in standard positions after resting animals and checking for unbroken boluses. Although we have no way of estimating the amount of recruitment of regional capillaries and anastomoses,11,43 it has been shown that physiologic changes have a negligible impact on nPTT.7,45 We therefore reason that changes in PBV as measured by nPTT under standard conditions are useful indicators of the cardiopulmonary function of dogs with MR.
In conclusion, we have shown that PBV increases in dogs with severe mitral regurgitation. Increased PBV, not decreased FSV, is the main cause of increased nPTT. Because increase in nPTT precedes signs of CHF, measurement of nPTT, either by FPRNA4 or contrast echocardiography,46 can add valuable information of cardiopulmonary function in dogs with MR both in the research and the clinical setting.7,47,48
aPicker SX 300, Picker International Inc, Cleveland, OH, or GE Maxicamera 400A, General Electric Medical Systems, Milwaukee, WI
bTechnescan DPTA (diethylenetriamine penta-acetate), Mallinckrodt Medical B.B., Petten, Holland
cNuclear Diagnostics, Hägersten, Sweden
dJMP, Version 7, SAS Institute Inc, Cary, NC
The study was supported by Finska Läkaresällskapet, Swedish Cultural Foundation in Finland, Agria Insurance Company, Swedish Medical Research Council (project no. 3392), and NorFA (Nordic Academy for Advanced Study).