Left ventricular remodeling is a complex process that is mediated in part by activation of the RAAS. In humans with mild to moderate1,2 and severe heart failure,3,4 blockade of the mineralocorticoid (aldosterone) receptor has been shown to improve survival times. This benefit is suggested to be attributable to a decrease in aldosterone-mediated interstitial fibrosis.5,6 In an experimental model of atrial fibrillation in dogs, treatment with spironolactone, which is a mineralocorticoid receptor antagonist, inhibited the development of atrial dilatation and fibrosis.7
In dogs with MMVD, survival time (from diagnosis to death or euthanasia due to cardiac disease) is negatively correlated with the myocardial fibrosis score at postmortem examination,8 and in dogs with heart failure, spironolactone administration appears to reduce the risk of death.9 Plasma aldosterone concentration is reportedly higher in the 6 months prior to cardiac decompensation than at the time of decompensation.10 Measurement of UAC has been suggested to estimate 24-hour aldosterone production and therefore may be more reflective of RAAS activation than measurement of the more labile plasma aldosterone concentration.11
Circulating PIIINP is a marker of collagen type III turnover and hence ECM turnover.12 In humans, higher circulating PIIINP concentrations have been associated with more advanced cardiac disease and are considered a marker of myocardial fibrosis.13,14 Mineralocorticoid receptor blockade is most beneficial in humans with circulating PIIINP concentrations at initiation of treatment that are higher than the median concentration.3 In experimental models of MMVD in dogs, left ventricular dilatation is associated with loss of interstitial collagen15; however, the degree to which such models simulate the naturally occurring disease is unclear.
The purpose of the study reported here was to investigate relationships among UAC, serum PIIINP concentrations, and clinical variables, including echocardiographic measurements, in healthy dogs and dogs with naturally occurring MMVD. Our hypothesis was that UAC and PIIINP concentrations would be associated with changes in left ventricular echocardiographic size due to left ventricular remodeling in the dogs with MMVD.
Materials and Methods
Animals—All dogs enrolled in the study were privately owned, and owner consent was obtained prior to enrollment. For inclusion in the control group, dogs were required to be healthy on the basis of their history and results of physical examination. Relevant mitral regurgitation was deemed to be absent on the basis of unremarkable auscultatory findings.
For inclusion in the MMVD group, dogs were required to have an echocardiographic diagnosis of MMVD, as made on the basis of characteristic abnormalities of the valve leaflets (thickening or prolapse) and evidence of regurgitant flow across the valve detected with Doppler techniques. Dogs with any other cardiac disease or major organ-related or systemic disease were excluded; however, dogs receiving medication for heart failure were included. The study protocol was approved by the Royal Veterinary College ethics and welfare committee.
Sample collection—Blood samples were collected from all dogs via jugular venipuncture into serum gel tubes. Owners collected urine samples from their dogs during natural voiding into sterile urine specimen containers on the morning of the examination (n = 161) or the previous evening (25). Urine samples collected the previous evening were stored at 2° to 4°C until centrifugation. Blood samples were stored at 2° to 4°C for up to 6 hours; urine samples were stored at 2° to 4°C for up to 18 hours. All samples were centrifuged for 15 minutes at 1,000 × g, and the serum and urine supernatants were stored at −80°C for subsequent analysis.
Laboratory analyses—Serum PIIINP concentration was measured with a commercially available RIA based on highly purified human PIIINPa that had been validated for use with canine serum and bronchoalveolar lavage fluid.16 Urinary aldosterone concentration was measured following mild acid hydrolysis and extraction into ethyl acetate. Each urine sample (0.25 mL) was incubated with 25 μL of 3.2N hydrochloric acid in the dark for 24 hours at room temperature (mean temperature, 22°C). Ethyl acetate (2.5 mL) was then added to each sample followed by mixing at 30 revolutions/min for 60 minutes in a rotary mixer at room temperature. The samples were centrifuged at 600 × g for 5 minutes at 4°C, and 0.5 mL of the solvent phase was evaporated to dryness under a gentle stream of nitrogen at 37°C. The residues were reconstituted in assay zero standard (0.5 mL) and aldosterone concentration measured via a commercially available RIAb validated for use with canine urine.17 All in-house assays were performed precisely in accordance with the manufacturers' instructions. Urinary creatinine concentrations were measured at a commercial laboratory.c
Samples from multiple dogs were pooled to provide serum with low, medium, and high PIIINP concentrations and urine with low, medium, and high aldosterone concentrations. Intra-assay variability was determined by measuring the sample pools 6 times within 1 RIA. Interassay variability was determined by measuring the sample pools in 6 RIAs. For measurements of aldosterone concentration, the extraction procedure was performed separately for each sample replicate for both intra- and interassay variability determinations. The results are reported as coefficients of variation, calculated as the SD as a percentage of the mean.18 Urinary aldosterone and creatinine values were used to calculate the UAC in grams per mole.
Echocardiography—All echocardiographic examinations were performed with dogs unsedated by a board-certified cardiologist. Dogs were positioned in right lateral recumbency and then left lateral recumbency on an ultrasonography examination table. The echocardiographic examination was performed with an ultrasonographic unitd equipped with 2- to 4-MHz–and 3- to 7-MHz–phased-array transducers and ECG monitoring. Standard imaging planes were digitally stored. Assessment of mitral valve structures was performed from the right parasternal long-axis view and the left apical 4-chamber view. Measurements of left atrial and aortic root diameters were made from the 2-D right parasternal short-axis view at the level of the aortic valve leaflets.
Measurements were made at the onset of the P wave and were used to calculate the LA:Ao.19 The LVEDD, left ventricular end systolic diameter, and LVFWd were measured from the right parasternal short-axis M-mode at the level of the papillary muscles20 and used to calculate the LVEDD:LVFWd ratio. The LVEDDN was determined by dividing by body weight (kg)0.294. The LVESDN was determined by dividing by body weight (kg)0.315.21
Dogs with MMVD were classified according to their echocardiographic measurements and clinical status, as recommended in the ACVIM consensus statement.22 Categories assigned were no cardiomegaly (LVEDDN ≤ 1.8 cm/[kg0.294] and LA:Ao ≤ 1.5; B1), cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) but no current or previous clinical signs of congestive heart failure (B2), and cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) plus current or previous clinical signs of congestive heart failure requiring medical management (C).
Data for dogs in class B1 that received treatment for congestive heart failure were excluded from between-group statistical comparisons because we believed congestive heart failure in the absence of cardiomegaly to be unlikely and any confounding effects of treatment on UAC and serum PIIINP concentration would be inappropriately increased by including these dogs in the analysis. These dogs were not excluded from linear regression analyses because the confounding effects of treatment were reduced by their inclusion as covariates in the multivariable models when appropriate.
ECG—All dogs with MMVD underwent ECG. Heart rate was measured from a 60-second ECG trace.
Measurement of MMVD progression—All procedures (blood and urine sample collection and assays, echocardiography, and ECG) were repeated for dogs with MMVD at approximately 6-month intervals unless the dogs died or were lost to follow-up. Each dog was examined on a maximum of 3 occasions, and to assess echocardiographic progression, a baseline visit was defined for comparison of findings with those of the previous or subsequent visit. For dogs examined 3 times, the second visit was designated as the baseline visit. For dogs examined only twice, either the first or second visit was randomly designated the baseline visit on the basis of a coin toss.
When the first visit was designated the baseline visit, subsequent percentage changes in LVEDDN and LVESDN per month were calculated. When the second visit was designated the baseline visit, prior percentage changes were calculated.
Prior percentage changes in LVEDDN and LVESDN per month were calculated by comparing the measurements from the baseline and previous examinations as follows:
where Vn is the measurement at visit n and T is the time between visits in months. Similarly, subsequent percentage changes in LVEDDN and LVESDN were calculated by comparing the measurements from the baseline and subsequent examinations as follows:
Statistical analysis—Statistical analyses were performed with commercially available software.e Values of P < 0.05 was considered significant. Data were assessed graphically for normality. Values for UAC, body weight, and LA:Ao were not normally distributed and were consequently logarithmically transformed prior to further analysis. The logarithmically transformed variables were normally distributed.
Comparisons of continuous variables were made for measurements obtained from healthy control dogs and dogs in the MMVD group at the baseline visit via the independent sample t test or 1-way ANOVA with Tukey post hoc comparisons, as appropriate. Proportions were compared via the χ2 test. Correlations were assessed via calculation of Pearson (r) correlation coefficients. Univariate linear regression analyses were performed to assess associations in dogs with MMVD among UAC and serum PIIINP concentration and clinical characteristics (age, breed [CKCS vs non-CKCS], body weight, heart rate obtained from the ECG, echocardiographic measurements [LA:Ao, LVEDD:LVFWd ratio, LVEDDN, LVESDN, prior percentage change in LVEDDN and LVESDN per month, and subsequent percentage change in LVEDDN and LVESDN per month] and treatment with ACE inhibitors [yes vs no], pimobendan [yes vs no], or diuretics [yes vs no]).
The CKCS breed was chosen as the comparator breed because this was the most common breed in the study. Cavalier King Charles Spaniels are particularly prone to MMVD, and another study23 revealed differences in the natural history of the disease in this breed. In the univariate analyses, the assumption of linearity was assessed graphically by plotting the independent variables against the dependent variable (PIIINP concentration or UAC). The distributions of residuals were graphically tested for normality by plotting the predicted values against the residuals.
Variables associated with PIIINP concentration and UAC with values of P < 0.20 in the univariate analysis were entered into multivariable linear regression analyses. Separate models were constructed for the outcomes UAC and serum PIIINP concentration. When treatment with furosemide (yes vs no) was not significant in the final model, the modeling process for UAC was repeated, forcing treatment with furosemide (yes vs no) into the final models. Multivariable models for UAC and serum PIIINP concentration were repeated, excluding dogs receiving medication for cardiac disease.
Not all dogs had been examined on > 1 occasion, so data for prior percentage change in LVEDDN and LVESDN per month and subsequent percentage change in LVEDDN and LVESDN per month were unavailable for dogs with only 1 evaluation. For a dog to have data for both prior percentage change in LVEDDN and LVESDN per month and subsequent percentage change in LVEDDN and LVESDN per month, it had to have been examined on at least 3 occasions. Therefore, separate multivariable models including and excluding these variables were constructed. In the multivariable regression models, analyses were performed in a backward stepwise manner. All variables were initially included, and the variable with the highest P value was removed until all remaining variables had a value of P < 0.05, which was considered significant.
The distributions of residuals in the multivariable analyses were graphically tested for normality by plotting the predicted values against the residuals. First-order interactions were assessed between variables that were significant in the multivariable analyses by including the product term of each pair of variables in the model. When a product term was found to be significant, it was retained in the final multivariable model.
Results
Animals—The control group consisted of 24 healthy dogs (14 neutered females, 3 sexually intact males, and 7 neutered males), with ages ranging from 3 to 13 years (mean ± SD, 8.9 ± 3.0 years) and body weights ranging from 4.2 to 26.9 kg (median, 11.2 kg [interquartile range, 7.1 to 13.8 kg]). There were 5 (21%) CKCSs and various other purebred and crossbred dogs. None of the dogs was treated with ACE inhibitors.
The MMVD group comprised 162 dogs (10 sexually intact females, 59 spayed females, 25 sexually intact males, and 68 neutered males), with ages ranging from 3 to 18 years (mean ± SD, 10.1 ± 2.9 years) and body weights ranging from 1.9 to 48.7 kg (median, 10.6 kg [interquartile range, 8.4 to 14.2 kg]). There were 68 (42%) CKCSs and various other purebred and crossbred dogs. Thirty-nine (24%) were undergoing treatment with ACE inhibitors, 24 (15%) were receiving diuretics, and 24 (15%) were receiving pimobendan. Seventy-two dogs with MMVD were evaluated 3 times during the study period, and 44 dogs were evaluated twice (13 had the first visit designated as the baseline visit, and 31 had the second visit designated as such). The remaining 46 dogs were examined on only 1 occasion.
Cardiopulmonary measurements at the baseline evaluation of dogs with MMVD were summarized (Table 1). At the baseline visit, there was no significant (P = 0.804) difference in body weight between the control and MMVD groups; control dogs were younger but not significantly (P = 0.059) so. Distributions of male and female dogs were similar (P = 0.220) between the control and MMVD groups.
Baseline median (range) cardiopulmonary values in privately owned dogs with MMVD.
| Variable | No. of dogs | Median (range) |
|---|---|---|
| LA:Ao | 159 | 1.29 (0.83 to 2.75) |
| LVEDD:LVFWd ratio | 158 | 3.92 (2.17 to 6.62) |
| LVEDDN (cm/[kg0.294]) | 158 | 1.71 (1.14 to 2.58) |
| Prior change in LVEDDN per month (%) | 85 | 0.47 (−2.30 to 4.45) |
| Subsequent change in LVEDDN in per month (%) | 103 | 0.48 (−3.05 to 4.73) |
| LVESDN (cm/[kg0.315]) | 158 | 1.01 (0.63 to 1.64) |
| Prior change in LVESDN per month (%) | 85 | 0.22 (−6.10 to 3.98) |
| Subsequent change in LVESDN per month (%) | 103 | 0.53 (−2.49 to 4.62) |
| HR measured from the ECG (beats/min) | 151 | 128 (61 to 204) |
Assay variation—The intra-assay coefficients of variation for low (5.9 μg/L), medium (10.2 μg/L), and high (15.7 μg/L) serum concentrations of PIIINP were 2.3%, 1.8%, and 1.5%, respectively. The interassay coefficients of variation for low, medium, and high serum concentrations of PIIINP were 8.2%, 3.3%, and 8.5%, respectively. The intra-assay coefficients of variation for low (168.3 pg/mL), medium (992.0 pg/mL), and high (7,378.4 pg/mL) urine concentrations of aldosterone were 1.7%, 3.4%, and 2.4%, respectively. The interassay coefficients of variation for low, medium, and high urine concentrations of aldosterone were 7.4%, 2.1%, and 4.6%, respectively. The lower limit of quantification was 14.2 pg/mL.
Groupwise comparisons—When dogs in the B1 group receiving treatment for congestive heart failure (n = 24) were excluded from the analyses, for the remaining 162 dogs (148 dogs in the MMVD group plus 24 healthy control dogs), an overall significant (P = 0.004) difference in serum PIIINP concentration was evident among groups (control, B1, B2, and C). Dogs in the B1 category (n = 75) had a higher mean concentration of serum PIIINP (12.2 ± 4.5 μg/L) than those in the B2 (33; P = 0.009) and C (30; P = 0.049) categories (9.7 ± 2.7 μg/L and 10.1 ± 3.3 μg/L, respectively; Figure 1). No significant differences were detected in serum PIIINP concentrations between any other group pairs. There were no significant (P = 0.820) differences in UAC among control, B1, B2, and C groups.
Box-and-whisker plots of serum PIIINP concentration (A) and UAC (B) in 162 dogs with MMVD and 24 healthy control dogs of comparable age and body weight. Dogs with MMVD were categorized by their echocardiographic measurements and clinical status. These categories were no cardiomegaly (LVEDDN ≤ 1.8 cm/[kg0.294] and LA:Ao ≤ 1.5; B1 group), cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) but no current or previous clinical signs of congestive heart failure (B2 group), and cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) plus current or previous clinical signs of congestive heart failure requiring medical management (C group). Groups were compared via 1-way ANOVA with Tukey post hoc comparisons. Top and bottom boundaries of each box represent the interquartile range, the horizontal middle line represents the median, whiskers represent the 2.5th and 97.5th percentiles, and circles represent outlying values. *Indicated comparisons are significant (P < 0.05).
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Box-and-whisker plots of serum PIIINP concentration (A) and UAC (B) in 162 dogs with MMVD and 24 healthy control dogs of comparable age and body weight. Dogs with MMVD were categorized by their echocardiographic measurements and clinical status. These categories were no cardiomegaly (LVEDDN ≤ 1.8 cm/[kg0.294] and LA:Ao ≤ 1.5; B1 group), cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) but no current or previous clinical signs of congestive heart failure (B2 group), and cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) plus current or previous clinical signs of congestive heart failure requiring medical management (C group). Groups were compared via 1-way ANOVA with Tukey post hoc comparisons. Top and bottom boundaries of each box represent the interquartile range, the horizontal middle line represents the median, whiskers represent the 2.5th and 97.5th percentiles, and circles represent outlying values. *Indicated comparisons are significant (P < 0.05).
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Box-and-whisker plots of serum PIIINP concentration (A) and UAC (B) in 162 dogs with MMVD and 24 healthy control dogs of comparable age and body weight. Dogs with MMVD were categorized by their echocardiographic measurements and clinical status. These categories were no cardiomegaly (LVEDDN ≤ 1.8 cm/[kg0.294] and LA:Ao ≤ 1.5; B1 group), cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) but no current or previous clinical signs of congestive heart failure (B2 group), and cardiomegaly (LVEDDN > 1.8 cm/[kg0.294] or LA:Ao > 1.5) plus current or previous clinical signs of congestive heart failure requiring medical management (C group). Groups were compared via 1-way ANOVA with Tukey post hoc comparisons. Top and bottom boundaries of each box represent the interquartile range, the horizontal middle line represents the median, whiskers represent the 2.5th and 97.5th percentiles, and circles represent outlying values. *Indicated comparisons are significant (P < 0.05).
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Dogs in the B1 group receiving treatment for congestive heart failure were included in subsequent analyses. There were no significant differences in UAC (P = 0.905) or serum PIIINP concentration (P = 0.163) between sexes. No correlation was detected between UAC and serum PIIINP concentration when all dogs (P = 0.391) or only dogs with MMVD (P = 0.483) were included in the analysis. In the healthy control group, there was no evidence for a difference in UAC (P = 0.211) or serum PIIINP concentration (P = 0.801) between CKCSs and other breeds.
Linear regression analyses—In the univariate analyses in dogs with MMVD, age, breed (CKCS vs non-CKCS), LA:Ao, LVEDDN, prior percentage change in LVEDDN per month, LVESDN, prior percentage change in LVESDN per month, heart rate, treatment with ACE inhibitors (yes vs no), and ACVIM heart disease class were correlated with PIIINP concentration (P < 0.20) and were submitted for consideration in the multivariable models (Table 2; Figure 2). The LVEDDN and LVESDN were strongly correlated (r = 0.758; P < 0.001), so separate models were constructed excluding each of these variables in turn. The model with the highest adjusted R2 value was reported in each situation. In the multivariable model excluding prior percentage change data, age (P = 0.007) and LVEDDN (P = 0.003) were negatively associated with PIIINP concentration and breed (CKCS vs non-CKCS) was positively associated with PIIINP concentration (Table 3; P = 0.007; adjusted R2 = 0.159). No significant interactions were identified among pairs of variables. Neither prior percentage change in LVEDDN nor prior percentage change in LVESDN was significant in a multivariable model including these variables (adjusted R2 = 0.267).
Results of univariate linear regression analysis of the relationships between serum PIIINP concentration and other variables in dogs with MMVD.
| Variable | No. of dogs | P value | β | 95% CI for β |
|---|---|---|---|---|
| Age | 162 | < 0.001 | −0.39 | −0.60 to −0.19 |
| Logarithm of body weight | 162 | 0.207 | 1.78 | −1.00 to 4.55 |
| CKCS (yes vs no) | 162 | 0.015 | 1.53 | 0.31 to 2.76 |
| Logarithm of LA:Ao | 159 | 0.024 | −7.56 | −14.10 to −1.02 |
| LVEDD:LVFWd ratio | 158 | 0.533 | −0.22 | −0.92 to 0.48 |
| LVEDDN | 158 | 0.031 | −2.33 | −4.45 to −0.21 |
| Prior change in LVEDDN per month (%) | 85 | 0.185 | −0.75 | −1.13 to 0.22 |
| Subsequent change in LVEDDN per month (%) | 103 | 0.723 | 0.10 | −0.44 to 0.63 |
| LVESDN | 158 | 0.092 | −3.07 | −6.65 to 0.51 |
| Prior change in LVESDN per month (%) | 85 | 0.070 | −0.34 | −0.71 to 0.03 |
| Subsequent change in LVESDN per month (%) | 103 | 0.639 | −0.11 | −0.55 to 0.34 |
| Heart rate measured from the ECG | 151 | 0.036 | 0.02 | 0.002 to 0.047 |
| Treatment with ACE inhibitors (yes vs no) | 162 | 0.075 | −1.30 | −2.72 to 0.13 |
| Treatment with diuretics (yes vs no) | 162 | 0.422 | −0.71 | −2.44 to 1.03 |
| Treatment with pimobendan (yes vs no) | 162 | 0.500 | −0.59 | −2.32 to 1.14 |
| ACVIM class | 162 | 0.018 | −1.07 | −1.95 to −0.19 |
β = Regression coefficient, representing the slope of the relationship between the variables. Values of P < 0.05 were considered significant.
Results of multivariable linear regression analysis of the relationships between serum PIIINP concentration and other variables in dogs (n = 158) with MMVD, by whether data on prior and subsequent percentage change of LVEDDN and LVESDN were included.
| Variable | P value | β | 95% CI for β |
|---|---|---|---|
| Percentage change data excluded | |||
| Age | 0.007 | −0.31 | −0.53 to −0.09 |
| LVEDDN | 0.003 | −3.27 | −5.43 to −1.12 |
| CKCS vs non-CKCS | 0.007 | 1.89 | 0.54 to 3.25 |
| Percentage change data included | |||
| Age | 0.006 | −0.42 | −0.72 to −0.13 |
| LVESDN | < 0.001 | −10.85 | −16.55 to −5.16 |
| CKCS vs non-CKCS | 0.009 | 2.20 | 0.56 to 3.84 |
The initial number of observations for the model including percentage change data was 85. However, once prior percentage changes in LVEDDN and LVESDN were eliminated from the backward stepwise analysis, the software program automatically included all data for the remaining variables. The number of observations for the final model was 158.
See Table 2 for remainder of key.
Box-and-whisker plots of serum PIIINP concentration in the dogs of the MMVD group that were CKCSs (n = 67) or any other breed (95). See Figure 1 for key.
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Box-and-whisker plots of serum PIIINP concentration in the dogs of the MMVD group that were CKCSs (n = 67) or any other breed (95). See Figure 1 for key.
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Box-and-whisker plots of serum PIIINP concentration in the dogs of the MMVD group that were CKCSs (n = 67) or any other breed (95). See Figure 1 for key.
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
The multivariable models were reconstructed to include data on only the 111 dogs remaining after those receiving medication for cardiac disease were excluded (Table 4). In the multivariable model excluding percentage change data, age (P = 0.015) and LVEDDN (P = 0.039) were negatively associated with PIIINP concentration and heart rate (P = 0.021) was positively associated with PIIINP concentration (adjusted R2 = 0.161). No significant interactions were identified. In the multivariable model including percentage change data, age (P = 0.006) and prior percentage change in LVEDDN per month (P = 0.012) were negatively associated with PIIINP concentration (adjusted R2 = 0.221). No significant interaction was identified.
Results of multivariable linear regression analysis of the relationships between serum PIIINP concentration and other variables in dogs with MMVD not receiving medications for cardiac disease, by whether prior and subsequent percentage change of LVEDDN and LVESDN data were included.
| Variable | P value | β | 95% CI for β |
|---|---|---|---|
| Percentage change data excluded (n = 99) | |||
| Age | 0.015 | −0.340 | −0.614 to −0.066 |
| LVEDDN | 0.039 | −3.126 | −6.083 to −0.169 |
| Heart rate measured from the ECG | 0.021 | 0.032 | 0.005 to 0.060 |
| Percentage change data included (n = 57) | |||
| Age | 0.006 | −0.487 | −0.826 to −0.148 |
| Prior percentage change in LVEDDN per month | 0.012 | −1.108 | −1.967 to −0.250 |
See Table 2 for key.
In the univariate analyses involving dogs with MMVD, age, breed (CKCS vs non-CKCS), LVEDD:LVFWd ratio, LVEDDN, prior percentage change in LVEDDN per month, LVESDN, subsequent percentage change in LVESDN per month, treatment with diuretics (yes vs no), and treatment with pimobendan (yes vs no) were correlated with the logarithm of UAC at a value of P < 0.20 and were taken further for consideration in multivariable modeling (Table 5; Figure 3). The LVEDDN and LVESDN (r = 0.758; P < 0.001) and LVEDDN and LVEDD:LVFWd ratio (r = 0.807; P < 0.001) were strongly correlated, so separate models were constructed excluding each of these variables in turn. The model with the highest adjusted R2 value is reported in each situation.
Results of univariate linear regression analysis of the relationships between the logarithm of UAC and other variables in dogs with MMVD.
| Variable | No. of dogs | P value | β | 95% CI for β |
|---|---|---|---|---|
| Age | 162 | < 0.001 | −0.04 | −0.06 to −0.02 |
| Logarithm of body weight | 162 | 0.402 | −0.11 | −0.36 to 0.14 |
| CKCS (yes vs no) | 162 | < 0.001 | 0.28 | 0.18 to 0.39 |
| Logarithm of LA:Ao | 159 | 0.493 | −0.21 | −0.80 to 0.39 |
| LVEDD:LVFWd ratio | 158 | 0.007 | 0.08 | 0.02 to 0.14 |
| LVEDDN | 158 | 0.188 | 0.13 | −0.06 to 0.32 |
| Prior change in LVEDDN per month (%) | 85 | 0.032 | 0.07 | 0.01 to 0.13 |
| Subsequent change in LVEDDN per month (%) | 103 | 0.330 | 0.03 | −0.03 to 0.08 |
| LVESDN | 158 | 0.157 | 0.07 | −0.09 to 0.55 |
| Prior change in LVESDN per month (%) | 85 | 0.770 | 0.01 | −0.03 to 0.04 |
| Subsequent change in LVESDN per month (%) | 103 | 0.015 | 0.06 | 0.01 to 0.10 |
| Heart rate measured from the ECG | 151 | 0.469 | 0.00 | 0.00 to 0.00 |
| Treatment with ACE inhibitors (yes vs no) | 162 | 0.283 | −0.07 | −0.20 to 0.06 |
| Treatment with diuretics (yes vs no) | 162 | 0.081 | 0.14 | −0.02 to 0.29 |
| Treatment with pimobendan (yes vs no) | 162 | 0.067 | 0.14 | −0.01 to 0.30 |
| ACVIM class | 162 | 0.218 | 0.05 | −0.03 to 0.13 |
See Table 2 for key.
Box-and-whisker plots of UAC in the dogs of the MMVD group that were CKCSs (n = 67) or any other breed (95).
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Box-and-whisker plots of UAC in the dogs of the MMVD group that were CKCSs (n = 67) or any other breed (95).
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
Box-and-whisker plots of UAC in the dogs of the MMVD group that were CKCSs (n = 67) or any other breed (95).
Citation: American Journal of Veterinary Research 73, 11; 10.2460/ajvr.73.11.1765
In the multivariable model excluding prior percentage change data, breed (CKCS vs non-CKCS; P < 0.001) and treatment with diuretics (yes vs no; P = 0.005) were positively associated with UAC and age was negatively associated with UAC (P = 0.002; Table 6) (adjusted R2 = 0.229). No significant interactions were identified. In the multivariable model including prior and subsequent percentage change data, breed (CKCS vs non-CKCS; P = 0.018), prior percentage change in LVEDDN per month (P = 0.036), and subsequent percentage change in LVESDN per month (P = 0.020) were positively associated with UAC, and age was negatively associated with UAC (P = 0.033; R2 = 0.263). No significant interactions were identified.
Results of multivariable linear regression analysis of the relationships between the logarithm of UAC and other variables in all dogs with MMVD, by whether prior and subsequent percentage change of LVEDDN and LVESDN data were included.
| Variable | P value | β | 95% CI for β |
|---|---|---|---|
| Percentage change data excluded (n = 162) | |||
| Age | 0.002 | −0.03 | −0.05 to −0.01 |
| CKCS vs non-CKCS | < 0.001 | 0.21 | 0.11 to 0.32 |
| Treatment with diuretics (yes vs no) | 0.005 | 0.20 | 0.06 to 0.34 |
| Percentage change data included (n = 72) | |||
| Age | 0.033 | −0.034 | −0.065 to −0.003 |
| CKCS vs non-CKCS | 0.018 | 0.203 | 0.035 to 0.370 |
| Prior percentage change in LVEDDN per month | 0.036 | 0.068 | 0.004 to 0.131 |
| Subsequent percentage change in LVESDN per month | 0.020 | 0.061 | 0.010 to 0.112 |
See Table 2 for key.
The multivariable model for UAC including prior and subsequent percentage change data was repeated, forcing treatment with furosemide (yes vs no) into the final model (Table 7). Breed (CKCS vs non-CKCS; P = 0.019) and subsequent percentage change in LVESDN per month (P = 0.023) were positively associated with UAC, and age was negatively associated with UAC (P = 0.028). The adjusted R2 for the model was 0.257.
Results of multivariable linear regression analysis of the relationships between the logarithm of UAC and other variables in all dogs with MMVD (data on prior and subsequent percentage change of LVEDDN and LVESDN included), with the variable furosemide treatment forced into the model.
| Variable | P value | β | 95% CI for β |
|---|---|---|---|
| Age | 0.028 | −0.035 | −0.066 to −0.004 |
| CKCS vs non-CKCS | 0.019 | 0.202 | 0.034 to 0.370 |
| Prior percentage change in LVEDDN per month | 0.063 | 0.062 | −0.003 to 0.128 |
| Subsequent percentage change in LVESDN per month | 0.023 | 0.059 | 0.008 to 0.111 |
| Treatment with furosemide (yes vs no) | 0.501 | 0.078 | −0.153 to 0.309 |
See Table 2 for key.
The multivariable models were repeated for 111 dogs remaining after excluding dogs receiving medication for cardiac disease (Table 8). In the multivariable model excluding percentage change data, breed (CKCS vs non-CKCS) was positively associated with UAC (P < 0.001; adjusted R2 = 0.201). In the multivariable model including percentage change data, breed (CKCS vs non-CKCS; P < 0.001) and subsequent percentage change in LVESDN per month (P = 0.024) were positively associated with UAC (adjusted R2 = 0.288). No significant interaction was identified.
Results of multivariable linear regression analysis of the relationships between the logarithm of UAC and other variables in dogs with MMVD not receiving medications for cardiac disease, by whether prior and subsequent percentage change of LVEDDN and LVESDN data were included.
| Multivariable model | P value | β | 95% CI for β |
|---|---|---|---|
| Percentage change data excluded (n = 111) | |||
| CKCS vs non-CKCS | < 0.001 | 0.313 | 0.197–0.429 |
| Percentage change data included (n = 69) | |||
| CKCS vs non-CKCS | < 0.001 | 0.352 | 0.196–0.508 |
| Subsequent percentage change in LVESDN per month | 0.024 | 0.057 | 0.008–0.107 |
See Table 2 for key.
Discussion
The present study was the first to demonstrate an independent relationship between UAC and prior percentage change in LVEDDN and subsequent percentage change in LVESDN over time in dogs with MMVD. It also revealed a negative relationship between serum PIIINP concentration and echocardiographic measurements of left ventricular diameter.
The UAC was found to be negatively associated with age and positively associated with diuretic treatment, as expected.24 The relationships in the multivariable regression analyses between UAC and prior percentage change in LVEDDN per month and subsequent percentage change in LVESDN per month, but not with absolute measurements of LVEDDN or LVESDN, suggested that aldosterone concentration may increase around times of active ventricular remodeling rather than continuously throughout the course of the disease. This potentially intermittent nature of aldosterone release may explain the lack of correlation between UAC and serum PIIINP concentration and the lack of an increase in aldosterone with ACVIM heart disease class. It may also explain the apparently conflicting findings in studies25–27 in which plasma aldosterone concentration was measured in dogs with MMVD.
Aldosterone is a recognized mediator of ventricular fibrosis.28 It is interesting to speculate whether a causal relationship exists between the high concentration of circulating aldosterone suggested by increasing UAC and the deteriorating ventricular systolic function suggested by an increase in LVESDN over time.29
As serum PIIINP concentration decreased in dogs with MMVD, left ventricular diastolic diameter increased. This negative association was independent of the reported age-related decrease in PIIINP concentration.16 Given that PIIINP concentration reflects collagen type III turnover, a decrease in turnover might be consistent with the decrease in myocardial collagen found in dogs with experimentally induced mitral regurgitation.15 However, given that PIIINP concentration reflects both production and degradation of collagen type III, any such relationship remains speculative.
We expected to find a relationship between increasing concentrations of fibrosis markers and increasing severity of MMVD, but this was not reflected in serum PIIINP concentrations. The lack of effect on PIIINP concentration may have been attributable to the balance in MMVD favoring a decrease in ECM turnover, despite an increase in fibrosis seen in some studies. Alternatively, turnover of collagen type III might not be an important component of fibrosis development in MMVD or fibrosis might not have developed in the study dogs. Although PIIINP concentration is strongly correlated with fibrosis in humans,14 whether such a relationship exists in dogs is not known. Additional studies are required to investigate the relationship between serum PIIINP concentration and fibrosis in dogs.
In dogs with MMVD but not healthy control dogs, PIIINP concentration and UAC were higher in CKCSs than non-CKCS breeds, despite statistically controlling for the influence of other variables known to affect concentrations of these analytes, such as age, heart size measurements, and treatment. This finding suggested that there might be breed-specific differences in collagen type III turnover and RAAS activation and hence differences in the pathophysiologic response to mitral regurgitation. No significant difference in UAC or serum PIIINP concentration was found between healthy control dogs and dogs with MMVD, which suggests that neither of these markers is suitable for use as a diagnostic test in dogs with clinical MMVD.
The UAC increases with furosemide administration in dogs24 and may decrease with ACE inhibitor administration.30 However, in the medium to long term, the effects of treatment with ACE inhibitors on UAC are diminished by the potential for angiotensin I to be converted to angiotensin II by enzymes other than ACE, such as chymase,31 a phenomenon known as aldosterone breakthrough.32 Short-term administration of pimobendan has no effect on UAC in healthy dogs,24 although the effects of long-term administration or administration to dogs with naturally occurring heart disease remain unclear. Serum C-terminal propeptide of procollagen type I concentrations do not change on administration of furosemide in human patients.33 In hypertensive humans treated with imidapril, serum PIIINP concentration decreases over time with regression of left ventricular hypertrophy.34
Any effects of treatment with furosemide, ACE inhibitors, or pimobendan on serum PIIINP concentration in dogs with MMVD remain to be elucidated. Therefore, measurements of UAC and serum PIIINP concentration might be confounded by the effects of treatment in studies that include dogs receiving medications for cardiac disease. Attempts were made to control for these effects in the present study by the inclusion of the following variables in the linear regression models: treatment with furosemide (yes vs no), treatment with ACE inhibitors (yes vs no), and treatment with pimobendan (yes vs no). Although these variables did not take into account any effect of dose on UAC or serum PIIINP concentration, no evidence for dose-dependent relationships was indicated by linear regression analysis (data not shown). As expected, treatment with furosemide was positively associated with UAC.
Treatment with furosemide was not significant in the multivariable model including percentage change over time data; however, to contribute data to this model, dogs with MMVD had to be examined on 3 occasions. Dogs with 3 evaluations were likely to be less severely affected and therefore less likely to be receiving furosemide, which might explain why treatment with furosemide was not significant in this model.
Treatment with neither ACE inhibitors nor pimobendan was significant in either multivariable model for UAC, and no treatment was significant in either multivariable model for PIIINP concentration. Those findings yield no evidence of important confounding effects of these treatments on the models.
To further investigate any potential effects of treatment on UAC and serum PIIINP concentration, the multivariable models were repeated, excluding dogs receiving medication for cardiac disease. Although some variables that were significant in the final multivariable model that included all dogs were not significant in the final multivariable model that included dogs receiving medication for cardiac disease, the regression coefficients for the variables that were significant in both multivariable models were similar. This suggests that the changes in the final models were most likely to be the result of the decrease in study power created by the elimination of data for 32% of the dogs. Additionally, it should be noted that the dogs excluded from these analyses were typically those with the most severe disease; therefore, their exclusion resulted in a decrease in the range of cardiac size over which the analytes were measured.
The low adjusted R2 values for the multivariable regression models in the present study may have reflected the fact that neither PIIINP concentration nor UAC is a specific marker of cardiac disease. For example, in humans, serum PIIINP concentration increases with hepatic disease,35 myelofibrosis,36 systemic sclerosis,37 and neoplasia.38 The excess amount of circulating aldosterone may be associated with primary hyperaldosteronism,39 coarctation of the aorta,40 renin-secreting tumors,41 disorders associated with low circulating volume, hepatic cirrhosis,42 or nephrotic syndrome.43 Plasma aldosterone concentration also increases in dogs fed a low-sodium diet.44 In addition, circulating aldosterone concentration may not be directly related to cardiac tissue aldosterone concentration.45,46 In our sample of middle-aged and older dogs, it was likely that other systemic processes influenced UACs and serum PIIINP concentrations, decreasing the strength of associations with cardiac variables.
The dogs in the multivariable analysis for which percentage change over time data were available needed to have been examined on at least 3 occasions, which might have biased findings toward less severely affected dogs, although death was not the only reason animals were lost to follow-up. The adjusted R2 value for the UAC model including the percentage change over time data was higher than that for the model excluding these data, suggesting that the model including percentage change more fully explained the relationships between the variables. However, the adjusted R2 remained < 0.5, suggesting that important factors influencing UAC were not included in the model.
The present study had a number of limitations. Only 1 marker of ECM turnover was measured, and serum procollagen type I concentration may have a different relationship with left ventricular echocardiographic measurements than serum PIIINP concentration. Histologic evaluation was not performed to investigate the relationship between UAC and serum PIIINP concentration and fibrosis; therefore, any putative relationship remains speculative.
Sodium-restricted diets are associated with increases in plasma aldosterone concentration in dogs.44,47 Although none of the dogs in the study reported here were fed such a diet, dietary sodium intake was not controlled and therefore may have influenced urinary aldosterone measurements. Intraobserver variability in the types of echocardiographic measurements made is reportedly approximately 10%,48 and this variability might have decreased the ability to detect significant associations between echocardiographic variables and PIIINP concentration or UAC.
Dogs were receiving various treatments, although the inclusion of these treatments in the multivariable model limited any potential for confounding of the results given these variables. The healthy control dogs were generally younger than the dogs with MMVD; therefore, an additional study to replicate the results of the present study would be ideal.
The healthy control dogs did not undergo echocardiography, so it was not possible to entirely rule out the presence of any cardiac disease in these dogs. However, given the types of heart disease to which dogs of this age and size are susceptible, a lack of abnormal physical examination findings made the presence of cardiac disease unlikely. Data from this group were included only for groupwise comparisons of PIIINP concentration and UAC and not in the construction of the multivariable models. Therefore, their inclusion did not affect the conclusions of the present study.
Finally, the primary veterinarians retained responsibility for all treatment decisions. Not all dogs with clinical signs of congestive heart failure underwent radiography prior to the initiation of treatment, meaning that diagnosis may have been inaccurate in dogs included in the C category of heart disease. Such inaccuracy would have had no effect on the major conclusions of the study because class was not a factor in the final multivariable models.
Regardless of the aforementioned limitations, serum PIIINP concentration and therefore, we believe, collagen type III turnover decreased with increasing measurements of left ventricular size in dogs with naturally occurring MMVD. Urinary aldosterone concentration increased with increasing rates of change in left ventricular measurements. Findings also suggested that there might be breed-specific differences in the pathophysiologic response to mitral regurgitation in dogs.
ABBREVIATIONS
| ACE | Angiotensin-converting enzyme |
| ACVIM | American College of Veterinary Internal Medicine |
| CKCS | Cavalier King Charles Spaniel |
| ECM | Extracellular matrix |
| LA:Ao | Ratio of left atrial diameter to aortic root diameter |
| LVEDD | Left ventricular end-diastolic diameter |
| LVEDDN | Left ventricular end-diastolic diameter normalized for body weight |
| LVESDN | Left ventricular end-systolic diameter normalized for body weight |
| LVFWd | Left ventricular free wall thickness in diastole |
| MMVD | Myxomatous mitral valve disease |
| PIIINP | N-terminal procollagen type III |
| RAAS | Renin-angiotensin-aldosterone system |
| RIA | Radioimmunoassay |
| UAC | Urinary aldosterone-to-creatinine concentration ratio |
UniQ PIIINP RIA, Orion Diagnostica Oy, Espoo, Finland.
Coat-a-Count Aldosterone, Siemens Healthcare Diagnostics, Deerfield, Ill.
IDEXX Laboratories, Wetherby, West Yorkshire, England.
Acuson Cypress, Siemens Medical Solutions, Oldbury, Bracknell, England.
PASW for Windows, version 18.0, SPSS Inc, Chicago, Ill.
References
1. Zannad F, McMurray JJ & Krum H et alEplerenone in patients with systolic heart failure and mild symptoms. N Engl J Med 2011; 364: 11–21.
2. Vizzardi E, D'Aloia A & Giubbini R et alEffect of spironolactone on left ventricular ejection fraction and volumes in patients with class I or II heart failure. Am J Cardiol 2010; 106: 1292–1296.
3. Pitt B, Zannad F & Remme WJ et alThe effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study Investigators. N Engl J Med 1999; 341: 709–717.
4. Pitt B, Remme W & Zannad F et alEplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N Engl J Med 2003; 348: 1309–1321.
5. Pitt B. Do diuretics and aldosterone receptor antagonists improve ventricular remodeling? J Card Fail 2002; 8: S491–S493.
6. Brilla CG, Matsubara LS, Weber KT. Anti-aldosterone treatment and the prevention of myocardial fibrosis in primary and secondary hyperaldosteronism. J Mol Cell Cardiol 1993; 25: 563–575.
7. Zhao J, Li J & Li W et alEffects of spironolactone on atrial structural remodelling in a canine model of atrial fibrillation produced by prolonged atrial pacing. Br J Pharmacol 2010; 159: 1584–1594.
8. Falk T, Jonsson L & Olsen LH et alAssociations between cardiac pathology and clinical, echocardiographic and electrocardiographic findings in dogs with chronic congestive heart failure. Vet J 2010; 185: 68–74.
9. Bernay F, Bland JM & Haggstrom J et alEfficacy of spironolactone on survival in dogs with naturally occurring mitral regurgitation caused by myxomatous mitral valve disease. J Vet Intern Med 2010; 24: 331–341.
10. Häggström J, Hansson K & Kvart C et alEffects of naturally acquired decompensated mitral valve regurgitation on the renin-angiotensin-aldosterone system and atrial natriuretic peptide concentration in dogs. Am J Vet Res 1997; 58: 77–82.
11. Gardner SY, Atkins CE & Rausch WP et alEstimation of 24-h aldosterone secretion in the dog using the urine aldosterone:creatinine ratio. J Vet Cardiol 2007; 9: 1–7.
12. Jensen LT. The aminoterminal propeptide of type III procollagen. Studies on physiology and pathophysiology. Dan Med Bull 1997; 44: 70–78.
13. Uusimaa P, Risteli J & Niemela M et alCollagen scar formation after acute myocardial infarction: relationships to infarct size, left ventricular function, and coronary artery patency. Circulation 1997; 96: 2565–2572.
14. Klappacher G, Franzen P & Haab D et alMeasuring extracellular matrix turnover in the serum of patients with idiopathic or ischemic dilated cardiomyopathy and impact on diagnosis and prognosis. Am J Cardiol 1995; 75: 913–918.
15. Stewart JA Jr, Wei CC & Brower GL et alCardiac mast cell- and chymase-mediated matrix metalloproteinase activity and left ventricular remodeling in mitral regurgitation in the dog. J Mol Cell Cardiol 2003; 35: 311–319.
16. Schuller S, Valentin S & Remy B et alAnalytical, physiologic, and clinical validation of a radioimmunoassay for measurement of procollagen type III amino terminal propeptide in serum and bronchoalveolar lavage fluid obtained from dogs. Am J Vet Res 2006; 67: 749–755.
17. Syme HM, Fletcher MG & Bailey SR et alMeasurement of aldosterone in feline, canine and human urine. J Small Anim Pract 2007; 48: 202–208.
18. Bland JM. Clinical measurement. In: An introduction to medical statistics. Oxford, England: Oxford University Press, 2000;268–293.
19. Hansson K, Haggstrom J & Kvart C et alLeft 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.
20. Thomas WP, Gaber CE & Jacobs GJ et alRecommendations for standards in transthoracic two-dimensional echocardiography in the dogs and cat. Echocardiography Committee of the Specialty of Cardiology, American College of Veterinary Internal Medicine. J Vet Intern Med 1993; 7: 247–252.
21. Cornell CC, Kittleson MD & Della Torre P et alAllometric scaling of M-mode cardiac measurements in normal adult dogs. J Vet Intern Med 2004; 18: 311–321.
22. Atkins C, Bonagura J & Ettinger S et alGuidelines for the diagnosis and treatment of canine chronic valvular heart disease. J Vet Intern Med 2009; 23: 1142–1150.
23. Haggstrom J, Hansson K & Kvart C et alChronic valvular disease in the Cavalier King Charles Spaniel in Sweden. Vet Rec 1992; 131: 549–553.
24. Sayer MB, Atkins CE & Fujii Y et alAcute effect of pimobendan and furosemide on the circulating renin-angiotensin-aldosterone system in healthy dogs. J Vet Intern Med 2009; 23: 1003–1006.
25. Sisson DD. Neuroendocrine evaluation of cardiac disease. Vet Clin North Am Small Anim Pract 2004; 34: 1105–1126.
26. Pedersen HD, Koch J & Poulsen K et alActivation of the renin-angiotensin system in dogs with asymptomatic and mildly symptomatic mitral valvular insufficiency. J Vet Intern Med 1995; 9: 328–331.
27. Pedersen HD. Effects of mild mitral valve insufficiency, sodium intake, and place of blood sampling on the renin-angiotensin system in dogs. Acta Vet Scand 1996; 37: 109–118.
28. Weber KT, Brilla CG. Pathological hypertrophy and cardiac interstitium. Fibrosis and renin-angiotensin-aldosterone system. Circulation 1991; 83: 1849–1865.
29. O'Gara P, Sugeng L & Lang R et alThe role of imaging in chronic degenerative mitral regurgitation. JACC Cardiovasc Imaging 2008; 1: 221–237.
30. Atkins CE, Rausch WP & Gardner SY et alThe effect of amlodipine and the combination of amlodipine and enalapril on the renin-angiotensin-aldosterone system in the dog. J Vet Pharmacol Ther 2007; 30: 394–400.
31. Urata H, Kinoshita A & Misono KS et alIdentification of a highly specific chymase as the major angiotensin II-forming enzyme in the human heart. J Biol Chem 1990; 265: 22348–22357.
32. MacFadyen RJ, Lee AF & Morton JJ et alHow often are angiotensin II and aldosterone concentrations raised during chronic ACE inhibitor treatment in cardiac failure? Heart 1999; 82: 57–61.
33. Lopez B, Querejeta R & Gonzalez A et alEffects of loop diuretics on myocardial fibrosis and collagen type I turnover in chronic heart failure. J Am Coll Cardiol 2004; 43: 2028–2035.
34. Sasaguri M, Noda K & Tashiro E et alThe regression of left ventricular hypertrophy by imidapril and the reduction of serum procollagen type III amino-terminal peptide in hypertensive patients. Hypertens Res 2000; 23: 317–322.
35. Leroy V, Monier F & Bottari S et alCirculating matrix metalloproteinases 1, 2, 9 and their inhibitors TIMP-1 and TIMP-2 as serum markers of liver fibrosis in patients with chronic hepatitis C: comparison with PIIINP and hyaluronic acid. Am J Gastroenterol 2004; 99: 271–279.
36. Hochweiss S, Fruchtman S & Hahn EG et alIncreased serum procollagen III aminoterminal peptide in myelofibrosis. Am J Hematol 1983; 15: 343–351.
37. Lee YJ, Shin KC & Kang SW et alType III procollagen N-terminal propeptide, soluble interleukin-2 receptor, and von Willebrand factor in systemic sclerosis. Clin Exp Rheumatol 2001; 19: 69–74.
38. Zhu GG, Stenback F & Risteli L et alOrganization of type III collagen in benign and malignant ovarian tumors. An immunohistochemical study. Cancer 1993; 72: 1679–1684.
39. Young WF Jr. Minireview: primary aldosteronism-changing concepts in diagnosis and treatment. Endocrinology 2003; 144: 2208–2213.
40. Alpert BS, Bain HH & Balfe JW et alRole of the renin-angiotensin-aldosterone system in hypertensive children with coarctation of the aorta. Am J Cardiol 1979; 43: 828–834.
41. Beevers DG, Maheshwari MB & Ryan PG et alHypertension due to a renin-secreting juxtaglomerular cell tumor. Am J Hypertens 2008; 21: 1359–1361.
42. Kuiper JJ, Boomsma F & van Buren H et alComponents of the renin-angiotensin-aldosterone system in plasma and ascites in hepatic cirrhosis. Eur J Clin Invest 2008; 38: 939–944.
43. Schrier RW, Masoumi A, Elhassan E. Aldosterone: role in edematous disorders, hypertension, chronic renal failure, and metabolic syndrome. Clin J Am Soc Nephrol 2010; 5: 1132–1140.
44. Freeman LM, Rush JE, Markwell PJ. Effects of dietary modification in dogs with early chronic valvular disease. J Vet Intern Med 2006; 20: 1116–1126.
45. Silvestre JS, Robert V & Heymes C et alMyocardial production of aldosterone and corticosterone in the rat. Physiological regulation. J Biol Chem 1998; 273: 4883–4891.
46. Weber KT. Aldosteronism revisited: perspectives on less well-recognized actions of aldosterone. J Lab Clin Med 2003; 142: 71–82.
47. Rush JE, Freeman LM & Brown DJ et alClinical, echocardiographic, and neurohormonal effects of a sodium-restricted diet in dogs with heart failure. J Vet Intern Med 2000; 14: 513–520.
48. Dukes-McEwan J, French AT, Corcoran BM. Doppler echocardiography in the dog: measurement variability and reproducibility. Vet Radiol Ultrasound 2002; 43: 144–152.
