The diagnostic relevance of NT-proBNP and proANP 31–67 measurements in staging of myxomatous mitral valve disease in dogs



Gerhard Wess, DVM, Dr. med. vet. habil., Dipl. ACVIM (Cardiology), Dipl. ECVIM-CA (Cardiology and Internal Medicine), Clinic of Small Animal Medicine, LMU University, Veterinaerstr. 13, 80539 Munich, Germany




There is no agreement in current publications regarding the reliability of serum concentrations of natriuretic peptides (NPs) to detect dogs with subclinical myxomatous mitral valve disease (MMVD) and to differentiate between asymptomatic stages.


We sought to compare N-terminal pro-B-type natriuretic peptide (NT-proBNP) and pro-atrial natriuretic peptide 31-67 (proANP) concentrations between various stages of canine MMVD and to investigate the influence of age, weight, and sex.


In this prospective study, dogs were classified in different disease stages using the modified Canine Heart failure International Expert Forum (CHIEF) system. Serum NP concentrations were compared between groups.


A total of 559 samples from 116 healthy dogs and 236 dogs with MMVD were analyzed. Using cut-off values (1207 pmol/L for NT-proBNP, 1578 fmol/mL for proANP), dogs with MMVD with and without congestive heart failure (CHF) could be differentiated with a sensitivity of 83% for both and specificities of 85% and 86%, respectively. Dogs staged in CHIEF B1 and B2 could not be distinguished based on NP concentrations due to wide variation within the groups. Intact females (means 598 pmol/L and 1036 fmol/mL, respectively) had significantly higher values of both NPs than intact males (315 pmol/L and 836 fmol/mL).


NPs in canine MMVD are useful to discriminate between asymptomatic dogs and dogs with CHF. Due to a large overlap of NP-concentrations between the groups, NPs do not seem to be useful to differentiate between dogs in stages B1 and B2. Interpretation of NT-proBNP and proANP values should include consideration of sex-specific differences.


Myxomatous mitral valve disease (MMVD) accounts for approximately 75% of all heart diseases in dogs.[1] The prevalence of the disease is higher in small dogs (< 20 kg) than in large breeds, and disease progression takes years.[2, 3] Various methods to stage the severity of mitral valve disease are used in the current literature. Modified New York Heart Association (NYHA),[4] International Small Animal Cardiac Health Council (ISACHC),[5] and Canine Heart failure International Expert Forum (CHIEF)[6] are classification methods based on clinical and radiographic changes. The CHIEF system is a newer approach based on recent guidelines of the American College of Cardiology and the American Heart Association.[7] In this system, and in contrast to previous staging methods,[5] asymptomatic dogs are divided into 2 groups, one group with previous signs of congestive heart failure (CHF) and the other group that never showed signs of CHF. A similar staging method subdividing asymptomatic dogs into 2 subgroups according to cardiac size is recommended by the consensus panel for therapy of MMVD of the American College of Veterinary Internal Medicine.[8] Traditionally, for the diagnosis of MMVD and assessment of disease severity, echocardiographic and radiographic examinations are relied on. To quantify the amount of mitral valve regurgitation, several echocardiographic approaches based on 2-dimensional (2-D) ultrasonography and Doppler-derived methods are available.[3, 9, 10]

In addition, several studies investigated concentrations of natriuretic peptides (NPs) in dogs at different stages of MMVD, and there is general consent that NPs are useful to detect patients with respiratory symptoms due to CHF.[11, 12] Both B-type natriuretic peptide (BNP) and atrial natriuretic peptide (ANP) are released by myocardial tissue in response to volume or pressure overload and consecutive increased ventricular and atrial wall stretch.[13, 14] In circulation, BNP and ANP increase natriuresis and diuresis, and decrease systemic vascular resistance.[15] BNP is synthesized as proBNP and further cleaved into BNP and the biological inactive N-terminal end NT-proBNP.[16] ANP is encoded as a 126-amino acid precursor molecule, which is cleaved into its 98-amino acid fragment amino-terminal pro-atrial natriuretic peptide (NT-proANP) and ANP, similar to BNP. NT-proANP is further cleaved into the 3 molecules proANP 1-30, proANP 31-67, and proANP 68-98,[17] which appear to have all physiologic functions similar to those of ANP,[18] while some authors do not describe any biologic activity.[19] Importantly, the half-lives of NT-proBNP and NT-proANP fragments are significantly longer than the ones of BNP and ANP, which make them preferable as diagnostic markers.[20, 21] Furthermore, studies in people suggest that proANP is more sensitive compared to ANP in detecting mild increases in atrial filling pressures[22] and that proANP is the most sensitive marker among ANP, BNP, C-type NP, and NT-proBNP in discriminating NYHA Class I individuals (patients with no limitations of physical activity) from healthy individuals.[23] Interestingly, NP concentrations in people seem to be dependent on body mass index (BMI), age, and androgen levels,[24-26] but no consistent veterinary reports about age-, gender- and weight-related variations in healthy dogs exist.[27-30]

So far, there is no consensus regarding the diagnostic power of NP concentrations to discriminate dogs without cardiac disease from dogs with subclinical stages of MMVD, and to differentiate between dogs in different asymptomatic stages.[31-35] Therefore, one aim of the present study was to investigate the diagnostic accuracy of NT-proBNP and proANP determinations in a large study population of dogs in different stages of MMVD, using a modified CHIEF system. In addition, we sought to compare the diagnostic power of NT-proBNP with that of proANP, and to investigate the influence of age, body size, and sex on NT-proBNP and proANP concentrations in a healthy dog population.

Materials and Methods

Animal recruitment

This prospective study included client-owned dogs of various breeds with MMVD and a healthy control group. The affected dogs were presented for cardiac workup and were prospectively enrolled at the Clinic of Small Animal Medicine, LMU University, Munich, Germany between February 2005 and November 2010. As other conditions such as renal failure or cardiac disease other than MMVD can affect NPs concentrations,[36-38] dogs diagnosed with such diseases were excluded from this study. In addition, systolic blood pressure over 160 mmHg was an exclusion criterion. Written owner consent was obtained. Every examination included a complete physical examination, blood pressure measurement, and an echocardiographic examination. If respiratory symptoms were reported by the owners or were noticed during examination, chest radiographs were taken. All data regarding history, physical examination, and diagnostic findings were recorded in a standardized datasheet. Control dogs were animals owned by faculty, staff members, or students, and were determined to be healthy based on history, physical examination, and a normal echo-cardiographic examination.

Diagnostic procedures and staging

Echocardiographic examinations were performed by experienced cardiologists (Dipl. ACVIM or residents) and were executed in nonsedated patients restrained in left and right lateral recumbency. To exclude any hereditary abnormalities, a complete echocardiographic examination of all valves with color, continuous and pulsed wave Doppler was performed. Diameters of left atrium and aorta were measured in the right parasternal short axis at the level of the heart base and the ratio of the left atrium to the aorta (LA/Ao) was calculated.[39] The left atrium was considered enlarged when LA/Ao was >1.5. To evaluate ventricular dimensions, motion-mode (M-mode) measurements using the right parasternal long-axis view were used. Criteria for the diagnosis of MMVD were used as described elsewhere.[40] The left apical 4-chamber view was used to assess size of mitral regurgitation with color Doppler. Images were analyzed frame by frame to determine the area of the regurgitant jet (ARJ), which was related to left atrium area (LAA; ARJ/LAA).[10] If tricuspid or pulmonic valve insufficiency was present, peak velocity of the regurgitant jet was measured by continuous wave Doppler as previously reported, to search for evidence of pulmonary hypertension.[41]

All dogs with MMVD were classified according to the CHIEF method on the basis of their clinical symptoms and cardiac size (Table 1). According to recent recommendations,[8] asymptomatic dogs (CHIEF B) were further subclassified into B1 (normal cardiac size) and B2 (enlarged heart), with cardiac enlargement defined by left ventricular internal end-diastolic dimension (LVIDd) and/or LA/Ao above the respective reference ranges.[42] If a patient was seen more than once, disease stage was re-evaluated at every recheck. Disease staging was performed before NP concentrations were analyzed.

Table 1. Classification of dogs with heart failure due to myxomatous mitral valve disease (MMVD) according to the Canine Heart failure International Expert Forum (CHIEF)
ADogs at risk of developing heart disease with no identifiable cardiac structural disorder
BDogs with structural heart disease, but never developed clinical signs caused by heart failure
C1Dogs with signs of heart failure in the past, but without current symptoms
C2Dogs with mild-to-moderate heart failure
C3Dogs with severe-to-life-threatening heart failure
DDogs with end-stage MMVD and heart failure that is refractory to standard therapy

Sample collection, handling, and stability

Blood samples were drawn from the cephalic or saphenous vein or were collected via jugular puncture into chilled tubes containing EDTA and serum tubes. EDTA-samples for NT-proBNP and proANP analysis were immediately centrifuged at 3600g at 4°C for 10 minutes, and the plasma was separated and frozen at −70°C to prevent enzymatic degradation of the analytes. Samples were analyzed in batches in our laboratory by a person trained by a laboratory specialist between December 2009 and January 2011 and blinded to the dogs' CHIEF classification. Serum was used for creatinine and urea measurements.

Plasma NT-proBNP and proANP were analyzed using commercially available assays (VETSIGN Canine CardioSCREEN NT-proBNP; Guildhay Ltd., Biomedica, Vienna, Austria; VETSIGN Canine CardioSCREEN proANP 31-67; Guildhay Ltd., Biomedica) according to the manufacturer's instructions. Both assays have been previously used and validated for diagnostic purpose in dogs.[43, 44] All samples were measured in duplicate and the means of the 2 results were used for further data analysis. If the NT-proBNP concentrations were above the upper detection limit of the assay (3306 pmol/L), samples were diluted at a ratio of 1:10 with sterile 0.9% sodium chloride solution and re-analyzed. For quality control, commercial human control samples with known NT-proBNP and proANP concentrations (Biomedica) were used. If the coefficient of variation between the 2 measurements of one sample was more than 10%, the sample was re-analyzed.

To evaluate the effect of storage on sample integrity, a short- (storage < 1 year, ie from 2010 on) and long-term (storage > 3 years, ie 2005–2007) storage group for the controls and the CHIEF groups were allocated from the available samples.


Asymptomatic dogs received no therapy, while dogs with CHF received diuretics (Furosemide, Hydrochlorothiazide, Spironolactone) and a calcium-sensitizer (Pimobendan). Some dogs also received, an angiotensin-converting-enzyme inhibitor was added (Ramipril).

Statistical analysis

Statistical analysis was performed using commercially available software (Statistical Package for the Social Sciences 18.0 for Windows; SPSS Inc., Chicago, IL, USA; MedCalc for Windows, Version; MedCalc Software, Mariakerke, Belgium). Normal distribution of the data was evaluated by Kolmogorov–Smirnov analysis. If data were not normally distributed, a natural log transformation was used. To adjust for repeat measurements, a linear mixed model with Bonferroni post-hoc correction was used to compare the NP concentrations between the control and the various MMVD groups. Receiver-operating characteristic (ROC) curve analysis and interactive dot diagrams were used to determine the diagnostic ability of proANP and NT-proBNP to predict CHF, to discriminate CHIEF B1 dogs from B2 dogs, to distinguish the control group from patients in stages B1 and B2, and to define optimal discrimination limits. The areas under the curve (AUC) were calculated. Optimal cut-off values were determined using ROC-curve analysis for different clinical questions. The software indicated the maximum of the Youden index: J = max(SEi + SPi − 1), where SEi and SPi are the sensitivity and specificity over all possible threshold values. This value corresponds to the point on the ROC curve farthest from the diagonal line. Only independent observations were included for ROC curve analysis and dot diagrams using the first examination of every patient. To test a possible influence of age, weight, sex, and neutering status on the NP concentrations in the healthy control group, a one-way analysis of variance with Bonferroni post-hoc correction was performed. The short- and long-term storage groups were compared using an unpaired t-test. Significance was defined at < .05.


A total of 559 samples of 352 dogs were collected as part of this study between February 2005 and December 2010. The control group consisted of 116 healthy dogs of various breeds (mean weight 22.5 kg, median 17.6 kg; mean age 5.7 years, median 3.9 years). All dogs in the control group were only examined once and by one blood sample. The group included 30 cross-bred dogs, 6 Labrador Retrievers, 6 Great Danes, 6 Cavalier King Charles Spaniels (CKCS), 6 Dachshunds, 6 Australian Shepherds, 5 Boxers, 5 Golden Retrievers, 4 Jack Russel Terriers, 4 Beagles, 3 Irish Wolfhounds, 3 Yorkshire Terriers, 2 Pugs, 2 Newfoundland Dogs, 2 Flat Coated Retrievers, and one each of 26 other breeds.

Mean NT-proBNP in the control group was 455.3 pmol/L (median 384.3) and proANP was 945.5 fmol/mL (median 308.3). Age and body size had no significant influence on NP levels. However, sex and neutering status had a significant influence on NT-proBNP and proANP concentrations. Intact female dogs had significantly higher NP values compared with intact male dogs (P = .03 and .01, respectively), whereas NP values of male neutered dogs were not significantly different from female spayed dogs. Also, no difference was found between intact and castrated male dogs (Table 2).

Table 2. Effect of sex and neutering on N-terminal pro-brain natriuretic peptide (NT-proBNP, pmol/L) and pro-atrial natriuretic peptide 31-67 (proANP, fmol/L) concentrations in male and female dogs
Sex n NT-proBNPaPercentile 05Percentile 95MinMax
Female31596 (83)427766761978
Female spayed32444 (81)280609332630
Male35315 (34)246385431035
Male neutered18528 (74)371685151181
  n proANPaPercentile 05Percentile 95MinMax
  1. a

    Data are means (SD).

  2. SD indicates standard deviation; Min, minimum; Max, maximum.

Female311036 (61)46519559341184
Female spayed32971 (54)86110815491881
Male35836 (50)7349383891583
Male neutered18923(49)82010265851205

In the MMVD group, a total of 443 NT-proBNP and proANP measurements of 236 dogs (mean weight 13 kg, median 9.7 kg, mean age 10.5 years, median 2.9 years) were performed at different time points. Breeds of the MMVD group included 75 cross-bred dogs, 45 Dachshunds, 16 CKCS, 12 Poodles, 9 Jack Russell Terriers, 6 Pekinese, 5 Yorkshire Terriers, 4 Beagles, 3 Berger des Pyrenees, 3 Airedale Terriers, 3 Border Collies, 3 Chihuahuas, 3 German Spitz, 3 Maltese, 3 Miniature Pinschers, 3 Shi Tzu, 2 each of 8 other breeds, and one each of 24 other breeds. The numbers of NP measurements within the different CHIEF classification groups are listed in Table 3.

Table 3. Ranges of N-terminal pro-brain natriuretic peptide (NT-proBNP, pmol/L) and pro-atrial natriuretic peptide 31-67 (proANP, fmol/L) concentrations in groups of dogs assigned to different stages of myxomatous mitral valve disease according to the Canine Heart failure International Expert Forum (CHIEF)
 CHIEF n Mean (SD)Percentile 05Percentile 95MinMax
  1. SD indicates standard deviation; Min, minimum; Max, maximum.

NT-proBNPControl116458 (36)387529152630
B1119711 (59)594827473513
B295987 (102)78511902586322
C11142177 (168)184425101898562
C2752766 (245)2278325423610056
C3334630 (507)35965663105213226
D74920 (946)2604723616058092
proANPControl116945 (29)88910023981955
B11191048 (38)97311242182422
B2951254 (48)115913494512450
C11142006 (76)185521574164571
C2752081 (89)190422586154050
C3332396 (133)2126266610554410
D72562 (172)2142298217813067

Concentrations of both NPs increased with advanced disease severity. NT-proBNP values of all disease stages (B1-D) were significantly higher than NT-proBNP of the control group (Figure 1). ProANP concentration was not significantly different between the control group and B1, while proANP concentrations in dogs at stages B2–D1 were significantly higher than the control and B1 group dogs (Figure 2). For both NT-proBNP and proANP, the differences between stage B1 and B2 were significant (P = .006 and < .001, respectively). Both NT-proBNP and proANP discriminated patients with acute heart failure (C2, C3, D) significantly from all asymptomatic stages (B1, B2, P < .001).

Figure 1.

Box plots illustrating the concentration of N-terminal pro-brain natriuretic peptide (NT-proBNP) in groups of dogs with myxomatous mitral valve disease categorized according to the Canine Heart failure International Expert Forum (CHIEF) classification. The whiskers indicate the range of values, the boxes represent the inter quartile range (IQR), and the lines within the boxes indicate the median. Outliers are indicated by individual points. *Statistically significant difference from all other stages. + Statistically significant difference from controls, B1, B2, C3. ¤ Statistically significant difference from controls, B1, B2, C1-2. ¥ Statistically significant difference from controls, B1, B2.

Figure 2.

Box plots illustrating the concentration of pro-atrial natriuretic peptide 31-67 (proANP) in groups of dogs with myxomatous mitral valve disease categorized according to the Canine Heart failure International Expert Forum (CHIEF) classification. The whiskers indicate the range of values, the boxes represent the inter quartile range (IQR) and the lines within the boxes indicate the median. Outliers are indicated by individual points. ˄ Statistically significant difference from B2, C1-3, D. ‡ Statistically significant difference from controls, B1, C1-3, D. ¥ Statistically significant difference from controls, B1, B2.

The ROC curve analyses were used to calculate cut-off values for reliably predicting different clinical stages. The best cut-off value to predict whether a dog with dyspnea but without a murmur (control vs C2, C3 and D) had CHF or not was > 1181 pmol/L for NT-proBNP, with a sensitivity of 83%, a specificity of 98%, and an AUC of 0.95. For proANP, the cut-off value was > 1366 fmol/mL with a sensitivity of 91%, a specificity of 90%, and an AUC of 0.96.

To distinguish between dogs with acute heart failure due to MMVD (C2, C3, and D) and dogs with MMVD (B1 and B2) but without heart failure, an NT-proBNP cut-off value of 1207 pmol/L and a proANP cut-off of 1578 fmol/mL provided the best sensitivity (83% and 83%, respectively) and specificity (85% and 86%, respectively; Figure 3). AUC was 0.89 and 0.9, respectively. Using the higher NT-proBNP cut-off value of > 1800 pmol/L recommended by the manufacturer increased the specificity to 91%, but resulted in a sensitivity of only 70%.

Figure 3.

Interactive dot diagrams displaying sensitivity (Sens) and specificity (Spec, in%) of pro-brain natriuretic peptide (NT-proBNP, top) and pro-atrial natriuretic peptide 31-67 (proANP, bottom) at diagnostic cut-off values (horizontal lines) of 1207 pmol/L and 1578 fmol/mL, respectively, to distinguish groups of dogs with myxomatous mitral valve disease categorized according to the Canine Heart failure International Expert Forum (CHIEF) as asymptomatic dogs (CHIEF B, n = 138) from dogs with acute congestive heart failure (CHIEF C2, C3, D, n = 78). The dots above the line in the CHIEF B group indicate false-positive test results, and dots below the line in the CHIEF C2, C3 and D groups represent false-negative test results.

Although there was a significant difference between B1 and B2, the calculated cut-off value for NT-proBNP (518 pmol/L) and proANP (1065 fmol/mL) could not reliably detect patients with cardiac enlargement, because sensitivity (83% and 64%, respectively) and specificity (50% and 55%, respectively) were comparatively low, and there was a large overlap between the 2 groups (AUC was 0.7 and 0.7, respectively; Figure 4). An NT-proBNP cut-off of 545 pmol/L and a proANP cut-off of 1246 fmol/mL were not useful to differentiate dogs in stage B from controls due to low sensitivity (60% and 39%, respectively) and specificity (75% and 87%).

Figure 4.

Interactive dot diagrams displaying sensitivity (Sens) and specificity (Spec, in%) of pro-brain natriuretic peptide (NT-proBNP, top) and pro-atrial natriuretic peptide 31-67 (proANP, bottom) diagnostic cut-off values (horizontal lines) of 518 pmol/L and 1065 fmol/mL, respectively, to distinguish dogs with myxomatous mitral valve disease categorized according to the Canine Heart failure International Expert Forum (CHIEF) with heart failure class B1 (n = 86) from dogs with heart failure class B2 (n = 52). The dots above the line in the CHIEF B1 group indicate false-positive test results, and dots below the line in the CHIEF B2 group represent false-negative test results.

No statistically significant differences were found between the NT-proBNP and proANP values in the long term (n = 149, controls: n = 31, CHIEF B1: n = 30, CHIEF B2: n = 29, CHIEF C1: n = 18, CHIEF C2: n = 24, CHIEF C3: n = 13, CHIEF D: n = 4) and short term (n = 99, controls: n = 41, CHIEF B1: n = 16, CHIEF B2: n = 13, CHIEF C1: n = 18, CHIEF C2: n = 6, CHIEF C3: n = 5, CHIEF D: n = 0) samples. CHIEF D groups were not compared.


Biomarkers might be beneficial in situations where conventional radiographic examinations are not sensitive enough to allow an accurate diagnosis or when echocardiography is not available or too expensive. The aim of this study was to evaluate whether NPs are a useful diagnostic variable to differentiate between the various disease stages in a comparatively large study population of dogs with MMVD and to investigate the influence of age, body size, and sex on NT-proBNP and proANP concentrations in a healthy dog population.

Similar to previous reports, the present study shows significantly higher concentrations of NT-proBNP and proANP in patients with CHF in comparison with the healthy control group or asymptomatic dogs with MMVD.[28, 32, 34, 45] In addition, our results suggest that a cardiac dysfunction as the cause of dyspnea can be ruled out at NT-proBNP concentrations of < 1181 pmol/L and proANP concentrations of < 1366 fmol/L with high specificity in dyspneic dogs presenting without a cardiac murmur. However, clinically, this question is rarely relevant, as small dogs without a murmur will not have CHF due to MMVD. Therefore, NPs could be useful as a biomarker for rare cases, such as dogs that are difficult to auscultate.

The results of this study indicate that using an NT-proBNP cut-off value of > 1207 pmol/L and a proANP cut-off value of > 1578 fmol/mL allows the differentiation of dogs with MMVD and CHF from dogs with MMVD without CHF with reasonable sensitivity and specificity. These cut-offs are higher than the ones discussed above likely due to cardiac enlargement present in some asymptomatic dogs with MMVD. Although a different assay was used in the present study, our cut-off value is very similar to those of 2 other reports, which reported a cut-off concentration of NT-proBNP > 1158 pmol/L to detect heart failure due to any kind of cardiac disease[11] and a cut-off of proANP > 1400 pmol/L to differentiate dogs with decompensated MMVD and DCM from dogs with primary respiratory disease.[46] However, as shown in Figure 3, there is some overlap in NP concentrations between symptomatic and asymptomatic patients, similar to the findings in other veterinary and human studies.[28, 47] Using a higher cut-off value of > 1800 for NT-proBNP resulted in an increased specificity of 91%, but lower sensitivity of only 71%. Therefore, under clinical circumstances, chances are better that higher NP concentrations correctly predict CHF. These results show that the measurement of serum NP concentrations in dogs with radiographic and echocardiographic examinations non-diagnostic for CHF can represent a useful diagnostic tool. Furthermore, a rapid (bed side) diagnostic test method for NT-proBNP might be helpful in the assessment of cardiac dysfunction in dyspneic dogs in critical care, as shown in human medicine.[48] However, NP levels must be interpreted with caution because increased concentrations can occur in dogs with primary respiratory disease and pulmonary hypertension.[11]

To the authors' knowledge, this is the first study using the CHIEF system for disease staging. NT-proBNP-concentrations of dogs in B1 and B2 and proANP concentrations of dogs in B2 were significantly higher compared with the healthy control group, which confirm results of another study demonstrating a significant difference between dogs in ISACHC Ia and b, and the healthy control group.[31] Class CHIEF B1 and ISACHC Ia are comparable, because they both refer to asymptomatic patients with no chamber enlargement. Using an NT-proBNP cut-off value of > 545 pmol/L dogs in CHIEF B1 and B2 can be differentiated from the healthy population with a low sensitivity and specificity. This is in contrast to 2 other studies showing comparable NT-proBNP and NT-proANP concentrations between in CKCS with mild and moderate MMVD or dogs of various breeds in ISACHC stage Ia, respectively, and a healthy control group.[21, 33] The reason for these conflicting results is not clear. However, in the present study, the study population was comparatively large, which may have increased the probability of detecting a significant difference between groups.

Due to the low sensitivity and specificity of the calculated cut-offs, NP measurements are not useful to differentiate between healthy and asymptomatic patients with MMVD. Rather, dogs with MMVD can easily be identified by a heart murmur based on auscultation, a very reliable, cost-efficient, and non-invasive diagnostic tool. Using NPs to detect asymptomatic heart disease seems to be more useful in the occult stage of DCM.[49] Many of these dogs have no obvious murmurs and may therefore benefit from NT-proBNP testing.[49]

The prediction of cardiac enlargement in asymptomatic patients is relevant as such dogs might benefit from early therapy. Both NT-proBNP and proANP concentrations were significantly different between CHIEF B1 and B2. An NT-proBNP cut-off value of 518 pmol/L and a proANP value of 1065 fmol/mL allowed discrimination of dogs in stage B1 from dogs in stage B2 with a reasonable sensitivity, but low specificity. Another recent study reported a significant difference in NT-proBNP concentration between patients in ISACHC class Ia and those in class Ib in a cohort of 72 asymptomatic dogs with MMVD.[31] In contrast, NT-proBNP could not discriminate between ISACHC stage Ia and Ib in another report with 44 dogs.[33] This discrepancy might again be attributable to the low number of dogs (44 vs 559) examined. As stage ISACHC Ib includes the CHIEF stages B2 and C1, a direct comparison with the results of the present study is not possible.

Although in our study there was a statistically significant difference of NT-proBNP and proANP concentrations between CHIEF B1 and B2, the cut-offs cannot be used as a basis for clinical decisions based on the extensive overlap between the groups, and radiographic or echocardiographic examinations remain the appropriate diagnostic modalities of choice to differentiate between CHIEF stages B1 and B2.

The present study indicates that NP testing could be useful to monitor disease progression, as NP concentrations increased with severity of disease. Serial testing of NPs might be useful for monitoring disease progress and identifying patients with clinically significant heart disease; however, the reported weekly or daily variability of NPs might be a limiting factor.[50] This, hypothesis has to be confirmed by further studies.

Another interesting and potentially clinically relevant result of our study were the statistically significant differences in NT-proBNP and proANP concentrations between intact male and female dogs. These results are in agreement with several studies in people, where BNP, ANP, and NT-proBNP concentrations were significantly higher in women than in men.[24, 51, 52] The reason for these results is not completely clear. One recent study suggests that androgens are mediating sex-related differences in BNP and NT-proBNP levels.[25]

There was no difference in NP concentrations between different body size and age groups, which is consistent with findings in other canine studies.[27, 28, 50] However, in one study including only healthy CKCS, NT-proANP was lower in larger breeds and higher in older dogs,[29] and in another study, a population of overweight dogs had significantly lower NT-proANP concentrations than controls with normal BMI.[53] Also, a study in Doberman Pinschers showed significantly increased NT-proBNP concentration in dogs > 8 years of age.[49] These results are similar to those of human studies, where NP levels decrease with increasing BMI and rise in older populations.[24, 26] These discrepancies might be explained by the fact that the dogs in the present study were classified according to their absolute weight and age. BMI and different life expectancy among various breeds were not taken into consideration.[54]

Some of the samples were stored for 3 or more years until analysis. This might represent a limitation of the present study as, to the authors' knowledge, there are no reports on long-term stability of NPs in veterinary medicine. In one human study, there was no substantial loss of immunoreactivity after a 2-year storage of NT-proBNP at −20°C.[55] We found no statistically significant difference between long-term and short-term stored samples. Although this suggests that NPs in our samples were stable at−70C, a validation study addressing storage stability is warranted.

Some of the apparently healthy dogs and some of the dogs with mild MMVD had NT-proBNP and proANP concentrations in a range usually observed in patients with CHF. About half of the patients with increased NT-proBNP concentrations had concurrent high proANP values. All other dogs had either increased NT-proBNP levels or increased proANP alone. Whether these dogs had a subclinical undiagnosed disease, falsely elevated levels due to assay interference,[56] or whether this represents a wide biological range leading to an overlap between healthy and affected dogs is not clear. No further screening for additional diseases was undertaken, except measurement of urea and creatinine concentrations, 2 parameters known to be of limited sensitivity to indicate chronic disease.

Although this study included a large dog population, the number of animals in class CHIEF D was very low, resulting in low statistical power. Furthermore, to define an optimal cut-off value to discriminate between patients with dyspnea and/or coughing due to CHF, and dogs with respiratory disease, a group with dyspnea related to primary respiratory disease and concurrent asymptomatic MMVD is needed. There was no such group in the present study resulting in the limited reliability of the cut-off value. Finally, mean age and body weight ranges of the control group compared with the MMVD group are limiting the power of the current study.

In summary, this study showed that NT-proBNP and proANP could differentiate dogs with CHF from patients with MMVD not yet affected by CHF. However, due to low sensitivity and specificity, NT-proBNP and proANP testing is not useful to detect asymptomatic dogs with MMVD, and to differentiate CHIEF stages B1 and B2 in asymptomatic dogs. Significant differences between healthy intact male and female dogs suggest that sex-specific reference intervals should be determined to help interpret concentrations of NT-proBNP and proANP.


We thank Idexx Laboratories for sponsoring the NT-proBNP and proANP test kits.

Disclosure: The authors have indicated that they have no affiliations or financial involvement with any organization or entity with a financial interest in, or in financial competition with, the subject matter or materials discussed in this article.