Clinical features and outcome in 25 dogs with respiratory‐associated pulmonary hypertension treated with sildenafil

Abstract Background Pulmonary hypertension (PH) can develop secondary to many common cardiopulmonary diseases, and the use of sildenafil has improved care of affected dogs. Objective To evaluate response to sildenafil in dogs with respiratory‐associated PH. Animals Twenty‐five dogs with PH. Methods Prospective clinical trial. Doppler echocardiography identified dogs with moderate to severe PH, and additional tests were performed to detect underlying diseases. A 17‐point quality of life (QOL) questionnaire was completed, and sildenafil was prescribed, along with other medications deemed necessary for the management of clinically diagnosed respiratory diseases. After 30 days, dogs returned to the hospital for repeat echocardiogram and QOL survey. Results The median age was 12.4 years, and most dogs were small breed dogs (median weight, 6.5 kg). Syncope (64%), cough (56%), and respiratory difficulty (32%) were the most common presenting complaints. Respiratory diseases associated with PH included tracheobronchomalacia, pulmonary fibrosis, inflammatory airway disease, and brachycephalic syndrome, with multiple diseases in some dogs. Eight of 25 dogs (32%) died or were euthanized within 1 month. In the remaining dogs, tricuspid regurgitation pressure gradient (83.0 ± 17.4 mm Hg before, 55.4 ± 17.4 mm Hg after) and QOL scores were significantly improved after 1 month of sildenafil. Fifty percent mortality was reached 6 months after study entry, with 4 dogs alive 5 years after diagnosis. Conclusions and Clinical Importance Sildenafil responsiveness is variable in dogs with respiratory‐associated PH, but improved QOL was demonstrated in dogs surviving >1 month, and long‐term survival was noted in some cases.

pressure exceeding 30 and 20 mm Hg, respectively. 1 Originally, PH was characterized as either a primary disease of the vasculature or as a secondary vascular disorder that developed in association with various cardiopulmonary diseases. The World Health Organization has identified 5 classes of diseases that can result in PH. In the recent classification, 2 Group 1 (pulmonary arterial hypertension) was ascribed to primary diseases of the vasculature, including idiopathic, familial, toxininduced, congenital heart disease, and veno-occlusive disease, among others. Group 2 (pulmonary venous hypertension) was related to left-sided heart disease and chronic increases in left atrial pressure. Group 3 PH occurred with lung diseases or hypoxemia, including obstructive pulmonary disease, interstitial lung disease, alveolar hypoventilation, and sleep apnea.
Group 4 comprised chronic thromboembolic PH, and group 5 included systemic and other disorders. Precise mechanistic explanations for the development of vascular remodeling in these disorders remain unclear, 2 with imbalance of endothelium-derived mediators, intermittent or sustained hypoxia-induced proliferation and inflammation, and epigenetic factors believed to play roles. [3][4][5][6][7][8] In veterinary medicine, group 1 disorders (pulmonary arteriopathy, capillary hemangiomatosis, and veno-occlusive disease) are rare and exclusively identified at necropsy, [9][10][11] although some of these disorders might be suspected clinically. The second classification includes animals with PH associated with left-sided heart disease, which is often the most commonly identified cause in dogs, [12][13][14][15] although the group of animals under investigation is likely important in determining the commonality of the underlying disease process associated with PH. Groups 3 through 5 also have been commonly described in veterinary studies of PH, 12,13 although definitive diagnosis of the underlying disease process related to PH for dogs in these groups is challenging because of the need for invasive testing including histopathology to confirm many of these disease processes. For example, pulmonary thromboembolism remains difficult to diagnose antemortem and can be found in dogs with normal D-dimer concentrations 16 ; tracheobronchomalacia (TBM) can require radiographic, fluoroscopic, and bronchoscopic assessment 17 ; and, pulmonary fibrosis requires histopathologic documentation, 18 although high-resolution computed tomography increasingly is used in the diagnosis. 19 Clinically, PH has been reported in a number of these diseases with diagnosis based on clinical suspicion, signalment, physical examination findings, and diagnostic imaging. 12,13,20 However, small numbers of dogs are generally available for inclusion in studies of clinical diseases in veterinary medicine, and diagnostic evaluation is limited by owner finances and consent. Dogs can suffer from multiple concomitant respiratory disorders as well as cardiac disease, further confusing categorization of specific disease processes that could result in PH. Finally, a miscellaneous category (group 5) including dogs with neoplasia, endocrinopathies, systemic inflammatory disease, and dogs with no obvious disease associated with PH has comprised 4%-38% of affected animals in various studies, 12,13,21 further complicating evaluation of the underlying diseases in dogs with PH.
Despite challenges in identifying the underlying disease processes associated with PH, the syndrome has been identified with increasing frequency because of advances in Doppler echocardiography, which provides a minimally invasive estimation of pulmonary artery pressure (PAP) by evaluation of the velocity of the tricuspid regurgitant jet. [12][13][14][15][20][21][22][23][24] Echocardiographic measures are impacted by operator and patient characteristics, 22 and it is important to remember that Doppler can provide only an approximation of PAP. Estimated PAP showed a significant and moderate correlation with results obtained during right heart catheterization in healthy Beagles, but variation increased at higher pressures. 23 It is likely that these variables also play a role in dogs with clinically-observed PH, although comparisons of Doppler echocardiography to more invasive methodology have not been performed in clinical patients to date.
Sildenafil is recommended for use only in group 1 human patients but it has been evaluated in both retrospective and prospective studies of dogs with PH associated with different cardiopulmonary diseases, with variable results reported. 21,22,24 Our prospective study sought to compare response to a standard 1-month regimen of sildenafil on the tricuspid regurgitation pressure gradient (TRPG) and quality of life (QOL) score in dogs with respiratory-associated causes of moderate to severe PH. In addition, we sought to refine our understanding of survival in dogs with PH related to respiratory causes. After diagnostic testing, sildenafil was prescribed at a target dosage of 2-4 mg/kg/day, with dosing adjusted as needed to accommodate available pill sizes. Owners were asked to observe dogs for hypotensive adverse effects potentially related to sildenafil such as weakness or lethargy after administration. The study protocol requested limited use of additional medications, but dogs were treated with antibiotics when infection was suspected or documented, corticosteroids when inflammatory or eosinophilic airway disease was diagnosed, and extended-release theophylline for potential assistance in management of TBM. After 1 month of sildenafil treatment, dogs returned to the hospital for repeat physical examination, cardiovascular and echocardiographic assessment, and completion of a post-treatment QOL questionnaire. Pre-and post-treatment questionnaires were submitted to the study manager, and results were not available to the primary clinician or the echocardiographer at the time of reevaluation.

| MATERIALS AND METHODS
Age, breed, weight, and sex were retrieved from the medical record. Presenting complaints and duration of clinical signs were retrieved for each dog. Survival was determined by contact with referring veterinarians and owners and was defined as the number of months beyond diagnosis to death or euthanasia for clinical signs related to cardiopulmonary disease. Dogs still alive or those euthanized for reasons unrelated to PH at the time of follow-up were rightcensored in the survival analysis.

| Statistical analysis
Statistical analysis was performed using a commercially available statistics program (GraphPad Prism 5.0f, San Diego, California). Visual inspection and D'Agostino and Pearson omnibus normality test were used to assess the distribution of results for age, body weight, duration of signs, QOL scores (before, after, and change), estimated TRPG (before, after, and change), and survival data. Normally-distributed data are presented as mean ± SD with inclusion of range when clinically relevant, and nonparametric data are presented as median with range. To account for missing data in dogs that died before 1-month follow-up, post-treatment QOL score was assigned the maximal value of 85, and the Wilcoxon signed rank test for paired data was used to compare pretreatment and posttreatment scores. In dogs that completed the 1-month treatment trial, QOL scores and TRPG pre-and post-sildenafil were compared using a paired t test. Tricuspid regurgitation pressure gradient and QOL scores pre-and post-sildenafil were compared between groups of dogs with and without a history of syncope. Correlation of TRPG with QOL score and survival was assessed using linear regression. Survival analysis was performed using a Kaplan-Meier estimate. For all analyses, significance was set at P < .05.

| RESULTS
Pulmonary hypertension associated with respiratory diseases was identified in 25  correlation was found between TRPG and QOL score (P = .29) at study entry. Tricuspid regurgitation pressure gradient and QOL scores did not differ between the groups of dogs with and without a history of syncope (P = .20 and .34; Table 1).
Multiple respiratory conditions were present in some dogs (Table 2) Despite worsened QOL scores, these dogs survived 2-9.5 months after diagnosis, with 3 dogs being managed using sildenafil alone. Note: Three dogs were considered to have respiratory-associated pulmonary hypertension but could not be assigned a specific etiology. These dogs had cough and syncope in association with normal left atrial size (n = 3), bronchointerstitial infiltrates (n = 2), and bronchiectasis (n = 1), and were treated with sildenafil alone. Abbreviation: BOAS, brachycephalic obstructive airway syndrome. a Two dogs in this category were ultimately diagnosed with pulmonary vascular disease on necropsy.
Tricuspid regurgitation pressure gradient measurements after 4 weeks of sildenafil treatment in 16 dogs available for follow-up were significantly lower than pretreatment (P = .0006) with a mean decrease of 24 ± 21 mm Hg (range, −26 to 61 mm Hg; Figure 2  Importantly, in that study, the estimated pressure gradient post-sildenafil did not differ from that of control dogs treated with placebo. This finding could suggest that factors other than sildenafil contributed to the decrease in estimated PA pressure in that study, and a similar effect might be responsible for some of the results observed in our study. It is notable however that the use of additional medications for management of respiratory conditions did not impact results in the dogs of our study. The decrease in estimated TRPG in dogs with respiratory-associated PH after 1 month of sildenafil treatment could provide support for an active or vasoconstrictive component in the pulmonary vasculature associated with hypoxia, as has been suggested previously. 28 Historically, dogs have been found to have a limited pulmonary vascular response to alveolar hypoxia, 29 but many clinical studies have reported increases in estimated TRPG in dogs with chronic pulmonary diseases, 12,13,21,22,28 despite a lack of blood gas data. Also, hypobaric hypoxia caused mild to moderate increases in systolic PA pressure (41-55 mm Hg) in healthy exercising sled dogs, 30 which suggests that global hypoxia triggers an increase in pulmonary vascular pressures even in normal dogs.
Not all studies have identified a decrease in TRPG after administration of sildenafil. A retrospective study of dogs with PH associated with cardiac or respiratory disease 22 reported that estimated TRPG pre-sildenafil did not differ from the result obtained 7-521 days posttreatment. Average PAP in that study 22 was in the moderate range (median, 62 mm Hg) compared to a majority of dogs having severe PH in the original study of sildenafil 21 and in our study. Therefore, perhaps the variable response to sildenafil in these different studies was partly related to the severity of PH in affected dogs, but inconsistency in Doppler estimates of TRPG must be considered. It is also possible that the decrease in TRPG detected in our study was related to the dosage of sildenafil administered (4.2 mg/kg/day), which seems to be higher than that used in previous studies. 21,22,24 This dosage was not only effective but also appeared to be safe because no adverse events were reported. Finally, genetic polymorphism in the phosphodiesterase 5A gene 31 could contribute to variable response of the TRPG to sildenafil in different groups of dogs.
Based on clinical data and lack of echocardiographic evidence of heart disease, a respiratory-associated cause of PH was assigned to all dogs in our study although, as indicated in Table 2 which might partly explain the high incidence of PH in these dogs. 20 More aggressive screening methods for dogs diagnosed with these conditions as well as with chronic bronchitis and TBM would allow determination of the overall prevalence of PH in these respiratory diseases, as has been done with mitral valve disease. 36 These investigations also could help elucidate the contribution of different respiratory disorders to PH, and the variable response to sildenafil.
As described in earlier studies, 12,21,22 syncope was a common finding in dogs in our study, with an overall occurrence of 64%. The percentage reported in our study was substantially higher than that previously described (32%) 22 and much higher than the 7% reported in dogs with PH associated with myxomatous mitral valve disease (MMVD). 14  Our study suffered from limitations common in studies involving clinical patients. Diagnoses in respiratory patients were not confirmed with airway sampling or histopathology in all cases, and subjective clinical variables were used to assign disease processes. Arterial blood gas evaluation rarely was performed in our study perhaps because of small patient size or concerns about stress to the patient, and definitive diagnoses sometimes were not sought because of concerns about invasive testing. Two dogs were clinically suspected to have interstitial fibrosis but on necropsy were determined to have pulmonary vascular and thromboembolic disorders. It is possible that other dogs in our study, as well as in other studies, were placed in the wrong classification for PH etiology. We did not assess caudal vena cava size 37 or consider the presence of right-sided heart failure in estimating PA pressure, 20 which has been recommended as a possible method to more closely approximate PA pressure. 23 Finally, echocardiography can overestimate or underestimate the pressure gradient as described earlier, and a recent study reported increased variance in echocardiographically derived PA pressure at higher pressures in comparison to direct catheterization studies. 23 Median PA pressure in that study was 42 mm Hg, which is substantially lower than the pressures determined here, and increased variation in estimated PA pressure might be expected in our study. Also, the ability to obtain maximal TR jet varies with the operator and compliance of the animal, 22 which despite best efforts potentially could hamper direct comparisons between measurements taken by different examiners on different days. Clearly, the difficulties in estimating PA pressure would impact all previous veterinary studies that have attempted to diagnose PH and assess response to sildenafil using echocardiographic measurements.
Finally, almost one third of the dogs enrolled in our study died or were euthanized before reevaluation, which limited follow-up assessment of TRPG and QOL scores.
Despite these limitations, our study confirmed improvement in QOL and an overall decrease in estimated PAP in dogs with respiratoryassociated causes of moderate to severe PH that were treated with sildenafil for 4 weeks. It also found a weak but significant effect of TRPG on survival, suggesting that the severity of PH as determined by Doppler echocardiography could be an important prognostic indicator in dogs with respiratory disease associated with PH, as it is in dogs with cardiacrelated PH. Our study also determined that PH can resolve if the underlying cause is identified and treated, as evidenced by the study dog that had resolution of PH and syncope with appropriate treatment for eosinophilic lung disease. With corticosteroid management and resolution of disease over 8-10 months, the TRPG decreased to normal and sildenafil ultimately was discontinued. We found that survival of dogs with PH treated with sildenafil often exceeded 6 months, although substantial mortality was encountered shortly after the diagnosis and before response to treatment could be assessed. Importantly, severe clinical signs related to syncope and severely increased TRPG did not influence response to sildenafil. Finally, the utility of QOL scores should be further evaluated in patients with respiratory disease, with or without PH.