Brachycephaly refers to canine breeds with “severe shortening of the muzzle, and therefore the underlying bones, and a more modest shortening and widening of the skull”. Conformational anomalies in brachycephalic dogs (BD) often are referred to as the brachycephalic syndrome (BS). The BS is characterized by increased upper airway resistance because of narrowed nostrils (58–85% of BD[3, 4]), an elongated and thickened soft palate (87–96%[3, 4]), everted laryngeal saccules (55–58%[3, 4]), and a hypoplastic trachea (46%[3-5]). Prominent nasopharyngeal turbinates have also been reported.
Studies have demonstrated laryngeal and pharyngeal dysfunction in BD.[7-9] However, little is known about the consequences of upper airway obstruction on the physiology of lower airways and lung parenchyma.[10, 11] Although De Lorenzi et al recently showed a high incidence of bronchial collapse in BD that most likely was because of chronic upper airway obstruction and the associated increase in negative intrathoracic pressures, the ability of BD to ventilate and oxygenate has been described clinically only in 5 French Bulldogs.
There is a similar paucity of literature on the effects of upper airway obstruction in the BS on the cardiovascular system. Research on increased upper airway resistance because of conformational anomalies (eg, narrowing of the air passage, obesity, prominent tongue, nasal congestion) in humans has indicated that these patients have an increased risk for the sleep apnea/hypopnea syndrome (SAHS). The hallmark of SAHS is complete upper airway collapse during sleep that can result in intermittent nocturnal hypoxemia and systemic hypertension.[14, 15] The incidence of hypertension in humans suffering from sleep apnea is high and SAHS is a known risk factor for development of hypertension. The mechanisms are unclear, but increased sympathetic activity, endothelial damage, decreased nitric oxide production, and negative intrathoracic pressures may play a role in the development of hypertension.[16, 17] Because English Bulldogs have been used as a spontaneous model for human SAHS, it would be reasonable to presume that other BD may similarly be at risk for sleep apnea. However, it is currently unknown if BD with BS are also at increased risk for hypertension.
The primary objective of this study was to evaluate ventilation and oxygenation in BD. The secondary objective was to measure arterial blood pressures (ABP) of BD. We hypothesized that BD would have higher arterial partial pressure of carbon dioxide (PaCO2), and ABP and lower arterial partial pressure of oxygen (PaO2) when compared to non-BD.
Materials and Methods
Owner consent was obtained. Patients were cared for according to institutional guidelines.
BD and Control Dogs
Pugs, Boston Terriers, French and English Bulldogs (Group B) that presented to our institution between March and May 2008 were included in the study if they were 0.5–8 years old, and if they were systemically healthy (ie, were brought in for wellness examination, routine vaccinations, or for the purpose of the study and had normal physical examination findings apart from potential BS). The control group (Group C) consisted of healthy (as evaluated by normal physical examination) meso- or dolicocephalic dogs (MDD). MDD are dogs with long or medium-sized muzzles. This group was comprised of dogs owned by veterinary students or obtained from laboratories (ie, Beagles not recently involved in any protocol). MDD were matched in age and weight with BD. Absence of overt stertor also was confirmed with the owners of the control dogs. Dogs presenting with evidence of illness, especially upper airway obstruction for MDD (eg, snoring, reverse sneezing, wheezing), were excluded from the study. Heights, thoracic circumferences, and 5-point body condition scores (BCS) were recorded.
To assess the daily impact of potential BS on BD, owners were asked about the subjective frequency of syncope, open-mouth breathing, and snoring experienced by their pets.
Arterial Blood Sampling and Analysis
If the dogs showed signs of stress or abnormal breathing on presentation (eg, tachypnea, hyperpnea), sampling was delayed and the animal was first given time to calm down, with its owner, in a quiet and cool room. The owner also was present in cases where the animal could not be restrained easily without inducing stress. A 1 mL polypropylene syringe was heparinized by aspirating 0.5 mL of liquid sodium heparin into the body of the syringe and then drawing the plunger back to the 1 mL mark. All air in the syringe then was expelled, followed by 10 repetitions of 1 mL of air being drawn into and expelled from the syringe.[19, 20] Arterial blood sampling was attempted from the dorsal pedal artery first. If this was unsuccessful, femoral arterial sampling then was performed. The skin was clipped and aseptically prepared. The artery was punctured with a 25 G needle, and blood was collected into the heparinized syringe. Samples were acquired in < 1 minute. If air was trapped in the syringe at the time of sampling, bubbles were immediately expelled from the body of the syringe with the bevel of the needle facing up. Samples were analyzed within 1–3 minutes of sampling. Temperature-corrected arterial blood gases (ABG) and hemoglobin concentrations were measured with commercially available cartridges.1 The bicarbonate concentration was derived by the machine with the formula [HCO3 −] = 0.0307 × PaCO2 × 10 (pH–6.129). The remaining blood was used to measure packed cell volume (PCV) and total protein concentration (TP) by refractometry. The alveolar-arterial oxygen gradient (A-a gradient) was calculated with the following formula: A-a gradient = (147 − PaCO2/0.8) − PaO2.
Blood Pressure Measurements
After a period of rest, 5 oscillometric systolic (SAP), mean (MAP), and diastolic (DAP) ABP measurements were performed.2 The cuff size was approximately 40% of the circumference of the leg above the tibiotarsal joint. The highest and lowest readings were rejected, and the average of the 3 remaining values then was calculated.
All statistical analyses were done by commercial software.3 Intergroup comparisons were performed after verifying homoscedasticity and homogeneity of the variances with a unilateral Student t-test for unpaired series. For A-a gradient data, 2 MDD had negative calculated values and were assigned the value zero for analysis purposes. Normal distribution was confirmed by Shapiro-Wilk testing. Differences in mean values for A-a gradients were assessed with a Student t-test. For results with variances that were not homogenous (ie, F test with P < .1), an Aspin-Welch or a Mann-Whitney test was used. A P value < .05 was considered significant. Results are presented as mean ± standard deviation unless otherwise indicated. Stepwise regression was used to identify factors associated with increased PaCO2 in BD.
Eleven BD met the inclusion criteria for Group B: 7 French and 4 English Bulldogs. Eleven MDD were in Group C: 6 Beagles, 2 mixed breed dogs, 1 Staffordshire Bull Terrier, 1 Parson Russell Terrier, and 1 Australian Cattle Dog. For dogs in Group B, age, weight, and BCS were 43 ± 19 months, 18 ± 7 kg, and 3.4/5 ± 0.5, respectively. For dogs in Group C, age, weight, and BCS were 38 ± 13 months, 15 ± 7 kg, and 3/5 ± 0.4, respectively. The difference in BCS between the 2 groups was statistically significant (P = .044). No statistical difference between the 2 groups was found in age and weight.
In Group B, the questionnaire indicated that 3 dogs (27%) had previously experienced syncopal episodes at least once in their lifetime, and 2 dogs (18%) breathed with open mouths most of the time. All owners reported a high frequency of snoring, even when the dogs were awake. One dog (9%) coughed frequently. Six dogs (55%) were diagnosed with the BS, which was surgically corrected in 2/6 dogs.
The temperatures, heart rates (HR), and respiratory rates (RR) in Group B were 38.5 ± 0.5°C, 118 ± 31 beats/min, and 47 ± 38 breaths/min, respectively. In Group C, they were 38.5 ± 0.3°C, 114 ± 31 beats/min, and 32 ± 16 breaths/min, respectively. There were no significant differences between the 2 groups in any of these parameters.
The dogs' heights and thoracic circumferences were 38.8 ± 8.1 cm and 60.6 ± 13 cm for Group B, and 40.5 ± 7.8 cm and 56.8 ± 9.1 cm for Group C, respectively. There was no statistical difference between the 2 groups.
Arterial Blood Gas Results
The findings are presented in Table 1. Arterial pH was not statistically different between the 2 groups (P = .34). Bicarbonate concentrations and PaCO2 were both significantly higher in Group B than in Group C (P = .021 and .019, respectively). The PaO2 and anion gap were significantly lower in Group B than in Group C (P = .017 and .012, respectively). No significant difference was found between the 2 groups for the A-a gradient (P = .07). Hemoglobin concentration was significantly higher in BD than in controls (P = .018). No statistically significant difference was found for the hemoglobin saturation value (P = .06).
Table 1. Arterial blood gas measurement results in brachycephalic dogs (group B, n = 11) compared to control, meso- or dolicocephalic dogs (group C, n = 11) (fraction of inspired oxygen 21%)
|Group B||7.40 ± 0.02||36.3 ± 4.6a||20.5 ± 2.3a||86.2 ± 15.9a||15 ± 12||95.2 ± 3.7||16.7 ± 1.3a||22.2 ± 6.3a|
|Group C||7.40 ± 0.04||32.7 ± 2.6||18.4 ± 2.2||100.2 ± 12.6||6 ± 11||97.1 ± 0.9||15.2 ± 1.9||27.3 ± 2.1|
Packed Cell Volume and Total Protein Concentration
The PCV was significantly higher in Group B than Group C (48 ± 4% versus 44 ± 5%, P = .026). No statistical difference was found between the 2 groups for TP (62 ± 8 g/L in Group B and 58 ± 5 g/L in Group C, P = .063).
Arterial Blood Pressure Measurements
Results are presented in Table 2. The SAP, MAP, and DAP were significantly higher in BD compared with controls (P = .013, .014, and .042, respectively).
Table 2. Mean systolic (SAP), mean (MAP), diastolic (DAP) arterial blood pressures in brachycephalic dogs (group B, n = 11) compared to control, meso- or dolicocephalic dogs (group C, n = 11)
|Group B||178 ± 25a||123 ± 17a||95 ± 19a|
|Group C||154 ± 22||108 ± 12||83 ± 11|
To investigate potential factors associated with high PaCO2, dogs in Group B were divided according to whether their PaCO2 was above or below the median PaCO2: BD with PaCO2 > 35 mmHg were assigned to the higher PaCO2 group (n = 5; range: 36–44 mmHg), where BD with PaCO2 ≤ 35 mmHg were assigned to the lower PaCO2 group (n = 6; range: 29–35 mmHg). All previous parameters were compared between the subgroups, and between each individual subgroup and the MDD group. Values are reported only when P < .05.
Animals in the higher PaCO2 group were significantly older than both the lower PaCO2 (P = .004) and control groups (P = .03). Mean age within the higher PaCO2 group was 58 ± 16 months, whereas for the lower PaCO2 and control groups ages were 30 ± 11 months and 38 ± 13 months, respectively. Moreover, stepwise regression showed that age, pH, and [HCO3 −] were associated with PaCO2 results in BD (P = .04, <.0001, and <.0001, respectively). The BCS also was higher in the higher PaCO2 group than in Group C (3.6 ± 0.5/5 versus 3 ± 0.4/5 in Group C, P = .018). When compared with Group C, the lower PaCO2 group had significantly higher SAP, DAP, and MAP (P = .031, .020, and .023, respectively), whereas the higher PaCO2 group only had higher SAP and MAP (P = .034 and .044, respectively; see Table 3). No statistical difference was found between the 2 subgroups of BD for ABP.
Table 3. Comparison of mean systolic (SAP), mean (MAP), and diastolic (DAP) blood pressures between brachycephalic dogs with higher PaCO2 (PaCO2 > 35 mmHg, n = 5), lower PaCO2 (PaCO2 ≤ 35 mmHg, n = 6), and control, meso- or dolicocephalic dogs (n = 11)
|Lower PaCO2 group||175.7 ± 21.7a||126.0 ± 21.7a||100.5 ± 20.9a|
|Higher PaCO2 group||180.0 ± 31.1a||120.0 ± 10.9a||89.0 ± 17.0|
The comparison between the 2 subgroups of Group B also indicated no significant difference in RR (72 ± 45 breaths/min versus 27 ± 9 breaths/min, respectively, P = .126). The HR was lower in the dogs with higher PaCO2 (93 ± 24 beats/min versus 138 ± 20 beats/min, respectively, P = .004).
Arterial blood gas results are summarized in Table 4. No statistical difference was found between dogs with lower PaCO2 and MDD. The higher PaCO2 group had significantly higher [HCO3 −] (P = .002) and lower PaO2 and anion gap (P = .004 and .03, respectively) when compared with the lower PaCO2 group. Compared with controls, the higher PaCO2 group had significantly higher [HCO3 −] and hemoglobin concentration (P = .001 and .041, respectively), and lower PaO2 and anion gap (P = .003 and .001, respectively).
Table 4. Comparison of arterial blood gas results between brachycephalic dogs with higher PaCO2 (PaCO2 > 35 mmHg, n = 5), lower PaCO2 (PaCO2 ≤ 35 mmHg, n = 6), and control, meso- or dolicocephalic dogs (n = 11) (fraction of inspired oxygen 21%)
|Lower PaCO2 group||7.41 ± 0.03||33.0 ± 2.1b||18.9 ± 1.5||94.0 ± 12.6b||96.8 ± 1.2||16.5 ± 1.1||25.4 ± 3.9b|
|Higher PaCO2 group||7.40 ± 0.02||40.2 ± 3.3a||22.4 ± 1.3ab||76.8 ± 15.2a||93.2 ± 4.8||17.0 ± 1.6a||18.4 ± 6.7a|
The main findings of this study were that our population of BD had significantly lower PaO2, higher PCV, PaCO2, and ABP when compared with MDD.
Upper airway obstruction in certain BD is a complex mechanism. The pathogenesis of upper airway dysfunction in BD may be similar to that seen in humans suffering from SAHS. Also known as obstructive sleep apnea (OSA) in its end stages, SAHS is a common disease in human patients.[22, 23] It has been correlated with a chronic hypoxemia and stimulation of the peripheral chemoreflex (PCR) because of complete upper airway obstruction during sleep. Hypoxemia, and to a much lesser extent, hypercapnia are the main stimuli for activation of peripheral chemoreceptors in the carotid body and aortic arch.[24, 25] Upon activation, peripheral chemoreceptors initiate the PCR, in which cranial nerves X (vagus nerve) and XI (accessory nerve) conduct impulses to the higher respiratory center in the medulla oblongata of the brainstem. There, the information is primarily integrated in the nucleus tractus solitarius, which in turn functions to coordinate input from other reflexes also governing cardiac and respiratory functions (ie, the baropulmonary stretch and central chemoreflex). Once this information has been processed in the brainstem, a cardio-respiratory response is elaborated: in the presence of hypoxemia and hypercapnia, the PCR will stimulate an increase in minute ventilation, sympathoexcitatory vascular tone, and vagal tone to the heart, signaling bradycardia and increased blood pressure. At the same time, the increased minute ventilation (tachypnea and hyperpnea) will stimulate the baropulmonary stretch receptors, resulting in a HR higher than normal.[24, 25] In vivo, changes in HR also are governed by other mechanisms and therefore may not be dictated only by the PCR. The English Bulldog has been well-established as a spontaneous model for SAHS. Predisposing factors identified in humans include narrowing of the air passage, obesity, and nasal congestion, and similar conformational abnormalities have been described in English Bulldogs. Studies have further demonstrated that the airway dilator muscles of English Bulldogs can suffer from decreased tone that will progressively permit closure of the upper airway. This has been documented on histology and magnetic resonance imaging.[8, 9] Muscle fibers are progressively converted from type a to type b (the latter constricting more strongly, but also more transiently, than the former). Once hypoxemia is sensed and the PCR is triggered, the animal abruptly awakens and the muscles involved in upper airway opening rapidly and strongly contract. Chronically, this abrupt change in muscular tone leads to inflammation, edema, and fibrosis.[8, 9] This further diminishes the ability of the muscles to maintain upper airway patency, leading to laryngeal collapse. Thus, with time BD are at risk for worsening upper airway obstruction.[7-9]
Only a few studies have addressed the link between upper airway obstruction and alveolar gas exchange.[10, 11] The absence of a statistically significant difference in A-a gradient between the 2 groups is contradictory with the significantly lower PaO2 in BD. Concluding that A-a gradient was not different between the 2 groups should be done cautiously because the small sample size increases risk of Type II error. The difference in PaO2 between the 2 groups with no difference in A-a gradient also could be evidence that hypoventilation is driving lower PaO2 in our BD population. Evidence of chronic mild subnormal PaO2 in BD previously has been reported: a causality between chronic hypoxemia and hyperplasia followed by neoplastic transformation of the chemoreceptors in dogs and cattle have been postulated. Correspondingly, BD (especially Boxers and Boston Terriers) are known to be prone to chemodectomas. We therefore suggest that the lower PaO2 might be associated with the upper airway obstruction observed in the BS. A recent study in humans showed that upper airway and bronchial inflammation, as well as endothelial dysfunction, was correlated with the severity of the upper airway obstruction. Although pleural pressures of BD have never been reported, a study showed that BD suffering from airway obstruction had significantly shorter expiratory-to-inspiratory times ratio compared with a historical control group, further supporting the hypothesis of high resistance to flow on inspiration. As supported by previous reports of bronchial collapse in BD, it can therefore be postulated that the lower airways of BD are exposed to recurrent and highly negative pressures because of the effort to overcome upper airway resistance (up to −65 mmHg in human patients with SAHS). Recurrent barotrauma has the potential to lead to inflammation, edema formation, and potentially fibrosis of the lower airways and pulmonary parenchyma. This may contribute to the lower PaO2 observed in our BD. To verify this hypothesis, trans esophageal pressure measurements would be needed to assess the transmural pressure gradient.
Interestingly, we showed that BD in our population had lower PaO2 but also higher hemoglobin concentrations and PCV than controls. This might be a compensatory mechanism to maintain normal arterial content of oxygen. The PaO2 is a minor component of the arterial content of oxygen but is a major regulated factor in the body.[33, 34] Chronic hypoxia is a strong stimulus for red cell production. Obstructive sleep apnea might have been present in some of our subjects and those dogs might experience episodes of hypoxemia at night, which might be sufficient to stimulate erythropoietin (EPO) production as discussed in human medicine. Serum concentration of EPO also has been shown to decrease with treatment of sleep apnea with continuous positive airway pressure (CPAP). Also, the renin-angiotensin-aldosterone system has been shown to be activated in humans with the SAHS, and renin can enhance red blood cell production.[37-41] In our study, young dogs did not show a significant difference when compared with controls, whereas older dogs had significantly higher hemoglobin concentrations. This could represent a worsening in their gas exchange disorder as they age.
We also found that BD had higher PaCO2 compared with MDD. This could be explained by probable upper airway obstruction. Increased PaCO2 also may trigger the central chemoreflex, which should have increased minute ventilation. Several hypotheses could account for the persistence of slightly higher PaCO2. First, PaCO2 may not have been high enough to trigger the central chemoreceptors and induce an increase in respiratory rate or tidal volume high enough to decrease PaCO2. Second, dogs in our study may have undergone habituation, and the PaCO2 threshold may have been reset to a higher value. Finally, respiratory muscles fatigue in conjunction with increased workload (eg, increased upper airway resistance, decreased pulmonary compliance), which also might be an explanation for this finding.
The increase in bicarbonate concentration may be a metabolic compensatory mechanism that allows BD to maintain normal arterial pH. Age appeared to be an important factor in the evaluation of ABG in the study population. Older dogs were more prone to higher PaCO2 with no significant difference in RR suggesting tidal volume reductions because of age-related changes such as decreased compliance or increased resistance, although other factors also may play a role.
Subgrouping results suggest that age could be a major contributor to the effect of BS on ABG. The BD with higher PaCO2 were significantly older than those with lower PaCO2. Older BD also had significantly lower PaO2 than MDD and younger BD. No difference was found between young BD and the MDD. This could represent progression of dysfunction of the pulmonary parenchyma in the older BD. Overall, BD in the present study did not show a significant increase in HR or RR in face of subnormal PaO2, but did demonstrate an increase in ABP. The sensitivity of the PCR to PaO2 is referred to as peripheral chemosensitivity. Most of the literature regarding peripheral chemosensitivity is based on observations made either from humans living at high altitude or from animal models (sheep,[43, 44] goats,[45, 46] and cats[43, 47-49]). A review of the literature showed that short exposure to hypoxemia would provoke an increased response of the PCR. In chronic hypoxemic situations, the organism can undergo a phenomenon called “habituation” or “desensitization to hypoxemia”. After chronic exposure to hypoxemia, the peripheral chemoreceptor develops decreased sensitivity and response to low PaO2.[50, 51] It is controversial whether or not changes in PaO2 or PaCO2 are the main stimuli for breathing. Both PaO2 and PaCO2 are implicated in the control of breathing and the level of one can alter the body's response to changes in level of the other. Although, changes in PaCO2 are the most significant driving force for ventilatory changes in air breathers, a decreased PaO2 can trigger a ventilatory response as well.
This study also demonstrated that ABP was significantly higher in BD than MDD. Mechanisms such as activation of the renin-angiotensin-aldosterone system, arterial wall stiffening,[56, 57] oxidative damage, endothelial dysfunction, and systemic inflammation have been shown to contribute to hypertension in SAHS patients. Reportedly, because of chronic exposure to catecholamines, SAHS patients are prone to systemic and pulmonary hypertension that could, if present in affected BD, contribute to oxygenation impairment as well.
The BD had slightly higher BCS than controls. In humans, obesity is a major risk factor for OSA and even has been associated with the obesity hypoventilation syndrome, which encompasses many perturbations in ventilatory drive. The effect of BCS has been mentioned as a potential aggravating factor for the BS, but no direct correlation could be made between the BCS and the severity of upper airway obstruction.[60, 61] Torrez et al suggested the potential benefit of weight loss for non-upper airway related reasons, which also has been discussed in human medicine.
This study provides pilot data about potential pulmonary and cardiovascular consequences of the BS. Advanced experimental procedures, such as pulmonary function testing, endoscopy, or histological evaluation of lung biopsies, would be necessary in the future to confirm the presence and quantify the severity of the BS in experimental subjects and also to further describe mechanisms of disease.
This study has several limitations. First, only 2 breeds of BD were represented, and the results may not directly be extrapolated to all other BD. Second, the study population was heterogeneous in terms of age, severity, and history of previous surgical correction of the BS. There are no data to address the effect of those factors on the parameters studied here. It must be noted, however, that this heterogeneity actually may provide an accurate representation of the heterogeneous general BD population. Also, we used a 5-point BCS, which was the standard of practice at our institution. The 9-point BCS is more commonly used in North America and has been shown to correlate best with actual body fat mass.
Also, it was unexpected to have negative values for A-a gradients. Nonetheless, negative values of A-a gradient previously have been reported.[63-65] This might be because of summations of errors inherent to the A-a gradient formula (variations in respiratory quotient, barometric pressure, water vapor pressure, pulmonary capillary temperature) or to measurement variation in patients with low A-a gradient. The results of the sub-group analysis must be interpreted with caution because the small sample size increases the risk for a type II error. Finally, other causes of venous admixture (eg, pneumonia or right-to-left shunting) were ruled out solely based on an unremarkable medical history and physical examination. Although it is unlikely that such conditions would have been present without associated clinical signs, these conditions cannot be definitively ruled out without further diagnostic evaluation.
For the first time, our study examined ABG in BD and provided evidence that BD have lower PaO2, higher PaCO2, and higher ABP when compared with MDD. Similar to the pathophysiology of exercise-induced pulmonary hemorrhage in horses and post-obstructive pulmonary edema, oxygenation impairment may be because of recurrent barotrauma to overcome upper airway resistance to airflow, which may lead to alveolar tearing, capillary damage, endothelial dysfunction, and interstitium damage.[66-68] We also found that BD had higher PCV when compared with MDD, potentially in response to their mild chronic lower PaO2. The results of this study may be useful in the clinical management of BD and patients suffering from the BS, because PaO2 results lower than the standard reference range may be within the reference range for individual BD. Additional studies will be necessary to determine the effect of breed, age, and BCS on variations in ABG among individual BD, and to assess the effect of surgical correction of the BS on ABG and ABP. Surgery has been shown to improve gastrointestinal signs induced by the BS, and it also may affect ABG and ABP. Because increased upper airway resistance may influence lower airway physiology and lung parenchyma, future studies may provide insight into the efficacy of surgery in older patients with the BS.