Abstract
Twenty-five cases of feline bronchial disease were identified retrospectively. The criteria for inclusion were consistent clinical signs or histopathology and no other identifiable aetiology. Patient records were analysed to determine historical, clinical, clinicopathologic and radiographic features. The main presenting complaints were coughing and dyspnoea. The most common physical finding was dyspnoea. The majority of radiographs had a bronchial pattern either as the sole change or as a component of a mixed pattern. Bronchoalveolar lavage cytology was neutrophilic or eosinophilic in the majority of cats. There was no association between age, breed, sex, clinical signs, bronchoalveolar lavage cytology or radiographic severity and disease severity.
Introduction
Bronchopulmonary disease in cats may be caused by infectious agents (viruses, bacteria, fungi, parasites), cardiac disease, neoplasia, trauma or toxins. Many cases, however, have inflammatory airway disease with no identifiable aetiology. These cases have been termed feline asthma, feline asthma syndrome, feline bronchitis, allergic bronchitis and feline bronchial disease (FBD) (Corcoran et al., 1995; Johnson, 1997; Moise et al., 1989). Some authors attempt to categorise FBD further as chronic bronchitis or feline asthma (Padrid, 2000). In humans, however, the term chronic bronchitis is no longer recommended (Haslett et al., 2002; Holgate and Frew, 2002) and examination of the definitions of both human asthma and chronic obstructive pulmonary disease (which encompasses chronic bronchitis and emphysema) would suggest that neither term is appropriate in cats. Pulmonary function testing and immunohistochemistry innaturally occurring FBD are not routine and the pathophysiology of FBD is unknown. To complicate matters further, other bronchopulmonary diseases such as bacterial infections (especially mycoplasmal infections) and parasitic infections (heartworm and lungworm infections) are frequently not rigidly excluded before making a diagnosis of FBD.
Previous studies have reported a wide range of clinical, clinicopathologic and radiographicfeatures for cats with FBD and failed to find any defining or prognostic criteria, other than breed: Siamese cats appear to be over-represented and have more severe progressive disease (Dye et al., 1996; Moise et al., 1989).
In previous studies of FBD, bacterial lower respiratory tract infections (LRTIs) were not necessarily excluded and mycoplasmal cultures ofbronchoalveolar lavage (BAL) specimens were not routine despite the fact that 44% of mycoplasmal cultures performed were positive in the study by Moise and others (1989). Dye and others (1996) eliminated the possibility of bacterial lower respiratory tract infections (LRTIs), other than mycoplasmosis, with a combination of BAL culture and response to therapy. In three other studies, bacterial cultures of BAL specimens were not performed in all cases and positive cultures did not exclude cats from analysis (Corcoran et al., 1995; Moise et al., 1989; Simard and Dube, 1992). Healthy lower airways are not sterile (Dye et al., 1996; Padrid et al., 1991) and the significance of the positive cultures in these studies is unknown as quantitative culture results and individual case histories were not provided. Cats were excluded from the study by Corcoran and others (1995) if their disease responded completely to antibiotic therapy but three cats had disease controlled with intermittent antibiotic therapy, which may or may not have been responsible for resolution of signs.
Parasitic LRTIs were also not necessarily excluded from previous studies. Dye and others (1996) performed routine faecal flotation and heartworm antigen tests on the majority of their cats. However, one cat with presumed lower respiratory tract (LRT) parasitism was not excluded. Parasitological investigations were not reported in the other studies (Corcoran et al., 1995; Moise et al., 1989; Simard and Dube, 1992).
In a retrospective study of unguided BAL cytology and microbiology in cats (Foster et al., 2004a), 25 cats in which bacterial LRTI had been excluded, were identified as having FBD on the basis of consistent ongoing clinical signs or histopathology and no other identifiable aetiology. Cases for which follow-up information was not available and cats that appeared to have self-limiting disease were excluded. By narrowing the selection criteria compared to previous studies, it was hoped that more definitive diagnostic and prognostic features of FBD would be identified without the need for pulmonary function testing.
Materials and methods
Twenty-five cats with FBD were identified in a previous retrospective study of unguided BAL cytology and microbiology performed between 1995 and 2000 at the University Veterinary Centre, Sydney (Foster et al., 2004a). The criteria forinclusion as FBD were lack of an identifiable cause (radiology, cytology, culture and/or response to empirical therapy) and either post-mortem histopathology or ongoing clinical signs.
Signalment was analysed for the 25 cases. Statistical comparisons for data on sex and breed were performed using SPSS for Windows version 11.0.0. Breeds were classified as domestic, Siamese, Burmese, purebred shorthair (other than Burmese and Siamese) and purebred longhair for comparison with the hospital population of cats at the time of the study (Gabor et al., 1998).
Historical and clinical data were analysed. Clinical data were recorded as (i) peracute, if signs had been present for 72 h or less, (ii) acute, if signs had been present for less than a month or (iii) chronic, if signs had been present for a month or longer. Tachypnoea was defined as a respiratory rate greater than or equal to 60 breaths per min as effects of temperament, transport and ambient temperature could not be assessed retrospectively. Date of presentation and, if known, date of onset of clinical signs were recorded. The seasons in Sydney are defined as summer (December to February), autumn (March to May), winter (June to August) and spring (September to November).
Haematology, serum biochemistry and serological test results from commercial ELISA or immunomigration kits for feline immunodeficiency virus (FIV) antibodies, feline leukaemia virus (FeLV) antigen and heartworm antigen were recorded if available. Bronchoalveolar lavage cytology and microbiology were analysed (Foster et al., 2004a). Bronchoalveolar lavage cytology was performed on squash preparations of mucus from each lavage specimen. Inflammatory cell numbers were subjectively assessed as many, moderate, few or none. Differential cell counts were classified by the predominant cell type (50% or more of the total) as neutrophilic, histiocytic, lymphocytic or eosinophilic. If no cell type predominated, the cytology was classified as mixed. Epithelial cells were not considered in the calculation of relative cell counts following the recommendations of the American Thoracic Society for reporting cell counts in humans (American Thoracic Society, 1990).
All available radiographs were reviewed retrospectively by a specialist radiologist who was blinded as to the clinical diagnosis (GA). Bronchial signs were defined as mild (first generation ofbronchi visible), moderate (second generation visible) and severe (third generation visible). Alveolar patterns were defined as mild (isolated fluffy infiltrates), moderate (well defined with air bronchograms) and severe (lobar sign). Reticular interstitial patterns were recorded as interstitial and classed as mild (mild interstitial framework visible but could be bronchial pattern), moderate (interstitial framework distinguishable from a bronchial pattern) and severe (undisputed reticular interstitial pattern). Nodular interstitial patterns wererecorded as nodular and classed as mild (less than five nodules, 1–2 mm in diameter), moderate (6–10 nodules, 2–5 mm in diameter) and severe (more than 10 nodules, greater than 5 mm in diameter). Overall radiographic severity, for the purposes of statistical analysis, was defined by the severity of the predominant pattern.
Therapeutic agents, response to therapy and long term follow-up in all survivors, were recorded. Disease severity was classified as mild (occasional signs requiring intermittent or no therapy), moderate (continuous treatment with bronchodilators or corticosteroid required) or severe (death due to disease or continuous treatment with multiple drugs required).
The association of disease severity with continuous data (age) was analysed with an analysis of variance. The association of disease severity (mild versus moderate and severe) with discrete data (breed, gender, clinical signs, BAL cytology, radiographic severity) was analysed with Fisher's exact test or Pearson's chi-square test of independence. Statistical data was analysed using SPSS for Windows version 11.0.0.
Results
Signalment, history, physical findings
Twenty-five cases of FBD were identified from a previous retrospective study (Foster et al., 2004a). The cats were two to 15 years old (median 9 years). There were 13 desexed females and 12 desexed male cats. The breeds for these cats were domestic (13), Siamese/Oriental/SiRex/Siamese cross (4), Burmese (3), British short-hair (3), Persian (1) and Australian Mist (1). Purebred cats were no more likely to have FBD than crossbred cats (P=0.14), however, purebred shorthair cats (excluding Burmese and Siamese) were four times more likely to be affected than domestic cats (95% confidence intervals 1.3, 12.6).
Historical findings included coughing (19), dyspnoea (10), abnormal respiratory noises (4), pruritic skin disease (4), upper respiratory tract (URT) signs (3), inappetence (3), ocular discharge (2), weight loss (2), tachypnoea (1), vomiting (1) and lethargy (1). However, the main presenting complaint was coughing in 19 cats (acute in one, chronic in 16, not recorded in two), dyspnoea in four cats (acute in two, chronic in two), lethargy in one cat and chronic URT disease in one cat.
The most common physical examination abnormalities were dyspnoea (10), periodontal disease (9), increased breath sounds (7), wheezing onauscultation (6), tracheal hypersensitivity (4), tachypnoea (4), cardiac murmur (3), crackles on auscultation (3) and obesity (3).
The seasons of presentation were summer (5), autumn (6), winter (10) and spring (2). Two cases were excluded from the seasonal analysis as presenting date was considered irrelevant when analysing seasonal trends of FBD: one was referred because of concerns about corticosteroid-induced hyperglycaemia and one had disease that was probably induced by potassium bromide therapy. Date of onset of clinical signs was known in nine cases: summer (5), autumn (2, both in March), winter (2). The initial signs for the cats in which disease onset was from December to March were dyspnoea (5), coughing (1) and coughing and dyspnoea (1). Coughing was the initial sign in the two cats that developed FBD in winter.
Of the 25 cats, seven had had no previous treatment and eight had been treated with doxycycline: four had had minimal or no response, two had had a good response initially and response in two could not be assessed as terbutaline or prednisolone had been prescribed concurrently. Eight cats had been treated with corticosteroids, seven with at least some improvement.
haematology, biochemistry and serology
Complete haematology was only available in four cases and complete biochemistry, in three cases. Packed cell volume was normal in 9/10 cases, increased in one. Total protein was increased in 5/10 cases with normal packed cell volumes and marginally decreased in 1/10 cases. The leucogram was normal in the four cats, for which results were available. Serological testing for FIV antibodies in five cats, heartworm antigen in six, and FeLV antigen in two, proved negative. One cat (previously diagnosed with mycoplasmal pneumonia) was positive for FIV antibody.
Radiology
Radiographic findings are detailed in Table 1. Radiographs were available for review in 22 cases. Eight cases had a bronchial pattern (mild in four, moderate in one and severe in three). One case had a moderate bronchial pattern and a lobar sign in the right middle lung lobe. Three cases had bronchoalveolar patterns and two cases had bronchointerstitial patterns. There was one case each of an alveolar pattern and an interstitial pattern. The remaining six cases had mixed patterns.
Disease severity, bronchoalveolar lavage cytology and thoracic radiographic details
Abbreviations: PA=pulmonary artery, VHS=vertebral heart scale (range 6.9–8.1; Litster and Buchanan, 2000).
Pulmonary overinflation was only noted in three cats. In two cats there was significant emphysema noted at necropsy shortly after the radiographs were performed and in neither case was this obvious on radiographs.
BAL cytology and microbiology
Twenty-seven BALs were performed in the 25 cats. Of the initial BALs, inflammatory cell numbers were recorded as many (22), moderate (2) or not recorded (1). Mucus was present in all specimens. Inflammation was neutrophilic in 12, eosinophilic in eight (all with at least 60% eosinophils), histiocytic in three and mixed in two (see Table 1). Infectious agents were not observed cytologically in any specimen.
A second BAL was performed in two cats. One cat (Case 25) had a mixed inflammatory response (macrophages, neutrophils and lymphocytes in equal numbers) on the first occasion and had a neutrophilic cytology on the second, 6 months later. The other cat (Case 11) initially had eosinophilic BAL cytology. The second BAL performed 5 months later (9 days prior to death) was neutrophilic but bacteria were observed cytologically and Pasteurella sp. and mixed anaerobes were cultured. On the basis of these findings and post-mortem histopathology, it was considered to have bacterial pneumonia superimposed on FBD.
Thirteen samples were cultured aerobically and there was no growth in four, light growth of mixed contaminants in eight and heavy mixed contaminants in one. Of the twelve cats that did not have BAL culture performed, ten had empirical antimicrobial therapy for mycoplasmas with minimal or no response. The other two cats had had chronic, continuous signs of respiratory disease, which required prednisolone, with or without bronchodilators, prior to BAL. Bacteria were not observed in either BAL and both cats had eosinophilic cytology which was not consistent with cytology from mycoplasmal LRTIs (Foster et al., 2004b).
Therapy and outcome
The outcome in these cases was influenced by the inclusion criteria which included a requirement for either continuous or intermittent signs of FBD and/or histopathology. Two cats died within a week of presentation: one cat (Case 17) was euthanased due to lack of response to bronchodilator and immunosuppressive therapy (after chronic corticosteroid therapy) and one (Case 21) died whilst hospitalised for treatment of FBD and a severe concurrent viral URT infection. Histopathology confirmed severe FBD in both cases. Twenty-three cats required oral prednisolone, bronchodilators (terbutaline, sustained-release theophylline or aminophylline) or both for improvement of clinical signs. Other adjuncts to therapy in some cases included cyproheptadine, dexchlorpheniramine and chlorambucil. Inhaled corticosteroids and bronchodilators were used in two cats after the year 2000 but only in addition to oral therapy.
Long term follow-up (5 months to 5 years, median 16 months) was available for the twenty-three surviving cats. Seventeen cats needed continuous bronchodilator and/or glucocorticoid therapy. Three cats required intermittent medication. Three cats were eventually weaned off medication: in two cases, the owners preferred mild ongoing clinical signs to the side-effects of the medications (both classified as mild disease); in one case, continuous treatment was required but treatment was regarded as unsatisfactory and withdrawn despite quite severe clinical signs (classified as moderately severe disease). Overall, 15 cases were classified as severe, five as moderately severe and five as mild (see Table 1).
There was no effect of age (P=0.64) or sex (P=0.43) on disease severity. There was also no effect of purebred versus crossbred on disease severity (P=0.11). Siamese were not more likely than non-Siamese cats to have severe disease (P=0.53). Of the four cats in the Siamese group, two were classified as severe, one as moderate and one, the Si-Rex, as mild. Clinical signs of chronic coughing (P=0.36) or dyspnoea (P=0.48) were not associated with disease severity. Radiographic severity was also not associated with disease severity (P=0.49).
Twenty percent of severe cases had eosinophilic cytology and 60% had neutrophilic cytology but overall, there were no significant correlation between cytology and disease severity (P=0.41). When the cytology was examined separately there were also no significant differences between groups.
At the time of writing, nine of the 23 survivor cats were known to have died or been euthanased. Of these, five died or were euthanased (5 to 31 months later) because of respiratory disease (4) or complications of therapy (1; diabetic ketoacidosis). Complications of therapy may have resulted in the euthanasia of one other case (tooth root abscess, mandibular osteomyelitis and possible pleural effusion reported by referring veterinarian). The other three cats died from unrelated diseases. A necropsy was performed on one of the cases euthanased due to respiratory disease.
Discussion
There have been three detailed clinical studies of FBD in cats: two from North America and one from Great Britain (Corcoran et al., 1995; Dye et al., 1996; Moise et al., 1989). These studies have demonstrated that FBD is not a homogenous syndrome and that clinical, clinicopathologic, radiographic features are variable as are pulmonary function tests and response to therapy.
This is the first study of FBD in Australian cats. The cats in this study were older than those in previous studies (Corcoran et al., 1995; Dye et al., 1996; Moise et al., 1989) with a median of 9 years. Many of these cases were referrals and they may not have been referred until disease control proved difficult. In addition, the rigid selection criteria which excluded mild and self-limiting cases may have influenced the age of presentation if disease was less severe when the cats were younger. However, Dye and others (1996) demonstrated that there were no marked differences in age among the cats with clinically mild, moderate or severe disease and airway reactivity, in healthy cats at least, has been shown to decline with age (Hirt et al., 2003).
Female (Moise et al., 1989) and male (Dyeet al., 1996) predispositions have previously been reported, however, in this study, as in that by Corcoran and others (1995), there was no sex predisposition. Female cats have been reported to be more severely affected (Dye et al., 1996) but there was no effect of sex on severity in this study. Siamese cats were reported as being over-represented in one study (Moise et al., 1989) and having more chronic or progressive disease (Dyeet al., 1996; Moise et al., 1989). Siamese cats were not over-represented or more severely affected in this study, however, small numbers of cases may have limited the statistical power of analysis.
Purebred shorthair cats (other than Burmese and Siamese) were four times more likely than domestic cats to have FBD in this study and four times more likely to have LRTIs than domestic cats in another study (Foster et al., 2004b). The LRTIs identified in this group of cats were all due to mycoplasmas (Foster et al., 2004b) and it is tempting to speculate that there may be an association between FBD and mycoplasmal LRTI in these cats. FBD may predispose to mycoplasmal LRTIs, however, it is equally possible that mycoplasmal LRTIs may cause airway hyper-reactivity and induce FBD (Fosteret al., 2004b).
The historical complaints and physical findings were consistent with those of previous studies(Corcoran et al., 1995; Dye et al., 1996; Moiseet al., 1989; Simard and Dube, 1992) with coughing the most common clinical sign in all studies. Only one previous study has analysed season of presentation (Moise et al., 1989) and no seasonal trends were identified. Dates of presentation in our study were considered of limited value as the same information for general feline hospital admissions was unknown. Date of onset of clinical signs was rarely known. However, in the cases in which accurate information was available, it appeared that the onset of signs was commonly between December and March and dyspnoea appeared to be the predominant feature. As coughing was the main presenting complaint identified in this study, the association between date of onset and dyspnoea is interesting. In addition, signs in one cat recurred every summer and treatment was only required in summer. It would appear that FBD, heralded by dyspnoea may be induced by seasonal environmental factors in Sydney. However, the number of cases for which the information was known was small and it may be that onset of clinical signs was more noticeable and well defined for dyspnoeic cats in comparison with coughing cats.
Cats with FBD show variable non-specific haematological abnormalities (Johnson, 1997). Changes in serum biochemistry and urinalysis are also non-specific (Johnson, 1997) and the positive predictive value of any serum biochemical abnormality other than hyperproteinaemia (Moise et al., 1989) would be low in the relatively healthy cats that typically present with FBD. Thus haematology and biochemistry were not performed routinely in these cats and there was insufficient data to draw conclusions about haematology and serum biochemistry in cats with FBD. Hyperproteinaemia was detected in 50% of the cats for which data was available. As albumin and globulin concentrations were not available, it is not known whether this was due to dehydration or gammaglobulinaemia. Hyperproteinaemia without dehydration was noted in a third of cats with FBD in a previous study and was attributed to immunologic responses (Moise et al., 1989) but hyperglobulinaemia was only detected in 8% of cats in another study (Dye et al., 1996).
Comparison of radiographic findings between this study and previous studies is difficult. A bronchial pattern was part of the inclusion criteria for one study (Moise et al., 1989), and radiographic findings were included as part of a disease severity score, and not discussed, in another (Dye et al., 1996). However, the variability of radiographicappearance and the lack of correlation between radiographic severity and disease severity would appear to be consistent with previous reports. Interestingly, pulmonary overinflation was uncommon and emphysema was not necessarily evident on radiographs.
BAL cytology was neutrophilic in 50% of cases and eosinophilic in 30%. BAL cytology in previous studies has varied according to method of collection and method of counting (with epithelial cells included in the differential cell counts in some studies) but neutrophilic BALs would appear to predominate in FBD. Neutrophilic BAL cytology is not specific for FBD whereas eosinophilic cytology (eosinophils comprising 50% or greater) in BALs is very suggestive of FBD (Foster et al., 2004a). In a retrospective study of BAL cytology during these years, FBD accounted for 73% of the eosinophilic samples (Foster et al., 2004a). The remaining eosinophilic cases were classified as inconclusive and could well have included cases of FBD.
Anecdotally, eosinophilic cases of FBD are regarded as being less refractory to therapy than neutrophilic cases. In this study, cytology did not predict disease severity. In the eight eosinophilic cases, three cats (37.5%) were classed as mild and five (62.5%) as moderate or severe. One of the severe cases (Case 11) died as a result of its disease. This would indicate that the range of therapeutic responses for cats with eosinophilic airway disease is similar to that of cats with neutrophilic or mixed inflammatory responses.
Potassium bromide was identified retrospectively as a likely cause of FBD in one of these cats (Case 11). Potassium bromide has been reported to induce potentially fatal FBD (Boothe et al., 2002; Wagner, 2001). In Case 11, signs of FBD occurred 3 weeks after initiating potassium bromide therapy for intractable seizures. The respiratory signs were refractory to therapy and the cat died as a result of its respiratory disease; severe emphysema and histological changes were identified in all lung lobes. Potassium bromide therapy would appear to be a potentially dangerous and not highly efficacious drug for seizure control in cats (Boothe et al., 2002).
Continuous therapy was required in 17 cats, with multiagent therapy required in 13. Death or euthanasia due to FBD or complications of therapy occurred in seven cats in this study at the time of writing. However, the inclusion criteria of ongoing signs of disease or death resulted in exclusion of some cats with mild or self-limiting disease. Whilst the inclusion criteria selected for more severe disease, the study highlights the morbidity and mortality that can occur with FBD. Most of these cats were diagnosed and treated prior to routine use of inhalant corticosteroid therapy and it may be that death due to complications of therapy would be less likely with inhalant therapy.
This is the first Australian study of FBD. It is also the first study of FBD in which bacterial (including mycoplasmal) LRTIs were excluded. Parasitic causes of bronchopulmonary disease are more difficult to exclude. Aelurostrongylus abstrusus and Eucoleus aerophilus can be diagnosed from BAL cytology but tracheal washing has been regarded as being less sensitive for A. abstrusus than faecal analysis by the Baermann technique. Larval numbers are relatively small, especially during the later stages of disease and may not be detected in a small amount of dilute BAL fluid (Pedersen, 1988). However, egg output decreases dramatically after 2–3 months and larvae are only excreted intermittently so even faecal analysis is insensitive and multiple (usually three) faecal samples have been recommended for analysis (Foster, 1998; Pedersen, 1988). Pedersen (1988)suggested that it can be very different to differentiate lungworm infection from chronic allergic bronchiolitis (FBD), which is why cases that had self limiting bronchopulmonary disease were excluded from this analysis. Heartworm infection can also mimic FBD. Heartworm antigen tests were negative in five cats but the sensitivity of antigen tests is poor. Heartworm antibody tests, which have greater sensitivity, were not available for the early years of this study. One cat with radiographic findings consistent with heartworm infection (Case 22) also had heartworm antibody testing and echocardiography performed to exclude the possibility of heartworm infection. Given the rarity of a clinical or necropsy diagnosis of feline heartworm at UCVS during these years and the chronic disease courses and long follow-up times for most of the cats, it is unlikely that heartworm disease mimicked FBD in these cats.
No attempt is usually made to exclude viral causes of FBD. Feline herpesvirus (FHV-1) can be isolated from the trachea and mainstem bronchi after acute infection (Killingsworth et al., 1990) and feline calicivirus can cause pneumonia (Gaskell and Dawson, 1998). In a retrospective post-mortem study of LRTIs, viral LRTIs were confirmed in 5/245 cases and suspected from histological appearance in a further 43 cases (Bart et al., 2000). As viral respiratory infections can produce transient increases in airway reactivity in healthy people (Haslett et al., 2002) and as FHV-1 has been demonstrated to increase tracheal hyperresponsiveness to vagal stimulation (Killingsworth et al., 1990), it is possible that feline URT viruses may cause signs consistent with FBD. In addition, as over 80% of childhood asthma exacerbations and about 50% of all adult asthma attacks are associated with viral URT infections (Micillo et al., 2000), it is possible that cats with mild FBD could have acute exacerbations of their disease with concurrent viral URT infections. Concurrent URT signs were noted in both this study and previous studies (Corcoranet al., 1995; Dye et al., 1996; Moise et al., 1989) and a temporal relationship between respiratory viral infection, or rhinitis, and the development of bronchopulmonary disease was noted in one study (Dye et al., 1996). Viral URT infection exacerbated the disease severity in one of the cats with chronic FBD that died within a week of presentation (Case 21). Further studies are required to investigate temporal or causal relationships of viral URT infections and onset, or exacerbation, of FBD.
The study's main limitation was its retrospective nature and the small number of cases that made the detection of significant statistical associations unlikely. The study defined features of FBD in Sydney cats although the inclusion criteria biased the selection to more severe cases which may also have impacted on the statistical analysis. The study highlights the need for a rigorous and large prospective study of cats with FBD. Ideally, such a study should involve cats in which date of onset of disease is known, no prior treatment has been given, and bacterial and parasitic LRTIs and viral URT infections have been thoroughly excluded. Until such time as this is done, FBD will remain a frustrating entity for practitioners and a cause of considerable morbidity and mortality for cats.
Footnotes
Acknowledgments
This study would not have been possible without the enthusiastic assistance of Rhian Foster and Karen Kilpatrick in medical records and Leanne Fitzsimmons from Veterinary Imaging Associates, Sydney. The clinicians who managed the cases are gratefully acknowledged. Richard Malik is supported by the Valentine Charlton Bequest of the Post Graduate Foundation in Veterinary Science of the University of Sydney.