Abstract
Melioidosis was diagnosed in two cats at necropsy. The first cat presented with jaundice and anaemia but died of overwhelming sepsis soon after admission, despite blood transfusion and other supportive measures. The second cat died several days after developing neurological signs; an infected digital wound may have been the primary focus of infection in this patient. The cats had presumably acquired the infection in Malaysia and northern Australia, respectively, and in both cases disease may have represented reactivation of a latent infection brought on by the stress of relocation. The epidemiology, clinical presentation, diagnosis and treatment of melioidosis are discussed from a feline perspective.
Case 1
A 5-year-old spayed domestic shorthaired cat was presented for inappetence and weakness on 4 January 2001. The cat, which had been born and acquired in Malaysia, had recently been transported to Australia. On admission to quarantine housing, the cat had been considered obese and accordingly was placed on a reduced calorie ration (Iams Cat Light, Iams Company, Dayton, Ohio). When released from quarantine (9 December 2000), it had a reduced appetite. Inappetence progressed to anorexia over the ensuing 4 weeks and the cat became so ill that it could not stand on the day prior to referral. It was taken to a veterinarian, who detected anaemia (haematocrit 0.13 l/l and icterus, prompting urgent referral to the University Veterinary Centre Sydney.
The cat was thin (3.0 kg) and obtunded on presentation. Physical findings included tachycardia (250 beats/min), a weak femoral pulse, rapid shallow respiration (50 breaths/min), hypothermia (36.5°C) and icterus. Supportive therapy was provided while further investigation was performed. The cat was given a balanced electrolyte solution intravenously (Compound Sodium Lactate, Baxter Healthcare; 15 ml/h) and nasopharyngeal oxygen (1 l/min). After results of blood typing were available, a fresh whole blood transfusion was administered (45 ml of type A blood; 20 ml by slow IV infusion over 10 min, the remaining 25 ml over 2 h).
Haematology demonstrated a non-regenerative anaemia (haematocrit 0.10 l/l [reference range (RR) 0.27 to 0.45 l/l]; haemoglobin 38 g/l [RR 80–140 g/l]; no reticulocytes detected), marked neutrophilia with a left shift (segmented neutrophils 39.0×109/l [RR 3.8–10.1×109/l]; bands 3.5×109/l [RR 0–0.42×109/l] and metamyelocytes 0.72×109/l, granulation and vacuolation of neutrophils and moderate numbers of Heinz bodies. The leukogram was suggestive of a severe inflammatory process, most likely an overwhelming infection. Serum biochemical values were normal apart from increased alkaline phosphatase activity (159 U/l RR <50 U/l, hyperbilirubinaemia (55.8 μmol/l RR 2.5–3.5 μmol/l) and hypokalaemia (2.4 mmol/l RR 4.0–4.6 mmol/l). Feline immunodeficiency virus antibodies and feline leukaemia virus antigen were not detected in serum using immunochromatography (Witness FIV and Witness FeLV, Agen Biomedical Ltd, Acacia Ridge, Queensland, Australia). Thoracic and abdominal radiographs taken 2 h after admission demonstrated sternal lymphadenomegaly (Fig 1), hepatomegaly (Figs 1 and 2) and reduced serosal detail within the abdominal cavity. Abdominal ultrasonography demonstrated a small amount of free fluid, mild mesenteric lymphadenomegaly and severalhypoechoic foci (up to 8 mm diameter) within the hepatic parenchyma. Material for cytology and microbiology was obtained via fine needle aspiration from the liver lesions. Despite oxygen therapy, intravenous crystalloids and blood transfusion the cat continued to deteriorate and suffered a respiratory arrest shortly after the ultrasound examination. Although it was easily resuscitated, it subsequently rearrested and died.
00094-3/unzip/10.1016_S1098-612X(02)00094-3-fig1.png)
Lateral thoracoabdominal radiograph of Case 1 on the day of referral. Note the large sternal lymph node (arrow) and the rounded border of the enlarged liver extending caudally beyond the costal arch.
00094-3/unzip/10.1016_S1098-612X(02)00094-3-fig2.png)
Ultrasonogram of the liver of Case 1 (5–10 MHz linear array transducer, ATL HDI Ultramark 9). Note the focal anechoic lesion (8 mm diameter) within the liver parenchyma. At necropsy, similar lesions were shown to be abscesses.
Necropsy demonstrated a small amount of yellow fluid (approximately 5 ml) within the abdominal cavity. The liver was large, friable and yellow, with rounded borders, a prominent lobular pattern and greasy consistency on handling. Gross sectioning of the liver revealed numerous white, nodular lesions as well as occasional dark red haemorrhagic areas. The spleen contained multiple small, dark red, raised nodules. The periportal and sternal lymph nodes were enlarged, and the lungs were consolidated. Material was collected from the liver and sternal lymph node for cytology and microbiology immediately after the cat had died. The antemortem aspirates were not examined.
Squash preparations of the liver were stained with DiffQuik (American Scientific Products, McGaw, Illinois) and Burke's modification of the Gram stain. DiffQuik stained smears demonstrated marked vacuolation of hepatocytes, interspersed with moderate numbers of neutrophils and macrophages (Fig 3). In Gram stained smears, Gram-negative rods could be seen extracellularly but not intracellularly (Fig 4). Smears made from sternal lymph node aspirates showed numerous lytic neutrophils, as well as increased numbers of plasma cells and macrophages, suggestive of lymphadenitis. Low numbers of Gram-negative rods were observed also in these specimens.
00094-3/unzip/10.1016_S1098-612X(02)00094-3-fig3.png)
Squash preparation of liver lesion from Case 1 obtained at necropsy. Note the suppurative inflammation and vacuolated hepatocytes. DiffQuik ×754.
00094-3/unzip/10.1016_S1098-612X(02)00094-3-fig4.png)
Squash preparation of liver lesion from Case 1 obtained at necropsy. Note the curved Gram-negative bacilli (arrow). Burke's modification of the Gram stain; ×754.
Histologically, the most severe changes were present within the liver. A severe vacuolar hepatopathy was evident and confirmed to be hepatic lipidosis with oil Red-O staining. Occasional portal triads were surrounded by a non-specific inflammatory response and there was some biliary hyperplasia. In addition, the hepatic parenchyma contained small abscesses (up to 10 mm diameter) and necrotic regions surrounding areas of haemorrhage. No bacteria were seen in haematoxylin and eosin (H&E) and Brown and Brenn (Gram) stained sections. There was severe purulent lymphadenitis of the sternal lymph node and chronic active pneumonitis affected much of the pulmonary parenchyma. Small micro-abscesses were present in the red pulp of the spleen. The bone marrow was hypocellular, with a deficiency of red cell progenitors and very little erythropoiesis.
Bacteriology
Samples of liver, bile and sternal lymph node collected at necropsy were submitted for bacteriological investigations. A pure, heavy growth of a Gram-negative rod was cultured on horse blood agar from both liver lesions and the sternal lymph node. No bacteria were isolated from the bile. Colonies were first visible aerobically after 24 h; at 4 days they were 3 mm in diameter, circular, creamish grey, low convex with a dull surface sheen. The surface of colonies became wrinkled with time and a pungent smell reminiscent of dirt was produced. The isolate did not grow anaerobically on tryptose agar (indicating it was an obligate aerobe), but was able to grow anaerobically on blood agar due to its ability to respire using nitrates in the media. It grew at 42°C but not 4°C. The isolate was motile, oxidase positive, catalase positive, reduced nitrates and produced gas when nitrates were reduced anaerobically. It possessed both urease and gelatinase activity and fermented lactose on McConkeys agar. Based on the agar disc diffusion method, the isolate was sensitive to amoxicillin/clavulanic acid, ticarcillin, enrofloxacin, ciprofloxacin, gentamicin, tetracycline and trimethoprim/sulphamethoxazole (TMP-SXT), but resistant to ampicillin.
Clinical, histological and bacteriological findings suggested the cat had a disseminated infection with Burkholderia pseudomallei (melioidosis) and died as a result of overwhelming sepsis.
Case 2
An 18-month-old castrated male cat died several days after the development of neurological signs. Three weeks after arriving from Queensland the cat was presented to a veterinarian in Darwin, with a 24-h history of staggering and a left forelimb lameness. The cat was bright, but had bilateral protrusion of its third eyelids. Recent probable catfight wounds were evident. The cat's temperature fluctuated over the ensuing 4 to 5 days, muscle trembling became evident and it developed left foreleg paralysis and right foreleg paresis without obvious hindlimb dysfunction. The cat remained bright and continued to eat, although in the last day before death it deteriorated dramatically. The cat was treated with antibiotics, anti-inflammatory drugs and fluids (details unobtainable), but died 5 days after presentation.
Necropsy examination by one of the authors (HP) revealed an abscess of the first digit of the left hind limb, discharging pus from around the nail bed. There was a single small abscess (1 to 2 mm diameter) in the spleen and several similar abscesses under the capsule of the liver. The lungs had patchy dark red discolouration, while the meninges were slightly dark and cloudy.
Swabs of the toe and meninges, and samples of fresh liver and spleen all grew Burkholderia pseudomallei. Histologically, acute abscesses were present in the spleen and liver. The cervicothoracic spinal cord and brain stem showed severe, focal necrosis with purulent inflammation. The pathological process extended into some spinal nerve roots and the spinal meninges. The final diagnosis was disseminated melioidosis with prominent meningoencephalitis.
Discussion
Melioidosis is a bacterial infection caused by the aerobic, Gram-negative bacillus Burkholderia pseudomallei (formerly Pseudomonas pseudomallei). This organism is a ubiquitous soil saprophyteendemic to Southeast Asia, northern Australia and the South Pacific. Epidemiological surveys have demonstrated that it is restricted mainly to areas of latitude 20° north and south of the equator (Leelarasamee & Bovornkitti 1989). The organism has been isolated from moist soil and waterholes in these areas and it is believed that these environments are the primary reservoirs from which most infections are acquired (Brett & Woods 2000). The condition has been recorded in humans as the commonest cause of fatal community-acquired sepsis during the wet season in northeast Thailand, and also in tropical regions of the Northern Territory of Australia (Currie et al 2000a), where it has recently become a notifiable disease (Low Choy et al 2000). It is thought that during the wet season, rising water tables result in the organism percolating up through the underlying soil, thereby increasing exposure of humans and animals (Brett & Woods 2000).
Melioidosis has been reported in many domestic and wild animals. Susceptibility to the infection varies between species, with sheep and goats being particularly susceptible. Porcine melioidosis is not uncommon, however bovine infectionis extremely rare. Melioidosis has also been recorded in dogs, camels, alpacas, horses, deer, laboratory animals, tree kangaroos, wallabies, koalas, crocodiles, various avian species (Low Choy et al 2000) and captive marine mammals (Hicks et al 2000). Dogs and cats appear relatively resistant to disease. To the best of our knowledge, only one other case report of feline infection has been published (Toh et al 1998), although the prevalence of infection in small domestic animals may be underestimated, as melioidosis is diagnosed with reasonable frequency in dogs and cats during the wet season in the Northern Territory of Australia (H Parkes, unpublished observations) and Malaysia (P Irwin, personal communication).
In people, the main modes of infection are thought to include (i) cutaneous inoculation via wounds or arthropod vectors, (ii) inhalation of dust containing the organism and (iii) ingestion of soil or contaminated carcases (Greene 1990). In susceptible hosts, especially human patients with predisposing conditions such as diabetes mellitus or renal disease (Woods et al 1999), the infection may then disseminate widely, resulting in abscess formation in a variety of organs (Greene 1990). Nosocomial transmission has been suspected and the bacteria has been demonstrated to survive in an intravenous anaesthetic agent used in cats(alphaxalone and alphadalone; Saffan, Schering-Plough Animal Health, North Ryde, NSW) and an antiseptic/cleaning agent used for surgical site preparation (cetrimide 3% and chlorhexidine 0.3%; Savlon, Novartis Animal Health Australasia, Pendle Hill, NSW) (Low Choy et al 2000). Zoonotic transmission is thought to beextremely uncommon, although there have been anecdotal reports of possible animal to human transmission in a meat worker, a veterinarian and a goat farmer, all of whom developed cutaneous Burkholderia infections (Low Choy et al 2000) and one report of transmission from a sheep (Idris et al 1998). Transplacental transmission has been established in goats (Thomas et al 1988a) andvenereal transmission was suspected in one human patient who developed a high indirect haemagglutination titre after sexual contact with an infected person (Chaowagul 2000).
Both of the cats in this report spent time in endemic areas and it must be presumed that the infections were acquired in these locations. The exact time course of the infections in these cats and the routes of entry of the bacteria cannot be determined with certainty. We hypothesise that Case 1 may have been subclinically infected whilst in Malaysia. The stress of quarantine, caloric restriction and the development of hepatic lipidosis presumably resulted in the conversion of the infection from subclinical to acute and fulminant. The distribution pattern of the lesions suggests spread of infection from the gut or mesenteric lymph nodes to the liver via the portal circulation. Lymphatic spread from the liver presumably accounted for involvement of the sternal lymph node. The time course and pathogenesis of the infection in Case 2 is similarly unclear. The cat may have acquired the infection via cutaneous wounds, particularly those on the left hind foot from which a positive culture of B pseudomallei was obtained, with secondary dissemination to internal organs and the central nervous system (CNS). Alternatively the infection may have started elsewhere, with late dissemination to the digits and CNS.
Melioidosis has been labelled the ‘great imitator’ as it can affect virtually any organ system. The spectrum of clinical syndromes range from (i) mild or inapparent infections, (ii) chronic abscessation and granuloma formation to (iii) acute, fulminant septicaemia and pneumonia. In the past, melioidosis was considered an uncommon condition resulting in severe and mostly fatal disease. The consensus now is that most infections in people living in endemic regions are probably asymptomatic (Leelarasamee & Bovornkitti 1989), with epidemiological studies demonstrating widespread occurrence of antibodies to B pseudomallei in apparently healthy individuals (Ashdown & Guard 1984, Currie et al 2000c). Chronic infections may be characterised by the formation of single or multiple abscesses or granulomata in any organ, although the liver, spleen and lungs tend to be involved preferentially (Low Choy et al 2000). Although a wide spectrum of signs can be seen among individuals, one study has suggested there may be a trend towards infection of particular organ systems in given species. Mastitis, for example, is seen more frequently in caprine melioidosis whereas the respiratory system is involved most commonly in sheep (Low Choy et al 2000). Melioidosis is also considered to be a cause of pyrexia of unknown origin in humans in endemic areas, even when comprehensive imaging studies have failed to demonstrate abscessation (Leelarasamee & Bovornkitti 1989). The acute form of the disease is associated with high morbidity and mortality and even with aggressive therapy utilising appropriate antibiotics, 40% or more of human patients with B pseudomallei septicaemia succumb (Woods et al 1999). A further worrisome feature of melioidosis is its propensity for re-activation after long periods of dormancy. Indeed there have been reports of the infection emerging as acute, fulminant disease up to 26 years after presumed exposure (Woods et al 1999).
Neither of the cats in this report had clinical signs strongly suggestive of an infectious aetiology. At the time of presentation Case 1 had severe signs of systemic disease, which could have been attributed to any number of causes. The diffuse loss of serosal detail in abdominal radiographs was attributed to the presence of free fluid, however as only a small volume was seen on ultrasound, a lack of intra-abdominal fat may have contributed also. The sternal lymphadenopathy added further weight to the likelihood of intra-abdominal disease, as it has been established that a portion of the lymphatic drainage of the abdominal cavity is returned to the systemic circulation via this route (Latimer et al 1997). However, the presence of an inflammatory leukogram with toxic neutrophils pointed towards an infection, and the observation of organisms within the multiple foci of suppuration within the liver of the Case 1 would have provided a definitive diagnosis had cytological and microbiological examination of liver aspirates been undertaken before the animal died. Establishing the pathogenesis of the severe anaemia in this case remains a challenge, however the absence of a regenerative response in the haemogram coupled with the bone marrow histology at necropsy suggest a failure of erythropoiesis, possibly with surperimposed haemolysis. The anaemia of inflammation and sepsis in cats has been documented, however exact mechanisms underlying these phenomena remain poorlyunderstood (Weisse et al 1983, Weisse & McClay 1988, Brady et al 2000).
Neurological melioidosis is well documented in veterinary and human patients, however its prevalence is low. It has been noted to have a predilection for the brainstem and spinal cord, with the formation of microabscesses within the neural parenchyma as well as diffuse meningoencephalitis. Bacteria may be cultured from central nervous system (CNS) tissue and cerebrospinal fluid (CSF), however is often difficult to visualise organisms in histological sections. CSF analysis typically demonstrates an elevated white cell count, variable differential cell populations and elevated protein concentrations. Computed tomography and magnetic resonance imaging show unequivocal changes in affected human patients (Currie et al 2000b). In the instance of Case 2, the CNS is most likely to have become affected via haematogenous spread from a primary focus elsewhere.
Melioidosis is traditionally diagnosed by isolating B pseudomallei from blood or lesions and this remains the diagnostic technique of choice (Sirisinha et al 2000). The organism requires no special media and is easy and inexpensive to isolate. Careful examination of Gram stained smears of pus from lesions can sometimes identify small numbers of bipolar ‘safety pin’ shaped Gram-negative rods, which may raise suspicion of melioidosis in patients from endemic areas (Leelarasamee & Bovornkitti 1989). Laboratory personnel unfamiliar with the organism may mistakenly identify it as a Pseudomonas species, the genus in which it was classified previously (Yabuuchi et al 1992). The main drawback with relying on culture is that it may take up to 3 or 4 days to obtain definitive results and this may be too slow for many patients, especially those with acute disease (Sirisinha et al 2000).
Immunological assays for the detection ofBurkholderia antigen in blood, urine, pus, biopsied tissues, sputum and throat swabs or antibodies in serum have been developed and have the advantage of providing a positive diagnosis much faster than culture (Sirisinha et al 2000). A number of easy to perform antibody tests are available and widely used for screening purposes in both human and veterinary patients in endemic areas (Thomas et al 1988b, Sirisinha et al 2000), although none have been developed specifically for use in cats. The main disadvantage of antibody testing is that it may reflect exposure rather than active disease, although a sufficiently high titre is highly suggestive of current infection (Leelarasamee & Bovornkitti 1989). Antigen detection is considered superior because it unequivocally indicates disease (Sirisinha et al 2000). This is particularly important in endemic areas where it is expected that a significant proportion of the population will have beenexposed to the organism (and mounted a successful immunological response) or are infected subclinically.
More recently DNA probes and polymerase chain reaction (PCR) tests have been developed and found to yield outstanding sensitivity (100%) but unsatisfactory specificity, with false positive results recorded in a third of patients in one study of where PCR was performed on a variety of specimens (Haase et al 1998). It has therefore been suggested that the most expedient route to diagnosing melioidosis, whilst maintaining an acceptable degree of accuracy, is a combination of culture and serology or PCR (Sirisinha et al 2000).
Treatment of melioidosis is protracted, expensive and often unsuccessful (Low Choy et al 2000). Acute severe melioidosis seems particularly difficult to cure and optimal antibiotic regimens have yet to be determined (Chaowagul 2000). B pseudomallei is generally resistant in vitro to penicillin, amino-penicillins, first and second-generation cephalosporins, most amino-glycosides, most macrolides and rifampicin (Chaowagul 2000). Generally, it is susceptible in vitro to some third generation cephalosporins, carbapenems, chloramphenicol, tetracyclines, TMP-SXT, fluoroquinolones and amoxycillin/clavulanic acid (Chaowagul 2000). Based onrecent trials the current recommendation for severe acute melioidosis in humans is intravenous ceftazidime (with or without TMP-SXT), or imipenem, given for at least 10 days or until there are definite signs of a clinical response (Chaowagul 2000). This is followed with a maintenance drug regimen consisting of orally administered doxycycline, TMP-SXT and chloramphenicol given for 8 to 10 weeks. This regimen would be cumbersome in a veterinary setting and as amoxycillin/clavulanic acid has been shown to be adequate for maintenance treatment(Rajchanuvong et al 1995), it is proposed that it may be used as first line of treatment in veterinary patients following a short concurrent course of intravenous imipenem. Although prolonged oral maintenance therapy should minimise the risk of relapse, it is recommended in the human literature that all patients should receive lifelong monitoring for signs of relapse (Chaowagul 2000) and this would seem appropriate also for cats.