Susan E. Shaw, BVSc(Hons), MSc, DACVIM, DECVIM-CA, FACVSc, MRCVS
Bartonellosis is caused by fastidious, gram-negative, intraerythrocytic, arthropod-transmitted bacteria of the genus, Bartonella. Several species infect cats (Bartonella henselae, B. clarridgeiae, B. koehlerae) while B. vinsonii subsp berkhoffii, and to a lesser extent B. henselae, B. clarridgeiae and B. elizabethae, have been reported in association with disease in dogs.
Bartonella infection has a world wide distribution in cats and is increasingly widely reported in dogs. Genotypic and phenotypic (serological) variations have been demonstrated among strains of B. henselae in domestic and wild cats and those from different geographical locations. At present, the most significant division within the species delineates strains into one of two subtypes on the basis of their 16S-rDNA gene sequence. In cats, the Houston1 strain predominates in the Far East while the Marseilles genotype is most common in Australia, western USA and Europe. Interestingly in Australia, it is the Houston1 genotype that predominates in human Bartonella cases.
Epidemiology and Risk Factors
Cat fleas are considered the main vector of B. henselae in cats and recent work has shown transmission by skin inoculation of infected flea faeces. However, the role of the cat flea in the transmission of B. henselae from cats to humans is not fully understood. In contrast, ticks (Rhipicephalus sanguineus, Amblyomma and Dermacentor spp.) have been variously incriminated as vectors for B. vinsonii subsp. berkhoffii in domestic and wild dog species. A significant correlation between Bartonella henselae and Borrelia burgdorferi seropositivity has also been reported in UK cats which may reflect common vector exposure. Both Bartonella and Borrelia spp. have been identified in Ixodes ricinus and I. scapularis tick populations feeding on small rodents in the US and the Netherlands.
The prevalence of asymptomatic infection with Bartonella henselae (and B. clarridgeiae) is common but variable in cats; 40-70% have seropositivity and 9-90% are bacteraemia. There have been relatively few surveys in Australia although the prevalence of B. henselae bacteraemia in 77 Sydney cats examined by PCR was 35%. Variability in reported figures may be a consequence of small survey sizes, differences in cat population characteristics (cattery, stray, feral and captive wild cats) and seasonal variation in prevalence as well as true differences in geographic prevalence. The prevalence of infection appears to be higher in young to middle-aged cats but there is no breed or gender predisposition. Although geographic environments with warm temperatures and high humidity are reportedly associated with the highest exposure; the prevalence in cool temperate climates is also relatively high. The effect of climatic factors on the ecology of Bartonella infection may be blurred as in colder countries, domesticated animals are kept in heated domestic or confined environments, facilitating the maintenance of the flea life cycle.
Pathogenicity and Clinical Signs
Most Bartonella-associated disease in dogs has been associated with B. vinsonii subsp. berkhoffii infection. However, as only a small number of cases have been recognised, the true spectrum of clinical manifestations induced by Bartonella species in dogs must be considered virtually unknown. Endocarditis and generalised granulomatous lymphadenitis and dermatitis have been reported most commonly. Both B. clarridgeiae and B. henselae have been associated with rare clinical disease in dogs. B. clarridgeiae has been is0olated from endocarditis in dogs particularly those predisposed to congenital valvular disease and B. henselae has been implicated in a rare case of peliosis hepatis.
Disease association with naturally occurring feline Bartonella infection is difficult to determine. Although clinical disease (fever, lethargy, transient anaemia, lymphadenomegaly, neurological dysfunction or reproductive failure) has been reported following experimental infections with B. henselae and B. clarridgeiae, naturally occurring disease associated with infection is more difficult to define because of its high prevalence in apparently asymptomatic cats. In naturally infected cats, there is a statistical correlation with stomatitis and urinary tract disease. Uveitis associated with intraocular Bartonella DNA and ocular IgG production has also been reported in cats. Although Bartonella infection has been associated with other clinical syndromes such as endocarditis based on positive blood culture, the association is difficult to evaluate unless the presence of lesional organisms is confirmed.
Difficulties in interpreting the significance of positive blood cultures and serology particularly in cats, necessitates the use of multiple diagnostic methods. Although isolation of a Bartonella species by blood culture from a non-reservoir host (e.g., B. henselae or B. clarridgeiae from an ill dog) is more supportive of its role as a causative agent, diagnosis of bartonellosis is best confirmed by demonstration of bartonellae in infected tissues using histological, immunohistological or molecular histological methods.
Blood culture of B. henselae and B. clarridgeiae from antibiotic-free cats is relatively simple microbiologically but requires prolonged incubation which limits its practicality for routine diagnosis. However, culture recovery from non-reservoir (accidental) hosts such as canines is extremely difficult.
Different serological methods have been used to detect circulating antibodies to Bartonella spp. In cats, IgG is persistent which limits the diagnostic usefulness of elevated antibody levels as an indicator of ongoing infection. The estimated positive predictive value of seropositivity as an indicator of bacteraemia is less than 50%. Significant antibody titres may also be present in apparently healthy dogs and the predictive value of serology for diagnosing clinical disease due to B. vinsonii subsp, berkhoffii is limited. In addition, interpretation of Bartonella serological results is compromised by cross-reactivity between Bartonella different species and the non-Bartonella α-subgroup Proteobacteria.
The value of histopathology in the diagnosis of naturally occurring Bartonella infections has only really been explored in dogs. The histopathological presentation of endocarditis and myocarditis due B. vinsonii subsp. berkhoffii is quite characteristic and is supported by the use of Warthin-Starry (WS) silver stain to demonstrate the presence of organisms within compatible lesions.
PCR-based methods have been described targeting a range of DNA fragments of the 16S rRNA-encoding gene, the 16S-23S intergenic spacer region, and the citrate synthase-encoding gene (gltA). These are rapid and specific but sensitivity for detection of Bartonella DNA in infected blood may be limited. This may be addressed by the introduction of the more sensitive real time PCR techniques.
The treatment of B. henselae bacteraemia in cats is problematic. Doxycycline, amoxicillin, amoxicillin/clavulanate used at higher than recommended dose rates have been reported to be successful in suppressing bacteraemia in experimental infections. However, more detailed study suggested that although enrofloxacin was more efficacious than doxycycline for the treatment of B. henselae or B. clarridgeiae, neither drug eliminated the infection in all animals, even when administered for 4 weeks. Data relating to the treatment of naturally infected cats are scant. However, because of the difficulty in eliminating bacteraemia, antibiotic therapy is only recommended for those cats that have confirmed Bartonella-associated disease or those in contact with immunosuppressed owners.
Treatment of canine endocarditis due to B. vinsonii subsp. berkhoffii is also difficult. There has been no clinical response reported to therapeutic protocols incorporating amoxicillin, enrofloxacin, cephalexin, doxycycline and amikacin in combination with diuretics and various combinations of cardiovascular drugs. Two dogs with granulomatous disease due to B. vinsonii subsp. berkhoffii appeared to respond well to antibiotics: a three week course of enrofloxacin (12.5 mg/kg q12 hr) in the first case and a 30 day course of doxycycline (5.4 mg/kg q12 hr) in the second.
Prevention and Control
Experimental studies have shown that cats develop resistance only after challenge with a homologous species of Bartonella or an homologous strain of B. henselae. This lack of cross-protection suggests that any vaccine would require incorporation of multiple Bartonella epitopes.
The prevalence of Bartonella bacteraemia in cats and the risk of Bartonella-associated disease in pet owners, should be decreased by a vigorous integrated flea control programme. When uninfected cats are housed with B. henselae bacteraemic SPF cats in an ectoparasite-free environment, there is no evidence of Bartonella transmission between cats.However, despite the availability of effective flea adulticide treatments, Bartonella infections remain common, even in the domestic cat populations of industrialised, affluent countries of Europe and North America. As yet no studies have been carried out examining the efficacy of different ectoparasiticides in the prevention of B. henselae transmission.
The zoonotic potential of B. henselae is enormous. For example, in the UK, 4.5 million households (or more than 1 in 4) house over 7.5 million cats, of which about 10% are B. henselae bacteraemic. The potential threat of this reservoir is reflected in the frequency with which humans acquire B. henselae infections, most commonly manifesting as cat scratch disease (CSD). In the USA about 24,000 cases of CSD are reported each year, of which about 2,000 require hospitalisation. Fortunately, this syndrome is usually benign and self-limiting, manifesting as a regional lymphadenopathy and affecting mainly children and young adults. However, systemic complications may arise leading to more profound diseases. Accurate diagnosis of CSD is important as it requires differentiation from other potentially more serious causes of lymphadenitis such as abscesses, lymphoma, mycobacterial infections, toxoplasmosis and Kawasaki disease.
When first characterised in the late 1980s, B. henselae was specifically associated with opportunistic infections in AIDS patients. The advent of more effective prophylactic therapy for these patients has seen the incidence of these infections decline in the USA and Europe, although they are likely to remain a significant health burden in Africa and other developing parts of the world where therapies are not currently affordable. Medical interest in zoonotic bartonellae, however, continues today as an increasing spectrum of syndromes among immuno-competent individuals is encountered. Perhaps of most relevance currently is the emergence of B. henselae in the aetiologies of ocular syndromes such as uveitis and neuro-retinitis.
B. clarridgeiae as also been implicated as an agent of CSD and B. vinsonii subsp. berkhoffii has, on one occasion, been identified as the aetiological agent of endocarditis.
1. Breitschwerdt, EB; Kordick, DL. Bartonella Infection in Animals: Carriership, Reservoir Potential, Pathogenicity, and Zoonotic Potential for Human Infection. Clin Micro Rev. (2000); 13: 428-438.
2. Chomel, BB; Kasten, RW; Floyd-Hawkins, K et al. Experimental transmission of Bartonella henselae by the cat flea. J. Clin Microbiol.1996; 34: 1952-1956.
3. Branley J, Wolfsen C, Waters P et al. Prevalence of Bartonella henselae bacteraemia, the causative agent of cat scratch disease in an Australian cat population. Pathology 1996,28:262-265.
4. Birtles RJ. Bartonellosis. In Arthropod-borne Infectious Diseases of the Dog and Cat. Shaw SE and Day MJ Eds. Manson Publishing, London. 2005: 110-119.
5. American Association of Feline Practitioners. 2006 Panel Report on Diagnosis, Treatment and Prevention of Bartonella spp. infections. In press.