Abstract
Practical relevance:
Bartonellae are small, vector-transmitted Gram-negative intracellular bacteria that are well adapted to one or more mammalian reservoir hosts. Cats are the natural reservoir for Bartonella henselae, which is a (re-)emerging bacterial pathogen. It can cause cat scratch disease in humans and, in immunocompromised people, may lead to severe systemic diseases, such as bacillary angiomatosis. Cats bacteraemic with B henselae constitute the main reservoir from which humans become infected. Most cats naturally infected with B henselae show no clinical signs themselves, but other Bartonella species for which cats are accidental hosts appear to have more pathogenicity.
Global importance:
Several studies have reported a prevalence of previous or current Bartonella species infection in cats of up to 36%. B henselae is common in cats worldwide, and bacteraemia can be documented by blood culture in about a quarter of healthy cats. The distribution of B henselae to various parts of the world has largely occurred through humans migrating with their pet cats. The pathogen is mainly transmitted from cat to cat by fleas, and the majority of infected cats derive from areas with high flea exposure. No significant difference in B henselae prevalence has been determined between male and female cats. In studies on both naturally and experimentally infected cats, chronic bacteraemia has mainly been found in cats under the age of 2 years, while those over 2 years of age are rarely chronically bacteraemic.
Evidence base:
This article reviews published studies and case reports on bartonellosis to explore the clinical significance of the infection in cats and its impact on humans. The article also discusses possible treatment options for cats and means of minimising the zoonotic potential.
Clinical signs in cats
It has often been suspected that Bartonella species may cause clinical signs in cats. Although several studies in experimentally and naturally infected cats have been undertaken to determine a cause–effect relationship between Bartonella species infections and clinical signs, a clear correlation has not been proven so far.
Bartonella henselae
Domestic cats are the natural reservoir for B henselae and are usually considered healthy carriers of this bacterium. However, some clinical signs have been connected to B henselae infection, based on studies involving naturally or experimentally infected cats. 1 Many studies rely on antibody detection,2–6 but are of limited value for establishing a correlation between infection and clinical signs because the presence of antibodies only demonstrates exposure and not necessarily active infection. Furthermore, because of the high prevalence of infection in healthy cats worldwide, it is not easy to demonstrate an association between clinical signs and the presence of B henselae.
Experimentally infected cats
Most cats experimentally infected by exposure to fleas do not develop clinical signs, 7 and it had been assumed that the mode of transmission would be important for clinical outcome. In a recent study, however, 3/6 cats experimentally infected by exposure to fleas showed lethargy and severe fever. One of these cats even developed myocarditis, pleural effusion, pericardial effusion, icterus and ataxia, and had to be euthanased. 8 In another study, 18 cats injected with B henselae-infected blood showed self-limiting febrile disease, transient mild anaemia, localised or generalised lymphadenopathy, mild neurological signs and reproductive failure. Necropsy revealed peripheral lymph node hyperplasia (in 13 cats), splenic follicular hyperplasia (nine cats), lymphocytic cholangitis/pericholangitis (nine cats), lymphocytic hepatitis (six cats), lymphoplasmacytic myocarditis (eight cats) and interstitial lymphocytic nephritis (four cats). 9 In a similar study, 12 cats received different numbers of colony-forming units (cfu)/ml (1010, 108 and 106, respectively) of B henselae intravenously. All cats became bacteraemic 2 weeks after inoculation. Cats infected with higher doses of B henselae developed lethargy, fever, partial anorexia and enlarged lymph nodes, but remained alert and responsive. Lymphoid hyperplasia was diagnosed on cytological evaluation of fine needle aspirates of the enlarged lymph nodes. 10
In most cases, therefore, persistent B henselae infection develops after experimental infection, but clinical signs, if present, are mild and cats recover without treatment.
Naturally infected cats
Most cats that are naturally infected with B henselae do not show clinical signs and appear to tolerate chronic bacteraemia without obvious clinical abnormalities. However, there have been several case reports and case- controlled studies that suggest a link between B henselae infection and a variety of clinical signs, as discussed below.

Endocarditis in a 3-year-old male castrated cat with B henselae infection. (a) Left parasternal long-axis view showing an echogenic vegetative lesion (Veg) associated with the anterior mitral leaflet and which was noted to move independently of the leaflet. LA = left atrium, Ao = aorta. (b) Image obtained 3 weeks after completing a 6-week course of antibiotic therapy. The vegetative mitral valve lesion is no longer present. Reproduced, with permission, from Perez et al (2010) 17

Gross and histological findings in two cats from a North Carolina shelter that died after a litter of flea-infested kittens was introduced to the shelter. (a) Coalescing granulomas distributed throughout the myocardium of an approximately 9-week-old female kitten. (b) Pyogranulomatous myocarditis in an 8-month-old castrated male cat, which had been co-housed with the flea-infested kittens. Macrophages, with a rare multinucleated giant cell (arrow), are particularly numerous towards the upper left of the image; haematoxylin and eosin stain. Inset: Cluster of short bacilli in an inflammatory focus are immunoreactive (brown) for B henselae-specific monoclonal antibody; immunohistochemistry with diaminobenzidine chromogen and haematoxylin counterstain. Reproduced, with permission, from Varanat et al (2012) 18
In another study, antibody titres in serum and aqueous humour and polymerase chain reaction (PCR) results from aqueous humour samples from 49 client-owned cats with uveitis, 49 healthy shelter cats and nine cats experimentally inoculated with either B henselae or B clarridgeiae were examined. Ocular production of Bartonella species IgG was detected in 7/49 cats with uveitis, 0/49 healthy shelter cats and 4/9 experimentally inoculated cats. The organism was detected by PCR in the aqueous humour of 3/24 cats with uveitis, 1/49 healthy shelter cats and 4/9 experimentally inoculated cats. The authors concluded that Bartonella species may infect the eyes of some cats following natural exposure or experimental inoculation and may be a cause of uveitis. 23 However, in another study that compared 113 cats with uveitis with 253 cats that were healthy or had non-ocular disease, there was no difference in serum antibody prevalence or titre magnitude between the groups. 24 The median B henselae titre was 1:64 for all cats, and healthy cats were even more likely to have higher titres than cats with uveitis or cats with non-ocular disease. 24 Despite the case report in which B henselae was strongly suspected as a cause of uveitis, 22 there is only weak evidence to substantiate B henselae as being a common cause of uveitis in cats.
Other Bartonella species
As a rule, Bartonella species rarely cause clinical signs in their reservoir host, but may cause disease in an accidental host. 32 As well as being the reservoir host for B henselae, cats can also be infected with B clarridgeiae, B koehlerae, B quintana, B vinsonii subspecies berkhoffii and B bovis (formerly B weissii).1,33
The role of B clarridgeiae in cats has not been clarified. Cats can be infected with this Bartonella species without clinical signs and transmit it to humans and, thus, may serve as reservoir host. B clarridgeiae has also been implicated as an agent of cat scratch disease (CSD) in humans and aortic valve endocarditis and hepatic disease in dogs. 34 In a study in Alabama and Florida, USA, not only B henselae DNA, but also B clarridgeiae DNA was isolated from blood, skin biopsies, gingival swabs and claw bed swabs of feral cats. 35 In a separate study in the USA, DNA of B clarridgeiae was amplified from the blood of cats or the fleas taken from those cats, and concurrent infection of cats and their fleas with both Bartonella organisms was common. 36 In an investigation into the association between B clarridgeiae infection and chronic rhinosinusitis in cats, no differences in antibody positivity or culture results were found in cats with chronic rhinosinusitis compared with healthy cats. 30 It is unclear whether cats are the main reservoir host for B clarridgeiae or are accidental hosts. It is also unclear how often cats develop clinical signs after B clarridgeiae infection.
B quintana DNA has been amplified from cat fleas 37 and ticks, 38 and the organism has been isolated from feral cats. These cats were healthy, but the pathogenicity of B quintana in cats is unknown, as is the potential for cats to transmit B quintana to humans. 39
Isolation of B koehlerae from the blood of domestic cats has been documented in two cases. In the first report, from the USA, two B koehlerae isolates were recovered from two kittens living in the same household. 40 The second report was from France, where B koehlerae was detected in the blood of a kitten suspected of causing CSD in its owner. 41 In addition, B koehlerae has been isolated from stray cats in Israel. 42
B bovis has been isolated from a few cats in the USA. The genome of the cat isolate was 80% identical to isolates found in more than 35% of 36 French dairy cattle. It was postulated that the cat isolate may have resulted from a transfer of B bovis from a bovine to a feline population some time ago. 43 It is unknown if B bovis may cause clinical signs in cats and/or whether cats may transmit the bacterium to cattle or to humans.
B vinsonii subspecies berkhoffii, the Bartonella species most commonly found in dogs, was detected by blood culture in a cat affected by recurrent osteomyelitis and polyarthritis, lameness, swollen limbs and limb pain (Figure 3). The cat was found to be hypercalcaemic, hyperglobulinaemic and thrombocytopenic. After treatment with antibiotics for 3 months, the cat was walking without pain, and blood culture was negative. Thus, Bartonella species infection may be considered in the differential diagnosis of feline osteomyelitis. 44

Dorsopalmar radiographs of the front feet of a 5-year-old spayed female cat with polyarthritis and osteomyelitis associated with B vinsonii subspecies berkhoffii infection. Note the aggressive expansile lesion of the fifth metacarpal bone and adjacent soft tissue swelling (arrow) in the right foot, and the severe osteolysis of the radial carpal bone (large arrow), osteoproliferation of the carpal and metacarpal bones (small arrow), and carpal joint effusion (arrowhead) in the left foot. Reproduced, with permission, from Varanat et al (2009) 44
Clinical signs in humans
The genus Bartonella comprises several important human pathogens that are associated with a wide range of clinical manifestations. These include long-recognised diseases such as Carrion’s disease, caused by B bacilliformis, and trench fever, caused by B quintana. CSD, as discussed, can be caused by B henselae and B clarridgeiae.45–47 Bacillary angiomatosis and peliosis hepatitis have been described due to B henselae and B quintana infection. Chronic lymphadenopathy has been associated with B quintana. In humans with endocarditis, B henselae, B quintana and B elizabethae have been identified.47–49
B henselae is the most common pathogen transmitted from cats to humans and is considered a zoonotic agent. In various studies performed in the early and mid-1990s, the prevalence of antibodies to B henselae in humans ranged from 4–6%.50–52 In a more recent (2003) study, 23% of examined people had antibodies against B henselae. However, no difference was seen between people with and without domestic pets; nor were differences noted in terms of age, sex, urban or rural residence, or concomitant diseases. 53
While the route of transmission from cat to cat, though not unequivocally clear, is thought to happen indirectly via flea transfer from one cat to another, 8 the transmission of B henselae from cats to humans can occur directly, usually via scratches and bites, or indirectly, by means of the cat flea (Ctenocephalides felis). Most commonly, transmission occurs via the following route:
Cat scratches and bites have been shown to be significantly associated with the development of CSD,51,54 which appears to be the most common clinical presentation of B henselae infection in humans (Figure 4) and has been reported worldwide. In the United States, epidemiological databases estimate that approximately 24,000 cases of CSD occur annually, with a calculated incidence of 9.3/10,000 ambulatory patients per year. 53

Immunohistochemical identification of B henselae in a case of cat scratch disease. With Warthin-Starry silver stain, these coccobacillary pathogens can be seen singly, in small clumps or in chains in necrotic foci. Courtesy of Dharam Ramnani, Webpathology.com
CSD is a persistent, necrotising inflammation of the lymph nodes. Its typical form is characterised by self-limiting regional lymphadenopathy that follows the development of a primary papular lesion and lasts from a few weeks to months. 1 Abscessation of the lymph node and systemic signs (such as fever) are occasionally reported. Atypical forms of CSD are also described; for example, encephalopathies in children and other neurological abnormalities. An unusual case of CSD has recently been reported in a veterinarian affected by persistent fever and back pain after an accidental needle puncture. 55
The immune status of a person infected with B henselae plays an important role in the outcome. In most immunocompetent people, B henselae infection has no clinical sequelae. 56 However, in some immunocompetent people, B henselae causes the typical form of CSD,57,58 which is seen most commonly in children (it is rarely seen in adults). 56 Immunocompromised people may develop atypical CSD and more severe forms of Bartonella-associated disease, which may be fatal if left untreated. These include bacillary angiomatosis, 59 parenchymal bacillary peliosis, relapsing fever with bacteraemia, 60 endocarditis,61–63 neuroretinitis, 64 aseptic meningitis 65 and uveitis. 66
What to do with a Bartonella-infected cat
Current therapeutic strategies in cats are based on in vitro studies, experience gained with human bartonellosis and case reports. Data from controlled efficacy studies in cats are lacking. While many antibiotics are effective in vitro, in vivo efficacy is limited. Elimination of Bartonella species is usually not achieved by treatment, though bacterial load decreases and clinical signs resolve. 68
Several drugs (doxycycline, amoxicillin, amoxicillin–clavulanic acid, enrofloxacin, erythromycin, rifampicin) have been used in cats naturally or experimentally infected with Bartonella species.69–71 Complete clearance of Bartonella species has not been seen with any of the drugs used.
Infected cats without clinical signs
Older cats (ie, over 2 years of age) are able to eliminate Bartonella species without antimicrobial therapy and develop an age-related resistance. In the light of this and the very rare occurrence of disease in cats, and despite the zoonotic risk, symptomless cats infected with B henselae should only be treated in particular circumstances. If young infected cats (under 2 years of age) live in a household with immunocompromised people or with children, they should be treated, regardless of whether or not they show clinical signs. In this scenario, treatment is aimed at decreasing bacterial load and thus the risk of transmission.
The treatment of choice in healthy infected cats living with immunocompromised people is doxycycline (5–10 mg/kg PO q12–24h). Treatment does not lead to total elimination of Bartonella species in all cats; in many cases it serves only to decrease the bacterial load. In one study, bacteraemia was only eliminated in 1/6 cats treated for 14 days and 1/2 cats treated for 28 days. 72
Healthy infected cats that do not live in a household with immunocompromised people or with children should not be treated with antibiotics.
Infected cats with clinical signs
Treatment is also recommended in the rare instances when Bartonella species actually cause disease in a cat. One study tested the in vitro efficacy of 17 antimicrobial compounds. Telithromycin, macrolides, doxycycline and rifampicin were found to be the most effective agents against Bartonella species. 73 There are several case reports documenting use of drugs in affected cats with some success. A cat with recurrent osteomyelitis and polyarthritis associated with B vinsonii subspecies berkhoffii infection and bacteraemia (see Figure 3) recovered after therapy with azithromycin (10 mg/kg PO q48h for 3 months) and amoxicillin–clavulanic acid (62.5 mg PO q12h for 2 months). 44 A cat with B henselae endocarditis was successfully treated with a combination of marbofloxacin (5 mg/kg PO q24h) and azithromycin (10 mg/kg PO q24h for 7 days and then q48h) for 6 weeks. 17

Fleas and ‘flea dirt’ represent the greatest risk factor. Courtesy of Dr Michael R Lappin and HESKA Corporation
Footnotes
Key Points
Funding
The authors received no specific grant from any funding agency in the public, commercial or not-for-profit sectors for the preparation of this review.
Conflict of interest
The authors declare that there is no conflict of interest.
