Article Text

Prevalence of Bartonella species infections in cats in Southern Germany
  1. M. Bergmann, DVM1,
  2. T. Englert, DVM1,
  3. B. Stuetzer, DVM1,
  4. J. R. Hawley, Research Associate2,
  5. M. R. Lappin, DVM, PhD2 and
  6. K. Hartmann, DVM, PhD1
  1. 1Clinic of Small Animal Medicine, Centre for Clinical Veterinary Medicine, LMU Munich, Veterinaerstrasse 13, Munich 80539, Germany
  2. 2Center for Companion Animal Studies, Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, Colorado, USA
  1. E-mail for correspondence: n.bergmann{at}


Bartonella species are zoonotic pathogens, and infections in cats are common. However, prevalence in cats in Southern Germany is still unknown. Therefore, prevalence of Bartonella species DNA in blood of 479 Southern German cats was determined using a previously published conventional PCR targeting a fragment of the 16S-23S rRNA intergenic spacer region. Associations between Bartonella bacteraemia, housing conditions, feline immunodeficiency virus (FIV) and feline leukaemia virus (FeLV) status, including progressive, regressive and abortive FeLV infection, were evaluated using Fisher's exact test. Prevalence of Bartonella species bacteraemia was 2.5 per cent (12/479; CI 0.01–0.04 per cent). Bartonella henselae DNA was amplified in 11 of the 12 cats. One cat was positive for Bartonella clarridgeiae DNA. Of the infected cats, 2/12 cats were ill; 6/12 cats had thrombocytopenia. There was a significantly higher risk of Bartonella species infection in young and shelter cats, but not in FIV-infected or FeLV-infected cats. Prevalence of Bartonella species bacteraemia is low in Southern German cats, but there is still a risk of zoonotic transmission associated with ownership of young cats. Most of the infected cats did not show clinical signs. Thrombocytopenia was common in Bartonella species-infected cats and further studies are required to define its clinical relevance.

  • feline
  • Cat Scratch Disease
  • PCR
  • bartonellosis
  • vector-borne
  • zoonosis
  • Accepted January 2, 2017.

Statistics from

Bartonella species are zoonotic Gram-negative intracellular bacteria with increasing importance in veterinary and human medicine (Brunt and others 2006, Chomel and others 2006, Breitschwerdt and others 2010). The family Bartonellaceae includes the single genus Bartonella with over 22 species. Bartonella species replicate intracellularly in red blood cells or endothelial cells and are usually transmitted by arthropods (Boulouis and others 2005, Chomel and others 2006, Bradbury and Lappin 2010, Lappin and others 2013).

Cats are the main reservoir hosts for Bartonella henselae, Bartonella clarridgeiae and likely for Bartonella koehlerae. B henselae and B clarridgeiae are the most common Bartonella species in cats (Lappin and others 2006, Lappin and Hawley 2009). They are transmitted among cats or from cats to people by Ctenocephalides felis or through its faeces (Chomel and others 2006, Bradbury and Lappin 2010). Ixodes species can also act as vectors (Cotté and others 2008).

Bartonella-infected cats can develop persistent or recurrent bacteraemia lasting for several months (Kabeya and others 2002). Typical for reservoir hosts, B henselae and B clarridgeiae rarely cause clinical signs in infected cats, and infections often remain undetected (Stuetzer and Hartmann 2012, Pennisi and others 2013). Bartonella species infections are associated with cat scratch disease in human beings, a number of chronic disease syndromes in immunocompetent veterinary healthcare providers or Bartonella researchers (Breitschwerdt and others 2007), bacillary angiomatosis and peliosis hepatis in immunocompromised people (Lange and others 2009) as well as endocarditis and other severe syndromes in human beings and dogs (Chomel and others 2006, Breitschwerdt and others 2010). To date, it is unclear whether immunosuppression in cats predisposes to Bartonella species infection or whether it influences the course of Bartonella species infection and the development of clinical signs (Buchmann and others 2010, De Bortoli and others 2012).

Worldwide, cats are commonly infected by Bartonella species. Prevalence of infection differs among cat populations and geographical region (Brunt and others 2006, Lappin and Hawley 2009, Breitschwerdt and others 2010, Barrs and others 2010). In Germany, three studies investigated the prevalence of Bartonella species bacteraemia so far (Buchmann and others 2010, Mietze and others 2011, Morgenthal and others 2012), but in all studies, cats originated from Northern Germany. Prevalence in cats in Southern Germany is unknown. Thus, the aim of this study was to determine the prevalence of Bartonella species infections in cats in Southern Germany. Associations between Bartonella species infections and housing conditions as well as feline immunodeficiency virus (FIV) and feline leukaemia virus (FeLV) infection status were also evaluated.

Materials and methods


The 479 cats evaluated in this study were either presented to the Clinic of Small Animal Medicine of the LMU Munich, Germany, or to private veterinary clinics in Southern Germany for various reasons. Health status of the cats was evaluated by physical examination. In each cat, the FIV and FeLV status were examined. FIV antibodies were detected using a commercial ELISA (SNAP Kombi Plus FeLV/FIV antibody test, IDEXX GmbH, Ludwigsburg, Germany). FeLV infection status was investigated by performing tests for free FeLV p27 antigen in serum using a commercial ELISA (SNAP Kombi Plus FeLV/FIV antibody test; IDEXX GmbH), FeLV provirus using PCR, as well as anti-FeLV p45 antibodies using an indirect ELISA, both as previously described (Englert and others 2012). Cats with progressive FeLV infection tested FeLV antigen-positive and provirus-positive. Regressively FeLV-infected cats tested antigen-negative and provirus-positive. Cats with abortive FeLV infection tested antigen-negative, and provirus-negative but had FeLV-specific antibodies.

Bartonella species PCR

Total DNA had previously been extracted from whole blood of the 479 cats using the MagNA Pure LC Total Nucleic Acid Isolation Kit (Roche Diagnostics AG) and stored at –80°C until assayed in this study. The samples were thawed at room temperature and prepared for PCR assays using commercially available reagents as described before (Quimby and others 2008).

DNA samples were assessed with a previously described conventional PCR assay for Bartonella species targeting a fragment of the 16S-23S rRNA intergenic spacer region. The amplified products were size-fractionated by electrophoresis on agarose gels and identified by their unique size (Jensen and others 2000). Genetic sequence analysis on Bartonella species amplicons was performed afterwards using a commercially available genetic sequencing service (Macromolecular Resources, Colorado State University, Fort Collins, Colorado).

Positive PCR controls were obtained from diagnostic whole blood samples received by the Center for Companion Animal Studies (Department of Clinical Sciences, Colorado State University). Negative controls consisted of PCR (molecular grade) water being added in lieu of DNA template.

Further laboratory testing

In all cats with Bartonella species bacteraemia, a complete blood cell count was performed. In addition, cats were tested for Anaplasma species, Ehrlichia species and hemotrophic Mycoplasma species by use of previously reported PCR assays (Jensen and others 2001, Lappin and others 2004).

Statistical analysis

Statistical analysis was performed with GraphPad Prism 6.0. CIs were determined by an exact binomial test (Clopper and Pearson 1934). The exact binomial test was one-tailed and was used to prove the alternative hypotheses that the prevalence of Bartonella species infections is within the 95 per cent CI. A significance level of <0.05 was chosen. Fisher’s exact test was used to assess associations between Bartonella species infections, cats’ age, housing conditions, FIV and FeLV status.



Of all cats, 298 cats were male (62.2 per cent) and 181 cats were female (37.8 per cent) (Table 1). Median age was 7.4 years. In total, 106 cats were purebred (22.6 per cent) and 362 cats were domestic shorthair (DSH) (77.4 per cent). The majority (378 cats) lived in multicat households (84.9 per cent) and 67 cats in single-cat households (15.1 per cent). Outdoor access was allowed in 230 of 434 cats (53.0 per cent). Fifty-six cats (11.7 per cent) originated from animal shelters.


Cats with and without Bartonella species infection, and analysis of the risk factors housing conditions, feline immunodeficiency virus (FIV) and feline leukaemia virus (FeLV) infection status

Progressive FeLV infection was detected in 9 of 479 cats (1.9 per cent), regressive FeLV infection was detected in 7 cats (1.5 per cent) and abortive FeLV infection was detected in 22 cats (4.5 per cent). A total of 7 cats were FIV antibody-positive (1.5 per cent), two of these 7 FIV-infected cats were also progressively FeLV-infected.

Prevalence of Bartonella species infections

In total, 12 of 479 cats were positive for Bartonella species DNA. Thus, prevalence of Bartonella species bacteraemia was 2.5 per cent (95 per cent CI 1.39–4.38 per cent). By genetic sequencing, 11 of the Bartonella species were shown to be B henselae; in 2 of these B henselae-positive cats, B henselae (Houston-I strain) complete genome was amplified. In 1/12 Bartonella species-positive cat, DNA homologous with B clarridgeiae was amplified. All PCR-negative and extract-negative controls were negative in all assays.

Risk factors

Bartonella species bacteraemia was significantly more often found in young cats under the age of two years (P=0.003; OR 6.481; 95 per cent CI 1.819–20.38). Cats from animal shelters had a significantly higher risk than cats living in private households (P<0.0001; OR 45.76; 95 per cent CI 8.49 –168.00). Cats in multicat households (P=0.228) and cats with outdoor access (P=0.230) had no higher risk (Table 1). There was no significant difference in the FIV status between Bartonella species-negative and species-positive cats (P=0.143). There was also no significant difference in the FeLV status, neither concerning progressive FeLV infection (P=0.206), regressive FeLV infection (P=1.000) nor abortive FeLV infection (P=0.224) between Bartonella species-negative and species-positive cats.

Bartonella species-infected cats

Table 2 summarises signalment, housing conditions and examination results of the 12 Bartonella species-infected cats.


Cats infected with Bartonella species

Three Bartonella species-infected cats were healthy and had no laboratory abnormalities. Four Bartonella species-infected cats were clinically healthy but had thrombocytopenia.

Two Bartonella species-infected cats were presented with a history of clinical illness. The first cat had vomiting, a bulging abdomen, gingivitis, flea infestation, thrombocytopenia (manually counted platelets: 60×109/l; reference range: 180–550×109/l) and Toxocara species infection. After treatment with flubendazole, the cat was in a good condition and the platelets were within the reference range. The second cat was presented with a history of poorly controlled diabetes, gingivitis, lymphopenia (0.71×109/l; reference range: 1–4×109/l) and low levels of glucose (2.3 mmol/l; reference range: 3.7–6.9 mmol/l), while fructosamine was elevated (479 µmol/l; reference range: 0–340 µmol/l). On ultrasound the cat’s pancreas was hyperechogenic; chronic pancreatitis and hypoglycaemia due to insulin overdose was suspected.

Overall, 6 of 12 (50 per cent) Bartonella species-infected cats were thrombocytopenic with platelet counts ranging from 7×109/l to 149×109/l. All thrombocytopenic cats tested negative for Anaplasma species, and Ehrlichia species. Also, 3 of the 12 Bartonella species-infected had lymphopenia ranging from 0.40×109/l to 0.80×109/l (Table 2). One cat showed monocytosis (monocytes: 0.94×109/l; reference range: 0.04–0.5×109/l).


Infection with Bartonella species was found in 2.5 per cent (12/479; CI 0.01 –0.04 per cent) of cats from Southern Germany in the present study, while in Northern Germany bacteraemia detected by PCR was present in 0.0 per cent, 8.0 per cent and 16.0 per cent of cats (Buchmann and others 2010, Mietz and others 2011, Morgenthal and others 2012). Differences in prevalence of Bartonella species in cats within certain regions are seen worldwide depending on investigated cat populations (Boulouis and others 2005). Studies on Bartonella species infections in shelter or stray cats often show higher prevalence (Heller and others 1997, Gurfield and others 2001). The prevalence of 2.5 per cent in the present study is markedly lower than those found in other studies in Europe with comparable cat populations (Gurfield and others 2001, Birtles and others 2002, Cabassi and others 2002). This is most likely due to the low prevalence of Bartonella species in cat fleas in Germany (4 per cent), the main vector for Bartonella species (Boulouis and others 2005). Prevalence of Bartonella species in cat fleas in other European countries ranges up to 28 per cent (Shaw and others 2004, Blanco and others 2006, Just and others 2008). In addition, prevalence of flea infestation in cats in Southern Germany was shown to be low (11 per cent) compared with other parts of Germany or Europe (Bond and others 2005; Beck and others 2006).

Bartonella species bacteraemia was significantly more often found in young cats (≤ 2 years) in the present study. This is in agreement with other studies. It is likely that the immune system of young cats infected for the first time does not repress growth of Bartonella species. In contrast, older cats are more likely to terminate bacteraemia, although many are probably still infected since antibody prevalence remains high (Guptill and others 2004). However, only bacteraemic cats are expected to transmit infection. Further risk factors for Bartonella species bacteraemia previously described in cats are multicat environment, especially animal shelters, and outdoor access as these cats have an increased risk of being exposed to arthropod vectors (Gurfield and others 2001, Guptill and others 2004, Boulouis and others 2005). In the present study, shelter cats had a significantly higher risk of Bartonella species infection, while outdoor access and multicat household were not significantly associated. Shelter cats are often housed in stressful conditions that can lead to a compromised immune system and predispose these cats for bacteraemia.

Association between Bartonella species infection and clinical disease in naturally infected cats is still not fully understood. Natural Bartonella species infection is mainly subclinical in cats, but it has been discussed that some cats might develop fever, lymphadenopathy, gingivitis, stomatitis, cardiac, renal or ocular disorders (Ueno and others 1996, Lappin and others 2000, Chomel and others 2003, Bradbury and Lappin 2010). It has also been discussed whether immunosuppression, such as through FIV and FeLV infection, could lead to enhanced pathogenicity of Bartonella species. Infection with B henselae or B clarridgeiae was indeed associated with FeLV infection in some studies (Buchmann and others 2010, de Bortoli and others 2012). There was also evidence that coinfection with Bartonella species and FIV could induce disease manifestations, such as lymphadenopathy (Ueno and others 1996, Rolain and others 2001, de Bortoli and others 2012). Therefore, the role of FIV and FeLV as risk factor was assessed in the present study. To diagnose FeLV infection, free FeLV p27 antigen is commonly used but this test does not identify cats that harbour the virus but are aviremic. To detect different courses of FeLV infection, various testing systems (antigen tests, PCR and antibody tests) were used. However, no significantly higher risk of coinfection with FIV and/or FeLV was detected. This is likely related to the low prevalence of FIV and FeLV infections in cats in this study and all over Germany today.

In the present study, most Bartonella species-infected cats did not show any clinical signs. This is in common with many other reports in naturally infected cats, and infection-associated illness has never been proven so far (Stuetzer and Hartmann 2012, Pennisi and others 2013). Only 2 of the B henselae-infected cats were ill with many of their problems likely related to other diseases. However, both cats were presented with gingivitis, and one of them with chronic pancreatitis. Bartonellosis has been proposed as a cause of gingivitis, as well as feline pancreatitis, but true associations could not be defined, and further studies are warranted (Lappin and others 2000, Bayliss and others 2009).

The most commonly reported haematological abnormality in experimental feline bartonellosis is transient anaemia and persistent eosinophilia (Kordick and others 1999). Only a few studies have evaluated laboratory changes in natural Bartonella species infections. These studies found lymphocytosis, haematuria and hyperglobulinaemia most commonly associated with the presence of Bartonella species antibodies (Breitschwerdt and others 2005, Whittemore and others 2012). Most common abnormality in the Bartonella species-infected cats in the present study was thrombocytopenia, found in 6/12 cats. History and physical examination in these cats revealed no evidence for other causes of thrombocytopenia, such as drug exposure, neoplasia or disseminated intravascular coagulation. No cat was infected with Anaplasma species or Ehrlichia species, which could alternatively explain the thrombocytopenia.

One cat was thrombocytopenic and coinfected with ‘Candidatus M. haemominutum’. This organism usually causes inapparent infections in cats or only results in mild anaemia. As no anaemia was present, ‘Candidatus M. haemominutum’ infection was likely subclinical and, thus, not the cause of thrombocytopenia in this cat (Foley and others 2001). Another cat had a severe thrombocytopenia, which is usually caused by a reduced production or a destruction of platelets. As this cat was a young cat, and as physical examination, other cell counts, FIV and FeLV testing were unremarkable, and Anaplasma species and Ehrlichia species PCR from whole blood were negative, B henselae infection might have indeed been the reason for the thrombocytopenia in this cat.

Thrombocytopenia caused by Bartonella species has been reported in human beings and dogs (Tuttle and others 2003, Goodman and Breitschwerdt 2005, Palumbo and others 2008). Bartonella species are endotheliotropic and can be located intracellularly in vascular endothelial cells of infected cats (Dehio 2001). Cell damage during bacteraemia could therefore have caused the thrombocytopenia. Alternately splenic hyperplasia caused by Bartonella species infection might have led to splenic sequestration of platelets (Kordick and others 1999).

A follow-up of the thrombocytopenic cats from the present study was not possible; this would have been interesting to assess whether the thrombocytopenia of the cats resolved or not.

The major limitation of the study was the relatively small number of cats with Bartonella species infections, making risk factor analysis difficult. Performing Bartonella species PCR but not antibody testing was another limitation. Cats with Bartonella species infections can be false negative by PCR, and performance of a single PCR might have led to an underestimation of the true Bartonella species prevalence rates in Southern Germany (Mietz and others 2011, Morgenthal and others 2012). Future studies should assess Bartonella species antibody prevalence in conjunction with molecular methods to determine the Bartonella species exposure rates in cats. In addition, standard PCR might be less sensitive than other PCR methods. Other PCR gene targets or real-time PCR can increase sensitivity of PCR (Kamrani and others 2008). To prove viable, Bartonella species infection blood enrichment and culture is the most reliable test that was not performed in the present study but could be considered for the future (Birtles and others 2002). Even though the prevalence rate for Bartonella species determined by PCR assay in this study was relatively low, the results indicate that flea control should be strictly recommended to minimise the risk of zoonotic transfer (Bradbury and Lappin 2010).

In conclusion, prevalence of Bartonella species infections was low in Southern German cats. The clinical relevance of Bartonella species as feline pathogen is still unclear. Concurrent infections with FIV and FeLV did neither appear to be a risk factor nor to alter the course of Bartonella infection. In the two B henselae-infected cats that were clinically ill, most changes were related to other disease. Thrombocytopenia was commonly present in Bartonella species-infected cats, and therefore, studies should be conducted to further define the clinical relevance of this finding as well as of Bartonella species infection in cats.


The authors thank Prof. Dr Ralf S. Mueller, Clinic of Small Animal Medicine, for his assistance in the statistical examination. Parts of the results were presented as a poster at the ECVIM Symposium in Lisbon, Portugal, 2014.

  • Accepted January 2, 2017.


View Abstract


  • Provenance: Not commissioned; externally peer reviewed

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.