An IgG heavy, chain-specific WB using blood-agar-grown B. henselae as the antigen source was optimized using sera collected over time from experimentally inoculated cats. Nine immunodominant
antigens were selected by analysis of the pre-infection and post-infection results. Matched serum and blood samples from client-owned
cats with (n = 20) or without fever (n = 19) that resided in high flea prevalence states were selected for analysis, using
the optimized WB and a previously published conventional PCR assay. Results of the serum WB were considered positive if any
two or more of the nine antigens were recognized and were greater than a predetermined appropriate density. Results of the
PCR assay on blood were considered positive if an amplicon of the appropriate size was detected.
Fisher's exact test was used to compare between groups for some parameters; significance was defined as p < 0.05.
WB results were positive in 10 of 20 cats (50 percent) with fever and six of 19 cats (31.6 percent) without fever; differences
between groups were not significant (p = 0.1998). PCR assay results were positive in five of 20 cats (25 percent) with fever
and one of 19 cats (5.3 percent) of cats without fever; differences between groups were not significant (p = 0.1022). Of 16
WB positive cats, three were PCR positive (percentage concordance = 59 percent). All three PCR positive but WB negative samples
were from acutely ill, febrile cats.
Three or more antigens were recognized by five of 10 (50 percent) WB positive cats with fever and five of six (83.3 percent)
WB positive cats without fever. Antigens with the apparent molecular masses of 48, 57, 62, 69 and 82 kD were each recognized
by at least four cats but there was no antigen recognition pattern that was specifically detected in cats with or without
We conclude that determination of Bartonella spp. antigen recognition patterns in this sample set could not be used to predict
which cats had fever associated with Bartonella spp. infection. In addition, presence of antibodies detected by WB did not reliably correlate with the presence or absence of
Bartonella spp. DNA in blood and so should not be used to predict the infection status of individual cats. Lastly, in peracute infections,
Bartonella spp. serum antibody test results can be falsely negative.
Dr. Hoskins is owner of DocuTech Services. He is a diplomate of the American College of Veterinary Internal Medicine with
specialities in small animal pediatrics. He can be reached at (225) 955-3252, fax: (214) 242-2200 or e-mail: