TEXT

Top Abstract Text References

Bartonella bacilliformis causes bartonellosis, an illness that is currently limited to high-altitude valleys of the Andes Mountains of Peru, Columbia, and Ecuador. It is one of several members of the genus Bartonella, along with B. elizabethae, B. henselae, and B. quintana, that are known to cause severe illness in humans. Bartonellosis is typically characterized by an acute phase of fever and hemolytic anemia followed by a second phase of cutaneous vascular lesions called "verruga peruana" (11). Recently there have been increases in the number of reported cases from areas where bartonellosis is endemic, along with an emergence of the disease in new locations and as a threat to travelers (1, 5, 6, 12). This increasing disease burden, coupled with the recognition of other Bartonella spp. as emerging pathogens of animals and humans, makes the study of South American bartonellosis increasingly important (4, 5, 10, 14).

Though bartonellosis has been recognized since pre-Columbian times, diagnosis remains problematic and is usually based on clinical impression and the demonstration of the intraerythrocytic bacilli on a Giemsa- or Wright-stained thin blood smear (3). The sensitivity of the thin-smear procedure has been shown to be only 36% (5). Culture of B. bacilliformis is difficult, requiring special media and techniques with up to an 8-week incubation time. There have been no generally accepted serologic assays available to confirm clinical suspicion of the disease or to conduct seroepidemiologic surveys of exposed populations. Crude-extract antigens, whole-cell antigens, and protein antigens have been used for the immunodiagnosis of bartonellosis (7, 9). Knobloch et al. (8) identified and prepared protein antigens of B. bacilliformis to overcome problems with nonspecific reactivity associated with the crude-extract and whole-cell antigen preparations. However, data on the sensitivity and specificity of these antigens for diagnostic testing have not been published.

We developed an indirect fluorescence antibody (IFA) test that uses an irradiated whole-cell antigen preparation cocultivated with Vero cells. Using this method of antigen preparation, the same IFA technique was previously found to be 88% sensitive and 95% specific for the serodiagnosis of another member of the genus Bartonella, B. henselae, which causes cat-scratch disease (15). This paper describes the development of an IFA test for B. bacilliformis and the subsequent performance of the test as an aid in the diagnosis of acute bartonellosis and as a diagnostic tool for epidemiologic surveys.

Antigen preparation. Two strains of B. bacilliformis, a Peruvian isolate from an area of Peru where bartonellosis is endemic (CON600-01) and an American Type Culture Collection isolate (ATCC 35685), were each cocultivated with Vero cells, to which individual Bartonella organisms readily adhere. A T-150 flask of Vero cells was inoculated with approximately 10⁶ to 10⁷ agar-grown B. bacilliformis organisms. The medium used was minimum essential medium supplemented with 10% fetal calf serum, 10 mM HEPES, 10 mM nonessential amino acids, and 2 mM L-glutamine. The cells and bacteria were incubated at 28°C in a sealed flask without additional CO₂ and harvested on day 3 postinoculation. At harvest, all but 2 ml of the medium was removed from the flask, and a sample of sterile glass beads was introduced and gently rocked to remove the Vero cell monolayer. The Bartonella-infected Vero cells were subsequently inactivated by gamma irradiation and frozen as single-use 0.2-ml aliquots at 70°C. Drops of the bacterial suspension were mounted on slides, air dried, fixed in acetone for 15 min, and, if not used immediately, stored at 70°C. Since the Peruvian isolate demonstrated higher antibody titers than the ATCC 35685 strain, it was used as the antigen in all tests.

IFA test. The IFA test was performed with twofold serum dilution steps using standard techniques (15). Fluorescein-labeled affinity-purified antibody to human immunoglobulin G (heavy plus light chains) (Kirkegaard & Perry, Gaithersburg, Md.) served as the conjugate in all tests. Incubation periods were for 30 min at 37°C. Slides were read using a 40× objective, 10× oculars, and a UV epifluorescence microscope (Olympus Optical Company LTD, Tokyo, Japan). The IFA test was scored by observing definite fluorescence of intact Bartonella bacilli, which is the standard technique for IFA testing.

Epidemiologic investigation. Study sites selected by Ministry of Health officials as being representative of areas of long-established bartonellosis endemicity were established in villages near Caraz City, Ancash, Peru (approximately 475 km northeast of Lima) in order to evaluate the usefulness of the IFA test as an aid in diagnosing bartonellosis cases. Community volunteers were asked to participate in a 2-year follow-up study designed to determine disease burden and risk factors for infection. Serum samples were obtained from 387 community volunteers and were used to estimate the point prevalence of infection in February 1998.

Preliminary test development. Twenty-eight of 33 patients with laboratory-confirmed bartonellosis (85%) had titers of 1/256 (Table 1). Titers ranged from 1/32 to 1/1,024. Paired convalescent-phase sera were available for three of the five patients with initial titers of <1/256, and there was a fourfold rise in titer in all three. Of the 101 healthy control sera, 93 (92%) had titers of 1/128. Figure 1 shows the distribution of B. bacilliformis-specific antibodies among healthy controls and patients with bartonellosis. Since the goals of this test are to aid in the diagnosis of patients with suspected bartonellosis and to serologically rule out the disease in epidemiologic surveys, using the 1/256 serum dilution end point as a positive cutoff value was chosen to yield an optimal combination of both sensitivity and specificity.

View larger version (36K): [in this window] [in a new window]   FIG. 1.   Distribution of B. bacilliformis antibodies among patients with bartonellosis (; n = 33) and healthy controls (; n = 101).

Point prevalence of infection. Of 387 volunteers from an area of bartonella endemicity in Peru, 175 (45%) were found to be seropositive for B. bacilliformis antibodies by the IFA test. Seventy-four percent of volunteers who had a history of bartonellosis within the last year had a positive IFA test, while 39% of people with a more distant or negative history of bartonellosis had a positive IFA test (prevalence risk ratio = 3.7; 95% confidence interval, 1.9 to 6.9).

Prospective application of the IFA test in acute bartonellosis patients. Eighty-six of 106 patients (81%) with laboratory-confirmed bartonellosis were initially seropositive for B. bacilliformis, with titers of 1/256 or higher. Convalescent-phase sera were available for 11 of the 20 patients whose acute-phase sera tested negative, and there was at least a fourfold rise in titer in 10 of them (91%). One patient's acute-phase serum had a titer of 1/128, and the paired convalescent-phase serum was positive at 1/256. However, since this represented only a twofold rise in titer, it was classified as a negative test.