Seroprevalence of Bartonellaspp. infection in HIV patients in Catalonia, Spain

The spectrum of Bartonella infections has expanded rapidly since the first HIV- infected patient with unusual, vascular proliferative lesions of bacillary angiomatosis (BA) was described in 1983 [1]. Of the 19 species of Bartonella described until now, only 10 were acknowledged as human pathogen species; B. bacilliformis, B. quintana, and B. henselae, are the most frequently described species [2–4], while B. elizabethae, B. vinsonii, B. washoensis, B. grahamii, B. clarridgeiae, B. koehlerae and B. alsatica were recently identified as responsible for a few cases of human infections [5–10]. Cat scratch disease (CSD), hepatic peliosis and some cases of bacteraemia and endocarditis are directly caused by some species of the genus Bartonella [11–14]. To determine the real incidence of Bartonella infections, we must study the seroprevalence in the general population and the principal reservoirs and vectors of infection transmission. The results yielded by different studies on seroprevalence vary depending on the type of population under study; thus, the research conducted on collective groups that present special characteristics or associated risk factors present a higher prevalence than that found in studies carried out on the normal population.

In patients with addiction to parenteral drugs, it ranged from 15% to 47.5% [15, 16]; in patients with HIV infection, it varied between 17.3% and 40% [17, 18]. The objective of this study was to evaluate the prevalence of Bartonella infections in patients with HIV in our catchment area and to assess related factors.

24.4% of women and 21.7% of men presented antibodies to Bartonella spp. A statistically significant increase of seropositivity against Bartonella spp. was observed as patient age increased (p < 0.05). Twenty-three (30.2%) patients with positive serology for Bartonella presented a past history of alcohol abuse. Of the 76 patients with positive serology, 47 (61.8%) were addicted to parenteral drugs and 22 (28.9%) had, at some time, being diagnosed with AIDS. Thirty-nine patients presented co-infection between Bartonella and HCV, whereas in 3 patients it was between Bartonella and HBV. No differences were found regarding the way of transmission of the human immunodeficiency syndrome.

Preliminary estimates of the prevalence of antibody to Bartonella among apparently healthy humans range from 5.88% to 24.7% [17, 20]. In a study carried out on a healthy population sample from the same area (83 were men and 78 women, and the mean age with positive serology was 45.18 ± 14.26 years), the seroprevalence was 10.6% [21], a figure that is very similar to that reported in other studies. The incidence of Bartonella-associated disease among HIV- infected people is less known. The studies carried out show seroprevalence rates that range from 16 to 40% [18, 22].

Epidemiological studies identified the major risk factors for acquiring Bartonella infections. B. quintana infection is associated with exposure to the human body louse, Pediculus humanus [13, 23], and homeless conditions [23, 24], whereas the major risk factor associated with B. henselae infection is cat exposure, especially cat scratches and fleas [13, 25–27]. B. henselae seroprevalence in cats in our area was 29.6%, this result confirmed the presence of infection in the main reservoir of this microorganism [28].

B. henselae infection, as in other infections caused by fastidious bacteria, gives rise for concern given the complexity of the diagnosis and the variety of presenting clinical pictures, mainly in HIV-infected patients [29]. In fact, coetaneous BA lesions can be clinically indistinguishable from Kaposi's sarcoma, being differentiated by biopsy only. Hepatic disease caused by Bartonella can be indistinguishable from others infectious or malignant conditions that cause hypodense lesions demonstrable by abdominal CT [30]. All these factors make serology as a useful tool for the diagnosis of acute infections caused by Bartonella spp. However, it must be taken into account that there can be important limitations when dealing with HIV infected patients [31, 32]. The serologic response in patients HIV with a good immunological status could be similar to that of the normal population, whereas possibly in situations of advanced immunodepression or in bacillary angiomatosis, the serologic response could show as a false negative [33].

Our study highlights in the first place an antibody seroprevalence of 22.4% to Bartonella species in HIV patients, which is superior to that observed in the healthy population [13, 19, 25], with the exception of that described in some other studies [25]. Some authors consider HIV patients as a risk group for Bartonella infection [34, 35] (characteristic personal and hygienic habits); however, other authors do not consider this infection risk to be greater [20, 34].

In our experience, a previous study carried out on a healthy population sample from the same area show a seropositive rate of 8.6% [21]. This study included 55 children and 161 adults (≥ 18 years). If children were not considered, the seroprevalence was 10.6%. In view of our HIV-infected patients, predominantly adult, seems to be higher, it could be considered HIV-infection as a risk factor.

Taking into account other studies, this difference among populations might be due to the fact that in HIV patients the infection can last longer or that these patients might become in contact more easily with some of the factors that have a direct impact on increased Bartonella infection, such as contact with cats (main reservoir of B. henselae) [16, 17, 24] or that they spend most of their time on the street [24, 34]. However, these informations are not available in our study. We could attribute the high rate of seropositivity observed to false positives that are not directly due to HIV infection (to date no crossed- reactions have been observed between both); however, there are studies describing crossed reactions with Coxiella burnetii [32, 35]. On the other hand, other studies suggest that HIV-infected patients present a higher risk of Q fever than immunocompetent patients [36]. The different factors that might play an important role in the transmission of Bartonella spp must also be taken into account. 57.6% of seropositive patients were intravenous drug users, these findings are most likely a result of several conditions that are frequent among IVDUs: living conditions with a high probability of close contact with bacteriaemic cats, fleas, lice, and other potential vectors, repeated parenteral exposures, insufficient medical care, and malnutrition [19, 33, 34, 37]. Other studies do not conclude that drug use might play such an important role [20].

An excessive consumption of alcohol constitutes an elevated risk factor for infection, not only for its capacity to physiologically deteriorate regular alcohol abusers (liver alterations, neurological disorders, etc.), but because in the majority of cases drinking can affect people at low socioeconomic levels (homeless) who might be in closer contact with infection-transmitting vectors [38, 39].

Of special interest in these patients was to find out whether the risk for infection increased with immunodepression. In our experience, we were unable to detect a greater incidence among patients who had been diagnosed with AIDS at some given time. In like manner, we were also unable to detect a greater infection incidence in patients with levels < 200 lymphocytes CD4/μl, in contrast with what has been reported by some authors who claim that the prevalence of antibodies against B. henselae is inversely proportional to the number of CD4 lymphocytes [30, 35].

Our study presented only a 13.8% of patients with low CD4 levels, which could significantly alter our results. If we take into account that HIV-induced immunodepression does not appear to significantly modify the serological response to Bartonella spp., serology could be useful in the diagnosis of diseases associated with these microorganisms in HIV-infected patients (acute infection and population-based epidemiological studies). PCR techniques might be of help to diagnose Bartonella infection in patients who are more immunodepressed, especially when the infection is subclinical. Bartonella infections are substantially underrecognized because of their non-specific symptomatology.

In conclusion, the seroprevalence of Bartonella spp. among HIV infected patients is greater than that of the healthy population of the same area and thus Bartonella infection should be considered in HIV patients.