Prevalence of Bartonella spp. Infection in Patients with Sickle Cell Disease

Introduction

Sickle cell disease (SCD) is the most common hematologic genetic disorder affecting millions of people worldwide. In Brazil, up to 10% of Afro-descendent population has the sickle gene (Cançado and Jesus 2007). SCD is associated with unpredictable, recurrent, acute vaso-occlusive (VOC) episodes with pain, inflammation, and oxidative stress, requiring frequent transfusions. Pain due to VOC is the hallmark of SCD and difficult to treat, requiring frequent hospitalization, distress, and poor quality of life (Kapoor et al. 2018). It can be triggered by temperature extremes, dehydration, emotional stress, and infection, but often there is no identifiable cause (Kapoor et al. 2018). Therefore, SCD patient infections can be related to severe pain, mainly pneumococcal infections in children and, in Brazil, hepatitis C, B viruses and Human T-lymphotropic virus type 1 (Neto et al. 2011, Brousse et al. 2014, Tewari et al. 2015). The main cause of death in children with SCD in two studies in Brazil was infection (Lobo et al. 2018, Rezende et al. 2018).

Bartonella spp. are emergent and neglected bacteria worldwide. They are responsible for—bartonellosis in Peru, trench fever, cat scratch disease, bacillary angiomatosis, chronic bacteremia, and endocarditis and can be potentially fatal, especially in immunodeficient patients (Chaudhry et al. 2018). Typical characteristics of Bartonella spp. infection such as hemolysis, bacteremia, fever, malaise, endothelial activation, and inflammation (Breitschwerdt 2014, Chaudhry et al. 2018) are also characteristic features of SCD. Thus, Bartonella spp. infection may be challenging to distinguish from the pathological symptoms of SCD. It is possible that chronic infection by these bacteria may contribute to the occurrence of VOC.

Bartonella spp. infection in humans is associated with exposure to animals such as cats and dogs (Breitschwerdt 2014). Using only one PCR assay, we found that 3.2% of blood donors in Campinas, Brazil, have asymptomatic bloodstream infections due to Bartonella spp. (Pitassi et al. 2015). Bartonella henselae can survive prolonged blood storage periods (Magalhães et al. 2008). In addition, frequent blood transfusions may increase the risk of bacterial transmission from infected donors to patients with SCD requiring periodic transfusions.

While blood transfusions are usually necessary in SCD patients, these patients may suffer from infections, and bartonella is an intraerythrocytic bacterium that can be found in donated blood.

The goal of this study was to evaluate the prevalence of Bartonella spp. infection in SCD patients' blood and the influence of risk factors or blood transfusion in these patients.

Materials and Methods

One hundred seven patients with homozygous (HbSS) or heterozygous (HbSC) patients (collectively called SCD henceforth) were invited to participate in this study when arriving for a routine visit at the Outpatient Clinic of the University of Campinas Blood Bank, HEMOCENTRO/UNICAMP. Informed written consent was obtained from all participants and the study was approved by the UNICAMP Institutional Research Board (CEP 145.233) in accordance with national guidelines.

An epidemiological questionnaire was given to each patient before blood sample collection. Only 89 of 107 patients answered the questionnaire. The remaining 18 patients refused to answer the questionnaire, stating lack of time. However, their samples were included in the study. The questionnaire included questions on demographics such as gender, age, ethnicity, and the general area where they lived (city or rural). Information on socioeconomic questions, patients' interactions with animals, and history of animals and arthropod bites was also requested.

Blood samples were collected in an EDTA tube followed by freezing lyse cells at −20°C and storage. A previously described approach that combines PCR detection of Bartonella spp. DNA and enrichment blood culture was used to test EDTA-anticoagulated whole blood with modifications (Bloch et al. 2012).

Bartonella bloodstream infection was defined by the presence of bartonella DNA in the patient samples: whole blood and/or 14-day blood culture in an enriched liquid growth medium (ELGM). This medium was produced by supplementing 900 mL of insect growth medium IPL-41 (Sigma-Aldrich) with 2 mg of sodium pyruvate (Sigma-Aldrich), 2 g of yeast extract, and a source of energy, namely 0.1 mg of NAD (Sigma-Aldrich), 1.25 mg of NADP (Sigma-Aldrich), and 2 mg of ATP (Sigma-Aldrich). After adjusting the ELGM pH to 7.4, it was sterilized using Stericup Filtration, pore size 0.22 μm (Millipore). No sheep blood was used for ELGM supplementation as described by Maggi et al. (2005), since the blood of the sample itself could provide a favorable growth medium for bartonella.

The samples in ELGM were incubated at 37°C and maintained with a constant shaking motion. After 14 days, an aliquot was subinoculated onto a plate with Bordet Gengou Agar Base (LAB) enriched with 30% of total sheep blood, previously testing negative for bartonella, which was incubated at 37°C for 45 days (Drummond et al. 2011).

DNA extraction was performed using QIAmp^® DNA Mini kit (QIAGEN, Inc.), according to the manufacturer's instructions. All procedures—DNA extraction, PCR preparation, PCR amplification, and electrophoresis—were performed in separate rooms with a unidirectional workflow to avoid DNA contamination as previously described (Pitassi et al. 2015). We performed PCR testing at two points, each representing a different component of the testing process for each patient sample: (1) DNA extracted from whole blood and (2) DNA extracted after blood culture in the ELGM. Bartonella DNA was amplified from blood and ELGM cultures using conventional Bartonella genus PCR primers targeting the 16S–23S rRNA Intergenic Transcribed Spacer (ITS) (Diniz et al. 2007). These samples were also submitted to nested PCR that amplifies the FtsZ region and is specific to B. henselae (Drummond et al. 2018). A second conventional PCR targeting another region, the ssrA region, was done for ITS and/or FtsZ ELGM-positive samples (Diaz et al. 2012). All amplicons obtained, for each of gene target analyzed, were sequenced. The positive control in all PCR assays were used B. henselae ATCC 49882^T, all reaction also heads negative controls.

Exploratory data analysis was performed through summary measures (frequency) of categorical data and descriptive statistics of quantitative data. To analyze the risk factors for positivity, the analysis of univariate logistic regression was used. The significance level adopted for this study was 5%. Statistical analysis was performed using SAS System for Windows (Statistical Analysis System) 9.4. SAS Institute, Inc., 2002–2008 (Cary, NC).

Results

Table 1 shows an overview of the results. Ten of the 107 patients were positive for B. henselae infection in at least one molecular test. Two patients had bartonella DNA detected by conventional PCR (ITS) from blood and, in one case, bartonella DNA was detected by nested PCR. The other eight patients tested positive for conventional PCR (ITS) after ELGM blood culture and five were also nested PCR positive. Seven patients of these eight were also positive for ssrA by conventional PCR. There was no growth of bacterial colonies on agar plates.

The logistic regression test showed no association between any of the risk factors addressed in the questionnaire and bloodstream infection with B. henselae (Table 2). Based on statistical results, association between presence of bartonella DNA and animal and arthropod contact was not found. The bloodstream infection with B. henselae also was not associated with socioeconomic conditions or blood transfusions.

All amplicons sequenced, from the three target gene, were 100% similar to B. henselae sequences deposited in GenBank (accession no. BX897699).

Of the SCD studied patients, 9.3% demonstrated B. henselae bloodstream infection, primarily urban females with animal contact, but without significant risk factors for bartonella infection.

Discussion

Pain crisis is the hallmark of SCD and can be a result of infection (Vera et al. 2014). Bartonella sp. infection had already been related to pain in SCD mice (Almeida et al. 2019). It is possible that infection is also related to pain in SCD patients (Almeida et al. 2019). Because of its fastidious nature, bartonella is extremely difficult to isolate and diagnose. In an effort to improve the diagnosis of this infection, current techniques include a 2–8 week culture of patients lysed blood cells, performed initially on a specific liquid medium (ELGM) for 7–14 days and plated on agar enriched with rabbit, sheep, or horse blood or cocultured with endothelial cells (Maggi et al. 2011). However, there is still a relevant chance of false negative results since bartonella bacteremia is cyclic and negative analysis may not preclude bartonella bloodstream infection (Breitschwerdt 2014).

Evidence of cat contact and tick bite was found in the blood donor study performed in the same facility HEMOCENTRO/UNICAMP (Pitassi et al. 2015), but these two risk factors were without significance in SCD patients. In these patients, blood transfusion was also not associated with B. henselae infection. However, Oteo et al. (2017) demonstrated high serological and molecular prevalences of exposure to or bloodstream Bartonella spp. infection in healthy veterinarian professionals from Spain as Pitassi et al. (2015) found in blood donors. These data reinforce the possibility that blood transfusions be related to Bartonella sp. transmission.

Asymptomatic individuals, such as in Staphylococcus aureus colonization, perhaps should not be treated for Bartonella sp. infection, but the treatment should be considered if the chronic asymptomatic infection may be related to pain crises, for example. The infected patients were notified and the option of treatment was offered to them.

Conclusion

The low detectability by most conventional, nested, or even real-time PCR assays imposes a technical limitation to laboratory diagnosis (Pitassi et al. 2015), calling for the necessity of developing tools for more sensitive diagnostic tests and for detecting the real prevalence of bartonella infection in SCD patients. Even though we found almost 1 in each 10 SCD patients with demonstrable B. henselae bloodstream infection, the clinical relevance of B. henselae infection in SCD patients requires further investigation.