Coxiella burnetii antibody seropositivity is not a risk factor for AIDS-related non-Hodgkin lymphoma

To the editor:

Coxiella burnetii is an important human pathogen and the causative agent of Q fever, a disease that can lead to life-threatening endocarditis and other serious conditions. About 3% of the US population has antibody evidence of exposure to this agent, which can persist for months to years after infection. Recently, chronic C burnetii infection has been linked to the development of non-Hodgkin B-cell lymphoma (NHL).¹  This study of 1468 patients from the French National Referral Center for Q Fever database found a 25-fold increase in the risk of NHL development among Q fever patients relative to the general population. The link between Q fever and NHL development is believed to be related to plasmacytoid dendritic cell infection within lymphoid tumors and interleukin-10 (IL-10) overproduction.

HIV is also linked to NHL as progression to AIDS puts patients at increased risk for development of NHL subtypes.^2,3  A link between HIV and increased C burnetii seroprevalence has been suggested. Two independent studies in France and Brazil found increased C burnetii seroprevalence in HIV-positive individuals relative to the general population.^4-6  Conversely, a study in Spain found seroprevalence among HIV-infected IV drug users similar to that of HIV-negative IV drug users.⁷  This same study and work in Tanzania found the percentage of HIV-positive individuals within cohorts of acute Q fever patients to be comparable to the general population.^7,8  Whether HIV infection alters antibody responses to C burnetii infection or plays a role in increased susceptibility or disease severity is not known.

The finding that HIV and chronic C burnetii infections are both risk factors for development of NHL, coupled with the fact that certain groups of HIV-positive individuals demonstrate increased C burnetii seropositivity, suggests a potential role for C burnetii in AIDS-related NHL. Therefore, we hypothesized that anti–C burnetii antibody seroprevalence would be higher in individuals with AIDS-related NHL relative to the general HIV-positive population. To test this, we performed a nested case control study using stored serum and plasma samples from 2 previously published prospective cohort studies, the District of Columbia/New York Gay Men’s Cohort (DCG)^9-11  and the AIDS Cancer Cohort Study (ACC).^10-15  Patients with incident (n = 29) and prevalent (n = 39) AIDS-NHL were matched by age, sex, and CD4 count to 67 HIV-positive controls (Table 1). Prelymphoma samples had been collected a median of 1.1 (interquartile range, 0.6-1.5) years prior to NHL diagnosis or from equivalent time points for controls. Institutional review boards at the National Cancer Institute and collaborating institutions approved each study, and all participants gave written informed consent.

Samples were analyzed for the presence of phase II anti–C burnetii immunoglobulin G antibodies using an enzyme-linked immunosorbent assay (ELISA; Virion/Serion, Wurzburg, Germany) according to the manufacturer’s instructions with the following modifications. Samples were diluted 1:100. Optical densities were read at 405 nm using an ELx800 (BioTek). Positive and borderline samples were additionally screened for both phase I and phase II anti–C burnetii immunoglobulin G by immunofluorescence assay (IFA), as described previously.¹⁶  Antibody titers ≥1:16 against phase I or phase II antigens were considered positive. Assay reproducibility was excellent based on 16 masked replicates, of which 15 were concordantly ELISA negative and 1 was concordantly borderline with IFA titers within a twofold dilution.

Nine percent (6 of 67) of HIV-positive controls were positive for anti–C burnetii antibodies, as compared with 7% (2 of 29) of incident and 8% (3 of 39) of prevalent AIDS-NHL cases. We used multinomial logistic regression to determine the associations between C burnetii seropositivity and incident or prevalent AIDS-related NHL, as well as binary logistic regression for both NHL outcomes combined. All regression models were adjusted for age at blood collection, sex, race/ethnicity (ie, white, African American, or Hispanic), CD4 count per µL and cohort (ie, DCG or ACC). Two-sided P values < 0.05 were considered statistically significant. All analyses were performed by using Stata version 13 (Stata Corp, College Station, TX). The adjusted odds ratios (ORs) were not statistically significant: 0.85 and 1.04 for incident and prevalent AIDS-NHL, respectively (Table 2). The overall seroprevalence combining incident and prevalent cases was 7% (5 of 68) for a nonsignificant OR of 0.82. These serologic findings indicate that the development of AIDS-associated NHL is not associated with exposure to C burnetii.

C burnetii infection of plasmacytoid dendritic cells within lymphoid tissues and IL-10 overproduction support the development of NHL.¹  HIV infection also leads to significantly higher IL-10 serum levels.¹⁷  As such, it is possible that IL-10 overproduction as a result of C burnetii infection in HIV-positive individuals is negligible relative to the HIV-induced cytokine response. Furthermore, HIV infection leads to depletion of plasmacytoid dendritic cells,¹⁸  perhaps impeding C burnetii infection of these cells and hindering C burnetii–related NHL development. Perhaps C burnetii is unable to elicit a comparable cytokine response in HIV-positive individuals relative to healthy hosts owing to HIV infection–related immune system dysfunction. As such, the disease characteristics that lead to NHL in healthy hosts with Q fever may be altered in HIV-positive individuals. The possibility exists that NHL-related B-cell dysfunction may lead to diminished antibody titers against C burnetii. However, previously published vaccine studies suggest that although antibody responses may be less robust than in healthy individuals, NHL patients are still able to produce antibodies to influenza.^19,20  We anticipate that measuring the presence of C burnetii antibodies rather than comparing antibody titers reduces the possibility that NHL-related B-cell dysfunction influenced our results.

Our findings suggest a potential for increased prevalence of anti–C burnetii antibodies among HIV-positive individuals in the United States compared with the general population. Of the 135 samples tested, 11 (8.1%) were positive for anti–C burnetii antibodies, which is higher than the 3.1% among adults in the National Health and Nutrition Examination Survey in the United States during 2003 to 2004.²¹  These findings support the French and Brazilian studies, which found the presence of anti–C burnetii antibodies to be higher in HIV-positive individuals.^4-6  Collectively, our findings suggest that rates of infection with C burnetii are similar among HIV patients with and without NHL, indicating that C burnetii infection is not a risk factor for NHL in this setting.