Histoplasmosis in HIV-Infected Persons, Yaoundé, Cameroon

To the Editor: In HIV-infected persons in Cameroon (Central Africa), histoplasmosis is still misdiagnosed as tuberculosis because of clinical similarities (1,2). These patients are automatically given presumptive antituberculous therapy, although tuberculosis is not confirmed. The patients subsequently die of probable disseminated histoplasmosis (DH), and the fungal infection might finally be detected in postmortem tissue samples (3). In this context, 3 cases of DH were detected in HIV-infected patients within a 1-year period (2007–2008) in Yaounde, Cameroon. We initiated this study to investigate the occurrence of histoplasmosis in HIV-infected patients in 4 medical centers for AIDS treatment in Yaounde from December 2008 through December 2011.

isolation, as required by microbiological criteria, was not possible because a second sputum sample was unavailable.
M. sherrisii is a relatively new species (6), closely related to M. simiae. Although most of the rare M. sherrisii infections reported since 2004 (online Technical Appendix Table 1, http://wwwnc.cdc.gov/EID/article/21/11/14-1809-Techapp1.pdf) were diagnosed in Europe or the United States, about half of the strains were isolated from patients in Africa. Because M. sherrisii infection probably is further underestimated by being misidentified as M. simiae infection by the commercially available line probe assays, the hypothesis that M. sherrisii infection is not so infrequent in the African setting seems therefore reasonable. In addition, the strategy recommended by World Health Organization and based on use of immunochromatographic tests (7), does not enable NTM identification. A leitmotiv of most M. sherrisii infections reported to date is HIV co-infection, which leads to dissemination of the mycobacterial disease.
This report, although it adds to the record of patients in Africa, does not support the association with HIV infection. Our findings are consistent with the view that the pathogenic potential of M. sherrisii is comparable to that of other well-known NTM species (e.g., MAC) responsible for disease both in HIV-positive and HIV-negative patients. The retrospective determination of the MICs of antimicrobial agents potentially active against slowly growing mycobacteria (online Technical Appendix Table 2) confirmed, for the 4 strains of M. sherrisii, the well-known multidrug resistance of the species (8). The therapeutic failure was thus not surprising because clarithromycin was the only drug among those administered during the treatment that had been shown to be active in vitro. This report provides evidence that conducting appropriate microbiological investigations is essential before initiating a treatment with second-line TB drugs (9). To the Editor: In HIV-infected persons in Cameroon (Central Africa), histoplasmosis is still misdiagnosed as tuberculosis because of clinical similarities (1,2). These patients are automatically given presumptive antituberculous therapy, although tuberculosis is not confirmed. The patients subsequently die of probable disseminated histoplasmosis (DH), and the fungal infection might finally be detected in postmortem tissue samples (3). In this context, 3 cases of DH were detected in HIV-infected patients within a 1-year period (2007)(2008) in Yaoundé, Cameroon. We initiated this study to investigate the occurrence of histoplasmosis in HIV-infected patients in 4 medical centers for AIDS treatment in Yaoundé from December 2008 through December 2011.

Histoplasmosis in HIV-Infected Persons, Yaoundé, Cameroon
We recruited patients with known HIV status who agreed to participate in the study. Inclusion criteria were CD4 cells <200/mm 3 , fever and cough of >2 weeks' duration, weight loss, asthenia, and histoplasmosis-like skin manifestations (i.e., ulcerative lesions and/or umbilicated papules or nodules and/or pustules). Patients under effective antituberculous therapy or antimicrobial drugs for any skin or pulmonary infectious disease were excluded from the study. CD4 cell counts were performed in all patients. Histoplasmosis was diagnosed in sputum, bronchoalveolar fluid (BALF), and bronchial and skin biopsies by direct staining with Gomori's methenamine silver and periodic acid Schiff stains and by culture of sputum and BALF samples on Sabouraud medium. Tuberculosis and bacterial infections were detected in sputum and BALF by using Ziehl-Neelsen and Gram staining and culture on Lowenstein-Jensen and Streptococcus pyogenes media. All laboratory examinations were performed at the Centre Pasteur du Cameroun in Yaoundé. Data were collected on an anonymous questionnaire. Means (and SDs) were calculated for quantitative variables, and frequencies were calculated for qualitative variables. The National Ethics Committee, the Ministry of Health of Cameroon, and the medical centers where the study took place approved the study. Patients approved and signed the informed consent form at the time of recruitment.
Our study comprised 56 patients. Histoplasma capsulatum was detected in 7 (13%) patients on 6 of 7 skin biopsies and 1 of 3 bronchial biopsies. The median CD4 cell count of H. capsulatum-positive patients was 40 cells/mm 3 . Similarly, some authors have reported diagnosis of severe DH by using direct staining of skin samples (4); in lowincome countries, skin involvement is the main presentation of DH because of limited laboratory facilities and/or late diagnosis. In Cameroon until recently, all DH cases in HIV-infected persons were diagnosed by skin biopsy or by chance on peripheral blood smear, thus revealing AIDS at the terminal stage (3,5). We did not detect H. capsulatum infection in sputum or BALF. These results are congruent with findings in Abidjan, Côte d'Ivoire, in 1999 (6). African histoplasmosis was not detected in any sample; although this type is endemic to areas with high rates of HIV infection, it is infrequently associated with AIDS patients (7).
We detected Mycobacterium tuberculosis in 18 (32%) patients and Candida albicans in 14 (25%) patients; 3 (0.5%) patients were co-infected with M. tuberculosis and C. albicans. M. tuberculosis was detected in sputum of 9 (21%) of 42 patients and in BALF of 9 (53%) of 17 patients; we detected C. albicans in sputum of 13 (31%) patients. Our detection of M. tuberculosis in 32% of patients confirms tuberculosis as the main AIDS-defining illness in Cameroon. We did not find tuberculosis and histoplasmosis co-infection, even though it occurs frequently in lowincome countries (1,8).
The limitation in our study was the unavailability of validated sensitive and specific tools for diagnosing histoplasmosis in Cameroon (e.g., detection of the H. capsulatum circulating antigen in body fluid using an enzyme immunoassay) (9). Thus, using direct staining methods and culture of biopsies and body fluid samples could possibly lead to false-negative results.
Our detection of H. capsulatum in 13% of the HIVinfected patients in this study suggests that histoplasmosis is an unknown public health problem in Cameroon that is misdiagnosed as tuberculosis. Accounting for the endemicity of tuberculosis, which is the main HIV-defining illness in Cameroon, and the fatal outcome of DH in HIV-infected patients, practitioners need a high index of awareness to differentiate between tuberculosis and histoplasmosis. A recent report showed major clinical and biologic factors discriminating between these infections (10). Knowing these factors may lead practitioners to early diagnosis and treatment of histoplasmosis and in turn reduce the death rate among HIV-infected patients.