Aetiology and presentation of HIV/AIDS-associated pulmonary infections in patients presenting for bronchoscopy at a referral hospital in northern Tanzania.

ABSTRACT Objectives: To determine the aetiological agents of pulmonary infections in HIV-infected Tanzanians and to correlate the causative agents with clinical, radiographic features, and mortality. Design : A prospective study. Setting: Kilimanjaro Christian Medical Centre (KCMC), Tanzania. Subjects: Bronchoalveolar lavage fl uid (BAL) were obtained from 120 HIV infected patients with pulmonary infections. BAL for causative agents was analysed and correlated with clinical and radiographic features, and one-month outcome. Results : Causative agents were identifi ed in 71 patients (59.2%) and in 16 of these patients, multiple agents were found. Common bacteria were identifi ed in 35 patients (29.2%), Mycobacterium tuberculosis in 28 (23.3%), Human Herpes Virus 8 (HHV8) in 12 (10%), Pneumocystis jiroveci in nine (7.5%) and fungi in fi ve (4.2%) patients. Median CD4 T cell count of the patients with identifi ed causes was 47 cells/μl (IQR 14-91) and in the 49 patients with undetermined aetiology was 100 cells/ μl (IQR 36-188; p=0.01). Micronodular chest radiographic lesions were associated with presence of M. tuberculosis (p=0.002). The one-month mortality was 20 (16.7%). The highest mortality was associated with HHV8 (41 .7%) and M. tuberculosis (32.1%). Mortality in patients with undetermined aetiology was 11.3%. No death occurred in patients with PCP. September 2007 EA S T AF R I C A N ME D I C A L JOURNAL 421 INTRODUCTION Sub-Saharan Africa is the region of the world most affected by HIV/AIDS, more than 70% of HIV infected individuals live in this African sub-continent, (1) and the HIV pandemic is still increasing. Despite this, data pertaining to the disease presentation from this region compared to the developed world are insuffi cient and require regular updating. The clinical presentation of HIV/AIDS and the occurrence of opportunistic infections depend on different factors such as the presence of endemic diseases, quality of health services, availability of and access to antiretroviral treatment, and levels of education of the population. Pulmonary infections are the leading causes of morbidity and mortality in HIV-infected individuals (2,3). A microbiological diagnosis is of crucial importance in HIV infected patients with pulmonary infections because of atypical clinical and radiographic manifestations and high mortality (4). Because in these settings resources are limited, treatment is largely empirical, but to be effective, it should be tailored to locally prevailing causative organisms. Therefore health facilities equipped with advanced diagnostic techniques should monitor the trend of diseases and obtain data on aetiological pathogens, so that empirical treatment regimens can be improved. The aim of this study was to fi nd the aetiological pathogens of pulmonary infections in HIV/AIDS patients who presented with features of pulmonary infection for bronchoscopy. Furthermore, we correlated the pathogens identified with clinical and radiographic presentations, level of immunity, and outcome after one month. MATERIALS AND METHODS At Kilimanjaro Christian Medical Center (KCMC) in northern Tanzania we enrolled 120 HIV infected patients aged 18 years and above, who presented with features of chest infection such as cough (dry or productive) and/or chest pain, dyspnoea, fever and chest radiographic abnormalities. These patients were recruited at the endoscopy unit where bronchoscopic procedures are done. The patients were referred by their attending clinicians for bronchoscopy as part of patient’s management after failing to establish the causative agent by other methods and/or following failure to respond to empirical treatment. Pregnant women and patients with oxygen saturation less than 90% under 61/mm of oxygen were excluded from the study. Bronchoscopy and bronchoalveolar lavage (BAL) was performed by standard procedure using fl exible fi beroptic bronchoscope. Briefl y, the bronchoscope was wedged in one of the heavily involved segmental bronchi as seen on the chest radiograph. In case of diffuse lung involvement, the scope was wedged in one of the segmental bronchi of the right middle lobe. Then aliquots of 50 ml or less of sterile saline at body temperature, up to a maximum of 150 ml, were instilled and at least 40 ml was sucked back into a sterile container. The sediments of BAL fl uid obtained by centrifugation at 1500 g for 10 minutes were used for laboratory tests to identify the aetiological agent of the chest infection. For identifi cation of M. tuberculosis, BAL samples were decontaminated with 4% NaOH, centrifuged and cultured using in-house made Lowenstein- Jensen (LJ) solid medium, with a maximum of eight weeks incubation. Diagnosis of M. tuberculosis was based on positive culture results showing acid fast bacilli (AFB) in ZN stain. Diagnosis of pneumocystis pneumonia (PCP) was based on identification of Pneumocystis (P) jiroveci with at least one of the following techniques: Giemsa stain, Gomori methenamine silver (GMS) stain or immunofl uorescence test (IFT). Also real time PCR was performed for detection of P. jiroveci DNA in the BAL fl uid, using 40 as the cycle threshold (CT) cut off value (5). Conventional PCR was used for the diagnosis of Human herpes virus 8 (HHV8) as described by Chang and collaborators (6). Fungi were diagnosed by direct smear and culture for seven days using Sabouraud dextrose Conclusion: In this population of severely immunosuppressed HIV-infected patients with pulmonary infection a variety of causative agents was identifi ed. Micronodular radiographic lesions were indicative of TB. High mortality was associated with M. tuberculosis or HHV8. No death occurred in patients with P. jiroveci infection.