Pulmonary tuberculosis and pneumocystis jiroveci pneumonia in HIV-infected patients in Ethiopia

Abstract: The objective of this study initially was to determine the prevalence of culture-verified pulmonary tuberculosis in TB suspects and investigate the impact of human immunodeficiency virus (HIV) infection on the prevalence, clinical and radiological presentation and diagnosis of tuberculosis. During the study, it soon became clear that the HIV sero-prevalence in tuberculosis (TB) suspects who could not be verified to be culture-positive was too high to deserve an explanation. The second objective, therefore, was to prospectively look for possible causes of other pulmonary opportunistic infections including pneumocystis pneumonia (PCP) to explain the excess HIV. In paper I, among 509 consecutive PTB suspects attending the outpatient department of a university hospital in Addis Ababa, 33% could be culture-verified as having PTB. PTB patients, non-TB patients (culture-negative PTB suspects) and controls were HIV-1 positive in 57.1%, 38.5% and 8.3% of cases respectively. Independent predictors of culture-verified TB were age <25yrs, male gender and the presence of HIV and fever whereas profound weight loss indicated HIV infection. Diagnosis of PTB based on clinical symptoms, direct sputum microscopy and chest radiography (CXR) was significantly less sensitive and specific in HIV+ patients. In paper II, a look at the relationship between disease pattern and disease burden by chest x-ray, mycobacterial load and HIV infection also demonstrated that: (1) atypical chest x-ray with interstitial infiltrates, pleural effusion, miliary mottling, normal CXR and absent cavitations occurred more frequently in HIV-infected than non-HIV-infected patients (2) Mycobacterial load as assessed by the number of colonies of Mycobacterium tuberculosis (MTB) culture was significantly less in HIV-infected patients than non-HIV-infected, (3) the occurrence of high number of culture-verified PTB cases with normal CXR was identified as one of the challenges in the diagnosis of PTB in patients with HIV infection. In paper III, amongst 119 culture-negative, HIV+ TB suspects, P. jiroveci was detected in 10.9% by single polymerase chain reaction (PCR) and immunofluorescence (IF) and 30.3% by nested PCR in expectorated sputum sample. In HIV- negative TB and Non-TB patients, the prevalence of P.carini was significantly lower. Besides, in the IF-positive and nPCR-positive HIV+ non-TB patients,more than 40% were interpreted as PTB by CXR whereas only one patient was diagnosed with clinical PCP. This observation prompted us to design a follow up prospective study to find out the relative importance of P. jiroveci and other pulmonary opportunistic diseases in smear-negative patients presenting with atypical chest x-rays. In paper IV, 131 consecutive HIV-1 infected patients presenting with respiratory symptoms and atypical chest x-rays who were sputum smear-negative for AFB and sero-reactive for HIV were enrolled into the study. They underwent clinical evaluation and investigation for P. jiroveci (using TBO and IF stain) and M. tuberculosis from expectorated sputum and BAL samples and fungal and bacterial culture from BAL alone. The results of this study showed that the prevalence of PCP was 29.8%, that of bacterial infection 33.6% and tuberculosis 23.7%. Pulmonary Kaposi sarcoma (PKS) and interstitial pneumonitis (NIP) occurred in 4 patients each. Double infection occurred in 18(13.7%) patients. Cryptococcal pneumonitis was conspicuously absent in this study population. In paper V, we evaluated the usefulness of a simple diagnostic method, Toluidine Blue O stain for the diagnosis of P. jiroveci in expectorated sputum sample and bronchoalveolar lavage (BAL) and compared it to immunofluorescence and PCR. Comparison of these diagnostic tests showed that the sensitivity of TBO in sputum and BAL samples was 71.4 % and 68% compared to immunofluorescence respectively. The overall sensitivity for the diagnosis of PJ was 42.7 % by PCR, 29.8% by IF and 20.6% by TBO. PCR was the most sensitive test and detected additional 18 patients than immunofluorescence. Considering cost, simplicity and efficacy, we recommend TBO as the most practical diagnostic test and expectorated sputum (ES) as the most practical biologic specimen for use in resource-constrained, high HIV-settings such as Ethiopia.