Screening for tuberculosis prior to isoniazid preventive therapy among HIV-infected gold miners in South Africa.

London School of Hygiene and Tropical Medicine, London, United Kingdom.
The International Journal of Tuberculosis and Lung Disease (Impact Factor: 2.76). 06/2006; 10(5):523-9.
Source: PubMed

ABSTRACT Human immunodeficiency virus (HIV) clinic for employees of a gold mine, Free State, South Africa.
To evaluate the process of screening for active tuberculosis (TB) prior to commencing TB preventive therapy in HIV-infected individuals.
Cross-sectional study comparing performance of various combinations of screening tests for TB against a gold standard diagnosis of TB based on symptoms, chest radiograph (CXR), sputum microscopy and culture.
Of 899 individuals, 44 (4.9%) had TB. The most sensitive symptom combination (59.1%) was any of night sweats, new or worsening cough or reported weight loss; measured weight loss > 5% or abnormal CXR increased sensitivity to 90.9%. Sputum microscopy did not increase sensitivity further, but including World Health Organization HIV clinical staging or CD4 count did. As the specificity of all these combinations was low, many individuals required further investigation to rule out TB. TB prevalence was high (11.7%) among individuals with a CD4 count < 200/mm3.
CXR greatly increased the sensitivity of screening for TB in this population. Sputum microscopy conferred no additional benefit among asymptomatic patients with a normal CXR. The high prevalence of TB amongst those with a low CD4 count underlines the importance of screening for active TB prior to commencing TB preventive therapy, and before antiretroviral therapy.

1 Bookmark
  • [Show abstract] [Hide abstract]
    ABSTRACT: The objective of the present study was to determine the diagnostic performance of the symptom-based tuberculosis (TB) screening questionnaire recommended by WHO for people living with HIV (PLWH) in resource-limited settings, among adults off and on antiretroviral therapy (ART). Cross-sectional study at two HIV clinics in South Africa. A total of 825 PLWH completed the screening questionnaire and underwent investigations [chest radiography (CXR) and microbiologic testing of sputa]. A positive screen was defined as presence of cough, fever, night sweats, or weight loss. Pulmonary tuberculosis (PTB) was defined as sputum smear positive for acid-fast bacilli or growth of Mycobacterium tuberculosis. Of 737 participants with at least one diagnostic sputum specimen, PTB was diagnosed in 31 of 522 (5.9%) on ART, and 34 of 215 (15.8%) not on ART. The questionnaire missed 15 of 31 (48.4%) PTB cases on ART, and three of 34 (8.8%) not on ART. Among participants on ART, post-test probability of PTB diagnosis (95% confidence interval) was 6.8% (4.0-10.9%) if screening positive, and 5.2% (2.9-8.4%) if screening negative, whereas among participants not on ART, post-test probabilities were 20.3% (14.2-27.5%) and 4.8% (1.0-13.5%), respectively. Among participants diagnosed with PTB, those on ART were significantly less likely to screen positive (adjusted odds ratio 0.04, 95% confidence interval: 0.01-0.39). In both groups (ART and no ART), screening was more sensitive when CXR was incorporated. For case detection and exclusion of PTB, the WHO-recommended questionnaire performed adequately among PLWH not on ART, and poorly among those on ART. Further research is needed to identify feasible and effective TB screening strategies for PLWH in resource-limited settings.
    AIDS (London, England) 03/2014; 28(10). DOI:10.1097/QAD.0000000000000278 · 6.56 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objectives We assessed the usefulness of the National TB and Leprosy Control Program (NTLP) symptom-based tuberculosis (TB) screening tool in identifying HIV-infected patients eligible for isoniazid preventive therapy in Muhimbili National Hospital, Dar es Salaam Tanzania.Methods Descriptive cross-sectional study. Data collected included socio-demographic and clinical data. Chest X-ray, sputum for acid-fast bacilli (AFB) microscopy, mycobacterial culture, CD4 + count and complete blood count were performed. Patients were considered not having active TB if they presented with no symptom in the screening tool, which comprised these symptoms: cough, fever and excessive night sweats for ≥2 weeks; weight loss of ≥3 kg in 4 weeks and haemoptysis of any duration. The reference standard was a negative culture for Mycobacterium tuberculosis.ResultsWe enroled 373 patients, of whom 72.1% were females. Active pulmonary TB was found in 4.1% (14/338) of the participants as defined by a positive culture. The sensitivity and specificity of the NTLP screening tool were 71.4% (10/14) and 75.9% (246/324), respectively. False-negative rate was 28.6% (4/10). Cough, fever for ≥2 weeks and weight loss were independent predictors of NTLP-defined TB. Cough ≥2 weeks predicted TB when a positive culture was used to define TB.Conclusion The screening tool had fairly good sensitivity and specificity for TB screening; however, there is a possibility that about 29% of the screened population will be given IPT while they are supposed to receive a full course of TB treatment.
    Tropical Medicine & International Health 03/2014; 19(6). DOI:10.1111/tmi.12307 · 2.30 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Setting: Urban, non-governmental HIV outpatient clinic in Mombasa, Kenya. Objective: To report outcomes and assess feasibility of serial clinical screening for active TB among adults enrolled in outpatient HIV care in a resource-limited setting. Design: Longitudinal analysis of screening conducted during routine clinic visits of HIV-infected Kenyan adults. The provider-initiated screen included TB symptom assessment and targeted physical exam. Participants with > 1 symptom/sign were to submit sputum for microscopy and undergo chest radiography. Results: Over 33 months, 4,854 HIV-infected outpatients were serially screened for active TB at a median interval of 3 months. Treatment for active TB was started in 127 (2.6%). Of those 127, 77 (60.6%) were diagnosed based on first screen, and 50 (39.4%) were diagnosed thereafter. Among those 50 diagnosed upon subsequent screens, 28 (56%) were identified in association with positive screens, suggesting that 22% (28 of 127) of TB diagnoses could be attributed to the serial screening protocol. Conclusion: Provider-initiated serial clinical screening during routine visits of HIV-infected outpatients continued to prompt treatment of active TB beyond initial screening. Serial screening strategies may lead to earlier TB treatment in patients receiving ongoing HIV care in resource-limited settings.


Available from