Publication Details

AFRICAN RESEARCH NEXUS

SHINING A SPOTLIGHT ON AFRICAN RESEARCH

medicine

Sputum smear concentration may misidentify acid-fast bacilli as mycobacterium tuberculosis in HIV-infected patients

Journal of Acquired Immune Deficiency Syndromes, Volume 63, No. 2, Year 2013

Background: Tuberculosis (TB) diagnosis in most resource-limited settings still depends on smear microscopy for identification of acidfast bacilli (AFB). However, recently developed molecular diagnostics that test for the presence of Mycobacterium tuberculosis (Mtb) DNA have been shown to be superior for confirmation of TB diagnosis. Methods: At regular clinical visits over a 12-month period, we collected sputa from HIV-infected patients presenting with signs or symptoms of TB at 2 Nigerian clinics. Sputa were stained for AFB and tested using the Genotype MTBDRplus to confirm the presence of Mtb. Other species were identified using 16S rRNA sequence. Results: In 56% (233/415) of AFB-positive patients, Mtb was confirmed. The patients on antiretroviral therapy were less likely than those not on antiretroviral therapy to be infected with Mtb [odds ratio (OR) = 0.25, P = 0.003]. In a multivariate logistic regression model using clinical features and diagnostic results, abnormal respiratory findings on auscultation (OR = 3.28, P = 0.03) and a direct sputum smear grade .3/100 (OR = 6.4, 4.6, P , 0.02) were significant predictors of Mtb infection. Concentrated sputum smear was predictive of Mtb infection only at the highest grades (2+, 3+). Interestingly, among 65 samples that could not be confirmed for Mtb, 32 (49%) were found to contain other, possibly novel, actinomycetes, including atypical Mycobacteria, Rhodococcus spp, Nocardia spp, and Corynebacterium spp. Conclusions: We conclude that concentrated sputum smears may misidentify other bacteria as Mtb in HIV-infected patients. The use of molecular diagnostics could reduce unnecessary or inappropriate treatment and improve identification of pathogens in resource-limited settings with high HIV burden. Copyright © 2013 by Lippincott Williams & Wilkins.
Statistics
Citations: 10
Authors: 10
Affiliations: 5
Research Areas
Genetics And Genomics
Infectious Diseases
Study Design
Case-Control Study