Validity of Acid-Fast Smear of Gastric Aspirates for the Diagnosis of Childhood Pulmonary Tuberculosis among Human Immunodeficiency Virus-Infected Children

Background/Aims: Gastric aspirate specimen is accepted as an alternative specimen of choice to sputum in the diagnosis of childhood pulmonary tuberculosis (TB); and specimen microscopy is usually the only available bacteriologic confirmatory test for streaming cases into the National TB Control Programme treatment protocol. Doubts expressed about the continued relevance of this test among the HIV-infected are based on observations that the bacteriologic yield of acid-fast smears of gastric aspirate specimens from these patients is markedly reduced. This study is aimed at determining the validity of acid-fast smear of gastric aspirates among the HIV-infected and compare with those of the HIV-uninfected.
Design, Place and Duration of Study: Diagnostic study. Suspected tuberculosis patients registered in the paediatrics department of University of Benin Teaching Hospital were prospectively recruited from January 2010 to March, 2011.
Methodology: A total of 263 children were recruited. Voluntary counseling and testing for HIV was done for each child. Two to three gastric aspirate specimens were also collected from each child using standard gastric aspirate (GA) collection protocols. On each specimen, Zeihl-Neelsen (ZN) staining and culture on Ogawa medium were done. After two or more weeks of incubation, typical acid-fast bacilli isolates on Ogawa medium, which failed to grow on para-nitrobenzoic acid-Ogawa (PNB-Ogawa) media were taken as positive culture for Mycobacterium tuberculosis complex (MTBC). MTBC Culture was taken as the “Reference test” for calculations of sensitivity and specificity.
Results: Of the 263 children surveyed 116 (44.0%) were HIV-infected while 147 (56.0%) of them were HIV-uninfected. Among HIV-infected patients, the sensitivity was 38.3% [95%CI:24.4 – 52.2] while the specificity was 95.7% [95% CI:90.9 – 100.0]. Among the HIV-uninfected patients, the sensitivity was 22.0%[95%CI:10.5 – 33.5] while the specificity was 99.0%[95% CI:96.9 -100.0]. The sensitivity of the acid-fast smear in the HIV-infected group of patients was significantly higher than that of the HIV-uninfected group (p = 0.0401).The difference in specificity between the two groups was not statistically significant (p=0.496). There was also no significant difference in their respective positive-predictive values (85.7 vs 91.7)(p = 0.60).
Conclusion: The sensitivity of GA acid-fast smear was significantly higher among the HIV infected group.