Volume of sputum to detect acid-fast bacilli as a measure of quality for the diagnosis of pulmonary tuberculosis at the Dr George Mukhari Hospital, South Africa

Abstract

BackgroundOptimum sputum results for acid-fast bacilli (AFB) microscopy are linked to a sputum quantity of at least 5.0 mL. This study was aimed at establishing the effect of sputum quantity in the pick-up rate of AFB microscopy by comparing sputum samples of 5.0 mL and 2.0 mL.MethodsAn analytical cross-sectional study was carried out at the Dr George Mukhari Hospital (DGMH) in Pretoria, South Africa, from 05 January 2007 to 04 January 2008. Two sputum samples, 5.0 mL and 2.0 mL, were collected from each of the 330 adult PTB (pulmonary tuberculosis) suspects. Fluorescence microscopy was used in the sputum analysis. The yield through microscopy of the 2.0 mL specimen versus the 5.0 mL specimen was compared and analysed, using culture results as the gold standard.ResultsFrom a sample of 330 specimens, 77 tested AFB positive on microscopy. In the 5.0 mL samples, the sensitivity was 76.6% (95% CI, 66.0% – 84.7%), specificity 99.6% (95% CI 97.8% – 99.9%), positive predictive value (PV+) 98.3% (95% CI 91.1% – 99.7%), negative predictive value (PV-) 93.3% (95% CI 89.7% – 95.7%), the likelihood ratio (LR) for a positive microscopy 192 and the LR for a negative test was 0.23. In the 2.0 mL specimens, the sensitivity was 75.3% (95% CI 64.6% – 83.6%), specificity 99.2% (95% CI 97.1% – 99.8%), positive predictive value (PV+) 96.7% (95% CI 88.6% – 99.1%), negative predictive value (PV-) 93.0% (95% CI 89.3% – 95.4%), the LR for a positive microscopy was 94 and 0.25 for a negative microscopy. There was a statistically significant association (p-value < 0.001) between the microscopy and culture tests in both the 5.0 mL and the 2.0 mL specimen categories. The strength of association between the microscopy and culture, as indicated by the kappa test was 0.83 and 0.81 in the 5.0 mL and 2.0 mL categories, respectively.ConclusionCompared to the 2.0 mL specimen category, the yield for AFB microscopy in the 5.0 mL specimen category was consistently superior, as indicated by the higher sensitivity, specificity, predictive values and the likelihood ratios in the 5.0 mL specimen category. It is recommended that sputum specimen collection for AFB microscopy should aim for a minimum volume of 5.0 mL.