Optimal number of samples required to diagnose tuberculosis by sputum culture among adult HIV infected smear negative suspects in Kampala, Uganda

Abstract

Background; The World Health Organization recommended sputum culture among smear negative HIV infected TB suspects; but the number of samples to be examined is still unspecified. A study was set out to determine the incremental yield and the number of sputum cultures needed to identify one additional tuberculosis case on serial sputum cultures among HIV-infected smear-negative suspects. Methods: In this cross sectional study, sputum samples (spot-Morning – spot) provided by participants in two days were examined. Each provided not more than three samples from February to March 2011. Each sputum sample was cultured both on Lowenstein Jensen (LJ) and Mycobacterium Growth Indicator Tube (MGIT). A pulmonary tuberculosis (PTB) case was defined as a positive culture on either one or both culture methods used. Data were analyzed to determine the Incremental yield and number of sputum sample culture needed to find one additional PTB case on serial sputum cultures. Results: A total of 170 sputum smear negative adult HIV infected participants were enrolled, of which 62 (36.5%) were culture positive for pulmonary tuberculosis (PTB). Of the 62, 52 (83.9%) were detected on the first, 7 (11.3%) on the second and 3 (4.8%) on the third sputum sample culture. Overall, the first sputum culture diagnosed 80.5%, 69.0% and 80.1% of the cases by Lowenstein Jensen (LJ), Mycobacterium Growth Indicator Tube (MGIT) and MGIT plus LJ respectively. The IY of the second and third sputum culture was 12.7% and 6.8% by LJ, 23.6% and 7.5% by MGIT, 12.6% and 7.3% by MGIT plus LJ culture methods respectively. The number of samples needed to find one additional TB case and their 95% CI on the second and third sputum culture were 30 (16.6, 156.5) and 56 (26.4, 500.4) by LJ, 11 (7.6, 21.9) and 36 (19.0, 313.8) by MGIT and 21 (12.5, 62.7) and 36 (19.1, 330.9) by MGIT plus LJ culture methods respectively. Conclusion: Among HIV infected smear negative PTB suspects in Kampala Uganda, more than 93% of PTB cases are diagnosed with two serial sputum samples hence, two sputum sample cultures may beoptimal for TB diagnosis. However, in resource limited settings, there is a need to look at the cost effectiveness of these findings.