Two vs. three sputum samples for microscopic detection of tuberculosis in a high HIV prevalence population.

SETTING: A busy urban hospital in Cameroon.

OBJECTIVES: To compare the yield in bacteriologically proven tuberculosis (TB) cases examining two morning vs. three spot-morning-spot sputum specimens (MM vs. SMS) by direct microscopy for acid-fast bacilli (AFB).

DESIGN: Repeated temporal cross-over between MM and SMS sampling for successive TB suspects, using culture as gold standard.

RESULTS: A total of 799 suspects were screened using the MM strategy, identifying 223 smear-positives, and 808 suspects with the SMS strategy, yielding 236 smear-positives. Of the MM, 256 were culture-positive, of whom 195 (76%) were smear-positive. For SMS, these figures were respectively 281 and 206 (73%), a non-significant difference. The MM and SMS strategies also detected respectively 28 and 30 smear-positive cases not confirmed by culture. No cases were lost to treatment with either strategy.

CONCLUSIONS: In this population with a high prevalence of human immunodeficiency virus (HIV) with late case presentation, smear microscopy of two morning specimens detected at least as many positive cases as the classical strategy, and no cases were lost before treatment. Two specimens for initial TB suspect screening can thus be recommended, also without excessive workload. Comparative studies in populations presenting with paucibacillary sputum are needed to determine the equivalent quality and yield of an alternative strategy with two spot specimens at consultation.