Setting: Treatment program for tuberculosis in a refugee camp in Thailand.
Objectives: To determine the cumulative frequency of conversion of sputum smears examined by direct microscopy by month of treatment and to identify factors predicting failure to convert.
Methods: Analysis of conversion based on three sputum smear examinations (performed monthly) in a cohort of patients with sputum smear-positive tuberculosis treated with a directly observed daily regimen containing rifampicin throughout. Nested case-control study of patients failing to convert definitively within four months compared to controls who did convert.
Results: Sputum conversion after the 2-month intensive phase was 75.0%, with a range from 61.7% to 90.9% in patients with initially strongly- and weakly-positive smears, respectively. The strongest predictor identified for no definitive conversion within four months of treatment was a positive sputum smear result at the end of the 2-month intensive phase (adjusted relative odds 4.2, 95% confidence interval 1.5-11.4). Of those patients who did not convert, positive smears were an isolated phenomenon in 15, repeatedly in four who definitely converted with a prolongation of treatment, and persistently positive in two requiring a re-treatment regimen.
Conclusions: Definitive sputum smear conversion is judged to be slower if a strict program of sputum smear examination is undertaken than under routine program conditions, but positive results late in the course are commonly an isolated phenomenon and possibly of little significance. Sputum smear results at two months strongly predict bacteriologic results beyond three months of treatment, and thus identify cases who might benefit from a prolongation of the intensive phase.