PURPOSE: International guidelines for the diagnosis of pulmonary tuberculosis (TB) recommend obtaining 3 sputum specimens for acid-fast bacilli (AFB) smear. Recent studies suggest that 2 specimens may be as sensitive as 3. If validated, this finding would relieve some financial and logistical pressures on patients and TB programs. The objective of the study is to assess the sensitivity of the AFB smear in culture-positive patients.
METHODS: Patients who had sputum samples for AFB smear and mycobacterial culture were prospectively included in Mexico, the Phillippines, England and the US.
RESULTS: Of 434 patients, 162 had ≥1 positive culture. The sensitivity was 52.85% for the first AFB sample, 57.85% for the first two samples, and 60% for all 3 samples. The proportion of cases diagnosed was not significantly different using the first 2 or all 3 samples (p=0.42). Only 9 (5.5%) patients had normal chest x-rays; 73 (45.06%) had a cavitary lesion, 24 with bilateral cavities (14.81%). Fifty patients had serology testing for HIV, with 7 reactive. There was no significant difference with AFB sensitivity in patients who were reactive or non reactive for HIV (83.7% vs. 85.7%; p=0.69). Among clinical characteristics, cavitary lesion was most predictive of positive AFB smears (p=0.11). The specificity of the AFB smear in 272 culture-negative patients was 98.5%, using 2 or 3 samples.
CONCLUSION: Reducing the number of sputum samples from 3 to 2 does not affect AFB smear accuracy.
CLINICAL IMPLICATIONS: Reducing the number of sputum samples to two will reduce the number of visits required for TB diagnosis, saving resources for the healthcare system and patients.
DISCLOSURE: Rafael Laniado-Laborin, None.