Yield of fluorescence microscopy versus culture for tuberculosis at a middle-income country referral hospital.

The aim of this study was to determine the usefulness of fluorescence microscopy (FM) at a referral centre in a middle-income country. Direct Ziehl-Neelsen (ZN) and direct, as well as concentrated, smear FM were performed on 2179 suspect sputa, with Löwenstein-Jensen (LJ) culture as the gold standard. ZN, direct FM and concentration FM detected 36.0, 38.6 and 37.0%, respectively, of 272 culture-positive specimens. Patient-wise, there were 8.1% (126/1553) positives on any smear compared with 12.0% (187/1553) on any culture. ZN, direct FM and concentrated FM smear were positive in 43.3, 46.5 and 45.5%, respectively, of culture-proven cases. All differences between microscopy and culture were significant (P<0.001), but not those between microscopy techniques. Acid-fast bacilli (AFB) were not rare in 60% of 48 duplicate smears, positive in ZN or FM only. Simple LJ culture, but not FM on direct or concentrated smears, was thus significantly more sensitive than ZN smears. The considerable numbers of AFB found in positive direct smears from discordant microscopy sets suggest that repeating smears can improve microscopy sensitivity more than variations of correctly executed technique, provided that overload is avoided. Thus FM could be particularly useful, as it is time-saving and could protect against the sensitivity loss associated with high workload.