Use of light-emitting diode fluorescence microscopy to detect acid-fast bacilli in sputum

Background. Fluorescence microscopy offers well-described benefits, compared with conventional light microscopy, for the evaluation of sputum smear samples for tuberculosis. However, its use in resource-limited settings has been limited by the high cost of the excitatory light source. We evaluated the diagnostic performance of fluorescence microscopy, using novel light-emitting diode (LED) technology as an alternative to the conventional mercury vapor lamp (MVP). Methods. Routinely collected sputum specimens from persons suspected to have tuberculosis who attended community clinics were stained with auramine O and were evaluated using 2 different excitatory light sources (MVP and LED); these specimens were then Ziehl-Neelsen stained and reexamined using light microscopy. Two microscopists independently evaluated all smears. Bacterial culture provided the gold standard. Results. Of the 221 sputum specimens evaluated, 36 (16.3%) were positive for Mycobacterium tuberculosis by culture. Sensitivity and specificity documented for the different modalities were 84.7% and 98.9%, respectively, for the LED assessment; 73.6% and 99.8%, respectively, for the MVP assessment; and 61.1% and 98.9%, respectively, for light microscopy. kappa values for interreader variation were 0.87 for the LED assessment, 0.79 for the MVP assessment, and 0.77 for light microscopy. The mean time to read a negative smear was 1.4 min with fluorescence microscopy and 3.6 min with light microscopy, reflecting a time savings of 61% with fluorescence microscopy. Conclusion. LED fluorescence microscopy provides a reliable alternative to conventional methods and has many favorable attributes that facilitate improved, decentralized, diagnostic services.