Research Highlights

Diagnosing tuberculosis using LED fluorescence

Published online 21 July 2011

Mohammed Yahia

Over nine million people develop tuberculosis (TB) each year, most live in poorer countries. To diagnose the disease, healthcare workers collect three sputum samples over two days and stain with Ziehl-Neelsen solution (ZN), which turns the disease-causing Mycobacterium tuberculosis bright red under a microscope — a lengthy procedure especially in countries with poor healthcare infrastructure.

An alternative method for detecting Mycobacterium involves adding a fluorescent dye to sputum samples and detecting the infective agent using a light-emitting-diode fluorescence microscopy (LED-FM). This can take a quarter of time the Zn smear test takes. A group of researchers, including Najla Al-Sonboli from Sana'a University in Yemen, collected sputum samples Ethiopia, Nepal, Nigeria and Yemen from 6,627 patients with a cough lasting more than two weeks to compare both techniques, and report their findings in PLoS Medicine.

The researchers found that, while LED-FM had a higher sensitivity than ZN smear microscopy, it had a higher rate of false positives. The accuracy of LED-FM after three sputum samples examined was 85%, compared to 91.8% in samples tested using ZN smear microscopy. The researchers suggest that despite this shortcoming, the combined increased sensitivity and shorter examination times might be worthwhile to diagnose more people from TB, which kills 1.7 million people per year, and reduce laboratory workloads.

The downside to LED-FM is the increased cost of putting some people who are falsely diagnosed on the six-month long TB treatment. The researchers caution that national control programmes that choose to introduce LED-FM smears must train healthcare workers on using the technique to ensure optimum performance.


  1. Cuevas, L. et al. LED Fluorescence Microscopy for the Diagnosis of Pulmonary Tuberculosis: A Multi-Country Cross-Sectional Evaluation. PLoS Med 8(7): e1001057 (2011). doi: 10.1371/journal.pmed.1001057 | Article |