Accuracy of smear microscopy and incremental yield of Xpert MTB/RIF in detection of pulmonary tuberculosis in laboratories in Rwanda

Abstract

Background: The National tuberculosis control Program (NTP) of Rwanda is currently phasing in the use of light emitting diode-fluorescent microscopy (LED-FM) as an alternative to the Ziehl-Neelsen technique (ZN). This, alongside the Xpert MTB/RIF technique is expected to improve diagnosis of tuberculosis (TB) and detection of Rifampicin resistance in patients at the various health facilities. There is limited information in Rwanda on the incremental yield of these new techniques in comparison to the ZN technique which has been in use for many years. Objective: To determine the accuracy of routine sputum smear microscopy and assess the incremental yield of Xpert MTB/RIF test in detection of pulmonary tuberculosis in peripheral and intermediate health facility laboratories in Rwanda. Methods: The study was conducted at four intermediate and four peripheral health facility laboratories across Rwanda. This study enrolled 648 participants with suspected pulmonary tuberculosis who visited any of the eight health facilities from April 1st to June 21st, 2014. Each participant gave one sputum sample which was processed as per the individual laboratory procedures and the results obtained from the direct sputum smear recorded. The remaining sample was processed using N-acetyl-cystein-Sodium hydroxide and inoculated a Lowenstein Jensen slant and mycobacterium growth indicator tube (MGIT) at the National Reference Laboratory (NRL). The pellet was used to prepare two smears, one for ZN and the other for LED-FM staining and microscopy. The original health facility slide was re-examined at NRL and the results recorded. The pellet was also used for testing using Xpert MTB/RIF. Results: The study found the overall sensitivity and specificity of smear microscopy to be 51.5% and 99.8% respectively, and positive and negative predictive values of 98% and 91.4% respectively. The increment of Xpert MTB/RIF test over the smear microscopy was 40.3% an 47.6% respectively in HIV-negative and HIV-positive tuberculosis suspects. The overall prevalence of negative pulmonary tuberculosis was 39.2%. Conclusion: This study has provided the baseline data on the performance of currently used routine methods and the incremental yield of Xpert MTB/RIF test compared to MTB culture as the gold standard. It is therefore imperative to implement measures which could increase the sensitivity of LED-FM microscopy before replacing ZN techniques in health facility laboratories. The use of smear microscopy alone misses more than 50% of pulmonary tuberculosis cases, consequently all smear negative cases with prominent signs and symptoms of pulmonary tuberculosis should be tested using Xpert MTB/RIF test in order to increase tuberculosis case detection and reduce the chain of transmissio