Performance of genexpert PCR on pooled sputum versus smear microscopy in tuberculosis

Abstract

INTRODUCTION/AIM: The diagnostic performance of GeneXpert MTB/RIF on combined sputum samples has not been systematically assessed in patients hospitalised with presumptive mycobacterium tuberculosis (TB). We aim to assess the diagnostic accuracy of a single GeneXpert MTB/RIF test performed on two combined sputa versus smear microscopy performed on three separate sputum samples, using sputum culture as the reference standard. METHODS: We assessed all adult patients admitted to Blacktown Hospital between July 2017 and January 2018 with presumptive pulmonary TB. The presence of acid fast bacilli (AFB) in expectorated sputum samples, collected on three separate days, was assessed using auramine staining and microscopy. For GeneXpert MTB/RIF testing we combined 0.5mL from each of the first two sputa. The same Sputa were also cultured in liquid medium (MGIT; Becton-Dickinson) for 8 weeks. The diagnostic accuracy, time in isolation, duration of hospital admission and final diagnosis were recorded. RESULTS: Forty patients met the inclusion criteria and 37 (92.5%) consented to study participation; 7 were excluded due to insufficient samples. Of 30 eligible patients, 4 (13.3%) had culture-confirmed pulmonary TB. Of the non-tuberculosis patients; 13 had pneumonia, 3 had non-tuberculous mycobacterium, 4 had TB in other sites and 6 had other pathology. GeneXpert MTB/RIF detected 3 of 4 (75%) cases with 100% specificity, while smear microscopy was positive in only 2 of 4 cases (50%) with 95.5% specificity. The average time in respiratory isolation was 10.9 days and the average length of hospital stay was 12.5 days. CONCLUSIONS: A single GeneXpert MTB/RIF test on pooled sputum had negative predictive value (NPV) and positive predictive value (PPV) of 96% and 100% respectively. This reflects improved diagnostic yield and accuracy compared to standard smear microscopy with NPV of 93% and PPV of 67%. We postulate this could translate into cost savings for the health system.