An assessment of nucleic acid amplification testing for active mycobacterial infection

Review of literature

Literature sources and search strategies

The medical literature was searched to identify relevant studies and reviews for the period between 1990 and June 2014. Searches were conducted for the databases described in Table . Search terms are described in Table to Table .

Due to the large volume of evidence for the diagnostic accuracy of NAAT compared with culture, only studies published after 2005 that provided 2x2 data suitable for meta-analysis for both AFB microscopy and NAAT compared with culture, were included in the final analysis. Studies on the only commercial NAAT product (Xpert) available in Australia were published in 2006 onwards. In-house NAAT, on the other hand, was available before 2005. However, as there have been significant changes in laboratory practice over the past 10 years (Boyle & Pai 2012; Moore, Guzman & Mikhail 2005; Nybo 2012; Public Health and Ambulatory Care 2012), it seemed reasonable to limit study eligibility to publications in the previous decade.

The diagnostic accuracy of in-house NAAT performed more than 10 years ago compared with culture was reported in two SRs. Pai et al. (2004) reported that the pooled sensitivity in pleural fluid specimens was 71% (95%CI 63, 76; k=26), and Pai et al. (2003) reported that the pooled sensitivity in CSF specimens was 76% (95%CI 57, 83; k=35). These values are much lower than that reported in this assessment for non-sputum specimens (90%; 95%CI 83, 94, k=44; see Figure ). Thus, the inclusion of only those studies published after 2005 in the final analysis provided more-accurate data on the accuracy of NAAT as currently performed in the diagnostic laboratory.

An SR by Takwoingi et al. (2013) showed that there was a > 2-fold discrepancy in the relative diagnostic odds ratio between non-comparative studies (that compared either the index test or the comparator with the reference standard) and comparative studies (that compared both the index test and the comparator with the reference standard). Thus, these studies provide the highest quality evidence available to assess the accuracy of NAAT and AFB microscopy compared with culture to diagnose MTB infections.

Table Electronic databases searched

Electronic database	Period covered
Cochrane Library – including, Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, the Cochrane Central Register of Controlled Trials (CENTRAL), the Health Technology Assessment Database, the NHS Economic Evaluation Database	1990 – 6/2014
Current Contents	1990 – 6/2014
Embase	1990 – 6/2014
PubMed	1990 – 6/2014
Web of Science – including Science Citation Index Expanded and Conference Proceedings Citation Index- Science	1990 – 6/2014
Cinahl	1990 – 6/2014
Econlit	1990 – 6/2014
Scopus	1990 – 6/2014

Table Search terms used for NAAT for MTB (direct evidence, accuracy and change in management)

Element of clinical question	Search terms
Population	(tuberculosis OR MTB OR MTB/RIF OR “tubercle bacillus” OR Tuberculosis [MeSH] OR Mycobacterium OR “M. africanum” OR “M. bovis” OR “M. microti” OR “M. canettii” OR “M. caprae” OR “M. pinnipedii” OR “M. mungi”)
Intervention	(Amplicor OR Amplified OR “Direct Test” OR “Direct Detection” OR TaqMan OR Xpert OR “nucleic acid amplification” OR NAAT OR “polymerase chain reaction” OR PCR OR “Nucleic Acid Amplification Techniques” [MeSH])
Comparator (if applicable)	N/A
Outcomes (if applicable)	N/A
Limits	1990 – June 2014; NOT (Other Animals NOT Humans)

MeSH = Medical Subject Heading, based on a Medline/PubMed platform; N/A = not applicable

Table Search terms used for NAAT for NTM (direct evidence, accuracy and change in management)