An assessment of nucleic acid amplification testing for active mycobacterial infection

Component	A Excellent	B Good	C Satisfactory
D Poor
Evidence-base^a			One or two level III studies with a low risk of bias, or level I or II studies with a moderate risk of bias
Consistency^b		Most studies consistent and inconsistency may be explained
Clinical impact				Slight or restricted
Generalisability			Population(s) studied in body of evidence differ to target population for guideline but it is clinically sensible to apply this evidence to target population
Applicability			Probably applicable to Australian healthcare context with some caveats

^a Level of evidence determined from the NHMRC evidence hierarchy (see Table ).

Source: Adapted from NHMRC (2009)

Ten studies reported the diagnostic accuracy of NAAT compared with culture, but only 6 of these studies also compared AFB microscopy with culture (Table ). Two of the studies also reported the diagnostic accuracy of NAAT and culture compared with a clinical reference standard. An additional two studies reported on the diagnostic accuracy of AFB, NAAT and culture compared with a clinical reference standard only (Table ). Culture methods included standard diagnostic laboratory procedures, including L-J or Ogawa solid media and/or liquid BACTEC media. Four studies included respiratory specimens (3 of which included sputum specimens and 1 also included extrapulmonary specimens), 4 used blood and bone marrow specimens (all from HIV-positive patients) and 4 used tissue biopsy specimens (3 used archived formalin-fixed paraffin-embedded specimens).

The NAAT used in these studies could be separated into three distinct categories. Five studies used NAAT to detect NTMs in general (NTM-NAAT) by targeting either the 16S–23S rRNA sequence (k=3) or the gene encoding the 65-kDa heat shock protein (k=2). Six studies used NAAT to specifically detect MAC strains (MAC-NAAT), which included all 4 studies involving HIV-positive patients (1 of which only used a clinical reference standard). One study used NAAT to detect M. ulcerans in patients suspected of having Buruli ulcer by targeting IS2404, but only compared NAAT with a clinical reference standard (Table ). However, many of these studies also identified patients (specimens) with MTB infections (Table and Table ). As MTB infections are much more common than NTM and would therefore affect the accuracy of NAAT in detecting NTM, MTB culture-positive specimens were excluded from the analysis wherever possible. MTB-positive results could not be excluded from the analysis for 2 studies that identified 6/46 (Bogner et al. 1997) and 2/36 (Mahaisavariya et al. 2005) positive cultures as MTB, and the study by Frevel et al. (1999) did not report the number of MTB-positive cultures included in the analysis.

The prevalence of patients with culture-positive NTM infections varied between 4% and 67%, with 5 studies reporting NTM-positive cultures in less than 10% of the tested specimens. Only 2 studies reported a prevalence rate greater than 30%; these were 42% in the study investigating the presence of M. ulcerans in biopsy specimens from a suspected Buruli ulcer (Phillips et al. 2005) and 67% in a study investigating the presence of MAC in blood and bone marrow aspirates from AIDS patients who were suspected of having disseminated mycobacterial infections (Gamboa et al. 1997). The reason for this high rate of culture-positivity when compared with other studies looking at disseminated mycobacterial infections could not be determined. Studies using NTM-NAATs reported a mean prevalence of NTM-positive cultures of 13% (range 4–30) compared with 25% (range 9–67) for those using MAC-NAATs. As expected, the prevalence of culture-positive NTM was higher in AFB-positive specimens (36%) compared with AFB-negative specimens (21%; Table ).

Table Prevalence of NTM culture-positive specimens in the included studies