TY  - JOUR
T1  - INterferon assay compared to tuberculin skin testing for latent tuberculosis detection
AU  - Hersh A, von Reyn C
Y1  - 2002/01/23
N1  - 10.1001/jama.287.4.447
JO  - JAMA
SP  - 450
EP  - 452
VL  - 287
IS  - 4
N2  - 
To the Editor: Dr Mazurek and colleagues1 concluded that a whole-blood interferon γ
(IFN-γ) assay was comparable to tuberculin skin testing (TST) in the
detection of latent tuberculosis infection (LTBI). However, the results fail
to support their conclusions about the sensitivity or specificity of this
assay. The operating characteristics of a new test should be derived by comparison
to a reference standard, TST in this case. Results of TST of 15 mm or greater
or less than 5 mm reliably identify the presence and absence of LTBI, respectively.2 Thus, the authors should demonstrate that the IFN-γ
assay is consistently positive when TST results are 15 mm or greater and negative
when TST results are less than 5 mm. These data are not presented, but the
absence of a positive assay result in 36% of patients with tuberculosis disease
(group 4) and the fact that half of the subjects in the asymptomatic "high
risk" group (group 2) have discordant TST and IFN-γ results suggest
that both sensitivity and specificity of the assay may be limited. Table 1
should be presented with corresponding data for the assay in each cell.

SN  - 0098-7484
M3  - doi: 10.1001/jama.287.4.447
UR  - http://dx.doi.org/10.1001/jama.287.4.447
ER  -