TABLE 1.

Estimation of MTD validity measures^f

Study	MTD resultsa of smear-positive samples						Sensitivityc (%)	Specificityc (%)
	TB positiveb			TB negativeb
	+	−	I	+	−	I
Chedore and Jamieson (6)	194	0	—d	1	318	—d	100	99.7
Gamboa et al. (11)	48	0	0	0	19	0	100	100
Bergmann et al. (2)e	13	0	0	0	9	0	100	100
Smith et al. (19)e	15	1	0	0	7	0	93.8	100
Scarparo et al. (17)	89	1	7	0	0	0	99.0	NA
Wang and Tay (20)	66	0	0	0	0	0	100	NA
Piersimoni et al. (16)	36	0	0	0	0	0	100	NA
Della-Latta and Whittier (8)	38	0	0	0	0	0	100	NA
Total	499	2	7	1	353
Mean							99.6	99.7

^a+, positive; −, negative with no inhibitors detectable; I, negative with inhibitors detectable.

^bTB status was determined by culture result, combined with clinical criteria when reported.

^cSensitivity and specificity are reported here after removal of specimens with MTD inhibitors. NA, not applicable.

^d—, authors did not report rates of inhibition.

^eStudy used a different diagnostic algorithm, whereby MTDs giving results between 30,000 and 300,000 RLU or 30,000 and 500,000 RLU were retested. The sensitivity of this alternative algorithm is lower and its specificity is higher than those of the diagnostic approach we have taken.

^fThe estimated total probability of sample inhibition was 2.3% (7 of 305) specimens if culture positive [the specimens tested by Chedore and Jamieson {6} were not included in this calculation because the presence or absence of inhibitory substances was not reported] and was 0% (0 of 35 specimens) if culture negative.