Using Reduced Inoculum Densities of Mycobacterium tuberculosis in MGIT Pyrazinamide Susceptibility Testing to Prevent False-Resistant Results and Improve Accuracy: A Multicenter Evaluation

Table 4

Comparison of predicted to actual MGIT PZA susceptibility test results stratified by inoculation method. Cumulative results of 9 laboratories each testing 20 isolates in duplicate.

Isolate; number	Predicted; result (category)	Actual number of results; by inoculation method
		BD		A		C
		S	R	S	R	S	R
1	S (A)	18	0	18	0	18	0
4	R (A)	0	18	0	18	0	18
7	S (A)	17	1	17	0	16	1
8	S (A)	18	0	18	0	18	0
11	S (A)	18	0	14	0	12	0
12	S (P)	1	16	13	5	15	3
13	S (P)	12	6	18	0	17	0
14	R (A)	0	18	2	15	2	16
16	S (P)	9	9	15	2	18	0
17	R (P)	2	16	9	9	7	8
18	S (A)	14	4	18	0	18	0
19	INC	13	5	18	0	18	0
20	S (A)	17	1	18	0	18	0
21	INC	6	12	13	5	14	4
22	S (A)	18	0	18	0	18	0
23	S (A)	17	0	16	0	12	0
24	S (P)	3	15	11	7	15	3
25	S (P)	12	6	18	0	17	0
26	S (P)	12	6	15	3	17	1
27	INC	7	11	14	4	13	5

Results for one laboratory excluded due to <70% reproducibility. Isolate numbers correspond to those in Table 1. S: susceptible; R: resistant; INC: inconclusive; A: always; P: predominately. Total results by method less than 18 (9 laboratories × 2 tests per isolate) indicate that some tests “Timed-Out” prior to a result. All 4 false-susceptible results occurred in the same laboratory.