Table 3.
Sensitivity of Xpert MTB/RIF Assay Versus 2 or 3 Acid-Fast Bacilli Smears for Identification of US Culture-Confirmed Tuberculosis Cases
| Smear | AFB Positive | 1 Xpert Positive Results | P Value | 2 Xpert Results, at Least 1 Positive | P Value |
|---|---|---|---|---|---|
| 2 AFB smears (n = 91a) | |||||
| All | 68.1%b (62/91) [58.0%–76.8%] |
85.2% (75/88)c [76.3%–91.2%] |
.001 | 91.1% (82/90)d [83.4%–95.4%] |
<.001 |
| AFB positive | 96.7% (59/61)d [88.8%–99.1%] |
100% (62/62) [94.2%–100%] |
|||
| AFB negative | 59.3% (16/27)e [40.7%–75.5%] |
71.4% (20/28)d [52.9%–84.7%] |
|||
| 3 AFB smears (n = 53) | |||||
| All | 60.4% (32/53)f [46.9%–71.4%] |
82.0% (41/50)c [69.2%–90.2%] |
.006 | 88.5% (46/52)d [77.0%–94.6%] |
<.001 |
| AFB positive | 96.8% (30/31)d [83.8%–99.4%] |
100% (32/32) [89.3%–100%] |
|||
| AFB negative | 57.9% (11/19)e [36.3%–76.9%] |
70.0% (14/20)d [48.1%–85.5%] |
|||
Data are presented as % (no./No.) [95% confidence interval]. P values are for comparison of AFB smear vs Xpert assay, using Fisher exact test.
Abbreviation: AFB, acid-fast bacilli.
a One missing Xpert due to site/laboratory error.
b AFB+ in 61 of 88 (69.3%) for 1 Xpert assay comparison and 62 of 90 (68.9%) for 2 Xpert assays comparison.
c Three invalid Xpert assays.
d One invalid Xpert assay.
e Two invalid Xpert assays.
f AFB+ in 31 of 50 (62.0%) for 1 Xpert assay comparison and 32 of 52 (61.5%) for 2 Xpert assays comparison.