Table 4.
Comparison of performance for Xpert MTB/XDR in sputum and LPA using pDST as gold standard
| N | TP | FP | FN | TN | Sensitivity | Specificity | PPV | NPV | |
|---|---|---|---|---|---|---|---|---|---|
| Isoniazid resistance | |||||||||
| MTB/XDR | 148 | 112 | 2 | 8 | 26 | 93.3 (87.1–96.7) | 92.9 (73.7–98.4) | 98.3 (93.3–99.6) | 76.5 (58.5–88.2) |
| MTBDRplus | 148 | 111 | 1 | 9 | 27 | 92.5 (86.1–96.1) | 96.4(76.3–99.6) | 99.1 (93.8–99.9) | 75.0 (57.6–86.9) |
| Fluoroquinolone resistance | |||||||||
| MTB/XDR | 147 | 140 | 0 | 2 | 5 | 96.6 (92.1–98.6) | 100 (-) | 100 (-) | 71.4 (64.2–78.5) |
| MTBDRsl | 147 | 144 | 0 | 2 | 1 | 99.3 (95.2–99.9) | 100 (-) | 100 (-) | 66.7 (53.1–99.9) |
| Amikacin resistance | |||||||||
| MTB/XDR | 147 | 145 | 0 | 2 | 0 | 98.7 (94.8–99.7) | * | 100 (-) | * |
| MTBDRsl | 147 | 147 | 0 | 0 | 0 | 100 (-) | * | 100 (-) | * |
| Kanamycin resistance | |||||||||
| MDR/XDR | 147 | 144 | 0 | 2 | 0 | 98.0 (93.8–99.3) | * | 100(-) | * |
| MTBDRsl | 147 | 147 | 0 | 0 | 0 | 100 (-) | * | 100(-) | * |
‘*’ All observations in culture were sensitive, hence specificity could be determined. (-) means that the confidence interval could not be determined because specificity/PPV/NPV was 100%. Analysis was limited to matched patients with results from both an Xpert MTB/XDR and LPA assays to ensure direct comparison