Table 2. Diagnostic accuracy of MTBDRplus and MTBDRsl for the direct detection of drug resistance in sputum samples using phenotypic culture-based susceptibility testing as a reference standard.
| All sputum samples |
Smear-positive sputum |
Smear-negative sputum |
|||||
|---|---|---|---|---|---|---|---|
| Sensitivity (%) | Specificity (%) | Sensitivity (%) | Specificity (%) | Sensitivity (%) | Specificity (%) | ||
| MTBDRplus‡(v 1.0) | RIFR* | 97.7 (86/88) | 91.8 (34/37) | 97.1 (69/71) | 91.2 (31/34) | 100 (17/17) (p = 0.484) | 100 (3/3) |
| INHR* | 95.4 (84/88) | 89 (33/37) | 95.6 (68/71) | 88.2 (30/34) | 94.1 (16/17) (p = 0.768) | 100 (3/3) | |
| MTBDRsl†(v 1.0) | OFXR | 58.9 (43/73) | 100 (38/38) | 79.2 (38/48) | 100 (34/34) | 20 (5/25) (p < 0.001) | 100 (4/4) |
| AMKR | 61.6 (45/73) | 100 (38/38) | 72.9 (35/48) | 100 (34/34) | 37 (10/27) (p < 0.001) | 100 (4/4) | |
‡ 7.4% (9/122) MTBDRplus results from smear-positive samples were indeterminate, compared to 17% (10/59) from smear-negative (p = 0.049).
†1.6% (2/122) MTBDRsl results from smear-positive samples were indeterminate, compared to 5.1% (3/59) from smear-negative (p = 0.185). Refer to the materials and methods for a description of what defines an indeterminate result. P-values are for comparisons between smear statuses.
*When the discrepant results were resolved by sequencing the sensitivities and specificities of MTBDRplus were 100% and 100% to detect RIFR, respectively and 97.7% and 97.4% to detect INHR, respectively. RIFR = rifampicin resistance, INHR = isoniazid resistance, OFXR = ofloxacin resistance, AMKR = amikacin resistance.