Comments
There is an error in the article by Zar and colleagues (1), published in the December 15, 2019, issue of the Journal. In Table 3, the first listed values in the third column (positive predictive value [PPV]) and fourth column (negative predictive value [NPV]) were inadvertently switched. The correct value for the PPV cell is “16/20 (80.0)”; the correct value for the NPV cell is “156/175 (89.1).” For the convenience of our readers, we reproduce the corrected table below (changes are indicated in bold).
| Sensitivity | Specificity | PPV | NPV | |
|---|---|---|---|---|
| NPA | ||||
| Ultra on NPA (n = 195) | ||||
| Number positive/number tested (%) | 16/35 (45.7) | 156/160 (97.5) | 16/20 (80.0) | 156/175 (89.1) |
| 95% CI | 28.8–63.3 | 93.7–99.3 | 83.6–93.3 | 56.3–94.3 |
| By HIV infection status | ||||
| HIV infected (n = 32) | ||||
| Number positive/number tested (%) | 3/6 (50.0) | 26/26 (100.0) | 3/3 (100.0) | 26/29 (89.7) |
| 95% CI | 11.8–88.2 | 87.1–100.0 | 43.9–100.0 | 72.6–97.8 |
| HIV uninfected (n = 162) | ||||
| Number positive/number tested (%) | 13/29 (44.8) | 129/133 (97.0) | 13/17 (76.5) | 129/145 (89.0) |
| 95% CI | 26.4–64.3 | 92.5–99.2 | 50.1–93.2 | 82.7–93.6 |
| By TB treatment | ||||
| Prior TB treatment (n = 15) | ||||
| Number positive/number tested (%) | 1/2 (50.0) | 12/13 (92.3) | 1/2 (50.0) | 12/13 (92.3) |
| 95% CI | 1.2–98.7 | 64.0–99.8 | 1.2–98.7 | 64.0–99.8 |
| No prior TB treatment (n = 177) | ||||
| Number positive/number tested (%) | 15/32 (46.9) | 142/145 (97.9) | 15/18 (83.3) | 142/159 (89.3) |
| 95% CI | 29.1–65.3 | 94.1–99.6 | 58.6–96.4 | 83.4–93.6 |
| IS | ||||
| Xpert MTB/RIF* on IS (n = 165) | ||||
| Number positive/number tested (%) | 24/35 (68.6) | 132/135 (97.8) | 21/24 (87.5) | 132/141 (93.6) |
| 95% CI | 50.7–83.1 | 93.6–99.5 | 67.6–97.3 | 88.2–97.0 |
| Ultra on IS (n = 195) | ||||
| Number positive/number tested (%) | 26/35 (74.3) | 155/160 (96.9) | 26/31 (83.9) | 155/164 (94.5) |
| 95% CI | 56.7–87.5 | 92.9–99.0 | 66.3–94.5 | 90.0–97.5 |
| By HIV infection status | ||||
| HIV infected (n = 32) | ||||
| Number positive/number tested (%) | 3/6 (50.0) | 26/26 (100.0) | 3/3 (100.0) | 26/29 (89.7) |
| 95% CI | 11.8–88.2 | 87.1–100.0 | 43.9–100.0 | 72.6–97.8 |
| HIV uninfected (n = 162) | ||||
| Number positive/number tested (%) | 23/29 (79.3) | 128/133 (96.2) | 23/28 (82.1) | 128/134 (95.5) |
| 95% CI | 60.3–92.0 | 91.4–98.8 | 63.1–93.9 | 90.5–98.3 |
| Ultra on NPA and IS | ||||
| Number positive/number tested (%) | 28/35 (80.0) | 152/160 (95.0) | 28/36 (77.8) | 152/159 (95.6) |
| 95% CI | 63.1–91.6 | 90.4–97.8 | 60.8–89.9 | 91.1–98.2 |
Zar and colleagues would like to thank Drs. Saurav Jain, Prawin Kumar, and Jagdish Prasad Goyal, who alerted us to this error in their letter (2) concerning our article.
| 1. | Zar HJ, Workman LJ, Prins M, Bateman LJ, Mbhele SP, Whitman CB, Denkinger CM, Nicol MP. Tuberculosis diagnosis in children using Xpert Ultra on different respiratory specimens. Am J Respir Crit Care Med 2019;200:1531–1538. |
| 2. | Jain S, Kumar P, Goyal JP. Utility of Xpert Ultra on different respiratory specimens in children [letter]. Am J Respir Crit Care Med 201:1452. |
