Reply to Jain et al.: Utility of Xpert Ultra on Different Respiratory Specimens in Children.

From the Authors:

Thanks for the opportunity to respond to these points. As noted, the semiquantitative results of the Xpert Ultra (Cepheid) test were mainly trace or very low in nasopharyngeal aspirates (NPAs) (1). We now provide the semiquantitative results on induced sputum; these showed a pattern similar to that of NPAs (Table 1). Most positive results in induced sputum were either trace or very low (71%; 22 of 31) compared with 75% on first (15 of 20) or second (9 of 12) NPAs, attesting to the low bacillary load in childhood pulmonary tuberculosis. Regarding the positive predictive value and negative predictive value, we apologize that these were inadvertently switched in Table 3; therefore, 156 of 175 (89.1%) is the negative predictive value.

Table 1.

Xpert Ultra Semiquantitative Results in Nasopharyngeal Aspirate and Induced Sputum

Sample	Negative	Trace	Very Low	Low	Moderate	High
First NPA (n = 195)	175	9	6	1	4	0
Second NPA (n = 130)	118	4	5	2	1	0
IS (n = 195)	164	8	14	5	3	1

Definition of abbreviations: IS = induced sputum; NPA = nasopharyngeal aspirate.

A comparison of the Xpert Ultra and Xpert MTB/RIF platforms was not the objective of this study, which was to investigate the incremental value of additional samples using Xpert Ultra testing. In addition, Xpert Ultra and Xpert MTB/RIF tests were performed using different induced sputum samples, so the results are not directly comparable. However, as noted in the study, there were more positive Xpert Ultra results than Xpert MTB/RIF results (74.3% vs. 68.6%) using matched induced sputum samples that were culture positive as the denominator. We have previously reported the comparison of the Xpert Ultra and Xpert MTB/RIF platforms using induced sputum (2), as have others (3, 4); the evidence suggests that the Xpert Ultra test is more sensitive, as would be expected, given the paucibacillary nature of the disease (Table 1).