OpiniãoChildren with severe acute malnutrition (SAM): four studies (with 259 participants, of whom 9 had TB disease) were identified for children with SAM using Xpert Ultra on gastric aspirates and three studies (428 participants, 19 with TB) using Xpert Ultra on stool specimens. The meta-analysis on stool samples in children with SAM showed similar accuracy as in the overall analysis (sensitivity 63.2%, specificity 98.5%). A meta-analysis on gastric aspirates could not be performed due to insufficient data.
Children living with HIV infection: four studies (99 participants, 8 with TB disease) were identified for children living with HIV using Xpert Ultra on gastric aspirates and two studies (100 participants, 3 with tuberculosis) using Xpert Ultra on stool specimens. A meta-analysis on stool or gastric aspirate samples could not be conducted for children living with HIV due to paucity of data.
Age subgroups: analyses on age subgroups were conducted by systematic reviewers.
- For gastric aspirate samples from children aged below 1 year, Xpert Ultra pooled sensitivity and specificity (95% CI) were 67.3% (43.5 to 84.6) and 94.0% (84.7 to 97.8) respectively (5 studies, 182 participants; low-certainty (sensitivity) and moderate-certainty (specificity) evidence). From children aged 1 to 4years, Xpert Ultra pooled sensitivity and specificity (95% CI) were 71.5% (40.0 to 90.4) and 94.0% (73.8 to 98.9), respectively (4 studies, 327 participants; low-certainty evidence).
- For stool samples from children aged below 1 year, Xpert Ultra pooled sensitivity and specificity (95% CI) were 65.2% (33.7 to 87.3) and 96.2% (88.9 to 98.7), respectively (4 studies, 295 participants; very low-certainty (sensitivity) and moderate-certainty (specificity) evidence). From children aged 1 to 4years, Xpert Ultra pooled sensitivity and specificity (95% CI) were 43.3% (27.1 to 61.2) and 97.1% (74.8 to 99.7), respectively (3 studies, 331 participants; very low-certainty (sensitivity) and low-certainty (specificity) evidence). The small number of children aged below 1 year and in the 1-4 years age group for both samples limited the confidence in the precision of the estimates for these age groups.
The systematic review did not identify studies that evaluated Xpert Ultra in gastric aspirate or stool samples of children with severe pneumonia.
Xpert Ultra on gastric aspirate, stool, NPA and sputum (expectorated or induced) samples can be used in all paediatric comorbidity and age subgroups mentioned in this section as the preferred initial test for the diagnosis of PTB in children.