Specimen collection and the quality thereof needs to be ensured to optimize accurate diagnosis for all specimen types.
Gastric aspirate samples: Gastric aspirate sample collection in children requires trained staff and access to supplies. It may also require hospitalization and may therefore not be feasible at lower levels of the health system. It is considered an invasive procedure, requires fasting and may be uncomfortable and less acceptable to children and caregivers, or have adverse effects. Referral systems need to be functional to ensure that children who need to be tested using gastric samples reach the appropriate level of care.
Stool samples: Stool sampling has the advantage of being non-invasive, with the collection generally perceived as easy and feasible by HCWs and caregivers, irrespective of the clinical condition of the child. The possible drawback is that children may not be able to pass stool on command causing delay in sample collection. Enhancing awareness and sensitizing caregivers on stool as a suitable sample for TB testing is important, to enhance acceptability and implementation. Stool processing is also an acceptable procedure for laboratory technicians and regarded as a good alternative to sputum samples. Relatively high rates of non-determinate Xpert Ultra results (including error, invalid, or no results) were reported in the studies included in the systematic review. These rates varied from less than 1% to 10% and may depend on the stool processing method used.
Stool processing methods: An analysis of preliminary results of a head-to-head comparison of three centrifuge-free stool processing methods combined with Xpert Ultra was conducted to inform implementation considerations on the use of Xpert Ultra in stool samples. The three processing methods were optimized sucrose flotation (OSF) (28), simple one step stool (SOS) (29) and stool processing kit (SPK) (25). The results at the time of the GDG meeting showed a similar performance of Xpert Ultra in combination with the three stool processing methods, in terms of sensitivity and specificity. All methods were found to be easy to process by laboratory staff at the reference laboratory level and had a high ease-of-use score. However, most users considered that these methods cannot be performed by non-laboratory personnel (such as nurses or HCWs) in PHC settings without access to a laboratory. Overall, the SOS method appeared to be the preferred method as it does not require additional equipment and is comparable to sputum processing using Xpert Ultra. The SPK is still at prototype stage but will not be commercialized; the optimized sucrose flotation is still under validation, and its development into a kit format to simplify some of the steps is envisaged in the near future. Therefore, either of the available centrifuge-free stool processing methods may be used, depending on local preference and laboratory infrastructure.
Xpert Ultra trace results: Trace results are common with the use of Xpert Ultra in all paediatric specimen types, reflecting the paucibacillary nature of TB disease in children. For children as well as people living with HIV who are being evaluated for PTB, and for persons being evaluated for EPTB, the "M. tuberculosis complex (MTBC) detected trace" Ultra result is considered as bacteriological confirmation of TB (30). This is an important implementation consideration, in view of the risk of morbidity and mortality in these populations. Trace results will have an indeterminate result for rifampicin resistance; therefore, alternative specimens may need to be collected for Xpert Ultra processing in persons with a high likelihood of drug resistance.