Nasopharyngeal aspirates and stool sample collection for molecular tuberculosis testing in children with severe pneumonia - a mixed methods study

Background: Child-friendly sample collection methods are needed to improve microbiological tuberculosis diagnosis in children. We assessed the feasibility, diagnostic yield, safety, tolerability, and acceptability of nasopharyngeal aspirates (NPA) and stool sample collection for XpertMTB/RIF Ultra (Ultra) testing in children <5 years with severe pneumonia in six high tuberculosis incidence countries. Methods: We conducted a cross-sectional mixed-methods study (10/2020-03/2021) nested within the TB-Speed Pneumonia trial (NCT03831906) in 15 tertiary hospitals. Uptake and diagnostic yield of NPA and stool, sampled by trained nurses, were computed. Nurses documented NPA feasibility and related adverse events. NPA tolerability was assessed through discomfort/distress/pain scales completed by nurses, parents/caregivers, and children >3 years. Individual interviews investigated nurses and parents/caregivers perceptions and experiences of both procedures. Results: Of 542 children enrolled, 541 (99.8%) and 423 (75.5%) had NPA and stool samples successfully collected, respectively; 8 (1.5%) had positive Ultra on any sample, 8 (1.5%) on NPA, and 3 (0.6%) on stool samples. Nurses reported that NPA was difficult to collect in 27 (5.0%) children. 130 (24%) children had bloodstained NPA samples and <5% experienced transient heart rate deceleration, digestive or respiratory distress during NPA. Median nurse-assessed tolerability scores were 1/10 [0,3] before and 5/10 [2,6] during NPA. Although most respondents perceived NPA as painful, all appreciated that NPA aimed to improve child health. Nurses reported that NPA and stool collection was quicker, less invasive than gastric aspiration. Parents/caregivers trusted nurses skills and valued the novelty of NPA. All respondents easily engaged in stool collection, but were frustrated to depend on children passing stool. Nurses underlined that continuous staff training and mentoring would sustain feasibility and self-efficacy in obtaining adequate samples. Conclusions: NPA and stool are viable specimen collection options for microbiological diagnosis of tuberculosis in young children <5 years with severe pneumonia in high tuberculosis incidence countries.