LAUSR

BACKGROUND This study assesses the impact of the Xpert MTB/RIF in the diagnosis of childhood tuberculosis (TB) in a rural hospital in a resource-constrained setting. METHODS Retrospective cross-sectional study in children evaluated for presumptive TB from 1 June 2016 to 31 May 2017 at the Gambo General Hospital in rural Southern Ethiopia. Children were evaluated according to a defined protocol based on national guidelines. Samples were submitted for Xpert MTB/RIF assay to the nearest reference laboratory. RESULTS Of the 201 children assessed for presumptive TB, 46.3% (93/201) were diagnosed with TB. Of these, 49.5% (46/93) were microbiologically confirmed, mostly by Xpert MTB/RIF (only one patient was diagnosed by smear alone). The rest were clinically diagnosed. Microbiologically confirmed patients had a higher mean age, longer duration of fever and cough and lymphadenopathy more frequently than those clinically diagnosed. Gastric aspirates were Xpert MTB/RIF-positive in 18.2% of the samples (26/143); none were smear-positive (0/140). Sputum samples were Xpert MTB/RIF-positive in 27.1% (13/35) of the samples and smear-positive in 8.6% (3/35). There were no HIV-positive patients and just one case of rifampicin-resistant TB. A long delay (median 15 days) was detected in returning the results. CONCLUSION Xpert MTB/RIF serves as an important adjunctive test for diagnosing childhood TB in rural settings, with microbiological confirmation in up to half the TB cases. Processes need to be optimized to achieve an early diagnosis. The diagnosis of childhood TB in high-burden countries such as Ethiopia still relies largely upon diagnostic algorithms and the clinician's skills.Lay summaryWorld Health Organization recommends the use of Xpert MTB/RIF to improve the microbiological diagnosis of childhood tuberculosis (TB) since 2014, but the impact of this test under real conditions in rural areas of low-income countries is not clear. We conducted a cross-sectional study in children evaluated for presumptive TB from 1 June 2016 to 31 May 2017 at the Gambo General Hospital in rural Southern Ethiopia. Children were evaluated according to a clinical protocol based on national guidelines and samples were submitted for Xpert MTB/RIF assay to the nearest reference laboratory.Of the 201 children assessed, 46.3% (93/201) were diagnosed with tuberculosis. Of these, 48.4% (45/93) were microbiologically confirmed by Xpert MTB/RIF [smear microscopy only diagnosed the 5.4% (5/93)]. Patients with microbiologically confirmed tuberculosis had a higher mean age, longer duration of fever and cough and had lymphadenopathy more frequently than those clinically diagnosed. A long delay in returning the results (median 15 days) was detected. Xpert MTB/RIF serves as an important test for diagnosing childhood TB in rural settings, with microbiological confirmation in up to half the cases. Processes need to be optimized to achieve an early diagnosis. The diagnosis of childhood TB in high-burden countries still relies largely upon diagnostic algorithms and the clinician's skills.