INT J TUBERC LUNG DIS 23(3):315–321 Q 2019 The Union http://dx.doi.org/10.5588/ijtld.18.0404 Initiation of anti-tuberculosis treatment in children following gastric aspirate testing, Botswana, 2008–2012 T. Q. Lo,* †‡ L. Matlhare, § K. Mugisha, ¶ T. D. Lere, ¶ A. Ho-Foster, §# R. Boyd,** J. Cavanaugh, ‡ R. Ncube, ¶ A. P. Steenhoff, §#†† T. Arscott-Mills §#†† *Epidemic Intelligence Service, † Division of Global Health Protection, and ‡ Division of TB Elimination, Centers for Disease Control and Prevention (CDC), Atlanta, Georgia, USA; § Botswana-UPenn Partnership, Gaborone, ¶ National TB Programme, Botswana Ministry of Health, Gaborone, Botswana; # University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA; **CDC, Gaborone, Botswana; †† The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA SUMMARY SETTING: Diagnosing pediatric tuberculosis (TB) is difficult; to improve diagnosis, gastric aspiration (GA) was performed in 121 Botswana health facilities. OBJECTIVE: To describe treatment initiation and out- comes in children with a positive GA result and those treated empirically. METHODS: Children with smear or culture-positive GA or those clinically diagnosed were referred for anti- tuberculosis treatment. Treatment initiation and out- comes were assessed from February 2008 to December 2012 using name-based matching algorithms of the GA database; treatment initiation was captured in the electronic TB registry. Analyses included descriptive statistics and regression models. RESULTS: GA was conducted in 1268 children. Among these, 121 (9.5%) were GA-positive; and treatment was initiated in 90 (74.3%). An additional 137 (11.9%) were treated empirically. More than a third (36.4%) had known human immunodeficiency virus status (positive or negative); this was significantly associated with TB treatment initiation (adjusted odds ratio [aOR] 1.8, 95%CI 1.3–2.5); P , 0.05). Among the 90 children with a positive GA result, nearly all either completed treatment (78.9%) or were on treatment (20.0%) at the time of data collection. CONCLUSION: We could not find documentation of treatment for more than a quarter of the children with laboratory-confirmed TB, an important gap that calls for further examination. The failure to initiate prompt treatment requires investigation and urgent action. KEY WORDS: gastric aspirate; pediatric TB; Botswana ALTHOUGH THERE WERE MORE THAN 1 million new cases of tuberculosis (TB) among children in 2015, representing 10–11% of the global total, 1 this burden is substantially underestimated. 2,3 Diagnosing childhood TB is difficult—children have weak cough and are unlikely to produce sputum. 4 Even in the case of a productive cough, sputum is swallowed before expectorating, and obtaining an adequate specimen is challenging. 4,5 Children fre- quently have paucibacillary pulmonary TB disease, which results in poor microbiological confirmation from sputum. 6,7 Gastric aspiration (GA) is a method of obtaining samples from children who cannot produce spu- tum. 8–10 GA involves insertion of a nasogastric tube into the stomach, with aspiration of the gastric contents. 11 GA is technologically simpler than sputum induction (SI) by an induction machine, and can be used for children with contraindications to SI. GA is recommended for children by the World Health Organization (WHO) 12,13 and, since 2011, has been part of Botswana’s National TB Programme (NTP) guidelines. 14 Diagnosis is the initial step in the treatment cascade. 15 To maximize the contribution of diag- nostic strategies and impact the pediatric TB epidemic, anti-tuberculosis treatment must be initi- ated in diagnosed children. Pre-treatment loss to follow-up in TB patients in low- to middle-income countries ranges from 4% to 38%. 15 Research describing treatment initiation and outcomes for diagnosed children is scarce. 16 We aimed to describe treatment initiation and outcomes in children with a positive GA result and those started on empiric treatment in the Botswana- UPenn Partnership (BUP) program. Correspondence to: Terrence Lo, Division of Global Health Protection Epidemiology, Informatics, Surveillance, and Laboratory Branch, Centers for Disease Control and Prevention, 1600 Clifton Road NE, MS-E93, Atlanta, GA 30329, USA. e-mail: terrencelo@cdc.gov Article submitted 11 June 2018. Final version accepted 27 August 2018.