Rapid Diagnosis of Pediatric Intrathoracic Tuberculosis From Stool Samples Using the Xpert MTB/RIF Assay

To the Editors: Early rapid confirmation of tuberculosis (TB) in young children can prevent disease progression and death. The Xpert MTB/RIF assay (Xpert; Cepheid, Sunnyvale, CA), an automated real-time polymerase chain reaction assay, detects the presence of Mycobacterium tuberculosis in sputum within 2 hours. There are no published studies regarding the use of Xpert on stool or gastric aspirates (GA) for diagnosing pediatric intrathoracic TB. GA samples are commonly obtained from children with suspected TB in resource-limited settings; stool sampling is attractive as it is noninvasive. From August to November 2011, we enrolled children less than 14 years of age with suspected intrathoracic TB from 2 hospitals in Cape Town in a cross-sectional pilot study. Suspected TB was ≥1 of the following: (1) symptoms suggestive of TB; (2) recent close contact with a TB source case; (3) reactive Mantoux skin test; (4) chest radiograph suggestive of TB, determined by 2 independent readers.1 From each participant, 1 GA and 1 stool sample, if obtained, were analyzed by smear microscopy, culture and Xpert. GA samples were neutralized with 4% sodium bicarbonate. Stool samples were homogenized by adding normal saline solution and by manual shaking. After decontamination and neutralization, 2 centrifugation steps of 20 minutes were performed. Supernatant was decanted; 30 µL of the concentrated pellet was used for auramine O-microscopy.2 The remaining pellet was resuspended in 1.5 mL phosphate buffer. Of this, 0.5 mL was inoculated into a supplemented Mycobacterial Growth Indicator Tube (Becton-Dickinson, Sparks, MD), and incubated at 37°C in a Bactec MGIT 960 machine. For Xpert, 0.7 mL of the resuspended pellet was mixed with 1.4 mL sample reagent and manually inverted before and during 15 minutes incubation. This mixture was transferred into 1 Xpert cartridge, loaded into the GeneXpert machine and processed automatically. Of 28 eligible children, we included 23 (15 female, 2 HIV-infected, median age 17.2 months). In 5, we failed to obtain consent and/or samples before discharge. Stool and GA samples were obtained from 14 children; 6 had only GA and 3 had only stool collected. Twelve children received a diagnosis of TB, of whom 6 were culture-confirmed and 2 had extrathoracic TB. Overall, Xpert on stool and GA detected 3 of 4 (75%) children with intrathoracic TB and positive GA cultures, and 3 of 6 (50%) children with M. tuberculosis cultured from any site. Time to culture positivity was 5–18 (GA) and 22–32 (stool) days. No contaminated cultures, indeterminate Xpert results or drug resistance were reported. This study suggests that Xpert on stool might be useful for the diagnosis of pediatric intrathoracic TB, as Xpert on stool and GA was equally sensitive and gave more rapid results than culture. We used longer initial centrifugation times to increase sample concentration and to improve the detection of M. tuberculosis in stool. Stool sampling is “child-friendly” and has minimal infectious risk, overcoming some barriers to bacteriologic investigation of TB in children. Rapid confirmation using noninvasive sampling may be particularly useful in young children with extensive TB or suspected drug resistance, and its potential clinical impact should be systematically investigated. Elisabetta Walters, MMed Robert Peter Gie, MD Anneke Catharina Hesseling, PhD Desmond Tutu TB Centre Stellenbosch University Sven Olaf Friedrich, PhD Andreas Henri Diacon, MD, PhD Division of Medical Physiology Department of Biomedical Sciences Faculty of Medicine and Health Sciences Stellenbosch University Robert Peter Gie, MD Paediatric Pulmonology Tygerberg Children’s Hospital Cape Town, South Africa