GeneXpert of stool versus gastric lavage fluid for the diagnosis of pulmonary tuberculosis in severely ill adults

Purpose Stool is an alternative specimen matrix for tuberculosis (TB) tests, because Mycobacterium tuberculosis (MTB) can be swallowed and detected in the samples from digestive tract. We aimed to assess the performance of GeneXpert on stool and gastric lavage fluid (GALF) in diagnosing TB among patients with severe pulmonary TB. Methods We enrolled adults with suspected pulmonary TB who were unable to produce sputum at visit between January 2016 and June 2018. Bacteriological samples consisted of one transtracheal aspirate sputum specimen, one stool specimen and/or one gastric lavage fluid specimen. Bacterial culture of transtracheal aspirate sputum provided the gold standard. Results Of 65 individuals recruited for analysis, MGIT culture identified the presence of MTB in 32 samples. Overall, 29 of 32 stool samples from culture-positive cases were detected by the GeneXpert test, demonstrating a sensitivity of 90.6%. For GALF, 13 patients were detected as infected with MTB by GeneXpert, yielding a sensitivity of 56.5%. The statistical analysis revealed that GeneXpert showed significantly better sensitivity in detecting MTB from stool samples than GALF samples (P = 0.003). Among individuals with GeneXpert-positive stool, the percentage of individuals with comorbid diabetes was significantly higher than among individuals with GeneXpert-negative stool (19.4% vs. 2.9%, P = 0.034). Conclusions In conclusion, our data reveal that GeneXpert provides a higher detection rate on stool compared to GALF, indicating stool should be considered as an alternative for adult TB patients unable to produce sputum. Individuals with diabetes are more likely to have positive GeneXpert stool than nondiabetic individuals.