Pulmonary tuberculosis diagnosis is difficult when patients cannot produce sputum. Most sputum is swallowed and tuberculosis DNA may survive intestinal transit, so stool specimens may be useful for detecting tuberculosis originating from the lungs. Both PCR and culture can be performed on stool samples, and IFHAD has studied the utility of both of these methods.  PCR is more sensitive than culture, but because culture is more feasible in many resource-poor settings, both techniques are important options

Using PCR to Diagnose TB from Stool Samples 

Methods:  Paired stool and sputum samples were collected from patients with sputum culture-proven pulmonary tuberculosis. Control stool samples were collected from patients without tuberculosis symptoms.  The diagnostic accuracy of the polymerase chain reaction (PCR) in stool was compared with sputum testing by PCR and culture.  A hemi-nested IS6110-PCR was used for tuberculosis detection and IS6110-PCR positive stool samples then had rifampicin sensitivity-testing by heteroduplex-PCR.

Results:  59/69 stools from newly diagnosed tuberculosis patients vs. 0/47 stools from controls were IS6110-PCR positive (sensitivity 86%; specificity 100%).  For these diagnostic samples, stool PCR had similar sensitivity for HIV-positive and HIV-negative patients (94% vs. 83% respectively, P=0.4) and PCR sensitivity was similar for stool vs. sputum samples (86 % vs. 91%, P=0.3).  There was 98% agreement between rifampicin susceptibility-testing by sputum culture vs. stool heteroduplex-PCR. Heteroduplex-PCR correctly predicted multi-drug resistant tuberculosis in 98% of cases. Detection and susceptibility-testing by PCR took 1-2 days compared with an average of nine weeks for traditional culture-based testing. Considering all 159 stool samples from patients before and during treatment, stool PCR was more sensitive for patients with sputum microscopy-positive tuberculosis (P<0.001) and remained positive for most patients after the first week of therapy.

Conclusions:   Stool PCR is a sensitive, rapid and relatively biosecure technique for the diagnosis and drug-susceptibility testing of pulmonary tuberculosis and should be considered when sputum samples are unavailable.

Using Culture to Diagnose TB from Stool Samples

Rapid diagnosis of pulmonary tuberculosis with the Microscopic Observation Drug Susceptibility (MODS) assay applied to human stool samples
Laura Martin, Louis Grandjean, R.H.Gilman, Luz Caviedes, Ron Shiloh, Vivian Kawai, Giselle Soto,  Paty Fuentes, Mirko Zimic, A. Roderick Escombe, David A. J. Moore, Carlton A Evans

Diagnosing pulmonary tuberculosis is difficult in patients who cannot produce sputum.  289 stool samples were collected to test whether Mycobacterium tuberculosis bacilli in swallowed sputum may be detected in stool to diagnose pulmonary tuberculosis.  All patients were sputum culture-positive and the diagnostic sensitivity of culturing paired stool samples was 60% with the MODS assay compared with 50% for Selective 7H10 agar, 43% for Löwenstein-Jensen agar, and 36% for auramine microscopy.  HIV-infection had no effect on sensitivity (P=0.8). MODS identified positive cultures most rapidly (median 11 days, P<0.001) and simultaneously indicated antibiotic susceptibility with 100% concordance with other tests. 47 stool specimens from healthy controls tested negative (specificity 100%).  Patients with multi-drug resistant tuberculosis remained stool culture positive throughout tuberculosis treatment. Sensitivity was lower for stool than sputum culture (P<0.001) despite similar contamination rates. IS6110 PCR confirmed speciation of Mycobacterium tuberculosis for 98% of stool cultures. Therefore, stool culture with MODS is effective for diagnosing pulmonary tuberculosis and should be considered when sputum is unavailable.