Yes, in rare cases mycobacteria can be isolated from stool. But that does not indicate tuberculosis at all.

The diagnosis requires clinical evidence (unspecific abdominal and/or general signs, enlarged abdominal  lymph nodes, positive IGRA or Mendel-Mantoux test) and species diagnosis of the AFB´s by conventional or molecular (PCR) methods.

The mere presence of acid fast bacilli in stool does not prove that the patient is infected, though it can happen. As Warner Solbach says, clinical co-relation along with other evidence (laboratory and otherwise) is required to confirm or refute the diagnosis. In the Indian context IGRAs will not help and actually they are banned by the Govt. of India (as with all other serological based assays). Secondly after culture, species identification is required. If it actually turns out to be M. tuberculosis, there is a reasonable chance of true infection.

Of course it is because this might be pulmonary or intestinal type or generalized TB case.

The diagnosis of Extrapulmonary Tb is a challenge ,requiring often invasive interventions. In susceptible individuals such as immunocompromised subjects, identification of MTB in stool could really help for the diagnosis and early treatment of of TB. Gastric aspiration for MTB detection is still used in younger patients or those not able to produce sputum.

data indicate that stools could be used in conjunction with sputum testing or as an alternative specimen upon which to base the diagnosis of  tuberculosis

Thankx for all for your valuable answers. I worked on HIV positive sputum smear negative cases and irrespective to that of clinical conditions i collected stool samples. Agreeing to Mr Lawrence and Werner , though smears were negative some showed culture positive and proved to be MTB  and somewhere NMTB when identified by LINE probe assay. Of course in these very few patients exhibited diarrhea & abdomen pain but others no symptoms. 

I agreed with Solbach and Mawlong. You shall check the clinical features and detect the non-tuberculosis mycobacteria isolated from stool.

Umamaheshwari,

If  M. tuberculosis was isolated from stool AND the patient is HIV positive, there are two common possibilities, GI infection or pulmonary infection (where the patient swallows expectorated sputum, re-routing it to the GI tract). I would treat the patient as TB and HIV co-infection is becoming increasingly common in India (also they are not necessarily symptomatic). Diarrhoea and abdominal pain are not specific for TB. For the NTM however i would be a bit more careful and have a more thorough work up done.

I agree with Mawlong, nowadays the cases of  HIV co-infection with TB are increasing especially in Southeast Asia & mostly asymptomatic TB infection. So if we can isolate M.tuberculosis from any spesimen including stool in HIV patients, we have to treat them.

Yes. Diagnosis in stool has very low sensitivity, but is specific. If positive is TB and should be treated.

yes,I agree with AlfonoseJ. idea

I want to stress my previous comment. Mere detection of mycobacteria does not justify Treatment. A thorough species diagnosis has to be made and MOTTs have to be excluded. If M. tuberculosis or M. bovis is found in stool a thorough evaluation of the patient decides whether Treatment is necessary or not. This includes IGRA testing or Mantoux Skin testing and exclusion of co-infections like HIV. Also, a high-Quality resistence testing is mandatory.

I do agree with Mr Werner. Species diagnosis is must to decide treatment

Can I request whoever had supporting evidence for conferring treatment diagnosed with Stool Smear for AFB. We are going to introduce GeneXpert MTB/RIF and Smear using stool specimen in our study. thank you

Please have a look at these useful RG links.

Article Detection of Mycobacterium tuberculosis in pediatric stool s...

Article Stool culture for the diagnosis of pulmonary tuberculosis in children