Introduction Abdominal tuberculosis (TB) is difficult to diagnose, due to non-specific symptoms and radiological, histological and endoscopic similarity to other conditions. It can mimic Crohn’s disease and should always be considered as a differential diagnosis. There are significant risk of incorrectly diagnosing TB and committing patients to a prolonged course of toxic chemotherapy; or missing TB with public health implications and causing life-threathening disseminated TB when immunosuppressing patients. We sought to review the route to diagnosis of patients treated for abdominal TB, their sites of disease and the yield of various diagnostic modalities.

Methods A retrospective review of patients treated at St George’s Hospital, London, for abdominal TB from June 2003 to August 2013 was conducted. Information was gained from electronic patient records and the hospital’s tuberculosis database.

Results 65 cases of abdominal TB were identified. Average age was 42 years (range 18–97), with 49.2% females.

Pre-diagnosis: 49.2% underwent endoscopy, 64.6% ultrasound, 70.8% CT, 3.1% MRI and 10.8% small bowel series.

TB was cultured in 47.7% of patients, in the remaining 52.3% the diagnosis of abdominal TB was based on radiology, symptoms, suggestive histopathology, exclusion of other conditions or TB at another site.

The site was: peritonitis in 35.4%, enteritis in 27.7%, solid organ TB in 3.1%, combination of sites in 33.8%. 24.6% had co-existent pulmonary TB isolated on sputum culture.

The rate of culture positivity varied from modality of specimen acquisition as outlined in the table 1 case was resistant to isoniazid and streptomycin.

Conclusion Confirming a diagnosis of abdominal TB is notoriously difficult, with the rate of positive culture below 50% in our series. Non-invasive imaging is commonly used and is useful to characterise the phenotype of abdominal TB and suggest sites for sampling, however it does not assist in obtaining a definitive diagnosis. Invasive testing is a cornerstone of diagnosis. Ascitic fluid and surgically acquired biopsies had a higher diagnostic rate than endoscopy. There was a low rate of endoscopic biopsies being sent for Microbiology. If TB is part of the differential diagnosis endoscopists must ensure microbiological samples are taken into normal saline solution and sent for mycobacterial culture.

Disclosure of Interest None Declared.