Routine diagnosis of intestinal tuberculosis and Crohn's disease in Southern India.

OBJECTIVE

To investigate whether routinely measured clinical variables could aid in differentiating intestinal tuberculosis (ITB) from Crohn's disease (CD).

METHODS

ITB and CD patients were prospectively included at four South Indian medical centres from October 2009 to July 2012. Routine investigations included case history, physical examination, blood biochemistry, ileocolonoscopy and histopathological examination of biopsies. Patients were followed-up after 2 and 6 mo of treatment. The diagnosis of ITB or CD was re-evaluated after 2 mo of antituberculous chemotherapy or immune suppressive therapy respectively, based on improvement in signs, symptoms and laboratory variables. This study was considered to be an exploratory analysis. Clinical, endoscopic and histopathological features recorded at the time of inclusion were subject to univariate analyses. Disease variables with sufficient number of recordings and P < 0.05 were entered into logistic regression models, adjusted for known confounders. Finally, we calculated the odds ratios with respective confidence intervals for variables associated with either ITB or CD.

RESULTS

This study included 38 ITB and 37 CD patients. Overall, ITB patients had the lowest body mass index (19.6 vs 22.7, P = 0.01) and more commonly reported weight loss (73% vs 38%, P < 0.01), watery diarrhoea (64% vs 33%, P = 0.01) and rural domicile (58% vs 35%, P < 0.05). Endoscopy typically showed mucosal nodularity (17/31 vs 2/37, P < 0.01) and histopathology more frequently showed granulomas (10/30 vs 2/35, P < 0.01). The CD patients more frequently reported malaise (87% vs 64%, P = 0.03), nausea (84% vs 56%, P = 0.01), pain in the right lower abdominal quadrant on examination (90% vs 54%, P < 0.01) and urban domicile (65% vs 42%, P < 0.05). In CD, endoscopy typically showed involvement of multiple intestinal segments (27/37 vs 9/31, P < 0.01). Using logistic regression analysis we found weight loss and nodularity of the mucosa were independently associated with ITB, with adjusted odds ratios of 8.6 (95%CI: 2.1-35.6) and 18.9 (95%CI: 3.5-102.8) respectively. Right lower abdominal quadrant pain on examination and involvement of ≥ 3 intestinal segments were independently associated with CD with adjusted odds ratios of 10.1 (95%CI: 2.0-51.3) and 5.9 (95%CI: 1.7-20.6), respectively.

CONCLUSIONS

Weight loss and mucosal nodularity were associated with ITB. Abdominal pain and excessive intestinal involvement were associated with CD. ITB and CD were equally common.