Li et al reported 2 models for differentiating Crohn's disease from intestinal tuberculosis. This can be an important differential diagnosis in developing countries. The authors are from 2nd Xiangya Hospital, Central South University and the Affiliated Hospital of University of Jishou in Hunan, China.
Patient selection: Crohn's disease vs intestinal tuberculosis
Exclusions: caseating granulomas or acid-fast bacilli seen in tissue
Parameters for clinical model:
(1) hematochezia
(2) surgical history
(3) perianal disease
(4) pulmonary tuberculosis
(5) ascites
(6) PPD positive
Parameter |
Finding |
Points |
hematochezia |
absent |
0 |
|
present |
1.409 |
history of surgery |
absent |
0 |
|
present |
2.798 |
perianal disease |
absent |
0 |
|
present |
2.713 |
pulmonary tuberculosis |
absent |
0 |
|
present |
-4.728 |
ascites |
absent |
0 |
|
present |
-2.066 |
PPD positive |
absent |
0 |
|
present |
-2.414 |
X =
= SUM(points for all 6 parameters) + 0.708
Parameters for endoscopic model:
(1) rectum involved
(2) longitudinal ulcers
(3) cobble-stone appearance
(4) fixed open ileocecal valve
(5) ring ulcer
(6) rodent ulcer
Parameter |
Finding |
Points |
rectal involvement |
absent |
0 |
|
present |
1.499 |
longitudinal ulcers |
absent |
0 |
|
present |
1.753 |
cobble-stone appearance |
absent |
0 |
|
present |
2.787 |
fixed open ileocecal valve |
absent |
0 |
|
present |
-1.432 |
ring ulcer |
absent |
0 |
|
present |
-2.379 |
rodent ulcer |
absent |
0 |
|
present |
-3.343 |
Y =
= SUM(points for all 6 parameters) + 0.283
probability of Crohn's disease based on clinical findings =
= 1 / (1 + EXP((-1) * X))
probability of Crohn's disease based on endoscopic findings =
= 1 / (1 + EXP((-1) * Y))
probability of tuberculosis =
= 1 - (probability of Crohn's disease)
Specialty: Gastroenterology, Infectious Diseases