Validation of a simplified clinical index to predict evolving patterns in Crohn's disease

Prior to this study we obtained a simple mathematical index that uses clinical variables to predict the evolution of Crohn's disease to either a stricturing or a penetrating type. This model was based on the following variables: duration of disease before diagnosis (DD), onset of symptoms (OS), anal disease (AD) and abdominal mass (AM). The aim of our study was to validate this model in an independent cohort of patients and to investigate the relationship between some of the variables and the actual pattern of Crohn's disease in the patients.

We prospectively evaluated 128 patients with Crohn's disease at the moment of diagnosis. We predicted the evolution of their disease using the mathematical model Z = -9.49 + 2.2643 (AD) - 0.0066 (DD) + 2.5282 (AM) + 1.3433 (OS). The cut-off value (reveiver operating characteristics curve) obtained in the training set of patients was P = 0.45. A value higher than this cut off discriminated patients who developed a stricturing pattern. The actual behaviour of the patients' Crohn's disease was observed after a median of 19 months from diagnosis. Of the 128 patients, 80 were classified into one of the two known patterns. Thirty-nine patients (48.8%) developed a stricturing pattern while 41 (51.2%) had a penetrating form of Crohn's disease.

The sensitivity of the model for predicting a stricturing type was 100% and the specificity was 31.7%. A P value of < 0.45 proved to be highly reliable in predicting the evolution to a penetrating pattern (positive predictive value was 100% and negative predictive value was 58%). No statistical differences were found between stricturing-type or penetrating-type groups in terms of anal disease, abdominal mass, duration of disease or onset of symptoms. Compared to patients with the penetrating form, initial ileal location was significantly more frequent than colonic location in patients with the stricturing type of Crohn's disease.

We have validated a simple mathematical model that is able to predict the behaviour of Crohn's disease in patients based on clinical variables collected at their initial evaluation. This model can be considered a useful tool for patient management. The anatomical location of the disease is related to the evolutive pattern.