Clinical pattern of corticosteroid dependent Crohn's disease

Corticosteroid dependency in Crohn's disease (CD) is characterized by the need for chronic use of corticosteroids to maintain clinical remission. Several definitions have been used. Depending on the definition, 10-30% of the patients are considered as corticosteroid dependent. The aim of the study was to define a clinical pattern of corticosteroid dependent CD patients.

Epidemiological and clinical characteristics were retrospectively compared between 20 corticosteroid dependent CD patients and 248 non-corticosteroid dependent CD patients by using univariate and multivariate analyses. Corticosteroid dependency was defined either by two successive relapses during the 2 months after discontinuing glucocorticoids (n=5) or by two successive relapses at dose tapering, after successful treatment of a flare-up by using glucocorticoids (n=15).

Corticosteroid dependent CD patients were younger at diagnosis (P < 0.001), and were characterized by a higher frequency of colonic location (P< 0.05), but lower frequency of ileal location (P < 0.01), and higher ano-perineal location (P < 0.05). They were also more frequently smokers (P< 0.05) and users of contraceptive pills (P< 0.01). The inflammatory type of the disease was increased (P < 0.01), while the fibrostenotic type was decreased (P < 0.001) in corticosteroid dependent CD patients. By multivariate analysis, a smoking habit (P < 0.01), a colonic location (P < 0.05), a non-fibrostenotic type (P< 0.05) and a younger age at diagnosis (P< 0.05) were shown to be independently associated with corticosteroid dependency.

This study suggests a clinical pattern associated with corticosteroid dependency. Whether this clinical pattern is simply associated with the dependency, or whether a primary decrease of corticosensitivity produces this clinical pattern, is not known. Further prospective studies will have to determine whether the presence of these clinical characteristics is predictive of corticosteroid dependency and whether this prediction will be useful for the management of this condition.