Glucocorticoid bioactivity does not predict response to steroid therapy in severe pediatric ulcerative colitis

Background: The pathophysiological basis for corticosteroid (CS) failure in ulcerative colitis (UC) is unknown. A transactivation glucocorticoid bioassay (GBA) was developed to measure the biological activity of CS by quantifying glucocorticoid response elements. This approach eliminates differences in bioavailability, chemistry, affinity, and other potential differences between the various steroids regarding their ability to activate the glucocorticoid receptor. In this multicenter prospective study, we aimed to evaluate whether CS bioavailability plays a role in CS refractoriness in severe pediatric UC. Methods: GBA (using COS-1 transfected cells) was measured in the serum of 50 children (52% males, age 13.4 ± 3.5 years) admitted for acute severe UC on the third day of CS treatment. Demographic, clinical, and laboratory data were prospectively recorded. Results: Of the children enrolled, 16 (32%) failed CS therapy and required infliximab (n = 14) or colectomy (n = 2) within a median of 10 days (interquartile range [IQR] 6.5-14.5). Reflecting internal validity of the assay, GBA was highly correlated with the last CS dose and the time interval to bloodletting (r = -0.41 and r = -0.54, respectively; P < 0.001). There was no statistically significant difference in the GBA levels between responders and nonresponders (249 nM versus 200 nM cortisol equivalent, P = 0.18). In a multivariate regression model adjusted for time elapsed from CS and the administered dose, GBA did not predict response to CS (P = 0.34). Conclusions: The lack of correlation of GBA level and treatment outcome lends support to the hypothesis that the bioavailability, type, and dosing of intravenous CS are not associated with response or failure to the drug.