Evaluation and validation of a Crohn's disease inflammatory activity index reflecting pattern of endoscopic severity

Background: This study was designed to investigate objective variables assessing the inflammatory activity of Crohn's disease accessible for routine management and their suitability to act as surrogate indicators for endoscopic alterations. Methods: Thirty-six patients were included in a prospective study and underwent endoscopic investigation, 18 with clinically exacerbated disease and 18 after acute-phase conservative therapy. The endoscopic findings were classified to define the dependent variable, yielding two categoric levels: acute active disease and remission. The extent of affected mucosal area was not taken into consideration. The serum variables alpha(1)-antitrypsin, acid alpha(1)-glycoprotein (AGP), C-reactive protein (CRP), sialic acids, prealbumin (PAB), and albumin were used as independent variables to predict the dependent variable. To compare the results with common indices, the Crohn's Disease Activity Index (CDAI) and van Hees Index were calculated. Results: The following model was calculated by stepwise logistic regression analysis: if AGP (mg/dl) - 4.2 x PAB (mg/dl) greater than or equal to 0.8, then endoscopically active disease will be predicted with a sensitivity of 100% and a specificity of 95% (P < 0.001). The predictive values of the single variables, CDAI, and van Hees Index were lower. For validation of results an analogous study was performed including 44 patients, 29 with active disease and 15 controls. The existence of the model was confirmed, again showing high values for sensitivity (86%) and specificity (100%). Conclusions: On a qualitative level focusing on clinical relevance, the endoscopic and biologic findings of Crohn's disease are highly associated. In addition to clinical assessment, usage of the developed index as a rationale contributing to therapeutic decisions in the short- and long-term management might be reasonable.