P198 Dipeptidyl peptidase 4: a strong predictive marker of disease activity and treatment escalation in inflammatory bowel diseases

Abstract

Background: Blood-based biomarkers provide a non-invasive estimation of the inflammatory burden in patients with inflammatory bowel diseases (IBD). We aimed to investigate the diagnostic and prognostic value of plasma Dipeptidyl peptidase 4 (DPP-4) in IBD patients. Methods: A total of 203 adult patients [n=149 IBD patients; n=42 healthy controls; n=12 immune controls - systemic lupus erythematosus (SLE) in remission] were prospectively recruited. Disease activity was assessed using the Harvey-Bradshaw Index (HBI) for Crohn's disease (CD), the partial Mayo Score (pMS) for Ulcerative colitis (UC) and the Systemic Lupus Erythematosus Disease Activity Index for SLE. A multi-biomarker model was derived using logistic regression to evaluate predictors of disease activity and Cox regression to evaluate predictors of treatment escalation (disease outcome). Treatment escalation was defined as the need for escalation to immunomodulatory/biologic therapies or intestinal resection surgery, as a consequence of a disease flare. Results: Median DPP-4 values were lower in active CD vs CD in remission [1043ng/mL, interquartile range (IQR): 824–1345 vs 1685ng/mL, IQR: 1519–2237; p<0.001]; as well as in active UC vs UC in remission (1323ng/mL, IQR: 1064–1766 vs 2149ng/mL, IQR: 1616–2478; p<0.001). IBD patients in remission showed DPP-4 values significantly lower than healthy subjects; no differences were found between immune and healthy controls. In CD, DPP-4 correlated strongly with faecal calprotectin (FC) (r=−0.61, p<0.001), C-reactive protein (CRP) (r=−0.60, p<0.001) and HBI (r=−0.56, p<0.001). In UC, DPP-4 correlated moderately with FC (r=−0.31, p<0.05) and CRP (r=−0.32, p<0.05), and strongly with pMS (r=−0.52, p<0.001). The multivariable logistic regression model revealed that DPP-4 and CRP (in CD) and DPP-4, FC and CRP (in UC) are independent predictors of disease activity (Table 1). Table 1. Multiple logistic regression model of predictors of disease activity At follow-up (median 578 days, IQR: 426–688), DPP-4 and CRP showed to independently predict treatment escalation in both CD and UC (Table 2). Table 2. Cox proportional hazards model for predictive factors of treatment escalation At 1 year, the proportion of patients who needed treatment escalation was 66% in CD and 41% in UC, if ≥2 biomarkers criteria were met. Conclusions: An activity assessment model and a prognostic model that combines DPP-4 and other biomarkers accurately predicts IBD activity and an aggressive disease course.