P452 Serum Amyloid A for Predicting Surgery and Disease Exacerbation in Patients with Newly Diagnosed Crohn’s Disease

Abstract

Abstract Background Crohn's disease (CD) is characterized by relapsing-remitting course with highly heterogenous prognosis, ranging from quiescent disease to complicated course with stricturing, penetrating or need of abdominal surgery. Identifying biomarkers that determine their disease course can enable implementing more aggressive preventive strategies. Serum amyloid A (SAA) has been demonstrated to involve in the pathogenesis of Crohn’s disease (CD); however, its prognostic value is still uncertain. This study was determined to explore the prognostic value of SAA for CD-related surgery and disease exacerbation in newly diagnosed CD. Methods We conducted a retrospective cohort study and recruited newly-diagnosed CD patients at the First Affiliated Hospital of Sun Yat-sen University. Level of SAA was detected at diagnosis. The primary outcome was CD-related surgery. The secondary outcome was disease exacerbation, defining as newly developed structuring or penetrating lesions. Association between variables and outcome was evaluated through Kaplan–Meier method and Cox proportional hazard regression analysis. For continuous variables, the optimal cutoff values were determined by using the survminer R package in the analysis of primary outcome. The prognostic model was developed by cox proportional-hazards model. Concordance-index (c-index) was used to assess the predictive ability of the prognostic model. Results During 2187.6 person-years (median age, 28 years, 72.4% male), 87 surgery and 153 disease exacerbation events were documented. After adjustment, a 14% higher risk for surgery (adjusted hazard ratio, aHR [95% confidence interval]: 1.14 [1.05–1.23]; p=0.001) and a 12% higher risk for disease exacerbation (1.12 [1.05–1.19]; p<0.001) were observed with a 100-unit increase in SAA level. Patients with SAA levels ≥89.2 mg/L had aHRs of 2.08 [1.31–3.28] and 1.72 [1.22–2.41] for CD-related surgery and disease exacerbation, respectively. The association between SAA and outcomes was assessed as linear by the restricted cubic spline. Adding SAA into the model including known risk factors, including age at diagnosis ≥40 years, disease behavior, disease location, and initial biological treatment, led to significant improvement in its predictive ability for surgery (net reclassification index and integrated discrimination index, p<0.001) and disease exacerbation (net reclassification index and integrated discrimination index, p<0.001). Sensitivity analyses showed robust results. Conclusion Our study shows that higher serum SAA level at diagnosis was associated with poorer prognosis, including risk of surgery and disease exacerbation. And prognostic models based on SAA and known risk factors could better predict prognosis.