A Panel to Predict Long-term Outcome of Infliximab Therapy for Patients With Ulcerative Colitis

BACKGROUND & AIMS: Infliximab is effective for patients with refractory ulcerative colitis (UC), but few factors have been identified that predict long-term outcome of therapy. We aimed to identify a panel of markers associated with outcome of infliximab therapy to help physicians make personalized treatment decisions. METHODS: We collected data from the first 285 patients with refractory UC (41% female; median age, 39 y) treated with infliximab before July 2012 at University Hospitals Leuven, in Belgium. We performed a Cox regression analysis to identify independent factors that predicted relapse-free and colectomy-free survival, and used these factors to create a panel of markers (risk panel). RESULTS: During a median follow-up period of 5 years, 61% of patients relapsed and 20% required colectomy. Independent predictors of relapse-free survival included short-term complete clinical response (odds ratio [OR], 3.75; 95% confidence interval [CI], 2.35-5.97; P <.001), mucosal healing (OR, 1.87; 95% CI, 1.17-2.98; P =.009), and absence of atypical perinuclear antineutrophil cytoplasmic antibodies (pANCA) (OR, 1.96; 95% CI, 1.23-3.12; P =.005). Independent predictors of colectomy-free survival included short-term clinical response (OR, 7.74; 95% CI, 2.76-21.68; P <.001), mucosal healing (OR, 4.02; 95% CI, 1.16-13.97; P =.028), baseline level of C-reactive protein (CRP) of 5 mg/L or less (OR, 2.95; 95% CI, 1.26-6.89; P =.012), and baseline level of albumin of 35 g/L or greater (OR, 3.03; 95% CI, 1.12-8.22; P =.029). Based on serologic analysis of a subgroup of 112 patients, levels of infliximab greater than 2.5 mg/mL at week 14 of treatment predicted relapse-free survival (P <.001) and colectomy-free survival (P =.034). A risk panel that included levels of pANCA, CRP, albumin, clinical response, and mucosal healing identified patients at risk for UC relapse or colectomy (both P <.001). CONCLUSIONS: Clinical response and mucosal healing were confirmed as independent predictors of long-term outcome from infliximab therapy in patients with UC. We identified additional factors (levels of pANCA, CRP, and albumin) to create a risk panel that predicts long-term outcomes of therapy. Serum levels of infliximab at week 14 of treatment also were associated with patient outcomes. Our risk panel and short-term serum levels of infliximab therefore might be used to guide therapy.