P1080 Real-world Study on Early Predictors of Infliximab Treatment Efficacy in Crohn's disease

Abstract

Abstract Background Infliximab (IFX) can effectively induce and maintain the clinical remission of Crohn's disease (CD). However, a substantial number of CD patients experience primary or secondary nonresponse to IFX during treatment. Therefore, it is essential to identify early predictors of IFX treatment efficacy. Methods This study included 147 CD patients. Based on clinical outcomes, patients were categorized into IFX nonresponse and response group. We collected and compared laboratory data on blood routine examination, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), fecal calprotectin and post-induction IFX drug concentration and antibody levels in both groups. Logistic regression models were employed to identify potential factors associated with the risk of IFX nonresponse. Machine learning using random forest analysis was utilized to quantitatively assess the predictive features for IFX treatment efficacy and ROC curves was used to evaluate the model’s accuracy. Additionally, we performed HE staining and double immunofluorescence staining for CD64 and CD206 on intestinal biopsy tissues obtained before IFX treatment to compare the difference of macrophage subtypes between the nonresponse and response groups. Results Data from both cohorts revealed that patients in the IFX nonresponse group had lower drug concentration (P < 0.001), higher antibody levels (P < 0.001), and increasing ESR during the induction therapy (P < 0.001). Univariate and multivariate Logistic regression models demonstrated that IFX drug concentration and the ratio of ESR before and after induction therapy were associated with the risk of nonresponse. After the induction period, for each unit increase in drug concentration (1 μg/ml), the risk of IFX nonresponse decreased by 23% (RR= 0.77, 95% CI = 0.68-0.89), while each doubling of the ESR ratio after induction was associated with a 1.43-fold increase in the risk of nonresponse (RR= 2.43, 95% CI = 1.48-4.00). After combining data from the two cohorts, random forest machine learning analysis revealed that drug concentration below 1.5 μg/ml and an increase in ESR during induction could predict IFX nonresponse, with ROC curve areas of 0.740 and 0.651, respectively. Furthermore, immunofluorescence staining of intestinal mucosal biopsy tissues before IFX treatment showed a higher proportion of M1-type macrophages in the nonresponse group compared to the response group (P < 0.05). Conclusion Our study suggests that lower post-induction IFX drug concentrations and an increase in ESR during the induction phase are predictive of IFX nonresponse. Additionally, a higher proportion of M1-type macrophages in the intestinal mucosa before IFX treatment is associated with IFX nonresponse.