P486 Assay of infliximab trough levels and of total antibodies to infliximab in the management of loss of response
L. Guidi*1, M. Marzo1, B. Tolusso2, C. Felice1, G. Andrisani1, S. Canestri2, D. Pugliese1, E. Gremese2, A. Papa1, G.L. Rapaccini1, A. Armuzzi1
1Catholic University, IBD Unit, Complesso Integrato Columbus, Rome, Italy, 2Catholic University, Rheumatology, Complesso Integrato Columbus, Rome, Italy
Optimization of infliximab therapy in inflammatory bowel disease (IBD) patients who lose the clinical response is currently performed empirically. The assay of antibodies to infliximab (ATI) and infliximab trough levels (TL) could be helpful in the clinical decision.
We collected sera from 102 IBD patients on maintenance infliximab. Fifty-five patients had experienced a loss of response (LOR) and most of them (45/55) had the infliximab dose already been optimized at 7.5 to 10 mg/kg. As a control group we studied 47 patients in stable clinical response (CR). Blood was drawn for ATI, TL and C-reactive protein (CRP) assay immediately before the next infliximab infusion. We assayed ATI and TL by an ELISA kit (Immundiagnostik AG, Bensheim, FRG). ATI were expressed as arbitrary units (AU)/ml (a level of 10 AU/ml is considered as positive for antibody presence in the serum). TL were expressed as ug/ml. The subsequent clinical decisions were made blinded to ATI and TL levels. Clinical activity and CRP were also assessed after 12 months.
Fifty-one patients had Crohn's disease and 51 had ulcerative colitis. Twenty-three were on concomitant immune modulators (IM). Median ATI were 24,5 AU/ml (IQR 7,4 to115,8) and 66 (65%) had ATI score higher than 10 AU/ml, while median TL were 2,3 ug/ml (IQR 0 to 5,6). ATI and TL were negatively correlated (Spearman's rho -0,579, p<0.0001). ATI were lower in the patients on concomitant IM, as compared to those not (median 10,6 AU/ml, IQR 5,5 to 48,6 versus 27,5 AU/ml, IQR 7,8 to 128,3, p= 0,045 by Mann-Whitney test). Mean TL were lower in the patients with LOR and standard infliximab dosage (0,49 ug/ml) as compared to those on optimized dosage (2,96 ug/ml) and those on CR (2,59 ug/ml) although these differences were not statistically significant. ATI levels were significantly higher in the LOR group as compared to the CR (40,8 AU/ml, IQR 9,8-189,7 versus 11,4 AU/ml, IQR 6,5 to 82,3, p = 0,037 by Mann-Whitney test). A ROC analysis detected a cut off value of ATI for CR at 12 months of < 26,3 AU/ml (AUC 0,693, 95% CI 0,594 to 0,780, p= 0,0033) with sensitivity of 64% and specificity of 75%. The same analysis for TL detected a cut off of >2,08 ug/ml for predicting CR at 12 months with 56,4% sensitivity and 70,8% specificity (AUC 0,628, 95% CI 0,527 to 0,722, p= 0,038). Among patients with LOR, 62% had ATI > 26,3 AU/ml, as compared to 32% of those maintaining the response (p=0,003 by Fisher's exact test). A significant correlation was detected between ATI and CRP levels both at baseline (Spearman's rho 0,218, p= 0,0287) and after 12 months (rho 0,284, p= 0,0043).
Prospective studies will determine the utility of these cut off values in the management of LOR to infliximab in IBD.