P697 Human-leukocyte antigen type is associated with duration on infliximab therapy in patients with antibodies to infliximab
T. Billiet*1, T. Van Stappen2, N. Vande Casteele2, I. Cleynen1, V. Ballet3, K. Claes1, F. Princen4, S. Singh4, A. Gils2, M. Ferrante3, G. Van Assche3, S. Vermeire3
1Department of Clinical and Experimental Medicine, KU Leuven, Translational Research in GastroIntestinal Disorders, Leuven, Belgium, 2KU Leuven, Laboratory for Therapeutic and Diagnostic Antibodies, Leuven, Belgium, 3University Hospitals Leuven, Gastroenterology - Translational Research Center for Gastrointestinal Disorders (TARGID), Leuven, Belgium, 4Prometheus Laboratories, Department of Research and Development, San Diego, United States
The contribution of antibodies to infliximab (ATI) to loss of response (LOR) in patients with IBD is well established. It was also shown that ATI may be transient and that infliximab (IFX) discontinuation is not always necessary in this occasion. The variables that influence discontinuation of IFX in patients who have developed ATI are unknown. The HLA system, responsible for processing antigens, might play an important role.
We hypothesized that HLA class II alleles influence the duration on IFX therapy in patients who develop ATI. We retrospectively identified 74 IBD (42 CD, 32 UC) patients who developed ATI during maintenance IFX. Of these, 61 (82%) discontinued IFX therapy because of LOR or side effects and 13 (18%) were still receiving IFX at the end of follow-up. All patients were anti-TNF naïve before start of IFX. A total of 1889 serial serum samples were measured for ATI with an improved bridging ELISA using monoclonal antibody MA-IFX10F9 as calibrator. For each patient, the highest ATI concentration measured, was used to create ATI quartiles. HLA-DRB1 was genotyped with sequence specific primers (Prometheus Laboratories Inc.). Patient and therapy variables (e.g. presence of IFX dose optimization, immunomodulator rescue), ATI quartiles and DRB1 alleles were included as possible confounders influencing total time on IFX using Cox proportional hazard regression.
The median time on IFX was 100 weeks (IQR 52 - 217) and did not differ significantly depending on ATI quartile (P=0.19, Kruskal-Wallis test). However, patients from quartile 4 showed a significant shorter time on IFX (median 72 weeks) compared to the other quartiles combined (median 111 weeks, P=0.049). We observed clear differences in time on IFX depending on the DRB1 allele with medians ranging from 43 weeks for DRB1*13 to 169 weeks for DRB1*15 (P=0.019, Log-rank test (Figure 1)).
“Figure 1: Survival analysis for the different DRB1 alleles (only those with a frequency > 5%, heterozygous allele carriers were counted twice, homozygous only once) and total time on IFX”
Cox proportional hazard regression identified albumin at start of IFX, ATI in quartile 4, presence of DRB1*11 and presence of DRB1*15 as independent predictors (P<0.05) of total time on IFX with hazard ratios (95% CI) of 0.34 (0.19–0.59), 3.3 (1.6–6.9), 0.45 (0.23–0.9) and 0.35 (0.16–0.77) respectively.
In patients who develop ATI, besides the magnitude of the titer, a higher concentration of albumin at start and the HLA-DRB1 genotype prolong the time patients will remain on IFX. The concomitant use of immunomodulators did not affect time on IFX in this study. These results clearly warrant further investigation in prospectively designed studies.