Dear Editor, We read with great interest the manuscript by Ungar and colleagues describing the temporal evolution of antibodies to infliximab (ATI) in patients with IBD treated with infliximab (IFX). 1 By prospectively following 125 patients with IBD, they showed that ATI formation is a dynamic process. Clinically relevant ATI were typically formed within the first 12 months but transient ATI, which are of little clinical significance, can be formed at any time during treatment. They furthermore demonstrated that the evolution of ATI correlates with clinical loss of response and that concomitant immunomodulator use prolonged ATI-free survival, which is in line with previous reports. Nevertheless, patient-related factors possibly influencing ATI formation weren’t considered and haven’t been studied extensively. We hypothesised that ATI formation may be triggered by HLA-DRB1 alleles, as was shown for immunogenicity to interferon-β therapy in multiple sclerosis. 2 We retrospectively analysed 192 patients with IBD: 76 patients developed ATI during IFX maintenance treatment (= ATI+)(44 Crohn’s disease (CD), 32 UC) and these were matched with 116 patients (64 CD, 52 UC) who never developed ATI (= ATI−). All patients were antitumour necrosis factor naïve before IFX initiation, and the ATI group required at least 2 years of IFX maintenance therapy. We measured drug concentrations and ATI at trough (in-house-developed ELISA, KU Leuven) and HLA-DRB1 alleles were typed using sequence-specific primer PCR (Prometheus Laboratories). Similar to the findings of Ungar et al, the majority of ATI (73.6%) developed during the first year of treatment. At IFX start, the use of a loading dose at weeks 0–2–6 and a higher albumin level were protective for ATI formation (p< 0.05). Carriage of the HLA-DRB1* 03 allele was more prevalent in ATI+(13%) than in ATI− patients (4%; p= 0.002, OR (95% CI) 3.6 (1.5 to 8.6)) and the HLA-DRB1* 13 was less prevalent in ATI+(8%) than in ATI− patients (16%; p= 0.02, OR 0.44 (0.22 to 0.91))(figure 1). After correction for multiple testing, DRB1* 03 remained significantly associated with ATI formation (pcorr= 0.02). Multiple stepwise logistic regression withheld the presence of DRB1* 03, absence of a loading dose and IFX monotherapy as independent predictors of ATI formation with OR 6.7 (2.3 to 19.5, p= 0.0005), 2.9 (1.4 to 6.3, p= 0.004) and 2.0 (0.98 to 4.1, p= 0.056), respectively.

HLA-DR is the protein of which the β1 subunit is encoded by HLA-DRB1. Different HLA-DRB1 alleles encode for different amino acid sequences in this β1-chain. We therefore also examined single amino acids in the DR β1-chain. These amino acid sequences were available through the IIBDGC HLA fine-mapping project 3 in 71%(136/192) of our patients. The presence of arginine at position 74 and absence of glutamate at position 71 were associated with ATI formation (OR