We read with great interest the recent article by Ben-Horin et al which concludes that the measurement of trough levels of infliximab (TLI) combined with the measurement of antibodies to infliximab (ATI) is highly correlated with the clinical response to the therapeutic.1

Infliximab (Remicade) and adalimumab (Humira), both monoclonal antibodies towards TNF-α, are widely used in the treatment of Crohn's disease and ulcerative colitis. Loss of efficacy, infusion and injection-site reactions and hypersensitivity reactions have been described due to the development of antibodies towards these agents.2 We agree with Ben-Horin et al that the interpretation of ATI with TLI is important to document the clinical response of a patient to infliximab. However, we urge that these measurements are performed early in the treatment. Previously it was stated that the TLI measured early after the first infusion of infliximab (week 4) is a good prognostic parameter for development of immunogenicity.3 Nevertheless, some patients will have low TLI at an early stage of the treatment due to a high TNF-α load, immunogenicity or a combination of both. Each situation demands an appropriate clinical approach. Therefore, we advise the measurement of TLI serially and in combination with ATI when TLIs are absent or low instead of at one single time point. Ben-Horin et al measured ATI levels only in the last sample available for a patient. We believe that ATI levels should be evaluated on a kinetic basis since the immunogenic response to a therapeutic can be a latent, transient or sustained effect.

We describe here the case of a Crohn's disease patient treated at our clinic where the retrospective analysis of serum samples for C reactive protein (CRP), antibody and trough levels put clinical management and disease outcome into a different perspective (table 1). The patient started infliximab 5 mg/kg at week 0 and responded clinically well till w27, although TLIs were detected only once at w5 (TLI=0.33±0.01 μg/ml) and CRP levels never completely normalised. At w27 due to ‘clinical loss of response’, the dosage interval was shortened and ATI levels rose during this period to 302±35 μg/ml at w42. Due to a clinically ineffective dose adjustment, at w42 a double dose of infliximab (10 mg/kg) was administered and the patient suffered an acute infusion reaction. On interpreting these data retrospectively, it is clear that all interventions (ie, shortening of the dosage interval and administration of a double dose) caused the immunogenic response to escalate in this patient.

At w44 the patient was switched to adalimumab 160/80 followed by 40 mg subcutaneously every other week; he showed good clinical response and CRP levels dropped. However 1 year after the start of adalimumab (w104), serum adalimumab levels (AL) were low (AL=0.8±0.05 μg/ml) and at w116 antibodies to adalimumab (ATA) were detected (ATA=0.045±0.01 μg/ml). At first, it seemed to be a transient immunogenic response with no clinical repercussions, but at w279 signs of loss of response appeared concomitantly with increased CRP levels (44 mg/l), indicating loss of response to adalimumab. For infliximab and adalimumab it can be argued that the lack of blood levels contributed to the formation of antibodies to the drug.

This case shows that adjusting drug administration only based on clinical evolution is often inaccurate. Therefore, we recommend the measurement of trough levels early in the treatment. When trough levels are low or absent, we advise the determination of antidrug antibodies and doing this on a serial basis. We also demonstrate here that ATIs can still be detected in patient serum up to 4.5 years after the last infusion of infliximab. Moreover, we hypothesise that the development of a sustained immunogenic response against one anti-TNF agent is a predictive factor for future loss of response towards other anti-TNFs over time, but this remains to be proven.