Comparison of Techniques for Monitoring Infliximab and Antibodies to Infliximab in Crohnʼs Disease Patients with Infliximab Treatment Failure

Abstract

Purpose: Several analytical techniques are currently used for measuring infliximab (IFX) and antibodies to infliximab (ATI) to follow proposed algorithms based on PK, PD and immunogenicity factors for therapeutic interventions in Crohn's disease patients with loss of response. We aimed to investigate how 3 of these assays compare in a clinical setting. Methods: IFX and ATI were measured in 67 patients with IFX treatment failure (53 luminal, 6 fistulizing, 8 both; median CDAI 284; median PDAI 9), by RIA (Biomonitor A/S) (1), ELISA (2), and homogenous mobility shift assay (HMSA) (3) (both from Prometheus Laboratories Inc.). Results: IFX: IFX was undetectable in 8 (12%) patients in RIA, 17 (25%) in ELISA, and 8 (12%) in HMSA. All patients with undetectable IFX by RIA or HMSA also had undetectable IFX by ELISA. Five patients had undetectable IFX in both RIA and HMSA. All assays showed linear correlations with R2=0.94 and α=0.90 [0.84-0.96] for ELISA and HMSA, R2=0.92 and α=1.44 [1.34-1.55] for RIA and HMSA, and R2=0.87 and α=1.52 [1.38-1.66] for RIA and ELISA (p<0.0001). However, assays disagreed on absolute IFX concentrations with a mean difference of 0.60 μg/ml [0.12-1.09] in ELISA and HMSA, -2.50 μg/ml [-3.22 to -1.78] in RIA and HMSA, and -3.10 μg/ml [-4.00 to -2.21] in RIA and ELISA. When comparing the two assays available in the U.S., the HMSA was marginally more sensitive compared to ELISA in detecting IFX levels. ATI: ATI were detectable in 18 (27%) patients in RIA, 6 (9%) in ELISA, and 23 (35%) in HMSA. All patients with detectable ATI in ELISA also had detectable ATI in RIA or HMSA. Seventeen patients had detectable ATI in both RIA and HMSA. Of the 17 additional samples detected by HMSA (and missed by ELISA), 14 had detectable levels of IFX and were therefore likely missed by ELISA due to interference from serum IFX. All assays generally showed linear correlations with R2=0.68 and α=0.08 [0.07-0.10] for RIA and ELISA, R2=0.65 and α=6.52 [5.30-7.73] for ELISA and HMSA, and R2=0.59 and α=0.65 [0.52-0.79] for RIA and HMSA (p<0.0001). However, the HMSA appears to be superior to the currently available ELISA and may be generally comparable to RIA, also a fluid-phase assay that is currently not available in the US. Conclusion: Commonly used assays for measuring IFX and ATI in patients on IFX agree on trends. All assays showed a good correlation for serum IFX levels, however, larger discrepancies were seen when comparing ATI measurements. Liquid phase assays seem to perform better in general compared to solid phase assays.