Search in the Abstract Database

Search Abstracts 2016

* = Presenting author

P486 Intrapatient variability in adalimumab drug levels within and across cycles in Crohn’s disease

M. Ward*1, 2, L. Beswick2, P. Thwaites1, J. Hogg1, G. Rosella3, J. Reynolds4, D. Van Langenberg2, P. Gibson1, 3, M. Sparrow1, 3

1Alfred Hospital, Gastroenterology, Melbourne, Australia, 2Eastern Health, Gastroenterology, Melbourne, Australia, 3Monash University, Faculty of Medicine, Nursing and Health Sciences, Melbourne, Australia, 4Monash University, Biostatistics, Faculty of Medicine, Nursing and Health Sciences, Melbourne, Australia

Background

Therapeutic drug monitoring (TDM) of infliximab (IFX) and adalimumab (ADA) is traditionally performed at trough. This is easy with IFX but impractical with ADA. We hypothesised that narrow peak/trough profile of ADA pharmacokinetics may correlate with little intrapatient drug level (DL) variation, permitting TDM to be performed at any time. Via intensive pharmacokinetic sampling, we aimed to build a model that could predict therapeutic trough DL from non-trough DL.

Methods

Prospective observational study of outpatient adult Crohn’s patients on ADA 40mg fortnightly at 2 centres. Study evaluations were at day 4–6 (visit 1), 7–9 (visit 2), and 13–14 (trough) across 2 consecutive ADA cycles where TDM (Matriks ELISA), CRP and Harvey–Bradshaw Index (HBI) were measured. Faecal calprotectin (FCP) was assessed once/cycle. HBI ≥ 5, CRP ≥ 3ng/mL, and FCP ≥ 150μg/g were defined as active disease and trough DL<4.9μg/mL as sub-therapeutic. We performed linear mixed model analyses for inter/intrapatient variation within/between cycles, regression analysis for modelling non-trough DL, covariates, and trough DL and logistic regression analysis to assess the association of DL at visit 1 or 2 with achieving therapeutic trough DL.

Results

In total, 20 patients underwent 117 evaluations (1 withdrew 2nd cycle). Active disease was seen in 14/38 (37%) of cycles according to FCp. 22/39 (56%) trough DL were sub-therapeutic. Mean patient DL are shown.

Table 1 Drug levels by visit and cycle

VisitCycle 1 DL (μg/mL), 
mean (SEM)Cycle 2 DL (μg/mL), 
mean (SEM)p- valueVisit main 
effect mean (SEM)Pairwise 
p- value
1 (D46)5.54 (0.86)5.56 (0.68)0.615.55 (0.54)V1 vs V2: 0.057
2 (D7–9)5.18 (0.58)4.88 (0.60)0.535.04 (0.41)V2 vs V3: <0.001
3 (trough)4.15 (0.50)4.05 (0.52)0.444.10 (0.36)V3 vs V1: <0.001

There were no significant differences in DL of paired visits across cycles (p = 0.61). DL did not significantly decline from visit 1 to 2 (p = 0.06), but declined between each of visits 1, 2 and trough (p < 0.001). DL showed good correlation between visits within cycles (1 to 2 r = 0.82, 1 to trough r = 0.72, 2 to trough r = 0.87). Trough DL showed a significant increase with DL at visit 1 (β = 0.4, p < 0.001) decline with smoking (-1.3, p = 0.02) and an increase with syringe (vs pen) device (2.1, p = 0.002) with R2 = 63%, and, in a similar model, a significant increase with DL at visit 2 (β = 0.7, p < 0.001) with R2 = 81%. Active disease (CRP, FCP or HBI) was not significantly associated with trough DL. DL of 7.9 (95% CI 6.5–15.8) at visit 1 and 7.2 (6.2–11.1) at visit 2 had an 80% chance of predicting therapeutic trough DL.

Conclusion

DLs vary little between cycles, suggesting the result of a single DL can be interpreted with confidence. DLs decline between day 4–6 and 7–9, and significantly between these early visits and trough. Larger population-based pharmacokinetic studies are needed before TDM with ADA can be accurately interpreted at any time point with confidence.