Pharmacokinetic studies on infliximab, an anti-TNF-α monoclonal antibody, in Crohn’s disease reported different pharmacokinetic (PK) parameters [1–4]. These discrepancies could be explained by (i) the inclusion of patients at distance of treatment initiation, (ii) the large intra-individual variability observed during long (>6 months) patient follow-up, or/and (iii) the inclusion of patients both with and without antibodies toward infliximab (ATI). We report here an analysis of infliximab pharmacokinetics in inflammatory bowel disease (IBD) patients that took into account these drawbacks.

This is an update of our previous study [3], and we retrospectively analyzed 133 patients treated by infliximab for IBD between 2006 and 2012 in Tours university hospital, in whom trough and peak infliximab concentrations were available during treatment initiation and ATI not detected during the first 6 months. These patients received 5 mg/kg infliximab at weeks 0, 2, 8, and 14. Median [range] body weight was 60 kg [41–120], and 79 (59 %) were women. Infliximab concentrations were measured using a validated ELISA technique [5].

Infliximab pharmacokinetics was described using a population approach (MONOLIX 4.3.2, Lixoft, Saclay, France). A two-compartment model was used. Central (V 1) and peripheral (V 2) volumes of distribution, and systemic (CL) and intercompartment clearances (Q) were estimated. Interindividual and residual models used were, respectively, exponential and mixed additive-proportional. Body weight (coded as a median-centered power function, β weight being the power parameter for body weight) and gender were tested as covariates for V 1 and CL.

All parameters were estimated with satisfactory accuracy, and no obvious model misspecification was observed. Typical parameters (relative standard error) were V 1 = 2.6 L (4 %), CL = 0.014 L/h (6 %), V 2 = 4.5 L (1 %), and Q = 0.083 L/h (3 %). Interindividual standard deviations for V 1 and CL (relative standard error) were ω V1 = 27 % (8 %) and ω CL = 47 % (7 %), respectively. Additive and proportional (relative standard error) standard deviations were σ add = 2.3 mg/L (16 %) and σ prop = 0.21 (6 %), respectively. Central volume of distribution (V 1) increased with body weight (β weight = 0.22, p = 0.00013). Both V 1 and CL were higher in men than in women: typical V 1 was 3.2 L in men and 2.6 L in women, and typical CL was 0.019 L/h in men and 0.014 L/h in women. Distribution and elimination half-lives were T½-α = 0.5 days and T½-β = 16.1 days, respectively.

Compared to other PK studies of infliximab in IBD patients (Table 1), our T½-α estimate was similar to that reported by Fasanmade et al. (0.2 days [2]) but lower than that reported by others (approximately 3 days [1, 3, 6]). Of note, our T½-α estimate was similar to values reported in rheumatoid arthritis (0.3 days [7]) and ankylosing spondylitis (0.3 days [8, 9]).

Table 1 Summary of pharmacokinetic studies of infliximab in inflammatory bowel disease (IBD)
Full size table

Among PK studies of infliximab in IBD, this study is the first to analyze patients from their first infusion of infliximab, with a short follow-up of less than 6 months and after exclusion of ATI+ patients. Our study has nevertheless limitations. First, our data were scarce, consisting in trough and peak concentrations. Second, the ATI assay used for this cohort was unable to detect ATI in the presence of concentrations of infliximab >2 mg/L, which probably led to an underestimation of the proportion of immunized patients.

Overall, our study is the first to describe infliximab pharmacokinetics at the time of treatment initiation, during the first 6 months of treatment, and in patient in whom ATI were not detected. It confirms the short distribution T½ of infliximab, in IBD patients as in other indications.