New steps in infliximab therapeutic drug monitoring in patients with inflammatory bowel diseases

Abstract

Aims

Therapeutic drug monitoring (TDM) of infliximab (IFX) appears to be beneficial for patients with inflammatory bowel disease (IBD). However, the recommended target concentrations depend partly on the method used to quantify IFX. Since we recently developed a liquid chromatography–tandem mass spectrometry method to quantify IFX, we aimed to determine IFX trough concentrations (Cmin) associated with biological remission.

Methods

We retrospectively measured IFX Cmin in sera from 55 patients with IBD, on IFX maintenance therapy, and for whom demographic, biological and clinical data were collected from medical records. A threshold of IFX Cmin associated with biological remission (defined by C‐reactive protein < 5 mg l^–1 and faecal calprotectin <150 μg g^–1) was determined using receiver operating characteristics analysis.

Results

IFX Cmin ranged from <1 mg l^–1 to 57.2 mg l^–1. IFX Cmin were higher (P = 0.038) in patients with biological remission and a cut‐off of IFX Cmin set to 6.2 mg l^–1 was associated with biological remission (sensitivity = 0.75, 95% confidence interval 0.58–0.75; specificity = 0.61, 95% confidence interval 0.39–0.83).

Conclusion

Liquid chromatography–tandem mass spectrometry measurement of IFX Cmin and the determination of a new threshold of IFX Cmin associated with biological remission are new steps towards IFX treatment personalization in patients with IBD.

What is Already Known about this Subject

• The recommended target residual concentrations of infliximab to reach biological remission in patients with inflammatory bowel disease depend partly on the method used to quantify infliximab, patient demographic factors and pharmacokinetics data.

• Liquid chromatography–tandem mass spectrometry emerges as new analytical tool for the personalization of biotherapy treatment.

What this Study Adds

• In patients with inflammatory bowel disease, this study is the first to determine that infliximab residual concentration cut‐off value set to 6.2 mg l^–1 and measured by liquid chromatography–tandem mass spectrometry was able to correctly discriminate patients with biological remission from patients without biological remission.

Introduction

Infliximab (IFX) 1 is a chimeric monoclonal antibody (Mab) targeting tumour necrosis factor (TNF‐α) approved for the treatment of inflammatory bowel disease (IBD) such as Crohn's disease (CD) or ulcerative colitis (UC). Despite improvement in clinical outcome, loss of response to anti‐TNF‐α Mabs remains a major concern in the management of IBD patients. Increasing evidences suggested that therapeutic drug monitoring (TDM) of anti‐TNF‐α Mabs could help to improve remission rate or mucosal healing achievement 2 but the clinical interest of TDM of anti‐TNFα Mabs in IBD patients remains to be further demonstrated 3. One of the limitations lies in the absence of standardization of the analytical methods used to measure IFX plasma concentration. Indeed, up to now, all previous studies used enzyme‐linked immunosorbent assay (ELISA) or homogeneous mobility shift assay, but the systematic bias 4, 5, 6 and lack of specificity 5, 6 of some ELISA methods could be of major concerns, and lead to the determination of different clinically relevant IFX thresholds 2, 7, 8 or different therapeutic strategies 9.

We 10, 11 and other independent teams 12, 13 recently developed and validated liquid chromatography/tandem mass spectrometry (LC–MS/MS) methods for the quantification of IFX in human serum. These methods are highly specific and reproducible, but require the determination of new clinically relevant IFX threshold, as recommended whenever a new analytical method is used for the measure of a biotherapy concentration 14.

In this study, we aimed to describe, in a population of IBD patients, the variability of IFX Cmin measured by LC–MS/MS. We studied the influence of demographic factors and biological remission on this variability. Then, we proposed a cut‐off of IFX Cmin that is associated with biological remission.

Methods

Study design and patient population

This was a monocentric retrospective study. Blood samples (one sample per patient) from 59 IBD patients followed from May 2014 to March 2017 for routine care in Grenoble University Hospital division of Hepatology and Gastroenterology were collected just before IFX infusion at day care. Residual serum samples were stored at −20°C in a biological sample collection (DRC‐2013‐1983). The study and the biobank were approved by the IRB 6705 (CPP Sud Est 5, Grenoble, France).

Demographic, clinical and biological data were collected from electronic medical records. Demographic data were: age, sex, weight, height and body mass index (BMI). Clinical data were: disease, date of diagnostic, IBD‐related surgery, date of treatment initiation, co‐medication, last administered doses of IFX and time since last dose. Biological data were: IFX Cmin (measured by LC–MS/MS), plasma C‐reactive protein (CRP) concentration, faecal calprotectin levels and presence of anti‐drug antibodies (ADA).

Inclusion criteria were: patients treated for IBD, on IFX maintenance therapy (treatment for at least 56 days (8 weeks) since treatment initiation, but irrespective of the duration of IFX use), available serum IFX Cmin concomitant to CRP plasma or faecal calprotectin concentrations. Fifty‐five patients filled inclusion criteria (see flow chart, Figure 1). IFX Cmin was measured by LC–MS/MS as previously described 11. The lower limit of quantification was 1 mg l^–1 (bias, −4%; CV = 12%) and upper limit of quantification was 26 mg l^–1 (bias, 3%; CV = 5%) 11.

Figure 1.

Flow chart. IBD, inflammatory bowel disease, CRP: C reactive protein

Outcomes

To determine the association between IFX Cmin and biological remission, patients were stratified as following: patients with biological remission (plasma CRP < 5 mg l^–1 and faecal calprotectin <150 μg g^–1 stools 15) and patients without biological remission (plasma CRP ≥ 5 mg l^–1 and/or faecal calprotectin ≥150 μg g^–1 stools; see flow chart, figure 1).

Statistical analysis

Continuous data were expressed as the median (25^th–75^th percentiles) and categorical variables were expressed as percentages. Non‐parametric tests were used to compare quantitative variables. Linear regressions were performed to study the association between IFX Cmin and individual factors of variability. These statistical analyses were performed using Statview (5.0; SAS Institute Inc 1992–1998 Cary, NC, USA). Receiver operating characteristics (ROC) curves were constructed using the R package pROC 16. The Youden index (sensitivity + specificity – 1) was computed to define the best threshold. The specificity was defined as the ratio of patients not on biological remission with an IFX Cmin lower than the threshold to the total number of patients not on biological remission. The sensitivity was the ratio of patients with biological remission and with an IFX Cmin higher than the threshold to the total number of patients with biological remission). The area under ROC curve is given with its confidence interval according to the DeLong method, the sensitivity and specificity of each threshold with their 95% confidence intervals. The associated likelihood ratios (LR) and their 95% confidence intervals are also specified. A P‐value ≤0.05 was considered statistically significant.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 17.

Results

Patients and data description

Clinical and biological characteristics of the study population are presented in Table 1.

Table 1.

Characteristics of the study population. Continuous data are presented as median (interquartile range). Qualitative data are presented as number (percentages)

Sex
Male	30 (54.5)
Female	25 (45.5)
Age (years)	35 (26.7–46.0)
Body weight (kg)	67 (52.5–80.7)
Body mass index (kg m–2)	22.5 (19.3–25.1)
Disease
Crohn's disease	34 (61.8)
Ulcerative colitis	21 (38.2)
Extent of mucosal damage
Confined to 1 site	20 (36.4)
Confined to 2 sites	14 (25.5)
3 or more sites; pancolitis	19 (34.5)
Treatment
Infliximab last dose (mg)	460 (312–608)
Time since last dose (weeks)	8 (6–8)
Duration of infliximab therapy (months)	32.1 (5.1–59.1)
Comedication
Immunosuppressantsa	20 (36.4)
Corticosteroids	33 (60.0)
mesalazin	1 (1.8)
Previous IBD related surgery
Yes	17 (30.9)
No	36 (65.5)
Biology
CRP < 5 mg l–1	40 (72.2)
Faecal calprotectin <150 μg g–1 stools	29 (74.5)
Presence of antidrug antibodies
Yes	8 (14.5)
No	47 (85.5)

CRP data were available for all patients (n = 55), faecal calprotectin data were available for 39 patients. IBD, inflammatory bowel disease; CRP, C‐reactive protein

^aImmunosuppressive drugs were azathioprine, 6‐mercaptopurine, methotrexate

IFX Cmin variability

IFX Cmin ranged from <1 mg l^–1 to 57.2 mg l^–1 with a median [25^th–75^th percentiles] of 6.7 mg l^–1 [4.5; 11.7]. In univariate analyses, IFX Cmin were not associated with age, weight, BMI, sex, IFX last dose, plasma CRP or faecal calprotectin levels, or with the presence of ADA (Table 2).

Table 2.

Univariate linear regression between infliximab trough concentrations and demographic, clinical and biological parameters

Variables	Estimate [95% CI]	P‐value
Demographic
Sex	1.028 [0.976; 1.082]	0.3044
Age	−0.148 [−0.518; 0.222]	0.4270
Weight	0.133 [−0.229; 0.495]	0.4658
BMI	0.0198 [−0.078; 0.117]	0.6860
Biological
Anti‐drug antibodies	1.034 [0.979; 1.093]	0.2330
Time since treatment initiation	6.33 [−21.5; 34.2]	0.6506
Last infliximab dose	3.81 [−0.696; 8.32]	0.0958
CRP	−0.25 [−0.814; 0.313]	0.3766
Faecal calprotectin	−6.05 [−18.8; 6.74]	0.3439

CI, confidence interval; BMI, body mass index; CRP, C reactive protein

Biological remission

Overall, 25 patients (13 CD and 12 UC) were considered to show biological remission. IFX Cmin were higher (P = 0.038) in patients with biological remission (IFX median = 8.8 mg l^–1 [6.2; 14.7]) compared to patients without biological remission (IFX median = 5.4 mg l^–1 [2.7; 8.3]; see Figure 2a). The ROC–area under the curve (AUC) analysis showed a sufficient performance of IFX Cmin to discriminate patients with or without biological remission [AUC = 0.674; 95% confidence interval (95% CI): 0.510–0.838; Figure 3]. The Youden index showed that the threshold of 6.2 mg l^–1 provided the best balance of sensitivity and specificity with a higher sensitivity = 0.75 (95% CI: 0.58–0.75) and specificity = 0.61 (95% CI: 0.39–0.83). The LR for a positive result was 1.92 (95% CI: 1.46–2.52) and the LR for a negative result was 0.41 (95% CI: 0.28–0.60). The estimated positive predictive value was 0.67 (95% CI: 0.48–0.84) and the estimated negative predictive value was 0.70 (95% CI: 0.49–0.90).

Figure 2.

Infliximab trough concentrations according to biological remission in all inflammatory bowel disease (IBD) patients (A) or in the subgroups of patients with Crohn's disease (B) or ulcerative colitis (C)

Figure 3.

Receiver operating characteristics curve for infliximab trough concentration and biological remission association in all inflammatory bowel disease (IBD) patients (A) and in the subgroup of patients with Crohn's disease (B)

When patients were stratified in terms of type of IBD, the association between IFX Cmin and biological remission remained significant for patients with CD (P = 0.0006, see Figure 2b), but not for UC (P = 0.35, see Figure 2c). Among patients with CD, IFX Cmin was higher in patients with biological remission (IFX median = 8.8 mg l^–1 [6.77; 12.6], n = 13) compared to patients without biological remission (IFX median = 4.9 mg l^–1 [1.8; 6.0], n = 15). In patients with CD, the optimal IFX Cmin threshold was 6.8 mg l^–1 (sensitivity = 0.83 [0.58–0.83], specificity = 0.81 [0.81–1]) and the ROC‐AUC analysis showed very good performance of IFX Cmin to discriminate patients with and without biological remission (ROC‐AUC = 0.87 [0.70–1.00]; Figure 3b), with an LR for a positive result of 4.36 (95% CI: 2.89–6.61) and a LR for a negative result of 0.21 (95% CI: 0.13–0.33).

Discussion

This is the first study to determine a cut‐off of IFX Cmin determined by LC–MS/MS and associated with biological remission. In agreement with previous studies 18, 19, the range of IFX Cmin was very wide (from <1 mg l^–1 up to 57 mg l^–1) highlighting the inter‐patient variability of IFX pharmacokinetic. However, age, sex, weight, IFX last dose and presence of ADA did not influence IFX Cmin (present study). These findings contrast with data from population pharmacokinetic studies suggesting that IFX clearance could be significantly influenced by high body weight or presence of ADA 20. However, in our population, no patient was obese and only eight (14.5%) presented with ADA, which could explain these discrepancies. Further studies on larger cohorts are therefore warranted to identify the determinants of the high IFX Cmin variability.

Our study demonstrated that at a random time point during IFX maintenance therapy, IFX Cmin was associated with biological remission especially in patients with CD. The ROC analysis allowed us to propose a threshold Cmin of 6.2 mg l^–1. In agreement with our finding, IFX Cmin (measured by ELISA) up to 6.8 mg l^–1 was 85% specific for CRP normalization 21 and IFX Cmin at week 14 (measured by a homogeneous mobility shift assay) below to 6.2 mg l^–1 was 82.4% specific for loss of response assessed by clinical scores at week 48 22 in IBD patients. Moreover, the American Gastroenterological Association reported that IFX Cmin at or above 5 mg l^–1 and 7 mg l^–1 were respectively associated with 7.9% and 3.9% of patients not on remission 23 and therefore recommended to reach IFX Cmin >5 mg l^–1. A recent study reported that CD patients with IFX Cmin ranging from 10.1 to 20.2 mg l^–1 had almost 3‐fold chance of achieving fistula healing when compared to patients with IFX Cmin <2.8 mg l^–1 19, and that IFX Cmin >10.1 mg l^–1 was the single independent predictor of mucosal healing 19. Conversely, several studies suggested the absence of clinical benefit when IFX Cmin >7 23 or >10 mg l^–1, 21 suggesting that the interest in reducing drug dosage or in switching to another therapeutic class when a lack of response to the Mab is acknowledged. All these data suggest that IFX Cmin target thresholds 21, 24 depend on the outcomes that have been chosen for the definition of loss of remission (clinical score 24, presence of ADA or combined clinical and biological remission 23) and the analytical methods used to measure IFX concentrations. Our ROC‐AUC analysis showed a very good performance of IFX Cmin to discriminate CD patients with and without biological remission, higher than those obtained with ELISA‐based measures previously reported by Yarur et al. 19 (AUC = 0.82), Ungar et al. 21 (AUC = 0.75) or Ward et al. 24 (AUC = 0.77). Moreover, the LR for positive (4.36) and negative (0.21) results were also very good in CD patients, highlighting the contribution of our work.

Conversely, our data showed no difference between IFX Cmin in UC patients with or without biological remission. However, we acknowledge that our study lacks statistical power with respect to the limited sample size of UC patients with biological remission and the high variability of IFX Cmin, and the need of further study on larger cohort of UC patients.

We acknowledge some limitations in our work. First our data were retrospectively collected and obtained on a limited number of patients. Second, clinical data allowing the determination of clinical score of disease activity (Mayo score or Crohn's disease activity index) were not available within the electronic medical records, limiting the study of the association of IFX Cmin with biological remission.

Further prospective studies on larger cohorts are warranted to strength our findings and provide distinct IFX Cmin threshold for patients with CD and UC.

Conclusion

We have demonstrated that the IFX Cmin cut‐off value set to 6.2 mg l^–1 and measured by LC/MS–MS was able to discriminate patients with IBD with biological remission from those without. Such a combination of the exact measure of IFX concentration by LC/MS–MS and the determination of an IFX Cmin cut‐off associated with biological response is a new step towards personalized IFX therapy. To reach this target concentration for each patient, further population pharmacokinetic studies are required to determine individual pharmacokinetic parameters.

Competing Interests

There are no competing interests to declare.

We thank Karine Scalabrino for her excellent relevant technical assistance.

Author contributions

F.S.‐L., study concept and design, drafting of the manuscript. B.N., acquisition of data, measurement of infliximab by LC–MS/MS, drafting of the manuscript. E.G.‐V., critical revision of the manuscript for intellectual content. B.B., critical revision of the manuscript for intellectual content, J.‐F.J., measurement of infliximab concentration by LC–MS/MS. S.B., statistical analyses, critical revision of the manuscript for intellectual content. D.T., critical revision of the manuscript for intellectual content. Guarantor of the article: F.S.‐L.

Nemoz B., Ternant D., Bailly S., Gautier‐Veyret E., Jourdil J.‐F., Bonaz B., and Stanke‐Labesque F. (2019) New steps in infliximab therapeutic drug monitoring in patients with inflammatory bowel diseases, Br J Clin Pharmacol, 85, 722–728. 10.1111/bcp.13845.