Early vedolizumab trough levels are not associated with short-term response in patients with inflammatory bowel disease

Summary

Background:

Data about the usefulness of monitoring vedolizumab therapy are sparse and conflicting. Here, the aim was to assess the association between early vedolizumab trough levels (VTLs) and responses to induction therapy in patients with inflammatory bowel disease (IBD).

Methods:

The study population comprised consecutive IBD patients from a prospective cohort of vedolizumab treated individuals at our centre, in whom VTLs and anti-vedolizumab antibodies (AVAs) were measured during the induction phase of therapy. Included patients received vedolizumab (300 mg) at weeks 0, 2 and 6, with an extra dose at week 10 in cases of inadequate response after the third infusion. Clinical response was evaluated by a physician at 1 month after the last induction dose (week 10 or 14). Measurement of VTL and AVA was performed by ELISA.

Results:

Eighty-seven patients, 31 with Crohn’s disease and 56 with ulcerative colitis, were included. Only 15% of patients were naïve to anti-tumour necrosis factor alpha therapy; 61% used concomitant systemic steroids and 26% used thiopurines. An additional dose at week 10 was given to 39% of individuals. Clinical response to induction was reported in 77% of IBD patients. The median VTL at week 6 was 30.6 µg/mL (range: 1.1–80.0). When comparing patients with and without a clinical response to vedolizumab, we found no significant difference in the median VTL at week 6 (29.4 vs. 34.4 µg/mL, respectively; p = 0.71). Likewise, VTL did not differ significantly between individuals receiving an additional dose at week 10 and those receiving standard induction (40.0 vs. 28.5, respectively; p = 0.69). Seven per cent of patients developed positive AVA up until weeks 10–14. Diagnosis type, concomitant immunosuppressants or previous biologic therapy had no impact on VTL.

Conclusion:

There was no association between early VTL and clinical response to induction therapy. Further studies should address the clinical relevance of therapeutic drug monitoring during long-term vedolizumab treatment.

Key words:

inflammatory bowel disease – vedolizumab – trough levels – clinical response

Introduction

Inflam­matory bowel dis­ease (IBD) comprises ulcerative colitis (UC) and Crohn‘s dis­ease (CD), which are characterised by idiopathic relapse-remitt­­ing inflam­mation of the gastrointestinal tract. These chronic disorders can result in ir­reversible mucosal damage, disability, and increased incidence of colitis-as­sociated neoplasia [1]. The mainstay of treatment for IBD is induc­­ing and maintain­­ing dis­ease remis­sion. New therapeutic goals also involve achiev­­ing mucosal healing, which has been shown to lead to improved clinical outcomes, includ­­ing lower rates of surgical intervention [2].

Traditional­ly, the drugs used in the treatment of IBD are mesalazine derivatives, glucocorticoids and im­munomodulators. With the advances in medical science and technology, a new group of drugs emerged, cal­led the bio­logics. The most widely used bio­logics are the tumour necrosis factor alpha (TNF-α) inhibitors such as infliximab and adalimumab, which are highly ef­fective in the treatment of both UC and CD. The newer bio­logic agents in IBD include the integrin receptor antagonist vedolizumab and IL-12 and IL-23 antagonist, ustekinumab [3].

Despite these advances, a significant proportion of patients either fail to initial­ly respond to bio­logics, or lose response with time. Therefore, there is an unmet need to optimise bio­logics in order to improve patient outcomes [4]. The role of therapeutic drug monitor­­ing (TDM) –  measur­­ing drug through levels and specific antidrug antibodies –  in guid­­ing therapeutic decisions in IBD patients treated with TNF-α inhibitors is now widely accepted since it has shown to deliver more clinical­ly ef­fective dosing [5,6]. TDM has now become the standard of care when prescrib­­ing anti-TNF-α ther­apy in patients with IBD, in particular for manag­­ing primary or secondary loss of response [4].

Vedolizumab is a humanised im­munoglobulin G1 monoclonal antibody that binds exclusively to the lymphocyte integrin α4 β7 and is indicated for the treatment of adult patients with moderately to severely active UC or CD. By bind­­ing specifical­ly to α4 β7, vedolizumab inhibits the interaction of α4 β7 –  expres­s­­ing cel­ls, in particular memory T lymphocytes, with mucosal addres­sin cell adhesion molecule-1 (MAdCAM-1) on endothelial cel­ls, thereby block­­ing the infiltration of these cel­ls into the gastrointestinal mucosa and gut-as­sociated lymphoid tis­sue and suppres­s­­ing gut inflam­mation [7– 9].

Published data regard­­ing the role of TDM for IBD patients treated with vedolizumab are limited and it is still unknown whether TDM of vedolizumab could improve patient management [10]. The aim of our study was thus to as­sess the as­sociation between early vedolizumab through levels (VTL) and response to induction ther­apy in patients with IBD.

Methods

Patient population

Study population comprised consecutive IBD patients from a prospective cohort of vedolizumab-treated patients at our centre, who initiated vedolizumab ther­apy between 2/ 2017 and 8/ 2018. Eligible patients for the study have had VTL (vedolizumab trough levels) and anti-vedolizumab antibodies (AVA) measured dur­­ing induction phase of ther­apy. All patients received ve­dolizumab 300 mg infusion at weeks 0, 2 and 6. In case of inadequate response after the 3^rd infusion, an additional dose at week 10 was administered upon discretion of the treat­­ing physician.

Data col­lection

The data on patients’ demographics, dis­ease characteristics and details on vedolizumab ther­apy were obtained primarily from the CREdIT registry which is a national database of IBD patients on bio­logic treatment. Data on bio­logic ther­apy, includ­­ing dosage, treatment regime, adverse events and dis­ease activity, are entered prospectively into the registry. Dis­ease activity is determined by Harvey-Bradshaw index for CD and partial Mayo score for UC based on patients’ question­naires that are fil­led out before each administration of bio­logic drug. Laboratory parameters such as C-reactive protein (CRP) and faecal calprotectin (FC) as well as mis­s­­ing information on clinical characteristics in the registry were obtained from patients’ medical records.

Clinical scores

Response to vedolizumab ther­apy was as­ses­sed one month after the last induction infusion (week 10 or week 14 in those receiv­­ing additional dose at week 10) and determined by physician global as­ses­sment (PGA). PGA was evaluated retrospectively us­­ing data from registry and medical records and defined as fol­lows:

• a) complete response –  complete improvement of clinical symp­toms and laboratory markers (CRP < 5 mg/ L; FC < 150 µg/ g);
• b) partial response –  improvement of clinical symp­toms and/ or laboratory markers.

Patients who did not experience any improvement by the end of the induction phase were clas­sified as primary non-responders.

Vedolizumab through levels and anti-vedolizumab antibodies measurement

At our centre, VTL and AVA are prospectively measured prior to each infusion start­­ing at week 2 of induction phase of the treatment. VTL were quantified us­­ing an enzyme-linked im­munosorbent as­say (ELISA) fol­low­­ing the manufacturer’s protocol (Im­munoGuide, Tani Medical). This ELISA is based on a vedolizumab-specific monoclonal antibody (catcher Ab, Im­munoGuide clone19F3). The level of detection of the as­say was 1.9 µg/ mL and the measurement range of VTL was 0– 600 µg/ mL. AVA mea­surement was performed by a bridg­­ing type ELISA (Im­munoGuide, Tani Medical) for a semi-quantitative determination of free antibodies to vedolizumab in serum and plasma. Cut-off value for the AVA levels was set at 3AU (arbitrary units)/ mL, samples with an equal and higher AVA than the cut-off value were considered to be positive.

Statistical analysis

Standard descriptive statistical analyses were performed, includ­­ing frequency distributions for categorical data and calculation of median and range for continuous variables. Dif­ference between outcome variables (VTLs, CRP and FC) was analyzed us­­ing a Man­n-Whitney test, not as­sum­­ing normal distribution of the variables. A receiver operat­­ing characteristic analysis was performed for analysis of the discriminatory accuracy of vedolizumab levels for PGA response outcome. Multivariate analysis was performed us­­ing logistic regres­sion to evaluate predictive value of selected factors for PGA response. The statistical tests were performed us­­ing GraphPad Prism (version 8.0.1) and IBM SPSS Statistics 24. A p-value (two-tailed) < 0.05 was considered statistical­ly significant.

Results

Baseline patient characteristics

A total of 87 IBD patients were included, 31 with CD and 56 with UC. Most of the patients (74, 85.1%) were anti-TNF-α experienced, only a minority were naïve to bio­logic ther­apy (13, 14.9%). Concomitant use of thiopurines was present in 23 (26.4%) patients and 53 (60.9%) patients had systemic corticosteroids. Dis­ease and baseline characteristics of included patients are shown in Tab. 1.

Median (range) CRP levels at the begin­n­­ing of ther­apy (week 0) were 7.6 (0.3– 97.9) mg/ L in CD patients and 2.0 (0.2– 62.8) mg/ L in UC patients. Week 0 FC levels were 1 419 (55– 6 000) µg/ g in CD patients and 2 207 (34– 6 274) µg/ g in UC patients.

Treatment response

One third of patients (33, 38%) needed additional dose of vedolizumab at week 10 (9 patients with CD and 24 patients with UC). Data for evaluation of clinical response at week 10– 14 were available for 77 out of 87 IBD patients (88.5%). Of them, 59 (76.6%) patients were considered responders to induction ther­apy, 8 (10.4%) had complete and 51 (66.2%) partial response while 18 (23.4%) experienced primary non-response (Graph 1).

Vedolizumab Trough Levels and Antibodies

At week 6, median VTLs in the entire study population were 30.6 µg/ mL (1.1– 80.0) with no significant dif­ference between CD and UC (27.8 vs 33.5 µg/ mL; p = 0.21).

Out of 85 patients with available AVA values, 6 (7.1%) developed positive AVA until week 10– 14.

Compar­­ing patients with and without clinical response to vedolizumab induction ther­apy, no significant dif­ference in median VTL was found (29.4 vs 34.4 µg/ mL; p = 0.71). A subgroup analysis of individuals with CD and UC separately revealed similar findings to the whole cohort (Graph 2a,b,c). Furthermore, week 6 VTLs of patients with additional dose at week 10 did not significantly dif­fer from those obtain­­ing standard induction only (40.0 vs. 28.5 µg/ mL; p = 0.69).

Analysis of factors with potential impact on VTL revealed no significant as­sociation with age, gender, dis­ease characteristics, previous anti-TNF-α ther­apy or use of concomitant im­munosuppres­sive treatment (data not shown).

Discus­sion

In our observational study, we explored the as­sociation between early VTL and response to induction ther­apy in patients with IBD. Around three quarters of patients were considered responders to induction ther­apy. However, approximately one third of them required an additional dose at week 10. VTLs at week 6 were neither predictive of short-term response to vedolizumab nor of a need for treatment intensification dur­­ing induction phase. Im­munogenicity was slightly higher with AVA be­­ing detected in 7% of patients.

The rate of short-term response observed in our cohort is in agreement with findings of other observational studies despite dif­ferent definitions used [11]. Slightly over one third of our patients needed an additional dose at week 10 of induction phase. Previous studies reported wide range of treatment intensification dur­­ing induction phase, from 11% to 54% of treated patients, which may reflect dif­ferent treatment policy of individual centres [10– 12].

We failed to demonstrate significant dif­ference in early VTLs compar­­ing patients with and without clinical response to vedolizumab induction ther­apy. In the literature, results on predictive value of early VTL on short-term but also long-term vedolizumab response are conflicting. While some studies suggested significant dif­ferences in VTL already at week 2 of vedolizumab ther­apy, others failed to prove any as­sociations of VTL dur­­ing induction phase with short- or long-term ef­ficacy [7,10– 14].

Furthermore, no dif­ference in VTL, stratified by need for additional dose at week 10, was found in our study. Similar results were reported in other observational cohort studies [10,11].

The role of TDM in optimis­­ing anti-TNFs agents to treat IBD has been long established, especial­ly in situations of treatment failure [15]. For vedolizumab, however, the ef­ficacy of TDM and its potential use for improv­­ing clinical outcomes is yet to be determined. It has been reported that serum concentrations of vedolizumab needed for complete receptor occupancy, either on circulat­­ing or mucosal T cel­ls, do not cor­respond to that required for therapeutic ef­ficacy [16]. Recent study from Israel showed that already low concentrations of vedolizumab led to full receptor saturation of peripheral blood T cel­ls and that the rate of receptor occupancy on both peripheral and mucosal T cel­ls was unrelated to treatment response [10]. So, as discus­sed by Ward et al., it is pos­sible that unlike anti-TNFs, serum through levels alone may be inadequate to predict clinical response with vedolizumab. Therefore, new bio­markers which more accurately reflect the vedolizumab pharmacokinetic-pharmacodynamic relationship are needed [4].

Seven percent of patients from our cohort developed positive AVA by week 10– 14. This is slightly higher than reported by studies us­­ing drug-sensitive as­says with AVA rates of less than 5% [4]. Interestingly, data from Leuven (us­­ing drug tolerant as­say) suggested that most of these early AVAs are only transient, have no impact on treatment response, and disappear dur­­ing maintenance phase [17]. Similar observation was reported also in the study from Israel [10].

No significant as­sociation was found between VTL and dia­gnosis, sex, age, dis­ease characteristics or use of im­munomodulators or systemic cortico­steroids. A recently published study by Al-Bawardy et al. supports our findings by hav­­ing observed there was no significant dif­ference in median VTLs between patients receiv­­ing and not receiv­­ing im­munosuppres­sants (16.1 µg/ mL vs. 14.8 µg/ ; p = 0.81) or between those treated or not treated with cortico­steroids (16.8 µg/ mL vs. 14.8 µg/ mL; p = 0.76) [18].

There are several limitations of our study. Firstly, our patient cohort was relatively smal­l, especial­ly regard­­ing CD patients. Furthermore, some data were col­lected retrospectively, includ­­ing as­ses­sment of dis­ease activity by PGA. Final­ly, we did not evaluate objective parameters of treatment response such as bio­markers alone or mucosal healing.

In conclusion, our study did not find any as­sociation between early VTL and a short-term response to induction ther­apy in IBD patients. There is a need for further studies with larger patient populations to address the usefulness of TDM in short and long-term vedolizumab treatment.

The authors declare they have no potential conflicts of interest concerning drugs, products, or services used in the study.

The Editorial Board declares that the manuscript met the ICMJE „uniform requirements“ for bio­­­­medical papers.

Funding

This study was supported by the IBD-COMFORT foundation.

 Submitted: 28. 1. 2019

 Accepted: 3. 2. 2019

MUC. Karolína Pudilová

The Clinical and Research Centre

for Inflam­matory Dis­eases

ISCARE I.V.F. a.s.

Jankovcova 1569/ 2c

170 00 Prague

Czech Republic

pudilova.karolina@seznam.cz