Therapeutic drug monitoring-based dosing of TNF inhibitors in inflammatory bowel disease: the way forward?

ABSTRACT

Introduction: Secondary loss of response to anti-tumor necrosis factor (TNF) therapy remains a challenge in the clinical management of inflammatory bowel disease (IBD) patients. A frequently observed reason for secondary loss of response to TNF blockers is inadequate drug exposure and sub-therapeutic serum drug concentrations.

Areas covered: This review presents an overview of recent research on therapeutic drug monitoring (TDM)-based dosing with anti-TNF agents in IBD. The role of reactive and proactive TDM and different approaches on how to optimize anti-TNF treatment are discussed.

Expert opinion: Due to variations within and between patients, the ‘one size fits all’ theory does not apply to all IBD patients receiving anti-TNF agents. Timing of TDM (i.e. reactive versus proactive) is a matter of debate. Both strategies might optimize anti-TNF treatment, although most trials did not show a clinical benefit compared to conventional dosing up to now. So-called dashboard systems might have an additive value in the optimization of anti-TNF treatment, since these tools enable clinicians to really personalize anti-TNF treatment.

KEYWORDS:

• Anti-tumor necrosis factor (TNF) therapy
• biosimilars
• dashboard system
• inflammatory bowel disease (IBD)
• therapeutic drug monitoring (TDM)

1. Introduction

New and promising treatments are appearing for patients suffering from inflammatory bowel disease (IBD). For Crohn’s disease (CD) and ulcerative colitis (UC), several therapeutic antibodies have received market authorization, including anti-tumor necrosis factor (TNF) antibodies, vedolizumab, and ustekinumab. The small molecule tofacitinib, a janus kinase (JAK) inhibitor, is the most recently approved agent for UC. Various therapeuticals are currently being evaluated in phase 2 and 3 clinical trials targeting inflammatory cytokines (such as anti-IL-23 and anti-IL36), leukocyte trafficking (anti-MadCam antibodies, anti-αEβ7 antibodies, and sphingosine-1-phosphate receptor agonists) or the intestinal microbiome [1–5]. Phosphodiesterase-4 inhibitors (such as apremilast and roflumilast) and JAK inhibitors (including filgotinib and upadacitinib) are promising small molecules [6–8]. Hence, treatment options for IBD patients are rapidly increasing. However, there is considerable variation in treatment discontinuation rates between different therapeutic agents, which can be caused by side effects, primary non-response or secondary loss of response [9–12]. Although these novel therapeutic agents may offer an additional benefit to treat IBD, TNF inhibitors still represent one of the most powerful agents [13–15]. In Europe, there are three anti-TNF agents available for IBD: infliximab (IFX), adalimumab (ADL) and golimumab (GLM). However, up to 50% of patients who initially respond to anti-TNF therapy lose their response over time (secondary loss of response), and this often leads to treatment discontinuation [16]. Loss of response to anti-TNF therapy due to inadequate drug exposure remains a problem in the treatment of IBD patients. Here, we will discuss different strategies on how to optimize anti-TNF treatment in IBD.

2. Anti-TNF agents and therapeutic drug monitoring (TDM)

Inadequate drug exposure and the formation of anti-drug antibodies (ADA) are important factors that contribute to secondary loss of response [17–19]. The importance of achieving adequate drug exposure of anti-TNF agents has been well established in IBD [20–24]. For both IFX and ADL, serum albumin, C-reactive protein (CRP), body weight and ADA are associated with increased drug clearance, resulting in inadequate drug exposure [25–27]. In order to optimize treatment outcomes with TNF blockers, all these factors should be taken into account. In line with these notions, therapeutic drug monitoring (TDM)-based dosing might increase therapeutic efficacy. TDM is defined as drug concentration measurements and adjusting the dose and/or dosing intervals in order to maintain serum concentration within a certain therapeutic range in order to optimize treatment outcomes [28]. For anti-TNF agents, higher trough levels (TLs, i.e. lowest drug concentration measured just before the next administration) are associated with improved treatment outcomes.

2.1. Reactive TDM-based IFX dosing

In clinical practice, TDM is usually performed in a reactive manner. This means that serum levels are measured when there is a suspicion of loss of response to anti-TNF therapy with confirmed active disease. In case of sub-therapeutic TLs, anti-TNF treatment can be intensified. In addition, when patients have elevated anti-drug antibodies, combination therapy with an immunomodulator (methotrexate or thiopurines) might be considered. When TLs are within the therapeutic range, a switch out of class should be considered [29].

Most data on optimizing treatment outcomes with anti-TNF agents using TDM are available for IFX. For CD and UC, standard IFX dosing is based on body weight (5 mg/kg) starting with an induction phase at weeks 0, 2, and 6, followed by an 8-weekly maintenance phase. However, there is substantial variability regarding drug exposure and treatment responses between patients who receive standard IFX doses. Hence, this underscores the need for personalized IFX dosing to reduce primary non-response and secondary loss of response rates. In general, IFX TLs between 3 and 7 µg/ml are associated with optimal clinical and endoscopic outcomes during maintenance treatment [22,30–32]. However, optimal target IFX TLs depend on desired treatment outcomes. Higher IFX TLs are, for example, needed to obtain endoscopic remission (≥9.7 mg/L) compared to biochemical remission (≥2.2 mg/L) [33]. For histological healing, IFX TLs ≥9.8 mg/L are required in CD and ≥10.5 mg/L in UC patients [33,34]. In addition, for CD patients with perianal fistulas, there is increasing evidence to support the idea that higher IFX serum concentrations are needed for complete disease control compared to patients without fistulizing disease [35–37]. Davidov et al. compared IFX serum concentrations during induction treatment between CD patients with and without fistula closure [36]. A positive association was found between IFX serum levels during induction treatment and clinical perianal fistula closure, indicating that these patients need higher IFX induction levels. Another study also showed higher IFX serum levels during maintenance treatment in patients with closed perianal fistulas compared to patients with persistently producing fistulas (15.8 versus 4.4 mg/L, respectively) [35]. The underlying mechanism should be investigated in future studies, but high local production of TNF within the fistula tract and in the surrounding tissue might explain the need for higher anti-TNF serum concentrations in this patient subgroup.

Standard ADL dosing consists of 160 mg at week 0 and 80 mg at week 2, followed by maintenance treatment 40 mg every other week. In clinical practice, ADL treatment can be intensified from 40 mg every other week to 40 mg every week, depending on clinical symptoms, inflammatory biomarkers (CRP, fecal calprotectin) or findings obtained with endoscopy and/or imaging. Also for ADL, higher ADL serum concentrations are associated with improved clinical outcomes [23,38]. In a recent prospective study performed by VandeCasteele et al., IBD patients with ADL TLs >7.3 mg/L at week 12 were more likely to achieve clinical remission. For endoscopic and histologic remission, ADL TLs ≥12 mg/L and ≥12.2 mg/L are associated with optimal outcomes, respectively [39]. The presence of anti-ADL antibodies is associated with loss of clinical response to ADL. Pharmacokinetic studies with ADL in CD patients showed that clearance is increased in case of detectable ADA, resulting in lower ADL serum concentrations and loss of response [27,40,41].

GLM represents another subcutaneous anti-TNF agent, which is approved for UC in Europe. Induction treatment consists of 200 and 100 mg at week 0 and 2, respectively, followed by a maintenance phase with 50 or 100 mg every 4 weeks, according to body weight (<80 kg or ≥80 kg, respectively). With the recent label change, patients with a body weight <80 kg who do not respond to induction therapy with GLM 200 and 100 mg at week 0 and 2, respectively, can also receive 100 mg GLM as maintenance treatment. Limited data are available about TDM-based dosing with GLM in UC. During maintenance therapy, GLM TLs >1.4 µg/ml are associated with clinical remission [42]. The GO-KINETIC trial investigated the pharmacokinetics of GLM in 20 patients with moderate to severe UC receiving induction and maintenance treatment [43]. In this prospective observational trial, clinical, biochemical and endoscopic outcomes were evaluated, and a thorough pharmacokinetic analysis was conducted. Patients with endoscopic response at week 8 had numerically higher median GLM serum concentrations at week 2 compared to endoscopic non-responders: 9.1 mg/L [5.9–12.3] vs 7.1 mg/L [5.2–9.0]; p = 0.384, respectively. Population pharmacokinetic analysis showed an inverse association between albumin concentrations and GLM clearance.

Furthermore, anti-TNF agents do not have a narrow therapeutic range and they cannot be easily overdosed. Moreover, higher IFX TLs do not seem to increase the risk of infections [44,45]. On the other hand, these agents are relatively expensive and they should be dosed in their most optimal way, not only from the patients perspective but also from an economic point of view. In addition, there is evidence that IFX TLs >5.5 mg/L and ADL serum concentrations >6.6 mg/L during maintenance treatment are associated with impaired quality of life [46].

2.2. Proactive TDM-based anti-TNF dosing

Although TDM-based personalized dosing seems to be an optimal strategy to prevent under- and overdosing, benefits of proactive dose adjustments with TNF blockers have not been demonstrated so far. Proactive TDM is performed at predefined time points, irrespective of symptoms. Proactive TDM might prevent a flare of the disease due to under-dosing or treatment might be safely de-escalated in case of supra-therapeutic TLs. The TAXIT trial included IBD patients in clinical remission receiving IFX maintenance treatment but failed to show higher clinical remission rates (primary outcome) in patients receiving TL-based IFX dosing compared to patients receiving clinically based dosing of IFX [47]. However, all patients in this trial initiated with dose optimization based on TDM to obtain an IFX TL between 3 and 7 mg/L. TAXIT did show that optimizing TLs between 3 and 7 mg/L resulted in longer relapse-free survival rates and it also proved to be cost-effective. In the TAILORIX trial, anti-TNF-naïve CD patients were randomized between IFX dose adaptations based on TDM, clinical symptoms, and biomarkers or dose adaptations based on clinical symptoms after first receiving standard induction treatment with IFX [48]. IFX dosing that was based on TDM, clinical symptoms and biomarkers proved not to be superior compared to treatment intensification based on clinical symptoms alone. However, possible differences between the two groups might be missed due to the fact that TAILORIX was underpowered. These two prospective trials that evaluated pro-active TDM-based IFX dosing in CD did not use an algorithm in which factors were incorporated influencing IFX pharmacokinetics [49]. A retrospective study performed by Papamichael et al. studied patients receiving proactive IFX monitoring after reactive testing, or reactive testing alone [50]. Patients receiving proactive IFX monitoring showed less treatment failure and fewer IBD-related hospitalizations compared to patients undergoing reactive TDM.

Up to now, no prospective TDM study has been conducted in IBD patients receiving ADL treatment. The open-label CALM study compared two treatment algorithms. Treatment could be intensified in a step by step manner (from no treatment, to ADL every other week, to weekly ADL, to weekly ADL plus azathioprine) in CD patients, who were naïve for immunomodulators and biologics, based on clinical symptoms alone or based on symptoms combined with inflammatory biomarkers (fecal calprotectin and CRP) [51]. This study showed that CD patients, with a relative short disease course, have better clinical and endoscopic outcomes when treatment escalation is based on clinical symptoms and biomarkers versus treatment escalation based on clinical symptoms alone.

2.3. TDM during anti-TNF induction treatment

As opposed to TDM-based IFX dosing during maintenance treatment, TDM might also have a beneficial effect during induction treatment. When already adjusting the dose or dosing interval in an early phase of treatment, loss off response might be prevented. It has been shown that an accelerated IFX induction regimen can reduce early colectomy rates in patients with severe UC [52]. Target IFX serum concentrations at week 2 associated with clinical outcomes range between 6.8 and 21.3 mg/L [53–55]. For endoscopic response, Papamichael et al. reported an association between optimal endoscopic outcomes and serum IFX concentrations ≥28 mg/L during the first 2 weeks after treatment initiation, and ≥15.0 mg/L between week 2 and 6 in UC patients receiving standard IFX induction treatment [56]. Similar for CD patients, IFX serum concentrations >23.1 mg/L at week 2 and >10.0 mg/L at week 6 are associated with endoscopic remission at week 12 [57].

Proactive TDM-based dose optimization might also improve therapeutic outcomes in IBD patients treated with ADL. A retrospective study showed that short-term mucosal healing in UC was associated with an ADL TL at week 4 of 9.4 mg/L.

3. Dashboard systems for personalized treatment

The importance of achieving therapeutic anti-TNF serum concentrations is well known and this leads to the hypothesis that TDM-based dosing might be superior to conventional dosing. However, prospective evidence to support this theory is lacking [47,48]. A possible explanation could be that none of the clinical trials that prospectively investigated TDM-based dosing used a population pharmacokinetic model to optimize dosing strategies. Population pharmacokinetic models can be used to identify parameters and sources of variability in a population. Furthermore, these models can be used to simulate dosing regimens and allow physicians to apply individualized dosing. A dashboard system is a valuable tool for non-pharmacologists to use population-based pharmacokinetic models [58]. This computer system comprehends a pharmacokinetic model that consists of a large dataset derived from a population similar to the target population. The dashboard system calculates the exact dose that a patient should receive at what time in order to maintain a certain serum drug concentration using a Bayesian approach. The more individual patient data are entered into the system, the more accurate the prediction will be for a particular patient. Hence, the first prediction is based on an average patient with typical factors that are known to affect the pharmacokinetics of IFX. Dubinski et al. compared dosing regimens recommended by a pharmacokinetic dashboard system versus standard care in pediatric IBD patient receiving IFX maintenance treatment [59]. The dashboard system recommended the standard IFX dose (5 mg/kg every 8 weeks) in only 22% of patients. Especially in patients with increased drug clearance, the dashboard system advised an intensified dosing regimen in order to maintain therapeutic TLs. To individualize anti-TNF dosing in daily practice, these dashboard systems might be a useful instrument for physicians. An in silico study, performed by Wojchichowski et al., showed a clinical benefit of personalized IFX dosing using a dashboard system [60]. These authors demonstrated that Bayesian forecasting resulted in reduced time to the first target trough achievement and less interpatient variability compared to standard dosing according to the label. When target serum levels are achieved earlier in the disease course, it is to be expected that less patients will develop clinically relevant ADA resulting in loss of response.

4. Biosimilars

In 2015, the European patent of originator IFX (Remicade®, Johnson and Johnson, New Jersey, USA) expired and this led to the introduction of biosimilar IFX CT-P13 (Celltrion, Incheon, South Korea). To date, several IFX biosimilars are available on the market (e.g. Remsima®, Flixabi®, Inflectra®, Zessly®). Although biosimilars are highly similar to the originator product, differences may appear due to the structural complexity and manufacturing procedures, leading to potential differences in biochemical characteristics. The approval of CT-P13 for the same indications as the originator was based on trials performed in patients with ankylosing spondylitis and rheumatoid arthritis [61,62]. In these trials, CT-P13 demonstrated equivalence to originator IFX in its pharmacokinetic, efficacy and safety profile. The decision of the European Medicines Agency (EMA) to extrapolate these results in other indications, including IBD, initially led to some hesitancy within the gastroenterology field. Soon after the introduction of biosimilar IFX, the first results of clinical trials investigating the efficacy of CT-P13 in IBD revealed similar efficacy, safety, and immunogenicity outcomes compared to originator IFX [63–72]. Also, long-term data show that switching to CT-P13 is well tolerated and feasible [73,74]. The main reason to use biosimilars is to reduce costs and to increase accessibility to these agents. As it turned out to be safe to start CT-P13 induction treatment in anti-TNF naïve patients, the next step was to switch patients from originator IFX to CT-P13. In the NOR-SWITCH trial, patients with various auto-immune diseases, including CD and UC, were randomized to continued IFX originator treatment or they were switched to CT-P13 [75]. No differences in clinical remission rates, pharmacokinetics, immunogenicity or adverse events were observed between the two groups. The results of this trial together with other trials demonstrated similar results and increased the confidence in prescribing IFX biosimilar agents amongst gastroenterologists. Since 2018, ADL biosimilars have entered the market (e.g. Amgevita®, Cyltezo®, Hyrimoz®, Imraldi®, Hulio®, Solymbic®). To date, clinical data with ADL biosimilars are only available from other chronic inflammatory diseases, such as rheumatoid arthritis and psoriasis, showing similar rates of immunogenicity, safety, and tolerability and efficacy outcomes [76–81].

5. Expert opinion

There is increasing data on how to optimize anti-TNF treatment in IBD. Although available evidence from prospective trials is limited, we strongly believe in the additive value of TDM-based treatment approaches. In an ideal clinical setting, anti-TNF treatment is started as a combination therapy together with methotrexate or a thiopurine and is already dose optimized during induction treatment. If needed, patients can receive IFX at higher doses (10 mg/kg) and/or at decreased dosing intervals, which will be mainly based on inflammatory biomarkers and IFX serum level measurements. A dose reduction of anti-TNF agents in IBD patients who are in remission with supra-therapeutic anti-TNF serum levels might not only be beneficial for patients but will also result in significant cost savings. To prevent secondary loss of response, pro-active measurement of anti-TNF TLs and dose adjustments are needed. In clinical practice, it can take up to two weeks after blood sample withdrawal before TDM laboratory results become available. By making use of point-of-care tests, results of anti-TNF TLs can be available within a couple of minutes. This enables clinicians to immediately perform dose adjustments if needed when the patient is at the out-patient clinic or at the infusion clinic. Also, in a more proactive TDM setting, the use of dried blood samples (DBS) for anti-TNF serum measurements can optimize anti-TNF treatment in a timely manner. With a finger prick, patients can perform DBS at home. Capillary blood is collected on a device, which can be sent to the laboratory. Results will be available before patients come to the out-patient clinic.

Factors that are associated with increased clearance of TNF inhibitors are indirect indicators of disease activity (inflammatory burden), such as CRP and fecal calprotectin. Patients with high CRP concentrations experience a higher clearance of anti-TNF treatment, presumably because of their higher inflammatory load. TDM-based dosing might be implemented in daily practice on a larger scale by the use of dashboard systems in which this pharmacokinetic information from a population can be used to individualize anti-TNF dosing. Optimal anti-TNF dosing should take into account both serum drug levels and inflammatory biomarkers, that are surrogate markers of disease activity. Altogether, this will likely improve treatment outcomes in IBD patients receiving treatment with TNF inhibitors. However, TDM also has some limitations. Treatment goals in IBD often differ between patients and might also vary throughout the disease course. The optimal serum anti-TNF concentration might be patient-specific, vary over time, and might also depend on specific treatment goals (such as mucosal healing, perianal fistula closure, histological remission, etc.). Furthermore, because syringes contain fixed doses, individualized dosing of subcutaneous anti-TNF agents (including ADL and GLM) will be limited to changing treatment intervals.

To date, TNF blockers are one of the most powerful therapeutic agents to treat IBD. Individualized dosing of these therapeutic antibodies, using pharmacokinetic modeling, will conceivably increase primary response rates. Identification of factors that influence patients’ individual clearance and defining patient-specific target serum concentrations might lead to lower primary and secondary loss of response rates ultimately improving treatment outcomes.

Article highlights

• Tumor necrosis factor (TNF) antibodies represent one of the most powerful agents in the treatment of inflammatory bowel disease (IBD).

• Loss of response to anti-TNF therapy represents a clinical problem that is frequently observed in daily practice.

• Therapeutic drug monitoring (TDM)-based dosing might improve anti-TNF treatment, since therapeutic anti-TNF trough levels are associated with optimal clinical and endoscopic outcomes in IBD.

• Based on patients’ characteristics (i.e. presence of active disease, body weight, trough concentrations, and anti-drug antibody status) patients should be dosed on an individual basis.

• Besides reactive TDM-based dosing, proactive TDM and TDM during anti-TNF induction therapy might prevent loss of response to these agents.

• The use of a dashboard system, point-of-care tests, and dried blood samples might further improve treatment outcomes with TNF inhibitors in IBD.

Declaration of interest

M Löwenberg has served as speaker and/or principal investigator for Abbvie, Celgene, Covidien, Dr. Falk, Ferring Pharmaceuticals, Gilead, GlaxoSmithKline, Janssen-Cilag, Merck Sharp & Dohme, Pfizer, Protagonist therapeutics, Receptos, Takeda, Tillotts, Tramedico. M Löwenberg has received research grants from AbbVie, Merck Sharp & Dohme, Achmea healthcare and ZonMW. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.