Review: optimizing rescue therapy in acute severe ulcerative colitis using infliximab pharmacokinetics—a practical guide for clinicians

4.1. Overview

Infliximab is administered as an IV infusion, bypassing the oral route. Elimination occurs via the reticuloendothelial system from phagocytosis and subsequent proteolytic degradation by immune cells.^26,27 The pharmacokinetic (PK) profile of infliximab after administration is influenced by its rate of distribution to body tissues as well as the rate of elimination. Serum infliximab concentrations follow a biphasic pattern: the distribution phase and the elimination phase (see Figures 2 and 3). In the distribution phase, measurable serum infliximab levels decrease more quickly with a “distribution half-life” of hours to days as the drug rapidly distributes into peripheral tissues.^28,29 The elimination phase commences after an equilibrium is reached between the concentrations in the central and peripheral compartments, such that serum levels drop more slowly with a “terminal half-life” of 11-15 days.^28,29 In this phase, the predominant driver of infliximab concentrations is elimination from the body. Initial peak concentrations are high and typically greater than 100 µg/mL, reducing to a week 8 trough level of 3-7 µg/mL.¹⁷ The unmeasurable tissue concentration is thought to be more stable and reach a steady state after approximately three to six infusions.

Mathematically, infliximab PK is typically modeled using a two-compartment model in order to account for its biphasic nature (see Figures 2-4).^26,28 The two-compartment model divides the body into a central compartment, representing the serum, and a peripheral compartment, representing the body tissues, which includes the intestinal mucosa.²⁸ Under this model, infliximab PK is governed by four PK parameters: clearance, central volume of distribution, peripheral volume of distribution, and intercompartmental clearance (see Figure 2). The central and peripheral volumes of distribution are the theoretical maximal volumes into which the drug can diffuse in the central and peripheral compartments respectively. Biologically, they can be simplistically interpreted as total blood volume and tissue volume respectively. Clearance is defined as the elimination rate divided by the central volume of distribution, and is thus directly correlated with elimination, and inversely correlated with central volume of distribution. Intercompartmental clearance governs the rate at which drug transfers between compartments.

4.2. Altered pharmacokinetics in ASUC

Infliximab exhibits altered pharmacokinetics in ASUC due predominantly to higher rates of elimination (see Table 2), which results from an increased inflammatory burden compared to healthy outpatients.^2,30 Infliximab clearance is highest in early ASUC and decreases over time as inflammation improves.^29,31 After standard induction at weeks 0, 2, and 6, serum infliximab concentrations have been found to be lower in ASUC patients compared to moderately severe UC patients in a single study.¹⁵ Biological mechanisms for a greater elimination in ASUC include increased proteolytic degradation in the heightened inflammatory state, target-mediated drug disposition, stool losses, and anti-drug antibody formation.

4.3. Predictors of infliximab pharmacokinetics

Population pharmacokinetic studies fit observed drug concentrations from a cohort of patients using mathematical models combined with a statistical covariate model to describe associations between pharmacokinetic parameters (clearance, volumes of distributions, and intercompartmental clearance) and clinical or biochemical factors such as sex, body weight, and inflammatory markers.²⁸ Covariate models can be interpreted as the collection of variables that most explains or determines the pharmacokinetic parameter of interest (see Table 3).

4.4. Limitations of current PK models

Despite the known relationship between infliximab PK and inflammation, attempts to study infliximab in ASUC are limited due to the failure to account for the reciprocal relationship between infliximab PK and pharmacodynamics (PD) due to the high burden of inflammation. Furthermore, such models fail to capture the real-time dynamics of inflammation and drug clearance because important parameters such as the elimination rate are treated as constant. The phenomenon in which a higher initial dose results in higher clearance suggests non-linear PK with concentration-dependent elimination, which is unaccounted for in current PK models. More robust models that combine PK and PD frameworks, as well as incorporate a time-varying elimination rate from the peripheral compartment, dependent on clinical and biochemical markers of inflammation, are likely to be more useful in optimizing infliximab dosing in patients with corticosteroid-refractory ASUC.

5.1. The biological propensity of infliximab response

Approximately 20%-30% of ASUC patients fail to respond to infliximab.^3,14 A patient’s biological propensity to respond to a sufficient exposure to infliximab can be theorized to depend predominantly on the degree to which inflammation can be arrested with anti-TNF therapy.⁵⁰ This, in turn, depends on the degree of total inflammation, as well as the proportion of the inflammation that is mediated by TNF-related pathways. A patient who has a modest degree of inflammation that is entirely mediated by TNF-α should in theory have a high propensity to respond to anti-TNF therapy. On the other hand, a patient with a high degree of inflammation that is predominantly mediated by non-TNF immunological pathways would have a low propensity to respond to anti-TNF therapy.

As a result, current literature purports that mechanisms of infliximab failure can be broadly categorized into drivers related to pharmacokinetic failure and drivers related to pharmacodynamic failure.⁵⁰ In practice, although the discrete categorization of infliximab failure mechanisms into two broad pharmacological processes can be useful, it is important to acknowledge that a patient’s propensity to respond exists on a spectrum, and thus a substantial overlap likely exists between the two categories of failure. For example, unfavorable PD leading to persistent inflammation would also drive unfavorable PK from increased infliximab clearance.

5.2. Pharmacokinetic failure: overcoming high clearance with greater dosing

For ASUC patients with a high biological propensity to respond to infliximab, actual response depends on whether sufficient drug can be delivered to achieve a stoichiometric excess relative to TNF-α in the setting of increased drug clearance. Patients who are inadequately dosed and fail to overcome the altered PK in ASUC are considered to have “pharmacokinetic failure.” Put simply, there are not enough infliximab molecules relative to TNF-α molecules, such that the predominant TNF-driven inflammation fails to arrest.

High clearance has been consistently shown to predict worse outcomes in ASUC, including persistent clinical symptoms, endoscopic activity, and colectomy.^{17,29,31,32,51,52} Similarly, greater stool infliximab losses are associated with infliximab non-response and colectomy.^17,36,37 One study demonstrated that the week 2-4 infliximab exposure to clearance ratio predicted colectomy beyond week 4.⁵¹ PREDICT-UC is currently the only prospective ASUC study to report on outcomes stratified by infliximab clearance and dose.¹⁷ Subsequent modeling work from PREDICT-UC showed that corticosteroid-refractory ASUC patients with high clearance (defined as ≥0.62 L/day) were 4.1 times more likely to fail infliximab clinically by 14 days and 8.2 times more likely to require colectomy by 3 months.¹⁷ However, an initial infliximab dose of 10 mg/kg vs 5 mg/kg was 50% more likely to capture an initial response and reduced colectomy rates by 79% in patients with high clearance. Further, in patients with high clearance who failed to respond to the initial dose, a second dose of 10 mg/kg was 3.4 times more likely to induce a response by 14 days compared to 5 mg/kg.

In lieu of clearance, which can be challenging to determine in the absence of PK modeling tools, lower serum infliximab levels are predictive of worse outcomes in both the moderate–severe and ASUC setting.^52–56 PREDICT-UC showed that a lower day 3 serum infliximab level was predictive of day 14 infliximab failure and colectomy, where a day 3 level ≤57.9 μg/mL predicted colectomy with 83.3% sensitivity, 67% specificity, 24.4% positive predictive value (PPV), and 96.9% negative predictive value (NPV).¹⁷ The TITRATE RCT explored the potential utility of dashboard-driven infliximab dosing in steroid-refractory ASUC, aiming for serum levels >28 µg/mL in the first 28 days and >15 µg/mL from days 29 to 42.¹⁸ Although TITRATE did not show a benefit to dashboard-driven dosing compared to standard infliximab induction, the lack of benefit may be related to the relatively modest early serum-level targets, compared to the higher thresholds identified in PREDICT-UC. Another recent study developed an interactive tool to guide individualized infliximab dosing based on a dose–exposure–response framework, though efficacy of this approach over a conventional approach has not been tested.⁵¹

Since true infliximab clearance can only be determined during treatment, baseline stratification by covariate models and their constituents such as albumin or CRP may be helpful in ­determining the first infliximab dose. A prior prospective study similarly demonstrated that estimated clearance ≥0.627 L/day using a baseline covariate model given by $\mathrm{Cl}=0.407\times {\left(\mathrm{ALB}/4.1\right)}^{-1.54}\times {1.471}^{\mathrm{ATI}}\times {0.764}^{\mathrm{SEX}}$ predicted colectomy, where ALB is albumin concentration in g/dL, ATI is the antibodies to infliximab status, which was given the value of 0 at baseline, and SEX = 1 for female sex or 0 for male sex.^28,57 However, the study was limited by using a drug-intolerance assay to measure ATI, such that patients with measurable infliximab levels were assumed to not have antibodies. Although the results were not statistically significant, PREDICT-UC estimated the rate of response to an initial dose of 10 mg/kg to be 43% higher compared to an initial 5 mg/kg dose in patients with day 0 pre-infliximab albumin <25 g/L, and 39% higher with 10 mg/kg vs 5 mg/kg in patients with day 0 pre-infliximab CRP ≥ 50 mg/L. Further post-hoc analyses suggested that patients with day 0 albumin <25 g/L were 2.5 times more likely to fail infliximab by 14 days if they received standard induction (total of 5-10 mg/kg depending on initial response, administered by 14 days) as opposed to either accelerated induction (10-15 mg/kg depending on initial response, by 14 days) or intensified induction (20 mg/kg by 14 days). In low albumin (<25 g/L) patients who failed to respond initially to the first dose, a second dose of 10 mg/kg was 4.5 times more likely to capture response by 14 days compared to a second dose of 5 mg/kg.

In theory, the extent and dynamic change of a patient’s response to the first infliximab dose may also offer insight into the mechanism of potential infliximab failure. For example, an initial response followed by early relapse may suggest impending PK failure and inadequate dosing. Similarly, an initial partial response raises the possibility of further improvement with greater infliximab dosing, although it is important to acknowledge that pharmacodynamic reasons for poor response are likely to be present to various degrees in both situations.

5.3. Pharmacodynamic failure: an opportunity for sequential therapy

Patients who have a low biological propensity to respond due to the presence of alternate immunological pathways tend to experience “pharmacodynamic failure,” characterized by an inability to suppress inflammation despite sufficient infliximab exposure. In such patients, alternate immunosuppressive agents are required to reduce the risk of requiring colectomy. The Janus kinase and signal transducer and activator of transcription (JAK/STAT) intracellular signaling mechanism mediates numerous pro-inflammatory cytokine pathways including IL-2, IL-6, interferon and TNF-α.^58–60 JAK inhibition may thus induce a more “global” immune suppression compared to isolated TNF-α blockade. Due to their potential as alternatives to infliximab, JAK inhibitors have gained traction over calcineurin inhibitors as sequential therapy in infliximab-refractory ASUC.^61,62

The concept of pharmacodynamic failure has been drawn from the outpatient setting and tends to refer to patients who exhibit primary non-response, or patients who have secondary loss of response despite optimized dosing and robust serum infliximab levels, suggesting an interval development of alternate immunological pathways. In the ASUC setting, pharmacodynamic failure can be inferred by the lack of a sustained clinical, biochemical, and/or endoscopic response to intensified or accelerated infliximab dosing, although it remains difficult to truly rule out concomitant pharmacokinetic failure without quantification of serum and/or mucosal TNF-α levels. Nevertheless, the residual failure rate even in patients receiving escalated infliximab induction regimens in ASUC suggests there exists a subgroup of patients who experience early pharmacodynamic failure.

CRP is a well-recognized serum inflammatory marker that is simultaneously used to monitor the early pharmacodynamic response to treatment, as well as a biomarker target, where a failure to suppress CRP during ASUC is frequently used as an indicator of non-response.²⁵ Fecal calprotectin dynamics also differ between infliximab responders and non-responders in ASUC.¹⁷ Results from PREDICT-UC, where two-thirds of patients received accelerated or intensified induction, found that fecal calprotectin decreased by 59% per week on average amongst patients who avoided colectomy, but failed to decrease in patients who required colectomy.¹⁷ Moreover, a greater percentage reduction in fecal calprotectin from baseline to day 14 was predictive of endoscopic remission at 3 months.¹⁷ Therefore, in patients who have already received escalated and/or accelerated infliximab dosing, a lack of early CRP and/or calprotectin response may suggest pharmacodynamic failure, necessitating a switch to sequential therapy.

In addition to biochemical markers, endoscopy and/or ultrasonography may also aid in identifying patients who may benefit from sequential therapy, though robust data are lacking. A lack of endoscopic response after a median of 9 days was found to be predictive of colectomy in a retrospective study, although the study relied on the treating endoscopists’ opinions and did not report validated scoring systems such as Mayo or Ulcerative Colitis Endoscopic Index of Severity (UCEIS).⁶³ However, another prospective study demonstrated that only the bleeding subscore of UCEIS, and not vascularity or erosion subscores, improves by day 7 after salvage (median subscore 1 at day 7 compared to median 3 at baseline),⁶⁴ suggesting that 7 days may be too soon for significant endoscopic improvement to occur. Sonographic improvement can be detected by as early as 2 weeks after commencement of immunosuppressive therapy in moderate to severe UC patients,⁶⁵ though no studies have reported on early dynamic sonographic assessment during ASUC.

Early data regarding JAK inhibitors such as tofacitinib and upadacitinib in ASUC have been promising, though a majority of the available evidence supports their use as up-front or first-line rescue therapy in ASUC, rather than their use as sequential or second-line rescue after infliximab failure.^66–72 A large retrospective study of 124 ASUC patients including 42% who received tofacitinib or upadacitinib sequentially found a 65% clinical response rate by day 7, though the risk of inpatient colectomy was higher in patients receiving a JAK inhibitor sequentially compared to as first-line rescue therapy.⁷³ A series of 11 infliximab-refractory ASUC patients treated with sequential tofacitinib 10 mg t.i.d. demonstrated a colectomy-free survival rate of 82% at both 3 and 12 months,⁶⁸ while a small case-series of four infliximab-refractory patients treated with sequential upadacitinib 45 mg daily demonstrated a 3-month colectomy-free survival rate of 75%.⁷⁴ Despite these early data, there is as yet no definitive measure to identify patients experiencing a predominant pharmacodynamic as opposed to pharmacokinetic mechanism of infliximab failure. Thus, it is important to acknowledge that switching from infliximab to a JAK inhibitor comes with the risk of abandoning infliximab prematurely, as well as the risk of adverse events of dual immunosuppression given infliximab’s long half-life.

6.1. Initial infliximab dose

Response assessment should be made between day 3 and 7 after initiation of IV corticosteroids. In corticosteroid-refractory ASUC patients in whom infliximab is the preferred rescue agent, a baseline risk assessment may help stratify patients based on their propensity to respond to infliximab. The choice of initial dose should be guided by considerations of inflammatory burden, expected pharmacokinetic profile, and any perceived risks of over-immunosuppression. The risk assessment should include baseline clinical assessment, CRP, albumin, and endoscopy. Additional pretreatment biomarkers that may predict response to infliximab include fecal calprotectin, serum IL-6, and fecal TNF.^75,76

All patients should receive an initial dose of at least 5 mg/kg infliximab. Intensified dosing such as 10 mg/kg may be considered in higher risk patients and patients expected to have high clearance, evidenced by higher Lichtiger score, higher CRP (ie, day 0 CRP ≥ 50 mg/L, where day 0 refers to the day of infliximab administration), lower albumin (ie, day 0 albumin <25 g/L) and more severe endoscopic disease activity (ie, higher UCEIS or Mayo endoscopic subscore). Pediatric patients, especially patients younger than 10 years, should receive an initial 10 mg/kg dose due to higher clearance.^77,78 Higher dosing could be considered in men compared to women due to their higher clearance, but evidence remains limited for adjustment of dosing based on gender alone: although males were found to have a lower probability of responding to anti-TNF therapy compared to females in moderate-to-severe UC,⁷⁹ no gender-based differences were found in PREDICT UC.¹⁴ A higher initial dose could also be considered in patients with a higher baseline infliximab clearance calculated based on a pre-existing formula, such as $\mathrm{Cl}=0.407\times {\left(\mathrm{ALB}/4.1\right)}^{-1.54}\times {1.471}^{\mathrm{ATI}}\times {0.764}^{\mathrm{SEX}}$ (ATI being set to 0 in infliximab-naive patients, SEX = 1 for female sex or 0 for male sex, ALB = albumin concentration in g/dL) and using a threshold of ≥0.62 or ≥0.627 L/day to inform dose intensification.^28,57 However, it is important to note that the suggested thresholds (ie, albumin <25 g/L, clearance ≥0.62 L/day) simplify a continuous disease state into binary categories, which may not reflect the underlying continuum of biological activity and dose–response relationships relevant to treatment decisions.

Patients predicted to have very high clearance may benefit from consideration of alternative small molecule rescue therapies such as calcineurin inhibitors or JAK inhibitors, which do not experience similar increased clearance in the setting of ASUC. However, all current pretreatment markers of increased infliximab clearance are also markers of disease severity (CRP, albumin, endoscopy, etc.), and may not predict relative efficacy of one agent over another.

6.2. Concomitant therapy

Patients on baseline thiopurines should continue the thiopurine and have their dose reviewed according to thiopurine metabolite testing for 6-TGN and 6-MMP. Early thiopurine commencement at the time of infliximab commencement may be considered in thiopurine-naive patients. IV corticosteroids are typically de-escalated to oral therapy (ie, prednisolone 40 mg daily) upon commencement of infliximab. Continuation of IV corticosteroids after commencement of infliximab should be carefully considered, balancing the risks of increased immunosuppression, concomitant infections such as Clostridium difficile infection, and the risk of masking early infliximab non-response. In pediatric ASUC, corticosteroids should either be ceased entirely or rapidly weaned (ie, to ≤0.2 mg/kg) to avoid excessive immunosuppression.⁷⁷ Concomitant prophylactic antibiotics are generally not required.⁸⁰ However, prophylactic antibiotic therapy for Pneumocystis jirovecii infection can be considered if there are additional risk factors, such as prolonged steroid use prior to hospitalization, sequential rescue therapy, or relative immunosuppression due to age and/or comorbidities. Although sometimes considered, there are currently no data supporting the use of concomitant albumin supplementation.

6.3. Subsequent infliximab dosing

Patients should be closely assessed for at least 2-3 days after infliximab commencement. A risk assessment on day 3 should evaluate whether clinical markers or CRP have improved.²⁵ Patients who have responded adequately could be considered for discharge for further outpatient dosing, although it is important to acknowledge that robust criteria to define infliximab response or failure have not yet been derived, and a degree of clinical acumen is thus required. For patients who have responded, accelerated induction with further infliximab dosing within 1 week, followed by a third dose at week 3, may offer advantages in terms of earlier response to therapy compared to standard induction at weeks 0, 2, and 6.¹⁴ In particular, accelerated induction (weeks 0, 1, and 3-4) is recommended in all pediatric patients due to their high clearance.⁷⁷ It may be practical to administer the second dose of infliximab on the day of discharge for patient convenience and to save costs associated with further hospital visits. Although data regarding alternative nonstandard dosing schedules such as week 0, 1, 4 or week 0, 2, 4 are lacking, such accelerated approaches are still likely advantageous compared to standard induction in patients with unfavorable PK.

Patients who fail to respond by day 3 but do not imminently need surgery should be considered for additional infliximab dosing or sequential therapy based on the individual risk profile and patient preferences. A partial response to infliximab suggests some pharmacodynamic benefit and raises the possibility of suboptimal response due to unfavorable PK. Repeat dosing with 10 mg/kg may be preferred over 5 mg/kg after 3 days, especially in higher risk patients with low day 3 infliximab levels (≤57.9 µg/mL), high clearance (≥0.62 L/day) and low albumin (<25 g/L). However, it is important to acknowledge that many centers may lack access to rapid analysis of infliximab concentrations and/or clearance calculation. In addition to considerations regarding PK, the choice of total cumulative dose of rescue infliximab should also account for the risk of infectious complications that may occur in the event of a switch to sequential medical therapy, due to the long half-life of infliximab.

6.4. Sequential therapy after infliximab failure

A failure to demonstrate an adequate clinical response despite successive infliximab doses (ie, ≥20 mg/kg cumulative dosing) should constitute “infliximab failure.” In the absence of clinical features warranting emergency colectomy, sequential therapy with a JAK inhibitor or calcineurin inhibitor could be offered, with careful discussion of the risks and benefits with the individual patient. However, there is a lack of consensus regarding the specific criteria constituting infliximab failure and the timing at which such an assessment should be applied. Given the rapid onset of action, 3 days after the most recent infliximab dose, or 14 days after the first dose may be reasonable time points to assess for infliximab failure. In addition to standard clinical markers and CRP, a decrease in fecal calprotectin at 14 days relative to baseline levels may support the diagnosis of infliximab failure and provide justification for sequential therapy, especially in patients experiencing only a partial response. Endoscopic reevaluation could be considered after 1-2 weeks, where a lack of endoscopic improvement may support the identification of infliximab failure, acknowledging that 7 days or fewer may be insufficient time for a noticeable endoscopic response.