Mistakes in therapeutic drug monitoring of biologics in IBD and how to avoid them

Reactive TDM has rationalised the management of PNR or SLR by identifying the underlying mechanisms of these unfavourable outcomes. Reactive TDM can help tailor and individualise treatment; for example, increasing the drug dose in patients with sub-therapeutic drug concentrations and undetectable or low-titer ADA. Testing for drug concentrations and ADA also would avoid giving more drugs to a patient with a mechanistic failure and adequate drug concentration. The latter would necessitate a switch in drug class.^{2, 3} Reactive TDM increases endoscopic remission rates and lessens hospitalisations compared to empiric treatment optimisation.⁴ In addition, reactive TDM is more cost-effective than empiric drug optimisation based only on clinical symptoms.⁵ Of note, the active disease should always be confirmed with objective measures of inflammation, including biomarkers, such as C-reactive protein (CRP) and faecal calprotectin, as well as endoscopy with histological evaluation. 

When utilising reactive TDM, a common mistake is abandoning treatment before optimising it. This is important as subsequent biologic therapies typically show less efficacy. It is most important when using anti-tumour necrosis factor (anti-TNF) therapy, specifically infliximab, as there are limited pharmacological options for some specific IBD phenotypes, such as perianal fistulising CD and acute severe ulcerative colitis. Of note, giving up on one anti-TNF due to adequate drug concentration suggests a mechanistic failure not just to that agent but all anti-TNFs. Thus, it is recommended that treatment discontinuation should not be considered until an infliximab or adalimumab concentration of at least 10-15 μg/ml is achieved.⁶ However, there may be occasions where these drug concentrations may not be attainable for various reasons, including very high drug clearance and insurance issues limiting dose intensification.

Cumulative evidence suggests that proactive TDM of anti-TNF therapy is associated with better outcomes than empiric treatment optimisation and/or reactive TDM. A recent meta-analysis, including retrospective studies and randomised controlled trials (RCTs), found that proactive TDM of anti-TNF therapy was associated with lower treatment failure rates than standard of care or reactive TDM. Moreover, proactive was associated with higher endoscopic remission rates than standard care.⁷ A recent RCT regarding a biologic naïve paediatric population with CD who had responded to induction infliximab therapy showed that proactive TDM compared to clinically based dosing was superior regarding sustained corticosteroid-free clinical remission and endoscopic healing.⁸ Other clinical scenarios that proactive TDM could efficiently guide clinical decisions are anti-TNF therapy de-escalation or even discontinuation and optimising infliximab monotherapy when combination therapy with an immunomodulator (IMM) is not an option due to patient preference or high risk of serious adverse events. Proactive TDM is also recommended after starting infliximab following a drug holiday (Figure 1).

Figure 1 | Other potential applications of proactive therapeutic drug monitoring in clinical practice. TDM: therapeutic drug monitoring; IFX: infliximab; TNF: tumor necrosis factor; IMM: immunomodulators. Refs (A) 31,56 and 60-62,Refs (B) 63-67, Refs (C) 68-71 and Refs (D) 72-79.

Optimal biologic drug concentrations to target can vary based on treatment phase, IBD phenotype, TDM assay used, targeted therapeutic outcome and route of drug administration (Figure 2). Most studies suggest that higher drug concentrations are needed to achieve more stringent therapeutic outcomes, including endoscopic and histologic healing. Recent data suggest that subcutaneous, compared to intravenous, administration of infliximab and vedolizumab produce multiple-fold higher serum drug concentrations due to PK differences.^{9, 10} Additionally, there may be discrepancies when measuring biologic drug concentrations among various assays, such as the enzyme-linked immunosorbent assay (ELISA), the homogenous mobility shift assay (HMSA) and point-of-care assays.^11-14 Importantly, higher drug concentrations are needed during the induction phase compared to the maintenance phase. Finally, higher drug concentrations are probably needed for patients with a more complicated phenotype, such as perianal fistulising CD.¹⁵ Consequently, applying the “one-size-fits-all” concept when performing TDM for optimising biologics is a mistake. A more personalised approach is needed.

Figure 2 | Factors associated with biologic drug concentrations threshold to target in therapeutic drug monitoring. TDM: therapeutic drug monitoring; HMSA: homogenous mobility shift assay; POC: point-of-care; ELISA: enzyme-linked immunosorbent assay; IBD: inflammatory bowel disease; CD: Crohn’s disease; UC: ulcerative colitis, ASUC: acute severe ulcerative colitis

Misinterpretation of ADA titers is a common mistake, mainly as titers are often described in different units across various assays, and these results cannot be directly compared (Table 1). A study assessing three commercially available ELISAs for the measurement of ATI showed that a clinically relevant cut-off titer of 200 ng/mL, previously associated with lack of response to treatment optimisation,¹⁶ when evaluated with the LISA-TRACKER assay (Theradiag) was equivalent to approximately 60 ng/mL on the RIDASCREEN assay (r-biopharm) and between 22.9 and 41 AU/mL on the Promonitor assay (Grifols).¹⁷ To make it even more complicated, these ADA titers have to be evaluated in the setting of a drug-tolerant versus a drug-sensitive assay, the latter of which can only measure ADA when drug concentrations are undetectable. As a result, physicians may wrongly interpret a result of being a high ADA titer and switch medications. If interpreted correctly, an attempt to overcome immunogenicity by dose optimisation and/or adding an IMM should be considered. It is critical to understand what high-level ADA are for each assay a provider may utilise. An association of antibody to infliximab titers evaluated with different assays with therapeutic outcomes in IBD is described in Table 1.

Table 1 | Association of antibody to infliximab titers evaluated with different assays with therapeutic outcomes 
in patients with IBD

Growing data suggest that proactive TDM can efficiently guide clinical decisions when anti-TNF therapy de-escalation is considered in patients with IBD due to clinical remission, including lengthening the dosing intervals, decreasing the dose, and stopping the IMM in case of combination therapy. The TAXIT RCT showed that dose reduction in patients with IBD and infliximab trough concentrations higher than 7 μg/mL was safe (no flares or increase of inflammatory markers) and cost-effective.¹⁸ A France study demonstrated that TDM-based infliximab de-escalation (drug concentrations higher than 7 μg/ml) in patients with IBD and clinical remission was associated with less relapse compared to empiric dose de-escalation based only on symptoms.¹⁹ The same group showed that proactive TDM is important after infliximab de-escalation to maintain an adequate trough concentration.²⁰ A study from Drobne and colleagues found that infliximab concentrations ≥5 μg/ml at the time of IMM withdrawal are related to long-term response in patients with CD after discontinuation of IMMs.²¹ Regarding adalimumab, concentrations higher than 12.2 µg/mL were associated with successful de-escalation in patients with IBD.²² We would like to point out that the ‘one-size-fits-all’ also should not apply when proactive TDM is used for treatment de-escalation. Higher drug concentration thresholds may be required for patients with a more complicated IBD phenotype, such as perianal fistulising CD. In the PRECISION RCT, three patients had a recurrence of an old perianal fistula after dosing de-escalation of infliximab based on proactive TDM using a PK dashboard for supposing supra-therapeutic drug concentrations of >3 μg/ml. As previously shown, supra-therapeutic infliximab concentrations are probably higher than 15-20 μg/ml for this IBD population.¹⁵ Furthermore, individual patients may require different drug concentration thresholds. 

TDM during induction may be even more critical than during the maintenance phase as patients typically have the active disease (with low albumin and high CRP levels) and increased drug clearance when an anti-TNF is initiated. 
High drug clearance puts patients at higher risk of early ADA formation.²³ A prospective study in UC showed that ATI could be developed as early as day 18 during induction therapy leading to treatment failure in patients with moderate to severe UC.²⁴ Numerous studies have found that higher biological drug concentrations during and early after induction therapy are associated with higher rates of favourable therapeutic outcomes.²⁵ A prospective study using a PK dashboard to guide infliximab dosing early during induction therapy proactively recognised the need for early accelerated infliximab dosing in 80% of patients who started on 5 mg/kg and 60% of patients who started on a 10 mg/kg dose.²⁶ Of note, adherence to the forecasts of the PK dashboard for the third, mainly the fourth, infliximab infusion was associated with higher treatment durability and decreased formation of ADAs.²⁶ A recent study showed that early treatment optimisation based on proactive TDM compared to standard induction infliximab therapy was associated with higher combined corticosteroid-free clinical and 
biomarker remission (CRP < 5 mg/L) at week 52 (83% vs 40%, respectively, p<0.001) in a paediatric population with IBD.²⁷ 

Identifying patients prone to develop ADA (Figure 3) is vital, as immunogenicity has been associated with treatment failure and drug discontinuation.²⁸ In this case, there is a need to use therapeutic strategies to prevent immunogenicity, such as combination therapy with an IMM or proactive TDM (especially in cases when combination therapy with IMM is not an option).^{29, 30} A recent meta-analysis showed that HLADQA1*05 variants were associated with increased risk of immunogenicity and SLR in patients with immune-mediated inflammatory disorders treated with anti-TNF therapy.³⁰ However, it seems that when proactive TDM is performed, the risk of immunogenicity and SLR is mitigated.^31-33 Proactive TDM could also be a valid therapeutic strategy to prevent ADA formation in paediatric patients with IBD and in patients with more severe diseases who typically have an increased drug clearance and a greater risk of inadequate drug exposure and immunogenicity including those with low albumin and high CRP levels. 

Figure 3 | Factors contributing to immunogenicity of anti-TNF therapy in IBD and other immune mediated  inflammatory disorders. BLI: β-lactamase inhibitors; TNF: tumor necrosis factor; ADA: anti-drug antibodies;  ESR: erythrocyte sedimentation rate; IMM: immunomodulators; WBC: white blood count; MTX: methotrexate.  Refs 28, 31, 33, 47 and 80-94.

Cumulative evidence suggests that PK dashboards incorporating factors such as type of IBD, type of drug, sex, CRP, albumin, weight, concomitant IMM use, previous drug concentrations, and anti-drug antibodies to individualise dosing can improve therapeutic outcomes.³⁴ The PRECISION trial showed that proactive TDM using a PK dashboard led to a higher rate of sustained clinical remission after one year of follow-up than conventional dosing (88% vs 64%, respectively, p=0.017).³⁵ In addition, patients in the proactive TDM group had lower faecal calprotectin levels compared to the control group (47 mg/g vs 144 mg/g, respectively, p=0.031).³⁵ In a study by Juncosa et al., the clinical remission rate increased from 65.7% to 80.4% after implementing PK dashboard-guided dose adjustments in patients with IBD treated with infliximab.³⁶ In another real-world cohort, Dubinsky et al. demonstrated that nonadherence to PK-driven infliximab dosing recommendations was a risk factor for immunogenicity and treatment discontinuation.²⁶

Preliminary data suggest that there may be quantitative and qualitative inconsistencies among different assays when evaluating drug concentrations, as previously shown between the ELISA and the HMSA for infliximab, adalimumab and ustekinumab^{12, 13} as well as the ELISA and point-of-care assays for adalimumab.¹⁴ This may also be the case for different commercial kits using the same quantification method. For example, significant differences were found among different ELISA commercial kits for both infliximab³⁷ and golimumab.¹¹ Even most importantly, ADAs are not easy to correctly interpret as titers are often expressed in different units across different assays, such as the ELISA, the HMSA and the electrochemiluminescence immunoassay (ECLIA) (Table 1). A study from Leuven showed that an infliximab ADA titer cut-off of 8 μg/ml evaluated with a first-generation ELISA had a similar impact as the cut-off of 374 ng/ml measured with the second-generation ELISA and a cut-off of 119 ng/ml in the ready-to-use ELISA kit.³⁸ Consequently, ADA levels cannot be directly compared among assays and thresholds for low and high titers cannot be adequately defined. Discrepancies among assays could lead to inappropriate clinical decisions as these often rely on drug concentration thresholds to target and ADA titer cut-offs that can be overcome. It would make sense to use the same assay for each patient, at least until harmonisation of assays and units of measurement is feasible.³⁹ We recommend that physicians be very comfortable interpreting ADA in their chosen assay.