Biologic therapies, including the anti–tumor necrosis factor-α and cell adhesion molecule inhibitor drugs, have revolutionized the treatment of moderate-to-severe inflammatory bowel disease. Since the introduction of anti–tumor necrosis factor therapies, the strategy of empiric dose-escalation, either increasing the dose or frequency of administration, has been used to recapture clinical response in inflammatory bowel disease. Disparate clinical outcomes have been linked to serum drug and antidrug antibody levels. Therapeutic drug monitoring has emerged as a framework for understanding and responding to the variability in clinical response and remission.
Key points
- •
Therapeutic drug monitoring (TDM) is an evidence-based strategy for managing biologic therapies in inflammatory bowel disease (IBD) after the loss of clinical response.
- •
Clinicians’ understanding of the pharmacokinetics of biologic therapy and the application of TDM is instrumental in the optimal management of moderate-to-severely active IBD.
- •
The newest assays for monitoring anti-TNF therapy consistently demonstrate that serum drug and antidrug antibodies (ADA) levels correlate with endoscopic and biomarkers of disease activity, and furthermore predict sustained remission versus later loss of response.
- •
More data are necessary to determine the cost-effectiveness and clinical use of proactive TDM to optimize the induction and maintenance phases of therapy.
Introduction
Biologic therapies, including the anti–tumor necrosis factor (TNF)-α and cell adhesion molecule inhibitor (CAMi) drugs, have revolutionized the treatment of moderate-to-severe inflammatory bowel disease (IBD). Monoclonal antibodies are designed to inhibit inflammatory pathways instrumental in the pathophysiology of IBD, but not all patients respond to biologic therapy and many lose response over time. Clinical trials report that primary nonresponse to anti-TNF therapies occurs in 20% to 40% of patients with IBD, and loss of response (LOR) to anti-TNF therapy ranges from 10% to 50% per year. Since the introduction of anti-TNF therapies, the strategy of empiric dose-escalation, either increasing the dose or frequency of administration, has been used to recapture clinical response in IBD. Clinical trials and series suggest that 60% to 90% will initially respond to dose-escalation, but 40% to 50% will again lose response to therapy over the next 12 months. Disparate clinical outcomes have been linked to serum drug and antidrug antibody (ADA) levels. Therapeutic drug monitoring (TDM) has emerged as a framework for understanding and responding to the variability in clinical response and remission.
Introduction
Biologic therapies, including the anti–tumor necrosis factor (TNF)-α and cell adhesion molecule inhibitor (CAMi) drugs, have revolutionized the treatment of moderate-to-severe inflammatory bowel disease (IBD). Monoclonal antibodies are designed to inhibit inflammatory pathways instrumental in the pathophysiology of IBD, but not all patients respond to biologic therapy and many lose response over time. Clinical trials report that primary nonresponse to anti-TNF therapies occurs in 20% to 40% of patients with IBD, and loss of response (LOR) to anti-TNF therapy ranges from 10% to 50% per year. Since the introduction of anti-TNF therapies, the strategy of empiric dose-escalation, either increasing the dose or frequency of administration, has been used to recapture clinical response in IBD. Clinical trials and series suggest that 60% to 90% will initially respond to dose-escalation, but 40% to 50% will again lose response to therapy over the next 12 months. Disparate clinical outcomes have been linked to serum drug and antidrug antibody (ADA) levels. Therapeutic drug monitoring (TDM) has emerged as a framework for understanding and responding to the variability in clinical response and remission.
Anti-TNF biologic agents
Pharmacokinetics
TNF-α is a proinflammatory cytokine that mediates gut recruitment of neutrophils, procoagulation, and fibrinolytic cascades, and promotes granuloma formation. Given the pivotal role of TNF-α in the pathogenesis of IBD, the available array of anti–TNF-α biologics continues to expand. Infliximab (IFX; Remicade; Janssen, Malvern, PA), the earliest and most well-studied TNF-α inhibitor, is an intravenously (IV) administered chimeric monoclonal antibody comprised of a human constant region IgG 1 κ light chain linked to a mouse variable region. IFX exhibits linear pharmacokinetics, with a direct relation between administered doses and the maximum serum concentration and area under the concentration-time curve. Conversely, volume of distribution at steady state and clearance are independent of dose. At standard dosing (5 mg/kg) the median half-life is 7.7 to 9.5 days. Steady state is achieved by week 14 of treatment, and maintenance dosing does not demonstrate systemic accumulation.
There are three subcutaneously (SC) administered anti-TNF biologic therapies for IBD: adalimumab (ADL) and certolizumab pegol (CZP) for Crohn disease (CD), and ADL and golimumab (GOL) for ulcerative colitis (UC). Similar to IFX, each demonstrates linear dose pharmacokinetics. ADL (Humira; Abbvie, Abbott Park, IL) is a recombinant IgG 1 antibody comprised of both a human constant region IgG 1 κ and human heavy and light chain variable regions. The drug distributes primarily in the extracellular space, with a bioavailability of 64% and half-life (T 1/2 ) of roughly 2 weeks. CZP (Cimzia; UCB, Smyrna, GA) is a humanized Fab’ fragment linked to polyethylene glycol resulting in high-affinity binding of TNF-α without an Fc region. It has a bioavailability of 80% after SC administration and T 1/2 of roughly 14 days. GOL (Simponi; Janssen, Radnor, PA), the latest anti-TNF biologic, is a fully human IgG 1 antibody that in vitro has demonstrated higher affinity for soluble human TNF-α than its predecessors, IFX and ADL. The reported median T 1/2 ranges from 7 to 20 days, with a volume of distribution that is body weight–dependent. Although all anti-TNF therapies exhibit linear pharmacokinetics, there is substantial interindividual variability in serum drug concentration across biologics, even with weight-based and IV dosing.
Assay technology
Traditional serum testing for anti-TNF drug and ADA concentrations has been performed with radioimmunoassay, solid-phase enzyme-linked immunosorbent assay, and bridging enzyme-linked immunosorbent assay technologies. The older assay technology has notable limitations. First, false-negative results for ADA presence and underestimation of ADA levels may occur in the setting of drug interference. Second, distinguishing specific ADA binding of TNF-α from nonspecific binding of serum IgG requires incubation with enzyme-labeled antidrug idiotype antibodies. The recent development of novel assays incorporating high-performance liquid chromatography and acid dissociation technology has enabled quantification of drug-complexed ADA in addition to “free” or excess ADA, thereby increasing overall diagnostic accuracy over prior solid-phase assays. Newer assays also demonstrate marked blocking of nonspecific binding serum IgG without the application of antidrug idiotype antibodies. Despite technologic advances, the absence of a universal assay across studies has been implicated in the discrepant findings regarding the association between ADAs and clinical outcomes.
Mechanisms of drug clearance
The mechanisms of anti-TNF drug clearance have not been fully elucidated, but there is increasing appreciation that disease severity accelerates drug clearance. Direct proteolytic catabolism of anti-TNF drug within the reticuloendothelial system (RES) is believed to be a primary route of drug clearance. Monoclonal antibodies cross-linked by antigen may undergo phagocytosis in the RES by receptors (Fc-γ) for the Fc portion of the IgG, with resultant lysosomal degradation. Increased protein catabolism within the RES is observed with increased systemic inflammation. In addition to clearance from systemic circulation, antibody clearance may also occur directly by membrane-bound antigen. TNF-α exists in soluble and membrane-bound form, with upregulation of serum and mucosal concentrations in active IBD. The concept of an “antigen sink” has been proposed, whereby circulating drug IgG becomes bound to membrane antigen, internalized into the cell, and degraded in lysosomes. Severe inflammation and diffuse ulcerative disease, marked by elevated C-reactive protein (CRP) and hypoalbuminemia, correlates with decreased serum drug concentration. A positive correlation between CRP levels and drug clearance has been demonstrated with IFX and ADL. Baseline serum hypoalbuminemia has been linked to low serum IFX concentrations in severe hospitalized UC and increased IFX clearance in CD. Albumin and IgG 1 antibodies may compete for RES-mediated drug clearance by FcRn receptors; thus, increased FcRn receptor binding and drug clearance may occur in the setting of hypoalbuminemia. Low albumin levels may also be a surrogate marker of high TNF-α and severe inflammation, a disease phenotype marked by extensive ulceration and massive weeping of protein and electrolytes. Disease phenotype, in addition to severity, has been implicated to mediate drug clearance. Whereas IV administered IFX and SC administered ADL exhibit comparable efficacy in CD, IV IFX induces higher remission rates than SC ADL in UC. Comparatively, a higher proportion of patients with UC than CD exhibit undetectable IFX trough levels. In a small study of patients with either severe CD or UC, elevated fecal IFX levels predicted low serum IFX levels. Thus, it is plausible that the pharmacokinetic differences observed between CD and UC may be conceptually explained by a phenotype characterized by extensive inflammation.
Disease-independent patient factors including age, body mass index, and genetics also affect drug clearance. In rheumatoid arthritis (RA), body weight has been positively associated with ADL clearance, whereas age has been negatively associated with ADL clearance. Increased body weight predicts earlier LOR to IFX in IBD, and the reduction of clinical efficacy seems more pronounced with ADL, which is not dosed by body weight. In a prospective study of patients with IBD treated with ADL, high body mass index predicted the need for dose escalation in a multivariate analysis. Further investigation is necessary to clarify whether obesity alters the bioavailability of anti-TNF therapy, or if increased mesenteric adiposity itself is a proinflammatory condition that augments TNF burden.
Intolerance to the anti-TNF biologic results in the development of ADA, which interferes with targeted antigen-binding and accelerates drug clearance by the RES. Factors that may affect immunogenicity include administration schedule, concomitant therapies, and individual pharmacogenetics. In patients with CD treated with IFX, the incidence of antibody formation has been reported between 36% and 61% with episodic therapy, and 5% and 18% with maintenance therapy. Immunotolerance to the monoclonal antibody seems to be promoted by scheduled infusions and maintenance of therapeutic drug concentrations. One prospective study of patients with CD treated with IFX suggests that low drug concentration after induction (IFX trough levels <2.5 μg/mL at 4 weeks postinitial infusion) predicts subsequent development of high-titer ADAs (positive predictive value, 86%). Fewer data are available regarding ADA formation with respect to the SC anti–TNF-α biologics, but available studies have reported a similar prevalence of ADA formation among patients with CD treated with ADL (9.2%–22%) and CZP (12%).
Concomitant immunomodulator (IMM) therapy with azathioprine or methotrexate may increase anti–TNF-α drug levels, presumably inhibiting ADA formation by altering RES clearance or impairing B-cell antibody development. A post hoc analysis of four randomized controlled trials administering maintenance IFX (5 mg/kg) with concomitant IMM therapy demonstrated a higher prevalence of ADAs with monotherapy (8%–23% vs 2%–7%). In the SONIC trial, ADA development was 15% with IFX monotherapy compared with less than 1% with concomitant azathioprine. In a recent 50-week randomized controlled trial of 126 patients with CD, IFX monotherapy, compared with IFX with concomitant methotrexate, was associated with an increased development of ADAs (20% vs 4%; P = .01) and a trend toward lower serum IFX trough (3.75 μg/mL vs 6.35 μg/mL; P = .08), but there was no difference in clinical remission between treatment groups. Lastly, in select cases, patient genotype may influence the development of ADA. Polymorphisms in interleukin-10 and the HLA-DR1 locus have been associated with a higher prevalence of ADA to ADL in patients with RA and IFX in patients with CD, respectively. Furthermore, polymorphisms of the rs1143634C allele of interleukin-1β and the TNF promoter region have been associated with nonresponse to IFX in CD and UC.
Serum Drug Concentration and Clinical Response
An abundance of literature has demonstrated that serum drug levels correlate with clinical response and endoscopic disease activity in CD and UC. In CD, high serum anti-TNF trough levels positively correlate with sustained response, fistula response, and mucosal healing, but low and undetectable levels are associated with elevated inflammatory markers, and predict LOR. Similarly, higher serum anti-TNF trough levels are positively correlated with clinical remission and mucosal healing in UC, but undetectable trough levels predict colectomy ( Table 1 ).
| Reference | Study Design | Subjects (N) | Anti–TNF-α | Follow-up | Clinical Impact of Serum Anti–TNF-α Drug Level |
|---|---|---|---|---|---|
| Crohn Disease | |||||
| Cornillie et al, 2014 | Post hoc analysis of ACCENT I | 147 | IFX | 54 wk | Week 14 serum TL >3.5 μg/mL associated with higher rates of remission |
| Fasanmade et al, 2003 | Subanalysis of ACCENT II | 282 | IFX | 30 wk | Higher serum TL correlated with complete fistula response |
| Bortlik et al, 2013 | Retrospective | 84 | IFX | 25 mo | Week 14 or 22 serum TL >3 μg/mL associated with sustained clinical response |
| Van Moerkercke et al, 2010 | Prospective | 210 | IFX | N/A | TL quartile positively predicted degree of mucosal healing (absence, partial, and complete) |
| Maser et al, 2006 | Retrospective | 90 | IFX | 23 mo | Detectable serum drug levels associated with remission, lower CRP, and endoscopic improvement |
| Colombel et al, 2010 | Prospective, SONIC | 338 | IFX | 46 wk | Higher serum TL correlated with clinical remission |
| Imaeda et al, 2014 | Prospective | 45 | IFX | 30 d | Higher serum TL required for mucosal healing (4 μg/mL) than for normalization of CRP, albumin, and fecal calprotectin |
| Li et al, 2010 | Subanalysis, CLASSIC I/II | 258 | ADL | 4, 24, 56 wk | Week 4 serum TL predicted clinical remission in CLASSIC I, but no dose-exposure-response relationship identified in CLASSIC II |
| Karmiris et al, 2009 | Prospective | 168 | ADL | 20.3 mo (median) | Low serum TL predicted LOR |
| Mazor et al, 2013 | Retrospective | 121 | ADL | N/A | Serum TL >5 μg/mL associated with higher clinical remission rates and normal CRP |
| Sandborn et al, 2012 | Subanalysis, PRECISE (open-label study) | 203 | CZP | 6 wk | Drug plasma concentrations positively correlated with clinical remission |
| Ulcerative Colitis | |||||
| Seow et al, 2010 | Prospective | 108 | IFX | 13.9 mo (median) | Detectable serum TL predicted clinical remission and endoscopic improvement; undetectable TL levels predicted colectomy |
| Roblin et al, 2014 | Cross-sectional | 40 CD/UC | ADL | — | Higher serum TL (median 6.5 μg/mL) associated with clinical remission and mucosal healing |
| Sandborn et al, 2014 | Prospective, PURSUIT | 625 (110) | GOL | 6 wk | Drug concentration quartile at week 6 positively predicted improvement in Mayo score and rates of clinical response and remission |
| Sandborn et al, 2014 | Prospective, PURSUIT | 157 | GOL | 30, 54 wk | Drug concentration quartile positively predicted higher rates of clinical remission |
Immunogenicity and Clinical Response
The formation of antibodies to anti-TNF biologics increases drug clearance and impairs binding of the target antigen, resulting in low or undetectable serum anti-TNF concentrations and clinical LOR. In a meta-analysis of 13 studies totaling 1378 patients with IBD treated with IFX, Nanda and colleagues determined a pooled risk ratio (RR) of 3.2 (95% confidence interval [CI], 2.0-4.9; P <.001) for loss of clinical response among patients with IBD with antibodies against IFX. Three of 13 studies (N = 243) also reported on serum IFX levels, which were uniformly lower in the presence of IFX ADA.
In patients with IBD treated with ADL maintenance therapy, multiple studies have demonstrated an inverse relation between the presence of detectable ADA (range, 9%–47%) and serum trough ADL concentrations, and a positive relation between ADAs and LOR. Yarur and colleagues reported a significant association between the presence of ADAs and lower serum ADL trough levels (odds ratio, 8.6; 95% CI, 2.3–31), higher CRP, mucosal inflammation on colonoscopy (odds ratio, 3.8; 95% CI, 1.1–13), and steroid use (odds ratio, 3.7; 95% CI, 1.1–13). Similarly, a prospective study of 113 patients with IBD (11 UC) by Velayos and colleagues reported lower median serum ADL concentration (4.78 μg/mL) among those with detectable ADA compared with those without (9.11 μg/mL). Moreover, detectable ADAs predicted worse clinical symptoms and higher CRP levels independent of serum ADL concentration. The presence of ADAs has also been shown to negate the effect of therapeutic IFX concentrations (>3 μg/mL) to control inflammation as measured by CRP level. Such findings suggest that significant immunogenicity disrupts drug efficacy beyond its direct effects on drug concentration. The clinical relevance of antibodies to CZP has been less studied, and a CZP assay is not commercially available. In the PRECISE 1 and 2 trials, ADAs against CZP was detected in 8% to 9% of patients with CD. The presence of ADA was mitigated by concurrent immunosuppressive therapy (4% vs 10%). Although ADA to CZP has not been demonstrated to decrease clinical efficacy in CD, the presence of ADA has been associated with decreased serum CZP levels and decreased clinical efficacy in RA randomized-controlled trials.
ADAs may also increase the risk for infusion reactions. In a systematic review by Chaparro and colleagues, the presence of ADAs was associated with a higher incidence of infusion reactions in six of eight studies. Among a consecutive cohort of 128 patients with IBD (105 CD, 23 UC) receiving IFX reinfusion after a median drug holiday of 15 months, 20% experienced an acute or delayed infusion reaction. Detectable ADA predicted the development of infusion reaction after reinfusion (hazard ratio, 7.7; 95% CI, 1.88–31.3; P = .004), but concomitant IMM therapy with retreatment attenuated ADA formation and predicted short-term response (hazard ratio, 6; 95% CI, 1.3–27; P = .019). In a prospective French trial, after discontinuation of IFX in patients with CD on IMM therapy, 39 (98%) of 40 patients who experienced relapse regained clinical response with retreatment, and none had detectable ADAs.
The presence of ADAs has not uniformly predicted LOR. In the pivotal ACCENT I trial, 67% of patients with detectable ADA still had a clinical response at week 54. Some studies have demonstrated successful dose escalation in the presence of ADAs. Heterogeneity in pharmacometrics across studies (assay technology, timing of measurement, unit of measurement, and cut-off level) may partially explain the inconsistency in the literature. LOR may be from noninflammatory causes of symptoms, there may be a lag-time between detection of ADA and later LOR, ADAs may be heterogeneous in their drug clearance and neutralizing effects, and alternative pathways of inflammation may emerge to circumvent anti-TNF therapy and thereby mitigate the relevance of ADA levels.
An emerging concept is that the clinical relevance of ADAs may be titer- and time-dependent. Although clinical response may occur in the setting of detectable ADA, higher median ADA titers significantly predict LOR. In a prospective study by Ungar and colleagues of 125 patients with IBD treated with IFX, 90% of patients who developed persistent ADA did so within the first 12 months of therapy (median, 22 weeks), and persistent ADA preceded clinical LOR by a median delay of 2 months. Transient ADA could be detectable at any time during the course of therapy and was without clinical relevance. In a retrospective study of 90 patients with IBD, Vande Casteele and colleagues reported detection of ADA at a median of 16 weeks after the start of IFX therapy (ie, after four infusions). Persistent ADAs exhibited significantly higher titer levels (median, 22 U/mL) than transient ADAs (median, 17 U/mL), and conferred a five times higher risk for discontinuation of IFX therapy because of LOR or hypersensitivity reaction compared with transient ADAs. ADA development was mitigated by concomitant IMM therapy; patients receiving IFX monotherapy during the first 6 months of treatment had an almost two-fold higher risk for ADA compared with those receiving combination therapy (RR, 1.8; 95% CI, 1.2–2.6; P = .0025).
Related posts:
Anti–Tumor Necrosis Factor-α Monotherapy Versus Combination Therapy with an Immunomodulator in IBD
Update on Anti-Tumor Necrosis Factor Agents in Crohn Disease
Pharmacokinetics of Biologics and the Role of Therapeutic Monitoring
Biologics of IBD
An Update on Anti-TNF Agents in Ulcerative Colitis
Update on Ustekinumab for the Treatment of Crohn’s Disease
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
