Department of Hepatogastroenterology, Hôpital Claude Huriez, Centre Hospitalier Universitaire de Lille, Lille, France, and Department of Gastroenterology, Mount Sinai School of Medicine, New York, New York, USA
Department of Gastroenterology, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1069, New York, NY 10029
Although tumor necrosis factor (TNF) antagonists are effective for the treatment of Crohn's disease and ulcerative colitis, lack and loss of clinical response is a clinical challenge. Accordingly, the use of therapeutic drug monitoring has been proposed as a means to optimize treatment. This article reviews the mechanisms of and factors which influence clearance of biologics, the relationship between serum drug concentrations and antidrug antibody presence and treatment efficacy, and identifies areas for future research needs regarding the use of therapeutic drug monitoring in clinical practice.
Publications regarding these topics were identified from literature searching and supplemented by review of gastroenterology meeting presentations and reference lists.
The clearance of monoclonal antibodies and pegylated antibody fragments is complex, and may be affected by demographic variables, concomitant medications, inflammatory burden, and immunogenicity, leading to high interpatient variability in plasma concentration of drug and clinical response. Several observational studies have demonstrated a relationship between anti-TNF agent serum drug concentrations and/or antidrug antibody presence and various symptomatic and objective clinical endpoints. However, these relationships are not absolute, and although some algorithms for the use of therapeutic drug monitoring in clinical practice have been proposed, none have yet been validated in a prospective clinical trial.
Further research to identify the most appropriate use of therapeutic drug monitoring is needed. (Inflamm Bowel Dis 2011;)
Tumor necrosis factor (TNF) antagonists are effective to induce and maintain response and remission in Crohn's disease (CD) patients.1 Approved agents for CD include the chimeric monoclonal IgG antibody infliximab (Remicade, Centocor, Malvern, PA) and the human monoclonal IgG antibody adalimumab (Humira, Abbott Laboratories, Abbott Park, IL).2, 3 Infliximab is also approved for the treatment of ulcerative colitis (UC).2 The pegylated humanized Fab' fragment certolizumab pegol (Cimzia, UCB, Smyrna, GA) is approved in the United States, Switzerland, and Russia for CD.4 The initial failure of some patients to respond and remit to anti-TNF therapy1 is a clinical challenge, as is the loss of response with continued treatment, with reported rates ranging from ≈10%–50% per year in CD.5
Because several analyses have implicated low serum anti-TNF concentrations (which are variably associated with the formation of antidrug antibodies [ADAs]) with lack or loss of response, some authors have recommended therapeutic drug monitoring (TDM, defined as of measurement of serum drug and/or ADA concentrations) to guide clinical decision-making for patients with inflammatory bowel disease (IBD).6, 7
The use of TDM for anti-TNF agents poses challenges, including assay standardization, interpretation of serum drug and/or ADA assay values, and validation of a treatment algorithm using TDM in clinical practice. This article reviews the pharmacokinetics of anti-TNF agents and evaluates existing evidence regarding the relationship between serum drug concentrations, ADAs, and therapeutic outcomes. The value of proposed treatment algorithms based on clinical and laboratory factors, including TDM, to guide individual patient management decisions is assessed.
PHARMACOKINETICS OF ANTI-TNF BIOLOGICS
Elimination half-lives for adalimumab, certolizumab pegol, and infliximab are shown in Table 1.2–4 All three agents demonstrate linear relationships between doses and serum concentrations.2–4 However, available pharmacokinetic data for adalimumab and infliximab demonstrate considerable interindividual variability in trough levels for a given dose, even when (in the case of infliximab) the dose is based on the patient's weight.8, 9 For example, the coefficient of variation (the standard deviation divided by the mean) calculated for adalimumab serum trough concentrations in the CLASSIC II study ranged from 52%–81%.8
Table 1. Half-life Values for Anti-TNF Agents Approved for IBD
Elimination mechanisms for monoclonal antibodies are not well described.10 One route may be proteolytic catabolism, but the exact locations of this have not been identified.11 The reticuloendothelial system (RES, i.e., the phagocytic cells of the immune system) may play a role (Fig. 1A); however, the importance of this pathway is incompletely understood and probably differs for different antibodies.10 Internalization of the antibody by phagocytic cells is primarily driven by Fc-γ receptors.10 Binding to and internalization by Fc-γ receptors is more efficient when the antibody is cross-linked, as when bound to a multivalent antigen.12
Binding to membrane-associated antigens, which leads to internalization of the antibody and subsequent degradation in lysosomes, is another potential route of clearance (Fig. 1B).10, 13 This pathway, also called the “antigen sink,” is saturable when sufficient amounts of therapeutic antibody are present.10, 13 TNF, a trimer which exists in both soluble and membrane-bound forms,14 is present in abnormally high concentrations in serum and gut mucosa in IBD.15, 16 Evidence of the influence of the contribution of the antigen sink has been demonstrated by analyses of the influence of C-reactive protein (CRP) concentrations on anti-TNF pharmacokinetics. CRP, an indirect marker of inflammation,17 correlates with perienteric findings of inflammation (increased fat density on computed tomography) in CD patients.18 Baseline CRP was significantly associated with infliximab clearance in UC patients.19 In rheumatoid arthritis (RA), pretreatment CRP levels negatively correlated with serum infliximab levels in patients treated with infliximab plus methotrexate,20 and CRP predicted minor increases in adalimumab clearance.3
Finally, neonatal Fc receptors (FcRn) also influence monoclonal antibody and IgG homeostasis and clearance. In adults, FcRn are expressed primarily in vascular endothelial cells or the RES, and function as a recycling or salvage mechanism to prolong IgG half-life. FcRn bind IgG bound to soluble antigens; the complex is internalized into the cell, the antigen is degraded, and the antibody is returned to the cell surface and released into the circulation (Fig. 1C).11, 13 In UC, baseline serum albumin concentrations (which may also be affected by FcRn-mediated recycling) correlated with infliximab clearance.21 As serum albumin has been shown to correlate with colonic endoscopic activity in IBD,22 low serum albumin may also reflect higher inflammatory burden in patients with active disease, which could influence biologic clearance via the antigen sink.
Fc receptor-mediated routes of elimination do not apply to certolizumab pegol, which lacks an Fc domain. Antibody fragments without the Fc domain are subject to renal clearance and proteolysis; the addition of polyethylene glycol increases the half-life to that approximating a full IgG1 molecule.23
In addition to its effect on clearance, antigen-biologic binding can influence the drug's pharmacokinetic profile.13 In the case of IBD, the amount TNF at the site of inflammation may vary between patients, and even within an individual patient over time, and likely influences the necessary dose of and potentially the minimum threshold concentration of an anti-TNF agent necessary for efficacy.24 TNF can also have concentration gradients in different compartments and binding of TNF can lead to subsequent redistribution, with further effects on the drug's pharmacokinetic profile.24
Other factors that may influence the clearance of biologics, including demographic variables, concomitant medications, and the presence of ADAs (which will be explored in more detail later), are shown in Table 2.3, 4, 19, 25–28 Concomitant drugs may influence clearance, possibly via an effect on Fc-γ receptor expression; for example, methotrexate may downregulate Fc-γ receptor expression on monocytes.11 Other potential mechanisms include modulation of the Fc-γ receptor-antibody interaction and reduction of immunogenicity development.13 The influence of thiopurines on clearance has been less well characterized; one study found little or no impact on adalimumab clearance,8 and another found no effect on adalimumab trough concentrations.29 In the SONIC trial, the combination of azathioprine and infliximab increased infliximab trough concentrations by more than 50% as compared with infliximab alone.30 In a study evaluating the effect of withdrawal of concomitant immunosuppressants (most of which were thiopurines) in patients on infliximab, patients who continued the immunosuppressant had higher median infliximab concentrations and were more likely to have detectable infliximab at trough than the discontinuation group.31
Table 2. Other Factors Influencing Anti-TNF Agent Clearance
• Low but significant association between body weight and adalimumab clearance (r2 = 0.14; P < 0.05) in RA
• Affected infliximab drug exposure in UC and CD
• Higher adalimumab clearance in men than women
• Age associated with lower clearance in patients over 40 years in RA
• Methotrexate reduced adalimumab clearance by 44%
• Methotrexate increased serum concentrations when used with infliximab in RA
• The effect of methotrexate on certolizumab pegol pharmacokinetics has not been studied
• Concurrent prednisone use reduced ADA formation in patients receiving episodic infliximab
Influence of Immunogenicity on Clearance
The development of ADAs, which have been demonstrated with all currently marketed biologics, can be influenced by several factors, including biologic structure, the patient's immune status, use of concomitant medications, route of administration, and dosing regimen.13 The influence of ADAs on the clearance of anti-TNF agents is complex and dynamic.32 Presence of immunogenicity against a therapeutic biologic may reduce its exposure through formation of immune complexes which accelerate RES-mediated clearance and/or impair binding.11, 13 The prescribing information for infliximab, adalimumab, and certolizumab pegol all note the association of ADAs with increased clearance.2–4 In a series of CD patients treated with episodic infliximab the concentration of ADAs was inversely correlated with infliximab concentrations measured 4 weeks postinfusion.33 Another CD study found that the presence of detectable ADAs to infliximab was associated with a 34% shorter half-life and 2.7-fold increased clearance.25 Patients treated with adalimumab for CD who developed ADAs demonstrated a lower median adalimumab trough serum concentration throughout the entire period of follow-up.29 Reductions in serum concentrations in patients who were positive for ADAs were seen with certolizumab pegol in the PRECiSE-2 and PRECiSE-4 trials.34, 35
OVERVIEW OF TECHNICAL ASPECTS
Current tests for anti-TNF and ADA concentrations are mostly based on enzyme immunoassays.6 Radioimmunoassay technology has also been developed.36 For existing commercially available assays for antibodies, the presence of persistent drug in the circulation generally interferes with the detection of ADAs, by some tests more than others.37 On the other hand, solid-phase enzyme-linked tests can also produce false-positive ADA results.37 Newer assays, based on high-performance liquid chromatography or the pH shift-anti-idiotype method, which can detect ADAs in the presence of circulating drug have been developed.38–40 These assays also can detect antibodies with weak affinities, eliminate nonspecific binding of irrelevant IgG, detect IgG4 ADAs, and overall have increased sensitivity and specificity compared to enzyme immunoassays.
Since assay methodology and sensitivity differ, comparisons between results obtained using different assays and/or assay techniques should be made with caution.
DRUG/ADA CONCENTRATIONS AND EFFICACY
The existing data regarding the relationship between anti-TNF serum concentrations and ADAs have limitations. Many of the analyses are retrospective. Often studies collected serum samples at predefined times, and not necessarily at the time a patient experienced a flare or lost response. Patients who fail treatment are likely to discontinue treatment, which could lead to bias if low concentrations were responsible for the lack of response. Often patients are not analyzed according to initial response status. Analysis of patients with disease that is not amenable to TNF blockade (at any dose) with treatment-responsive patients (in whom a concentration–response relationship might be more likely to be observed) could blur the ability to detect a link between concentrations and outcomes. Some studies evaluated clinical efficacy via only symptom-based endpoints, such as the Crohn's Disease Activity Index (CDAI), which are relatively subjective and could also be influenced by symptoms arising from disorders other than the underlying disease.41, 42 Finally, the follow-up of patients in studies is often limited; it is possible that there may be a time lag from the detection of low drug concentrations and/or ADAs and the development on clinical symptoms.
In this section we review data evaluating anti-TNF serum concentrations and ADAs and clinical outcomes. The focus is IBD, but selected studies in RA are included for additional evidence.
Clinical Impact of Serum Drug Concentrations
Several analyses have found numerically greater rates of attainment of efficacy endpoints in subjects with higher serum trough drug concentrations. An analysis from ACCENT I found that early treatment responders with higher infliximab trough concentrations had a greater CDAI decline than patients in lower concentration categories.43 In ACCENT II, patients with lower trough serum infliximab concentrations were less likely to achieve complete fistula response (25.0% in subjects with concentration <0.06 μg/mL vs. 64.3% with >10 μg/mL).44 In the SONIC study, week 30 rates of corticosteroid-free clinical remission were greater in patients with higher serum concentrations of infliximab at trough but were still high among patients with lower trough levels.30 Similar results were seen at week 46. An analysis of the COMMIT study found that patients with detectable infliximab concentration at trough were more likely to achieve treatment success than subjects with undetectable troughs, although this was not statistically significant (P = 0.08).45
In the previously mentioned infliximab immunosuppressant discontinuation study, CRP and CDAI correlated weakly with infliximab troughs (CRP: r = 0.387; CDAI: r = 0.205, P < 0.01), and subjects in the lowest quartile of infliximab troughs had significantly higher median CRP values.31 While patients in the lowest infliximab trough concentration quartile had the highest CDAI (P < 0.03), there was substantial overlap between quartiles. Low serum trough concentrations early after immunosuppressant discontinuation were also associated with more frequent infliximab dose intensification (P < 0.01).
An observational single-center study evaluated the relationship between infliximab trough concentrations and CD clinical outcomes.46 In 105 patients receiving scheduled or episodic infliximab, a positive relationship was seen between infliximab troughs and the percentage of time between infusions that patients had a Harvey–Bradshaw Index (HBI) ≤2 (R2 = 0.61; P < 0.001). Patients with detectable infliximab at trough were less likely to discontinue therapy before one year (odds ratio [OR] 23.1, P < 0.001). A subanalysis of 90 patients who received scheduled maintenance therapy for more than 52 weeks found strong relationships between detectable infliximab troughs and clinical remission, CRP, and endoscopic improvement (P < 0.001 for all parameters). Detectable infliximab at trough was a significant positive predictor for maintaining HBI ≤2 100% of the time between infusions (OR, 38.1; 95% confidence interval [CI] 9.1–60.5; P < 0.001)
In a small retrospective series of CD patients who had received at least two infliximab infusions, 95% (18/19) of patients who maintained response had detectable infliximab (as measured by TNF-alpha binding capacity) at trough, as compared to only 12.5% (1/8) of patients with secondary loss of response.36
An association between infliximab serum trough concentrations and endoscopic mucosal healing was seen in an analysis of 210 patients with CD. Patients with complete mucosal healing had the highest troughs (median 5.77 μg/mL, interquartile range [IQR] 1.05–10.72), followed by those with partial healing (3.89 IQR 0.35–8.28). Patients without healing had the lowest median troughs (0.95 μg/mL 0.35–6.56; P = 0.013 for the comparison of all three groups)47; however, there was significant overlap in concentrations between the groups. A limitation of this analysis is that “trough” sampling could have occurred over a 4-week span (from 2–6 weeks after dosing).
An observational analysis of 108 UC patients treated with scheduled infliximab maintenance therapy found detectable trough serum infliximab concentration to be a significant positive predictor for clinical remission (OR, 12.5; 95% CI: 4.6–33.9; P < 0.001) and endoscopic improvement (OR, 7.3; 95% CI: 2.9–18.4; P < 0.001).48 Undetectable trough serum infliximab concentration was a significant predictor for colectomy (OR, 9.3; 95% CI: 2.9–29.9; P < 0.001).
An analysis of adalimumab trough concentrations from the CLASSIC I and II studies found slightly higher mean concentrations in patients who achieved remission after induction and during maintenance treatment than those who did not.49 However, there was high variability in the trough serum concentration values and substantial overlap between concentration ranges between groups. Troughs achieved after induction dosing (week 4) were weakly correlated with remission (Spearman correlation = 0.173; P = 0.01), but no significant correlation was seen during maintenance therapy (Spearman correlation values ranged from 0.068–0.136; P-values all > 0.09). Analysis of remission status at different serum concentration threshold values (including undetectable concentrations) was not able to find a concentration that would distinguish patients by remission status.49
On the other hand, an observational cohort study in a Belgian referral center of 191 patients treated with adalimumab (168 of whom had previously failed infliximab) found no relationship between adalimumab serum concentrations or ADAs and short-term clinical outcome, although median adalimumab concentrations were lower in patients who discontinued therapy by 6 months than those who continued.29 Patients who responded to dose escalation had an increase in their adalimumab serum levels after dose escalation.
The mean troughs of patients treated with certolizumab pegol have been published.35, 50 In a study evaluating the efficacy of certolizumab pegol in patients previously treated with infliximab, serum drug concentrations were mentioned to not have a major influence on clinical responses or HBI.35, 51
Clinical Impact of ADAs
The clinical influence of ADAs was evaluated in a cohort of 125 CD patients in Leuven, Belgium, who were treated with episodic infliximab infusions.33 A high proportion (61%) of patients had detectable ADAs after the fifth infusion, and there was a negative relationship between ADA concentration and the duration of response (P < 0.001). While patients with infliximab concentrations (measured 4 weeks postinfusion) lower than the median concentration of 12.0 μg/mL had a shorter median duration of response (68.5 days) than patients with higher levels (81.5 days; P < 0.01), infliximab concentrations were not associated with duration of response by logistic regression (P = 0.70). Presence of ADAs was independently associated with a shorter duration of response (P < 0.001). The authors concluded that the development of ADAs correlates with a shorter duration of response as a result of their effect on infliximab concentrations.
In ACCENT I, episodic infliximab use was associated with a higher incidence of ADA formation than scheduled therapy (30% vs. 8%, P < 0.001) and higher ADA titers.52 In the pooled analysis of both methods of administration, ADAs did not affect week 54 clinical response and remission rates. In the episodic retreatment group, patients with early (week 14) ADA-positive assays had lower infliximab concentrations 2 weeks postinfusion, less improvement in CDAI, and shorter duration of response than patients without detectable ADAs. Despite this, 67% of ADA-positive patients had a clinical response at week 54. The proportion in clinical remission at week 54 was substantially lower (33%) in the ADA-positive subjects than the ADA-inconclusive group (73%, P = 0.078).
In an analysis of CD patients receiving infliximab, all patients (8/8) who lost response had ADAs present, whereas 53% (10/19) who maintained response had them, indicating that ADA presence is not always associated with lack of response. On the other hand, patients with loss of response had higher median ADA titers than the patients who maintained response.36 A study of CD patients treated with episodic infliximab found that ADA-positive patients with a titer <8 μg/mL had similar serum infliximab levels 4 weeks postinfusion as those without antibodies, whereas the infliximab concentration for patients with higher ADA titers was significantly lower than either of these two groups (P < 0.001).53 A retrospective, single-center study of 62 infliximab-treated patients with IBD found ADAs in 47% of patients.54 While 83% (10/12) of the patients with complete loss of response were ADA-positive, 22% (6/24) of treatment responders were also ADA-positive, although the titer was low in the majority (5/6) of patients.
A small (N = 30) retrospective single-center study of patients with CD treated with adalimumab found ADAs in 5 of 30 patients (17%); the presence of ADAs was associated with a lack of response to adalimumab (OR, 13.1, 95% CI: 1.7–99.2, P = 0.006).55
In the adalimumab CLASSIC II study, 7 of 269 (2.7%) patients tested positive for ADAs.56 The proportion of patients in clinical remission at week 56 was lower in the ADA-positive patients (29%), although the small number of patients precludes a definitive conclusion about the impact of ADAs to adalimumab.
A recent cohort study of 272 adalimumab-treated patients with RA followed for up to 3 years found that 28% developed ADAs (some as early as 4 weeks after starting therapy).57 Patients without ADAs had higher adalimumab concentrations (median, 12 mg/L; IQR, 9–16 mg/L) compared with patients with “intermediate” antibody titers (median, 5 mg/L; IQR, 3–9 mg/L) and those with “high” ADA titers (median, 0 mg/L; IQR, 0–3 mg/L). Patients with ADAs were more likely to discontinue treatment due to treatment failure (38% vs. 14% for ADA-negative subjects, P < 0.001), less likely to have minimal disease activity (48% vs. 13% for ADA-negative subjects), and less likely to achieve sustained minimal disease activity or sustained remission compared with ADA-negative patients.
In the pivotal studies for certolizumab pegol, ADAs were detected in 8% (26/331) of patients randomized to the drug arm in the PRECiSE 1 study58; a similar value (8%, 17/213) was seen in PRECiSE 2.50 Clinical response status through week 26 was similar in subjects regardless of ADA status.50
Available evidence suggests that there is a relationship between clinical efficacy endpoints (both objective and symptom-related) in IBD patients and anti-TNF trough serum concentrations, although the strength of the relationship varied and was not consistently statistically significant. For ADAs, available data suggest that their presence has a potential, but not absolute, negative impact on clinical outcomes, which is probably related to their influence on drug clearance.
USING TDM IN CLINICAL PRACTICE
TDM has been traditionally used for drugs such as anticonvulsants, antiarrhythmics, aminoglycoside antibiotics, and theophylline that possess a narrow therapeutic index (i.e., a small range between the minimum effective concentration and the maximum safe concentration).59 Criteria that have been suggested as necessary for serum concentration monitoring to be useful include difficulty in interpretation of clinical evidence of therapeutic or toxic effects, established relationship between drug concentration and therapeutic and/or toxic effect of the drug (i.e., “therapeutic window”), narrow therapeutic index, and a drug that lacks important active metabolites.60
Interpreting Serum Drug Concentration Values
While there are several commercially available tests to measure serum concentrations of anti-TNF agents (and ADAs), most manufacturers do not offer guidance for interpretation of the results. An exception is Prometheus Laboratories in the U.S., which offers serum concentration and ADA testing for infliximab. The Prometheus product data sheet61 references the Belgian study reviewed above33 and suggests sampling infliximab levels 4 weeks postinfusion, although the type of patient who should receive testing (i.e., stable patients or patients with limited or loss of response) is not specified. Patients with a serum level less than the median level seen in the Belgian study with episodic dosing (12 μg/mL) are suggested to have an anti-infliximab antibody test and suggestions are made to alter the infliximab dose or interval based on the ADA titer.
Two recent publications have proposed a different approach for patients who have ongoing inflammation and are not responding to therapy (Fig. 2). For patients with low drug serum concentrations who are negative for ADAs, increasing the anti-TNF dose is suggested. In the event that high drug concentrations are present in patients with documented inflammation, shifting to another class of drug is proposed (Fig. 2).6, 7
A retrospective evaluation of TDM following this algorithm from the Mayo Clinic was recently published.7 In this study a subtherapeutic infliximab concentration was defined as <12 μg/mL 4 weeks postinfusion or undetectable (<1.4 μg/mL) at trough. In patients with subtherapeutic drug concentrations (and no ADAs), increasing the infliximab dose achieved complete or partial response more frequently (25/29 of patients) than changing to another anti-TNF agent (2/6, P < 0.016). Ten patients with subtherapeutic concentrations continued on the same dose; information on their clinical course was not reported.
A potential limitation of the above algorithm is that the therapeutic range for anti-TNF agents may not be a “one size fits all” scenario. Many of the analyses cited above found increasingly higher rates of efficacy with higher concentration ranges, suggesting that some patients will only respond at a level higher than that sufficient for a different individual. For example, a retrospective analysis from France of CD patients who had infliximab serum concentrations and ADAs measured at the time of loss of response (although the treating physician was blinded to the test results) found that half of patients who had dosage intensification had a high (>1.5 μg/mL) infliximab trough level at inclusion and 70% of these patients had a clinical response after intensification.62 The median infliximab concentration in patients who responded to intensification was not different (3.2 μg/mL, n = 28) than those who did not respond (2.3 μg/mL, n = 12; P > 0.3). Four of the five patients who were moved to adalimumab responded; two of the responders had measurable infliximab trough concentrations (1.7 and 9.4 μg/mL) at the time of loss of response to infliximab. Another potential interpretation of these data are that the trough cutoff value to use to define patients in whom dose escalation may be of benefit may be higher than “undetectable.”
The referral center analysis of adalimumab use in CD found that patients who dose-escalated had a median adalimumab concentration of 4.8 μg/mL on every other week dosing; patients who responded to escalation to weekly dosing had a median increase of 5.9 μg/mL.29 An analysis of the effects of infliximab dose adjustment in patients with loss of response from the same center also found large interindividual variability in trough levels, suggesting that troughs needed to maintain remission may be patient-specific.63
Interpreting the Presence of ADAs
The TDM algorithm in Figure 2 proposes moving to a different anti-TNF in the setting of detectable ADAs and low serum drug concentrations, due to the suggestion that patients have been “immunized” to the drug and that subsequent dose escalation will not be beneficial.6, 7 In the Mayo series, 11/12 (92%) of patients with detectable ADAs achieved a complete or partial response to a different anti-TNF agent, whereas only 1/6 (17%) of patients with dose escalation responded (P < 0.004), and none of the patients managed with dose escalation regained infliximab concentrations defined as therapeutic.7 The titer of the ADAs was not reported.
On the other hand, the French study found that 6/10 patients with ADAs to infliximab responded to infliximab intensification, including 4/6 patients with elevated (>200 ng/mL) ADA concentrations.62 Seven patients had an ADA measurement after intensification; concentrations decreased in three patients and were no longer detectable in two.
A previously mentioned study evaluating the pharmacokinetics of infliximab in CD patients found that dosage intensification in two ADA-positive patients led to a reversal in infliximab clearance toward that in patients without ADAs. The authors suggest that the higher infliximab exposure after dosage increase binds and hence reduces the influence of the ADAs on drug clearance.25 Similarly, in an analysis of RA patients treated with adalimumab, five nonresponding ADA-positive patients moved to weekly dosing. Afterwards, the ADAs were no longer detectable and the mean adalimumab trough concentrations increased from 2.0 mg/L to 15.0 μg/L (P = 0.043).63 An update of patients treated at the same institution similarly found that 6 of 20 ADA-positive subjects who underwent dose escalation became ADA-negative; however, none of these patients achieved minimal disease activity.57
An analysis of 30 adalimumab-treated RA patients found ADAs in 21/30 of the subjects using a pH-shift-anti-idiotype antigen binding test (which is able to detect ADAs in the presence of adalimumab), whereas the traditional antigen binding test detected ADAs in only 5 of the same 30 patients.40 However, the presence of the “hidden” ADAs were associated with only a minor reduction in circulating adalimumab, indicating that the majority of patients who make anti-adalimumab antibodies do not make sufficient quantities to clear or neutralize all of the drug, or that perhaps the ADAs have low binding affinity or neutralizing capacity. The titer of ADAs may affect the degree to which they influence clearance; RA patients with “intermediate” concentrations of anti-adalimumab antibodies had similar clinical response as patients without detectable ADAs.64 The authors theorize that these patients have adequate serum drug concentrations for most of the dosing interval, with levels dropping below the threshold to allow ADA detection just prior to the next injection.
A further challenge with the issue of evaluating ADA status is that they are not always present in patients with low drug concentrations and lack of response.31, 46 The lower rate of responses seen in the presence of ADAs in some datasets may be more of a reflection of the fact that patients who have detectable ADAs with traditional assays have very low or undetectable circulating drug at the time of measurement. The summary conclusion is that the loss of response is often related to increased clearance; ADAs may not all influence clearance to the same degree and ADAs are not the only factors that can influence clearance.
USING TDM TO AVOID TOXICITY
There are few data regarding the use of TDM to prevent toxicity with anti-TNF agents. Large clinical trials of all three anti-TNF agents did not show meaningful differences in adverse reactions between the recommended initial maintenance doses and higher doses that were tested.51, 65, 66 On the other hand, a study of infliximab therapy in combination with methotrexate in RA found a higher relative risk versus placebo plus methotrexate (3.1; 95% CI: 1.2–7.9, P = 0.013) of serious infections with a regimen of three doses of 10 mg/kg followed by 10 mg/mg maintenance therapy administered every 8 weeks that seen with a 3 mg/kg regimen (relative risk [RR] 1.0, 95% CI: 0.3–3.1, P = 0.995).67 An analysis of skin-related adverse events during anti-TNF therapy for IBD found that patients experiencing an event were less likely to have undetectable infliximab concentrations at trough than those who did not.68
FUTURE RESEARCH NEEDS
While the concept of TDM is appealing, more research is needed to clarify its utility and to document its ability to improve patient outcomes. While the detection of ADAs or undetectable serum drug concentrations in a patient who is losing response should probably lead to a change in management, the optimal intervention (e.g., dose escalate, add an immunomodulator, or change to a different anti-TNF, or change to a drug with a different mechanism of action) remains to be determined. Additionally, studies designed to evaluate the concentration-effect relationship of anti-TNFs in order to define the “therapeutic window” are needed,25 this is also true for the concentration-toxicity relationship. Future evaluations should include assessment of the relationship with symptomatic as well as more objective measures of response (such as mucosal healing or change in CRP). The thresholds for serum drug and/or ADA concentrations should demonstrate an acceptable predictive value to reliably discriminate patients based on efficacy status. As seen from an analysis of RA patients treated in Sweden, a composite biomarker index (defined as the presence of undetectable infliximab concentrations and/or ADAs) with a strong association with loss of response (OR, 5.9, 95% CI: 1.3–26.6) was seen in only 42% of patients who lost response.69 The best approach for the remaining patients (all with detectable infliximab trough concentrations) remains to be determined in order to determine if dose escalation would be of benefit or if the patient should be moved to a different class of therapy altogether.
The best approach to TDM of biologic agents may not be only monitoring of trough concentrations. Other parameters of drug exposure (peak concentration, time above a minimum threshold, area under the curve, etc., all of which are affected by drug frequency and route of administration as well as the patient's individual clearance) could be related to outcomes and should be explored. It may also be that drug exposure variables associated with efficacy may vary depending on the point in time of treatment; for example, optimal pharmacokinetic parameters during induction therapy may differ from that those needed during maintenance therapy. Subcutaneously administered agents, which have more uniform concentration-time profiles as compared to intravenously administered agents, may have different application of TDM.
Optimal use of TDM might not be limited to one-time measurements in patients losing efficacy. In the Belgian referral center analysis, adalimumab-treated patients who ended up discontinuing therapy had significantly lower serum trough levels as early as weeks 2–4.29 Based on this, a potential approach for TDM could be modeled after the individualized dosing used for aminoglycoside antibiotics,70 wherein dosing is determined and adjusted based on the patient's specific clearance and the minimum inhibitory concentration of the organism being treated. In the case of anti-TNF agents, the dose could be changed or concomitant medications could be added or adjusted in patients with rapid or increasing clearance in order to boost concentrations, reduce immunogenicity, and potentially avoid loss of response. Adoption of this approach will first require determination and validation of the therapeutic range for the agent being used.
The pharmacokinetic profile of anti-TNF agents exhibits substantial interindividual variability. Anti-TNF pharmacokinetics can be affected by the use of concurrent medications such as methotrexate and thiopurines, which affect drug clearance and the development of immunogenicity. Current literature shows variable associations of the relationships between serum drug levels and ADA status and clinical efficacy. The concept of TDM is appealing to optimize therapy in order to help patients achieve long-term IBD control. Currently the use of TDM has not been prospectively validated and a therapeutic range for anti-TNF biologics has not been established. Well-designed, prospective, blinded, controlled trials to definitely answer the question of the utility and positive predictive value of use of TDM versus standard symptom-based approaches are needed.
This article was independently developed by the authors. Based on employment of one of the authors (A.M.R.), Abbott Laboratories was provided an opportunity to review and provide feedback, but the authors had the ultimate decision for final content.