Introduction

The traditional approach to medical therapy in IBD consisted of progressive escalation of medications as the disease progressed in severity, in order to induce and maintain clinical remission. With the advent of biologic medications and a greater understanding of the pathogenesis underlying IBD, this treatment paradigm subsequently shifted toward a new goal of reversing macroscopic inflammation in order to modify the natural course of IBD. A growing body of evidence suggests that gastroenterologists should strive for mucosal healing in addition to clinical remission. In this review, we aim to summarize recent studies that assessed strategies to enhance the initial response to, and maintain remission with, biologic agents for patients with IBD.

Currently Available Biologic Agents

Tumor necrosis factor (TNF) has a pivotal role in the mucosal inflammation that characterizes IBD [1], so it is no surprise that anti-TNF antibodies have revolutionized the treatment of luminal and fistulizing Crohn’s disease (CD) and severe ulcerative colitis (UC) [2]. Infliximab, adalimumab, and certolizumab are all FDA-approved for CD, and infliximab, adalimumab, and golimumab are all approved for UC. Typical rates of clinical response and remission to these drugs in IBD appear similar in clinical trials and observational studies, in addition to secondary outcomes like hospitalizations and quality of life [3–6].

The other approved class of biologic agents are the anti-integrins: natalizumab (humanized monoclonal antibody directed against alpha-4 integrin) and vedolizumab (humanized monoclonal antibody directed against alpha-4-beta-7 integrin) [7–9]. Use of natalizumab has been restricted due to the risk of JC virus-associated progressive multifocal leukoencephalopathy (PML), a severe demyelinating central nervous system disease [10]. Vedolizumab targets lymphocytes selective for the intestinal tract (α4β7), and no cases of PML have been reported with vedolizumab to date.

In addition, biologics in development for potential use in IBD include ustekinumab, etrolizumab, and anrukinzumab. Ustekinumab is a human IgG1k monoclonal antibody that blocks the biologic activity of IL-12 and IL-23 by inhibiting receptors for these cytokines on T cells, natural killer cells, and antigen-presenting cells [11] and has been shown to induce and maintain remission of CD [12]. Subcutaneous ustekinumab was specifically shown to be an effective therapy in CD patients who lost response to anti-TNF [13]. Ustekinumab is currently FDA-approved for the treated of psoriasis.

It should be noted that the vast majority of studies of optimizing biologics in practice have focused on infliximab, the pioneer biologic in IBD, and evidence with other drugs is only recently emerging. Consequently, most of the evidence on optimization of biologics is based on infliximab, with assumptions made on the likely effects in using biologics in general.

Evidence that Biologics May Improve the Natural History of IBD

A growing body of evidence suggests that the use of biologic agents may alter the natural history of IBD. This may be attributed to the ability of biologics to induce mucosal healing at greater rates than seen with steroids, for example, [14]. Anti-TNF therapy has been associated with reductions in hospitalizations, risk of surgery, and health care costs when compared to non-biologic approaches to IBD management. In the ACCENT I RCT, 11 out of 385 patients who received combined, scheduled infliximab treatment (5 or 10 mg/kg) underwent significantly fewer CD-related surgical procedures throughout the 54-week trial period when compared with 14/188 receiving episodic dosing [15]. In ACCENT II, patients with fistulizing Crohn’s disease on maintenance therapy 5 mg/kg every 8 weeks had significantly reduced hospitalizations, surgeries, and procedures compared with placebo maintenance [16]. In the CHARM RCT (where 778 CD patients were induced with adalimumab and then randomized to placebo, adalimumab 40 mg every other week or adalimumab 40 mg weekly) patients with CD treated with adalimumab had lower 1-year risks of hospitalization and surgery than placebo patients [17]. In the SONIC trial of 508 patients with moderate to severe CD, infliximab-based strategies were more effective than azathioprine alone in maintaining clinical and endoscopic remission [18••]. Observational studies have also confirmed these results; a French study of 296 patients with CD found that long-term anti-TNF treatment was associated with a lower risk for surgery, whereas azathioprine only modestly lowered this risk [19]. It thus appears likely that targeting inflammatory disease (rather than symptoms alone) may prevent irreversible damage and the complications that typically occur from transmural disease.

Limitations of Biologic Agents in Practice

Although anti-TNF targeting makes pathophysiological sense, the rates of clinical remission, endoscopic response, and mucosal healing achieved in clinical trials is less than ideal. Reported rates of clinical remission, a drug approval end-point, in UC are only 20–30 % after induction therapy [ULTRA, ACT] [20–22].

Similarly, mucosal healing rates consistently occur in the 30–60 % range, depending on the definition of “healing” [23]. Why do we not see ~100 % remission rates after induction therapy with anti-TNF agents in practice? In some cases, patients’ symptoms are not inflammatory in nature, despite an established diagnosis of IBD. Objective evidence of active macroscopic inflammation has not been an entry requirement for clinical trials in CD, allowing patients with overlap irritable bowel syndrome or chronic pain to enroll in these studies. In addition, TNF may be a primary driver of early inflammatory processes in IBD, but with chronic inflammation, the cytokine repertoire likely becomes diversified and less amenable to single-cytokine agents [24]. Finally, complications such as strictures, bacterial overgrowth, and visceral hypersensitivity can develop over many years of chronic inflammation, and these do not respond to anti-TNF agents.

In the long term, observational studies and maintenance trials highlight that even in patients who initially obtain remission with anti-TNF agents, many subsequently lose response over time. In clinical trials, only ~40 % of responder patients are in remission at 1 year after induction, and in clinical practice, less than 50 % of initial responders are in remission at this timepoint [25]. There are many reasons for this failure, and these are summarized in Table 1. The relative contribution of each cause is difficult to determine, but cohort studies suggest that immunogenicity and overlap functional symptoms account for most cases of “loss of response” [26•]. Immunogenicity refers to the propensity for patients to develop anti-drug antibodies (ADAs) against the monoclonal agents, which can accelerate drug clearance by the reticuloendothelial system (RES) [27]. Once they develop, ADAs are associated with a higher risk of loss of clinical response, and risk of infusions reactions, in this patient population [28–30].

Table 1 Reasons for biologic failure in patients with inflammatory bowel disease
Full size table

Cumulatively, these data suggest that optimal outcomes from biologic therapy may be influenced by patient selection, focusing on objective end-points, minimizing immunogenicity and ensuring adequate drug levels.

Optimizing Biologic Therapy

Given the benefits, and limitations, of biologic therapy noted above, a number of approaches have evolved to enhance the initial response and durability of this drug class in patients with IBD. These include earlier introduction in the disease course, intensified dosing schedules, concurrent immunomodulators, and therapeutic drug monitoring.

Early Initiation of Biologic Therapy—the “Top-Down” Approach

Epidemiological studies have highlighted that some phenotypes of IBD are associated with a greater risk of complicated disease over time. Patients with early onset disease, perianal disease, initial requirement for steroid use, early penetrating or structuring disease, severe endoscopic lesions, and those who smoke are all at risk of a complicated disease course of up to 60 % within 3 years of diagnosis [31, 32]. In these groups, the early introduction of biologics (top-down approach) has been recommended to increase the likelihood of achieving mucosal healing and reduce the risk of complications. Initial indirect evidence to support this came from post hoc analysis of the anti-TNF trials according to duration of disease; subjects less than 3 years from diagnosis had higher response and remission rates than those with more prolonged disease [33, 34].

Similarly, in a randomized trial of 133 newly diagnosed CD patients, those receiving biologics earlier in their disease course had higher mucosal healing rates compared to historical rates in those who received biologics later in established disease [35••]. This top-down approach of early infliximab and azathioprine makes sense in the high-risk patients noted above. In contrast, the 10–20 % of CD patients who achieve sustained clinical remission in the absence of maintenance therapy [36] do not warrant such an approach, given their lower risk of a complicated disease course.

Accelerated Induction for Acute Severe Colitis

The efficacy of infliximab (IFX) in UC was demonstrated in the ACT I (induction) and II (maintenance) placebo controlled trials, where standard dosing of 0-, 2-, and 6-week induction was used [22]. In practice, however, many patients hospitalized with UC have more severe disease and are steroid-refractory. Response rates in this setting are much lower than those seen in RCTs; in one study, only 50 % of patients had responded 4 weeks after receiving infliximab [37]

Severe ulcerative colitis is associated with higher levels of circulating TNF, higher CRP, and lower albumin; all factors associated with more rapid drug clearance and leakage of drug into stool [38–40]. Hypoalbuminemia increases IFX clearance, with subsequent diminished clinical response [41, 42•], and accelerated drug clearance also occurs with higher CRP levels [43]. Data from 728 patients with moderate to severe UC enrolled in the ACT I and ACT II trials were analyzed to assess if infliximab levels influenced clinical outcomes. Median serum concentrations of infliximab at weeks 8, 30, and/or 54 were significantly higher in patients with clinical response, mucosal healing, and/or clinical remission than in patients who failed to respond [44]. In light of this, some groups have advocated an “accelerated” dosage regimen in this patient population. Gibson et al. retrospectively analyzed 50 inpatients with UC receiving IFX for acute severe colitis and compared outcomes of standard dosing (n = 35) and accelerated dosing (n = 15) [45•]. Those in the accelerated arm had significantly lower rates of colectomy during induction therapy (1/15 % vs 14/35 p = 0.039). Factors independently associated with successful induction therapy in this study were albumin level at beginning of treatment (p = .003) and the accelerated dosing regimen (p = .03).

Concurrent Therapy

In the maintenance phase of biologic therapy, secondary loss of response has been associated with the development of anti-drug antibodies, a phenomenon that occurs less frequently in patients who are also taking immunomodulators such as azathioprine or methotrexate. This finding led to interest in the role of concurrent use of azathioprine or methotrexate with biologics to improve their durability. The Randomized Evaluation of an Algorithm for Crohn’s Treatment (REACT) trial studied early combined immunosuppression (with an anti-TNF and an immunomodulator) in 1084 CD patients compared to conventional step-up therapy in 898 CD patients attending 39 IBD centers. At 1 year, 15.1 % of patients in the early combined group received dual immunospupression, compared to 6.5 % in the conventional treatment arm (p < 0.001). Highly significant and clinically relevant differences in the rates of complications, surgeries, and the combined outcome of hospitalizations, complications, and surgeries were detected, favoring the use of early combined immunosuppression over 24 months [46]. Rates of surgery and hospitalization within a year of introduction of infliximab or immunomodulator were studied retrospectively in a cohort of greater than 20,000 veterans with IBD and found a 73.1 and 92 % reductions in risk of hospitalization and surgery, respectively, after 9 months of dual therapy. The authors concluded that patients are more likely to benefit from combined therapy if instituted early in their disease course [47]. In addition, two large randomized, double-blind, controlled trials have shown that CD and UC patients receiving IFX plus azathioprine (AZA) have significantly higher rates of corticosteroid-free clinical remission and mucosal healing when infliximab is administered with azathioprine, than when used alone [18••, 48, 49].

Further support came from cohort studies of patients treated with IFX which suggested that concurrent treatment with immunomodulators was associated with reduced IBD activity, IFX dose requirements and need for a switch to adalimumab [4, 50], and improved perianal fistula closure [51]. There is less data so far supporting the use of combined therapy with adalimumab. A modest benefit was seen in 207 CD patients treated with concurrent therapy compared to ADA monotherapy; a slight decrease in adalimumab failure rates and lower need for adalimumab dosage escalation were observed [52]. Data supporting concurrent therapy with methotrexate is less compelling than with thiopurines; the Combination Of Maintenance Methotrexate and Infliximab Trial (COMMIT) concluded that addition of methotrexate did not improve long-term efficacy beyond infliximab alone in patients with CD treated with prednisone [53].

Although there appear to be some efficacy benefits to concurrent therapy, both infections and lymphoma have been associated with use of thiopurines in these patients [54]. A rare and fatal hepatosplenic T cell lymphoma (HSTCL) was reported in 20 in young men receiving both IFX with a thiopurine [55]. A FDA study of 19 patients who developed HSTCL concluded that the risk was higher in patients treated with TNF-α inhibitor use in concurrent with thiopurines (95 % confidence interval (CI) 4.98–354.09; p < 0.0001) and thiopurines alone (95 % CI 8.32–945.38; p < 0.0001) but not with TNF-α inhibitor use alone (95 % CI 0.13–10.61; p = 1.00) [56, 57•]. In absolute terms, however, the risks of adverse events with concurrent therapy do not seem to be significant. In a long-term study of the safety of IFX in 6273 CD patients, serious infections were related to treatment with IFX, prednisone therapy, narcotic use, and moderate to severe disease activity, and increased risk of death was only associated with prednisone and narcotic use [58]. This study also reported that baseline age, disease duration, and smoking were independently associated with the risk of malignancy but that no increased risk was associated with immunosuppressive therapy alone, IFX therapy alone, nor their concurrent use [59]. Furthermore, a meta-analysis comparing safety and efficacy of concurrent thiopurine and anti-TNF therapy to anti-TNF monotherapy identified no increased risk of total adverse events, including infection and severe infection [49, 60].

Therapeutic Drug Monitoring

Monoclonal antibody therapies are cleared from the circulation by the action of proteases in the reticuloendothelial system [58]. Physiological factors within individuals that influence this process can lead to significant variations in serum drug levels between subjects receiving similar doses of anti-TNF therapy. Patients who exhibit lower trough levels of anti-TNF agents have a lower rate of clinical remission and mucosal healing than those with higher drug levels. Low drug levels have also been associated with a higher rate of anti-drug antibody development and infusion reactions. This link between serum drug levels and outcomes, which is presumed to be causative, has led to much interest in the role of therapeutic drug monitoring in clinical practice [61, 62].

A retrospective review of 247 IBD patients (42 UC) with suspected loss of response to biologics examined trough levels of drug and anti-drug antibodies (ADA). Trough levels of adalimumab >4.5 mcg/ml and infliximab >3.8 mcg/ml identified patients who failed to respond to an increase in drug dosage or a switch to another anti-TNF agent with 90 % specificity. Adequate trough levels identified patients who responded to expectant management or out-of-class interventions with >75 % specificity. Levels of antibodies against adalimumab >4 mcg/ml-eq. or antibodies against infliximab >9 mcg/ml-eq. identified patients who did not respond to increased drug dosage, with 90 % specificity [57•]. A retrospective analysis of the ACCENT I trial (maintenance infliximab for Crohn’s disease) found that patients who sustained response to infliximab 5 mg/kg had significantly higher trough levels of infliximab early in their course of treatment compared to those who lost response; furthermore, a greater than 60 % reduction in CRP at week 14 as well as baseline thiopurine use were also associated with a more durable response. The authors recommend an optimal post-induction level of ≥3.5 mg/mL and that infliximab levels, as well as change in CRP, could be used as a guide to titrate infliximab doses to obtain sustained clinical response [63].

A related observation is the impact of immunomodulators on serum infliximab levels. In a retrospective study of 223 patients, trough levels of infliximab and levels of ATI were measured in blood samples collected from 117 patients throughout concurrent therapy, as well as after withdrawal of immunomodulators. Patients on dual therapy had higher trough levels of infliximab than those on monotherapy (adjusted mean increase of 1.44-fold). Less patients on combined therapy developed ADAs than those on concurrent therapy (22 vs 38 %). However, in those on concurrent therapy, withdrawal of immunomodulators did not impact serum infliximab levels during the subsequent 6 months [64].

In view of these association studies, many groups have assessed the impact of dose escalation in patients with low drug levels and loss of response. Paul et al. prospectively studied the effect of therapeutic drug monitoring and mucosal healing in 52 patients (34 CD, 18 UC) on maintenance IFX treatment (5 mg/kg), with secondary loss of response. Patients with low drug levels had their dose of IFX intensified, and the effects on mucosal healing noted; mucosal healing was achieved in approximately 50 % of patients after dose intensification and was associated with an increase in trough infliximab levels [65]. Afif et al. similarly reported their results with altering drug dose or within-class switch in patients losing response to IFX; patients with no ADAs and low drug levels had higher probability of response to dose escalation than ADA+ patients [66]. More recently, a prospective trial was completed by Steenholdt et al. to compare empiric dose escalation to level-directed dose escalation in patients losing response to IFX. They concluded that this approach led to similar outcomes at less cost than simply dose escalating all secondary non-responders [26•]. It should be noted that the costs of testing IFX drug levels in Europe (20–50 euros) are dramatically less expensive than commercial tests in the US ($500–$2500).

These studies all addressed reactive testing to clinical non-response. Another approach is to proactively screen patients for low drug levels and dose escalate those who fall below a certain threshold. The Trough Level Adapted infliximab Treatment (TAXIT) trial prospectively studied drug level-tailored dosing of infliximab. During the optimization phase, better disease control was seen in patients with a level <3 μg/ml whose dose was intensified; furthermore, dose reduction in those with a level of >7 μg/ml was less costly and exposed the patients to less drug while maintaining disease control [67]. During the maintenance phase, no benefit was seen in adjusting dosing based on ongoing level measurement compared to dosing based on traditional clinical parameters in terms of maintenance of remission but was associated with less antibody formation [68]. This group has also assessed treatment with ADA in 720 patients with active CD. Thirty-four percent of patients required escalation to weekly dosing, and factors associated with dose escalation were prior anti-TNF use, no concomitant azathioprine, and abnormal CRP [69]. A retrospective study by our group reported that proactive therapeutic drug monitoring led to greater duration of time on IFX than reactive testing to loss of response [70•]

Limitations to therapeutic drug monitoring include a delay in obtaining the result, costs of commercial tests, and the lack of direct correlation between drug levels and clinical state [71, 72]. The development of point-of-care testing with less expensive kits should allow more timely modifications to anti-TNF treatment [73].

Switching Medications

In those who have failed to respond or lose response to their first anti-TNF agent, it is reasonable to switch to another drug within this class. Afif et al. noted their success rates in recovering clinical remission when adalimumab was used in IFX failures [66]. Thirteen patients had previously been treated with IFX, eight of whom were anti-drug antibody (ADA) positive. Increased efficacy was noted in the ADA+ group, with 63 % of patients responding to adalimumab compared to a 20 % rate in the ADA-negative patients. The authors postulated that patients who initially respond to IFX but then develop antibodies may be more likely to respond to adalimumab compared to patients who fail to respond to IFX for unknown reasons.

It should be noted that patients who have lost response to one anti-TNF are likely to develop antibodies against the second anti-TNF. Frederiksen et al. studied how immunogenicity to infliximab influenced antibody formation and clinical response to adalimumab. Patients who developed antibodies to IFX were significantly more likely to develop antibodies to adalimumab after switching medication for therapeutic failure. Anti-adalimumab abs were measured in 57 patients treated with adalimumab; 12 (21 %) were positive, 33 % of these had developed antibodies to IFX, conversely no patients who tested negative for IFX antibodies patients subsequently developed anti-adalimumab antibodies (OR 11). Furthermore, antibody positivity was associated with a undetectable adalimumab levels with subsequent treatment failure. This data supports dual therapy to optimize adalimumab and prevent treatment failure in those with previous anti-IFX antibodies when they switch to another anti-TNF [74].

Another common scenario is when a patient in remission is electively switched from IFX to a self-administered subcutaneously administered preparation for convenience. Van Assche et al. prospectively studied outcomes in 73 patients in clinical remission for at least 6 months on infliximab randomized to continue IFX therapy or switch to adalimumab. He found that elective switching from infliximab to adalimumab is associated with loss of tolerance and loss of efficacy within 1 year. Dose optimization or interruption of treatment occurred in 17/36 patients (47 %) in the adalimumab group and in 6/37 patients (16 %) in the infliximab group (p = 0.006). One patient interrupted infliximab treatment and ten patients interrupted adalimumab treatment (p = 0.003), mostly for loss of tolerance [75].

Conclusion

Optimizing biologic therapy in IBD involves patient selection, dosing strategies, monitoring of serum drug levels, concurrent immunomodulators, and realizing when the time is right to switch to a different drug. The best results are seen when biologics are introduced early in the disease course, continued as maintenance therapy with thiopurines until mucosal healing has been achieved, and adjusting dosing regimens to an individual’s pharmacokinetics. Future developments in IBD therapy will provide companion diagnostics to select patients for biologic therapy based on specific pathogenic pathways, point-of-care drug monitoring, and physician software tools to assist in achieving optimal drug levels.