Mentoring in IBD is an innovative and successful educational program for Canadian gastroenterologists that includes an annual national meeting, regional satellites in both official languages, www.mentoringinibd.com, an educational newsletter series, and regular electronic communications answering key clinical questions with new research. This issue is based on the presentation made by the contributing editor, Dr Brian Feagan, at the annual national meeting, Mentoring in IBD XIX: The Master Class, held November 2, 2018 in Toronto, Ontario.
The TNF antagonist revolutionized the care of inflammatory bowel disease (IBD); however, even with these highly effective medications, care of our IBD patients is a long way from optimal. Accordingly, medical science continues to deliver new medications, both small molecules and biologics, to expand the armamentarium and provide patients with more effective choices. This review will discuss: the limitations of the currently available biologics for IBD; review the next-generation biologics that are in development in IBD; and speculate on treatment paradigms that we might see in the future.
While the TNF antagonists were a true game changer for IBD therapy, analysis of key trial data clearly shows that these are still not ideal agents. For example, in the pivotal CHARM trial that evaluated adalimumab maintenance therapy for patients with Crohn’s disease (CD), after 56 weeks, the response rate was 41% for patients assigned to adalimumab every other week and 48% for the 40 mg weekly regimen.(1) Corresponding remission rates were 36% and 41%.(1) This circumstance leaves substantially more than half of treated patients needing alternative approaches. Even with relatively early initial combination of a TNF antagonist and a thiopurine that was evaluated with infliximab and azathioprine in the SONIC study in CD, a substantial proportion of patients did not achieve corticosteroid-free clinical remission.(2) For the primary analysis at week 26, this endpoint was achieved by 57% of patients.
An additional limitation of current biologics are the high rates of IBD complications associated with their use due to incomplete control of disease activity and drug-related adverse effects. In the REACT trial, in which CD patients received early, combined immunosuppression with a TNF antagonist and a thiopurine, the two-year rates of the combined endpoint of first hospitalization, surgery or therapy-related complication was 27.4% for this approach.(3) Similarly, in the CALM study, where a treat-to-target approach guided by biomarkers was evaluated and found to be more effective than symptom-based therapy, the majority of patients in the biomarker group did not achieve the optimal endpoint of CDEIS <4 and no deep ulcerations.(4)
An important limitation of TNF-antagonist-based treatment algorithms is an increased risk of serious infections. In the TREAT registry, the incidence of serious infections per 100 patient-years was 1.69 among infliximab-treated patients and 0.69 for patients who were not exposed.(5) Additional statistically significant independent risk factors for serious infection in this cohort were use of prednisone, use of narcotics and increased disease activity.(5) Thus, clinicians should be aware of these risks and adjust their management accordingly to minimize morbidity from infection.
Vedolizumab, a monoclonal antibody directed towards the α4β7 integrin inhibitor, selectively inhibits lymphocyte trafficking to the gut.(6) The efficacy of vedolizumab has been well documented in both CD and ulcerative colitis (UC).(7,8) Vedolizumab is a welcome addition to our therapeutic armamentarium, given that it has been shown effective in both biologic-naïve and anti-TNF-experienced patients. However, similar to the TNF antagonists, a substantial proportion of patients will not respond to treatment or will not achieve remission. For example, in the pivotal GEMINI 1 trial, among UC patients with no prior anti-TNF exposure, the proportion in clinical remission at week 52 was 45.8% with vedolizumab every 8 weeks and 47.9% with vedolizumab every four weeks. Corresponding rates for patients with previous TNF antagonist exposure were 36.0% and 40.4%.(9) The most relevant advance provided by vedolizumab is an unrivalled safety profile with a lower risk of serious infections than systemic immunosuppressives. Remarkably, an analysis based upon randomized controlled trial data (2830 IBD patients with 4811 patient-years of vedolizumab exposure) showed that vedolizumab therapy was actually associated with a reduced risk of infection compared with placebo.(10) This finding can be explained by the possibility that if there is no attributable risk from active drug, placebo patients could be expected to have a greater likelihood of infection due to a relatively greater prevalence of high disease activity and increased corticosteroid exposure—known risk factors for infection.(10)
Vedolizumab is currently available as an intravenous (i.v.) infusion.(11) However, an investigational subcutaneous (s.c.) formulation has been compared to the i.v. formulation in a randomized controlled trial performed in patients with active UC (VISIBLE 1).(12) At week 52 the rates of clinical remission were similar with both formulations (42.6% in the i.v. every 8 weeks group and 46.2% in the s.c. every 2 weeks group, P=NS; Figure 1) and both formulations were statistically superior to placebo (clinical remission rate 14.3%).(12) Notably, these remission rates were similar to those reported in the original pivotal trial (GEMINI 1), where the rates were 41.8%, 44.8% and 15.9% for the i.v. formulation administered every 8 weeks, the formulation administered every 4 weeks and placebo, respectively (Figure 2).(8)
This agent, which is a monoclonal antibody directed towards the p40 subunit of IL-12 and IL-23, has been used for many years for the treatment of psoriasis and psoriatic arthritis.(13) More recently, ustekinumab has been shown to be effective in a broad population of patients with CD in the UNITI-1 and UNITI-2 studies.(14) UNITI-1 included only patients who had failed a TNF antagonist (primary or secondary nonresponse or unacceptable side effects), while UNITI-2 included patients who had failed conventional agents. Induction consisted of a single i.v. dose of ustekinumab (130 mg or a higher dose that approximated 6 mg/kg of body weight) or placebo. The proportions of patients who responded at week 6 in UNITI-1 were 34.3% for ustekinumab 130 mg, 33.7% for ustekinumab ~6 mg/kg and 21.5% for placebo (P≤0.003 for both comparisons with placebo). In UNITI-2, the corresponding rates were 51.7%, 55.5% and 28.7% (P<0.001 for both doses vs. placebo).(14) All responders from these studies were then re-randomized to receive maintenance with ustekinumab 90 mg s.c. every 8 weeks, every 12 weeks or placebo. After 44 weeks, clinical remission rates were 53.1% for ustekinumab every 8 weeks, 48.8% for ustekinumab every 12 weeks and 35.9% for placebo.(14) These findings were the basis for the approval of ustekinumab therapy for CD with a recommended dose of 45 mg administered at weeks 0 and 4, then every 12 weeks thereafter. Alternatively, 90 mg may be used in patients with a body weight greater than 100 kg.(15)
In UC, ustekinumab is also being investigated in a phase 3 study (UNIFI).(16) Patients in this study were randomized to receive a single i.v. induction dose of ustekinumab 130 mg or ~6 mg/kg or placebo. At week 8, clinical response was seen in 51% of those treated with ustekinumab 130 mg, 62% of those who received ~6 mg/kg and 31% assigned to placebo.(16) Responders were re-randomized to either placebo or active drug in the maintenance phase of the trial, results of which are not yet available.
While use of this agent in IBD is not yet as extensive as TNF antagonists or vedolizumab, ustekinumab has extensive safety experience in psoriasis, where it is well established as a safe agent, with no signal of increased risk of malignancy, major adverse cardiovascular events, serious infections or mortality compared to patients not treated with biologics.(13) This observation is somewhat surprising given the mechanism of action of the drug that is based upon suppression of both the TH1 and TH2 pathways. Further investigations on the mechanism of action of ustekinumab in IBD are needed.
Notwithstanding access issues, it is likely that TNF antagonists will evolve to second-line biologics within the next five years because of the more favourable safety profiles of the other options. It is likely that vedolizumab will become the preferred first biologic choice for UC, while ustekinumab may take that position for patients with CD.
In the interim, it is noteworthy that treatment targets may need to be re-calibrated—downwards in CD, where the current STRIDE recommendation incorporates the difficult-to-attain goal of absence of large ulceration and upwards in UC, where histopathology may be an aspirational treatment target beyond the recommendation for reaching a Mayo endoscopic sub-scale score of 0 or 1.
Furthermore, an unequivocal conclusion is that we still have low absolute corticosteroid-free remission rates with our best IBD therapies, and payors are exhausted from adding expensive treatments with small incremental benefits. We need new solutions that are both clinically effective and cost-effective.
Anti-integrin biologics modulate leukocyte migration to the gut mucosa. This process is primarily mediated by interactions between mucosal addressin cell adhesion molecule (MAdCAM-1) and α4β7.(17) However, lymphocyte retention in the epithelium is primarily mediated by another integrin-based interaction between E-cadherin, a protein present on the basement membrane of epithelial surfaces and αEβ7, an integrin on the surface of approximately 1–2% of lymphocytes in the peripheral circulation.(17)
Etrolizumab is a humanized monoclonal antibody with high affinity for the β7 subunit of both the α4β7 and the αEβ7 integrins that might convey relatively greater efficacy than vedolizumab because of the dual mechanism of action.(18) In a phase 2 study in UC, etrolizumab 100 mg s.c. every four weeks was associated with a 43.8% clinical remission rate among TNF-antagonist-naïve patients at week 10 compared to a rate of 0% with placebo (P<0.01). Among those with prior TNF antagonist failure, the remission rate was much lower and not significantly different from placebo.(18) Multiple phase 3 studies in TNF-antagonist-naïve and -experienced patients with UC are in progress, including a direct comparison to infliximab.(19-23) A phase 3 development program is also underway in CD.(24)
Given that that ustekinumab blocks the p40 subunit of both IL-12 and IL-23 a critical question has evolved as to which effect is responsible for efficacy. In this respect, several agents that block only the p19 subunit of IL-23 have been developed (e.g., brazikumab, mirikizumab, risankizumab) for the treatment of inflammatory diseases. In psoriasis these agents have been shown to be consistently superior to ustekinumab, suggesting that inhibition of the TH17 pathway through blocking P19\interleukin-23 is the dominant mechanism. In CD, both brazikumab and risankizumab have shown promising results in phase 2 studies.(25,26) In the risankizumab study, clinical remission rates were as high as 36.6% at week 12 (compared to 15.0% with placebo, P<0.05), while endoscopic remission at the same time point was as high as 19.5% compared to 2.6% with placebo (P<0.05).(25) Brazikumab demonstrated efficacy (49.2% clinical response at week 12 vs. 26.7% with placebo; P=0.01) among CD patients who had previously failed anti-TNF therapy.(26)
In a phase 2 study in UC, mirikizumab therapy was associated with as high as 22.6% clinical remission rate at week 12 (vs. 4.8% with placebo, P=0.004), including a 36.4% rate among biologic-naïve subjects (vs. 8.7% with placebo, P=0.035) and a 15.0% rate in those with prior biologic experience (vs. 2.5% with placebo, P=0.108).(27)
In addition to novel biologic therapies, the future treatment landscape in IBD may also include oral peptides. As an example , PTG-100 is an oral, GI-restricted α4β7 integrin antagonist peptide that works by binding to the MAdCAM-1 protein in the gastrointestinal tract. A phase 2 study with this agent was terminated early due to failure to meet the primary endpoint of clinical remission at an interim analysis.(28) However, further analysis of the data did show dose-dependent decreases in endoscopic and clinical activity.(28) It is not clear whether development of this agent will continue, but the encouraging findings suggest that the oral peptide approach is one to continue investigating.
The paradigm of care that is most well known in IBD involves a step-wise approach, where patients are treated early with less potent agents (e.g., aminosalicylates), relying on corticosteroids for induction, then perhaps a thiopurine for maintenance before stepping up to a biologic with or without another agent in combination. Factors that determine rate of progression though this algorithm include disease severity at presentation and response to initial therapy as defined by improvement in symptoms.
In recent years, there has been a movement towards a treat-to-target paradigm, including the published recommendations of the STRIDE group, which include both clinical and endoscopic targets for CD and UC.(29) In addition, this group endorses the use of biomarker targets as adjunctive measures for monitoring IBD. Furthermore, the results of clinical trials comparing algorithmic approaches to management have confirmed the value of using both symptom-based (REACT)(3) and biomarker-based targets (CALM).(4) Evolution has also occurred in the regulatory landscape with the U.S. FDA endorsing both patient-reported outcomes and objective measures of disease activity (e.g., endoscopy) as preferred primary endpoints of clinical trials.
Despite some of the advances that have been outlined, it is clear that in the current treatment environment, many patients with IBD will fail to achieve an optimum treatment target. Based upon this reality, what opportunities are most promising to change this circumstance? Two potential solutions are part of the way forward. The first is use of combination therapy to take advantage of the complementary mechanisms of action that exist with the best available and experimental therapies. The SONIC study showed the way forward by demonstrating the value of combined therapy with two established agents. Accordingly, even greater benefits may result from combining some of the highly promising investigational drugs. A parallel model that has proven to be very successful in another therapeutic area is combined antiretroviral therapy for hepatitis C. In that field, multiple agents with complementary mechanisms of action are combined to optimally target the virus, and cure rates approaching 100% in clinical trials are now the expectation.(30)
While we don’t currently have evidence to suggest that combination of any of our currently available IBD therapies could lead to 100% remission rates, it is certainly desirable to try to determine which advanced agents might have complementary effects when used in combination.
The second possibility is the use of clinical prediction rules to identify high-risk patients for early treatment. This is an area of active research; several clinical prediction models have been developed to aid in patient selection. Using data from the GEMINI-2 study of vedolizumab in CD (and validated with data from the VICTORY consortium clinical practice cohort), Dulai and colleagues developed a prediction score that stratifies patients into low or high probability of response to vedolizumab (Table 1).(31)
Despite important advances in IBD therapeutics, many of our patients still do not achieve the aspirational goal of durable clinical and endoscopic corticosteroid-free remission. Although many promising therapies are in development, including novel biologics and oral peptides, we should not expect that any of these therapies will be universally successful in achieving this goal.
The path forward includes optimization of our current therapies through the development of agent-specific clinical prediction tools. In addition, we must continue to explore combination therapy for IBD, such that most patients will have a high likelihood of achieving durable and deep remission.
Carmela is a 72-year-old widow diagnosed five years ago with colonic and perianal fistulizing CD, complicated by small pelvic abscess. Other medical history includes breast cancer at age 52, treated with surgery and radiation. For her CD, she was initially treated with antibiotics and surgical drainage, followed by infliximab monotherapy. A small squamous cell carcinoma was excised from her cheek shortly thereafter. Around the same time, her twin sister was diagnosed with lymphoma. After one year of treatment, she was asymptomatic with no perianal pain or discharge. Her CRP and fecal calprotectin were normal. At that point, infliximab treatment was withdrawn because of multiple concerns related to an increased risk of malignancy.
Repeat colonoscopy one year ago showed scarring throughout the colon with scattered inflammatory polyps but no overt disease activity. Segmental biopsies revealed chronic inflammation with deep plasma cells and neutrophils. You and she are concerned about her prognosis off biologic therapy.
Discontinuation of therapy remains a source of great controversy in IBD therapeutics. Research has shown that discontinuation of TNF-antagonist therapy among CD patients in remission conveys a high risk of clinical relapse. Published studies have reported 1-year relapse rates from approximately 20% to 50%.(32-36). Thus discontinuation of treatment in this patient is likely to result in recurrent disease. This is especially problematic because she is a very high-risk patient who has responded nicely to infliximab therapy.
With respect to the history of malignancy (squamous cell carcinoma), a discussion with the patient is warranted to discuss the pros and cons of re-initiating infliximab therapy. It should be recognized that interrupted therapy in the setting of infliximab monotherapy has an approximate 20% risk of sensitization with formation of anti-drug antibodies. Furthermore, the potential role of TNF antagonism in malignancy should be considered.
Current data indicates that for most malignancies, there is no increased risk of cancer associated with use of TNF antagonists.(37,38)
For example, long-term registry data with infliximab (the TREAT registry) showed that the global incidence of neoplasia (malignant, benign, and unspecified) was similar between patients receiving infliximab and the other treatments with no differences seen for solid tumours, non-melanoma skin cancers, or lymphoma.(38)
In contrast, some studies have reported an increased risk of melanoma and lymphoma with TNF antagonists.(39,40)
When used in combination with a thiopurine, there are well established elevated risks of lymphoma and skin cancer due to these agents.(41) This potential risk might have led to the initial decision to use infliximab monotherapy in this patient.
The biopsy findings are important to consider. Despite clinical remission, the tissue biopsies document continuing inflammatory disease activity. Limited data suggest that histopathologic assessment may provide important prognostic information beyond what is provided by endoscopic assessment.(42) Specifically, increased eosinophils or neutrophils in the lamina propria and cryptitis are strongly associated with an increased risk of clinical flare within 1 to 2 years.(40)
Carmela did not resume therapy after colonoscopy. She now presents to your office with increasing symptoms over the last six months. She is passing five stools per day (up from two) with moderate urgency. She has also noted scant discharge from fistula tract on alternate days. Examination shows the fistula to be open, but without any signs of local sepsis. Her CRP is 6.3 mg/L. Blood work is otherwise normal.
The increase in clinical disease activity is an indication for further investigation and, potentially, resumption of biologic therapy. Given that the first biologic (infliximab) was not a failure of efficacy—indeed, the patient was in clinical remission—it would be reasonable to restart infliximab at the same dose.
A meta-analysis reported that among patients who had achieved remission on a TNF antagonist and subsequently relapsed following withdrawal of the medication, retreatment with the same agent led to recapture of remission in 80% of patients.(36) However, a cautionary note regarding these studies is that very few of them used objective measures such as ileocolonoscopy to document remission status. As noted previously, re-introduction of infliximab could lead to sensitization.
Use of adalimumab might also be considered, which would mitigate the risk of sensitization but not the potential for increased risk of cancer. Ustekinumab or vedolizumab could also be considered for this patient.
Importantly this patient should be re-staged before re-starting therapy with both ileocolonoscopy and pelvic MRI.
Her fecal calprotectin concentration is elevated at 876 mcg/g. Colonoscopy shows linear ulceration in the rectosigmoid. An MRI of the pelvis shows a single trans-sphincteric fistula with no abscess.
As discussed previously, because this patient previously responded to TNF-antagonist therapy, one could resume the infliximab or switch to adalimumab, depending on patient preferences and in consideration of the approximately 20% risk of sensitization. The presence of perianal disease is also a factor that might favor use of a TNF antagonist.(43)
With respect to biologics with different mechanisms of action (i.e., the integrin inhibitor vedolizumab or the anti-IL-12/-23 agent ustekinumab), there are limited data to guide decisions in this scenario. However, among patients who have failed prior therapy, there is some evidence that suggests that ustekinumab is associated with better rates of response and remission relative to other biologics.(44) To date, ustekinumab has not been associated with increased risk of serious infection, malignancy or mortality. Again, however, the extent of safety information with this agent is limited largely to the clinical trial experience in IBD.(45) Note that in psoriasis, ustekinumab been used extensively and continues to have a favorable safety profile over the long term.(46)
Vedolizumab may theoretically be preferable to other biologics in a patient with a history of non-GI malignancy, as its activity is restricted to the gut. Furthermore, the patient is older and has a greater risk of infection than average on that basis.(47)
After extensive discussion, Carmela chooses to start ustekinumab as monotherapy. She receives an intravenous induction dose of 390 mg and then three subcutaneous maintenance doses of 90 mg before you see her in follow-up. She reports a slight reduction in stool frequency but no change in perianal discharge. She tells you that her dermatologist recently treated a number of actinic keratoses on her back with liquid nitrogen.
Just before her last dose, her CRP was 5.8 mg/L and fecal calprotectin level was 957 mcg/g. She also had therapeutic drug monitoring (TDM), with a ustekinumab trough level of 4.5 mcg/mL. Anti-drug antibodies were not tested.
At this point, the only additional investigation one might consider is TDM. However, there is little available guidance for clinicians regarding the use of TDM with this agent in CD.(48)
The largest source of data from the ustekinumab pivotal trials indicates a clear exposure-response association, although other authors have published conflicting findings.(49,50) Based on the clinical trial data, authors have suggested a threshold of drug concentration of 1 mcg/mL, below which dose optimization may be required.(51)
A subsequent observational study suggested that a drug concentration of 4.5 mcg/mL was associated with optimal biomarker and endoscopic response.(52) However, it should be noted in that study subcutaneous dosing was used for induction, raising the question of whether the reported trough concentration target was at steady state. In any event, based upon trough concentration of 4.5 mcg/mL in this case, it is likely that the patient has adequate exposure; a switch of therapy (back to an anti-TNF or to vedolizumab) should be considered.
Carmela refuses to take higher doses of ustekinumab as she has seen no improvement in her symptoms. Accordingly, you start her on vedolizumab in combination with oral methotrexate 12.5 mg weekly to prevent sensitization. She returns for follow up just before her fourth dose and reports that her symptoms persist, with increasing drainage from her fistula. She sees a poster in your waiting room advertising a clinical trial for a new JAK inhibitor and likes the idea of an oral treatment.
With vedolizumab in CD, response should be evaluated after 14 weeks of therapy; it may be too early after the fourth dose to see the full effects of the treatment.(48) If, however, there continues to be no/insufficient response at that time, one might then consider switching back to an anti-TNF.
However, given the patient’s preference for an oral therapy, enrolment in a JAK inhibitor clinical trial is a reasonable option. Phase 3 studies of both filgotinib and upadacitinb for CD are currently recruiting.(53,54) Tofacitinib is already approved for treatment of UC, but not CD.
Additionally, ozanimod, an oral S1P modulator, is also enrolling patients for two phase 3 induction studies in CD.(55,56) A maintenance study is planned as well, but is not yet recruiting.(57)
John K. Marshall, MD MSc FRCPC AGAF, Director, Division of Gastroenterology, Professor, Department of Medicine, McMaster University, Hamilton, ON
Brian G. Feagan, MD FRCPC, Professor of Medicine, Epidemiology and Biostatistics, Senior Scientific Officer, Robarts Clinical Trials Inc., Robarts Research Institute, Western University, London, ON
Alain Bitton, MD FRCPC, McGill University, Montreal, QC
Brian Bressler, MD MS FRCPC, University of British Columbia, Vancouver, BC
Anne M. Griffiths, MC FRCPC, University of Toronto, Toronto, ON
Remo Panaccione, MD FRCPC, University of Calgary, Calgary, AB
Cynthia Seow, MBBS (Hons) MSc FRACP, University of Calgary, Calgary, AB
Hillary Steinhart, MD MSc FRCPC, University of Toronto, Toronto, ON
Jennifer Stretton, ACNP MN BScN, St. Joseph’s Healthcare, Hamilton, ON
Published by Catrile & Associates Ltd., 1B-391 Berkeley Street, Toronto, ON M5A 2X8
(c) Catrile & Associates Ltd., 2019. All rights reserved. None of the contents may be reproduced in any form without prior written permission from the publisher. The opinions expressed in this paper are those of the authors and do not necessarily reflect the opinions or recommendations of the sponsors, the grantor, or the publisher.