Vedolizumab (VDZ) is an anti-integrin monoclonal antibody effective in ulcerative colitis (UC) and Crohn's disease (CD). The aim of this study was to examine the “real world” efficacy and safety of VDZ in a large national patient cohort.
Patients with inflammatory bowel disease treated with VDZ were prospectively followed for 14 weeks. Patients who completed the induction protocol (week 0/2/6/14) or discontinued the treatment before week 14 for adverse events (AEs) or primary nonresponse were included. The primary outcome was induction of clinical remission at week 14; secondary outcomes included clinical response and corticosteroid-free clinical remission.
A total of 204 patients (CD-130, UC-69, inflammatory bowel disease–unclassified-5) from 8 centers in Israel were included. Fifteen (7.4%) of the patients were anti–tumor necrosis factor naive and 46 (35.4%) had a previous surgery. For patients with CD, 69/130 (53.1%) responded to treatment; 45 (34.6%) achieved clinical remission; and 38 (29.2%) achieved corticosteroid-free remission at week 14. Fourteen (10.7%) patients discontinued VDZ before week 14 due to primary nonresponse or AEs. For UC, 32/74 (43.2%) responded to treatment; 20 (28.4%) achieved clinical remission, and 18 (24.3%) achieved corticosteroid-free remission at week 14. Fifteen (20.3%) patients with UC did not complete the induction due to primary nonresponse or AEs. AEs were reported by 29 (14.2%) patients (CD and UC combined), most common being nasopharyngitis and skin eruptions.
In a large real-world Israeli cohort of anti–tumor necrosis factor–experienced patients with inflammatory bowel disease, VDZ was effective and safe in induction of clinical remission and steroid-free clinical remission.
Vedolizumab (VDZ) is a monoclonal antibody which blocks lymphocyte trafficking by binding to the α4β7 integrin. In clinical trials, VDZ was demonstrated to be effective in the induction and maintenance of remission in both ulcerative colitis (UC) and Crohn's disease (CD).1,2 In GEMINI I study, patients with moderate-to-severe UC, who received 300 mg VDZ intravenously, had a significantly higher (47.1%) response rate at week 6 compared with 25.5% in the placebo arm (P < 0.001). Remission rates at week 52 for patients continuing VDZ every 8 (41.8%) or 4 weeks (44.8%) were significantly higher compared with those switching to placebo after induction (15.9% P < 0.001).1 In the GEMINI II study, clinical remission at week 6 in patients with active CD was 14.5% compared with 6.8% in the placebo arm (P = 0.02). Remission rates at week 52 for patients continuing VDZ every 8 (39.0%) or 4 weeks (36.4%) were significantly higher compared with those switching to placebo after induction (21.6%).2 By contrast, in the Gemini III study, results of induction of remission in anti–tumor necrosis factor (TNF)–experienced patients with CD at week 6 did not demonstrate a definite benefit over placebo (primary outcome), although clinical superiority of VDZ over placebo was demonstrated for later time points at week 10 and beyond.3 Thus, data on clinical efficacy of VDZ, especially with respect to anti-TNF–experienced patients with CD are scarce. The safety profile of VDZ seems to be favorable, as indicated by the pooled data from 6 VDZ trials; specifically, no increase in the prevalence of infectious complications in comparison with placebo was reported.4 Several real-world data series have been published to date5,–8 demonstrating efficacy comparable with that reported in the randomized controlled trials.
The aim of this study was to describe our real-world experience in treating patients with inflammatory bowel disease (IBD) with VDZ in a multicenter national Israeli cohort.
Methods
This was a prospective observational Israeli multicenter study aimed to evaluate the efficacy and safety of VDZ in IBD. All patients who accumulated at least 14 weeks of follow-up between January 2015 and February 2016 were included. Patients discontinuing VDZ before week 14 for primary nonresponse or adverse events (AEs) were included as well. Patients with ileostomy or ileoanal pouch were excluded. VDZ was administered at the outpatient clinics of the participating centers. The patients were examined by a physician at each visit. Clinical remission was defined as a Harvey–Bradshaw index (HBI) <5 for CD and a partial Mayo score <2 or Simple Clinical Colitis Activity Index <4 for UC; physician's global assessment (PGA) was used when clinical scores were unavailable. Clinical response was defined as a drop of at least one severity category as defined by the relevant clinical score (HBI for CD and partial Mayo score for UC). AEs, including IBD-related hospitalization, surgery, infections, immediate and delayed hypersensitivity reactions and miscellaneous adverse effects, were recorded at each infusion.
The primary objective of the study was to evaluate the efficacy of VDZ for induction of clinical remission at week 14. Secondary objectives included clinical response and induction of steroid-free clinical remission at week 14.
Statistical Analysis
Descriptive statistics were presented as mean ± SD for continuous variables and percentages for categorical variables. Categorical variables were analyzed by Chi Square or Fisher's exact test and continuous variables by T-test or Mann–Whitney test. A 2-tailed P value <0.05 was considered statistically significant. For multivariate analysis, variables with significance level <0.1 were included. The analysis was performed using IBM SPSS statistic (version 20.0) (Armonk, NY).
Results
Two hundred four patients (130 CD, 69 UC, and 5 inflammatory bowel disease–unclassified [IBD-U]) from 8 centers were enrolled in the study between January 2015 and February 2016. Fifteen (7.4%) patients were anti-TNF naive. The clinical and demographic characteristics of the included patients are detailed in Table 1 (CD) and Table 2 (UC and IBD-U).
Clinical and Demographic Characteristics of Patients with Crohn's Disease Treated with Vedolizumab
 |
|
Clinical and Demographic Characteristics of Patients with Crohn's Disease Treated with Vedolizumab
|
|
Clinical and Demographic Characteristics of Patients with Ulcerative Colitis Treated with Vedolizumab
 |
|
Clinical and Demographic Characteristics of Patients with Ulcerative Colitis Treated with Vedolizumab
|
|
Efficacy—CD
One hundred thirty patients with CD were enrolled in the study; 116/130 (89.2%) completed VDZ induction (14 weeks). The mean disease duration was 11 ± 8 years; 46 (35.4%) had a previous surgery. Only 4 patients (3.1%) were anti-TNF naive, whereas 94/116 (72.3%) failed 2 anti-TNF agents. Forty-four (33.5%) patients had a moderate and 41 (31.8) severe clinical disease activity. HBI scores were available for 88 (75.6%) patients; PGA was used for the rest of the patients. At week 14, 69/130 (53.1%) responded, 45 (34.6%) achieved clinical remission, and 38 (29.2%) achieved corticosteroid-free remission (Fig. 1). When the cohort was limited to patients with available clinical scores only (excluding patients evaluated with PGA), 42/88 (47.7%) responded, 37/88 (42%) achieved clinical remission, and 14 (15.9%) achieved clinical remission by week 14.
The clinical outcomes of patients with IBD treated with VDZ at week 14.
C-reactive protein (CRP) was normalized in 21/76 (27.6%) patients with available CRP levels both at treatment initiation and during follow-up. Fourteen (10.7%) discontinued the treatment earlier because of primary nonresponse or AEs (12 and 2 patients, respectively).
The only clinical or demographic parameter that was associated with the likelihood of achieving clinical remission at week 14 was mild clinical activity (HBI <0 or mild disease defined by PGA) at treatment onset (P = 0.0001 and P = 0.001 on univariate and multivariate analyses, respectively).
Efficacy—UC/IBD-U
Seventy-four (69 UC and 5-IBD-U) patients were enrolled in the study; 59/74 completed VDZ induction (14 wk). The mean duration of disease was 10 ± 6 years; 38 (51.8) had left-sided colitis and 34 (46%) pancolitis. Twenty-four (32.4%) had a moderate and 26 (35.1%) severe disease activity. Sixty-three (85.1%) patients failed previous anti-TNF treatment, including 25 (33.8%) who failed 2 anti-TNF agents. Partial Mayo score was available for 46/69 (66.6%), the Simple Clinical Colitis Activity Index in 5 (7.2%) PGAs was used for the rest of the patients. At week 14, 32/74 (43.2%) responded; 20 (28.4%) achieved clinical remission, and 18 (24.3%) achieved corticosteroid-free remission (Fig. 1). When the cohort was limited to patients with available clinical scores only (excluding patients evaluated with PGA), 23/46 (50%) responded, 13/46 (28.2%) achieved clinical remission, and 9 (14.6%) patients treated with corticosteroids at treatment onset achieved corticosteroid-free clinical remission, CRP normalized in 14/34 (41.1) patients to the available CRP levels at treatment onset and clinical follow-up. Fifteen (20.3%) discontinued VDZ before week 14 because of nonresponse (14 patients) or AEs (1 patient). The only clinical or demographic parameter that was associated with the likelihood of achieving clinical remission was mild disease activity (partial Mayo score <5, Simple Clinical Colitis Activity Index <6, or mild disease as defined by PGA) on treatment onset (P = 0.04 and P = 0.02 for univariate and multivariate analyses, respectively).
Safety
As VDZ is considered to have a good safety profile in randomized controlled trials and long-term follow-up,1,–7,9,–11 we wished to assess whether this would be demonstrated in our real-world cohort of patients with complex IBD. By week 14, 13 (6.3%) required surgery, including 6 patients who underwent total colectomy and 7—ileocecectomy or small bowel resection. AEs were reported by 29 (14.2%) patients, leading to discontinuation of the treatment in 7 patients. One patient died. This was a 69-year-old patient who received VDZ and full-dose prednisone for severe pancolitis with deep ulcerations; after his second infusion of VDZ, he developed necrotizing fasciitis, followed by perforation caused by colonoscopy. After surgery, the patient developed multiorgan failure that resulted in death. One case of encephalitis occurred. The patient underwent a complete neurological workup, and progressive multifocal leukoencephalopathy was ruled out. VDZ treatment was reinstituted. One patient was hospitalized for an acute psychotic episode that was attributed to concomitant use of medical cannabis. As expected, the most frequently reported AE was nasopharyngitis. Importantly, we observed 5 gastrointestinal infections—3 cases of Clostridium difficile infection, 1 case of cytomegalovirus colitis, and 1 case of campylobacter infection. No acute infusion reactions occurred in this study. One patient developed lip edema and itching 3 days after VDZ infusion, classified as a possible delayed hypersensitivity reaction. He continued his treatment uneventfully with a premedication of prednisone and an antihistamine. Mild thrombocytopenia (60–100 K/μL) was observed in 1 patient after the second infusion of VDZ; the treatment was continued with no subsequent changes in platelet count; bone marrow biopsy was normal. Another patient complained of hearing loss that was reversed after treatment discontinuation. The main AEs described in the entire study cohort are detailed in Table 3.
Discussion
In this national real-world cohort of Israeli patients with IBD, VDZ was effective for induction and maintenance of clinical response, remission, and steroid-free remission at week 14 in patients with both CD and UC. The magnitude of the effect, similar to the previous studies, is specifically impressive because patients had complex IBD, as reflected by long disease duration, abundance of patients with previous surgeries, and previous exposure to at least one anti-TNF alpha agent in most patients. Our results are largely consistent with previously published real-world series.5,6,8,11 Amiot et al5 described the French experience with premarketing compassionate use VDZ program. Similar to our study, the cohort was CD-predominant and included anti-TNF experienced patients almost exclusively. At week 14, 31% of patients with CD were in steroid-free clinical remission and 51% responded; among patients with UC, 36% were in steroid-free clinical remission and 50% responded.5 In a recent German national cohort study including patients with CD and UC, 23.7% and 23.5% patients achieved clinical remission, 19.6% and 19.1% achieved steroid-free clinical remission, and 60.8% and 57.4% responded to the treatment at week 14 for CD and UC, respectively.6 In a recently published multicenter American study, clinical remission was achieved in 1/3 patients with moderate-to-severe CD.12
In GEMINI I/II studies, the primary outcome was defined at week 6.1,2 However, VDZ seems to be a relatively slow-acting agent, and a later endpoint for assessment of response may be more pertinent. The rates of clinical remission observed in our cohort are higher than those reported in the GEMINI/II/III study,1,–3 probably reflecting the longer duration of follow-up allowing for a slower acting mode of action of VDZ to be fully exposed. Our study cohort was relatively similar to that of the GEMINI III study that specifically addressed the anti-TNF–experienced patient population with CD.3 In GEMINI III, 10 weeks of VDZ treatment was required to achieve a statistically significant effect, while the efficacy at week 6 was not different from that of placebo. In our study, in similarity to the available real-world series, the response of patients with CD and UC to VDZ was comparable. In the GEMINI I/II studies, no benefit of concomitant immunomodulator treatment was detected, despite a significant difference in the emergence of anti-VDZ antibodies.1,2 Likewise, in our study, concomitant treatment with immunomodulators was not associated with a likelihood of response to induction and beyond.
A noticeable difference between the CD and UC cohorts in our study was the percentage of biologic-experienced patients. Although most patients with CD treated with VDZ were treated previously with 2 anti-TNF agents, this was the case for only a third of the UC cohort. This, however, did not affect short-term efficacy outcomes. In similarity to other real world experience (RWE) series and in contrast to the GEMINI studies' results, the efficacy of VDZ was quite similar in UC and CD in our cohort. The explanation for this may stem from the lower proportion of anti-TNF–naive patients with UC in the RWE series (15.1% in our study, 33% in the German series,6 and 2.5 in the Massachusetts cohort8 as opposed to 48.2% in GEMINI I1). It remains to be determined whether longer term follow-up will demonstrate that previous biologic therapies, especially when multiple, are associated with decreased efficacy, as shown for the anti-TNF agents.13
The safety profile of VDZ in our cohort was favorable and consistent with the current experience. In a recent summary of AEs reported in 6 randomized controlled studies of VDZ in IBD,4 no increased risk of any infection or serious infection was associated with VDZ exposure. Serious C. difficile infections, sepsis, and tuberculosis were reported infrequently (≤0.6% of patients). The most commonly reported AE was nasopharyngitis, which was also the most commonly reported AE in our cohort. Of note, we did observe several (6, 2.9%) gastrointestinal infections including 4 C. difficile infections, 1 case of cytomegalovirus colitis, and 1 campylobacter infection. As these pathogens are common in patients with IBD,14 it remains to be determined whether VDZ poses any additional risk for gastrointestinal infections, possibly due to its gut-selective mode of action. Similar rates of AEs were reported in the French and German cohort studies.5,6 Hearing loss and thrombocytopenia reported in our cohort were previously not described to the best of our knowledge. There was 1 death in our cohort, which occurred in a patient with very severe disease complicated by nosocomial infectious AEs.
The real-world experience with new therapies, specifically biologic therapy of patients with IBD, may differ from results reported in randomized controlled trials. This is intuitive, as patients included in real-world cohorts tend to have more complicated disease, and follow-up may be of longer duration. In this sense, it is important to mention that, similarly to several other real-world cohorts, our patients had complicated IBD, and most would not satisfy GEMINI I/II inclusion criteria. Despite this, VDZ efficacy in ours, and in other cohorts, is comparable to that of the randomized controlled trials. Moreover, safety seemed to be good despite this complexity, although several AEs such as mortality, encephalitis, hearing loss, and intestinal infection deserve further follow-up and attention. Thus, overall, real-world experience with VDZ points to a trend different from the anti-TNF alpha agents, where efficacy tends to be sustained or even increase, while safety is not hampered. These may be related to the unique mechanism of actions and may be applicable to other anti-integrin agents as well.
The main limitations of our study include a lack of regular endoscopic evaluation (pretreatment endoscopy was available in only a small proportion of patients) and a very low number of patients with available fecal calprotectin results. Clinical scores and CRP levels were also missing for some of the patients. However, these limitations are inherent to real-world, as opposed to clinical trial, setting, and are similar to many other similar series.
In summary, our study confirms the efficacy of VDZ for treatment of patients with complicated anti-TNF resistant IBD. VDZ was associated with a favorable safety profile consistent with the results of previously published randomized controlled trials and real-world series.
Acknowledgments
U. Kopylov reports research support form Takeda, Janssen, Medtronic and speaking fees from AbbVie, Takeda, Janssen, and CTS. I. Dotan reports consulting fees from Takeda, Janssen, AbbVie, Genentech, Pfizer, Ferring, Rafa laboratories, Given Imaging, and Protalix and speaking and teaching for Takeda, Janssen, AbbVie, Ferring, Falk Pharma, Given Imaging, Genentech, Pfizer, and MSD. A. Bar Gil Shitrit reports consulting fees from Takeda, Janssen, and AbbVie; speaking and teaching for Takeda, Janssen, AbbVie, Neopharm, and Rafa laboratories; and research support from Takeda and Janssen. S. Ben-Horin reports consultancy and/or advisory board fees from AbbVie, Schering Plough, Celltrion, Takeda, Janssen, and Novartis and research support from Celltrion and AbbVie. M. Waterman reports speaker and consulting fees from JC Healthcare, Israel, AbbVie, Israel, and Takeda, Israel and consulting fees from Given Imaging, Israel. Y. Chowers reports grant support from AbbVie, Takeda, Janssen, Medtronic, and Protalix and speaker fees from AbbVie, Takeda, Janssen, and Ferring. E. Zittan reports consulting fees from Janssen. R. Eliakim reports consulting/speaker fees from AbbVie, Takeda, Janssen, Rafa, Ferring, Medtronic, and Given Imaging. The remaining authors have no conflict of interest to declare.
References
Author notes
U. Kopylov and Y. Ron have contributed equally.
Author disclosures are available in the Acknowledgments.
