Efficacy and safety of vedolizumab for pediatrics with inflammatory bowel disease: a systematic review

Background Vedolizumab use in pediatrics is still off-label and the data are limited. We conducted a systematic review evaluating the efficacy and safety of vedolizumab in children and adolescents with inflammatory bowel disease (IBD). Methods PubMed, EMBASE and Cochrane databases were systematically searched for studies of vedolizumab in children and adolescents with IBD reporting clinical remission, response, corticosteroid-free (CS-free) remission, mucosal healing, or safety up to December 3rd 2021. Results Ten studies, comprising 455 patients were included. For CD, the pooled clinical remission rates were 25% (19/75) at 6 weeks, 28% (25/85) at 14 weeks, 32% (17/53) at 22 weeks, and 46% (43/92) at 1 year. For UC/IBD-U, the pooled clinical remission rates were 36% (25/70) at 6 weeks, 48% (52/101) at 14 weeks, 53% (24/45) at 22 weeks, and 45% (50/112) at 1 year. Mucosal healing was found in 17%-39% of CD and 15%-34% of UC/IBD-U respectively. Six percent of patients reported serious adverse events. Conclusions According to low-quality evidence based on case series, approximately one-third and one-half of patients for CD and UC/IBD-U respectively achieved remission within 22 weeks, and about half of patients achieved remission at 1 year with reasonable safety profile. Long-term benefit profile data and high quality evidence are still needed. Supplementary Information The online version contains supplementary material available at 10.1186/s12887-022-03229-x.


Background
Medical therapies commonly used for inflammatory bowel disease (IBD) include aminosalicylates, corticosteroids, immune modifiers, biologic agents, antibiotics and probiotics [1]. As IBD relapse rate is high, some patients might become corticosteroid-dependent or corticosteroid-resistant. It is reported that the rate of steroid dependency is much higher in children than in adults (45% vs. 8% respectively) [2]. Besides, although anti-TNF agents have been a significant breakthrough in the treatment of IBD, approximately ~ 10%-40% patients do not improve after therapy (primary non-response), and ~ 20%-40% may lose response to therapy overtime (second loss of response) [3][4][5][6][7]. Therefore, there is still a great need for new drugs with other mechanisms of action that act on different inflammatory pathways involved in the pathogenesis of IBD [8].
Vedolizumab is a novel, fully humanized immunoglobulin G1 monoclonal antibody selective for the gut. It can block only α4β7 integrin that inhibits adhesion of a gut-homing subset of T lymphocytes to mucosal addressing cellular adhesion molecule-1 (MAdCAM-1) [9]. For adults, the efficacy and safety of vedolizumab on moderate-to-severely active UC or CD have been established by GEMINI clinical trials [10][11][12], and got marketing approval in May 2014 in the USA and later in Europe [13,14]. Guidelines suggested vedolizumab could be used in the treatment of UC where anti-TNF therapy had failed [15], and UC or CD who was refractory to steroids or anti-TNF [16,17]. For children, European Crohn's and Colitis Organization (ECCO) and European Society for Pediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) recommended vedolizumab for UC in chronically active or steroid-dependent patients as second line biologic therapy after anti-TNF failure [18], and for CD in patients who fail to achieve or maintain clinical remission on anti-TNF agents, despite anti-TNF dose optimization and immunomodulator use [19]. Nevertheless, vedolizumab use in pediatric population is still offlabel and the efficacy evidence is insufficient. Given that the increasing use of vedolizumab in pediatrics, safety monitoring is essential, as it is suggested that drug safety must be demonstrated independently from adult studies and couldn't be extrapolated [20].
The aim of this study was to summarize the current evidence and to assess the efficacy and safety of vedolizumab for children and adolescents with IBD.

Methods
The systematic review was registered in PROSPERO (registration number: CRD42020222828) and was performed in accordance with the guidelines established by the PRISMA statement [21].

Literature search
We systematically searched PubMed, EMBASE and Cochrane Library databases from inception to December 3 rd 2021 using the following search terms: "inflammatory bowel disease", "vedolizumab", "child" and "adolescent". The full search strategy is detailed in Supplementary Data (Table S1). Language or publication type was without restriction.

Inclusion criteria
Studies that met the following criteria were included in this systematic review: (a) studies carried out in children and adolescents with pediatric onset (< 18 years) IBD (CD, UC, unclassified), remaining under pediatric monitoring, and evaluation up to 21 years old were included; (b) treatment with vedolizumab alone or combination with other agents; (c) studies written in English.

Exclusion criteria
Studies were excluded according to the following criteria: (a) studies on non-human subjects; (b) studies conducted on adults subjects; (c) the number of case series is less than five; (d) studies that were letters or editorial; (e) studies that lacked sufficient raw data; (f ) studies that were duplicated; (g) studies that were ongoing or not finished; (h) abstract only.

Outcomes and endpoints
The primary outcome measure of this systematic review was clinical remission; second outcome measures included: (a) clinical response; (b) corticosteroidfree (CS-free) clinical remission; (c) mucosal healing; (d) safety (any adverse event that was judged related to vedolizumab by authors of the primary study). Clinical remission, clinical response and CS-free rates were collected after first dose where available. The definition of clinical remission, clinical response, CS-free clinical remission and mucosal healing varied in different studies and were summarized in Table 1.

Data extraction
All the potentially related articles were retained by two authors (FSB, SYQ) independently, and the full texts were strictly reviewed according to inclusion/exclusion criteria regarding to preset outcomes. Any disagreements were resolved by consensus or consulted with a senior author (WLB). For the included studies, the following items were extracted: study characteristics (author, year of publication, country, study design), patients characteristics (age, type of IBD, disease behavior, percentage of anti-TNF experienced), vedolizumab dosage, clinical efficacy and adverse events (AEs).

Methodological assessment
For quality assessment, a validated quality appraisal tool developed by the Canadian Institute of Health Economics (IHE) was used for case series [32], including study objectives, population, interventions and co-interventions, outcome measures, statistical analysis, results and conclusions and competing interests. A study with 14 or more yes responses (≥ 70%) was considered to be of acceptable quality [33]. The grade of evidence was showed in Table S2.

Statistical analysis
We provided descriptive statistics. Continuous parametric data are presented as mean and standard deviation (SD), while nonparametric data presented as median followed by range or interquartile range (IQR), unless otherwise specified. The categorical data of the outcome measures are expressed as percentage of total cases with 95% confidence interval (95% CI).
Six studies reported CS-free remission rates [22-25, 28, 31] and 3 reported mucosal healing [24,27,29]. Nine studies reported safety outcomes for CD or UC/ IBD-U combined, rather than by separate indication [22][23][24][25][26][28][29][30][31]. Characteristics of the included studies are listed in Table 1, and patient demographics were showed in Table 2. These studies were mostly reported by institutions from the USA and differed with respect to patients' age, number of patients included, concomitant treatment, vedolizumab dose, duration of treatment and follow-up, and definition of outcomes. Most patients received 300 mg vedolizumab, and others received 3.6-10.3 mg/kg vedolizumab.
The pooled results for clinical remission rates, CS-free clinical remission, and response rates were presented in Table 3.

Mucosal healing
Three studies investigated mucosal healing, but one study did not draw clear conclusion due to the small sample size (n = 8) [29]. Another two studies (87 patients in total) reported mucosal healing results [24,27]. Mucosal healing was found in 17%-39% of CD (n = 39) and 15%-34% of UC/IBD-U (n = 48) respectively, with various evaluation time. Details of number of patients assessed and evaluation time were presented in Table 4.

Study quality
A 20-item validated quality appraisal tool for case series were used for quality assessment. The median of quality score was 17 (range 13-18), with only one study quality score less than 14 [26].This study was case series which only safety data were involved in our study, and did not affect the quality of the whole analysis. The grade of evidence was showed in Table S2.

Discussion
The results of this systematic review showed that most of the pediatric data on the effectiveness and safety of vedolizumab for the treatment of IBD were descriptive and the evidence were inadequate, as all the studies included were case series without randomized controlled trails (RCTs).
Overall, we found 0%-35% of CD patients achieved clinical remission in short-term therapy, compared to that of 20%-64% in UC patients. During maintenance therapy, 17%-73% of CD patients and 20%-77% of UC/ IBD-U patients achieved clinical remission. Approximately 33%-75% of CD patients and 20%-78% of UC/ IBD-U patients had clinical response with quite small sample size. These findings suggested similar therapeutic response were obtained in CD and UC, which were not consistent with previously published studies in adults. Randomized controlled trials of GEMINI 1 and 2 found that compared with CD, the response and remission rates in UC were higher at both 6 weeks (47.1% and 16.9% vs. 31.4% and 14.5%) and 52 weeks (56.6% and 41.8% vs. 39.0% and 43.5%) [10,11]. Canadian and Hungarian realworld cohorts also showed significantly greater clinical remission and response rate for UC compared with CD [34,35]. However, opposite results reported by Dragoni et al., cohort in Italy showed better results for CD patients, with higher clinical response and remission rate compared with UC at 14 weeks (85% and 69% vs. 52% and 30%), 24 weeks (84% and 61% vs. 56% and 26%) and 52 weeks (59% and 45% vs. 25% and 20%) [36]. The difference in clinical response and remission rate could be attributed to quite small sample size and differences in patients baseline characteristics variability: the characteristic of patients involved varied in IBD phenotype, disease severity at vedolizumab initiation, disease duration.
Steroid-free remission, whether clinically or endoscopically is an important treatment goal for pediatric IBD [20,37], as corticosteroids have potentially serious side effects associated with long term use including linear growth restriction, and osteopenia amongst many others [38]. In a meta-analysis on adult population, approximately one-quarter of CD or UC achieved CS-free clinical remission at 14 weeks, while 31% of CD and 42% of UC of that at 12 months [39]. Our study seemingly showed similar results. In our study, we found 0%-19% of CD and 20%-44% of UC/IBD-U patients achieved CS-free clinical remission at 14 weeks. Higher rate was identified for UC (40%-71%) compared with CD (13%-33%) at 22 weeks, and similar rate was found at 1 year [45%(CD) vs. 41%(UC)]. However, opposite results have also been reported. In a real-world study by Zingone et al., better results for CD were identified at any follow-up time [ie, between 8-12 weeks 53.6% vs.18.7% (UC); 30 weeks 56.5% vs. 25% (UC); 52 weeks 53.6% vs. 35.4%(UC)], as much more CS ongoing UC were initially involved [45.8% vs. 24.6%(CD)] [40].
Although mucosal healing is a critical IBD therapy goal associated with sustained clinical remission, it is too burdensome for children to frequently undergo endoscopy. Therefore, only two studies of small sample size reported mucosal healing rates of 17%-39% for CD and 15%-34% for UC/IBD-U with median follow-up time over 6 months. In addition, the definition of mucosal healing is still controversial. Most investigators agree that an endoscopic Mayo subscore of 0 for UC, and simple endoscopic score for Crohn's disease (SES-CD) 0-2 for CD [20]. However, one study involved in our review defined mucosal healing more strictly, including a composite of both endoscopic (macroscopic) and histologic indices [27]. In adult population, mucosal healing rates are reported as 21.2-41.9% for CD and 15%-57.1% for UC regardless of patients' baseline characteristics [34][35][36]41].
It is discussed, whether a previous treatment with anti TNF might influence the outcome of treatment with vedolizumab. Two post hoc analyses from the GEMINI studies assessed the efficacy of vedolizumab in CD and UC based on previously anti-TNF experienced patients [42,43]. Results showed for CD, there were higher response and remission rates in patients who were anti-TNFnaïve compared with anti-TNF-experienced, and the advantages persisted to week 52 [42]. In UC patients, similar outcomes were found. Compared to placebo, patients naïve to anti-TNF had higher rates of response than patients with anti-TNF failure at week 6, whereas during maintenance therapy, there were no significant difference with placebo in both groups and between the two groups [43]. In contrast, however, some real-world clinical studies indicated that there was no impact of previous anti-TNF exposure on response or maintenance of remission though the sample size of TNF-naïve patients were small [34,44]. Chaparro et al. found the remission rates of patients who were anti-TNF naïve, with failure to 1 anti-TNF and failure to > 1 anti-TNF at week 14 were 57.6%, 51.2% and 44% respectively [44]. And a study by Kotze et al. demonstrated that previous failure to anti-TNF agents was not associated with the efficacy of vedolizumab [34]. More interestingly, a study by Mader showed previous treatment with anti-TNF agents was associated with a significantly lower efficacy of VDZ in UC but not in CD patients [45]. This might be attributed to longer disease duration for anti-TNF-experienced UC patients. In pediatric population, Jossen et al. found higher rates of both endoscopic and histologic remission in anti-TNF-naïve patients compared to those who were anti-TNF-experienced (66% vs. 42%, 52% vs 33%, respectively) [27]. However, authors admitted these anti-TNFnaïve patients had slightly less severe disease at baseline compared with the anti-TNF-experienced patients [wPC-DAI 26.2 (19.4-35.6) vs. 35 (25-57.5); pMayo 3.5(2-5) vs. (3-6.5)]. Therefore, this question deserves further investigation to determine whether the differences are due to true biological effects of anti-TNF exposure or the severity and duration of the disease reflected in patients who started using vedolizumab.
With respect to safety, phase 2 and 3 trials showed a favorable safety profile of vedolizumab, with similar AEs incidence rate compared with placebo [46,47]. Safety data from real-world cohort studies reported the total AE incidence rate was 23.6%, with infectious complication rate 7.8% [47]. In pre-marketing clinical  The subject who had diverting ileostomy due to severe perianal disease, developed bowel-associated dermatosis-arthritis syndrome and was treated with antibiotics and corticosteroids with subsequent resolution of symptoms and continued on vedolizumab without further recurrence of these manifestations g The subject who initially had erythema nodosum, later developed synovitis, acne, pustulosis, hyperostosis, and osteitis syndrome, characterized by dermatologic and osteoarticular findings without clear etiology that has been associated with IBD in previous case reports h The subject with CD, who had a history of recurrent acute kidney injury due to hypovolemia with disease flares, developed obstructing nephrolithiasis with associated pyonephritis, then underwent drainage and ureteral stent placement as well as intravenous antibiotic treatment, and was continued on vedolizumab achieving remission without further kidney involvement i The subject with CD, who had worsening symptoms and distal colonic inflammation, required a diverting ileostomy j The subject with UC treated with the combination of vedolizumab 300 mg every 8 weeks and tofacitinib 10 mg twice daily, in addition to prednisone 30 mg daily, developed septic arthritis of the right knee 2 months after dual therapy initiation, requiring inpatient hospitalization with incision and drainage and a prolonged course of intravenous antibiotic therapy k The subject above subsequently developed a deep vein thrombosis in the right leg 5 months after dual therapy initiation There are several limitations in our review. Initially, there was a significant heterogeneity in study design, including the threshold criteria of patients involved and definitions of remission, response and mucosal healing. Most studies used Pediatric Crohn's Disease Activity Index (PCDAI) or Pediatric Ulcerative Disease Activity Index (PUCAI), but weighted Pediatric Crohn's Disease Activity Index (wPCDAI), short Pediatric Crohn's Disease Activity Index (shPCDAI) or partial Mayo score were also used. As to mucosal healing, endoscopic assessment alone was agreed by majority of investigators, but Jossen et al. also evaluated histological changes [27]. Moreover, all the studies included were case series, some reported the data prospectively while the others used a retrospective approach, which may result in significant differences in clinical decision. In addition, there was no placebo-controlled trial with a standard protocol, which meant the effectiveness was not necessarily attributed to the intervention. The recurrent nature of CD additionally weakens the assessment of causal relationships between interventions and outcomes. Nevertheless, vedolizumab for pediatric patients is usually applied to patients with severe disease or those who are refractory to conventional therapies, which are unlikely to have spontaneous relief.
In spite of the above-mentioned shortcomings, we performed a comprehensive literature search. Although no RCTs were included, case series of vedolizumab therapy seemed to represent 'real-world' experience of pediatric population in different areas and medical centers and provide a deeper understanding of vedolizumab in heterogenous and more complex patient populations. Besides, the role of case series evidence in systematic reviews of health care interventions is especially suitable for reviews of rapidly developing pharmacological interventions and supporting evidence on safety, when case series are usually the only available clinical evidence [48].

Conclusions
Based on low-quality evidence provided by case series, approximately one-third and one-half of patients for CD and UC/IBD-U respectively, achieved remission within 22 weeks with favorable safety profile, and about half of patients achieved remission at 1 year with reasonable safety profile. Long-term benefit profile data and more robust evidence are still needed.