Management of Paediatric Ulcerative Colitis, Part 1: Ambulatory Care—An Evidence-based Guideline From European Crohn's and Colitis Organization and European Society of Paediatric Gastroenterology, Hepatology and Nutrition

Abstract

Background: 

The contemporary management of ambulatory ulcerative colitis (UC) continues to be challenging with ∼20% of children needing a colectomy within childhood years. We thus aimed to standardize daily treatment of pediatric UC and inflammatory bowel diseases (IBD)-unclassified through detailed recommendations and practice points.

Methods: 

These guidelines are a joint effort of the European Crohn's and Colitis Organization (ECCO) and the Paediatric IBD Porto group of European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN). An extensive literature search with subsequent evidence appraisal using robust methodology was performed before 2 face-to-face meetings. All 40 included recommendations and 86 practice points were endorsed by 43 experts in Paediatric IBD with at least an 88% consensus rate.

Results: 

These guidelines discuss how to optimize the use of mesalamine (including topical), systemic and locally active steroids, thiopurines and, for more severe disease, biologics. The use of other emerging therapies and the role of surgery are also covered. Algorithms are provided to aid therapeutic decision-making based on clinical assessment and the Paediatric UC Activity Index (PUCAI). Advice on contemporary therapeutic targets incorporating the use of calprotectin and the role of therapeutic drug monitoring are presented, as well as other management considerations around pouchitis, extraintestinal manifestations, nutrition, growth, psychology, and transition. A brief section on disease classification using the PIBD-classes criteria and IBD-unclassified is also part of these guidelines.

Conclusions: 

These guidelines provide a guide to clinicians managing children with UC and IBD-unclassified management to provide modern management strategies while maintaining vigilance around appropriate outcomes and safety issues.

What Is Known

• The previously published European Society of Paediatric Gastroenterology, Hepatology and Nutrition–European Crohn's and Colitis Organization guidelines were published in 2012 and are updated herein.

What Is New

• The diagnosis section has been replaced by the inflammatory bowel disease-classes criteria; a discussion of inflammatory bowel disease-unclassified has been added; fecal calprotectin has been given more emphasis; new drugs (eg, vedolizumab, golimumab) have been incorporated as off-label medications; recommendations for therapeutic drug monitoring have been provided; a treat to target algorithm has been added and other sections updated.

INTRODUCTION

Ulcerative colitis (UC) is a disease with a less heterogeneous phenotype than Crohn disease (CD) but it still poses many unique challenges. The incidence of pediatric onset UC, which constitutes roughly 15% to 20% of all UC, ranges at 1 to 4/100,000/year in most North American and European regions ⁽¹⁾. It is extensive in 60% to 80% of all cases, twice as often as in adults ⁽²⁾. Since disease extent has been consistently associated with disease severity, it is not surprising that children with UC more often require hospitalization for an acute severe exacerbation (25%–30% over 3–4 years) ^(3,4) and more often undergo colectomy for medically refractory disease (up to 30%–40% in 10-year follow-up ^(2,5), although lower colectomy rates have also been reported ^(6–8)). Canadian population-based health administrative data showed no reduction of colectomy rate from 1994 to 2007 before the widespread use of biologics ⁽⁹⁾. In addition to more severe colitis, children also have unique age-related issues, such as growth, pubertal development, nutrition, and bone mineral density accretion, as well as differing psychosocial needs. Finally, although mortality in pediatric UC has become rare, a retrospective case collection across Europe over 6 years reported 19 deaths in children with UC mainly due to infections and cancer (1 case of colorectal cancer [CRC]), including 1 with toxic megacolon ⁽¹⁰⁾.

The revised Porto criteria ⁽¹¹⁾ proposed explicit guidance for diagnostic workup in pediatric inflammatory bowel diseases (IBDs). Consequently, the Paediatric IBD Porto group of European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) published the “PIBD-Classes” criteria that standardized the differentiation of pediatric IBD into 5 categories: typical UC, atypical UC, IBD-unclassified (IBDU), Crohn colitis and CD ⁽¹²⁾; the first 3 categories will be covered in these guidelines.

The PIBD-classes system is based on 23 features that are typical of CD, grouped in 3 classes: those that are totally incompatible with UC and thus should be diagnosed as CD; those that may be present in UC but rarely (<5%; class 2); and those that may be present in UC uncommonly (5%–10%; class 3). Accumulation of the different features, weighted by the classes, standardized the diagnosis of PIBD (Fig. 1, Table 1). The sensitivity and specificity of the PIBD-classes to differentiate UC from CD and IBDU was 80% and 84%, and CD from IBDU and UC 78% and 94%, respectively ⁽¹²⁾.

FIGURE 1: Classification algorithm into paediatric inflammatory bowel disease (IBD) subclasses based on the “PIBD-classes” features of Table 1 (reproduced from reference 12).

TABLE 1: PIBD-classes features

The use of the Paris classification is advocated for phenotyping pediatric UC, with E1-E4, A1a-A2, and S0-S1 denoting disease extent, age of diagnosis and severity, respectively ⁽¹³⁾. Additional labels of very-early onset IBD (≤6 years of age at diagnosis) and infantile IBD (<2 years of age) may also be added ⁽¹⁴⁾.

We aimed to develop guidelines for managing UC in children based on a systematic review of the literature and a robust consensus process of an international working group composed of specialists in pediatric IBD from the ESPGHAN and the European Crohn's and Colitis Organization (ECCO). We focus on the principles, pitfalls, and pediatric considerations related to the diagnosis and care of children and adolescents with UC. These guidelines supplement those published for adults ^(15,16); similar topics are covered only in brief, referencing the extensive ECCO review. The pediatric UC guidelines are divided into 2 parts but should be read as 1 manuscript: Part 1: ambulatory UC (updating the previous 2012 ECCO-ESPGHAN guidelines ⁽¹⁷⁾) and Part 2: acute severe colitis (ASC; updating the previous 2011 ECCO-ESPGHAN guidelines ⁽¹⁸⁾).

In addition to providing an update of new literature, several major topics have changed from the previous guidelines. The diagnosis section has been replaced by the aforementioned IBD-Classes criteria; a discussion of IBDU has been added; fecal calprotectin has been given more emphasis; new drugs (eg, vedolizumab, golimumab, and locally active steroids) have been incorporated as off-label medications; practical recommendations for therapeutic drug monitoring have been provided; the use of thrombotic prophylaxis has been revisited based on predicting variables; sequential therapy has been newly presented; a treat to target algorithm has been added; and other sections updated and changed.

METHODS

Following an open call in ECCO and the Porto plus the Interest Paediatric IBD groups of ESPGHAN, 22 international experts in pediatric IBD were selected by the steering committee, including 2 pediatric surgeons. A list of 23 questions addressing the management of UC in children was first developed (composing the subtitles of the current manuscript and the next one on ASC). Next, a systematic review of the literature was performed centrally by 2 of the authors (E.O.M. and C.S.) with the aid of an experienced librarian searching for all combinations of UC and pediatrics (Supplemental Table 1, Supplemental Digital Content 1, https://links.lww.com/MPG/B393). Electronic searches were performed in Oct 2016 using Medline, Embase, and web of science. Clinical guidelines, systematic reviews, clinical trials, cohort studies, case-control studies, diagnostic studies, surveys, letters, narrative reviews, case series, and highly relevant selected abstracts published after 1985 were all utilized if performed in children. Following elimination of duplicates, 10,096 abstracts were reviewed by EOM for eligibility. A total of 8,996 abstracts were excluded, mainly for the following reasons: clear irrelevance to the pre-defined topics, manuscripts published before 1985, review manuscripts, manuscripts focusing on CD, or on molecular/genetic pathways. Although we aimed to base our adult literature on the recently updated ECCO UC guidelines ^(15,16), salient adult RCTs identified in the initial search were not excluded for perusal and reference. The decision regarding questionable eligibility was made by one of the senior authors (D.T.). Finally, 1100 full-text manuscripts were retrieved and circulated to the relevant subgroups for writing their sections. Highly relevant manuscripts published after the search date were included individually.

Each of the 23 questions was allocated to a subgroup of 2 experts for drafting of the first text. The subgroup's text and recommendations were iterated by e-mail with the steering committee until refined. The guidelines include both recommendations and practice points that reflect common practice where evidence is lacking or provide useful technical details, including grading of evidence according to the Newcastle-Ottawa assessment scales for case-control and cohort studies ⁽¹⁹⁾ and according to the Cochrane Handbook for clinical trials ⁽²⁰⁾ (Supplemental Table 2: tables of evidence with grading, Supplemental Digital Content 2, https://links.lww.com/MPG/B394). The group then voted on all recommendations and practice points while adding specific comments using a web-based voting platform. A second round of electronic voting and revisions was done, including all members of the Paediatric IBD Porto group of ESPGHAN. In addition, the draft was circulated for comments to ECCO (national representatives and governing board) and to members of the IBD Interest group of ESPGHAN.

The group met twice face-to-face: during UEGW annual meeting (Barcelona, October 2016) before drafting the initial topics and during ESPGHAN annual meeting (Prague, May 2017) after the 2 voting rounds were completed. The meetings were supplemented by an e-mail Delphi process with the entire group until agreement was reached. In total 43 pediatric IBD experts voted on all recommendations and practice points: 35 Porto group members (of whom 14 were authors) and 8 non-Porto group authors (Supplemental Table 3: names of the 43 voting experts, Supplemental Digital Content 1, https://links.lww.com/MPG/B393). All statements and practice points were supported by at least 88% of the group. Recommendations were graded according to Oxford Centre for Evidence-Based Medicine (see table at https://www.cebm.net/wp-content/uploads/2014/06/CEBM-Levels-of-Evidence-2.1.pdf⁽²¹⁾).

EVALUATION AND PREDICTION

Assessing and Predicting Disease Activity

Recommendations

• 1. Disease activity should be monitored at every visit utilizing the PUCAI [EL2] and treatment should be revisited when PUCAI ≥ 10 points [EL2]. (93% agreement) (Fig. 2)
• 2. Colonoscopic evaluation is recommended at diagnosis [EL4, adults EL4], before major therapeutic modifications [EL5, adults EL5], for cancer surveillance [EL5, adults EL3], and when it is not clear if symptoms are disease-related especially if calprotectin is elevated [EL5, adults EL5]; it is not routinely indicated during relapses that are not severe [EL5, adults EL5]. (100% agreement)
• 3. If available, fecal calprotectin should be obtained while in sustained clinical remission and endoscopic evaluation should be considered when calprotectin is high, as defined below and in Figure 2 [EL2, adults EL2]. (88% agreement)

Practice Points

• 1. Clinical remission is defined as PUCAI < 10 points, mild disease as 10 to 34 points, moderate disease 35 to 64 points and severe disease ≥65 points (Appendix 1). Clinically significant response is defined by a PUCAI change of at least 20 points, or entering remission. (95% agreement)
• 2. Long-term prognosis is better in patients who achieve complete clinical remission (ie, PUCAI < 10) during the first 3 months after diagnosis. (95% agreement)
• 3. There is currently no evidence whether measuring calprotectin in a child who is in a PUCAI-defined remission has an added value for predicting disease course. Given the fact that significant endoscopic disease may, however, be present in ∼20% of children with PUCAI < 10, it is reasonable to measure calprotectin once sustained clinical remission has been achieved to verify mucosal healing and select those who require endoscopic assessment. Other fecal markers (eg, lactoferrin) may have a comparable diagnostic value, but less supportive data are available. (93% agreement)
• 4. There is no ideal cutoff value of fecal calprotectin to reflect mucosal inflammation and predict disease outcome (Tables 2 and 3). Values differ substantially in the different studies using different reference standards. Cutoff value <100 μg/g usually reflects remission while >250 μg/g more accurately predicts mucosal inflammation. The value that should trigger an endoscopic evaluation or a change in treatment should be thus individualized based on these values, especially when values increase over time. (98% agreement)
• 5. An episode of acute severe colitis (ie, PUCAI ≥ 65) is a risk factor for a more aggressive disease course and thus this should be incorporated in the management scheme. (100% agreement)
• 6. Blood tests (CBC, albumin, transaminases, gGT, CRP, and ESR) should be performed regularly depending on symptoms and therapy and at least every 3 months while on immunosuppressive medications and at least every 6 to 12 months otherwise. It is a common practice to include testing for renal function in patients taking mesalamine and annual urinalysis; however, there is no evidence that this prevents adverse outcomes. (98% agreement)
• 7. Before treatment modification, it is essential to consider other clinical conditions such as non-adherence, irritable bowel syndrome, celiac disease, medication-related adverse events, and infections (especially Clostridium difficile, which should be excluded in any acute exacerbation, but also bacterial infections and CMV). (95% agreement)
• 8. A standardized endoscopic activity index, including the Mayo endoscopic subscore or Ulcerative Colitis Endoscopic Index of Severity (UCEIS), should be used during colonoscopic examinations. (95% agreement)
• 9. CRC surveillance by a trained endoscopist is recommended following 8 to 10 years of disease duration, dictated by risk factors such as disease extent, disease severity over the course of disease and family history. Surveillance recommendations in children with primary sclerosing cholangitis (PSC) can be found in the PSC section. According to the adult guidelines, chromoendoscopy with targeted biopsies has been shown to increase dysplasia detection rate. If not available, random biopsies (quadrantic biopsies every 10 cm) and targeted biopsies of any visible lesion should be performed using high definition endoscopes. (93% agreement)

In 2 pediatric inception cohorts, disease severity during the first 3 months after diagnosis and the occurrence of an episode of ASC were associated with increased risk of refractory disease ^(22,23). Thus, by using constructs of disease severity it is possible to characterize children who are at high risk for a more complicated disease course and to guide management and tight monitoring.

FIGURE 2: Algorithm for monitoring paediatric ulcerative colitis (UC) during the maintenance phase. (1) Assessments earlier than 3 months are usually required and in any significant disease or deterioration, early intervention is required. (2) The decision whether to escalate therapy based on a Mayo—0 or 1 endoscopic findings should be individualized such as based on the current treatment (eg, it is easier to increase mesalamine dose or add rectal therapy than starting thiopurines), symptoms and extent (short Mayo 1 segment may be closely monitored whereas extensive disease may require escalation). (3) Proceeding to colonoscopy should preferably be based on at least 2 independent measurements of calprotectin. (4) Obtaining calprotectin may be delayed to 4 to 6 months since histological remission lags after macroscopic improvement.

TABLE 2: The utility of different cutoff values of faecal calprotectin to predict endoscopic disease in UC (selected references)

TABLE 3: Cutoff values of faecal calprotectin in prediction poor outcome in ulcerative colitis (selected references)

Endoscopy is the reference standard to evaluate mucosal inflammation. Mayo endoscopic score of none, mild, moderate, or severe (0–3 points) with number of involved colonic segments (rectum, sigmoid, and descending, transverse, and ascending colon) may be used in pediatric UC ⁽²⁴⁾. The modified Mayo endoscopic score is an easy to use, non-validated tool, which combines disease extent with Mayo Endoscopic score ⁽²⁵⁾. The Ulcerative Colitis Endoscopic Index of Severity (UCEIS) is a convenient and validated index which includes vascular pattern, bleeding, and ulcers at the worst part ^(26,27). These indices are described in the coming ESPGHAN Porto group guidelines of endoscopy utilization in IBD (JPGN 2018).

Mucosal healing in UC is associated with a favorable disease outcome in adult patients ^(28–31). Nevertheless, clinical remission has been proven to predict long-term outcomes in UC, with no less accuracy than endoscopic evaluation, both in children using the PUCAI ^(23,32) and in adults ⁽³³⁾. An adult study showed that UCEIS predicted relapse in 155 patients who were in clinical remission; however, clinical remission was not stringently defined (ie, partial Mayo score of 0–1, allowing for streaks of blood for instance) and the number needed to test was high ⁽³⁴⁾. A post-hoc analysis of the adult ACT trials showed that while endoscopic inflammation predicted colectomy, this was not the case in the subgroup of patients who were in clinical remission ⁽²⁸⁾. A PUCAI-defined remission at 3 months following diagnosis predicted 1-year sustained steroid-free remission (AUROC 0.7, 95% CI 0.6–0.8) and colectomy by 2 years (AUROC 0.75, 0.6–0.89). It was superior to both CRP and ESR ⁽²³⁾ and predicted choice of treatment ^(35,36). Furthermore, in the prospective multicenter PROTECT pediatric cohort study, failure to achieve clinical remission (PUCAI < 10) 4 weeks after discharge of children who required intravenous corticosteroids at disease onset was highly associated with need for additional medical therapy by week 12 ⁽³⁷⁾.

PUCAI cutoff scores of remission, mild, moderate, and severe disease have been validated in several cohorts ^(35,38,39) and were successfully utilized in the PROTECT study to guide the choice of initial treatment at disease onset, as outlined in Figure 3⁽³⁷⁾. PUCAI at diagnosis was associated with steroid-free remission rates at week 12 and with long-term outcomes (at 54 weeks). Selected children with moderate disease activity were, however, treated with 5-aminoscalicylic acid (5-ASA) and not with oral steroids, and on average had similar outcomes at week 12; this supports our algorithm that 5-ASA may be considered also in the lower range of the moderate disease activity group (Fig. 3). The PUCAI correlates well with endoscopic appearance of the colonic mucosa, showing similar remission rates in multiple studies ^(38–43). In addition, the correlation of the PUCAI with Mayo score has been reported to be as high as 0.95 ^(32,38,39). While most aforementioned studies report a group average, on an individual basis there is a likelihood of ∼20% for a significant mucosal inflammation even in the presence of a PUCAI-defined complete remission ⁽⁴⁴⁾. Therefore, biomarkers should be used to confirm endoscopic remission in those who are in sustained clinical remission, particularly in the presence of PSC where the PUCAI does not correlate well with mucosal inflammation ⁽⁴⁵⁾ (Fig. 2).

FIGURE 3: Summary flowchart of managing pediatric ulcerative colitis (UC). Comments: Medical therapies in UC should be divided into those that induce remission (5-ASA, corticosteroids, anti-tumur necrosis factor (TNF) therapy, calcineurin inhibitors and likely probiotics) and those that maintain remission (5-ASA, thiopurines, anti-TNF therapy, vedolizumab and selected probiotics). 1. The following should be considered in active disease: infectious colitis (including CMV and C. difficile), 5-ASA related colitis, lactose intolerance, irritable bowel syndrome, celiac disease. In case of discrepancy between PUCAI and endoscopic grading of colitis, endoscopy should prevail. 2. 5-ASA is dosed 60-80 mg/kg/day up to 4.8 grams daily. Once daily dosing may be as effective as twice daily dosing. 3. 5-ASA enemas (1g daily is as effective as higher doses) are more effective than steroid enemas. Enemas should be administered in the left decubitus position. Liquid enemas are more difficult to tolerate than foams and suppositories but are more suitable for extensive colitis. 4. If lack of improvement (i.e. PUCAI decrease of <20 points) after 7–10 days, or increase in PUCAI ≥20 points at any time, consider treatment escalation. Steroid dependency must be avoided. 5. See Part 2 of these guidelines. 6. Response is defined as a drop in PUCAI of at least 20 points. However, the goal of induction therapy is eventually complete remission (Figure 2). 7. e.g. previous intolerance or resistance to steroids, or when infliximab is indicated anyway for maintenance treatment after failing thiopurines. 8. Measuring TPMT (genotyping or enzymatic activity) at baseline, and 6-TG and 6-MMP levels after 2–3 months, may aid in optimizing thiopurine dosing. 9. If infliximab has been used in thiopurine-naïve disease, thiopurines may be added and infliximab discontinued after 4–8 months if complete remission has been achieved. Stepping down to 5-ASA may be considered in selected cases, if 5-ASA did not fail previously, and after a period of sustained deep remission.

Routine laboratory parameters (platelets, CRP, albumin, hemoglobin) are more frequently normal in UC than in CD during mild to moderate flares ^(46,47). In contrast to adult UC, high sensitivity (hs)-CRP was not suitable to differentiate between remission and relapse in children with normal standard CRP ⁽⁴⁸⁾. In pediatric UC, ESR and CRP should be measured at least initially, since at times only 1 measure is elevated ⁽⁴⁹⁾. Initial albumin was the only significant laboratory test that was predictive for acute severe colitis in 1 follow-up study ⁽²³⁾. Similarly, earlier surgery was necessary in children with initially low serum albumin (HR 6.05, 99% CI 2.15–17.04) in 57 children who ultimately required colectomy (median time to surgery was 3.8 years) ⁽⁵⁰⁾. In another study, elevated white blood cell and low hematocrit measured at diagnosis were associated with colectomy rate at 3 years ⁽⁵¹⁾.

Fecal biomarkers reflect especially histological activity ^(52–54). High correlation of calprotectin with clinical disease activity, endoscopic, and histological indices has been described in both children and adults ^(52,55–59) (Table 2). In a retrospective pediatric study, calprotectin value of 275 μg/g achieved sensitivity and negative predictive value of 97% and specificity and positive predictive value of 85% in evaluating histological activity ⁽⁶⁰⁾. A few studies have indicated that calprotectin can be useful to predict relapses in UC patients ^(56,61,62), but its added predictive utility while in clinical remission is less clear (Table 3).

Roughly 60% of children with UC are pANCA positive at the time of diagnosis ⁽⁶³⁾. pANCA positivity was not associated with disease activity in 1 pediatric study ⁽⁶⁴⁾, or with early relapse (1-year follow-up) ⁽⁶⁵⁾. In a recent Porto group multicenter retrospective study of 801 children with colonic IBD, pANCA predicted the need for biologics in UC (P = 0.026) ⁽⁶³⁾. In an adult UC population, pANCA status was associated with higher risk of pouchitis after colectomy ⁽⁶⁶⁾.

Data supporting CRC surveillance recommendations can be found in extensive adult guidelines ^(15,16,67). Of note, a Swedish nationwide cohort study of pediatric IBD confirmed that CRC was almost non-existent during the first 5 years of follow-up, but incidence was higher after 10 years of follow-up ⁽⁶⁸⁾. Interestingly, the incidence of CRC in the first 20 years of follow-up was considerably lower in childhood-onset IBD than in disease with onset at other ages.

MEDICAL MANAGEMENT

5-ASA and Enemas

Recommendations

• 1. Oral 5-ASA compounds are recommended as first-line induction and maintenance therapy for mild-moderate UC [EL2, adults EL1]. (100% agreement)
• 2. Combined oral and rectal 5-ASA therapy is more effective than oral 5-ASA monotherapy [EL2, adults EL1]. (98% agreement)
• 3. Rectal monotherapy should be reserved for mild-to-moderate ulcerative proctitis, an uncommon pediatric phenotype [EL2, adults EL1]. (100% agreement)
• 4. When rectal therapy is used, 5-ASA is preferred over steroids [EL5, adults EL1]. (100% agreement)

Practice Points

• 1. No mesalamine delivery system has proven clearly superior for induction or maintenance of remission. Sulfasalazine may be somewhat superior to mesalamine for maintenance of remission in adult studies. Only sulfasalazine is available in liquid formulation and may be also effective for arthritis, but it is associated with more adverse events. (100% agreement)
• 2. Suggested dosing: oral mesalamine 60 to 80 mg · kg⁻¹ · day⁻¹ to 4.8 g daily; rectal mesalamine 25 mg/kg up to 1 g daily; sulfasalazine 40 to 70 mg · kg⁻¹ · day⁻¹ up to 4 g daily. Higher rectal doses up to 4 g are being used but evidence suggest that it is no more effective than 1 g. (98% agreement)
• 3. Suppositories are useful for limited proctitis, while foam and liquid mesalamine enemas are also suitable in more extensive colitis. (95% agreement)
• 4. Dosing 5-ASA once-daily can be considered for induction of remission and for maintenance. (95% agreement)
• 5. Gradual sulfasalazine dose augmentation over 7 to 14 days may mitigate against dose-dependent side-effects (see text). (93% agreement)
• 6. The effective induction dose should be continued also as the maintenance dose. Dose reduction, within the suggested dose range, may be considered after several months of sustained remission. Maintenance therapy should be continued in pediatric patients. (93% agreement)
• 7. Most children with mild-moderate UC will not achieve remission with oral mesalamine monotherapy alone. Treatment modification should be considered in those who do not show initial meaningful response within 2 to 3 weeks of therapy. (95% agreement)
• 8. Acute mesalamine intolerance could present as an exacerbation of the UC, usually within the first month of treatment. Symptoms resolve within days of cessation. Recurrence on re-challenge is diagnostic and precludes its future use. Symptoms usually recur also following rectal administration. (100% agreement)
• 9. Rectal tacrolimus may be considered in patients with ulcerative proctitis who are either refractory or intolerant to mesalamine and steroids topical therapies (suggested dose 0.07 mg · kg⁻¹ · day⁻¹; maximum dose in adult trials 3 mg/day). (88% agreement)

Strong evidence, mostly from adult trials, supports the use of 5-ASA for induction and maintenance of remission in mild-moderate UC ^(41,69–72). In that context, mesalamine induces remission in 35% to 55% of children, as defined by the PUCAI ^(73,74).

The MUPPIT trial randomized children with mild-moderate UC into once versus twice daily oral mesalamine (Pentasa), with comparable outcomes ⁽⁷⁴⁾. Once daily dosing achieved clinical response in 25/43 (60%) and remission in 13/43 (30%), compared with 25/40 (63%) and 16/40 (40%), respectively, for the twice daily group. Most responders did so by week 2 and no further response was seen after week 3. While the groups were statistically comparable, more patients in the once-daily study arm had pancolitis and were already on immunomodulators. Endoscopic remission was not assessed and the study was not powered for non-inferiority. Long-term studies of 5-ASA of once daily maintenance in pediatrics are currently ongoing. In another pediatric RCT, low versus higher dose balsalazide induced remission in 3/35 (9%) versus 4/33 (12%) of children, respectively ⁽⁷²⁾. Clinical improvement in mild-moderate UC was seen in nearly twice as many children randomized to sulfasalazine (22/28, 79%) compared to olsalazine (11/28, 39%) ⁽⁷¹⁾.

There are no pediatric maintenance comparative trials of 5-ASA, but only ∼ 40% (86/213) of children treated with 5-ASA within 1 month of diagnosis were in steroid-free remission by 1 year in the North American registry ⁽⁷⁵⁾. Similar data were reported from the prospective Italian pediatric IBD registry, with 45% of patients in remission at 1 year on 5-ASA therapy alone ⁽⁷⁶⁾. EPIMAD data reported that 32% (36/113) of children with UC remained on 5-ASA therapy without steroids, by maximum follow-up ⁽⁵⁾. In a recent Cochrane analysis of adult trials, the relative risk of successful induction of clinical and endoscopic remission with 5-ASA was 1.16 (95% CI 1.12–1.21) and 1.29 (95% CI 1.16–1.69), respectively ⁽⁶⁹⁾. No specific 5-ASA compound was superior for inducing remission, although sulfasalazine was statistically superior to other 5-ASA compounds for maintenance of remission ^{(69,70,77,78)}.

The pharmacokinetics of 5-ASA are comparable between children and adults ^(79–81). Adult trials have shown somewhat greater efficacy of higher induction mesalamine dose in patients with severe or extensive disease, phenotypes more commonly seen in children ^(82–84). In a multicenter RCT, 81 children with mild-moderate UC were, however, randomized to high dose (53–118 mg · kg⁻¹ · day⁻¹) or lower dose (27–71 mg · kg⁻¹ · day⁻¹) delayed release mesalamine with similar PUCAI-defined remission rates after induction (55% and 56% respectively) ⁽⁷³⁾. While greater reductions in fecal biomarkers were seen in the higher dose group, this did not reach significance. This trial enrolled on average children with milder disease which may explain the higher remission rates compared to other aforementioned pediatric trials.

Oral mesalamine may be better tolerated than sulfasalazine (relative risk of adverse effects 0.48, 95% CI 0.36–0.63), but the latter is cheaper and remains the only 5-ASA available in liquid formulation ^(70,71). Moreover, except for the uncommon allergic reaction (<0.1%), the vast majority of events are mild (eg, headache and gastrointestinal symptoms) ^(85,86). Serious adverse events with 5-ASA treatment are rare and include renal, pancreatic, pulmonary, and cardiac complications ^(87–93). Withdrawal due to intolerance in adult studies is in the range of 2% to 5% ^(69,70). Intolerance to 5-ASA medications may mimic a colitis flare, and when clinically proven by re-challenge, it precludes further use of 5-ASA compounds ⁽⁹⁴⁾. Regular laboratory monitoring of full blood count, renal function, and urinalysis, though not supported by evidence, remains the practice of many clinicians.

Rectal therapy (as suppositories) is indicated for ulcerative proctitis, an infrequent phenotype in pediatrics ⁽⁹⁵⁾. In order to allay concerns and ensure optimal compliance, children and their caregivers require support and reassurance when topical rectal therapies are proposed.

In a pediatric ulcerative proctitis trial, mesalamine suppositories (0.5 g daily) were associated with improved disease activity at 3 and 6 weeks in children with mild-moderate proctitis ⁽⁹⁵⁾. Combining oral and rectal 5-ASA therapy improves clinical outcomes ^(96–98). Remission was reported in 16/38 children (42%) in a prospective uncontrolled trial of 3 weeks’ rectal mesalamine in patients unresponsive to oral high dose mesalamine ⁽⁹⁹⁾. Adult studies with larger numbers and a higher evidence level have shown that rectal mesalamine foam, gel or liquid enema formulations have comparable tolerance, safety and outcomes ^(100–103). Once daily rectal therapy is as effective as divided daily dosing ⁽¹⁰⁴⁾. In adults, daily doses in excess of 1 g of rectal mesalamine do not enhance outcomes including clinical, endoscopic and histological remission ^(100,101). Rectal steroid preparations are useful for patients who are 5-ASA intolerant. They are superior to placebo in children and adults for inducing remission of proctitis, but meta-analysis data consistently support the superiority of rectal mesalamine over rectal steroids (symptomatic remission OR 1.65 [95% CI 1.1–2.45]) ^(100,101).

Rectal tacrolimus has been reported in children and adults as a successful third-line treatment of ulcerative proctitis ^(105,106). In a recent double-blind placebo-controlled trial, 8/11 adult patients receiving rectal tacrolimus ointment (1.5 mg twice daily) achieved mucosal healing by week 8, compared with 1/10 receiving placebo ⁽¹⁰⁷⁾. Although usually well tolerated, rare toxicity episodes have been reported ⁽¹⁰⁶⁾.

Oral Steroids

Recommendations

• 1. Oral steroids should be used as second-line treatment for mild-moderate UC not responding to 5-ASA (oral ± rectal) and may be considered as first line in the higher end of the moderate disease range [EL3, adults EL1]. (100% agreement)
• 2. Severe UC should normally be treated with intravenous steroids [EL2, adults EL1]. (98% agreement)
• 3. Second-generation oral steroids with lower systemic effect such as beclomethasone dipropionate (BDP) [EL2, adults EL1] and budesonide-MMX [EL5, adults EL2; the evidence for budesonide-MMX is supportive only for left-sided colitis] may be considered in patients with mild disease refractory to 5-ASA before oral prednisolone. (93% agreement)
• 4. Steroids are not recommended for maintaining remission; steroid sparing strategies should be applied [EL5, adults EL4]. (100% agreement)

Practice Points

• 1. Regarding recommendation #2, a short trial of oral steroids could be considered in selected children with severe colitis (ie, PUCAI ≥ 65) who appear well with normal or near-normal lab values. (93% agreement)
• 2. The recommended daily dose for oral prednisolone/prednisone is 1 mg · kg⁻¹ · day⁻¹ (max 40 mg) once daily for 2 to 3 weeks followed by a tapering period of up to 8 to 10 weeks (Table 4). (98% agreement)
• 3. Once daily administration of steroids in the morning is as effective as the same dose given in multiple divided doses. (100% agreement)
• 4. In patients >30 kg the dosing schedule of BDP is 5 mg once daily for 4 weeks and for budesonide-MMX 9 mg for 8 weeks. Dosing for children <30 kg has not been established and no liquid formulation is available. There is no evidence to support tapering of either drug. While abrupt discontinuation has been practiced in the RCTs, alternate day tapering over 2 to 4 weeks has been proposed by some. (93% agreement)
• 5. The term “steroid-dependency” applies to patients who are unable to stop steroids within 3 months without recurrent active disease, or who have a relapse requiring steroids within 3 months of stopping steroids. (95% agreement)
• 6. High glucocorticoid dose and long duration of the therapy (>3 months) has been associated with adrenal suppression (ie, present after gradual weaning off) in 20% of children with IBD. (98% agreement)
• 7. If symptoms of adrenal suppression (eg, weakness/fatigue, malaise, nausea, vomiting, diarrhea, headache, arthralgia, and abdominal pain) are present while weaning steroids, adrenal insufficiency should be excluded by first testing cortisol level at 08:00 am before drug intake, and, if abnormal, consult with a pediatric endocrinologist. (93% agreement)

Studies of oral steroids for treating children with active UC report short-term (1-3 months) remission rates of 50% to 64% ^(108–110); at 1 year 49% to 61% had prolonged response, 14% to 49% were steroid-dependent, and 5% to 29% required surgery ^{(5,7,108,110)}. Mucosal healing lags behind clinical improvement; in a non-randomized study after 8 weeks of steroids or 5-ASA, 87% had clinical remission, 40% endoscopic remission and 15% histological remission with no significant difference in outcomes between the 2 therapies ⁽¹¹¹⁾.

TABLE 4: Steroids tapering schedule (doses are in mg/day prednisone equivalent): the goal is to discontinue steroids by week 10

Steroid dependency has been reported to be higher in children than in adults (45% vs 8%, respectively) ⁽⁷⁾. Strategies to avoid steroid dependency include optimization of 5-ASA, adjuvant therapy with enemas, and escalation to thiopurines or biologics.

Second-generation topical steroids have a more favorable safety profile and may be considered before systemic steroids in selected patients ⁽¹¹²⁾. BDP uses gastro-resistant film coatings to target delivery to the distal small intestine and the colon. Studies in adults demonstrate the effectiveness of BDP compared with both prednisolone and mesalamine ^(15,113). An RCT of 30 children (weight >30 kg) with mild-to-moderate UC showed that oral BDP, 5 mg/day for 4 weeks, was well tolerated and more effective than 5-ASA in achieving both clinical remission (80% vs 33%, P < 0.025) and endoscopic remission (73% vs 27%, P < 0.025), respectively ⁽⁴¹⁾.

A Cochrane systematic review of older selective release budesonide in adults showed that it was less likely to induce remission than mesalamine (relative risk [RR] 0.72, 95% confidence interval [CI] 0.57–0.91) with no benefit over placebo (RR 1.41, 95% CI 0.59–3.39) ⁽¹¹⁴⁾. Budesonide-MMX is a novel oral formulation designed to extend release of the drug to the colon. Two adult trials showed significant benefit in the intention to treat (ITT) population (combined left sided and extensive), but subanalysis based upon disease extent was only significant for left-sided disease ^(115,116). Indeed, in a recent case series of 16 children, 15 of whom with pancolitis, budesonide-MMX showed minimal clinical effectiveness ⁽¹¹⁷⁾. Another recent RCT in adult UC refractory to 5-ASA showed superiority over placebo but with a disappointing 6% effect size difference; no subgroup analysis of disease extent was performed ⁽¹¹⁸⁾.

Children with UC may have more steroid-related complications, including osteopenia, acne, glaucoma, and cataracts, than adults even when adjusted for weight ⁽¹¹⁹⁾. Even low steroid doses (0.1–0.4 mg · kg⁻¹ · day⁻¹) can suppress growth ⁽¹²⁰⁾. The ECCO statement in adults suggests supplementation with vitamin D while on steroid therapy ⁽¹²¹⁾, but we could not find clear evidence to support supplementing vitamin D in those who are not deficient.

AS may present with non-specific symptoms (including abdominal pain, malaise, weakness/fatigue, nausea, diarrhea, headache, fever, arthralgia) or rarely adrenal crisis (hypotension, lethargy, decreased consciousness/coma, hyponatraemia, hypoglycemia, seizures) ⁽¹²²⁾. There are no published consensus guidelines that advise who should be screened for AS. In 1 recent review, it was recommended to screen patients who received steroids for >3 weeks and after gradual weaning have persistent symptoms that may be attributable to AS ⁽¹²³⁾. The range of 8 am morning cortisol value at which AS is confirmed varies between studies. In 1 recent review a value of <100 nmol/L was used while >500 nmol/L virtually excluded AS ⁽¹²⁴⁾. Another manuscript suggested that <85 nmol/L should be used to diagnose AS ⁽¹²³⁾. In a study of consecutive children with IBD about to stop steroids (ie, on physiological doses of oral steroids meaning 5 to 10 mg daily prednisolone) 20% had biochemical AS using a value <69 nmol/L and of these half had an undetectable cortisol ⁽¹²⁵⁾. Higher glucocorticoid dose and longer duration of the therapy were associated with increased risk ⁽¹²⁵⁾. In the only study of children with IBD, all children treated with steroids for <3 months did not have biochemically confirmed AS (after gradually weaning to physiological doses of steroids) ^(125–127).

Immunomodulators

Recommendations

• 1. Thiopurines are recommended for maintaining remission in children who are corticosteroid-dependent or relapsing frequently (≥2 relapses per year) despite optimal 5-ASA treatment and in 5-ASA intolerant patients [EL3, adults EL1]; thiopurines should be considered following discharge from acute severe colitis episode [EL4, adults EL3]. (98% agreement)
• 2. Thiopurines should not be used for induction of remission in pediatric UC patients [EL5, adults EL2]. (100% agreement)
• 3. Measuring thiopurine metabolites is recommended in patients with incomplete response on a stable thiopurine dosage, in patients who present with leucopenia or elevated transaminases, or if poor compliance is suspected [EL2, adults EL2]. (95% agreement)

Practice Points

• 1. Thiopurines may be somewhat more effective than 5-ASA for maintaining remission in UC, but considering their safety profile, they should generally be reserved as second-line therapy after 5-ASA has failed. (93% agreement)
• 2. Determination of TPMT genotype or phenotype (ie, TPMT activity) is encouraged to identify patients at greater risk of profound myelosuppression. Dose should be reduced in heterozygous patients or in those with low activity. Thiopurines should not be used in children homozygous mutants for TPMT or those with very low TPMT activity as defined at each laboratory. (93% agreement)
• 3. Regular monitoring of blood counts and liver enzymes is recommended in all cases every 1 to 2 weeks during the first month then every month up to 3 months followed by every 3 months thereafter. (100% agreement)
• 4. Families should be instructed to use sun protection with the use of thiopurines and other immunosuppressive drugs. (100% agreement)
• 5. Given its excellent safety profile, it is reasonable to continue 5-ASA with thiopurines, at least initially, despite lack of evidence. 5-ASA inhibits the enzyme TPMT thus increasing the active metabolite 6-thioguanine (6-TGN). (88% agreement)
• 6. The maximal therapeutic effect of thiopurines may not be evident until 10 to 12 weeks of treatment. (98% agreement)
• 7. Thiopurine dose should be approximately 2 to 2.5 mg/kg of azathioprine and 1 to 1.5 mg/kg of mercaptopurine, in a single daily dose in patients with a normal TPMT. Measuring thiopurine metabolites may assist in further dose adjustments and reduce adverse events while considering 6-TGN level of 235 to 450 pmol/8 × 10⁸ RBCs and 6-methylmercaptopurine ribonucleotides (6-MMP) < 6700 pmol/8 × 10⁸ RBCs as optimal (note that cutoff values may vary between methods ⁽¹²⁸⁾). (95% agreement)
• 8. Patients who show gastrointestinal intolerance or flu-like reaction to 1 thiopurine compound may tolerate lower doses or a “switch” to another thiopurine (azathioprine to 6-mercaptopurine and vice versa). Limited data suggest that splitting the daily dose into 2, may alleviate gastrointestinal and hepatic toxicity in patients who have hyperactive TPMT activity. (95% agreement)
• 9. Thiopurines should be discontinued in clinically significant myelosuppression or pancreatitis. Reintroduction of thiopurines after leucopenia (but not usually pancreatitis) can be considered at a lower dose after carefully assessing the risks and benefits and after measuring thiopurine metabolites and/or TPMT. (95% agreement)
• 10. Change in treatment should be considered in patients with active disease despite adequate 6-TGN level after at least 12 weeks of thiopurine treatment. (98% agreement)
• 11. Concomitant use of allopurinol) 50 mg once daily in patients <30 kg and 100 mg once daily is patients ≥30 kg, maximum 5 mg/kg) with reduced dose of azathioprine (to approximately 25% to 30% of initial dose), may provide a valid therapeutic option in cases of hyperactive TPMT resulting in high 6-MMP (often associated with elevated transaminases) and low 6-TGN, in suitably experienced units. Children must be closely monitored given the increased risk of toxicity. (95% agreement)
• 12. Benefits of withdrawal should be carefully weighed against an increased risk of relapse. Thiopurine withdrawal could be considered in patients in sustained clinical remission following long-term treatment (at least 1 year) after ensuring complete mucosal healing and preferably also histological remission. In the case of thiopurine withdrawal, 5-ASA treatment may assist in maintaining remission (particularly in patients naïve to 5-ASA). (91% agreement)
• 13. Methotrexate may rarely be considered in UC patients who fail to respond or are intolerant to thiopurines, when other alternatives are not possible or available. (91% agreement)
• 14. Oral tacrolimus (FK-506) may be considered in selected outpatient UC children as another option to steroids for bridging to thiopurines or vedolizumab (given the longer time to onset of action). At initiation, high target serum trough level (10–15 ng/mL) should be achieved with a gradual titration to lower trough levels (5–10 and eventually 2–5 ng/mL) in order to avoid serious adverse events. Selected patients may benefit from a long-term, low-dose treatment (ie, drug level target of 2 ng/mL), but the potential toxicity should be carefully considered, as well as noting the limited supportive evidence. (93% agreement)

The efficacy of thiopurines (azathioprine and 6-mercaptopurine) was evaluated systematically for both induction and maintenance of remission in adult UC patients. Meta-analyses of adult data concluded that azathioprine is not more effective than placebo for induction of remission but is superior to placebo in preventing relapse ^(129–131). In a recent prospective cohort study, sustained clinical benefit was achieved in 60% of 255 adult UC patients receiving azathioprine following 5-ASA failure, at a median follow-up of 30 months ⁽¹³²⁾.

Prospective pediatric studies reported steroid-free remission rates of 49% at 1 year ⁽¹³³⁾ and 72% at 2 years ⁽¹³⁴⁾ in thiopurine treated children with no difference in either clinical or endoscopic end-points between early or late initiation of treatment. A few retrospective studies ^(135–138) in children supported the benefit of thiopurines in maintaining remission and steroid sparing with a median time to achieve steady state of thiopurine levels of 55 days ⁽¹³⁹⁾. Cox proportional hazard modeling of retrospective data from 1175 incident children and young adults, did not demonstrate a benefit to early thiopurine use in reducing the risk of colectomy ⁽¹⁴⁰⁾.

Despite 1 negative small adult study ⁽¹⁴¹⁾, it is not unreasonable to combine 5-ASA with thiopurines given the excellent safety profile of the former and its possible additive effect, including chemoprotection. 5-ASA may partially inhibit TMPT activity and therefore may increase 6-TGN levels ^(142,143).

Most adult studies used doses of 2.5 mg/kg for azathioprine and 1.5 mg/kg for 6-MP. There was, however, no clear dose-response effect for azathioprine, implying that low-dose azathioprine (1.5 mg/kg) may not be inferior to standard dose ⁽¹⁴⁴⁾. Children younger than 6 years may require higher doses of azathioprine per body weight with doses of up to 3 mg · kg⁻¹ · day^{−1(145,146)}.

The relative risk of serious adverse events with thiopurines was found to be 2.82 in a Cochrane meta-analysis of adult data ⁽¹³⁰⁾. Thiopurine withdrawal rate due to adverse events in large pediatric cohorts was 18% ⁽¹⁴⁷⁾ and 30% ⁽¹⁴⁸⁾. Dose-independent adverse reactions include fever, pancreatitis, rash, arthralgias, nausea, vomiting, and diarrhea, while dose-dependent toxicities included leucopenia (up to 5%), thrombocytopenia, infections, and hepatitis ^(149,150). A meta-analysis of studies of 6-MP found that it was tolerated in 68% of 455 adult patients who were azathioprine-intolerant, lending support to switching between these drugs in cases of specific dose-independent adverse events ⁽¹⁵¹⁾. Switching in the case of pancreatitis has traditionally not been recommended but some recent case series have challenged this notion ⁽¹⁵¹⁾.

A meta-analysis (25,728 IBD patient-years) demonstrated that patients younger than 30 years have a high relative risk for non-Hodgkin lymphoma (SIR = 6.99) with younger men being at the highest risk. However, the absolute risk is much higher in the elderly. In patients younger than 30 years, the absolute risk is estimated at only 1 in 4000 to 5000 ⁽¹⁵²⁾. Hepatosplenic T-cell lymphoma (HSTCL) is a very rare but fatal complication of thiopurine therapy. Of over 40 reported cases of IBD-related HSTCL, almost all received thiopurines, with or without anti-TNF and almost all were males; there are only extremely rare and anecdotal case reports of children with HSTCL who were treated solely with anti-TNF ^(153,154).

TPMT assay (either phenotype or genotype) can be used before initiation of thiopurines to identify some of the patients who are at risk for dose-dependent myelosuppression, and in whom this drug should either not be used (if homozygous for variant alleles or have very low TPMT activity) or administered at lower dosage (if heterozygous for variant alleles or having low TPMT activity). TPMT testing does not, however, replace the need for mandatory monitoring of complete blood cell count especially during initiation of treatment. In an adult study ⁽¹⁵⁵⁾, a significantly smaller proportion of carriers of a TPMT variant with adjusted dose developed hematologic adverse events (RR = 0.11). In a pediatric study, 7 of 46 (15%) carriers of at least 1 variant allele or low/intermediate TPMT activity developed myelosuppression compared to 0/62 in the wild type/high activity group ⁽¹⁵⁶⁾. In contrast, a study of 72 children showed no association between TPMT polymorphisms and the occurrence of thiopurine-related adverse events ⁽¹⁵⁷⁾.

In the case of hyperactive TPMT resulting in high 6-MMP and low 6-TGN, concomitant use of allopurinol with reduced dose of azathioprine may provide a valid therapeutic option ^(158,159) but needs to be used with caution. Adequate dose reduction and repeated monitoring of CBC and 6-TGN/6-MMP is essential to avoid myelosuppression related side effects. In adult trials, allopurinol was used at 100 mg once daily ^(158,160) whereas in the few pediatric case series lower doses (50 or 75 mg once daily) were utlized in younger children ^(159,161).

Therapeutic drug monitoring, namely measurement of thiopurine metabolites, specifically 6-TGN and 6-MMP levels, has been implemented as a means of optimizing efficacy and avoiding myelosuppression. In a meta-analysis which included 1026 IBD children ⁽¹⁶²⁾, higher 6-TGN concentrations were not consistently associated with leucopenia, while marginally associated with greater likelihood of clinical remission. High 6-MMP levels correlated with hepatotoxicity, and low thiopurine metabolite levels with non-compliance. In a retrospective study including 86 IBD children, 6-TGN levels of >250 pmol per 8 × 10⁸ red blood cells correlated with a higher response rate (OR = 4.14) ⁽¹⁶³⁾. The association between both bone marrow toxicity and clinical response with 6-TGN levels was demonstrated in prospective adult studies ^(164,165) as well as several retrospective pediatric cohort studies ^{(163,166,167)}. Dose adjustment following measurement of metabolites was reported to increase disease remission rate ⁽¹⁶⁸⁾. Children with IBD were shown to experience fewer exacerbations when thiopurine metabolites were measured ⁽¹⁶⁹⁾. In a study of 78 IBD children, 6-TGN level above 405 pmol/8 × 10⁸ RBCs was the only predictor for azathioprine resistance (OR 10.8) implying that patients with active disease and adequate 6-TGN level should receive alternative therapies ⁽¹⁷⁰⁾.

Thiopurine withdrawal after attaining sustained remission is controversial. In a retrospective study of 127 UC patients in remission, approximately one-third relapsed within 12 months following withdrawal, and two-thirds within 5 years ⁽¹⁷¹⁾. Moderate/severe relapse rate of 26% at 2 years was observed in 108 UC patients who withdrew treatment following prolonged thiopurine treatment ⁽¹⁷²⁾.

Cochrane meta-analyses of methotrexate (MTX) for induction (2 RCTs, 101 patients) ⁽¹⁷³⁾ or maintenance (3 RCTs, 165 patients) ⁽¹⁷⁴⁾ of remission in adult UC concluded that there is no evidence supporting the use of MTX for either induction or maintenance of remission in UC. Nevertheless, this conclusion relies on low-quality evidence. In the METEOR double-blind, placebo-controlled trial of 111 steroid-dependent UC adults, steroid-free remission at week 16 was not statistically different than placebo (32% vs 20%, respectively; P > 0.05) though clinical remission did differ (42% vs 24%, respectively; P = 0.04) ⁽¹⁷⁵⁾. In a retrospective study of 32 UC children unresponsive or intolerant to thiopurines, response/remission was achieved in 72%, 63% and 50% of patients treated with parenteral MTX at 3, 6, and 12 months, respectively ⁽¹⁷⁶⁾.

Tacrolimus has been studied in ambulatory UC patients ⁽¹⁷⁷⁾. An RCT comparing high target trough level of tacrolimus (10–15 ng/mL) versus low trough level (5–10 ng/mL) versus placebo in adult moderate-to-severe UC patients who were hospitalized for the study, reported a significantly higher response rate in the high trough group (68% vs 38% vs 10%, respectively) ⁽¹⁷⁸⁾. A retrospective cohort study of 25 ambulatory moderate-to-severe adult UC patients reported 52% clinical improvement and 44% clinical remission at 6 months ⁽¹⁷⁹⁾. Three small retrospective pediatric studies, including 18 steroid refractory/dependent UC patients ⁽¹⁸⁰⁾, 10 ⁽¹⁸¹⁾, and 8 ⁽¹⁸²⁾ steroid-resistant patients treated with tacrolimus, reported 50% to 95% response rate; however, colectomy was eventually performed in most patients during the follow-up period. In a subgroup analysis, steroid-dependent patients had a significantly higher long-term colectomy free rate when compared with steroid refractory patients (78% vs 0%) ⁽¹⁸⁰⁾.

Biologics

Recommendations

• 1. Infliximab (IFX) should be considered in chronically active or steroid-dependent UC, uncontrolled by 5-ASA and thiopurines, for both induction and maintenance of remission [EL2, adults EL1]. (100% agreement)
• 2. Adalimumab [EL4, adults EL4] or golimumab [EL4, adults EL3] could be considered in those who initially respond but then lose response or are intolerant to IFX, based on serum levels and antibodies (Fig. 4). (95% agreement)
• 3. Adalimumab and golimumab have no role in patients with primary non-response to IFX [EL4, adults EL4]. (93% agreement)
• 4. Vedolizumab should be considered in chronically active or steroid-dependent patients as second-line biologic therapy after anti-TNF failure [EL4, adults EL2]. (95% agreement)

FIGURE 4: Practical interpretation of drug levels and antibodies for infliximab and adalimumab. Different countries use different measuring kits with different cutoff values; absolute drug and antibodies levels should be adapted accordingly. (1) Some studies (and the AGA recommendation in adults ⁽²¹⁰⁾) suggest that higher levels for infliximab (>5 μg/mL) and adalimumab (>8 μg/mL) should be the goal (see text).

Practice Points

• 1. Screening for latent tuberculosis with combination of patient history, chest x-ray, tuberculin skin test, or interferon-gamma release assays (quantiferon) is essential before initiating anti-TNF. The quantiferon test is preferred in patients under immunosuppressive therapy and in BCG immunized patients. Screening for hepatitis B and C viruses, varicella zoster virus, and HIV when appropriate, is also recommended if not done recently. (95% agreement)
• 2. In ambulatory patients with UC, IFX should be administered initially at 5 mg/kg per dose (at weeks 0, 2, and 6 followed by 5 mg/kg every 8 weeks for maintenance). Higher initial dosing should be considered in children with low body weight (<30 kg) or high BMI, and in the presence of higher inflammatory burden and hypoalbuminemia. Target trough levels post induction (week 14) and subsequent doses are reported in different studies as >4 to 5 μg/mL. Rapid infusion (over 1 hour) seems as safe and effective as traditional slower infusions, if the induction doses were well tolerated and dose is stable. (98% agreement)
• 3. IFX is recommended to be used preferably in combination with an immunomodulator (IMM) (with the most evidence in UC being thiopurines) in order to reduce the likelihood of developing antibodies to IFX (ATI) and in thiopurine-naïve patients to enhance effectiveness. Discontinuation of the IMM may be considered after 6 months, especially in boys, preferably after ensuring trough IFX level >5 μg/mL, since levels may decrease after stopping IMM. (98% agreement)
• 4. The utility of combination adalimumab, golimumab, and vedolizumab with thiopurines is more controversial and they are most commonly prescribed as monotherapy in children. (100% agreement)
• 5. Golimumab recommended doses for induction are 200 mg at week 0 followed by 100 mg at week 2 for those weighing ≥45 kg. Children with lower weight should be dosed based on body surface area (115 and 60 mg/m² at weeks 0 and 2). Maintenance doses q4w are 60 mg/m² if weight <45 kg and 100 mg if weight ≥45 kg. Target trough levels during maintenance are >2 μg/mL. (100% agreement)
• 6. Extrapolating from pediatric Crohn disease, adalimumab should be started at 160 mg, followed by 80 mg after 2 weeks and then 40 mg every other week in adolescents with weight >40 kg. Optimal dosing in younger children has not been well defined, but BSA-based dosing could be considered taking as a base an adult BSA of 1.73m² (ie, induction with 92 mg/m² followed by 46 mg/m² followed by 23 mg/m² every other week for maintenance). Adalimumab target levels during maintenance are reported in different studies as >5 to 8 μg/mL. (100% agreement)
• 7. Measurement of drug levels and anti-drug antibody levels following induction (ie, at the week 14 infusion for IFX and at 8–10 week for adalimumab) can assist in optimizing treatment. Measuring drug levels is also useful in the assessment of unsatisfactory response to anti-TNF to guide dose escalation or a switch to another biologic (see text). (98% agreement)
• 8. Standard vedolizumab dosing in adults has been adapted in pediatric studies (5 mg/kg up to 300 mg per dose at weeks 0, 2, 6 followed by every 8 weeks thereafter). For those weighing <30 kg, higher dose per kg is required, but BSA-based calculation may be preferred (ie, 177 mg/m²). The effect of vedolizumab in UC has been described to occur by week 6 of treatment, but complete response may not be apparent until week 14. Shortening of interval between infusions to 4 weekly may be required during maintenance in partial responders. (93% agreement)
• 9. In patients with persistent symptomatic distal inflammation despite adequate optimal anti-TNF treatment, addition of rectal therapies (preferably 5-ASA) could be beneficial. (98% agreement)

A Cochrane systematic review of 7 adult UC trials concluded that IFX is effective in inducing clinical remission, promoting mucosal healing, and reducing the need for colectomy in patients with active UC ⁽¹⁸³⁾. Combination therapy with IFX and azathioprine was shown to be superior in the SUCCESS trial in adult UC to monotherapy with azathioprine or IFX alone, while there was no superiority of IFX monotherapy over azathioprine ⁽¹⁸⁴⁾.

In the pediatric UC regulatory RCT (ie, the T-72 study), 45 of 60 (75%) ambulatory children with moderate-severe UC responded to a standard induction protocol of IFX ⁽⁴⁰⁾. Both clinical remission (PUCAI < 10 points) and complete mucosal healing (Mayo endoscopic subscore = 0) were achieved in 33% at week 8. Dose escalation to 10 mg/kg was required in 44% of the patients in the maintenance phase.

Different studies in children have shown a pooled long-term success rate of IFX in UC of 64% ⁽¹⁸⁵⁾, and a corticosteroid-free remission of 38% and 21% at 12 and 24 months, respectively, with a likelihood of avoiding colectomy at 2 years of 61% ⁽¹⁸⁶⁾. A relationship between the increased use of anti-TNF agents and the reduction of surgery risk for UC children has also been suggested ⁽¹⁸⁷⁾.

Adalimumab has shown efficacy and safety for induction and maintenance in the adult moderate-to-severe active UC. In the adult ULTRA-1 trial, clinical remission was obtained in 18.5% of patients in the 160/80 mg group, 10% in the 80/40 mg group and 9.2% in the placebo group ⁽¹⁸⁸⁾. In the ULTRA-2 trial, overall rates of clinical remission for active drug at week 52 were 17.3%, with better results among anti-TNF-naïve patients (22%) as compared to those anti-TNF experienced (10.2%) ⁽¹⁸⁹⁾. A network meta-analysis of 5 RCTs in moderate-to-severe adult UC suggested that while IFX is more effective than adalimumab in the induction of remission, response and mucosal healing, both are comparable in efficacy at 52 weeks of maintenance treatment ⁽¹⁹⁰⁾. Another meta-analysis showed superiority of IFX over adalimumab in inducing and maintaining endoscopic healing in UC ⁽¹⁹¹⁾. In a propensity score adjusted analysis, a study of 419 adults with UC found no difference in the effectiveness of these agents, but the adalimumab group was relatively small ⁽¹⁹²⁾.

In a retrospective cohort study of 188 children, Vahabnezhad et al showed that 60% of UC children who discontinued IFX were commenced on adalimumab, with 83% of these remaining on adalimumab at last follow-up ⁽¹⁹³⁾. In another retrospective study, 55% of UC children switched to adalimumab after IFX failure, achieved and maintained clinical remission at a median of 25 months while 36% underwent colectomy ⁽¹⁹⁴⁾. There are no published data on adalimumab in UC children naïve to anti-TNF.

A second subcutaneously administered, fully human anti-TNF agent, golimumab, has been studied in placebo-controlled trials among anti-TNF-naïve adults with moderately-severely active UC in the PURSUIT-SC ⁽¹⁹⁵⁾ for induction, and PURSUIT-M ⁽¹⁹⁶⁾ for maintenance. Golimumab use in pediatric UC was studied in an open-label pharmacokinetic study of 35 children with moderate-severe UC ^(197,42). Doses given subcutaneously at weeks 0 and 2 were 90 and 45 mg/m² for children weighing <45 kg and 200 mg followed by 100 mg for those weighing ≥45 kg. Maintenance doses of 45 mg/m² if weight <45 kg and 100 mg if weight ≥45 kg were given every 4 weeks. Among week 6 Mayo clinical responders (60%) who continued to receive 4 weekly golimumab maintenance, 57% were in PUCAI remission at week 14. Complete mucosal healing at week 6 was achieved in 23%, slightly higher than reported in the adult trials. While the pharmacokinetics data of the entire pediatric cohort were comparable with those previously reported in the golimumab adult trials, drug levels in the subgroup of children weighing <45 kg were numerically lower than those ≥45 kg. This likely stems from the under-dosing of the former group. The equivalent dosing of 200 mg in adults and adolescents would translate to 115 mg/m² in BSA (considering 200 mg/1.73 m²) followed by 60 mg/m² for maintenance. Given the lower drug levels in the pediatric study, these higher doses should be considered in practice.

In general, response to anti-TNF medication can occur as early as 1 to 4 weeks and peaks by week 12 to 16 of treatment ^{(188,189,198)}. During induction, trough level of ≥15 μg/mL at week 6 best predicted likelihood of short-term mucosal healing (area under the ROC of 0.69) ⁽¹⁹⁹⁾. Recommended optimal levels for IFX during maintenance therapy for improved clinical outcomes has been defined as >4 μg/mL ^(200–202), for adalimumab >5 μg/mL ^(203,204), and for golimumab >1.4 μg/mL ⁽²⁰⁵⁾. For mucosal healing, adult studies from both UC ^(206,207) and CD ^(207,208), however, suggested that higher adalimumab level ≥7.1 to 9.4 μg/mL may be more appropriate. Similarly, IFX trough levels >5 μg/mL were associated with mucosal healing in adult IBD ⁽²⁰⁷⁾ and with a decreased risk for loss of response when withdrawing concomitant immunomodulators ⁽²⁰⁹⁾. The American Gastroenterological Association guidelines thus recommend using higher cutoff values of ≥5 μg/mL for IFX and ≥7.5 μg/mL for adalimumab ⁽²¹⁰⁾.

Drug and antibody levels should dictate the course of action in patients with secondary loss of response ⁽²¹¹⁾ (Fig. 4). Ongoing symptoms despite adequate drug levels, mandates switching therapy “out of class.” High antibodies titer predicts failure of dose intensification ⁽²¹¹⁾ (Fig. 4).

Factors predicting lower drug levels (and thus possibly dictating higher dosing) include higher body mass index ⁽²¹²⁾, low body weight <30 kg ^(213–215), male gender ⁽²¹⁶⁾, high inflammatory burden (extent and severity of disease) ⁽²¹⁷⁾, hypoalbuminemia ⁽²¹⁸⁾, the presence of anti-drug antibodies, and the absence of a concomitant immunomodulator ^{(184,219–221)}.

Safety issues of anti-TNF include acute infusion reactions (within 4 hours of infusion), delayed hypersensitivity reactions (beyond 4 hours in both lines of infusion and up to 14 days), serious and opportunistic infections ⁽²²²⁾, and a potential risk of skin cancer; evidence to date does not indicate that anti-TNF is associated with lymphoma if prescribed as mono-therapy, but a recent study challenges this concept ⁽²²³⁾. Psoriasis has been well documented as an adverse class effect of anti-TNF, but it is usually mild and controllable in the majority of patients with topical therapy ⁽²²⁴⁾. Other very rare adverse events, such as demyelination events and optic neuritis, have been reported ⁽²²⁵⁾.

There is no clear evidence that pre-medication with any drug prevents the development of acute infusion reaction ^(226,227). A self-reporting system in the United States with >5000 documented patients calculated a rate of infusion reactions of 3% (1.1% immediate and 1.7% delayed) in IBD-treated patients ⁽²²⁸⁾.

Required infectious screening before initiation of anti-TNF treatment includes testing for HBV, HCV, HIV, VZV, and tuberculosis according to local prevalence and national recommendations ⁽²²⁹⁾. The risk of reactivation of other viruses (eg, CMV, EBV) is not clear. A recent systematic review and meta-analysis including 49 RCT comprising >14,000 patients treated with biologics (anti-TNF, natalizumab, and vedolizumab) concluded that their use has a moderate risk of any infection (OR 1.19 (95% CI 1.1–1.29)) and a significant risk of opportunistic infections in IBD (OR 1.90 (1.21–3.01)) ⁽²³⁰⁾. In another study, the estimated risk of severe infections in IBD patients treated with anti-TNF has been reported as 2.2% ⁽²³¹⁾. Concomitant immunosuppressant treatment, particularly steroids, is an additional risk for opportunistic and other infections. Surprisingly, the meta-analysis found a reduced risk of serious infections (OR 0.56 (0.35–0.9)) and no increased risk of malignancies (OR 0.9 (0.54–1.5)), but for the latter outcome the data were insufficient in terms of exposure and follow-up period ⁽²³⁰⁾. Studies report conflicting results regarding the risk of anti-TNF and the risk for melanoma and non-melanoma skin cancer ^(232,233).

DEVELOP is a prospective post-marketing industry-initiated safety registry for pediatric IBD, which includes both patients exposed and never exposed to IFX ⁽²³⁴⁾. In 5766 patients (29% UC; 24,543 patient years follow-up; median 4.5 years per patient follow-up) there were 15 malignancy events (13 exposed to thiopurines (10 with IFX; 3 thiopurine only); 1 only to IFX; 1 to neither biologics nor thiopurines). Comparison with rates from the SEER database of healthy controls indicated a standardized incidence rate (SIR) for neoplasia of 2.43 (95% CI 1.29–4.15) for thiopurine exposure (with or without biologic exposure), but no significant increase in neoplasia with IFX exposure in the absence of thiopurine exposure (SIR 1.49, 95% CI 0.04–8.28). Five children in total experienced hemophagocytic lymphocytic histiocytosis (HLH), 4 with primary EBV infection, one with CMV infection, and all during thiopurine monotherapy.

Vedolizumab is a humanized anti-α4β7 integrin that downregulates intestinal inflammation by specifically inhibiting intestinal T-lymphocyte migration into the tissue. In the adult GEMINI-1 study in UC, 47% of patients responded to 2-dose induction (300 mg per dose) by week 6 and were re-randomized to continued vedolizumab 300 mg intravenously (4 weekly vs 8 weekly vs placebo). The 52-week remission rates among initial week 6 responders were 42% (q8w) and 45% (q4w) ⁽²³⁵⁾, regardless of the prior anti-TNF exposure status ^(236,237). This is supported by pharmacokinetic data demonstrating significant correlation between higher vedolizumab drug levels and clinical response in IBD patients ^(238–240).

Experience with vedolizumab in pediatric UC is currently limited to small retrospective cohorts, almost all with prior anti-TNF failure. The 14-week remission rates were 37% (n = 41 definition of remission included steroid-free and utilized ITT rates) ⁽²⁴¹⁾, 40% (n = 5; ⁽²⁴²⁾), and 76% (n = 22 ⁽²⁴³⁾). The 22-week corresponding remission rates in the 3 studies were 34%, 40%, and 71%, respectively. There is no evidence that combination therapy with IMM is superior over sole vedolizumab treatment based on very limited data from adults ⁽²⁴⁴⁾ and children ⁽²⁴¹⁾.

Limited safety data are reported for vedolizumab in children. Conrad et al reported 29 adverse events in children including upper respiratory tract infections, nausea, fatigue, headaches, nasopharyngitis, skin infections, and sinusitis ⁽²⁴²⁾. Pruritus, infusion reaction, and nasopharyngitis (1 each) was also reported by the Paediatric IBD Porto group of ESPGHAN ⁽²⁴¹⁾.

In a recent network meta-analysis in adults, IFX, adalimumab, golimumab and vedolizumab were all superior to placebo for maintenance of remission and response; however, superiority of 1 agent over another could not be clearly established ⁽²⁴⁵⁾.

Other Interventions

Recommendations

• 1. Granulocyte/monocyte apheresis should not be routinely used in pediatric UC [EL4, adult EL2]. (100% agreement)
• 2. Fecal microbiota transplantation (FMT) should not be routinely used in pediatric UC [EL4, adult EL1]. (100% agreement)
• 3. Antibiotics should not be routinely used for induction or maintenance of remission of pediatric UC [EL5, adult EL2]. (100% agreement)
• 4. Probiotic agents (eg, VSL#3 [Table 5], Escherichia coli Nissle 917) may be considered in mild UC as an adjuvant therapy or in those intolerant to 5-ASA [EL2, adult EL2]. (100% agreement)
• 5. Curcumin may be considered as an add-on therapy for inducing and maintaining clinical remission of mild-to-moderate UC [EL4, adult EL1]. (91% agreement)
• 6. Germinated barley foodstuff, omega-3, aloe vera, herbal medicine, and intravenous immunoglobulin, are not recommended as primary treatment [EL5, adult EL2]. (98% agreement)

Practice Points

• 1. If apheresis is considered, then the most commonly utilized scheme involves 1 session per week of granulocyte/monocyte apheresis for 5 to 10 consecutive weeks. (93% agreement)
• 2. VSL#3 dosing may be seen in Table 5. E coli Nissle 1917 strain is prescribed as 200 mg/day in adults and adolescents. No dosing recommendation is available for young children. (98% agreement)
• 3. Neither the formulation nor dosage of curcumin (the active ingredient of tumeric/curcum) are established for children but evidence suggests that it can be safely used up to 4 g/day for induction and up to 2 g/day during maintenance. The induction dosing of an ongoing pediatric trial is as follows: (all doses are daily, prescribed as 2 divided doses): 4 g for children over 30 kg, 3 g for 20 to 30 kg and 2 g for those under 20 kg (safety has not been established in infants). Doses may be halved for maintenance treatment. (98% agreement)

Apheresis acts by an extracorporeal removal of leukocytes and other cells of the immune system (granulocytes, granulocyte/monocyte) through an adsorptive system of cellulose acetate beads (Adacolumn, Otsuka Pharmaceuticals, UK), or a polyester fiber filter (Cellsorba, Asahi Medical Company). Overall, pediatric data suggest a possible clinical efficacy of apheresis in children with both steroid-dependent and resistant UC, with reported response rates ranging between 60% and 85%, although they are mainly small case series or cohort studies ^(246–251). Data in adults are conflicting, with some observational and randomized clinical trials suggesting benefit ^(252–257), others, among them a large randomized, double-blind clinical trial evaluating active versus sham apheresis, showing no efficacy ⁽²⁵⁸⁾. A systematic review published in 2010 reported that, although there may be some efficacy in specific settings, concerns about methodological quality of identified studies prevent a rigorous meta-analysis and definitive conclusions ⁽²⁵⁹⁾.

TABLE 5: Paediatric VSL#3 dosing (by Miele et al ⁽²⁷⁹⁾)

FMT is based on the transfer of stool from a healthy donor, with a presumed healthy diverse microbiome, to a patient. Related or unrelated donors can be used, and they must undergo an accurate clinical and laboratory screening before the procedure. Some studies have used specifically prepared fresh stools, although frozen stools seem to have the same efficacy and safety ⁽²⁶⁰⁾, with delivery both to the upper gastrointestinal tract through nasogastric tube or to the lower gastrointestinal tract through colonoscopy or serial enemas. A few case series on the efficacy of FMT in pediatric UC have been published, reporting inconclusive results ^(261–263). The largest pediatric series (9 children with UC) showed a 33% clinical remission (PUCAI < 10) with serial enemas ⁽²⁶¹⁾. One small pediatric study reported no clinical improvement after FMT delivered via nasogastric tube ⁽²⁶²⁾. Overall, the safety profile appears acceptable, although mild-to-moderate side effects were common, and a case of transitory systemic reaction (profuse sweating, vomiting, paleness, tachycardia, and fever) has been reported ⁽²⁶⁴⁾. There may be a theoretical risk pertaining the transfer of an adult microbiome to a child, particularly very young with a developing microbiome, to quickening of immune aging and developing immune-related consequences ⁽²⁶⁵⁾. Rapid weight gain and the development of autoimmune disease have been reported after FMT in adults and in animal models ^(266–268).

Two small RCTs in adults with active UC reported different results: one showed clinical and endoscopic benefit of FMT administered via enema compared to sham ⁽²⁶⁹⁾; the other reported no difference between FMT using healthy donors or autologous feces administered via naso-duodenal tube, although the limited number of patients and the route of administration may have impacted on these results ⁽²⁷⁰⁾. Interestingly, patients who responded to FMT from a healthy donor restored their altered microbiota toward the healthy donor composition, while non-responders had no changes. Recently, the results of a third large, randomized, placebo-controlled trial in active UC resistant to conventional treatment have been reported ⁽²⁷¹⁾. Eighty-one adults with UC were randomized to receive a single FMT or placebo colonoscopic infusion on day 1, followed by FMT or placebo enemas 5 days per week for 8 weeks. Each active enema was derived from 3 to 7 unrelated donors. Steroid-free clinical remission with endoscopic response was achieved in 11/41 (27%) patients receiving FMT compared to 3/40 (8%) patients receiving placebo (P = 0.02). Microbial diversity increased and persisted after FMT while Fusobacterium spp was associated with lack of remission. Although FMT is gaining increased enthusiasm, the ideal donor and method of administration should be first determined before this can be incorporated outside the research setting.

Probiotics have been evaluated for induction and maintenance of remission in UC. One pediatric and 3 adult trials found E coli Nissle 1917 to be as successful as mesalamine in maintaining remission ^(272–275). The dosage used in all these studies, including the pediatric one, is 200 mg/day (100 mg contains 25 × 109 viable E coli bacteria), administered as capsules. A recent systematic review and meta-analysis suggests that E coli Nissle is equivalent to mesalamine to prevent relapse, while its efficacy is comparable to placebo in the induction of remission ⁽²⁷⁶⁾. A previous Cochrane systematic review, however, highlighted several methodological limitations in the maintenance studies, preventing any conclusion ⁽²⁷⁷⁾.

A small randomized, placebo-controlled trial of 29 children treated with 5-ASA reported that the combination of VSL#3 in conjunction with concomitant steroid induction and mesalamine maintenance treatment was superior to placebo in inducing and maintaining 1-year remission ⁽²⁷⁸⁾. A small open-label study in 18 children with mild-moderate UC evaluated the efficacy of VSL#3 added to standard treatment with 56% remission rate ⁽²⁷⁹⁾. Overall, adult data suggest a therapeutic benefit of VSL#3 in the maintenance of remission, supported by a systematic review ⁽²⁸⁰⁾. Studies on VSL#3 in IBD patients were performed on the original formulation containing 8 bacterial strains (Lactobacillus paracasei DSM 24733, Lactobacillus plantarum DSM 24730, Lactobacillus acidophilus DSM 24735, Lactobacillus delbrueckii subspecies bulgaricus DSM 24734, Bifidobacterium longum DSM 24736, Bifidobacterium infantis DSM 24737, Bifidobacterium breve DSM 24732, and Streptococcus thermophilus DSM 24731). Changing the manufacturing processes by different manufacturers may not have the same clinical efficacy and safety. Scarce published data report varying content of live/dead bacteria in various VSL#3 products and differences in effect on intestinal epithelial cell status ^(281,282). More studies are, however, needed to confirm these data and no changes in the recommendations are warranted at this stage. One randomized pediatric trial showed that rectal enemas of Lactobacillus reuteri ATCC 55730, added to oral mesalamine, were superior to placebo for inducing remission in left-sided UC ⁽²⁸³⁾.

Antibiotics have been evaluated as a therapy for UC both in the induction of remission and to prevent disease relapses as shown in 2 systematic reviews and meta-analyses ^(284,285). Both included 9 RCTs and concluded that antibiotics may improve outcomes in UC, but further studies are required to confirm this benefit since the included trials were very heterogeneous in their methodology and the type of drug intervention. The use of antibiotics in treating pediatric UC outside the research setting awaits further trials.

Recently, a small case series on the tolerability of curcumin added to standard therapy in pediatric IBD has been published, reporting an acceptable tolerability and a possible signal of benefit ⁽²⁸⁶⁾. Two placebo-controlled trials conducted in adults suggested the possible efficacy of curcumin in achieving and maintaining sustained clinical remission ^(287,288). Moreover, endoscopic remission was observed in 38% (8/22) patients treated with curcumin, compared with 0% (0/16) in the placebo group ⁽²⁸⁸⁾. A recent randomized, placebo-controlled, pilot study reported efficacy of topical curcumin as enema, added to oral mesalamine, compared to placebo, in 45 adults with mild-moderate proctitis/proctosigmoiditis ⁽²⁸⁹⁾.

Systematic review of complementary and alternative medicine treatments in IBD, including aloe-vera, andrographis paniculata, artemisia absinthium, barley foodstuff, boswellia serrata, cannabis, evening primrose oil, Myrrhinil intest, plantago ovata, silymarin, sophora, tormentil, wheatgrass-juice, and wormwood reported a possible benefit of some interventions, although, given the small number of trials and their heterogeneous methodological quality, no definite conclusions could be drawn ^(290,291). Of note, oral aloe vera has been evaluated in a double-blind, randomized, placebo-controlled trial as an adjuvant therapy in 44 adults with mild-to-moderate UC ⁽²⁹²⁾. Higher remission and response rates with improvement of the histological score were reported in the aloe vera group. These encouraging but preliminary findings await confirmation before aloe vera can be recommended for clinical practice.

Other complementary therapies, including germinated barley foodstuff and herbal medicine have been studied in adult case series or prospective cohorts. Because of sample size, study design, concomitant therapies and methodological limitations, these agents cannot currently be recommended for treating pediatric UC ^{(290,291,293)}.

A systematic review and a meta-analysis reported no efficacy of omega-3 supplement for maintaining remission in UC ^(293–295). Recently, a retrospective individual cohort study of 24 adults with IBD suggested efficacy and safety of intravenous immunoglobulin in the short-term management, when standard therapies are contraindicated ⁽²⁹⁶⁾. There are, however, no RCTs on its role both in adults and children.

Inflammatory Bowel Disease-unclassified

Recommendation

• 1. Treatment of IBDU patients should broadly follow that of UC patients of a similar disease severity [EL4, adult EL5]. (98% agreement)

Practice Points

• 1. A diagnosis of IBDU should only be made after a complete assessment including ileocolonoscopy, gastroscopy, and small bowel imaging. (100% agreement)
• 2. A lower threshold for disease reassessment should be adopted in patients with IBDU before treatment change. (95% agreement)
• 3. Although not validated for this indication, it is reasonable to use the PUCAI score to assess disease activity also in IBDU given the similarity of IBDU, clinically, to UC. (98% agreement)
• 4. A multi-item algorithm should be used to standardize the diagnosis of IBDU (Fig. 1; Table 1). (98% agreement)
• 5. While ASCA+/ANCA− profile is more suggestive of CD, and ASCA−/ANCA+ of UC, their diagnostic accuracy is too low to be used in isolation in the setup of IBDU. (98% agreement)

Patients with IBDU represent approximately 5% to 10% of pediatric IBD without a decline in incidence over time despite improved diagnostic measures. The rate is even higher in very early onset IBD. Complete examination is important, however, and the proportion of patients with IBDU is reduced if a full diagnostic work up is performed ⁽²⁹⁷⁾. IBDU is not a misclassification but rather a true overlap diagnosis within the spectrum of phenotypes between UC and Crohn colitis ⁽¹²⁾. Historically, patients with IBDU have often been poorly classified with no specific guidance available for detailed diagnostic criteria. The PIBD-Classes criteria were validated on a large multicenter dataset of 749 patients with colonic IBD from the Paediatric IBD Porto group of ESPGHAN ⁽¹²⁾. A diagnostic algorithm combining 23 features of different weightings (grouped in class 1, 2, and 3 features) (Table 1) may differentiate between patients with UC, atypical UC, IBDU, Crohn colitis and ileal/ileocolonic Crohn disease (Fig. 1) ⁽¹²⁾.

Given the rarity of IBDU and the hitherto lack of standardized diagnosis, there are very few studies which have been able to collect treatment information on significant numbers of patients. The aforementioned retrospective study from the Porto group of ESPGHAN utilized the data of 537 children with colonic IBD, including 260 IBDU, to explore common treatment schemes and to compare the treatment outcomes ⁽²⁹⁸⁾. This study demonstrated that treatment for IBDU and UC were broadly similar with the most common treatment used initially being 5-ASA. The use of steroids was lower than in UC; thiopurines and IFX use was broadly similar to patients with UC and lower than for patients with Crohn disease. Rates of surgery were lower than in Crohn disease and UC and the disease was more likely to be mild at follow-up compared to the other IBD subtypes, despite the similar use of medications as in UC. This suggests that treatment can follow that of UC initially with a 5-ASA regimen.

SURGICAL CONSIDERATIONS (RELATED TO BOTH PART 1 AND 2 OF THESE GUIDELINES)

The Surgeon's Perspective

Recommendations

• 1. Elective colectomy should be considered in children with active, or steroid-dependent, UC despite optimized medical therapy, and in those with colonic dysplasia [EL4, adult EL3]. (98% agreement)
• 2. Restorative proctocolectomy with ileal pouch-anal anastomosis (IPAA; J pouch) and a covering loop-ileostomy is the recommended elective surgery for pediatric UC [EL3, adult EL3]. (93% agreement)
• 3. Three-stage procedure (subtotal colectomy with ileostomy first) is recommended for patients with acute severe colitis, treated with high-dose steroids, or recent anti-TNF therapy, severe malnutrition, or IBDU; however, the final choice of the surgical approach should be individualized [EL4, adult EL3]. (98% agreement)
• 4. A minimally invasive laparoscopic approach is recommended in children as there are equivalent outcomes to open surgery both for urgent and elective cases and possibly superior outcomes regarding fertility in girls [EL4, adult EL3]. (100% agreement)
• 5. Pouch surgery for children with UC should be performed by experienced pediatric or adult surgeons in high volume centers preferably performing at least 10 pouches per year. (100% agreement)

Practice Points

• 1. Crohn disease must be excluded before the time of surgery, through a diagnostic workup including ileocolonoscopy, gastroscopy, and small bowel imaging, before colectomy, as clinical status allows. (100% agreement)
• 2. Functional outcomes and surgical complications are comparable after hand-sewn and stapled IPAA. The length of remaining anorectal mucosa between the dentate line and the anastomosis should not exceed 2 cm, regardless of anastomotic technique. (100% agreement)
• 3. IPAA without a covering loop ileostomy (ie, 1-stage procedure) may be considered in selected children with mild disease and good nutritional status without anti-TNF or steroid treatment, provided that no technical difficulties or anastomotic tension occur during surgery; however, the final choice of the surgical approach should be individualized. (98% agreement)
• 4. There is no published evidence on whether postponing pouch surgery after subtotal colectomy, for example, until after puberty, influences long-term outcomes after IPAA. If pouch surgery is delayed, a strategy to maintain the rectal stump free of inflammation should be discussed, based on topical treatment. (100% agreement)
• 5. The role of ileorectal anastomosis (IRA) remains controversial. It may be offered to selected female patients, who are particularly concerned about the risk of reduced fertility associated with IPAA. Information on higher failure rate and the need for lifelong cancer surveillance should be provided. (98% agreement)
• 6. Treatment with steroids (prednisolone >0.25 mg · kg⁻¹ · day⁻¹ or >20 mg/day) is associated with an increased risk of surgical complications, whereas thiopurines and calcineurin inhibitors are not. There are insufficient data regarding vedolizumab. Anti-TNF increases the surgical risk in Crohn disease and according to the precautionary principle, colectomy should be preferably performed 4 to 6 weeks after the last IFX infusion if it can be safely postponed. (93% agreement)

Surgery for pediatric UC may require up to 3-staged procedures—first stage, subtotal colectomy with end-ileostomy; second stage, restorative procto-colectomy with ileal pouch-anal anastomosis or ileo-rectal anastomosis (with or without covering ileostomy); third stage, closure of the covering ileostomy. The decision concerning the best combination of procedures is dictated by the clinical status of the patient. Restorative procto-colectomy and IPAA/IRA with covering ileostomy can be performed as a combined first stage for most “ambulatory” elective UC cases. The covering ileostomy is reversed several months later after confirmed healing of the pouch ^(299–303). Three-stage surgery (subtotal colectomy and ileostomy first) is recommended for ASC, for example, where the pre-operative PUCAI is >45, or in those on high-dose pre-operative steroids (prednisolone >0.20 mg · kg⁻¹ · day⁻¹) ^(36,301). Although single-stage restorative procto-colectomy IPAA without a covering ileostomy was not associated with increased anastomotic complications in some retrospective pediatric series ^{(300,304–306)}, this cannot be recommended before more studies are available given the retrospective design of the studies and the inherent confounding by indication bias.

Emergency surgery for ASC is an initial subtotal colectomy (leaving a rectal stump) with end-ileostomy formation only. Creation of IPAA/IRA should be deferred until the clinical status of the patient has normalized, followed by stoma closure as the third stage. Laparoscopic colectomy/ileostomy for both ASC and ambulatory UC is safe and feasible in experienced hands also in children ^(36,307). The PUCAI has been reported in a retrospective analysis to be a useful tool when considering 1- versus 2- versus 3-stage procedures for pediatric UC ⁽³⁶⁾.

As significant complication rates are reported after colectomy for both ASC and ambulatory UC in children, in particular infectious and thromboembolic events ^(8,308), peri-operative antibiotic and thromboembolism prophylaxis should be routine. The rectal stump can be fashioned as a mucous fistula (open or within the subcutaneous tissue) if there is significant proctitis. A more commonly used alternative is to close the rectal stump within the abdomen and place a temporary trans-anal drain ⁽³⁰⁹⁾. Length of hospital stay, short-term surgical complications and functional outcomes seem similar after open and laparoscopic procedures ^{(300,310–312)}.

Steroid treatment, hypoalbuminemia and malnutrition are also associated with increased surgical complication rates ⁽³¹³⁾. In ASC, children are likely to have been on recent steroid therapy and may be in a relatively poor nutritional state, but surgery when needed should not be delayed to correct this. Thiopurine and calcineurin inhibitors were not associated with postoperative surgical complications ^(313–315), while current retrospective pediatric data on anti-TNF regimens are controversial ^{(313,314,316)}. Meta-analysis of adult data shows an increased risk of surgical complications in patients who had been on pre-operative anti-TNF therapy in CD but not in UC ^(317,318).

According to a meta-analysis of 5 pediatric studies (306 patients), straight ileo-anal anastomosis (SIAA) was associated with a higher failure rate (15% vs 8%) and perianal sepsis (20% vs 10%), as well as a higher stooling frequency as compared with a J pouch ileo-anal anastomosis (JPAA) ⁽³¹⁹⁾. A more recent multicenter study, including 112 children with SIAA, and 91 with JPAA, reported comparable postoperative complication rates ⁽³²⁰⁾. Both day-time and night-time stooling frequency were higher after SIAA, although the difference became less apparent by 2 years (mean 24 hours stooling frequency 8.4 vs 6.2 at 2 years). This difference may still be clinically important, because quality of life in children after restorative proctocolectomy is inversely associated with stooling frequency ⁽³⁰²⁾.

JPAA, on the other hand, carries a risk of pouchitis, which clearly exceeds the incidence of enteritis following SIAA (49% vs 24%, OR 4.5; see henceforth detailed chapter on the pouch) ⁽³²⁰⁾. Surgical complications and functional outcomes are comparable after hand-sewn or stapled J-pouch anastomosis. For example, in 1 series, stool frequency was 4 per day after both techniques ^{(299,321,322)}. A common complication of stapled IAA is, however, an undesirably long rectal stump with excessive remaining anorectal mucosa above the dentate line (>2 cm). Chronic inflammation of the rectal mucosal remnant is called “cuffitis” and discussed further below. One study reported a lower rate of small bowel obstruction during 4 post-operative years after laparoscopic IPAA compared to open procedures ⁽³¹⁰⁾, while no difference was found in another ⁽³⁰⁰⁾.

In those undergoing IPAA, the diagnosis of UC may change to CD; ∼15% in adult series ^{(299,302,303,323)} and 11 of 128 children (9%) in a recent multicenter pediatric study from the Paediatric IBD Porto group of ESPGHAN ^(324,325). Three-stage IPAA has been used to reduce these complications in children with IBDU. Histology of a colectomy specimen or pre-operative diagnosis of IBDU, however, poorly predicts the long-term outcomes of IPAA in adults with UC ^(326,327). In most studies, the incidence of pouchitis and post-operative diagnosis of CD is similar after IPAA in patients with UC and IBDU ⁽³²¹⁾. There is no published evidence on whether postponing pouch surgery after subtotal colectomy for an extended period influences the rate of complications or long-term outcome after IPAA. Overall, results from pediatric series of IPAA in terms of later pouch abandonment (<15% at median 10–20 years follow-up) are similar to adult reports, albeit with shorter length of follow-up in most series ^{(300,302,328)}. A multicenter, retrospective study from the Paediatric IBD Porto Group of ESPGHAN included 129 children who underwent IPAA, showed an increased rate of surgical complications in children undergoing colectomy under the age of 10 years but there was no difference in complications rate whether the pouch surgery was delayed or not ⁽³²⁴⁾.

While IPAA has been shown to reduce female fecundity and fertility in adult studies (e.g. reduction of fertility rate by 52% among women aged 15 to 44 years ⁽³²³⁾), most used the non-stringent definition of inability to become pregnant within 1-year of intent ^(322,323). This should be discussed with female patients and their family before any surgical procedures. Laparoscopic IPAA, as is increasingly performed, may ameliorate the risk of subfertility due to reduced adhesion formation, pelvic scarring and Fallopian tube obstruction ^(329–331). In 1 adult series, spontaneous pregnancy rate was higher after laparoscopic IPAA (70%) compared to open IPAA (39%, P = 0.023) among 50 women who attempted to conceive ⁽³²⁶⁾. Fertility is also much better preserved after IRA ⁽³⁰⁰⁾. Fecundity remained similar to the general population after IRA, but dropped to 54% after IPAA among women with familial adenomatous polyposis ⁽³²⁷⁾. In a recent follow-up study of 343 adults with UC, 10- and 20-year IRA failure rate was 27% and 40%, respectively ⁽³²⁸⁾. Secondary proctectomy was required for refractory proctitis (66%), dysplasia (11%) and for cancer (10%) ⁽³³²⁾. At the end of the follow-up, 18% had undergone secondary IPAA and 13% had permanent ileostomy. Although fecal continence and stooling frequency is better preserved after IRA compared to IPAA, most patients require anti-inflammatory medication and urgency rate is higher, while quality of life similar to that after IPAA ⁽³²⁸⁾.

Data from the Porto group of ESPGHAN suggest that the experience of the surgeon is associated with the likelihood of development of chronic pouchitis; (15%) in surgeons with ≥10 surgeries/year versus (41%) in surgeons with <10 per year, P = 0.013 ⁽³²⁵⁾. This is in line with a large study from the UK showing the pouch outcome was superior if done in centers performing at least 9 to 10 procedures annually ⁽³³³⁾.

Pouchitis and Cuffitis

Recommendations

• 1. Pouchoscopy with mucosal biopsies should be performed at the first suspected episode of pouchitis [EL3, adult EL3]. (95% agreement)
• 2. A 14-day course of ciprofloxacin and/or metronidazole is recommended as first-line therapy for pouchitis while the former may be more effective [EL5, adult EL1]. (100% agreement)
• 3. Combined metronidazole and ciprofloxacin or oral/topical budesonide can be used in persistent cases [EL5, adult EL2]. (98% agreement)
• 4. In recurrent and/or chronic pouchitis, VSL#3 is recommended for maintaining remission [EL5, adult EL1]. (98% agreement)
• 5. Topical mesalamine is recommended for treating cuffitis [EL 5, adult EL4]. (100% agreement)

Practice Points

• 1. A clinically useful categorization of pouchitis is “antibiotic-responsive” (ie, infrequent episodes (<4 per year) each with a rapid response to a 2-week course of a single antibiotic, “antibiotic-dependent” (ie, frequent episodes (≥4 per year) or persistent symptoms which require long-term antibiotic therapy to maintain remission) and “antibiotic-refractory” (ie, failure to respond to a 4-week course of antibiotics, necessitating an alternative therapy of 4 weeks or longer). Duration of pouchitis can be categorized as acute (<4 weeks) or chronic (≥4 weeks) and frequency may be described as infrequent, relapsing, or continuous. (100% agreement)
• 2. In chronic, recurrent or refractory pouchitis-like symptoms, other diagnoses, such as cuffitis, missed Crohn disease, anastomotic ulcer, irritable pouch syndrome, infectious pouchitis, and anastomotic stenosis, should be excluded. (100% agreement)
• 3. Fecal calprotectin may be used to assess pouch inflammation to minimize repeated pouchoscopies in recurrent pouchitis and to monitor response to treatment. Calprotectin >300 μg/g is suggestive of pouchitis while lower levels do not preclude pouchitis (57% sensitivity, 92% specificity). (95% agreement)
• 4. The common antibiotic dosing strategies for pouchitis are ciprofloxacin (30 mg · kg⁻¹ · day⁻¹ up to 1 g/day in 2 divided doses) and/or metronidazole (20–30 mg · kg⁻¹ · day⁻¹ in 3 divided doses up to 1.5 g/day) for 14 days. (98% agreement)
• 5. VSL#3 can be used once daily at an age- or weight-dependent dose (Table 5). (95% agreement)
• 6. VSL#3 may be also effective for preventing the first episode of pouchitis, but this is not justified since many children will never develop pouchitis. (100% agreement)
• 7. Thiopurines may be considered in refractory pouchitis, not responding to antibiotic therapy or in the presence of budesonide dependence, despite the lack of good evidence. The effectiveness of IFX for this indication has been demonstrated only in adult case series with a response rate of ∼50%. (98% agreement)

Pouchitis, a non-specific and idiopathic inflammation of the ileal reservoir, is the most common complication of IPAA, occurring in 24% to 67% of pediatric UC patients ^{(299,300,302,320,323,334–337)}. A recent multicenter, retrospective cohort study from the Paediatric IBD Porto Group of ESPGHAN included 129 children who underwent IPAA (93% UC and 7% IBDU) and showed that 86 children (67%) developed pouchitis during follow-up ⁽³²⁵⁾. In 33 (26%) the pouchitis was chronic, 10 of whom (8%) had Crohn-like disease of the pouch. Median time from pouch formation to the first episode of pouchitis was 10.5 months (IQR 6–22); in 54% of cases the first episode occurred within 1 year. In an older cohort of 399 UC children with a mean age of 18 ± 3 years at colectomy, 121 (36%) had at least 1 episode of acute pouchitis, and 29 (9%) pouch failure ⁽³⁰⁰⁾. Pouch type, age, and operative technique had no impact on whether patients developed pouchitis.

Symptoms and severity of pouchitis vary, but typically include increased stool frequency and urgency, tenesmus, incontinence, abdominal pain, and rectal bleeding ⁽³³⁸⁾. Cuffitis, residual rectal cuff inflammation, may cause symptoms similar to those of pouchitis, especially bleeding. The cuff is the remaining rectal mucosa between the dentate line and the anastomosis after restorative procto-colectomy. Symptoms of pouch dysfunction in patients with IPAA may be caused by conditions other than pouchitis, including CD of the pouch, anastomotic ulcer or stenosis. In children, the occurrence of terminal ileitis, or “pre-pouch ileitis,” has also been reported ⁽³³⁹⁾, and does not necessarily confirm the diagnosis of CD if it involves only mild inflammation in a short segment. Other differential diagnoses include ischemia and, rarely, infections such as CMV and C difficile. A diagnosis of irritable pouch syndrome is suspected when symptoms are present without endoscopic inflammation ⁽³⁴⁰⁾. Thus, endoscopic and histological evaluation of the pouch should be performed at the first episode of pouchitis and periodically thereafter.

Endoscopic features of pouchitis may include hyperemia, diminished vascular pattern, friability, hemorrhage, and ulcers. Abnormalities may be focal or diffuse, and unlike in UC, they may be discontinuous. Often, they are more severe in the distal compared to the proximal pouch ^(341–343). Mucosal biopsies typically demonstrate partial to complete villous blunting with crypt hyperplasia and increased mononuclear inflammatory cells and eosinophils in the lamina propria, crypt abscesses, and ulcerations. Mucosal biopsies should be obtained from the pouch and from the afferent ileal loop, but not from the staple line, as erosions and/or ulcers along the staple line do not necessarily indicate pouchitis ⁽³⁴⁴⁾.

Two main scoring systems exist for the diagnosis of pouchitis but their utility in clinical practice is limited as they await further validation to associate the scores with clinical outcomes ^(345,346). The Pouchitis Disease Activity Index (PDAI) evaluates symptoms, endoscopic findings, and histological patterns in a composite score, with a score of ≥7 indicating pouchitis ⁽³⁴⁷⁾. The Pouchitis Activity Score (PAS) incorporates similar elements to the PDAI and a score >13 is suggestive of pouchitis ⁽³⁴⁸⁾. A modified PDAI (mPDAI), omits the histology component ⁽³⁴⁹⁾.

Several variables may predict the risk of pouchitis. A small pediatric study reported that the only predictive factor associated with risk of pouchitis was a higher PUCAI score at the time of diagnosis ⁽³³⁷⁾. As discussed above, data suggest that the surgeon's experience is associated with risk for pouchitis ⁽³²⁵⁾. Chronic pouchitis was also associated with Ashkenazi Jewish ethnicity, while any-pouchitis was associated with age at diagnosis and longer disease duration. Several adult studies have reported an increased incidence of pouchitis in patients with a younger age at onset, backwash ileitis, PSC, extensive colonic disease, positive pANCA, preoperative steroid use, being a non-smoker, and carriage of genetic polymorphisms in NOD2/CARD15, which is more prevalent in Ashkenazi Jews ^{(66,350–358)}.

The probiotic mixture VSL#3 was effective in maintaining remission in adult patients with chronic pouchitis as shown in 2 double-blind placebo-controlled trials from Italy ^(359,360). Results regarding the effectiveness of VSL#3 in preventing the first episode of pouchitis are more controversial ^(361,362).

Antibiotic treatment is considered first-line treatment for pouchitis. Only small placebo-controlled trials have, however, been conducted to support this practice and none in children ^(363,364). Ciprofloxacin may be slightly more effective than metronidazole, with fewer adverse events. Shen et al have shown the superiority of ciprofloxacin over metronidazole in inducing remission ⁽³⁶⁵⁾. In antibiotic-refractory pouchitis, Gionchetti et al used oral budesonide for 8 weeks and achieved remission in 75% of 20 patients ^(366,367). A case series of 28 patients with refractory pouchitis were treated with IFX of whom 88% responded after 10 weeks, and 56% after a median follow-up of 20 months ⁽³⁶⁸⁾; other case series also support the use of IFX in refractory pouchitis ^(369,370) as well as adalimumab ⁽³⁷¹⁾ and alicaforsen (an inhibitor of intercellular adhesion molecule-1) enemas ⁽³⁷²⁾.

In an open study, topical treatment with metronidazole induced clinical improvement within a few days without systemic side effects and with a decrease in concentrations of anaerobic bacteria ⁽³⁷³⁾. Furthermore, uncontrolled studies have suggested that 5-ASA either as suppositories or enemas may help in the treatment of pouchitis ⁽³⁷⁴⁾.

OTHER MANAGEMENT CONSIDERATIONS

Extraintestinal Manifestations

Recommendations

• 1. Treatment of peripheral arthritis should be directed at inducing remission of the luminal disease [EL4, adult EL3]; sulfasalazine should be considered as first-line treatment for peripheral arthritis, followed by anti-TNF [EL4, adult EL2]. (93% agreement)
• 2. Transaminases and γGT should be monitored at least annually in all UC patients, to screen for PSC and autoimmune hepatitis [EL4, adult EL4]. (100% agreement)
• 3. Chronic elevation of liver enzymes in the presence of cholestasis, should be investigated with ultrasound followed by MR-cholangiopancreatography (MRCP), in addition to liver biopsy when indicated (see practice point); endoscopic retrograde cholangiopancreatography (ERCP) is recommended for therapeutic interventions [EL3, adult EL3]. (95% agreement)
• 4. Patients with PSC and IBD are at increased risk for colorectal carcinoma (CRC) and thus annual or bi-annual surveillance colonoscopy should be initiated from the time of PSC diagnosis. However, surveillance could be deferred in pre-pubertal children while individualizing based on risk factors (disease duration, family history, severity of the disease over time, and disease extent), since CRC is extremely rare under the age of 12 years even in the presence of PSC. (95% agreement)

Practice Points

• 1. Acute peripheral arthritis affecting the large joints is usually associated with active IBD and thus treatment should be directed to the gut. (98% agreement)
• 2. The diagnosis of axial spondylo-arthritis or sacro-ileitis is based on typical clinical signs such as progressive low back pain, gluteal and thigh pain combined with radiological abnormalities (most often MRI). Treating sacro-ileitis requires close collaboration with a rheumatologist. (100% agreement)
• 3. If required for the treatment of articular inflammation, non-steroidal anti-inflammatory drugs may be used for a short course and at low doses to minimize the risk of aggravating IBD. (98% agreement)
• 4. Since some degree of autoimmune hepatits/overlap syndrome is not uncommon in children with PSC, a low threshold should be practiced when considering a liver biopsy in this setup. (95% agreement)
• 5. No medication has been proven to reduce the time from PSC diagnosis to liver transplant or the development of cholangiocarcinoma. The benefit of ursodeoxycholic acid remains questionable, and if used, doses should be preferably low (10–15 mg · kg⁻¹ · day⁻¹). Alternatively, oral vancomycin may be considered (usual total daily dose 35 mg/kg (maximum 1500 mg) divided into 3 times daily), for 12 weeks but long-term data are lacking. (95% agreement)
• 6. PSC is a significant risk factor for cholangiocarcinoma also during childhood. Serial CA19.9 and liver ultrasound/MRCP testing may thus be considered every 1 to 2 years to screen for cholangiocarcinoma but there is no pediatric evidence to support this practice. (95% agreement)

We would like to refer the reader to comprehensive ECCO guidelines on extraintestinal manifestations (EIM) and highlight here only pertinent points common in children ⁽³⁷⁵⁾. Some EIM are associated with intestinal disease activity (ie, erythema nodosum, peripheral arthritis), whereas others occur independently (ie, pyoderma gangrenosum, uveitis, ankylosing spondylitis, and PSC) ⁽³⁷⁶⁾. Data from 2 pediatric registries in the USA ^(377,378) and 1 in Europe ⁽³⁷⁶⁾ indicate that 1 or more EIMs are present at diagnosis in 6% to 17% % of children with UC, especially those older than 5 years, with an increase to almost 50% with disease evolution ^(379–382), and more so with extensive colitis ⁽³⁷⁸⁾.

Joint disease in IBD may be axial (sacro-ileitis or ankylosing spondylitis), causing lower back pain or peripheral arthritis, which is usually acute and self-limiting, seronegative and not deforming. In children, the prevalence of arthritis seems to be twice as high as in adults, ⁽³⁷⁷⁾ with a clear female predominance. There are some concerns about aggravating the bowel disease by using NSAIDs; however, the risk seems to be low if prescribed for a short course and at low doses ⁽³⁸³⁾. The sulfapyridine component of sulfasalazine has an anti-inflammatory effect on both the colonic mucosa and the joints ⁽³⁸⁴⁾. MTX is the cornerstone disease-modifying anti-rheumatic drug in juvenile arthritis ⁽³⁸⁵⁾ but anti-TNF regimes have emerged in the last 2 decades ⁽³⁸⁶⁾.

PSC is 3 times more likely to occur in UC compared to CD ⁽³⁷⁸⁾, and is associated with older age in children ⁽³⁷⁸⁾. PSC may precede the onset of IBD by years but may occur even after colectomy. The prevalence of PSC in pediatric IBD is 1.6% at 10 years after diagnosis ⁽³⁷⁷⁾, but higher at 3% ⁽³⁸⁷⁾ if systematic screening tests are performed. In a recent multicenter report of 781 children with PSC (4277 person-years of follow-up), overall event-free survival was 70% at 5 years and 53% at 10 years but PSC-IBD was associated with a favorable prognosis; cholangiocarcinoma occurred in 1% ⁽³⁸⁸⁾.

Being non-invasive, MRCP is the most appropriate imaging modality for diagnosing PSC in children. A pattern of irregular bile ducts, with zones of narrowing and dilatation is characteristic of PSC ⁽³⁸⁹⁾. PSC may progress to liver cirrhosis, ultimately necessitating liver transplantation. Patients with PSC and UC have a greater risk of malignancies such as CRC and cholangiocarcinoma (8%–30% of UC patients with long standing PSC) ^(390,391). A recent study on the cancer and mortality in children in Europe has demonstrated several cases associated with PSC ⁽¹⁰⁾, but CRC in UC children younger than 12 years is extremely rare. PSC is associated with more extensive disease and thus has a greater cancer risk ⁽³⁹¹⁾ but also with milder disease course. The higher colectomy rate in these patients is secondary to dysplasia and CRC. In adults with PSC, ursodeoxycholic acid is reported to improve abnormal liver tests ⁽³⁹²⁾ and to reduce the risk of CRC ⁽³⁹³⁾, although this has not been shown by all ^(394,395). No therapy has been shown to reduce time to liver transplantation, cholangiocarcinoma or death ^{(394,396,397)}. Recent recommendations for adult patients suggest ursodeoxycholic acid at a dose of 10 to 15 mg · kg⁻¹ · day⁻¹ and warn against high dose treatment (>20 mg · kg⁻¹ · day⁻¹), which may increase mortality ^{(394,395,398)}.

Oral vancomycin may be considered for 12 weeks as it has been shown to reduce and even normalize serum liver enzymes and gGT ^(399–405). Both vancomycin and metronidazole have been efficacious in recent small studies; however, only patients in the vancomycin groups reached the primary endpoint, and with fewer adverse effects ⁽⁴⁰³⁾. Oral vancomycin re-treatment when needed has been associated with a rise in T regulatory cells (Treg) and normalization of liver function tests ⁽⁴⁰⁶⁾.

Older age at PSC diagnosis increases the risk of colonic neoplasia ⁽⁴⁰⁷⁾. Targeted biopsies aimed at abnormal areas identified by newer colonoscopic techniques (chromoendoscopy, confocal microendoscopy) should be preferred ⁽⁴⁰⁸⁾. The optimal follow-up method is still debatable ⁽⁴⁰⁹⁾.

Nutrition, Growth, and Bone Health

Practice Points

• 1. High intake of red or processed meat, protein, alcoholic beverages, sulfur, and sulfates have been associated with disease exacerbations. Due to the lack of solid evidence, exclusion diets, however, should not be used to induce or maintain remission in pediatric UC, and could lead to nutritional deficiencies. (100% agreement)
• 2. DEXA (corrected for height and age to produce age- and sex-matched z scores ⁽⁴¹⁰⁾) should be considered in high-risk patients such as those with severe disease, prolonged malnutrition, amenorrhea, delayed puberty, and/or steroid dependency. (98% agreement)
• 3. Promoting mucosal healing, adequate nutrition, weight-bearing exercise, avoiding smoking, and steroid-sparing strategies should be employed to facilitate bone health. The rare use of bisphosphonates should be reserved to those with pathological fractures, in consultation with a pediatric bone specialist. (100% agreement)
• 4. Growth impairment is rare in children with UC who are not steroid-dependent. Therefore, Crohn colitis or primary growth hormone deficiency should be considered when significant growth impairment is present. (100% agreement)
• 5. Vitamin D should be supplemented if 25-OH vitamin D is <50 nmol/L, regardless of steroid use. (93% agreement)
• 6. There are different strategies in treating vitamin D deficiency in addition to daily treatment (>2000 IU/day). A commonly applied strategy is to prescribe a “loading dose” (50,000 IU of vitamin D3 orally once weekly for 2 to 3 months, or 3 times weekly for 1 month). Single high-dose oral vitamin D3 300,000 to 500,000 IU (ie, stoss dosing) has been reported ⁽⁴¹¹⁾ to be effective and safe. (98% agreement)

While nutritional deficiencies can develop quickly during periods of active UC ⁽⁴¹²⁾, normal growth is maintained in >95% of children with UC who are not steroid dependent ^(413–415). A more detailed review of all nutritional issues in children with IBD can be found in the recently published guideline from the Paediatric IBD Porto group of ESPGHAN ⁽⁴¹⁶⁾. Patients with active UC often reduce fiber in their diets without supportive evidence. Corn and corn products, nuts, milk, and bran were avoided by >20% of UC patients ⁽⁴¹⁷⁾. Soluble fiber is, however, the best way to generate short-chain fatty acids such as butyrate, which has anti-inflammatory effects ⁽⁴¹⁸⁾. In addition, many UC patients avoided tomato, dairy products, chocolate, wheat, tomato sauces, and fruit juice ⁽⁴¹⁷⁾, but there is no nutritional intervention clearly supported in UC and the reader is referred to an excellent recent summary on the topic ⁽⁴¹⁸⁾.

Peak bone mass attained during adolescence is the most important determinant of lifelong skeletal health. Some osteopenia is present in up to 22% of UC children ⁽⁴¹⁹⁾, but severe osteopenia is only present in 3% to 6% in UC, as compared with 12% to 18% in CD ^(420–422). Nutritional status seems to have a greater impact on bone status than corticosteroid therapy ⁽⁴²³⁾. Children with IBD are at particularly risk for vitamin D deficiency, but this was not found to be directly associated with osteopenia ⁽⁴²⁴⁾. Nonetheless, vitamin D deficiency should be treated especially in children with decreased bone mineral density. A recent meta-analysis showed that low vitamin D is associated with a more active disease ⁽⁴²⁵⁾. Age-appropriate nutrition support, weight-bearing exercise, and adequate disease control using steroid-sparing strategies ^{(410,421,426)} have been suggested as means to improve bone formation but without supportive evidence. Indeed, a prospective study that followed 58 children with CD for 2 years did not show significant improvement in bone mineral density despite increased height z score and reduced disease activity ⁽⁴²¹⁾.

The most important determinant of treating osteopenia, besides avoiding steroids, is efficient treatment aiming at mucosal healing since osteopenia may typically be a consequence of pro-inflammatory cytokines ⁽⁴²⁷⁾. Indeed, interventions that lead to mucosal healing such as anti-TNF therapy and exclusive enteral nutrition showed rapid improvement of serum bone markers in children with CD ^(428–432). Bisphosphonates are effective to improve bone mineral density in IBD but pediatric use should be reserved for extreme circumstances, typically when pathological fractures are present, an uncommon situation in UC.

Psychosocial Support, Adherence to Therapy, and Transitional Care

Recommendations

• 1. Adolescents should be included in transition to adult care programs, which can be adapted according to the local organization of the pediatric and adult facilities [EL4]. (100% agreement)

Practice Points

• 1. Paediatric IBD centers should offer psychological support according to local resources. (100% agreement)
• 2. Adherence should be regularly evaluated by patient interviews, drug monitoring (eg, serum drug level), and prescription refill rates. (100% agreement)
• 3. Adherence may be improved by providing comprehensive information regarding the prescribed medication, keeping the pill burden as low as possible, using single daily dosage when possible, utilizing electronic reminders and providing pill boxes. (100% agreement)

Several systematic reviews concluded that adolescents with IBD, especially boys, have reduced health-related quality of life, including anxiety, depression, social problems, and low self-esteem ^(433–436). The altered quality of life of children with IBD can affect the entire family, who often lack the appropriate strategies to deal with this complicated reality ⁽⁴³⁷⁾. The rate of depression may be as high as 25% and it is often under-recognized both by parents and health care professionals. Anxiety and depression appear to be risk factors for early recurrence of the disease and adversely affect the disease course but may also commonly be a reactive response to active disease ⁽⁴³⁸⁾. Cognitive behavioral therapy has been shown to be especially effective in improving depressive symptoms and functioning in children with IBD ⁽⁴³⁹⁾.

Non-adherence in IBD is reported in 50% to 66% of children ^(433,440), especially during adolescence. Pediatric-specific barriers include fear of adverse events of medication, feeling that the disease is inactive, belief that the medication is not working, >1 daily medication, forgetting, interfering with other activities, difficulty in swallowing pills ⁽⁴⁴¹⁾, lack of motivation, and parent-child conflict ⁽⁴⁴²⁾.

Transition is defined as the planned move of adolescents and young adults with long-term physical conditions from child-centered to adult-orientated healthcare. The optimal timing of transition from pediatric to adult management of UC has to be decided on an individual basis by a joint team of pediatric and adult gastroenterologists ⁽⁴⁴³⁾. Several suggestions for transition programs have been published, but none has been formally evaluated ⁽⁴⁴⁴⁾. The transition period usually starts from the age of 14 to 18 years depending on the development of the patient and availability of qualified pediatric and adult gastroenterologists. The time of transition should be individually adapted according to the psychosocial readiness. Whenever feasible, at least 1 joint clinic with both the pediatric and the adult gastroenterologist is recommended during the transition process. The adolescent should be encouraged to assume increasing responsibility for treatment and to visit the clinic room at least once without being accompanied by the parents. The ECCO topical review on transition to adult care addresses in detail all aspects related to the steps to be followed during transition ⁽⁴⁴⁵⁾.

Very Early-Onset Inflammatory Bowel Disease Presenting As Colitis

Practice Points

• 1. In infants younger than 2 years, allergic colitis, immunological disorders and monogenic forms of colitis should be excluded. (100% agreement)
• 2. Unusual disease evolution, history of recurrent infections, HLH, and non-response to multiple IBD medications may indicate an underlying genetic defect which should prompt genetic and/or immunological analyses at any age during childhood (Table 6). (100% agreement)

TABLE 6: Diagnostic workup of very early onset IBD to be adapted according to the clinical presentation (see text)

The colitis phenotype is the most common in the VEOIBD group (6 years of age and younger) ⁽⁴⁴⁶⁾, and even Crohn disease frequently resembles UC. Therefore, the term IBDU rather than UC may be more appropriate in this earlier age group, reported in 34% and even 71% of very young children ^(447,448). The differential diagnostic spectrum for this age group is challenging ^(448,449) since the colitis may be caused by various immunological disorders: classical immune defects (such as combined immune-deficiencies), subtle immune defects or defects of the regulation of immune responses due to a monogenetic disorder including defects in interleukin (IL)10-signaling, X-linked inhibitor of apoptosis protein (XIAP) deficiency, defective neutrophil function and many others (Table 6) ⁽⁴⁴⁹⁾. Since no specific biological test confirms allergic colitis, only a successful trial of elimination diet is useful diagnostically ⁽⁴⁵⁰⁾ and may be proposed according to the clinical context especially in those younger than 1 year.

A large percentage of children with IBD developing before age 6 years present with relatively typical colitis, including mild disease which can be easily managed with 5-ASA ⁽⁴⁵¹⁾. Many monogenic forms of VEOIBD may, however, initially appear as typical polygenic IBD but then prove resistant to standard therapy ^(448,449). Over 50 different monogenetic defects causing an IBD-like disorder have been described. A complete review of the genetic workup of VEOIBD and treatments is beyond the scope of these guidelines and the reader is referred to a previous comprehensive review ⁽⁴⁴⁹⁾. Briefly, appearance of perianal disease with skin folliculitis during the first few months of life is a strong indicator of a defect in the IL10 axis ^(452–454). Repeated bacterial and fungal infections orientate toward defective neutrophil functions, (eg, chronic granulomatous disease ^(455,456)). Recurrent skin infections, and EBV or CMV-induced HLH may indicate the presence of XIAP-defect ⁽⁴⁵⁷⁾. This X-linked defect can affect boys and in rare cases also girls ⁽⁴⁵⁸⁾. The presence of multiple intestinal atresia, or evidence of increased rate of epithelial cell apoptosis on small bowel biopsy may hint toward TTC7A, especially when observed in the presence of low IgG levels, T- and B-cell lymphopenia and mild reduction in NK cells ^(459–461). Woollen, fragile hair, and facial abnormalities (small chin, broad flat nasal bridge, and prominent forehead), immune defects, liver disease, and colitis (referred to as trichohepatoenteric syndrome or phenotypic diarrhea) may be due to mutation in SCIVL2 ⁽⁴⁶²⁾ or TTC37 ⁽⁴⁶³⁾. If signs of autoimmunity are associated with intestinal inflammation with high rates of epithelial cell apoptosis, IPEX-syndrome or IPEX-like disorders should be considered ^(464–467).

If the molecular defect is caused by a mutation affecting predominantly immunological cells (eg, IL10 signaling defects, XIAP and chronic granulomatous disease), hematopoietic stem cell transplantation may be curative ^{(452,453,458,468)}. Inhibition with IL1-antagonists may be a way to stabilize patients with IL10 signaling defects while awaiting hematopoietic stem cell transplantation (HSCT), but more confirmation is required before this can be utilized in clinical practice ⁽⁴⁶⁹⁾. Early HSCT improves life expectancy of IL10-deficient patients since they are at risk for developing lymphoma ⁽⁴⁷⁰⁾. HSCT is not always the ultimate treatment option, as shown in patients with TTC7A mutations, which involve the epithelial gut barrier rather than immunological cells. This highlights the importance of a rapid and precise molecular diagnosis in children with colitis starting early in life.

SYNTHESIS AND SUMMARY

This Part 1 of the pediatric UC guidlines yielded 40 formal recommendations and 86 practice points along with practical tables and figures, based on systematic review of the literature. Guidance for the management of pediatric UC is summarized in algorithms to be used in conjunction with reading this document (Figs. 2 and 3). The goal of treatment in active UC should be complete clinical remission (PUCAI < 10 points), and usually this can be assessed without the need for endoscopic verification. Nonetheless, ∼20% of children in clinical remission can still have endoscopic inflammation, and thus calprotectin may aid in selecting those who require endoscopic evaluation to ensure mucosal healing has been achieved. The choice of treatment in adults is a factor of both the disease severity and disease extent ^(15,16), but since limited distal disease is uncommon in children, pediatric treatment strategy mainly depends on disease severity. Mesalamine regimes are considered first line for inducing and maintaining remission of mild-to-moderate UC. Non-response to oral mesalamine may be treated with the addition of mesalamine enemas and/or switching to locally active steroids, with budesonide-MMX only in left-sided colitis. In ambulatory children with moderate-to-severe UC, or in those with mild to moderate disease, who have failed optimized mesalamine therapy, oral steroids should be used, but only as induction agents. If the patient does not clearly respond to oral steroids within 1 to 2 weeks, consider admission for intravenous corticosteroids (see Part 2 of these guidelines). In refractory non-severe cases, an alternative to admission may include outpatient treatment with IFX (especially in those who failed thiopurines and mesalamine); in selected patients, oral tacrolimus may be considered.

Patients who received intravenous corticosteroids should be usually weaned to thiopurines. Almost all children with UC must be treated with a maintenance therapy indefinitely. Anti-TNF is indicated for non-response to corticosteroids, and in loss of response or intolerance to mesalamine and thiopurines. Patients needing anti-TNF induction should continue this therapy and if thiopurine-naïve, may be subsequently stepped down to thiopurines after a period of 6 to 12 months of deep remission. Golimumab or adalimumab should be considered in secondary loss of response to IFX due to antibody formation. Vedolizumab is a valid option in primary non-response to anti-TNF, in secondary loss of response in the presence of adequate drug level, and in anti-TNF related adverse events, such as refractory psoriasis. Endoscopic evaluation is recommended before any significant treatment change. Finally, colectomy is always a viable option, which must be discussed whenever treatment escalation is considered.

These clinical management guidelines were developed to assist practitioners at all levels of health care, while recognizing that each patient is unique. The recommendations may, thus, be subject to local practice patterns, and serve merely as a general framework for the management of UC in children. The development of the guidelines should now be followed by dissemination of the information to clinical practice.

DISCLAIMER

ESPGHAN and ECCO are not responsible for the practices of physicians and provide guidelines and position papers as indicators of best practice only. Diagnosis and treatment is at the discretion of physicians.

QUALIFYING STATEMENT and ACKNOWLEDGMENT

Please refer to the end of Part 2 of these guidelines