Secondary Logo

Journal Logo

Short Communications: Gastroenterology: Inflammatory Bowel Disease

Early Treatment Response Predicts Outcome in Paediatric Ulcerative Colitis: Gastroenterology: Inflammatory Bowel Disease

Gasparetto, Marco∗,†; Wong-Spracklen, Vivien; Torrente, Franco; Howell, Kate; Brennan, Mary; Noble-Jamieson, Gabriele; Heuschkel, Robert; Zilbauer, Matthias∗,†

Author Information
Journal of Pediatric Gastroenterology and Nutrition: August 2018 - Volume 67 - Issue 2 - p 217-220
doi: 10.1097/MPG.0000000000001941


What Is Known

  • Published data on predictors of disease outcomes in children with ulcerative colitis remain scarce.
  • Recent studies performed on small cohorts have suggested that early response to treatment correlates with longer-term outcome.

What Is New

  • Children with pancolitis, anaemia, and low albumin at diagnosis, who fail to achieve steroid-free remission or response to steroids at 3 months, have a >80% likelihood to require biologics and/or major surgery within 18 months.
  • We recommend that a 3-month check is used as a guide for future management including closer monitoring.

Ulcerative colitis (UC) causes chronic relapsing mucosal inflammation of the large bowel (1,2). The condition is increasingly being diagnosed in young children, where disease outcome tends to be significantly worse compared to onset in adult life (1–3). Disease outcome can, however, still vary substantially, with a few children responding well to minimal treatment, whereas approximately 5% to 10% will go on to undergo colectomy within 2 years follow-up (4–6).

Importantly, our ability to predict disease outcome at the point of diagnosis, or even in the early stages of disease, remains limited (6–8). Reliable predictors of disease outcome would enable a more personalized, and hopefully more effective, treatment approach. This should reduce the occurrence of unnecessary adverse events to medication and minimize unexpected life-threatening complications such as acute severe colitis.

Recent studies have suggested that early disease behaviour, particularly response to treatment within the first 6 months, strongly correlates with longer-term outcome (9–11).

Although only performed on small cohorts (maximum n = 117), these reports are encouraging and suggest that specific clinical, phenotypic, and serological markers can better predict long-term disease outcome during the early stages of paediatric UC (3,4,9–12). The best model so far at predicting a severe disease course in children does so with a certainty of 70% (1).

With the present study we extend the currently available outcome data in children with newly diagnosed UC, by analysing a further well characterized paediatric cohort.


Between September 2006 and February 2014, 93 children newly diagnosed with UC (or UC-like inflamamtory bowel disease (IBD)-unclassified) were captured in the departmental registry and clinical data were entered prospectively on a comprehensive electronic medical record. The study included all children who had a minimum of 18 months follow-up data, although the cohort had mean follow-up of 59.2 months ± SD 33.4 (median 49.7 months [range 18–190 months]) (Supplemental Table, Supplemental Digital Content 1,

Diagnosis of UC was made on clinical suspicion supported by appropriate macroscopic findings on colonoscopy, typical histological findings on biopsy, and negative stool examinations for infectious agents, as recommended by international guidelines (13).

Patients whose clinical presentation and macroscopic findings were consistent with a diagnosis of UC (eg, bloody diarrhoea, continuous colitic changes on endoscopic assessment), whose histological examination did not include typical findings of Crohn colitis (eg, granulomas, transmural inflammation) although it was also lacking characteristic features of UC, were included as UC-like IBD-U.

All children were diagnosed and treated in our regional centre, in which a strict standardized step-up treatment protocol was used for all children with UC. Accordingly, induction treatment at the point of diagnosis was limited to mesalazine (5-aminosalycilates [5-ASA]), a reducing course of systemic steroids, or both (Fig. 1). Escalation to treatment with biologics (ie, infliximab [IFX]) required corticosteroid resistance/dependence and/or immunosuppressant resistance/intolerance.

Tree chart with patient numbers showing groups 1, 2, and 3, defined according to requirement of treatment escalations. 5-ASA = 5-aminosalycilates, ASC = acute severe colitis, AZA = azathioprine, CAC = chronic active colitis, 6-MP = 6-mercaptopurine.

All patients requiring any treatment escalation including commencing biologics were discussed during a weekly multidisciplinary team meeting; therefore, the same indications to treatment escalations applied to all children in the study population.

According to the latest ESPGHAN guidelines, we defined steroid refractoriness as nonresponse to oral steroids within 7 to 14 days, and steroid dependency as remission with corticosteroids but recurrence of symptoms when the dose is lowered or within 3 months after complete taper, or if steroids cannot be stopped within 14 to 16 weeks (7).

In order to objectively assess the outcome and disease behaviour in our patient cohort, children were categorized into 3 clinically relevant, but distinct, groups. Specifically, children with a mild course were defined as those needing a maximum of 1 course of steroids in 18 months (ie, induction treatment at diagnosis with a weaning course of steroids) and maintenance treatment with 5-ASA only (Fig. 1, group 1). Patients with a moderate course required maintenance treatment with 5-ASA and azathioprine or 6-mercaptopurine (Fig. 1, group 2). Any child requiring escalation treatment to biologics and/or surgical intervention was categorized as having a severe course (Fig. 1, group 3).

No ethical approval was required as data were collected and analysed as part of a clinical service evaluation in advance of the implementation of the ImproveCareNow QI initiative (14).

To maximize the discriminating potential of our predictors, we compared the severe group (group 3) with the combined mild and moderate groups (groups 1 and 2). We analysed their disease distribution at diagnosis and the response to steroids at 3 months from diagnosis.

The Mann-Whitney test was used to compare serological markers (full blood count (FBC), albumin, erythrocyte sedimentation rate (ESR)) between all groups. P values ≤0.05 were considered significant. The score log-rank test (Cox regression) was used to analyse the difference in timeline for the escalation to azathioprine in the severe and moderate groups. The Kaplan-Meier analysis was used to determine how long after diagnosis patients required escalation to third-line treatment of biologics and/or surgery.

Receiver operating characteristic (ROC) curves were used to show the performance (sensitivity and specificity) of logistic regression models based on specific clinical variables to predict disease severity, including clinical response to steroids at 3 months, disease location at diagnosis (Paris classification), haemoglobin, ESR, and albumin at diagnosis.


Overall, 81.8% (n = 76) of children experienced either a mild (ie, 40.9%, n = 38) or a moderate (40.9% n = 38) disease course and 18.2% (n = 17) fell into the severe group (Fig. 1).

Survival analysis of the outcomes of our cohort revealed that 80% of children requiring escalation to treatment with biologics or colectomy, did so within 18 months from diagnosis (Supplemental Fig. 1, Supplemental Digital Content 2, Within the latter group, 2 children underwent colectomy without prior escalation to IFX. Amongst those who went on to receive IFX (n = 15), 12 were escalated because they had acute severe colitis with half of these going on to require colectomy (mean and median time to colectomy from diagnosis was 17 months).

We next compared clinical phenotypic parameters (ie, disease distribution), serological markers (haemoglobin, albumin, ESR) at diagnosis, PUCAI score (at diagnosis and at 3 months from diagnosis), and response to steroids at 3 months, between group 3 (severe) and groups 1 and 2 (mild and moderate). We observed significant differences in disease location, albumin, and haemoglobin levels at diagnosis, as well in response to treatment with steroids at 3 months (Supplemental Fig. 2, Supplemental Digital Content 3, Specifically, pancolitis, low albumin, and low haemoglobin levels at diagnosis, and nonresponse to steroids at 3 months, were all indicative of a severe disease course. In contrast, ESR levels at diagnosis did not show any significant difference between patient groups.

The PUCAI score in our cohort was not significantly different at diagnosis in the severe group compared to the mild and moderate groups (P 0.06). Nevertheless, there was a significant difference in the PUCAI score between the 2 groups at 3 months from diagnosis (P < 0.0001) (Supplemental Table, Supplemental Digital Content 1,

Lastly, we subjected significantly different parameters to logistic regression model analysis. As demonstrated in Figure 2, clinical response to treatment at 3 months combined with haemoglobin levels at diagnosis showed the strongest predictive value with an AUC of 0.87, predicting a severe outcome with a sensitivity of 0.82 and specificity of 0.9 (positive predictive value 0.64, negative predictive value 0.96).

ROC curves to show the performance (sensitivity and specificity) of logistic regression models based on specific clinical variables to predict disease severity. A, Clinical response to steroids at 3 months and haemoglobin. B, Prediction of severity for the “late severe” group using the model from A, 7 out of 9 patients would have been classified as severe. The colour scale along the ROC curve represents the model threshold, which allows the choice of an appropriate threshold to achieve the desired model sensitivity and specificity. Lines are drawn at the 0.85 sensitivity mark to aid in the choice of model threshold.


Our ability to predict the course of disease either at diagnosis or even during the early stages of disease remains limited. A small number of highly encouraging reports, however, suggest that distinct clinical phenotypes and biochemical parameters at diagnosis and/or during the first months after diagnosis may be indicative of future outcomes both in adults and children (1,2,4,8,11,12).

Our study adds further to these early reports by providing novel insights and validating a number of previous findings. Specifically, we confirm that disease extent at diagnosis was indicative of future outcome, with patients presenting with pancolitis being at a higher risk of suffering from a severe disease course (1–4).

In addition, systemic inflammatory markers at diagnosis did not differ significantly between groups (ie, P 0.1 for ESR), suggesting that serologically measured systemic inflammation at diagnosis is not indicative of later disease behaviour (2). In contrast, we observed significant differences in haemoglobin and albumin levels at diagnosis between patient groups.

Our findings also confirm reports by Schechter et al (1) that response to treatment at 3 months was found to be the strongest predictor of future disease outcome.

Similarly, the significant difference in PUCAI scores between the severe and the mild to moderate groups at 3 months from diagnosis, also validates the data showed by Schechter et al. This finding overall reflects how achieving remission at 3 months correlates with a milder disease course, whereas the severity of disease (by PUCAI) at diagnosis is not predictive of disease course.

Lastly, applying logistic regression analysis to our data revealed the highly predictive value of combining “haemoglobin levels at diagnosis” and “response to steroids at 3 months.” This combination predicts disease severity with a sensitivity of 0.82 and specificity of 0.9.

In summary, our data show that steroid-free clinical remission at 3 months is the most reliable clinical predictor of a severe disease course over 5 years.

Based on our study, failure to achieve remission at 3 months implies a >80% likelihood to require biologics and/or major surgery within 18 months. We therefore recommend that a specific 3-month review point is used to guide future management, with ongoing active disease prompting earlier escalation of therapy.


1. Schechter A, Griffiths C, Gana JC, et al. Early endoscopic, laboratory and clinical predictors of poor disease course in paediatric ulcerative colitis. Gut 2015; 64:580–588.
2. Yoon JY, Cheon JH, Park JJ, et al. Clinical outcomes and factors for response prediction after the first course of corticosteroid therapy in patients with active ulcerative colitis. J Gastroenterol Hepatol 2011; 26:1114–1122.
3. Gower-Rousseau C, Dauchet L, Vernier-Massouille G, et al. The natural history of pediatric ulcerative colitis: a population-based cohort study. Am J Gastroenterol 2009; 104:2080–2088.
4. Aloi M, D’Arcangelo G, Pofi F, et al. Presenting features and disease course of pediatric ulcerative colitis. J Crohns Colitis 2013; 7:e509–e515.
5. Aloi M, D’Arcangelo G, Capponi M, et al. Managing paediatric acute severe ulcerative colitis according to the 2011 ECCO-ESPGHAN guidelines: efficacy of infliximab as a rescue therapy. Dig Liver Dis 2015; 47:455–459.
6. Ruemmele FM, Turner D. Differences in the management of pediatric and adult onset ulcerative colitis—lessons from the joint ECCO and ESPGHAN consensus guidelines for the management of pediatric ulcerative colitis. J Crohns Colitis 2014; 8:1–4.
7. Turner D, Levine A, Scher JC, et al. European Crohn's and Colitis Organization; European Society for Paediatric Gastroenterology, Hepatology and Nutrition. Management of pediatric ulcerative colitis: joint ECCO and ESPGHAN evidence-based consensus guidelines. J Pediatr Gastroenterol Nutr 2012; 55:340–361.
8. Rinawi F, Assa A, Eliakim R, et al. Risk of colectomy in patients with paediatric-onset ulcerative colitis. J Pediatr Gastroenterol Nutr 2017; 65:410–415.
9. Turner D, Otley AR, Mack D, et al. Development, validation, and evaluation of a pediatric ulcerative colitis activity index: a prospective multicentre study. Gastroenterology 2007; 133:423–432.
10. Ledder O, Turner D. Seeing without looking: predicting mucosal healing without endoscopic evaluation in pediatric ulcerative colitis. J Pediatr Gastroenterol Nutr 2016; 64:848–849.
11. Diedern K, Hoekman DR, Leek A, et al. Raised faecal calprotectin is associated with subsequent symptomatic relapse, in children and adolescents with inflammatory bowel disease in clinical remission. Aliment Pharmacol Ther 2017; 45:951–960.
12. Waterman M, Knight J, Dinani A, et al. Predictors of outcome in ulcerative colitis. Inflamm Bowel Dis 2015; 21:2097–2105.
13. Levine A, Koletzko S, Turner D, et al. European Society of Pediatric Gastroenterology, Hepatology, and Nutrition. ESPGHAN revised porto criteria for the diagnosis of inflammatory bowel disease in children and adolescents. J Pediatr Gastroenterol Nutr 2014; 58:795–806.
14. Siegel CA, Allen JI, Melmed GY. Translating improved quality of care into an improved quality of life for patients with inflammatory bowel disease. Clin Gastroenterol Hepatol 2013; 11:908–912.

children; clinical outcomes; colectomy; steroid-free remission; ulcerative colitis

Supplemental Digital Content

Copyright © 2018 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition