Presenting features and disease course of pediatric ulcerative colitis

Abstract

Clinical variables and disease course of pediatric ulcerative colitis (UC) have been poorly reported. The aim of this study was to retrospectively describe the phenotype and disease course of pediatric onset UC diagnosed at a tertiary referral Center for Pediatric Gastroenterology.

Patients and methods: 110 patients with a diagnosis of UC were identified at our Department database. Records were reviewed for disease location and behavior at the diagnosis, family history for inflammatory bowel disease, pattern changes at the follow-up, need of surgery and cumulative risk for colectomy.

Results: Thirty-five % of patients had an early-onset disease (0–7 years). At the diagnosis, 29% had proctitis, 22% left-sided colitis, 15% extensive colitis and 34% pancolitis. Fifteen % presented with a rectal sparing, while a patchy colonic inflammation was reported in 18%. Rectal sparing was significantly related to the younger age (p: < 0.05). Disease extension at the follow up was reported in 29% of pts. No clinical variables at the diagnosis were related to the subsequent extension of the disease. The cumulative rates of colectomy were 9% at 2 year and 14% at 5 years. An extensive disease as well as acute severe colitis and corticosteroid therapy at the diagnosis were significantly associated with an increased risk of colectomy.

Conclusions: Pediatric UC is extensive and severe at the diagnosis, with an overall high rate of disease extension at the follow-up. Endoscopic atypical features are common in young children. The colectomy rate is related to the location and severity of the disease at the diagnosis.

Ulcerative colitis, Children, Atypical findings, Disease course, Early-onset

1 Introduction

Ulcerative colitis (UC) and Crohn's disease (CD) are chronic, unremitting, immune-mediated disorders, known as inflammatory bowel diseases (IBD). Up to 25% of IBD first develop in childhood and adolescence, with recent studies suggesting an increasing incidence of pediatric CD in the last 20–30 years, whereas the incidence of UC seems to be remained stable or slightly decreased.¹–⁵

UC is classically defined as a chronic inflammation characterized by a continuous involvement of the colonic mucosa in the absence of histological granulomas, affecting the rectum and a variable extent of the colon in continuity, with a course that is characteristically relapsing and remitting.⁶ However, it is widely known that pediatric UC may present with atypical phenotypes, thus the above mentioned definition is not always applicable to the pediatric-onset disease.⁷

The clinical features at the diagnosis and the natural history of pediatric UC are poorly described, and only few population-based studies are available, most of them having been carried out before the era of immunomodulators and biological agents. Some data do suggest that most of children with UC present with moderate/severely active disease already at the diagnosis and 25–30% will become corticosteroid-dependent by 1 year.⁸–¹⁰ The aggressive nature of the pediatric disease is marked by a relatively high rate of early colectomy: studies in 1990 showed a rate of colectomy around 25%¹¹ and, despite medical progresses, recent data have shown similar colectomy rates,¹² thus, raising the question as to whether the conventional therapeutic strategy in pediatric UC should be revised.

In the context of pediatric disease, early-onset (EO) UC is the most poorly described because this age group represents a small subgroup of IBD patients. However, few data suggest that it is more severe than the late-onset disease both in terms of phenotype and course.¹³,¹⁴

Given the high risk of a disabling course of disease and risk for surgery, children with UC have much to gain from therapeutic strategies altering the predicted course of the disease.¹⁵ To date, there are no data on the history of pediatric UC following the wide use in clinical practice of immunomodulators and biologics in early phases of the disease. It is thus warranted in this population to identify predictive factors of severe disease course in order to optimize the therapeutic armamentarium.

The aims of this study, performed in a cohort of UC children followed at a single tertiary referral center for Pediatric Gastroenterology, were to describe clinical features of early- and late-onset pediatric UC, to evaluate the use of current therapy, including infliximab, and responses to treatment, and to identify factors that may predict extension of the disease and need of surgery.

2 Patients and methods

This is a retrospective longitudinal study conducted at a tertiary referral Center for Pediatric Gastroenterology and Hepatology. The department database was screened for all children diagnosed with UC by 2006 to 2011. Diagnosis of UC was made using widely agreed clinical, endoscopic and histological criteria.¹⁶ All patients had a full colonoscopy with retrograde ileocolonoscopy. Until 2007, upper gastrointestinal endoscopy was mainly performed in patients who had not a definite diagnosis of UC based on colonoscopy. Since 2008, according to Porto criteria,¹⁶ all patients at the diagnosis had both lower and upper gastrointestinal endoscopy. After the diagnosis, all patients were followed according to a standardized protocol including an annual ileocolonoscopy.

Case records were reviewed for clinical variables at the diagnosis and during follow-up. EO-UC was defined as a diagnosis of UC made in children aged between 0 and 7 years. Disease location at diagnosis and at follow-up was defined according to the Paris classification.¹⁷ Ulcerative proctitis (E1) was defined as an involvement limited to the rectum (i.e., proximal extent of inflammation distal to the rectosigmoid junction). Left-sided UC (E2) as an involvement limited to the portion of the colorectum distal to the splenic flexure. Extensive UC (E3) as a disease extending proximally to the splenic flexure but distally to the hepatic flexure, while pancolitis (E4) included a colitis extended proximally to the hepatic flexure. Rectal sparing was defined as a normal rectum (endoscopically and/or histologically) when definite proximal disease was present. Patchy colonic inflammation was defined as areas of normal mucosa (endoscopically and/or histologically) amid areas of apparent inflammation. Disease extension was defined as the involvement, during follow-up, of at least one additional colonic segment. Disease activity was scored by the Pediatric UC Activity Index (PUCAI)¹⁸; an episode of severe colitis was defined as a PUCAI score > 65.¹⁹ Extra-intestinal manifestations (EIMs) included skin, joint and ocular manifestations, and primary sclerosing cholangitis. A family history for IBD was defined by the presence of CD or UC in first-degree relatives only. Laboratory tests included full blood count, C-Reactive Protein (CRP), Erythrocyte Sedimentation Rate (ESR), perinuclear Anti-Neutrophil Cytoplasmic Antibodies (p-ANCA), nutritional, renal and liver function parameters. Medical (i.e. mesalamine, corticosteroids, immunomodulators, infliximab) and surgical treatment received were also evaluated as well as intolerance or failure to therapy. One year after induction of the first course of corticosteroids (CS) treatment, CS-dependency and resistance were reported. The former was defined as a relapse of the disease within 30 days of cessation of treatment or relapse when dose reduction was attempted; the latter as a persistent active disease under CS therapy.¹⁵ The study protocol has been approved by the ethical committee of the hospital.

2.1 Statistical analysis

All data were summarized and displayed as the mean ± SD for the continuous variables. Categorical data were expressed as frequencies and percentages. Comparison of groups was performed using Student's t test for unpaired data in two group comparison and one way analysis of variance (ANOVA) with Bonferroni's test for multiple group comparison. Chi square test with Fisher's correction was used to evaluate the differences for categorical variables wherever needed. A p value of 0.05 or less was considered significant. The Kaplan–Meier survival method was used to estimate the interval free from colectomy and from disease extension during follow-up. Differences between curves were tested using the Log-Rank test. Factors associated with disease extension were analyzed in patients who may have extension of the disease, i.e., those with an initial diagnosis of proctitis or left-sided colitis and were identified using logistic regression analysis. Unifactorial analyses were performed using Fisher's exact test to calculate odds ratios (OR) and 95% confidence intervals (CI).

Assessment of the crude relation between colectomy and clinical variables was performed using the Pearson partial correlation for continuous variables, and the Spearman correlation procedure was used for categorical variables. The SPSS statistical package was used to perform all statistical evaluations (SPSS Inc., Chicago, IL).

3 Results

From 2006 to 2011 we identified 110 patients with a diagnosis of UC, 38% of them were males and median age was 10.2 years (range: 1.2–18.3). Mean follow-up was 48 months (range: 28–94). Median time between onset of symptoms and diagnosis was 4 months (range: 1–12), and 18% of children had a diagnostic delay exceeding 6 months. All patients had a full exploration of the gastrointestinal tract performed at diagnosis, including a complete ileocolonoscopy under general anesthesia or deep sedation. The most common location at diagnosis was pancolitis (E4) in 34% of patients, while 15% reported an extensive colitis (E3); 29% had a left-sided colitis (E2), while a proctitis (E1) was detected in 22% (Fig. 1). Fifteen % of patients presented with a rectal sparing, while a patchy colonic inflammation was reported in 18%. A family history for IBD was present in 22% of patients; EIMs were detected in 16% at the diagnosis: skin involvement (n = 7), axial arthropathies (n = 2), peripheral arthritis/arthralgia (n = 7), pancreatic involvement (n = 1) and osteopenia (n = 1). Table 1 lists the clinical characteristics at the diagnosis of UC patients. Thirty-five % (n = 39) of patients had an EO-disease (0–7 years). The latter significantly differed from those with late onset for a higher rate of rectal sparing (p: 0.04) and lower prevalence of left-sided colitis (p: 0.04) at the diagnosis, and for more frequent courses of corticosteroids at follow-up (p: 0.05). Other variables at the diagnosis, i.e. sex, family history for IBD, delay of the diagnosis (> 6 months), EIMs, mean PUCAI, CRP, and p-ANCA, did not differ between the two groups. Moreover, antimetabolite and infliximab use as well as colectomy rate at the follow-up were not significantly different between younger and older children (Table 2 ).

Ten % of patients presented a severe colitis at the diagnosis, while 23% presented at least one episode of severe colitis at the follow-up. Disease extension at the follow up was reported in 29% of patients. Among the latter with an initial E1 location, 12% and 38% progressed to E2 and to E3, respectively; among patients with an initial E2 location, 50% progressed to E3, 29% to E4 (Fig. 2). No clinical variables at the diagnosis, i.e. duration of symptoms before the diagnosis, age, gender, family history for IBD, location and endoscopic atypical findings, PUCAI, CRP, acute severe colitis and CS use, were related to the subsequent extension of the disease. At maximal follow-up, 74 patients (67%) had experienced EIMs: axial arthropathies (n = 20), peripheral arthritis/arthralgia (n = 35), skin manifestations (n = 3), and primary sclerosing cholangitis (n = 10). The prevalence of EIMs was significantly higher at maximal follow-up than at diagnosis (p < 0.0001).

At the diagnosis, 98 patients (89%) started 5-aminosalicylic acid (5-ASA) or sulfasalazine, while 41% received oral or intravenous corticosteroids and 10% were treated with topical steroids; 16% initiated thiopurines and 3% methotrexate; 3 patients (3%) began infliximab. At 1 year, 24 patients (22%) were steroid-dependent and 27 (25%) steroid-resistant. Ten % of 5-ASA treated patients were intolerant, and treatment was discontinued in 4% of them, most of whom under sulfasalazine. At 1 year, twenty-five % of patients were receiving immunomodulators and 6% were under infliximab; of the latter, 28% received combined therapy with thiopurines. At 2-year follow-up, 40% of patients were under antimetabolites and 15% under infliximab. At latest follow-up, 60% of patients were under corticosteroids, 52% were receiving thiopurines and 17% methotrexate. Twenty % of children were under infliximab, 23% of whom were in combination therapy with thiopurines. Only one patient was treated with cyclosporine during the follow-up.

Fifteen patients underwent surgery at least once between diagnosis and maximal follow-up, resulting in a crude colectomy rate of 14%. The cumulative probability of colectomy was 0.09 (95% CI: 0.05–0.14) at 2 years, and 0.14 (95% CI: 0.09–0.17) at 5 years after the diagnosis of UC (Fig. 3). At univariate analysis, the presence of extensive disease (Odds Ratio (OR) = 4.0 (95% CI: 0.8–18.2)), severe colitis (OR = 3.6, 95% CI: 1.1–11.6) and need of corticosteroids (OR = 4.7; 95% CI: 1.2–18.2) at the diagnosis were associated with an increased risk of colectomy (Table 3 ).

4 Discussion

In this retrospective study we report clinical data of a well-defined large cohort of pediatric patients with UC; moreover, we have described the presentation and course of early- (age 0–7 years) versus late-onset (> 8 years) disease.

In 49% of our patients UC was extensive at the diagnosis, while 29% of those with limited disease at the diagnosis had an extension at the follow-up. The initial location of the disease was similar to other pediatric studies,¹¹,¹²,²⁰ although recent data report much higher rate of extensive disease at the diagnosis.⁷,²¹,²² We did find a high overall rate of proximal disease extension at the follow-up: thus, at the end of the follow-up, 61% of patients had an extensive disease. These data are consistent with previously reported studies on the natural history of pediatric UC.⁷,¹² Disease extension has been reported as an important characteristic of UC that might alter long-term prognosis.²³ Interestingly, we did not identify clinical variables able to predict which patients will undergo disease extension. However, we cannot exclude that the small sample size limited the power of our study when variables associated with risk factors related to disease extension were evaluated. Sixty-two percent of our patients were females. This figure is higher than most of pediatric UC data, reporting an equal prevalence of males and females. However, as in our population, few studies indicate a slightly higher prevalence of the disease in females even in pediatric UC, as in the adult-onset disease.²¹ An explanation for the high proportion of females in our study may be that many patients in the age group > 8 years were adolescents in the postpubertal period, thus close to adulthood.

Atypical features at the initial colonoscopy were detected in 32% of patients, 15% of them had rectal sparing and 18% patchy colonic inflammation. Atypical features at the endoscopy in UC children have been reported to vary between 0 and 30%, particularly in those with EO-UC.⁷,²⁴–²⁷ According to these data, the prevalence of atypical findings in our cohort was inversely associated with age. These results highlight the importance of a complete diagnostic work-up in children with bloody diarrhea, always performing colonoscopy with ileal inspection (and multiple biopsies).

Our data suggest a high prevalence of extraintestinal manifestations associated with pediatric UC. Recent population-based studies report extraintestinal manifestations in IBD ranging from 6 to 47%,²⁸–³⁰ with several reports indicating even a greater prevalence in children.³¹,³² Our results are probably due to the inclusion of all reported extraintestinal manifestations associated with IBD, including arthralgia, which was very frequent in our patients.

An EO-UC was found in 25% of patients. Clinical data on the EO-UC are poor because this age group represents a small proportion of the pediatric IBD population. Heyman reported on 1370 IBD cases of age less than 16 years enrolled in the Pediatric IBD Consortium from January 2000 to November 2002³³: in total, 393 were classified as UC, and 85 UC cases (22%) of age < 5 years at diagnosis. Ten percent of patients were less than 5 years of age at diagnosis in Griffiths'³⁴ report on children with UC diagnosed from 1980 to 1999. The highest percentage in our study is likely due to the inclusion in the EO-UC of patients aging in the range from 0 to 7 years. Interestingly, we did not find a significant difference in terms of family history for IBD between early- and late-disease onset patients. Previous studies have reported an increased rate of familial aggregation in EO-UC, suggesting that this subgroup may represent a more highly penetrant and genetically distinct form of UC.¹³ Outside of cases of very EO forms of severe colitis where a strong genetic susceptibility is demonstrated (mutations in IL-10 and IL-10 receptor genes),³⁵–³⁷ further studies are needed to assess whether EO-UC is more genetically determined than the late onset forms.

Some data report a higher severity in terms of phenotype of EO-UC,³³ whereas data on disease course are scarce. In our cohort, children with EO showed a high prevalence of extensive colitis, although not significantly different from older children. Conversely, the prevalence of left-sided colitis was significantly lower in younger children. Our results are consistent with those of Paul et al.¹³ reporting a similar distribution of disease locations when comparing children 0–5 years old with older pediatric patients. During follow-up, our younger patients received significantly more courses of corticosteroids, although there were no significant differences in the use of antimetabolites and infliximab compared to older children. This may suggest a more aggressive disease course of EO-UC, although, at the maximal follow-up, the colectomy rate was not significantly different.

The majority of children in our cohort were treated with mesalamine at least once over the disease course, in accordance with recently published ECCO and ESPGHAN guidelines that recommend oral 5-ASA regimens as first-line induction therapy for mild to moderately active pediatric UC and for maintenance of remission, regardless of other initial treatments.³⁸ Systemic corticosteroids were given in 60% of our patients at same time in the disease course, a figure similar to that reported by Gower-Rousseau et al.¹² in a cohort of 113 children with UC newly diagnosed in France. At 1 year, corticosteroid-dependence occurred in 22% of cases, whereas corticosteroid-resistance was reported in 25%. These data are consistent with previous North American report on the rate of corticosteroid dependence in children and adults.¹⁵ At the diagnosis, 19% of patients started an immunomodulator, while 3% were treated with infliximab. This percentage increased to 40% and 15%, respectively, at 2-year follow-up. At the maximal follow-up, 52% of patients were under antimetabolites and 20% under infliximab. This high rate of “aggressive” therapy could be related to the colectomy rate of 14% at 5-year follow-up. Previous French and North American studies, conducted in the pre-infliximab era, reported higher colectomy rates in cohorts of children with UC.¹¹,¹²,³⁹ To our knowledge, this is the first study reporting data on the impact of the wide and early use of immunomodulators and biologics on the natural history of pediatric UC. Our data seem to suggest that an early use of immunomodulators and biologics might reduce the need of surgery in a 5-year follow-up. However, this study has the limitations of being retrospective and single-center, thus these results must be further confirmed by prospective, large and with longer follow-up studies, specifically aimed at evaluating the impact of early aggressive therapies in modifying the natural history of pediatric UC.

Furthermore, our data show that an extensive and severe disease at the diagnosis are the main risk factors for colectomy in early and late-onset pediatric UC, suggesting that an aggressive therapeutic strategy in these forms of UC could impact on the need of surgery.

In conclusion, early and late-onset pediatric UC are characterized by an extensive and severe phenotype at the diagnosis with a high rate of disease extension. Rectal sparing and patchy colonic inflammation are commonly recognized at the diagnosis in EO-UC. Aggressive therapy seems to be effective in reducing the need of surgery, but prospective and controlled studies are warranted to evaluate the impact of new therapeutic strategies in pediatric UC.

Conflict of interest

There are no prior publications or submissions with any overlapping information, including studies and patients. All the authors declare no financial relationships with a commercial entity producing health-related products and or services related to this article. No honorarium, grant, or other form of payment was given to anyone to write and to produce the manuscript.

Each author has seen and approved the submission of this version of the manuscript and takes full responsibility for the manuscript.

Acknowledgments

The specific contribution of each author to this manuscript was the following:

• Marina Aloi, M.D., Ph.D., designed the study and wrote the manuscript.

• Giulia D'Arcangelo, M.D., contributed to design the study and to write the manuscript.

• Fabrizio Pofi, M.D., acquired the data and revised the manuscript.

• Francesca Vassallo, Valentina Rizzo, Federica Nuti, M.D., performed the acquisition, analysis and interpretation of data.

• Giovanni Di Nardo, M.D., and Maria Pierdomenico, Ph.D., interpreted data and made the statistical analysis.

• Franca Viola, M.D., contributed to draft and revise the article critically for important intellectual concept.

• Salvatore Cucchiara, M.D., Ph.D., the Head of the Pediatric Gastroenterology Unit, strongly revised a draft of the paper and approved the version to be published.

Figure 1

Disease location at the diagnosis according to Paris classification in 110 children with ulcerative colitis. E1: proctitis; E2: left-sided colitis; E3: extensive colitis; E4: pancolitis.

Figure 2

Location of the disease at the diagnosis and at latest follow-up according to Paris classification in 110 pediatric patients with ulcerative colitis. E1: proctitis; E2: left-sided colitis; E3: extensive colitis; E4: pancolitis.

Figure 3

Cumulative risk of colectomy in 110 pediatric patients with ulcerative colitis.

Table 1

Baseline clinical characteristics of 110 children with ulcerative colitis.

Mean follow up time (months)	48 (range: 28–94)
Sex (F)	68 (62%)
Diagnostic delay > 6 months	20 (18%)
Family history for IBD (first degree relatives)	24 (22%)
PUCAI at the diagnosis (mean ± SD)	43.5 ± 11.8
Location of the disease at the diagnosis (Paris classification)
E1	32 (29%)
E2	24 (22%)
E3	17 (15%)
E4	37 (34%)
Rectal sparing	17 (15%)
Patchy inflammation	20 (18%)
Extraintestinal manifestations	18 (16%)

IBD: inflammatory bowel disease; PUCAI: Pediatric ulcerative colitis activity index.

Table 2

Clinical characteristics of early- and late-onset UC.

	0–7 years N = 34	8–18 years N = 76	p
Sex (F)	22 (65%)	46 (61%)	0.83
PUCAI (mean ± SD)	45.5 ± 12.4	38.4 ± 12.4	0.12
Family history for IBD (first degree relatives)	9 (26%)	15 (20%)	0.45
CRP (mean ± SD)	4.8 ± 4.7	5.7 ± 9.1	0.74
pANCA	5 (13%)	15 (21%)	0.46
Delay of the diagnosis (> 6 months)	7 (20%)	12 (16%)	0.58
Location of the disease at the diagnosis (Paris classification)
E1	11 (35%)	21 (28%)	0.65
E2	3 (9%)	21 (28%)	0.04
E3	5 (15%)	12 (16%)	0.98
E4	14 (41%)	23 (30%)	0.28
Rectal sparing	9 (27%)	8 (11%)	0.04
Patchy colonic inflammation	5 (15%)	15 (20%)	0.63
Acute severe colitis at the diagnosis	3 (9%)	7 (9%)	0.95
Extraintestinal manifestations	7 (18%)	24 (35%)	0.15
Courses of steroids at the follow-up	21 (56%)	25 (35%)	0.05
Antimetabolite use at the follow-up	10 (25%)	24 (33%)	0.58
Infliximab at the follow-up	8 (21%)	7 (10%)	0.17
Colectomy at 5-year follow-up	5 (15%)	10 (13%)	0.77

PUCAI: Pediatric ulcerative colitis activity index; CRP: C-reactive protein; pANCA: perinuclear anti-neutrophil cytoplasmic antibodies.

Table 3

Risk factors for colectomy in a population of 110 pediatric patients with ulcerative colitis.

	Univariate analysis
	OR	95% CI	p
Age (≥ 8 vs ≤ 7 years) at the diagnosis	2.73	0.89–8.37	0.13
Family history for IBD (first degree relatives)	0.45	0.095–2.16	0.49
Acute severe colitis at the diagnosis	3.63	1.13–11.62	0.05
EIMs vs no EIMs at diagnosis	0.66	0.17–2.54	0.77
Gender (male)	0.72	0.22–2.34	0.82
Use of systemic steroids at diagnosis	4.78	1.26–18.1	0.02
Disease extension at follow-up	2.4	0.8–7.4	0.12
Disease location at the diagnosis (Paris classification)
E1	0.14	0.01–1.1	0.06
E2	1.3	0.39–4.7	0.73
E3	7.4	2.2–24.8	0.001
E4	0.44	0.11–1.7	0.37

IBD: inflammatory bowel disease; EIMs: extraintestinal manifestations.