THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 97, No.

8, 2002
© 2002 by Am. Coll. of Gastroenterology ISSN 0002-9270/02/$22.00
Published by Elsevier Science Inc. PII S0002-9270(02)04272-7

Inflammatory Bowel Disease in Children 5 Years of

Age and Younger
Petar Mamula, M.D., Grzegorz W. Telega, M.D., Jonathan E. Markowitz, M.D., Kurt A. Brown, M.D.,
Pierre A. Russo, M.D., David A. Piccoli, M.D., and Robert N. Baldassano, M.D.
Division of Gastroenterology and Nutrition, The Children’s Hospital of Philadelphia, Philadelphia,
Pennsylvania

OBJECTIVES: Clinicians are becoming increasingly aware inflammatory bowel disease (IBD) combines both environ-
that inflammatory bowel disease (IBD) can affect all age mental factors and an altered immune response in geneti-
groups, although it has not been well described in infants cally predisposed patients, which then leads to chronic in-
and young children. Our aim was to evaluate early onset flammation of the intestinal tract (1). Recently, a mutation in
IBD in patients 5 yr of age and younger. the gene known as Nod2, which resides on chromosome 16
and encodes the protein recognizing lipopolysaccharides
METHODS: Medical records of patients diagnosed with early
(LPS) of the bacterial outer cell wall membrane, has been
onset IBD at The Children’s Hospital of Philadelphia be-
identified twice as frequently in patients with CD as in the
tween 1977 and 2000 were reviewed. Patients were divided
general population (2, 3). In patients with CD, the inability
into three categories: those with Crohn’s disease (CD), those
to recognize LPS may lead to an exaggerated inflammatory
with ulcerative colitis (UC), and those with indeterminant
response without the innate immune system modulation.
colitis (IC).
Living in a more sterile environment has been shown to
RESULTS: A total of 82 patients fulfilled the criteria. In 12 delay exposure to enteric infections in early life, possibly
patients (15%), the IBD diagnosis was changed during the resulting in failure of the normal maturation process neces-
course of illness. At the end of the follow-up period, linear sary to develop normal intestinal tolerance (4 – 6).
growth failure was present in 10 of 35 (29%) children with During the last several decades, the incidence of IBD in
CD, one of 30 (3%) with UC, and three of 17 (18%) with IC. adults is increasing (7). Pediatric epidemiological studies
Failure to thrive was a frequent presenting symptom in indicate the same patterns of increased incidence for both
children with CD (44%) and IC (39%), whereas in all four types of IBD (8 –11). The reasons for this increase are
patients with UC and failure to thrive the diagnosis was unclear, and the contributing factors may occur early in life.
subsequently changed to CD or IC. A high proportion of Very young patients with IBD require special attention
patients with CD had large bowel (89%), and perianal (34%) because of the potentially large impact of the disease on
disease. None of the tested patients were positive for anti– their growth and development.
Saccharomyces cerevisiae antibody (ASCA), and 10 tested The aim of the present study was to describe the present-
positive for perinuclear antineutrophil cytoplasmic antibody ing signs and symptoms along with the disease progression
(three of five patients with CD, five of seven with UC, and in children 5 yr of age and younger with early onset inflam-
two of three with IC). matory bowel disease (EO-IBD). This information may help
to improve care for children who present with IBD at this
CONCLUSIONS: Failure to thrive, at the time of presentation,
early age.
is indicative of a final diagnosis of CD or IC, not UC. Linear
growth failure is a common finding in patients with early
onset CD. A high proportion of patients with CD have
MATERIALS AND METHODS
failure to thrive, colonic, and perianal disease. The IBD
serology panel is of limited clinical relevance in providing A retrospective analysis of the database with all IBD pa-
definitive diagnostic information in this pediatric tients followed at the Children’s Hospital of Philadelphia
population. (Am J Gastroenterol 2002;97:2005–2010. between 1977 and 2000 was performed. A total of 94
© 2002 by Am. Coll. of Gastroenterology) patients diagnosed with EO-IBD were identified. Twelve
patients were excluded from the study: three did not meet
the histological criteria for IBD after review by one pathol-
INTRODUCTION
ogist (P.R.), and medical records were incomplete for six
Ulcerative colitis (UC) and Crohn’s disease (CD) are patients and were not available for three. Among the six
chronic inflammatory bowel diseases of unclear etiology. patients with incomplete records whose diagnoses could not
The prevailing hypothesis regarding the pathogenesis of be confirmed, four carried the diagnosis of UC and two of
2006 Mamula et al. AJG – Vol. 97, No. 8, 2002

CD. In all, 82 patients were included in the study. The

diagnoses were confirmed by standard endoscopic, histo-
logical, and radiographic criteria (12). The patients were
divided into three categories: those with CD, those with UC,
and those with indeterminant colitis (IC). The diagnosis of
IC was assigned to patients with chronic inflammation lim-
ited to the large intestine for whom (on the basis of clinical,
laboratory, radiological, endoscopic, or histological criteria)
it was not possible to distinguish between CD and UC. All
patients had colonoscopy or flexible sigmoidoscopy per-
formed at the time of diagnosis, and all but one patient had
a radiographic study during the course of the disease (upper
GI series with a small bowel follow-through, or barium
enema). The remaining patient had the diagnosis made at the
time of surgery. The medical records and x-ray reports were
reviewed by the two investigators (P.M. and G.T.) and by
Figure 1. Changes in diagnoses among study patients over time.
the patients’ primary gastroenterologists. The diagnosis of
the small intestinal CD was made based on the radiographic
study. of 19 patients (6%) diagnosed with EO-IBD below the age
The following characteristics were investigated: age, sex, of 2 yr was diagnosed with UC.
diagnoses and change in diagnoses over time, length of In all, 23% of patients with CD and UC and 18% of
follow-up, presenting symptoms, diagnostic time lag, med- patients with IC had a family history of IBD. When 13
ications, surgical procedures, anatomic location of the dis- patients whose family history was not available were ex-
ease, growth data, family history, and results of anti–Sac- cluded, 28% of patients with CD and UC and 20% of
charomyces cerevisiae antibody (ASCA) and perinuclear patients with IC had a family history of IBD. The age of
antineutrophil cytoplasmic antibody (p-ANCA) testing. The onset of disease in relatives was not available.
family history was defined as a history of IBD in parents, Linear growth failure was present in 10 of 35 children
grandparents, aunts or uncles, and first cousins. Linear (29%) with CD, one of 30 (3%) with UC, and three of 17
growth failure was defined as height below the fifth percen- (18%) with IC at the end of the follow-up period.
tile on a height for age growth curve developed by the At the time of initial diagnosis, 26% of patients with CD
National Center for Health Statistics in 1979, and failure to had inflammation in the stomach, 22% in the duodenum,
thrive (FTT) as weight below the fifth percentile on a weight 19% in the small bowel, 52% in the terminal ileum, and 89%
for age growth curve. Measurements of linear growth were in the large bowel. In all, 34% had perianal disease. Of the
obtained at the end of the follow-up period. Weight mea- CD patients, 60% had nonstricturing nonpenetrating (in-
surements for definition of failure to thrive were recorded at flammatory) type of the disease, 34% penetrating (fistuliz-
the time of diagnosis. Perianal disease was defined as peri- ing), and 6% stricturing type. Of the patients with UC, 60%
anal fistula or perianal abscess. had isolated left-sided colitis, and 40% had pancolitis.
The association between presenting symptoms and initial Presenting symptoms are shown in Table 1. Blood in the
diagnoses were evaluated using univariate analysis with the stool (hematochezia) was more commonly associated with
␹2 test. Statistical analysis was performed using Stata ver- UC than with IC and CD combined (p ⫽ 0.0002). Failure to
sion 6.0 software (Stata, College Station, TX). This study thrive at the time of initial presentation was more common
was approved by the Institutional Review Board of the in patients with CD or IC than in those with UC (p ⫽ 0.004).
Children’s Hospital of Philadelphia. All four patients initially diagnosed with UC who had FTT
as a presenting symptom had their diagnosis changed to IC
or CD (p ⬍ 0.0001). Chronic fever was associated with CD
RESULTS and not with UC or IC (p ⫽ 0.015). Vomiting was associ-
ated with CD or IC but not UC (p ⫽ 0.01). The median
A total of 82 patients (47 male and 35 female; sex ratio, diagnostic lag (i.e., median time between onset of symptoms
1.34) were diagnosed with EO-IBD. The median follow-up and time of diagnosis) was 4.5 months for patients with CD,
was 7.5 yr (range, 6 months to 23 yr). Initially, 36 children 2 months for UC, and 6.5 months for IC.
(44%) were diagnosed with UC, 27 (33%) with CD, and 19 Serological testing for pANCA and ASCA was per-
(23%) with IC. formed in 15 children before the age of 6 yr. All samples
Figure 1 shows the changes in diagnoses. The majority of were tested by the Prometheus Laboratories (San Diego,
patients (nine of 12) in whom the diagnosis was changed CA). None of the patients had positive ASCA test results,
had the initial diagnosis made before 1991. The age distri- but 10 patients had positive results for pANCA. Three of
bution and final diagnoses are shown in Figure 2. Only one five patients (60%) with CD, five of seven (71%) with UC,
AJG – August, 2002 IBD in Children 5 Years of Age and Younger 2007

Figure 2. Age at time of diagnosis and final diagnosis.

and two of three (66%) with IC had positive pANCA results. DISCUSSION
The presence of pANCA antibodies was not significantly
more prevalent in UC than in CD or IC. Reports of young children with IBD have been published for
several decades (13–17). Most case series have involved
Corticosteroids were used before the age of 6 yr in 67%
children of different ages, and only a few have concentrated
of patients with CD, 67% of patients with UC, and 52.6% of
exclusively on young children (18, 19). Recently, evidence
patients with IC. 6-Mercaptopurine was used before the age
for increased incidence of IBD not only in adults but also in
of 6 yr in 22%, 11%, and 26% of patients with CD, UC, and
children and adolescents has caused renewed interest in the
IC, respectively.
younger age group (19). So far, this is the largest study of
During the course of their disease, 10 patients (12%) children aged 5 yr and younger who have been diagnosed
required surgical intervention: five patients (14%) with CD, with EO-IBD.
four (13%) with UC, and one (6%) with IC. The patient with The genetic factors in the etiology of IBD are well rec-
IC underwent colectomy because of toxic megacolon. Three ognized, with a high rate of concordance between monozy-
patients with CD had resection of a stricture, one had a gotic twins (44.4%) compared with dizygotic twins (3.8%)
diverting colostomy secondary to severe perianal disease among patients with CD (20). A recent report described
and stricture, and one had perforation requiring partial co- multiple siblings affected with CD (21). In our series, 28%
lectomy and recto-vaginal fistula repair. Four patients with of patients with CD had a family history of IBD. This
UC had a colectomy performed: one for fulminant colitis correlates with the data from a combined adult and pediatric
unresponsive to i.v. corticosteroids and cyclosporin, and study, in which 29.9% of patients diagnosed with CD before
three for severe disease with failure of chronic immuno- age 20 yr had a positive family history compared with only
modulatory therapy. 13.6% of patients whose diagnoses were made at a later age

Table 1. Presenting Symptoms of Study Patients

Symptom CD (n ⫽ 27) UC (n ⫽ 36) IC (n ⫽ 18)
Diarrhea 81% 79% 89%
Blood in the stool* 67% 94% 67%
Abdominal pain 67% 33% 33%
Failure to thrive* 44% 11% 39%
Perianal disease 34% 0% 0%
Vomiting* 15% 0% 22%
Constipation 4% 0% 6%
Chronic fever* 11% 0% 0%
* p ⬍ 0.05.
2008 Mamula et al. AJG – Vol. 97, No. 8, 2002

(22). Genetic anticipation, which was described in parent- Presenting symptoms can be helpful in differentiating
child pairs in Huntington’s disease (23), may explain this between CD and UC. Hematochezia was more common in
finding. Genetic anticipation is a concept whereby the af- UC than in CD or IC. Vomiting was associated with CD and
fected offspring manifest the disease at an earlier age, and IC, and chronic fever was associated exclusively with CD.
sometimes in a more severe form, than does the affected Most importantly, in this cohort with early onset CD and IC,
parent. In the future, it will be of interest to investigate how failure to thrive (FTT) as a presenting symptom was present
many children with EO-IBD are positive for the recently in 44% and 39% of patients, respectively. Four patients
discovered gene Nod2 associated with CD (2). (12%) initially diagnosed with UC had FTT as a presenting
Our finding of only one patient diagnosed with UC within symptom. Interestingly, all four patients had their diagnosis
the first 2 yr of life differs from the report by Gryboski, in subsequently changed from UC to IC or CD. One of these
which UC was found to be more common than CD at this patients was subsequently found to have granulomas in the
age (19). The reason is unclear. Of the patients with CD, large intestine, one had severe inflammation of the terminal
60% were diagnosed during the period between 1991–2000 ileum on a repeat colonoscopy, one had persistent inflam-
in our study, whereas the study period in Gryboski’s report mation of the ileum after the colectomy, and one was found
ended in 1990. It is possible that changes in the epidemiol- to have patchy inflammation on subsequent colonoscopies.
ogy of the disease, as well as changes in the practice of At the same time, none of the remaining 32 patients initially
performing full colonoscopies as opposed to flexible sig- diagnosed with UC had FTT. We therefore conclude that
moidoscopies, thus allowing histological sampling of the failure to thrive is a presenting symptom that is predictive of
terminal ileum, contributed to the difference. CD or IC and not UC. The only other available study that
A striking number of patients with CD (29%) were noted examined FTT as one of the symptoms of IBD (33) de-
to have linear growth failure at the final follow-up, indicat- scribed it in 25% of patients with CD, which is less than in
ing that despite early diagnosis and treatment, a significant our series, and in none in UC, which corresponds to our
proportion of patients are not able to achieve appropriate results. This may possibly be due to more severe disease at
growth. Previous studies have reported growth failure in the time of diagnosis in our group of patients with CD.
7–30% of patients with CD (24). In a study by Kanof et al. More than 20% of patients with EO-IBD had the diag-
(25), 25% of patients with CD developed severe linear nosis of indeterminant colitis. In a large, multicenter study
growth failure (height below the fifth percentile), and in a of European adults, 5% of newly diagnosed patients were
study by McCaffery et al. (26), 18% of patients with IBD diagnosed with IC (35). The large proportion of children
demonstrated height below the third percentile. Growth with large bowel involvement could possibly explain this
failure most likely occurs because of a combination of difference. In pediatric series of older children, 14 –23%
several detrimental factors: malnutrition, the effects of in- were diagnosed with IC (8), which is similar to our results.
flammatory cytokines, and iatrogenic causes (e.g., cortico- Changes in diagnoses occurred more frequently in patients
steroid therapy). Nutritional therapy has previously been whose diagnoses were made before 1990. This could be
shown to be beneficial in improving growth (27). However, explained both by the longer duration of follow-up, allowing
nine of 10 patients with CD and linear growth failure in this the establishment of correct diagnosis, or by improvements
group received either nasogastric or gastric tube feeding in the technical aspects of pediatric colonoscopy during the
during the course of their illness. A new form of biological last decade, allowing better visualization and tissue sam-
therapy directed against tumor necrosis factor–␣ with a pling of the terminal ileum.
potential for improved histological healing may have a Serological assay is a potentially important addition to the
beneficial effect on the final growth in patients with CD diagnostic armamentarium in IBD. Combined measurement
(28 –32). of pANCA and ASCA has been advocated as a valuable
The anatomic distribution of the disease in this group of diagnostic approach in older children and adults. The com-
patients with CD is different from that in previously pub- bination of positive pANCA and negative ASCA had 57%
lished studies in older children and adolescents (33). In our sensitivity and 97% specificity for the diagnosis of UC, and
study, isolated small bowel disease was seen in only 11% of the combination of positive ASCA and negative pANCA
patients, isolated large bowel disease in 30%, and small and had 47% sensitivity and 97% specificity for the diagnosis of
large bowel disease in 59%, resulting in a total of 89% of CD (36). No study with long term follow-up has been
patients with large bowel disease. Pooled data of 14 pedi- performed to validate the use of these tests to differentiate
atric studies involving a total of 1153 older children re- between UC and CD in a setting of indeterminate colitis
vealed that only 58% of patients had large bowel disease (37). In a study of older children, Ruemmele et al. reported
(34). Another study reported that patients younger than 20 that the combined test had 57% sensitivity and 92% speci-
yr were more likely to have small intestinal disease when ficity for UC, and 55% sensitivity and 95% specificity for
compared with patients 40 yr and older (22). The reasons for CD (38). None of the patients with EO-IBD in our study had
the noted difference are unclear. However, patients with CD a positive ASCA. One can postulate that several years of
diagnosed at a very early age may represent a new subgroup exposure to Saccharomyces cerevisiae in an individual with
with a distinct anatomic disease distribution. increased intestinal permeability are necessary to produce
AJG – August, 2002 IBD in Children 5 Years of Age and Younger 2009

detectable ASCA levels (38). The percentage of positive 7. Logan RF. Inflammatory bowel disease incidence: Up, down
pANCA was similar among UC and CD patients. Because or unchanged? Gut 1998;42:309 –11.
8. Hildebrand H, Fredrikzon B, Holmquist L, et al. Chronic
only 15 patients had testing performed before the age of 6
inflammatory bowel disease in children and adolescents in
yr, the sample size may be too small to draw reliable Sweden. J Pediatr Gastroenterol Nutr 1991;13:293–7.
conclusions. A larger study is necessary to confirm these 9. Cosgrove M, Al-Atia RF, Jenkins HR. The epidemiology of
important findings. paediatric inflammatory bowel disease. Arch Dis Child 1996;
Finally, surgery was performed in 12.2% of patients dur- 74:460 –1.
10. Barton JR, Gillon S, Ferguson A. Incidence of inflammatory
ing the course of their disease, a higher proportion than the
bowel disease in Scottish children between 1968 and 1983;
5% of patients reported in a pediatric series of patients under marginal fall in ulcerative colitis; three-fold rise in Crohn’s
10 yr of age (19). Another study of Crohn’s disease in disease. Gut 1989;30:618 –22.
children reported 50% and 70% of patients requiring sur- 11. Askling J, Grahnquist L, Ekbom A, et al. Incidence of paedi-
gery within the first 10 and 15 yr of diagnosis, respectively atric Crohn’s disease in Stockholm, Sweden. Lancet 1999;354:
1179.
(39). Among adult patients with UC, almost one half will
12. Ogorek CP, Fisher RS. Differentiation between Crohn’s dis-
undergo surgery within the first 10 yr of their illness, and ease and ulcerative colitis. Med Clin North Am 1994;78:
more than 75% of patients with CD will have surgery during 1249 –58.
the first 20 yr of the disease (40). Length of the follow-up is 13. Farmer RG, Michener WM. Prognosis of Crohn’s disease with
most likely the reason for the small proportion of children onset in childhood or adolescence. Dig Dis Sci 1979;24:
752–7.
requiring surgery in our series. Continued follow-up may
14. Ein SH, Lynch MJ, Stephens CA. Ulcerative colitis in children
provide further information. Also, a small number of pa- under one year: A twenty-year review. J Pediatr Surg 1971;
tients in this group were found to have small intestinal CD, 6:264 –71.
which may be associated with a greater need for surgery 15. Guttman FM. Granulomatous enterocolitis in childhood and
because of stricturing disease. adolescence. J Pediatr Surg 1974;9:115–21.
In summary, we describe a unique subgroup of young 16. Miller RC, Larsen E. Regional enteritis in early infancy. Am J
Dis Child 1971;122:301–11.
patients with EO-IBD. Accurate differentiation between CD 17. O’Donoghue DP, Dawson AM. Crohn’s disease in childhood.
and UC in this age group is very difficult. We noted a high Arch Dis Child 1977;52:627–32.
proportion of patients with Crohn’s disease with linear 18. Gryboski JD. Ulcerative colitis in children 10 years old or
growth failure and large bowel disease. Failure to thrive at younger. J Pediatr Gastroenterol Nutr 1993;17:24 –31.
the time of presentation was predictive of a final diagnosis 19. Gryboski JD. Crohn’s disease in children 10 years old and
younger: Comparison with ulcerative colitis. J Pediatr Gastro-
of CD or IC and not UC. The combined testing of pANCA enterol Nutr 1994;18:174 – 82.
and ASCA is of limited clinical use for differentiating 20. Tysk C, Lindberg E, Jarnerot G, et al. Ulcerative colitis and
between CD and UC in children 5 yr of age or younger. Crohn’s disease in an unselected population of monozygotic
and dizygotic twins. A study of heritability and the influence
of smoking. Gut 1988;29:990 – 6.
Reprint requests and correspondence: Petar Mamula, M.D., 21. Cohen Z, Weizman Z, Kurtzbart E, et al. Infantile colonic
Division of GI & Nutrition, The Children’s Hospital of Philadel- Crohn’s disease: A report of four cases in one family. J Pediatr
phia, 34th Street & Civic Center Boulevard, Philadelphia, PA Gastroenterol Nutr 2000;30:461–3.
19104. 22. Polito JM II, Childs B, Mellits ED, et al. Crohn’s disease:
Received Oct. 10, 2001; accepted Jan. 29, 2002. Influence of age at diagnosis on site and clinical type of
disease. Gastroenterology 1996;111:580 – 6.
23. Snell RG, MacMillan JC, Cheadle JP, et al. Relationship
REFERENCES between trinucleotide repeat expansion and phenotypic varia-
tion in Huntington’s disease. Nat Genet 1993;4:393–7.
1. Papadakis KA, Targan SR. Current theories on the causes of 24. Ballinger AB, Camacho-Hubner C, Croft NM. Growth failure
inflammatory bowel disease. Gastroenterol Clin North Am and intestinal inflammation. Q J Med 2001;94:121–5.
1999;28:283–96. 25. Kanof ME, Lake AM, Bayless TM. Decreased height velocity
2. Ogura Y, Bonen DK, Inohara N, et al. A frameshift mutation in children and adolescents before the diagnosis of Crohn’s
in NOD2 associated with susceptibility to Crohn’s disease. disease. Gastroenterology 1988;95:1523–7.
Nature 2001;411:603– 6. 26. McCaffery TD, Nasr K, Lawrence AM, et al. Severe growth
3. Hugot J-P, Chamaillard M, Zouali H, et al. Association of retardation in children with inflammatory bowel disease. Pe-
NOD2 leucine-rich repeat variants with susceptibility to diatrics 1970;45:386 –93.
Crohn’s disease. Nature 2001;411:599 – 603. 27. Polk DB, Hattner JA, Kerner JA Jr. Improved growth and
4. Gent AE, Hellier MD, Grace RH, et al. Inflammatory bowel disease activity after intermittent administration of a defined
disease and domestic hygiene in infancy. Lancet 1994;343: formula diet in children with Crohn’s disease. JPN 1992;16:
766 –7. 499 –504.
5. Montgomery SM, Pounder RE, Wakefield AJ. Infant mortality 28. Baldassano RN, Vasiliauskas E, Braegger CP, et al. A multi-
and the incidence of inflammatory bowel disease. Lancet center study of infliximab (anti-TNF alpha antibody) in the
1997;349:472–3. treatment of children with active Crohn’s disease. Gastroen-
6. Duggan AE, Usmani I, Neal KR, et al. Appendicectomy, terology 1999;116:A665.
childhood hygiene, Helicobacter pylori status, and risk of 29. Kugathasan S, Werlin SL, Martinez A, et al. Prolonged dura-
inflammatory bowel disease: A case control study. Gut 1998; tion of response to infliximab in early but not late pediatric
43:494 – 8. Crohn’s disease. Am J Gastroenterol 2000;95:3189 –94.
2010 Mamula et al. AJG – Vol. 97, No. 8, 2002

30. Hyams JS, Markowitz J, Wyllie R. Use of infliximab in the laborative Study on Inflammatory Bowel Disease (EC-IBD).
treatment of Crohn’s disease in children and adolescents. Gut 1996;39:690 –7.
J Pediatr 2000;137:192– 6. 36. Quinton JF, Sendid B, Reumaux D, et al. Anti-Saccharomyces
31. D’Haens G, Van Deventer S, Van Hogezand R, et al. Endo- cerevisiae mannan antibodies combined with antineutrophil
scopic and histological healing with infliximab anti-tumor cytoplasmic autoantibodies in inflammatory bowel disease:
necrosis factor antibodies in Crohn’s disease: A European Prevalence and diagnostic role. Gut 1998;42:788 –91.
multicenter trial. Gastroenterology 1999;116:1029 –34. 37. Moum B, Ekbom A, Vatn MH, et al. Inflammatory bowel
32. Murch SH, Lamkin VA, Savage MO, et al. Serum concentra- disease: Re-evaluation of the diagnosis in a prospective pop-
tions of tumour necrosis factor alpha in childhood chronic ulation based study in south eastern Norway. Gut 1997;40:
inflammatory bowel disease. Gut 1991;32:913–7. 328 –32.
33. Krebs C, Nixon HH. Inflammatory bowel disease in child- 38. Ruemmele FM, Targan SR, Levy G, et al. Diagnostic accuracy
hood. Presentation and diagnosis. Z Kinderchir 1983;38:387– of serological assays in pediatric inflammatory bowel disease.
91. Gastroenterology 1998;115:822–9.
34. Barton JR, Ferguson A. Clinical features, morbidity and mor- 39. Davies G, Evans CM, Shand WS, et al. Surgery for Crohn’s
tality of Scottish children with inflammatory bowel disease. Q disease in childhood: Influence of site of disease and operative
J Med 1990;75:423–39. procedure on outcome. Br J Surg 1990;77:891– 4.
35. Shivananda S, Lennard-Jones J, Logan R, et al. Incidence of 40. Becker JM. Surgical therapy for ulcerative colitis and
inflammatory bowel disease across Europe: Is there a differ- Crohn’s disease. Gastroenterol Clin North Am 1999;28:
ence between north and south? Results of the European Col- 371–90, viii-ix.