Auvin, Stéphane*; Molinié, Florence†; Gower-Rousseau, Corinne†; Brazier, Franck¶; Merle, Véronique§; Grandbastien, Bruno§; Marti, Raymond‡; Lerebours, Eric∥; Dupas, Jean-Louis¶; Colombel, Jean-Frédéric‡; Salomez, Jean-Louis†; Cortot, Antoine‡; Turck, Dominique*
The epidemiology of inflammatory bowel disease (IBD) has been the subject of numerous studies in adults, whereas less information is available in children and adolescents. Geographic or temporal variations in the incidence of IBD may help identify environmental, genetic and/or immunologic causal factors. In adults, the incidence of IBD has increased in industrialized countries in the last century. Studies from Scotland and Wales have shown an increase in the incidence of pediatric Crohn disease (CD) in the last 20 to 30 years, whereas the incidence of pediatric ulcerative colitis remains stable or marginally decreased (1-9). Data remain scarce, however, and whether the incidence of CD is really increasing in children remains a matter of debate. There are no regional data available on the incidence of IBD in children in northern France, and no previous prospective longitudinal study has been performed in France. A registry of IBD (EPIMAD) started in northern France in 1988. The registry has allowed us test the hypothesis that there is an increased incidence over the last 15 years. The aim of this study was to gain further insight into the incidence and location of IBD in children and adolescents aged less than 17 years in northern France from 1988 to 1999.
Northern France has 5,790,526 inhabitants (1999 National Population Census). The study area is divided into four departments.
1. Nord, population 2,554,449 and population density 445/km2;
2. Pas-de-Calais, population 1,441,422 and population density 216/km2;
3. Somme, population 555,479 and population density 90/km2;
4. Seine-Maritime, population 1,239,176 and population density 197/km2.
The pediatric population under 17 years of age is: Nord = 593,837; Pas de Calais = 332,228; Somme = 115,969; Seine-Maritime = 270,107; total = 1,312,141. These areas include urban and rural populations (urban/rural ratios: 8.9 in Nord, 4.5 in Pas-de-Calais, 1.4 in Somme and 3.0 in Seine-Maritime). This region is a well-defined geographic entity bordering Belgium in the north and the English Channel in the west (Fig. 1). The population is stable. The percentage of population moving per year in each department was 0.8 for Nord, 0.8 for Pas-de-Calais, 1.1 for Somme and 0.9 for Seine-Maritime (according to the French National Statistical Institute).
This prospective study started on January 1, 1988. The study protocol is described in detail elsewhere (10). Only patients residing in the study region at the time of IBD diagnosis were included. The study was conducted by the gastroenterologists and pediatric gastroenterologists of the four departments. Before the study, all the gastroenterologists (n = 232, 160 in private practice and 72 in hospital practice) and pediatric gastroenterologists (n = 6) of the four departments were informed about the methods and aims of this work by letter and during a meeting. Each gastroenterologist and pediatric gastroenterologist was required to report any patient consulting for the first time with clinical symptoms compatible with IBD. Physicians were contacted by phone at least three times a year by an interviewer practitioner. The interviewer went to the gastroenterologist's office and collected the following data from patient charts in a standardized questionnaire for each new case of IBD from January 1, 1988 to December 31, 1999: age, sex, year of diagnosis, interval between onset of symptoms and diagnosis and clinical, radiologic, endoscopic and histologic findings at the time of diagnosis. Children and adolescents younger than 17 years of age were analyzed.
Two expert gastroenterologists from two different academic institutions assessed each case. Each expert gastroenterologist ascertained a diagnosis without knowing the diagnosis made by the other expert and made a final diagnosis of CD, ulcerative colitis (UC) or indeterminate colitis (IC). In case of discrepancy, discussion between the two experts led to the final diagnosis. The diagnosis of CD and UC was classified as definite, probable or possible according to the presence or absence of the criteria published by Gower-Rousseau et al. (10).
The diagnostic criteria used by the experts were as follows:
Definite CD: Epithelial granulomas in biopsy or surgical specimen.
1) Lesions in the large bowel without involvement of the small bowel and at least three of the following four diagnostic criteria:
a) history of diarrhea or abdominal pain for 6 weeks or longer;
b) radiologic and/or endoscopic findings with segmental lesions of cobblestone appearance and/or stenosis in the large bowel;
c) histologic findings compatible with CD;
d) fistulas and/or abscesses in relation to the intestinal lesions.
2) Lesions in the small bowel with or without involvement of the colon and, irrespective of the duration of clinical symptoms, the presence of at least two of the above four criteria.
1) Lesions in the large bowel without involvement of the small bowel and two of the above four diagnostic criteria.
2) Lesions in the small bowel with or without involvement of the colon and, irrespective of the duration of clinical symptoms, one of the above four criteria.
1) history of diarrhea and/or rectal bleeding for 6 weeks or longer and at least two of the following three criteria:
a) typical endoscopic findings including granular, friable mucosa and/or superficial ulcerations;
b) radiologic findings including ulcerations, narrowing and shortening of the colon;
c) typical histologic findings on surgical or biopsy specimens.
2) surgical specimens grossly typical of ulcerative colitis with typical histologic findings, irrespective of the duration of clinical symptoms.
1) history of diarrhea and/or rectal bleeding for 6 weeks or longer and one of the above three criteria.
2) history of 6 weeks or longer of digestive symptoms without diarrhea and rectal bleeding and two of the above three criteria.
3) Surgical specimens grossly typical of UC with no typical histologic findings.
Possible UC: history of more than 6 weeks but with no morphologic and histologic findings.
IC: history of chronic colitis compatible with both the diagnosis of CD or UC.
Ulcerative proctitis was defined according to the same criteria as UC but with macroscopic evidence that the sigmoid was normal above the rectum. Patients with a history of colitis shorter than 6 weeks were classified as acute colitis.
The location of disease was reported only for patients who underwent a complete bowel investigation (small and large bowel visualized for CD and large bowel visualized up to the cecum for UC). For CD, three locations were considered: ileocolonic, isolated colonic and isolated small bowel. For UC, three locations were considered: ulcerative proctitis, defined as involvement up to the rectosigmoid junction, arbitrarily estimated to be 20 cm from the anal margin; pancolitis, defined as an involvement up to the cecum; left-sided colitis including involvement above the rectosigmoid junction without involvement of the cecum.
Only patients with certain and probable CD or UC and patients with IC were included in the study. We did not include possible cases of IBD and acute colitis to be able to compare our data with those of most published reports.
Quality Assurance Procedures
The Steering Committee of the study included five academic gastroenterologists, two academic epidemiologists, one academic pediatric gastroenterologist, one private gastroenterologist and five interviewer practitioners. The committee established written instructions for the interviewer practitioners to ensure the questionnaire was correctly completed. It also set up a framework to help the expert gastroenterologists in their final diagnosis. The committee met every 3 months. The aims of these meetings were, by review of test cases, to verify that the protocol was accurately followed by all interviewer practitioners and to minimize inter-expert variations.
To validate the completeness of the case collection by the gastroenterologists, we undertook a 1-year validation study in the Somme Departement in 1989. All general practitioners (n = 498) and radiologists, histopathologists and surgical and pediatric specialists (n = 36) received a questionnaire that served to notify incident cases of IBD. If they did not reply, they were contacted by telephone. All specialists and 450 of 498 general practitioners (91%) replied. In total, 85 cases were notified by these practitioners. In contrast, the recording system by the gastroenterologists notified only 82 cases. The three missing cases were reported by two pathologists and one general practitioner (one possible UC, one possible CD and one probable ulcerative proctitis). These patients had consulted a gastroenterologist who had forgotten to declare their cases to the interviewer practitioner. Thus, the completeness of our case collection was 96.5%.
Age-adjusted incidence rates of CD, UC and IC were calculated by combining definite as well as probable cases. Cases of UC and ulcerative proctitis were pooled. To identify changes in the incidence of IBD, we divided the 12-year study into four equal 3-year periods (1988-1990, 1991-1993, 1994-1996 and 1997-1999). The mean annual incidence rates were calculated for each 3-year period and for the entire study period. For each department, population data by age and sex were obtained for each 3-year period from linear interpolation of the 1982, 1990 and 1999 National Population Census. Age direct standardization was done using the French population weights of 1990 for each age group (0 to 4 years: 26%; 5 to 9 years: 31%; 10 to 14 years: 31%; 15 to 16 years: 12%) (11). Confidence intervals for the age-adjusted rates were calculated according to the method proposed by Breslow and Day (12).
Age at diagnosis (years) and time elapsed between symptoms and diagnosis (months) were calculated as medians with interquartile range. Group comparisons were performed with the Wilcoxon rank sum test or the Kruskal-Wallis test. The distributions of categorical data were compared with the χ2 or Fisher exact test. Tests for trend were performed with a χ2 for qualitative data. The change in incidence rates by age-at-diagnosis and calendar year of diagnosis (3-year intervals) were estimated with a Poisson regression model adjusted for age and sex. Data were analyzed with SAS software version 8.2 (SAS, Chicago, IL) (13).
From 1988 to 1999, 7066 patients presented with IBD: 4013 CD (57%), 2665 UC (38%) and 388 IC (5%). There were 509 cases in children younger than 17 years of age (7.2% of total patients) divided as follows: CD (n = 367; definite CD = 156 and probable CD = 211), UC (n = 122; definite UC = 93 and probable UC = 29), of whom ulcerative proctitis (n = 30; definite ulcerative proctitis = 21 and probable ulcerative proctitis n = 9) and IC (n = 20). The pediatric cases of CD and UC represented 9.2% and 4.6% of all CD and UC patients, respectively. Among pediatric IBD patients, the percentages of CD, UC and IC were 72%, 24% and 4%, respectively. The pediatric standardized mean annual incidence was 3.1/105 (95% confidence interval, 2.9-3.4) for IBD as a whole, and the incidence of CD, UC and IC was 2.3/105 (2.0-2.5), 0.8/105 (0.6-0.9) and 0.12/105 (0.07-0.18), respectively. The mean incidence of IBD according to age is shown in Table 1.
Clinical Presentation at Diagnosis
A family history of IBD was found in 12% of children with IBD. The prevalence of positive family history of IBD was 13.1% for CD and 8.2% for UC. The sex ratio (male:female) for CD was 1.1 and 0.7 for UC. The median age at diagnosis was 14 years for IBD as a whole, and age ranges were as follows: CD, 11-16 years; UC, 11-15 years; and IC, 10-16 years (not significant). The median interval from onset of symptoms to diagnosis was 3 (1-7) months for IBD. This interval was significantly higher in CD, 4 (2-9) months, than in UC, 2 (1-4) months (P < 0.001).
The location of CD was studied only in patients in whom both small and large bowel had been investigated at diagnosis. The initial locations of CD were: 19% in small bowel, 10% in colon and 71% in both small bowel and colon (Fig. 2). A granuloma was found in 42.5% of children. The location of UC was studied only in patients who underwent total colonic exploration at diagnosis. The initial locations of UC were: 11% proctitis, 57% left colon and 32% pancolitis (Fig. 3).
Evolution of CD and UC Throughout the Study Period
There was an increase in the incidence of CD from 2.1 to 2.6 (+20% between 1988 to 1990 and 1997 to 1999) that did not reach statistical significance (P = 0.2) (Fig. 4). The UC incidence remained stable throughout the study period (P = 0.8). There was no change of location and type of investigations performed for CD and UC throughout the 12-year study. When the entire population with CD (n = 4013) was considered, the incidence of CD significantly increased from 5.2 to 6.4/105 (+23% between 1988-1990 and 1997-1999) (P < 0.0001). When the entire population with UC (n = 2665) was considered, the incidence of UC significantly decreased from 4.2 to 3.5/105 (−17% between 1988 to 1990 and 1997 to 1999) (P < 0.0001).
Our study shows that the incidence of pediatric CD in Northern France is comparable to previous reports from Brittany (1), Wales (2,3), Scotland (4), Great Brittain (5), Norway (6) and Sweden (7,8), where the incidence ranged from 1.2/105 to 4.9/105 (Fig. 5). Only the study from Denmark reported a much lower incidence of CD of 0.7/105 (9).
Although differing slightly among studies, the diagnostic criteria were very similar to ours, taking into account definite and probable CD and UC. The upper age limit at enrollment ranged from 14 years (9) to 15 years (2,5,6-8), 16 years (4,14,15) and 17 years in the present study and that of Brittany (1). Most studies used hospital-discharge-linked databases. It cannot be excluded that patients with mild or moderate CD or UC might not have been registered in a discharge diagnosis database. Few prospective studies have been published (1,5-8). A registration system was used only in our study and in the studies from Brittany and Great Brittain, with the collection of data made through interviewer practitioners at the gastroenterologist consulting room in our study and that of Brittany, and through a postal questionnaire in the study from Great Brittain (5).
In contrast to CD, the incidence of UC in our study was much lower than that reported in previous studies (Fig. 5), with the exceptions of Brittany (0.57/105) and Wales (0.71-0.75/105); the incidence of UC ranged from 1.56/105 in Scotland to 4.3/105 in Norway. It has been recently shown that the protective effect of appendectomy against UC only occurs when surgery is performed before the age of 20 years (16). The mean rate of appendectomy among French children and adolescents is known to be fourfold higher than that in the other countries of the European Union (17). This high rate of appendectomy among children and adolescents may play a role in the lower incidence of UC in Northern France and Brittany.
The lower incidence of UC compared with CD was observed in most pediatric studies, the CD/UC ratios being 0.1 in Denmark, 1.46 in Scotland, 1.82 in Wales, 2.3 in Stockholm-Sweden, 2.8 in Brittany, 3.09 in our study and 3.11 in Southwest Wales (1983-1993). In most epidemiologic studies of adult patients, the CD/UC ratio is less than 1.
Most reviews and texts indicate that the proportion of childhood cases among all IBD cases is 20% to 30% (18). Our study shows that the percentage of pediatric cases among IBD patients is lower than previously thought. To our knowledge, the only study giving similar information is that of Langholz et al. (9) in Denmark, in which IBD started before the age of 15 years in 6% of CD patients and 7% of UC patients. In our population, 91% of CD patients and 86% of UC patients were older than 10 years, showing that IBD mainly involves older children and adolescents.
It is known that the incidence of CD is higher among females than among males. The predominance of boys in pediatric CD observed in the present study has been reported in the Canadian province of Manitoba and in Great Brittain, Norway, Sweden and Scotland (5-8,15,16,19). This suggests that hormonal changes related to age could be involved in this phenomenon. However, other etiologic factors should be considered.
The question of an increased incidence of CD and UC in children over the last 10 to 20 years remains controversial. While the incidence of UC remained stable in our study, the incidence of CD increased from 2.1 in 1988-1990 to 2.6 in 1997-1999 (+20%), but this increase was not significant. A similar trend has been reported in Wales (1983-1993) (2), Scotland (1981-1992) (14) and Great Brittain (5). Only a Swedish study (1990-2001) found a significant increase of CD incidence that increased from 1.7 in 1990-1992 to 8.4 in 1999-2001 (P < 0.0001 for trend 1990-2001) (8). In our study, the absence of statistical significance is probably attributable to the lack of power related to the small population of CD pediatric patients. Indeed, the incidence of CD in the entire population (children and adults) of Northern France significantly increased (+23%) from 1988-1990 to 1997-1999, whereas the incidence of UC in the entire population significantly decreased (−17%) (20). Although not of statistical significance, the increased CD incidence in children from 1988-1990 to 1997-1999 suggests that factors influencing the frequency of the disease are still at work in this part of Europe. Tobacco consumption is a well-known factor influencing the occurrence of CD (21). However, tobacco consumption has not increased in French adolescents over the last 15 years (22). Case-control studies might give further insight into the possible factors involved in this increased incidence.
The authors thank the interviewer practitioners who collected the data: B. Lemaire, N. Guillon, M. Inglard, I. Rousseau, N. Wauquier, P. Fosse, M. Lecomte, D. Panis, L. Yzet, C. Dias, S. Richon. EPIMAD is organized under an agreement between the Institut National de la Santé et de la Recherche Médicale (INSERM), the Direction Générale de la Santé (DGS) and the Institut National de Veille Sanitaire (InVS), and is also supported by the François Aupetit Association, the Ferring Laboratories, IRMAD (ASTRA Company), the Observatoire Régional de la Santé du Nord-Pas-de-Calais, the Caisse Régionale d'Assurance Maladie de Nord Picardie and the Lille University Hospital.
The authors thank all the gastroenterologists and all specialists who participated to this study: J. M. André, M. Antonietti, A. Armand, I. Aroichane, J. P. Aubet, E. Auxenfants, B. Barbry, N. Bardoux, P. Baron, A. Baudet, B. Bazin, A. Bebahani, J. P. Becqwort, V. Benet, C. Benguigui, E. Ben Soussan, A. Bental, I. Berkelmans, J. Bernet, C. Bernou-Dron, P. Bertot, N. Biron, M. Bleuet, F. Blondel, F. Bohon, E. Boniface, P. Bonnière, E. Bonvarlet, P. Bonvarlet, A. Boruchowicz, R. Bostvironnois, B. Bouche, C. Boudaillez, C. Bourgeaux, A. Bourguet, A. Bourienne, G. Bray, F. Brazier, P. Breban, V. Brung-Lefebvre, P. Burgiere, J. Butel, J. Y. Canva, V. Canva-Delcambre, J. P. Capron, F. Cardot, P. Carpentier, E. Cartier, J. F. Cassar, J. F. Castex, L. Catteau, B. Caujolle, G. Cayron, M. Chantre, J. Charles, T. Charneau, J. F. Claerbout, P. Y. Clergue, G. Cohen, R. Collet, J. F. Crinquette, I. Dadamessi, V. Dapvril, T. Davion, J. Debas, F. Dehont, C. Delatre, R. Delcenserie, O. Delette, T. Delgrange, L. Delhoustal, J. S. Delmotte, G. Deregnaucourt, P. Descombes, J. P. Desechalliers, P. Desmet, G. Desseaux, P. Desurmont, A. Devienne, M. Devred, A. Devroux, A. Dewailly, R. Dubois, P. Ducatillon, J. Duclay, B. Ducrocq, F. Ducrot, P. Ducrotte, J. Dufilho, C. Duhamel, D. Dujardin, F. Dupont, Y. Duranton, M. C. Elie-Legrend, J. P. Evrard, B. Filoche, L. Finet, D. Foissey, M. C. Foutrein-Comes, P. Foutrein, T. Frere, P. Gallet, C. Gamblin, G. Geslin, Y. Gheyssens, T. Gilbert, P. Godard, J. M. Godchaux, R. Godchaux, O. Goria, F. Gottrand, P. Gower, B. Grandmaison, C. Guedon, J. F. Guillard, L. Guillem, F. Guillemot, D. Hanon, P. Heckestweiller, J. P. Hedde, H. Hellal, B. Heyman, M. Heraud, S. Herve, P. Hochain, P. Houcke, A. Ivanovic, E. Janicki, J. Jeu, C. Jonas, A. Kerleveo, A. Kiriakos, J. Kiriakos, O. Klein, R. Kornhauser, D. Koutsomanis, J. E. Laberenne, G. Laffineur, P. Lannoy, J. Lapchin, M. Laprand, D. Laude, P. Lecieux, N. Leclerc, C. Le Couteulx, J. Ledent, J. Lefebvre, A. Le Grix, P. Lelong, C. Lenaerts, A. Leplat, H. Leroi, M. Y. Leroy, J. P. Lesage, X. Lesage, J. Lesage, I. Lescanne-Darchis, J. Lescut, D. Lescut, B. Leurent, M. Lhermie, A. Lion, B. Lisambert, F. Loire, M. Luciani, D. Lucidarme, J. J. Lugand, O. Macaigne, D. Maetz, D. Maillard, H. Mancheron, R. Marti, F. Martin, G. Martin, E. Marzloff, C. Mathieu-Chandelier, J. Mauillon, V. Maunoury, J. L. Maupas, B. Mesnard, P. Metayer, B. Meurisse, F. Meurisse, L. Michaud, X. Mirmaran, T. Modaine, L. Morel, E. Moulin, O. Mouterde, J. Mudry, R. N'GuyenTack, B. Notteghem, A. Ostyn, D. Ouvry, N. Panien-Claudot, C. Paoletti, A. Papazian, B. Parent, J. C. Paris, P. Patrier, L. Paupart, B. Pauwels, M. Pauwels, R. Petit, M. Piat, S. Piotte, C. Plane, B. Plouvier, E. Pollet, P. Pommelet, G. Pouchain, P. Prades, A. Prevost, J. C. Prevost, A. M. Queuniet, J. F. Quinton, A. Rabache, P. Rabelle, V. Razemon, N. Reix, C. Richez, P. Robinson, J. Rodriguez, J. Roger, J. M. Roux, A. Rudelli, G. Savoye, P. Schlosseberg, M. Segrestin, A. Seryer, F. Sevenet, J. Silvie, V. Simon, C. Spyckerelle, N. Talbodec, J. L. Thelu, J. M. Thorel, J. Toisin, J. Tonnel, J. Y. Touchais, J. L. Tranvouez, C. Triplet, E. Vaillant, C. Valmage, D. Vanco, E. Vanderbecq, P. Vandermolen, P. Vandevenne, C. Vandewalle, J. P. Vanhoove, G. Verbiese, P. Vermelle, C. Verne, M. Vincendet, J. Viot, Y. M. Voiment, L. Waeghemaecker, J. Y. Wallez, M. Wantiez, J. Weber, J. L. Willocquet, E. Wolschies, A. Zellweger, C. Ziade.
© 2005 Lippincott Williams & Wilkins, Inc.