Inflammatory bowel disease in children and adolescents: Working Group Report of the First World Congress of Pediatric Gastroenterology, Hepatology, and Nutrition.

Abstract

I. SUMMARY OF THE PROBLEM Inflammatory bowel disease (IBD) encompasses two related but distinct disorders of as yet unknown cause. Ulcerative colitis (UC) is a chronic, idiopathic, diffuse inflammation of the rectum that frequently may extend continuously over variable lengths of more proximal colon. The inflammation is usually restricted to the mucosa and submucosa. Crohn's disease (CD) is a chronic, idiopathic, transmural inflammation that can affect one or several segments of the digestive tract. The predominant sites of disease are the terminal ileum, colon, and perianal region. The inflammation is often focal with involved areas adjacent to normal areas (skip lesions). In most cases, these two conditions can be distinguished after clinical, endoscopic, histopathologic, and radiologic examinations. The term indeterminate colitis is reserved for cases of colitis in which findings are not sufficient to allow differentiation between Crohn's colitis and UC. Whether these entities represent different disease processes with related phenotypic expressions or a single disease entity with variable phenotypic expression remains to be determined. The first description of UC dates from 1859. The landmark paper by Crohn et al describing a series of patients including a 17-year-old boy dates to 1932, but patients with chronic ileitis had been previously described. The symptoms of IBD are protean, but commonly include intestinal symptoms such as abdominal pain, bloody diarrhea, poor appetite, and weight loss, and at times a wide range of extraintestinal manifestations including fever, arthritis, skin rashes, renal calculi, and hepatobiliary disease, all of which are not significantly different from adults. There are, however, distinct differences in pediatric IBD compared with adult disease, particularly in its impact on growth and its potential to delay puberty. Current opinion regarding the cause of IBD favors the hypothesis that IBD results from an interaction between immunologic, genetic, and environmental factors. A natural model for UC has been described in the cotton top tamarin. Although no natural animal CD model exists, a large number of biologic (e.g., peptidoglycan polysaccharide), chemically induced (e.g., TNBS), and genetically derived (e.g., IL-I0 knockout mice) animal models of CD have been developed. These models offer insights into the immunologic basis of IBD, which in turn have led to advances in the study of human disease. It appears that it some cases CD and UC can be differentiated immunologically on the basis of serologic findings (e.g., pANCA in UC, ASCA in CD) and cytokine profiles. CD typically is characterized by a T lymphocyte helper 1 (Thl) cytokine pattern, in which ILl, IL2, TNFα and interferon-y are predominant. UC tends to be characterized by more of a T lymphocyte helper 2 (Th2) pattern. Although these models have advanced our understanding of the immunology of IBD, the cause of the human disease remains elusive. The role of pathogens as inducers of disease, such as mycobacteria or measles, remains controversial. Genetic factors play an important role, with about 15% of affected individuals having first-degree relatives with disease. However, IBD appears to be a complex, polygenic disorder that cannot be explained by a simple Mendelian model. Current therapeutic goals in children and adolescents seek to induce and maintain clinical and histologic remission and allow children to achieve their full potential for growth, social interaction, and educational attainment. Treatment modalities and clinical responses are similar to those seen in adult patients, although there are important differences in the treatment of pediatric patients. In particular, therapies for children must be designed to promote rather than interfere with growth, sexual development, and bone mineral accretion. These therapeutic requirements have resulted in a number of differences in the treatment of children and adolescents compared with adults, such as the increased use of nutritional interventions and the earlier use of immunomodulators. II. MAJOR ISSUES IN NEED OF INVESTIGATION OR IMPLEMENTATION Epidemiology The relatively few epidemiologic studies of IBD in the pediatric and adolescent age-groups have largely focused on the incidence, prevalence, and clinical features of IBD and how these features differ based on geographic location, demographic influences, ethnicity, and over time (1). The reported incidence of pediatric and adolescent IBD varies from 2.2 to 6.8 per 100,000, but there are significant differences between continents, within countries, and in particular areas over time (2,3). Although explanations for the variations in disease incidence have not been fully determined, a number of environmental factors such as hygiene, diet, breast feeding, smoking, the contraceptive pill, and infectious agents such as measles, mumps, Epstein-Barr virus, and Mycobacterium paratuberculosis have been explored (3,4). In adult patients, cigarette smoking is the factor most consistently demonstrated to be associated with the development of IBD. Smoking appears to predispose to CD and protect against the development of UC. By contrast, ex-smokers appear to have an enhanced risk of developing UC. The only pediatric study to examine the influence of passive smoking was a case control study, where passive smoking exposure at birth, and to a lesser extent at diagnosis, was associated with an increased risk of developing both forms of IBD (CD more than UC)(2). Interestingly, in contrast to the adult literature investigating personal smoking history, passive smoking was not protective against the development of UC. Other controversial data suggest that the early measles infection may be important in the development of a granulomatous vasculitis and CD. Epidemiologic studies require population-based data. Focusing studies on children potentially has great advantages compared with studying adults, as many confounding variables (such as smoking and the use of birth control pills) are much less prevalent in the pediatric population. The following represents the major epidemiologic issues that need to be addressed: Population-based databases from all parts of the world should be developed to determine and track IBD incidence and prevalence rates, disease behavior, and effect of therapy. Identifying areas of increased or rapidly changing incidence should allow more focused investigation of possible environmental and genetic factors. Prospective, standardized, and regularly updated databases should include health histories of first-degree and more distant relatives focusing specifically on whether any also have IBD. Complete descriptions of IBD phenotype in these relatives, as well as data on associated illnesses such as autoimmune disease and malignancy, should be compiled. Databases should be designed to identify potential risk factors for developing IBD and its complications. They should also seek to identify which factors influence the course of IBD. The data collected could be used to develop large case-controlled studies. Early immunologic influences such as infectious history, tonsillectomy, appendectomy, antibiotic usage, vaccination, diet, and atopy need to be examined further. As infectious agent(s) may be directly pathogenic, or may trigger an immunologically mediated and perpetuated response, there is a need for population-based studies to determine the role of suspected microorganisms in the pathogenesis of IBD. The organism(s) may be ubiquitous or even a part of the normal flora. Studies evaluating differences in the bowel flora between IBD and controls are needed. Similarly, evaluation of the functional aspects of the bowel flora, within IBD patient groups and clinical subgroups, should be undertaken. Such studies should include examination of the epidemiology of Mycobacterium paratuberculosis and measles and its association with CD. Genetics The importance of genetic susceptibility to IBD is evident from twin and family studies (5,6). Among 80 Swedish twin pairs, the concordance rate for CD was 58.3% in monozygotic twin pairs but only 6.3% for UC. In dizygotic twin pairs, most were discordant for IBD (96% for CD and 100% for UC). The increase in IBD prevalence among siblings is 10- to 15-fold compared with the general population. This sibling risk is much higher than is seen in other complex genetic illnesses such as diabetes mellitus, schizophrenia, and asthma. The lifetime risk for developing IBD for first-degree relatives of IBD patients is estimated at 7.8% in patients of Jewish ancestry and 5.25% in non-Jews. For CD, the risk is 16.8% for siblings of CD patients in Jewish families. Particularly high risks of susceptibility occur in children when both parents have IBD. It is increasingly evident that IBD represents one or more complex genetic diseases. This has led to an increasingly sophisticated search for the genes responsible for the development and evolution of both CD and UC. Genetic Anticipation The question of whether true genetic anticipation is present in IBD needs to be more clearly elucidated, as current studies are often in conflict. In one study of two-generation pedigrees with CD, affected individuals from the second generation were often younger at diagnosis than were their parents. However, once it is recognized that IBD is “in the family,” children in the next generation may be rapidly brought for consultation with the express desire to “rule out IBD.” This “first ascertainment bias” can influence the age at which CD is diagnosed, resulting in the appearance of genetic anticipation when there is none. Recent European studies provide evidence that this may be the case (7). Gene Linkage Studies Initially, investigations involving genes and IBD focused on the major histocompatibility complex (MHC) because of its role in immune function. In California, associations were found between HLA-DR2 and UC and the combination ofDR1 and Dqw5 and CD. However, the association of HLA-DR2 and UC was not confirmed in Pittsburgh. In England, linkage was found between DRB 1 (but not DR2) and UC but no association with CD was evident. By contrast, in Japan, an HLA-linked Crohn's disease susceptibility gene was associated primarily with DQB 1 *04. These studies illustrate the important ethnic differences that can influence genetic studies. Subsequently, gene associations have been sought through linkage studies and genome-wide searches. This method led to the identification of the CD susceptibility locus, IBD 1, on chromosome 16 and to a second putative locus, IBD2, associated with both CD and UC on chromosome 12 (8,9). However, the association with chromosome 12 was not confirmed in the United States (using genome-wide screening) or in Canada (using selected microsatellite markers). Additional loci have been suggested on chromosomes 1, 3, 4, and 7 as well as a link between the X chromosome and UC. However, these associations are not universally confirmed when new populations are studied, demonstrating the polygenic nature of IBD and the importance of ethnic differences. These studies provide evidence that non-MHC genes are important susceptibility genes for IBD that need to be identified and their gene products and functions characterized. Candidate Gene Studies In contrast to genome-wide searches, directed studies of non-HLA genes important in the function of the immune system have been initiated. In particular, cytokine gene polymorphisms are being investigated. A specific TNFα allele (TNFα2b1c2d4e1) has been associated with DRl/Dqw5 and appears to represent a risk factor for CD. Others have assessed the role of different allelic frequencies of TNFα, interleukin-l receptor antagonist (IL-1RA) gene, an allele of the C3 gene and the B 1 bradykinin receptor gene promoter (8). Other genes important in the development and function of the immune system such as the CD11 integrin cluster, complement receptor type 3, B-lymphocyte marker CD19, adhesion molecule sialophorin, and the interleukin 4 receptor, all of which are located on chromosome 16 (the site of the putative IBDI susceptibility gene), may be important (9). The study of candidate genes represents a fertile area for future research. The interaction between these genetic alterations and potential environmental triggers may one day help explain why our patients are susceptible to developing IBD. Family Studies and Ethnic Variation Several studies demonstrate the utility of using multiply affected pedigrees for gene linkage studies to detect potential IBD susceptibility genes. Including patients with onset during childhood and adolescence may provide an enhanced opportunity for gene linkage associations (6,7). Lod scores are higher in pedigrees when an affected member is <20 years versus >30 years. Differences in the frequency of different genetic alleles between countries and within diverse ethnic groups within a country must also be investigated. Recognizing that certain ethnic groups (such as American Ashkenazic Jews) have a particularly high prevalence of IBD underscores the importance of identifying subgroups of patients. IBD susceptibility genes show different strengths of association depending on subject ethnicity: genes on chromosome lp and 3q are strongly associated with IBD in non-Ashkenazic Caucasians, whereas linkage on 4q and 16 is more commonly linked with Ashkenazim. These particular subgroup associations have not, however, always been evident in other studies. Some of the discrepancies among studies may be explained by ethnic variability in HLA class II allelic frequencies. DRB1*1502 accounts for 55% to 57% of DR2 in Jewish and Japanese populations but <5% in white non-Jewish patients studied in Pittsburgh and Oxford, in whom DRBI* 1501 accounts for 96% of DR2. International cooperative studies are needed to clarify concordance and discordance between countries and ethnic groups within one country. Genotype/Phenotype Correlation Several studies document strong concordance within families for site and type (e.g., inflammatory or stenosing/fistulizing) of CD. Families with more than two affected members show enhanced concordance for disease site and type compared with those with only two persons. Evaluating candidate genes in multiply affected families should prove a particularly powerful method of investigation. These studies require precise clinical characterization of patients. Findings should be compared with large populations of patients with sporadic disease. Phenotypic Expressions of Disease These have also been linked with genetic markers. For instance, the association of HLA DR B 1*0103 and UC is greatest among patients with extraintestinal manifestations such as uveitis, arthritis, mouth ulcers, and especially, pauciarticular arthropathy. Further studies are needed to determine which susceptibility genes, HLA types or modifying genes, alone or in combination, are responsible for the varied manifestations of CD and UC. Correlation of Genotype with Natural History and Response to Therapeutic Intervention Once phenotype/genotype associations are identified, the influence of genetic differences on response to therapy can be evaluated. Linkage between the DR3 DQ2 haplotype and extensive UC has been reported. Although this did not coincide with the need for surgery, it was suggested that MHC genes may influence the course of UC. In Japan, an HLA subtype, DQAI*0102, has been linked to resistant CD, whereas in the United States, a specific TNFα microsatellite haplotype appears to correlate with the need for colectomy in UC (5,9). Preliminary data suggests that specific immunologic profiles may be associated with clinical response to anti-TNF antibody therapy in CD. If alterations in cytokine gene polymorphisms are consistently associated with specific forms of IBD, targeted therapy may become a valuable asset in patient management. Thus, future investigations may need to analyze patients' responses to various therapies, taking into consideration their genetic profiles. Newer Diagnostic Modalities and Prognostic Indicators The precise differentiation of CD and UC is important to children and adolescents facing a lifetime of illness, as clinical course, risk of complication, and response to both medical and surgical therapy depends on accurate diagnosis. The traditional differentiation of IBD from other illnesses of the gastrointestinal tract, and of CD from UC, has depended on excluding potentially confounding illnesses (e.g., no evidence of enteric pathogens) while demonstrating characteristic (although not diagnostically definitive) clinical, radiographic, endoscopic, and histologic abnormalities. It is now well accepted that any child suspected of IBD should undergo a full colonoscopy with biopsies from terminal ileum (when possible) and colon. In the case of possible CD, this should be followed by upper endoscopy with biopsy and small bowel radiography. Unfortunately, such a diagnostic approach only defines clinical phenotypes, which often can overlap. In particular, differentiating ulcerative from Crohn's colitis is at times impossible without a surgical specimen, and even then indeterminate forms of IBD can defy classical definitions. Other diagnostic modalities such as ultrasonography, computerized tomography, magnetic resonance imaging, and tagged leukocyte scan have not proved to be diagnostically more accurate than the endoscopic and radiographic modalities more commonly employed for diagnosis. As a consequence, future studies must focus on the following: Immunogenetic Disease Profiles As a clearer understanding of the immunologic and genetic characteristics of IBD evolves, the definitions of the disease processes themselves will likely need to change. Serologic markers such as anti-neutrophil cytoplasmic antibody (pANCA) and antisaccaromyces cerevisiae antibody (ASCA) and genetic markers such as tumor necrosis factor (TNF) gene microsatellites and specific HLA class II alleles have the potential to define new subgroups of illness (10,11,12). Additional markers need to be identified to better characterize each individual's form of IBD. Just as the cell surface markers of a leukemia or lymphoma may predict the prognosis and response of a particular tumor, identification of the genetic and immunologic “profile” of a particular patient may allow a clearer definition of the illness and an improved understanding of how it can best be treated. Disease Activity Scores Improved characterization of a child's illness also requires developing better measures of disease activity for use both in research and in clinical practice. Multicenter studies are increasingly important given the large numbers of subjects required for adequate statistical power. These studies demand objective disease activity scores with minimal interobserver variability. Age and disease-appropriate activity scores which minimize the use of subjective data need to be evaluated on a worldwide basis, and if possible standardized. The Pediatric Crohn's Disease Activity Index (PCDAI) has been shown to be a reliable and reproducible outcome measure, but its utility outside of North America needs to be assessed. A similar pediatric scoring system for UC does not exist, and its development should be considered a high priority. Disease-specific quality of life (QOL) measures also need further development and validation as outcome measures in pediatric therapeutic trials (13). The utility of objective QOL measures in clinical practice also need to be assessed. Prognostic Indicators Factors determining risk of complications also need to be evaluated. The ability to accurately identify individuals at high risk for specific extraintestinal complications (e.g., hepatobiliary disease, severe perianal disease) or gastrointestinal malignancy are needed (14,15). Risk factors for colon cancer (duration of colitis, extent of disease, concomitant sclerosing cholangitis) are well established, but other potentially additive risk factors (such as genetic proclivity to cancer from mutations in mismatch repair or tumor suppressor genes) need to be evaluated (16). Identification of these factors should facilitate more focused trials of potentially preventative treatments. Such studies are particularly relevant to children and adolescents with IBD, as epidemiologic studies have already demonstrated that children's lifetime burden of disease and tendency for more fulminant illness put them at particular risk of many of these complications. Therapy Important progress has been made in recent years in the treatment options for IBD, although there is still no curative therapy (15). However, with rare exceptions, pediatric therapeutic trials are limited, and treatments for children are commonly extrapolated from studies in adults. Before initiating therapy, it is essential to establish the diagnosis, location, extent, and severity of disease. Subsequent therapeutic interventions often require an individualized approach combining nutritional, medical, and surgical considerations. Several approaches need to be discussed. Enteral Nutrition A major distinction from adult IBD is the efficacy of enteral nutrition (15). This response is poorly understood, but may include alteration of gut flora, enterocyte nutrition, removal of dietary antigen, or the presence of growth factors in the nutrition itself or the modulation of endogenous growth factors. There is evidence for efficacy in Crohn's colitis, but not for UC. Enteral therapies result in comparable remission rates compared with steroids, and accumulation of lean body mass is higher and linear growth better-maintained than with steroids. Chronic supplementary enteral therapy also reduces relapse rates. Throughout the world there is great variability in the use of enteral nutrition, with its use as first-line treatment in some centers and as last resort in others. Trials are still needed to establish the optimal role of nutritional therapy in children and adolescents with IBD. These studies should include optimal feed types, the use of immunomodulating diets and dietary antioxidant therapy and the possible role of food intolerances in some selected patients. Anti-inflammatory Therapy Most children and adolescents with IBD are at some stage treated with 5-aminosalicylate (5-ASA) derivatives, although there have been few formal studies of their use in children. The choice of an individual 5-ASA derivative is often empirical, and there are no convincing pediatric data to suggest that any one form is superior. The lack of a liquid preparation for younger children is a problem, however, and the efficacy of removing mesal amine from capsules or dissolving tablets in extemporaneously prepared suspensions has not been evaluated. The wide international experience of 5-ASA use has, however, provided ample clinical experience and it is doubtful whether even a major collaborative study of these drugs would provide significant advance in overall management. In addition, there is recent evidence to suggest that the long-term risk of colorectal cancer may be reduced by 5-ASA treatment. An optimal and maximal dose in children and adolescents needs to be established. Studies of other potential anti-inflammatory agents, such as n-3 fatty acids in fish oils, need to be evaluated. Immunomodulatory Therapy Despite widespread use, there remains uncertainty about the exact role of corticosteroids in treating children with IBD because of their adverse metabolic effects and because clinical well-being is often not associated with mucosal healing. For many clinicians, however, corticosteroids remain the cornerstone of treatment, especially in moderate to severe disease. Dosing regimens remain empiric, and while clinical practice has demonstrated reduced toxicity with alternate-day treatment regimens, large studies in children are lacking. The relative efficacy, safety and side effects of newer derivatives such as budesonide should be established in randomized double-blind placebo-controlled studies (17). Because of concerns regarding prolonged steroid use, in particular the deleterious effect on bone mineralization and growth, there is a trend toward the earlier and more widespread use of immunosuppressive agents such as azathioprine and 6-mercaptopurine to maintain remission in children and adolescents with IBD (18). This changing pattern of medication use has occurred before the completion of pediatric trials, and there remains a lack of consensus about monitoring requirements. Other immunosuppressants including cyclosporine, tacrolimus, and methotrexate have been evaluated only in small open-label studies. The use of the anti-TNF monoclonal antibody infliximab appears safe and effective in children and adolescents, at least in the short term, but there remain concerns about late adverse effects (19). The further use of this monoclonal antibody and other new biologic agents such as anti-inflammatory cytokines, anti-sense molecules, and receptor antagonists will require close monitoring. Safety and efficacy should be determined in pediatric clinical trials using multidose regimens in adequate numbers of patients. Antibiotic and Probiotic Therapy These forms of therapy appear attractive in pediatric practice because of the relatively low risk of adverse effects, and have gained scientific credibility from studies of the role of the enteric flora in disease pathogenesis. Controlled studies are lacking, however, and the composition of the enteric flora along the gastrointestinal tract in IBD needs to be clarified. Similarly, the role of immune responses to the endogenous enteric flora is also unknown. These issues require future investigation. Other Medical Therapies Many children are given remedies for IBD that have not been prescribed by traditional physician. These may include herbal products such as aloe vera or evening primrose oil, or alternative therapeutic approaches such as homeopathy or cranial osteopathy. This area remains largely unstudied, but has been attracting increased attention recently. As appropriate studies have not yet been performed, safety as well as efficacy in children with IBD remains to be determined. Large-scale controlled trials need to be encouraged using the correct design and methodology. Surgical Therapy Despite advances in the medica] treatment of IBD, there remain distinct indications for surgical resection in both CD and UC. Appropriate timing of surgery and choice of procedure may allow growth acceleration and pubertal advance that cannot yet be achieved medically. Delay in colectomy for toxic megacolon remains a potential cause of fatality. Optimizing surgical outcome, minimizing the need for ileostomy and preventing disease recurrence in CD or chronic pouchitis in UC remain areas for future pediatric research. III. PROPOSED PLAN TO ACHIEVE GOALS Epidemiology An international forum should be installed to design a worldwide database or, more realistically, a mechanism for free interaction and dialogue between national databases for prospective, longitudinal, population-based epidemiologic studies. Well-prepared hypotheses should be put forward regarding epidemiologic aspects of IBD in children. The forum must define the minimum set of features, such as disease characteristics, genetics, complications, and effect of therapy, that need to be included in any proposed database. Issues regarding anonymity of records and other privacy concerns need to be addressed. This forum should consider how technology such as the Internet might be used to facilitate collaborative epidemiologic studies worldwide. The funding for this very important, ongoing worldwide database should come from government agencies in combination with industry. In addition, the participants should identify mechanisms by which the long-term maintenance of such a database can be assured, and determine rules for researcher participation and access to these databases. Cooperative efforts with existing and future epidemiologic studies in adult populations need to be planned as well. Genetics Investigations of candidate gene polymorphisms and genome-wide searches have already been fruitful in identifying several potential IBD susceptibility and disease-modifying genes. New candidate genes will be found as research progresses. In the future, the functions and products of these genes will have to be investigated, toward the goal of providing targeted therapy and possible cure for IBD. Large numbers of sporadic IBD cases as well as non-IBD controls are required to confirm the importance and specificity of each gene locus initially identified by linkage studies. Given the relatively limited number of children available for study at any individual investigative center, the successful promotion of such studies will require a worldwide collaborative effort. The potential importance of specific gene linkages can only be assessed through the study of large numbers of patients from different ethnic and geographic backgrounds. It will only be through the ongoing collaboration of centers committed to unraveling the genetics of IBD that rapid advances will take place. Toward that end, an international working group of clinicians and geneticists involved in IBD investigations needs to be established. This group should establish a central library of clinical material for collaborative research efforts. Including patients with onset of IBD during childhood in these investigations is necessary as studies have already demonstrated that such individuals may be the most informative with regard to identifying IBD susceptibility gene loci. A necessary adjunct to such a repository must be the development of a comprehensive, regularly updated clinical database. Such an undertaking, with appropriate privacy safeguards, must include comprehensive data regarding disease characteristics, family health history, and effect of therapy and should be developed to allow genetic correlation with disease phenotype, natural history, and response to therapy. Diagnostic Modalities and Prognostic Indicators The improved diagnosis of IBD in children demands a more accurate definition of the disease process and delineation of extent of disease using less invasive, more “child-friendly” diagnostic modalities. Current definitions of disease describing a patient's clinical phenotype need to be enhanced by disease definitions based on genotype and immunologic profile. To accomplish this, studies will need to be done in different populations around the world to identify the genes predisposing to IBD, and the specific immunologic and cytokine profiles that characterize the different forms of illness. Improved imaging studies are also necessary if they are to replace more invasive, but currently more sensitive, endoscopic techniques. Once this process is successful, it is conceivable that the diagnosis and characterization of IBD in children could eventually be noninvasive; genetic and serologic markers will define the specific illness, and a noninvasive imaging study such as spiral CT, MRI, or tagged leukocyte scan might the be adequate to determine the extent of disease. There is a great need for objective, reproducible disease activity scores to be developed and validated that reflect the clinical problems of IBD in children. Ideally, to foster multicenter, worldwide studies, these scores need to be validated in different populations around the world. The PCDAI appears to be an appropriate instrument for pediatric CD, but it needs further validation using longitudinal observations, as well as evaluation in populations from outside of North America. A pediatric UC score and disease-specific quality of life instruments need to be developed and tested as well. An international collaboration between members of the various Pediatric Gastroenterology Associations worldwide should be convened to develop and test these activity scores. As objective, validated measures of disease activity offer reproducible outcome measures for clinical trials, support for the development and validation of these scores should come from industry. Investigation into the risk factors associated with the development of extraintestinal complications and intestinal cancer needs to continue. Correlation between immunologic and genetic profiles and the development of particular extraintestinal manifestations must be explored. The interplay between environmental factors, diet, coincident infection, and therapy, and how these factors affect the development of intestinal cancer or other clinical characteristics in individuals with IBD, needs to be unraveled. The relationship between other concomitant genetic characteristics (such as mutations in tumor suppressor or DNA mismatch repair genes) and IBD complications must also be undertaken. Children with IBD represent a valuable resource for such investigations, for most do not have the coexisting risk factors for IBD and its complications commonly seen in adult populations. These issues should be included in the proposed prospective worldwide databases for epidemiologic and genetic studies. Therapy It is important to note the lack of international consensus concerning major management stratagies in children. Major variations in basic management protocols exist between respected tertiary units. In particular, there needs to be consensus on the indications for enteral nutrition as opposed to steroids in CD, as the former may offer a metabolic response better suited to the developing child. Novel therapies such as anti-TNF antibody therapy will also have to be placed rapidly in therapeutic context after appropriate large pediatric trials. There are challenges provided by the dramatic pace of technical advance. Molecular genetic mapping may allow stratification of disease subtypes, and may even predict response to therapy. Therefore, whenever possible, children in therapeutic trials should be genotyped as well, and their DNA deposited in the proposed central genetic repository. By so doing, the clinical trial will not only provide data on response to a particular therapy, but will also provide the best opportunity to evaluate correlation between genotype, disease course and therapeutic response. One obvious problem we face is the ethical difficulty of conducting placebo-controlled trials in ill children. Another is the relatively small population available to study compared with the number of subjects required to provide any individual investigation with sufficient statistical power. As academic clinicians, we must commit ourselves to actively participate in collaborative clinical trials. We must adapt our clinical practice to provide both for the needs of our patients and the rather inflexible requirements of controlled trials. Every effort must be made to have a series of clinical investigations ongoing simultaneously so that as many children as possible can be enrolled in an appropriate clinical investigation. There is thus a great need to establish an international pediatric IBD clinical trials bureau, focused on promoting appropriate drug and nutrition trials in children. This bureau should facilitate collaborative studies among international clinical centers, and serve as a clearinghouse for industry sponsored investigations. It should include a team of statisticians as well help in the development of appropriate study design as well as the interpretation of data. Such a bureau could be funded by a combination of government, industry, and national patient organizations. Should major collaborative efforts be directed toward determination of safety and efficacy of individual agents, or toward comparison of broad therapeutic approaches? Is mucosal healing more important than clinical remission? Should studies of combination therapy (the “real world” approach) be undertaken, even though these are currently unfashionable in the adult literature? Should future studies attempt to modify the natural course of pediatric IBD by aggressive early treatment with potent immunomodulators? Rather than the common careful step-up approach in the initial therapy of children and adolescents with IBD, should we give these patients aggressive broad-spectrum treatment, followed by a careful step-down? Should we now move beyond belated “trickle-down” therapeutic trials carried out only after adult studies have been accomplished? These and other questions should be answered by the proposed international pediatric IBD clinical trials bureau, in partnership with industry and national regulatory authorities.