Answers are grounded in this paper and cite their sources.
“If we knew what we were doing, it wouldn't be called research.” —Albert Einstein “Your child has inflammatory bowel disease.” Any of us who have said these words recognize the import to the child and the family. Childhood and adolescence have their own challenges and the diagnosis of any chronic medical condition such as inflammatory bowel disease (IBD) adds a completely new and unwelcome dimension to the lives of all those involved. In many cases, the diagnosis of IBD comes after months of gastrointestinal symptoms, doctors' visits during which reassurance is given while concerned parents maintain that “Johnny is not getting better” and, ultimately, referral and subspecialist investigation. Along with the diagnosis of IBD come many questions, most of which center around “will my child be alright?” The good news in the 21st century is that our therapeutic arsenal has never been larger. Clinical trials and drug discovery have given us new and improved formulations of topical agents, a variety of immunomodulators, and an ever-growing number of biologic agents. With increasing therapeutic potency and immunologic directed mechanisms of actions come increased potential risks that have to be weighed against the potential benefits. In addition to choosing between the many medical options, those who care for IBD patients recognize that surgery is often times the appropriate intervention. When making such therapeutic decisions for their IBD patients, clinicians often find themselves with their own set of questions centering on “how can I best assure that Johnny will be alright?” To answer this pivotal question for any one patient, we need a deep understanding of the expected natural history of their IBD. If we were able to predict more aggressive disease, we would be in a better position to decide on who should receive more aggressive therapy. So what do we know? In general terms, it is recognized that pediatric-onset IBD usually involves a greater percentage of the gastrointestinal tract and has a more aggressive course than adult-onset IBD.1 In pediatric patients with ulcerative colitis (UC), the aggressive nature of the disease is marked by a relatively high rate of early colectomy. In fact, studies in the 1990s showed the rate of colectomy to be about 25%2 and, despite medical advances, recent investigators have shown a similar colectomy rate3 raising the question as to whether we are optimizing the use of our therapeutic armamentarium. For pediatric UC patients the current literature does suggest that those who present with mild disease have a tendency to remain clinically mild over time.3 Unfortunately, close to 80% present with moderate/severely active disease and 31% will be steroid-dependent by 1 year.4 While a proportion of these will ultimately respond to medical therapy, the rest are destined to become part of that 25% colectomy cohort under our current treatment paradigms. How nice it would be to have the ability to reliably predict, on the day of diagnosis, who will have aggressive disease or who will not respond to conventional agents, especially corticosteroids. Such knowledge would allow us to abandon such concepts as “top down” or “bottom up” therapy—approaches that often feel like one size fits nobody. Instead, we could use “step in” therapy and give the right treatment to the right patient. The positive impact on clinical outcome and the quality of life of our patients and their families would be quite gratifying to all involved. In this issue of the Journal, Moore et al5 explore whether, at the time of diagnosis, readily accessible and inexpensive clinical variables could predict colectomy risk in pediatric UC patients. In their single-center, retrospective cohort review from Primary Children's Medical Center (PCMC) in Salt Lake City, Utah, the authors looked at all pediatric patients (younger than or equal to 18 years) with UC who were cared for at their institution from 1997 to 2007. Taking prespecified clinical variables, the authors attempted to generate a colectomy risk score. Using a multivariate regression model, white blood cell count and hematocrit at diagnosis were found to be clinical variables of significance. Using these variables the authors created a UC risk score for predicting the need for colectomy. The authors found that their score had negative predictive values of 0.95 and 0.89, but low positive predictive values of 0.22 and 0.38 at 1 and 3 years following diagnosis, respectively. The authors conclude that prospective multicenter studies need to be performed to better predict pediatric UC outcome and that those studies should further investigate the utility of including simple clinical variables such as the white blood cell count and hematocrit at diagnosis. At first glance, the study seems to only confirm what is already known. Patients who had mildly active disease at presentation, as measured by relatively low white blood counts and normal or near-normal hematocrits, were unlikely to progress to colectomy. What is needed is a model that can predict which of the patients who present with more active disease will go on to demonstrate an aggressive, medication-unresponsive course. The presented model did not provide such risk stratification. However, as the authors correctly point out, their study's limitations may have significantly affected the outcome, thereby limiting the apparent significance of their model. This was a single-center and retrospective experience with only 37 colectomy patients. Perhaps a larger, prospective, powered study would have shown more significance. Inclusion of other significant variables such as age may have strengthened their risk model, as other investigators have also found age to be an important variable.3 Additionally, perhaps colectomy was too strict an endpoint and using the need for salvage therapy (infliximab or calcineurin inhibitor) may have yielded more convincing, and no less important results. It should be pointed out that Moore et al raise a very valid point. As medicine becomes more complicated, we should not come to think that data points are inherently useless if they are simple to obtain. An excellent example of this already comes from the study of pediatric UC. The Pediatric Ulcerative Colitis Activity Index (PUCAI) was developed by consensus using the Delphi technique and is composed of six common and easily measured clinical symptoms of pediatric UC.6 A recent prospective, multicenter trial of severe, hospitalized pediatric UC patients receiving intravenous corticosteroids showed that the PUCAI on days 3 and 5 of such therapy accurately predicted the likelihood of colectomy as well as response to therapy 1 year after discharge.7 The ability to use the PUCAI in this manner is likely to shorten the hospital course and decrease the prolonged use of futile therapy in this setting. Having a similar tool to predict outcome in newly diagnosed children with UC was the aim of Moore et al. Unfortunately, the very low positive predictive value of their risk score precludes its current “prime time” use in the clinical arena. In real terms, it does not help predict colectomy in high-risk pediatric UC patients—which are the majority. But, the authors' work should be appreciated for its depth as well as its novelty. The field of risk stratification in pediatric IBD is in its infancy. Indeed, the Crohn's and Colitis Foundation has funded a multicenter, prospective risk stratification project in pediatric Crohn's disease through the Pediatric Research Organization For Kids with IBD (PRO-KIDS). The planning and launch of such a project in pediatric UC is too important to not be imminent. Moore et al have raised a thought-provoking question—perhaps simple clinical variables should be studied as part of this project along with newer biomarkers,8 serology,9 and genetics.10 In an era of cost consciousness, it behooves us to answer the call of this study and to look at simple and inexpensive potential markers of disease risk. The author thanks Rami Bustami, PhD, and Eric Lazar, MD, for insights and discussions in the preparatrion of this editorial.
No takes yet. Share an insight, caveat, or question.
Joel R. Rosh (2010) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: