Els van de Vijver

116 This dissertation addresses two particular knowledge gaps out of the many that exist in paediatric inflammatory bowel disease (IBD). In the first part of this thesis, we evaluated diagnostic strategies to assess whether gastrointestinal complaints are due to IBD, for appropriate triage for endoscopic evaluation. In the second part, we quantified and characterised fatigue in IBD. In this chapter, we will discuss the main results and their clinical implications. PART I - Triage for endoscopy In our effort to develop an appropriate strategy whether or not endoscopy is indicated to evaluate IBD in a child with abdominal complaints, we first evaluated faecal calprotectin (FC) as an isolated triage test. In Chapter 2 we described a cohort of 117 children with chronic diarrhoea and nonspecific abdominal pain. The treating physicians had to base their decision whether or not to perform endoscopy on the standard practice of that time: a combination of signs, symptoms and blood results. Without the knowledge of the FC result, 62% of the patients that were selected for endoscopy were diagnosed with IBD. If they would have based the selection for endoscopy on the combination of raised FC levels (i.e. >50 µg/g) and negative stool cultures, the yield of ileocolonoscopy towards diagnosing IBD would have improved to 78%, without missing any IBD patient. At the same time, FC levels below this cut-off point would have prevented a considerable proportion of patients being subjected to an endoscopic procedure that would not have led to the diagnosis of IBD, and, arguably, could then even have been labelled ‘futile’. Even though adding FC results to the decision strategy improved the diagnostic yield compared to the standard diagnostic strategy of that time, still 22% of the patients would have been subjected to an IBD-negative ileocolonoscopy. In Chapter 3 we evaluated whether another faecal biomarker for mucosal inflammation, calgranulin-C, is better than FC in predicting IBD in children and teenagers. When predefined test thresholds were used (50 µg/g for FC and 0.75 µg/g for calgranulin-C), the diagnostic accuracy of calgranulin-C indeed appeared to be better. However, when receiver-operator characteristic (ROC) curves were used to identify the optimal test threshold for each test separately, what appeared to be 400 µg/g for FC and 0.75 µg/g for calgranulin-C, the superiority of calgranulin-C relative to FC disappeared. We therefore Chapter 7 116