Els van de Vijver

117 concluded that the diagnostic accuracy of the calgranulin-C test was not superior to the FC test. The cohort evaluated in Chapter 2 included patients with rectal blood loss and perianal disease. These red flag symptoms provide sufficient reasons for immediate endoscopic evaluation to obviate the need for additional diagnostic testing. Inclusion of these patients increases the pre-test probability and causes an overestimation of the discriminating power of FC relative to the practical situation, where a test seems particularly useful to discriminate between those with IBD and those with functional abdominal pain. Children and teenagers presenting with non-bloody diarrhoea and abdominal pain, in other words without red flag symptoms, are a spectrum of patients more commonly seen in general paediatric practice. These patients constitute the most challenging group to discriminate IBD from Irritable Bowel Syndrome (IBS) because the pre-test probability for IBD is low. Previously published meta-analyses pooled studies which included patients with red flag symptoms and may have exaggerated the diagnostic accuracy of FC to diagnose IBD. We therefore set out to determine the optimal test strategy in patients without red flag symptoms (Chapter 4). This time we used a FC threshold of 250 µg/g, which was, according to new insights, (1) considered to be the optimal cut-off point to discriminate IBD from functional abdominal disorders. We compared four diagnostic strategies to predict the need of endoscopy based on (A) symptoms alone, (B) symptoms + blood markers, (C) symptoms + faecal calprotectin, and (D) symptoms + blood markers + faecal calprotectin. Triaging with strategy C resulted in 20 of 100 patients undergoing endoscopy, and triaging with strategy D further limited this number to 14 of 100 patients. Eleven out of 14 had IBD and three did not have IBD. No IBD-affected child was missed. Clinical Implications Our search for the optimal diagnostic approach to triage paediatric patients with gastrointestinal complaints and absence of red flags for endoscopy culminated in a combination of meticulous history taking with measuring C-reactive protein in blood and calprotectin in stool (Chapter 4) . This strategy provides an easy and effective way to correctly selecting those who appeared to have IBD. Clinical practitioners can be General discussion 117