ORIGINAL ARTICLES Non-Invasive Markers for Early Diagnosis and Determination of the Severity of Necrotizing Enterocolitis Geertje Thuijls, MD,* Joep P. M. Derikx, MD,* Kim van Wijck, MD,* Luc J. I. Zimmermann, MD, PhD,† Pieter L. Degraeuwe, MD, PhD,† Twan L. Mulder, MD, PhD,† David C. Van der Zee, MD, PhD,‡ Hens A. A. Brouwers, MD, PhD,§ Bas H. Verhoeven, MD, PhD,* L. W. Ernest van Heurn, MD, PhD,* Boris W. Kramer, MD, PhD,† Wim A. Buurman, PhD,* and Erik Heineman, MD, PhD* Objectives: To improve diagnosis of necrotizing enterocolitis (NEC) by noninvasive markers representing gut wall integrity loss (I-FABP and clau- din-3) and gut wall inflammation (calprotectin). Furthermore, the usefulness of I-FABP to predict NEC severity and to screen for NEC was evaluated. Methods: Urinary I-FABP and claudin-3 concentrations and fecal calpro- tectin concentrations were measured in 35 consecutive neonates suspected of NEC at the moment of NEC suspicion. To investigate I-FABP as screening tool for NEC, daily urinary levels were determined in 6 neonates who developed NEC out of 226 neonates included before clinical suspicion of NEC. Results: Of 35 neonates suspected of NEC, 14 developed NEC. Median I-FABP, claudin-3, and calprotectin levels were significantly higher in neonates with NEC than in neonates with other diagnoses. Cutoff values for I-FABP (2.20 pg/nmol creatinine), claudin-3 (800.8 INT), and calprotectin (286.2 g/g feces) showed clinically relevant positive likelihood ratios (LRs) of 9.30, 3.74, 12.29, and negative LRs of 0.08, 0.36, 0.15, respectively. At suspicion of NEC, median urinary I-FABP levels of neonates with intestinal necrosis necessitating surgery or causing death were significantly higher than urinary I-FABP levels in conservatively treated neonates. Of the 226 neonates included before clinical suspicion of NEC, 6 developed NEC. In 4 of these 6 neonates I-FABP levels were not above the cutoff level to diagnose NEC before clinical suspicion. Conclusions: Urinary I-FABP levels are not suitable as screening tool for NEC before clinical suspicion. However, urinary I-FABP and claudin-3 and fecal calprotectin are promising diagnostic markers for NEC. Furthermore, urinary I-FABP might also be used to predict disease severity. (Ann Surg 2010;251: 1174 –1180) N ecrotizing enterocolitis (NEC) is a severe gastrointestinal dis- order with high morbidity and mortality (20%– 40%), affecting predominantly premature neonates. 1 NEC continues to present a diagnostic challenge to clinicians. Especially, early detection of intestinal necrosis requiring surgical treatment is still a key problem. The initial clinical manifestations of NEC are nonspecific and indistinguishable from other gastrointestinal disorders and sepsis. 2 Diagnosis is further hampered by limited diagnostic accuracy of laboratory tests and currently used imaging modalities. 3,4 The pathophysiology of NEC is characterized by defects in the intestinal epithelial barrier and gut wall inflammation. 1,5 There- fore, we sought noninvasive tests to identify gut wall integrity loss and gut wall inflammation in neonates with gastrointestinal symp- toms suspected of NEC, to differentiate NEC from other neonatal diseases that present with abdominal signs. We used urinary markers to study the loss of integrity of enterocytes and tight-junctions, the 2 components comprising the intestinal epithelial lining, and a fecal inflammation marker to evaluate intestinal wall inflammation in neonates suspected of NEC. Intestinal fatty acid binding protein (I-FABP) has been re- ported to be a useful plasma and urinary marker for enterocyte damage. 6–8 The presence of I-FABP, a small (14 –15 kDa) cytoso- lic, water-soluble protein, is limited to mature enterocytes of small and large intestine. It is released into the circulation as soon as cell membrane integrity is compromised. 6–9 Due to its small molecular weight, plasma I-FABP passes the glomerular filter (fractional renal excretion: 28%, half-life time: 11 minutes) and can readily be detected in urine. 10 Therefore, urinary I-FABP values can provide specific and actual information about the extent of intestinal epithe- lial cell injury. Loss of tight junction strands, the paracellular barrier, is part of intestinal integrity loss. These tight junctions consist of a large complex of intra- and extracellular proteins, including claudins (22 kDa). 11,12 Claudin-3 is an important sealing tight junction protein and is expressed in high quantity in the intestine. 11 Claudin-3 disappears rapidly from tight junctions following hemorrhagic shock and inflammatory bowel disease. 13 Intestinal inflammation is characterized by sequestration of neutrophils into the gut wall. 1,5 Intestinal neutrophil influx followed by activation results in release of calprotectin, a heterodimeric peptide (36 kDa) constituting 60% of the cytosolic content of neutrophils. In intestinal inflammation, calprotectin is readily detect- able in feces and plasma. Furthermore, calprotectin is remarkably resistant to degradation by fecal bacteria, making fecal calprotectin a suitable marker for gut wall inflammation as reported for inflam- matory bowel disease. 14 –16 This study aims at improving diagnosis of NEC, by using noninvasive tests to identify gut wall integrity loss and gut wall inflammation in infants with gastrointestinal symptoms suspected of NEC. Furthermore, the usefulness of urinary I-FABP to predict severity of NEC and to screen for NEC was evaluated. From the *Department of Surgery, Maastricht University Medical Centre, and Nutrition and Toxicology Research Institute (NUTRIM), Maastricht, The Netherlands; †Department of Paediatrics, Maastricht University Medical Cen- tre, and School for Oncology and Developmental Biology (GROW), Maas- tricht, The Netherlands; and the Departments of ‡Surgery and §Neonatology, Wilhelmina Children’s Hospital, University Medical Centre, Utrecht, The Netherlands. Supported by AGIKO-stipendium 920 – 03– 438 from The Netherlands Organisa- tion for Health Research and Development (to J.P.M.D.); and by Stichting Sint Annadal, Maastricht, The Netherlands; and, also, partially funded by a “Profileringsfonds” grant from the University Hospital Maastricht. All authors contributed to analysis and interpretation of data; revised the manu- script critically for important intellectual content; and gave final approval of the version to be published. Erik Heineman, MD, PhD, is currently at Department of Surgery, Groningen University Medical Centre, Groningen, The Netherlands. Reprints: W.A. Buurman, PhD, Department of Surgery, Maastricht University Medical Centre, Universiteitssingel 50, 6229 ER, Maastricht, The Nether- lands. E-mail: w.buurman@AH.unimaas.nl. Copyright © 2010 by Lippincott Williams & Wilkins ISSN: 0003-4932/10/25106-1174 DOI: 10.1097/SLA.0b013e3181d778c4 Annals of Surgery • Volume 251, Number 6, June 2010 1174 | www.annalsofsurgery.com