Necrotizing enterocolitis (NEC) is a common and critical gastrointestinal disease in neonates, especially in preterm infants, with typical manifestations such as abdominal distention, vomiting, and blood stools (1). With the development of medical technology, the incidence of NEC has been decreasing over the past decade (2). However, it is still a serious problem in preterm infants. In those neonates born less than 32 weeks, the incidence rate ranges from 5% to 22% (3) and that of mortality ranges from 18.9% to 23.5% (4, 5). In addition, complications such as intestinal stenosis, short bowel syndrome, and neurobehavioral retardation subsequently affect the quality of life of NEC infants, especially those reaching Stage III (6, 7). Thus, for those infants with NEC, timely recognition of disease deterioration is important. However, the Bell staging criteria that are used nowadays are based on clinical manifestations and imaging presentations (8), and when infants meet the criteria of Stage III, peritonitis, septic shock, and even ensuing intestinal perforation, the disease will be life-threatening (9). Therefore, it is important to explore new indications to prevent the deterioration of NEC, especially from Stage II to Stage III.

Human β-defensin 2 (HBD-2), Claudin-3, high-mobility group box-1 protein (HMGB-1), and resistin-like molecule β (Relmβ) are potential biomarkers in intestinal inflammation, and they have been proposed as such in the diagnosis of NEC (10–15). HBD-2 is an antibacterial peptide in the innate immune system of the intestinal mucosa that can regulate the immune response with Toll-like receptor 4 and other recognition receptors to maintain the normal function that is expressed in inflammation but not in normal epithelial tissues (16–18). A previous study showed that serum HBD-2 increased in Crohn’s disease because of its systemic antimicrobial and immunomodulatory roles (19), and infants with moderate NEC showed significantly increased fecal HBD2 concentrations before the occurrence of clinical symptoms (11). Claudin-3 is an important subtype of the Claudin family, one of the most important tight conjunction proteins in the intestinal tract. It forms a dense structure between intestinal epithelial cells to reduce the permeability material exchange (20). It has been shown that Claudin-3 is a potential diagnostic biomarker for intestinal permeability and it increases in the intestinal tissues of NEC rats (10, 14).HMGB-1 in the nucleus is actively secreted by intestinal epithelial cells or inflammatory cells stimulated by inflammatory response signals or released into the intestinal lumen by necrotic cells. It can bind to non-histone chromosomes, activate Toll-like receptor 4 to inhibit intestinal cell migration, and delay the repair of intestinal mucosa (21–23). In an animal model of colitis, the expression of HMGB-1 increased in the intestinal tissue (24), and compared with those infants without NEC, the fecal level of HMGB-1 of those with NEC was significantly higher (15). Relmβ is a bactericidal protein produced by colonic epithelial goblet cells, playing an important role in the local mucosal immunity of the intestine, maintaining the integrity of the intestinal barrier, and inducing the differentiation and proliferation of epithelial cells in intestinal inflammation (25, 26). It has been pointed out that colitis mice with silencing Relmβ genes suffer less mucosa injury (27) and that serum Relmβ in neonates with NEC is significantly higher (13).

HBD2, Claudin-3, HMGB-1, and Relmβ all have potential efficacies to predict the onset of NEC (10–15). However, a few studies focused on whether they could act as potential biomarkers to predict the deterioration of NEC. In addition, the collection of fecal samples is noninvasive in nature, and compared with blood sampling, there is less pain and feces can more directly reflect intestinal injury (28). Therefore, the aim of this study is to explore the value of fecal HBD2, Claudin-3, HMGB-1, and Relmβ to predict the deterioration of NEC from Stage II to Stage III.

Infants with NEC treated in the Neonatal Diagnosis and Treatment Center at Children's Hospital of Chongqing Medical University between March 2019 and November 2020 were enrolled in this study. This study was approved by the ethics committee of Children's Hospital of Chongqing Medical University (No. 2019-284), and consent forms were obtained from parents or guardians of the infants.

During the study period, a total of 81 infants with NEC at Stage II were enrolled. Eight infants were excluded for other gastrointestinal diseases, with three for appendicitis (n = 3), three for gastric wall perforation (n = 3), and two for intestinal atresia (n = 2). Seven infants were discharged against medical advice (n = 7) and six refused to participate in this study (n = 6). Finally, 60 infants diagnosed with NEC at Stage II were enrolled in our study. Among them, twenty-seven progressed to Stage III rapidly and were included in the Stage III group (n = 27). Thirty-three infants were enrolled into the Stage II group (n = 33).

NEC is a severe intestinal disease in neonates (3) and it is important to recognize the progression of diseases in time to reduce mortality and the incidence of sequelae, including peritonitis, intestinal adhesion, and sepsis, as shown in our study (35). We investigated the value of fecal biomarker as well as some non-specific laboratory metrics in the progression of NEC from Stage II to Stage III and found that fecal HBD-2 and Claudin-3 may be potential biomarkers in the prediction of deterioration of NEC.

Jenke et al. found that HBD-2 increased in the intestinal tissues from very-low-birth weight infants in moderate NEC, while in six infants with severe NEC, HBD-2 decreased when surgeries were performed on them (11). Thuijls et al. explored the urinary Claudin-3 in 14 NEC infants and found that it could be promising biomarkers for diagnosis, but they did not possess the data of urine Claudin-3 in severe NEC infants (36). To our knowledge, few clinical studies have focused on fecal HBD-2 and Claudin-3 to predict the disease progression of NEC. In our study, we found that when the infants were diagnosed with NEC, the levels of HBD-2 and Claudin-3 were higher in those who finally progressed to Stage III. Moreover, in our study, ROC curves were established and both showed medium value to predict the deterioration of NEC, and we also evaluated the combined utility of HBD-2 and Claudin-3. We found that when the fecal concentrations of both HBD-2 and Claudin-3 increased once NEC was diagnosed at Stage II, there was a high possibility of Stage II progressing to Stage III. Thus, to prevent disease progression, when HBD-2 concentration is higher than 1649.02 ng/g and Claudin-3 concentration is over 2488.71 pg/g, in the infants, extra attention should be paid to them and positive treatment should be given clinically.

Also, in our study, neutrophil count was found decreased in infants progressing to Stage III. However, in the advanced regression analysis, when we adapted the widely accepted standard of neutropenia as an independent risk factor, it showed no significant difference. Thus, the value of neutrophil as well as its cutoff value in the prediction of deterioration of NEC should be further explored in clinical trials with larger sample sizes in the future. In addition, we measured the fecal concentration of HMGB-1 and Relmβ in our study. Fecal HMGB-1 has been reported to be helpful in the diagnosis of NEC (15), and our previous study found that in those needing surgeries, the level of serum Relmβ was even higher (12). HMGB-1 participates in intestinal inflammation of NEC by inhibiting the NLR pyrin domain containing 3 via Toll-like receptor 4 and nuclear factor-kappaB signaling pathways and decreases intestinal microcirculatory perfusion by rescuing nitric oxide production and eliminating oxygen production through endothelial nitric oxide synthase activation (23, 37). Thus, HMGB-1 might increase in NEC, in which both infection and hypoxia play an important role. Relmβ is a bactericidal protein that is increasingly secreted in intestinal inflammation expression in the tissue (38). Relmβ stimulates resident immune cells within the mucosa to secrete significant amounts of proinflammatory cytokines by nuclear factor kappaB in ileitis (39, 40) and takes part in the maintenance of the mucosal defense barrier in colitis by upregulating the gene expression of mucin between intestinal epithelial cells (41). However, HMGB1 and Relmβ showed no statistically significant difference. The small sample size might be the main reason for this, warranting further exploration in clinical studies with larger sample sizes.

There were some limitations in our study. First, fecal biomarkers should be dynamically monitored in the progression from Stage II to Stage III. Second, there are several outliers of the biomarkers and they might impact the present results of our study with their small sample sizes. Therefore, a larger sample size study is needed to determine whether these outliers represent the internal changes of these biomarkers. Third, the gestational age of the infants in the Stage III group was smaller than that of those in the Stage II group. It has been found that concentrations of fecal HBD-2 increase with increasing gestational age (42), and it seems that it was not influenced by the difference of gestational age in our study. However, it has not been reported how fecal claudin-3 changes with gestational age in neonates, and whether gestational age is a significant confounder in our study still needs further exploration. Finally, as the center of the Neonatal Emergency Transport System in southwestern China, our department serves neonates form several provinces, and many of the NEC patients enrolled in our study were late preterm and term infants, because most infants admitted to our department were of larger gestational age. Thus, multicentral clinical trials with larger sample sizes enrolling infants with different gestational ages, especially preterm or extremely preterm infants, are needed to verify the predictive value of these biomarkers in the study.

In this study, we found that fecal HBD-2 and Claudin-3 increased, while neutrophil decreased when NEC infants progressed to a worsening condition. Fecal HBD-2 and Claudin-3 may be potential biomarkers to predict the deterioration of NEC from Stage II to Stage III, and when their levels increase in some patients, more attention should be paid to them.