Risk factors of enterostomy in neonates with Hirschsprung disease

This study was aiming to explore the risk factors contributing to enterostomy in neonates with Hirschsprung disease (HD) to provide a reference for clinicians to make treatment decisions. Medical records of 284 patients diagnosed with HD during the neonatal period were retrospectively analyzed. The patients were divided into 2 groups based on operative intervention (one stage transanal pull-through, versus enterotomy and staged transanal pull-through). Univariate and multivariable logistic regression analysis was performed to identify risk factors contributing to enterostomy. The incidence of enterostomy was 12.0% (34/284) in neonates with HD. Univariate and multivariate logistic regression analysis showed that serum albumin < 25.4 g/L, radiographic results as subphrenic free air, and level of aganglionosis with long-segment or total colonic aganglionosis (TCA) were independent risk factors of enterostomy in neonates, with OR of 42.045 (6.131, 288.319), 285.558 (26.651, 3059.694) and 15.573 (4.319, 56.157), respectively. The low serum albumin level, bowel perforation, and level of aganglionosis with long-segment or TCA could influence the occurrence of enterostomy in neonates with HD.


Introduction
Hirschsprung disease (HD) is a congenital bowel disorder characterized by the absence of intrinsic parasympathetic ganglia of the intestinal tract with an incidence of 1:5000 [1]. Neonates commonly present with delayed passage of meconium, abdominal distention, and vomiting due to functional intestinal obstruction [2]. Previously, a lot of surgeons preferred diverting stoma (colostomy or ileostomy) during the neonate period and followed by staged surgeries. These staged operations brought physical and psychological burdens on patients, and stoma management may be associated with risk of stoma complications and financial burden [3]. In recent years, most pediatric surgeons have accepted single stage surgery for treatment of HD, which can shorten the length of hospital stay, reduce the number of anesthesia, reduce the cost, and avoid a series of complications [4,5]. However, there are still a few patients with HD who cannot avoid staged surgery. Especially for those with bloating that cannot relived by colonic irrigation, and/or severe enterocolitis, and/or bowel perforation, enterostomy is a proven lifesaving method [6] during their neonate period. Therefore, quantitative assessment of the risk factors leading to enterostomy can help surgeons choose the right time of enterostomy.
There are few studies that focus on the risk factors of enterostomy in neonates with HD, so this study retrospectively reviewed neonates with HD diagnosed with HD in our hospital, to explore the relevant risk factors of enterostomy.

Patients and clinical characteristics
Beijing Children's Hospital. The criteria for inclusion in the study were as follows: (1) infants under 28 days of age when diagnosed of HD, (2) preoperative diagnosis of HD based on the results of suction rectal biopsy (SRB). The exclusion criteria were as follows: (1) postoperative histological findings did not support diagnosis of HD, or (2) infants over 28 days of age when diagnosed of HD.
Clinical and laboratory data were collected and analyzed. Patients were, therefore, divided into two groups according to surgical technique performed: one stage transanal pull-through, versus enterotomy and staged transanal pullthrough. The study was approved by the Medical Ethics Committee of Beijing Children's Hospital (2020-Z-082), and the patient informed consent requirements were waived.

Statistical analysis
Statistical analysis was performed by SPSS 26.0. Continuous variables were presented as the mean and standard deviation for normal distribution, or median and interquartile range (IQR) for non-normal distribution. Categorical variables were reported as counts and percentages. Pearson's χ 2 test, Fisher's exact test, two independent samples t-tests and the non-parametric Mann-Whitney U test was used to compare characteristics between the enterostomy and non-enterostomy groups. Receiver operating characteristic (ROC) curve analysis was performed to determine the most appropriate cut-off values. Univariate and multivariable logistic regression analysis were performed to identify risk factors of enterostomy. p < 0.05 was considered statistically significant.
In total, 235 neonates (82.7%) with HD had undergone one-stage transanal endorectal pull-through (TEPT) with a median age at surgery of 17 days (range: 2-28 days); 15 neonates (5.3%) who were diagnosed with HD by clinical manifestations and SRB had been managed with colon enema in the early stage, and then, they received one-stage TEPT, at a median age of 118 days (range: 41-202 days); 34 neonates (12.0%) with HD received enterostomy prior to definitive surgery, of which ileostomy was performed in 12 cases with TCA and 2 cases with long-segment HD, and colostomy was performed in 12 cases with long-segment HD and 8 cases with short-segment HD. The level of aganglionosis was determined to be short-segment (197/284, 69.4%), longsegment (71/284, 25.0%), and total colonic aganglionosis (TCA) (16/284, 5.6%).

Comparison of clinical characteristics between the enterostomy and non-enterostomy groups
Of the 284 neonates in this group, enterostomy was performed in 34 cases (12.0%). As shown in Table 1, by comparing clinical characteristics between enterostomy (n = 34) and non-enterostomy groups (n = 250), we found significant differences in fever, preoperative mechanical ventilation, level of aganglionosis, C-reactive protein (CRP), neutrophil ratio (NEUT%), plasma albumin, plasma fibrinogen (FIB), radiographic results as subphrenic free air, and radiographic results as intestinal wall pneumatosis (all p < 0.05). There were no differences between the two groups for other clinical characteristics (all p > 0.05).

Risk factors of enterostomy
In order to find the risk factors of enterostomy in neonates with HD, we conducted a multivariate analysis. ROC curve analysis was used to determine the stratification value for NEUT%, plasma albumin and FIB according to the maximum combined sensitivity and specificity values, and the cut-off values for the above characteristics were 44.4, 25.4 g/L, and 2.4 g/L, respectively. As shown in Table 2, multivariate analysis showed that serum albumin < 25.4 g/L, radiographic results as subphrenic free air and level of aganglionosis with longsegment or TCA were independent risk factors of enterostomy in neonates with HD, with OR of 42.045 (

Discussion
Most HD children are diagnosed and treated in the neonatal period and early infancy, and the main treatment principle of HD is resection of the diseased bowel segment [7]. Despite the fact that one-stage operation has become the preferred practice in managing HD, some children require an enterostomy in neonatal period. Patients undergoing staged surgery had higher rates of readmissions and additional operations [4], therefore, to make the appropriate treatment decisions, it's essential to understand the factors that influence enterostomy. According to literature reports, the incidence of enterostomy in infants with HD is 15.4-53%, and presents a gradually decreasing trend [4,[8][9][10]. In our study, the incidence of enterostomy was 12.0%, which was relatively lower. Few studies have reported the influencing factors of enterostomy. A retrospective US analysis of a multi-institutional cohort of neonates with HD showed that patients with low birth weights, premature delivery, and non-HD gastrointestinal anomalies were more likely to undergo a stoma before definite surgery [4]. Previous studies showed that Down syndrome was strongly associated with stoma formation [8,11]. In this study, the factors mentioned above were not independently related to the enterostomy, and we found that low serum albumin level, bowel perforation, and level of aganglionosis with long-segment and TCA were significant independent risk factors of enterostomy in neonatal HD patients.
We found that the serum albumin level was significantly lower of patients with stoma than that of patients without stoma. The serum albumin was usually used to assess acute versus chronic malnutrition, but some studies showed that it was likely more predictive of inflammation and morbidity rather than nutritional status [12,13]. In our study, there was no significant difference in birth weight or hospital weight between the enterostomy group and the nonenterostomy group, but there were significant differences in many inflammation-related indexes such as CRP, WBC, NEUT% between the two groups, which seemed to indicate that the decrease of albumin was related to the poor state of the patient.
Commonly reported factors influencing the decision for enterostomy include the presence of preoperative WBC white blood cell count, PLT platelet count, RBC red blood cell count, BUN blood urea nitrogen, ALP alkaline phosphatase, AST aspartate amino transferase, ALT alanine aminotransferase, TBIL total bilirubin, DBIL direct bilirubin, PT prothrombin time, APTT activated partial thromboplastin time enterocolitis [9,14,15]. However, Bradnock et al. [8] indicated that enterocolitis was not independently correlated with stoma formation. This could be because the clinical definition of enterocolitis is still elusive, making it difficult to capture the case. We did not collect cases of enterocolitis, but we collected and analyzed clinical factors associated with the diagnosis of enterocolitis [16], such as diarrhea, distended abdomen, fever, CRP, WBC, NEUT%, radiographic results as gas-fluid level, and radiographic results as intestinal wall pneumatosis. The results showed that fever, CRP, NEUT%, and radiographic results were associated with enterostomy, but they were not significant in multivariate logistic regression analysis, possibly because they were predictors of enterocolitis that might lead to enterostomy. Since previous studies have used different definitions of HDassociated enterocolitis, factors for the relationship between enterocolitis and enterostomy are not absolute, and the conclusions are still controversial. According to the European Reference Network for Rare Inherited Congenital Anomalies (ERNICA) guidelines for the management of rectosigmoid HD, a stoma is indicated if there is bowel perforation [6]. In our study, the incidence of bowel perforation was 5.3%, and the multivariate logistic regression analysis showed that radiographic result as subphrenic free air was a vital risk factor of enterostomy. Although most HD patients mainly present with low intestinal obstruction in the neonatal period, there are also some children with intestinal perforation as the first symptom, mostly located in the cecum and appendix [1]. However, we found that 53.3% of patients (8/15) had perforations in the sigmoid colon, which was similar to another research about perforated HD [17]. The intestines of newborns with HD-associated enterocolitis are very fragile and easy to perforate during colon enema. In this group, 4 cases had obvious abdominal distension on admission, which could not be relieved by diet resistance, and received colon enema. Unfortunately, the 4 patients developed intestinal perforation after colon enema. Therefore, for institutions with less experience in colon enema in the neonatal period, in order to avoid iatrogenic perforation and aggravation of the patient's condition, temporary enterostomy could be selected. The atypical manifestations of neonatal intestinal perforation make clinical diagnosis and treatment extremely difficult. Once a diagnosis has been made, an exploratory laparotomy should be performed as soon as possible, and pathological biopsy should be performed during the operation on children suspected of HD [18].
In addition, two patients needed preoperative mechanical, and both had bowel perforation. We found that preoperative mechanical was associated with enterostomy, consistent with the previous study [4], which might mean that patients were seriously ill and needed emergency surgery. However, the number of events in this study was limited, necessitating additional research.
Compared to rectosigmoid HD, long-segment HD or TCA conferred a higher rate of stoma formation [9,15,19]. In this study, type of HD with long-segment and TCA was an independent risk factor in multivariate logistic regression analysis, which was in agreement with a prospective study [8]. The intestinal obstruction in children with long-segment or TCA often appeared earlier, and the effect of colon enema was poor, which was easy to be complicated with enterocolitis [20]. As a result, if neonates with long-segment HD or TCA are in poor general condition, an enterostomy may be preferable. And in our experience, patients with TCA and long-segment HD whose intestinal morphology was atypical received the ileostomy and patients with short-segment and long-segment HD whose intestinal morphology was typical received the colostomy. However, it is worth mentioning that not all long-segment HD require functioning stoma in the neonatal period, which is also reflected in our data.
The retrospective nature of this study limited it, and it was subjected to several confounding variables. Although some of the analyses in this study did not reach statistical significance, we believe they may be clinically important, and these risk factors should be closely monitored. Additionally, cases of HD that were treated on an outpatient basis during neonatal period would not be captured in our cohort. A prospective study is required to explore the relationship between additional risk factors and enterostomy.

Conclusions
In conclusion, temporary enterostomy is an important means to relieve the condition of HD patients with low serum albumin level, intestinal perforation, and long-segment HD or even TCA suggested by preoperative examination during the neonatal period.
Author contribution Jinshi Huang, Yanan Zhang, and Zhaozhou Liu designed the study. Zhaozhou Liu and Yanan Zhang wrote the main manuscript text. Yongwei Chen, Weihong Guo, Dawei Hou, and Yingzi Li prepared Tables 1 and 2. All authors reviewed the manuscript.