Necrotising enterocolitis–A 15‐year outcome report from a UK specialist centre

Necrotising enterocolitis (NEC) is a disease associated with high mortality and morbidity, low birthweight and prematurity are risk factors. This study reports outcomes of babies having emergency laparotomy for NEC, examining institutional trends and exploring impact of multiple variables on mortality at 30 days and 1 year post‐operatively.

term infants and in these cases may co-exist with other co-morbidities reportedly with a more fulminant onset. 5,6 Operative survival rates for NEC vary widely in the literature.

Battersby et al's 7 systematic review study initially identifies 1888
papers reporting NEC outcomes in high-income countries, in those chosen for review mortality rates ranged from 21.9% to 38%. Notable amongst those cited is a report from Youn et al (Korea) who analysed survival outcomes for surgically managed VLBW (<1500 gm) NEC newborns (n = 77) recording a 29% mortality in this subgroup. 8 A study from Finland has further shown a 27% overall mortality rate for pre-term NEC infants having operation. 9 Alder Hey Children's NHS Foundation Trust is a UK tertiary specialist centre serving a catchment population of over 4.5 million where our regional service operates and manages an integrated neonatal network to triage all newborns requiring neonatal surgery.
Within the network, decision to transfer a neonate for surgical review and operation is made between neonatologists and paediatricians in referring hospitals in consultation with the 'on-call' attending paediatric surgeon. Criteria defining indication(s) for operative intervention at our paediatric surgery centre include (a) cases of intestinal perforation and pneumoperitoneum, (b) bowel obstruction, (c) persistent inflammatory mass (d) and/or failure of maximum medical NEC conservative therapy(s). Operative strategy is determined by the duty surgeon and includes central venous catheter insertion (Bard ® Broviac) or a secure peripheral venous access long line to facilitate resuscitation (antibiotics, blood products, inotropes, etc), laparotomy, bowel resection and/or stoma formation, or the use of the 'clip and drop' back technique when babies are very unstable. Postoperatively, we practise a care pathway where infants are re-located after immediate post-operative stabilisation back to tertiary-level NICUs within the regional network. A system is successfully undertaken and reported by other centres worldwide. [10][11][12][13][14] This study herein examines trends in survival and outcomes in a 15-year period for all babies having emergency laparotomy for NEC at Alder Hey Children's NHS Foundation Trust. We survey the influence of common variables such as year of birthweight, gestational age, congenital cardiac anomaly(s) requiring intervention, presence of pneumoperitoneum at presentation and area of diseased bowel segment affected on mortality (30 days and 1 year) in an effort to develop a mortality predictor. We have chosen these defined timescales as mortality at 30 days post-operatively is a well-established standard surgical metric for 'operative complication(s)'. Large UKwide cohort studies likewise utilise 1-year mortality as a long-term indicator of co-morbid outcome(s). 15 This study also explored the effect of gestational age on the timing (postbirth) of emergent operation for NEC and its relationship (if any) to infant survival. Data were analysed using SPSS. Descriptive statistics of the cohort were produced, and univariate logistic regression was then deployed to further analyse whether the recorded variables of gestation, birthweight, free air (at presentation), and type of bowel region involvement (focal, multifocal, pan-intestinal NEC) and/ or need for cardiac surgery had any significance on 30 days and 1-year mortality rate(s).

| ME THODS
Base on the results from the univariate analysis, variables were chosen to perform a multivariate multiple regression model which was then used to produce a graphical post-operative NEC predictive mortality score.

| Descriptive statistics and mortality figures
The data are non-normally distributed (median and interquartile ranges have been used where appropriate). The population rangethe male:female ratio was 3:2, median (IQR) gestation was 28 weeks

| Operative strategy
Of the 243 patients having operations for NEC, 3 cases underwent    Table 1 which was statistically significant; negative Pearson correlation (P-value 0.027) (Figure 1).

| Statistical analysis and mortality predictor
A univariate linear regression model was conducted using SPSS looking at all collected variables (including year of surgery) and their effect on 1-year mortality.

TA B L E 3 Multivariate multiple regression model
predict mortality for gestational age by incorporating the variables of free air and region of bowel involvement-see Figure 2

| D ISCUSS I ON
This UK single-centre study shows good survival outcomes for NEC babies with low 30-day and 1-year mortality metrics consistent with large UK nationwide reports. 15,16 These outcomes have been achieved with a clinical network utilising co-location to regional NICUs to facilitate post-operative recovery demonstrating the effectiveness of integrated service led pathways. This UK cohort study is therefore widely supportive of data published elsewhere from other centres of this safe practice. 14 We demonstrate a significant relationship with gestational age and birthweight to NEC 30-day and 1-year mortality rate(s). Our analysis showed that anatomically localised NEC disease yields improved survival outcomes compared to more regional intestinal involvement. 17 Previous reports exist to suggest that pneumoperitoneum does not adversely affect infant survival though we have found in this current study an increased odds ratio for mortality in patients demonstrating 'free air'. 18 Death at 30 days is most often attributable to the disease consequence(s) of NEC. This direct link thereafter diminishes as late mortality at 1 year which we have shown is often due to other system co-morbidities and the consequences of prematurity.
In this study, particular note should be taken of the findings we report relating to term NEC. A statistically significant correlation was observed in this current study between increasing gestation age and a younger age at emergency laparotomy likely demonstrating a more rapid fulminant progress of NEC in the term newborn. These data findings thus support previously published work by Andrews et al 19 and Ostlie et al 20 Our term NEC data set results (Table 1) also further confirm the previous recorded findings of an increased incidence of colon bowel sector involvement 19 and the added burden of cardiac surgery in the mature infant. 21 We speculate term NEC may well have a very different aetiology and pathogenesis to its pre-term disease counterpart.
This study also included babies undergoing cardiac surgery as this cohort clearly showed that significant cardiac pathology requir-

| CON CLUS ION
In summary, this study shows good survival outcomes for advanced NEC in babies managed in a single UK centre adopting a care pathway for emergency laparotomy in the setting of a collaborative neonatal network. 26 The increased mortality observed in infants during the first year of life (from 30 days to 1 year) at all gestational ages demonstrates the ongoing challenges these vulnerable infants sustain from the additional co-morbidities of prematurity and chronic illness. 3,27 Gestational age and birthweight are once again confirmed as the largest contributing factor(s) to high mortality in NEC newborns.
Extent of anatomical region of bowel sector involvement and presence of 'free air' are added variables that lend themselves to future predictive analysis.
In this study, we further uncovered the interesting observation notably a statistically significant link between gestational age of babies acquiring NEC and their time course to emergent operation, that is a rapid fulminant progress of disease with failure of conservative medical therapy(s). The study findings also demonstrate that NEC even in the term newborn additionally carries late mortality (5% at 1 year follow up). This late mortality in a term cohort group appeared linked to the consequence(s) of cardiac co-morbidities. We conclude that NEC survival metrics may be usefully predicted by exploring statistical modelling tools that may be of value in guiding future NEC collaborative multicentre studies.

CO N FLI C T O F I NTE R E S T
There are no conflicts of Interest for any authors.

E TH I C A L A PPROVA L
Ethical approval was not required for this study.