Enteral feeding in duct-dependent congenital heart disease

1 Introduction

Infants with duct-dependent congenital heart lesions are treated pre-operatively with a continuous prostaglandin E1 infusion to maintain patency of the ductus arteriosus, to ensure an adequate pulmonary or systemic blood flow. The feeding strategy of these patients remains controversial, and there is widespread variation in practice worldwide [1]. Data regarding the safety and risks of enteral feeding in this group are limited.

The primary concern regarding enteral feeding in prostaglandin-dependant infants is that of necrotising enterocolitis (NEC). Some studies have shown that term infants with congenital heart disease are at increased risk of NEC compared with healthy term infants [2, 3]. It is due to this risk that many units choose not to introduce enteral feeding in this patient group, although there is no high-quality evidence showing that the risk of NEC in this patient group is modified by feeding strategy.

We aimed to describe the feeding strategies and gastrointestinal complications in infants in our institution with duct-dependent congenital heart lesions, to investigate whether enteral feeding is associated with a higher risk of NEC in this patient group.

2 Methods

Using an institutional cardiology database, all patients diagnosed with hypoplastic left heart syndrome, coarctation of the aorta, pulmonary atresia, or transposition of the great arteries born between September 2013 and December 2014 were identified. Further data were gathered from the hospital records on demographic factors, clinical course, capillary blood gas data (lowest pre-operative pH, highest pre-operative lactate), prostaglandin use, and gastrointestinal complications. Necrotising enterocolitis was defined as ‘confirmed’ if classical radiographic changes were seen, or if the diagnosis was confirmed during laparotomy (i.e. Bell’s classification stage II or III). It was defined as ‘suspected’ (i.e. Bell’s classification stage I) if the medical team caring for the baby documented in the clinical notes that NEC was suspected and treated by stopping all enteral feeds (if applicable) and starting antibiotic therapy. ‘Small for gestational age’ was defined as a birth weight less than the 10th centile for gestational age using standard World Health Organisation sex-specific growth charts. Feeding data were collected for the period the infants were receiving a prostaglandin infusion. ‘Full enteral feeding’ was defined as the entire daily fluid requirement being met using enteral feeds (either human milk or formula milk), with no additional intravenous nutrition or fluids. ‘Partial enteral feeding’ was defined as some enteral feeds being administered on at least one occasion, with additional intravenous feeds or fluids to make up the total daily fluid requirement. ‘No enteral feeding’ was defined as the infant never receiving any enteral fluids or feeds for the entire period. Premature infants, those with pre-existing gastrointestinal disease, and those who never received prostaglandin were excluded. The data were analysed with summary statistics. Univariable and multivariable logistic regression models were constructed with NEC as the outcome variable, to investigate possible risk factors for the development of NEC in this patient group.

3 Results

We identified 177 term infants with congenital heart disease who received an infusion of prostaglandin over the defined period. Of these, 62 had the diagnosis of coarctation of the aorta, 23 hypoplastic left heart syndrome, 23 pulmonary atresia, and 69 transposition of the great arteries. 108 (61%) were male. 123 (69.5%) were diagnosed with congenital heart disease antenatally.

In our cohort there were 18 cases of NEC (10.1%). Of these, 16 were suspected and 2 were confirmed. All were treated with antibiotics, and one infant underwent gastrointestinal surgery, at 8 days of age.

The prostaglandin infusion was maintained for a median of 3 days (interquartile range 2–5 days), with a median age at cardiac operation of 7 days (interquartile range 4–11 days). Whilst receiving the prostaglandin infusion, 43 infants (24.4%) received at least one day of full enteral feeds. Six infants (3.4%) received at least some enteral feeds but never reached full enteral feeds. 127 infants (72.2%) did not receive any enteral feeds.

Univariable logistic regression models found no association between feeding strategy and risk of NEC (Table 1). No association was found when full enteral feeds and partial enteral feeds were considered separately, or when they were combined. In addition, none of the other factors analysed were found to have a significant association with risk of NEC (Table 1). Further univariable analyses were performed with the duct-dependent systemic circulation diagnoses (hypoplastic left heart syndrome and coarctation of the aorta) considered as a single group, and no significant associations were found. A multivariable model was also constructed to adjust for possible confounding effects, but again no significant association between NEC and any other variable was found.

We found an incidence of NEC (suspected or confirmed) in our cohort of 10.1%, which is in keeping with previous studies demonstrating that infants with congenital heart disease are at increased risk of NEC compared to healthy term infants [2]. However, our findings do not support the theory that enteral feeds increase the risk of NEC in this patient group. We found no difference in risk of NEC in infants that were either fully or partially enterally fed compared to infants that received no enteral feeds. The type of cardiac diagnosis also had no effect on NEC risk.

NEC is a neonatal disease in which in gut mucosa is damaged by a mechanism that is not fully understood, but prematurity is the most well-described risk factor [2]. Gut hypoperfusion is known to be a contributory mechanism in the development of NEC in term infants [3]. In patients with duct-dependent congenital heart disease, such as our cohort, this may either be caused by the underlying heart defect itself, or because of diastolic steal due to the patent arterial duct. It is logical therefore to assume our patient group is at a relatively high risk of NEC. What is uncertain is the impact of enteral feeding on this risk.

Most of the evidence regarding feeding and NEC risk is based on premature infants. In this group of patients, there is a clear association between feeding strategy and NEC risk [4]. We have not replicated this finding in our cohort. Furthermore, there are potential benefits to enteral feeding in our patient group. Enteral feeding has been associated with improved outcomes in both term and pre-term critically ill infants, but no study has demonstrated such benefits specifically in patients with congenital heart disease [1, 5]. If found to be safe, enteral feeding in this preoperative period would simplify management as well as contributing to normal gut flora and physiology.

In our patients the interval to surgery was relatively short however this may not be the case in all centres. The use of TPN in centres where longer waiting periods are necessary are not without cost nor risk of complications.

Our study is limited by its retrospective design, and relatively small sample size. Although NEC is more common in patients with congenital heart disease, it still only occurs in a minority of those with an already rare disease, making it difficult to study a large number of patients in a single centre. In view of the potential benefits of enteral feeding, we feel our data would justify the use of enteral feeds in infants receiving prostaglandin infusions. However, a randomised controlled trial would be the next step to improve the evidence base for this practice.

Disclosure statements

All authors have no actual of potential interests to declare.

The authors affirm that the research involving human subjects submitted to the journal was conducted in accordance with the ethical standards of all applicable national and institutional committees and the World Medical Association’s Helsinki Declaration.

Funding

No funding was received.