Antenatal corticosteroids and outcomes in gastroschisis: A multicenter retrospective cohort study

In gastroschisis, there is evidence to suggest that gut dysfunction develops secondary to bowel inflammation; we aimed to evaluate the effect of maternal antenatal corticosteroids administered for obstetric reasons on time to full enteral feeds in a multicenter cohort study of gastroschisis infants.

Conclusion: Maternal antenatal corticosteroids use, under current antenatal steroid protocols, in gastroschisis is not associated with an improvement in neonatal outcomes such as time to full enteral feeds or length of hospital stay.

| INTRODUCTION
The most common postnatal gastroschisis-related morbidity is intestinal dysmotility with 50% of neonates taking >28 days to attain full enteral autonomy, 1,2 which carries a risk of central line sepsis and liver dysfunction from prolonged parenteral nutrition (PN). Although improved neonatal care and the advent of PN has reduced gastroschisis mortality from 60% in the 1960s to 3% to 10% in the 1990s, 1 there have been no advances in reducing gastroschisis-related intestinal dysfunction (GRID) and the time to achieve full enteral feeds.
A study mapping the natural history of fetal gastroschisis throughout gestation showed that bowel wall inflammation develops and increases during the third trimester of pregnancy. 3 In addition, amniotic fluid collected at 37 to 38 weeks gestation from fetuses with gastroschisis showed evidence of inflammation, including an elevation in neutrophil and mononuclear cell counts, and raised interleukin-8 concentration compared to age matched normal controls. 4 It has therefore been hypothesized that GRID is secondary to bowel wall inflammation during the third trimester. Animal models of gastroschisis (rabbit, rat and chick) treated with in utero dexamethasone injections have shown reduced bowel wall thickness, increased glucose uptake and increased bowel protein, DNA and enzyme content 5-7 suggesting a beneficial effect of dexamethasone on intestinal function. In humans, data are very limited but a single center prospective study of gastroschisis cases reported that maternal antenatal corticosteroids reduced duration of PN, time to full enteral feeds and length of hospital stay (Polnik et al. [unpublished] presentation at the Joint European Paediatric Surgeons' Association/British Association of Paediatric Surgeons Congress, Rome). Finally, a study of duodenal contractility, in otherwise normal premature neonates, showed that antenatal corticosteroids were associated with improved duodenal contractility 8 and antenatal corticosteroids are also known to be associated with lower rates of necrotizing enterocolitis in premature infants, 9 suggesting that corticosteroids may have a maturational effect on the intestine.
We aimed to determine whether or not maternal antenatal corticosteroids were associated with an improvement in postnatal intestinal function (measured as time to full enteral feeds) in infants with gastroschisis.

| METHODS
We performed an international, multicenter retrospective cohort study of infants with gastroschisis who were managed at the following linked fetal and pediatric surgical centers: Bell's stage 2 or more and sepsis (defined as isolation of a pathogenic organism from blood).
The data were described as median (range) and analyzed using Mann-Whitney for univariate analysis and Cox regression or binary logistic regression for multivariate analysis. Those infants who failed to achieve full enteral feeds (death, long term PN dependence, transfer to another hospital on PN) remained in the Cox regression analysis but were censored at the point of discharge/ death. We identified complex gastroschisis, birth GA and source hospital as potential confounders and adjusted for these variables in our regression models alongside maternal antenatal corticosteroid administration. Statistical analysis was performed using Gra-phPad Prism v6 and SPSS v22 software. A P-value of <.05 was considered statistically significant.
Of the 69 infants who received maternal antenatal corticosteroids, 56 received a full course (total of 24 mg of betamethasone or dexamethasone in divided IM doses) and 13 a partial course (less than 24 mg of betamethasone or dexamethasone in divided IM doses). The administration of the corticosteroids varied across the group with two infants receiving corticosteroids <24 hours before birth, 42 infants between >24 hours and ≤7 days before birth, 23 infants at >7 days before birth and for two infants the time of administration was unknown. Maternal antenatal corticosteroids were administered for the following reasons; elective caesarean section delivery at 36 weeks GA (n = 3), premature labor (n = 15), threatened premature labor (n = 18), planned delivery at <37 weeks for bowel dilatation (n = 20), premature delivery for obstetric reasons (n = 10) or unknown (n = 3).
As expected, infants whose mothers had received corticosteroids were born at a significantly lower GA and birth weight compared to those whose mothers had not received corticosteroids (Table 1).

| Univariate analysis
A significantly higher proportion of "complex" patients (22%) had received maternal steroids compared with "simple" patients (12%, P = .03). Table 2 indicates how many simple and complex patients had Gastroschisis study population and demographics

| DISCUSSION
Our analysis of linked maternal and neonatal data from 500 cases of gastroschisis indicated that maternal antenatal corticosteroid administration was not associated with any effect on time to full enteral feeds, or length of hospital stay. In keeping with the published literature, we identified that both complex gastroschisis 1,11 and lower GA at birth [12][13][14] were associated with a significant increase in the time to full enteral feeds and length of hospital stay.
Amniotic fluid composition significantly changes after 25 weeks gestation. 15 These changes in amniotic fluid composition coincide with the increase in bowel wall inflammation in fetuses with gastroschisis in the third trimester 3 although we do not know whether changes in amniotic fluid composition are associated with GRID. We speculate that our findings of null effect could be due to late administration of steroids in our cohort during gestation. It is possible that if administered at the start of third trimester steroids might prevent the onset of bowel inflammation and in turn the gastrointestinal damage leading to intestinal dysfunction. In addition, it is also important to note that antenatal corticosteroids are fast acting with a short half-life, 16 suggesting the effect of early corticosteroid administration would be short lived. This implies that multiple courses might be required throughout the third trimester to exert a prolonged anti-inflammatory effect. However, concerns have been raised regarding multiple courses of antenatal corticosteroids even for threatened preterm births. [17][18][19] Given this uncertainty and the lack of apparent benefit identified in our cohort, it will be difficult to conceive a trial of multiple courses of maternal antenatal corticosteroids to improve postnatal intestinal function in fetuses with gastroschisis unless more data are available from relevant animal models. Of relevance, the recent randomized controlled trial of amnioexchange to decrease amniotic fluid-derived inflammation in gastroschisis also found no benefit. 20 Although in a multicenter French study there was an apparent benefit of corticosteroid administration on number of surgeries and time on PN, 21 these data were not controlled for major confounders such as complexity and birth GA, both of which we showed to be significantly associated with time to full enteral feeds in our study.
Our data suggested that corticosteroid administration to all mothers with gastroschisis fetuses has no effect. We undertook further hypothesis-generating analyses to explore whether there might be a subset in whom there might be a benefit. None of these analyses revealed any associated differences in gastroschisis postnatal outcomes.
These data suggest that simply reducing inflammation or an intestinal maturational effect from a course of maternal antenatal corticosteroids does not prevent or reverse the intestinal changes that are responsible for postnatal GRID. It is possible that steroid administration is inadequate to prevent or reverse inflammation resulting in GRID, or that GRID is due to other causes such as mechanical factors and/or ischemic changes in the intestine.
We However, the strength of this study comes from the large number of patients from three centers in two countries, yielding data that very consistently shows no association between antenatal corticosteroid administration and improved postnatal outcomes. Although overall our data collection included 500 gastroschisis infants, only 69 pregnancies received steroids, which limits the power of our study to detect differences in time to full feeds. With this number of patients, and ratio of pregnancies in which steroids was administered to those in which they were not administered, we would have had 80% power to detect a hazard ratio of 1.5 (alpha = .05), which is close to the upper limit of the 95% confidence interval of our estimate if the allocation had been random. This was clearly not the case in our study, nevertheless, our data do not provide support for conducting a randomized controlled trial of steroid administration.
Although there was no evidence for benefit of steroids in our study, there was also none for adverse effects in the outcomes measured: time to full feed, hospital stay, ICU stay, NEC or sepsis, so our data do not suggest that current maternal steroids should protocols for prematurity should be changed when gastroschisis has been diagnosed.

| CONCLUSIONS
Maternal antenatal corticosteroids use, under current antenatal steroid protocols, in gastroschisis might not be associated with an improvement in neonatal outcomes such as time to full enteral feeds or length of hospital stay.
There is a need for further research into the pathophysiology underpinning gastroschisis-related intestinal dysfunction in order to develop either prenatal or postnatal therapies to improve intestinal function.