Antenatal Magnesium Sulfate Reduced Intestinal Morbidities Requiring Surgery in Preterm Infants With Extremely Low Gestational Age: A Retrospective Cohort Study

Background Antenatal magnesium sulfate is widely used as a tocolytic, for maternal seizures, and for seizure prophylaxis in preeclampsia. Recent studies have suggested that antenatal magnesium sulfate use is associated with favorable neurodevelopmental outcomes in preterm infants. However, there are concerns regarding the neonatal effects of antenatal magnesium sulfate, especially regarding gastrointestinal morbidities. This study aims to explore the effects of antenatal magnesium sulfate on intestinal morbidities requiring surgery in preterm infants. Methods This was a retrospective cohort study of 181 preterm infants who were born at less than 28 weeks of gestational age. Subjects were categorized as infants exposed to antenatal magnesium sulfate and those not exposed to antenatal magnesium sulfate. Results Antenatal magnesium sulfate was associated with a lower incidence of surgical conditions of the intestine (OR 0.393, 95% CI 0.170–0.905). Multivariate analysis showed that the duration of antenatal magnesium sulfate use was associated with surgical conditions of the intestine (adjusted OR 0.766, 95% CI 0.589–0.997). In the < 26 weeks of gestational age subgroup, use of antenatal magnesium sulfate was signicantly associated with decreased intestinal morbidities requiring surgery (adjusted OR 0.234, 95% CI 0.060–0.922). Conclusion Antenatal magnesium sulfate use appears to have a protective effect on intestinal morbidities requiring surgery in preterm infants in a duration-dependent manner. Association of antenatal magnesium sulfate use and decreased intestinal morbidities requiring surgery was more distinct in preterm infants < 26 weeks of gestational age.


Background
Antenatal magnesium sulfate (MgSO 4 ) has been widely used as a tocolytic and for the treatment of eclampsia. [1] Since randomized control trials showed that antenatal MgSO 4 is associated with a reduced risk of CP in surviving infants without major adverse outcomes [2][3][4][5], the short-term use of MgSO 4 for fetal neuroprotection before preterm delivery has been recommended. [6,7] Recent systematic reviews concluded that the administration of intravenous MgSO 4 to mothers who are at risk of early preterm delivery reduced the risk of neurodevelopmental disability in their surviving children. [8,9] However, the US Food and Drug Administration has advised against the use of MgSO 4 injections for more than 5-7 days when intending to stop preterm labor in pregnant women because MgSO 4 has the potential to cause skeletal abnormalities in developing babies. [6] There have also been several concerns regarding the adverse effects of antenatal MgSO 4 on the neonatal outcomes of preterm infants. [10][11][12] In a small randomized controlled study, higher levels of umbilical cord ionized magnesium at delivery were associated with adverse outcomes. [11] A secondary analysis of neonates < 26 weeks of age in a multicenter randomized trial for neuroprotection showed that antenatal MgSO 4 was associated with an increased risk of death or severe necrotizing enterocolitis (NEC). [10] However, the association between antenatal MgSO 4 and morbidities of the gastrointestinal (GI) tract in preterm infants is still controversial in extremely preterm infants. [13] Therefore, our study aims to explore the effects of antenatal MgSO 4 on the neonatal outcomes of preterm infants, including intestinal morbidities requiring surgical interventions.

Methods
This was a retrospective cohort study of preterm infants who were born at less than 28 weeks of gestation and who were admitted to the Seoul National University Children's Hospital neonatal intensive care unit between January 2011 and December 2015. Infants with major congenital anomalies and maternal preeclampsia were excluded from the study population. One patient who experienced iatrogenic gastric perforation was excluded. They were categorized as infants with antenatal MgSO 4 and without antenatal MgSO 4 . Indications of antenatal MgSO 4 included preterm labor and neuroprotection. The protocol for the antenatal MgSO 4 treatment of preterm labor was a loading dose of 4 g infused over 30 minutes followed by a maintenance dose of 2 g per hour, which could be titrated according to the effect.
For neuroprotection of the fetus, a loading dose of 6 g infused over 30 minutes followed by a maintenance dose of 2 g per hour was used.
Medical records of perinatal factors, including the cumulative dose of antenatal MgSO 4 and the duration of antenatal MgSO 4 during the last 7 days of pregnancy, as well as the maternal level of serum magnesium before delivery, were reviewed and analyzed. Data on the neonatal factors and outcomes, including the neonatal blood magnesium level and intestinal injuries, such as NEC, spontaneous intestinal perforation (SIP), and meconium obstruction (MO) requiring surgery, were collected. Data on respiratory distress syndrome, patent ductus arteriosus (PDA), bronchopulmonary dysplasia (BPD), sepsis, intraventricular hemorrhage (IVH), retinopathy of prematurity (ROP), hypotension during the rst week, ibuprofen use during the rst week, postnatal steroid use, duration of invasive ventilation and time to full feeding were also collected.
Intestinal morbidity requiring surgery was considered when surgical intervention was conducted for NEC, SIP or MO. NEC was diagnosed according to Bell's staging criteria and SIP when isolated bowel perforation was present without evidence of NEC or MO [14]. MO was suspected when premature infants had problems with meconium passage accompanied by abdominal distension and feeding intolerance. It was diagnosed if persistent or progressive gaseous bowel distension was noted on plain radiography, along with the presence of hypoechoic meconium-lled bowel loops and distended proximal bowel loops observed by sonography. [15,16] Glycerin enema was used as the rst-line therapy for a few days, and ultrasonography-guided water-soluble contrast enema was performed if the condition was not resolved. If contrast enema failed to relieve the obstruction, surgical intervention was carried out.
Statistical analysis was performed with STATA version 11.0. Wilcoxon rank-sum tests for continuous data and chi square tests for categorical data were used. A P value < 0.05 was considered statistically signi cant in all analyses. Univariate logistic regression analysis was conducted for intestinal injury requiring surgery, and factors with p < 0.10 in the univariate analysis were included in multivariate analysis to evaluate the association between maternal MgSO 4 use and intestinal injury requiring surgery.
Next, duration of MgSO 4 , cumulative dose of MgSO 4 , maternal Mg level and neonatal Mg level were used in the multivariate logistic regression analysis. Subgroup analysis of those born before 26 weeks of gestation was also conducted.

Results
During the study period, 202 infants born at less than 28 weeks of gestational age (GA) were admitted to the Seoul National University Hospital neonatal intensive care unit. Four infants with major congenital anomalies and sixteen with preeclampsia were excluded from the study population. Among 181 infants, 109 (60.2%) were exposed to antenatal MgSO 4 , including 100 infants for preterm labor and 9 infants for neuroprotection ( Fig. 1

Discussion
The current study showed that antenatal MgSO 4 was associated with low incidence of intestinal morbidities requiring surgery in preterm infants with GA less than 26 weeks. In the total study population, duration of exposure to antenatal MgSO 4 was associated with decreased risk of intestinal injury needing surgery. There were no differences in mortality and other morbidities such as RDS, IVH and BPD between infants with and without antenatal MgSO 4 .
Antenatal MgSO 4 has been used widely according to maternal and fetal conditions. Since bene ts in terms of the neurodevelopmental outcomes were found in prospective randomized trials [2,3], antenatal MgSO 4 is now recommended for fetal neuroprotection in anticipation of preterm delivery in very preterm infants [17]. As the use of antenatal MgSO 4 has been highlighted, concerns have been raised regarding the neonatal adverse effects of antenatal MgSO 4 because magnesium can block calcium entry into cells, leading to muscle weakness and intestinal atony in the offspring. [10] A clinical study showed a signi cant negative relationship between the mean blood ow velocity and the time from birth to the blood ow velocity measurement in antenatal MgSO 4 -exposed infants. [18] In the secondary analysis of a prospective study, antenatal exposure to MgSO 4 was associated with death or severe NEC in infants with a GA less than 26 weeks. [2,10] Furthermore, a study of historical comparisons by Rattray et al reported that antenatal MgSO 4 was associated with SIP and death. [12] However, another secondary analysis of a prospective trial reported that newborns with higher magnesium levels did not show any adverse non-neurological effects. [19] Moreover, a meta-analysis of 5 randomized controlled studies concluded that there were no statistically signi cant differences in neonatal secondary outcomes, such as IVH, periventricular leukomalacia, NEC, RDS, ROP or chronic lung disease. [20] In the aforementioned secondary analysis study, although the combined outcome of death and severe NEC was associated with maternal antenatal MgSO 4 administration in infants who were GA less than 26 weeks, NEC alone was not associated with antenatal MgSO 4 . [10] Caution should be taken in the interpretation of the combined outcome of NEC and death because the two conditions were not competitive as death and BPD. Although Rattray et al reported the association between antenatal MgSO 4 and SIP, but SGA was not included in the analysis. [12] As the development of intestinal morbidities may be in uenced by various causes, factors such as GA, SGA and medication history should be considered and coanalyzed to investigate factors associated with these conditions. For instance, SIP was associated with intrauterine growth restriction and PDA with delayed meconium passage in VLBW infants. [21,22] Although this was a retrospective study, the cumulative dose of antenatal MgSO 4 , duration of antenatal MgSO 4 and level of magnesium were calculated and analyzed thoroughly. Factors associated with intestinal morbidities such as GA, SGA, and treated PDA were also reviewed. The results of the present study are consistent with the report from a nationwide database from North America, reporting that antenatal MgSO 4 exposure in extremely preterm neonates was associated with reduced risk of a combined outcome of death, NEC and SIP. [23] In a retrospective study from a tertiary-level NICU, cumulative dose of MgSO 4 was associated with a combined outcome of death, NEC and SIP. [24] Although a protective role of antenatal MgSO 4 has been raised, the physiologically bene cial effect of magnesium on the intestine of preterm infants is not yet fully understood. One study of grass carp showed that magnesium de ciency suppressed the growth and damaged the intestinal structural integrity of the sh. [25] A preclinical study of mice demonstrated that prophylactic oral administration of magnesium ameliorates induced colitis through the inhibition of colonic mast cell activation. [26] This was a single-center retrospective design, and the indications for antenatal MgSO 4 included not only neuroprotection but also preterm labor. However, the dosages of the antenatal MgSO 4 treatments for preterm labor and for neuroprotection were similar in our institution, and the data regarding duration, cumulative dosage and level of serum magnesium were reviewed. Mothers with preeclampsia, an important indication for antenatal MgSO 4 , were excluded because the maternal environment in this condition can affect vascular development of the intestine. [27,28] Conclusions This retrospective cohort study suggested the bene t of antenatal MgSO 4 in the reduction of intestinal injury requiring surgical intervention in extremely preterm infants in a duration-dependent manner, especially in those with a GA less than 26 weeks at birth. Further study with a large population might be