Although the survival rates of preterm newborns have increased significantly in the past 20 years, the incidence of cerebral palsy (CP) in children has remained stable through time, probably due to the increased survival of extremely preterm NBs and the associated morbidity and neurosensory impairment.1,2

The main risk factor for CP in children is preterm birth, and 35% of cases occur in children born at less than 34 weeks’ gestation.3 Cerebral palsy is 70 times more frequent in children born at less than 28 weeks and 40 times more frequent in children born at 28–32 weeks compared to children born to term.4–7 The risk is inversely proportional to gestational age.

A Cochrane systematic review8 published in 2009 concluded that administration of magnesium sulphate (MgSO4) to women at risk of preterm birth, regardless of its purpose or route of administration, significantly reduces the incidence of cerebral palsy and severe motor dysfunction at age 2 years in children born before 32 weeks’ gestation.

Two metaanalyses9,10 published in 2009 suggest that the number needed to treat to prevent one case of CP is 63 pregnant women.

However, few studies have assessed the initial outcomes of newborns exposed to MgSO4 before birth or described their clinical course during admission in the neonatal period. Studies in the literature have reported an increased incidence of hypotonia, intubation in the delivery room, a higher need for mechanical ventilation and patent ductus arteriosus, among other respiratory, haemodynamic,11 neurologic and gastrointestinal7,12 adverse effects.

On the other hand, there is currently no consensus regarding the dose, standard regimen, therapeutic window and safety of the use of MgSO4 for neuroprotection. The cumulative dose of magnesium received by the mother that is safe to the newborn is not known, as its association with the serum levels of magnesium in the newborn has not been sufficiently studied.13 Thus, although there is a therapeutic window in which the neuroprotective effects of magnesium are observed, it is known that neonatal morbidity and mortality increase after serum magnesium exceeds certain levels in the newborn.14

The main aim of our study was to assess the initial clinical outcomes of patients admitted to a tertiary level neonatal intensive care unit (NICU) that received prenatal MgSO4 for neuroprotection, to analyse the potential side effects of its administration, and last of all to determine the association between the dose of magnesium given to the mother and the serum levels of magnesium in the newborn.

We conducted a prospective cohort study from December 2012 to July 2015. We included newborns delivered at less than 29 weeks’ gestation to mothers that were given MgSO4 for neuroprotection admitted to the NICU of the Complejo Hospitalario Universitario de Vigo. We compared these preterm newborns to a control group of the same gestational age born during the same period. We excluded newborns with other risk factors leading to immediately poor outcomes, which in our series included one patient with polymalformation syndrome and one that died a few hours after birth following placental abruption. We chose newborns of less than 29 weeks of gestational age because this group is more likely to require advanced resuscitation measures in the delivery room or respiratory support with invasive mechanical ventilation, and is at higher riskof neurologic sequelae.1,5

We collected data for the following perinatal variables: gestational age, weight, small for gestational age (defined as birth weight below the 10th percentile), sex, twin pregnancy, caesarean delivery, exposure to prenatal corticosteroids, chorioamnionitis (diagnosed by histological examination), preeclampsia, gestational diabetes and timing of rupture of membranes.

We analysed the need in the first hour of life of advanced or intensive cardiopulmonary resuscitation (CPR) (defined as any of the following: intubation,chest compressions, administration of epinephrine), the Apgar score at 1 and 5min, and the umbilical artery pH.

We studied a series of haemodynamic, respiratory, neurologic, gastrointestinal, ophthalmologic and infectious variables.

The haemodynamic analysis included the sonographic assessment for patent ductus arteriosus, the need for volume expansion and the administration of inotropic agents for low blood pressure.

To assess the impact of MgSO4 from a respiratory perspective, we took into account the need for invasive mechanical ventilation, surfactant administration, the total duration of oxygen therapy (in days), the presence of bronchopulmonary dysplasia (BPD) and its severity (based on the definition given by the Standards Committee of the Sociedad Española de Neonatología [Spanish Society of Neonatology], 2013)15 and administration of intravenous corticosteroids for BPD.

The study of gastrointestinal variables included the presence of necrotising enterocolitis (NEC),16 the day since birth at which enteral feedings were initiated, and the day when exclusive enteral feeding without need for any type of parenteral supplementation was achieved.

We also analysed neurologic outcomes (development of intraventricular haemorrhage [HIV] and periventricularleukomalacia), ophthalmologic outcomes (retinopathy of prematurity), infectious outcomes (nosocomial sepsis, defined as elevated acute phase reactants combined with clinical manifestations of sepsis) and mortality.

In our hospital, the MgSO4 regimen used for neuroprotection is the same that has been used for the treatment of preeclampsia and/or prevention of eclampsia.13 Treatment with MgSO4 is initiated with a loading dose of 4g delivered over half an hour and continues with infusion of 1g/hour until delivery. Administration of MgSO4 should be discontinued if delivery does not occur within 24h of initiating infusion. If the risk of imminent preterm birth recurs after discontinuation, treatment with MgSO4 should be resumed. If less than 6h have elapsed from discontinuation, treatment should resume with the maintenance dose, and otherwise should be reinitiated with a loading dose.

To study the association between the cumulative dose of magnesium received by mothers and the levels of magnesium in the newborn, we measured these levels in the first 24h from birth in all preterm newborns whose mothers had received MgSO4 by adding this parameter to blood tests performed for other reasons.

We analysed the data with SPSS version 21. We used the Kolmogorov–Smirnov test to analyse distribution of the variables, and based the analysis of quantitative variables on its results, using Student's t test and ANOVA for normally distributed variables and the Mann–Whitney U test for nonparametric variables. We analysed differences between nominal variables by means of the chi square test. We conducted linear regression analysis to determine the statistically significant correlations between qualitative variables. We considered differences to be statistically significant when the p-value obtained in any of the tests mentioned above was less than 0.05.

The total sample consisted of 42 newborns of less than 29 weeks of gestational age, of whom 28 were exposed to MgSO4 prenatally, and 14 were not.

Both groups had similar perinatal characteristics, which allowed the comparison of the treatment and control groups. The percentage of caesarean deliveries (85.7%) and twin pregnancies (46.4%) was higher in the group treated with MgSO4 (Table 1), and these differences were statistically significant.

We did not find any differences in the outcomes of the two groups in terms of need for advanced CPR, Apgar score, umbilical artery pH, respiratory variables, neurologic variables, retinopathy (grade >2),and nosocomial sepsis (Table 2). However, the mortality was greater in the group that was not treated with MgSO4 (21.4%), a difference that was statistically significant. Furthermore, while the incidence of BPD was similar in both groups, the subset of patients born before 29 weeks’ gestation and not treated with MgSO4 had a higher incidence of moderate-to-severe BPD (five cases of moderate-severe BPD compared to 2).

All newborns and their mothers were included in the analysis of the association between maternal magnesium dose and neonatal magnesium levels. In the group treated with MgSO4, the prenatal exposure was of 4g in eight newborns, more than 8g in twelve newborns, and between 4 and 8g in the rest. The mean cumulative dose of MgSO4 given to mothers was 10.09g, with a median of 7.5g.14 The mean serum levels of magnesium in exposed newborns was 2.68mg/dL (median, 2.65mg/dL), which differed significantly from magnesium levels measured the first day of life in newborns that were not treated with MgSO4 (Table 3). We found a statistically significant linear correlation between the MgSO4 dose received by mothers and the magnesium serum levels measured in newborns in the first 24h of life (r2=0.436; p<0.001). In our series, the highest serum level of magnesium was 4.04mg/dL and corresponded to a newborn whose mother had received 20g of MgSO4.

Using the cumulative dose of magnesium received by the mother and the serum levels of magnesium in the newborn (Table 3), we conducted an analysis to assess the correlation between birth weight and serum magnesium levels, but found no statistically significant differences.

We also compared the levels of magnesium in preterm newborns whose mothers had only received the loading bolus with 4g of MgSO4 with the levels in newborns of mothers that were not treated with MgSO4. The group with mothers that received the loading dose of 4g of MgSO4 had a mean serum level of magnesium of 2.39mg/dL (median, 2.3mg/dL), while the group that was not exposed to magnesium had a mean serum level of magnesium of 1.89mg/dL (median, 1.86mg/dL), a difference that was statistically significant.

When it came to the initial outcomes in the delivery room, we found no significant differences in the need for advanced CPR, Apgar score or umbilical artery pH. These results were consistent with those of other cohort studies that already supported the safety of the use of MgSO4 under these circumstances.17

From a respiratory perspective, there were no differences in the need for invasive mechanical ventilation, use of surfactant and total days of oxygen therapy, so we could not attribute any respiratory side effects to MgSO4, contrary to what has been assumed historically, as treatment with MgSO4 was associated with an increased incidence of respiratory depression.18 Bronchopulmonary dysplasia was less severe in the treated group.

From a haemodynamic perspective, the absence of adverse effects was also consistent with the finding of other studies, as we did not observe an increased prevalence of patent ductus arteriosus or increased use of inotropic agents.19 We did not find differences in the remaining outcome variablesduring the stay in the NICU, and we observed that, at the neurologic level, the incidence of IVH and leukomalacia were similar in both groups. In this regard, it would be interesting to conduct a long-term followup of these patients to study their psychomotor development. We also found no differences at the gastrointestinal level in either the development of NEC, initiation of enteral nutrition or the number of days post birth at which full enteral nutrition was achieved. Furthermore, we observed similar outcomes in relation to the development of retinopathy of prematurity from an ophthalmologic perspective, and of nosocomial sepsis from an infectious perspective.

We did find differences in mortality in the group treated with MgSO4. Once again, the small sample size requires that we interpret this result with caution and continue collecting data for future comparisons. We must keep in mind that the Cochrane systematic review of 200913 established that when MgSO4 was administered for the sole purpose of neuroprotection, there was a reduction in the composite outcome of death and cerebral palsy at age 2 years in exposed children.9,20

It would be worth exploring whether the association of MgSO4 exposure with a decrease in childhood mortality starts as early as the neonatal period. Other studies in the literature have also found a lower mortality in the group treated with MgSO4,17 although its mechanism of action remains unclear. Studies conducted in recent years have found an association between antenatal MgSO4 and decreased oxygen consumption in the newborn brain, with a decreased cerebral fractional tissue oxygenation extraction (cFTOE).21

When it came to the dose of MgSO4 administered to the mothers, we found a linear correlation between the cumulative dose and magnesium levels in newborns of less than 29 weeks of gestational age in the first 24h post birth. This finding allows the establishment of safe maternal doses to prevent the development of side effects in preterm newborns.

The mean level of magnesium in the newborns was 2.68mg/dL, below the 4.5mg/dLthreshold potentially associated with the development of side effects and poor clinical outcomes in the neonatal period.22 It is very likely that the absence of side effects from the use of MgSO4 in our series was due to the generally low doses administered to the mothers. In the group of exposed children in our study, the highest level of magnesium was 4.04mg/dL and corresponded to a boy whose mother had received 20g of MgSO4. Therefore, we could establish a safety threshold of 20g for the cumulative dose and 16h of continuous infusion, after which the dose would come dangerously close to levels associated with worse outcomes.

In eight out of the twenty-eight newborns of mothers treated with MgSO4, the mother was only given the loading bolus with 4g due to the short time elapsed to delivery. Even these newborns had a statistically significant increase in magnesium levels compared to the newborns of mothers that had not received MgSO4. For this reason, we recommend that at least the loading dose of 4g be given to mothers at risk of giving birth at less than 29 weeks’ gestation, even if the birth is imminent.

We also recommend the measurement of magnesium levels in all newborns of mothers treated with MgSO4 before birth. The reasons for this are, first, the increased mortality detected in association with serum levels of magnesium greater than 4.5mg/dL22 and, second, the adjustments that need to be made in the composition of the parenteral feeds administered to these newborns in the first days of life until their magnesium levels normalise.

In conclusion, our study found that MgSO4 was safe at the administered dose, with a linear correlation between serum levels of magnesium in the newborn in the first 24h of life and the cumulative dose of magnesium received by the mother.

Long-term follow-up studies of the patients included in this study are requiredto assess the impact of MgSO4 therapy on future neurodevelopmental outcomes.

The authors have no conflict of interests to declare.