Keywords neuroprotection - preterm birth - magnesium sulfate - evidence - umbrella review - PRISMA - AMSTAR score
Introduction/Background
Epidemiology of premature births and cerebral palsy
Globally, more than 1 in 10 newborn is born prematurely [1 ]
[2 ]. In Germany, the
number of premature infants is over 8%, with 1.5% of all children being born as extremely or very premature infants before 32+0 weeks of pregnancy [3 ]
[4 ]
[5 ]. Premature infants have a significantly higher
morbidity and mortality risk compared to full-term infants. Among other things, the probability of premature infants suffering from cerebral palsy is increased compared to children born
full-term [6 ]. The prevalence of cerebral palsy depends on gestational age. For children with a gestational age > 36+0 weeks of
pregnancy, the literature gives the prevalence as 1.35/1000, for a gestational age between 32+0 and 36+0 weeks of pregnancy it is 6.75/1000, for premature infants between 28+0 and 31+0 weeks
of pregnancy 43.15/1000, and before 28+0 weeks of pregnancy it is even 82.25/1000 [6 ]. In the United States, cerebral palsy is diagnosed
annually in 1 in 345 children [7 ], in Germany 3042 patients were treated in 2016 for the ICD-10 main diagnosis G80 (infantile cerebral
palsy) [8 ], the prevalence of cerebral palsy in premature infants is tending to decrease (with the exception of extremely premature
infants) in Europe [9 ]. In countries with a lower standard of living (developing and emerging countries) there is a higher prevalence
than in industrialized countries [10 ]
[11 ].
Definition and etiology of cerebral palsy
Cerebral palsy describes a group of differently pronounced symptoms, which always include movement disorders and spasticity and originate in the brain. The clinical manifestation of
cerebral palsy differs in the type of movement disorder, the degree of functional abilities and limitations, and the body parts affected. In 10% of cases, the cause of cerebral palsy is
impaired brain development. 90% result from a damaging effect on the brain just before, during or just after birth, which results in lesions of healthy brain tissue. Because the area of the
periventricular white matter is a highly active proliferative zone in the immature brain, this area is particularly prone to injury in premature infants. In addition, development-related
metabolic and molecular factors increase the susceptibility of the periventricular white matter in the prematurely born child [12 ].
Disorders of the metabolism in the brain caused by ischemia, inflammatory processes, or oxygen-deficient states, which can often occur in premature infants, trigger a cascade that
ultimately leads to apoptosis or necrosis of glial cells and neurons [13 ]. Due to the reduced supply, the function of the adenosine
triphosphate (ATP)-dependent processes decreases, resulting in a loss of membrane potential. As a result, voltage-gated calcium channels open. The high intracellular level of calcium causes
glutamate to be released and accumulate in the extracellular space, thereby activating glutaminergic n-methyl-D-aspartate channels (NMDA channels). Additional calcium flows into the cell
through the activated NMDA channels. The excess calcium activates proteolytic mechanisms and produces free oxygen radicals, which activates inflammatory and apoptotic automatisms. Since
magnesium blocks these NMDA receptors, the administration of magnesium can prevent tissue destruction through excitotoxicity and the resulting apoptosis [12 ]
[14 ]
[15 ]. Since cerebral
palsy is irreversible damage to the brain, which has a significant impact on the lives of the children affected, the prevention of cerebral palsy is particularly important.
Objective and research question
In 2009, a Cochrane review on the topic “Magnesium sulphate for women at risk of preterm birth for neuroprotection of the fetus” was published [16 ]. This review summarized the state of research at the time. The aim of this paper is to develop an update on the current state of knowledge on antenatal magnesium
sulfate administration in the event of imminent premature birth, i.e. to determine whether new findings have emerged since the Cochrane review was published and what the current evidence
base for this measure is. Therefore, the state of knowledge of all systematic reviews and meta-analyses published since 2009 should be summarized and evaluated. The research question was
developed using the PICO model (see [Fig. 1 ]). The aim was to determine the neuroprotective effects of magnesium sulfate in premature infants. The
occurrence of infantile cerebral palsy is of particular interest. The magnesium sulfate is administered antenatally to the mother in the event of imminent premature birth. This leads to the
following research question: Does antenatal administration of magnesium sulfate to the mother in the event of imminent premature birth reduce the occurrence of infantile cerebral palsy in
the child?
Fig. 1 PICO model.
Materials and Methods
Study design, search strategy, and selection of studies
A systematic literature search was attempted in order to answer the research question. On 21 February 2022 (search update on 13 November 2022), an electronic database search was carried out
in PubMed. The search was structured using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) statement [17 ] (see [Fig. 2 ]). Since the study was designed as an umbrella review, only systematic reviews and meta-analyses published
after the publication of the Cochrane review by Doyle et al. (2009) [16 ] were considered. The aim of umbrella reviews is to compile the
results of existing systematic reviews and to evaluate their results with regard to their evidence [18 ]
[19 ]. Primary literature in the form of randomized controlled trials (RCT) on the issue was reviewed, but not taken into account in the
scientific evaluation.
Fig. 2 Flowchart for systematic literature research according to the PRISMA statement (Moher D, Liberati A, Tetzlaff J, Altman DG, the PRISMA Group. Preferred Reporting Items for Systematic
Reviews and Meta-analyses: The PRISMA Statement. PLoS Med 2009; 6(7): e1000097 doi:10.1371/journal.pmed1000097 ).
The search was carried out using previously defined inclusion and exclusion criteria as a collaboration between the two first authors. The results were merged in the next step.
Search term
The search terms “neuroprotection”, “cerebral palsy”, “preterm birth” and “magnesium sulfate” were combined (for complete search term see Supplementary Material). The filters
“Meta-Analysis”, “Review”, and “Systematic Review” and Meshterms were applied. In addition, only publications after 2009 were considered in order to expand the research state of the Cochrane
review [16 ].
Inclusion and exclusion criteria
In order to obtain an overview of the current state of research on the subject of neuroprotection through antenatal magnesium administration, systematic reviews and meta-analyses were
included that show a systematic literature search and were published after the Cochrane review (“Magnesium sulphate for women at risk of premature birth for neuroprotection of the fetus”) in
2009 [16 ]. In addition, the studies included in the reviews and meta-analyses should meet the criteria of a randomized controlled
trial. The participants should be women with an acute risk of premature birth (< 37+0 weeks of pregnancy). Further inclusion criteria were defined as follows:
Neuroprotection through magnesium administration
Infantile cerebral palsy as the primary outcome,
Studies conducted on humans,
Language of publication German or English.
Data extraction
First, the methodological characteristics of the reviews and meta-analyses considered, as well as the RCTs included there, were analyzed and presented. These included the type of study, the
country of publication, the dosage of magnesium sulfate, the number and weeks of pregnancy of the subjects, the study objective, the primary outcomes, and the inclusion/exclusion criteria.
In addition, the results of the studies for the primary outcome “occurrence of cerebral palsy” were presented. The relative risk (RR) or odds ratio (OR) (95% confidence interval) was
considered as a measure of the results of the included reviews and meta-analyses.
Assessment of evidence
The AMSTAR score [20 ] was used to systematically evaluate the evidence of the included studies.
Results
Results of the systematic literature search
The literature search in PubMed yielded 43 results. Based on the title, 21 studies could be excluded. Criteria for this were that the publications do not deal with neuroprotection using
magnesium sulfate, the development of infantile cerebral palsy was generally treated, neuroprotection was observed in children at term, or risk factors for the negative neurological
development in premature infants or very low-birthweight children were the focus. After reviewing the abstracts, another eight studies could be excluded. The reasons for this were
duplications and publications in French and Danish, the use of substances other than magnesium sulfate for neuroprotection and the description of biochemical processes in the development of
cerebral palsy. The full text of the 14 remaining publications were assessed, of which a further 10 were excluded due to language, a missing or unspecified search strategy, or an
inappropriate study design. Four articles were ultimately included in this review [16 ]
[21 ]
[22 ]
[23 ]. This process is shown in [Fig. 2 ], a representation of the PRISMA statement [17 ]. One of these articles presented the
Cochrane review from 2009 [16 ], which was intended to serve as a comparative study to determine whether the state of research has
changed since then.
General characteristics of the included studies
The four publications are two systematic reviews with meta-analysis, one systematic review and one individual participant data meta-analysis from Denmark, the USA, Germany, and
Australia/New Zealand. Some of the articles overlap completely in the included studies. The studies by Mittendorf et al. (2002) [24 ],
Crowther et al. (2003) [25 ], Marret et al. (2007) [26 ], and Rouse et al. (2008)
[27 ] were cited in each of the four publications and their data were used in the meta-analyses. Duley et al. (2007) [28 ] is cited in three of the four papers. Wolf et al. (2020) [29 ] is only referenced
in Wolf et al. (2020) [21 ]. All included studies are RCTs with the aim of investigating neuroprotection in premature infants with
magnesium sulfate. The exceptions here are the publication by Marret et al. (2007) [26 ], which also integrates a tocolysis group, as
well as the publication by Duley et al. (2007) [28 ], which primarily deals with the topic of preeclampsia. All six studies use a
different dosage of magnesium sulfate. Administration always begins with a bolus (4–6 g MgSO4 ) followed by a maintenance dose of 0–3 g/h. In the four reviews or meta-analyses, the
number of subjects varied from n = 4025 to n = 6178. The aim of the publications is to examine the efficacy and safety of antenatal magnesium sulfate administration for neuroprotection in
the event of imminent premature birth. Crowther et al. (2017) [23 ] examine this issue with regard to different effects of
neuroprotection depending on the cause of the premature birth, the initial reason for the magnesium sulfate administration, the gestational age, the dosage, and the timing of the
administration. The occurrence of infantile cerebral palsy is the primary outcome in each of the publications. In addition, mortality, other neurological impairments in the child, and
negative maternal events are primary outcomes. The general characteristics of the reviews or meta-analyses considered and the RCTs included in these are presented in detail in [Table 1 ].
Table 1
Methodical characteristics of the included reviews and meta-analyses and the RCTs considered there, as well as primary outcomes of the included
reviews and meta-analyses.
Study
Study type/type of review
Country and intervention of each included study
MgSO4 administration/dosage
Characteristics of the subjects
Study objective
PICO model
Endpoints/outcomes
Inclusion/exclusion criteria
* Only 1593 women who meet the criteria for a premature birth are considered in the meta-analysis.
Wolf et al. 2020 [21 ]
Systematic review, meta-analysis
Mittendorf 2002 [24 ], United States
4 g bolus, 0 or 2–3 g/h
149 women, 25–33 weeks of pregnancy
Update of the Cochrane review [2 ] including own study, which increases the number of subjects in a meta-analysis by 10%
P = premature birth
I = antenatal MgSO4 administration
C = premature birth without antenatal MgSO4 administration
O = Risk for CP
Primary outcome: Occurrence of CP
Studies that compare i.v. treatment with MgSO4 and with placebo in imminent premature birth
Study outcomes must include: Neonatal mortality, neurological outcomes with follow-up ≥ 12 months
Crowther 2003 [25 ], Australia/New Zealand
4 g bolus, 2 g/h
1062 women, < 30 weeks of pregnancy
Duley 2006 [28 ], 33 countries on 6 continents
4 g bolus, 1 g/h
10141 women, all weeks*
Marret 2006 [26 ], France
4 g bolus, 0 g/h
573 women, < 33 weeks of pregnancy
Rouse 2008 [27 ], United States
6 g bolus, 2 g/h
2241 women, 24–31 weeks of pregnancy
Wolf 2020 [29 ], Denmark
5 g bolus, 1 g/h
560 women, 24–31 weeks of pregnancy
Doyle et al. 2009 [16 ]
Systematic review, meta-analysis
Mittendorf 2002 [24 ], United States
4 g bolus, 0 or 2–3 g/h
149 women, 25–33 weeks of pregnancy
Examining the efficacy and safety of MgSO4 administration as a neuroprotective agent for women at risk of premature birth using the best available evidence
P = pregnancy; adults 19–44 years; premature birth; mean age 45–64 years; young adults 19–24 years; adolescents 13–18 years
I = intramuscularly; intravenously; magnesium sulfate; orally
C = placebo
O = fetal death; motor disorders; neurological impairment; hearing impairment; blindness; respiratory arrest; cardiac arrest; death; death in newborn; cerebral palsy;
adverse event
Primary outcomes child: Fetal, neonatal and subsequent death, neurological impairments/disabilities (such as cerebral palsy, blindness, deafness, gross motor impairment),
infantile mortality combined with cerebral palsy, combined with gross motor impairment, combined with neurological impairment, combined with neurological impairment
Maternal outcomes: Maternal death, respiratory arrest, cardiac arrest, discontinuation of treatment
Randomized trials that compare treatment with MgSO4 and other therapy (e.g. placebo) in the case of imminent premature birth and investigate at least one of the desired
outcomes
Crowther 2003 [25 ], Australia/New Zealand
4 g bolus, 2 g/h
1062 women, < 30 weeks of pregnancy
Duley 2006 [28 ], 33 countries on 6 continents
4 g bolus, 1 g/h
10141 women, all weeks*
Marret 2006 [26 ], France
4 g bolus, 0 g/h
573 women, < 33 weeks of pregnancy
Rouse 2008 [27 ], United States
6 g bolus, 2 g/h
2241 women, 24–31 weeks of pregnancy
Jacquemyn et al. 2015 [22 ]
Systematic review
Mittendorf 2002 [24 ], United States
4 g bolus, 0 or 2–3 g/h
149 women, 25–33 weeks of pregnancy
An evaluation of the effects and side effects of antenatal i.v. MgSO4 administration for women without preeclampsia and with imminent premature birth
Presentation of possible biological mechanisms of action
P = premature birth
I = antenatal, intravenous MgSO4
C = premature birth without antenatal, intravenous, MgSO4 administration
O = Risk for CP
Primary outcome: Fetal or neonatal neurological outcomes
Randomized controlled trials and meta-analyses with fetal or neonatal neurological outcomes as primary or secondary outcome with intravenous MgSO4 administration in
imminent premature birth
Crowther 2003 [25 ], Australia/New Zealand
4 g bolus, 2 g/h
1062 women, < 30 weeks of pregnancy
Marret 2006 [26 ], France
4 g bolus, 0 g/h
573 women, < 33 weeks of pregnancy
Rouse 2008 [27 ], United States
6 g bolus, 2 g/h
2241 women, 24–31 weeks of pregnancy
Crowther et al. 2017 [23 ]
Individual Participant Data (IPD) meta-analysis
Mittendorf 2002 [24 ], United States
4 g bolus, 0 or 2–3 g/h
149 women, 25–33 weeks of pregnancy
Evaluation of the effects of antenatal MgSO4 administration for women with imminent premature birth and how the effects differ with different causes of imminent
premature birth, initial reason for MgSO4 administration, gestational age, dosage, and timing of MgSO4 administration
P = premature birth (< 37 weeks)
I = antenatal or prenatal MgSO4 administration
C = premature birth without antenatal MgSO4 administration
O = risk for CP or neuroprotection
Primary outcomes child: Mortality, cerebral palsy
Primary outcome mother: Serious maternal complication
Randomized trials examining neuroprotection by MgSO4 in imminent premature birth (< 37 weeks of pregnancy) with neurological outcomes as endpoints
Crowther 2003 [25 ], Australia/ New Zealand
4 g bolus, 2 g/h
1062 women, < 30 weeks of pregnancy
Duley 2006 [28 ], 33 countries on 6 continents
4 g bolus, 1 g/h
10141 women, all weeks*
Marret 2006 [26 ], France
4 g bolus, 0 g/h
573 women, < 33 weeks of pregnancy
Rouse 2008 [27 ], United States
6 g bolus, 2 g/h
2241 women, 24–31 weeks of pregnancy
Results of the included studies with regard to the defined outcomes
The three meta-analyses included refer to a significant risk reduction when using magnesium sulfate in imminent premature birth with comparable outcome measures (RR 0.68, 95% CI 0.54–0.85
[21 ], RR 0.68, 95% CI 0.54–0.87) [16 ]
[23 ]. This applies to any gestational age. The risk of moderate to severe cerebral palsy is also reduced in the magnesium sulfate group compared to the placebo group. The
included review states that the five randomized controlled trials and four meta-analyses considered indicate a significant 32% reduction in cerebral palsy with the administration of
magnesium sulfate in premature births [22 ]. An review of the results can be found in [Table 2 ].
Table 2
Key results of the included studies.
Study
Endpoints
Results/conclusion
Heterogeneity
Wolf et al. 2020 [21 ]
Occurrence of cerebral palsy
Risk of cerebral palsy significantly reduced in the MgSO4 group compared to the placebo group (3.4% vs. 5.0%, RR 0.68, 95% CI 0.54–0.85)
Risk of moderate to severe cerebral palsy lower in the MgSO4 group compared to the placebo group (2.1% vs. 3.2%, RR 0.63, 95% CI 0.45–0.89)
MgSO4 has the same preventive effect for all gestational ages.
I2 = 0%, not statistically significant
Doyle et al. 2009 [16 ]
Fetal, neonatal, subsequent death; neurological impairments (including cerebral palsy); severe neurological disability; severe maternal side effects
MgSO4 reduces the risk of developing cerebral palsy (RR 0.68; 95% CI: 0.54–0.87; 5 studies; 6145 newborn)
More moderate to severe cerebral palsy occurred in the placebo group compared to the MgSO4 group (RR 0.64; 95% CI 0.44 to 0.92; 3 studies; 4387 neonates).
for CP: I2 = 0%-25%
Jacquemyn et al. 2015 [22 ]
Neuroprotection of the newborn
significant 32% reduction in cerebral palsy
not calculated
Crowther et al. 2017 [23 ]
Child’s death or CP, severe maternal side effects
The risk of developing cerebral palsy is reduced in the MgSO4 group compared to the placebo group (RR 0.68, 95% CI 0.54–0.87, 4601 neonates, 5 studies).
Reduction of moderate and severe CP (RR 0.63, 95% CI 0.44–0.90), as well as only severe CP (RR 0.54, 95% CI 0.30–0.94) in the MgSO4 group compared to the placebo
group.
I2 not stated, but indication that heterogeneity is not statistically significant
Evidence assessment of the included studies
Three of the four included reviews [16 ]
[21 ]
[23 ] are of high quality according to the AMSTAR rating, resulting in a high level of evidence. In one review, seven out of 11 criteria must
be answered with unclear/no/not applicable. The quality can therefore be classified as rather low [22 ]. The results of this review are
therefore of low evidence. [Table 3 ] shows the detailed evidence assessment according to the AMSTAR score.
Table 3
Evidence assessment of the included studies according to the AMSTAR score.
Study
A priori design
Study selection and data extraction by two independent persons
Systematic literature search
Unpublished and gray literature included
References of included and excluded literature
Characteristics of the included studies specified
Quality of studies assessed according to risk of bias
Adequate consideration of the risk of bias in the interpretation of the results
Appropriate methodology/statistically adequate evaluation
Potential publication bias assessed
Disclosure of conflicts of interest
Wolf et al. 2020 [21 ]
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Doyle et al. 2010 [16 ]
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Uncertain
Yes
Jacquemyn et al. 2015 [22 ]
Yes
Uncertain
Yes
Uncertain
No
No
No
Yes
Not applicable
No
Yes
Crowther et al. 2017 [23 ]
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Uncertain
Yes
Discussion
Summary of results
A total of two systematic reviews with meta-analysis [16 ]
[21 ], one systematic
review [22 ], and one individual participant data meta-analysis [23 ] were
included in this paper. These deal with the question to what extent premature infants benefit from neuroprotection by magnesium sulfate compared to a placebo group. The occurrence of
infantile cerebral palsy was chosen as the primary outcome. The reviews/meta-analyses came to a clear conclusion: In the magnesium sulfate group, the risk of suffering from cerebral palsy
was significantly reduced compared to the placebo group. According to the AMSTAR score, three of the four publications are of high quality, which means that the results can be concluded to
be highly evident.
Side effects in the child
Jaquemy et al. (2015) [22 ] describe the potential negative outcomes for the child after magnesium sulfate therapy: Hypermagnesemia
theoretically poses a risk of respiratory arrest, hypotension, reduced or absent peripheral reflexes, and stupor or coma to the newborn. Such effects on the newborn have been described after
extremely high doses of magnesium sulfate to prevent eclampsia [22 ]. In the Mittendorf study [24 ], administration of high doses of magnesium sulfate for long-term tocolysis or long-term preeclampsia treatment was shown to result in higher infant mortality, more
brain hemorrhage, and more brain lesions. Doses of more than 50 g magnesium are critical for the newborn [22 ]. However, with an
administration schedule of 4 g bolus and 1 g/h maintenance dose over 12 h and repeated if necessary, this critical value is not even remotely reached. Neonatal hypotension was also only
observed after administration of high doses of magnesium sulfate [22 ]. Wolf et al. (2020) [21 ], Doyle et al. (2009) [16 ], and Crowther et al. (2017) [23 ]
were able to show in their meta-analyses that the secondary neonatal outcomes show no signs of damage from antenatal magnesium sulfate administration. Doyle et al. (2009) [16 ] found no significant statistical differences between groups for intraventricular hemorrhage (RR 1.01; 95% CI 0.87–1.18), cystic
periventricular leukomalacia (RR 0.99; 95% CI 0.68–1.45), Apgar 5′ < 7 (RR 1.12; 95% CI 0.89–1.40), neonatal seizures (RR 0.77; 95% CI 0.49–1.21) or sustained respiratory support (RR
0.99; 95% CI 0.89–1.11). Crowther et al. (2017) [23 ] also found no significant differences between groups for the neonatal morbidity
such as Apgar 5‘ < 7, active resuscitation at birth, intraventricular hemorrhage, cystic periventricular leukomalacia, neonatal convulsions, neonatal encephalopathy, chronic lung
diseases, NEC, patent ductus arteriosus, and retinopathy. Likewise, Wolf et al. (2020) [21 ] found no negative side effects for the
newborn. Rather, positive effects could be observed: The risk of a major intraventricular brain hemorrhage was borderline significantly reduced in the magnesium sulfate group (RR 0.77; 95%
CI 0.60–1.00). The risk of intubation, cardiopulmonary resuscitation, or endotracheal intubation being necessary during the initial steps of newborn care was also marginally significantly
reduced in the magnesium sulfate group. Giving magnesium sulfate at the lowest dose that has been shown to be effective is therefore safe for the newborn and has no negative consequences if
an overdose does not occur.
Side effects in the mother
Doyle et al. (2009) [16 ] found significantly more frequent maternal hypotension (RR 1.51; 95% CI 1.09–2.09) and tachycardia (RR 1.53;
95% CI 1.03–2.29) in the magnesium group compared to the placebo group, and significantly more treatment discontinuations due to side effects in the magnesium sulfate group (RR 3.26; 95% CI
2.46–4.31). They found no significant differences for maternal respiratory depression, postpartum hemorrhage, Caesarean section delivery, or ICU transfer. Also, for serious maternal
outcomes, such as maternal death (RR 1.25; 95% CI 0.51–3.07), cardiac arrest (RR 0.34; 95% CI 0.04–3.26), and respiratory arrest (RR 1.02; 95% CI 0.06–16.25) no significant differences were
shown. Crowther et al. (2016) [23 ] reported no case of severe maternal consequences of the therapy (death, respiratory arrest, cardiac
arrest). Negative side effects that led to premature discontinuation of therapy were significantly more frequent (RR 1.95; 95% CI 1.44–2.65). There were no differences in type of delivery,
postnatal hemorrhage, or maternal infection. Jaquemy et al. (2015) [22 ] pointed out that when magnesium sulfate is combined with
nifedipine, the risk of serious maternal side effects such as hypotension and respiratory depression is increased. The data on this are insufficient, so caution and strict indications are
required when combining nifedipine and magnesium sulfate.
Dosage
The studies included in the meta-analyses included different administration regimens, which is why the total dose administered also varies. In subgroup analyses it could be determined that
there are no significant differences in the efficacy of the doses and no linear trends that speak in favor of a higher dose. Therefore, in the event of imminent premature birth, the smallest
effective dosage should be selected, i.e., a bolus of 4 g over 15–30 minutes and a maintenance dose of 1 g/h for 12 hours or until delivery [16 ]
[21 ]
[22 ]
[23 ].
Gestational age
The inclusion criterion of weeks of pregnancy varied considerably between the studies considered in the meta-analyses. Crowther et al. (2017) [23 ] created subgroups by gestational age (< 26 weeks of pregnancy, 26–27 weeks of pregnancy, 28–29 weeks of pregnancy, 30–31 weeks of pregnancy, 32+ weeks of pregnancy).
In the subgroup analysis, there were no obvious trends of a difference in the effect of therapy on the main outcomes and no statistically significant differences. Magnesium sulfate thus has
similar effects over a wide range of gestational ages up to 34 weeks of pregnancy. Also, Doyle et al. (2009) [16 ] reported a reduction
in cerebral palsy for all studies with women under 34 weeks of pregnancy and no clear differences in the subgroups by gestational age. Jaquemy et al. (2015) [22 ] recommended the therapy for all impending premature births under 32 weeks of pregnancy and pointed out the decreasing effect with
increasing weeks of gestation. Wolf et al. (2020) [21 ] showed that the optimal range of gestational ages for therapy is a matter of
interpretation, since the rate of cerebral palsy decreases with increasing gestational age and thus the number needed to treat (NNT) increases. The weighing of the benefit against the
maternal side effects can therefore not only be answered with the available data. With the gestational-age-dependent prevalence of infantile cerebral palsy according to Oskoui et al. (2013)
[6 ] and the RR calculated by Doyle et al. (2009) [16 ] for cerebral palsy after
magnesium sulfate administration, the NNT can be calculated for the respective gestational age: For premature births < 28 weeks of pregnancy, it is 38, for 28–31 weeks of pregnancy 73 and
for 32–36 weeks of pregnancy 463. This consideration is weighed differently at the international level: The AWMF [4 ], the ACOG [30 ], and the World Health Organization (WHO) [31 ] recommend therapy for the acute
risk of premature birth < 32 weeks of pregnancy, the National Institute for Health and Care Excellence (NICE) Guideline [32 ] and the
Queensland Clinical Guidelines [33 ] recommend therapy for women < 30 weeks of pregnancy and considering therapy for women < 34
weeks of pregnancy with an acute risk of premature birth. The disagreement about the therapy between 30 and 34 weeks of pregnancy could not be conclusively clarified by the present study
either. In 2013 the (still ongoing) MAGENTA study was announced, which aims to answer this question [34 ].
Limitations
This paper contains some limitations, which result from both the included studies and the methodical procedure. The included studies were evaluated using the AMSTAR score. As already shown
in the Results section, three of the four studies considered were of very high quality. In three of the four studies, however, the category “potential publication bias assessed” could not be
answered with yes. There is a possibility that potentially existing publication bias is not sufficiently considered in the conclusions and the reported effect of antenatal magnesium sulfate
administration appears greater than it possibly is. Also, for the point “References of the included and excluded literature”, the answer to two studies was not yes, which makes it difficult
to check whether the specified inclusion and exclusion criteria have been applied correctly. One of the included studies is of low quality, as only four of the 11 AMSTAR endpoints could be
answered with yes. The poor quality of this study was taken into account in the results section. Deficiencies in the methodological approach in the present study are the fact that no
two-fold, independently performed data extraction was carried out, but the literature search was carried out in teamwork. The evaluators were therefore not independent of each other. In
addition, the literature search was carried out exclusively in PubMed and therefore does not meet the AMSTAR criterion of a systematic literature search, which has only a limited influence,
however, since an umbrella review was carried out in the present paper. In addition, the probability of publication bias could not be assessed. Finally, possible distortions also resulted
from the fact that the search was exclusively in PubMed and only English and German-language studies were considered.
Answering the research question
Does antenatal MgSO4 administration to the mother in the event of impending premature birth reduce the occurrence of infantile cerebral palsy in the child? – Due to the
consistent results of the included studies and their good quality, the above research question can be answered positively. The antenatal administration of magnesium sulfate reduces the risk
of cerebral palsy in premature infants. There is very good evidence for this, as the included meta-analyses and reviews are mostly of very high quality and contain a reasonably large number
of cases.
Conclusions
The meta-analyses published since 2009 confirm the results of the Cochrane review [16 ]. Due to the high level of evidence on the
protective effect of antenatal magnesium sulfate in preventing cerebral palsy, there is a clear recommendation to protect the child’s brain by administering magnesium sulfate to the mother
in the event of an acute risk of premature birth. Since the side effects depend on the dose, the lowest dose that has been shown to be effective of an initial bolus of 4 g over 15–30 minutes
followed by a maintenance dose of 1 g/h for 12 h or until the birth of the child should be chosen. If birth of the child does not occur, the therapy can be repeated at a later date if
necessary.
There is also sufficient high-quality data on the potential side effects for mother and child. The therapy is safe for the child. The mother may experience undesirable side effects such as
hypotension and tachycardia, some of which require discontinuation of therapy. However, severe maternal side effects do not occur more frequently with the therapy than in the placebo group,
which is why the therapy can also be classified as safe for the mother. Maternal vital signs and their reflex status should be monitored during administration. Checking maternal magnesium
levels can also be useful. Particular care should be taken when combining nifedipine and magnesium sulfate. However, the combination of the two drugs is rated more critically in the AWMF
guideline “Prevention and therapy of premature birth” [4 ] than in the AWMF guideline “Hypertensive pregnancy diseases: Diagnostics and
therapy” [35 ]. Based on the data, no clear statement can be made as to the gestational age up to which neuroprotection with magnesium
sulfate should take place. Due to the high prevalence of cerebral palsy < 30 weeks of pregnancy, therapy up to this point is always indicated. Therapy should be offered and recommended
between 30 and 32 weeks of pregnancy. Even between 32 and 34 weeks of pregnancy, an effect of neuroprotection with magnesium sulfate can still be demonstrated, but the prevalence is so low
that it should only be offered here after the mother has been given detailed information (NNT, side effects for the mother). The current AWMF guideline “Prevention and therapy of premature
birth” [4 ] does not recommend this. The question of the conditions under which neuroprotection with magnesium sulfate should take place
is therefore still relevant for research and clinical practice [36 ].
