Maternal Receipt of Magnesium Sulfate Does Not Seem to Reduce the Risk of Neonatal White Matter Damage
Objective. To investigate whether in utero exposure to magnesium sulfate is associated with a lower incidence of cranial ultrasonographic abnormalities that predict cerebral palsy in infants who weigh less than 1501 g at birth.
Design. For a prospective study of the antecedents of cranial ultrasonographic abnormalities, we enrolled infants who weighed 500 to 1500 g when born at five institutions. Data were collected by interview of the mothers and review of medical records. Protocol cranial ultrasonograms were obtained as close as possible to postnatal days 1, 7, and 21. Abnormality on cranial ultrasound scans was determined by a consensus committee of three sonologists.
Results. Of the 1518 infants for whom we knew whether the mothers received magnesium sulfate, the first protocol cranial ultrasound scan was available for 1409 infants, the second for 1274 infants, and the third for 1050 infants. Forty-five percent of infants were exposed to magnesium sulfate before delivery. The major correlates of magnesium sulfate exposure were receipt of antenatal corticosteriods and a diagnosis of preeclampsia and/or pregnancy-induced hypertension. Maternal magnesium receipt was not associated with a reduced incidence of hypoechoic or hyperechoic images of white matter parenchyma, intraventricular hemorrhage, or ventriculomegaly, even when the sample was stratified by each of six potential confounders. When adjustment was made for gestational age, a measure of birth weight for gestational age, antenatal corticosteroid exposure, preeclampsia and pregnancy-induced hypertension, route of delivery, and the occurrence of any labor, the risk ratios for each cranial ultrasonographic abnormality associated with magnesium sulfate exposure hovered close to 1.
Conclusion. Maternal receipt of magnesium sulfate does not seem to be associated with an appreciably reduced risk of cranial ultrasonographically defined neonatal white matter damage, intraventricular hemorrhage, or ventriculomegaly. low birth weight, magnesium sulfate, periventricular leukomalacia, antenatal corticosteriods, intraventricular hemorrhage, pregnancy-induced hypertension.
Based on their recently published case-control study, Nelson and Grether1 concluded that maternal receipt of magnesium sulfate was associated with a greatly reduced risk of cerebral palsy in very low birth weight infants. They characterized magnesium sulfate as an inexpensive intervention that seems to be relatively safe. One additional study has also found a reduced incidence of cerebral palsy in those born before term who were exposed antenatally to magnesium sulfate,2 but another has not.3
Hypoechoic images (ie, echolucencies or what some consider cysts) in the cerebral white matter parenchyma, some of which are referred to as “periventricular leukomalacia,” are the cranial ultrasonographic abnormalities that best predict cerebral palsy in premature infants.4 Hyperechoic images (ie, echodensities), sometimes interpreted as infarction or grade IV hemorrhage, also predict an increased risk of cerebral palsy.
In one sample, 54% of prematurely born infants in whom disabling cerebral palsy developed had parenchymal echodensity and lucency or ventricular enlargement seen on cranial ultrasound scans obtained between 1 and 5 weeks after birth.9 This has led to the inferences that cerebral palsy in those born before term probably represents two disorders; only one of these disorders is characterized by a cranial ultrasound signature, and as much as half of disabling cerebral palsy in preterm infants is a consequence of this disorder. The study of the antecedents of etiologically heterogeneous disorders such as cerebral palsy should be enhanced by study of more homogeneous subgroups, such as those with early sonographic expression.
If maternal receipt of magnesium sulfate reduces the risk of cerebral palsy in the prematurely born, then magnesium sulfate exposure in utero could be expected to reduce the risk of hyperechoic and hypoechoic images. This report explores the relationships between magnesium sulfate and hyperechoic and hypoechoic images of parenchymal white matter (which we refer to jointly as ultrasonographic expressions of white matter damage) and other cranial ultrasound abnormalities, both before and after adjustment for potential confounders.
The mothers of 1665 infants were recruited for this study. To be eligible, infants had to weigh 500 to 1500 g when born at five participating hospitals between January 1991 and December 1993. A total of 60 infants died before cranial ultrasound scans could be obtained or had scans that, despite intensive efforts, could not be found months to years later. Information about receipt of magnesium sulfate and potential confounders was available from maternal medical records for 1518 of the 1605 infants for whom scans were available (Table1). These infants and their 1331 mothers constitute the sample for this set of analyses. The distributions of maternal, obstetric, and newborn characteristics are presented in Table 1. The gestational age estimate was based on the following hierarchy: fetal ultrasound scan estimate obtained before the end of the 13th week of gestation (32% of the sample), dates in the prenatal record (62%), maternal interview (4%), and the admission log book of the neonatal intensive care unit (2%).
We divided gestational age into three groups, younger than 26 weeks, 26 to 28 weeks, and older than 28 weeks. The group of infants younger than 26 weeks represents infants near the borderline of viability. Because we excluded infants weighing more than 1500 g, fetal growth retardation is overrepresented in the group of infants older than 28 weeks.11
Cranial Ultrasound Scans
Manuals were created to standardize the scanning procedure and the interpretation of scans. The six standard coronal views were those advised by Teele and Share.12 Protocol scans were obtained once during the first 4 postnatal days (median, day 1), once between postnatal days 5 and 14 (median, day 7), and once between days 15 and 60 (median, day 22). The first protocol cranial ultrasound scans were available for 1409 infants. The 109 infants who were not scanned during the first protocol scan interval tended to be gestationally older and more stable physiologically than their peers. All infants who died and some who were clinically well or who were transferred to other institutions did not have later scans. As a consequence, the second protocol set of scans was available for 1274 infants, and the third set was available for 1050 infants.
Each set of scans was read independently by two sonologists. Scans were brought to a consensus committee if either reader identified: (1) a hyperechoic or hypoechoic lesion anywhere in the cerebral white matter, (2) moderate or severe ventriculomegaly (visually assessed against a template), or (3) intraventricular hemorrhage (IVH; ie, hyperechoic material in the lateral ventricles). Nine of the 10 study sonologists constituted the consensus committees. All consensus committees consisted of a minimum of three sonologists (each from a different institution) reading together. After every 2-hour session, a break was taken, and one member of the committee was rotated. All members of each consensus committee had to agree about the presence, size, and location of all white matter lesions and about the presence of at least moderate ventricular dilation or IVH. If they could not agree, then the scans were brought to a larger consensus group (consisting of the three or four readers who could not achieve consensus plus the three or four readers from another consensus committee). The majority decision, needed in approximately 8% of all sets of scans presented to a consensus committee, was accepted as the consensus reading for these scans.
For the purposes of this study, the five major sonographic entities are IVH, early parenchymal echo abnormality (PEA; defined as a parenchymal hyperechoic or hypoechoic image evident on the first protocol scan), late PEA (defined as PEA not seen on the first protocol scan and first evident on the second or third protocol scan), any PEA (defined as either early or late PEA), and ventriculomegaly (defined as moderate or severe ventricular dilation on any of the three protocol scans). Hypoechoic images of the white matter parenchyma, which probably reflect a number of different disorders13,14 that are not readily distinguished sonographically,15 are also included in Table 4 as sonographic disorders of major interest.
Pregnancy-induced Hypertension and Magnesium Sulfate
A woman was considered to have pregnancy-induced hypertension or preeclampsia if the diagnosis of either was found in the prenatal or delivery charts, if during an interview shortly after delivery, the mother acknowledged that she was told she had preeclampsia, toxemia, or pregnancy-induced hypertension, or if the diastolic blood pressure during pregnancy was 90 or higher, but prepregnancy hypertension was not cited in the medical records and was denied during the interview. Classification of magnesium sulfate receipt (ie, yes or no) was based on information in the mother’s medical record. Details about dose and duration of magnesium sulfate were not collected.
Confounders and Correlates
For the purposes of this report, the main potential confounders of assessments of the relationship between magnesium sulfate receipt and intracranial sonographic abnormalities of the newborns were deemed a priori to be preeclampsia and pregnancy-induced hypertension, receipt of any antenatal corticosteroids, multiple gestations (ie, singletons versus twins and triplets), the occurrence of any labor, route of delivery, gestational age (ie, <26, 26 to 28, and >28 weeks), and birth weight z score (ie, −1 or greater, less than −1 to −2, and less than −2). The birth weight z score is the number of SDs the infant’s birth weight is from the median for gestational age. Infants whose birth weight z scores are less than −2 are in the lower 2.5% of all infants classified by birth weight for gestational age. Subsequent examinations of our data did not identify other potential confounders.
The null hypothesis evaluated is that infants born to women who received magnesium sulfate shortly before delivery have the same risk of IVH, early PEA, late PEA, any PEA, a hypoechoic image in white matter parenchyma, and ventriculomegaly as do infants born to women who did not receive magnesium sulfate. The relationships among variables were evaluated in univariate analyses and in different strata defined by the presence or absence, or level, of each potential confounder. We created logistic regression models to adjust for all the potential confounders at the same time. The separate analyses of singletons and multifetal gestations provided such similar results that only findings of the entire sample are presented. Given a sample size of 1050 infants who had third protocol scans, a magnesium sulfate exposure prevalence of 44% and a prevalence of any PEA of 12%, this study has a power of 0.998 to perceive an odds ratio of 0.5 and a power of 0.75 to perceive an odds ratio of 0.67.
The mothers of 45% (678 of 1518) of infants were given magnesium sulfate before delivery. These infants were more likely than their peers to be exposed antenatally to corticosteroids (60% vs 46%), and their mothers were more likely to carry a diagnosis of pregnancy-induced hypertension or preeclampsia (37% vs 18%). Women with pregnancy-induced hypertension were more likely than their peers to give birth to gestationally older infants who had evidence of fetal growth retardation (ie, birth weight z score less than −2) (Table 1).
The risks of IVH, PEA, and ventriculomegaly were reduced in infants born to women who received antenatal corticosteroids or were given a diagnosis of preeclampsia or pregnancy-induced hypertension. Receipt of magnesium sulfate was not associated with a reduced risk of these cranial ultrasonographically defined disorders (Table2).
When infants were stratified according to the presence or absence of a potential confounder, receipt of magnesium sulfate was not associated with appreciable reductions of any cranial ultrasound abnormality among the infants of mothers with pregnancy-induced hypertension or among the infants of mothers who had labor (Table 3). The risk of IVH was slightly reduced among those exposed to magnesium sulfate in strata defined by the absence of labor, abdominal delivery, and maternal receipt of corticosteroids, as well as in the gestationally oldest infants and those who were growth retarded. Magnesium sulfate receipt was not associated with a reduced risk of the other cranial ultrasonographic abnormalities in any of the strata.
When logistic regression adjustment was made for the six potential confounders (ie, gestational age divided into three groups, birth weight z score divided into three groups, antenatal corticosteroids, preeclampsia and/or pregnancy-induced hypertension, route of delivery, and labor) magnesium sulfate was not associated with a reduced risk of IVH, PEA of any form, or ventriculomegaly (Table4).
We did not find evidence that magnesium sulfate reduces the incidence of those cranial ultrasound abnormalities that predict cerebral palsy in preterm infants. Our analyses, stratified by the presence or absence of preeclampsia and pregnancy-induced hypertension, as well as by other potential correlates of magnesium receipt and cranial ultrasound abnormalities, indicate that our findings are unlikely to be attributable to confounding.
Because the cranial ultrasonographic expression of neonatal white matter disorders predicts cerebral palsy in very low birth weight infants, we assumed that if magnesium sulfate reduced the risk of cerebral palsy, then it might do so by reducing the risk of neonatal white matter disorders. Our failure to find a protective effect of magnesium sulfate on the risk of neonatal white matter disorders suggests that if magnesium sulfate does prevent cerebral palsy in those born before term, it does so by a mechanism not associated with preventing sonographically evident neonatal white matter damage. The alternative inference is that magnesium sulfate does not reduce the risk of cerebral palsy in very low birth weight infants.
Explanations for a reduced risk of cerebral palsy without a reduced risk of neonatal white matter damage might be plausible if some cerebral palsy in those born before term is caused by gray matter damage. Indeed, in laboratory animal studies, magnesium does seem to reduce the frequency and severity of neuronal damage caused by excitotoxic amino acids.16
IVH and ventriculomegaly are correlates of white matter damage.21 For this reason we sought, but did not find, an inverse relationship between maternal magnesium sulfate receipt and risk of IVH and ventriculomegaly in the newborn. Our failure to find a reduced risk of IVH in association with maternal receipt of magnesium sulfate is in keeping with the observation of some investigators28,29 but not of others.30,31
Our study has a number of potential limitations. First, approximately one third (468 of 1518) of the infants did not have third protocol cranial ultrasound scans. This might have limited our ability to identify hypoechoic images in some infants. On the other hand, infants who did not have third protocol scans were the most mature and physiologically stable, precisely those least likely to have had hypoechoic images.4
Second, half of the infants had their third protocol cranial ultrasound scans obtained by postnatal day 22. In our sample, however, the proportion of third protocol scans with new hypoechoic images remained constant during each 10-day interval, between days 22 and 60. Thus, it is unlikely that we would have had a sizable increase in the number of infants with hypoechoic images had more of the third protocol scans been obtained closer to postnatal day 60.
The third limitation of our study is our failure to collect information about the amount and timing of magnesium sulfate administration. The fourth limitation is the relatively modest power to perceive modest reductions in the odds ratios for white matter echo abnormality. The fifth limitation is that this is an observational study and not a randomized clinical trial.
On the other hand, our study has a number of strengths, including consensus reading of every set of scans deemed abnormal by either of two readers, data collected prospectively about magnesium sulfate and correlates, and as large a sample studied in such detail as is likely to be available. Our findings need to be assessed in light of the strengths and potential limitations of this study, as well as in light of the findings, strengths, and potential limitations of other studies of magnesium sulfate. The ideal study, a randomized clinical trial, is unlikely to be published soon.
Since this paper was accepted for publication, two relevant articles have been published:
Schendel DE, Berg CJ, Yeargin-Alsopp M, Boyle CA, Decoufle P. Prenatal magnesium sulfate exposure and the risk for cerebral palsy or mental retardation among very low-birth-weight children aged 3 to 5 years.JAMA. 1996;276:1805–1810
Rouse DJ, Hauth JC, Nelson KG, Goldenberg RL. The feasibility of a randomized clinical perinatal trial: maternal magnesium sulfate for the prevention of cerebral palsy. Am J Obstet Gynecol. 1996;175:701–705
Funds for this project were provided by grant NS 27306 from the National Institute of Neurological Disorders and Stroke.
We are grateful to the women who not only agreed to be interviewed for this study, but also allowed data to be collected from their infants’ charts. We also express appreciation to our colleagues who contributed to the success of this project.
- Received April 1, 1996.
- Accepted August 20, 1996.
Reprint requests to (A.L.) Carnegie 207, Children’s Hospital, 300 Longwood Ave, Boston, MA 02115.
- IVH =
- intraventricular hemorrhage •
- PEA =
- parenchymal echo abnormality
- Nelson KB,
- Grether JK
- Hauth JC,
- Goldenberg RL,
- Nelson KG,
- DuBard MB,
- Peralta MA,
- Gaudier FL
- ↵Paneth N, Jetton J, Pinto-Martin J, Susser M. Magnesium sulfate in labor and risk of neonatal brain lesions and cerebral palsy in low birth weight infants. Pediatrics. 1997;99(5). URL:http://www.pediatrics.org/cgi/content/full/99/5/e1
- Stewart AL, Turcan D, Rawlings G, Hart S, Gregory S. Outcome for infant at high risk of handicap. In: Elliot K, O’Connor M, eds.Major Mental Handicap: Methods and Costs of Prevention. CIBA Foundation Symposium 59. Amsterdam, Netherlands: Elsevier Science Publishers; 1990:151–164
- Campbell MK,
- Halinda E,
- Carlyle MJ,
- et al.
- Pinto-Martin JA,
- Riolo S,
- Cnaan A,
- et al.
- Aziz K,
- Vickar DB,
- Sauve RS,
- Etches PC,
- Pain KS,
- Robertson CMT
- Arnold CC,
- Kramer MS,
- Hobbs CA,
- et al.
- ↵Teele R, Share J. Ultrasonography of Infants and Children.Philadelphia, PA: WB Saunders Co; 1991
- ↵Volpe JJ. Neurology of the Newborn. 3rd ed. Philadelphia, PA: WB Saunders Co; 1994
- ↵Paneth N, Rudelli R, Kazam E, Monte W. Brain Damage in the Preterm Infant. London, England: Mac Keith Press; 1994
- FineSmith RB,
- Yellin P,
- Ambrosino M,
- Khan A
- Rushton DI,
- Preston PR,
- Durbin GM
- Guzzetta F,
- Shackelford GD,
- Volpe S,
- et al.
- Perlman JM,
- Rollins N,
- Burns D,
- Risser R
- ↵Lemons J, Stevenson D, Verter J, et al. In utero magnesium exposure: risk of death and of intraventricular hemorrhage (IVH) in very low birth weight (VLBW) infants. Pediatr Res. 1996;39:225A. Abstract
- Kuban KCK,
- Leviton A,
- Pagano M,
- et al.
- ↵Perlman J, Fernandez C, Gee J, Leveno K, Risser R. Magnesium sulfate (Mg) administered to mothers with pregnancy-induced hypertension (PIH) is associated with a reduction in periventricular-intraventricular hemorrhage. Pediatr Res.1995;38:231A. Abstract
- Copyright © 1997 American Academy of Pediatrics