Alexander, James M. MD; McIntire, Donald D. PhD; Leveno, Kenneth J. MD; Cunningham, F Gary MD
Magnesium sulfate has been used for the treatment of eclamptic seizures since the treatment was first described in 1925.1 Over the next several decades, its use was gradually incorporated into obstetric practice for seizure prevention in women with preeclampsia. Although the drug was presumed to be efficacious, it was not until more than 50 years of use that clinical trials proved magnesium sulfate to be effective for treatment as well as prevention of eclampsia.2,3 These studies showed magnesium to be superior to both placebo and traditional anticonvulsant drugs for women with preeclampsia and eclampsia, but debate ensued concerning risks versus benefits for women with mild hypertension. Specifically, although the multicenter European Magpie Trial demonstrated clear benefits of magnesium sulfate for eclampsia prophylaxis in women with severe gestational hypertension, its use for those with mild hypertension remains controversial.4 This is because the risks of seizures in these latter women is predicted to be relatively low, and the toxicity of magnesium, albeit uncommon, is potentially life threatening. Beginning in 2000, we changed our practice of administering magnesium sulfate prophylaxis at the time of delivery for all women with gestational hypertension to one of selective prophylaxis reserved for women with evidence of severe gestational hypertension or preeclampsia. In this report we describe the impact of this change on the incidence of eclampsia and its associated morbidity for these women and their infants.
PARTICIPANTS AND METHODS
This is a population-based observational study of the impact of selective magnesium sulfate prophylaxis on the incidence of eclampsia in indigent women presenting for obstetric care at Parkland Health and Hospital System in Dallas, Texas. Obstetric and neonatal outcomes for women who deliver at Parkland Hospital are entered into a computerized database. Nurses present at delivery complete data sheets that are checked by research nurses for consistency and completeness before electronic storage. Data on infant outcomes are abstracted from discharge records. Before beginning the study, approval was obtained from the Institutional Review Board of the University of Texas Southwestern Medical Center and Parkland Hospital. In December 1999, we instituted a standardized protocol for the intravenous administration regimen of magnesium sulfate for seizure prophylaxis. Before this, all women with gestational hypertension were administered magnesium intramuscularly as first described by Pritchard in 1955.5 Coincidental with adoption of the intravenous magnesium regimen, we changed our practice of universal seizure prophylaxis for all women with gestational hypertension to one of selective prophylaxis given only to women who met our criteria for severe hypertension.
Our criteria for defining severe gestational hypertension that required administration of magnesium sulfate were based on those described in the Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy6 and Practice Bulletin Number 33 of the American College of Obstetricians and Gynecologists (ACOG).7 These criteria, which are shown in the box, “Criteria used for Magnesium Sulfate Prophylaxis,” called for magnesium prophylaxis for women whose blood pressure was at least 140/90 mmHg and who had criteria for severe disease as defined by Practice Bulletin 33 or whose findings “increased the certainty” of the diagnosis of preeclampsia as set forth by the Working Group. Thus, we included women with 2+ protein or greater as measured by dipstick in a catheterized urine specimen. Although not considered a criterion for severe disease in Practice Bulletin 33, the Working Group used this to diagnose preeclampsia, and it offered consideration for this finding with other traditional signs of severe disease.6
The first 6 months of the intravenous magnesium protocol were used as a run-in period to determine an optimal intravenous dosing regimen. The magnesium infusion was given at the commencement of spontaneous labor or at the time of labor induction and for 24 hours postpartum. Our initial protocol included a 4-g loading dose over 20 minutes, followed by an infusion of 2 g per hour. The first serum magnesium level was measured 4 hours after the infusion started, with a goal of achieving a therapeutic range of 4–7 mEq/L (4.8–8.4 mg/dL). In women whose serum magnesium level was less than 4.0 mEq/L, and who were thus considered to be subtherapeutic, the magnesium sulfate infusion was increased to 3 g/h. At the end of the 6-month run-in period, we increased the loading dose of magnesium sulfate to 6 g because subtherapeutic levels were consistently found at 4 hours using the 4-g loading dose. We found that women whose magnesium dose was increased to 3 g/h were more likely to show signs of toxicity, and we now routinely repeat serum magnesium levels after 12 hours in these women.
Women whose serum creatinine exceeded 1 mg/dL or whose urine output was less than 30 mL/h were given the standard loading dose of 6 g followed by an infusion of 1 g/h, and serum magnesium levels were measured every 4 hours. In all women, if respiratory distress developed or if deep tendon reflexes became unresponsive or if any serum magnesium level exceeded 8 mEq/L, the infusion was stopped and serum magnesium levels measured at 1- to 2-hour intervals. The infusion was restarted at 1 g/h after the serum level decreased to less than 7 mEq/L (8.4 mg/dL), and serum levels were then determined every 4 hours.
For the purpose of this report, we analyzed outcomes of women with gestational hypertension who were treated between July 1, 2000, and December 31, 2004. Obstetric and neonatal outcomes were entered into a computerized database. Nurses present at delivery completed data sheets, which were checked for completeness and accuracy before electronic storage. Neonatal outcomes data were abstracted from discharge records by specially trained research nurses. The accuracy of this database is monitored and is verified to be over 99% accurate.
Statistical analyses of comparisons included Pearson χ2, Student t test, and multiple logistic regressions. For categorical variables, results are presented as counts and percentages (or per 1,000) and odds ratios with 95% confidence intervals. Results for continuous variables were presented as mean±standard deviation. Data analysis was performed with SAS 9 (SAS Institute, Cary, NC) and a two-sided P≤.05 was required for statistical significance.
A total of 72,004 women were delivered between July 1, 2000, and December 31, 2004. Shown in Figure 1 are the characteristics of the population by diagnosis using the criteria in the box, “Criteria Used for Magnesium Sulfate Prophylaxis.” There were 87 cases of eclampsia during the study period, for an overall incidence of eclampsia of 1 in 828 women. These are detailed as follows: 1) Eclampsia developed in 17 women before arrival at the hospital. 2) There were 32 women who developed eclampsia despite blood pressures less than 140/90 mmHg. Twelve of these 32 had an increase in blood pressure from their midpregnancy baseline of at least 30 mmHg systolic and/or 15 mmHg diastolic. One of these 32 had an eclamptic seizure more than 24 hours after delivery and another two women had 2+ or higher proteinuria with no other evidence of preeclampsia. 3) There were 27 women with eclampsia in whom hypertension was diagnosed but who did not have any of the criteria shown in the box to receive magnesium sulfate. 4) There were 11 women with severe gestational hypertension who developed eclampsia despite magnesium prophylaxis.
In Table 1 the maternal outcomes of all women with eclampsia are compared with those of women who had hypertension but no seizures. Women with eclampsia were more likely to be younger and nulliparous and more likely to undergo labor induction. Importantly, women who developed eclampsia with or without magnesium were significantly more likely (23% versus 4%, P<.001) to receive general anesthesia for emergent cesarean delivery, but they had no additional morbidity. Indications cited for the use of general anesthesia use in eclamptic women were varied and included emergent cesarean for nonreassuring fetal status, airway protection for women with recurrent convulsions, and concerns for coagulopathy in women with preeclampsia-associated thrombocytopenia or a placental abruption. There were no maternal deaths or severe morbidity in either group. Their other characteristics were similar.
A neonatal adverse outcomes composite assessment compared infants born to women with eclampsia with those born to women who had hypertension but no seizures. The composite included any one of the following: cord pH less than 7.0, 5-minute Apgar score less than 4, stillbirth or neonatal death, and an unanticipated admission of a term infant to the intensive care nursery. This composite was seen more commonly in infants born to eclamptic women than in those born to noneclamptic hypertensive women (12 versus 1%, P=.04). Infant outcomes of women with eclampsia are summarized in detail in Table 2. Five of the 17 infants born to mothers who had eclampsia before admission to labor and delivery had adverse composite outcomes, and another three were perinatal deaths. In women diagnosed with gestational hypertension but not given magnesium prophylaxis, three infants had an adverse composite outcome. These infants did well and were discharged home without sequelae. One of the 11 mothers who developed eclampsia while receiving magnesium had an infant with an adverse composite outcome as did one infant of a mother not diagnosed with gestational hypertension but who had eclampsia. These two infants did well.
As shown in Figure 1, the overall eclampsia rate for the study period was 1 in 828 deliveries. This represents a 50% increase compared with the 5 years preceding the change in magnesium sulfate protocol to one of selective administration, when the incidence of eclampsia was 1 in 1,678 deliveries (P=.04). Of the women diagnosed with hypertension who did not meet the criteria (see box “Criteria Used for Magnesium Sulfate Prophylaxis”) for magnesium prophylaxis, 27 developed eclampsia. If one assumes that these 27 cases had been prevented with prophylaxis, then the eclampsia rate during the study period would be 1 in 1,200 deliveries. Furthermore, if the eclamptic seizures in 12 women with a 30/15 mmHg or greater increase from baseline blood pressure had been prevented if they had been given magnesium prophylaxis, then the incidence of eclampsia during the study period would have been 1 in 1,500, a rate remarkably similar to that seen before the change in protocol.
After almost 50 years of successful use of intramuscular magnesium sulfate to treat or prevent eclampsia at Parkland Hospital, our decision to switch to an intravenous infusion protocol was based on its virtually universal use in the United States. The intravenous regimen chosen has been shown to result in serum levels similar to those obtained with the intramuscular regimen we had used for almost 50 years as described by Pritchard.5–8 At the same time, we changed our protocol to provide magnesium prophylaxis only to women with our criteria for severe gestational hypertension as shown in the box. Together, these changes afforded us the unique opportunity to study any effects of not giving magnesium prophylaxis to women with mild gestational hypertension. There are at least four important findings from this 4½-year observational study. First, the incidence of eclampsia more than doubled during this time compared with the preceding 5 years when magnesium prophylaxis was given to all women with gestational hypertension. Second, there were a number of adverse maternal and neonatal outcomes directly related to excess seizures. Third, only 60% of women with gestational hypertension had criteria for severe hypertension and thus received magnesium sulfate. And, finally, no serious magnesium toxicity was observed from intravenous magnesium administration.
The first observation regarding the increased incidence of eclampsia was anticipated. The rate of one case in 828 deliveries was slightly more than twice that compared with one in 1,678 deliveries in the preceding 5-year period when all women with gestational hypertension were given magnesium prophylaxis. Of the 87 eclamptic women, 27 had gestational hypertension not considered severe by the criteria in the box, and thus, they were not given magnesium prophylaxis. Another 12 of these 87 eclamptic women had elevations of baseline blood pressure of either 30 mmHg systolic or 15 mmHg diastolic, but they either did not have hypertension or had no criteria for severe disease. This group of women is of some interest because in the past these increases in blood pressure from baseline, even when absolute values did not reach 140/90 mmHg, were used as diagnostic criteria for gestational hypertension. The Working Group excluded these criteria in its second report because there was no evidence that women with these blood pressure changes have excessive adverse outcomes. For these reasons, we did not use these criteria for our study, but it seems obvious that some of these women have preeclampsia and subsequently have seizures. If we hypothetically assume that magnesium prophylaxis would have prevented eclampsia in these 39 women with either mild hypertension or a 30/15 mmHg increase in blood pressure, then the incidence of eclampsia would be about one in 1,500 deliveries. This incidence is remarkably similar to that observed during the 5-year period that preceded our change of protocol from one of universal to one of selective prophylaxis.
The second observation of increased maternal and perinatal morbidity in women who developed eclampsia is worrisome. A fourth of the women who developed eclampsia required general anesthesia for emergent cesarean delivery. Although none of these women had serious complications because of this, it is of concern because eclamptic women are generally accepted to be at higher risk for airway edema, difficult or failed tracheal intubation, and aspiration of acidic gastric contents.9 Until recently, it was commonly held that seizures with preeclampsia have no significant long-term sequelae. Zeeman and colleagues10 and Loureio and colleagues11, however, have reported that a fourth of these women have associated cerebral infarctions with subsequent gliosis, a finding of unknown prognostic significance. Another concern was the tenfold increase in composite neonatal morbidity in infants born to women with eclampsia. Of the 10 infants with adverse neonatal composites, five were in women who had eclampsia at home, and three of these infants were stillborn or died after birth. The remaining five women had eclampsia while in the hospital, and although 3 of their 5 neonates had a cord arterial pH less than 7.0, all did well. Taken in toto, we were unable to directly relate maternal or neonatal morbidity to the excess seizures, but these outcomes are of concern.
The third finding that we treated only 60% of women with gestational hypertension deserves comment. Because we presumed that withholding prophylaxis would result in eclampsia in some women with nonsevere hypertension, we used more liberal criteria for defining “severe” hypertension than recommended by most.6,7 For example, our criteria shown in Table 1 included 2+ proteinuria to define severe gestational hypertension for which magnesium prophylaxis was given. According to ACOG,6 proteinuria of at least 3+ by dipstick is necessary to classify gestational hypertension as “severe” without other findings. Because of these liberal criteria for defining severe disease, only 40% of women with gestational hypertension were not given magnesium sulfate.
The fourth important observation was the lack of serious morbidity associated with intravenous magnesium sulfate administration. Even so, 2.4% of treated women had at least one serum magnesium level that exceeded 8 mEq/L, and this is a level considered to be potentially toxic. Although we measured serum magnesium to verify “therapeutic levels,” this is not done in most hospitals. Despite such monitoring, the eclamptic seizure rate was 1 in 358 in women given prophylaxis for severe gestational hypertension. It seems reasonable to assume that they likely are part of “irreducible minimum” of unpreventable cases of eclampsia.12
Although magnesium sulfate has been used for decades in American obstetrics for both the treatment and the prevention of eclampsia, it has been only recently that well-controlled comparative studies of its use have been reported. These studies have proved magnesium to be the most effective agent for prevention of primary and recurrent eclamptic seizures. Specifically, the Collaborative Eclampsia Trial showed a 50% reduction of recurrent seizures in eclamptic women given magnesium sulfate compared with those given either phenytoin or diazepam.3 In another study, we reported a tenfold decrease in convulsions in hypertensive women treated with magnesium (0 of 1,000) compared with those given phenytoin (10 of 1,000).2 Subsequently, the multicenter Magpie Trial showed that the eclampsia rate in women with severe gestational hypertension was decreased by 60% with prophylactic magnesium compared with placebo.4
Although these studies established the effectiveness of magnesium to treat eclampsia and its use as prophylaxis in women with severe gestational hypertension, its efficacy in preventing seizures in women with mild gestational hypertension remained unclear. We are aware of only two randomized trials done to evaluate this premise and, combined, they enrolled a total of 357 women and were thus underpowered to answer this question.12,13 In an editorial, Scott14 concluded it to be unlikely that an adequately powered trial would ever be conducted because of the large numbers of women needed for such a comparative study. Our data are in agreement with this sentiment, and we compute that 8,600 women with mild gestational hypertension would have to be randomized to ascertain whether there are any significant risks or benefits of magnesium prophylaxis for mild gestational hypertension.
These findings seem straightforward. If only those women with gestational hypertension who meet the criteria for “severe” as shown in Table 1 are given intravenous magnesium sulfate, the rate of eclampsia in the population will increase. If 3+ or greater proteinuria by dipstick is chosen as the criteria for diagnosing severe hypertension with no other findings, then we assume that the seizure rate will be even higher. Using the criteria shown in Table 1, the overall use of intravenous magnesium and its potential toxicity is decreased by 40%, and about 1 in 100 women with “nonsevere” disease develop eclampsia. These women are at risk for general anesthesia and their neonates are at risk for transient acidosis. Although not necessarily associated with serious morbidity, the potential for such is concerning.
Criteria Used for Magnesium Sulfate Prophylaxis
* Blood pressure of 140/90 mmHg or greater after 20 weeks of gestation in a woman not known to be chronically hypertensive
* Proteinuria of 2+ or greater as measured by dipstick in a catheterized urine specimen
* Serum creatinine more than 1.2 mg/dL
* Platelets less than 100,000/μL
* Aspartate transaminase elevated two times above upper limit of normal range
* Persistent headache or scotomata
* Persistent midepigastric or right-upper quadrant pain
1. Lazard EM. A preliminary report on the intravenous use of magnesium sulphate in puerperal eclampsia. Am J Obstet Gynecol 1925;9:178–88.
2. Lucas MJ, Leveno KJ, Cunningham FG. A comparison of magnesium sulfate with phenytoin for the prevention of eclampsia. N Engl J Med 1995;333:201–5.
3. The Eclampsia Trial Collaborative Group. Which anticonvulsant for women with eclampsia? Evidence from the Collaborative Eclampsia Trial. Lancet 1995;345:1455–63.
4. Altman D, Carroli G, Duley L, Farrell B, Moodley J, Neilson J, et al; The Magpie Trial Collaborative Group. Do women with preeclampsia, and their babies, benefit from magnesium sulfate? The Magpie Trial: a randomized placebo-controlled trial. Lancet 2002;359:1877–90.
5. Pritchard JA. The use of the magnesium ion in the management of eclamptogenic toxemias. Surg Gynecol Obstet 1955;100: 131–40.
6. Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol 2000;183:S1–S22.
7. Diagnosis and management of preeclampsia and eclampsia. ACOG Practice Bulletin No. 33. American College of Obstetricians and Gynecologists. Obstet Gynecol 2002;99:159–67.
8. Sibai BM, Graham JM, McCubbin JH. A comparison of intravenous and intramuscular magnesium sulfate regimens in preeclampsia. Am J Obstet Gynecol 1984;150:728–33.
9. Obstetric analgesia and anesthesia. ACOG Practice Bulletin No. 36. American College of Obstetricians and Gynecologists. Int J Gynaecol Obstet 2002;78:321–35.
10. Zeeman GG, Fleckenstein JL, Twickler DM, Cunningham FG. Cerebral infarction in eclampsia. Am J Obstet Gynecol 2004;190:714–20.
11. Loureiro R, Leite CC, Kahhale S, Freire S, Sousa B, Cardoso EF, et al. Diffusion imaging may predict reversible brain lesions in eclampsia and severe preeclampsia: initial experience. Am J Obstet Gynecol 2003;189:1350–55.
12. Witlin AG, Sibai BM. Magnesium sulfate therapy in preeclampsia and eclampsia. Obstet Gynecol 1998;92:883–9.
13. Livingston JC, Livingston LW, Ramsey R, Mabie BC, Sibai BM. Magnesium sulfate in women with mild preeclampsia: a randomized controlled trial. Obstet Gynecol 2003;101:217–20.
14. Scott JR. Magnesium sulfate for mild preeclampsia. Obstet Gynecol 2003;101:213.