The Lancet Correspondence

Volume 360, Number 9342 26 October 2002

The Magpie Trial

Sir--The results of the Magpie Trial (June 1, p 1877)1 confirm claims that magnesium sulphate given during labour and for at least 24 h post partum will reduce the risk of seizures in women with severe pre-eclampsia or imminent eclampsia.2 This study has generated much enthusiasm with the lay and medical press, including a Commentary3 in your journal suggesting that magnesium sulphate has the potential to prevent millions of deaths in developing nations. However, a more critical look at this trial raises several methodological concerns, questions related to its conclusions, and potential for magnesium to do more harm than good, especially in the third world.

This trial was done at 175 hospitals in 33 nations, resulting in substantial heterogeneity in clinical characteristics, obstetric care, and availability of maternal and neonatal care units. Complications (tables 4 and 5) were only listed after randomisation, and are not well defined (eg, what was judged as renal or hepatic failure, coagulopathy, respiratory depression?). Nor do we know how many of these events occurred before entry into the trial. Magnesium therapy decreased the incidence of eclamptic convulsions, but maternal and perinatal mortality and morbidity were similar in both groups. There were nine more deaths in the placebo group, but they were in the renal failure, embolism, and infection categories and these differences can hardly be ascribed to magnesium, whose actions to reduce mortality are by preventing status epilepticus or aspiration. The trial failed to show differences in these latter outcomes, but the women receiving magnesium had serious side-effects including respiratory depression and hypotension, each of which could prove fatal if not identified and managed properly.

Shirish S Sheth and Iain Chalmers' suggestion3 that magnesium sulphate will prevent millions of deaths in the third world is misleading and probably untrue. Most of the morbidity and mortality associated with eclampsia is associated with out-of-hospital seizures, and such events are more frequent among patients in developing countries without prenatal care.4 Thus the introduction of magnesium sulphate in the third world is unlikely to have much effect on maternal mortality from pre-eclampsia. Given that this trial, and others in developed countries, fail to show beneficial effects of magnesium on perinatal mortality and morbidity,2,4,5 we must look elsewhere to improve statistics in the underdeveloped world. Of primary importance would be to focus resources on improvement of access to care and provide the clinical resources to improve diagnosis and management of pre-eclampsia.

Unfortunately, this trial did not answer the most important question regarding magnesium sulphate therapy--ie, whether its use is indicated in women with mild pre-eclampsia. Almost all the enrolled patients had severe disease: 50% received antihypertensives before randomisation, 75% received antihypertensives after randomisation, and the remainder had severe pre-eclampsia or imminent eclampsia. Even so, magnesium was not beneficial among those enrolled in developed countries and the number needed to treat to prevent one seizure was 385. Thus more women might die from magnesium toxicity than from seizures if magnesium was universally recommended for all women with pre-eclampsia, especially in developing countries where its disciplined use may be harder to achieve.

Magnesium sulphate is the ideal agent to use in women with eclampsia and severe pre-eclampsia. It reduces the rate of eclamptic seizures only during labour and immediately post partum (<35% of all cases of eclampsia).4 Before it is made available to women with pre-eclampsia and their care providers, the indications, contraindications, availability of trained individuals, and resources to manage complications must be well defined. Finally, magnesium will not prevent most maternal and perinatal mortality and morbidity related to pre-eclampsia and eclampsia in the third world. This can only be prevented by improving the health-service infrastructure in these countries or preventing severe pre-eclampsia.

I thank Marshall D Lindheimer of the University of Chicago for his advice and review of these comments.

Baha M Sibai

--------------------------------------------------------------------------------

Department of Obstetrics and Gynecology, University of Cincinnati, Cincinnati, OH 45267, USA (e-mail:[email protected])

1 The Magpie Trial Collaborative Group. Do women with pre-eclampsia, and their babies, benefit from magnesium sulphate? The Magpie Trial: a randomised placebo-controlled trial. Lancet 2002; 359: 1877-90. [Text]

2 Coetzee EJ, Dommisse J, Anthony J. A randomized controlled trial of intravenous magnesium sulphate versus placebo in the management of women with severe pre-eclampsia. Br J Obstet Gynaecol 1998; 105: 300-03. [PubMed]

3 Sheth SS, Chalmers I. Magnesium for preventing and treating eclampsia: time for international action. Lancet 2002; 359: 1872-73. [Text]

4 Katz VL, Farmer R, Kuller JA. Preeclampsia into eclampsia: toward a new paradigm. Am J Obstet Gynecol 2000; 182: 1389-96. [PubMed]

5 Mattar F, Sibai BM. Eclampsia VIII. Risk factors for maternal morbidity. Am J Obstet Gynecol 2000; 182: 307-12. [PubMed]

Sir--Although pre-eclampsia is seen throughout the world, its prognosis varies somewhat with geography and race. The Magpie Trial1 aimed to determine whether women with pre-eclampsia should receive magnesium sulphate as prophylaxis against eclampsia. In the trial population as a whole, there was indeed a significant reduction in the incidence of eclampsia (1·9% in the placebo group, 1·1% with magnesium), although there was no significant difference in maternal or perinatal mortality. There were few complications of eclampsia, and death was attributed to the disorder in only three cases out of the 10 141 pre-eclamptic women studied.

The investigators ensured that treatment allocation was balanced for several variables including severity of pre-eclampsia and country. Moreover, just before the final analysis, three subgroups of the participating countries were defined as having "low, moderate, or high" perinatal mortality. A subgroup analysis was then done to determine whether the treatment effect was uniform and independent of socioeconomic and racial variables. It was not.

1560 women were randomised in countries with low perinatal mortality (including Australia, the UK, and Canada). There were no deaths in these patients. Eclampsia occurred in four of 778 in the magnesium group and six of 782 in the placebo group--a difference that was not significant. The eclampsia rate overall in countries with low perinatal mortality was one in 156 pre-eclamptic patients, which is consistent with the rate inferred from a definitive study of eclampsia in the UK.2 In that study, 4·9 of 10 000 pregnancies (one in 163 cases of pre-eclampsia assuming a pre-eclampsia rate of 8%) were complicated by eclampsia at a time when magnesium sulphate was not used in the UK.

In the women who delivered in countries with high perinatal mortality (including Bangladesh, Brazil, Malawi, Uganda, and South Africa), the eclampsia rate with magnesium treatment was one in 128 pre-eclampsia pregnancies (22 of 2814 patients), whereas the rate after placebo treatment was one in 44 (64 of 2812 patients). This finding suggested a clear benefit of magnesium treatment.

Why does there seem to be such a transnational difference in the response to magnesium? Perhaps racial characteristics are important in determining propensity to eclampsia. Perhaps treatment of hypertension is more effective in "first world" nations, or perhaps delivery is delayed in countries with high perinatal mortality. 25% of the women in each group received no antihypertensive treatment. Were magnesium-treated and placebo-treated women equally represented in this untreated group?

There is a danger in applying the results of this trial to the treatment of women in countries with low perinatal mortality. In these countries, the "side-effects and problems at injection site", which afflicted 25% of the magnesium group and 5% of the placebo group, represent a significant cost when no benefit is apparent. In less developed nations, the analysis of cost and benefit favours the use of magnesium.

We agree with the authors that "the results of this trial should be made available to women with pre-eclampsia, and to those responsible for their care". The conclusions drawn will vary from nation to nation. Should anyone be surprised?

*Barry N J Walters, Dorothy Graham, Dale Hamilton

--------------------------------------------------------------------------------

*Obstetric Medicine, King Edward Memorial Hospital for Women, and School of Women's and Infants' Health, University of Western Australia, Perth, Western Australia 6008, Australia (e-mail:[email protected])

2 Douglas KA, Redman CWG. Eclampsia in the United Kingdom. BMJ 1994; 309: 1395-400. [PubMed]

Sir--Our research group has long acknowledged the growing evidence supporting the use of magnesium sulphate for seizure prophylaxis in the setting of pre-eclampsia. Moreover, we agree with the Magpie authors1 that magnesium sulphate, compared with placebo, seems to be better for the prevention of eclampsia, and may reduce attendant maternal mortality. With regard to their data relating to fetal and infant outcomes, however, we are not completely reassured by the conclusion that magnesium sulphate "does not appear to be harmful".

Our own work (the Magnesium and Neurologic Endpoints Trial [MagNET])2 using magnesium sulphate under the different clinical circumstance of preterm labour, unexpectedly found a significant association (risk difference 15·2%, 95% CI 4·4-26·0; p=0·01) between tocolytic magnesium sulphate and total paediatric mortality (fetal, neonatal, and postneonatal death). More recently, we have shown that higher concentrations of ionised magnesium in maternal serum at the time of infant delivery were associated with an increased risk of neonatal intraventricular haemorrhage.3

Women diagnosed with pre-eclampsia were excluded from the MagNET trial to avoid confounding. Nevertheless, the unexpected paediatric complications in our tocolytic magnesium trial led us to scrutinise the Magpie data set, enquiring about possible adverse paediatric effects when magnesium sulphate was used instead for the prevention of eclampsia. Indeed, their table 9 reveals that, after excluding fetuses who were probably dead before randomisation, there was a numeric excess of paediatric mortality (519 magnesium group deaths-488 placebo group deaths=31 excess deaths) among pregnancies randomised to magnesium. This difference does not reach significance. However, it does represent a substantially larger excess of paediatric deaths, in magnesium treated pregnancies, than the nine excess maternal deaths seen among placebo randomised women who did not receive magnesium. Those nine excess maternal deaths, although also not significant, were of sufficient magnitude to lead the authors to conclude that magnesium sulphate "probably reduces the risk of maternal death". Why then not speculate possible fetal adverse effects as well?

We strongly support an international public-health initiative to make magnesium sulphate widely available for the prevention of eclampsia, as called for in the Commentary4 accompanying the Magpie report. However, we are concerned that, once introduced, the drug may also be used as a tocolytic. Such an expansion of obstetric indications for magnesium sulphate might occur even in the absence of any proof for tocolytic efficacy, and despite growing evidence that it may be harmful, just as it continues to be used in parts of North America. In mathematical models developed to calculate the possible effect of magnesium tocolysis in the USA, tocolytic magnesium sulphate was estimated to be associated with several thousand excess infant deaths annually.5 If the same assumptions were applied worldwide, the effect on infant mortality of widespread magnesium sulphate use (for tocolysis) could be substantial. Thus, as magnesium sulphate is poised to become the world's first-line agent for the prevention and treatment of eclampsia, we urge that paediatric effects continue to be monitored, and that obstetric indications for its use are not expanded to include tocolysis.

*Robert Mittendorf, Peter G Pryde, Marguerite Herschel, Kwang-Sun Lee

--------------------------------------------------------------------------------

*Department of Obstetrics and Gynecology, Stritch School of Medicine, Loyola University Medical Center, Maywood, IL 60153, USA; Department of Pediatrics, University of Chicago Children's Hospital, Chicago, IL. (e-mail:[email protected])

2 Mittendorf R, Covert R, Boman J, Khoshnood B, Lee K-S, Siegler M. Is tocolytic magnesium sulphate associated with increased total paediatric mortality? Lancet 1997; 350: 1517-18. [Text]

3 Mittendorf R, Dambrosia J, Dammann O, et al. Association between maternal serum ionized magnesium levels at delivery and neonatal intraventricular hemorrhage. J Pediatr 2002; 140: 540-46. [PubMed]

4 Sheth SS, Chalmers I. Magnesium for preventing and treating eclampsia: time for international action. Lancet 2002; 359: 1872-73. [Text]

5 Mittendorf R, Pryde P, Khoshnood B, Lee K-S. If tocolytic magnesium sulfate is associated with excess total pediatric mortality, what is its impact? Obstet Gynecol 1998; 92: 308-11. [PubMed]

Sir--In their Commentary, Shirish S Sheth and Iain Chalmers1 state that "it is now up to those responsible for maternal health services . . . to ensure that this effective, apparently safe, and inexpensive drug is available to women everywhere when needed." But the provision of magnesium sulphate for a condition that affects 5% of all pregnancies worldwide poses a substantial logistical challenge. In our hospital, where we deliver 22 000 women yearly, it would cost over US$6000 per annum in drug costs alone. If we are to ensure that women worldwide have access to this important drug, then the magnesium sulphate protocols may need to be reviewed.

In the standard magnesium sulphate regimen (and that used in the Magpie Trial), the treatment is continued for a period of 24 h, or six injections. The original decision to continue it for this long was made on the basis of the eclampsia treatment regimen where it is continued for the duration of the period in which recurrent fits are most likely. For eclampsia prophylaxis, however, the same arguments do not apply, especially when one is treating women with mild disease in whom immediate delivery is not anticipated.

Fortunately, the results of the Magpie Trial suggest that a shorter course of treatment may be adequate. 875 women received anticonvulsants before recruitment into the trial.2 In more than half the cases, this treatment was magnesium, and most of these women probably received only the loading dose injection before being referred on to the recruiting hospital. For these women, there was no difference in outcome between those given further magnesium sulphate or placebo (relative risk 1·24, 95% CI 0·49-3·11). The argument for a short course is further supported by the trial data, which suggest that the drug may continue to be beneficial long after the treatment has been given. In the trial, the median time from recruitment to delivery was 12 h. Since eclamptic fits can occur at any time up to 7 days after delivery, for most of the time at which the women were at risk, they would have had subtherapeutic serum concentrations of magnesium sulphate. Despite this situation, prophylaxis seems to have been successful.

To make magnesium sulphate available to women most at risk of death from eclampsia, a short regimen suitable for use in under-resourced countries needs to be produced. Ideally this would include a single magnesium sulphate injection, but this hypothesis needs to be explored in a further randomised trial.

*Andrew D Weeks, Samuel Ononge

--------------------------------------------------------------------------------

Department of Obstetrics and Gynaecology, Makerere University, PO Box 7072, Kampala, Uganda (e-mail:[email protected])

1 Sheth SS, Chalmers I. Magnesium for preventing and treating eclampsia: time for international action. Lancet 2002; 359: 1872-73. [Text]

2 The Magpie Trial Collaborative Group. Do women with pre-eclampsia, and their babies, benefit from magnesium sulphate? The Magpie Trial: a randomised placebo-controlled trial. Lancet 2002; 359: 1877-90. [Text]

Sir--In their Commentary,1 Shirish S Sheth and Iain Chalmers state that "5 years after the report of the Eclampsia Trial, the importance of magnesium sulphate was formally acknowledged in a document published by WHO, UNFPA, UNICEF, and the World Bank".2 Further statements and suggestions express that "publication cannot ensure that magnesium sulphate is available to women whose lives may depend on it", and that "WHO could make sure that evidence about the important beneficial effects of the drug is disseminated effectively to ministries of health, and that it is registered and available in all countries".

Having worked in WHO for many years, I would like to add that promotion of magnesium sulphate is one of the major strategies of the Department of Reproductive Health and Research. For example, magnesium sulphate is actively promoted in countries and Regional Offices through a multitude of workshops, meetings, and consultations. Such promotion began with the issuing of the Mother-Baby Package in 1994,3 in which an addendum was added referring to the results of the Eclampsia Trial. In addition to formal publications, magnesium sulphate is promoted in WHO's Safe Motherhood newsletters and advocacy brochures. Magnesium sulphate is one of the "best practices" recommended in WHO's Reproductive Health Library--an evidence-based CD-ROM actively promoted worldwide through seminars, workshops, and meetings. Once magnesium sulphate was proven to be the drug of choice for the treatment of eclampsia, WHO made sure to add it to their Essential Drug List. Support to countries to ensure that these drugs are made available is a major approach of WHO's Department on Essential Drugs and Medicines Policy.

Although I agree with the authors that a publication alone cannot in itself change practices, I would like to reassure your readers that WHO has gone beyond publications and is actively promoting magnesium sulphate in securing "universal access to quality care".4

Richard Guidotti

--------------------------------------------------------------------------------

Department of Reproductive Health and Research, World Health Organization, 1211 Geneva 27, Switzerland (e-mail:[email protected])

1 Sheth SS, Chalmers I. Magnesium for preventing and treating eclampsia: time for international action. Lancet 2002; 359: 1872-73. [Text]

2 WHO/UNICEF/UNFPA/World Bank. Managing complications in pregnancy and childbirth: a guide to midwives and doctors. WHO/RHR/00.7. Geneva: WHO, 2000.

3 WHO. Mother-Baby Package. WHO/FHE/MSM/94.11. Geneva: WHO, 1994.

4 Horton R. WHO: the causalities and compromises of renewal. Lancet 2002; 359: 1605-11. [Text]

Authors' reply

Sir--Baha Sibai mentions methodological concerns about our study, but fails to say what they are. Sibai has also misunderstood several important aspects of our study. We gave magnesium sulphate for 24 h only, from randomisation. Most women did not have severe pre-eclampsia; 26% met our criteria for severe pre-eclampsia at trial entry. Having an antihypertensive drug does not equate with severe pre-eclampsia. Table 4 (compliance) and table 5 (side-effects and problems at injection site) can only refer to events after randomisation. Outcomes were clinical, and defined as mortality and measures of severe morbidity after trial entry. The striking consistency of our results across the heterogeneous range of women and settings strengthens, rather than weakens, their clinical value and applicability. Total mortality is a more reliable guide to the likely true effect of magnesium sulphate than any analysis based on cause of death. As we reported, remarkably few women experienced worrying side-effects. 51 of 4999 women allocated magnesium sulphate had respiratory depression compared with 26 of 4993 allocated placebo--an excess of 25 in 9992 women (0·3%). For hypotension, the excess was 18 women (38 vs 20). This reassurance about safety applies only to dose, duration of therapy, and clinical monitoring used within the trial.

It was intentional, rather than "unfortunate", that the trial did not address the question of whether magnesium sulphate is indicated for mild pre-eclampsia. This question is of little relevance in most countries outside the USA.

Sibai also argues that, in developing countries, the morbidity and mortality associated with eclampsia occurs mostly outside hospital, but supports this statement with reference to a retrospective case series from two tertiary hospitals in North Carolina, USA.

Contrary to the statements by Sibai and Barry Walters and colleagues, the effect for magnesium sulphate in different perinatal mortality countries was uniform: there is almost complete overlap in the CIs. The lack of significance in two strata is merely due to smaller numbers. Once again, the best guide to the true effect is the overall statistic. In countries with high perinatal mortality, the proportion of women recruited with severe pre-eclampsia was higher (28% vs 24% in countries with middle-to-low perinatal mortality), and these women probably also had pre-eclampsia for longer. Severity of pre-eclampsia would seem a better criterion for deciding whether to use magnesium sulphate than country of residence. Our forthcoming cost-effectiveness analysis will further inform future decision making.

We recommended magnesium sulphate "for women with pre-eclampsia, for whom there is concern about the risk of eclampsia". Clinicians now have reliable data, on both benefits and side-effects, to present to women and to inform their decision-making.

In response to Robert Mittendorf, there are three key differences between the Magpie Trial and MagNET. MagNET recruited only 150 women. MagNET used 2-3 g/h magnesium sulphate for an unspecified duration and the median dose was 50 g,1 compared with 29 g (IQR 29-39) in the Magpie Trial. Finally, less than 0·5% of our baby deaths were postneonatal. Our follow-up study continues, and when complete we will report survival at 12-24 months. Nevertheless, as discussed in our paper, our data do not support a clinically important tocolytic effect for magnesium sulphate.

*Lelia Duley, Guillermo Carroli, Barbara Farrell, Jack Moodley, Jim Neilson, on behalf of the Magpie Trial Collaborative Group

--------------------------------------------------------------------------------

Resource Centre for Randomised Trials, Institute of Health Sciences, Headington, Oxford OX3 7LF, UK (e-mail:[email protected])

1 Mittendorf R, Roizen N, Siegler M, Khoshnood B, Lee KS. Tocolytic magnesium sulphate and paediatric mortality. Lancet 1998; 351: 293.

Hosted by www.Geocities.ws