Severe pre-eclampsia and eclampsia

James J. Walker MD, FRCOG, FRCP (Edin, Glas)

ProfessorDepartment of Obstetrics and Gynaecology, St James University Hospital, Beckett Street, Leeds, UK

The mainstay of the management of severe pre-eclampsia is early referral, stabilization of themother with antihypertensive therapy and anticonvulsants if required, full assessment of themother and the baby, and delivery on the best day in the best way. It is to be rememberedthat delivery is the long-term cure, but most women get worse after delivery and mostmaternal deaths occur postpartum. It is important that doctors have the training to be awareof the dangers of this condition, guidelines to follow and senior support. Lowering bloodpressure has been associated with a reduction in the mortality from cerebrovascular accidentand early use of antihypertensive agents is bene®cial to both mother and baby. The main causeof death is now pulmonary oedema, with renal failure a rare complication. It is important that,after delivery, vigilance is maintained and ¯uid replacement is given with care. It is better to`run them dry' than to give ¯uid replacement that may encourage pulmonary oedema. Follow-up is required with counselling about what has happened and the prospects of recurrence.

Key words: pre-eclampsia; eclampsia; hypertension; pulmonary oedema; maternal mortality;antihypertensive drugs; anticonvulsants; ¯uid therapy.

The latest con®dential enquiry into maternity mortality in the UK con®rmed thathypertension in pregnancy remains one of the two most common causes of directdeath.1 The number of deaths in the last triennium was 20, exactly the same as wasreported for the previous 3 years. However, since 1950, there has been a dramatic fallin the maternal mortality associated with pre-eclampsia/eclampsia (Figure 1), and anyfurther improvement will require even higher levels of vigilance and care. Comparedwith many other parts of the world, the complications of pre-eclampsia have steadilylessened, the incidence of eclampsia in the UK continually falling2 and now being only4.9 per 10 000 deliveries.3

These reductions in mortality and morbidity coincided with introduction of theNational Health Service and free, generalized antenatal care for all at the point ofaccess. Thus, one of the main reasons for the reduction of maternal morbidity is notthe improvement of the management of the acute disease, but the screening andintervention that comes with organized antenatal care and the improved health of thenation. This is the mainstay of obstetric practice, and any changes in antenatal caremust maintain this level of vigilance.

At the present time in the UK, most hypertensive women are diagnosed early inthe disease process when intervention can reduce the incidence of severe disease andeclampsia. A consequence of the reduction in the incidence of the severe disease is theconcomitant reduction of the experience of its management obtained by the majority

1521±6934/00/010057+15 $35.00/00 *c 2000 Harcourt Publishers Ltd.

BaillieÁ re's Clinical Obstetrics and GynaecologyVol. 14, No. 1, pp 57±71, 2000doi:10.1053/beog.1999.0063, available online at http://www.idealibrary.com on

4

Figure

1.Maternalmortalityassociated

withpre-eclam

psia/eclampsiasince

1950.

58 J. J. Walker

of obstetricians. The average consultant will only see a woman with eclampsia onceevery 4 years and, even then, usually after the event. This increases the importance ofthe development of regularly updated consensus guidelines to ensure an adequate levelof care. The Con®dential Enquiry persistently reports a high level of inappropriatemanagement. Much of the problem is related to a lack of understanding of the diseaseprocess. Pre-eclampsia is more than just hypertension.4 There are many otherproblems encountered that put the mother's life at risk (Table 1)5, and the guidelinesmust cover all aspects of the disease, particularly pulmonary oedema, which has nowbecome the main cause of maternal death and morbidity (Table 2).1 As far as the babyis concerned, perinatal mortality and morbidity are related to placental insu�ciencyand the complications of premature delivery.6 The level of maternal blood pressureand the severity of disease are less signi®cant except when they necessitate apremature delivery.

Because of these multifaceted problems, the management of severe pre-eclampsia/eclampsia requires multiple, separately targeted strategies that are part of a step-wisesystem of management of pregnancy hypertension (Table 3). There needs to be an easymethod of referral from the community into a monitoring system7 so that any a�ectedwoman can be seen early and intervention therapy commenced to reduce thematernal risk and damage associated with more severe disease.8

It is not just high blood pressure that requires intervention: other clinical signs mustalso be screened for. This is a multisystem disorder with many similarities to an

Table 1. The maternal sequelae of severe hypertension in pregnancy.

MajorCerebrovascular accident Occipital lobe blindnessConvulsions Pulmonary oedemaHaemolysis, elevated liver enzymes and lowplatelets (HELLP syndrome)

Aspiration syndromeRenal failure

Liver failure Deep venous thrombosisLiver rupture Postpartum haemorrhage

Complications of treatmentSedation and aspiration Caesarean sectionFluid overload

Table 2. The causes of maternal death documented in the Con®dential Enquiries into Maternal Mortalityover the past 23 years.

Year Cerebral Pulmonary Hepatic Renal Other

70±72 25 8 5 3 673±75 23 7 14 1 476±78 21 4 5 0 379±81 17 8 8 0 382±84 21 3 0 0 185±87 11 12 1 1 288±90 14 10 1 0 291±93 5 11 0 1 394±96 7 9 3 0 1

Total 144 72 37 6 25

Severe pre-eclampsia and eclampsia 59

in¯ammatory disease.9 This leads to a multiplicity of clinical signs and symptoms. Thedevelopment of proteinuria, increasing oedema or the addition of clinical symptomssuch as headache, visual disturbance or abdominal pain should lead to immediateadmission and further investigation. Similarly, any pregnant woman with a presentingsymptom of headache, abdominal pair or nausea, or who is just feeling unwell, shouldhave her blood pressure checked and her urine tested for proteinuria as pre-eclampsiacan be present in many guises. If there is concern, immediate referral to the obstetricunit is mandatory to allow full assessment and a consideration of intervention therapyand the appropriate timing of delivery.

IMMEDIATE ASSESSMENT AND CARE

On ®rst presentation, it is the maternal hypertension that is the primary concern andneeds to be controlled. It was once thought that lowering the blood pressure waspotentially dangerous for the baby, and there were reports to support this.10,11 Since1980, there has been an increase in the use of antihypertensive drugs byobstetricians12, and this has been associated with a reduction in maternal and neonataldeath from pre-eclampsia, particularly in the number of maternal deaths related tocerebral pathology (see Table 2 above).1

A recent meta-analysis of antihypertensive therapy in pregnancy has shown thatearly treatment reduces not only the incidence of hypertensive crisis, but also neonatalcomplications such as respiratory distress syndrome.13 Therefore, an early aggressiveuse of antihypertensive therapy to stabilize the maternal condition would appear to bebene®cial not only to the mother, but also to her baby. The meta-analysis does notcompare di�erent drugs. It probably does not matter which drugs are used as long asattendant sta� are experienced in their use and the drugs are thought to be safe.

The most common drugs used in the UK are hydralazine, methyldopa, labetalol andnifedipine.11,12 No one therapy can control all patients, and increasing doses andcombinations of drugs are usually required.14 Care should be taken not to reduce theblood pressure too quickly as side-e�ects can occur. Intravenous drugs should beavoided if the woman can tolerate oral therapy. The latter is commonly used inhypertensive crises in non-pregnant patients.15 In randomized trials, oral labetalol16

and nifedipine17 have been shown to be as e�ective as and more predictable thanparental hydralazine. Oral therapy appears to provide a smoother, more predictable

Table 3. The step-wise management of pregnancy hypertension.

ScreeningScreening women for the risk or signs of hypertension at the antenatal clinicRefer those at risk for monitoring in an antenatal day unit

Prior to deliveryFollow regularly updated management guidelinesEarly involvement of senior medical sta�Lower blood pressure if requiredProphylactic steroid administration if the pregnancy is of less than 34 weeks' gestationConsider the need for magnesium sulphateContinuing monitoring for signs of disease progressionDelivery of the baby on the best day, in the best way and in the best place

60 J. J. Walker

fall in blood pressure. The use of intravenous hydralazine is associated with a suddendrop in blood pressure, with abnormalities of fetal heart rate.18 Studies have suggestedthat this may be due to a combination of acute vasodilatation in women with a reducedplasma volume. To reduce this complication, hydralazine has been used in combinationwith plasma expansion, but the results are no better than those with oralantihypertensive therapy alone.19 Oral labetalol and nifedipine appear to be at leastas good as hydralazine16,18 without the need for plasma expansion and invasivemonitoring.

In Yorkshire, there is a consensus protocol that uses oral labetalol with a startingdose of 200 mg three times daily (Table 4). This can be steadily increased to amaximum of 300 mg four times a day. If control is not achieved, 10 mg oral nifedipineretard can be given twice a day, increasing to a maximum of 60 mg per day. Thesedrugs have been shown in animal experiments not to reduce placental or maternalperipheral blood ¯ow in human pregnancy.11 Nifedipine can also be used as theprimary therapy, but care should be taken as, especially with sublingual usage, suddenfalls in blood pressure can sometimes be seen.20 The oral route is safer and as e�ective.The combination of nifedipine and magnesium sulphate was thought to be potentiallydangerous because of synergism, but many centres have had extensive experience withthis combination without any problems. If oral therapy is unsuccessful, intravenouslabetalol 50 mg given by slow bolus injection followed by labetalol infusion will lowerthe blood pressure in the majority of patients.21

The aim of therapy is not to normalize the level, but to stop the rise of bloodpressure and achieve a moderate fall, a drop of around 10 mmHg diastolic pressurebeing the primary aim. Once this has been achieved, there is usually a need for anincrease in drug dosage every few days as this is a progressive disease. Since womenwith severe pre-eclampsia/eclampsia often present acutely out of clinic hours, it isimportant that all sta� are familiar with the management guidelines and the aims oftherapy. Blood pressure control must, if at all possible, be achieved prior to delivery,but especially before intubation for general anaesthetic, before caesarean section as anacute blood pressure rise can occur.22

Table 4. A suggested antihypertensive regimen.

Acute managementLabetalol 200 mg orally repeated hourly until control is achievedorNifedipine 10 mg orally (not sublingual)orLabetalol 50 mg by slow intravenous bolusfollowed byLabetalol infusion starting at a rate of 60 mg per hour doubling every 15 minutes until control is achievedor a maximum of 480 mg is reached

Starting daily dose for chronic therapyLabetalol 200 mg three times dailyincreasing toLabetalol 300 mg four times dailywith the addition ofNifedipine retard 10 mg twice a dayincreasing toNifedipine retard 60 mg per day in divided doses

Severe pre-eclampsia and eclampsia 61

THE USE OF ANTICONVULSANT AGENTS

Although eclampsia is a common presentation in many parts of the world, it is now arare occurrence in the UK, occurring in 1 in 2000 of all pregnant women.3 If people dosu�er a convulsion, the majority of patients convulse only once.5 The risks to themother and baby from eclampsia appear to be more related to the degree of pre-existing pre-eclampsia than to the eclampsia itself. Although evidence of prodromalsigns, particularly headache and abdominal pain, is present in many women prior tothe convulsion, about 20% of patients convulse unexpectedly, often with normal bloodpressure, and no sign is speci®c for the development of eclampsia.5 The widely usedclinical assessment of hyperre¯exia has never been demonstrated as being an accuratepredictor even though it is widely used as such. However, it is of great value as aclinical measure of magnesium toxicity. If the re¯exes are absent, magnesium toxicityshould be suspected and the medication stopped.

If convulsions have occurred, anticonvulsant therapy is required. A large multi-centre study has shown that magnesium sulphate is superior to both phenytoin anddiazepam.5 Both the reconvulsion and maternal death rates were found to besigni®cantly reduced compared with either diazepam or phenytoin. Although there isnow no doubt what preparation to use (Table 5), it is important to note that notreatment will completely prevent seizures, reconvulsion rates varying between 5%and 20%.5

Both the intravenous and intramuscular regimens used in the study were similar tothose previously published. There is no evidence to suggest that one route is superiorto any other. Both regimens use a slow intravenous loading dose of 5 g magnesiumsulphate that can be used to treat the convulsion. This should be given over 20 minutesas a faster bolus may produce cardiac arrest. There is no need to use diazepam initiallyto stop the seizure. Whether the intramuscular or the intravenous maintenancedosage is used depends on the preference and facilities of the centre. The intra-muscular route has the advantage of ease of use, although intravenous therapy mayprovide better blood levels. Many obstetricians are concerned about the possibility ofmagnesium toxicity. The study cited here used clinical evaluation alone, showing thatthe toxicity did not occur and the side-e�ects from magnesium were no worse than

Table 5. Anticonvulsant therapy to be used in the eclamptic patient. Either the intravenous or theintramuscular regimen can be used.

Intravenous regimenLoading dose of 5 g intravenously over 20 minutesInfusion of 1 g per hour for 24 hours

Intramuscular regimenLoading dose of 4 g intravenously over 20 minutes5 g into each buttock intramuscularlyA further 5 g intramuscularly every 4 hours(Provided the respiratory rate was over 16 per minute, the urinary output over 25 ml per hour and kneejerks were present)

For recurring convulsionsA further 2±4 g given intravenously over 10 minutes

If magnesium toxicity is suspectedSuggested by the absence of re¯exes1 g calcium gluconate should be given

62 J. J. Walker

those from diazepam, being signi®cantly less than those found with phenytoin.Therefore, using these dose regimens, there is no need to check magnesium levels orto have this facility available at all times. Magnesium levels are useful in themanagement of treatment failure to test whether an increased dosage might be ofbene®t.

There is thus good evidence to support the e�cacy of magnesium sulphate therapyfor women with eclampsia.23 However, the role of prophylactic magnesium sulphate inwomen with pre-eclampsia is less clear. If in doubt, it is probably safer to give it thannot, but magnesium is not without its own risks and, in some series, the only deathshave arisen from magnesium toxicity. The risk of eclampsia in women with pre-eclampsia is probably around 1 in 200, and there is no evidence that magnesium willreduce this.24 However, if a prophylactic anticonvulsant is to be used, magnesiumsulphate is the drug of choice.25 There have been concerns expressed over the e�ect ofmagnesium therapy on the e�cacy of labour. Compared with phenytoin, magnesiumsulphate given for the intrapartum treatment of pregnancy-induced hypertension didnot signi®cantly a�ect labour outcome26, but did necessitate a higher dose ofoxytocin.27

WHAT TO DO ONCE THE PATIENT HAS BEEN STABILIZED

Once the patient has been stabilized, a plan of action should be made by a seniordoctor. After convulsion has occurred, the continuation of pregnancy cannot bejusti®ed in the UK setting, but there is no hurry to deliver the baby. The motherneeds to be assessed and stabilized prior to delivery as this will reduce the severity ofher postpartum exacerbation. The fetus may present with acute distress after amaternal convulsion, but will quickly recover. A carefully planned delivery over thenext few hours should be carried out so that all the relevant carers are in place andready to look after both mother and baby. Transfer of the mother to a moreappropriate setting may further improve the ability to care for her and her baby afterdelivery.

In the absence of convulsions, prolongation of the pregnancy is possible, in mostcases, for an average of 15 days.11 This has been shown to improve the outcome for thebaby without detriment to the mother. Rushing the delivery in the maternal interestmay not be of bene®t to the mother and can lead to increased morbidity for the baby,as it is the gestation at delivery more than any other parameter that in¯uences thebaby's outcome.6 Even in cases of HELLP syndrome, there is no di�erence in outcomecompared with gestation-matched controls with severe pre-eclampsia alone19 or thoseborn prematurely for other reasons.28 However, it is important that the mother'scondition is stable so that the prolongation of pregnancy does not jeopardize her life.The situation should be constantly reassessed and the management plan reviewed by asenior doctor (Table 6). A woman who is stable may not remain so. The plan ofmanagement is more along the lines of `not needing delivery now' rather than `let'sdeliver her tomorrow'. If delivery within the next few days is thought to be apossibility and the gestation is less than 34 weeks, prophylactic steroids should begiven to induce fetal lung maturity. Two doses of dexamethasone 12 mg intra-muscularly given 12 hours apart is the regimen used in Yorkshire guidelines. The aimis to prolong the pregnancy for at least 48 hours to achieve maximum e�ect, althougheven if the delivery is earlier, some bene®t will be gained.

Severe pre-eclampsia and eclampsia 63

Delivery should be carried out before 48 hours have elapsed if there are strongmaternal or fetal reasons. After 48 hours, if a further prolongation for more than aweek is not thought to be achievable, delivery should be carried out when the motheris stable and the baby is well and has received maximum bene®t from the prophylacticsteroids. Transfer of the mother prior to delivery may be necessary for the optimumcare of the mother or her newborn, possibly premature baby.

ASSESSMENT OF THE HYPERTENSIVE WOMAN (TABLE 6)

Antihypertensive therapy lowers blood pressure and should not be expected to doanything else. It is important to emphasize that if pregnancy is going to be continued,close monitoring of the mother is required using blood pressure monitoring, protein-uria, maternal uric acid levels29 and platelet count30 to monitor disease progression.Fetal well-being should be followed using cardiotocography and ultrasound for fetalgrowth, umbilical artery Doppler velocity waveforms and liquor volume estimations.

Blood pressure measurement

Abnormality of blood pressure is, for obvious reasons, the primary diagnostic criterionin hypertension in pregnancy. However, blood pressure is a variable and can ¯uctuateon a minute-to-minute basis.31 A single, elevated blood pressure reading thus needs tobe con®rmed before management decisions are made.14 The average of severalreadings allows a more accurate estimation of the true blood pressure measurement.Automatic blood pressure recorders often measure a diastolic blood pressure severalmmHg lower than a normal sphygmomanometer, and this should be taken intoaccount when these machines are being used in the acute situation.

Table 6. Management plan for the continuing care of the woman with severe pre-eclampsia.

Maternal managementKeep blood pressure controlled with antihypertensive drugsUse the average of four blood pressure readings for monitoringAt least twice-weekly serum uric acid level, platelet count and liver function testsTest for the presence of proteinuria with or without quanti®cation

Fetal managementProphylactic steroids given if the gestation is below 34 weeksInitial ultrasound assessment of fetal weightDoppler ultrasound assessment of umbilical blood ¯ow velocity twice a weekDaily cardiotocographsAt least twice-weekly ultrasound for liquor volume

Care after deliveryContinuing close monitoring of the mother within a closely monitored environment by experienced

carersCareful ¯uid balance and early use of diuretics if requiredReduction of antihypertensive agents as indicatedStop anticonvulsant therapy after 48 hours if stable

Follow-upLong-term follow-up to make sure that the problem with blood pressure resolvesDiscussion concerning what has gone on and its signi®cance for the futureReferral for counselling to local experts and future planning

64 J. J. Walker

The presence of hypertension can, in itself, put the mother at increased risk ofcerebrovascular accident, but it is the presence of systemic involvement that is thehallmark of pre-eclampsia as it implies that the hypertension is part of a progressivedisorder. The major morbidity factors such as pulmonary oedema, liver failure andcoagulation defects are associated with these systemic changes.

Proteinuria

The testing of the urine for proteinuria is the simplest and most available test. Theassociation of protein with eclampsia has been known for over 100 years. Urine sticktesting will overestimate the presence of proteinuria, and this should always becon®rmed by a 24-hour collection for quanti®cation. The importance of proteinuria isthat it helps to con®rm the diagnosis of pre-eclampsia with the concomitant increasedrisks.6 The proteinuria is a sign not of renal damage but of the presence of a capillaryleak that is associated with the development of oedema ¯uid throughout the body.Once protein is present, a change in the amount does not re¯ect increasing severity ofdisease and should not in¯uence the decision to deliver on its own, but it can signal anincreased risk of pulmonary oedema. Lowering blood pressure is associated with areduction in proteinuria; similarly this does not imply that the risks have diminished,but instead re¯ects a reduced renal perfusion pressure.

Uric acid measurement

Uric acid rises in pre-eclampsia and is a better predictor of fetal morbidity than isblood pressure itself.29 The reason for the rise in uric acid is not totally clear butappears partly to be caused by renal tubular function impairment, reducing itsexcretion. This renal involvement is probably the result of renal medularly ischaemiaand is not a sign of renal failure, urea and creatinine levels both remaining normal untillate in the disease process. Some of the rise in uric acid is due to increased productionsecondary to tissue damage caused by ischaemia. Therefore, uric acid is partly asurrogate marker of tissue involvement in pre-eclampsia and a useful adjunct to othermonitoring tests. Serial testing is of particular bene®t as a rising level is indicative ofdisease progression and increased risk.

Platelet count

Platelet counts fall in pre-eclampsia and are associated with progressive disease andworsening outcome.30 In recent years, the combination of Haemolysis, Elevated Liverenzymes and Low Platelets has been described as the HELLP syndrome.32 Althoughthis is undoubtedly associated with increased maternal morbidity, it is not a newphenomenon, but instead a new description of previously described and acceptedpathological ®ndings and is a marker of severe systemic involvement. Serial samplesshould be taken to observe developing trends. Falling platelet counts are of concernand can give some guidance on the timing of the delivery. If a platelet count of below100 is found, a full coagulation screen, blood ®lm and liver function tests should becarried out as this can give further information about the progression of the diseaseand the future maternal risk.28 The presence of HELLP syndrome is particularlyworrying and usually implies the need for delivery, although cases of conservativemanagement have been reported.33 If the platelet count is higher than 100, it is

Severe pre-eclampsia and eclampsia 65

unlikely that the coagulation screen will be abnormal. In such a case, there is no needto carry out a coagulation screen and no risk in the use of epidural analgesia.

Liver function tests

These are not true liver function tests, but are measurements of enzymes containedlargely in liver cells. The disruption of these cells leads to release of enzymes into thecirculation and hence a rise in their levels. The most sensitive change is seen inaspartate transaminase (Ast), but there is also a rise in alanine transaminase (Alt),which is easier to measure in an emergency laboratory. A rise in liver enzyme levels isalways signi®cant and needs to be followed carefully, Severe liver impairment can beassociated with liver swelling, which causes the epigastric pain and can lead to liverrupture, a rare but potentially fatal outcome. In pre-eclampsia, disseminated intra-vascular coagulation is a rare complication in the absence of an abruption. Thecoagulation defects seen are usually related to HELLP syndrome, liver functionimpairment leading to an intrinsic coagulation defect and low platelet count. The lossof red cells is due to intravascular haemolysis caused by an increased red cell fragilitypossibly associated with free radical damage.34

ASSESSMENT OF THE FETUS (TABLE 6)

The assessment of the maternal disease does not give accurate information about thewell-being of the baby. Once the mother has been stabilized, speci®c fetal assessmenttests should be carried out. The main risks to the fetus are placental insu�ciency andprematurity. Investigations are directed at the assessment of fetal health, fetal growthand placental blood ¯ow.

Cardiotocography

Cardiotocography is the mainstay of fetal monitoring in most units. It gives inform-ation concerning fetal well-being at that time, but has little predictive value. It can berepeated regularly and easily without the need for expensive equipment or highlyskilled personnel, and can be particularly reassuring to the mother herself. Althoughtracings from fetuses less than 30 weeks' gestation can be misleading, it is the change inthe trace that is most informative.

As long as the tracing is reactive with good variability, there is no evidence of fetalhypoxia. If accelerations or variability are lost, this implies a deterioration in the fetalstate, and the presence of decelerations must be seen as sinister at any time.

Cardiotocography should be repeated as often as thought necessary depending onthe severity of the situation, but taking daily tracings from women hospitalized onantihypertensive therapy would seem sensible.

Fetal ultrasound assessment

An ultrasound assessment of fetal size is valuable as a one-o� measurement to assessfetal growth at the time of the initial admission. It also gives the paediatricians usefulinformation about the expected fetal weight if the baby is to be delivered, as well asallowing estimates of the chances of survival of the baby that can be used in adiscussion with the parents about management decisions. Serial growth studies are of

66 J. J. Walker

less value as they should not be carried out more often than every 2 weeks and theaverage length of prolongation of pregnancy in severe disease is only 15 days.35 Liquorvolume estimation is of less value in earlier gestations where close fetal assessment isusually being carried out36, although reduced liquor volume is associated with fetalgrowth restriction. Serial estimations are of value as a decreasing volume of liquor canbe indicative of increasing fetal compromise.

Placental blood ¯ow

Reduced uterine artery Doppler end diastolic ¯ow is associated with an increasedincidence of placental insu�ciency and pre-eclampsia.37 It can be used as a screeningtest in normal pregnancy. In established disease, it is of less value38, but it still can givean indication of whether fetal growth restriction is likely to be present. UmbilicalDoppler artery assessment is of greater value39 where serial investigations can be usedto follow pregnancies under treatment. If end-diastolic ¯ow is present, therapy can becontinued in an attempt to prolong the pregnancy as the baby will normally be able totolerate a prolongation of pregnancy and maternal antihypertensive therapy.39

Monitoring protocol (Table 6)

Although single tests can give information about maternal and fetal well-being, serialsampling adds to this by demonstrating the rate of progression. Proteinuria should bequanti®ed and serial uric acid levels and platelet counts taken at least on a twice-weekly basis. A cardiotocograph should be carried out daily, and at least twice-weeklyultrasound estimations of liquor volume and umbilical artery Doppler velocitywaveforms should be made. A particular area of concern is that of liver impairmentand the development of HELLP syndrome. Liver function tests or a minimum of atleast Alt or Ast level should be carried out on a twice-weekly basis. This, along withthe platelet count, will screen for early signs. If the liver function tests and plateletcount remain normal, HELLP syndrome and coagulation abnormalities are notpresent. If HELLP syndrome is thought to be developing, a full coagulation screen anda blood ®lm, looking for fragmented red cells and evidence of haemolysis, should becarried out.

PLANNING THE DELIVERY

Delivery is the ultimate cure for pre-eclampsia, the timing of delivery a�ecting theoutcome for both mother and baby. However, most maternal deaths occur post-partum. A rushed delivery in an unstable patient adds to rather than reduces her riskand is often associated with maternal mortality.1 A delay in delivery in a sick patientmay also be associated with maternal death. Any transfer of thewoman should be at anearly stage while she is stable, and communication should occur between seniordoctors in each hospital.

If delivery has been decided upon before 32 weeks, it should be carried out byelective caesarean section unless this is more dangerous for the mother. A longinduced labour can add to the problems of ¯uid overload and the di�culties of bloodpressure control, vaginal delivery being successful in fewer than 50% of cases.40 If itoccurs after 34 weeks' gestation, a vaginal delivery is aimed for, particularly when theprobability of this being successful has been greatly increased by the use of vaginal

Severe pre-eclampsia and eclampsia 67

prostaglandins. Antihypertensive therapy should be continued throughout labour tohelp to control blood pressure.

The choice of anaesthetic in caesarean section is important. General anaesthetic canadd to the risk to the mother since intubation and extubation can cause a rise in bothsystolic and diastolic blood pressure as well as heart rate.22 An epidural or spinal blockwould seem far safer, allowing delivery to take place with the patient awake, and alsospeeds recovery. Care should be taken that the ¯uid load is kept to a minimum, and acolloid solution may be preferable to a crystalline one which passes quickly across thecapillaries into the interstitial space. Women with pre-eclampsia already have anincrease of ¯uid in this compartment and are extremely sensitive to excessive ¯uid.41,42

During labour, an epidural is also bene®cial, but not to lower blood pressure as thisis a pathological rather than a physiological e�ect. By allowing adequate pain relief, itcan stop the rise in blood pressure often associated with labour. It also allows aplanned delivery and easy transition to caesarean section if this is deemed necessary.There is no reason routinely to shorten the second stage of labour by instrumentaldelivery as long as the mother is pushing well and remains stable. After delivery, drugscontaining ergometrine should not be used in the third stage of labour, intravenousSyntocinon being the preferred option.

POST-PARTUM CARE

Since most maternal death and signi®cant morbidity occur after delivery1,2, continuedclose monitoring is required in a suitable environment. An initial improvement isoften seen, but around 60% of women will to some extent worsen within 48 hours.Women with severe disease should be kept in the high-dependency areas and not betransferred to the postnatal wards until their parameters begin to return to normal.

With the increased use of blood pressure control and the reduction of cerebralbleeding as a cause of death, the main cause of maternal mortality in this condition isnow pulmonary oedema (see Table 2 above)1, probably caused by various underlyingmechanisms within the woman with pre-eclampsia, including a low albumin level andan increased capillary leakage.41,42 Although labetalol has been blamed, since it mayinterfere with cardiac output, studies have shown that it is an increase in interstitial¯uid volume rather than a reduction in cardiac function that is at fault.42 The risk ofpulmonary oedema is aggravated by exogenous ¯uid given in the belief that thesewomen are at risk of renal failure.

Renal failure is, in fact, unusual in pure pre-eclampsia and is usually associated withadditional problems such as haemorrhage and sepsis.43 If it does occur, the morbidityand mortality are low (see Table 2 above) as it is usually an acute tubular necrosis thatcan be managed easily.43 Post-delivery, a relative oliguria is not uncommon, occurringin about 30% of patients with severe disease. This does not require therapy as urinaryoutput will recover in its own time. If there is any doubt about renal function, urinaryosmolality can be checked. If the urine is concentrated, renal function is satisfactoryand the reduced output is related to a reduced renal perfusion, which will graduallyimprove over time. If the urine is not concentrated, renal failure is present and a renalfailure regimen should be put in place. There is no evidence, however, that oliguria,especially in the presence of concentrated urine, leads to renal failure. Fluid challengesare potentially dangerous in pre-eclampsia as much of the ¯uid will be lost from thevessels into the interstitial ¯uid, aggravating the existing tissue oedema. If ¯uid is to begiven, colloid should be used up to a maximum of 500 ml and not repeated unless an

68 J. J. Walker

adequate urinary response has been achieved. In Yorkshire, our initial analysis of the®rst 40 women with severe pre-eclampsia managed by our consensus guidelinesshowed that the only women who experienced pulmonary oedema were those given a¯uid challenge because of a reduced urinary output.

Invasive monitoring is usually not necessary, and centre venous pressure (CVP) linescan be misleading.44 In pre-eclampsia, pulmonary oedema, because of increasedinterstitial ¯uid, can occur in the presence of a low CVP. If a central line is used, it isimportant that the CVP level is no higher than 5 mmHg or 7 cm H2O. If the CVP ishigher than this, diuretics should be used to reduce it. One of the best methods ofassessing pulmonary oedema is the continuous measurement of oxygen saturationusing a pulse oximeter. If there is a requirement for facial oxygen to maintain anadequate oxygen saturation, there is probably pulmonary oedema present and adiuresis will improve the oxygen carriage.

It is far safer and more sensible to run a patient `dry' and restrict intravenous ¯uidsthan to run the risk of pulmonary overload. Total ¯uid replacement should not exceedan hourly rate of 80 ml. Most of the problems of ¯uid overload begin around 16 hoursafter delivery, severe problems occurring between 24 and 48 hours. This is associatedwith a failure of the normal postpartum diuresis. If the patient is in positive balance orthere is evidence of pulmonary oedema, 40 mg frusemide should be given, whichusually produces an adequate urine output but can be repeated if necessary. This canbe followed by 20 g mannitol to reduce tissue ¯uid. This will lower the probability ofpulmonary oedema, increase oxygen saturation, reduce cerebral oedema and improveblood pressure control.

Therefore, after delivery, management should be aimed at maintaining bloodpressure control and carefully monitoring the ¯uid balance. If convulsions occur, anti-convulsant therapy can be added. The majority of postnatal convulsions occur withinthe ®rst 24 hours, so anticonvulsant therapy is usually continued for 48 hours post-delivery. If the patient is well, especially if a prophylactic anticonvulsant is being usedin the absence of any convulsions, this therapy can be stopped at 24 hours. Antihyper-tensive therapy should be reduced after delivery depending on the blood pressure.There may be a signi®cant drop within the ®rst 24 hours with a rise again after48 hours. Antihypertensive drugs may be necessary for some weeks after delivery.

It has been claimed that postpartum curettage can be curative in women withpersistent disease.45 However, follow-up studies have demonstrated that nifedipinetherapy is as good as uterine curettage in accelerating recovery from severe pre-eclampsia, although uterine curettage may be more e�ective in rapidly resolving thethrombocytopenia associated with severe pre-eclampsia.35 An acute re-presentationcan occur some days after apparent improvement, although this is uncommon.46

Women with persistent headaches, malaise, nausea and vomiting are more likely tohave a recurrence of their symptoms and should be assessed accordingly.47

FOLLOW-UP

The patient should be followed up until normal blood pressure has returned. Thediagnosis of the underlying cause of the hypertension in pregnancy may not beapparent until after the pregnancy is over. If blood pressure has not returned tonormal by 6 weeks postpartum, the patient should be referred for further investi-gation as the cause is unlikely to have been pre-eclampsia. A postnatal visit should becarried out to assess the recovery of the woman and to discuss the signi®cance of what

Severe pre-eclampsia and eclampsia 69

happened and what might happen in the future. It should be remembered that womenwith severe pre-eclampsia have a 30% chance of a recurrent problem, but it is usuallyless severe and occurs 2±3 weeks later than in the previous pregnancy.

REFERENCES

* 1. Department of HealthWelsh O�ce, Scottish Home and Health Department, Department of Health andSocial Security, Northern Ireland.Why Mothers Die. Report on Con®dential Inquiries into Maternal Deaths inthe United Kingdom 1994±96. London: HMSO, 1999.

2. Leitch CR, Cameron AD & Walker JJ. The changing pattern of eclampsia over a 60-year period. BritishJournal of Obstetrics and Gynaecology 1997; 104: 917±922.

* 3. Douglas KA & Redman CWG. Eclampsia in the United Kingdom. British Medical Journal 1994; 309:1395±1400.

* 4. Roberts JM & Redman CWG. Preeclampsia ± more than pregnancy-induced hypertension. Lancet 1993;341: 1447±1451.

* 5. Duley L, Carroli G, Belizan J et al. Which anticonvulsant for women with eclampsia ± evidence from thecollaborative eclampsia trial. Lancet 1995; 345: 1455±1463.

6. Friedman EA & Ne� RK. Pregnancy, outcome as related to hypertension, edema, and proteinuria.Perspectives in Nephrology and Hypertension 1976; 5: 13±22.

7. Walker JJ. Day-care obstetrics. British Journal of Hospital Medicine 1993; 50: 225±226.8. Saphier CJ & Repke JT. Hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome: a

review of diagnosis and management. Seminars in Perinatology 1998; 22: 118±133.9. Walker JJ. Antioxidants and in¯ammatory cell response in preeclampsia. Seminars in Reproductive

Endocrinology 1998; 16: 47±55.10. Vink GJ, Moodley J & Philpott RH. E�ect of dihydralazine on the fetus in the treatment of maternal

hypertension. Obstetrics and Gynecology 1980; 55: 519±522.*11. Walker JJ. Hypertensive drugs in pregnancy. Antihypertension therapy in pregnancy, preeclampsia, and

eclampsia. Clinics in Perinatology 1991; 18: 845±873.12. Hutton JD, James DK, Stirrat GM et al. Management of severe preeclampsia and eclampsia by UK

consultants. British Journal of Obstetrics and Gynaecology 1992; 99: 554±556.*13. Collins R & Duley L. Any antihypertensive therapy for pregnancy hypertension. In Enkin MW, Keirse

MJNC, Renfrew MJ & Neilson JP (eds). Pregnancy and Childbirth Module, Cochrane Database ofSystematic Reviews. Review No. 04426, Disk Issue 1. Oxford: Update Software, 1994.

14. Walker JJ. Care of the patient with severe pregnancy induced hypertension. European Journal ofObstetrics, Gynaecology, and Reproductive Biology 1996; 65: 127±135.

15. Abdelwahab W, Frishman W & Landau A. Management of hypertensive urgencies and emergencies.Journal of Clinical Pharmacology 1995; 35: 747±762.

16. Walker JJ, Greer I & Calder AA. Treatment of acute pregnancy-related hypertension: labetalol andhydralazine compared. Postgraduate Medical Journal 1983; 59: 168±170.

17. Kwawukume EY & Ghosh TS. Oral nifedipine therapy in the management of severe preeclampsia.International Journal of Gynaecology and Obstetrics 1995; 49: 265±269.

18. Visser W & Wallenburg HCS. A comparison between the hemodynamic-e�ects of oral nifedipine andintravenous dihydralazine in patients with severe preeclampsia. Journal of Hypertension 1995; 13: 791±795.

19. Abramovici D, Friedman SA, Mercer BM et al. Neonatal outcome in severe preeclampsia at 24 to 36weeks' gestation: does the HELLP (hemolysis, elevated liver enzymes, and low platelet count) syndromematter? American Journal of Obstetrics and Gynecology 1999; 180: 221±225.

20. Impey L. Severe hypotension and fetal distress following sublingual administration of nifedipine to apatient with severe pregnancy-induced hypertension at 33 weeks. British Journal of Obstetrics andGynaecology 1993; 100: 959±961.

21. Mabie WC, Gonzalez AR, Sibai BM & Amon E. A comparative trial of labetalol and hydralazine in theacute management of severe hypertension complicating pregnancy. Obstetrics and Gynecology 1987; 70:328±333.

22. Ramanathan J, Sibai BM, Mabie WC et al. The use of labetalol for attenuation of the hypertensiveresponse to endotracheal intubation in preeclampsia. American Journal of Obstetrics and Gynecology 1988;159: 650±654.

23. Witlin AG & Sibai BM. Magnesium sulfate therapy in preeclampsia and eclampsia. Obstetrics andGynecology 1998; 92: 883±889.

70 J. J. Walker

24. Moodley J & Moodley VV. Prophylactic anticonvulsant therapy in hypertensive crises of pregnancy ± theneed for a large, randomized trial. Hypertension in Pregnancy 1994; 13: 245±252.

*25. Lucas MJ, Leveno KJ & Cunningham FG. A comparison of magnesium-sulfate with phenytoin for theprevention of eclampsia. New England Journal of Medicine 1995; 333: 201±205.

26. Leveno KK, Alexander JM, McIntire DD & Lucas MJ. Does magnesium sulfate given for prevention ofeclampsia a�ect the outcome of labor? American Journal of Obstetrics and Gynecology 1998; 178: 707±712.

27. Witlin AG, Friedman SA & Sibai BM. The e�ect of magnesium sulfate therapy on the duration of labor inwomen with mild preeclampsia at term: a randomized, double-blind, placebo-controlled trial. AmericanJournal of Obstetrics and Gynecology 1997; 176: 623±627.

28. Magann EF, Perry KGJ, Chauhan SP et al. Neonatal salvage by week's gestation in pregnanciescomplicated by HELLP syndrome. Journal of the Society for Gynecologic Investigation 1994; 1: 206±209.

29. Redman CW, Beilin LJB & Wilkinson BH. Plasma urate measurements in predicting fetal death inhypertensive pregnancy. Lancet 1976; 1: 1370±1374.

30. Redman CW, Bonnar J & Beilin L. Early platelet consumption in pre-eclampsia. British Medical Journal1978; 1: 467±469.

31. Halligan A, Shennan A, Thurston H et al. Ambulator blood-pressure measurement in pregnancy ± thecurrent state-of-the-art. Hypertension in Pregnancy 1995; 14: 1±16.

32. Weinstein L. Syndrome of hemolysis, elevated liver enzymes, and low platelet count: a severeconsequence of hypertension in pregnancy. American Journal of Obstetrics and Gynecology 1982; 142:159±167.

33. Visser W & Wallenburg HCS. Temporising management of severe preeclampsia with and without theHELLP-syndrome. British Journal of Obstetrics and Gynaecology 1995; 102: 111±117.

34. Spickett CM, Reglinski J, SmithWE et al. Erythrocyte glutathione balance and membrane stability duringpreeclampsia. Free Radical Biology and Medicine 1998; 24: 1049±1055.

35. Magann EF, Bass JD, Chauhan SP et al. Accelerated recovery from severe preeclampsia: uterine curettageversus nifedipine. Journal of the Society for Gynecologic Investigation 1994; 1: 210±214.

36. Kahra K, Draganov B, Sund S & Hovig T. Postpartum renal failure: a complex case with probablecoexistence of hemolysis, elevated liver enzymes, low platelet count, and hemolytic uremic syndrome.Obstetrics and Gynecology 1998; 92: 698±700.

*37. Bower S, Bewley S & Campbell S. Improved prediction of preeclampsia by 2-stage screening of uterinearteries using the early diastolic notch and color Doppler imaging. Obstetrics and Gynecology 1993; 82:78±83.

38. Hanretty KP, Whittle MJ & Rubin PC. Doppler uteroplacental waveforms in pregnancy-inducedhypertension: a re-appraisal. Lancet 1988; 1: 850±852.

39. Fairlie FM. Doppler ¯ow velocimetry in hypertension in pregnancy. Clinics in Perinatology 1991; 18:749±778.

40. Nassar AH, Adra AM, Chakhtoura M et al. Severe preeclampsia remote from term: labor induction orelective cesarean delivery? American Journal of Obstetrics and Gynecology 1998; 179: 1210±1213.

41. Brown MA, Zammit VC & Lowe SA. Capillary permeability and extracellular ¯uid volumes inpregnancy-induced hypertension. Clinical Science 1989; 77: 599±604.

42. Gilson GJ, Kramer RL, Barada C et al. Does labetalol predispose to pulmonary edema in severepregnancy-induced hypertensive disease? Journal of Maternal and Fetal Medicine 1998; 7: 142±147.

43. Ventura JE, Villa M, Mizraji R & Ferreiros R. Acute renal failure in pregnancy. Renal Failure 1997; 19:217±220.

*44. Cotton DB, Gonik B, Dorman K & Harrist R. Cardiovascular alterations in severe pregnancy-inducedhypertension: relationship of central venous pressure to pulmonary capillary wedge pressure. AmericanJournal of Obstetrics and Gynecology 1985; 151: 762±764.

45. Magann EF, Bass D, Chauhan SP et al. Accelerated recovery from severe preeclampsia±uterine curettageversus nifedipine. Journal of the Society for Gynecologic Investigation 1994; 1: 210±214.

46. Miles JFJ, Martin JNJ, Blake PG et al. Postpartum eclampsia: a recurring perinatal dilemma. Obstetrics andGynecology 1990; 76: 328±331.

47. Atterbury JL, Groome LJ, Ho� C & Yarnell JA. Clinical presentation of women readmitted withpostpartum severe preeclampsia or eclampsia. Journal of Obstetric Gynecological and Neonatal Nursing1998; 27: 134±141.

Severe pre-eclampsia and eclampsia 71