Recent developments in obstetrics

Prediction

Many biochemical substances, principally of placental and endothelial origin, increase in pre-eclampsia. However, their predictive value to detect susceptible women before clinical presentation is poor. As the disorder originates in the placenta, blood flow in the uterine artery is abnormal in most women destined to present with the syndrome. A population at risk can be targeted for intervention (fig 1) by using Doppler ultrasound to measure blood flow in the uterine artery in the second trimester. This biophysical test is currently the best predictor of pre-eclampsia. Positive predictive values are similar in both high and low risk groups, so 20% of women with an abnormal Doppler result will develop pre-eclampsia.1 2

Fig 1

Flow velocity waveforms in the uterine artery as shown by Doppler analysis. a: abnormal flow velocity waveforms—high resistance index, early diastolic match. b: normal flow velocity waveforms—low resistance index, no match

• Download figure
• Open in new tab
• Download powerpoint

Prevention

Aspirin at low dosage has been evaluated extensively as an agent to prevent pre-eclampsia, with a sound rationale of reversing the imbalance of prostacyclin and thromboxane that is found in women with the disease. In 32 randomised controlled trials including nearly 30 000 women, low dose aspirin results in a 15% decrease in pre-eclampsia, with a similar reduction in the chance of fetal death.3 4 Babies are also less likely to be delivered prematurely, and the benefit is independent of risk status. Few interventions in perinatal medicine actually reduce the risk of the baby dying. Targeting women with an abnormal Doppler ultrasound can reduce the number needed to treat (NNT) considerably. Figure 2 shows how the positive predictive value of a test and the reduction in relative risk of an intervention will influence the number needed to treat. Even with a positive predictive value of around 20% (as found with uterine Doppler ultrasound) and a relative risk reduction of 15%, the number needed to treat can still be reduced substantively and make clinical application worth while. The value of screening women at low risk has, however, not been established.

Fig 2

Influence of positive predictive value and relative risk reduction (RRR) on the numberneeded to treat (NNT)

• Download figure
• Open in new tab
• Download powerpoint

The benefit of the intervention may also be better in women identified by abnormal Doppler results. In five randomised controlled trials including women treated with aspirin after abnormal uterine Doppler ultrasounds, the reduction of relative risk in preeclampsia was nearer 45%.5 Recently antioxidants in the form of high dosages of vitamins C (1000 mg) and E (400 IU) have also been used successfully in reducing pre-eclampsia by more than 50%.6 Figure 2 also shows the numbers needed to treat when the intervention results in a relative risk reduction of 50%. Again the rationale for using these vitamins is scientifically sound as oxidative stress seems to be a key feature of the pathophysiology. This study was performed in women principally identified by abnormal Doppler ultrasounds and needs confirmation in other groups before widespread use of vitamins C and E can be recommended.6 However, the combination of predicting women at risk and targeting them with a safe intervention seems to be a sound strategy to limit the number needed to treat to prevent one case.

Management

The management of pre-eclampsia entails close surveillance and delivery. Otherwise it is empirical, aimed at reducing morbidity or mortality. It includes antihypertensive agents, which can be used to halve severe hypertension (relative risk reduction 0.52, confidence interval 0.41 to 0.64).7 This is likely to reduce the risk of cerebral haemorrhage, which occurs in half the women who die.8 Controlling blood pressure does not, however, alter progression of the disease.

The eclampsia trial showed that magnesium sulphate was the drug of choice to prevent further fits after eclampsia,9 but the value of prophylactic treatment with anticonvulsant drugs has been uncertain. As large numbers of women would need to be treated the benefits of prophylactic treatment are important to ascertain. The Magpie study compared giving magnesium sulphate with giving placebo to more than 10 000 women with pre-eclampsia.10 The chance of eclampsia was 0.8% if allocated magnesium sulphate compared with 1.9% in the controls; this represented 11 fewer fits per 1000 women treated. No benefit to the baby was found, although placental abruption may be less common in the group treated with magnesium sulphate. An eclamptic fit indicates serious underlying disease and is usually self limiting and not inherently life threatening, although undesirable. The importance of these findings will therefore depend on previous practice and the population treated. For example, in the United States where magnesium sulphate is used extensively in women at relatively low risk its use may now decline. However, women at high risk are likely to benefit as a trend has become obvious towards fewer maternal deaths in the Magpie study (relative risk 0.55, 95% confidence interval 0.26 to 1.14), and there are physiological reasons why magnesium sulphate might benefit other aspects of sequelae of pre-eclampsia.

Prediction

A short cervix identified on vaginal ultrasound has a good predictive value for early delivery, even in a low risk group, and a cervical length of less than 15 mm will result in a 50% chance of delivery before 32 weeks' gestation.11 The predictive value in multiple pregnancies is similar.12 Identifying fetal fibronectin in vaginal secretions after 22 weeks' gestation also has a strong association with preterm delivery.13 Both these tests perform far better than previous history as predictors of preterm delivery, which has traditionally been the way to identify a group at risk, as most women with recurrent prematurity will achieve a term pregnancy in their subsequent pregnancy. Although it is impractical to screen all women, populations at risk could be targeted with these tests to identify groups for intervention.

Prevention

Few agents can stop uterine contractions effectively, and acute tocolysis has never been shown to reduce perinatal mortality. Delivery may be prevented in the short term, which will allow steroids to be given to the mother or enable transfer to a suitable neonatal unit. A new tocolytic, atosiban (an oxytocin antagonist), is no more effective than β agonists but has fewer side effects and is therefore likely to be safer.14 Treatment has therefore focused on prophylactic measures.

Cervical cerclage—Elective cervical cerclage has previously shown only marginal benefit in preventing preterm delivery, and attention is now focused on cerclage in women identified at risk from a short cervix. Preliminary results show that this may be beneficial, although they are based on very small numbers.15 Some retrospective evidence shows that transabdominal cerclage may have a role in cases where transvaginal cerclage has failed.16

Antibiotics—Although preterm labour is multifactorial, it is often associated with and caused by infection, particularly if very preterm. Both bacterial vaginosis and vaginal infection with group B streptococci increase the risk of preterm delivery. Unfortunately randomised controlled trials have shown that using metronidazole in women at low risk who are positive for bacterial vaginosis does not reduce this risk.17 Targeting women at risk has shown variable results, but some randomised controlled trials show that antibiotics can even increase the risk of preterm delivery.18 More recently oral clindamycin given early to women with abnormal vaginal flora has shown some value in reducing late miscarriage and preterm delivery.19 Therefore the timing, type, and population targeted may determine if treatment is beneficial. The ORACLE trial considered if antibiotics were valuable in preterm labour and prelabour rupture of the membranes.20 Erythromycin significantly reduced the chance of delivery within one week while improving neonatal outcome in women with prelabour rupture of the membranes.20 Importantly co-amoxiclav increased the risk of necrotising enterocolitis and is not recommended with prelabour rupture of the membranes. Asymptomatic bacteruria should be treated as the treatment reduces prematurity.

Tocolytic agents—Although acute tocolysis is of limited benefit, its use is generally restricted by unwanted side effects. Atosiban can be given for prolonged periods, and its use as a prophylactic agent is now being assessed. Calcium channel blockers such as nifedipine also have fewer side effects and are equally efficacious to atosiban, although they are unlicensed for use in pregnancy. Non-steroidal antiinflammatory drugs can cause renal and ductal problems in the fetus. Current studies are evaluating the cyclo-oxygenase-2 inhibitors, which theoretically may be as efficacious at preventing delivery while limiting these unwanted side effects, thus allowing their use as a prophylactic measure.

Other strategies—Progesterone, given either intramuscularly or by vaginal pessary to women at risk, has shown some benefit in reducing preterm delivery in at risk women, and confirmatory trials are eagerly awaited.21 22 Initial reports show that fish oils may reduce the incidence of preterm labour in women with a previous history, although the scientific rationale underpinning this has not been robustly tested.23 Confirmation is required in prospective trials. In the mean time the biggest impact to prevent prematurity is probably simple: avoid smoking.24

Trial of scar

Concerns have arisen that the rate of caesarean sections has been increasing steadily over the past two decades; the results of the term breech trial will be an added pressure on this. The National Sentinel Caesarean Section Audit Report in 2001 (www.rcog.org.uk) showed that in the United Kingdom, 21.3% of women now deliver by caesarean section, a similar rate to the United States. This rate has almost doubled in the past decade. The audit commission has estimated that every 1% increase in the rate of caesarean sections costs the NHS an additional £5m ($8m; €7m). Even elective caesarean sections, which are recognised to be the safest, probably have at least a doubling in risk of maternal death.8 Research has therefore recently focused on the safety of a vaginal delivery after a previous caesarean section and in what circumstances it should be encouraged.

Some 80% of women will achieve a vaginal delivery after a previous caesarean section, even when the cause of the original operation was well defined cephalopelvic disproportion.29 Reasonable retrospective evidence now supports the commonly held belief that both the interval between pregnancies (> 6 months) and a previous vaginal delivery reduce the risk of scar failure.30 31

The chance of successful vaginal delivery is, however, reduced to 60% if labour is induced.32 Rupture of the scar occurs more commonly in these women (1 in 50 women induced with prostaglandins, compared with 1 in 200 if the labour is spontaneous).33 This study of more than 20 000 women with one prior caesarean section defined outcome accurately; only one in 20 scar ruptures caused fetal death. Therefore the overall fetal mortality associated with a trial of scar is extremely small and comparable to the risk of an intrapartum stillbirth in all women (1 in 1500), assuming safe intrapartum management. Only a minority of women with ruptured scars require a hysterectomy (4.4%) and other serious morbidity such as infection, bladder injury, or paralytic ileus occurs in fewer than 10% of women. These data support the policy of encouraging women to deliver vaginally after one previous caesarean section, which is probably the single most important intervention to prevent caesarean section rates escalating further.

Intrapartum care

The recent increase in operative intervention has not been associated with improved perinatal morbidity or mortality. Increasing expectations and medicolegal fears probably contribute to iatrogenic morbidity in pregnancy. Routine investigations in all women are known to increase intervention, sometimes without clear benefit. An example of this is the routine use of electronic fetal monitoring, which is now not recommended in women at low risk. A cardiotocogram performed on admission in women at low risk was assumed to be desirable to identify fetuses at risk. A recent randomised controlled trial comparing this with Doppler auscultation has shown that cardiotocograms result in a notable increase in intervention, including oxytocin augmentation and use of epidurals, and cause an increase in operative deliveries without any evidence of improvement in the wellbeing of the newborn. Given the number of women involved in this intervention, withdrawing the routine use of the cardiotocogram on admission is likely to reduce unnecessary intervention considerably.

Where available, more than one in five women will request epidurals for pain relief in labour. These are associated with longer labours, more use of oxytocin, and increased operative vaginal deliveries.34 Low dose epidurals are being used increasingly, which allow some mobility while maintaining adequate analgesia. This is achieved by adding opiates and reducing the bupivacaine dose. A recent randomised controlled trial has shown that low dose epidurals reduce the need for operative vaginal deliveries, so that one in four could be prevented.35 Several studies have also shown that women prefer them, so there seems little justification in continuing to use the higher dose procedure.