Use of tocolytics: what is the benefit of gaining even more time?
Prof Y Jacquemyn, Department of Obstetrics and Gynaecology, Antwerp University Hospital UZA, Wilrijkstraat 10, B-2650 Edegem, Belgium. Email Yves.Jacquemyn@uza.be
Today’s dogma states that tocolytics can be used to prolong pregnancy for just 48 hours, allowing corticosteroids to be administered and transportation of the mother to a tertiary care centre. Surveys have shown that up to 30% of practitioners use maintenance tocolysis. Theoretically, maintenance tocolysis should be able to improve neonatal outcome by avoiding preterm birth and allowing delivery in the community maternity hospital or even at home, minimising social difficulties created by long distances between the mother and/or her baby and the rest of the family. This should result in fewer neonatal intensive care unit admissions, less respiratory distress syndrome and fewer long-term neurological sequelae. Such an effect has never been proven, probably because we do not know which women benefit from treatment, which do not require treatment because they are not in labour and which babies would better be born because chorioamnionitis and other insults jeopardise intrauterine development. Most studies on long-term tocolysis have been performed with beta-agonists. No improvement has been shown. On the contrary, a trend towards fetal harm with an increased risk for periventricular leucomalacia exists. Results from studies of one tocolytic should not be generalised; one published study on maintenance therapy with atosiban showed prolonged uterine quiescence and prolonged gestation, but was too small to detect differences in neonatal outcome. In the future, we need larger studies, not only to detect whether long-term tocolysis with newer tocolytics (oral oxytocin antagonists, prostaglandin receptor blockers) results in better neonatal outcome, especially at the lower gestational ages but also to discover methods that allow us to identify those women who will benefit from treatment and those for whom prolongation of pregnancy may cause harm.
The original rationale of administering tocolytics was to prolong pregnancy and to aim for term delivery, and this is still what women and the lay public expect from us. This idea has been superseded by a more recent dogma that can be found in most articles on tocolysis: tocolytics are given to delay delivery for 48 hours, to allow the administration of corticosteroids to induce lung maturation and to transport the pregnant woman to a tertiary care centre with neonatal intensive care facilities. To date, no trial has shown a difference in the number of women completing a course of antenatal corticosteroids following the use of tocolytics. It is generally accepted that tocolytics do not result in a reduction of preterm birth, but that 2 to 3 days can be gained,1 except for indomethacin where a reduction in birth before 37 weeks of gestation has been shown, but without a difference in maternal or neonatal outcome.2 This is why the Royal College of Obstetrics and Gynaecology states ‘…it is reasonable not to use tocolytic drugs, as there is no clear evidence that they improve outcome’.3 So why even consider long-term maintenance therapy and defend this apparently hopeless case?
Studies and systematic reviews on maintenance tocolysis are abundant, and were reviewed in 2005 by Thornton.4 To summarise, betamimetics have not proven to be beneficial for perinatal outcome and can result in serious fetal and neonatal adverse effects. Although 10 of 16 studies included in a Cochrane review used maintenance tocolysis, no decrease in the number of births within 7 days, was shown.5 Long-term use of beta-agonists in pregnancy has been associated with an increased risk of neonatal periventricular leucomalacia6 and maternal peripartum cardiomyopathy.7 Neither oral nor subcutaneous beta-agonists have been shown to decrease the risk of preterm birth.8,9 A more recent, non-randomised gestational-age matched study, including 558 women comparing oral tocolytics with continuous subcutaneous terbutaline infusion showed a gain of 5 days (28.4 ± 19.8 versus 33.9 ± 19.0 days, P < 0.001) and a cost of $5286 less per pregnancy for the subcutaneous compared with the oral group.10 This study shows that obstetric outcomes are measured no longer solely in neonatal and/or maternal morbidity but also in cost-effectiveness. Magnesium has no tocolytic benefit, so it is not surprising that magnesium maintenance therapy is ineffective.11 As far as we were able to determine, no data on maintenance tocolysis with nitric oxide donors are available. Recent studies have shown a reduced number of preterm births with the use of intramuscular or vaginal progesterone, but no studies are available on the use of progesterone as a maintenance therapy after initial treatment of spontaneous preterm labour.
Two trials are available using calcium channel blockers for maintenance therapy.12,13 In the study by Carr, oral nifedipine was compared with no treatment and no difference in time gained during pregnancy was noted.12 Since all women were initially ‘treated’ with intravenous magnesium sulphate (which is not an effective tocolytic), this implies that the women studied probably had false spontaneous preterm labour, a condition that does not require treatment. Sayin et al. randomised 73 women to receive either maintenance nifedipine or no treatment after intravenous ritrodrine and verapamil tocolysis.13 Time gained to delivery (mean difference ten days, P = 0.03) and gestational age at delivery (37 versus 35 weeks, P = 0.03) were higher in the nifedipine group, but there were no differences in perinatal outcome. This can be explained because gestational age at admission was 32.3 ± 3.5 (SD) weeks in the maintenance group and 31.3 ± 3.1 weeks in the treatment group. Gestational age at delivery was 37.03 ± 2 weeks in the nifedipine group and 35.1 ± 3 weeks in the control group. At such late gestational ages, a very large number of cases are required to show extremely small differences in perinatal outcome.
Cyclooxygenase inhibitors have not been promoted for maintenance tocolysis due to fear of constriction of the ductus arteriosus, but, while these adverse effects were published in case reports, they were not shown in a systematic review.14
Maintenance treatment with subcutaneous atosiban (oxytocin receptor antagonist) was studied in one trial in women with spontaneous preterm labour who achieved successful tocolysis with previous intravenous atosiban treatment.15 The median time from the start of maintenance treatment for the first recurrence of labour was 32.6 days with atosiban and 27.6 days with placebo (P = 0.02). Maternal and neonatal outcomes were comparable and there was no difference in the number of preterm births. The efficacy of maintenance therapy with atosiban is proven for women who have already responded to the drug.
Despite this lack of proof of effect for maintenance tocolysis on perinatal outcome, a few more days seem to be gained both with oral nifedipine and subcutaneous atosiban, but lack of proof of effect is not synonymous with proof of lack of effect. All the studies and reviews conclude that numbers are too small to compare relevant fetal outcomes such as long-term neurologic sequelae, on which almost no data are available. Since the adverse effects of intravenous atosiban are almost non-existent, one can wonder why no study has been performed using long-term intravenous treatment. In this way, the therapeutic 100 microgram/minute could be maintained instead of lowering the dose to 30 microgram/minute as was performed in the study by Valenzuela et al.,15 which might be insufficient to achieve tocolysis and introduces the risk of studying subtherapeutic doses.
Despite the general tendency in the literature to doubt the effect of maintenance tocolysis, it is frequently administered: 34% of practitioners in Australia and New Zealand gave maintenance tocolysis in 2002.16 If practitioners continue to do so, what can they be expecting to gain? The use of maintenance tocolysis must be based on the simple belief that we can prolong pregnancy. Obstetricians probably react as intuitively as do their patients and try to stop each and every uterine contraction, even if this is not related to labour (as a process of progressive cervical dilation resulting in delivery). However, our actions are not evidence based. Intuition tells us that each day that can be gained is important. At 23–26 weeks, ±3% survival is gained per day.17 In a recent report from Belgium, survival increased from 49% at 24 weeks to 72% at 26 weeks. Short-term intact survivors (without neurologic sequelae and without specific intensive care complications at the time of hospital discharge) were 4, 7 and 28% at 21, 25 and 26 weeks, respectively.18
Every day gained is precious at very early gestational ages. At later gestational ages, survival or the avoidance of major neurological deficits are no longer major problems, but at this time we must take the economic perspective into account. The cost of neonatal care at 29 weeks are ten-fold those at 35 weeks, but at 35 weeks they are still ten times higher than at 38 weeks.19 Maintenance tocolysis may result in exactly the contrary of ‘gaining time to transport the mother to a tertiary care centre’. Patients and doctors believe that by prolonging pregnancy, intrauterine transport and eventually neonatal intensive care will not be necessary. Arguments for this could be that care and delivery in the local hospital diminish social and familial problems caused by the distance to a tertiary care centre. Economic implications can play a role in this scenario, since sending a woman to another hospital can result in loss of income. We know of no studies that have quantified this aspect.
There are some caveats in defending maintenance tocolysis. It is possible and even probable that some of the mechanisms that trigger spontaneous preterm labour also cause fetal neurologic (and other) injury. A clear link exists between chorioamnionitis and (cytokine induced) fetal-neonatal cerebral palsy as conceptualised by Romero’s Fetal Inflammatory Response Syndrome.20 The incidence of neonatal white matter lesions is appreciably lower in preterm infants born after pregnancy-induced hypertension compared with those born due to spontaneous preterm labour or preterm prelabour rupture of the membranes.21 Term children exposed to tocolysis showed a higher rate of psychiatric disorders and poorer cognitive and motor performance than controls, but no such effect has been shown following preterm birth.22 These conclusions should not be generalised to the newer tocolytics, they only concern beta-agonists, and it is not clear whether the effects are a result of the medication or of an underlying disease leading to spontaneous preterm labour and tocolysis.
What we really need is a test that discriminates between those women in spontaneous preterm labour associated with an inflammatory and/or infectious process, and who are in need not only of tocolysis but also of a more aetiologic therapy such as antibiotics or anti-inflammatory agents and those women without inflammatory or infectious problems. Such a test is not readily available. Possible candidates include detection of cytokines and interleukins in serum or amniotic fluid, culture of Ureaplasma urealyticum and leucocytes or glucose in amniotic fluid. By continuing tocolysis after the golden first 48 hours and the administration of corticosteroids in cases of chorioamnionitis, we are doing more harm than good, but the same is probably true if we stop tocolysis in cases where no other threat to the fetus exists than preterm birth.
In conclusion, every new study on maintenance tocolysis must be stratified by gestational age, as more effect can be expected at lower gestational ages. Groups should be as homogenous as possible regarding the cause of spontaneous preterm labour, excluding those with an inflammatory/infectious aetiology. Economic, social and familial costs must be included. The route and dosage of administration of tocolytics must be rationalised by avoiding subtherapeutic dosages. Not only surrogate outcome measures (such as the number of days gained) but also the real and definitive result, i.e. long-term neurological and behavioural outcome must be evaluated.
As long as we are unable to discriminate within 48 hours between the inflammatory/infectious and the noninflammatory groups of women in spontaneous preterm labour, new studies are condemned to be inconclusive. It is probable that pregnancy can also be prolonged with nifedipine after previous tocolysis with a betamimetic. At the moment, all we do know is that continuing atosiban, even at a lower dose, in women who previously responded to the drug can prolong pregnancy.