Tocolytics and preterm labourBMJ 2009; 338 doi: http://dx.doi.org/10.1136/bmj.b195 (Published 06 March 2009) Cite this as: BMJ 2009;338:b195
- Andrew Carlin, director of maternal fetal medicine1,
- Jane Norman, chair of maternal and fetal health2,
- Stephen Cole, consultant in maternal-fetal medicine3,
- Roger Smith, director of mothers and babies research centre4
- 1Department of Obstetrics and Gynaecology, Mothers and Babies Research Centre, John Hunter Hospital, NSW 2310, Australia
- 2Reproductive and Developmental Sciences, University of Edinburgh, Edinburgh EH16 4TY
- 3Maternal-Fetal Medicine, Royal Womens’ Hospital, Melbourne, Parkville VIC 3052, Australia
- 4Mothers and Babies Research Centre/Hunter Medical Research Institute, John Hunter Hospital
Preterm labour is still the main cause of perinatal morbidity and mortality in the developed world. Despite a better understanding of the pathophysiology, and the recognition that it is a syndrome with multiple causes, rates continue to rise.1 In the linked cohort study (doi:10.1136/bmj.b744), de Heus and colleagues assess the incidence of serious maternal complications with the use of tocolytic drugs for the treatment of preterm labour.2
Corticosteroids given to women at risk of preterm birth increase fetal lung maturity and can reduce fetal death, intraventricular haemorrhage, and respiratory distress by up to 50%.3 In contrast, it is still unclear whether tocolysis improves neonatal or longer term outcomes.4 Thus, the Royal College of Obstetricians and Gynaecologists’ clinical guideline on tocolysis concludes that, “it is reasonable not to use tocolytic drugs.”5
Despite the lack of evidence, several tocolytic drugs are commonly used worldwide. They are used primarily to delay delivery for up to 48 hours to allow for administration of corticosteroids or transfer to a unit with neonatal intensive care facilities (or both). Little consensus exists as to which tocolytic agent is the best. Beta agonists, atosiban, and indometacin all reduce the incidence of delivery within 48 hours compared with placebo, but none has been shown to improve neonatal outcomes.4 Nifedipine has not been studied in placebo controlled trials, but it is more effective at delaying delivery than β agonists, and has been associated with improved neonatal outcomes.6 Atosiban and nifedipine have not been directly compared.
The study by de Heus and colleagues assesses maternal safety. They recorded adverse maternal effects in an unselected population of 1920 consecutive women given tocolytic drugs in routine clinical settings in the Netherlands and Belgium over an 18 month period. The most commonly used agent was atosiban (42%), followed by nifedipine (34%), β agonists (14%), and indometacin (8%).
The overall incidence of adverse effects was reassuringly low, but the study highlights the dangers of multidrug regimens, which have no proven benefit either sequentially or in combination. Four women needed intensive care after being subjected to such regimens—three of which included β agonists. But before such practices are condemned outright, a thorough analysis of each case is warranted, so that further research into safer combination treatments is not compromised.
Because of the definitions used in the study design, postpartum haemorrhage was not included as a maternal adverse event. Some people may think this a surprising omission considering the basic mechanism of action of all tocolytics.
Consistent with previous studies, the β agonists had a higher incidence of serious adverse drug reactions (1.7%) than nifedipine (0.9%); no such reactions were reported for atosiban.
Hypotension was the most serious adverse drug reaction associated with nifedipine, which is not surprising considering its licensed indication. Hypotension was defined as a systolic blood pressure of <100 mm Hg and a drop of more than 20%. Associated symptoms did not need to be treated. It is debatable whether this reaction would be considered “life threatening” according to the Council for International Organizations for Medical Science definition used as a guideline in the article, and no justification is given for using these particular values. Indeed, the authors themselves question the importance of these events and point out that no associated episodes of fetal compromise occurred.
These findings highlight the relative maternal safety of tocolytics, although β agonists have the highest adverse event rate and their continued use is hard to justify. Atosiban stands out as the drug with the safest maternal adverse event profile, but it has not been shown to improve neonatal outcomes and is considerably more expensive than nifedipine. Using the number needed to treat analysis provided in the article, we calculated that 108 patients would need to be treated with atosiban to prevent one serious adverse drug reaction associated with nifedipine. This means that, allowing for the spread across the confidence intervals, an extra £30 000 (€33 000; $44 000) to £45 000 would need to be spent for each serious adverse drug reaction prevented (costs sourced from British National Formulary, September 2008). Clearly a detailed cost-benefit analysis is warranted.
This study also serves as a timely reminder that the decision to use tocolysis should not be taken lightly, especially as we have no good test for differentiating “true” from “spurious” preterm labour, so many women are treated unnecessarily.4
After more than 30 years of research we still do not know whether tocolysis benefits the fetus, so the choice of which drug to use remains a secondary question. The real dilemma is whether or not we should treat at all. Not only may tocolysis lack any real benefit, but it could be harmful, as highlighted in the recent publication of the long term outcomes of the ORACLE II trial.7 In the original study, antibiotics were given to women with intact membranes who were at risk of preterm labour, and no improvement was seen in neonatal outcomes. But of major concern is the increased incidence of cerebral palsy at 7 years seen in the recently published follow-up data.
We agree with de Heus and colleagues that a randomised controlled trial of nifedipine and atosiban is needed; however, any such study should include a placebo arm. Accumulating evidence such as that from the ORACLE II trial and from groups examining the associations between infection, inflammation, preterm labour, and brain injury,8 9 10 11 alongside improvements in neonatal care, mean that the old assumption that “keeping the baby inside longer must be a good thing” can no longer go unchallenged.
Cite this as: BMJ 2009;337:b195
Competing interests: RS and JN have received grants from government and charitable organisations for research into understanding the mechanism of term and preterm labour and investigating treatments. JN has acted as a consultant to a small drug company that was considering developing treatments for preterm labour.
Provenance and peer review: Commissioned; not externally peer reviewed.