Correspondence: Dr R. Lamont, Northwick Park and St Mark's NHS Trust, Watford Road, Harrow, UK.
Since the 7th and 13th Study Groups of the Royal College of Obstetricians and Gynaecologists met in 1977 and 1985, respectively, no meeting of this magnitude has convened to discuss the problems of spontaneous preterm labour and delivery and the associated fetomaternal mortality and morbidity. In the 17 years or so since that time, advances have been made in our understanding of the mechanisms of labour, the role of infection, the benefit of antepartum corticosteroids and the development of safer more specific tocolytics. In the future, an understanding of the genetic risk of spontaneous preterm labour and preterm birth is essential, particularly with respect to the predisposition to produce potentially damaging pro-inflammatory cytokines. The examination of the tissue damage will require pathologists specifically trained in perinatal pathology if the aetiology is to be ascertained and future management tailored to the risks. A greater understanding of fetomaternal immunology and response to antigen exposure in pregnancy may help us to understand which fetomaternal pairs are at greatest risk of responding by delivering preterm, with greater or lesser tissue damage than others with similar risk. Specifically, the relation between spontaneous preterm labour and proteinuric pre-eclampsia with their common immunology, inflammatory response and tissue damage leading to either spontaneous preterm labour or iatrogenic preterm birth will need to be addressed. This meeting has been very clinically and obstetrically orientated, in future we will need to involve epidemiologists, neonatologists, microbiologists, genito-urinary medicine physicians, immunologists, geneticists, biochemists, physiologists and endocrinologists. Although spontaneous preterm labour and preterm birth are the major causes of perinatal mortality and morbidity in the developed world, the definition and management protocols for spontaneous preterm labour varies from unit to unit and country to country. A process has already begun, hopefully fuelled by this meeting and those attending, to develop an international consensus on definitions and evidence-based practical guidelines on the management of spontaneous preterm labour. Perhaps in the longer term it may be possible to influence standards of care, outcome measures and training across international boundaries.
As a result of advances in obstetric ultrasound over the last 20 to 30 years, followed by induced abortion of the malformed fetus, preterm birth has replaced congenital malformation as the major cause of death and handicap in neonates in the developed world. Of babies born between 22 and 26 weeks gestation, 65% die in the delivery suite or on the neonatal intensive care unit (NICU). Of those who survive and are available for follow up to 30 months, half will be handicapped and in 50% of these, the handicap will be severe. This means that only 12–13% of babies born between 22 and 26 weeks gestation will be alive and intact at the age of 2 1/2 years1. The social, psychological, emotional and family costs of preterm birth are immeasurable but it is estimated that in the USA, 10,000 US dollars are required to keep a baby in the NICU for one week. This amounts to 5 billion US dollars annually and if a baby is handicapped and requires life long residential care, the cost is approximately 450,000 US dollars per baby2.
With the contribution that spontaneous preterm labour and preterm birth makes to perinatal mortality and morbidity, it might be assumed that we would know a great deal more about the condition. Unfortunately, we still do not know what causes spontaneous preterm labour and preterm birth so it is difficult to predict or prevent. As a result, we have to manage spontaneous preterm labour as it arises at which stage we have difficulty making the diagnosis. Even if we can make the diagnosis, we are unsure whether or not we should interfere. If intervention is considered appropriate there is still debate about the use of corticosteroids, tocolytics and antibiotics. If we decide not to interfere or are unsuccessful with intervention, there is still doubt about the best way to deliver the preterm infant and there are still some difficulties persuading clinicians to transfer babies to centres with NICU facilities.
In the last 25 years, on only two previous occasions have a group of opinion leaders in the field of spontaneous preterm labour and preterm birth been assembled to address these issues. At the 6th study group of the Royal College of Obstetricians and Gynaecologists (RCOG; 1977)3, spontaneous preterm labour and preterm birth had to be defined and at that stage 40% of cases of spontaneous preterm labour and preterm birth were thought to be idiopathic. We know now that there is probably no such thing as idiopathic spontaneous preterm labour since at least one fetal, maternal or placental cause will be identified in 96% of cases and two or more causes in 58% of cases4. At the 13th study group of the RCOG (1985), the difficulties associated with conducting a well-controlled tocolytic study were identified. Using Bishop's score as a means of diagnosing spontaneous preterm labour and also as a means of ensuring comparability between two treatment groups, a time period of 18 months was thought to be sufficient to recruit 150 patients from two centres. In the end, using these strict criteria, it took four centres three years to recruit only 58 patients, which was insufficient for statistical analysis5.
In the 17 years or so since that time, there have been major advances in our understanding of the biochemical, endocrine, paracrine, and molecular biology of the mechanisms of physiological and pathological preterm labour.
Heterogeneity of Preterm Birth
A number of concepts of spontaneous preterm labour and preterm birth need to be considered. Firstly, is preterm labour a physiological process occurring too early in pregnancy or is it a pathological process? It is likely that the nearer to 37 weeks gestation that a woman delivers as a result of spontaneous preterm labour and preterm birth the more likely this is to be due to physiology (i.e. slightly early normal term labour) as opposed to pathology. Conversely, the nearer a woman is to 24 weeks gestation, the more likely her spontaneous preterm labour and preterm birth is likely to be due to pathology and less likely to be due to physiology (Fig. 1). The heterogeneity of preterm birth is also manifest in the following four scenarios.
1A woman at 36 completed weeks of gestation who delivers preterm as a result of spontaneous preterm labour and preterm birth.
2A woman at 33 weeks gestation who requires elective iatrogenic preterm delivery because of fulminating proteinuric pre-eclampsia.
3A Caucasian woman at 30 weeks gestation who has had two previous term deliveries, but as a result of an antepartum haemorrhage due to either placental abruption or placenta praevia, develops spontaneous preterm labour and preterm birth but is likely to deliver at term in a subsequent pregnancy if that pregnancy is not complicated again by antepartum haemorrhage.
4An Afro-Caribbean woman at 24 to 26 weeks gestation who recurrently experiences preterm prelabour rupture of the membranes and delivers extremely early preterm.
While each of these examples are considered under the overall heading of preterm birth they are clearly due to different aetiologies and require different management in the index pregnancy and have a different risk of preterm birth in the future.
Reactive Versus Proactive Approach
Historically, the approach to the management of spontaneous preterm labour and preterm birth has been reactive using tocolytics, corticosteroids, in utero transfer and considering the possibility of infection.
Until the meta-analysis of Crowley et al. in 19906, the use of antepartum corticosteroids had been inconsistent. This showed that the use of antepartum corticosteroids was able to reduce the incidence and severity of respiratory distress syndrome especially after a course of optimal treatment. The incidence of periventricular haemorrhage, necrotising enterocolitis and neonatal death were also statistically significantly reduced. Since that time, antepartum corticosteroids have been used much more extensively, but in future, more data are required with respect to the use of repeat courses of antepartum corticosteroids and the detrimental effects this may have on the fetus.
We have a greater understanding of the mechanism of term and preterm labour. A greater understanding of the role of oxytocin has led to the development of new tocolytic agents with a specific effect on the target organ, with fewer side effects, which have been specifically developed to treat spontaneous preterm labour and are licensed for use. In future, a better understanding of the mechanisms of term and preterm labour may result in the introduction of other tocolytic agents, such as cyclo-oxygenase (COX)-2 inhibitors, anti-inflammatory cytokines, or more potent oxytocic antagonists which are available by the oral route. Further research may lead to the use of prophylactic tocolytics or long term maintenance therapy and/or re-treatment, and perhaps the development of combined treatment using either antibiotics with tocolytics or combined tocolytics ± antibiotics. This will necessitate a debate on cost versus safety versus benefit. Clinicians often consider cost versus benefit but rarely consider cost versus safety. How unsafe does a tocolytic have to be before cost becomes an issue? Conversely, how costly should a tocolytic be before safety might be compromised? If there is compromise on safety or benefit because of cost, this will have clinical governance implications, including consent to the use of non-licensed agents, which in itself will have risk management and medicolegal implications. There will be debate about the ethics of using such tocolytics in the developing world/countries. Should these more expensive drugs, which are safer for the mother and the baby, be used only in the developed world because of their cost? What are the ethics of developing world countries being expected to use non-licensed agents which have not undergone clinical trials empowered by sufficient sample size, with strict inclusion/exclusion criteria to satisfy regulatory bodies and so are not used in accordance with the best evidence-based practice, but used in the developing world solely because they are cheap and orally administered.
In Utero Transfer
Over the last 20 years or so no one has doubted the benefit of in utero transfer with respect to neonatal morbidity and mortality compared with neonatal transfer, but there still appears to be some difficulty persuading clinicians to transfer babies in spontaneous preterm labour to centres with NICU facilities. Perhaps we should be trying to develop better means of communication between units where obstetricians and midwives are not required to spend precious time on the telephone trying to find vacant intensive care cots. It would be interesting to try to develop an Internet web-based system updated every four hours or so of the availability of neonatal cots, maternity beds and paediatric surgery, so that it can be instantly recognised where any one or more of this combination of services are available.
The ORACLE study7 in which antibiotics were used in women in suspected spontaneous preterm labour with intact membranes has failed to answer the question, ‘should antibiotics be used in women in spontaneous preterm labour of infective aetiology’? The ORACLE study has been criticised for using the wrong antibiotics, in the wrong women, too late in pregnancy8 and should have been more objective in the identification of women at risk of preterm birth due to infection. This emphasises the need for smaller more focused trials of women given antibiotics because they have been identified using objective measures as being at real risk of preterm birth due to infection9.
In the long term, it will be helpful to reduce the incidence of spontaneous preterm labour and preterm birth by reducing risk factors. Those clinicians involved in medically assisted conception techniques may need to justify the number of embryos re-implanted after in vitro fertilisation (IVF) and also to reduce the incidence of multiple higher order births as a result of IVF and ovulation induction techniques. It would be helpful to improve socio-biological status, but it has to be recognised that societal changes will take tens of years to accomplish. The future of early detection using patient education and home monitoring needs to be further explored.
Prediction of Spontaneous Preterm Labour and Preterm Birth
Since 1985 there have been advances in the prediction of spontaneous preterm labour and preterm birth using abnormal genital tract flora in the form of bacterial vaginosis diagnosed on Gram stain of vaginal secretions, transvaginal ultrasound scanning (TV USS) to measure cervical length and the introduction of oncofetal fibronectin for the prediction of preterm birth.
Women diagnosed as having bacterial vaginosis or being colonised by bacterial vaginosis-related organisms in pregnancy are at statistically significantly increased risk of late miscarriage and preterm birth. The earlier in pregnancy at which these abnormal genital tract flora are detected, the greater is the risk of an adverse outcome10. Transvaginal ultrasound scanning has replaced digital examination or transabdominal scanning as the best way to measure cervical length. Although, there remains doubt for the future as to whether cervical length should be considered a categorical variable or a continuous variable, and what length of the cervix should be accepted as a cutoff point above and below, which there is an increased risk of spontaneous preterm labour and preterm birth. Future research should also indicate what intervention if any is indicated, e.g. prophylactic cerclage or surveillance followed by emergency cerclage in those women whose cervix shortens. Like bacterial vaginosis and TV USS, oncofetal fibronectin has a low sensitivity and poor positive predictive value, but a high specificity and negative predictive value. In future, better guidance will be needed as to what precautions might be taken, e.g. admission to hospital, prophylactic antibiotics, prophylactic tocolytics, or transfer to a centre with NICU facilities. Further information is required as to whether single or sequential sampling is most beneficial. Rather than considering bacterial vaginosis, TV USS or oncofetal fibronectin individually for prediction of preterm birth, it would be interesting to know in the future whether those women who have abnormal genital tract flora diagnosed by bacterial vaginosis prior to 16 weeks gestation are those who will have shortened cervical length on TV USS between 20 and 24 weeks gestation and who will have a positive oncofetal fibronectin test at 24 weeks gestation. In this sense, the cumulative benefit of these three predictors may turn out to be better than each one individually. In addition, the benefit of these predictors at different gestations should be considered and the method which can be found to be the most predictive at the earliest gestation may determine the most helpful approach.
The use of Antibiotics Prophylactically for the Prevention of Spontaneous Preterm Labour and Preterm Birth
The use of antibiotics prophylactically in the prevention of preterm birth in women considered to be at high risk, either because of a pre-existing risk factors or abnormal genital tract flora in early pregnancy, remains unanswered10. We still do not know whether we should screen and treat or re-screen and retreat or whether we should give antibiotics prophylactically. We do not know at which gestation antibiotics should be given, but it is likely these will be most beneficial at early gestations. There is still doubt about what is the best antibiotic and its dosage and duration and route of administration. It is likely that studies of combined intravaginal and systemic antibiotics active against ureaplasmas, mycoplasmas, anaerobes and other bacterial vaginosis-related organisms given in early pregnancy to those women with the greatest degree of abnormal genital tract flora will be necessary to show antibiotic benefit. We need to find a better way of measuring the outcome of antibiotic studies, avoiding the endpoint of preterm or term delivery per se, when what we are trying to achieve is the avoidance of the birth of an infected baby, be it either term or preterm. We also need to decide whether we treat those women who are of low or high risk.
The detection of abnormal genital tract flora per se in early pregnancy has a low sensitivity for the prediction of preterm birth. The risk of preterm birth is clearly a balance between susceptibility and exposure. There may be some women who have abnormal colonisation, but do not possess the genetic predisposition to be high producers of pro-inflammatory cytokines, under which circumstances spontaneous preterm labour and preterm birth does not ensue. Alternatively there may be some women who have a genetic predisposition to be high producers of pro-inflammatory cytokines, but if they are not exposed to abnormal genital tract flora, spontaneous preterm labour and preterm birth does not ensue. It may be only those women who have abnormal genital tract flora who also have a genetic predisposition to produce pro-inflammatory cytokines who develop spontaneous preterm labour and preterm birth.
Better Communication Between the Imagers and the Pathologists
There is now convincing evidence that babies born preterm as a result of spontaneous preterm labour of infectious aetiology are more likely to be associated with elevated concentrations of pro-inflammatory cytokines in vaginal secretions, amniotic fluid and fetomaternal serum. Under these circumstances, there is greater risk of tissue damage in the form of bronchopulmonary dysplasia, periventricular leucomalacia and, at long term follow up, cerebral palsy. While there are exciting advances in imaging of such tissue damage, in trying to relate these images to the insult and the eventual degree of disability and handicap, magnetic resonance images and their relation to ultrasound images and the subsequent perinatal pathology requires future evaluation.
We have spent many years quantifying and identifying abnormal genital tract flora without considering the host defence and host response to such an insult. The use of preterm birth as an outcome parameter per se is a poor one, since many babies will be born preterm who are healthy and will survive intact. A better way of measuring outcome would be to show a reduction in the incidence of babies born infected as opposed to those non-infected at any gestation whether preterm or term. A better understanding of neonatal immunology and innate response is required. With the protection of the mother and the placental barrier, most babies would be born immunologically naıuml;ve apart from the passive immunity received from the mother. In this way, babies should be born without evidence of exposure to bacterial antigens. By studying antigen-presenting cells (dendritic cells), which dictate the type and degree of neonatal immune response, we may understand how this interacts with the maternal immune response and what role this has in the production of pro-inflammatory cytokines and other agents as part of this response, which may lead to fetal tissue damage and mortality or long-term morbidity.
Final End Common Pathway with Pre-Eclampsia
Finally, there is a recent suggestion that toxaemia, pre-eclampsia and spontaneous preterm labour and preterm birth may have common links through an immunological response. Many cases of fulminating pre-eclampsia will result in elective preterm birth. There are some who say that elective delivery under these circumstances simply brings forward by some two weeks or so spontaneous preterm labour and preterm birth. Similar to bacteria or other infective antigens, it may be that in pre-eclampsia, small microvilli from the placenta break off into the maternal circulation, either predisposing a woman to, or resulting from, the endothelial damage associated with pre-eclampsia. In some women there may be no immune response to these circulating foreign bodies, but in some the immune cytokine response may be initiated resulting in the final cascade of prostaglandin production and spontaneous preterm labour and preterm birth.
Finally, those clinicians interested in spontaneous preterm labour and preterm birth will need to communicate more closely to try to achieve a consensus within and across national boundaries with respect to many of these controversies in the diagnosis, prediction, prevention and management of spontaneous preterm labour. This might lead to more fruitful consensus solutions and research collaboration within and across national boundaries.