Management of preterm labour 

Preterm birth, defined as birth before the 37th week, can be further subdivided according to gestational age as shown in Table 1. Over the past decade survival rates have dramatically improved due to improvements in neonatal care rather than improvements in obstetric care. Babies born at 26 weeks of gestation and above now have a survival rate of approximately 80%. However, approximately 50% will suffer from some form of disability.

Incidence of preterm birth

The incidence of preterm birth is increasing in both the UK and USA. The rate of preterm delivery in the UK has recently increased from 7% over the last 15 years to a rate of 8.6%. In the USA, the preterm delivery rate has similarly increased, from 9.5% in 1981 to 12.8% in 2006. These increases are despite advancing knowledge of risk factors related to preterm labour and the introduction of many public health and medical

Sarah A Hamilton MBChB is a Clinical Research Fellow, ST4-Maternal and Fetal Health Research Centre, St Mary’s Hospital, University of Man- chester, Manchester, UK. Conflicts of interests: none declared.

Clare L Tower MBChB PhD MRCOG is a Clinical Lecturer, Subspeciality Trainee in Maternal and Fetal Medicine at the Maternal and Fetal Health Research Centre, St Mary’s Hospital, Manchester, UK. Conflicts of interests: none declared. 

Preterm deliveries subdivided by gestation

Table 1

interventions, such as tocolysis, e.g. atosiban and nifedipine, designed to delay preterm birth. 

Causes of preterm birth

The principal pathways leading to preterm birth are spontaneous preterm labour (PTL), preterm prelabour rupture of the membranes (PPROM) and iatrogenic causes. Approximately 45% of births are due to spontaneous PTL, 30% are iatrogenic and 15% are due to PPROM. PPROM is defined as preterm sponta- neous rupture of membranes, at least 1 h before the onset of contractions. The major concerns with PPROM are prematurity, maternal sepsis and chorioamnionitis. Iatrogenic causes are deliveries (labour induction or Caesarean section) for maternal or foetal indications, such as preeclampsia and IUGR.

Maternal characteristics associated with preterm labour are numerous (Table 2). Maternal ethnicity has a significant impact on risk of preterm delivery. In the USA in 2003, the preterm birth rate for African-American women was 17.8%, compared to 10.5% in Asian and Pacific Islander women and 11.5% for Caucasian women. Previous preterm delivery increases the risk of a subsequent preterm delivery 2.5-fold, with the risk of subsequent preterm delivery inversely related to the gestational age of the previous preterm birth.

PTL is a complex process with many factors thought to contribute to it (Figure 1). The management of PTL can be split into three areas, prediction of those women at high risk of PTL, prevention in high risk women and finally diagnosis and treat- ment of those women in PTL.

Prediction e in the general population and those at increased risk

There are several studies that have been carried out to investigate methods of predicting PTL in women at high risk. The main methods used are transvaginal ultrasound and foetal fibronectin (FFN).

Transvaginal ultrasound

Transvaginal ultrasound to measure cervical length and funnel- ling has been studied as a screening test for preterm labour. It has been shown to be safe, acceptable, and reproducible. Cervical length at 24 weeks has been shown to be normally distributed with a mean length of 35.2 mm ` 8.3 mm. In normal pregnancies delivered at term, the cervical length stays relatively constant until the third trimester.

There is an inverse relationship between cervical length and incidence of preterm delivery. Iams (1996) showed that relative risks could be assigned to a particular cervical length. For example, a lady with a cervical length of 22 mm has a relative 

Risk factors for spontaneous preterm labour 

Table 2

risk of PTL of nine while a women with a cervical length of <13 mm has a relative risk of 14, when compared with longer cervical length. In the general population, only 1.7% of women have a cervical length less that 15 mm and these women account for 100% of births prior to 26 weeks, 80% of births prior to 30 weeks of gestation and 60% of births prior to 32 weeks of gestation. Therefore, a cervical length less than 15 mm is a sensitive predictor of severe prematurity, as it is associated with a 50% risk of delivery prior to 32 weeks’ gestation.

Funnelling, which is opening of the internal os with closed cervix below, has also been shown in some studies to be asso- ciated with an independent risk factor for PTL, while other studies have contradicted this. However, the relevant studies have been small and further research is needed.

In order for a screening test to be effective, there needs to be an effective available treatment. At present there is no conclusive evidence that any intervention helps to prevent preterm labour following the identification of cervical shortening or funnelling. Therefore, the main benefit of transvaginal ultrasound screening may be its high negative predictive value of 90% for cervical length over 3 cm at 24 weeks. Women may be reassured, avoiding further clinic visits and intervention.

Foetal fibronectin

Foetal fibronectin (FFN) is an extracellular matrix glycoprotein localized at the maternalefoetal interface of the amniotic membranes, between chorion and decidua (Figure 2). Thus, it is found in the cervico-vaginal secretions prior to labour in both 

Mechanisms of preterm labour 

Figure 1 

preterm and term labours. In a normal pregnancy, FFN should not be present in cervico-vaginal secretions between 20 and 36 weeks’ gestation. Only 3e4% of cervico-vaginal secretions between 21 and 37 weeks are positive for foetal fibronectin, suggesting disruption between the membranes and the decidua has occurred. FFN levels greater than or equal to 50 ng/mL at 22 weeks have been associated with a 40% increased risk of spontaneous preterm birth.

In order to be accurate, the swab for FFN must be taken correctly. The swab should be taken from the posterior fornix or ectocervix. False positive results can occur with the use of lubricant gel, recent sexual intercourse, recent vaginal bleeding, and rupture of membranes.

At present available FFN kits are non-quantitative, giving posi- tive or negative results only. However, it has been shown that there is a correlation between preterm birth and quantitative assessment of foetal fibronectin. The higher the level of foetal fibronectin, the higher the relative risk of delivery prior to 28 weeks.

Therapeutic interventions to prevent PTL in those at high risk

Cervical suture

Cervical suture or cerclage has been widely used in the manage- ment of pregnancies at high risk of preterm delivery. It was initially introduced as a treatment for ‘cervical incompetence’ where the cervix was believed to have some form of inherent weakness. However, true cervical incompetence is very difficult to diagnose. The current uses of cervical cerclage are as an elective or preven- tative procedure, usually conducted at around 14 weeks’ gestation, 

Schematic presentation of location of fetal fibronectin 

Figure 2 

or as an emergency, when the cervix is found to be effacing and dilating at a previable gestation (discussed below).

There are three main forms of cervical suture, the Shirodkar suture described in 1954, the Macdonald suture described in 1957 and abdominal cervical cerclage. The most common method of cerclage used is the Macdonald suture, which is a purse string suture around the cervix. This is electively removed at approximately 37 weeks’ gestation, to allow for a normal vaginal delivery. A Shirodkar suture is placed at the level of the internal os and requires dissection of the vaginal mucosa and bladder, with the vaginal mucosa then closed over the suture. A meta-analysis of Shirodkar vs Macdonald cervical cerclage was performed in women at high risk of preterm birth and no differences were seen in the rates of preterm birth. Further analysis was carried out for those women with short cervices diagnosed on ultrasound scan and again no difference was seen in outcome between the two types of suture. Abdom- inal sutures are used less often as they require more specialist expertise, and tend to be used in women with an extremely short or scarred cervix or when transvaginal cerclage has previously failed. Approximately 80e90% of those women who have an abdominal suture will deliver at term and they are currently the focus of a randomized trial. Abdominal sutures are usually inserted around 11e12 weeks. All types of cervical cerclage have associated risk factors, such as rupture of membranes, bleeding, pregnancy loss, bladder injury and anaesthetic risks.

Randomized controlled studies comparing elective cervical suture with expectant management have been performed and have shown conflicting results. A study carried out on 253 women at high risk of PTL, due to a cervical length of less than 15 mm, showed a non-significant reduction of PTL when using elective cervical suture vs expectant management (22 vs 26%).

Genital swabs and use of antibiotics

Several organisms have been linked with PTL such as bacterial vaginosis, Ureaplasma urealyticum, Mycoplasma Hominis, Trichomonas vaginalis, Chlamydia trachomatis, Neisseria gonorrhoeae. Antenatal screening and treatment of asymptomatic vaginal infection have shown conflicting results. One random- ized control trial showed that treatment of bacterial vaginosis found on routine screening prior to 22 weeks’ gestation can reduce the rate of PTL by 10%. Furthermore a large study with 2058 women in the intervention group has shown that infection screening and treatment in early second trimester, between 15 weeks and 19 weeks and 6 days, for bacterial vaginosis, T. vaginalis and candidiasis can significantly reduce the preterm birth rate when compared with control group (3 vs 5.3%). Furthermore, screening and treatment of those women with infection were found to be cost effective due to its reduction in neonatal care. Other studies disagree although most of these have looked at treating bacterial vaginosis up to 26 weeks of gestation. The RCOG therefore suggests that treatment of BV prior to 20 weeks’ gestation maybe beneficial in reducing PTL.

A more recent study has been carried out using prophylactic azithromycin at 16 weeks and again at 24 weeks in a population at high risk of preterm labour. This study has shown no signifi- cant reduction in PTL in these women when compared with no treatment. 

Progesterone

Progesterone has been of interest for the prevention of preterm labour for many years. Progesterone is an anti-inflammatory agent, which acts by inhibiting myometrial contractions and cervical ripening, down regulating gap junctions and inhibiting chemokine production (see, Drugs acting on the pregnant uterus, by Arrowsmith, Kendrick and Wray, in this issue, pp 241e247). Progesterone has been studied extensively to assess whether it is effective at preventing preterm labour. A recent meta-analysis has shown that progesterone reduces the risk of PTL in women with a singleton pregnancy and a previous history of a preterm delivery when compared with placebo, with the number needed to treat to prevent one PTL being 16. Progesterone has also been shown to reduce preterm birth when administered to women with a short cervix.

Although it may be beneficial in reducing the risks of preterm labour in women at high risk, there is currently little evidence to indicate an improvement in neonatal outcomes. One randomized control trial, which included neonatal factors as secondary outcomes, suggested that treatment with progesterone signifi- cantly reduced the rates of necrotizing enterocolitis and intra- ventricular haemorrhage, and the need for supplemental oxygen. Only one study has attempted longer-term follow-up of children whose mothers were administered progesterone. This study showed no effect of progesterone on either physical examination or development. However, there was significant loss to follow-up and the study was underpowered.

In contrast to singleton pregnancy, there is no evidence of benefit from the use of progesterone in multiple pregnancies. The largest double blinded randomized control trial of progesterone in twin pregnancy was the STOPPIT trial. The primary outcome was delivery or intrauterine death prior to 34 weeks’ gestation. The results demonstrated that 24.7% of women in the proges- terone group had either a delivery or foetal death in utero prior to 34 weeks compared to 19.4% in the placebo group, showing that progesterone does not reduce the combined risk. Other studies in twin pregnancies using different progesterone regimes have had similar results. Therefore, progesterone is of no benefit in multiple pregnancy, and may confer some harm.

Given the lack of evidence for long-term benefit of proges- terone treatment, and the hypothesis that maintaining a foetus in adverse environment may be harmful, the RCOG Preterm Birth study group issued a statement that progesterone use should be restricted to randomized controlled trials. A UK multicentre randomized controlled trial is currently ongoing to investigate both the effect on PTL and long-term neonatal outcome (OPTIMMUM). There is also currently limited evidence to guide the best dosing regimes and routes of administration. The dosage being used by OPTIMMUM is a 200 mg vaginal pessary daily.

Diagnosis of PTL

The presentation of women to the labour ward with symptoms suggestive of threatened PTL is common, but diagnosis is hampered with inaccuracy. Only a small proportion, 8e24%, of those who present with symptoms will go on to deliver prema- turely. The diagnosis is usually made on the clinical basis of regular uterine contractions associated with cervical change, as assessed on vaginal examination. The poor association between clinical symptoms and the likelihood of delivery means that a large number of women receive treatment unnecessarily, and this also causes significant problems for trials of potential treat- ments. Therefore, a better means of diagnosis is needed to prevent women receiving steroids, tocolysis and possible transfer to a tertiary centre unnecessarily, all of which represent a significant financial cost to NHS resources.

FFN has been shown in some studies to be of predictive value in women presenting with symptoms of preterm labour with intact foetal membranes. A recent meta-analysis of 32 studies suggested that FFN is a good short term predictor of preterm birth with a sensitivity of 76% and specificity of 81% for delivery within the next 7 days. A positive foetal fibronectin test has been shown to be a better predictor of PTL than clinical assessments alone. However, the main benefit of FFN is the high negative predictive value of 99%, which can allow clinicians the ability to avoid unnecessary interventions such as tocolysis and to reas- sure women.

Transvaginal ultrasound can also be used in the diagnosis of PTL, although this requires expensive equipment and expertise. A study performed by Gomez et al looked at the use of FFN and transvaginal ultrasound in the diagnosis of preterm labour. The results of the study are shown below (Table 3). They showed that the combining the use of these two diagnostic tests can identify 75% of those women who will deliver within 7 days.

Treatment of PTL

Tocolysis

Several drugs have been investigated for their tocolytic proper- ties, but, to date, no study has shown that tocolysis reduces rates of preterm delivery or improves neonatal outcome. However, pregnancy can be prolonged for up to 48 h in approximately 80% of cases, which beneficial in allowing time for administration of corticosteroids and in-utero transfer. The main tocolytics used in the UK are COX inhibitors (e.g. indomethacin), calcium channel blockers (e.g. nifedipine) and oxytocin antagonists (e.g. atosi- ban). Ritodrine, a b-2 adrenergic receptor agonists, which induced uterine relaxation, was previously used but is no longer recommended due to significant adverse maternal side effects (see Drugs Acting on the Pregnant Uterus in this issue, pp 241e247).

Nifedipine: nifedipine is not licensed for use in threatened preterm labour and there have been no randomized control trials 

Rates of diagnosis of PTL with FFN and Transvaginal scanning 

Table 3 

of nifedipine vs placebo in the treatment of threatened preterm labour. One small randomized multicentre study has looked at maintenance therapy with nifedipine vs placebo once an initial ‘successful tocolysis’ has been performed. This showed that nifedipine did not alter preterm delivery rates. Most studies have compared nifedipine efficacy with that of ritodrine or terbutaline, with efficacy being the same with both types of drugs, but less maternal side effects with nifedipine. There has been a small randomized study (n 1⁄4 63) directly comparing atosiban with nifedipine; there was no difference in the efficacy of either of the drugs, though there was a non-significant increase in adverse reactions, such as palpitations and headaches, with nifedipine compared with atosiban.

Advantages of nifedipine over other tocolytics are that it is cheap and can be given orally. There is no standard protocol for the administration of nifedipine but in most studies it has been administered as 10 mg sublingually every 15 min for an hour or until cessation of contractions followed by 60e160 mg slow release nifedipine daily for maintenance therapy in four divided doses.

Atosiban: atosiban is the only drug licensed for treatment of threatened preterm labour in the UK. It is a competitive oxytocin antagonist that acts at the uterine oxytocin receptors. It is given as an initial bolus of 6.75 mg over 1 min, followed by an infusion of18mg/hfor3hthenreducedto6mg/hforupto45h.Ato- siban should not be continued for more than 48 h.

Atosiban has not been shown to reduce the preterm labour rate when compared with placebo or to improve neonatal outcomes. There is a report of a higher number of neonatal deaths in the atosiban compared with placebo group, but this could have been due to a higher number of women at less than 26 weeks’ gestation randomized to the atosiban group. Atosiban has also been compared with betamimetics, such as salbutamol, terbutaline and ritodrine. In these studies, atosiban efficacy was similar to beta- mimetics, in that it did not significantly alter delivery before 48 h. However, atosiban was better tolerated than betamimetics.

COX inhibitors: there are several cyclo-oxgenase inhibitors used for tocolysis, such as ketorolac, celecoxib, indomethacin and sulindac. The one most commonly used in the UK is indometh- acin. When COX inhibitors have been compared with other tocolytics such as ritodrine and magnesium sulphate they have equal efficacy at prolonging gestation for 48 h. COX inhibitors inhibit uterine contractions, are easily administered and have few maternal side effects. However, adverse effects have been reported in the newborn following exposure to COX inhibitors, including premature closure of the ductus arteriosus, renal and cerebral vasoconstriction, and necrotizing enterocolitis (associ- ated with high dose or prolonged exposure).

In summary, there is no evidence that tocolysis works. At best, it delays delivery by between 48 h and 7 days, giving enough time for corticosteroids to be administered. Choice of drug varies with unit policy but first line agents are usually atosiban or nifedipine, with indomethacin only being adminis- tered at less than 32 weeks’ gestation.

Emergency cervical sutures

Emergency cervical sutures are performed when the cervix is objectively open and the membranes are at or below the external os prior to 26 weeks’ gestation. In a retrospective study over an 8-year period, 46 emergency cervical cerclages were inserted with a 44% success rate for delivery after 36 weeks; this is similar to other studies quoting a 50% success rate. Indicators of poor outcome are membranes below the external os, dilatation over 3 cm, signs of infection (raised C-reactive protein or white blood cell count) and bleeding of over 100 ml at time of opera- tion. Failure of the suture was also closely associated with post delivery chorioamnionitis.

Antibiotics

Extreme preterm birth is usually associated with infection, most commonly ascending infection from the vagina and several studies have assessed antibiotic use in the prevention of preterm labour. The largest study performed to date was the ORACLE II study which investigated women presenting with symptoms of spontaneous preterm labour with intact membranes. The primary outcome was a reduction in neonatal death with the use of antibiotics. The routine prescribing of antibiotics to women in spontaneous preterm labour did not reduce neonatal death, but did reduce the risk of maternal infection. Further follow-up of the participants of the ORACLE II study at 7 years found an increased risk of cerebral palsy in the children who received antibiotics (odds ratio 1.93 (95% confidence interval 1.21e3.09) for eryth- romycin and 1.69 (1.07e2.67) for co-amoxiclav). When antibi- otics were combined, risks were higher still than with erythromycin alone (4.55 vs 2.29%). It has been suggested that the use of antibiotics could be masking a subclinical infection and keeping a baby within a hostile environment longer, thus increasing the risk of cerebral palsy.

Corticosteroids

Corticosteroids are used in PTL to increase foetal surfactant and accelerate foetal lung maturity. They have been shown to be beneficial in reducing neonatal death, respiratory distress syndrome (RDS), necrotizing enterocolitis, cerebrovascular hae- morrhage and neonatal intensive care admissions. For maximum benefit, the optimum time between administration of steroids and delivery is from 24 h to 7 days, though there has been a trend towards benefit following 7 days. A single course of corticoste- roids, two injections of 12 mg betamethasone 24 h apart, has been shown to confer no harm to the foetus in long-term follow- up studies and every effort should be given to administer steroids to all women at risk of preterm labour between 24 and 34 weeks’ gestation. It has also been shown that there is benefit in treating women up to 36 weeks’ gestation, though the number needed to treat to prevent one case of RDS is considerably increased. Repeat doses of corticosteroids are associated with increased risks of neurodevelopmental problems in the baby, and thus should be avoided.

Consequences of preterm birth

Preterm birth contributes to substantial neuro-cognitive, pulmonary, and ophthalmologic morbidity and globally accounts for 28% of neonatal deaths. In the US, preterm birth is the most frequent cause of infant death, accounting for one third of infant mortality in 2002. Prolonging a pregnancy from 30 weeks to 34 weeks’ gestation decreases the in-hospital mortality from 8.1 to 0.4%. Respiratory distress syndrome also reduces from 43.8% at 30 weeks’ gestation to 2.6% at 34 weeks, even with steroid use in both groups. Therefore, the ability to prolong a pregnancy has the potential to have a huge impact on the health of the child. Understanding more about the mechanisms of term and preterm labour is essential to identify targets for novel therapies to prevent PTL.

Despite significant advances in neonatal care, preterm delivery is still the largest cause of neonatal morbidity and mortality. A study carried out on 144 live-born infants prior to 28 weeks of gestation showed that all received neonatal resuscitation. Thirty-four percent died prior to discharge home and the rest were assessed at 2 years and 6 years of age. Psychomotor development was normal for 52%, borderline for 20%, moderately handicapped in 20%, and severely handi- capped in 8%. The UK multicentre study Epicure collected data on 4044 births in 1995 and between 20 weeks and 25 weeks, 6 days. Of these births, 811 babies survived to be admitted to neonatal units and of these 39% survived to be discharged. Follow-up at their expected due date showed that 62% had cerebral parenchymal cysts, hydrocephalus, retinopathy of prematurity or continuing need of oxygen. Further follow-up was carried out at approximately 2.5 years of age. At this stage 50% had no disability but approximately 25% had severe disability, either cerebral palsy, neurological problems or developmental delay.

The Nuffield Council on Bioethics has developed guidelines on the critical care in foetal and neonatal decisions. In this report it is stated that normal practice of full intensive care and support should be offered to all neonates from 24 weeks’ gestation, unless the patient or clinicians decide it is not in the baby’s best interests due to her/his condition. They also state that between 23 weeks, 0 days and 23 weeks, 6 days, precedence should be given to the wishes of the parents as to whether they wish for full intensive care for the child. However, this should not be completely against clinical judgement, if treatment would be futile. Prior to 23 weeks, it suggested that resuscitation should not be offered.

In conclusion, PTL remains poorly understood. Further research is needed within this area to improve preventative strategies and treatment of PTL. 

FURTHER READING

Berghella V, Hayes E, Visintine J, Baxter JK. Fetal fibronectin testing for reducing the risk of preterm birth. Cochrane Database Syst Rev; 2008 Oct 8: CD006843.

Berghella V, Baxter JK, Hendrix NW. Cervical assessment by ultrasound for preventing preterm delivery. Cochrane Database Syst Rev; 2009 Jul 8: CD007235.

Costeloe K, Hennessy E, Gibson AT, Marlow N, Wilkinson AR. The EPICure study: outcomes to discharge from hospital for infants born at the threshold of viability. Pediatrics October 2000; 106: 659e71.

Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008 Jan 5; 371: 75e84.

Kenyon S, Pike K, Jones DR, et al. Childhood outcomes after prescription of antibiotics to pregnant women with spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet 2008 Oct 11; 372: 1319e27. 

Norman JE, Mackenzie F, Owen P, et al. Progesterone for the prevention of preterm birth in twin pregnancy (STOPPIT): a randomised, double- blind, placebo-controlled study and meta-analysis. Lancet June 09; 373(9680): 2034e40.

Nuffield Council of Bioethics. Critical care decisions in fetal and neonatal medicine: ethical issues, www.nuffieldbioethics.org (published Thu 16 Nov 2006).

RCOG Green top guidelines 1b preterm labour, tocolytic drugs. London: RCOG, Oct 2002.

RCOG Green top guidelines 7 Respiratory Distress Syndrome, antenatal corticosteroids. London: RCOG, Feb 2004.

Rode L, Langhoff-Roos J, Andersson C, et al. Systematic review of progesterone for the prevention of preterm birth in singleton preg- nancies. Acta Obstet Gynecol Scand 2009; 88: 1180e9. 

Practice points 

• The aetiology of PTL is multifactorial.
  C Betamethasone is the only drug proven to improve neonatal

• outcome in PTL.
  C Tocolysis can at best delay PTL for long enough for the

• maximum benefit of steroids.
  C Diagnosis of PTL is difficult but can be improved with use of

• FFN and cervical scanning.
  C Further research is needed to investigate the use of proges-

• terone for the prevention of PTL in women at risk prior to its widespread use.