Labour as an inflammatory process

Throughout pregnancy the uterine cervix needs to remain firm and closed whilst the body of the uterus grows by hypertrophy and hyperplasia but without significant fundal dominant contractions. For labour to be successful the cervix needs to be converted into a soft and pliable structure that can efface and dilate and the uterus needs to become a powerful contractile organ. There is no single endocrine or biochemical switch in the human that changes the uterus from its not‐in‐labour state to its in‐labour state. The onset of labour is a gradual process which begins several weeks before delivery itself with changes in the lower pole of the uterus which cause cervical ripening and effacement. The onset of clinically identifiable contractions is a relatively late event in this process. Cervical ripening occurs through breakdown of collagen, changes in proteoglycan concentrations and an increase in water content. The lower segment of the uterus also stretches and relaxes and behaves physiologically more like the cervix than like the contractile upper segment of the uterus. These changes in the lower segment of the uterus are associated with an increase in the production of inflammatory cytokines, particularly interleukin (IL)‐8 and prostaglandins from the overlying fetal membranes and decidua and from the cervix itself. Cervical ripening is associated with an influx of inflammatory cells into the cervix which release matrix metalloproteins that contribute to the anatomical changes associated with ripening. The later increase in fundally dominant contractility in the upper segment of the uterus is associated with an increase in the expression of receptors for oxytocin and prostaglandins, in gap junction proteins, which mediate electrical connectivity between myocytes, and in more complex changes in the intracellular signalling pathways which increase the contractility of the myocytes.

Roles of progesterone, corticotrophin‐releasing hormone and oxytocin

In many species progesterone is thought to play an important role in suppressing the onset of labour [6]. Progesterone has a generally anti‐inflammatory action within the uterus. As discussed above, many of the biochemical events associated with cervical ripening and the onset of labour are similar to those seen at sites of inflammation. In most species the onset of labour is heralded by withdrawal of progesterone.

So, for example, in the rodent, prostaglandin‐mediated regression of the corpus luteum leads to a fall in progesterone concentrations immediately prior to the onset of labour. In the sheep increased production of cortisol from the fetal adrenal signals fetal maturation and induces placental 17α‐hydroxylase, which increases synthesis of oestrogen at the expense of progesterone, again leading to progesterone withdrawal immediately prior to the onset of labour. There is no systemic withdrawal of progesterone in the human prior to the onset of labour, although there is an increase in the expression of genes formerly repressed by progesterone, which has led to the hypothesis of a ‘functional progesterone withdrawal’ mediated by changes in the expression or function of progesterone receptors or of cofactors needed for the function of the progesterone receptor. Another hypothesis is that inflammatory events seen within the uterus at the time of labour are associated with increased activity of the transcription factors nuclear factor (NF)‐κB and AP‐1 (transcription factors strongly associated with inflammation in other contexts such as asthma, inflammatory bowel disease or arthritis). NF‐κB and AP‐1 repress the function of the progesterone receptor and so could mediate functional progesterone withdrawal. Although in the mouse progesterone concentrations fall due to luteolysis just prior to labour, there is still sufficient circulating progesterone concentrations to activate progesterone receptors. In the mouse it appears that the final event leading to parturition is the increased production of surfactant protein A from the fetal lung, which stimulates the activity of NF‐κB within the uterus leading to an influx of inflammatory cells, an increase in inflammatory cytokine synthesis and depression of the residual function of the progesterone receptor. It is an attractive hypothesis that pulmonary maturation in the human may signal the final phase of the onset of labour but there is at present no direct evidence that this mechanism applies in the human.

Circulating levels of corticotrophin‐releasing hormone (CRH), synthesized in the placenta, increase progressively throughout pregnancy and especially during the weeks prior to the onset of labour. CRH‐binding protein concentrations fall with advancing gestational age such that, approximately 3 weeks prior to the onset of labour, the concentration of CRH exceeds that of its binding protein. Unlike in the hypothalamus, placental CRH is upregulated by cortisol. Several studies have linked placental production of CRH with the timing of birth and have demonstrated that a premature rise in CRH is associated with preterm delivery. The upregulation of CRH by cortisol suggests a mechanism by which the fetus, through increased adrenal cortisol production, may signal its maturation and control the timing of birth [7]. For much of pregnancy the CRH receptor expressed by the myometrium is linked to second messenger systems that promote relaxation. Near to term, however, CRH may enhance the contractile response to oxytocin and may stimulate the production of prostaglandins from the fetal membranes and the placenta.

In the monkey, uterine contractions occur only at night. In the days preceding labour and delivery there are nocturnal non‐fundal dominant contractions which have been termed ‘contractures’. The conversion from contractures to contractions is mediated by an increase in the production of oxytocin from the maternal posterior pituitary gland [8]. In the monkey, therefore, while the fetus might signal its general readiness to be born through increased cortisol production from the adrenal, the precise timing of birth is signalled by the mother. This may be a mechanism of defence against predators which ensures that delivery is always at night. Contrary to the experience of many obstetricians, this phenomenon does not apply to the human. There is no increase in the production of oxytocin associated with the onset or progression of either preterm or term labour. There is, however, an increase in the expression of oxytocin receptors within the uterus and there is local production of oxytocin in the uterus, decidua and fetal membranes. Although oxytocin probably does not play an important role in the precise timing of parturition in the human, increases in the density of oxytocin receptors suggests that oxytocin does play a role in mediating contractility. Recent studies have shown that oxytocin acts not only to stimulate the uterus to contract, but also to upregulate inflammatory mediators within the uterus, therefore adding an additional ‘pre‐labour’ mechanism of action for the hormone. Oxytocin also plays important postnatal functions in mediating the milk‐let down reflex, contracting the uterus to prevent postpartum haemorrhage and having a effect on maternal bonding with the baby.

Cervical function

With improved survival at early gestational ages, there is now overlap between second‐trimester pregnancy loss and early preterm delivery. Historically, cervical incompetence was diagnosed in women who experienced persistent, often rapid and painless, late second‐trimester pregnancy loss. More recently, the concept of cervical competence as a continuum has evolved. It is probable that cervical length and strength, together with the quality of the cervical mucus, contribute towards cervical function, both to retain the pregnancy within the uterus and to exclude potential bacterial pathogens from ascending from the vagina. Numerous studies have demonstrated a strong relationship between cervical length and the risk of preterm delivery. The cervix may be damaged (or completely removed) by surgery in the treatment of cervical cancer or, rarely, during a difficult instrumental vaginal delivery, or caesarean section at full dilatation. Historically, there were associations between diethylstilbestrol exposure in utero and developmental anomalies in the genital tract and cervical weakness. This ceased to be a problem in modern obstetric practice since the cohort of women exposed to the drug in the 1960s are now beyond reproductive age. A short or partially dilated cervix may allow bacteria to ascend into the lower pole of the uterus where, acting through the Toll‐like receptors of the innate immune system which recognize bacterial components, they stimulate production of inflammatory cytokines, prostaglandins and the inflammatory response. This then leads to cervical ripening and shortening, which in turn decreases the ability of the cervix to act as either a mechanical or microbiological barrier and, ultimately, leads to the development of either localized or generalized chorioamnionitis and to preterm delivery. A short or weak cervix may therefore contribute to preterm delivery not only by leading to simple second‐trimester miscarriage but also by contributing to a risk of ascending infection leading to a more classical spontaneous preterm labour. Delivery by caesarean section at or close to full dilatation of the cervix is now recognized as a risk factor for preterm birth. The probability is that difficult delivery leads to mechanical damage to the cervix, through the trauma from failed instrumental delivery, through a uterine incision made within cervical rather than lower segment tissue, or through damage to the cervix caused by the need to disimpact a deeply engaged fetal head.

There is an association between risk of preterm delivery and cervical intraepithelial neoplasia (CIN) [11]. The greatest risk is in those women with CIN who have had a particularly deep large loop excision of the transformation zone (LLETZ) or a cold knife cone biopsy. In women who have had a deep LLETZ or a cold knife cone biopsy, mechanical damage to the integrity of the cervix is probably a major aetiological factor in their risk of preterm labour. However, there is a smaller underlying risk associated with CIN alone. It may be that human papillomavirus (HPV) infection is an independent risk factor for preterm birth. It is also possible that the underlying factors associated with the development of CIN following HPV infection in an individual woman may also be factors which increase her risk of preterm birth.

Genital tract infection

There is a strong correlation between infection within the uterus and the onset of spontaneous preterm labour. As discussed, activation of inflammatory mediators is a central part of the normal biology of parturition. Therefore infection within the uterus has the potential to activate all the biochemical pathways, ultimately leading to cervical ripening and uterine contractions. It has been estimated that approximately 40% of all preterm births are associated with bacterial infection. The most likely source of infection is bacteria ascending from the vagina through the cervix into the lower part of the uterus. However, bacteria may also gain access to the amniotic cavity through haematogenous spread or by introduction at the time of invasive procedures. Following preterm delivery histological chorioamnionitis is usually more common and severe at the site of membrane rupture than elsewhere, such as overlying the placenta or umbilical cord. In virtually all cases of congenital pneumonia, inflammation of the fetal membranes is also present. Bacteria identified in the majority of cases of congenital infection are often also found in the maternal lower genital tract and, following twin preterm delivery, chorioamnionitis is more common and severe in the presenting twin than in the second twin (although this is not always the case). These factors all suggest that ascending infection from the lower genital tract is the commonest mechanism for chorioamnionitis.

The most common microbes isolated from the amniotic cavity of women in preterm labour are Ureaplasma urealyticum, Fusobacterium and Mycoplasma hominis. More than 50% of patients in preterm labour will have more than one microorganism isolated from the amniotic cavity. Microorganisms can be identified in the fetal membranes of the majority of women delivering both preterm and at term. It is probable that some cases of spontaneous preterm delivery are due to an excessive inflammatory response to a lesser degree of bacterial invasion of the amniotic cavity. So, for example, bacterial vaginosis (see below) may be a greater risk factor for preterm labour in women who carry a high secretory form of the tumour necrosis factor (TNF)‐α gene.

There is now considerable interest in the role of the microbial communities in the vagina in the aetiology of preterm birth. The collective term for the range of bacterial species in the vagina is ‘vaginal microbiota’. The collective term for all the bacterial genes present is ‘vaginal microbiome’ (although the term ‘microbiome’ is often used interchangeably with ‘microbiota’ to define a microbial community occupying a reasonably well defined habitat which has distinct physicochemical properties). The study of the bacterial genes present in the vaginal microbiome is described as metagenomics.

In reproductive life the vaginal microbiota is usually dominated by the presence of lactobacilli, representing more than 90% of bacterial species present. Lactobacilli secrete lactic acid, which maintains a low pH hostile to other microorganisms and which has anti‐inflammatory actions. Lactobacilli also excrete specific antimicrobial proteins. A minority of women will have a Lactobacillus‐depleted vaginal microbiota, and this may allow overgrowth of bacterial vaginosis (BV)‐associated anaerobic organisms such as Gardnerella vaginalis, which creates a biofilm that allows other opportunistic bacteria to thrive. The increased oestrogen concentrations of pregnancy increase the availability of vaginal mucosal glycogen, a source of energy for lactobacilli. Therefore, in general, the proportion of lactobacilli increases in the vagina during pregnancy. The relationship between the structure of the vaginal microbiota and the risk of preterm birth varies from population to population. In some but not all populations in the USA, where Lactobacillus depletion is common, a dysbiotic Lactobacillus‐depleted BV‐like vaginal microbiota is a risk factor for preterm birth. In the UK, prevalence of a dysbiotic vaginal microbiota in pregnancy is low but is probably still a risk factor. However, the dominance of one particular species, Lactobacillus iners, appears to be a risk factor for both cervical shortening and preterm birth. Lactobacillus iners has less ability to excrete anti‐inflammatory isomers of lactic acid or antimicrobial proteins, and may represent a transitional organism between healthy vaginal microbiota and vaginal dysbiosis or bacterial vaginosis.

Haemorrhage

Placental abruption may lead to the onset of preterm labour. This is thought to be through release of thrombin, which stimulates myometrial contractions by protease‐activated receptors but independently of prostaglandin synthesis. This may explain the clinical impression that preterm labour associated with chorioamnionitis is often rapid whereas that associated with placental abruption is less so because there is no pre‐ripening of the uterine cervix. Generation of thrombin may also play a role in preterm labour associated with chorioamnionitis when it is released as a consequence of decidual haemorrhage.

Fetal and maternal stress

There is evidence that both fetal and maternal stress may be risk factors for preterm labour. Fetal stress may arise in association with abnormal placentation and growth restriction. Maternal stress could be due to environmental factors. In both cases it is postulated that over‐secretion of cortisol leads to upregulation of CRH production in the placenta.

Ultrasound measurement of cervical length

There is very good evidence that transvaginal sonographic measurement of cervical length can be used to identify women at risk of preterm labour in both low‐ and high‐risk pregnancies and in women who are symptomatic [12] (Fig. 28.5). Transabdominal measurement of cervical length is unreliable because of the need for a full bladder, which may compress the cervix leading to an overestimate of its length, and because it is more difficult to obtain adequate views of the cervix with this technique. Transvaginal ultrasound should be performed with the bladder empty. The probe is placed in the anterior fornix of the vagina without undue pressure on the cervix and optimally the internal and external os and the echogenic endocervical mucosa should be identified along the length of the canal. For identification of risk in asymptomatic women (those who do not have symptoms of labour) two broad strategies are currently in common use: a single measurement in the mid‐second trimester, or serial measurement of cervical length throughout the second and early third trimester of pregnancy.

A single measurement of cervical length, usually at the time of a routine ultrasound scan between 18 and 22 weeks, has been widely used to identify subjects at high risk of preterm birth for inclusion into intervention trials. If a screening strategy using a single ultrasound measurement of cervical length is used, then assessment between 21 and 24 weeks’ gestation appears to be better than assessment prior to 20 weeks’ gestation in predicting the risk of preterm labour. However, this is to a certain extent a self‐fulfilling prophesy since clearly the closer to the actual onset of preterm labour the assessment of cervical length is made, the more likely it is that the cervix will be found to be short. It is arguable that identification of a risk of preterm labour as late as 23 weeks may be too late for any potential prophylactic therapies to be fully effective. In addition, such a strategy is unable to detect any of the women whose pregnancy loss or preterm delivery occurs prior to 23 weeks. A large number of studies have examined the relationship between gestational age, cervical length and the risk of preterm delivery (Fig. 28.6). Many studies have used single cut‐off values. So, for example, a cervical length of 15 mm or less at 20–24 weeks predicts a risk of preterm delivery prior to 34 weeks’ gestation of approximately 50% in a low‐risk population. It is absolute cervical length rather than the presence or absence of funnelling which is the principal predictor of spontaneous preterm birth (although clearly the presence of funnelling will lead to a shorter cervical length).

It has been suggested that the introduction of routine measurement of cervical length at the time of the second‐trimester anomaly ultrasound scan would enable screening of low‐risk populations. This concept is greatly predicated on the assumption that an effective intervention is available (see section on progesterone and cervical cerclage). The value of routine measurement of cervical length also depends on the prevalence of a short cervix and the incidence of preterm birth in the background population. In UK populations this approach will only detect about 15% of all preterm births, a reflection of the multi‐aetiological nature of the syndrome.

Women at high risk of preterm birth may be offered serial measurement of cervical length to assess their risk of preterm labour. This approach appears to be superior to a single measurement in assessing the risk of preterm delivery. It has been widely advocated as an approach for the detection of women who would benefit from progesterone prophylaxis during pregnancy. It is also a particularly useful approach in women with a history of a previous preterm birth or second‐trimester pregnancy loss in whom a diagnosis of cervical insufficiency or incompetence is not clear and can be used to reduce the number of unnecessary cervical cerclage procedures performed. In this management strategy, cervical cerclage would be indicated either when cervical length reduces to a fixed cut‐off, commonly 25 mm, or falls below the 10th or 3rd centile for cervical length at that gestational age. In continental Europe it is common practice to perform a vaginal assessment of cervical length at each antenatal consultation, although multicentre trials have shown that this policy is of no benefit in predicting the risk of preterm delivery.

Bacterial vaginosis

As alreadt discussed, BV is a risk factor for preterm birth, although most studies have shown that treating BV with antibiotics does not change the risk. Studies of the risk of preterm labour associated with BV have reported widely varying results. However, it seems that, overall, BV approximately doubles the risk of preterm delivery. It also appears that there is a relationship between the gestational age at diagnosis of BV and the risk of preterm delivery, in that if BV is diagnosed earlier in pregnancy this appears to be associated with a higher risk of preterm delivery. Routine screening for BV is not therefore undertaken in low‐risk populations. Some obstetricians do include screening for BV in the management of high‐risk populations, and this is currently undertaken by non‐genetic techniques, although the future introduction of DNA sequence‐based bacteriology may change this situation. Currently, diagnosis of BV can be made on Gram staining of vaginal fluid using either Nugent’s or Spiegel’s criteria, by gas–liquid chromatography of vaginal fluid (finding a high ratio of succinate to lactate) or on clinical grounds based on a high vaginal pH, a fishy odour in a thin homogeneous vaginal discharge and the presence of clue cells in the discharge on a wet mount. There is no significant difference in the ability of each of these diagnostic tests to predict preterm birth.

Although there is reasonably good evidence that BV is a risk factor for preterm delivery, it is less clear that treating it with antibiotics is beneficial. This may be in part because various studies of BV have used different antibiotics in different regimens and at different times, but it may also reflect the fact that antibiotics may not necessarily result in the re‐establishment of normal bacterial flora. The two antibiotics commonly used in the treatment of BV are metronidazole administered orally or clindamycin, which may be given either orally or vaginally. Clindamycin may have advantages over metronidazole since it has better activity against anaerobic bacteria and Mycoplasma hominis and Ureaplasma urealyticum which are often associated with BV. While screening of pregnant women who are at high risk for preterm delivery based on their past obstetric history or other factors might be justified, there is currently no strong evidence to recommend the routine screening and treatment of the general obstetric population.