The preterm parturition syndrome and its implications for understanding the biology, risk assessment, diagnosis, treatment and prevention of preterm birth

Introduction

The syndromic nature of spontaneous labour was a concept first proposed in 1994 [1]. It has important implications for understanding the biology of preterm parturition, and also, setting realistic expectations about what programs for the prediction and prevention of preterm birth can and cannot do.

The traditional view of preterm parturition has been that it is fundamentally the same process as spontaneous labour at term, and that the crucial difference is the gestational age at which these processes occur [2-4]. Another implicit assumption is that all causes of preterm labour are the same, because patients with preterm labour and intact membranes are treated with the combination of tocolytic agents and glucocorticoids worldwide. This article outlines the potential flaws in this concept, which we believe has halted progress in the prevention of preterm birth and in the treatment of preterm labour.

Preterm birth: definition, classification and burden of disease

A preterm birth is one that occurs between 24 weeks (fetal viability) and 37 weeks of gestation (menstrual age). The term “fetal viability” is intended to identify a fetus capable of extra-uterine life. When a conceptus is delivered before viability, this is considered a spontaneous abortion rather than a preterm birth. We have discussed the issues surrounding the definition of “viability” and the gold standards used for such a definition, and the interested reader is referred to such an article for further details [5].

Approximately two-thirds of all preterm births occur after the spontaneous onset of labour. The remaining one-third is the result of “indicated preterm birth”. The term “indicated” connotes that labour and/or delivery has been the result of induction and/or abdominal delivery for maternal or fetal indications. The most common maternal indication is pre-eclampsia, and the fetal indication is a small-for-gestational-age fetus with evidence of compromise. Other causes of “indicated preterm birth” include congenital anomalies or maternal diseases that require delivery (e.g. acute fatty liver of pregnancy or other disorders, which would improve after delivery) [6-10]. This distinction between spontaneous and indicated preterm birth is based upon the clinical circumstances which determine preterm delivery, namely, the presence of spontaneous labour. It is important to note that this classification does not take into account the cause of the “spontaneous” or “indicated preterm birth”, and indeed, there may be an overlap between the two [9],[11].

Preterm birth accounts for 5–12% of all births, and it is a leading cause of perinatal morbidity and mortality worldwide. The risk of death of a preterm neonate is 120 times greater than that of an infant born at term [6],[12]–[16]. Moreover, survivors are at risk of short-term morbidity (respiratory distress syndrome, intraventricular haemorrhage, necrotizing enterocolitis, sepsis, retinopathy of prematurity) [17-29], and long-term morbidity [30,31] such as cerebral palsy [32-36] learning disabilities [23],[37,38], blindness [39,40] and crippling respiratory disease [41-48]. Recent evidence suggests that preterm neonates may also be at risk for altered metabolic states in adult life [49-52].

Preterm births have been classified according to the gestational age at which they occur, into very early (those with a gestational age at delivery of ≤28 weeks), early (≤32 weeks), and late preterm birth (33–36 weeks). The frequency of these three types is 0.82%, 2.2%, and 8.9%, respectively [53]. Therefore, most preterm births occur late, yet serious morbidity and mortality affect disproportionately those born before the 28^th week [42],[44,45],[54-64]. Recent observations suggest that late preterm births are at a higher risk for health and developmental problems than infants born at term [65-74]. Indeed, infant mortality, defined as “deaths at the age of one year” is significantly greater in late preterm birth than in term birth [16],[67,68]. Of interest is that this excess includes Sudden Infant Death Syndrome (SIDS) [16], suggesting that a pathological process may operate in both inducing preterm birth and also increasing the susceptibility of infant mortality. The clinical implications of these observations is that late preterm birth should not be neglected as unimportant, or a short-term neonatal problem.

The common pathway of parturition

The common pathway of parturition has been defined as the anatomical, physiological, biochemical, endocrinological, immunological and clinical events that occur in the mother and/or fetus at the time of parturition regardless of whether this occurs at term or preterm. The most well-known components of the common pathway of parturition are the uterine components because they are clinically apparent to obstetricians and patients. Such components include: 1) increased myometrial contractility; 2) cervical ripening/dilatation and effacement; and 3) membrane/decidual activation [1,2]. Each of these components is visibly and clinically activated in spontaneous labour at term, where uterine contractions become regular and painful, cervical dilatation and effacement easily detectable, and rupture of membranes is a well-recognised part of the parturitional process. Therefore, in most cases of spontaneous labour at term, there is synchronous activation of the common pathway.

We have proposed that the onset of spontaneous labour at term is the result of physiological signals that are set into motion when the mother cannot sustain fetal growth any longer. This is predicated on the basis that human gestation has an inherent limit related to the ability of the mother and/or placenta to sustain growth. When such a limit is reached, parturition is initiated. An unlimited duration of pregnancy would result in taxation of maternal resources, and the growth of a fetus that cannot pass through the birth canal. It is unclear if the duration of normal pregnancy is determined by the maternal and/or fetal genome. Clearly, there can be conflicting interests between these two genomes, and the time of the onset of labour most likely represents a balance between these conflicting interests. Whether this should be framed in terms of David Haig's Conflict Hypothesis, or alternatively, viewed as an issue of cooperation between the maternal and the fetal genome is an interesting biological question.

We view preterm parturition as the consequence of pathological signals that activate the common pathway of parturition. Moreover, the activation may be asynchronous. Therefore, some patients may present with preterm increased uterine contractility. Other patients may present with acute cervical insufficiency (a dilated cervix with membranes protruding through the cervix) and others may present with preterm premature rupture of membranes. It is also possible that preterm labour presents with synchronous activation of the common pathway. This represents the patient with increased uterine contractility, cervical dilatation, and membrane rupture. Figure 1 illustrates the components of the common pathway of parturition, and also the phenotypic expression of the preterm parturition syndromes (preterm uterine contractions, cervical insufficiency, preterm premature rupture of membranes).

Figure 1. The common pathway of term and preterm parturition. Reproduced with permission from reference 2.

Symptoms and signs of activation of the common pathway of parturition

The activation of the uterine components of the common pathway is manifested clinically with signs and symptoms. Mothers who have pain during uterine contractions, pelvic pressure in cases of cervical dilatation with prolapse of the membranes, or leakage of fluid have symptoms of the preterm parturition syndrome. The objective demonstration of these symptoms (uterine contractions demonstrated by palpation or a uterine strip, visual inspection with a speculum examination, or demonstration of fluid in the posterior fornix) would represent the signs. It is important to note that none of the signs and symptoms provide any information as to why preterm parturition has begun.

Subclinical evidence of activation of the common pathway

It is possible to detect activation of the different components of the pathway using sophisticated tools. For example, electromyography has been employed to detect changes in uterine electrical activity associated with labour [75-78]. Cervical ultrasound can detect preterm effacement of the uterine cervix. The colloscope has been used to identify changes in collagen organization of the cervix underlining the process of cervical ripening. Finally, a positive fetal fibronectin or other markers represent evidence of degradation of the extracellular matrix. Such a process is important for membrane rupture, and therefore, it has been interpreted to be a biochemical marker of membrane/decidual activation. This conceptual framework allows a link between the clinical manifestations of the preterm labour syndrome, and the tools being developed to detect untimely activation of the different components of the common pathway of parturition.

Preterm parturition as a syndrome with multiple aetiologies

We have proposed that preterm parturition should not be considered a monolithic entity, but rather as a syndrome. The term “syndrome” is defined in the Oxford Medical Dictionary as “a combination of symptoms and signs that form a distinct clinical picture indicative of a particular disorder”. Therefore, these symptoms and signs tend to occur together. The crucial point is that the signs and symptoms that define the preterm parturition syndrome are caused by multiple aetiologies or pathological processes.

Many have referred to preterm parturition as a multifactorial condition. This term is used to indicate that there may be more than one factor responsible for the disorder. However, we find this characterization incomplete, because there are no unifactorial diseases in medicine. For example, sickle cell disease, which is caused by a single base pair mutation in the nucleotides encoding for haemoglobin, may be expressed with a wide range of phenotypes which depend to a large extent on environmental factors.

We have argued that obstetrical disorders are really syndromes, and refer to them as “The Great Obstetrical Syndromes”[79]. The features of “The Great Obstetrical Syndromes” are the following: 1) multiple aetiology; 2) long preclinical stage; 3) frequent fetal involvement; 4) clinical manifestations which are often adaptive in nature; and 5) predisposition to a particular syndrome is influenced by gene–environment interaction and/or complex gene–gene interactions involving maternal and/or fetal genotypes [80-108].

Preterm parturition meets all the criteria for the syndrome, and we refer the reader to a previous article for the examination of the arguments employed in support of this thesis [109].

Aetiologies of the preterm parturition syndrome

The following mechanisms of disease have been implicated: 1) intrauterine infection/inflammation [110-292]; 2) uterine ischaemia [293-307]; 3) uterine overdistension [308-327]; 4) abnormal allogenic recognition [328-335]; 5) allergic-like reaction [336-340]; 6) cervical disease [341,342]; and 7) endocrine disorders [343-359]. Most of these mechanisms of disease operate in non-pregnant women. For example, ischaemia is a major pathophysiologic process in cardiovascular disease. Fluid overload may be responsible for acute congestive heart failure. Rejection is the main problem in transplantation medicine. Allergy and endocrine disorders are well-established mechanisms of disease in childhood and adult life. It is also possible that preterm parturition may be caused by mechanisms of disease that are unique to the maternal–fetal relationship, because of its unique anatomy, physiology, immunology and metabolic demands. Consequently, we remain open to the discovery of undescribed mechanisms of disease during pregnancy.

Figure 2 illustrates the causes of the preterm labour syndrome. Thus far, infection is the only mechanism of disease for which a causal link has been established. However, there is substantial evidence in support of the others, but causality still needs to be proven. A full discussion of the scientific evidence which supports a role for each mechanism of disease in the preterm parturition syndrome is available in recent publications by our group [109].

Figure 2. Pathological processes implicated in the preterm parturition syndrome. Reproduced with permission from reference 2.

What are the consequences of the concept that preterm parturition is a syndrome?

The implicit consequence that preterm parturition is a syndrome with multiple aetiologies is that there would not be a single treatment, diagnostic method or preventive strategy that will work for all preterm births. For example, tocolysis has been tested in women with preterm labour and intact membranes, and has been shown to prolong pregnancy for 48 hours to 7 days [360-364]. This prolongation of pregnancy has not been associated with a change in perinatal morbidity and mortality [360-364]. It is now clear that tocolytic agents do not work in this setting of subclinical intra-amniotic infection, and indeed, that treatment with beta-adrenergic agents may increase the risk of pulmonary oedema in such patients [365]. It is possible that some of the judgements made about tocolytic effectiveness could be the result of administering these drugs to patients who may not benefit (patients with subclinical infection). This does not mean that there may not be a subgroup that would benefit from efforts to arrest uterine contractions and prolong pregnancy.

The same concept applies to methods for risk assessment or preventive strategies. A positive fetal fibronectin test identified patients at risk for preterm delivery [366-372]. Similarly, a short sonographic cervical length is a risk factor for early preterm delivery [373-381]. Moreover, there is evidence that patients with a short cervix tend to have a positive fetal fibronectin test in vaginal fluid, and such patients are more likely to deliver preterm [369],[382]–[385]. We interpret this evidence as suggesting that the insult responsible for preterm birth may act in an asynchronous fashion, and affect one component of the pathway of parturition more than the others. Similarly, such insult may be of a temporary nature, and may be resolved without resorting to preterm parturition. For example, patients with a positive fibronectin often have a second test which is negative [386]. Although the first test may be interpreted as a false positive, an alternative explanation is that an insult to the utero-placental unit resulted in activation of matrix degradation detectable by an increased concentration of fibronectin in vaginal fluid. However, if such insult is self-limited, this would allow homeostasis to handle the insult and return the membrane-decidual state to its baseline. On the other hand, persistence of the stimulus and further insult may lead not only to persisting activation of one pathway, but recruitment of the others. This would explain why some patients begin with a positive fibronectin, develop a short cervix, and eventually have increased uterine contractility. The more components activated, the more likely it is that the process will evolve from a reversible to an irreversible state of preterm labour.

The methods of prevention may also have a different effectiveness, depending upon the mechanism of disease. For example, a cervical cerclage may be effective in the subpopulation with cervical disease, which is amenable to early diagnosis (e.g. a sonographic short cervix and a prior history of preterm birth, or a patient with active shortening of the cervix in the current pregnancy) [387,388]. However, this intervention may be totally ineffective in patients who have had a previous preterm delivery unrelated to cervical disease. Indeed, a randomized clinical trial of cervical cerclage in women with a cervical length of ≤15 mm demonstrated that this operation is not effective in reducing preterm birth in nulliparous women with an isolated short cervix [389]. If the primary mechanism of disease is infection/inflammation, these patients may benefit from antimicrobial/anti-inflammatory treatment. However, if infection/inflammation is not the mechanism of disease, there is little hope that antimicrobial treatment would prevent preterm birth. This is probably at the heart of why antimicrobial treatment to prevent preterm birth has largely failed. Patients with Group B streptococcus (GBS) or Ureaplasma urealyticum[390] are not at risk for preterm delivery, and therefore, treatment of colonized women has not reduced the rate of preterm birth. The same applies to patients with asymptomatic Trichomonas vaginalis[391]. The odds ratio for preterm delivery is only 1.4, which means that most women with asymptomatic Trichomonas vaginalis will deliver at term, and therefore, cannot benefit from antibiotic administration. Bacterial vaginosis also represents a complex story in which the problems include the adequacy of the diagnosis and the host response [255],[392]–[395]. The interested reader is referred to a previous article in which we have explored the question in depth [109].

We predict that a successful strategy to prevent preterm birth must be based on an understanding of the mechanisms of disease for a particular patient, the development of a method for early detection of this specific pathologic process, as well as treatment targeted to this process. In other words, diagnosis, treatment and prevention aimed at the specific cause of the syndrome rather than to the manifestations of the syndrome.

We do not exclude that there may be primary disorders of the terminal components of the pathway. For example, cervical insufficiency may result from subclinical intrauterine infection, but may also be a primary disease of the cervix. The latter can be identified with cervical sonography and treated with a cervical cerclage, which may prevent preterm delivery even though it does not cure the underlying mechanism of disease responsible for the primary disorder. Additionally, we believe that a primary myometrium disorder may be operative in some patients. Professor Stephen Lye has described that the myometrium undergoes an organized and sequential change of phenotype as it progresses from early pregnancy to labour [317,318],[320],[327], [396-400]. Acceleration of such a timetable may be a cause for preterm delivery. We envision that if a similar program exists in humans, pharmacologic interventions may be used to treat and prevent preterm delivery. For example, tocolysis may be used acutely to achieve uterine rest, while progesterone could be employed to reverse the untimely progression away from a contractile phenotype of the myometrium and towards a quiescent one.

Inflammation as a mechanism for term and preterm parturition

Liggins [401] was the first to liken cervical ripening to an inflammatory response. Since then, accumulating evidence supports the idea that inflammation can be detected in the cervix, myometrium, chorioamniotic membranes and amniotic cavity of women in labour [402-409].

Spontaneous labour at term is associated with the infiltration of inflammatory cells in these tissues and increased production of pro-inflammatory cytokines (IL-1β, IL-6, TNF-α and IL-8) [202],[410] and chemokines (GROα, granulocyte colony stimulating factor [GCSF], GMCSM, neutrophil attractant/activating peptide –1/IL-8, etc) [410]. Recently, using a genome-wide screen, we have been able to demonstrate that genes involved in the control of inflammation are upregulated in the chorioamniotic membranes after labour at term, even in the absence of histological chorioamnionitis [403,404]. Of interest is that an inflammatory signature was not observed in the maternal circulation of patients with spontaneous labour at term [403,404]. This suggests that an inflammatory process is localized to the membranes. These findings may reflect the fact that spontaneous parturition is associated with an increased pro-inflammatory cytokine and chemokine response. We have made similar observations in the myometrium (Romero, unpublished observations), and this is consistent with those made by Professor Roger Smith from Newcastle, Australia [402], as well as Professor Murray Mitchell from Auckland, New Zealand (in the membranes) [410-414]. Thus, the emerging picture is that inflammation is important in term spontaneous labour.

In addition, pathologic inflammation, which is that elicited by the presence of microbial products, microorganisms and other mechanisms of tissue injury, has been implicated in preterm parturition. Thus, the evidence is now compelling that inflammation in the uterine components of the common pathway is present in term and preterm labour, although the severity of the inflammatory process is far greater in the context of subclinical intrauterine infection in preterm labour [406,407].

Microbial invasion of the amniotic cavity and intrauterine inflammation have been associated with microbial invasion of the human fetus [415] and the development of the Fetal Inflammatory Response Syndrome (FIRS) [249],[407],[416-418]. This condition has been operationally defined as an elevation in fetal plasma IL-6 concentration, and is a risk factor for spontaneous preterm labour, neonatal morbidity and mortality, as well as multisystemic involvement (cardiac dysfunction [419,420], adrenal stress response [421], neurologic injury [422-438], as well as lung disease [128],[439-445]) (Figure 3). The relevance of this is that the identification of the patient with intra-amniotic inflammation has become a priority to select a patient in whom tocolysis would be futile, and who may benefit from antimicrobial and anti-inflammatory agents. Of interest is that the lower the gestational age at presentation, the higher the frequency of intra-amniotic infection and inflammation [110],[446]. This has implications for the management of patients with preterm labour, because we have demonstrated that treatment with a tocolytic agent was less effective at gestational ages of <28 weeks of gestation, when infection is more common [360,361]. A full discussion of the role of infection and inflammation in preterm labour and fetal injury is available in recent reviews by our group [407],[418],[447].

Figure 3. Fetal target organs during the fetal inflammatory response syndrome (FIRS): haematopoietic system, adrenals, heart, brain, lungs and skin. CSFs: colony stimulating factors; MMPs: matrix metalloproteinases. Reproduced with permission from reference 269.

Acknowledgement

This research was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), NIH, DHHS.