Is an episode of suspected preterm labor that subsequently leads to a term delivery benign?

The most common cause for admission to the hospital of pregnant women is an episode of suspected preterm labor (1–4) with a frequency ranging from 9% (234/2534) in a prospective cohort study of patients with this diagnosis (3) to 24% (884/3619) based on a database of a large managed care organization in the USA (administrative database) (5). Of patients with suspected preterm labor, approximately one-half deliver at term (3). This information is derived from both observational studies (6–9), randomized clinical trials (10–20), and meta-analysis of such trials (16, 20–26) to which patients were allocated as placebo or acute intravenous tocolysis. The management of women with suspected preterm labor and the decision about the need for hospitalization affect the women and their families as well as the health care system. The estimated cost to the U.S. healthcare system for the treatment of suspected preterm labor exceeded $800 million in 2005 (27, 28). It is likely that this cost has increased because there is no evidence that the prevalence of suspected preterm labor has decreased (29), and the cost of hospitalization and medical care has increased in the last decade.

There is a widespread belief that if an episode of preterm labor resolves spontaneously or after treatment with acute tocolysis, the major risk is recurrence of an episode of suspected preterm labor (24) and preterm delivery. In contrast, if the patient delivers at term, it is thought that the episode of preterm labor was benign and sometimes labeled as “false preterm labor,” a designation that appears inappropriate for many reasons to be discussed herein. It is also thought that there are no long-term effects for the fetus, neonate, or infant if birth occurs at term. Accumulating evidence reported in recent years and published in this issue of the Journal suggests that this conclusion may not be correct and requires scrutiny (30–32).

The first evidence that an episode of preterm labor leading to term delivery may not be entirely benign came from a study by Espinoza et al. (31) who reported that the prevalence of small-for-gestational-age (SGA; defined as <10th percentile of birthweight) neonates was significantly higher for those who experienced an episode of suspected preterm labor and were born at term than for those who were delivered preterm [21.5% (64/298) vs. 12.7% (60/474); p=0.001; odds ratio (OR), 2.2; 95% confidence interval (CI), 1.3–3.9, after controlling for confounding variables] (31). Patients were included in this study if they had an episode of suspected preterm labor (20–36 weeks) defined as increased uterine contractility with intact membranes requiring hospitalization and follow-up until delivery. Two subsequent studies confirmed the higher rate of SGA neonates for patients with an episode of preterm labor who subsequently delivered at term (30, 32).

In a large retrospective study including deliveries occurring over a period of 20 years (between 1989 and 2009) at the Soroka Medical Center, Ben Gurion University, Beer-Sheva, Israel, the frequency of SGA birth at term in patients who delivered after an episode of suspected preterm labor was also higher than in patients who delivered preterm [10.1% (191/1897) vs 5.6% (22/393); p=0.005; crude OR, 1.88; 95% CI, 1.19–2.98] (32). Further confirmation derives from a prospective cohort study conducted in Canada, in which patients with an episode of suspected preterm labor had an OR of 3.9 (95% CI, 1.3–11.4) to deliver an SGA neonate at term (30). Collectively, the clinical epidemiological evidence is consistent, indicating that a patient with an episode of suspected preterm labor who subsequently delivers a neonate at term (with or without tocolysis) is at increased risk for delivering an SGA neonate. Why?

Placental vascular lesions in patients with preterm labor who deliver an SGA neonate at term

To address the mechanisms whereby an episode of preterm labor may be associated with the delivery of an SGA neonate at term, Espinoza et al examined the frequency of placental histological findings for women with an episode of preterm labor who delivered an SGA vs. an appropriate-for-gestational-age (AGA) neonate at term: patients who delivered at term had a higher frequency of maternal vascular lesions of underperfusion or fetal thrombotic lesions than those who delivered preterm [29.1% (43/148) vs. 18.9% (65/344); p=0.02] (31).

These results were novel and contrary to expectation (indeed, placentas are normally not examined in patients who had an episode of preterm delivery and subsequently delivered at term). What was known at the time was that patients who delivered preterm after spontaneous preterm labor or preterm prelabor rupture of the membranes (PROM) had an excess rate of vascular lesions of the placenta (33, 34). However, this was interpreted as a potential etiology for spontaneous preterm delivery. The specific lesions, found in excess in placentas delivered after spontaneous preterm labor, were diagnosed as maternal decidual vasculopathy (defined as the presence of failure of physiological transformation of the decidual portion of the spiral arteries and organized thrombi and multiple placental infarcts) (35–38). The proposed operative mechanism of disease responsible for preterm labor in these cases is utero-placental ischemia (39). There are multiple lines of evidence in support of this view, but perhaps the most compelling is that clamping the uterine arteries in non-human primates to generate an animal model of preeclampsia often failed because the animals went into preterm labor (40).

Further evidence that vascular lesions are implicated in preterm labor was strengthened by studies of the placental bed (to detect disorders of deep placentation and failure of physiological transformation of the spiral arteries). Kim et al. (38) showed that failure of physiological transformation of the myometrial segment of the spiral arteries is more common in patients with an episode of preterm labor or with preterm PROM who delivered preterm than in patients who delivered at term. Although originally thought to be typical of preeclampsia and SGA (41–48), it is now well established that a disorder of deep placentation with failure of physiological transformation of the spiral arteries, and even atherosis, occurs in the “great obstetrical syndromes,” including preterm labor and preterm PROM (49). The reader is referred to reference 34 for a discussion of why similar vascular lesions may present differently among the great obstetrical syndromes. Issues of time, severity, widespread involvement, timing, and additional insults/pathological processes all may play a role (50). Altogether, it seems that the most severe vascular lesions would be associated with preterm labor progressing to preterm delivery, while less severe lesions may predispose to underperfusion of the uterus with increased uterine contractility, which can be compatible with the continuation of pregnancy, but this may come at a price, namely, the deceleration of fetal growth predisposing to the birth of an SGA neonate.

Decelerations of fetal growth following an episode of preterm labor

The evidence that SGA was more common in patients who delivered after spontaneous preterm labor originally came from cross-sectional studies (51–55). However, the most compelling evidence derives from longitudinal studies in which fetuses experiencing an episode of suspected preterm labor were followed with fetal biometry. Lampl et al. (56, 57) reported a series of seminal observations indicating that, in patients who presented with an episode of preterm labor, neonates delivered at term were relatively smaller than their peers at birth, compared to the end of the second trimester (after stratification by sex, internal z-scores, birth standard scores) (56). Moreover, Lampl et al. (56) were able to show a significant downward crossing in the percentile of growth, indicating a disturbance of individual growth trajectories – meaning that the proportion of fetuses experiencing downward centile crossing was greatest among those who experienced an episode of preterm labor (p=0.05) (56).

In a separate study, Lampl et al. (57) called attention to the complexity of the relationship between fetal growth and preterm labor. The authors reported that, in early pregnancy, acceleration of fetal growth predisposes to an episode of preterm labor (57). This could be interpreted as a mismatch between nutrient supply and fetal demands for growth (58). These findings were consistent with an earlier work reported by our team in which fetal growth acceleration was associated with preterm labor and delivery in a subset of patients (59). Collectively, this suggests that there is a relationship between fetal growth and the initiation of parturition at term (60) as well as preterm.

Is a fetus exposed to an episode of suspected preterm labor at risk for multiple neurodevelopmental disorders in childhood?

In this issue of the Journal, Paules and her colleagues (61) from the University of Zaragoza and Johns Hopkins University report the results of a prospective cohort study in which the investigators examined the outcome of infants at two years of age who had experienced an episode of preterm labor during gestation but were delivered at or near term. The cohort included three groups: 1) women without an episode of preterm labor who delivered at term (≥37 weeks; control group) (n=42); 2) women with an episode of preterm labor who delivered late preterm (32–36 weeks) (n=22); and 3) women who had an episode of preterm labor and delivered at term (n=23).

The key finding was that infants at the age of two years who had been born at term after an episode of suspected preterm labor scored lower in the global cognitive index, cognition, fine and gross motor skills, memory, receptive language, speed of processing, and visual motor coordination than children who were born at term who had not experienced an episode of preterm labor (after adjustment for gestational age at delivery, birthweight, neonatal gender, cesarean delivery, and maternal education) [see Table 3 from the article by Paules et al (61)].

What other factors may explain the changes in neurodevelopmental outcomes in babies born at term after an episode of suspected preterm labor reported by Paules et al.?

One possibility is that the association between an episode of suspected preterm labor and adverse outcome is due to the presence of intra-amniotic inflammation. Recent evidence indicates that increased preterm uterine contractility, even in the absence of the cervical change [which does not mean the traditional definition of preterm labor (62)], is associated with an increased frequency of intra-amniotic inflammation determined by amniocentesis and an elevation in amniotic fluid matrix metalloproteinase-8 (MMP-8) concentrations (≥23 ng/mL). In a cohort study by Kim et al. (62) of 132 patients with an increased frequency of premature contractions without cervical changes who underwent amniocentesis to rule out the presence of intra-amniotic infection and/or inflammation, 12.1% (16/132) had proven intra-amniotic inflammation. Interestingly, only 32.8% (38/116) of these patients delivered preterm (p<0.001), indicating that not all patients with intra-amniotic inflammation progress to spontaneous preterm delivery and that some may experience chronic intra-amniotic inflammation, a condition that has not been recognized until recently and may represent a hostile intra-uterine environment for the fetus.

Is there evidence that chronic intra-uterine inflammation without preterm delivery may be associated with neurodevelopmental disorders?

The consequences of chronic exposure to subclinical intra-amniotic inflammation have only begun to be recognized. First, in women undergoing mid-trimester amniocentesis and who have intra-amniotic inflammation determined by MPP-8 or interleukin (IL)-6, a fraction delivered at term (63). Although these processes haven largely overlooked, there is now evidence that these infants are at increased risk for motor and cognitive disabilities (62–64). The association between intra-amniotic inflammation and possibly neonatal brain injury has been demonstrated in an animal model in which intra-uterine inflammation was induced by injection of low-dose lipopolysaccharides. At this dose, preterm labor did not occur and the animals delivered at term. However, the neonates exposed to endotoxin in utero had severe neuroinflammation detected with positron emission tomography and microglial activation, which was proven by histological examination of the brain after euthanasia. Moreover, the neonates showed evidence of severe motor disorder (65). Similar observations have been made by Elovitz et al. (66).

The totality of the evidence suggests the following: 1) patients with an episode of suspected preterm labor may have subclinical intra-amniotic inflammation of unknown etiology, which has recently been attributed to danger signals (67–69); 2) some of these patients do not progress to preterm delivery and may enter a state of chronic intra-amniotic inflammation to which the fetus may be exposed; 3) experimental evidence in animal models shows that this process can lead to fetal neuroinflammation and damage (65); and 4) interestingly, this process can be subject to treatment with targeted nanodevices (i.e., dendrimers) if detected early (70–75).

Are infants born at term after an episode of preterm labor at risk for neurodevelopmental disorders?

The findings of this study published in this issue of the Journal by Dr. Cristina Paules and her associates (61) are novel, potentially important, and interesting for the understanding of the most common complication requiring admission of pregnant women to the hospital. The authors are correct in calling for replication and further research, given the potential clinical implications. The investigators point out that a larger sample size is desirable and that infants should be evaluated at more than one single time, as this may limit the ability to determine the full capabilities of a child examined in a professional setting than at home. Also, careful consideration should be given to comorbidities and the potential effect of interventions such as tocolysis and steroids. It is noteworthy that 100% of the children born at term after a suspected episode of preterm labor and only 55% of those who experienced an episode of preterm labor and were delivered late preterm received antenatal glucocorticosteroids. Recent experimental and observational studies in human necropsies on preterm babies who died after receiving glucocorticosteroids in utero showed that these powerful agents induce dramatic apoptosis in hippocampal neurons in animals as well as human fetuses, an effect that is dramatic, and its impact is not understood (76, 77). This effect is potentially important, given the inclination to repeat steroid administration during fetal life at time of rapid neuronal/glial development.

Another factor that may be important is the potential effect of the tocolytic agents. For example, the administration of terbutaline has been implicated in the genesis of autism in infants exposed because of an episode of preterm labor (78). Although this has been a controversial subject that has not gained popularity, it is prudent to remember that birth is associated with a dramatic change in the γ-aminobutryic acid secretion neurotransmitters, and that this process may be susceptible to endocrine and paracrine regulation during parturition (79, 80).

Should the findings lead to a change in clinical management?

The report of Paules et al. (61) is important and hence has been highlighted in the pages of the Journal. However, we agree with the authors that no change in the clinical management is required other than being prudent in the way we counsel patients about the long-term prognosis after an episode of suspected preterm labor and delivery at term. We believe that it is appropriate to counsel patients that they may be at risk for delivering an SGA neonate based on three epidemiological studies (30, 32) reporting consistent results; however, any counseling about the risk of neurodevelopmental disorders must await replication of these findings. It is important to note that the absolute magnitude of risk is small and most of the infants would be unaffected. There is a need to refine the nomenclature for “suspected preterm labor,” “threatened preterm labor,” and “true vs. false preterm labor,” which have been used in the literature, not always with clarity.

Are there biomarkers that identify patients with suspected preterm labor and maternal vascular lesions who subsequently deliver SGA neonates?

The fundamental mechanism of disease responsible for the excess rate of vascular lesions in the placenta of patients with spontaneous preterm labor has been unknown. However, hints have begun to emerge. The balance of angiogenic and anti-angiogenic factors appears to be key in the establishment of a successful and functional utero-placental circulation (81–86). A deficiency in angiogenic factors was first implicated in the pathogenesis of preeclampsia (87–95) and subsequently discovered to play a role in SGA/fetal growth restriction (93, 96), fetal death(97, 98), spontaneous preterm labor (99), Ballantyne syndrome (100), twin-to-twin transfusion syndrome (101), etc. During the course of prospective studies, we observed that some patients with a high concentration of endoglin (an anti-angiogenic factor) had a spontaneous preterm delivery (99). This anti-angiogenic factor is higher in patients destined to deliver an SGA neonate (93). Moreover, among women with preterm labor who delivered preterm, those with placental maternal vascular lesions of underperfusion, but not inflammatory lesions, had a higher median maternal plasma endoglin concentration 5–10 weeks prior to the clinical presentation of preterm labor that led to preterm birth (99). Recently, a case-cohort study identified that the ratio of placental growth factor to soluble vascular endothelial growth factor receitor-1 below 2.5 centile is a biomarker for maternal vascular lesions of underperfusion, and that this biomarker is abnormal not only in women who subsequently develop preeclampsia and an SGA neonate, but also in those who subsequently develop preterm labor with maternal vascular lesions of underperfusion (85). Therefore, this ratio could be used as a biomarker in patients admitted with an episode of suspected preterm labor. Whether this biomarker can identify women with suspected preterm labor who develop an SGA neonate remains to be determined. If this proves to be the case, then therapeutic interventions and intense surveillance may be warranted to prevent adverse pregnancy outcomes, even at term. The emerging picture is that biomarkers and improved understanding of the mechanisms of disease responsible for the great obstetrical syndromes are beginning to yield benefits in the prevention of obstetrical disorders.