CHAPTER 42: Preterm Birth

In this textbook’s first edition, very little was mentioned regarding preterm birth. Indeed, preterm birth was not incorporated as a stand-alone topic until the 13th edition in 1966. And, this content totaled only three sentences that cited use of isoxsuprine as a tocolytic agent. In contrast, present-day research now produces more than 3000 articles published annually. Data derive from study of animal models, translational research, clinical trials, and genetic investigations. Despite efforts, elucidating the biology of human parturition and the subsequent efforts to prevent preterm birth remain elusive (Martin, 2017).

Low birthweight defines neonates who are born too small. Preterm or premature birth describes neonates who are born too early. With respect to gestational age, a newborn may be preterm, term, or postterm. With respect to size, a newborn may be normally grown and appropriate for gestational age; undersized, thus, small for gestational age; or overgrown and consequently, large for gestational age. Small for gestational age categorizes newborns whose birthweight is <10th percentile for gestational age. Other frequently used terms have included fetal-growth restriction or intrauterine growth restriction. The term large for gestational age describes newborns whose birthweight is >90th percentile for gestational age. The term appropriate for gestational age designates newborns whose weight is between the 10th and 90th percentiles.

Thus, neonates born before term can be small or large for gestational age, but still preterm by definition. Low birthweight refers to neonates weighing 1500 to 2500 g; very low birthweight are those between 1000 and 1500 g; and extremely low birthweight refers to those between 500 and 1000 g.

Before the 15th edition of this textbook, a preterm or premature newborn was defined by a birthweight <2500 g. With that edition, preterm neonates were considered to be those delivered before 37 completed weeks, that is, <36^6/7 weeks (Pritchard, 1976). This definition, which has now been in use for more than 40 years, was first promulgated in 1976 by the World Health Organization (WHO) and the International Federation of Gynecology and Obstetrics (FIGO). The definition derived from a statistical analysis of gestational age distribution at birth (Steer, 2005). Importantly, the denotation lacks a specific functional basis and should be clearly distinguished from the concept of prematurity. Prematurity represents incomplete development of various organ systems at birth. For example, the lungs are particularly affected, leading to the respiratory distress syndrome (RDS) (Chap. 34, Respiratory Distress Syndrome).

In 2013 in the United States, 23,446 infants died in their first year of life, and a third of infants died from preterm-related causes (Matthews, 2015). Gestational age at delivery and the risk of neonatal morbidity and mortality are inversely related (Frey, 2016). Namely, neonates born in the early-preterm period make up the smallest proportion of births, but these infants experience disproportionately higher rates of prematurity-related complications (Table 42-1).

Beginning in 2005, in recognition that neonates born between 34^0/7 weeks and 36^6/7 weeks experience morbidities and mortality characteristic of premature newborns, preterm births were subdivided. Those before 33^6/7 weeks are labeled early preterm, and those occurring between 34 and 36 completed weeks are late preterm. Indeed, compared with births at 39^0/7 weeks through 40^6/7 weeks, these late-preterm infants experience morbidities that are also associated with prematurity (Spong, 2013). Recently, this concept has expanded to births 37^0/7 weeks through 38^6/7 weeks, which are now defined as early term, and those 39^0/7 weeks through 40^6/7 weeks, which are defined as term.

This revised terminology has led some to redefine a short gestation as those <39^0/7 weeks. By doing so, more than a third of live births in the United States in 2015 would be defined as having a shortened period of gestation (Martin, 2017). One implication is that only 65 percent of births in the United States occurred during the optimal 39 to 41 weeks’ gestation. This emphasizes the realization that fetal maturation in humans is a continuum that is completed later in human pregnancy than previously appreciated. As a result, adverse neonatal sequelae from neonatal immaturity with elective delivery before 39 completed weeks are appreciable (Reddy, 2009; Tita, 2009).

This knowledge resulted in the development and application of the “39-week rule” to deter nonmedically indicated deliveries before 39 weeks (Spong, 2011). An unintended consequence of this health-care strategy has been a rise in stillbirth rates in the United States. One concern is that the rule may be misapplied to gestations with true medical indications for early delivery (Hill, 2017; Nicholson, 2016). Spong (2016) has emphasized the need to perform necessary obstetrical interventions when indicated.

Methodology

In the United States, the preterm birth rate rose slightly from 9.57 percent in 2014 to 9.63 percent for 2015 (Martin, 2017). This marks the first rise in this percentage since 2007. Although concerning, some argue that the drop in preterm birth rates from 2007 to 2014 reflected systematic bias associated with changes in obstetrical dating (Frey, 2016).

Specifically, beginning with the 2014 data year, the National Vital Statistics Reports from the National Center for Health Statistics transitioned to a new standard for estimating newborn gestational age for birth certificate completion (Martin, 2015). The new measure—obstetrical estimate of gestational age at delivery—replaced calculations based on the date of the last normal menses (Chap. 44, Fetal Growth versus Birthweight). As shown in Figure 42-1, these measures differ and do not provide equivalent absolute numerical comparisons of preterm birth rates. For example, the 2015 obstetrical estimate-based preterm birth rate was 9.6 percent compared with the last menstrual period-based rate of 11.3 percent (Martin, 2017). Thus, current national data are now not directly comparable to previously reported rates of preterm birth due to differing gestational age calculation methodologies. The national data are now reported starting with year 2007, which coincides with the year that this information became available.

Trend Factors

Ferré and colleagues (2016) using National Vital Statistics System data showed that the overall reduction of preterm birth before 2014 was in part due to the changes in maternal age distribution. Specifically, teen birth rates declined. This translated into a drop in preterm birth rates over the same epoch and could explain the lower infant mortality rates (Callaghan, 2017).

One disturbing aspect of preterm birth rate trends in the United States is persistent racial and ethnic disparities. Rates of preterm birth among black women are markedly elevated above those for white and Hispanic women in every year recorded (Martin, 2017). Moreover, rates of births before 32 completed weeks in black women are higher than those in white and Hispanic women combined. Some investigators attribute this disparity to socioeconomic circumstances (Collins, 2007; Leveno, 2009).

Internationally, the rates of preterm birth in the United States are also higher compared with those in other industrialized countries (Ananth, 2009; Delnord, 2017; Martin, 2017).

Newborns born before 37 weeks suffer various morbidities, largely due to organ system immaturity (Table 42-2). That said, remarkable strides have been made in neonatal survival for those born preterm. This is especially true for neonates born after 28 weeks. In a study of more than 18,000 newborns weighing between 400 and 1500 g or aged between 22 and 32 weeks’ gestation, survival rates were analyzed as a function of both birthweight and gestational age (Fanaroff, 2007). After achieving a birthweight of ≥1000 g or a gestational age of 28 weeks for females, or 30 weeks for males, survival rates reach 95 percent.

Threshold of Viability

Births once considered to be “abortuses” because the fetus weighed <500 g are now classified as live births. In the United States in 2014, 5863 live births <500 g were recorded (Martin, 2017). For those newborns delivered before 33 weeks’ gestation, perinatal and neonatal care has advanced tremendously. As a result, the threshold of viability, which is the lower limit of fetal maturation compatible with extrauterine survival, has been reassessed. Currently, the threshold of viability lies between 20 and 26 weeks’ gestation.

Neonates born in this periviable period have been described as fragile and vulnerable because of their immature organ systems. Many of these are described in Chapter 34 (Retinopathy of Prematurity) and include brain injury from hypoxic-ischemic injury and sepsis. In this setting, hypoxia and sepsis start a cascade of events that lead to brain hemorrhage, to white-matter injury that causes periventricular leukomalacia, and to poor subsequent brain growth eventuating in neurodevelopmental impairment. Associated morbidities include intellectual disability, cerebral palsy, blindness, seizures, and spastic quadriparesis that can result in the need for a lifetime of medical care (Annas, 2004). Because active brain development normally occurs throughout the second and third trimesters, those born <25 weeks are believed to be especially vulnerable to brain injury.

To clarify obstetrical care of these fetuses, the Society for Maternal-Fetal Medicine, the Eunice Kennedy Shriver National Institute for Child Health and Human Development (NICHD), the American Academy of Pediatrics, and the American College of Obstetricians and Gynecologists convened a joint workshop in 2013 (Raju, 2014). The executive summary statement from this meeting served as the underpinnings for an Obstetric Care Consensus document from the American College of Obstetricians and Gynecologists (2017e).

Periviable Neonatal Survival

The Obstetric Care Consensus summary provides a review of outcomes for those born in the periviable period. Delivery before 23 weeks typically results in death, and survival rates approximate only 5 percent (Fig. 42-2). Among those that live, morbidity is nearly universal. Notably, the authors highlight the wide variation in practices regarding active resuscitation and suggest that these variations may explain the differing perinatal outcomes among different institutions. An important caveat, however, is ascertainment bias. For example, the mean survival rate is 45 percent if the denominator is all live births compared with 72 percent if the denominator is only newborns admitted to neonatal intensive care (Guillen, 2011). Another source of bias is use of multicenter datasets with considerable differences in obstetrical and early neonatal interventions, particularly at 22 and 23 weeks’ gestation (Stoll, 2010).

To evaluate contemporaneous outcomes of neonates born at 22 to 24 weeks, the NICHD Neonatal Research Network reported both survival and neurodevelopmental outcomes assessed across consecutive birth-year epochs of 2000 to 2003, 2004 to 2007, and 2008 to 2011 in infants aged 18 to 22 months (Younge, 2017). The percentage of infants who survived rose significantly from 30 percent in 2000 to 2003 to 36 percent in 2008 to 2011. The percentage of infants who survived without neurodevelopmental impairment also significantly grew from 16 percent to 20 percent during the same time period (Fig. 42-3). Although rates of survival without neurodevelopmental impairment increased over time among infants born at 23 and 24 weeks, only 1 percent of infants born at 22 weeks survived without neurodevelopmental impairment (Younge, 2017).

Somewhat similar results were published from Sweden (Serenius, 2013). This report details a national population-based prospective study of all neonates born before 27 weeks. Shown in Table 42-3 are the survival and disability rates for 707 Swedish infants born alive from 22 to 26 weeks’ gestation between 2004 and 2007 in Sweden. Compared with rates in the United States, rates of survival without neurodevelopmental impairment were higher in the Swedish cohort for infants born at 24 weeks during 2004 to 2007.

Clinical Management

The Obstetric Care Consensus document also addresses management options based on the clinical characteristics of a given pregnancy. Nonmodifiable factors are fetal gender, weight, and plurality. Potentially modifiable antepartum and intrapartum factors include the location of delivery, intent to intervene by cesarean delivery or labor induction, and administration of antenatal corticosteroids and magnesium sulfate. Postnatal management addresses the initiation or withdrawal of intensive care after birth. Areas of general guidance were then reviewed for each week of gestation (Table 42-4).

Mode of delivery represents another dilemma because cesarean delivery at the threshold of viability is controversial. For example, if the fetus-neonate is perceived to be too immature for aggressive support, then cesarean delivery for common indications such as breech presentation or nonreassuring fetal heart rate patterns might be preempted. Moreover, observational studies have failed to document a benefit of cesarean delivery for the sole indication of periviability (Alfirevic, 2013).

In a study of 2906 singletons between 24^0/7 and 31^6/7 weeks eligible for attempted vaginal birth, 84 percent of cephalic presenting fetuses were delivered vaginally (Reddy, 2012). Neonatal mortality rates did not differ compared with those associated with planned cesarean delivery. For breech presentations, however, relative risk for mortality was threefold higher with attempted vaginal delivery. In another study, Werner and colleagues (2013) analyzed 20,231 newborns delivered at 24 to 34 weeks. Cesarean delivery did not protect against poor outcomes such as neonatal death, intraventricular hemorrhage, seizures, respiratory distress, and subdural hemorrhage. From these findings, the Obstetric Care Consensus proposes that cesarean delivery be considered for fetal indications at 23^0/7 to 24^6/7 weeks. However, before 22 weeks, this route is reserved only for maternal indications.

It is difficult to summarize the current practices of obstetrical care in the management of the periviable pregnancy given its rapid evolution. For example, since January 2016, the American College of Obstetricians and Gynecologists has penned three iterations of its Practice Bulletin “Management of Preterm Labor,” and three versions of the “Obstetric Care Consensus” document have been published since November 2015. In this uncertain environment, individualized, patient-centered care with a multidisciplinary team remain as bedrock for the clinician.

While we do not have the answers, we describe our strategies from Parkland Hospital as one approach to management. Our policies were developed in conjunction with the Division of Neonatal Medicine. Importantly, the decision not to perform cesarean delivery does not necessarily imply that care for the fetus is discounted. Neonatologists are consulted before delivery, and a trilateral discussion of survival and morbidity rates ensues with the woman and her family. A neonatologist attends each delivery and determines subsequent management.

In our institution, traditional fetal indications for cesarean delivery are practiced in women at 25^0/7 weeks or beyond. Cesarean delivery is not offered for fetal indications before 24^0/7 weeks. At 24^0/7 weeks, cesarean delivery is not offered unless fetal weight is estimated at 750 g or greater. Aggressive obstetrical management is practiced in cases of growth restriction, wherein gestational age is used to guide management rather than fetal size.

Late-Preterm Birth

Neonates born between 34 and 36 weeks account for more than 70 percent of all preterm births (Fig. 42-4). This group has been the fastest rising proportion of singleton preterm births in the United States (Raju, 2006). Nationally, the late-preterm birth rate rose from 6.82 percent to 6.87 percent from 2014 to 2015 (Martin, 2017).

To estimate the risks associated with late-preterm births, investigators analyzed neonatal mortality and morbidity rates at 34, 35, and 36 weeks compared with those of births at term between 1988 and 2005 at Parkland Hospital (McIntire, 2008). Approximately 3 percent of all births during the study period were between 24 and 32 weeks, and 9 percent were during the late-preterm weeks. Thus, and similar to the national rates, late-preterm births accounted for three fourths of all preterm births. Approximately 80 percent of these resulted from idiopathic spontaneous preterm labor or prematurely ruptured membranes (Fig. 42-5). Other obstetrical complications were implicated in the remaining 20 percent of cases. Rates of morbidity and mortality were greater in these late-preterm newborns compared with rates in term ones (Table 42-5 and Fig. 42-6). Similarly, Tomashek (2017) also reported higher neonatal mortality rates for late-preterm newborns. Rates of adverse neurodevelopment outcomes are also increased in these late-preterm infants (Petrini, 2009).

Taken together, these findings suggest that a health-care focus on prematurity should include these late-preterm births. Moreover, approximately 80 percent of affected women begin labor spontaneously—similar to births before 34 weeks—and attempts to interrupt preterm labor have been insufficient (Institute of Medicine, 2007). Because of this, a national strategy aimed at prevention of late-preterm births is unlikely to provide discernible benefit without new developments in the prevention and management of preterm labor. In the meantime, the American College of Obstetricians and Gynecologists (2017c) emphasizes that intentional late-preterm delivery should occur only if an accepted maternal or fetal indication for delivery exists.

Four direct causes for preterm births in the United States include: (1) spontaneous unexplained preterm labor with intact membranes, (2) idiopathic preterm premature rupture of membranes (PPROM), (3) delivery for maternal or fetal indications, and (4) twins and higher-order multifetal births. Of all preterm births, 30 to 35 percent are indicated, 40 to 45 percent are due to spontaneous preterm labor, and 30 to 35 percent follow preterm membrane rupture (Goldenberg, 2008). Indeed, much of the increase in the singleton preterm birth rate in the United States is explained by rising numbers of indicated preterm births (Ananth, 2005). Last, more than one of every two twins and more than nine of every ten triplets are born preterm or with low birthweight in the United States (Chap. 45, Unique Fetal Complications) (Martin, 2017).

Reasons for preterm birth have multiple, often interacting, antecedents and contributing factors (Esplin, 2016). This is particularly true for PPROM and spontaneous preterm labor.

Analogous to other complex disease processes, multiple coexistent genetic alterations and environmental factors may lead to preterm birth (Esplin, 2005; Velez, 2008; Ward, 2008). For example, inherited mutations in genes regulating collagen assembly may predispose to cervical insufficiency or prematurely ruptured membranes (Anum, 2009; Wang, 2006; Warren, 2007). And, whole blood gene expression and proteonomic biomarkers are now being used to help identify predictors of preterm birth (Cantonwine, 2016; Heng, 2016).

Spontaneous Preterm Labor

For both clinical and research purposes, pregnancies with spontaneous preterm labor yet intact fetal membranes must be distinguished from those complicated by preterm prematurely ruptured membranes. Even so, those with spontaneous preterm labor do not constitute a homogeneous group. Among the more common associated findings are multifetal pregnancy, intrauterine infection, bleeding, placental infarction, premature cervical dilation, cervical insufficiency, hydramnios, uterine fundal abnormalities, and fetal anomalies. Severe maternal illness from infections, autoimmune diseases, and gestational hypertension also raises preterm labor risks.

Despite their diversity, these processes culminate in a common end point—premature cervical dilation and effacement and premature activation of uterine contractions. Importantly, the actual process of preterm labor should be considered a final step stemming from progressive or acute changes that could be initiated days or even weeks before labor onset. Indeed, many forms of spontaneous preterm labor that result from premature initiation of phase 2 of parturition may be viewed in this light (Chap. 21, Phase 2: Preparation for Labor). Although the end result in preterm birth is the same as at term with cervical ripening and myometrial activation, recent studies in animal models support the idea that preterm birth is not always an acceleration of the normal process. Diverse pathways to instigate parturition exist and are dependent on the etiology of preterm birth. Four major causes include uterine distention, maternal–fetal stress, premature cervical changes, and infection.

Uterine Distention

Multifetal pregnancy and hydramnios are well-recognized risks for preterm birth. Early uterine distention likely acts to initiate expression of contraction-associated proteins (CAPs) in the myometrium. The CAP genes that are influenced by stretch include those coding for gap-junction proteins such as connexin 43, for oxytocin receptors, and for prostaglandin synthase (Korita, 2002; Lyall, 2002; Sooranna, 2004). More recent reports suggest that levels of gastrin-releasing peptides (GRPs) are increased with stretch to promote myometrial contractility. GRP antagonists can inhibit uterine contractility (Tattershell, 2012). Also, a stretch-induced potassium channel—TREK-1—is upregulated during gestation and downregulated in labor. This suggests a potential role in uterine relaxation during pregnancy (Buxton, 2010).

Excessive uterine stretch also leads to early activation of the placental–fetal endocrine cascade shown in Figure 21-10. The resultant early rise in maternal corticotropin-releasing hormone and estrogen levels can further enhance the expression of myometrial CAP genes (Warren, 1990; Wolfe, 1988). Finally, the influence of uterine stretch should be considered with regard to the cervix. Prematurely increased stretch and endocrine activity may initiate events that shift the timing of uterine activation, including premature cervical ripening.

Maternal–Fetal Stress

Stress is defined as a condition or adverse circumstance that disturbs the normal physiological or psychological functioning of an individual. Examples of stressors are nutrient restriction, obesity, infection, and diabetes. Psychological duress can include racial discrimination, childhood stress, depression, or posttraumatic stress syndrome (Gillespie, 2017; Goldstein, 2017; Shaw, 2017). A quantitative measure of “stress” is difficult. Yet, considerable evidence shows a correlation between some degree of maternal stress and adverse birth outcomes that include stillbirth, preterm birth, and abnormal fetal development (Hobel, 2003; Ruiz, 2003). Factors that activate this cascade likely are broad and influence the stress response.

One potential mechanism for stress-induced preterm birth is premature activation of the placental–adrenal endocrine axis. One trigger may be elevations in cortisol from maternal psychological stress (Lockwood, 1999; Petraglia, 2010; Wadhwa, 2001). Activation of this axis yields rising maternal serum levels of placental-derived corticotropin-releasing hormone (CRH). This raises adult and fetal adrenal steroid hormone production and promotes early loss of uterine quiescence (Fig. 21-10).

If preterm delivery is associated with early activation of the fetal adrenal–placental endocrine axis, maternal estrogen levels would likely be prematurely elevated. Indeed, an early rise in serum estriol concentrations is noted in women with subsequent preterm labor (Heine, 2000; McGregor, 1995). Physiologically, this premature rise in estrogen levels may alter myometrial quiescence and accelerate cervical ripening.

Another mechanism by which stress may translate to preterm birth is premature cellular senescence. As part of normal physiology, aging of fetal and decidual cells precipitates the release of uterotonic signals for uterine activation at term (Menon, 2014a). Animal studies supported by correlative studies in women demonstrate that accelerated senescence of the decidua results in preterm birth (Cha, 2013; Hirota, 2010). In addition, premature cellular senescence may contribute to PPROM (Menon, 2016).

Cervical Dysfunction

In most cases, premature cervical remodeling precedes premature labor onset. In some instances, cervical dysfunction of either the epithelia or stromal extracellular matrix is the underlying cause. For example, an intact cervical epithelial barrier is critical to prevent ascending infection. Disruption of this barrier, such as that in mice lacking the glycosaminoglycan hyaluronan, predisposes mice to ascending infection and preterm birth (Akgul, 2014). Interestingly, the enhanced risk of preterm birth from group B streptococcal colonization may be in part due to the bacteria’s ability to secrete hyaluronidase. This enzyme degrades hyaluronic acid in the cervicovaginal epithelia to aid bacterial ascension (Vornhagen, 2017).

Second, the mechanical competence of the cervix can be reduced. For example, genetic mutations in components of collagen and elastic fibers or proteins required for their assembly are risk factors for cervical insufficiency, PPROM, and preterm birth (Anum, 2009; Nallasamy, 2017; Pyeritz, 2000).

Infection

A patent female reproductive tract, although essential for conception and delivery, is theoretically problematic during phase 1 of parturition. Bacteria can gain access to intrauterine tissues through: (1) transplacental transfer of maternal systemic infection, (2) retrograde flow of infection into the peritoneal cavity via the fallopian tubes, or (3) ascending infection with bacteria from the vagina and cervix. Because the lower pole of the fetal membrane–decidual junction is contiguous with the cervical canal orifice, this anatomical arrangement provides a passageway for microorganisms. Ascending infection is considered to be the most common entry route. Ascending microorganisms colonize the cervix, decidua, and possibly the membranes, where they then may enter the amnionic sac.

Intraamnionic infection as a primary cause of preterm labor in pregnancies with intact membranes accounts for 25 to 40 percent of preterm births (Goncalves, 2002; Iams, 1987). In some instances, histological evidence of inflammation is found in the fetal membranes, decidua, or umbilical cord. Other cases are deemed “subclinical.” Current data suggest that microbial invasion of the reproductive tract is sufficient to induce infection-mediated preterm birth. Affected women are more likely to develop clinical chorioamnionitis and PPROM compared with women with sterile cultures. Moreover, their neonates are also more likely to have perinatal complications that include RDS, intraventricular hemorrhage, and necrotizing enterocolitis (Hitti, 2001). Although the clinical course is more severe when intraamnionic infection is obvious, inflammation in the absence of detectable intraamnionic microorganisms—termed sterile intraamnionic inflammation—is also a risk factor for an inflammatory response, described in the next section (Lee, 2007, 2008; Romero, 2014). In sum, the earlier the onset of preterm labor, the greater the likelihood of underlying infection (Goldenberg, 2000; Goncalves, 2002; Watts, 1992).

The incidence of culture-positive amnionic fluid collected by amniocentesis during spontaneous term labor is similar to that with preterm labor (Gomez, 1994; Romero, 1993). It has been suggested that at term, amnionic fluid is infiltrated by bacteria as a consequence of labor, whereas in preterm pregnancies, bacteria represent an inciting cause of labor. Thus, fetal infection, as defined by bacteria detected within the amnionic fluid, has both differing etiologies and consequences.

Despite these observations, considerable data associate chorioamnionitis with preterm labor (Goldenberg, 2002; Üstün, 2001). With chorioamnionitis, microbes may invade maternal tissue only and not amnionic fluid. Despite this, endotoxins can stimulate amnionic cells to secrete cytokines that enter amnionic fluid. This scenario may serve to explain the apparently contradictory observations concerning an association between amnionic fluid cytokines and preterm labor in cases in which microbes are not detected in the amnionic fluid.

Inflammatory Responses

Inflammatory responses drive the pathogenesis of infection-induced preterm labor. Lipopolysaccharide (LPS) or other toxins elaborated by bacteria are recognized by pattern-recognition receptors such as toll-like receptors (Janssens, 2003). These receptors are present on mononuclear phagocytes, decidual cells, cervical epithelia, and trophoblasts (Chuang, 2000; Gonzalez, 2007; Holmlund, 2002). Loss of specific toll-like receptors results in delayed parturition in mouse models (Montalbano, 2013). Instead, activation of toll-like receptors induces a signaling cascade that activates production of chemokines such as interleukin 8 (IL-8) and cytokines such as IL-1β. Activation also recruits immune cells into the reproductive tract. Cytokines are produced by immune cells and by cells within the cervix, decidua, membranes, or fetus itself.

LPS-induced production of IL-1β in turn promotes a series of responses that include: (1) increased synthesis of others, that is, IL-6, IL-8, and tumor-necrosis factor alpha (TNF-α); (2) proliferation, activation, and migration of leukocytes; (3) modifications in extracellular matrix proteins; and (4) mitogenic and cytotoxic effects such as fever and acute-phase response (El-Bastawissi, 2000). Also, in many tissues, including myometrium, decidua, and amnion, IL-1β promotes prostaglandin formation that induces cervical ripening and loss of myometrial quiescence (Casey, 1990; Challis, 2002; Keelan, 2003). The importance of prostaglandins to infection-mediated preterm birth is supported by the observation that prostaglandin inhibitors can reduce the rate of LPS-induced preterm birth in both the mouse and nonhuman primate (Gravett, 2007; Timmons, 2014). Inhibition of cyclooxygenase-2 prevents inflammation-mediated preterm labor in the mouse. And, immunomodulators plus antibiotics delay preterm delivery after experimental intraamnionic infection in a nonhuman primate model.

Proteases such as matrix metalloproteinases (MMPs) are also induced by IL-1β and function to break down extracellular matrix components such as collagen or elastic fibers. This disrupts the structural integrity of fetal membranes and the cervix. Current evidence from animal and human studies suggests that many aspects of infection-mediated preterm birth differ from pathways that regulate term parturition (Hamilton, 2012; Holt, 2011; Shynlova, 2013a,b; Timmons, 2014).

Origin of Cytokines

Uterine cytokines are likely important for preterm labor. For example, it appears that cytokines produced in maternal decidua and myometrium will have effects on that side, whereas cytokines produced in the membranes or in cells within the amnionic fluid will not be transferred to maternal tissues. The transfer of cytokines such as IL-1β from decidua across the membranes into amnionic fluid, however, appears to be severely limited. Additionally, the human myometrium expresses chemokine receptors that decline during labor (Hua, 2013).

The requirement of leukocytes for initiation of term labor in women remains inconclusive. In general, resident and invading leukocytes produce the bulk of cytokines in cases of inflammation resulting from infection. Indeed, with infection, leukocytes—mainly neutrophils, macrophages, and T lymphocytes—infiltrate the cervix, lower uterine segment, fundus, and membranes at the time of labor. Invading leukocytes and certain parenchymal cells produce cytokines and appear to be the primary source of myometrial cytokines (Young, 2002). By contrast, in the decidua, both stromal cells and leukocytes likely contribute. In the cervix, glandular and surface epithelial cells appear to produce cytokines.

The presence of cytokines in amnionic fluid and their association with preterm labor is well documented. But, their exact cellular origin—with or without recoverable microorganisms—is not well defined. Amnionic fluid cytokines are most likely secreted by mononuclear phagocytes or neutrophils activated and recruited into the amnionic fluid. Thus, the amount of amnionic fluid IL-1β would be determined by the number of leukocytes recruited, their activational status, or the effect of amnionic fluid constituents on their IL-1β secretion rate.

Vaginal Microbiota

Factors that predispose to ascending infection and then preterm birth are a key focus of ongoing basic research. Mucosal immunity and barrier function of the cervicovaginal epithelia, the microbiota composition in the vaginal tract, and their interplay are a major topic (Smith, 2017). From animal studies, mucosal immunity of the lower reproductive tract can be disrupted by viral infection with a subsequent enhanced susceptibility to ascending bacterial infection (Racicot, 2013, 2017). Along with the ability of the cervicovaginal epithelia to respond to environmental insults, the composition of the microbe ecosystem in the vaginal tract may also determine susceptibility to ascending infection.

New technologies based on genomic analysis have shown that the nonpregnant vaginal tract hosts a complex microbial community (Gajer, 2012; White, 2011). Also described in Chapter 65 (Vaginitis), these community state types can differ widely among women who are all healthy. And, the vaginal microbiome changes during normal pregnancy (Aagaard, 2012; Stout, 2017). Namely, the diversity and richness of microbe populations are reduced during pregnancy and become more stable. Compared with nonpregnant controls, Lactobacillus species show an enhanced dominance. Some but not all studies report an increased population of certain microbes—for example, Gardnerella vaginalis and Ureaplasma urealyticum—in women with preterm birth (Donders, 2009; Nelson, 2014). However, differences in populations of pregnant women studied, in definitions of preterm birth, and in data analysis may complicate interpretation of these data.

Some specific microorganisms are detected more frequently than others in amnionic fluid of women with preterm labor (Gerber, 2003; Hillier, 1988; Yoon, 1998). These include G vaginalis, Fusobacterium species, Mycoplasma hominis, and U urealyticum. Their identification was interpreted by some as presumptive evidence that specific microorganisms are more commonly involved as pathogens in the induction of preterm labor. Another interpretation, however, is that given direct access to the membranes after cervical dilation, selected microorganisms, such as fusobacteria, are more capable of burrowing through these exposed tissues and will do so. Fusobacteria are found in the vaginal fluid of only 9 percent of women but in 28 percent of positive amnionic fluid cultures from pregnancies with preterm labor and intact membranes (Chaim, 1992).

Preterm Premature Rupture of Membranes

This term defines spontaneous rupture of the fetal membranes before 37 completed weeks and before labor onset (American College of Obstetricians and Gynecologists, 2016d). Such rupture likely has various causes, but intrauterine infection, oxidative stress-induced DNA damage, and premature cellular senescence are major predisposing events (Dutta, 2016; Gomez, 1997; Mercer, 2003). Associated risk factors include lower socioeconomic status, body mass index <19.8, nutritional deficiencies, and cigarette smoking. Women with PPROM carry an enhanced risk for recurrence during a subsequent pregnancy (Bloom, 2001). Despite these known risk factors, none is identified in most cases of preterm rupture.

Molecular Changes

Increased apoptosis or necroptosis of membrane cellular components and greater levels of specific proteases in membranes and amnionic fluid are related to PPROM. Most tensile strength of the membranes is provided by the amnionic extracellular matrix and interstitial amnionic collagens that are produced in mesenchymal cells (Casey, 1996). Thus, collagen degradation has been a focus of research. The MMP family is involved with normal tissue remodeling and particularly with collagen degradation. Some members are found in higher concentrations in amnionic fluid from pregnancies with PPROM (Maymon, 2000; Park, 2003; Romero, 2002). MMP activity is in part regulated by tissue inhibitors of matrix metalloproteinases—TIMPs. Several of these inhibitors are found in lower concentrations in amnionic fluid from women with ruptured membranes. Elevated MMP levels found at a time when protease inhibitor expression declines further supports that their expression alters amnionic tensile strength.

Studies of amniochorion explants show greater MMP expression following treatment with certain cytokines (Fortunato, 1999a,b, 2002). With membrane rupture, thrombin activity rises, which activates MMPs and prostaglandin synthesis. Studies by Mogami (2013) provide a mechanism by which bacterial endotoxin or TNF-α elicits release of fetal fibronectin (fFN) by amnion epithelial cells. The fFN then binds toll-like receptor 4 in the amnion mesenchymal cells to activate signaling cascades. These result in augmented prostaglandin E (PGE₂) synthesis and elevated activity of MMPs. Higher prostaglandin levels promote cervical ripening and uterine contractions. Greater MMP concentrations allow collagen breakdown in the fetal membranes resulting in premature rupture.

In pregnancies with PPROM, the amnion exhibits a higher degree of cell death and more apoptosis markers than that in term amnion (Arechavaleta-Velasco, 2002; Fortunato, 2003). In vitro studies indicate that apoptosis is likely regulated by bacterial endotoxin, IL-1β, and TNF-α. In addition, oxidative stress initiated by events other than infection can induce DNA damage, premature senescence, and subsequent inflammation and proteolysis that leads to PPROM (Menon, 2014a,b). Last, there are proteins involved in the synthesis of mature cross-linked collagen or matrix proteins that bind collagen and thereby promote tensile strength. These proteins are altered in membranes with premature rupture (Wang, 2006).

Infection

Several studies have investigated the incidence of infection-induced PPROM. Bacterial cultures of amnionic fluid support a role for infection in a significant proportion. One review of 18 studies and almost 1500 women with PPROM found that bacteria were isolated from amnionic fluid in a third of cases (Goncalves, 2002). Accordingly, some have given antimicrobial treatment to women in spontaneous preterm labor with intact membranes, however, results have been disappointing as discussed later in Antimicrobials (Kenyon, 2008b).

The inflammatory response that leads to membrane weakening and mediators of this process are currently areas of research. One goal is to identify early risk markers for PPROM.

Multifetal Pregnancy

Twins and higher-order multifetal births account for approximately 3 percent of neonates born in the United States (Martin, 2017). Preterm delivery continues to be the major cause of the excessive perinatal morbidity and mortality with multifetal pregnancies. The effects of uterine stretch discussed in Causes of Preterm Birth are obvious. Many of these interrelationships are discussed in Chapter 45 (Preterm Birth).

Pregnancy Factors

Several genetic and environmental factors affect the frequency of preterm labor. Of these, threatened abortion in early pregnancy is associated with higher rates of later adverse outcomes. Weiss (2004) reported outcomes with vaginal bleeding at 6 to 13 weeks’ gestation in nearly 14,000 women. Both light and heavy bleeding were associated with subsequent preterm labor, placental abruption, and pregnancy loss before 24 weeks.

Birth defects in the fetus may also predispose to preterm birth. In a secondary analysis of data from the First- and Second-Trimester Evaluation of Risk (FASTER) Trial, birth defects were associated with preterm birth and low birthweight neonates (Dolan, 2007).

Lifestyle Factors

Cigarette smoking, inadequate maternal weight gain, and illicit drug use affect the incidence and outcome of low-birthweight neonates (Chap. 44, Accelerated Lung Maturation). Extremes of maternal weight—both underweight and obese mothers—have an enhanced risk of preterm birth (Cnattingius, 2013; Girsen, 2016). Other maternal factors implicated include young or advanced maternal age, poverty, short stature, and vitamin C deficiency (Casanueva, 2005; Gielchinsky, 2002; Kramer, 1995; Leveno, 2009; Meis, 1995).

As discussed in Causes of Preterm Birth, psychological factors such as depression, anxiety, and chronic stress are associated with preterm birth (Hoffman, 2016; Venkatesh, 2016). In one review of more than 50 studies, Donovan and coworkers (2016) found a significant link between low birthweight and preterm birth in women injured by physical abuse (Chap. 47, Trauma).

Studies of work and physical activity related to preterm birth have yielded conflicting results (Goldenberg, 2008). Some evidence suggests that working long hours and hard physical labor are probably linked to a higher risk of preterm birth (Luke, 1995). However, aerobic exercise in normal-weight women with uncomplicated singleton pregnancies appears to be safe and not associated with preterm birth (American College of Obstetricians and Gynecologists, 2017d; Di Mascio, 2016). One metaanalysis of physical activity found that leisure-time physical activity was associated with a reduced risk of preterm birth (Aune, 2017).

Genetic Factors

The recurrent, familial, and racial nature of preterm birth suggests that genetics may play a causal role. Accumulating literature on genetic variants buttresses this concept (Gibson, 2007; Hampton, 2006; Macones, 2004; Velez, 2009). Several such studies have also implicated immunoregulatory genes in potentiating chorioamnionitis in cases of preterm delivery due to infection (Varner, 2005).

Periodontal Disease

Gum inflammation is a chronic anaerobic inflammation that affects as many as 50 percent of pregnant women in the United States (Goepfert, 2004). Vergnes and Sixou (2007) performed a metaanalysis of 17 studies and concluded that periodontal disease was significantly associated with preterm birth. To better study this relationship, Michalowicz (2006) randomly assigned 813 pregnant women between 13 and 17 weeks’ gestation who had periodontal disease to treatment during pregnancy or postpartum. Treatment during pregnancy improved periodontal disease and was safe. However, treatment failed to significantly alter preterm birth rates. This position was reaffirmed by a joint workshop of the European Federation of Periodontology and the American Academy of Periodontology (Sanz, 2013).

Interval between Pregnancies

The intervals between pregnancies are linked with adverse perinatal outcomes. In a metaanalysis, intervals <18 months and >59 months were associated with greater risks for both preterm birth and small-for-gestational age newborns (Conde-Agudelo, 2006). The causal effect of short interpregnancy intervals, however, has been questioned (Ball, 2014).

Prior Preterm Birth

The most important risk factor for preterm labor is a prior preterm delivery. Data from nearly 16,000 women delivered at Parkland Hospital are instructive. Namely, the recurrent preterm delivery risk for women with a preterm first delivery was threefold greater than that of women whose first neonate was born at term. More than a third of women whose first two newborns were preterm subsequently delivered a third preterm newborn. Most—70 percent—of the recurrent births in this study occurred within 2 weeks of the gestational age of the prior preterm delivery. The causes of prior preterm delivery also recurred. Although women with prior preterm births are clearly at risk for recurrence, they represented only 10 percent of the total preterm births in this study. Expressed another way, 90 percent of the preterm births at Parkland Hospital could not be predicted based on a history of preterm birth. Laughon and associates (2014) confirmed the importance of prior spontaneous preterm birth. Notably, these investigators also found that prior indicated preterm birth was strongly associated with subsequent spontaneous preterm birth. Variations in the definition of spontaneous and indicated used may explain this association.

Ultimately, risk of recurrent preterm birth is influenced by three factors: the frequency of prior preterm deliveries, severity as measured by gestational age, and the order in which the prior preterm delivery occurred (McManemy, 2007). That is, an individual woman’s risk for recurrent preterm birth is influenced by her past number and sequence of preterm and term births. For example, a risk of recurrent preterm birth for a gravida 3 para 2 woman with a prior preterm birth followed by a term birth differs from a woman with a prior term birth followed by preterm birth. Thus, the influence of reproductive history has a profound prognostic significance for risk of recurrence. Moreover, this may also influence the supposed benefit attributed to various interventions described later.

Infection

Antibiotic Prophylaxis

As discussed in Cervical Dysfunction, a link between some cases of preterm birth and infection seems irrefutable (Goldenberg, 2008). In several studies, antimicrobial treatment has been given to prevent preterm labor thought to be due to microbial invasion. These strategies especially targeted Mycoplasma species. Morency and colleagues (2007) performed a metaanalysis of 61 articles and suggested that antimicrobials given in the second trimester may prevent subsequent preterm birth. Andrews and associates (2006) reported results of a randomized trial in which they gave a course of azithromycin plus metronidazole every 4 months to 241 nonpregnant women whose last pregnancy resulted in spontaneous delivery before 34 weeks. Approximately 80 percent of the women with subsequent pregnancies had received study drug within 6 months of their subsequent conception. Such interconceptional antimicrobial treatment did not reduce the rate of recurrent preterm birth. Using a subgroup analysis of these same women, Tita and coworkers (2007) concluded that such use of antimicrobials may be harmful. In another randomized study, 2661 women received placebo or metronidazole plus erythromycin between 20 and 24 weeks’ gestation followed by ampicillin plus metronidazole during labor (Goldenberg, 2006). This antimicrobial regimen did not reduce the rate of preterm birth or that of histological chorioamnionitis. At this time, antibiotic prophylaxis to prevent preterm birth is not recommended in women with preterm labor and intact membranes (Flenady, 2013).

Bacterial Vaginosis

In this condition, normal, hydrogen peroxide-producing, lactobacillus-predominant vaginal flora is replaced with anaerobes (Hillier, 1995; Nugent, 1991). Its diagnosis and management are discussed in Chapter 65 (Vaginitis). Using Gram staining, relative concentrations of the bacterial morphotypes characteristic of bacterial vaginosis are determined and graded by the Nugent score or assessed clinically with Amsel criteria.

Bacterial vaginosis has been associated with spontaneous abortion, preterm labor, PPROM, chorioamnionitis, and amnionic fluid infection (Hillier, 1995; Kurki, 1992; Leitich, 2003a,b). Environmental factors appear to be important in bacterial vaginosis development. Exposure to chronic stress, ethnic differences, and frequent or recent douching are associated with higher rates of the condition (Culhane, 2002; Ness, 2002). A gene-environment interaction has also been described (Macones, 2004). Women with bacterial vaginosis and a susceptible TNF-α genotype had a ninefold increased incidence of preterm birth.

From all of these studies, adverse vaginal flora seems associated with spontaneous preterm birth. Unfortunately, to date, screening and treatment have not prevented preterm birth. These are discussed further in Chapter 65 (Trichomoniasis). Indeed, microbial resistance or antimicrobial-induced change in the vaginal flora results from regimens intended to eliminate bacterial vaginosis (Beigi, 2004; Carey, 2005).

Symptoms

Early differentiation between true and false labor is difficult—especially before demonstrable cervical effacement and dilation. Uterine activity alone can be misleading because of Braxton Hicks contractions. These irregular, nonrhythmical contractions can cause considerable confusion in the diagnosis of true labor. Not infrequently, women who deliver before term have uterine activity that is attributed to Braxton Hicks contractions that prompt an incorrect diagnosis of false labor. Accordingly, the American College of Obstetricians and Gynecologists (2016b) defines preterm labor to be regular contractions before 37 weeks that are associated with cervical change.

In addition to contractions, pelvic pressure, menstrual-like cramps, watery vaginal discharge, and lower back pain have been empirically associated with impending preterm birth. Such complaints are thought by some to be common in normal pregnancy and are therefore often minimized by patients and obstetrical care providers.

The importance of these symptoms as a harbinger of labor has been emphasized by some but not all investigators (Iams, 1990; Kragt, 1990). Iams and coworkers (1994) found that the signs and symptoms signaling preterm labor, including uterine contractions, appeared only within 24 hours of preterm labor.

Chao (2011) prospectively studied 843 women with a singleton fetus who presented to Parkland Hospital with preterm labor symptoms between 24^0/7 and 33^6/7 weeks, intact membranes, and cervical dilation <2 cm. Those whose cervix remained <2 cm were sent home with a diagnosis of false preterm labor. When analyzed against the general obstetrical population, women sent home had a similar rate of birth before 34 weeks—2 versus 1 percent. However, these women did have significantly higher rates of birth between 34 and 36 weeks—5 percent compared with 2 percent. Women with cervical dilation of 1 cm at discharge were significantly more likely to deliver before 34 weeks compared with women without cervical dilation—5 percent versus 1 percent. Importantly, almost 90 percent of the 1-cm group delivered within 21 days of the initial presentation.

Cervical Change

Researchers have evaluated asymptomatic cervical changes that may predict preterm labor. Asymptomatic cervical dilation after midpregnancy is suspected to be a risk factor for preterm delivery, although some clinicians consider it to be a normal anatomical variant. Moreover, study results have suggested that parity alone is not sufficient to explain cervical dilation discovered early in the third trimester.

Cook (1996) longitudinally evaluated cervical status with transvaginal sonography between 18 and 30 weeks in nulliparas and multiparas who all subsequently gave birth at term. Cervical length and diameter were identical in both groups throughout these critical weeks. In a study from Parkland Hospital, routine digital cervical examinations were performed between 26 and 30 weeks in 185 asymptomatic women. Approximately 25 percent of women whose cervix was dilated 2 or 3 cm delivered before 34 weeks (Leveno, 1986a). Other investigators have verified cervical dilation as a predictor of increased preterm delivery risk (Copper, 1995).

Although women with dilation and effacement in the third trimester are at greater risk for preterm birth, detection does not necessarily improve pregnancy outcome. Buekens and associates (1994) randomly assigned 2719 women to undergo routine cervical examinations at each prenatal visit and compared them with 2721 women in whom serial examinations were not performed. Knowledge of antenatal cervical dilation did not affect any pregnancy outcome related to preterm birth or the frequency of interventions for preterm labor. The investigators also reported that cervical examinations were not related to PPROM. Thus, it seems that prenatal cervical examinations in asymptomatic women are neither beneficial nor harmful.

Ambulatory Uterine Monitoring

An external tocodynamometer belted around the abdomen and connected to an electronic waist recorder allows a woman to ambulate while uterine activity is recorded. Results are transmitted via telephone daily. Women are educated concerning signs and symptoms of preterm labor, and clinicians are kept apprised of their progress. The 1985 approval of this monitor by the Food and Drug Administration (FDA) prompted its widespread clinical use. Subsequently, it was proven that the use of this expensive and time-consuming system does not reduce preterm birth rates (Collaborative Home Uterine Monitoring Study Group, 1995; Iams, 2002; Urquhart, 2017). Despite improvements in technology with the internet and cellular telephones, use of such monitoring is discouraged (American College of Obstetricians and Gynecologists, 2016c).

Fetal Fibronectin

This glycoprotein is produced in 20 different molecular forms by various cell types, including hepatocytes, fibroblasts, endothelial cells, and fetal amnion cells. Present in high concentrations in maternal blood and amnionic fluid, fFN is thought to function in intercellular adhesion during implantation and in maintenance of placental adherence to uterine decidua (Leeson, 1996). Detected in cervicovaginal secretions in women who have normal pregnancies with intact membranes at term, fFN appears to reflect stromal remodeling of the cervix before labor.

Lockwood (1991) reported that fFN detection in cervicovaginal secretions before membrane rupture was a possible marker for impending preterm labor. Qualitative and quantitative fFN levels are measured using enzyme-linked immunosorbent assays, and values exceeding 50 ng/mL are considered positive. Sample contamination by amnionic fluid and maternal blood should be avoided. Interventional studies based on the use of fFN screening in asymptomatic women have not demonstrated improved perinatal outcomes (Andrews, 2003; Esplin, 2017; Grobman, 2004). The American College of Obstetricians and Gynecologists (2016c) does not recommend screening with fFN tests. Its use in conjunction with cervical length measurement is discussed next.

Cervical Length Measurement

Progressively shorter cervical canals assessed sonographically are associated with increased rates of preterm birth (Iams, 1996). The technique to measure cervical length with sonography is described in Chapter 10 (Cervical Length Assessment). When performed by trained operators, cervical length analysis using transvaginal sonography is safe, highly reproducible, and more sensitive than transabdominal sonographic screening (American College of Obstetricians and Gynecologists, 2016c). The Society for Maternal-Fetal Medicine (2016b) has provided guidance for the performance of proper cervical length measurement. The Society recommends that sonographers and practitioners obtain specific training in the acquisition and interpretation of cervical length imaging through accreditation programs.

Transvaginal cervical sonography is not affected by maternal obesity, cervix position, or shadowing from the fetal presenting part. Because of the inability to easily distinguish the lower uterine segment from the cervix in early gestation, transvaginal cervical length assessment is typically performed after 16 weeks’ gestation. Such interrogation is currently limited to singleton gestations and not recommended for multifetal gestations outside of research trials (American College of Obstetricians and Gynecologists, 2016c).

Indications for cervical length measurement are somewhat controversial. For those women with a history of prior spontaneous preterm birth, the Society for Maternal-Fetal Medicine (2016b) recommends transvaginal cervical length screening. But, the American College of Obstetricians and Gynecologists (2016c) only recommends consideration of screening for this indication. In women with singleton pregnancies but without a history of prior preterm birth, the Society for Maternal-Fetal Medicine (2016b) views cervical length screening as reasonable yet acknowledges that this remains an area of debate.

A first concern with screening surrounds the efficacy of interventions to improve perinatal outcomes once cervical length screening has isolated at-risk gravidas. Of interventions, cervical cerclage and administration of vaginal progesterone have both been evaluated. Vaginal progesterone for this indication is discussed in Progesterone Use without Prior Preterm Birth. For prophylactic cerclage in women with prior preterm births and shortened cervices, many studies have failed to show superior primary outcomes. But, piecemeal subgroup analyses from the trials have been subsequently used as the basis for recommendations in current practice guidelines regarding cervical length assessment and consideration for cerclage placement. A second concern is the accuracy and utility of the screening test, especially in low-risk women, who represent most of the population with preterm birth (Interval between Pregnancies).

To address this issue in low-risk women, Esplin and colleagues (2017) prospectively studied 9410 nulliparas with singleton pregnancies. Universal screening of sonographically measured cervical length and quantitative measurement of vaginal fFN levels were evaluated as predictors of women who would spontaneously deliver before 37 weeks. These measures had poor predictive performance as a screening test. In fact, all screening modalities had relative low sensitivity and low positive-predictive values. Based on these findings, routine use of these screening tests in this low-risk population is not recommended. Bloom and Leveno (2017) subsequently critiqued the use of transvaginal cervical length screening in low-risk women and the promulgation of consensus guidelines. As described in Chapter 1 (Repeal and Replace), they highlighted the staggering costs encumbering the health-care system in the United States as a result of such strategies.

Cervical Cerclage

Prevention of preterm birth remains an elusive goal. Still, recent reports suggest that prevention in selected populations may be achievable.

Of options, cerclage placement may be used to prevent preterm birth in at least three circumstances. First, the procedure may benefit women who have a history of recurrent midtrimester losses and who are diagnosed with cervical insufficiency. A second instance is the woman identified during sonographic examination to have a short cervix. The third indication is a “rescue” cerclage, done emergently when cervical incompetence is recognized in women with threatened preterm labor.

As is the case for virtually all obstetrical conditions, an accurate history is critical for management decisions. For recurrent abortion from cervical incompetence, historical clues are outlined in Chapter 18 (Management). For women with a short cervix incidentally detected by sonography, the benefit of cerclage placement appears directly related to whether the woman has a history of prior preterm birth. In those without a prior preterm birth, cerclage for a sonographically detected short cervix alone offers no advantage. To and associates (2004) screened 47,123 women and randomly assigned the 253 women with cervices <15 mm, with or without a history of preterm birth, to cerclage or no cerclage. The frequency of preterm delivery before 33 weeks did not differ significantly. In contrast, women with a sonographically diagnosed short cervix and a history of preterm birth may benefit. Owen and colleagues (2009) randomly assigned 302 women with prior preterm birth and with a short cervix—defined as length <25 mm—to cerclage or no procedure. The primary study outcome was not supported by the intervention. However, women with a cervical length <15 mm delivered before 35 weeks significantly less often following cerclage compared with women with no cerclage—30 versus 65 percent. This study suggested that recurrent preterm birth could be prevented in a subset of women with asymptomatic singleton gestations with both previous preterm birth and short cervical length.

These findings prompted a reassessment by Berghella and coworkers (2011), who performed a metaanalysis using individual patient data (Fig. 42-7). The primary outcomes from the included trials did not support cerclage placement. However, these investigators concluded that cerclage significantly prevented preterm birth and improved composite perinatal mortality and morbidity in women with prior spontaneous preterm birth, singleton gestation, and cervical length <25 mm.

One caveat in the interpretation of this cerclage data is the influence of obstetrical history. For example, all of the trials comprising the metaanalysis included preterm birth as early as 16 to 17 weeks’ gestation. Defining these early second trimester losses as preterm births, rather than cervical incompetence, is problematic. Thus, it is difficult to distinguish whether these women were treated in the context of cervical incompetence or of preterm labor at 16 weeks. Nonetheless, based on these findings, the American College of Obstetricians and Gynecologists (2016c) concluded that in women with a singleton pregnancy, prior spontaneous preterm birth before 34 weeks, cervical length <25 mm, and gestational age <24 weeks, cerclage placement may be considered.

Prophylaxis with Progestogen Compounds

In most mammals, progesterone withdrawal is considered to be a parturition-triggering event. During human parturition, however, maternal, fetal, and amnionic fluid progesterone levels remain elevated. It has been proposed that human parturition involves functional progesterone withdrawal mediated by decreased activity of progesterone receptors (Chap. 21, Sex Steroid Hormone Role). It follows conceptually that the administration of progesterone may block preterm labor. This hypothesis has stimulated several studies of both 17-alpha hydroxyprogesterone caproate (17-OHP-C) and vaginally administered progesterone in women with varying risks for preterm birth.

At present, the reported benefits of either of these progestogen therapies are limited to women with singleton pregnancies. Progesterone prophylaxis specifically in multifetal gestations has not lowered preterm birth rates (Caritis, 2009; Rouse, 2007). Accordingly, both the American College of Obstetricians and Gynecologists (2016c) and the Society for Maternal-Fetal Medicine (2017a) approve the use of progestogen therapy for prevention of preterm birth in select women with singleton pregnancies. Criteria are a history of prior preterm birth or no prior preterm birth but a sonographically identified short cervix.

Prior Preterm Birth and Progestogen Compounds

17-OHP-C is a synthetic progestogen, and the first and only drug approved by the FDA for prevention of recurrent preterm birth. The approval in 2011 was supported by a study by the Maternal-Fetal Medicine Units (MFMU) Network (Meis, 2003). In this trial, 463 women with a prior preterm birth were randomly assigned to receive weekly intramuscular injections of inert oil or 17-OHP-C from 16 through 36 weeks’ gestation. They reported a significantly reduced recurrence of preterm birth in 36 percent of women receiving 17-OHP-C compared with 55 percent of those given placebo.

This MFMU study has been challenged because of the unexpectedly high preterm delivery rate in the placebo arm (Romero, 2013). One explanation for this high rate in the placebo group was asymmetry in the risks of recurrence. Indeed, 41 percent of the control group had ≥2 prior preterm births compared with only 28 percent in the 17-OHP-C group. Another concern was that the injection dosage of 17-OHP-C, which was 250 mg weekly, was empirically chosen (Caritis, 2014). Only later reports described the pharmacokinetics of 17-OHP-C (Caritis, 2012). Nonetheless, the Society for Maternal-Fetal Medicine (2017a) recently reaffirmed the use of 17-OHP-C, rather than vaginal progesterone, for prevention of recurrent preterm birth.

Metabolism

Sharma and associates (2008) reported that the metabolism of 17-OHP-C was predominantly mediated by the CYP3A enzymatic system. Thus, other agents that induce or inhibit this enzymatic system as well as hepatic impairment may alter drug levels. They also showed that 17-OHP-C is not converted after administration to the primary progesterone metabolite, 17α-hydroxyprogesterone. The relative binding affinity of 17-OHP-C to progesterone receptors approximates only 30 percent of that by progesterone (Attardi, 2007). Because synthetic 17-OHP-C is not converted to a naturally occurring progestogen and is not superior to progesterone in eliciting a hormonal response via the classic steroid-receptor mediated pathway, alternative pathways are now being considered to explain its efficacy (Manuck, 2011).

Caritis and colleagues (2012) examined 61 women receiving 17-OHP-C therapy and found that the half-life was relatively long (median 16.2 days). Pharmacokinetic parameters were affected by maternal body habitus and varied widely between subjects. In addition, 17-OHP-C crossed the placental barrier and was detectible in cord plasma 44 days after the last maternal injection (Caritis, 2012). Despite this, evidence to date suggests that 17-OHP-C is safe for the fetus. No abnormalities, including abnormal genitalia, were found in a 48-month follow-up study of infants exposed in the 2003 MFMU Network trial (Northen, 2007).

Pricing Concerns

There are special concerns involving 17-OHP-C and subsequent price-gouging claims (Cohen, 2011; Romero, 2013). In 2011, the FDA gave temporary approval to KV Pharmaceutical to market 17-OHP-C under the brand name Makena. Because regulations prohibited compounding, there was no competitor for this relatively inexpensive drug, and Makena was priced at $1500 per injection. This caused widespread concern because the cumulative cost of Makena would be more than $30,000 per pregnancy.

Use of 17-OHP-C at Parkland Hospital

A program for implementation of 17-OHP-C was incorporated at Parkland Hospital in 2012. Given the concerns above, a local compounding pharmacy provided 250-mg, single-dose vials of 17-OHP-C in sesame oil at a cost of $25 per dose. Nelson and colleagues (2017) recently reported their findings from this program in a prospective study of 430 women given compounded 17-OHP-C. Use of 17-OHP-C was ineffective for prevention of recurrent preterm birth at 35 weeks or less compared with a historical cohort from Parkland. As shown in Table 42-6, 17-OHP-C did not significantly reduce the rates of recurrent preterm birth regardless of prior preterm birth number or sequence. Moreover, plasma concentrations of 17-OHP-C were not different at 24 weeks or 32 weeks between women delivered ≤35 weeks and those delivered later. Interestingly, levels were consistent with those previously reported using castor oil as a vehicle (Caritis, 2014). Last, the gestational age interval at which preterm birth recurred did not differ after use of 17-OHP-C. A side effect of 17-OHP-C was a significant increase in the rate of gestational diabetes. Taken together, 17-OHP-C use was ineffective to prevent recurrent preterm birth and was associated with a significant side effect.

From the foregoing studies, evidence to support the use of 17-OHP-C to prevent recurrent preterm birth is problematic (Young, 2017). The mechanism of action remains unknown, and the pharmacological properties have yet to be established. Evidence of clinical effectiveness has yet to be replicated. A condition of the FDA approval for 17-OHP-C was that a confirmatory multicenter, double-blind randomized controlled trial be conducted with a preferred primary end point of delivery <35 weeks. This international trial—PROLONG—is currently underway with scheduled completion in 2018 and with an estimated enrollment of 1707 participants (PROLONG, 2014).

Progesterone Use without Prior Preterm Birth

Three randomized trials are at the center of whether progestogen therapy should be used in women without prior preterm births. These trials, shown in Table 42-7, hinge on sonographically determined cervical length. In the first trial, Fonseca and colleagues (2007) randomly assigned 250 women with short cervices measuring ≤15 mm identified during routine prenatal care. Women were given nightly 200-mg micronized progesterone vaginal capsules or placebo from 24 to 34 weeks’ gestation. Spontaneous delivery <34 weeks was significantly reduced by progesterone therapy. Importantly, this trial included not only nulliparas but also those with twins or prior preterm birth.

In the second trial, Hassan and coworkers (2011) randomly assigned 465 women with a short cervix—10 to 20 mm—to vaginal progesterone gel, 90 mg daily, or placebo. This trial also included nulliparas and women with prior preterm births.

From these studies, the FDA rejected progesterone gel for use because the results did not meet the level of statistical significance required to show efficacy in the subjects recruited in the United States. According to Likis and colleagues (2012), the heterogeneity of these first two studies that included women with varied indications for progestogen treatment, combined with the fact that outcomes were not reported by risk factors such as nulliparity, made it impossible to interpret the efficacy of progesterone for specific indications.

The third study randomly assigned administration of 17-OHP-C intramuscular injection or placebo between 16 and 22^3/7 weeks to nulliparas with a singleton gestation and a cervical length <30 mm detected sonographically (Table 42-8) (Grobman, 2012). Treatment with 17-OHP-C given weekly did not reduce the frequency of preterm birth before 37 weeks. Regardless of cervical length, 17-OHP-C was ineffective.

Thus, vaginal progesterone, but not 17-OHP-C, appears to benefit women with a sonographically measured short cervix. Romero and Stanczyk (2013) provided a review to explain the conflicting evidence and argued that naturally occurring progesterone, which is used in the vaginal preparations, is not the same as synthetic 17-OHP-C. Likewise, Furcron and coworkers (2015) found that 17-OHP-C did not have local antiinflammatory effects at the maternal-fetal interface or cervix. Further, 17-OHP-C did not protect against endotoxin-induced preterm birth.

From all these studies, the American College of Obstetricians and Gynecologists (2016c) concluded that universal cervical length screening in women without a prior preterm birth is not mandatory. However, this screening strategy could be considered in the context of treatment with vaginal progesterone.

The OPPTIMUM Study

This study of 1228 high-risk women with singleton pregnancies is the largest to date for vaginal progesterone prophylaxis (Norman, 2016). This randomized trial of vaginal progesterone, 200 mg daily from 22–24 weeks to 34 weeks of gestation, was termed the OPPTIMUM study—dOes Progesterone Prophylaxis To prevent preterm labor IMprove oUtcoMe? High-risk women were defined as those with a prior spontaneous birth ≤34 weeks or with a cervical length ≤25 mm or a positive fFN test result combined with other clinical risk factors for preterm birth.

The primary outcomes of OPPTIMUM were unique in that both immediate obstetrical and childhood outcomes were examined. These were fetal death or birth <34 weeks; a composite of death, brain injury, or bronchopulmonary dysplasia; and a standardized cognitive score at 2 years of age. Contrary to earlier reports, vaginal progesterone was not associated with a lower risk of preterm birth or composite neonatal adverse outcomes. In children at 2 years of age, vaginal progesterone also had no long-term benefit or harm.

Thus, evidence is conflicting as to the efficacy of progestogens across the spectrum of the various specific indications. Some have attempted to resolve these issues through systematic review and metaanalysis (Prior, 2017; Romero, 2016, 2017). As outlined in this entire section, virtually all evidence supporting use of progestogens for a specific indication can be challenged in some way. We agree with the conclusions from the OPPTIMUM trial (Norman, 2016) that the results of recent studies should prompt a major review of progesterone use for preterm birth prophylaxis, a search to identify specific women who might specifically benefit, and a redoubling of efforts to find alternative strategies to prevent preterm birth in women at risk.

Geographic-Based Public Health-Care Programs

A well-organized prenatal system lowers the preterm birth rate in high-risk indigent populations (Creasy, 1980). One example is the Parkland Hospital prenatal clinic system (Leveno, 2009). As shown in Figure 42-8, the declining preterm birth rate between 1988 and 2006 coincided with a substantial rise in prenatal care attendance. In the early 1990s, a concerted effort was made to improve access to prenatal care by creating seamless care that began with antenatal enrollment and extended through delivery and the puerperium. Prenatal clinics were placed strategically throughout Dallas County to provide convenient access for our patients. Prenatal protocols are used by nurse practitioners at all clinic sites to guarantee homogeneous care. Women with high-risk pregnancy complications are referred to our hospital-based central clinic system. Here, Maternal-Fetal Medicine clinics operate each weekday and are staffed by residents and midwives and supervised by fellows and faculty.

Thus, prenatal care is considered one component of a comprehensive and orchestrated public health-care system that is community-based. We believe that the drop in preterm births experienced at our inner-city hospital is at least partially attributable to a geographically based public health-care program specifically targeting minority populations of pregnant women. A similar obstetrical care system for indigent women at the University of Alabama at Birmingham has also produced salutary results (Tita, 2011).

Methods used to diagnose ruptured membranes are detailed in Chapter 22 (Identification of Labor). A history of vaginal leakage of fluid, either as a continuous stream or a gush, should prompt a speculum examination to visualize gross vaginal pooling of amnionic fluid, clear fluid from the cervical canal, or both. Confirmation of PPROM is usually accompanied by sonographic examination to assess amnionic fluid volume, to identify the presenting part, and if not previously determined, to estimate gestational age. Once PPROM is identified, the general scheme shown in Table 42-9 can guide management.

Natural History

Cox and associates (1988) described pregnancy outcomes of 298 consecutive women who gave birth following spontaneously ruptured membranes between 24 and 34 weeks’ gestation at Parkland Hospital. Although this complication was identified in only 1.7 percent of pregnancies, it contributed to 20 percent of all perinatal deaths. By the time they presented, 76 percent of the women were already in labor, and 5 percent were delivered for other complications. Thus, only 19 percent initially were permitted expectant management. Ultimately, delivery was delayed 48 hours or more after membrane rupture in only 7 percent of the total study cohort. There was benefit noted from delayed delivery, however, as none of the neonates died in this group. This contrasted with a neonatal death rate of 80 per 1000 in preterm newborns delivered within 48 hours of membrane rupture. Nelson and colleagues later (1994) reported similar results.

The time from PPROM to delivery is inversely proportional to the gestational age when rupture occurs (Carroll, 1995). As shown in Figure 42-9, very few days are gained when membranes ruptured during the third trimester compared with midpregnancy.

Hospitalization

Most clinicians hospitalize women with ruptured membranes. Concerns regarding the costs of lengthy hospitalizations are usually moot, because most women enter labor within a week or less after membrane rupture. Carlan and coworkers (1993) randomly assigned 67 women with ruptured membranes to home or hospital management. No benefits were found for hospitalization, and maternal hospital stays were reduced by 50 percent in those sent home—14 versus 7 days. Importantly, the investigators emphasized that this study was too small to conclude that home management was safe in regard to umbilical cord prolapse.

Intentional Delivery

Before the mid-1970s, labor was usually induced in women with preterm ruptured membranes because of fear of sepsis. Maternal infection risk and fetal prematurity risk vary according to the gestational age at membrane rupture, and management decisions hinge on this balance. With regard to periviable pregnancy, Morales (1993b) expectantly managed 94 singleton pregnancies with ruptured membranes before 25 weeks. The average time gained was 11 days. Although 41 percent of infants survived to age 1 year, only 27 percent of the original cohort were neurologically normal. Similar results were reported by Farooqi (1998) and Winn (2000) and their colleagues. Management of these early pregnancies is discussed in Late-Preterm Birth.

For PPROM in general, two randomized trials in the 1990s compared labor induction with expectant management (Cox, 1995; Mercer, 1993). In both of these studies, the balance of risk and benefit was difficult to ascertain, as neither immediate delivery nor expectant management were proven to be superior for neonatal outcomes. Lieman and associates (2005) found that neonatal outcomes did not improve with expectant management beyond 33 weeks. McElrath and coworkers (2003) found that prolonged latency after membrane rupture was not associated with a greater incidence of fetal neurological damage. An important correlate is that infection—specifically chorioamnionitis—is recognized as a risk factor for development of neonatal neurological injury (Gaudet, 2001; Wu, 2000).

Bond and colleagues (2017) recently compared planned early birth with expectant management for women with PPROM before 37 weeks’ gestation. They evaluated 12 randomized trials totaling 3617 women and 3628 newborns. No clinically important differences in the incidence of neonatal sepsis between women who immediately delivered and those managed expectantly were identified. Although the incidence of chorioamnionitis was lower, neonates of women randomized to early birth were more likely to be born at an earlier gestational age and encountered attendant perinatal sequelae. The authors concluded that in women with rupture of membranes before 37 weeks’ gestation without contraindications to continuing the pregnancy, a policy of expectant management with careful monitoring was associated with better outcomes for both the mother and newborn.

The American College of Obstetricians and Gynecologists (2016d) has recognized the controversies of immediate delivery compared with expectant management. Clearly, gestational age is an important consideration. At 24^0/7 to 33^6/7 weeks, expectant management in the absence of nonreassuring fetal status, clinical chorioamnionitis, or placental abruption is recommended. At 34^0/7 weeks of gestation or greater, delivery is still recommended by the College for all women with ruptured membranes. Our current practices at Parkland Hospital are consistent with these recommendations.

Considerations with Expectant Management

Several scenarios during expectant management merit consideration. One is performance of digital cervical examination. Alexander and colleagues (2000) analyzed findings in women with PPROM expectantly managed between 24 and 32 weeks’ gestation. They compared those who had one or two digital cervical examinations with women who were not examined. Those who were examined had a rupture-to-delivery interval of 3 days compared with 5 days in those not examined. This difference did not worsen maternal or neonatal outcomes.

Rupture of membranes following second-trimester amniocentesis is uncommon (Chap. 14, Chorionic Villus Sampling). Compared with women with spontaneous rupture during the second trimester, Borgida and associates (2000) found that pregnancies complicated by PPROM after genetic amniocentesis resulted in significantly better perinatal outcomes. The perinatal survival rate was 91 percent. After counseling, affected women are typically managed expectantly as outpatients with serial surveillance of amnionic fluid volume (American College of Obstetricians and Gynecologists, 2016d). In the series cited above, the mean time to documentation of a normal amnionic fluid volume after amniocentesis approximated 2 weeks.

Tocolysis has been used in few studies. In women with ruptured membranes and lack of labor, prophylactic tocolysis does not improve neonatal outcomes but is associated with greater rates of chorioamnionitis (Mackeen, 2014). Similarly, therapeutic tocolysis—for those with ruptured membranes and labor—has also not provided significant perinatal benefit (Garite, 1987).

There is uncertainty regarding PPROM in the woman who has undergone cervical cerclage. McElrath and associates (2002) studied 114 women with a cerclage in place who later had ruptured membranes before 34 weeks. They were compared with 288 controls. Pregnancy outcomes were equivalent in both groups. Cerclage retention for more than 24 hours after preterm rupture of membranes may be associated with pregnancy prolongation, however, there is risk of intrauterine infection and its consequences (Giraldo-Isaza, 2011; Laskin, 2012). As discussed in Chapter 18 (Induced Abortion), such management is controversial.

With PPROM in general, the volume of amnionic fluid remaining after rupture appears to have prognostic importance in pregnancies before 26 weeks. Hadi and colleagues (1994) described 178 pregnancies with ruptured membranes between 20 and 25 weeks. Almost 40 percent developed oligohydramnios, defined by the absence of fluid pockets measuring 2 cm or more. Virtually all women with oligohydramnios delivered before 25 weeks, whereas 85 percent with adequate amnionic fluid volume were delivered in the third trimester. Carroll and coworkers (1995) observed no cases of pulmonary hypoplasia in fetuses born after membrane rupture at 24 weeks or beyond. This suggests that 23 weeks or less is the threshold for development of lung hypoplasia (Chap. 7, Respiratory System). Further, when contemplating early expectant management, consideration is also given to oligohydramnios and resultant limb compression deformities (Chap. 11, Management).

Other risk factors have also been evaluated. First, in neonates born to women with active herpetic lesions who were expectantly managed, the infectious morbidity risk appeared to be outweighed by risks associated with preterm delivery (Major, 2003). Second, Lewis and associates (2007) found that expectant management of women with PPROM and noncephalic presentation was associated with a higher rate of umbilical cord prolapse, especially before 26 weeks.

Clinical Chorioamnionitis

As discussed, infection is a major concern with membrane rupture. While some cases remain subclinical, if chorioamnionitis is diagnosed, prompt efforts to effect delivery, preferably vaginally, are initiated. Because maternal leukocytosis alone is not a consistent finding, fever is the only reliable indicator for the diagnosis of chorioamnionitis. Institutional practices and protocols vary in defining the temperature threshold. Traditionally, a temperature ≥38°C (100.4°F) accompanying ruptured membranes has implied infection. At Parkland Hospital, we still adhere to this criterion.

In 2015, a workshop sponsored by the NICHD was convened and suggested renaming this condition “intraamnionic infection and inflammation” (Higgins, 2016). The merits and clinical utility of this novel “triple I” terminology have been questioned (Barth, 2016). Nonetheless, the American College of Obstetricians and Gynecologists (2017b) recently revised both the definitions and temperature thresholds for intraamnionic infection. Using these new definitions, the diagnosis of suspected intraamniotic infection is made when the maternal temperature is ≥39.0°C or when the maternal temperature is 38.0 to 38.9°C and one additional clinical risk factor is present. Suggested factors include low parity, multiple digital examinations, use of internal uterine and fetal monitors, meconium-stained amnionic fluid, and the presence of certain genital tract pathogens. Examples are group B streptococcus and sexually transmitted agents. Isolated maternal fever is defined as any maternal temperature between 38.0°C and 38.9°C with no additional risk factors present, and with or without persistent temperature elevation.

With chorioamnionitis, fetal and neonatal morbidity are substantively increased. Alexander and colleagues (1998) studied 1367 very-low-birthweight neonates delivered at Parkland Hospital. Approximately 7 percent were born to women with overt chorioamnionitis, and their outcomes were compared with similar newborns without clinical infection. Those in the infected group had higher incidences of sepsis, RDS, early-onset seizures, intraventricular hemorrhage, and periventricular leukomalacia. The investigators concluded that these very-low-birthweight newborns were vulnerable to neurological injury attributable to chorioamnionitis. Yoon and colleagues (2000) found that intraamnionic infection in preterm neonates was related to increased rates of cerebral palsy. Petrova and associates (2001) studied more than 11 million singleton live births in the United States from 1995 to 1997. During labor, 1.6 percent of all women had fever, and this was a strong predictor of infection-related death in both term and preterm neonates.

Antimicrobial Therapy

The proposed microbial pathogenesis for spontaneous preterm labor or ruptured membranes has prompted investigators to give various antimicrobials to forestall delivery. Mercer and associates (1995) reviewed 13 randomized trials performed before 35 weeks. Their metaanalysis indicated that only three of 10 outcomes were possibly benefited: (1) fewer women developed chorioamnionitis, (2) fewer newborns developed sepsis, and (3) pregnancy was more often prolonged 7 days in women given antimicrobials. Rates of neonatal survival, necrotizing enterocolitis, RDS, or intracranial hemorrhage, however, were unaffected.

To further address this issue, the MFMU Network designed a trial to study expectant management combined with placebo or with a 7-day antibiotic regimen. Treatment included intravenous ampicillin plus erythromycin every 6 hours for 48 hours, which was followed by oral amoxicillin plus erythromycin, every 8 hours for 5 days. The women had membrane rupture between 24 and 32 weeks’ gestation. Neither tocolytics nor corticosteroids were given. Antimicrobial-treated women had significantly fewer newborns with RDS, necrotizing enterocolitis, and composite adverse outcomes (Mercer, 1997). The latency period was also significantly longer. Specifically, 50 percent of women given an antimicrobial regimen remained undelivered after 7 days of treatment compared with only 25 percent of those given placebo. Also, a significantly greater number of treated pregnancies were undelivered at 14 and 21 days. Cervicovaginal group B streptococcal colonization did not alter these results.

Other studies have examined the efficacy of shorter treatment lengths and different antimicrobial combinations. Three-day treatments compared with 7-day regimens using either ampicillin or ampicillin-sulbactam appear equally effective in regard to perinatal outcomes (Lewis, 2003; Segel, 2003). Similarly, erythromycin compared with placebo offered a range of significant neonatal benefits. The amoxicillin-clavulanate regimen was not recommended, however, because of its association with an increased incidence of neonatal necrotizing enterocolitis (Kenyon, 2004).

Some predicted that prolonged antimicrobial therapy in such pregnancies might have unwanted consequences (Carroll, 1996; Mercer, 1999). Stoll and associates (2002) studied 4337 neonates weighing from 400 to 1500 g and born from 1998 to 2000. Their outcomes were compared with those of 7606 neonates of similar birthweight born from 1991 to 1993 and prior to the practice of antibiotic prophylaxis. The overall rate of early-onset sepsis did not change between these two epochs. But, the rate of group B streptococcal sepsis dropped from 5.9 per 1000 births in the 1991 to 1993 group to 1.7 per 1000 births in the 1998 to 2000 group. Comparing these same epochs, the rate of Escherichia coli sepsis, however, rose from 3.2 to 6.8 per 1000 births. Almost 85 percent of coliform isolates from the more recent cohort were resistant to ampicillin. Neonates with early-onset sepsis were more likely to die, especially if they were infected with coliforms. Long term, Kenyon and coworkers (2008a) found that antimicrobials given for women with PPROM had no effect on the health of children at age 7 years.

Corticosteroids to Accelerate Fetal Lung Maturity

The use of antenatal corticosteroids in the setting of PPROM was once considered controversial as the magnitude of the benefits was not as great as when the membranes were intact. A single course of corticosteroids, however, is now recommended for pregnant women with ruptured membranes between 24^0/7 and 34^0/7 weeks’ gestation (American College of Obstetricians and Gynecologists, 2017a). As with periviability (Late-Preterm Birth), a single course of corticosteroids as early as 23^0/7 weeks in those who are at risk for preterm delivery within 7 days may be considered (American College of Obstetricians and Gynecologists, 2017e). A similar controversy is found at the other end of the gestational age spectrum, wherein corticosteroid administration in the late-preterm period is also under consideration (Corticosteroids for Fetal Lung Maturation).

Membrane Repair

Tissue sealants are used for various purposes in medicine, including achieving surgical hemostasis. As discussed in Chap. 18 (Inevitable Abortion), there are limited reports of sealants in the repair of fetal membranes. Crowley and coworkers (2016) recently reviewed the available evidence and concluded that data are currently insufficient to evaluate sealing procedures for ruptured membranes. At Parkland Hospital, we do not currently use these agents for this indication.

Women with signs and symptoms of preterm labor with intact membranes are managed similarly to those with PPROM. If possible, delivery before 34 weeks’ gestation is delayed. Drugs used to abate or suppress preterm uterine contractions are subsequently discussed.

Amniocentesis to Detect Infection

Several tests have been used to diagnose intraamnionic infection (Andrews, 1995; Romero, 1993; Yoon, 1996). Although such infection can be confirmed with a positive test result, there is little utility for routine amniocentesis (American College of Obstetricians and Gynecologists (2017b).

Corticosteroids for Fetal Lung Maturation

Because glucocorticosteroids were found to accelerate lung maturation in preterm sheep fetuses, Liggins and Howie (1972) evaluated them to treat women. Corticosteroid therapy was effective in lowering the incidence of RDS and neonatal mortality rates if birth was delayed for at least 24 hours after initiation of betamethasone. Infants exposed to corticosteroids in these early studies have now been followed to age 31 years with no ill effects detected. In 1995, a National Institutes of Health (NIH) Consensus Development Conference panel recommended corticosteroids for fetal lung maturation in threatened preterm birth. In a subsequent meeting, another NIH Conference (2000) concluded that data were insufficient to assess corticosteroid effectiveness in pregnancies complicated by hypertension, diabetes, multifetal gestation, fetal-growth restriction, or fetal hydrops. It was concluded, however, that it was reasonable to administer corticosteroids to these women.

A recent metaanalysis by Roberts and associates (2017) of 30 studies totaling 7774 women and 8158 infants quantified the benefit of a single course of corticosteroids. Treatment was associated with lower rates of perinatal death, neonatal death, RDS, intraventricular hemorrhage, necrotizing enterocolitis, need for mechanical ventilation, and systemic infection in the first 48 hours of life. No obvious benefits were gained for chronic lung disease, death in childhood, or neurodevelopmental delay in childhood. Therapy was not associated with chorioamnionitis. Parenthetically, corticosteroids given prophylactically to women at risk of preterm birth in low- and middle-income countries actually increase perinatal mortality rates (Althabe, 2015).

A single course of corticosteroids is currently recommended by the American College of Obstetricians and Gynecologists (2017a) for women between 24 and 34 weeks who are at risk for delivery within 7 days. This recommendation for premature twins has been challenged (Viteri, 2016). Boundaries of gestational age for corticosteroid administration are also now being explored. For pregnancies at 23 weeks and at risk of delivery within 7 days, a single course of corticosteroids may be considered (Late-Preterm Birth). Administration of corticosteroids during the periviable period is linked to a family’s decision regarding resuscitation and should be considered in that context (American College of Obstetricians and Gynecologists, 2017e).

Betamethasone and dexamethasone appear to be equivalent for fetal lung maturation (Murphy, 2007). These two drugs are comparable in reducing rates of major neonatal morbidities in preterm newborns (Elimian, 2007). A treatment course may be two 12-mg doses of betamethasone, and each dose is given intramuscularly 24 hours apart. With dexamethasone, 6-mg doses are given intramuscularly every 12 hours for four doses. Because treatment for less than 24 hours may be beneficial and reduce neonatal morbidity and mortality rates, a first dose of antenatal corticosteroids is administered regardless of the ability to complete additional doses before delivery (American College of Obstetricians and Gynecologists, 2017a).

Women at Risk for Late-Preterm Delivery

The MFMU Network conducted a randomized trial to assess whether administration of antenatal betamethasone to women who were likely to deliver in the late-preterm period would decrease respiratory and other neonatal complications (Gyamfi-Bannerman, 2016). Even though only 60 percent of the study cohort of 2831 women received both injections, the rate of respiratory complications measured as a composite outcome was lower with corticosteroid use compared with placebo—11.6 versus 14.4 percent. Because of these findings, consideration for administration of a single course of betamethasone for women between 34^0/7 and 36^6/7 weeks has been recommended by both the American College of Obstetricians and Gynecologists (2017a) and the Society for Maternal-Fetal Medicine (2016a).

Adoption of this practice has not been universal. Both short- and long-term neonatal safety are concerns (Crowther, 2016; Kamath-Rayne, 2016). Specifically, in the newborns receiving betamethasone, rates of hypoglycemia were significantly greater (Gyamfi-Bannerman, 2016). Neonatal hypoglycemia is particularly worrisome for possible adverse long-term consequences that include developmental delay (Kerstjens, 2012). Another caveat is that the largest effects of betamethasone included a reduction in transient tachypnea of the newborn—a self-limited condition with little clinical significance (Kamath-Rayne, 2016). Specifically, the rates of transient tachypnea of the newborn were 6.7 and 9.9 percent in those given betamethasone and placebo, respectively. These rates are three- to fourfold higher than those reported by the Consortium on Safe Labor (2010), which was a retrospective, observational study that abstracted detailed labor and delivery information from 19 hospitals across the United States on 233,844 deliveries. Because of these issues, we do not provide corticosteroids beyond 34 weeks at Parkland Hospital at this time.

Repeated Courses

Single versus additional courses of intramuscular corticosteroids for lung maturation has been the topic of two major trials. Although both found repeated courses to be beneficial in reducing neonatal respiratory morbidity rates, the long-term consequences were much different. In one randomized study by Crowther and associates (2007), all women at risk for preterm birth were given a primary course of betamethasone. Women were then given serial weekly doses of 11.4 mg of betamethasone for persistent risk or were given placebo. These investigators found no adverse effects in the infants followed to age 2 years. Wapner and coworkers (2007) studied infants born to 495 women who were randomly assigned to receive a single corticosteroid course that contained two doses or assigned to repeated courses that were given weekly. A nonsignificant rise in the cerebral palsy rate was identified in infants exposed to repeated courses. The doubled betamethasone dose in this study was worrisome because some experimental evidence supports the view that adverse effects are dose dependent (Bruschettini, 2006). Stiles (2007) summarized these two studies as “early gain, long-term questions.” We agree, and at Parkland Hospital, we follow the recommendation by the American College of Obstetricians and Gynecologists (2017a) for single-course therapy.

Rescue Therapy

This refers to administration of a second corticosteroid dose when delivery becomes imminent and more than 7 days have elapsed since the initial dose. In one randomized trial, 326 women received placebo or a single 12-mg dose of betamethasone (Peltoniemi, 2007). Paradoxically, the rescue dose of betamethasone increased the risk of RDS. In another randomized study of 437 women with gestations <33 weeks, Garite and associates (2009) reported significantly lower rates of respiratory complications and neonatal composite morbidity with rescue corticosteroids versus placebo. Rates of perinatal mortality and other morbidities, however, did not differ. Last, McEvoy and colleagues (2010) found that treated infants had improved respiratory compliance.

Garite and coworkers (2009) randomly assigned 437 women with singletons or twins <33 weeks’ gestation and with intact membranes to one rescue course of either betamethasone or dexamethasone or placebo. These women had all previously completed a single course of corticosteroids before 30 weeks’ gestation and at least 14 days before the rescue course. RDS developed in 41 percent of the newborns given rescue corticosteroids compared with 62 percent of those randomized to placebo. Rates of other morbidities attributable to prematurity did not differ. In a metaanalysis, Crowther and colleagues (2011) concluded that a single course of corticosteroids should be considered in women whose prior course was administered at least 7 days previously and who were <34 weeks’ gestation. The American College of Obstetricians and Gynecologists (2017a) has taken the position that a single rescue course of antenatal corticosteroids be considered in women before 34 weeks whose prior course was administered at least 7 days previously. Effects of rescue therapy beyond 34 weeks are currently unknown. At Parkland Hospital, we currently do not provide additional courses of corticosteroids beyond the initial single-course therapy.

Magnesium Sulfate for Neuroprotection

Very-low-birthweight neonates whose mothers were treated with magnesium sulfate for preterm labor or preeclampsia were found to have a reduced incidence of cerebral palsy at 3 years (Grether, 2000; Nelson, 1995). Because of this, randomized trials were designed to investigate this hypothesis. In one trial, 1063 women at risk of delivery before 30 weeks were given magnesium sulfate or placebo (Crowther, 2003). Magnesium exposure improved some perinatal outcomes. Namely, rates of both neonatal death and cerebral palsy were lower in the magnesium-treated group—but this study was not sufficiently powered. The multicenter French trial reported by Marret and associates (2008) had similar problems.

More convincing evidence for magnesium neuroprotection came from the MFMU Network study—Beneficial Effects of Antenatal Magnesium Sulfate—BEAM—Study (Rouse, 2008). This was a placebo-controlled trial in 2241 women at imminent risk for preterm birth between 24 and 31 weeks. Women randomized to magnesium sulfate were given a 6-g bolus over 20 to 30 minutes followed by a maintenance infusion of 2 g per hour. Magnesium sulfate was actually infusing at the time of delivery in approximately half of the treated women. A 2-year follow-up was available for 96 percent of the children. Results are shown in Table 42-10. This trial can be interpreted differently depending on statistical methodologies employed. Some interpret these findings to mean that magnesium infusion prevents cerebral palsy regardless of the gestational age at which therapy is given. Those with a differing view conclude that this trial only supports use of magnesium sulfate for prevention of cerebral palsy before 28 weeks.

Subsequent to these studies, Doyle and associates (2009) reviewed five randomized trials to assess neuroprotective effects. A total of 6145 infants were studied, and these reviewers concluded that magnesium exposure compared with no exposure significantly lowered risks for cerebral palsy. Rates of other neonatal morbidity did not differ significantly. It was calculated that treatment given to 63 women would prevent one case of cerebral palsy.

Controversy surrounding magnesium efficacy for neuroprotection prompted a debate at the 2011 annual meeting of the Society for Maternal-Fetal Medicine. Rouse (2011) spoke for the benefits of magnesium sulfate, whereas Sibai (2011) challenged that the reported benefits were false positive due to random statistical error in the metaanalysis by Doyle (2009). Another peculiarity is the apparent lack of dose-response for efficacy (McPherson, 2014). Because none of the individual studies found a benefit from magnesium sulfate for fetal neuroprotection, the American College of Obstetricians and Gynecologists (2016a) concluded that those electing prophylaxis should develop specific guidelines. To guide such therapy, the American College of Obstetricians and Gynecologists (2012) issued a Patient Safety Checklist for use of magnesium sulfate for neuroprotection. For those with PPROM, prophylaxis may similarly be considered. At Parkland Hospital, we provide magnesium sulfate for neuroprotection with threatened preterm delivery from 24^0/7 to 27^6/7 weeks.

Antimicrobials

Results have been disappointing in studies of antimicrobials given to arrest preterm labor. From one Cochrane metaanalysis, antimicrobial prophylaxis given to women with intact membranes did not reduce preterm birth rates or affect other clinically important short-term outcomes (Flenady, 2013). However, rates of short- and longer-term harm were higher for children of mothers exposed to antibiotics. Kenyon (2001) reported the ORACLE Collaborative Group study of 6295 women with spontaneous preterm labor and intact membranes, but without evidence of infection. Women were randomly assigned to receive antimicrobial or placebo therapy. The primary outcomes of neonatal death, chronic lung disease, and major cerebral abnormality were similar in both groups. In a follow-up of the ORACLE II trial, fetal exposure to antimicrobials in this clinical setting was associated with an increased cerebral palsy rate at age 7 years compared with that in children without fetal exposure (Kenyon, 2008b). Importantly, antimicrobial use described here is distinct from that given for group B streptococcal prophylaxis (Chap. 64, Methicillin-Resistant Staphylococcus aureus).

Bed Rest

This is one of the most often prescribed interventions during pregnancy, yet one of the least studied. One systematic review concluded that evidence neither supported nor refuted bed rest for prevention of preterm birth (Sosa, 2004). Goulet and coworkers (2001) randomly assigned 250 Canadian women to either home care or hospitalization after treatment of an acute episode of preterm labor and found no benefits. There have, however, been reports of possible harm. Kovacevich and associates (2000) reported that bed rest for 3 days or more increased thromboembolic complications to 16 per 1000 women compared with only 1 per 1000 with normal ambulation. Promislow and colleagues (2004) observed significant bone loss in pregnant women prescribed outpatient bed rest. More recently, Grobman and associates (2013) noted that women with activity restriction were nearly 2.5 times more likely to have a preterm birth before 34 weeks. This finding, however, may reflect ascertainment bias. That is, women with restricted activity may have been assigned to bed rest because they were viewed to be at more imminent risk of preterm delivery. McCall and coworkers (2013) summarized the literature on bed rest, and they found insufficient evidence to support its use. The American College of Obstetricians and Gynecologists (2017d) suggests that, although frequently prescribed, bed rest is only rarely indicated, and ambulation should be considered in most cases.

Cervical Pessaries

Silicone rings, such as the Arabin pessary, are being used to support the cervix in women with a sonographically short cervix. For 385 Spanish women with a cervical length ≤25 mm, Goya and associates (2012) provided a silicone pessary or expectant management. Newborns spontaneous delivered before 34 weeks’ gestation in 6 percent of women in the pessary group compared with 27 percent in the expectant management group. Another trial randomly assigned almost 100 women with a cervix <25 mm at 20 to 24 weeks to silicone pessaries or expectant management (Hui, 2013). The pessary did not lower the rate of delivery <34 weeks. Similar findings were reported by Nicolaides and colleagues (2016). The Society for Maternal-Fetal Medicine (2017b) recently recognized the conflicting published reports and lack of an FDA-approved pessary for the indication of preterm birth prevention. They currently recommend pessary prophylaxis only within research protocols.

Emergency or Rescue Cerclage

Some evidence supports the concept that cervical incompetence and preterm labor lie along a spectrum leading to preterm delivery. Consequently, investigators have evaluated cerclage placement after preterm labor begins to manifest clinically. Althuisius and colleagues (2003) randomly assigned 23 women with cervical incompetence before 27 weeks to bed rest, with or without emergency McDonald cerclage. Delivery delay was significantly greater in the cerclage group compared with those assigned to bed rest—54 versus 24 days, respectively. Terkildsen and coworkers (2003) studied 116 women who underwent second-trimester emergency cerclage. Nulliparity, membranes extending beyond the external cervical os, and cerclage before 22 weeks were associated with a significantly decreased chance of significant pregnancy continuation. For women facing a poor pregnancy prognosis due to cervical dilation at midgestation, it seems reasonable to offer emergency or rescue cerclage with appropriate counseling. However, it is unclear if such interventions truly confer a benefit or merely increase the risk of membrane rupture and infection (Hawkins, 2017).

Tocolysis to Treat Preterm Labor

Although several drugs and other interventions have been used to prevent or inhibit preterm labor, none is completely effective. The American College of Obstetricians and Gynecologists (2016b) has concluded that tocolytic agents do not markedly prolong gestation but may delay delivery in some women for up to 48 hours. This may allow transport to an obstetrical center with higher-level neonatal care and permit time for a course of corticosteroid therapy. Although delivery may be delayed to administer corticosteroids, treatment has not improved perinatal outcome rates (Gyetvai, 1999).

Beta-adrenergic agonists, magnesium sulfate, calcium-channel blockers, or indomethacin are the recommended tocolytic agents for this short-term use. The gestational age range for tocolytic use is debatable. But, because corticosteroids are not generally used after 34 weeks, and because the perinatal outcomes in preterm neonates are generally good after this time, most do not recommend use of tocolytics after 33 weeks’ gestation (Goldenberg, 2002).

In many women, tocolytics stop contractions temporarily but rarely prevent preterm birth. The College (2016b) notes that maintenance therapy with tocolytics is ineffective for preventing preterm birth. Importantly, no trial has ever convincingly shown reductions in rates of any important adverse outcome by a tocolytic drug compared with placebo (Walker, 2016). Maintenance tocolysis after acute therapy is not recommended.

β-Adrenergic Receptor Agonists

Several compounds react with β-adrenergic receptors to reduce intracellular ionized calcium levels and prevent activation of myometrial contractile proteins (Chap. 21, G-Protein–Coupled Receptors). Of β-mimetic drugs in the United States, ritodrine and terbutaline have been used in obstetrics, but only ritodrine is approved for preterm labor by the FDA.

Ritodrine was voluntarily withdrawn from the United States market in 2003, but a discussion of ritodrine is included here to present issues with β-mimetic drug use. In one early trial, neonates whose mothers were treated with ritodrine for threatened preterm labor had lower rates of preterm birth and its complications (Merkatz, 1980). In a randomized trial at Parkland Hospital, intravenous ritodrine delayed delivery for 24 hours but without other benefits (Leveno, 1986b). Additional studies confirmed a delivery delay up to 48 hours (Canadian Preterm Labor Investigators Group, 1992).

β-Agonist drug infusion has resulted in serious and even fatal maternal side effects. Pulmonary edema is a special concern, and its contribution to morbidity is discussed in Chapter 47 (Acute Pulmonary Edema). In one early study, tocolysis was the third most common cause of acute respiratory distress and death in pregnant women during a 14-year period in Mississippi (Perry, 1998). The cause of pulmonary edema is multifactorial. Risk factors include tocolytic therapy with β-agonist drugs, multifetal gestation, concurrent corticosteroid therapy, tocolysis for more than 24 hours, and intravenous infusion of large volumes of crystalloid. β-Agonist agents cause retention of sodium and water, and with time—usually 24 to 48 hours—these can cause volume overload (Hankins, 1988). The drugs have been implicated in increased capillary permeability, cardiac rhythm disturbances, and myocardial ischemia.

Terbutaline is commonly used in the United States to forestall preterm labor. Like ritodrine, it may cause pulmonary edema (Angel, 1988). Low-dose terbutaline can be administered long-term by subcutaneous pump (Lam, 1988; Perry, 1995). But, randomized trials have shown no benefit for terbutaline pump therapy (Guinn, 1998; Wenstrom, 1997). Oral terbutaline given to prevent preterm delivery is also ineffective (How, 1995; Parilla, 1993). In one trial, 203 women with arrested preterm labor at 24 to 34 weeks’ gestation were randomly assigned to receive 5-mg terbutaline tablets or placebo every 4 hours (Lewis, 1996). Of outcomes, delivery rates at 1 week, median latency duration, mean gestational age at delivery, and incidence of preterm labor relapse were similar in both groups. Because of reports of serious maternal side effects, the FDA (2011) issued a warning regarding the use of terbutaline to treat preterm labor. The American College of Obstetricians and Gynecologists (2016b) recommends only short-term inpatient use of terbutaline as a tocolytic or as acute therapy of uterine tachysystole. Subcutaneous dosages of 0.25 mg are commonly used for the latter indication. Terbutaline, used as a tocolytic prior to external cephalic version, is discussed in Chapter 28 (Tocolysis).

Magnesium Sulfate

Ionic magnesium in a sufficiently high concentration can alter myometrial contractility. Its role is presumably that of a calcium antagonist, and when given in pharmacological doses, it may inhibit labor. Intravenous magnesium sulfate, given as a 4-g loading dose and followed by a continuous infusion of 2 g/hr, usually arrests labor (Steer, 1977). Like β-mimetic agents, magnesium treatment can cause pulmonary edema (Samol, 2005). However, this has not been our experience at Parkland Hospital in the treatment of tens of thousands of preeclamptic women with intravenous magnesium sulfate. Pharmacology and toxicology of magnesium are considered in more detail in Chapter 40 (Pharmacology and Toxicology).

Only two randomized studies have evaluated tocolysis with magnesium sulfate. Cotton and colleagues (1984) compared magnesium sulfate, ritodrine, and placebo in 54 women with preterm labor. They identified few differences in outcomes. Cox and coworkers (1990) randomly assigned 156 women to receive magnesium sulfate or infusions of normal saline. Magnesium-treated women and their neonates had identical outcomes compared with those given placebo. Because of these findings, this method of tocolysis was abandoned at Parkland Hospital. Similarly, Crowther and associates (2014) reviewed magnesium sulfate as a tocolytic agent and concluded it was ineffective and potentially harmful. Last, the FDA (2013) has warned against prolonged use of magnesium sulfate given to arrest preterm labor because of bone thinning and fractures in fetuses exposed for more than 5 to 7 days. This was attributed to low calcium levels in the fetus.

Prostaglandin Inhibitors

These compounds are intimately involved in contractions of normal labor (Chap. 21, Uterotonins in Parturition Phase 3). Antagonists act by inhibiting prostaglandin synthesis or by blocking their action on target organs. A group of enzymes collectively termed prostaglandin synthase is responsible for the conversion of free arachidonic acid to prostaglandins. Several drugs block this system, including acetylsalicylate and indomethacin.

Indomethacin, a nonselective cyclooxygenase inhibitor, was first used as a tocolytic in one study of 50 women (Zuckerman, 1974). Studies that followed reported the efficacy of indomethacin in halting contractions and delaying preterm birth (Muench, 2003; Niebyl, 1980). Morales and coworkers (1989, 1993a), however, compared indomethacin with either ritodrine or magnesium sulfate and found no difference in their efficacy to forestall preterm delivery. Berghella and associates (2006) reviewed four trials of indomethacin given to women with a sonographically determined short cervix and found such therapy to be ineffective.

Indomethacin is administered orally or rectally. Most studies have limited indomethacin use to 24 to 48 hours because of concerns for oligohydramnios, which can develop with therapeutic doses. If amnionic fluid is monitored, oligohydramnios can be detected early, and it is reversible with drug discontinuation.

In a study of neonates born before 30 weeks, Norton and coworkers (1993) identified necrotizing enterocolitis in 30 percent of 37 indomethacin-exposed newborns compared with 8 percent of 37 control newborns. Higher incidences of intraventricular hemorrhage and patent ductus arteriosus were also documented in the indomethacin group. Several investigators have challenged the association between indomethacin exposure and necrotizing enterocolitis (Muench, 2001; Parilla, 2000). Similarly, Gardner (1996) and Abbasi (2003) and their colleagues found no link between indomethacin use and intraventricular hemorrhage, patent ductus arteriosus, sepsis, necrotizing enterocolitis, or neonatal death. Two metaanalyses of the effects of antenatal indomethacin on neonatal outcomes had conflicting findings (Amin, 2007; Loe, 2005). Reinebrant and colleagues (2015) in a review of 20 studies reported no clear benefit from cyclooxygenase inhibitors, including indomethacin, compared with placebo or any other tocolytic agent.

Nitric Oxide Donors

These potent smooth-muscle relaxants affect the vasculature, gut, and uterus. In randomized clinical trials, nitroglycerin administered orally, transdermally, or intravenously was ineffective or showed no superiority over other tocolytics. In addition, maternal hypotension was a common side effect (Bisits, 2004; El-Sayed, 1999; Lees, 1999).

Calcium-Channel Blockers

Discussed in Chapter 21 (Actin-Myosin Interactions), myometrial activity is directly related to cytoplasmic free calcium, and reduced calcium concentrations inhibit contractions. Calcium-channel blockers act to inhibit, by various mechanisms, calcium entry through cell membrane channels. Although they were developed to treat hypertension, their ability to arrest preterm labor has been evaluated.

From study results, calcium-channel blockers, especially nifedipine, are safer and more effective tocolytic agents than β-agonist drugs (King, 2003; Papatsonis, 1997). Lyell and colleagues (2007) randomized 192 women at 24 to 33 weeks’ gestation to either magnesium sulfate or nifedipine and found no substantial differences in efficacy or adverse effects. In another randomized study, 145 women with preterm labor between 24 and 33 weeks received nifedipine or atosiban. Neither proved superior to delay delivery, and neonatal morbidity was equivalent (Salim, 2012).

Flenady and coworkers (2014b) reviewed 38 trials of calcium- channel blockers (mainly nifedipine) for preterm labor. These investigators suggested that calcium-channel blockers have benefits compared with placebo or no treatment. But, this conclusion stemmed from a trial with unclear risk of selection bias and a three-arm study of 84 women that was not blinded (Ara, 2008; Zhang, 2002). We are currently performing a randomized, double-blind, placebo-controlled trial of nifedipine for acute tocolysis of preterm labor at Parkland Hospital.

Importantly, the combination of nifedipine with magnesium for tocolysis is potentially dangerous. Ben-Ami (1994) and Kurtzman (1993) and their coworkers reported that nifedipine enhances the neuromuscular blocking effects of magnesium, which can interfere with pulmonary and cardiac function. In one small study of 54 women with preterm labor who received either magnesium sulfate plus nifedipine or no tocolytic, neither benefit nor harm was found (How, 2006).

Atosiban

This nonapeptide oxytocin analogue is an oxytocin-receptor antagonist (ORA). Goodwin and colleagues (1995) described its pharmacokinetics in pregnant women. In randomized clinical trials, atosiban failed to improve relevant neonatal outcomes and was linked with significant neonatal morbidity (Moutquin, 2000; Romero, 2000). The FDA has denied approval of atosiban because of concerns regarding efficacy and fetal–newborn safety. Further, in 2014, a metaanalysis did not demonstrate superiority of ORAs (largely atosiban) as a tocolytic compared with placebo, β-mimetic drugs, or calcium-channel blockers in terms of pregnancy prolongation or neonatal outcomes. But, ORAs were associated with fewer maternal adverse effects (Flenady, 2014a). Recently, van Vliet and coworkers (2016) conducted a randomized trial comparing nifedipine with atosiban in 510 women with threatened preterm birth. Using a composite of adverse perinatal outcomes, no differences were reported between the two study groups.

Labor

Whether preterm labor is induced or spontaneous, abnormalities of fetal heart rate and uterine contractions are sought. We prefer continuous electronic monitoring. Fetal tachycardia, especially with ruptured membranes, is suggestive of sepsis. Some evidence supports that intrapartum acidemia may intensify some of the neonatal complications usually attributed to preterm delivery. For example, Morgan and associates (2017) found that metabolic acidemia significantly raised the risks related to prematurity in neonates delivered prior to 34 weeks’ gestation. Low and colleagues (1995) observed that intrapartum acidosis—umbilical artery blood pH <7.0—had an important role in neonatal complications (Chap. 33, Neonatal Encephalopathy). Group B streptococcal infections are common and dangerous in the preterm neonate, and antimicrobial prophylaxis should be provided (Chap. 64, Methicillin-Resistant Staphylococcus aureus).

Delivery

In the absence of a relaxed vaginal outlet, an episiotomy for delivery may be necessary once the fetal head reaches the perineum. Perinatal outcome data do not support routine episiotomy or forceps delivery to protect the “fragile” preterm fetal head. Staff proficient in resuscitative techniques commensurate with the gestational age and fully oriented to any specific problems should be present at delivery. Principles of resuscitation described in Chapter 32 (Resuscitation Protocol) are applicable. The importance of specialized personnel and facilities for preterm newborn care is underscored by the improved survival rates of these neonates when delivered in tertiary-care centers.

Prevention of Intracranial Hemorrhage

Preterm newborns frequently have intracranial germinal matrix bleeding that can extend to more serious intraventricular hemorrhage (Chap. 34, Retinopathy of Prematurity). It was hypothesized that cesarean delivery to obviate trauma from labor and vaginal delivery might prevent these complications. This has not been validated by subsequent studies. Malloy (1991) analyzed 1765 newborns with birthweights <1500 g and found that cesarean delivery did not lower the risk of mortality or intracranial hemorrhage. Anderson and colleagues (1988), however, made an interesting observation regarding the role of cesarean delivery in intracranial hemorrhage prevention. These hemorrhages correlated with exposure to active-phase labor. However, they emphasized that avoidance of active-phase labor is impossible in most preterm births because decisions for delivery route are not required until active labor is firmly established.