Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Obstetric Factors Associated with Infections of the Fetus and Newborn Infant
Overview
Obstetric infections remain an important and potentially preventable contributor to maternal and neonatal morbidity and mortality. Intraamniotic infection (IAI) is associated with 20% of all preterm births (PTBs) and 50% of extreme PTBs at 28 weeks of gestation or less. PTB, defined as birth at less than 37 completed weeks of gestation, is one of the leading causes of adverse outcomes of pregnancy. In the United States, in 2011, nearly one in every eight neonates was born prematurely. Prematurity accounts for 70% of all perinatal deaths and half of long-term neurologic morbidity in the United States, and globally prematurity is now the second leading underlying cause of under-5 childhood mortality. In most cases, the underlying cause of preterm labor is not apparent, but evidence from many sources points to a relationship between PTB and genitourinary tract infections. In addition to the genitourinary tract, infection leading to PTB may arise in the placenta or from a more remote site, such as the periodontal tissues. Newer information has suggested that subclinical infection is responsible not only for PTB but also for many serious neonatal sequelae, including periventricular leukomalacia, cerebral palsy, bronchopulmonary dysplasia, and necrotizing enterocolitis. In addition, one quarter to one third of PTBs are preceded by preterm premature rupture of the membranes (PPROM), which is associated with an increased risk of maternal endometritis and neonatal sepsis. Maternal endometritis can develop into puerperal sepsis, which is an infection of the genital tract that begins any time between rupture of membranes or onset of labor and the 42nd postpartum day. Puerperal sepsis is the third leading cause of maternal mortality globally, accounting for 10% to 12% of all maternal deaths. This chapter will focus on these entities and their consequences.
Infections not related specifically to pregnancy (e.g., human immunodeficiency virus [HIV], curable sexually transmitted infections, tuberculosis, and malaria) are also important contributors to maternal and neonatal mortality and morbidity. HIV is the leading non-obstetric infectious cause of maternal death in sub-Saharan Africa and also increases the risk for puerperal sepsis compared with HIV-negative women. Tuberculosis is an important comorbidity among HIV-infected women in Africa. For example, an autopsy study from Mozambique found that 8.5% of maternal deaths were due to tuberculosis in the context of a 53% HIV prevalence rate. Similarly, malaria is an important contributor to maternal mortality in endemic regions; placental malaria is associated with stillbirth, preterm delivery, and low birth weight. Finally, the World Health Organization estimates that 340 million treatable sexually transmitted infections occur each year (e.g., syphilis, gonorrhea, chlamydial infections, trichomoniasis) and are associated with a wide range of adverse pregnancy outcomes. These and other non-obstetric infections are covered in detail elsewhere in this book. In this chapter, we focus on infection as a cause of PTB and neonatal morbidity and mortality.
Infection as a Cause of Preterm Birth
Intraamniotic infection (IAI), prematurity, and PPROM are important risk factors for neonatal infectious morbidity and mortality. IAI occurs in 1% to 4% of all pregnant women and is the main cause of the earliest PTBs that impart the greatest risk of neonatal mortality and morbidity. Synonymous terms applied to IAI include clinical chorioamnionitis, amnionitis, and amniotic fluid infection. We use the term IAI to encompass a clinical syndrome, distinct from bacterial colonization of amniotic fluid (also referred to as microbial invasion of the amniotic cavity) and from histologic inflammation of the placenta (i.e., histologic chorioamnionitis). Histologic chorioamnionitis is characterized by a polymorphonuclear leukocyte infiltration of the chorioamnion and occurs more frequently than clinical IAI, especially at term.
IAI is an important contributor to PTB, occurring in approximately 20% (range, 9%-39%) of all women in preterm labor when assessed by standard laboratory culture of the amniotic fluid ( Table 3-1 ). Amplification of bacterial DNA encoding 16S ribosomal RNA (rDNA) is a more sensitive test and detects IAI in a greater number of women with preterm labor (range, 11%-56%). The rate of IAI also appears to increase in pregnancies with a lower gestational age. In a study of 105 women in preterm labor, IAI was diagnosed by culture in 67% (4/6) at 23 to 24 weeks, 36% (5/14) at 25 to 26 weeks, 17% (2/12) at 27 to 28 weeks, and 11% (4/36) at 33 to 34 weeks of gestation.
Table 3-1
Prevalence of a Positive Culture and/or PCR Test for Bacteria in the Amniotic Fluid of Women in Preterm Labor with Intact Membranes
Modified from Table 1 in DiGuilio DB: Diversity of microbes in amniotic fluid. Semin Fetal Neonatal Med 17:2-11, 2012.
| Author (year) | N | Culture | PCR | Culture and PCR Combined | Reference |
|---|---|---|---|---|---|
| Watts (1992) | 105 | 19% | — | — | |
| Romero (1989) | 264 | 9% | |||
| Hitti (1997) | 69 | 23% | 30% | 32% | |
| Markenson (1997) | 54 | 9% | 56% | 56% | |
| Oyarzun (1998) | 50 | 12% | 46% | 46% | |
| DiGiulio (2008) | 166 | 10% | 11% | 15% | |
| Han (2009) | 26 | 39% | 50% | 50% | |
| Marconi (2011) | 20 | — | 40% | — |
PCR, Polymerase chain reaction.
Pathogenesis
Before labor and membrane rupture, amniotic fluid is usually sterile. The physical and chemical barriers formed by intact fetal membranes (chorioamnion) and the cervical mucus are usually effective in preventing entry of bacteria. With the onset of labor or with membrane rupture, bacteria from the lower genital tract may enter the amniotic cavity. The origin of IAI was originally postulated to occur in four discrete stages of an ascending IAI ( Fig. 3-1 A). The first stage represents a shift in vaginal or cervical microbial flora with trafficking of bacteria into the cervix. An example of stage I is bacterial vaginosis (BV), a heterogeneous vaginal condition associated with perturbed vaginal flora and loss of healthy Lactobacillus organisms. In stage II, bacteria ascend from the vagina or cervix into the choriodecidua, the specialized endometrium of pregnancy. The inflammatory response here facilitates trafficking of organisms into the chorioamnion, leading to chorioamnionitis. In stage III, bacteria invade chorionic vessels (choriovasculitis) and migrate through the amnion into the amniotic cavity to cause IAI. Bacteria recovered from the amniotic fluid and fetal membranes generally consist of organisms that colonize the vagina, including gram-positive (group B streptococci [GBS]), gram-negative (e.g. Escherichia coli, Gardnerella vaginalis ), and anaerobic ( Mycoplasma hominis ) bacteria ( Table 3-2 ). Once bacteria invade the amniotic cavity, bacteria may then gain access to the fetus through several potential mechanisms, culminating in stage IV; fetal bacteremia, sepsis, and pneumonia.
Table 3-2
Microbial Diversity Within the Amniotic Fluid From Women in Preterm Labor with Intact Membranes as Detected by Culture and PCR
| Culture | PCR | |||
|---|---|---|---|---|
| Bacterial Species | % Positive (No. Positive /Total N ∗ ) | Reference | % Positive (No. Positive/Total N ∗ ) |
Reference |
| Genital Mycoplasmas | ||||
| Ureaplasma spp. | 25% (6/24) | 10% (2/21) | ||
| 35% (7/20) | ||||
| 13% (4/30) | ||||
| 61% (14/23) | ||||
| 19% (3/16) | 16% (3/19) | |||
| 24% (5/21) | ||||
| Mycoplasma spp. | 17% (4/24) | 5% (1/19) | ||
| 5% (1/20) | ||||
| 7% (2/30) | ||||
| 6% (1/16) | 10% (2/21) | |||
| Anaerobes | ||||
| Bacteroides spp. | 13% (3/24) | 14% (3/21) | ||
| 40% (8/20) | ||||
| Fusobacterium spp. | 21% (5/24) | 33% (7/21) | ||
| 35% (7/20) | ||||
| 31% (5/16) | 4% (2/50) | |||
| 25% (4/16) | 26% (5/19) | |||
| Aerobes | ||||
| Group B streptococci | 4% (1/24) | 5% (1/19) | ||
| 0% (0/20) | ||||
| 7% (2/30) | ||||
| 6% (1/16) | 10% (2/21) | |||
| 13% (2/16) | ||||
| Escherichia coli | 0% (0/24) | 26% (13/50) | ||
| 0% (0/20) | ||||
| 13% (2/16) | ||||
| Gardnerella vaginalis | 13% (3/24) | — | ||
| 15% (3/20) | ||||
| 13% (2/16) | — | |||
| Fastidious, Noncultivatable Bacteria | ||||
| Leptotrichia amnionii | — | 11% (2/19) | ||
| 5% (1/21) | ||||
| Sneathia sanguinegens | — | 21% (4/19) | ||
| 14% (3/21) † | ||||
| Fungi | ||||
| Candida albicans | 10% (2/20) | 11% (2/19) | ||
| 7% (2/30) | ||||
| 6% (1/16) | ||||
PCR, Polymerase chain reaction.
∗ Total N represents all the positive cultures or PCR tests reported in the study.
† A positive PCR test for L. sanguinegens ( N = 1) and S. sanguinegens ( N = 2) in the original study were combined into S. sanguinegens because of a recent change in nomenclature.
Alternatively, an ascending IAI has been hypothesized to occur, which emphasized rapid trafficking through the fetal membranes into the amniotic fluid with subsequent colonization of the membranes from bacteria in the amniotic fluid ( Fig. 3-1 B). In this model, widespread colonization of the chorioamnion occurs as a secondary event after microbial invasion of the amniotic cavity. Animal models have also suggested that PTB can occur in the absence of microbial invasion of the amniotic cavity as a result of choriodecidual colonization of a low-dose of bacteria or lipopolysaccharide (LPS), leading to an inflammatory response culminating in both clearance of the bacteria and PTB ( Fig. 3-1 C). From this perspective, choriodecidual inflammation is a transitional stage of ascending infection. Regardless of mechanism, IAI is likely dependent on bacterial inoculum, microbial pathogenicity and host response.
Investigation of bacterial trafficking into the pregnant uterus has been limited in humans for ethical reasons; no animal model has been used to study these events. However, as early as the 1960s, bacterial trafficking from the lower genital tract into the uterus was known to occur in nonpregnant women based on the rapid movement of carbon particles into the abdomen from the vagina and likely occurs during pregnancy as well. No single lower genital tract bacteria studied has been associated with a higher risk for PTB with the possible exception of Trichomonas vaginalis , which increases the risk of PPROM. Factors that facilitate an ascending bacterial infection during pregnancy remain elusive. Immune responses within the vagina, cervix, and placenta likely play an important role to eliminate pathogenic bacteria before invasion of the chorioamnion and amniotic cavity. Although a robust inflammatory response may trigger preterm labor, not all individuals will deliver preterm, underscoring the role of host-pathogen interactions in the process.
The inflammatory response triggered by bacteria is a central mechanism of PTB and fetal injury. Perinatal research has focused on small immunologic proteins called cytokines and chemokines, which are frequently detected in the amniotic fluid and cord blood of preterm neonates with an IAI. Frequently studied cytokines associated with PTB include interleukin-1β (IL-1β), IL-6, IL-8, tumor necrosis factor-α (TNF-α), and C-X-C motif chemokine ligand (CXCL) 10. Many different placental tissues produce cytokines in response to bacterial stimulation, including amniotic epithelium, chorion, decidua and trophoblast cells. IL-1β and TNF-α are considered key inflammatory cytokines because amniotic fluid infusion of either is capable of inducing PTB in a nonhuman primate model. Although infusion of IL-6 does not induce PTB, murine IL-6 knockouts deliver 1 day later than expected and are refractory to LPS stimulated PTB. This suggests that IL-6 may play an important role in controlling the progression of labor and facilitating PTB. IL-6 and other cytokines can also stimulate prostaglandin production by amniotic epithelium and the decidua, which further drive labor.
Another model of placental infection may be specific to Listeria monocytogenes , an aerobic gram-positive rod associated with PTB and stillbirth. In this case, hematogenous spread of the bacteria occurs after ingestion of colonized food (e.g., unpasteurized cheese). Next, bacteria invade the extravillous trophoblast cells, which are specialized invasive placental cells that form the placenta vascular bed. Within the extravillous trophoblast cells, L. monocytogenes is efficiently confined for a period of time within vacuolar compartments destined for lysosome degradation. Over time, the placenta becomes a dangerous nidus of infection and continuously reseeds maternal organs until the placenta is expulsed with delivery. Maternal deaths may occur as a result, which is also the case for other virulent organisms with hematogenous spread such as group A streptococci. The mechanism of stillbirth may result from an impaired response of the normally immunosuppressive maternal T-regulatory (Foxp3 + ) cells to fetal antigens.
IAI may also develop as a consequence of obstetric procedures such as cervical cerclage, diagnostic amniocentesis, cordocentesis (percutaneous umbilical cord blood sampling), or intrauterine transfusion. After cervical cerclage, data regarding infectious complications are reported to range from 1% to 18%, but this may also be a consequence of cervical shortening and dilation. After diagnostic amniocentesis, rates of IAI range from 0% to 1%. Chorioamnionitis is a rare complication of chorionic villus sampling. Although IAI is very rare after percutaneous umbilical blood sampling, and the fetal loss rate accompanying this procedure is only 1% to 2%, infection is responsible for a high percentage of losses and may lead to life-threatening maternal complications.
Clinical risk factors for IAI have implicated low parity, a greater number of vaginal examinations in labor, longer labor, greater duration of membrane rupture, and internal fetal monitoring. The most significant clinical risk factor is membrane rupture longer than 12 hours. Meconium staining of the amniotic fluid has also been associated with an increased risk of chorioamnionitis (4.3% vs. 2.1%). Prior spontaneous and elective abortion (at <20 weeks) in the immediately preceding pregnancy has also been associated with development of IAI in the subsequent pregnancy.
Histologic Chorioamnionitis
A consistent observation is that placentas in PTBs are more likely to show evidence of inflammation (i.e., histologic chorioamnionitis). In a series of 3500 consecutive placentas, Driscoll found infiltrates of polymorphonuclear cells in 11%. Clinically evident infection developed in only a few of the women in the study, but the likelihood of neonatal sepsis and death was increased. An association has also been established between histologic chorioamnionitis and chorioamnion infection (defined as the recovery of microorganisms from the chorioamnion). Overall, the organisms found in the chorioamnion are similar to organisms found in the amniotic fluid in cases of clinical IAI. This array of organisms supports an ascending route for chorioamnion infection in most cases. The rate of histologic chorioamnionitis increases with decreasing gestational age at delivery. In one study, when birth weight was greater than 3000 g, the percentage of placentas showing histologic chorioamnionitis was less than 20%; when birth weight was less than 1500 g, the percentage was 60% to 70%.
Animal Models
Pregnant mice, rabbits and sheep have been used to identify inflammatory mediators in the PTB pathway and potential therapeutic targets that might improve neonatal outcome. Together, data from multiple animal models have been critical for understanding the inflammatory cascade occurring in the placenta, amniotic fluid, fetus and mother, which is necessary to prevent development of IAI, PTB, and early-onset neonatal infections. However, many animal models are sufficiently different from humans in terms of hormonal events associated with parturition and placental structure to limit their translational potential.
The model that most closely resembles human pregnancy in both placentation and hormonal events of pregnancy is the chronically instrumented nonhuman primate (NHP) model, using either pregnant rhesus or pigtail macaques. A major strength of the NHP model lies in the ability to longitudinally correlate data from maternal, fetal, and amniotic fluid samples with uterine contractility over time in individual animals. Experiments in the NHP model have elucidated the primary role for maternal and fetal inflammation in mediating infection-associated PTB. For example, experiments in the NHP model demonstrated that immunomodulators (dexamethasone, indomethacin) that can suppress the inflammatory response in combination with antibiotics could prolong pregnancy in women with IAI from GBS. Further study is required to demonstrate fetal safety of this approach, as well as efficacy, in animal models.
Animal models have also delineated the role of specific inflammatory mediators in PTB and therapeutic potential of their antagonists.
- ▪
IL-1β and TNF-α: IL-1β and TNF-α are cytokines elevated in the amniotic fluid of women with IAI and appear to play a critical role in infection-induced PTB in animal models. Infusion of IL-1β into the amniotic fluid of NHP leads uniformly to PTB, whereas high-dose TNF-α caused PTB in about half the animals. However, the overexpression of the natural IL-1 receptor antagonist in transgenic mice could not inhibit IL-1β induction of PTB. Pretreatment with an antibody to TNF-α was shown to reduce PTB and fetal death.
- ▪
IL-6 and IL-8: In human studies, amniotic fluid IL-6 and IL-8 has been implicated in the cytokine cascade, culminating in PTB. In NHP, intraamniotic infection or inflammation induced by GBS, LPS, and other pathogens elevates IL-6 and IL-8 in amniotic fluid. However, isolated infusions of IL-6 or IL-8 into the amniotic fluid do not cause PTB. This suggests a hierarchy of cytokine effects on preterm labor, with IL-1β and TNF-α playing a more important role than either IL-6 or IL-8. Recently, IL-6 has been postulated to facilitate labor progression, because there is a slight delay in normal parturition and LPS-induced PTB in IL-6 knockout mice.
- ▪
N -acetylcysteine (NAC): Oxidative stress is implicated in PPROM by promoting apoptosis and collagen weakening through production of reactive oxygen species (ROS). ROS are unstable molecules released by immune cells during bacterial killing and from mitochondria that are capable of widespread membrane damage. NAC counteracts oxidative stress and has shown promise in delaying PTB in mice and rats. Administration of NAC in a murine model of LPS-induced PTB was associated with a significant delay in delivery, reduction in oxidative stress, and greater fetal survival. In rats, NAC was shown to significantly attenuate an LPS-induced cytokine response in maternal serum and amniotic fluid.
- ▪
Toll-like receptors (TLRs): TLRs are a family of pattern recognition receptors that activate the innate and adaptive immune systems after recognition of bacterial pathogens. PTB has been shown to occur in mice after activation of either TLR2, TLR3, TLR4, or TLR9 by their respective ligands: lipotechoic acid or peptidoglycan, polyinosinic:polycytidylic acid (poly[I:C]), LPS, and cytosine-phosphate-guanosine (CpG) dinucleotide, respectively. Research has focused on TLR4 because gram-negative bacteria are commonly associated with PTB, and TLR4 recognizes LPS from the outer cell membrane of gram-negative microbes. LPS-induced PTB does not occur in TLR4 knockout mice. Furthermore, pretreatment with a TLR4 antagonist in the NHP model completely suppressed LPS-induced uterine activity and amniotic fluid cytokines and prostaglandins.
- ▪
Matrix metalloproteinases (MMPs): Elevated amniotic fluid MMP-8 is a powerful predictor of spontaneous PTB in humans. MMP-1 is also elevated in the placenta during labor. A nonspecific MMP inhibitor (GM6001; EMD Biosciences, San Diego, CA) significantly inhibited LPS-induced PTB in a murine model, suggesting that at least one (or more) MMP is critical for inducing labor in the setting of inflammation or infection. Endothelin-1 (ET-1) has recently been shown in the mouse to act in the same molecular pathway as MMP-1 for infection-associated PTB. Inhibition of ET-1 through endothelin receptor antagonists or RNA interference also inhibits PTB.
Microbiology
The cause of IAI is often polymicrobial, involving aerobic and anaerobic organisms (see Table 3-2 ). In a microbiologic controlled study, amniotic fluid cultures from women with and without IAI were compared. Patients with IAI were more likely to have 10 2 colony-forming units (CFU)/mL of any isolate, any number of high virulence isolates, and more than 10 2 CFU/mL of a high-virulence isolate (e.g., GBS, Escherichia coli, and enterococci). Although GBS and E. coli were isolated with only modest frequency (15% and 8%, respectively), if detected in the amniotic fluid, they are strongly associated with maternal or neonatal bacteremia (25% and 33%, respectively). The isolation of low-virulence organisms, such as lactobacilli, diphtheroids, and Staphylococcus epidermidis, was similar in the IAI and control groups. Common amniotic fluid isolates and odds ratios (ORs) for the association of specific bacterial species with PTB are shown in Table 3-3 . Several sexually transmitted infections also appear to be risk factors for PTB, including Neisseria gonorrhea , Chlamydia trachomatis , and Trichomonas vaginalis .
Table 3-3
Association of Lower Genitourinary Tract Infections with Preterm Birth
| Infection | Odds or Hazard Ratio for Preterm Birth (95% Confidence Interval) | Reference |
|---|---|---|
| Ureaplasma urealyticum | 1.0 (0.18-1.2) | |
| Chlamydia trachomatis | 2.42 ∗ (1.37-4.27) | |
| Neisseria gonorrhea | 1.77 ∗ (1.05-3.00) | |
| Trichomonas vaginalis | 1.59 ∗ (1.18-2.14) | |
| Bacterial vaginosis | 1.3 (1.1-1.4) | |
| Bacteriuria | 1.64 (1.35-1.78) |
∗ These hazard ratios were calculated for a “very preterm birth” defined as delivery before 33 weeks. Chlamydia, Trichomonas, and gonorrhea were also significantly associated with later preterm births.
Mycoplasma hominis and Ureaplasma species have been consistently recovered from amniotic fluid at a higher frequency than other bacteria but usually in association with other bacteria of known virulence. Some have questioned whether these bacteria cause PTB or may secondarily invade the amniotic cavity in the setting of another primary infection. Amniotic fluid inoculation of either M. hominis or Ureaplasma species induced PTB, chorioamnionitis, and fetal pneumonia in NHP. PTB occurred between 2.5 to 15 days after Ureaplasma parvum inoculation (10 7 CFU) and 17 to 30 days after M. hominis inoculation (10 5 -10 7 CFU). Genital mycoplasmas ( Ureaplasma urealyticum and M. hominis ) have also been associated with fetal and neonatal morbidity in human studies. In one study of neonates born between 23 and 32 weeks’ gestation, positive umbilical cord blood cultures for genital mycoplasmas were detected in 23%. Patients with spontaneous preterm delivery had a significantly higher rate of blood cultures positive for U. urealyticum or M. hominis or both than patients with indicated preterm delivery (34.7% vs. 3.2%; P < .0001) The earlier the gestational age at delivery, the more likely the culture was positive. In addition, newborns with a positive blood culture had a higher frequency of a neonatal systemic inflammatory response syndrome, higher serum concentrations of IL-6, and more frequent histologic evidence of placental inflammation than neonates with negative cultures. Bacteria within the amniotic fluid have also been identified by broad-range polymerase chain reaction (PCR) targeting the bacterial ribosomal unit 16S. This technique allows the identification of fastidious or uncultivatable microbes and has confirmed the heterogeneous and polymicrobial nature of amniotic fluid infection.
Maternal genital tract colonization with GBS may lead to neonatal sepsis, especially when birth occurs prematurely, or when the membranes have been ruptured for prolonged intervals. Several studies have found an association between GBS colonization of the lower genital tract and PTB or an earlier term birth. In many other studies, no association between GBS genital colonization and preterm labor or delivery was found. In contrast with the conflicting data regarding genital colonization with GBS, GBS bacteriuria has been consistently associated with preterm delivery, and treatment of this bacteriuria resulted in a marked reduction in prematurity (37.5% in the placebo group vs. 5.4% in the treatment group). In a randomized treatment trial of erythromycin versus placebo in women colonized with GBS, erythromycin use was not shown to be effective in prolonging gestation or increasing birth weight.
BV is a vaginal infection characterized by perturbations of the vaginal flora, high concentrations of certain bacterial species (e.g., G. vaginalis), and reductions in vaginal lactobacilli that has also been associated with IAI and PTB. The prevalence of BV varies by population, with rates of 8% in pregnant U.S. white women , 23% in pregnant U.S. black women, and 44% in nonpregnant Kenyan women. BV is associated with an increased PTB risk (1.4- to 3-fold or greater) thought to occur from an ascending vaginal infection resulting from loss of normal, healthy Lactobacillus flora and acquisition of a particularly pathogenic bacterial species. Another possibility is that BV acts locally in the lower genital tract to facilitate ascending infection by other organisms.
Bacteria associated with BV have been detected in amniotic fluid and placenta, suggesting that trafficking of vaginal microbes into the uterus occurs during pregnancy. Molecular techniques now enable identification of recognized bacterial species that are highly specific for BV but had not previously been detected using standard cultivation methods. These bacteria are fastidious or uncultivable and include species such as Leptotrichia amnionii that are detected using PCR for bacterial DNA encoding 16S rRNA. There is striking heterogeneity in the microbial communities within BV, which has led to the hypothesis that one or more specific bacterial species may be responsible for the increased PTB risk. In contrast, the vaginal microbiota from women with normal, healthy vaginal flora is typically fairly homogeneous with between one and six bacterial species in the vagina dominated by Lactobacillus.
BV may interact with genetic susceptibility and environmental factors to increase the risk of spontaneous PTB. Maternal carriers of a single nucleotide polymorphism in the TNF-α gene 2 (TNF2) were at significantly increased risk of spontaneous PTB (OR, 2.7; 95% confidence interval [CI], 1.7 to 4.5) in a case-control study. However, the association between TNF2 and PTB was modified by the presence of BV; mothers with a “susceptible” genotype and BV had an increased odds of PTB compared with mothers who did not. This association has not been confirmed at present but represents an interesting hypothesis for the interaction between a genetic predisposition and vaginal microbes in the susceptibility to PTB.
Viruses have also been detected in the amniotic fluid of asymptomatic women in the second trimester, suggesting that the hematogenous spread of viruses to the placenta occurs with a low frequency. Several studies have determined that between 2.2% and 8.4% of low-risk women with normal fetuses on ultrasonography in the second trimester have detectable genome sequences from at least one of eight viruses: human herpesvirus 6 (HHV6), adenoviruses, Epstein-Barr virus, varicella zoster virus, herpes simplex virus, parvovirus B19, human cytomegalovirus (HCMV), and enteroviruses. The most commonly detected virus in amniotic fluid in these studies was HHV6, recovered from 1.0% of patients. No significant changes in amniotic fluid white blood cell count, glucose, or IL-6 level were associated with viral invasion of the amniotic cavity. Of the inflammatory mediators tested, only CXCL 10 was associated with HCMV viral load. In general, however, viruses have not been linked to IAI, and fetal morbidity or mortality appears rare.
Diagnosis of Intraamniotic Infection
The diagnosis of IAI requires a high index of suspicion because the clinical signs and symptoms may be subtle and occur late in the course of the infection. Clinical features of IAI include maternal fever greater than or equal to 37.8° C, maternal leukocytosis greater than 15,000/mm 3 , maternal tachycardia greater than 100 beats per minute, fetal tachycardia greater than 160 beats per minute, uterine tenderness, and foul-smelling amniotic fluid. Maternal leukocytosis supports the diagnosis of clinical IAI, although recent administration of antenatal corticosteroids may cause a mild leukocytosis. The presence of a left shift (i.e., an increase in the proportion of neutrophils, especially immature forms), however, is particularly suggestive of clinical IAI. Malodorous amniotic fluid and uterine tenderness, although more specific for IAI, occur in a minority of cases. Other causes of fever in the parturient, such as epidural analgesia and concurrent infection of the urinary tract or other organ systems, must be considered.
Amniotic Fluid Testing
Direct examination of the amniotic fluid is desirable to confirm the diagnosis of IAI because clinical signs are inconsistent. Samples can be collected transabdominally by amniocentesis or transvaginally by aspiration of amniotic fluid through an intrauterine pressure catheter. Although amniotic fluid microbial cultures are important in ascertaining the etiology of the infection, they require 48 to 72 hours to provide results and thus cannot be used immediately to guide therapy. Adjunctive, rapid diagnostic tests, including Gram stain, amniotic fluid white cell count, and amniotic fluid glucose concentration, are therefore frequently used. In one study directly comparing these tests with amniotic fluid culture, amniotic fluid glucose level and Gram stain were the most specific for predicting a positive amniotic fluid culture ( Table 3-4 ). The combination of Gram stain with amniotic fluid glucose concentration is superior to any individual rapid adjunctive test currently widely available. The combination of Gram stain and an amniotic fluid glucose of less than or equal to 14 mg/dL had a sensitivity of 91% and a specificity of 81% when compared with amniotic fluid culture and those are the two most widely used adjunctive tests to diagnose IAI. Other diagnostic tools reported include measurement of amniotic fluid IL-6, MMP-8, or the detection of inflammatory mediators by proteomic analysis of amniotic fluid or cervicovaginal secretions. These tests, although sensitive and specific for the diagnosis of IAI, are not widely available or used at this time. Finally, detection of microorganisms by PCR, to detect DNA encoding 16S rRNA, may be useful for the rapid diagnosis of IAI in the future. The PCR assay has a higher sensitivity than culture for detection of microorganisms in the amniotic fluid, particularly in patients whose amniotic fluid is culture negative, but other markers indicate evidence of an inflammatory response.
Table 3-4
Diagnostic Value of Amniotic Fluid Testing in Detection of Positive Amniotic Fluid Culture in Patients with Preterm Labor and Intact Membranes
Modified from Romero R, Yoon BH, Mazor M, et al: The diagnostic and prognostic value of amniotic fluid white blood cell count, glucose, interleukin-6, and Gram stain in patients with preterm labor and intact membranes. Am J Obstet Gynecol 169:805-816, 1993.
| Diagnostic Test | Sensitivity | Specificity | Positive Predictive Value | Negative Predictive Value |
|---|---|---|---|---|
| Gram stain | 7/11 (64%) | 108/109 (99%) | 7/8 (88%) | 108/112 (96%) |
| IL-6 | 11/11 (100%) | 90/109 (83%) | 11/30 (37%) | 90/90 (100%) |
| WBC count | 7/11 (64%) | 103/109 (95%) | 7/13 (54%) | 103/107 (96%) |
| Glucose | 9/11 (82%) | 80/109 (82%) | 9/29 (31%) | 89/91 (98%) |
| Gram stain plus WBC count | 10/11 (91%) | 102/109 (94%) | 10/17 (59%) | 102/103 (99%) |
| Gram stain plus glucose | 10/11 (91%) | 88/109 (81%) | 10/31 (33%) | 88/89 (99%) |
| Gram stain plus glucose plus WBC count | 10/11 (91%) | 85/109 (78%) | 10/34 (29%) | 85/86 (99%) |
IL-6, Interleukin-6; WBC, white blood cell.
WBC ≥ 50 cells/mm 3 on Gram stain.
Glucose threshold ≤ 14 mg/dL.
Recently, the detection of amniotic fluid “sludge,” hyperechogenic material within the amniotic fluid identified by ultrasonography, has also been described as a diagnostic marker for microbial colonization of amniotic fluid and IAI. In a retrospective case-control study, patients with “sludge” had a significantly higher rate of spontaneous preterm delivery; a higher frequency of clinical chorioamnionitis, histologic chorioamnionitis, and funisitis; a higher frequency of PPROM; and a shorter median ultrasound-to-delivery interval. In addition, the combination of a cervical length less than 25 mm and “sludge” conferred a 15-fold increased odds for spontaneous preterm delivery at less than 28 weeks when compared with women with a normal cervical length and no amniotic fluid sludge (OR, 14.8; 95% CI, 3.9 to 56.5). By comparison, the increased risk of preterm delivery at less than 28 weeks was 6.8 (95% CI, 1.2 to 39.5) for those with a short cervix alone and 9.1 (95% CI, 2.9 to 28.6) for those with amniotic fluid sludge alone, respectively.
In the setting of premature rupture of the membranes (PROM), diagnosis of IAI is usually based on clinical criteria because oligohydramnios may preclude successful sampling of amniotic fluid. The incidence of microbial colonization and IAI are therefore likely underestimated. Risks of colonization and of IAI vary with gestational age and with labor. The rate of positive amniotic fluid cultures for microorganisms is higher with PPROM (32.4%) than with preterm labor with intact membranes (12.8%). In addition, among women with PROM the risk of microbial invasion of the amniotic after the onset of labor is as high as 75%.
Proteomics
Timely interventions to prevent neonatal sepsis in the setting of IAI would require early diagnosis, which is challenging because the clinical signs and symptoms tend to occur as late manifestations, and adjunctive laboratory tests have limited predictive value or require an invasive amniocentesis. Proteomics has been applied to investigate differentially expressed proteins in the amniotic fluid, cervicovaginal fluid, and maternal serum predictive of IAI ( Fig. 3-2 ). Multiple proteins in the amniotic fluid were differentially expressed in NHF as early as 12 hours after experimental IAI and before clinical signs or symptoms of infection, including azurocidin, calgranulin B, and a proteolytic fragment of insulin-like growth factor binding protein-1 (IGFBP-1). Validation of these biomarkers among a cohort of women in preterm labor with intact fetal membranes yielded a sensitivity of 100% and a specificity of 91%. A limitation of this approach is the requirement for an amniocentesis, which many providers are reluctant to perform in women with preterm labor. In the nonhuman primate, 27 proteins were differentially expressed in the cervicovaginal fluid after experimental U. parvum infection. In a cohort of 170 women in preterm labor with intact fetal membranes, 15 differentially expressed proteins in vaginal fluid were associated with subclinical IAI. A four-analyte immunoassay panel developed from these differentially expressed proteins was able to correctly classify 89% of patients as infected or not infected and included α1-acid glycoprotein, IGFBP-1, calgranulin C, and cystatin A. In maternal serum, proteomics identified a reduction in three peptides associated with the inter-α-trypsin inhibitor heavy chain–4 protein in women destined to have a PTB.
Treatment of Intraamniotic Infection
Management guidelines for the treatment of women in preterm labor and the use of antibiotics in preterm labor and PPROM are presented in Box 3-1 and Table 3-5 . Antibiotics have not been shown to improve neonatal outcomes or prolong pregnancy in women in preterm labor with intact membranes. However, antibiotic therapy is important to prevent postpartum endometritis and has been studied in the prevention of IAI. Progesterone therapy has been discovered to prevent PTB in a small subset of women, but it is unknown whether it prevents IAI and is, therefore, not discussed.
Box 3-1
Management Principles of Preterm Labor
Reproduced from American College of Obstetrics and Gynecology; Committee on Practice Bulletins—Obstetrics: ACOG practice bulletin no. 127: management of preterm labor, Obstet Gynecol 119:1308-1317, 2012.
The following recommendations and conclusions are based on good and consistent scientific evidence (Level A):
- ▪
A single course of corticosteroids to promote fetal lung maturation is recommended for pregnant women between 24 weeks of gestation and 34 weeks of gestation who are at risk of preterm delivery within 7 days.
- ▪
Accumulated available evidence suggests that magnesium sulfate reduces the severity and risk of cerebral palsy in surviving infants if administered when birth is anticipated before 32 weeks of gestation. Hospitals that elect to use magnesium sulfate for fetal neuroprotection should develop uniform and specific guidelines for their departments regarding inclusion criteria, treatment regimens, concurrent tocolysis, and monitoring in accordance with one of the larger trials.
- ▪
Evidence supports the use of first-line tocolytic treatment with β-adrenergic agonist therapy, calcium channel blockers, or nonsteroidal antiinflammatory drugs for short-term prolongation of pregnancy (up to 48 hours) to allow the administration of antenatal steroids.
- ▪
Maintenance therapy with tocolytics is ineffective for preventing preterm birth and improving neonatal outcomes and is not recommended for this purpose.
- ▪
Antibiotics should not be used to prolong gestation or improve neonatal outcomes in women with preterm labor and intact membranes.
The following recommendations and conclusions are based on limited and inconsistent scientific evidence (Level B):
- ▪
A single course of repeat antenatal corticosteroids should be considered in women whose prior course of antenatal corticosteroids was administered at least 7 days previously and who remain at risk of preterm birth before 34 weeks of gestation.
- ▪
Bed rest and hydration have not been shown to be effective for the prevention of preterm birth and should not be routinely recommended.
- ▪
The positive predictive value of a positive fetal fibronectin test result or a short cervix alone is poor and should not be used exclusively to direct management in the setting of acute symptoms.
Table 3-5
Suggested Guidelines on the Use of Antibiotics to Prevent Preterm Birth or for Treatment of Preterm Labor with Intact Membranes or Preterm Premature Rupture of Membranes
| Opinion | Comment |
|---|---|
| During Prenatal Care | |
| Treat Neisseria gonorrhoeae and Chlamydia trachomatis infection. | Screening and treatment of these two sexually transmitted organisms should follow standard recommendations to prevent spread to sexual partner(s) and the newborn. Published nonrandomized trials show improved pregnancy outcome with treatment. |
| Treat bacteriuria, including group B streptococcal bacteriuria. | Screening and treatment for bacteriuria is a standard practice to prevent pyelonephritis. A meta-analysis concluded that bacteriuria is directly associated with preterm birth. |
| Screen for and treat bacterial vaginosis in patients at high risk for preterm birth. In these high-risk women, treat with oral metronidazole for ≥1 week. | A meta-analysis has shown benefit with this treatment in women with high-risk pregnancies. |
| Treat symptomatic Trichomonas vaginalis infection to relieve maternal symptoms, but do not screen for or treat asymptomatic trichomoniasis. | This opinion is based on randomized trials in asymptomatic infected women. |
| Do not treat Ureaplasma urealyticum genital colonization. | One double-blind treatment trial that corrected for confounding infections showed no benefit. |
| Do not treat group B streptococcal genital colonization. | One double-blind treatment trial showed no benefit. |
| With Preterm Labor and Intact Membranes | |
| Give group B streptococcal prophylaxis to prevent neonatal sepsis. | As recommended by Centers for Disease Control and Prevention and American College of Obstetricians and Gynecologists. |
| Do not give antibiotics routinely to prolong pregnancy. | A meta-analysis concluded that antibiotics gave no neonatal benefit. |
| With Preterm Premature Rupture of the Membranes | |
| Give group B streptococcal prophylaxis to prevent neonatal sepsis. | As recommended by Centers for Disease Control and Prevention and American College of Obstetricians and Gynecologists. |
| Give additional antibiotics (ampicillin or amoxicillin plus erythromycin) in pregnancies at 24 to 34 weeks. | Meta-analyses concluded that there was substantial benefit to the neonate. Controversy exists as to whether there is benefit between 32 to 34 weeks. |
Antenatal Antibiotics to Prevent Intraamniotic Infection
Antibiotics given during prenatal care to patients at increased risk of preterm delivery have not shown benefit in preventing IAI. Early studies found a decrease in low-birth-weight infants delivered of women who received oral erythromycin for 6 weeks in the third trimester compared with placebo. A larger multicenter study of more than 1100 women with genital U. urealyticum found no change in any measured outcomes (gestational age at birth, low birth weight) when women were randomly assigned to receive placebo or erythromycin beginning at 26 to 30 weeks of gestation and continuing until 35 weeks. Treatment of U. urealyticum in pregnancy to prevent prematurity remains experimental.
Two retrospective, nonrandomized studies have reported reductions in preterm labor, PROM, and low birth weight through antenatal treatment of C. trachomatis infection. In the first study, patients successfully treated for C. trachomatis had significantly lower rates of PROM and premature labor than patients who failed to have C. trachomatis eradicated. In the second study, adverse outcome was assessed among three large groups: C. trachomatis –positive but untreated ( N = 1110), C. trachomatis –positive and treated ( N = 1327), and C. trachomatis –negative ( N = 9111). The C. trachomatis –positive but untreated group had higher rates of PROM, low birth weight, and perinatal mortality than the other two groups. The only randomized treatment trial for C. trachomatis in pregnancy led to conflicting results, however. In this latter study, the rate of pregnancies resulting in low-birth-weight infants was reduced in three of the five centers but not significantly reduced in the remaining two. A more recent large study found that treatment of C. trachomatis in midpregnancy was not associated with a decreased frequency of PTB. At the present time, it is the standard of care to treat women with C. trachomatis infection, not so much to prevent preterm labor but to prevent spread of the sexually transmitted disease.
Treatment with metronidazole should be offered to women who have symptomatic T. vaginalis infection in pregnancy, to relieve maternal symptoms and prevent spread of a sexually transmitted disease. Metronidazole is safe for use in the first trimester of pregnancy. However, in one study when pregnant women with asymptomatic T. vaginalis infection at 24 to 29 weeks of gestation were randomly assigned to receive either metronidazole or placebo, rates of PTB were increased in the group given metronidazole. Caution is therefore advised if treating asymptomatic trichomoniasis in the late second or third trimester.
To date, antibiotic treatment trials of BV have generally not reduced PTB despite overwhelming evidence that BV is consistently related to PTB, amniotic fluid infection, and chorioamnionitis. Several methodologic problems with prior trials include (1) BV treatment too late in pregnancy to prevent significant microbial trafficking into the uterus, (2) antibiotics tested were ineffective against many uncultivable pathogens, (3) failure to establish a Lactobacillus -dominant flora to prevent recurrence, and (4) lack of follow-up to document resolution. In Tanzanian women, a single course of metronidazole resulted in a shift from one BV-associated microbiota profile to another but rarely resulted in return of a Lactobacillus -dominant flora. The route of antibiotic administration may also be a key factor in resolution; a recent report suggests that oral metronidazole is effective at clearing BV-associated species, such as Leptotrichia and Sneathia, but intravaginal metronidazole is not. Vaginal microbial communities are also dynamic during pregnancy and may derive from an intestinal reservoir. With advancing gestation, there is an increasing representation of proteobacteria (gram-negative bacteria) in the third trimester, an overall reduction in species richness (number of taxa), and a reduction in within-subject (alpha) diversity. There may also be a unique window of PTB susceptibility in early pregnancy during which vaginal bacteria can colonize the endometrial cavity, which would favor earlier therapy of BV to prevent preterm labor.
An American College of Obstetricians and Gynecologists practice bulletin on assessment of risk factors for PTB has advocated that BV screening and treatment of high-risk or low-risk women would not be expected to reduce the overall rate of PTB. In certain populations of high-risk women, such as women with prior PTB and BV early in pregnancy, many experts still recommend treatment of BV diagnosed early in pregnancy.
Lack of consistent findings in antibiotic trials for BV to prevent PTB raises the question of why antibiotics have been effective in so few clinical situations. One explanation is that bacterial vaginosis is microbially very heterogenous, and only a subset of the women with bacterial vaginosis may harbor particularly pathogenic microbes for preterm labor. Second, treatment trials may have targeted therapy too late in pregnancy to be effective in preventing ascending infection that leads to PTB. Finally, antibiotics administered may not have been completely effective in eradicating fastidious bacteria.
For reasons other than prevention of PTB, detection and treatment of N. gonorrhoeae, C. trachomatis, and bacteriuria are appropriate. Future research is urgently needed, however, to identify markers in women who are in preterm labor as a result of infection, in whom intervention with antibiotics or other novel therapies is most likely to be beneficial. In addition, detection of women genetically predisposed to infection-induced PTB is important. Some investigators have identified associations between polymorphisms (single nucleotide polymorphisms [SNPs]) in the cytokine gene complexes, including TNF-α and PPROM or spontaneous PTB. Further research is needed, however, to determine if screening for these SNPs can inform or impact clinical management.
Prevention of Postpartum Endometritis After Intraamniotic Infection Diagnosis
Several principles guide the treatment of IAI. First, IAI is frequently a polymicrobial infection, involving both facultative and anaerobic microorganisms; broad-spectrum parenteral antibiotics are indicated mainly to eradicate the infection and prevent postpartum endometritis. Antibiotic therapy should be begun in the intrapartum period, as soon as the diagnosis is confirmed. Delivery, except under very unusual circumstances (e.g., listeriosis), is warranted by usual obstetric indications.
Antibiotic Choice
Several broad-spectrum antibiotics are appropriate for the treatment of IAI to prevent postpartum endometritis ( Table 3-6 ). A combination of parenteral ampicillin and an aminoglycoside, such as gentamicin, has traditionally been used because of excellent activity against common neonatal pathogens, including GBS and E. coli . However, these agents provide little anaerobic coverage. Therefore, clindamycin may be added to this regimen for patients with IAI who undergo cesarean delivery to prevent postpartum endometritis. Because dysfunctional labor leads to cesarean delivery in many patients with IAI, the consequence of this treatment protocol is that a high proportion of patients with IAI will ultimately require triple therapy, with its associated expense and potential toxicity. For most infections, single-agent therapy with a broad-spectrum antibiotic is equally efficacious and cost effective. Recommended intravenous regimens include cefotetan, cefoxitin, or penicillins combined with β-lactamase inhibitors, such as timentin/clavulanate or piperacillin/tazobactam. Because of the greater risk of postpartum endometritis after cesarean delivery, the chosen antibiotic regimen should be continued until the patient has been afebrile and symptoms resolved for at least 24 hours in those patients delivered by cesarean. For patients delivering vaginally, the duration of antibiotic therapy is arbitrary. Most patients defervesce promptly and more than one dose of antibiotics postpartum may be unnecessary.
Table 3-6
Antibiotic Regimens to Prevent Endometritis in Women with IAI
From Hopkins L, Smaill F: Antibiotic regimens for management of intraamniotic infection, Cochrane Database Syst Rev (3):CD003254, 2002.
| Antibiotic | Dose | Comment | Reference |
|---|---|---|---|
| Ampicillin + Gentamicin | Ampicillin: 2 g IV every 6 hours, plus gentamicin: 5.0 mg/kg IV once daily (normal renal function) | Fetal levels are threefold higher with once-daily versus standard (1.5 mg/kg) dosing of gentamicin
|
|
| Ampicillin-sulbactam | 3 g IV every 6 hours | ||
| Ticarcillin-clavulanate | 3.1 g (3 g ticarcillin + 100 mg clavulanate) IV every 4 hours | ||
| Cefoxitin | 2 g IV every 6 hours | ||
| Cefotetan | 2 g IV every 12 hours | ||
| Piperacillin-tazobactam | 3.375 g IV every 6 hours |
IAI, Intraamniotic infection; IV, intravenous.
Data does not support an ideal antimicrobial regimen for the treatment of IAI to prevent postpartum endometritis. Ampicillin and gentamicin are inexpensive and highly effective.
Among women in preterm labor with intact membranes, there have been several studies and meta-analyses studying the effect of various antibiotic regimens. The ORACLE II study showed no delay in delivery and no improvement in a composite outcome that included neonatal death, chronic lung disease, or cerebral anomaly. In the Cochrane meta-analysis, 7428 women in 11 trials were assessed. The relative risk (RR) for neonatal death in the antibiotic treatment group was 1.52 (95% CI, 0.99 to 2.34). There was a significant reduction in postpartum intrauterine infection with use of antibiotics, but this reduction was not seen as sufficient justification for widespread use of antibiotics in preterm labor. In a subanalysis, the reviewers looked at trials using antibiotics that were active against anaerobes (i.e., metronidazole or clindamycin). There were significant benefits in delivery within 7 days and in neonatal intensive care unit admissions. These benefits were not accompanied, however, by significant reductions in major end points, such as PTB, perinatal mortality, or neonatal sepsis.
In two studies with varying durations of antibiotic therapy after delivery complicated by IAI, there was no significant difference in rate of endometritis or postpartum fever. In one study, however, there was a 2.5-fold increase in wound infection rate in patients who did not receive scheduled postpartum antibiotics (5% vs. 1.8%). Based on these studies, it seems that when antibiotic treatment is initiated early, a short course of therapy in the puerperium is sufficient therapy in most patients.
More recent studies have addressed the issue of duration of antibiotic therapy postpartum in cases of IAI. One randomized trial compared single-dose versus multidose postpartum treatment of mothers and reported that single-dose treatment was accompanied by a shorter time to discharge (33 vs. 57 hours; P = .001). The single-dose group had a nearly threefold increase in failure of therapy, but this did not achieve statistical significance (11% vs. 3.7%; P = .27). Although not statistically significant, this threefold increase in “failed therapy” elicits concern regarding single-dose postpartum therapy for IAI.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here