Clinical chorioamnionitis at term: definition, pathogenesis, microbiology, diagnosis, and treatment

Abstract

Clinical chorioamnionitis, the most common infection-related diagnosis in labor and delivery units, is an antecedent of puerperal infection and neonatal sepsis. The condition is suspected when intrapartum fever is associated with two other maternal and fetal signs of local or systemic inflammation (eg, maternal tachycardia, uterine tenderness, maternal leukocytosis, malodorous vaginal discharge or amniotic fluid, and fetal tachycardia). Clinical chorioamnionitis is a syndrome caused by intraamniotic infection, sterile intraamniotic inflammation (inflammation without bacteria), or systemic maternal inflammation induced by epidural analgesia. In cases of uncertainty, a definitive diagnosis can be made by analyzing amniotic fluid with methods to detect bacteria (Gram stain, culture, or microbial nucleic acid) and inflammation (white blood cell count, glucose concentration, interleukin-6, interleukin-8, matrix metalloproteinase-8). The most common microorganisms are Ureaplasma species, and polymicrobial infections occur in 70% of cases. The fetal attack rate is low, and the rate of positive neonatal blood cultures ranges between 0.2% and 4%. Intrapartum antibiotic administration is the standard treatment to reduce neonatal sepsis. Treatment with ampicillin and gentamicin have been recommended by professional societies, although other antibiotic regimens, eg, cephalosporins, have been used. Given the importance of Ureaplasma species as a cause of intraamniotic infection, consideration needs to be given to the administration of antimicrobial agents effective against these microorganisms such as azithromycin or clarithromycin. We have used the combination of ceftriaxone, clarithromycin, and metronidazole, which has been shown to eradicate intraamniotic infection with microbiologic studies. Routine testing of neonates born to affected mothers for genital mycoplasmas can improve the detection of neonatal sepsis. Clinical chorioamnionitis is associated with decreased uterine activity, failure to progress in labor, and postpartum hemorrhage; however, clinical chorioamnionitis by itself is not an indication for cesarean delivery. Oxytocin is often administered for labor augmentation, and it is prudent to have uterotonic agents at hand to manage postpartum hemorrhage. Infants born to mothers with clinical chorioamnionitis near term are at risk for early-onset neonatal sepsis and for long-term disability such as cerebral palsy. A frontier is the noninvasive assessment of amniotic fluid to diagnose intraamniotic inflammation with a transcervical amniotic fluid collector and a rapid bedside test for IL-8 for patients with ruptured membranes. This approach promises to improve diagnostic accuracy and to provide a basis for antimicrobial administration.

Introduction

Clinical chorioamnionitis, a syndrome defined by the presence of maternal and fetal signs of local and systemic inflammation, is mainly caused by intraamniotic infection or by sterile intraamniotic inflammation.^{1, 2} This condition is the most common infection-related diagnosis in labor and delivery units, affecting 1% to 6% of term gestations.^3–10 Moreover, the disorder is an antecedent of puerperal infections such as endometritis,^11–13 pelvic peritonitis,¹⁴ wound infections,^{4, 13, 15} pelvic abscess,⁴ septic pelvic thrombophlebitis,^4,16 and necrotizing fasciitis.^17,18 Clinical chorioamnionitis is the leading cause of maternal ^19–22 and neonatal sepsis.^4,23–26 Patients with this diagnosis are at risk for uterine atony, postpartum hemorrhage, and other complications listed in Table 1.^4,9,10

Table 1.

Adverse maternal outcomes in relationship to the diagnosis of clinical chorioamnionitis at term

Adverse maternal outcomes	No clinical chorioamnionitis(N = 14,685)	Clinical chorioamnionitis(N = 1,965)	RR(95% CI)
Uterine atony	5.7%	14.2%	2.50 (2.20–2.84)
Blood transfusion	2.0%	4.5%	2.25 (1.78–2.83)
Pelvic abscess	0.1%	0.3%	3.74 (1.28–10.92)
Septic pelvic thrombophlebitis	0.2%	0.6%	2.74 (1.38–5.46)
Thromboembolic disease	0.1%	0.1%	1.49 (0.33–6.82)
Wound complication	1.1%	1.5%	1.39 (0.95–2.05)
Necrotizing fasciitis	0.0%	0.1%	2.49 (0.26–23.93)
ICU admission	0.5%	0.6%	1.03 (0.55–1.93)
Exploratory laparotomy	0.2%	0.3%	1.07 (0.42–2.72)
Hysterectomy	0.3%	0.4%	1.57 (0.74–3.37)
Death	0.1%	0.0%	–

CI: confidence interval; ICU: intensive care unit; RR: relative risk.

Adapted from Rouse DJ, Landon M, Leveno KJ, Leindecker S, Varner MW, Caritis SN, et al. The Maternal-Fetal Medicine Units cesarean registry: chorioamnionitis at term and its duration-relationship to outcomes. Am J Obstet Gynecol. 2004;191(1):211–6.⁴

The term “clinical” distinguishes this entity from “histologic” chorioamnionitis, which is diagnosed by pathologic examination of the placenta (Figure 1).²⁷ Often, the term “intraamniotic infection” is used interchangeably with “clinical chorioamnionitis”—this is unfortunate because they are not synonymous. The designation of intraamniotic infection should be reserved for cases in which there is laboratory evidence of microorganisms in the amniotic fluid coupled with intraamniotic inflammation. This article reviews advances in the understanding of clinical chorioamnionitis at term and its association with intraamniotic infection, focusing on the definition, pathogenesis, microbiology, inflammatory response, diagnosis, and treatment.

Figure 1. Definitions of chorioamnionitis: clinical, histologic, and microbiologic.

Chorioamnionitis can be defined as a clinical or a histopathologic identity or from a microbiologic point of view: (1) the clinical diagnosis of chorioamnionitis is based on clinical signs, eg, maternal fever, uterine tenderness, malodorous discharge, and maternal and fetal tachycardia, and on laboratory abnormalities, ie, leukocytosis; (2) histologic chorioamnionitis is diagnosed by histopathologic examination of the placenta; the characteristic morphologic feature of acute histologic chorioamnionitis is diffuse infiltration of neutrophils into the chorioamniotic membranes; and (3) microbiologic chorioamnionitis, or intraamniotic infection, refers to the presence of microorganisms in the amniotic fluid, retrieved by amniocentesis. These terms are not synonymous, and confusion is caused when they are used interchangeably.

The amniotic cavity does not contain bacteria in normal pregnancy

Under normal circumstances, the fetus lives in an environment devoid of bacteria^{28, 29} Therefore, birth represents a critical stage for the acquisition of the first microbiota (the collection of living microorganisms present in a defined site). During the passage through the birth canal, the fetus is exposed to the vaginal microbiota,^30–32 and these bacteria become the pioneer microorganisms that invade the formerly sterile body of the infant,³³ thus establishing the first neonatal microbiota.^34,35 This microbiota differs according to the mode of delivery (vaginal vs cesarean delivery)^36–38 although mother-to-infant microbial transmission routes develop across several body sites after birth.^39,40

Microbial invasion of the amniotic cavity (MIAC) by bacteria, fungi, and viruses can predispose to fetal infection. Microorganisms can enter the fetus through different sites, ie, the mucous membranes of the airway, gastrointestinal tract, tympanic membrane, conjunctiva, or skin. Once bacteria are in contact with the mucous membranes, microorganisms can gain access to the fetal circulation and cause bacteremia, which can lead to sepsis, septic shock, and fetal death.⁴¹ The fetal immune system is capable of mounting innate and adaptive immune responses, which are key in host defense against infection.^42–47 Indeed, recent studies show that the amniotic fluid harbors cells of the innate immune system, largely of fetal origin, that participate in host defense against microbes invading the amniotic cavity.^48–56

Nomenclature of conditions related to the presence or absence of microorganisms and intraamniotic inflammation

Microbial invasion of the amniotic cavity is defined as the presence of microorganisms in amniotic fluid retrieved by transabdominal amniocentesis or at the time of cesarean delivery and detected by cultivation or molecular microbiologic techniques (Figure 2). The term MIAC indicates the presence of microorganisms regardless of the presence or absence of an inflammatory response. Intraamniotic inflammation refers to the presence of an inflammatory response in the amniotic cavity, which can be detected by the presence of inflammatory cells such as neutrophils^57–61 or by an elevated concentration of an inflammatory mediator, eg, interleukin (IL)-6^{58−60, 62–69} or matrix metalloproteinase (MMP)-8.⁷⁰ A wide range of inflammatory products can be detected in biological fluids, and their concentrations can be used for diagnostic purposes.. The most frequently employed analytes in the diagnosis of intraamniotic inflammation are cytokines (IL-6^{58−60,62–69,71} and IL-8⁷²⁻⁷⁴), acute-phase reactants (C-reactive protein),^61,75–79 or other compounds eg, MMP-8 (also known as neutrophil collagenase).^70,80–82

Figure 2. Nomenclature of states related to intraamniotic inflammation and intraamniotic infection.

Microorganisms in amniotic fluid can be detected by cultivation methods and/or by molecular microbiologic techniques. Intraamniotic inflammation is defined by the presence of inflammatory cells (white blood cell count ≥ 50 cells/mm³) or by an elevated concentration of a biomarker of inflammation, eg, an interleukin-6 concentration ≥2.6 ng/mL or a matrix metalloproteinase-8 concentration >23 ng/mL. On the basis of the results of the presence of microorganisms in amniotic fluid and intraamniotic inflammation, patients can be classified into four subgroups; (1) no intraamniotic infection (negative amniotic fluid by culture and PCR and the absence of intra-amniotic inflammation); (2) microorganisms in amniotic fluid without intraamniotic inflammation (positive amniotic fluid by either culture or PCR but the absence of intraamniotic inflammation); (3) sterile intraamniotic inflammation (negative amniotic fluid by culture and PCR but the presence of intraamniotic inflammation); and (4) intraamniotic infection (positive amniotic fluid by either culture or PCR and the presence of intraamniotic inflammation).

PCR: polymerase chain reaction.

The term intraamniotic infection refers to the combination of MIAC and intraamniotic inflammation (Figure 2). Most cases in which microorganisms are present in amniotic fluid without evidence of inflammation are caused by specimen contamination with skin flora or with bacteria in the laboratory or in reagents.⁸³ When intraamniotic inflammation is present in the absence of microorganisms, the condition is referred to as sterile intraamniotic inflammation.^{1, 2,84–87} This term should ideally be reserved for samples that are negative for microorganisms, ascertained by cultivation and molecular microbiologic techniques.^1,2,84–86

Clinical chorioamnionitis refers to a syndrome characterized by the presence of a maternal fever (temperature ≥37.8°C or ≥38.0°C) and 2 or more of these 5 clinical signs: (1) maternal tachycardia (>100 beats/min); (2) fetal tachycardia (>160 beats/min); (3) uterine tenderness; (4) purulent or malodorous amniotic fluid or vaginal discharge; and (5) maternal leukocytosis (white blood cell count >15,000/mm³).^20,88–95

Acute histologic chorioamnionitis indicates the presence of neutrophils in the chorioamniotic membranes or in the chorionic plate and represents a maternal host response,^27,96–106 whereas funisitis and chorionic vasculitis are evidence of a fetal response (Figure 3). Funisitis is defined as inflammation of the umbilical cord (umbilical vein, artery, and Wharton’s jelly).^107,108 Chorionic vasculitis consists of inflammation of the fetal vessels on the surface of the chorionic plate. Acute histologic chorioamnionitis is the pathologic expression of intraamniotic inflammation (Figure 3).^{97,99,109−115}Acute histologic chorioamnionitis is not synonymous with intraamniotic infection. Only a fraction of patients diagnosed with histologic chorioamnionitis has proven intraamniotic infection. The rest have sterile intraamniotic inflammation, which may be accompanied by funisitis. The term acute chorioamnionitis refers to a specific inflammatory lesion in which neutrophils are the predominant cell type. By contrast, chronic chorioamnionitis refers to inflammatory lesions characterized by the infiltration of lymphocytes. Chronic inflammatory lesions occur when specific infectious agents, eg, viruses, are present, yet the main cause of chronic inflammatory lesions is maternal anti-fetal rejection.^116–123 Interested readers are referred to a recent review by the authors on this subject.¹²³

Figure 3. Anatomy of the placental disc and chorioamniotic membranes and acute inflammatory lesions of the placenta: chorioamnionitis, chorionic vasculitis, and funisitis.

A. The placenta is composed of 3 major structures: the chorioamniotic membranes, the placental disc, and the umbilical cord. Acute inflammatory lesions of the placenta are characterized by the infiltration of neutrophils in any of these structures. Specifically, when the inflammatory process affects the chorion and amnion, the term is acute chorioamnionitis. If the inflammatory process involves the umbilical cord (umbilical vein, umbilical artery, and the Wharton’s jelly), this is referred to as acute funisitis, the histologic counterpart of the fetal inflammatory response syndrome. B. Acute chorioamnionitis (stage 2 acute inflammation of the chorioamniotic membranes): neutrophilic migration into the amniotic connective tissue is shown (asterisk). C. Chorionic vasculitis is inflammation on the surface of the fetal vessels within the chorionic plate. Acute chorionic vasculitis (asterisk) is a stage 1 fetal inflammatory response. D. Funisitis: its characteristic feature is concentric, perivascular distribution of degenerated neutrophils (asterisk). Modified from Kim CJ et al. Acute chorioamnionitis and funisitis: definition, pathologic features, and clinical significance. Am J Obstet Gynecol. 2015;213(4 Suppl):S29–52.²⁷

A panel proposed to replace the term “clinical chorioamnionitis” with the term “intrauterine inflammation or infection or both,” abbreviated as “Triple I.”¹²⁴ However, this proposal has not been adopted because it implies that the inflammatory status of the amniotic cavity and the presence of microorganisms can be determined only by clinical signs or symptoms. There is solid evidence that this is not the case.^125–127 Indeed, a subset of patients that meets the clinical criteria of chorioamnionitis does not show evidence of intraamniotic infection or intraamniotic inflammation.^{1, 2} Therefore, we favor the use of the term “clinical chorioamnionitis” to refer to the clinical syndrome before the analysis of amniotic fluid. Examination of amniotic fluid allows ascertainment of whether intraamniotic inflammation is present in patients with a fever and likely etiology, ie, microbial or sterile.

The definition of clinical chorioamnionitis according to the reVITALize Initiative (a USA multidisciplinary collaboration formed to standardize obstetrical data definitions) includes unexplained fever (>38°C or 100.4°F) with one or more of the following symptoms or signs: (1) uterine tenderness and irritability, (2) leukocytosis, (3) fetal tachycardia, (4) maternal tachycardia, and (5) malodorous vaginal discharge.¹²⁸ Amniotic fluid analysis in patients with intact membranes requires an amniocentesis. If the membranes are ruptured, a transcervical amniotic fluid collector and an assessment of inflammatory products, eg, IL-8, makes possible the diagnosis of intraamniotic inflammation (for details, see the section on diagnosis).^{129, 130} Amniotic fluid could also be retrieved for microbiologic studies at the time of cesarean delivery.

Clinical chorioamnionitis as a syndrome

Patients diagnosed with clinical chorioamnionitis at term, can be categorized into 3 groups according to the results of amniotic fluid analysis of bacteria and inflammation: (1) intraamniotic infection, ie, the presence of microorganisms and intraamniotic inflammation, (2) sterile intraamniotic inflammation, and (3) maternal signs or symptoms of systemic inflammation without evidence of intraamniotic inflammation or intraamniotic infection.^{1, 2} The prevalence of the three conditions relates to the status of membrane rupture and the administration of an epidural.^{1, 2} The relative frequency of each of the three conditions is intraamniotic infection, 65%; sterile intraamniotic inflammation, 15%; and absence of intraamniotic inflammation and/or intraamniotic infection, 20% (Figure 4).^{1, 2} The etiology of maternal systemic inflammation in the latter is unclear and has been attributed to a neuroinflammatory process after the administration of epidural analgesia.

Figure 4. The syndrome of chorioamnionitis at term: intraamniotic infection and sterile intraamniotic inflammation.

Patients with clinical chorioamnionitis at term, when studied with molecular microbiologic techniques and assessed for intraamniotic inflammation, are grouped into 3 categories: (1) intraamniotic infection (presence of microorganisms and intraamniotic inflammation); (2) sterile intraamniotic inflammation; and (3) maternal signs or symptoms of systemic inflammation without evidence of intraamniotic infection. The relative frequencies of these 3 conditions were calculated as 65% (58/89) in intraamniotic infection; 15% (13/89) in intraamniotic inflammation; and 20% (18/89) for no evidence of intraamniotic infection.

Pathophysiology of intraamniotic infection

Microorganisms may gain access to the amniotic cavity through three main pathways^131–133: (1) ascending from the vagina and cervix^{109, 134–137}; (2) hematogenous dissemination through the placenta (transplacental infection)^138–145; and (3) accidental introduction at the time of invasive procedures such as amniocentesis, percutaneous fetal blood sampling, chorionic villus sampling, or shunting.^146–152

The most common pathway of intraamniotic infection is the ascending route^{131, 132, 137} based on the following evidence: (1) the nucleotide sequences of microorganisms grown in culture from amniotic fluid samples of patients with intraamniotic infection are similar to those from the vagina in the same patients¹³⁷; (2) in twin gestations, acute histologic chorioamnionitis, more common in the first-born twin, is rarely observed in the second twin. Moreover, when intraamniotic infection is detected, the presenting sac is nearly always involved¹⁵³; and (3) in virtually all cases of congenital pneumonia, inflammation of the chorioamniotic membranes is present.^{138, 154}

Risk factors for clinical chorioamnionitis

Ruptured fetal membranes

The chorioamniotic membranes represent a physical and a biochemical barrier to microorganisms, as the fetal membranes produce antimicrobial peptides^155–157 and contain cells of the innate immune system capable of protecting the host against bacteria.^158–161 The frequency of MIAC in patients with ruptured membranes at term is 34%,¹⁶² which is higher than in those at term without labor (1%)¹⁶³ or in those with spontaneous labor with intact membranes (19%) (Figure 5).¹⁶³ In patients with clinical chorioamnionitis at term, the frequency of intraamniotic infection is higher when the membranes are ruptured than when they are intact [ruptured membranes, 78% vs intact membranes, 38%; p=0.01 (Figure 6)].² Rupture of the membranes increases the risk of a postpartum uterine infection, eg, endometritis, endomyometritis, and endoparametritis.^{14, 164} The increased risk for postpartum endometritis conferred by rupture of the membranes has been documented in studies performed before the routine use of antimicrobial prophylaxis for cesarean delivery.¹⁶⁵ In patients whose membranes had been ruptured for longer than six hours, bacteria were detected in all cases for which amniotic fluid had been obtained at the time of cesarean delivery; 95% of patients developed postpartum myometritis.¹⁶⁵

Figure 5. Frequency of microorganisms in the amniotic cavity of patients at term without labor, patients in labor with intact membranes, and those with prelabor rupture of membranes at term.

In the presence of ruptured membranes, 34% of patients at term have microorganisms in amniotic fluid, which is higher compared to those without labor (1%) and to those with spontaneous labor with intact membranes (19%). Patients at term, not in labor, and with intact membranes rarely have bacteria detected in the amniotic cavity (left side of the figure, with a frequency of 1%). Patients in labor at term with intact membranes have a higher frequency of bacteria in the amniotic cavity even if the membranes are intact (middle part of the figure, with a frequency of 19%). Rupture of the chorioamniotic membranes at term prior to the onset of labor has a higher frequency of bacteria in the amniotic cavity (right side of the figure and table, with a frequency of approximately 34%). This frequency is similar to that observed in patients with preterm PROM.

PROM: premature rupture of the membranes.

Figure 6. Frequency of intraamniotic infection and sterile intraamniotic inflammation in patients with clinical chorioamnionitis at term according to membrane status.

Prevalence of intraamniotic infection and sterile intraamniotic inflammation in patients with clinical chorioamnionitis at term according to the status of the membranes at the time of amniocentesis. Modified from Romero R et al. Clinical chorioamnionitis at term X: microbiology, clinical signs, placental pathology, and neonatal bacteremia—implications for clinical care. J Perinat Med. 2021;49(3):275–98.²

Prolonged labor

Uterine contractions increase the risk of microbial invasion of the amniotic cavity. An experimental study indicates that uterine contractions exert a “suction-like effect” whereby vaginal fluid ascends into the uterine cavity (demonstrated by hysterosalpingoscintigraphy and radio-labeled carbon particles).¹⁶⁶ Similar observations have been reported in animals after the placement of fluorescent bacteria in the vagina (unpublished observations). The frequency of microorganisms in the amniotic fluid of patients with spontaneous labor at term with intact membranes is 19%, which is higher than 1% observed in patients at term without labor¹⁶⁷ (Figure 5). Seong et al ¹⁶³ reported that the frequency of microorganisms in amniotic fluid at term increased over time during labor—3.5% during early labor (cervical dilatation <4 cm)—whereas the frequency increased 4-fold (13%) when patients were in active labor (cervical dilatation ≥ 4cm),¹⁶³ suggesting that the longer the labor and the greater the cervical dilatation, the higher the risk of microorganisms in the amniotic cavity.

Digital examination

Digital examination of the uterine cervix before or during labor has been implicated in the genesis of intraamniotic infection as microorganisms may be carried from the vaginal ecosystem into the lower pole of the chorioamniotic sac. This has led to the view that digital examination should not be performed in patients with preterm prelabor rupture of the membranes (PROM) or even PROM at term because of the adage that “once an examination has been performed, the clock of infection starts to tick”.¹⁶⁸ Several studies have addressed the clinical consequences of digital examination of the cervix during labor.^169–176 Figure 7 illustrates that the risk of clinical chorioamnionitis at term increases with each additional cervical examination.¹⁷⁶ Moreover, the risk of clinical chorioamnionitis is greater when the number of digital examinations increases in the setting of ruptured membranes during the intrapartum period.^{170, 176}

Figure 7. Probability of clinical chorioamnionitis by the number of digital cervical examinations and by the number of hours after rupture of membranes.

Each additional cervical examination confers an incremental risk of clinical chorioamnionitis. A greater length of time from rupture to delivery is associated with an increased risk of clinical chorioamnionitis. Modified from Gomez Slagle HB et al. Incremental risk of clinical chorioamnionitis associated with cervical examination. Am J Obstet Gynecol MFM. 2022;4(1):100524.¹⁷⁶

History of clinical chorioamnionitis in a previous pregnancy

A history of clinical chorioamnionitis increases the odds of recurrence in a subsequent pregnancy approximately by 2- to 3-fold compared to patients who did not have this diagnosis in a previous pregnancy.^177–179 The frequency of recurrent clinical chorioamnionitis is the highest in patients with clinical chorioamnionitis in the first pregnancy who delivered at 20 to 24 weeks of gestation. The most likely explanation of this is that the rate of intraamniotic infection is extremely high the lower the gestational age at birth.

Microbiology of intraamniotic infection in clinical chorioamnionitis at term

Microorganisms responsible for intraamniotic infection are shown in Table 2. Polymicrobial infections are detected in 70% of patients with intraamniotic infection and clinical chorioamnionitis.^1,2 The microorganisms most commonly found in the amniotic fluid are Ureaplasma species (U. parvum or U. urealyticum), Gardnerella vaginalis, Mycoplasma hominis, Streptococcus agalactiae, Staphylococcus aureus, and Bacteroides spp.^{1,2,20,180–183} Sneathia spp. can be readily identified by the use of molecular microbiologic techniques. Some microorganisms are typical inhabitants of the mouth, eg, Porphyromonas spp., Fusobacterium spp., and Streptococcus oralis.^1,2 Fungi (Candida spp.) can cause intraamniotic infection, particularly in patients who become pregnant while using an intrauterine device.¹⁸⁴ Viruses such as herpes simplex virus-1 and roseolovirus have been identified in the amniotic fluid, but it is unclear whether they have an etiologic role. To date, Chlamydia spp. has not been detected in the amniotic cavity of patients diagnosed with clinical chorioamnionitis at term; this is consistent with the observation that Chlamydia trachomatis is rarely detected in patients diagnosed with preterm PROM.¹⁸⁵

Table 2.

Microorganisms identified in the amniotic fluid of patients with clinical chorioamnionitis at term

Ureaplasma species (Ureaplasma urealyticum, Ureaplasma parvum)

Gardnerella vaginalis

Mycoplasma hominis

Streptococcus agalactiae

Lactobacillus species

Bacteroides species

Fusobacterium species (Fusobacterium nucleatum, Fusobacterium varium)

Sneathia species

Streptococcus viridans

Porphyromonas species

Veillonella species

Peptostreptococcus

Escherichia coli

Pseudomonas aeruginosa

Staphylococcus aureus

Candida species

Enterococcus faecalis

Acinetobacter species

Others: Candida albicans, Staphylococcus epidermidis, Propionibacterium acnes

Detected with the use of cultivation and molecular microbiologic techniques in the amniotic fluid of patients with clinical chorioamnionitis at term.

Microbial burden has been studied with molecular microbiologic techniques, and the range of microbial load is broad (6–978 genome equivalents per PCR well).² A consistent finding in our studies is that the greater the microbial burden, the stronger the intensity of the intraamniotic inflammatory response, as reflected by the concentration of IL-6 in amniotic fluid.^1,2

Genital mycoplasmas as a cause of intraamniotic infection, maternal infection, and neonatal sepsis

Genital mycoplasmas (U. urealyticum, U. parvum, M. hominis) are the most frequently detected microorganisms in the amniotic cavity in a wide range of obstetrical complications, including clinical chorioamnionitis at term.^{1, 2, 68, 162, 186–197} These organisms have been implicated in postpartum infections, such as endometritis, and in wound infections in nearly 62% of cases.¹⁹⁸ For example, Andrews et al¹⁹⁹ reported that patients who had U. urealyticum in the chorioamniotic membranes had a 3-fold increased risk of endometritis. In another study, genital mycoplasmas were detected in 30% of endometrial samples obtained from patients with puerperal endometritis.²⁰⁰ Although some investigators have proposed that genital mycoplasmas are of low virulence and may not be pathogenic, our studies have clearly shown that the presence of genital mycoplasmas in amniotic fluid is associated with intraamniotic, maternal, and fetal inflammatory responses,^196,201,202 and, in some cases, that the magnitude of the inflammatory process is more severe than that associated with other bacteria.¹⁹⁶ Experimental evidence indicates that the incubation of genital mycoplasmas with chorioamniotic membranes elicits the release of inflammatory mediators, eg, tumor necrosis factor (TNF)-α, into the culture medium.²⁰³Importantly, intraamniotic administration of Ureaplasma spp. leads to the onset of labor and fetal infection in nonhuman primates and mice.^204–208

The role of genital mycoplasmas in neonatal disease has also been subject to intensive investigation. There is good evidence that genital mycoplasmas can be isolated in cases congenital pneumonia,^{154, 209–211} neonatal meningitis,^212–220 and sepsis.^{154, 210, 211} Such findings are not unexpected, given that 23% of preterm neonates born between 23 and 33 weeks of gestation have demonstrable bacteremia with genital mycoplasmas at the time of birth when appropriate cultures for these organisms are obtained from umbilical cord blood.^{221, 222} Therefore, the role of genital mycoplasmas in neonatal infections is well supported despite initial skepticism. Regrettably, these organisms are not routinely cultured, in obstetrics or in neonatal medicine, nor are they effectively treated by standard antimicrobial regimens (ampicillin and gentamicin) in neonatal intensive care units (NICU) or obstetrical units.^223–225 Key issues that need to be addressed include the rapid and sensitive diagnosis of infections caused by Ureaplasma spp. in neonates and the modification of the current antimicrobial therapy in the neonatal period. Antibiotics effective against genital mycoplasmas should be considered.

The inflammatory response in the amniotic, maternal, and fetal compartments in clinical chorioamnionitis at term

Microbial or non-microbial stimuli can induce an inflammatory response in 1) the amniotic cavity; 2) the maternal compartment; or 3) the fetal compartment. Typically, in cases of ascending infection from the lower genital tract, the inflammatory process begins in the amniotic cavity and can extend into the fetus and/or the mother. In cases of hematogenous infection, such as maternal listeriosis, the mother is affected first and the fetus and the amniotic compartment secondarily.^{139, 140, 142–145} The frequency of relative involvement of the amniotic, maternal, and fetal compartments has not been subject to rigorous investigation. Clinical signs of infection are unreliable indicators of involvement; most infections in neonates are subclinical and the same is the case for mothers. The advent of new molecular microbiologic techniques and high-dimensional methods to assess inflammation can be leveraged to ascertain the rate of maternal and fetal involvement.

Inflammation has cellular and soluble components. The typical changes in the soluble mediators are those that occur in the inflammatory cytokine network, which may extend to other inflammatory products such as prostaglandins.^226–235 The cellular components involve mainly those of the innate immune system such as neutrophils and macrophages.²³⁶

The intraamniotic inflammatory response

Microbial products or alarmins engage the innate immune system to induce an intraamniotic inflammatory response. Neutrophils are the most common subset of leukocytes in the amniotic fluid (Figure 8), followed by monocytes and macrophages. Other immune cells such as T and B lymphocytes, as well as natural killer cells, are less frequent in the amniotic fluid.^{48, 183} The absolute count of neutrophils, monocytes, and macrophages in amniotic fluid is significantly higher in patients with intraamniotic infection than in patients with sterile intraamniotic inflammation.^{48, 52} These cells in amniotic fluid can be of maternal, fetal, or mixed origin, as demonstrated by DNA fingerprinting.^{49, 52} The earlier the gestational age, the more likely they are predominantly of fetal origin (capable of inducing a fever). The closer to term, the more likely that neutrophils, monocytes, and macrophages would be of maternal origin.^{49, 52} Neutrophils produce TNF-α and macrophage inflammatory protein (MIP)-1β predominantly, whereas monocytes and macrophages express mainly IL-1β.⁴⁸ These cytokines are implicated both in the process of parturition and in the host response against intraamniotic infection.^{73, 237–251} The amniotic fluid concentrations of pro-inflammatory cytokines, eg, IL-1β, interferon (IFN)-γ, TNF-α, TNF-β, IL-2, and IL-6, are significantly higher in patients with clinical chorioamnionitis at term than in those with spontaneous labor at term without clinical chorioamnionitis (Figure 9A).²⁵¹ Among patients with intraamniotic inflammation, the changes in the amniotic fluid cytokine concentrations are similar in magnitude between those with intraamniotic infection and those with sterile intraamniotic inflammation (Figure 9A).²⁵¹

Figure 8. Amniotic fluid neutrophils and monocytes in patients with clinical chorioamnionitis at term.

Hematoxylin and eosin staining shows the typical morphology of neutrophils (red arrow) and monocytes (green arrow) in the amniotic fluid of patients with clinical chorioamnionitis. Magnification 400X. Scale bars: 50 μm. Modified from Martinez-Varea A. et al. Clinical chorioamnionitis at term VII: the amniotic fluid cellular immune response. J Perinat Med. 2017;45(5):523–38.⁴⁸

Figure 9. The amniotic, maternal, and fetal inflammatory responses in patients with clinical chorioamnionitis at term.

(A) Amniotic fluid concentrations of inflammatory cytokines (IL-1β, IFN-γ, TNF-α, and TNF-β) are higher in patients with clinical chorioamnionitis at term than in those with spontaneous labor at term. Modified from Romero R et al. Clinical chorioamnionitis at term II: the intra-amniotic inflammatory response. J Perinat Med. 2016;44(1):5–22.²⁵¹

(B) Maternal plasma concentrations of pyrogenic cytokines (IL-2, IL-6, IL-1β, and IL-17α) are higher in patients with clinical chorioamnionitis at term than in those with spontaneous labor at term. Modified from Romero R et al. Clinical chorioamnionitis at term IV: the maternal plasma cytokine profile. J Perinat Med. 2016;44(1):77–98.²⁵²

(C) Umbilical cord plasma concentrations of inflammatory cytokines and chemokines (IL-12p70, IL-6, IL-16, and IL-8) are higher in fetuses with clinical chorioamnionitis at term than in those delivered by women with spontaneous labor at term. Modified from Romero R et al. Clinical chorioamnionitis at term V: umbilical cord plasma cytokine profile in the context of a systemic maternal inflammatory response. J Perinat Med. 2016;44(1):53–76.²⁷⁶

* p < 0.05

IFN: interferon; IL: interleukin; TNF: tumor necrosis factor

Maternal inflammatory response

Microbial products, cytokines, or alarmins can gain access to the maternal compartment and then induce a local inflammatory response, such as deciduitis, followed by a systemic response that can be detected by changes in the concentrations of inflammatory mediators in maternal blood or by clinical signs of inflammation. Indeed, most of the signs of clinical chorioamnionitis (except fetal tachycardia) are thought to reflect a maternal inflammatory response.⁹² Maternal plasma concentrations of pyrogenic cytokines such as IL-2, IL-6, IL-1β, and IL-17A are higher in patients with clinical chorioamnionitis at term than in those with spontaneous labor at term without clinical chorioamnionitis (Figure 9B).²⁵² In addition, patients with clinical chorioamnionitis at term without intraamniotic infection have a greater maternal plasma concentration of pyrogenic cytokines than those with spontaneous labor at term without this clinical condition (Figure 9B).²⁵² However, maternal plasma concentrations of cytokines are not significantly different among the 3 major subgroups of patients with clinical chorioamnionitis at term (Figure 9B).²⁵² These findings indicate that the absolute concentration of cytokines in the maternal plasma cannot be used to identify patients who have intraamniotic infection.²⁵²

The fetal inflammatory response

When microbial and non-microbial stimuli reach the human fetus, a local and then a systemic inflammatory response are elicited. The pattern of engagement of the fetal immune system depends upon the pathway of infection. For example, in cases of ascending infection, microorganisms can reach the fetal skin and tympanic membranes or can be aspirated into the respiratory tract (explaining the occurrence of fetal dermatitis, otitis media, pneumonia, etc., in cases of intraamniotic infection). If the infection is hematogenous, microorganisms can elicit a fetal systemic inflammatory response by engaging the fetal endothelium during the course of bacteremia. This systemic inflammatory response is a precursor of fetal and neonatal sepsis. Indeed, there is a dose–response relationship between the concentrations of cytokines, such as TNF-α, and early neonatal sepsis.²⁵³

Neonates born to mothers diagnosed with clinical chorioamnionitis are at risk for sepsis,^{4, 23–26, 254–259} meconium aspiration syndrome,^{254, 260–264} neonatal encephalopathy,^265–268 and long-term neurodevelopmental disabilities, including cognitive impairment,^269–273 cerebral palsy,²⁷² and neonatal death.^{6, 10, 112, 259, 274, 275} The concentration of inflammatory cytokines and chemokines (IL-12p70, IL-6, IL-16, and IL-8) in the umbilical cord is higher in fetuses born to mothers diagnosed with clinical chorioamnionitis at term than in those whose mothers underwent spontaneous labor at term without this clinical condition (Figure 9C).²⁷⁶ The fetal inflammatory response syndrome (FIRS; defined as an elevated fetal plasma IL-6 concentration ≥11pg/ml) is present in fetuses with clinical chorioamnionitis (Figure 10).²⁷⁷ Of note, the fetuses affected with FIRS generally are born to mothers with proven intraamniotic infection,²⁷⁶ a finding reflecting the severity of the fetal inflammatory response tends to be greater in such cases.²⁷⁶ Interestingly, the fetuses of mothers with clinical chorioamnionitis at term without intraamniotic infection or sterile intraamniotic inflammation also have a higher concentration of inflammatory cytokines in umbilical cord blood than those without clinical chorioamnionitis, suggesting that maternal fever in the absence of intraamniotic infection leads to a change in the fetal cytokine network (Figure 9C).²⁷⁶ Furthermore, there are significant positive correlations between maternal and umbilical cord plasma IL-6 and IL-8 concentrations; these observations are consistent with the possibility of placental transfer of cytokines.^{276, 278}

Figure 10. Fetal inflammatory response syndrome in clinical chorioamnionitis.

A. Fetal inflammatory response syndrome can be diagnosed by an increased concentration of umbilical cord plasma IL-6 ( ≥ 11pg/mL).

B. The median concentration of umbilical cord plasma IL-6 is higher in fetuses with clinical chorioamnionitis than in those delivered by patients with near-term labor without clinical chorioamnionitis [27.46 pg/mL vs. 2.13 pg/mL; p<0.001]. Sixty-two percent of fetuses (16/26) with clinical chorioamnionitis have a fetal plasma concentration of IL-6 >11 pg/mL. Modified from Chaiworapongsa T et al. Evidence for fetal involvement in the pathologic process of clinical chorioamnionitis. Am J Obstet Gynecol. 2002;186(6):1178–82.²⁷⁷

IL: interleukin.

Criteria for the diagnosis of clinical chorioamnionitis

The current criteria for the diagnosis of clinical chorioamnionitis^{20, 89, 279} are based on clinical signs, including fever (±37.8°C or ±100.4°F) and two or more of the following: A) maternal tachycardia (>100 beats per minute); B) maternal leukocytosis (white blood cell count >15,000 cells/mm³; C) uterine tenderness; D) fetal tachycardia (>160 beats per minute); or E) purulent or malodorous cervical discharge.^{1, 19, 20, 89–92, 94, 279–284} The rationale for the precise cut-off values to define fever, maternal tachycardia, and fetal tachycardia was reviewed in detail by Newton,⁹² who proposed the 90^th percentile threshold for both maternal tachycardia and fetal tachycardia and the 80^th percentile for a white blood cell count (Table 3).

Table 3.

The origin of the criteria of clinical chorioamnionitis

Parameter	Mean ± standard deviation	95th percentile	Diagnosis of intraamniotic infection
Temperature, °C	36.9 ± 0.5	37.85	37.8
Maternal heart rate, beats/min	85 ± 13	110	100
Fetal heart rate, beats/min	139 ± 10	159	160
White blood cell count, cells/mm3	12,500 ± 3.9	20,100	15,000

Adapted from Newton ER. Chorioamnionitis and intraamniotic infection. Clin Obstet Gynecol. 1993;36(4):795–808.⁹²

The origin and evolution of the criteria for the current definition of clinical chorioamnionitis is interesting and has been influenced by the work of Gibbs et al.²⁰ They reported that patients who met the clinical criteria, proposed by their group, were more likely to have a higher concentration of microorganisms in amniotic fluid, using quantitative microbiology, and had isolates with higher virulence. This observation was based on a study of 52 patients diagnosed with clinical chorioamnionitis at term and matched with 52 controls in whom amniotic fluid was retrieved by a transcervical catheter.²⁰ The key observation was that 81% (42/52) of patients with clinical chorioamnionitis had >10² colony-forming units (CFU)/mL compared to 31% (16/52) in the control group (p<0.001). Organisms considered to be of high virulence were present in 69% (36/52) of patients with clinical chorioamnionitis and in 8% (4/52) of controls (p<0.01). Moreover, maternal bacteremia was detected in 10% (4/39) of patients diagnosed with clinical chorioamnionitis who had a blood culture.²⁰ Of note, a limitation recognized by the authors at the time of publication was that amniotic fluid for microbiologic studies had been obtained through a transcervical catheter and that despite discarding the first 7 mL of each sample, 31% of patients in the control group presented more than 10² CFU/mL, suggesting contamination by microorganisms in the lower genital tract.²⁰ Indeed, in studies of patients in term labor when amniotic fluid collection was performed by transabdominal amniocentesis at the time of cesarean delivery, the frequency of positive amniotic fluid cultures ranged between 13% and 19%.^163,167 Therefore, an accurate assessment of microbial status requires that the amniotic fluid sample be obtained transabdominally. The clinical criteria developed by Gibbs et al.²⁰ and Newton⁹² have continued to be applied because they have been perceived as useful in the differential diagnosis of intrapartum fever.

How well do the criteria for clinical chorioamnionitis perform in the identification of intraamniotic infection?

Studies of patients with the diagnosis of clinical chorioamnionitis at term in which cultivation and molecular microbiologic methods were used to detect intraamniotic infection show that the traditional criteria do not accurately distinguish between patients with and without intraamniotic infection.^{2, 95} The sensitivity of maternal tachycardia, fetal tachycardia, and maternal leukocytosis ranged from 75% to 90%; however, the specificity was poor for these criteria, ranging from 0% to 30%. By contrast, malodorous amniotic fluid and uterine tenderness had high specificity (95% and 95%, respectively) but a low sensitivity (8% and 12%, respectively) for the identification of intraamniotic infection.^{2, 95} Collectively, the diagnostic accuracy for each clinical sign ranges between 40% and 58% for the diagnosis of intraamniotic infection (Table 4).⁹⁵ The disappointing performance of clinical criteria has substantial implications: a large number of mothers and neonates are treated with antibiotics and undergo evaluation for suspected sepsis, given the fear that infection will be missed. Further studies are required to identify biomarkers that reliably and rapidly differentiate the patients with a fever due to intraamniotic infection or sterile intraamniotic inflammation, or those without intraamniotic infection or intraamniotic inflammation.

Table 4.

The diagnostic accuracy of clinical criteria for the identification of intraamniotic infection in patients with clinical chorioamnionitis at term

Clinical signs	Sensitivity		Specificity		Positive likelihood ratio(95% CI)	Negative likelihood ratio(95% CI)	Diagnostic accuracy% (n/N)
	% (n/N)	(95% CI)	% (n/N)	(95% CI)
Maternal tachycardia	88.0 (22/25)	(68.75–97.31)	5.0 (1/20)	(0.83–24.95)	0.93 (0.78–1.10)	2.40 (0.27–21.35)	51.1 (23/45)
Fetal tachycardia	80.0 (20/25)	(59.29–93.09)	30 (6/20)	(11.97–54.27)	1.14 (0.81–1.62)	0.67 (0.24–1.87)	57.8 (26/45)
Maternal leukocytosis	76 (19/25)	(54.87–90.58)	30 (6/20)	(11.97–54.27)	1.09 (0.76–1.56)	0.80 (0.30–2.10)	55.6 (25/45)
Malodorous vaginal discharge	8.00 (2/25)	(1.22–26.07)	95 (19/20)	(75.05–99.17)	1.60 (0.16–16.40)	0.97 (0.83–1.13)	46.7 (21/45)
Uterine tenderness	12.0 (3/25)	(2.69–31.25)	95 (19/20)	(75.05–99.17)	2.40 (0.27–21.35)	0.93 (0.78–1.10)	48.9 (22/45)
≥ 3 criteria	56.00 (14/25)	(34.94–75.57)	55 (11/20)	(31.55–76.90)	1.24 (0.69–2.26)	0.80 (0.44–1.45)	55.6 (25/45)
> 4 criteria	8.00 (2/25)	(1.22–26.07)	100 (20/20)	(83.01–100)	0.92 (0.82–1.03)	–	48.9 (22/45)

Maternal tachycardia (> 100 beats/min); fetal tachycardia (> 160 beats/min); and maternal leukocytosis ( > 15,000 cells/mm³).

CI: confidence interval.

Adapted from Romero R, Chaemsaithong P, Korzeniewski SJ, Kusanovic JP, Docheva N, Martinez-Varea A, et al. Clinical chorioamnionitis at term III: how well do clinical criteria perform in the identification of proven intra-amniotic infection? J Perinat Med. 2016;44(1):23–32.⁹⁵

Laboratory diagnosis of intraamniotic inflammation and intraamniotic infection

The diagnosis of clinical chorioamnionitis is based on the presence of maternal fever and two signs of maternal or fetal inflammation. Table 4 shows the sensitivity and specificity of clinical criteria to detect the patient with intraamniotic infection.

Amniotic fluid analysis o is indicated in cases of diagnostic certainty. Indeed, analysis of amniotic fluid is the gold standard for the diagnosis of intraamniotic inflammation and intraamniotic infection (Figure 11). Clinical criteria only allow a presumptive diagnosis. A conclusive, or definitive, diagnosis of infection requires identification of microorganisms and inflammation. Fluid can be retrieved at the time of cesarean delivery or by transabdominal amniocentesis. Tests could include the Gram stain, white blood cell count, and glucose concentration. These tests are employed to analyze cerebrospinal fluid in cases of suspected meningitis and are widely available in clinical laboratories all over the world.

Figure 11. Laboratory diagnosis of intraamniotic inflammation and intraamniotic infection.

Definitive diagnoses of intraamniotic inflammation and intraamniotic infection require a transabdominal amniocentesis to collect amniotic fluid. Amniotic fluid should be assessed by (1) a WBC count and differential; (2) a glucose concentration; (3) a Gram stain and a bacterial culture that include aerobic/anaerobic bacteria and genital mycoplasmas; (4) MMP-8; (5) cytokines, eg, IL-6; and (6) other tests (eg, rapid tests, PCR). Intraamniotic inflammation is the presence of an inflammatory response in the amniotic cavity, which can be diagnosed by a WBC count ≥ 50 cells/mm³, an amniotic fluid glucose concentration <14 mg/dL, an IL-6 concentration ≥ 2.6 ng/mL, or an MMP-8 concentration >23 ng/mL. Rapid tests for amniotic fluid MMP-8 (Yoon’s MMP-8 Check®; OBMed Co., Ltd., Seoul, Republic of Korea) and IL-6 (Milenia QuickLine®; Milenia Biotec, Bad Nauheim, Germany) are available to be performed at the bedside. Microorganisms in amniotic fluid can be identified by Gram stain, cultivation methods, or molecular microbiologic techniques.

IL: interleukin; MMP: matrix metalloproteinase; PCR: polymerase chain reaction; rRNA: ribosomal RNA; WBC: white blood cell.

Under normal circumstances, bacteria are not visible in amniotic fluid with a Gram stain, and a white blood cell count should be <50 cells/mm³ (criteria listed in Figure 11).^57–60 The amniotic fluid glucose concentration is used to assess the likelihood of intraamniotic inflammation in preterm gestation; however, the concentration of glucose decreases with gestational age, and a clear threshold has not been established at term, although amniotic fluid glucose is extremely low [median (interquartile range), 1 (1–4) mg/dL] in patients with intraamniotic infection at term.^{1, 2} Amniotic fluid concentrations of MMP-8 or IL-6 are excellent tests for the detection of inflammation. A rapid assay for MMP-8 is available, which can be performed at the bedside (Yoon’s MMP-8 Check®; OBMed Co., Ltd. Seoul, Republic of Korea); its diagnostic accuracy in identifying intraamniotic inflammation in patients with clinical chorioamnionitis at term is 84% (sensitivity, 82%; specificity, 90%; positive likelihood ratio, 8.2; negative likelihood ratio, 0.2).²⁸⁵ A rapid test to determine amniotic fluid IL-6 is commercially available in Europe (Milenia QuickLine®; Milenia Biotec, Bad Nauheim, Germany) (Figure 11).

A transcervical amniotic fluid collector is a non-invasive method that assesses the inflammatory status of the amniotic cavity in patients with rupture of the membranes (Figure 12A).^{129, 130} Amniotic fluid analysis can be performed by utilizing point-of-care assays such as an IL-8 test (Figure 12B).¹³⁰ Future studies to determine whether a transcervical amniotic fluid collector can be useful in the differential diagnosis of patients with clinical chorioamnionitis with intraamniotic inflammation vs neural inflammation related to an epidural are needed.

Figure 12. Amniotic fluid collector and rapid test to assess the presence of intraamniotic inflammation by determining interleukin-8 at the bedside.

A. A transcervical collector allows sampling of amniotic fluid after rupture of the membranes.

B. Rapid analysis of an interleukin-8 concentration at the bedside can be used to diagnose intraamniotic inflammation.

Cultures for aerobic and anaerobic bacteria as well as for genital mycoplasmas are performed to diagnose intraamniotic infection; however, results for these tests can take days to become available. Molecular microbiologic techniques allow the detection of the 16S ribosomal RNA gene to identify bacteria within a few hours.²⁸⁶ Indeed, successful use of nanopore sequencing, a long-read, real-time DNA sequencing technique, to diagnose intraamniotic infection within 5 to 9 hours has been recently reported.²⁸⁷ We envision that this technology will replace culture, allow the identification of intraamniotic infection, and guide the selection of appropriate antimicrobial agents for mothers and neonates in the future. It is possible to identify the microorganisms and to determine whether they have antimicrobial resistant genes in less than 24 hours.

Effect of clinical chorioamnionitis on uterine contractility

Patients with clinical chorioamnionitis are more likely to experience decreased or ineffective uterine contractility,^{3, 9, 288–290} prolonged labor,^{3, 9, 288–290} and cesarean delivery for failure to progress or for non-reassuring fetal heart rate tracing,^{3, 4, 8, 9, 280, 288, 290, 291} and to receive oxytocin for augmentation of labor^{3, 280, 288, 290} than those without clinical chorioamnionitis. Duff et al²⁸⁸ reported that 75% of patients with clinical chorioamnionitis have decreased uterine contractility and that, despite oxytocin augmentation, 34% require cesarean delivery because of the diagnosis of failure to progress in labor. A careful study by the group of Goetzl et al reported a significant decrease in uterine activity two hours after the onset of a fever (Figure 13). This occurred in nulliparous and multiparous patients.²⁹²

Figure 13. Relationship between maternal fever and uterine contractility.

Uterine contractility significantly and steadily declined 2 hours after the onset of maternal fever. Modified from Zackler A et al. Suspected Chorioamnionitis and Myometrial Contractility: Mechanisms for Increased Risk of Cesarean Delivery and Postpartum Hemorrhage. Reprod Sci. 2019;26(2):178–83.²⁹²

The mechanisms whereby maternal inflammation predisposes to myometrial dysfunction have not been determined. Proinflammatory cytokines, eg, IL-1β or TNF-α, can stimulate uterine contractility and have been implicated in the mechanisms responsible for the onset of preterm parturition in the context of intraamniotic infection^{293, 294} as well as spontaneous labor at term.^{295, 296} This concept would appear to conflict with the clinical findings that maternal fever and clinical chorioamnionitis are associated with impaired uterine contractility at term^{296, 297} and uterine atony in the postpartum period.^{280, 292, 298} How can these findings be reconciled? One possibility is that labor is stimulated during the early stages of intraamniotic infection. However, when the infection and the inflammatory process extend to the myometrium, pathologic inflammation impairs uterine contractility. Cierny et al ²⁹⁶ reported that the relationship between the concentration of maternal plasma cytokines and the progression of term labor is complex and likely to be bimodal. A low maternal plasma concentration of IL-6 has been associated with a prolonged latent phase of labor, suggesting that an inadequate inflammatory process slows labor (normal labor is considered a sterile inflammatory process), as demonstrated by transcriptomic analysis of the myometrium,^{299, 300} cervix,³⁰¹ and chorioamniotic membranes.^{302, 303} However, once active labor is established, patients with elevated maternal plasma cytokines (IL-1β and TNF-α) may experience prolonged labor.²⁹⁶ This suggests that pathologic inflammation can lead to an impairment of myometrial contractility. Indeed, an arrest of dilatation³⁰⁴ and an arrest of descent³⁰⁵ have features of exaggerated inflammation compared to normal labor.

Fetal heart rate in clinical chorioamnionitis

Abnormal fetal heart rate patterns observed in patients with clinical chorioamnionitis include tachycardia, absence of accelerations, variable and late decelerations, a sinusoidal pattern, persistently reduced variability, and absence of cycling.^{288, 306–309} Of these abnormalities, fetal tachycardia is the most common finding (~81%),^{1, 2, 95} and it is thought to reflect an increase in maternal temperature^{310, 311} and/or in fetal systemic inflammation (Figure 14).^{277, 312–315}

Figure 14. Fetal tachycardia in a case of clinical chorioamnionitis at 40 weeks of gestation.

A. Fetal heart rate of 190 beats per minute (bpm). The associated loss of variability is also noteworthy.

B. Tachycardia is associated with decelerations. The amniotic fluid culture was positive for Ureaplasma urealyticum and Staphylococcus aureus. The concentration of interleukin-6 in amniotic fluid was 37.49 ng/mL (cut-off 2.6 ng/mL), which is consistent with intraamniotic inflammation. The fetus presented fetal systemic inflammation, as demonstrated by the presence of funisitis. The newborn was septic, and the neonatal blood culture was positive for Staphylococcus aureus, the same bacterium identified in the amniotic fluid.

The frequency of fetal acidemia assessed by umbilical cord pH and blood gases is similar in patients with and without clinical chorioamnionitis.^{306, 316, 317} A study of 197 patients diagnosed with clinical chorioamnionitis showed no significant association between fetal acidemia (umbilical artery pH <7.20) and abnormal fetal heart rate patterns, including loss of variability, absence of accelerations, and tachycardia.³⁰⁶ Another study of 139 patients with intrauterine bacterial infection (defined as clinical chorioamnionitis plus a positive bacterial amniotic fluid culture or a neonatal infection) reported that fetal heart rate deceleration patterns (eg, decreased variability and absence of accelerations) were not significantly associated with the risk of cerebral palsy at two years of age.³⁰⁷

Placental pathology: acute histologic chorioamnionitis and funisitis

Histologic examination of the placenta is more helpful in excluding intraamniotic infection/inflammation than in making a positive diagnosis. In the absence of funisitis and chorionic vasculitis, fetal inflammation is unlikely to be an etiologic factor in neonatal complications and long-term handicap. Acute placental inflammatory lesions are observed in 79% of the placentas of patients with clinical chorioamnionitis at term (acute histologic chorioamnionitis: 79%; funisitis: 67%).² Most patients with proven intraamniotic infection have acute chorioamnionitis (93%) or funisitis (67%), and 52% of such cases have severe lesions (defined as stage 3 and grade 2 acute histologic chorioamnionitis and/or funisitis).² Patients with clinical chorioamnionitis but without intraamniotic infection and intraamniotic inflammation generally do not show evidence of severe acute histologic chorioamnionitis.²

Differential diagnosis of intraamniotic infection

Sterile intraamniotic inflammation: a new entity

A subset of patients with clinical chorioamnionitis in term and preterm gestation shows evidence of intraamniotic inflammation without the presence of bacteria in amniotic fluid, ie, sterile intraamniotic inflammation.^{1, 2} The presence of microorganisms has been excluded by a combination of methods (culture, presence of nucleic acids, and metagenomics).³¹⁸ The magnitude of inflammatory response in the mother, fetus, and amniotic fluid in patients with sterile intraamniotic inflammation is similar to or of a lesser magnitude than that found in patients with intraamniotic infection (Figure 9).²⁵¹

Sterile intraamniotic inflammation is thought to result from danger signals, or alarmins,^{283, 319–322} or from infections that escaped detection with current cultivation and molecular microbiologic techniques. Sterile inflammatory processes are well-recognized in other fields of medicine and occur when the injurious agent is of non-microbial origin. For example, a burn induces a sterile inflammatory process,³²³ urate crystals in the case of gout,^{324, 325} and calcium pyrophosphate dihydrate crystals in pseudogout.³²⁶ Urate crystals or calcium pyrophosphate engage the innate immune system by activating inflammasomes.

Damage-associated molecular patterns (DAMPs, also referred to as danger signals or alarmins) are molecules released by cells undergoing stress, necrosis, or senescence.^327–330 Cell death of the chorioamniotic membranes (ie, pyroptosis) or in fetal tissues could lead to the accumulation of alarmins and generate sterile intraamniotic inflammation in a wide range of obstetrical complications.³²² For example, in gastroschisis, exposure of the fetal bowel to amniotic fluid results in a sterile intraamniotic inflammatory process characterized by an elevated MMP-8 concentration in amniotic fluid.³³¹ Fetal surgery can also result in a transient sterile intraamniotic inflammatory process as surgical trauma elicits an inflammatory response. Alarmins are detectable in the amniotic fluid of patients with an episode of preterm labor with intraamniotic inflammation but without detectable bacteria,^294,332 and the intraamniotic administration of alarmins can induce preterm labor in vivo.³²¹ Sterile intraamniotic inflammation generates a chemotactic gradient, thus acute histologic chorioamnionitis similar to that generated by intraamniotic infection. The same can be said about funisitis and chorionic vasculitis.

Several interventions, including the blockade of the NLRP3 inflammasome,³³³ macrolides,³³⁴ and betamethasone,³³⁵ can be used to treat sterile intraamniotic inflammation. In animal models, these interventions reduced spontaneous preterm delivery and neonatal morbidity. Blockade of the NLRP3 inflammasome can be accomplished with a specific inhibitor which is currently being tested in non-pregnant subjects. N-acetylcysteine (NAC), a small molecule used to treat acetaminophen intoxication during pregnancy, has anti-inflammatory effects that have been attributed to attenuation of oxidative stress (ie, as a scavenger of reactive oxygen species). NAC also inhibits the NLRP3 inflammasome.³⁶⁸

Clinical chorioamnionitis without intraamniotic inflammation: a clinical enigma

Twenty percent of patients with the diagnosis of clinical chorioamnionitis at term do not show evidence of intraamniotic infection or intraamniotic inflammation after thorough examination of amniotic fluid samples obtained by transabdominal amniocentesis.^1,2 Interestingly, these patients have maternal plasma concentrations of cytokines similar to those with proven intraamniotic infection or sterile intraamniotic inflammation,²⁵¹ and this raises the question of the source of the systemic inflammatory response. One explanation for the fever is neuroinflammation caused by epidural analgesia. This condition is important, given that 73% of patients in the United States receive epidural analgesia during labor^336–338 and that 14% to 33% of patients who receive this anesthetic develop a fever.^339–345

Epidural analgesia and maternal fever

The association between epidural analgesia and maternal fever is supported by the following observations: 1) maternal body temperature increases within 4 to 6 hours after an epidural placement (Figure 15A),^346–350 and 2) an epidural increases the odds ratio (OR) of fever (≥37.5°C or ≥99.5°F) substantially [OR, 5.26; 95% confidence interval (CI), 4.98–5.56].³⁵¹ Not widely known is that an epidural-related fever is a particular phenomenon observed in pregnant women in labor. Indeed, an epidural administered to pregnant women not in labor is rarely associated with a fever (eg, an epidural for cesarean delivery).^352–354

Figure 15. Epidural analgesia, maternal temperature, and serum interleukin-6 concentrations.

A. Mean vaginal temperature (°C) of patients during labor according to intrapartum pain control. Pregnant women who received epidural analgesia (red) had an increased mean temperature after the administration of anesthetic agents. By contrast, the temperature of women receiving pethidine (non-epidural, blue) remained constant. Modified from Fusi L et al. Maternal pyrexia associated with the use of epidural analgesia in labour. Lancet. 1989 Jun 3;1(8649):1250–2.³⁴⁶

B. Maternal serum interleukin-6 (IL-6) concentrations significantly increase with the duration of epidural analgesia. Modified Goetzl et al. Elevated maternal and fetal serum interleukin-6 levels are associated with epidural fever. Am J Obstet Gynecol. 2002 Oct;187(4):834–8.³⁵⁸

In the case of epidural-induced maternal fever, the process is thought to represent a case of sterile systemic inflammation.^{355, 356} Riley et al ³⁵⁵ reported that 95% of epidural-related fever cases had a negative placental culture, and a systematic review and meta-analysis showed no association between epidural analgesia and neonatal sepsis (risk ratio, 0.56; 95% CI, 0.08–3.9),³⁵⁶ strengthening the case for a non-infection-related etiology. Systematic studies have shown that women who receive epidural analgesia have a higher mean maternal serum IL-6 concentration than those without epidural analgesia³⁵⁷ and that the longer the duration of epidural analgesia, the higher the maternal serum IL-6 concentration (Figure 15B).³⁵⁸

The precise etiology of the inflammatory process may be related to local anesthetics (eg, ropivacaine or bupivacaine) that can induce a maternal systemic inflammation after absorption from the epidural space into the vasculature.^{354, 359} During pregnancy, the density of the vascular network of epidural veins, ie, Batson’s plexus, is higher than that of non-pregnant subjects as shown by endoscopic studies of the epidural space, Figure 16A).³⁶⁰ Compression of the inferior vena cava by the pregnant uterus is thought to cause congestion of the epidural veins and a decrease in the volume of the epidural space.³⁶¹ Thus, a similar volume of local anesthetic agents would spread more extensively in pregnant women than in non-pregnant patients. Recently, Wohlrab et al ³⁶² reported that in vitro ropivacaine causes endothelial cell apoptosis and the release of alarmins, or danger signals, that, in turn, could induce the production of pyrogenic cytokines, eg, IL-1β and IL-6 (Figure 16B).^{345, 354, 362–364} For further details about epidural analgesia and maternal fever, the reader is referred to the review articles by Goetzl³⁶⁵ and Patel et al ³⁶⁶ in this Supplement.

Figure 16. A needle in the epidural space, optical imaging, and the mechanisms of fever.

A. Fiberscope of the epidural space. During pregnancy, the density of the epidural vessels is higher than that of non-pregnant subjects. It has been proposed that during pregnancy the inferior vena cava becomes compressed by the pregnant uterus and that this causes the epidural veins to engorge and the volume of the epidural space (*, blue) to decrease. Thus, a similar volume of local anesthetic agents can spread more extensively in the pregnant state than in the non-pregnant state. Modified from Eltzschig HK et al. Regional anesthesia and analgesia for labor and delivery. N Engl J Med. 2003 Jan 23;348(4):319–32³⁵⁹ and from Igarashi T et al. The fiberscopic findings of the epidural space in pregnant women. Anesthesiology. 2000 Jun;92(6):1631–6.³⁶⁰

B. Local anesthetics such as ropivacaine used during epidural analgesia can induce maternal systemic inflammation. Ropivacaine can directly cause endothelial cell apoptosis, and this induces the release of damage-associated molecular patterns (DAMP), also known as alarmins, that trigger the production of pyrogenic cytokines (eg, interleukin (IL)-1β, IL-6, IL-8), which may be responsible for fever in pregnant women following epidural analgesia.

Management

Antibiotic administration

The standard treatment of clinical chorioamnionitis at term is the administration of antibiotics and augmentation of labor.^{91, 279, 282, 367, 368} There is consensus that patients with clinical chorioamnionitis should receive intrapartum antibiotic therapy to prevent adverse perinatal outcomes.^{124, 369–372} This recommendation is based on the findings of a randomized clinical trial that compared intrapartum antibiotic treatment to immediate postpartum treatment.²⁸¹ This influential trial changed clinical practice. Therefore, the context of the study, design, conduct, and outcome are worth reviewing.

This study was conceived at a time when intrapartum antibiotic administration for clinical chorioamnionitis was not the standard of practice. At that time, neonatologists often requested that obstetricians delay antibiotic treatment until the postpartum period so that a blood culture of the neonate could be obtained to assess the likelihood of sepsis and to identify the organisms.

The trial included patients with clinical chorioamnionitis after 34 weeks of gestation²⁸¹ who were randomized to receive the combination of ampicillin (2 grams IV every 6 hours) and gentamicin (1.5 mg/kg IV every 8 hours) intrapartum versus postpartum (the postpartum treatment began after clamping of the umbilical cord). Patients who underwent cesarean delivery also received clindamycin (900 mg intravenously every 8 hours) after cord clamping. Intravenous antibiotics were continued until patients were afebrile for more than 48 hours. Timing and route of delivery were based on obstetrical criteria.

The primary endpoint of the trial was neonatal sepsis, defined as either bacteremia or death with the clinical diagnosis of sepsis and a positive culture. Evaluation of neonates after birth was based on one set of blood cultures and a chest x-ray, read by a radiologist unaware of the clinical course. Neonates received IV antibiotic treatment with ampicillin (75 mg/kg) and gentamicin (2.5 mg/kg) every 12 hours. If the neonatal blood culture and chest x-rays were negative, and the clinical course uncomplicated, antibiotics were discontinued after 72 hours. If sepsis or pneumonia was diagnosed, treatment was continued for a minimum of 10 days.

The trial was stopped by the Data and Safety Monitoring Committee after the enrollment of 48 patients (26 in the intrapartum treatment and 22 in the postpartum treatment), given that antibiotic administration was associated with a significantly lower rate of neonatal sepsis (21% in the postpartum treatment group and 0% in the intrapartum group).²⁸¹ In addition, intrapartum treatment was associated with a significant decrease in the risk of neonatal pneumonia or sepsis (32% vs 0.0%; p = 0.03), length of neonatal stay (mean: 5.7 days vs 3.8 days; p = 0.02), and maternal postpartum hospital stay (mean: 5.0 vs 4.0 days; p = 0.05)²⁸¹ compared to immediate postpartum treatment.

Some aspects of the trial deserve commentary. First, the rate of neonatal sepsis in the control group was 21% (4/19),²⁸¹ which is high compared to the rate of neonatal sepsis in clinical chorioamnionitis at the time, or now.^{373, 374} Second, the observed beneficial effects of antibiotics, namely a decrease in the rate of neonatal sepsis, could be due to suppression of bacterial growth in culture—rather than a true decrease in the rate of neonatal sepsis. This possibility must be considered in light of the evidence that antimicrobial administration can reduce the rate of positive blood cultures.³⁷⁵

This trial was not replicated after publication. Subsequently, other studies examined the effects of different antimicrobial agents on postpartum febrile morbidity and endometritis.^376–379 However, neonatal endpoints have not been revisited. In essence, the overall evidence for efficacy of antibiotics in medicine is so persuasive and the risks of neonatal sepsis so high that withholding antibiotics is difficult to justify.³⁷³Three-and-a-half decades after this important trial, a number of questions has emerged about the optimal diagnosis and treatment of clinical chorioamnionitis and suspected neonatal sepsis. Observational studies have shown that microorganisms isolated from infection sites such as wound infections or pelvic abscesses are the same as those found in amniotic fluid at the time of cesarean delivery.¹⁶⁵ These observations suggest that antibiotic choice needs to be modified to consider the microbiology of intraamniotic infection in clinical chorioamnionitis at term, which is now well-known, using state-of-the-art culture techniques and molecular microbiologic methods.^1,2 Several antibiotic regimens have been proposed for the treatment of clinical chorioamnionitis, and they include different generations of cephalosporins and other antimicrobials listed in Table 5.

Table 5.

Empiric antibiotic regimens proposed for the treatment of clinical chorioamnionitis

Antibiotic regimen proven to eradicate intraamniotic infection 119–122

Ceftriaxone 1 g IV every 24 hours plus clarithromycin 500 mg IV every 12 hours plus metronidazole 500 mg IV every 8 hours (clarithromycin is sometimes not available for intravenous administration but rather for oral administration)

Antibiotic regimen by recommended by ACOG (alternatives for patients allergic to penicillin are described in the text) 371

Ampicillin 2 g IV every 6 hours plus gentamicin 2 mg/kg load followed by 1.5 mg/kg every 8 hours (or gentamicin 5 mg/kg IV every 24 hours)a,b (clindamycin is added if patients have a cesarean delivery after clamping of the umbilical cord)

Alternative regimens used for the treatment of clinical chorioamnionitis 368

Cefotetan 2 g IV every 12 hours

Cefoxitin 2 g IV every 6 to 8 hours

Ceftizoxime 2 g IV every 12 hours

Cefotaxime 2 g IV every 8 to 12 hours

Cefuroxime 1.5 g IV every 8 hours

Cefazolin 1 g IV every 8 hours plus gentamicin 5 mg/kg IV every 24 hours or 1.5 mg/kg IV every 8 hours

Cefuroxime 750 mg IV every 8 hours plus metronidazole 500 mg IV every 8 hours

Mezlocillin 3–4 g IV every 6 hours

Piperacillin-Tazobactam 3.375 g IV every 6 hours

Piperacillin-Tazobactam 4 g IV every 6 hours plus clarithromycin 500 mg orally every 12 hours

Ticarcillin-clavulanic acid 3.1 g IV every 6 hours

Ertapenem 1 g IV every 24 hours

Meropenem 1 g IV every 12 hours

Imipenem-cilastatin 500 mg IV every 6 hours

Modified from Conde-Agudelo et al.³⁶⁸ (specific citations for each combination are in the original Table 2 in the article)

ACOG, American College of Obstetricians and Gynecologists; IV, intravenous.

^aFor mild penicillin allergy: cefazolin (2 g IV every 8 hours) as an alternative to ampicillin; For severe penicillin allergy: clindamycin (900mg IV every 8 hours) or vancomycin (1g IV every 12 hours as an alternative to ampicillin. Vancomycin should be used if the woman is colonized with group B streptococci resistant to either clindamycin or erythromycin (unless clindamycin-inducible resistance testing is available and is negative) or if the woman is colonized with group B streptococci and antibiotic sensitivities are not available.

^bPost-cesarean delivery antibiotic regimen: One additional dose of the chosen regimen is indicated. Add clindamycin 900 mg IV or metronidazole 500 mg IV for at least one additional dose. Post-vaginal delivery: No additional doses required; but if given, clindamycin is not indicated.

^*Vancomycin should be used if the woman is colonized with Group B Streptococci resistant to either clindamycin or erythromycin (unless clindamycin-inducible resistance testing is available and is negative) or if the woman is colonized with Group B Streptococci and antibiotic sensitivities are not available.

Empiric antibiotic treatment of clinical chorioamnionitis

Antimicrobial treatment is initiated as soon as the diagnosis is made. Patients can be treated with ampicillin 2 g IV every 6 hours combined with gentamicin 5 mg/kg every 24 hours or 1.5 mg/kg every 8 hours or ampicillin plus sulbactam 3 g IV every 6 hours. In case of a cesarean delivery, patients receive clindamycin 900 mg IV at the time of umbilical cord clamping. Based on expert opinion, metronidazole 500 mg IV has been proposed as an alternative to clindamycin in the event of cesarean delivery. In penicillin-allergic patients, clindamycin 900 mg IV every 8 hours, or vancomycin 1 g IV every 12 hours, or erythromycin 500 mg to 1 g IV every 6 hours can be used instead of ampicillin.

Other alternatives to ampicillin plus gentamicin have emerged. A major advance in antimicrobial treatment has been the development of newer beta-lactam antibiotics, a term describing agents containing a beta-lactam ring that binds with enzymes in the cell wall of the bacteria, such as penicillin-binding proteins, thereby disrupting the crosslinking of peptidoglycan, a key component of the bacterial cell wall. A weak bacterial cell wall causes osmotic stress; therefore, these types of antibiotics are bactericidal instead of bacteriostatic.

Since bacteria can produce beta-lactamases, which inhibit the beta-lactam ring and lead to antibiotic resistance, beta-lactamase inhibitors have been developed for use in combination with beta-lactam antimicrobial agents. Examples include clavulanic acid, sulbactam, tazobactam, and cilastatin. Popular combinations are ampicillin plus sulbactam, which appears to be equivalent to the combination of ampicillin and gentamicin (for details, see reviews of antimicrobial agents in ³⁶⁸ and ⁴⁰⁷), piperacillin and tazobactam with and without clarithromycin as well as ticarcillin plus clavulanic acid. Piperacillin, a penicillin with a polar side chain that enhances penetration into Gram-negative bacteria, is frequently used for the treatment of Pseudomonas aeruginosa infection (often called antipseudomonal penicillin) but has little activity against Staphylococcus aureus. Ticarcillin is also effective against Gram-negative bacteria such as Pseudomonas aeruginosa and Proteus vulgaris. Ticarcillin is one of the few antibiotics effective against Stenotrophomonas maltophilia, which has been implicated in infections in immunocompromised hosts, including preterm neonates.

Cephalosporins are a suitable alternative to ampicillin and gentamicin. Several generations of cephalosporins have been used. Initially, cefazolin, a first-generation cephalosporin, was an alternative to ampicillin in combination with gentamicin. Subsequent-generation cephalosporins have been used without gentamicin (eg, cefotetan, cefoxitin, ceftizoxime, etc.). Newer cephalosporins are of interest because of their broader antimicrobial coverage. Another reason to explore alternative antimicrobial agents is that gentamicin can be nephrotoxic and ototoxic.

Carbapenems, a class of antibiotics usually reserved for known or suspected multi-drug-resistant bacterial infections, include ertapenem, meropenem, and imipenem. These antibiotics are often used for the treatment of multi-resistant Gram-negative bacteria such as Pseudomonas spp. and Acinetobacter spp. Ertapenem is administered intravenously every 24 hours, which makes it an attractive agent. The antibiotics mentioned here have been used for the treatment of patients with clinical chorioamnionitis, and the dose, route, and schedule of administration are shown in Table 5. Some of the antibiotics can also be administered intramuscularly; however, the intravenous route is preferable.

An important consideration is to administer antimicrobials effective against genital mycoplasmas, such as macrolides (azithromycin or clarithromycin). We have used the combination of ceftriaxone, clarithromycin, and metronidazole. The rationale is described in the next section of this article.

There is no evidence that one antimicrobial regimen is superior to others. There is evidence demonstrating that antibiotic use in pregnancy is associated with the risk of antibiotic-resistant strains on the infant gut microbiota.³⁸⁰ The current obstetrical practice has evolved so that mothers with intrapartum fever are given antibiotics and their neonates observe or undergo septic workups. Since a fraction of patients with intrapartum fever has no evidence of infection, a sizable group of term neonates is exposed to antibiotics unnecessarily. This practice is costly, uncomfortable (because of blood culture and lumbar punctures in neonates), creates risk for the emergence of resistant strains of bacteria, and exposes infants to the long-term risks of early antimicrobial exposure (eg, obesity).³⁸¹

New concepts about the microbiology of intraamniotic infection and implications for antimicrobial agent selection

The use of ampicillin and gentamicin (ampicillin 2 g IV every 6 hours combined with gentamicin 5 mg/kg every 24 hours) is recommended by the American College of Obstetricians and Gynecologists whenever an intraamniotic infection is suspected or confirmed.³⁷¹ However, these antibiotics are not effective against Ureaplasma spp. or Mycoplasma hominis. These bacteria lack a cell wall; therefore, β-lactams (penicillins and cephalosporins) and glycopeptides (vancomycin) are not effective antimicrobial agents.^{382, 383} Similarly, gentamicin is also not effective against U. parvum and U. urealyticum.³⁸⁴

Macrolides, such as erythromycin, clarithromycin, and azithromycin, provide good antibacterial activity against Ureaplasma spp.^385–387; experimental evidence in non-human primates has shown that azithromycin can eradicate Ureaplasma spp. from the amniotic cavity and can reduce fetal lung injury.^{388, 389} Macrolides have been used to treat intraamniotic infection as a single agent or in combination with other antibiotics.^{207, 334, 368, 388, 390–398} In preterm gestations, eradication of intraamniotic infection by macrolides has been shown by serial amniocenteses.^392–395 In addition, adjunctive azithromycin has been shown to reduce the rate of surgical site infection in patients undergoing cesarean delivery. Specfically, in a large randomized clinical trial reported by Tita et al, patients at ≥24 weeks of gestation who were to undergo cesarean delivery during labor or those with rupture of the membranes for ≥4 hours were randomly assigned to adjunctive treatment with azithromycin or to a placebo.³⁹⁹ Patients with clinical chorioamnionitis were excluded from this study. All patients received cefazolin, a standard antibiotic prophylaxis. Allergic patients received clindamycin alone or clindamycin plus gentamicin. Patients allocated to receive adjunctive azithromycin had a reduction in the rate of postpartum endometritis [6.1% (61/994) vs 3.8% (39/1,019); p=0.02], wound infection [6.6% (66/994) vs 2.4% (24/1,019); p<0.001], and serious maternal adverse events [2.9% (29/994) vs 1.5% (15/1,019); p=0.03].³⁹⁹ Although there was no difference in the neonatal composite outcome, the findings of this study were considered sufficiently persuasive by professional organizations in the United States to recommend adjunctive azithromycin treatment.⁴⁰⁰

Clarithromycin has emerged as an effective treatment against Ureaplasma spp. Indeed, it has a higher rate of transplacental passage than azithromycin.⁴⁰¹ U. urealyticum isolated from pregnant patients is less resistant (0.9%–25%) to clarithromycin than to other antibiotics,^{385, 386, 402} and a recent report suggests that clarithromycin reduces the burden of Ureaplasma spp. DNA in the amniotic fluid of patients with preterm PROM.³⁹⁷ Randomized clinical trials comparing ampicillin to gentamycin vs other regimens, which include antibiotics effective against genital mycoplasmas (eg, clarithromycin and azithromycin) in patients with clinical chorioamnionitis, should be considered. Clarithromycin can be given intravenously or orally.

Given that most intraamniotic infections in clinical chorioamnionitis are polymicrobial, we have explored a combination of antimicrobial agents consisting of ceftriaxone, clarithromycin, and metronidazole in preterm gestations. This combination has successfully eradicated intraamniotic infection caused by genital mycoplasmas as well as mixed anaerobic and/or aerobic bacteria.^392–395 The rationale for the administration of this antibiotic regimen is the following: 1) ceftriaxone has enhanced coverage of aerobic bacteria and a high rate of transplacental passage^{403, 404}; 2) clarithromycin is effective against Ureaplasma spp.^{385, 386, 402}; and 3) metronidazole has shown optimal coverage of anaerobic microorganisms.^{405, 406} This combination has been effective in eradicating intraamniotic infection and sterile intraamniotic inflammation in patients with preterm PROM,^{392, 393} preterm labor with intact membranes,³⁹⁴ cervical insufficiency,^{395, 396} and threatened midtrimester miscarriage.³⁹⁸ Macrolides not only inhibit bacterial proliferation but also have anti-inflammatory properties, hence their value in treating infections with Ureaplasma spp. and also sterile intraamniotic inflammation.³⁴⁴ In the absence of strong evidence for a specific antibiotic regimen to treat patients with chorioamnionitis at term,^{368, 407} this new antibiotic regimen can be considered and should be the subject of investigation.

For how long should patients with clinical chorioamnionitis be treated with an antibiotic regimen?

The duration of antibiotic treatment has been the subject of discussion and of clinical trials.^408–412 Most professional societies do not recommend continuing antibiotic administration in the postpartum period in cases of vaginal delivery.^{371, 407, 410} Presumably, this suggestion is based on the notion that intraamniotic infection is confined to the amniotic cavity and the chorioamniotic membranes and that the infectious process is considered resolved after the expulsion of the placenta. Others recommend continuing antibiotics in patients who undergo a cesarean delivery for at least one dose or for 24 to 48 hours.³⁷¹

Studies about the administration of antibiotics in the postpartum period have generally used a febrile index or febrile morbidity as an endpoint for the cure.^{413, 414} However, clinicians need to consider that clinical chorioamnionitis is a risk factor for secondary infertility,⁴¹⁵ and inadequate treatment of infection may result in this complication. We do not believe that fever, an indication of a systemic host response, is an adequate endpoint to assess the cure of a soft-tissue infection, such as endomyometritis, or the resolution of pelvic peritonitis caused by the spillage of bacteria into the perineal cavity at the time of cesarean delivery. The World Health Organization recommends that patients with puerperal endometritis be treated for at least 24 to 48 hours after the complete resolution of signs and symptoms, ie, fever, uterine tenderness, purulent lochia, and leukocytosis.⁴⁰⁷

Antipyretics

Maternal hyperthermia, even in the absence of infection, seems to have deleterious effects on the fetus. Experimental studies have shown that maternal hyperthermia leads to changes in fetal hemodynamic and metabolic parameters.^{416, 417} For example, in non-human primates, hyperthermia, induced with heat from a lamp (in the absence of infection), is associated with the development of fetal tachycardia,⁴¹⁷ hypoxia,⁴¹⁶ metabolic acidosis,⁴¹⁶ and hypotension.⁴¹⁶ Adverse neonatal outcomes reported in patients who developed intrapartum fever include low Apgar scores,^418–421 acidosis,³ hypotonia,^{418, 419} requirement of resuscitation immediately after delivery,⁴¹⁸ oxygen therapy or mechanical ventilation,^{3, 418, 419, 422} NICU admission,^420–422 neonatal sepsis,^{3, 420, 421} neurological morbidities,^{3, 268, 418, 419, 421–426} and neonatal mortality.⁴²⁷ The extent to which these complications are a result of maternal fever is uncertain. A systematic review and meta-analysis show that intrapartum maternal fever increases the odds of adverse neurological outcomes, including neonatal seizures, encephalopathy, or cerebral palsy (OR, 2.48; 95% CI, 2.28–2.70).³⁵¹ Of note, intrapartum fever remains a significant risk factor for neonatal brain injury (OR, 2.79; 95% CI, 2.54–3.06), after excluding those with a diagnosis of clinical chorioamnionitis.³⁵¹

Antipyretics have been recommended in patients with clinical chorioamnionitis.^{94, 369, 371} Acetaminophen is the most frequently used antipyretic and can be administered orally, rectally, or intravenously. Serum peak levels (~12 μg/mL) and half-life (~1.5 hours) of acetaminophen in pregnant women are similar to those observed in non-pregnant adults.⁴²⁸ The conventional oral dose of acetaminophen is 325–650 mg every 4 to 6 hours; total daily doses should not exceed 4 g. Studies on the effect of acetaminophen on maternal and fetal temperatures during labor, as well as on adverse obstetrical and neonatal outcomes, are sparse and show conflicting results. In 1989, a case series described the effects of acetaminophen administration (650 mg rectally) in 8 febrile patients diagnosed with clinical chorioamnionitis.⁴²⁹ When the temperature remained >38.3°C, the dose was repeated for one to two hours. Acetaminophen administration resulted in a mean decrease in temperature of 1.2°C.⁴²⁹ In addition, the fetal heart rate tracings improved once the maternal fever was reduced.⁴²⁹Significant improvements in acid-base status were noted at birth as compared to that of the fetal scalp blood at the peak of the maternal fever.⁴²⁹

A retrospective observational study reported that acetaminophen administration in patients with intrapartum fever was associated with a significant decrease in the frequency of failure to progress in labor in comparison to no administration of acetaminophen [18% (22/122) vs 32% (12/38); p = 0.04].⁴³⁰ On the other hand, there was no evidence that acetaminophen reduced the frequency of meconium-stained amniotic fluid, fetal distress, or NICU admission.

Oral administration of 1000 mg of acetaminophen to patients with intrapartum fever ≥38.0°C (n=18) has been reported to decrease neither maternal axillary nor fetal scalp temperatures even though acetaminophen appeared to halt ongoing increases in fetal temperatures.⁴³¹ Re-analysis of the study data demonstrated that maternal and fetal temperatures decreased after acetaminophen administration.⁴³² A more recent study of 54 patients who presented with an intrapartum fever ≥38°C, of whom only 3 were diagnosed with clinical chorioamnionitis, reported the effects of oral administration of 650 mg of acetaminophen.⁴³³ Overall, no significant differences were observed between patients who received acetaminophen (N=41) and those who did not (N=13) in the frequency of cesarean delivery, presence of meconium, requirement for neonatal bag/mask ventilation or continuous positive pressure ventilation, and NICU admission.⁴³³

Intravenous administration of acetaminophen may be an alternative when patients are unable to tolerate oral administration or when an earlier onset of action is desirable. Indeed, intravenous acetaminophen has increased bioavailability and more rapid onset of action.⁴³⁴ A recent report showed that two patients with intrapartum fever and fetal tachycardia had a reduction of maternal temperature and resolution of fetal tachycardia 20 minutes after intravenous administration of 1 g of acetaminophen.⁴³⁵ Each patient had an uncomplicated vaginal delivery of a healthy neonate. The authors argued that the resolution of maternal fever and fetal tachycardia may have prevented cesarean delivery secondary to a non-reassuring fetal heart rate status. A randomized clinical trial is in progress to compare two regimens: acetaminophen 1g IV every 6 hours (up to 2 doses) vs acetaminophen 1 g orally every 6 hours (up to 2 doses) in women in active labor who present with a fever >38.0°C (NCT02625454).⁴³⁶

In summary, although there is no clear evidence that the treatment of intrapartum fever reduces the risk of adverse obstetrical and neonatal outcomes, antipyretics, mainly acetaminophen, can be used to treat hyperthermia as well as to reduce fetal tachycardia. Randomized controlled trials evaluating the effects of acetaminophen on intrapartum fever and on obstetrical and neonatal outcomes in patients with clinical chorioamnionitis are necessary.

Conclusions

Clinical chorioamnionitis is a syndrome defined by intrapartum fever combined with the presence of maternal and fetal signs of systemic inflammation, mainly caused by intraamniotic infection or sterile intraamniotic inflammation. A subset of patients with clinical chorioamnionitis does not have intraamniotic infection or sterile intraamniotic inflammation, and some of these cases are attributed to a systemic maternal inflammatory process related to epidural anesthesia and analgesia. The differential diagnosis of this syndrome and optimal treatment are important issues in clinical obstetrics, and biomarkers for the detection of intraamniotic infection are needed. Genital mycoplasmas are the most frequent microorganisms found in amniotic fluid. The standard treatment is intrapartum administration of antibiotics to reduce the rate of neonatal sepsis. The fetal attack rate by microorganisms in the amniotic cavity is low; however, neonatal sepsis in patients with clinical chorioamnionitis remains an important problem. Antipyretics, mainly acetaminophen, can be used to treat hyperthermia as well as to reduce fetal tachycardia; however, there is no clear evidence that the treatment of intrapartum fever reduces the risk of adverse obstetrical and neonatal outcomes. Important questions for future investigations are the following: 1) can amniotic fluid retrieved by a transcervical collector be used to diagnose intraamniotic inflammation?; 2) can this information be used to select patients who could benefit from antibiotic treatment and those who do not?; 3) is treatment with antimicrobial agents effective against genital mycoplasmas (azithromycin or clarithromycin) warranted to improve maternal and neonatal outcomes?; 4) should neonatologists include tests to identify genital mycoplasmas in septic workups of neonates born to patients with clinical chorioamnionitis?; and 5) what is the etiology of clinical chorioamnionitis in the absence of intraamniotic infection and intraamniotic inflammation, particularly if epidural analgesia or anesthesia has not been used?