Placental Infection With Ureaplasma species Is Associated With Histologic Chorioamnionitis and Adverse Outcomes in Moderately Preterm and Late-Preterm Infants

Abstract

Objective. The human Ureaplasma species are the microbes most frequently isolated from placentae of women who deliver preterm. The role of Ureaplasma species has been investigated in pregnancies at <32 weeks of gestation, but currently no studies have determined the prevalence of ureaplasmas in moderately preterm and late-preterm (hereafter, “moderate/late preterm”) infants, the largest cohort of preterm infants.

Methods. Women delivering moderate/late preterm infants (n = 477) and their infants/placentae (n = 535) were recruited, and swab specimens of chorioamnion tissue, chorioamnion tissue specimens, and cord blood specimens were obtained at delivery. Swab and tissue specimens were cultured and analyzed by 16S ribosomal RNA polymerase chain reaction (PCR) for the presence of microorganisms, while cord blood specimens were analyzed for the presence of cytokines, chemokines, and growth factors.

Results. We detected microorganisms in 10.6% of 535 placentae (443 were delivered late preterm and 92 were delivered at term). Significantly, Ureaplasma species were the most prevalent microorganisms, and their presence alone was associated with histologically confirmed chorioamnionitis in moderate/late preterm and term placentae (P < .001). The presence of ureaplasmas in the chorioamnion was also associated with elevated levels of granulocyte colony-stimulating factor (P = .02).

Conclusions. These findings have important implications for infection and adverse pregnancy outcomes throughout gestation and should be of major consideration for obstetricians and neonatologists.

Preterm birth (ie, delivery <37 weeks of gestation) is the leading cause of neonatal death and disability [1]. Globally, 15 million infants are born prematurely each year, accounting for 10% of all births [2]. Preterm birth can be grouped according to gestational age, with infants born at <28 weeks of gestation considered to be extremely preterm; infants delivered between 29 and 32 weeks of gestation are considered to be very preterm, and infants delivered between 32 and 36 weeks of gestation are referred to as moderately or late-preterm (M/LPT) [1]. Of these, the M/LPT infants are the most prevalent and account for up to 85% of all preterm births worldwide.

Infection of the female upper genital tract is one of the most frequent causes of preterm birth [3, 4] with as many as one in three women experiencing an upper genital tract infection during their pregnancy [2]. Currently, there is overwhelming evidence implicating upper genital tract infection with early preterm births (ie, birth at <32 weeks of gestation) and adverse neonatal outcomes [5–8]; however, there is limited evidence as to the prevalence of infection later in gestation (ie, at >32 weeks of gestation). Studies have demonstrated that 25%–40% of all preterm births are associated with infection, and the most common pathogens are the human Ureaplasma species [9–25]. While rates of ureaplasma infection are reported to be as high as 42% in some early preterm birth studies [23, 26, 27], the role of ureaplasmas in disease is complicated, as upper genital tract infections that occur at <32 weeks of gestation are often polymicrobial [24, 28–30].

In this study, we investigated the prevalence and diversity of microorganisms infecting the chorioamnion of M/LPT placentae and those delivered at term. We hypothesized that the presence of microorganisms would be associated with histologically confirmed chorioamnionitis and adverse pregnancy outcomes, and we aimed to provide novel insight into the role of infection late (ie, >32 weeks) during gestation.

MATERIALS AND METHODS

Study Population

From July 2010 to April 2013, patients delivering at the Good Samaritan Hospital (Cincinnati, Ohio) were recruited, and informed consent was obtained. Women with known congenital infections, human immunodeficiency virus infection, or fetal malformations were excluded. Ethics approval was granted by the review boards of the Good Samaritan Hospital (approval 09105-09-067) and Cincinnati Children's Hospital Medical Center (approval 2009-0236). All patients gave permission for the collection of their placentae upon delivery and for their medical records to be accessed and recorded in a deidentified database.

Specimen Collection

Following delivery, placentae were refrigerated in sterile containers and transported to a procedure room for sampling under strict aseptic conditions within 24 hours of delivery. The external placental surface was decontaminated using 70% alcohol, and areas where the amnion had detached from the placenta were avoided, to minimize contamination. An incision was made into the amnion (ie, uppermost) membrane, and the interface between the amnion and chorion was identified. The membranes were then peeled apart, and the membrane interface was swabbed using sterile cotton swabs and placed into BBL Port-a-Cul medium (for isolation of bacteria and yeast; Becton, Dickenson, and Company Franklin Lakes, New Jersey) and into universal viral transport medium (for isolation of ureaplasmas/mycoplasmas; Becton, Dickenson, and Company). Chorioamnion tissue was also excised and placed into sterile cryogenic vials. All samples were stored at −80°C prior to analysis.

Culture of Microorganisms From Clinical Samples

Swabs stored in BBL Port-a-Cul medium were inoculated into thioglycolate broth and onto a range of media, including horse blood agar, chocolate agar, Schaedler anaerobe agar, Sabauraud's dextrose agar, deMan Rogosa Sharpe agar, and MacConkey no. 1 agar (Thermo Fisher Scientific, Queensland, Australia). Media were incubated either aerobically, under 5% CO₂ or anaerobically in jars (Anaerogen gas pack, 2.5 L; Thermo Fisher Scientific) as appropriate. Agar plates were checked daily for signs of growth for up to 5 days.

Swabs stored in universal viral transport medium were cultured and quantified for the presence and numbers of Ureaplasma species colony-forming units as previously described [31], using 10B broth and A8 agar [32].

Tissue specimens were weighed and homogenized in sterile cryogenic vials containing 1 mL of sterile saline and sterile glass beads, using the Mini Beadbeater-16 cell disruptor (Daintree Scientific; Tasmania, Australia). Tissues were homogenized in 1-minute cycles for up to 4 minutes and inoculated onto bacteriological media (as described above), using a 10 µL sterile loop, into ureaplasma selective broth and thioglycolate broth. The number of colony-forming units was measured semiquantitatively to determine the numbers of microbes in the chorioamnion tissue. Numbers of ureaplasmas were quantified as described above.

Extraction of DNA From Cultured Microorganisms and Chorioamnion Tissue Homogenates

DNA extractions were performed using the QIAamp mini DNA extraction kit (Qiagen; Victoria, Australia). Extraction of DNA from cultured microorganisms (3–5 colonies per colony type) and tissue homogenates (approximately 500 µL) were performed using the manufacturer's instructions for tissues. Extracted DNA was stored at −20°C until required.

Polymerase Chain Reaction (PCR) Targeting the 16S Ribosomal RNA (rRNA) Gene

16S rRNA PCR assays were performed in 20 µL reactions containing 4 µL of extracted DNA, 100 µM dNTPs (Roche; Queensland, Australia), 1 X PCR buffer (Tris HCl, KCl, and [NH₄]₂SO₄, pH 8.7, Invitrogen; Victoria, Australia), 2 mM MgCl₂ (Invitrogen), 0.5 µM of forward (5′-C[C/T]GG[C/T]AG[C/T]CCACGCCG[C/T]AAA-3′) and reverse (5′-ACA[C/T]C[C/T]CACGACACGAGC[C/T]G-3′) primers (Sigma-Aldrich), 2.5 U of Platinum taq DNA polymerase (Invitrogen), and sterile DNAse/RNAse-free dH₂O (Thermo Fisher Scientific). PCR cycling consisted of initial denaturation at 95°C for 5 minutes, followed by 30 cycles of 94°C for 30 seconds, primer annealing at 60°C for 30 seconds, and extension at 72°C for 30 seconds. DNA extracted from pure cultures of gram-positive and gram-negative organisms served as the template for positive-control PCR reactions. Negative controls included replicates of master mix only and water substituted for template, to identify contamination.

PCR Targeting the Multiple Banded Antigen (mba) Gene of Ureaplasma species

To distinguish the 2 different Ureaplasma species in the clinical isolates, PCR assays targeting the upstream conserved portion of the mba gene were performed as previously described [33].

Sequencing of PCR Amplicons

PCR amplicons were sequenced to confirm the identity of each microorganism. PCR products were purified using the PureLink PCR purification kit (Invitrogen) according to the manufacturer's instructions.

PCR amplicons were labeled using the BigDye Terminator (BDT) v3.1 cycle sequencing kit (Thermo Fisher Scientific) as per the manufacturer's instructions. Briefly, this included a PCR assay to incorporate the BigDye Terminator in the PCR amplicon, and the labeled amplicons were then purified using ethylenediaminetetraacetic acid (125 mM, pH 8.0), sodium acetate (3 M), and 100% ethanol. Sequencing was carried out using the ion personal genome machine (PGM) molecular sequencer (Life Technologies, Victoria, Australia). Sequence data was imported and manipulated using the Geneious bioinformatics software (Biomatters, Auckland, New Zealand) and the identity of each clinical isolate was obtained using the basic local alignment search tool (BLAST; National Centre for Biotechnology Information [NCBI]).

Detection of Cytokines Within Cord Blood

Cord blood was collected by sterile needle and syringe at the time of delivery. Blood from each placenta was centrifuged and the serum was snap frozen in cryogenic vials until required. Cord blood was assayed using the Luminex Milliplex Map human cytokine/chemokine magnetic bead panel (Millipore, Billerica). Concentrations of interleukin 6 (IL-6), interleukin 8 (IL-8), monocyte chemoattractant protein 1 (MCP-1), and granulocyte colony-stimulating factor (G-CSF) were calculated from standard curves using recombinant proteins.

Statistical Analysis

Data are presented as mean values ± standard error of the mean (SEM). All data in this study were analyzed using the SPSS software program, and data were plotted using GraphPad Prism. Analyses included both binary logistic regression analysis and analysis of variance to determine statistical significance between cohorts. Statistical significance was accepted as a P value of <.05.

RESULTS

Study Population

A total of 477 women and 535 infants (n = 535 placentae) were recruited during the study period. The majority of women in this study were white (66.4%) or African American (25.8%), with 26 women of mixed ethnicity and 8 women of Asian descent. Of the 535 infants included in the study, 421 were singleton births, and 114 infants were part of multiple births (54 sets of twins and 2 sets of triplets). Of the 477 women, 385 delivered 443 M/LPT infants, while 92 women delivered at term (all term deliveries were singleton births). Almost all women (>97%) in our study had evidence of at least 1 prenatal care visit during their current pregnancy.

Chorioamnion Infection in M/LPT and Term Placentae

Microorganisms were detected by culture and/or molecular methods in 57 of 535 placentae (10.6%). Of the organisms isolated, all microbes were detected by molecular methods, while 52 of 61 (85.2%) were identified by traditional culture techniques. Of these, all organisms were isolated from both swab and tissue specimens, with the exception of 2 placentae, from which microorganisms were isolated from the tissue specimen but were not isolated from the swab specimen. Of the placentae tested, the majority of infections (53 of 57 [93.0%]) were caused by a single microbial species, whereas 44/57 (7.0%) were found to contain polymicrobial infections. A total of 61 microorganisms were isolated from 57 placentae, and these included Ureaplasma parvum, Ureaplasma urealyticum, Streptococcus agalactiae (group B streptococcus), Bacteroides fragilis, Bifidobacterium species, Gardnerella vaginalis, Propionibacterium species, and Escherichia coli. Noncultivable bacteria were detected by 16S rRNA PCR in 6 placentae (9.8%; Table 1). Of the organisms isolated, the human Ureaplasma species were the most prevalent, with U. parvum alone accounting for 36/61, 59.0% of all microorganisms in our study. No Lactobacillus species were detected by culture or PCR, nor were any fungi or yeasts detected.

There was no difference in the incidence of microorganisms isolated from placentae of women who delivered M/LPT infants or at term (44/443, 9.9% and 13/92, 14.1%, respectively; P = .12; Table 1).

Association of Infection With Adverse Pregnancy and Neonatal Outcomes in M/LPT and Term Pregnancies

Placentae infected by microorganisms were more likely than noninfected placentae to exhibit histologically confirmed chorioamnionitis (31/47, 54.4% vs 90/427, 18.8%; P < .001; Table 2), irrespective of gestational age (Figure 1A and 1B) and ethnicity (P = .528; data not shown). A greater percentage of neonates delivered by mothers with chorioamnion infection required oxygen or positive pressure respiratory support for prolonged periods (ie, >6 h; 12/57, 21.0%), when compared with infants whose placentae had no detectable microorganisms (43/478, 9%; P = .009; Table 2).

Ureaplasma species and Other Microorganisms in Association With Adverse Pregnancy or Neonatal Outcomes

Because the Ureaplasma species were the microorganisms most frequently isolated in our study, we also correlated the presence of these species with adverse pregnancy or neonatal outcomes. Owing to the small number of microorganisms other than Ureaplasma species present, it was not possible to analyze these other species individually; therefore, we compared the outcomes of women with placentae infected or colonized with Ureaplasma species and those with placentae infected with microorganisms other than Ureaplasma (referred to as ‘other’ microorganisms throughout the paper). In some cases, women delivered multiple infants/placentae in which 1 placenta was infected but the other was not (n = 3) or had placentae that contained both ureaplasmas and other microorganisms (n = 4); data from these women were excluded from these analyses.

Women testing positive for the presence of Ureaplasma species within a placenta (n = 32) had a decreased maternal age (mean age, 24.2 ± 0.8 years; P = .002), compared to women - positive for the presence of other microorganisms (n = 12; mean age, 28.2 ± 1.4 years) and those women who had an absence of microorganisms in placentae (n = 427; mean age, 27.8 ± 0.3 years). There were no other differences in maternal demographic data (Table 2) or incidence of adverse pregnancy outcomes between these groups of women.

The presence of Ureaplasma species (n = 38), but not other microorganisms in placentae (n = 535) correlated with histologically confirmed chorioamnionitis (26 of the 38 Ureaplasma - positive placentae; 68.4% vs 4/15 of the placentae infected with other microorganisms, 26.7%, respectively; P < .001). Of the placentae infected with ureaplasmas, 12/38 (31.6%) showed no evidence of inflammation, 14/38 (36.8%) showed only mild evidence of chorioamnionitis (grade 1), and 12/38 (31.6%) demonstrated histologically confirmed severe chorioamnionitis (grades 2 or 3; Figure 1B). There was no difference in the prevalence of U. parvum or U. urealyticum clinical isolates in placentae with or without histologically confirmed chorioamnionitis (P = .10) or in placentae that were delivered late preterm or at term (P = .076; data not shown). Interestingly, the presence of histologically confirmed chorioamnionitis was significantly associated with spontaneous (but not medically indicated) M/LPT birth when Ureaplasma species infection was present (22.5% vs 3.1%; P < .001 data not shown). There were no other significant correlations between the presence of other microorganisms and spontaneous delivery in the presence and absence of histologically confirmed chorioamnionitis (3.4% and 1.8%, respectively).

Placental Infection With Ureaplasma species and Other Microorganisms Was Associated With Elevated Cord Blood Cytokine Levels

Chorioamnion infection, irrespective of etiology, was associated with elevated cord blood factors. Placentae with chorioamnion infection demonstrated significantly higher concentrations of IL-8 (536 pg/mL) and G-CSF (403 pg/mL), compared with cord blood from pregnancies in which no chorioamnion infection was detected (56 pg/mL [P = .03] and 231 pg/mL [P = .04], respectively; data not shown).

We also investigated whether specific cord blood cytokines were associated with infection with particular microorganisms. Detection of Ureaplasma species in the chorioamnion was associated with elevated levels of G-CSF (P = .02; Figure 2D) but not IL-6 (Figure 2A), IL-8 (Figure 2B), and MCP-1 (Figure 2C). By contrast, cord blood collected from pregnancies in which the chorioamnion was infected by microorganisms other than Ureaplasma demonstrated higher levels of IL-8 (P = .01; Figure 2B).

DISCUSSION

Infections of the female upper genital tract are associated with preterm birth, and up to 40% of all preterm births are attributed to infection [3]. While previous studies have identified an association between upper genital tract infections in pregnancy and early preterm births (ie, <32 weeks of gestation) [3, 4, 24, 35, 36], to our knowledge our large study of 535 placentae is the first to investigate the prevalence of upper genital tract infections in late gestation (ie, >32 weeks of gestation). We identified chorioamnion infection in 10.6% of all M/LPT and term pregnancies. We also found no differences in the incidence of chorioamnionitis in M/LPT or term placentae and no major differences in the prevalence of signs/symptoms of infection in women experiencing chorioamnionitis. This highlights why it is so difficult to identify and treat women with asymptomatic upper genital tract infection and chorioamnionitis during pregnancy.

Within our study, colonization and infection of placentae were predominantly caused by a single microorganism. The human Ureaplasma species were the most prevalent microorganisms identified in the chorioamnion, accounting for >68% of all isolates in the study. While 70% of the women in our study delivered vaginally, we isolated no Lactobacillus species in any chorioamnion tissue specimens tested. Lactobacillus species have been isolated in up to 90% of vaginal specimens [37], and the absence of this microorganism in placental tissue confirms that our methods of collection and sampling have eliminated or circumvented vaginal microflora contamination. Furthermore, the majority of the microorganisms detected in these placentae are consistently associated with intraamniotic infection. Some studies have demonstrated the presence of these species in early preterm births (<32 weeks) [21, 38], and other studies have linked their presence to bacterial vaginosis and a 2-fold increase in the likelihood of delivering preterm [39].

Perhaps the most significant finding of this study was that infection of the chorioamnion was associated with histologically confirmed chorioamnionitis, regardless of gestational age. We identified an increased prevalence of chorioamnionitis in placentae that were infected with microorganisms, compared to placentae with no detectable microorganisms. These findings also confirm the findings of others, who demonstrated that upper genital tract infection during pregnancy was associated with chorioamnionitis [24, 28, 29]; however, our study is the first to show such high rates of infection and chorioamnionitis in the M/LPT and term gestational periods, indicating that chorioamnionitis is a significant finding, regardless of gestational age.

This study further investigated the role of individual microorganisms with adverse pregnancy or neonatal outcomes and identified that infection with Ureaplasma species but not other microorganisms was significantly associated with chorioamnionitis and spontaneous M/LPT delivery. Previous studies have reported associations between infection with ureaplasmas and preterm birth accompanied by histologically confirmed chorioamnionitis; however, these studies focused on early preterm births, in which the majority of infections are polymicrobial [17, 24]. As a result, researchers have not yet been able to confidently claim that these microorganisms are true etiological agents of preterm birth or chorioamnionitis. However, in this large study, we demonstrated that infection is predominantly caused by a single microorganisms and that Ureaplasma species infection was independently associated with chorioamnionitis, regardless of gestational age.

Significantly, higher concentrations of IL-8 and G-CSF were detected in cord blood in association with chorioamnion infection (regardless of the etiology), compared to pregnancies with no infection. These results are similar to the findings of others who demonstrated that infection of the amniotic fluid or placentae was associated with elevated levels of cytokines in amniotic fluid [30, 40–42]. In the current study, the presence of microorganisms other than Ureaplasma species was associated with a higher concentration of IL-8 in cord blood, whereas the presence of Ureaplasma species was associated with an elevated level of G-CSF. G-CSF is a cytokine that affects the proliferation and differentiation of neutrophil progenitors, and previous studies have identified elevated levels of G-CSF in association with chorioamnionitis [43]; however, ours is the first study to investigate the concentrations of cord blood cytokines in relation to placental infection. These findings suggest that G-CSF may be a potential biomarker of asymptomatic chorioamnionitis or ureaplasma infections during pregnancy.

While we identified placental infection in 10.6% of M/LPT and term pregnancies, we found little evidence that upper genital tract infection was associated with M/LPT delivery. Therefore, the question remains as to the other likely causes or etiologies of M/LPT birth. Some known precursors and risk factors include spontaneous preterm labor, preterm premature rupture of membranes, cervical incompetence, preeclampsia, multiple gestation pregnancies, intrauterine growth restriction, prior Cesarean delivery, chorioamnionitis, and fetal distress [44]. In our study, many of these factors (which may predispose to preterm birth) were present. Given the presence of these factors, it is probable that a combination of these various precursors/risk factors contributed to delivery in the M/LPT period, and as such, the etiology of M/LPT birth is likely to be multifactorial.

To the best of our knowledge, this is the first study to report the prevalence of chorioamnion infection in M/LPT placentae. Greater than 70% of all preterm births worldwide occur in the M/LPT period [2]; therefore, the number of infants potentially exposed to chorioamnion infection in this period is far greater (estimated to be 1.3 million per year) than the number of infants delivered at <32 weeks of gestation (702 000; Table 3). This also highlights that, irrespective of gestation duration, upper genital tract infections during pregnancy are associated with adverse outcomes, and these findings should be a major consideration for microbiologists and clinicians, including obstetricians and neonatologists.

While this study has extended our knowledge of chorioamnion infection, it has some limitations. More than 50% of the women in this study were treated with intrapartum antibiotics. While antibiotic dose and type were not available for analysis, the use of antibiotics for prolonged periods (>3 hours) may have decreased the overall numbers of microorganisms isolated from these placentae, and therefore the rate of chorioamnion infection may be higher than is reported in our study. While we used both culture and molecular detection methods and isolated both cultivable and noncultivable microorganisms, additional investigation using more-sensitive techniques, such as deep sequencing, may be useful for determining the true prevalence of infection, to fully characterize infection throughout gestation.

In conclusion, the presence of chorioamnion infection is an important finding, regardless of gestation. This study demonstrated that chorioamnion infection directly correlates with chorioamnionitis and with adverse pregnancy and neonatal outcomes. This is also the first study to demonstrate that Ureaplasma species are independently associated with histologically confirmed chorioamnionitis and elevated levels of the cord blood cytokine G-CSF, regardless of gestation. These findings suggest that ureaplasmas are a cause of inflammation and that other immune factors may be involved in the severity of inflammation in the upper genital tract during pregnancy.

Notes

Acknowledgments. We thank the nurses at the Good Samaritan Hospital, particularly Peggy Walsh and Rita Doeger, for their assistance in the collection of our high-quality placental samples; the Research Flow Cytometry Core facility and Casey Wells, for performing and analyzing the cord blood cytokine data; and Manuel Alvarez Jr, for his assistance in the transfer of clinical samples from Cincinnati Children's Hospital Medical Center to Queensland University of Technology.

Financial support. This work was supported by the National Institute of Health (grant 1R01HL097064-105956).

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes