Objective To study the correlation between the bacteriological and histopathological findings in placentas from women with suspected or proven chorioamnionitis (CA).
Methods Over a 1-year period, 376 placentas were prospectively collected and processed for bacteriological and pathological studies in cases of confirmed or suspected maternal or neonatal infection.
Results Histological CA was diagnosed in 26.9% of placentas (101/376), and 27.7% (28/101) of these placentas had positive bacteriological cultures. A monomicrobial culture, mainly represented by Gram-positive cocci and Gram-negative bacilli, was identified in 27% of the positive bacterial cultures. The proportion of positive cultures was higher (p=0.03) when CA was associated with funisitis, as compared with placental samples with early CA. In placentas without histological CA, bacteriological cultures were mostly negative (230/275), although pathogenic bacteria were identified in 16.3% of them (45/275).
Conclusions The histological and bacteriological results were concordant in about 70% of the examined placentas, with 61.1% negative cases (CA absent and negative bacterial cultures), and only 7.4% placentas with positive histological and bacteriological results. Discordant results (positive histology with negative bacteriology) were obtained in placentas with early CA documented by histology although possibly in relation with antibiotic prophylaxis and the presence of fastidious bacteria. Conversely, negative histology with positive bacteriology could be explained by the presence of an early-stage bacterial infection that has not yet led to detectable microscopic lesions.
Statistics from Altmetric.com
Acute chorioamnionitis (CA) represents about 20% of fetal and neonatal pathology.1 It usually develops during the second, or at the beginning of the third trimester of pregnancy and is an important cause of preterm labour and delivery.2 CA is associated with strong fetal and intra-amniotic inflammatory responses and might represent a better independent predictor of neonatal sepsis than funisitis, which is considered as the final stage of the infection.3 Although it is accepted that CA is an acute infection due to micro-organisms that ascend along the endocervical canal, its physiopathology remains only partially clarified.
To this aim, we compared the results of bacterial cultures of subchorionic plate swabs with the histopathological data of 376 placentas that were collected in cases of confirmed or suspected maternal or neonatal infection.
Materials and methods
This prospective study was conducted on placentas from women who delivered at the Lausanne University Hospital and that were collected by the Lausanne University Institute of Pathology from January to December 2011. Among 1262 placentas sent to the department of pathology within this period, only 376 were processed for both bacteriological and pathological studies for clinical infection or high-risk situations of maternal and/or neonatal infection (table 1).
Unfixed placentas were sent to the department of pathology just after delivery, and were first processed for bacteriological studies. Swabs were taken from the undersurface of the amnion, after the fetal surface was sterilised using the hot blade of a scalpel, and sent to the Laboratory of Clinical Bacteriology, Institute of Microbiology (University of Lausanne and University Hospital Centre) to be processed as described below.
Macroscopic analysis of each placenta was performed after fixation in 10% buffered formalin solution for 48 h. After examination of the appearance, colour and insertion of the amniotic membranes, a 10 cm-wide membrane roll was rolled with the amnion inside, from the rupture site to the placental margin. Samples were taken at two levels of the roll, each containing the rupture site and the placental margin. Similarly, the length, diameter and insertion of the umbilical cord were examined, and three samples were taken (at the two extremities and in the middle). The placental disk was measured and weighed after removal of the membranes and cord, and the fetal and maternal surfaces examined. Placental tissue was sampled, each 5–10 mm, including the area at the cord insertion, and at least three full-thickness villous tissue samples. All samples were less than 1 cm thick. Microscopic examination was performed using H&E-stained sections.
The presence of CA was based on the detection of polymorphonuclear infiltrates within the amnion of the peripheral fetal membranes and/or the chorionic plate.4 Early CA was diagnosed when polymorphonuclear infiltrates were found within the subchorionic plate (figure 1A). In ‘classical’ CA (figure 1B), polymorphonuclear infiltrates were detected within the amnion, from the deep chorion through the fetal surface, and were sometimes associated with necrosis. Vasculitis was diagnosed when at least one chorionic plate vessel was invaded by polymorphonuclear cells (figure 1C). Funisitis was diagnosed when inflammatory cells were detected in the umbilical cord, infiltrating the cord vessels and migrating into the Wharton's jelly (figure 1D).
Swabs were used for Gram-stained slide preparation and to inoculate the following media: Columbia agar with 5% sheep blood, chocolate agar, Schaedler agar and thioglycolate broth. Bacterial isolates were identified using Matrix-assisted laser desorption-ionization time-of-flight (MALDI-TOF) mass spectrometry, and when necessary, with standard biochemical tests, including Vitek cards (BioMerieux, France).
Bacteria were grouped in three categories: (i) pathogens, (ii) conditional pathogens (colonising) and (iii) contaminants (table 2) according to the definition proposed.5 Besides pathogens, such as Haemophilus ducreyi, Listeria Mmonocytogenes, Trichomonas vaginalis that are always significant, other bacteria (Staphylococcus aureus, β-hemolytic streptococci) were considered as pathogenic based on the presence of polymorphonuclear neutrophils, and the absolute quantity of bacteria by direct microscopic examination and/or inspection of cultures on solid media. These bacteria were considered as colonising micro-organisms when obtained only after subculture in thioglycolate broth or in a low amount in primary culture. Environmental bacteria, including those that habitually live on human skin, were considered as contaminants.
For statistical analysis, we used the StatView software (StatView 512, Brain Power, Inc, Calabas, California, USA). Baseline variables were compared using the Pearson's χ² test (categorical data) or, when appropriate, the Fisher's exact test. p<0.05 Was considered statistically significant.
The population's characteristics are summarised in table 1. About 54% (203/376) of the 376 placentas included in the study were from preterm spontaneous deliveries, classified as reported.6 Delivery data were available only for 67% of women. Caesarean section was performed in 50% of these patients and labour was induced in 16.5%. Most of the patients (61%) received epidural anaesthesia and per partum antibiotics. Independently of the clinical context, antibiotics were given to all women delivered by a Caesarean section (129 cases), performed either under epidural (88/129) or general anaesthesia (41/129). Antibiotics were also given in cases of vaginal deliveries performed under peridural anaesthesia in a maternal context of infection (28 cases). Following the decision to prescribe or not, a placental swab was taken by the paediatric pathologist (CR, PY) in a context of suspected (135/226) or proven maternal infection (91/226), according to local guidelines (box). Neonatal indication (117/226) was mostly represented by respiratory distress (77/117).
Box Local guidelines for requesting placental swabs
Situations of suspected or proven maternal and/or neonate infection
Premature rupture of membranes
Maternal infection (sepsis, fever, urinary tract infection)
Maternal context of risk of infection (chronic endocervicitis, prior cervical surgery, positive screen for specific micro-organisms such as group B Streptococcus)
In utero fetal death after 14 weeks of gestation
Status postchorionic villous sampling
Neonatal infection (respiratory distress, wet lungs)
A diagnosis of histological CA was reached in 26.9% of cases (101/376). These placentas with CA were from term (early, plus full term) and preterm deliveries (extremely, plus very, plus moderate, plus late preterm) in a proportion of 42% (43/101) and 52% (53/101) of cases, respectively. Most of them were classical CA (54/101; 53.4%), whereas, early and necrotising CA were diagnosed in 28% (29/101) and 17.8% (18/101) of placentas, respectively. CA was associated with vasculitis and/or funisitis in 60% (60/101). Funisitis was observed in 61% of placentas with classical CA (95% CI 0.49 to 0.75), and 94% with necrotising CA (95% CI 0.84 to 1.04), more frequently (p<0.01) than in those with early CA (31%; 95% CI 0.14 to 0.48). CA was more often associated with maternal infection (74/101; 73%). Most of the women with histological CA were treated with antibiotics (75%; 95% CI 0.64 to 0.83), whereas, a smaller proportion (p=0.009) of women without chorioamnionitis was treated with antibiotics (56%, 95% CI 0.49 to 0.63). All women treated with antibiotics had histological CA, classical or necrotising CA (41/55, 74%) and early CA (9/55, 16%).
Bacterial cultures from the 376 placentas were positive with at least one pathogenic bacterium in 19.4% of cases (73/376). A monomorphic flora was identified in 65.7% of these positive cultures (48/73). Half the placentas with positive cultures were from preterm deliveries (38/73; 52%). Maternal infections accounted for 64.3% of the positive placental cultures (47/73). In addition, colonising bacteria from the perineal and vaginal florawere documented in 8.8% of cases (33/376), whereas, environmental contaminant bacteria were present in 12.2% of cases (46/303). Finally, a negative bacterial culture was obtained in 59% of cases (224/376).
A total of 193 micro-organisms were identified from 152 positive cultures. Pathogenic bacteria (table 2) represented 49.7% of all bacteria that have been documented (96/193), including L monocytogenes (1/96) as well as conditionally pathogenic bacteria from the P/V flora, among which Enterobacteriaceae (18.7%) and α-haemolytic streptococci (10.4%). Contaminant bacteria were found in 27.9% of cases (54/193) and coagulase-negative staphylococci were identified in about one-third of these cultures (19/54). Contaminant bacteria also included different non-fermentative Gram-negative bacteria, such as Pseudomonas aeruginosa and Stenotrophomonas maltophilia, different Gram-positive bacilli, such as Corynebacterium sp. and Propionibacterium acnes).
Correlation between placental bacterial cultures, histological and clinical findings
Overall, bacterial cultures were positive for 27.7% of placentas with CA (95% CI 0.19 to 0.35) and for 16% of placentas without CA (95% CI 0.12 to 0.2, p=0.01) (table 3). The proportion of positive bacterial cultures was higher (p=0.03) when CA was associated with funisitis (35%; 95% CI 0.22 to 0.47) in comparison with placentas with early CA (14%; 95% CI 0.01 to 0.26). The proportion of monomicrobial flora, mainly represented by Gram-positive cocci and Gram-negative bacilli, was similar in placentas with and without CA (box).
Bacterial cultures were negative in 72% of placentas with CA (73/101), and in 83% of placentas without histological evidence of CA (230/275). In placentas with CA, the proportion of sterile cultures was 69% in early CA (20/29; 95% CI 0.51 to 0.85), whereas, it was 41% (p=0.01) in placentas with classical CA or necrotising forms (30/72; 95% CI 0.29 to 0.52). Cultures with colonising bacteria from the P/V flora, but negative for pathogenic bacteria, were observed in 20.5% of placentas with CA (15/73; 95% CI 0.11 to 0.29) more frequently (p=0.002) than in 6% of those without CA (18/230; 95% CI 0.03 to 0.09).
The distribution of histological and bacterial findings with regard to gestational age (GA) is presented in table 4. In placentas from term deliveries (early, plus full term), bacterial cultures were positive and negative, respectively, in 30% (13/42) and 70% (29/42) of cases with histological CA, very similar to those from preterm deliveries (extremely, plus very, plus moderate, plus late preterm) in which the proportion of CA with positive and negative cultures was 25% (14/56) and 75% (42/56), respectively. In placentas from women that received antibiotic prophylaxis, bacterial cultures were positive and negative in 14% (23/157) and 85% (134/157) of cases, respectively. The proportion of histological CA with positive and negative cultures in women that were treated (positive cultures: 26% (15/56); negative cultures: 73% (41/56)) was similar as compared with the untreated ones (positive cultures: 21% (4/19), negative cultures: 78% (15/19)).
This work investigated the correlation between the placental histological data and the bacteriological findings from swabs taken from the amnion undersurface in patients with increased likelihood of CA.
Among the 376 placentas that were prospectively collected, histological CA was diagnosed in 101 placentas (26.9%). This finding is comparable with what has been previously described,7 ,8 with the exception of the work by Dong and coworkers who reported a CA rate of 85% in a series of 123 placentas from women at high risk of infection, with rupture of membranes that occurred more than 10 h before delivery.9
In the placentas with histological CA, 27% of bacterial cultures were positive, and the proportion of positive cultures was significantly higher, when CA was associated with funisitis (p=0.04). Conversely, the percentage of positive cultures was significantly lower in the presence of early CA (p=0.04). These results are comparable with those by Kim and coworkers 10 who showed a correlation between the severity of histological CA and bacterial quantity in the amniotic membranes. However, they quantified bacteria in the amniotic membranes by Q-PCR amplification, whereas, here we used both qualitative and semiquantitative criteria to interpret the results of bacterial cultures. With regard to GA, the proportion of positive cultures was very similar in placentas from preterm deliveries (14/56, 25%) compared with those coming from term deliveries (13/42, 30%). This differs from a recent study by Roberts et al11 who reported, in a series of term placentas with CA, that the incidence of infection would be very low, suggesting the possibility of a non-infectious disease. L monocytogenes was considered an obligatory pathogen based on the clinical data and the histopathological findings (microabscesses in the placental parenchyma) suggestive of a haematogenous spread. In addition, the identified bacteria were mainly conditional pathogens of the P/V flora, including Gram-positive cocci (24%) and Gram-negative bacilli (19%). Similarly, Bhola et al7 reported that, in a series of 252 placentas, the most frequently identified bacteria were group B Streptococcus and Escherichia Coli.
In placentas without histological CA, 16% of the bacteriological cultures were positive. Moreover, the proportion of monobacterial cultures and the micro-organism distribution were comparable with those observed in the group of placentas with histological CA. Several reasons might explain these observations: (1) absence of CA is a diagnosis that is based on the analysis of the available samples. These should be representative, but segmental/focal CA lesions might be overlooked due to sampling bias; (2) the absence of histological CA lesions could correspond to the very early phase of a bacterial infection before the development of a neutrophilic inflammatory response that may be observed by histology. Indeed, acute inflammatory reactions in human tissues are characterised by the migration of inflammatory cells, particularly, polynucleated neutrophils, with a peak 24 h after infection.12
Bacterial cultures were negative (sterile or presence of colonising and/or contaminant micro-organisms) in 303 placentas (table 3). In placentas with CA, sterile cultures were associated with early CA lesions in 40% of cases that corresponded to the stage 1 of the classification of the Society for Paediatric Pathology (Perinatal Section).13 The infectious nature of these early CA lesions is debatable, particularly because not all inflammatory reactions are due to an infection, especially when polynucleated leucocytes are not numerous.9 Indeed, deciduitis and early changes close to the rupture site may be observed in the absence of apparent intra-amniotic infection.13 The association with funisitis, which is considered as the final stage of ascending intrauterine infections,3 was 2.6 times less frequent (p=0.03) in these placentas than in placentas with classical and/or necrotising CA. The proportion of negative bacterial cultures with colonising micro-organisms from the P/V flora (box) was higher (p=0.002) in placentas with CA than in those without histological CA, suggesting a possible contamination of the placenta during the passage through the genital tract.
Discordant results (histologically documented CA and negative bacterial cultures) could be due to antibiotic prophylaxis that may have affected the viability of bacteria as early as 1 h after treatment, and thus, reduced the rate of bacterial recovery. Discordant results could also be explained by the presence of early CA lesions, the infectious nature of which is debatable, or by the presence of fastidious bacteria, such as Mycoplasma or strict intracellular bacteria, such as Chlamydia trachomatis. Both have been associated with adverse pregnancy outcomes and/or chorioamnionitis.14 ,15 However, their identification is difficult16 and was not routinely performed in the current study. In addition, the precise role of these bacteria remain unproven, or even speculative.17
In our prospective study on 376 placentas, the results of the histological and bacteriological analyses were concordant in about 70% of cases: in most of the placentas, histological CA was absent and the bacterial cultures were negative (n=230; 61.1%), whereas, in a minority of placentas (7.4%), both histological and bacteriological assays were positive. In 31.5% of placentas, the histological and bacteriological results were discordant; these cases corresponded mostly to placentas with early CA, the infectious nature of which was debatable (19.4%), or to placentas with early-stage bacterial infections without histological lesions.
The main weakness of this study is the long interval between the arrival of the placentas at the Institute of Pathology, and the combined delivery of the histological and bacteriological findings to the obstetricians (on average, 5 days). This might explain the inconstant feedback often observed from the obstetricians to the pathologists. The low culture yield in CA could be improved by the newer technologies for identifying pathogens.
In cases of suspected or proven chorioamnionitis (CA), the histological and bacteriological results were concordant in most of the examined placentas (70%), with 61.1% negative cases (CA absent and negative bacterial cultures) and only 7.4% positive cases (CA positive with positive bacteriological results).
Discordant results (placentas with positive histology and negative bacteriology) were obtained in the case of early CA documented by histology although possibly in relation with antibiotic prophylaxis and the presence of fastidious bacteria; conversely, negative histology and positive bacteriology could be explained by the presence of an early-stage bacterial infection that has not yet led to detectable microscopic lesions.
The low culture yield in CA could be improved by the newer technologies for identifying pathogens.
We thank Anna Maria Teixeira and Benjamin Massignac for their technical help. We are indebted to Elisabetta Andermarcher for editing the manuscript.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors Pathological studies were performed by VQM, PYN and CR. Most placentas were sent by PH, who is in charge of the Department of Obstetrics. Bacteriological studies were performed and interpreted by GP and GG. VQM and CR wrote the paper. CR is the corresponding author.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement As the corresponding author of this manuscript, I certify that this report is an original work, has not been published elsewhere, and will not be submitted to any other journal while it is under the consideration of the Journal of Clinical Pathology.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.