Placental histopathology in preterm birth with confirmed maternal infection: A systematic literature review

Four in five neonatal deaths of preterm births occur in low and middle income countries and placental histopathology examination can help clarify the pathogenesis. Infection is known to play a significant role in preterm birth. The aim of this systematic review is to explore the association between placental histopathological abnormality and preterm birth in the presence of confirmed infection. PubMed/Medline, Scopus, Web of Science and Embase were searched using the keywords related to preterm birth, placental histopathology and infection. Titles and abstracts were screened and the full texts of eligible articles were reviewed to extract and summarise data. Of 1529 articles, only 23 studies (13 bacterial, 6 viral and 4 parasitic) were included, and they used 7 different gestational age windows, and 20 different histopathological classification systems, precluding data pooling. Despite this, histopathological chorioamnionitis, and funisitis (when examined) were commonly observed in preterm birth complicated by confirmed bacterial or viral, but not parasitic, infection. The presence of malaria parasites but not pigment in placenta was reported to increase the risk of PTB, but this finding was inconclusive. One in three studies were conducted in low and middle income countries. An array of: definitions of preterm birth subgroups, histological classification systems, histopathologic abnormalities and diagnostic methods to identify infections were reported in this systematic review. Commitment to using standardised terminology and classification of histopathological abnormalities associated with infections is needed to identify causality and potential treatment of preterm birth. Studies on preterm birth needs to occur in high burden countries and control for clinical characteristics (maternal, fetal, labor, and placental) that may have an impact on placental histopathological abnormalities.

Introduction Preterm birth (PTB), occurring before 37 completed weeks' gestation, is recognised globally as the main cause of neonatal mortality and morbidity [1,2]. In 2014, an estimated 14.8 million babies (1 in 10), were born preterm; of these, 80% were born in Asia and sub-Saharan Africa [3]. However, accurate estimation of PTB rates, especially in low-and middle-income countries (LMICs) is often difficult because of the unavailability of ultrasonography for early and precise dating of the pregnancy [4].
PTB can be classified as: 1) spontaneous PTB with intact membranes, 2) preterm premature/pre-labour rupture of membranes (PPROM) or 3) indicated or induced due to maternal and/or fetal complications [5]. PTB is further categorised in the following subgroups: late preterm (�32 to <37 weeks' gestation), very preterm (28 to <32 weeks' gestation) and extremely preterm (<28 weeks' gestation), with the last group having the highest mortality and morbidity [6]. The associated morbidity includes a wide range of conditions with long-term impairment of motor-and neurocognitive development. Preterm neonates in high-income countries (HICs) have a much better chance of survival than those in LMICs [4,7]. In addition, the support services to care for related sequelae such as visual, hearing or learning difficulties, frequently do not exist in LMIC settings [8].
The placenta plays a key role in creating the optimal environment required to sustain a pregnancy until term, which is crucial for optimal neonatal outcomes. Changes in placental microanatomy can lead to severe consequences (e.g., PTB); hence, historically, histopathological examination has been an important method to understand the possible cause of PTB [9]. The placental inflammatory changes detected by histopathological examination are usually classified into acute or chronic inflammatory responses of the placenta, which can be further divided into maternal inflammatory response (chorioamnionitis) and fetal inflammatory response (umbilical vasculitis) [10]. Chorioamnionitis can be graded histologically by severity, chronicity and by fetal inflammatory responses that are observed in the chorionic plate and umbilical cord. Unfortunately, however, there is no consensus on placental sampling, tissue processing methods, or the recording and reporting of histopathological results [11]. The need for standardisation arises because common placental histopathological lesions are seen in PTB cases with and without infection. These include: villitis of unknown origin (destructive villous inflammatory lesion), acute or chronic chorioamnionitis (inflammation of chorioamnionic membrane), and chronic deciduitis (inflammation of decidua basalis) [12][13][14][15].
Placental histopathological abnormalities from intra-amniotic infection can be caused by "1) ascending bacterial infection from lower female reproductive tract (major pathway), 2) haematogenous spread of a bacterial, viral or parasitic infection through the placenta, 3) invasive procedures like amniocentesis, and 4) retrograde spread via the fallopian tubes, which is less likely to occur" [5]. The risk of spontaneous PTB also depends on the local vaginal microbial flora: microbiome studies have shown that elevated Ureaplasma or Gardnerella levels increase the risk of spontaneous PTB [5,16,17], while a vaginal flora rich in Lactobacilli is associated with a reduced spontaneous PTB risk [18]. Systemic infection in pregnancy, in particular infection of the urogenital tract and severe viral and malarial infections, are causally associated with PTB. Most notably, microbial genital tract infections are responsible for up to 25-40% of PTBs [19]. While the acute inflammatory response mostly occurs with ascending infection due to bacterial and fungal infections, haematological spread of viral and parasitic infections are more commonly a cause of chronic inflammatory response of the placenta resulting in chronic chorioamnionitis and chronic villitis [20,21]. While intervillous inflammatory infiltrates by mononuclear cells are associated with acute and chronic malaria infection [22], histopathological findings of malaria infection are more frequently characterised by the presence of malaria parasites and/or pigment in the intervillous space of the placenta but not histologic chorioamnionitis or funisitis [22,23].
Placental histopathological abnormalities can also arise due to PPROM and the length of time elapsed between membrane rupture and birth, which can make it difficult to differentiate PTB phenotypes based on the presence or absence of intrauterine acute inflammation/infection, immunologically mediated processes and uterine ischaemia [24]. This systematic literature review explores: (1) the association between placental histopathological abnormalities and PTB in the presence of confirmed bacterial, viral or parasitic infection, and (2) the possible association between the identified microorganism and specific placental histopathological abnormalities.

Materials and methods
The rationale, objective and search strategy of this systematic review were registered in the International Prospective Registers of Systematic Reviews (PROSPERO) under the registration number CRD42019137099 [25].

Search strategy
A systematic literature review following PRISMA guidelines [26] was conducted to identify studies reporting placental histopathology in PTB with a confirmed infection. Four databases: (PubMed/Medline, Scopus, Web of Science and Embase) were searched. The original search was conducted in March 2019 and continually updated until February 2021 focusing on three components: placental histopathology, PTB and infection. The search terms (S1 Table) included: "pregnan � AND (preterm � OR premature � ) AND placenta � AND (histopatholog � OR histology � patholog � ) AND (Infect � OR microorganism OR bacteria � OR virus � OR parasite � )". No temporal, geographical or language restrictions were applied to the search.
Interventional and observational studies were included if the histopathological findings were related to PTB and any type of infection (bacterial, viral or parasitic). The inclusion criteria used for full text screening were English language, full text availability, original research describing PTB/PPROM, placental histopathology assessed, micro-organisms reported, and most importantly that the study described the association between placental histopathology and infection in PTB/PPROM. In the event of discrepancies during the selection process, a second assessment was conducted to resolve the inconsistency. The bibliographic management software, EndNote (version X9.3.3), was used to organise the articles.
Studies that did not unambiguously include the three components in triangulation, i.e. histopathological examination, infection and PTB, were excluded. Histopathological placental findings reported in isolation or as part of a series of investigations were included if the other conditions were satisfied. Systematic reviews on histopathology and PTB were checked for any possible articles that may have been missed by the primary search.
After inclusion, the following data were extracted: year, country, economic country profile (HIC or LMIC), study design, timing of inclusion of participants (prospective at time of infection or at birth), estimated gestational age (EGA) of included PTBs, method of gestational age assessment, presence of rupture of membranes (ROM), sample size, histopathological classification and reporting system, confirmation of infection, and methodology and blinding of pathologist. Reference to the histopathological abnormality classification system was extracted when provided. High numbers of papers were excluded because they compared only two of three inclusion criteria in their analysis: placental histopathology and PTB, or microbial infection and histopathology, or microbial infection and PTB.
The methodological quality of the studies was assessed by evaluating study design, inclusion criteria and the blinding of investigators to pregnancy outcomes or causation. To address the risk of bias, the Newcastle-Ottawa quality assessment scale for case-control and cohort studies was used [27]. Two independent review authors (AMM and RM) appraised the selected studies; in the event of a discrepancy, a third reviewer (FHN) was consulted. Any encountered quality issues or concerns were addressed by referring to the particular section and grading the severity of the bias (moderate, serious or critical). Twenty selected clinical characteristics (maternal, fetal, delivery and infection) associated with PTB were extracted and a binary score was applied for presence or absence.

Subgroup allocation
As "infection" encompasses a broad range of causative pathogens and studies typically had a narrow range of interest, the term "infection" was divided into three groups according to the following infectious agents: bacteria, viruses or parasites.

Study characteristics
A total of 3277 records were identified through database searching. After removing duplicates, 1529 articles were screened against the defined eligibility criteria. Of these, 1407 were excluded after screening the study title and abstract: animal studies = 135; case reports, case series or review articles = 495; articles that did not report confirmed infection = 164; articles focused on microbiome, proteomic or serum markers = 366; multiple pregnancies = 13; not preterm = 91; placental histopathology not reported = 143. This left only 23 studies for inclusion (bacteria, viral and parasitic) (Fig 1) [28 -50].
Of included research articles that were published between 1988 and 2020, 15/23 studies (65.2%) were reported in the last 10 years. Only 8 (34.8%) were conducted in LMIC settings (Fig 2).
Most studies mentioned or clarified only a few of the clinical characteristics (maternal, fetal, delivery and infection) associated with PTB: median score of 6 [range 1-12] out of a possible total of 20 (S2 Table).

Variations of study participants
While an estimated gestational age of less than 37 weeks was the most common PTB definition, various lower cut-offs were used: seven different gestational age windows were described

Pathogen detection methods
Culture technique was used in 11 bacterial studies [31, 35-38, 41, 44, 46-49] and molecular methods were performed in three studies to identify the infective bacterial organisms [30, 39, [37,46,47]. For studies targeting human immunodeficiency virus (HIV) infection, two studies used an enzyme linked immunosorbent assay (ELISA) as a diagnostic test [33,42], while another two studies did not mention the diagnostic methods [29,45]. In another two studies that targeted viral infection, polymerase chain reaction (PCR) was performed to detect adenovirus, human papilloma virus (HPV) and enterovirus [32,50]. Two of three malaria studies confirmed infection by peripheral blood films, placental blood microscopy, and Loop-mediated isothermal application (LAMP) [28,40] but the histopathological detection of malaria parasites and pigment in placenta was used in all three malaria studies [28, 40,43].

Placental histopathologic findings in studies of preterm birth with bacterial infection
There were 13 studies that reported placental histopathology and PTB with bacterial infection. A wide range of bacterial species (n = 54) were reported with Ureaplasma species (7/13), Group B streptococcus (5/13) and Mycoplasma hominis (6/13) being the most common. Four articles specifically investigated only Ureaplasma and Mycoplasma species [31,38,41,44]. Another two studies examined the association between selected microorganisms (Chlamydia trachomatis and group B Streptococcus (GBS)) and PTB [39,46]. Histological chorioamnionitis was the most common placental histopathological abnormality in PTB with bacterial infection [29-31, 33, 35-39, 41, 42, 46-49]. Four studies reported both chorioamnionitis and funisitis [31,33,35,47]. Ingrid et al. used PCR to identify C. trachomatis and studied its association with placental histopathology in PTB. They reported that a chlamydia infection was two times as likely to be associated with placental histological inflammation in preterm cases compared to absence of chlamydia infection [39]. The rates of histological chorioamnionitis between GBS positive and negative groups of PPROM were not significantly different, as described by Patel et al. [46]. Pettket et al. reported that in preterm cases with a positive amniotic fluid culture, acute histological chorioamnionitis and funisitis were more pronounced when compared to preterm amniotic fluid culture negative cases [47]. Querios da Mota et al. showed that the proportion of histological chorioamnionitis with positive bacteriological culture in preterm placentas was 25% and more than 50% of microbial culture results came from preterm placentas [48]. Sweeney et al. also concluded that histological chorioamnionitis was more likely to be associated with moderate/late PTB, complicated by an infection with Ureaplasma species, the most common organisms identified in placental culture [49]. More detailed information on these studies is provided in Table 2.

Placental histopathology in studies of preterm birth with viral infections
The main placental abnormality in four of six viral infection studies was histologically confirmed chorioamnionitis [29,33,42,50]. Ategeka et al. reported that the risk of PTB was significantly increased with histological evidence of severe "maternal acute chorioamnionitis" in HIV infected pregnancies but not associated with "fetal acute chorioamnionitis" (maternal and fetal acute chorioamnionitis refers to maternal & fetal acute inflammatory response to infection in this article) [29]. Obimbo et al. reported that fibrinoid deposition with villous degeneration, syncytiotrophoblast delamination, increased red blood cell adhesion to terminal villi and increased number of capillaries were significantly associated with preterm placentas in HIV infection and/or antiretroviral therapy [45]. Gichangi et al. reported that the likelihood of having moderate to severe histological chorioamnionitis and moderate to severe funisitis (Continued ) was significantly increased in preterm placentas from HIV seropositive pregnant women compared to HIV negative women, and concluded that HIV infection was a risk factor for histological chorioamnionitis in preterm deliveries [33]. Feist et al. did not find an association between HPV and enteroviruses in the development of placental histological abnormalities (i.e., villitis of unknown origin and chronic deciduitis) in both preterm and term pregnancies [32]. Tsekoura et al. reported a significant association between histological signs of chorioamnionitis and an adenovirus infection in preterm placentas compared to adenovirus negative samples [50]. Other viruses commonly causing placental lesions and adverse birth outcomes include cytomegalovirus, herpes simplex virus, and parvovirus B19. However, no study investigating these viruses met the inclusion criteria of this review. More information on placental histological findings and viral infection can be seen in Table 3.

Placental histopathology in studies of preterm birth with parasitic infection
The placental histopathological abnormality with P. falciparum malaria infection in pregnancy was the presence of malarial parasites and/or pigment, while no histological correlate of chorioamnionitis was described in three malaria studies [28, 40,43]. Ategeka et al. [28] concluded that the risk of PTB was not increased with histological evidence of placental malaria pigment in malaria infected pregnant women from Uganda compared to those not-infected. However, placental malaria pigment is significantly associated with small for gestational age and low birth weight, and it was recommended to measure adverse birth outcomes in malaria intervention studies, especially in malaria high-transmission areas [28]. Kapisi et al. mentioned that the risk of PTB was significantly higher if malaria parasites were detected in the placenta in comparison with non-malaria infected cases, and placental malaria pigment was not associated with preterm deliveries [40]. Lufele et al. also reported an association between histopathologically confirmed placental malaria infection and adverse birth outcomes. A chronic infection, defined as the presence of malaria parasites and pigment deposition in monocytes and/or fibrin, was significantly associated with increased risk of PTB, while acute placental infection, defined as the presence of malaria parasites but not pigment, increased the risk of PTB at a non-significant level [43]. Saad et al. investigated an association between a Toxoplasma gondii infection and PTB by using the mean histopathological score based on abnormalities such as fibrin deposition, intervillous haemorrhage, villous degeneration and fibrosis, haemorrhage and nearby inflammatory cellular infiltrate and infarction [34]. They reported that the highest pathological score was observed in anti-toxoplasma IgM positive cases compared to anti-toxoplasma IgM/IgG negative cases [34]. No other studies on parasites such as soil transmitted helminths, Schistosomiasis, Trichomonas vaginalis, Giardia, Cryptosporidium were identified for this review. The detailed information on included studies is shown in Table 4.

Discussion
To our knowledge, this is the first systematic literature review to summarise placental histological abnormalities in PTB cases with a confirmed infection. The strength of this review is that it includes studies reporting the triad of placental histopathology, infection and PTB, allowing an inference on the causal pathways. However, the heterogeneity across studies meant it was not possible to combine the results in a meta-analysis as the included studies had different study designs, employed various diagnostic techniques to confirm causative pathogens with different test sensitivities, and used various systems for placental histopathological classification. Additionally, several PTB subcategories (spontaneous PTB with intact membranes, PPROM and indicated or induced due to maternal or fetal complications) were investigated in most of the reviewed studies, again preventing data pooling. Only 12 articles specified in the Methods section that gestational age was determined by ultrasound. Notably, only one in three studies were conducted in an LMIC, where neonatal mortality from PTB is highest [69]. Further limitations of this systematic review are that it fails to explore the association between specific causative microorganisms and specific placental histopathological changes due to study diversity; and that title and abstract screening of more than 1500 manuscripts has a possibility that some studies were inadvertently overlooked. Histological examination of the placenta is a very useful tool for assessing the multifactorial aetiology of PTB [70] as well as other adverse pregnancy outcomes (e.g., stillbirth and SGA) [71]. While placental tissue sample collection and the classification system of placental histological abnormalities varied, histological chorioamnionitis was common and strongly associated with PTB in the presence of both bacterial and viral infection in this systematic review. In malaria, histopathology can confirm infection, although it is not usually described as chorioamnionitis. The placental malaria confirmed by histopathological examination (presence of malaria infected red blood cells or malaria pigment in the intervillous space) is a better diagnostic tool for malaria infection rather than peripheral blood film due to subpatent infections and it is frequently used to assess the adverse pregnancy outcomes in malaria studies [21-23, 61, 63]. Lufele et al. found that chronic placental malaria infection increased the risk of PTB but the risk of PTB was not increased in patients with acute placental malaria infection [43]. Similarly, Kapisi et al. found that the risk of PTB was significantly higher if malaria parasites, but not pigment, were detected in the placenta. Although the association between PTB and placental malaria remains inconclusive, it was significantly associated with other adverse birth outcomes such as SGA and miscarriage [28, 40,43]. The only other study in the parasitic subgroup was of toxoplasmosis in which placental histopathological findings were more similar to acute and chronic inflammatory changes reported for bacteria, although the authors created their own detailed scoring system [34]. A methodological limitation in the process of diagnosing histological chorioamnionitis is that selection of a piece of placental tissue may lead to sampling of a segment with missing focal lesions in the placenta. Additionally, there is sparse evidence in humans on the time it takes for ascending intrauterine infection to result in detectable histologic chorioamnionitis which could have resulted in negative histopathology in the presence of confirmed infection [72]. It should be considered in future studies [48]. Pettker et al. concluded that histopathological examination of the placenta and microbiological tests fail to diagnose intra-amniotic inflammation and that caution is required in vaginal deliveries due to bacterial contamination or colonisation [47]. For an accurate interpretation of histopathological findings, clinical and procedural details should be reported. This includes exposure to antibiotics or antivirals, length of membrane rupture, number of vaginal examinations, and ascending bacterial infection. These important clinical factors were not adequately described in most of the included studies (S2 Table) and remain an area for improvement in future PTB studies.
Causality between histopathological changes in the placental tissue, PTB and infection can only be established if the causative pathogen is identified. The development of more sensitive, non-selective diagnostic tools over the last few years, provides researchers with additional options to unravel the link between the triad of PTB, infection and placental histopathological abnormalities [73]. Diagnostic tools reported in the reviewed studies included bacteriological culture, nucleic amplification techniques, immunohistochemical staining of placental tissue, with only one study using 16S rRNA sequencing [49]. Improved methods will increase the range of detected pathogens and lead to an improved understanding of gestational tissue infection. Contamination of the placenta and membranes can occur due to PPROM or during passage through the vagina, or skin during Caesarean section, and sequencing methods do not distinguish between viable and non-viable bacteria [74,75]. Some authors in this review indicated that the role of infection in PTB may be underestimated due to limitations in the diagnostic methods [31,47,49], the difficulty of distinguishing between colonisation and contamination [35], and decreased bacterial abundance due to antibiotic treatment prior to tissue sampling [32,34,35]. No manuscript in this review reached the ideal "gold standard" to link the pathogen and placental histopathology of "amniotic fluid culture/polymerase chain reaction (PCR) and placental pathology" [70]. Only one study included amniotic fluid culture [47] and only four used a PCR test to establish the presence of microorganisms [30,39,44,50].
Current consensus is that PTB is a syndrome, not a single clinical entity, that involves inherited predisposition, and maternal, fetal factors and placental factors, as well as signs of initiation of parturition and mode of delivery [76,77]. Suggestions to improve future research on this subject include: i) adherence to internationally recognised classification of PTB subgroups; ii) adherence to standard methods of placental sampling and histopathology classification (including severity); iii) a careful approach to study design as well as sensitive and specific infection detection methods, and iv) an effort to associate placental histopathology, PTB and infection.
Supporting information S1