Placental Pathology, Perinatal Death, Neonatal Outcome, and Neurological Development: A Systematic Review

Background The placenta plays a crucial role during pregnancy for growth and development of the fetus. Less than optimal placental performance may result in morbidity or even mortality of both mother and fetus. Awareness among pediatricians, however, of the benefit of placental findings for neonatal care, is limited. Objectives To provide a systematic overview of the relation between placental lesions and neonatal outcome. Data sources Pubmed database, reference lists of selected publications and important research groups in the field. Study appraisal and synthesis methods We systematically searched the Pubmed database for literature on the relation between placental lesions and fetal and neonatal mortality, neonatal morbidity and neurological outcome. We conducted three separate searches starting with a search for placental pathology and fetal and neonatal mortality, followed by placental pathology and neonatal morbidity, and finally placental pathology and neurological development. We limited our search to full-text articles published in English from January 1995 to October 2013. We refined our search results by selecting the appropriate articles from the ones found during the initial searches. The first selection was based on the title, the second on the abstract, and the third on the full article. The quality of the selected articles was determined by using the Newcastle-Ottawa Quality Assessment Scale. Results Placental lesions are one of the main causes of fetal death, where placental lesions consistent with maternal vascular underperfusion are most important. Several neonatal problems are also associated with placental lesions, whereby ascending intrauterine infection (with a fetal component) and fetal thrombotic vasculopathy constitute the greatest problem. Conclusions The placenta plays a key role in fetal and neonatal mortality, morbidity, and outcome. Pediatricians should make an effort to obtain the results of placental examinations.


Introduction
The placenta is the organ that links mother and fetus during pregnancy. It plays a crucial role in fetal growth and development by enabling the exchange of nutrients and oxygen from the mother to the fetus and removing fetal waste products. [1] The placenta is an endocrine organ, a site of synthesis and selective transport of hormones and neurotransmitters. In addition, the placenta forms a barrier to toxins and infective organisms. [2,3] In recent years, findings based on placental lesions have contributed to a better understanding of how the placenta functions. Less than optimal placental performance may result in morbidity or even mortality of both mother and fetus. Indeed, there are indications that placental lesions are the main cause of fetal death. [4] It is also becoming increasingly clear that impaired placental functioning can have major implications for the live-born infant. Awareness among pediatricians, however, of the benefit of placental findings for neonatal care, is limited. Usually, the results of placental examinations are only reported back to the obstetrician instead of also passing it on to the pediatrician. In our opinion, this is a missed opportunity. Information on placental lesions can often be helpful towards explaining an abnormal neonatal outcome and might have consequences for treatment.
This article provides a systematic review of the relation between placental lesions and neonatal mortality, morbidity, and neurological development. We summarized the literature published on this topic during the past 18 years. Our hypothesis is that placental examination provides useful information about the pathophysiological mechanisms that lead to neonatal mortality and morbidity. Should this prove to be the case, this information is important for the pediatrician who should, therefore, be aware of and take into consideration the placental findings of their patients.

Literature search
This systematic review was conducted following the PRISMA guidelines for systematic reviews. A registered systematic review protocol is not available. Two independent researchers (AMR and AFB) searched the PubMed database for literature on the relation between placental lesions and perinatal mortality, neonatal morbidity, and neurological development. We limited our search to full-text articles published in English from January 1 st 1995 to October 31 st 2013. We conducted three separate searches starting with a search for placental lesions and fetal and neonatal mortality, followed by placental lesions and neonatal morbidity, and finally placental lesions and neurological development.
Subsequently, we refined our search results by selecting the appropriate articles from the ones found during the initial searches in three stages. The first selection was based on the title, the second on the abstract, and the third on the full-text article. Review articles on the subject of placental lesions and outcome were indicated as background articles. We did not use these articles in the tables, but we did use them in the text of our article. We were mainly interested in single births, therefore articles focusing primary on multiple births were excluded. In addition to the database search, we screened the reference lists of the selected articles, and the publications of important research groups in the field.

Quality assessment
We assessed the quality of all the selected studies by means of the Newcastle-Ottawa Quality Assessment Scale for cohort and case-control studies. This assessment scale consists of three parts. For cohort studies these parts include selection, comparability, and outcome, for case-control studies selection, comparability, and exposure. The selection part consists of 4 items, with a maximum of 1 point per item. The comparability part has 1 item, with a maximum of 2 points for this item. Both the outcome and exposure parts consist of 3 items, with a maximum of 1 point per item. This provides a score, ranging from 0-9 points, with 9 points for the highest quality.

Placental pathology
Examination of the placenta can reveal a wide range of pathologies. For good reproducibility it is necessary that placental lesions are well defined. Committees of the perinatal section of the Society for Pediatric Pathology have proposed definitions for maternal vascular underperfusion, fetal vascular obstructive lesions (fetal thrombotic vasculopathy), and the amniotic infection syndrome. [59][60][61] Definitions and descriptions of additional pathologies can be found in various textbooks on the pathology of the placenta. [62][63][64][65][66][67] Since we acknowledge the fact that most pediatricians are unfamiliar with placental lesions and because a wide variety of terminology is used in the literature, we classified placental lesions according to the underlying pathology as previously proposed together with their pathological descriptions in Table 4. [35,42,[59][60][61][68][69][70][71] Placental lesions and perinatal mortality Perinatal mortality is defined as death during the perinatal period. In the 10 th Edition of the World Health Organization's International Classification of Diseases, the perinatal period is defined as death from 22 completed weeks of gestation up to 7 days after birth. [72] Fetal deaths form the largest group of perinatal mortality. In high-income countries one in every 200 infants that reaches 22 weeks' gestation or more, is stillborn. [73] Recently, the important role of the placenta in fetal deaths has become increasingly clear and several studies suggested that placental pathology is one of the main causes of fetal death (Table 5). This underscores the importance of examining the placenta, a fact sorely underestimated by obstetricians and general pathologists. [16] In 30% of the cases the cause of stillbirth is unknown. [73] In the remainder, i.e. the proportion of cases with known cause, most stillbirths are caused by placental lesions (12-65%, Table 5), followed by infections and umbilical cord abnormalities. [73] For lower gestational ages (GAs) (20 to 24 weeks), an unknown cause of death is most prominent, followed by placental lesions. At higher GAs, the relative importance of unknown causes decreases and placental causes increase. [73] Placental pathology consistent with maternal vascular underperfusion is the main contributor to fetal death, ranging from 34 to 38 percent. [4,8,14] This is most prominent during the preterm period, in pregnancies complicated by hypertensive disorders, with a strong decline thereafter. During the term period, fetal death is mainly caused by developmental pathology of placenta parenchyma. [4] We can conclude that a pathological examination of the placenta is essential for clarifying causes of stillbirths. [5,13,14,19] The older classification systems for perinatal mortality did not address placental pathology, or specific placental lesions, as a separate group. Only in the more recent classification systems is placental pathology included as a cause of fetal death. In all recent stillbirth studies placental pathology is designated as one of the main causes of fetal death. [74,75] The introduction of classification systems with placental pathology included as a separate group might be one of the reasons why recent studies identify placental pathology as one of the main causes of fetal death. Most of the placental lesions found in stillbirths, however, are also seen regularly after live, preterm or term, births. [76] The question arises whether placental lesions are also related to neonatal and neurological morbidity.
To summarize, in recent years the role of the placenta in fetal deaths has become increasingly clear. Placental pathology is one of the main causes of fetal death, with placental pathology consistent with maternal vascular underperfusion as the main contributor.

Placental lesions and neonatal morbidity
It has been suggested that placental lesions are also associated with neonatal morbidity, but the association is less clear than for fetal mortality. Placental lesions are suggested to be associated with illness severity shortly after birth, and with a wide range of neonatal problems (Table 6).
Illness severity shortly after birth can be determined by the presence of asphyxia, Apgar scores during the first minutes after birth, and by several clinical variables during the first 24 hours after birth. Perinatal asphyxia is described to be associated with placental lesions affecting fetal vascular supply. These lesions were umbilical cord complications (disrupted velamentous vessels, cord tear, hypercoiled cord, cord hematoma), chorioamnionitis with fetal vasculitis, and fetal thrombotic vasculopathy. [31,35] Low Apgar scores at 1 and 5 minutes are associated with ascending intrauterine infection and maternal vascular underperfusion. [22,26] Higher illness severity during the first 24 hours after birth, determined by the Score of Neonatal Acute Physiology Perinatal Extension (SNAPPE), is associated with placental pathological findings of fetal thrombotic vasculopathy and elevated nucleated red blood cells (a sign of hypoxia). [38]    Lung development and neonatal respiratory problems, such as neonatal respiratory distress syndrome (RDS) and bronchopulmonary dysplasia (BPD), are associated with placental inflammation. There are indications that the incidence of RDS is reduced in infants exposed to chorioamnionitis (ORs 0.1-0.6, 95% CI: 0.02-0.8). [23,29,37,77] This beneficial effect may be explained in several ways. It can be explained by advanced lung maturation in terms of an early elevation of interleukin-1 beta (IL-1b) in lung lavage fluid in the presence of chorioamnionitis, which stimulates the release of corticotrophin-releasing factor and corticotrophin. [78,79] These hormones enhance the production of cortisol which results in accelerated lung maturation and, therefore, a decrease in the incidence of RDS. [80] Lung maturation is also explained with animal models of fetal inflammation. Chorioamnionitis in the fetal lung induces elevated IL-1, which in turn increases the amounts of surfactant proteins in parallel with increases in surfactant lipids in bronchoalveolar lavages. The lung mesenchymal tissue decreases, which increases the epithelial surface area and airspace volume of the lung. This results in a more mature lung structure that contains more surfactant, has increased compliance, and supports better gas exchange. [77,81,82] Besides potentially a beneficial effect on lung function immediately after birth, an ascending intrauterine infection can also have a detrimental effect on the preterm lung, particularly in the long-term. [77] Chorioamnionitis can promote BPD, with ORs ranging from 2.0-7.4 (95% CI: 1.2-31.2). [23,26,37,40,77,83] BPD results from multiple antenatal and postnatal factors (hits) contributing to disease progression. [84] Despite a healthier initial condition (less RDS), the pulmonary status worsens during the postnatal period. [83] This is explained by an increased susceptibility of the lung to postnatal injurious events (second hits). [83][84][85][86] Even so, the relation between respiratory problems and chorioamnionitis is difficult to assess, since it is confounded by a variety of prenatal factors. [85] Necrotizing enterocolitis (NEC) is a challenging problem in the neonatal care of, mainly, preterm infants. The etiology of NEC is still poorly understood, but it is believed to be multifactorial. [87] Several studies found an association between NEC and placental lesions, in particular fetal vascular obstructive lesions (fetal thrombotic vasculopathy, congested villi, coagulation-related lesions) with ORs ranging from 2.6 to 9.10 (95% CI: 1.13-15.08). [26,32,33] The presence of ischemia has been proposed as an explanation for the etiology of NEC. Placental vasculopathy, which causes uteroplacental insufficiency, may cause fetal circulatory adaptive changes to hypoxia, which may result in bowel ischemia predisposing to NEC. [26] Retinopathy of prematurity (ROP) is also associated with placental lesions, in particular with inflammatory lesions with ORs ranging from 1.8 to 3.1 (95% CI: 1.02-9.5). [26,36,37,39,88] ROP affects preterm infants and is caused by disorganized growth of retinal blood vessels which may result in scarring and retinal detachment. The etiology of ROP is likely to be a multihit phenomenon. At least part of the multihit is an inflammationrelated pathogenesis, which is thought to be mediated by cytokines and growth factors present in the newborn's systemic circulation. [39] The severity of ROP also correlates positively with ascending intrauterine infection. [88] Fetal cardiac abnormalities are also thought to be associated with placental lesions. A six-fold increase in fetal cardiac abnormalities is reported in the presence of fetal thrombotic vasculopathy. [34] The most common cardiac abnormalities found in its presence are ventricular and atrial septal defects, cardiomegaly, and coarctation of the aorta. It is hypothesized that the relation may be explained by a causal link between the two   lesions. [34] The presence of one lesion may lead to the establishment of the other, through abnormal blood flow which serves as the common denominator. Another theory is that a common genetic variation underlies both placental fetal thrombotic vasculopathy and abnormal development of the heart. [34] This theory is supported by studies in mice which have shown placental and cardiac functions to be intimately linked, both through secretion of placental factors which affect maternal and fetal circulation and through genes which contribute to the development of both organ systems. [89] In addition, ascending intrauterine infection with both maternal and fetal response is associated with an increased risk for patent ductus arteriosus with ORs ranging from 1.7 to 5.53 (95% CI: 1. 1-19.27). [26,36,40] To summarize, the most important placental lesions in neonatal morbidity seem to be ascending intrauterine infection and fetal thrombotic vasculopathy. Nevertheless, caution is required in order to interpret these findings properly. Many studies on neonatal outcome only focus on infectious placental lesions and fetal thrombotic vasculopathy to the exclusion of other placental lesions. Thus there may be a bias towards these two lesions, because chorioamnionitis is a placental lesion well-known to both gynecologists and pediatricians. Even so, four of the larger studies including a wide range of placental lesions identified ascending intrauterine infection and fetal thrombotic vasculopathy as the most important placental finding with respect to neonatal morbidity. [22,26,32,40] This may pave the way for early interventions serving to prevent morbidity. Before such interventions can be defined, however, detailed knowledge of the pathophysiological mechanisms that lead to neonatal morbidity is required.
Regarding white matter injury, several studies failed to find a relation with histological AIUI (maternal and fetal response). [43,51] Nevertheless, in a meta-analysis, ascending intrauterine infection (clinical and histological) was indicated as a risk factor for white matter injury in preterm infants, with a relative risk of approximately 2.1. [91] The authors hypothesized that elevated cytokine levels play a role in the etiology of white matter brain lesions. The reason for the inconsistency of the results may be ascribed to differences in adjusting for potential confounders. Wu et al. [91] explained the effect of adjusting for gestational age. Although gestational age appears to be a possible confounder, it may also lie directly in the causal pathway between maternal infection and cerebral palsy (CP). Chorioamnionitis is associated with preterm delivery, and low gestational age is in turn associated with a host of intrinsic vulnerabilities within the brain that have been implicated in the pathogenesis of cystic PVL and CP. Therefore, if low gestational age resulting from maternal infection in itself plays a direct role in the pathogenesis of CP, then adjusting for its effect will falsely diminish the observed association between chorioamnionitis and CP. [91] Neonatal encephalopathy has mainly an antepartum, rather than an intrapartum, etiology. An important antepartum factor is placental pathology. [45,58] Placental lesions consistent with fetal thrombotic vasculopathy (OR 4.63, 95% CI: 2.01-10.68) and AIUI with a fetal response (funisitis) (OR 22.54 95% CI: 11.07-45.91) are both associated with neonatal encephalopathy. [45,58] Another less strongly associated placental lesion is accelerated villous maturation (disturbed uteroplacental flow) with an OR of 3.86 (95% CI: 1.36-10.92). [45] Elbers et el. studied placental pathology in relation to neonatal stroke. [53] They systematically described their findings in twelve cases of neonatal stroke, ten of which had placental lesions. They found the following types of lesions: thromboinflammatory process in six cases, sudden catastrophic event in five cases, decreased placental reserve in three cases, and stressful intrauterine environment in two cases. They suggested that multiple risk factors are involved in neonatal stroke, and that placental pathology may be a contributing factor. [53] The Extremely Low Gestational Age Newborns (ELGAN) investigators studied the predictive value of placental pathology in regard to white matter damage and later CP. They found histologic inflammation to be predictive of ventriculomegaly and diplegic CP, with ORs ranging from 1.4 to 1.5 (95% CI: 1.0-2.4) and ORs 2.3-3.4 (95% CI: 1.1-7.4), respectively. Placental inflammation was not predictive for echolucent lesions. [52] Also fetal thrombotic vasculopathy is found to be associated with CP. In the presence of FTV and CP, obstructive umbilical cord abnormalities have been identified. These umbilical cord abnormalities can lead to fetal placental vascular stasis resulting in fetal thrombotic vasculopathy. [42,46,92] Macroscopic examination of the placenta can also identify an increased risk of CP. Placental infarction thus identified is associated with an increased risk of the spastic quadriplegic subtype of CP (OR 2.6, 95% CI: 1.2-5.6). [56] The pathophysiological mechanism of placental infarction leading to CP is not clear. It is stated that because of the many functions and substantial functional reserve of the placenta, it cannot be assumed that placental infarction acts mainly by interference with gas exchange. A hypothesis is that, whatever the underlying   process that harmed the vasculature of the placenta causing infarction, the same process may also have directly harmed either the fetal cerebral vasculature or the brain. [56] Results on the association between placental pathology and long-term neurological outcome, including developmental tests and functional outcome, are also inconsistent between studies. In preterm infants it is thought that neurological impairment is associated with recent non-occlusive thrombi of the chorionic plate vessels in combination with chorioamnionitis and severe villous edema. Chorioamnionitis alone is not associated with neurological impairment. [41] This was attributed to the strong and consistent relationship between neurologic impairment and chorionic plate thrombi that occur only in placentas with chorioamnionitis. [41] Placental pathology consistent with maternal vascular underperfusion was also found to be a risk factor for neurological impairment, with ORs ranging from 7.4 to 10.1 (95% CI: 1.6-46.3). [48] For term infants, Redline et al. reported an association between neurological impairment and ascending intrauterine infection with a fetal response (OR 2.9-13.2, 95% CI: 1.2-144). [42,46] In addition to AIUI with a fetal response, they found that the following lesions are present significantly more often in placentas of infants with neurological impairment: meconium associated vascular necrosis, chorionic vessel thrombi, increased nucleated red blood cells (sign of fetal hypoxia), findings consistent with abruption placenta, diffuse chronic villitis, extensive avascular villi, diffuse chorioamnionic hemosiderosis, and perivillous fibrin. [42] Neurodevelopmental outcome of preterm-born children at toddler age is also associated with ascending intrauterine infection with a fetal response (funisitis). In the presence of funisitis, a higher incidence of moderate to severe disability is present with an OR of 4.08 (95% CI: 1. 16-14.44). [54] In addition, speech abnormalities and hearing loss are associated with AIUI (ORs 2.9-5.1, 95% CI: 1.2-19.4 and OR 11.6, 95% CI: 1.3-105.9, respectively. A study comparing neurodevelopmental outcome at two years of age between very preterm infants with maternal vascular underperfusion and very preterm infants with histological chorioamnionitis found poorer mental development in infants with maternal vascular underperfusion compared to infants with chorioamnionitis. [57] Neurocognitive outcome of preterm-born children at school age is associated with villous edema (OR 4.7, 95% CI: 1.1-19.2). Lower scores on mental processing and on neuropsychological assessment are found in its presence. In this study, ascending intrauterine infection is not predictive of impaired neurodevelopmental outcome in the population as a whole, but a severe funisitis is associated with lower scores on neurocognitive tests in the subpopulations with ascending intrauterine infection. [48] In summary, despite the difficulties in studying the relation between placental lesions and neurological morbidity, and the inconsistent results, some conclusions can be drawn. For those studies finding a relation with poor neurological outcome, the placental lesion is ascending intrauterine infection with a fetal response. Furthermore, in term infants a larger variety of placental lesions seem to be associated with poor neurological outcome compared to preterm infants. Knowledge on the pathophysiological mechanisms leading to long-term neurological deficits may lead to possible interventions to improve outcome. The fact that the placenta is available for histological examination immediately after birth and that it may reveal valuable information for pediatricians, leads to an early opportunity to intervene to the benefit, hopefully, of ill neonates.

Discussion/Conclusion
The placenta plays a key role in fetal and neonatal mortality, morbidity, and outcome. Placental lesions are one of the main contributors to fetal death. In these cases placental lesions consistent with maternal vascular underperfusion are most important. Although less clear-cut, several neonatal problems are also associated with placental lesions. Regarding neonatal morbidity and neurological outcome, placental lesions with ascending intrauterine infection (with a fetal component) and fetal thrombotic vasculopathy, constitute the greatest problem.
To our surprise we noticed a difference in the description of placental lesions between studies on perinatal death and studies on neonatal outcome. The majority of studies on placental pathology and stillbirth only focus on the presence or absence of placental lesions in general, but they do not examine the relation between specific placental lesions and stillbirth. Studies concerning placental lesions and neonatal or neurological outcome do specify the lesions, finding several relations between specific placental lesions and outcome. Characterizing placental lesions in more detail in stillbirth studies may provide additional information concerning the cause of death.
Most studies report on associations between placental lesions and outcomes but this does not necessarily reflect a causal relation. There is still need to clarify pathophysiological mechanisms. One of these proposed mechanisms include gene-environment interactions. [92] Placental lesions might already have their onset early in pregnancy, due to changes in placental genes, leading to epigenetic alterations. Causes for these placental epigenetic changes may include a non optimal intrauterine environment, due to a maternal disease or adverse insults to the intrauterine environment. [93] This may in turn cause placental dysfunction and hence adverse neonatal outcome. We thus have to take into account that multiple interactions from maternal, placental, and fetal health play a role in the etiology of perinatal death and neonatal morbidity. Future research must consider statistical tools to better address interactions among these multiple variables, such as a mixed-effect regression analyses for example.
There are several limitations to our systematic review. Firstly, there is a potential risk of publication bias. Studies finding negative results regarding placental lesions and outcome might not be published. This may lead to an overestimation of associations between placental lesions and outcomes. Secondly, we included studies from the past 18 years. Earlier studies might have had different results. Finally, most studies included in this review were conducted in high-risk populations. Studies in a low-or moderaterisk group may reveal different results.
A final point we would like to address is an urgent need for increasing awareness among pediatricians for placental lesions and neonatal outcome. The obstetrician sends the placenta to the pathologist for histological examination. The results of the examination are reported back to the obstetrician. In most cases the pediatrician is unaware of the results of the placental examination. In the light of the accumulating evidence, however, that placental pathology is associated with perinatal mortality, neonatal morbidity, and neurological outcome, pediatricians should make an effort to obtain the results of placental examinations. Placental pathology, ascending intrauterine infection, and fetal thrombotic vasculopathy in particular, may help to identify the group of neonates at risk of adverse neonatal outcome. Monitoring these infants more closely could be helpful. Knowledge of the pathophysiological mechanisms leading to neonatal mortality and morbidity may lead the way to finding early intervention strategies to improve infants' morbidity and outcome.

Supporting Information
Checklist S1 PRISMA flowchart of identified articles published between January 1995 and October 2013. (DOC)