Authors: Stephanie A. FISHER (1Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, Feinberg School of Medicine, Northwestern University, Chicago, IL), Xiaoshuang XUN (2Department of Epidemiology, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD), Alison GEMMILL (3Department of Population, Family and Reproductive Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD), Lynn M. YEE (1Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, Feinberg School of Medicine, Northwestern University, Chicago, IL), Leena MITHAL (4Department of Pediatrics, Division of Pediatric Infectious Diseases, Ann and Robert H. Lurie Children’s Hospital of Chicago and Feinberg School of Medicine, Northwestern University, Chicago, IL), Aaron HAMVAS (5Department of Pediatrics, Division of Neonatology, Ann and Robert H. Lurie Children’s Hospital of Chicago and Feinberg School of Medicine, Northwestern University, Chicago, IL), Raye-Ann de REGNIER (5Department of Pediatrics, Division of Neonatology, Ann and Robert H. Lurie Children’s Hospital of Chicago and Feinberg School of Medicine, Northwestern University, Chicago, IL), Judy L. ASCHNER (6Department of Pediatrics, Hackensack Meridian School of Medicine, Center for Discovery and Innovation, Nutley, NJ), Nathalie L. MAITRE (7Department of Pediatrics, Emory University School of Medicine, Atlanta, GA), Thomas G. O’CONNOR (8Department of Psychiatry, University of Rochester Medical Center, Rochester, NY), Whitney COWELL (9Departments of Pediatrics and Population Health, New York University Grossman School of Medicine, New York, NY), Linda G. Kahn (9Departments of Pediatrics and Population Health, New York University Grossman School of Medicine, New York, NY), Roger B. NEWMAN (10Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, Medical University of South Carolina, Charleston, SC), Richard K. MILLER (11Departments of Obstetrics and Gynecology, Environmental Medicine, Pathology and Laboratory Medicine, and Pediatrics, University of Rochester Medical Center, Rochester, NY), Carolyn SALAFIA (12Department of Pediatrics, Placental Analytics, LLC, New Rochelle, NY), Amy J. ELLIOTT (13Avera Research Institute & Department of Pediatrics, University of South Dakota School of Medicine, Sioux Falls, SD), Anne M. SINGH (14Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, Madison, WI; 15Department of Dermatology, University of Wisconsin School of Medicine and Public Health, Madison, WI), Nicole BAUMANN-BLACKMORE (14Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, Madison, WI), Jeffery GOLDSTEIN (16Department of Pathology, Northwestern University Feinberg School of Medicine, Chicago, IL)
Categories: Article, placenta, histopathology, perinatal pathology, maternal-fetal interface, neurodevelopment
Source: American journal of obstetrics and gynecology
Authors: Stephanie A. FISHER, Xiaoshuang XUN, Alison GEMMILL, Lynn M. YEE, Leena MITHAL, Aaron HAMVAS, Raye-Ann de REGNIER, Judy L. ASCHNER, Nathalie L. MAITRE, Thomas G. O’CONNOR, Whitney COWELL, Linda G. Kahn, Roger B. NEWMAN, Richard K. MILLER, Carolyn SALAFIA, Amy J. ELLIOTT, Anne M. SINGH, Nicole BAUMANN-BLACKMORE, Jeffery GOLDSTEIN
Prenatal exposures influence childhood neurodevelopment. Placental histopathology has been associated with abnormal early childhood neurodevelopment, albeit often confounded by prematurity and/or fetal growth restriction. Most pregnant people, however, have term births, and some of these children have abnormal neurodevelopment despite the absence of adverse birth outcomes. Leveraging placental histopathology may help distinguish infants at higher risk of subsequent neurodevelopmental impairment following a term birth.
To investigate the association of placental histopathology with a high-risk screen for abnormal early childhood neurodevelopment following a term birth.
The sample included singleton births at ≥37 weeks 0 days between 2020–2023 in the prospective, longitudinal multisite Environmental Influences on Child Health Outcomes (ECHO) Cohort. Children with available placental histopathologic data and whose birthing parent had completed at least one Ages & Stages Questionnaire-Third Edition (ASQ^®^-3) between 2–18 months of life were eligible for inclusion. Children diagnosed with hypoxic-ischemic encephalopathy after birth were excluded. Exposures were chronic placental inflammation, maternal or fetal acute inflammatory response, and maternal or fetal vascular perfusion. The primary outcome was a high-risk composite ASQ^®^-3 screen, defined as a high-risk score (≥2 standard deviations below the mean) on at least one of the five individual domains (communication, gross motor, fine motor, personal-social, and problem solving) on any ASQ^®^-3 questionnaire performed between 2–18 months of life. Individual ASQ^®^-3 domains were secondarily assessed. Generalized estimating equation models were used to calculate the odds of a high-risk screen for each outcome in children exposed versus unexposed to each placental histopathologic finding, adjusted for maternal age, education, insurance, depression, parity, child sex, and birthweight.
At ECHO sites performing placental collection and histopathologic evaluation, assessment of at least one ASQ^®^-3 domain was performed in 7353 children ages 2–18 months during the study period. Of these, 486 (13%) were born at term and met additional eligibility criteria. Pregnant participants self-identified predominately as non-Hispanic White (57%), exceeded a high school education (78%), and were multiparous (70%). The frequency of each placental histopathologic exposure ranged from 16.5–59.5%, and the primary outcome of a high-risk composite ASQ^®^-3 screen was present in 26% of children. In multivariable analyses, none of the placental exposures were associated with a high-risk composite ASQ^®^-3 screen (aOR 1.43, 95%CI 0.95–2.15) at 2–18 months. However, chronic placental inflammation was associated with high-risk communication (aOR 2.84, 95%CI 1.09–7.40) and fine motor (aOR 2.26, 95%CI 1.02–5.04) domain scores at 2–18 months, and with a high-risk screen for the composite ASQ^®^-3 score (aOR 2.07, 95%CI 1.05–4.07) and gross motor domains (aOR 3.89, 95%CI 1.25, 12.10) at 12–18 months. In post-hoc sensitivity analyses, associations between chronic placental inflammation and high-risk ASQ^®^-3 screens varied by child sex and were not present in individuals without obesity (BMI <30kg/m^2^).
After a term birth, placental histopathology was not associated with a high-risk composite ASQ^®^-3 screen in children assessed at 2–18 months. However, chronic placental inflammation was positively associated with a high-risk composite score in children aged 12–18 months. This population may warrant enhanced surveillance, screening, and diagnostic follow up for neurodevelopmental impairment in early childhood.
Prenatal exposures influence childhood health, including neurodevelopmental, behavioral, and psychological development.^1^ The placenta is a multifaceted exchange organ that supplies the fetus, particularly the fetal brain, with oxygen and nutrients.^2^ Placental and fetal brain development occur in parallel, and clinical studies suggest that in utero stressors may lead to both placental abnormalities and neurobehavioral disorders.^3^ Placental structure and function thereby play a crucial role in determining normal and abnormal fetal, neonatal, and early childhood neurological development.^4,5^ Examining placental abnormalities may elucidate processes underlying abnormal childhood cognitive, behavioral, and emotional development.^1^
Using heterogeneous screening tools, previous studies have identified that select placental histopathologic abnormalities are associated with diverse adverse neurodevelopmental outcomes in early childhood, including hearing and speech disabilities; motor, cognitive and/or executive dysfunction; autism spectrum disorder (ASD), attention deficit hyperactivity disorder, and other behavioral and/or psychiatric disturbances.^6,7^ Associations vary by placental histopathology type. For example, acute and chronic placental inflammation have been associated with cerebral palsy, low neurodevelopmental indices, and long-term neurologic impairment.^8–12^ Maternal vascular malperfusion has been associated with decreased Bayley II mental development index at 2 years and Mullen Scales of Early Learning composite scores from 10–40 months.^13–15^ Fetal vascular malperfusion has been associated with lower scores on the Griffiths’ Mental Developmental Scales–Extended Revised at 24 months.^16^ However, these studies are limited by small sizes and the retrospective nature of these cohorts.^6,7^ Notably, most of these studies evaluated placental pathology from pregnancies complicated by preterm birth, fetal growth restriction, low birthweight, and/or neonatal intensive care unit admission, all of which likely bias the associations observed between placental histopathology and neurodevelopmental outcomes.^6,7^
Evaluation of the association of placental histopathology with the risk of abnormal early childhood neurodevelopment within larger prospective cohorts of term births with rich clinical and covariate data is necessary to account for these potential biases, as the majority of pregnant people give birth to healthy, term infants with appropriate-for-gestational age birthweight. Yet, some of these children experience abnormal neurodevelopment in the absence of an adverse birth outcome. Leveraging data from the Environmental Influences on Child Health Outcomes (ECHO) Cohort, the central objective of this study is to investigate the association of placental histopathologic lesions in term births with a high-risk screen for abnormal early childhood neurodevelopment and explore if histologic evaluation in term births offers insight into risk of future neurodevelopmental impairment. We hypothesized that acute and/or chronic inflammatory and vascular lesions in the placenta would be associated with aberrant cerebral development in utero and a subsequent high-risk screen for neurologic impairments in early childhood.
This study is an analysis of data collected between 2020–2023 in Cycle 1 of the ECHO Cohort, a United States national, multisite, prospective longitudinal cohort assembled with the aim to improve knowledge of early environmental factors that influence child health and development.^17^ We included pregnancies from the ECHO Cohort in which live birth occurred at ≥37 weeks and 0 days gestational age based on best available obstetric dating information available, had placental histopathologic data available, and had at least one Ages & Stages Questionnaire-Third Edition^18^ (ASQ^®^-3) completed by a parent between 2–18 months of life. We excluded multifetal gestations, pregnancies with uncertain pregnancy dating or a severe fetal malformation(s), and children diagnosed with hypoxic-ischemic encephalopathy or a chromosomal or other genetic condition.
Of five ECHO sites that collected placental data in Cycle 1, this analysis included three study Childhood Allergy and the Neonatal Environment (CANOE)^19^, Safe Passage Study (PASS)^20^, and The New York University Children’s Health and Environment Study (NYU CHES).^21^ The Boricua Youth Study cohort had placental histopathologic evaluation and ASQ^®^-3 data available for 12 maternal-child dyads, but was excluded from this analysis due to small sample size (Figure).^22^ The Fair Start Study birth cohort had placental data available from 97 additional pregnancies, but did not perform ASQ^®^-3 assessments.^23^
The study protocols were reviewed and approved by the ECHO Institutional Review Boards (IRBs, local or central). All participants provided written informed consent. The IRB at Johns Hopkins Bloomberg School of Public Health approved the involvement of the ECHO Data Analysis Center (DAC).
Exposures of interest were chronic placental inflammation, maternal acute inflammatory response, fetal acute inflammatory response, maternal vascular perfusion, and fetal vascular malperfusion. At the selected ECHO Cohort study sites, placenta samples were obtained at birth according to their local protocol,^19–21^ fixed in formalin, then shipped to the ECHO Cohort’s center for processing. Hematoxylin and eosin-stained slides were prepared and reviewed centrally by an experienced perinatal pathologist (CS). Eight sections of the placental disc, two in each quadrant, were taken to maximize sensitivity, in addition to two samples from the umbilical cord, two samples directly under the placental cord insertion site, two membrane rolls, and one additional sample taken from any gross lesions.^24^ Histopathologic lesions were identified in accordance with the Amsterdam Placental Workshop Group Consensus Statement, and coded as present or absent.^25^ Chronic placental inflammation was defined as the presence of one or more of the following six histopathologic 1) chronic deciduitis (i.e., plasma cell infiltrates present in the decidua), 2) chronic chorioamnionitis (i.e., lymphoplasmacytic infiltration of the chorion and amnion), 3) basal chronic villitis (i.e., infiltration of anchoring villi attached to the basal plate by lymphocytes, histiocytes, and/or plasma cells) 4) chronic chorionic vasculitis (i.e., eosinophilic T-cells seen crossing fetal vessels in the chorionic plate), 5) chronic villitis of unknown etiology (i.e., infiltration of fetal stem villi by lymphocytes, histiocytes, and/or plasma cells), and 6) chronic chorionitis (i.e., lymphocytic or lymphoplasmacytic infiltration of the chorionic plate alone).^26,27^ Maternal acute inflammatory response was defined as the presence of neutrophils in the extraplacental membranes and/or chorionic plate, whereas fetal acute inflammatory response was defined as the presence of neutrophils in the umbilical cord and/or chorionic plate vessels.^26^ Maternal vascular malperfusion was defined as the presence of accelerated villous maturation or decidual arteriopathy, intervillous thrombi, villous infarcts, and/or chronic or acute abruption within the placenta.^28^ Fetal vascular malperfusion included the presence of avascular villi, villous stromal vascular karyorrhexis, hemorrhagic endovasculitis, fetal stem villous mural thrombosis, and/or fetal large vessel mural thrombosis in the chorionic plate vessels.^28^
Each study site administered the ASQ^®^-3 questionnaire, completed by the child’s birthing parent typically within 10–15 minutes, to screen children for risk of abnormal early childhood neurodevelopment in the first 18 months of life. The CANOE^19^ and PASS^20^ sites administered the ASQ^®^-3 in English only; NYU CHES^21^ administered the ASQ^®^-3 in English, Spanish, and Chinese. The primary outcome was an abnormal, or high-risk, composite ASQ^®^-3 screen, defined as having a high-risk score on at least one of the five individual domains (communication, gross motor, fine motor, personal-social, and problem solving) on any ASQ^®^-3 questionnaire completed between 2–18 months of life. The individual ASQ^®^-3 domains were secondarily assessed, also as binary outcomes (high-risk versus low-risk screen). A high-risk score was defined as greater than or equal to two standard deviations below the mean; all other scores were considered low-risk. If a a child had multiple ASQ^®^-3 questionnaires completed over this period, the lower score was used for the composite and individual domain scores.
Each ECHO Cohort site collected covariate information, which was harmonized by the ECHO DAC. We identified potential confounders and precision variables for consideration in analytic models based on expert input and literature review. We considered the following variables as potential covariates for site, maternal age at delivery, highest level of maternal education, health insurance (public, private, or both), family structure, prenatal tobacco, alcohol or other substance use during pregnancy, maternal depression or anxiety (at any point pre-pregnancy, during pregnancy, or through 8 weeks postpartum by medical records or self-report), parity, antenatal corticosteroid exposure, delivery indication (spontaneous or medically indicated), group B streptococcus status, child sex, and neonatal birthweight. Based on data availability, the following variables were included in the final multivariable maternal age at delivery, highest level of maternal education, insurance, maternal depression, parity, child sex, and neonatal birthweight. Neonatal birthweight was included as a precision variable given that low birthweight, fetal growth restriction, and/or small-for-gestational age (SGA) is not always related to placental dysfunction and may be constitutional in approximately 20% of cases. Yet, birthweight has been identified as an important prenatal factor associated with abnormal early childhood neurodevelopment.^6,7,29–31^ Although we report the distributions of hypertensive disorders (chronic hypertension, gestational hypertension, or preeclampsia), pre-pregnancy body mass index (BMI, both as a continuous and categorical measure), and pregestational or gestational diabetes in the analytic cohort, these maternal factors were not included as covariates as they may contribute to placental histopathology.^32–37^ We additionally report on paternal age, self-reported maternal race and ethnicity, annual household income, gestational weight gain, gestational age at birth, mode of birth (vaginal or cesarean), neonatal birthweight percentile (SGA, birthweight <10%ile; large-for-gestational age, birthweight >90%ile), and 5-minute Apgar score.
We report descriptive statistics of all demographic and clinical characteristics for the overall cohort and for each individual cohort. Pearson’s Chi-square or one-way ANOVA tests, as appropriate, were used to compare baseline characteristics across the three included cohorts. Additionally, we report the median (interquartile range, IQR) child age at time of completion of the ASQ^®^-3 questionnaire. For each placental histopathologic exposure, the proportion of children with, compared to those without, the primary composite and individual secondary outcomes was assessed using Pearson’s Chi-square test. We further report the proportion of low- and high-risk ASQ^®^-3 screens in ECHO participants with at least one ASQ^®^-3 assessment performed at 2–18 months regardless of availability of placental exposure data) with those in the analytic subset.
Using generalized estimating equation (GEE) regression models with a binomial link function and clustering based on ECHO Cohort study site, univariable and multivariable analyses were performed to generate the covariate-unadjusted and adjusted odds and 95% confidence intervals (CI), respectively, of a child having a high- versus a low-risk screen for the primary and secondary outcomes based on the presence of each placental histopathologic exposure. We imputed missing covariates using multiple imputation by chained equations (MICE) with ten imputed datasets and five iterations. Univariate and multivariable analyses were repeated in a pre-specified sensitivity analysis evaluating only the subset of children who were appropriate-for-gestational age with respect to size at birth (i.e., neonatal birthweight 10^th^-90^th^ percentile) according to the INTERGROWTH-21^ST^ fetal growth standards to exclude SGA infants given heterogeneous factors associated with early childhood neurodevelopment in the SGA population.^29–31,38^
We further conducted a sensitivity analysis excluding individuals with pre-pregnancy BMI ≥30kg/m^2^ to exclude any potential influence of obesity given the established links between maternal obesity and placental inflammation.^32,39^ Based on expert recommendation, clinical differences observed in early childhood development by child age, and improved validity of the ASQ^®^-3 in children at older testing ages, we then examined models stratified by age at ASQ^®^-3 assessment (<12 versus 12–18 months).^40,41^ As offspring sex differences have been associated with differential risk of adverse pregnancy and child neurodevelopmental outcomes, with greater morbidity observed among males, we separately examined models stratified by child sex (male versus female).^42,43^ Finally, we performed univariable and multivariable regression models using GEE for all outcomes based on exposure to each of the six individual histopathologic subtypes of chronic placental inflammation.
All statistical analyses were conducted using R (version.4.4.0; R Core Development Team).^44^ A p-value <0.05 was used to determine statistical significance. Due to the exploratory nature of this analysis, correction for multiple comparisons testing was not performed.
In the ECHO Cohort, 7353 children underwent assessment of at least one domain of the ASQ^®^-3 questionnaire during the study period. Among these, 503 also had placental evaluation, conducted from March 2020 to July 2023 (Supplementary Table 2), of which 486 had the ASQ^®^-3 assessed between 2–18 months and met all other eligibility criteria for inclusion in the analytic cohort (Figure). In the analytic cohort, 57% (273/483) of pregnant participants self-identified as non-Hispanic White, 78% (367/470) reported a greater than high school educational level, 41% (191/468) had full or partial public insurance coverage, and 70% (329/469) were multiparous. Twenty percent reported a history of depression (97/477), 74% were non-obese (322/437), and 77% (370/483) gave birth to a neonate with appropriate-for-gestational age birthweight. Most baseline characteristics differed among participants across the three individual sites (Table 1).
The frequency of placental histopathologic exposures evaluated in the analytic cohort ranged from 16.5–59.5%, with chronic placental inflammation representing the most common histopathologic exposure (Table 2) and fetal inflammatory response the least common. One-fourth (25.7%, 125/486) of participants in the analytic cohort had the high-risk composite ASQ^®^-3 outcome (i.e., a high-risk score in at least one domain), and the frequency of a high-risk score for individual ASQ^®^-3 domains ranged from 5.8–11.5%. After accounting for all eligibility criteria, notably exclusion of infants born preterm and/or diagnosed with hypoxic-ischemic encephalopathy after birth, the frequency of a high-risk score for each of the individual ASQ^®^-3 domains was lower in the analytic cohort than in the three individual cohorts that contributed ASQ^®^-3 data for this analysis (Supplementary Table 2).^19–21^ The overall median (IQR) gestational age of assessment was 12 (8, 18) months, and the median age at assessment of nearly all ASQ^®^-3 domains were similar in children with high-risk versus low-risk ASQ^®^-3 screens (Supplementary Table 3).
In regression models, none of the placental histopathologic lesions were associated with the composite ASQ^®^-3 outcome (aOR 1.43, 95% CI 0.95–2.15). In evaluation of the individual ASQ^®^-3 domains, chronic placental inflammation was associated with more than double the odds of a high-risk communication (aOR 2.84, 95% CI 1.09–7.40) and fine motor (aOR 2.26, 95% CI 1.2–5.04) domain score. Fetal vascular malperfusion was associated with 55% lower adjusted odds of an abnormal high-risk screen in the communication domain (aOR 0.45, 95% CI 0.21–0.95; Table 2). In the pre-specified sensitivity analysis limited to children who were appropriate-for-gestational age by birthweight, chronic placental inflammation was associated with higher unadjusted odds of a high-risk fine motor domain score; however, this finding was no longer significant in adjusted analysis (Table 3).
Post-hoc analysis stratified by age at ASQ^®^-3 assessment identified an association between chronic placental inflammation with a high-risk composite ASQ^®^-3 score (aOR 2.07, 95% CI 1.05–4.07) and a high-risk screen in the gross motor domain (aOR 3.89, 95% CI 1.25–12.10) at 12–18 months of age. Maternal vascular malperfusion was also associated with higher odds of a high-risk ASQ^®^-3 screen performed at 12–18 months in the gross motor domain (aOR 3.20, 95% CI 1.26–8.11) (Supplementary Table 4). No associations were observed between placental histopathology and ASQ^®^-3 assessments performed prior to 12 months. In analyses stratified by fetal sex, chronic placental inflammation was associated with more than twelve times the odds of a high-risk screen in the communication domain among male children (aOR 12.13, 95% CI 2.02, 72.97), and more than three times the odds of a high-risk screen in the fine motor domain (aOR 3.93, 95% CI 1.01–15.27) among female children (Supplementary Table 4). However, considering the wide confidence intervals noted for the individual ASQ^®^-3 domains in analyses stratified by fetal sex, the significance of these associations with chronic placental inflammation should be interpreted with caution.
In the sensitivity analysis limited to children born to birthing parents without obesity, no significant associations were noted between the placental exposures and ASQ^®^-3 scores, either overall or in the individual domains. We did not identify an association between placental histopathology and the problem-solving or personal social ASQ^®^-3 domains in any sensitivity or subgroup analyses. Finally, among the individual subtypes of chronic placental inflammation, only basal chronic villitis was associated with a high-risk ASQ^®^-3 screen in the communication (aOR 2.26, 95% CI 1.01–5.08) and fine motor (aOR 2.20, 95% CI 1.09–4.44) domains (Supplementary Table 5).
In this analysis of term singleton births in the ECHO Cohort, placental histopathology was not associated with a high-risk composite ASQ^®^-3 screen in children assessed at 2–18 months. However, we identified an association of chronic placental inflammation with a high-risk ASQ^®^-3 screen in the communication and fine motor domains at 2–18 months of life. In the sensitivity analysis limited to children who were appropriate-for-gestational age by birthweight, the associations between chronic placental inflammation and a high-risk ASQ^®^-3 screen in these domains were no longer significant. Additional sensitivity analyses signaled potential associations between chronic placental inflammation and abnormal ASQ^®^-3 screens in children ages 12–18 months and by child sex, as well as between basal chronic villitis, as a subtype of chronic placental inflammation, and abnormal ASQ^®^-3 domain scores.
The true prevalence of placental histopathologic lesions is difficult to measure, considering the lack of routine placental histopathology, especially in term infants.^45^ Features of chronic placental inflammation – chronic deciduitis, chorioamnionitis, and villitis (inclusive of basal chronic villitis, chronic chorionic vasculitis, chronic villitis of unknown etiology, and chronic chorionitis) – have been reported in 2–33% of pregnancies and are relatively common in term and late preterm placentas (5–15%).^27,46^ Our study showed a much higher rate of chronic placental inflammation (59.5%) and other histopathology, likely attributable to the large number of blocks collected from each placenta and thus greater detection of histopathologic lesions than is typically observed in placental evaluation for clinical indications alone.^24,28^ Maternal vascular malperfusion has typically been associated with 14–32% of term placentas,^28,47,48^ and fetal vascular malperfusion in 1–11% of term placenta,.^28,48–50^ whereas in the ECHO Cohort maternal or fetal vascular malperfusion occurred in 38.3% and 56.2%, respectively. Accordingly, how these findings translate to the clinical setting and inform risk of abnormal child neurodevelopment requires further study.
Inconsistent associations between placental histopathological features and child neurodevelopmental factors have been reported in prior literature, related in part to limited short- and long-term neurodevelopmental follow-up of children and difficulty identifying the timing of the neurologic insult.^45^ Prior studies evaluating placental histopathologic lesions with adverse early childhood neurodevelopment have most extensively focused on infants with adverse birth outcomes (e.g., very preterm, SGA, very low birthweight, and/or with neonatal encephalopathy^5,7,12,13,51–54^) who are already at greater risk of adverse neurodevelopmental outcomes, regardless of placental pathology.^8,11,14^ Despite these challenges, prior literature does suggest some features of placental pathology are more strongly associated with perinatal morbidity and occur more frequently in children who subsequently exhibit adverse neurologic outcomes, as supported by our findings.^45^
In Japan’s Hamamatsu Birth Cohort of predominately term (86%) singleton births from 2007 to 2011, Ueda et al. identified that markers of maternal vascular malperfusion, present in 26–46% of 258 placentas evaluated, was associated with lower composite scores at 10–40 months on the Mullen Scales of Early Learning, a provider-completed assessment of gross motor, fine motor, visual reception, receptive language, and expressive language subscales. They also identified higher composite Mullen Scales of Early Learning scores associated with fetal vascular malperfusion.^14,15^ Evaluation of chronic placental inflammation, identified in 5–6% of placentas in the Hamamatsu cohort, was comparatively limited, and associations with individual subscales were not evaluated.^14^ In the ECHO Cohort, we identified similar signals for an association of maternal vascular malperfusion with greater odds of an abnormal screen for gross motor development at ages 12–18 months, and fetal vascular malperfusion with lower odds of an abnormal screen for development of communication skills in the overall cohort. Maternal vascular malperfusion suggests a low supply of maternal blood into the placental intervillous space, potentially resulting in relative hypoxic conditions in utero that may alter neurodevelopmental programming that manifests as slow early infantile neurodevelopment.^14,55,56^ Fetal vascular malperfusion suggests insufficient fetal blood supply to specific villous areas, similarly contributing to conditions of relative hypoxia with lesser exchange of oxygen and nutrients.^14,56^ Exposure to prenatal hypoxia leads to oxidative damage, mitochondrial dysfunction, endocrine axis dysfunction, epigenetic modifications, alteration of intracranial blood flow, and anatomical changes in the brain and have been implicated in abnormal motor and neurocognitive development in offspring.^55,56^ Similar to Ueda et al., we do not have a clear explanation for the paradoxical association suggested between fetal vascular malperfusion and lower risk of an abnormal ASQ^®^-3 screen in the communication domain, and cannot exclude unmeasured confounding or false discovery with type 1 error.
In case-control analyses of children born from the 1990s to early 2000s in Cleveland, Ohio at term and subsequently diagnosed with cerebral palsy or other neurologic impairment, Redline and O’Riordan identified associations with two subtypes of chronic placental inflammation among cases referred for medicolegal diffuse chronic villitis^8^ (5/40 [13%] of cases with neurologic impairment diagnosed at least 20 months postnatally, versus 6/176 [3%] of controls) and villitis of unknown etiology with obliterative fetal vasculopathy^11^ (9/42 [21%] of cases with neonatal encephalopathy versus 7/250 [3%] of controls; age at diagnosis not specified). Comparison of our findings with these prior studies is limited by employment of heterogeneous assessment tools and definitions of neurodevelopmental impairment. Additionally, the lower rate of high-risk neurodevelopmental screens in this study suggests future study in other cohorts is warranted. Nonetheless, they are similarly suggestive of a role for harnessing placental histopathology, especially when chronic placental inflammation is identified, to identify children at risk of abnormal early childhood neurodevelopment.
Maternal immune activation has been implicated in animal and human studies as influencing changes in placental function and offspring brain development.^7^ This underlying upregulation of the maternal immune response involves maternal anti-fetal cellular rejection (i.e., a breakdown of maternal immune tolerance leading to maternal immune response against paternal antigens expressed by the fetus) and leads to destruction of the placental architecture, disruption of the intervillous space where maternal-fetal gas exchange occurs, and idiopathic chronic placental inflammatory lesions on histopathology.^27,57^ These changes secondary to the maternal immune response represent one proposed mechanism predisposing to perinatal morbidity and abnormal childhood development.^27,46,57^ The higher proportion of chronic placental inflammation identified in cases of a high-risk ASQ^®^-3 screen at 2–18 months in this analysis augments the body of literature suggesting a role for maternal anti-fetal cellular rejection in contributing to central nervous system injury and higher risk of abnormal early childhood neurodevelopment.
Albeit an uncommon event, extensive efforts have been made to identify maternal risk factors, antenatal events, and biomarkers that can effectively predict adverse neurologic outcomes in term infants.^8,14^ These efforts further seek to identify infants at high-risk of abnormal neurocognitive development who would most benefit from early intervention during the first several months of life, a period of significant neuroplasticity that provides the opportunity for the greatest impact of early intervention.^6,8,14,58^ Significant investment into the application of advanced ‘omics’ technologies is underway to identify such biomarkers, but these methods are not easily translatable into routine health care practice because of cost and accessibility.^14,59^ Comparatively, placental histopathology is an available tool in clinical practice and offers a host of biomarker data that sheds light on pathophysiologic mechanisms that occur throughout gestation and may influence early childhood neurodevelopment.^8,14,45^ Enhanced communication of placental histopathology results by obstetric providers to pediatricians, and heightened awareness by pediatricians when placental histopathologic lesions are identified, may inform postnatal management such 1) closer monitoring of infants identified to have these placental histopathologic biomarkers (e.g., chronic placental inflammation), 2) linkage to diagnostic evaluation following an abnormal neurodevelopmental screening assessment, and 3) earlier implementation of evidence-based interventions when indicated to improve early childhood neurodevelopmental outcomes.^6,53^ Finally, given the potential for placental histopathology to inform postnatal counseling and management, our findings may be used to justify performing placental histopathology more broadly in clinical practice, not only in pregnancies with a maternal or fetal pregnancy complication, but also in uncomplicated or low-risk pregnancies.
The multisite ECHO Cohort, which began its second cycle of funding in September 2023 (Cycle 2), will allow for confirmation of the signals identified herein, and further in-depth exploration of understudied subgroups, and more extensive evaluation of individual placental histopathologic subtypes and domains of abnormal early childhood neurodevelopment.^60^ Beyond an anticipated larger sample size, ECHO Cycle 2 has planned broad prospective placental sampling, harmonized placental sampling and assessment protocols, longer follow-up of infants up to 7 years of age, and formal assessments of infant neurodevelopment administered by trained professionals (i.e., Mullen Scales of Early Learning^15^, Bayley Scales of Infant and Toddler Development-Third Edition^61^). Recognizing that adverse neurologic outcomes are often not immediately evident after birth, as suggested in our sensitivity analyses in which positive associations between placental lesions and high-risk ASQ^®^-3 screens were identified at 12–18 months but not before 12 months, the long-term follow up of these infants is critical.^62^
The lack of significant associations between placental histopathologic lesions and high-risk ASQ^®^-3 screen in our sensitivity analyses of appropriate-for-gestational age infants and in individuals without obesity suggests additional study of small- and large-for-gestational age infants born at term, and in individuals with obesity, is warranted. Further evaluation of the association of placental histopathology with sex differences in early childhood neurodevelopment is indicated as the placenta has been an important mediator of sexual dimorphism in adaptive responses to insults and offspring health outcomes.^7,63–65^ The relationship between basal chronic villitis, which encompasses severe and extensive lymphocyte and plasma cell infiltration and suggests an exaggerated maternal immune response in utero,^27,28,66,67^ and fine motor development also requires confirmation in ECHO Cycle 2. More broadly, as suggested by associations with chronic placental inflammation in our various sensitivity analyses, additional histopathologic phenotyping of maternal anti-fetal rejection is needed in conjunction with other methods of clinical and molecular placental phenotyping to inform monitoring algorithms and/or immunomodulating therapies to improve short- and long-term outcomes.^27,68^
This study, the first analysis of placental data harmonized from the ECHO Cohort, is the largest analysis of prospectively collected placental data using current histopathologic classifications defined by the Amsterdam criteria with corresponding early childhood neurodevelopmental evaluation in children born at term.^25^ Abnormal neurologic development is a rare but clinically devastating outcome following a term birth. The longer-term follow-up and rich covariate data collected in ECHO allows us to account for a number of potential confounders and precision variables. As nearly 90% of singleton pregnancies result in term births, our findings are highly generalizable and clinically important.^69^
We recognize several limitations inherent to our study design. Although the ASQ^®^-3 is a widely used, accessible, and validated tool used to screen for developmental concerns, it is not clinically diagnostic for neurodevelopmental diagnoses; however, our analysis based on this screening tool, which is available to both parents and providers, enhances the clinical utility of our results, especially should placental histopathology become more broadly used to guide surveillance of at-risk infants. Despite the large size of the overall ECHO Cohort, the small sample size for some exposures and outcomes likely contributed to wide confidence intervals for certain analyses performed, and the significance of such findings should be interpreted with caution given the possibility of type 1 error. The small sample size for certain exposures and outcomes additionally may have limited our ability to detect some significant differences in the stratified and sensitivity analyses performed. A larger cohort with more universal placental sampling, which ECHO Cycle 2 will enable, is needed to affirm the associations identified.
We further recognize the potential for false discovery, and did not adjust for multiple comparisons given the exploratory nature of this analysis which is intended to inform future analyses in Cycle 2 of the ECHO Cohort, and considering our broad singular hypothesis that a higher frequency of placental lesions would be associated with greater risk of abnormal neurodevelopment. Moreover, each individual test performed can stand on its own with an individual hypothesis. However, we recognize that each placental exposure, as well as each outcome, are not completely independent, and are correlated to some degree. As we did not account for multiplicity or correlation between the multiple comparisons performed, it is possible that the significant findings identified are due to type 1 error and thus our findings are hypothesis-generating rather than confirmatory.
Finally, high rates of histopathologic lesions in this cohort are greater than what has been previously reported and is suggestive of sampling bias. The fifteen slides per placenta evaluated in the ECHO Cohort is well beyond the number typically evaluated for clinically indicated placental sampling, which may explain the higher frequency of histopathologic lesions in this cohort than typically reported clinically. Conversely, placentas from high-risk pregnancies at ECHO Cohort sites were often required to go to clinical pathology at their respective site, and placental collection was missed as part of the study protocol, which have been expected to underestimate the frequency of histopathologic lesions. Furthermore, we report only the presence or absence of each histopathologic category, and did not assess stage or grade of histopathologic findings, which are included as subclassifications in the Amsterdam criteria and should be further evaluated in ECHO Cycle 2 analyses.^25^ Despite these limitations, these data represent an important contribution to the literature and lay the foundation for future analyses of term infants.
These findings from the diverse, multisite, prospective ECHO Cohort improve our understanding of placental pathologic lesions and potential underlying mechanisms associated with an elevated risk of abnormal early childhood neurodevelopment following a term birth. Findings of chronic placental inflammation on histopathologic evaluation may be used to inform parental counseling regarding risk of abnormal early childhood development. These data may inform postnatal surveillance strategies to facilitate linkage to follow-up evaluation and early intervention therapies in at-risk infants to support lifelong neurodevelopmental function and well-being when placental histopathologic lesions are identified.