Chorioamnionitis and risk of long-term neurodevelopmental disorders in offspring: a population-based cohort study

BACKGROUND: Evidence indicates that in utero exposure to chorioamnionitis might increase the risk of neurodevelopmental disorders in the offspring. However, ﬁndings on this topic have been inconsistent. OBJECTIVE: To examine the association between chorioamnionitis and neurodevelopmental disorders in offspring. STUDY DESIGN: This was a retrospective population-based cohort study in Sweden. A total of 2,228,280 singleton live births and stillbirths between 1998 and 2019 were included in our study population. Data on maternal characteristics and neurodevelopmental disorders in offspring were obtained by individual record-linkages of nationwide Swedish reg-istries. Chorioamnionitis was identiﬁed using the National Medical Birth Register. Inpatient and outpatient diagnoses were obtained for cerebral palsy, autism, attention deﬁcit hyperactivity disorder, epilepsy, and intellectual disability. Multivariable Cox proportional hazards regression was used to estimate the association between chorioamnionitis and each neurodevelopmental disorder with adjusted hazard ratios and 95% conﬁdence intervals. A causal mediation analysis of the relationship between chorioamnionitis and neurodevelopmental disorders with preterm delivery ( < 37 weeks) was performed. RESULTS: A total of 5770 (0.26%) offspring were exposed to chorioamnionitis during pregnancy. During the study’s follow-up time there were 4752 (0.21%) cases of cerebral palsy, 17,897 (0.80 %) cases of epilepsy, 50,570 (2.27 %) cases of autism, 114,087 (5.12%) cases of attention deﬁcit hyperactivity disorder, and 14,574 (0.65%) cases of intellectual disability. After adjusting for potential confounders, exposure to chorioamnionitis increased the hazard ratios of cerebral palsy (adjusted hazard ratio, 7.43; 95% conﬁdence interval, 5.90 e 9.37), autism (adjusted hazard ratio, 1.43; 95% conﬁdence interval, 1.21 e 1.68), attention deﬁcit hyperactivity disorder (adjusted hazard ratio, 1.17; 95% conﬁdence interval, 1.03 e 1.33), and intellectual disability (adjusted hazard ratio, 1.99; 95% conﬁdence interval, 1.53 e 2.58), whereas chorioamnionitis was not signiﬁcantly associated with higher rates of epilepsy in offspring. Mediation analysis revealed that these associations were mainly explained through preterm delivery; however, increased risk was also observed among term infants. CONCLUSION: Chorioamnionitis increases the risk of neurodevelopmental disorders, particularly cerebral palsy, autism, attention deﬁcit hyperactivity disorder, and intellectual disability. These associations were mainly mediated through preterm delivery. Efforts for timely identi-ﬁcation and appropriate interventions to treat infections during pregnancy will have sustained beneﬁts in reducing the burden of neurologic complications in children at the population level.


Introduction
Neurodevelopmental disorders represent a significant public health issue worldwide and are responsible for a considerable proportion of the global burden of disease. 1,2 Over the last decades, there has been a rise in the prevalence of neurodevelopmental disorders and an increase in the number of in-dividuals acquiring such diagnosis. 3e5 Many risk factors are likely to operate during fetal life and infancy, the earliest and most sensitive stages of brain development. 6,7 Evidence indicates that maternal infections, in particular chorioamnionitis, might negatively affect the sensitive fetal brain and lead to brain injury, adverse neurodevelopmental outcomes, and an increased lifetime risk of specific psychiatric diseases. 8e11 Clinical chorioamnionitis is globally the most common infection-related complication in labor and delivery wards 12,13 and is estimated to have a prevalence of 1% to 6% in all pregnancies in the United States, whereas intraamniotic infection might be present in 10% of patients with preterm labor. 12,14 Clinical chorioamnionitis has been characterized as a syndrome rather than a single entity, linked with proven intraamniotic infection, sterile intraamniotic inflammation, or signs of a maternal systemic inflammatory process without intraamniotic inflammation. 13,15 The diagnostic criteria involve the presence of fever with 2 or more of the following: maternal or fetal tachycardia, maternal leukocytosis, tenderness of the uterus, and purulent or malodorous amniotic fluid. 13,14 Chorioamnionitis has been shown to be associated with long-term neonatal outcomes, such as cerebral palsy. 11,16 However, epidemiologic evidence about the association between chorioamnionitis and risks of other long-term neurologic disorders in offspring is limited. Therefore, in this Swedish nationwide population-based cohort study of >2 million singleton births, we aimed to investigate the association between chorioamnionitis and risk of longterm neurodevelopmental disorders in offspring, in particular cerebral palsy, autism, attention deficit hyperactivity disorder (ADHD), epilepsy, and intellectual disability. We also examined the extent to which preterm delivery mediates the effect of chorioamnionitis on neurodevelopmental disorders in offspring. We hypothesized that the exposure to chorioamnionitis is associated with increased risks of long-term neurodevelopmental disorders in offspring.

Materials and Methods
Using the Swedish Medical Birth Register, our cohort comprised all singleton births at 22 completed gestational weeks in Sweden from January 1, 1998, through December 31, 2019. Using the unique personal national registration numbers of mothers and their offspring the Medical Birth Register was crosslinked with the nation-wide National Patient, National Prescribed Drug, Total Population 17 and Education Registers. Since 1997, diseases have been coded according to the Swedish version of the International Classification of Diseases,

Tenth
Revision (ICD-10). The Anatomical Therapeutic Chemical Classification System and the Drug Identification Numbers were used to retrieve the prescription medications for ADHD.

Exposure
Women with chorioamnionitis were identified from the Medical Birth Register with diagnosis records for the ICD-10 code O41.1 (including diagnoses of infection of amniotic sac and membranes, chorioamnionitis, and amnionitis) and their infants' records including the ICD-10 code P02.7 (fetus or newborn affected by complications of placenta, cord, and membranes: chorioamnionitis). This case definition refers to clinical chorioamnionitis in the assessment of the clinician treating the mother and/or infant. Results of pathologic placental investigations were not available in our data sources.

Disorders
Adverse neurodevelopmental disorders included all clinically ascertained diagnoses of cerebral palsy, epilepsy, autism, ADHD, and intellectual disability. Children with epilepsy who had also cerebral palsy were not included in the epilepsy group. Diagnoses were identified from birth until December 31, 2020, in the National Patient Register and the prescription registry using ICD-10 codes (Supplemental Table 1 shows specific codes). In Sweden, all infants and preschool children regularly undergo routine medical and developmental examinations. At 4 years of age, a mandatory assessment of motor, language, cognitive, and social development is conducted. 18 Children who are suspected of having a developmental disorder are referred to a specialist team for further assessment, with any diagnostic information reported to the National Patient Register. 18

Covariates
We obtained maternal information about the country of birth, age at child's birth, early-pregnancy body mass index (BMI), height, parity, years of education, smoking during pregnancy, cohabitation with a partner, history of psychiatric disorders, prepregnancy hypertension, and diabetes mellitus. Infant's information included the calendar year of birth, sex, gestational age at birth, birthweight for gestational age, and major congenital malformation. The percentiles of birthweight for gestational age were based on the Swedish fetal growth reference obtained from the Medical Birth Register. 19 Maternal BMI (kg/m 2 ) was classified according to the World Health Organization as underweight (BMI <18.5), normal weight (18.5e24.9), overweight (25.0e29.9), obesity class I (30.0e34.9), and obesity class II and III (35.0). Mothers who reported daily smoking at the first antenatal visit and/or at 30 to 32 gestational weeks were classified as smokers. 20 The mode of delivery was obtained from obstetrical records and categorized as vaginal noninstrumental, vaginal instrumental, elective cesarean delivery, and emergency cesarean delivery.

Statistical analysis
Baseline demographic characteristics of children born to mothers with and without chorioamnionitis were compared as presented in Table 1.

AJOG at a Glance
Why was this study conducted? Evidence indicates that in utero exposure to chorioamnionitis might increase the risk of neurodevelopmental outcomes in the offspring. However, findings on this topic have been inconsistent. This study aimed to examine the association between chorioamnionitis and long-term neurodevelopmental disorders in the offspring.

Key findings
Chorioamnionitis increases the risk of neurodevelopmental disorders, particularly cerebral palsy, autism, attention deficit hyperactivity disorder (ADHD), and intellectual disability. These associations were mainly mediated through preterm birth; however, increased risk was also observed among term infants.
What does this add to what is known? Chorioamnionitis was associated with increased hazard ratios of 7.43 for cerebral palsy, 1.43 for autism, 1.17 for ADHD, and 1.99 for intellectual disability in the offspring (compared with offspring not exposed to chorioamnionitis), even after adjusting for several potential confounders. Preterm delivery accounted for a large proportion of the neurodevelopmental disorder risk associated with chorioamnionitis.  Summaries of categorical variables were presented in absolute numbers and proportions (%), whereas statistical significance was assessed by the Pearson chi-square test. Cumulative hazard curves were used to compare risks of each neurodevelopmental outcome over time according to chorioamnionitis status. The differences between the curves were assessed using the log-rank test. We calculated hazard ratios (HRs) and the corresponding 2-sided Wald-type 95% confidence intervals (CIs) using Cox proportional hazards regression models, which allowed detailed adjustment for censoring depending on the length of follow-up of each child. Each child was followed up from birth until the diagnosis of the outcome, death, emigration, or end of follow-up on December 31, 2020, whichever occurred first. Adjusted HRs were obtained from multivariable Cox models in 3 steps, gradually adjusting for additional potential confounders. In model 1, we adjusted for maternal age at child's birth, parity, maternal educational level, country of mother's birth, smoking during pregnancy, maternal height, early-pregnancy BMI, any psychiatric disorders, child's sex, calendar year of birth, and cohabitation with a partner. In model 2, we also adjusted for major congenital malformations, and in model 3, additional adjustment was made for the mode of delivery. The robust sandwich estimate of the covariance matrix was used to calculate 95% CIs in all Cox models to account for the sequential births to the same mother.

Causal mediation analysis
We considered preterm birth (<37 weeks) as a potential mediator for the effect of chorioamnionitis on cerebral palsy, autism, ADHD, and intellectual disability (Supplemental Figure, Supplemental Table 2, Supplemental  Table 3). Therefore, we undertook causal mediation analyses based on a counterfactual framework 21 to disentangle the association between chorioamnionitis and the outcomes (ie, total effect) into the natural direct effect (the association between chorioamnionitis and the outcomes [cerebral palsy, autism, ADHD, and intellectual disability] in the absence of preterm birth) and the natural indirect effect (the association operating through the mediators). 22 We also estimated the controlled direct effect, which provided an estimate of the effect of chorioamnionitis on the outcomes that is not mediated through preterm birth (ie, among term births). We also assessed the proportion of the total effect (on the HR scale) between chorioamnionitis and the outcome(s) that was mediated through preterm delivery. Furthermore, we created a composite mediator of preterm delivery and neonatal infection, and preterm delivery and respiratory distress syndrome (RDS), diagnosed at 0 to 27 days of age, to examine the joint mediation effect of neonatal morbidity and preterm delivery on the association between chorioamnionitis and the disorders.

Sensitivity analyses
We performed several sensitivity analyses. First, because mediation methods were developed under a strict nounmeasured-confounding assumption, 21 we examined the robustness of causal effects to unmeasured confounders by estimating an E-value (defined as the maximal strength of association that an unmeasured confounder would need to have with the exposure and the outcome to fully explain away an observed exposureeoutcome association). 23 Second, given that death before the diagnosis of outcome would preclude a child ajog.org OBSTETRICS Original Research from being diagnosed with such conditions in the future, we also quantified the adjusted association between chorioamnionitis and a composite outcome including any of the following: stillbirth, infant mortality (ie, death within the first year after birth), or any neurodevelopmental disorder (ie, stillbirth, infant death, or epilepsy). Logistic regression analyses were used to assess the association between chorioamnionitis and each composite outcome, adjusting for the same confounders noted in model 2.
Third, to focus only on potentially clinically relevant cases of neurodevelopmental disorders, we restricted the age at diagnosis: 3þ years of age for ADHD and intellectual disability, 1þ years of age for diagnoses of autism, and 26þ days of age for epilepsy. Fourth, to address missing values of covariates in our cohort and the possible bias that it could introduce, we performed multiple imputation with chained equations under the assumption of missing at random. All analyses were performed using Stata statistical software, version 16 (StataCorp, College Station, TX) and SAS, version 9.4 (SAS Institute, Cary, NC).

Results
Between January 1, 1998 and December 31, 2019, the Medical Birth Register recorded information of about 2,228,290 singleton live births and stillbirths with valid national registration numbers for mothers and children. After excluding 10 births with missing information on child's sex and maternal age, the final study cohort included 2,228,280 singleton births.

FIGURE
Lifelong risk for neurodevelopmental disorders in the offspring after exposure to chorioamnionitis A, Unadjusted cumulative hazard curves show the cumulative hazard for intellectual disability among offspring exposed and not exposed to chorioamnionitis. Log rank P value<.001. B, Risk for ADHD among offspring exposed and not exposed to chorioamnionitis. Log rank P value<.001. C, Risk for epilepsy among offspring exposed and not exposed to chorioamnionitis. Log rank P value¼.98. D, Risk for autism among offspring exposed and not exposed to chorioamnionitis. Log rank P value<.001. E, Risk for cerebral palsy among offspring exposed and not exposed to chorioamnionitis. Log rank P value<.001 Shading around the lines indicates the 95% CI.  a Rate is calculated as number of cases per 10,000 person-years; b Model 1 adjusted for maternal age at child birth, parity, maternal educational level, country of mother's birth, smoking during pregnancy, maternal height, early-pregnancy BMI, any psychiatric disorders, child's sex, calendar year of birth, and cohabitation with a partner; c Model 2: in addition to the factors noted in model 1, also adjusted for major malformation; d Model 3: in addition to the factors noted in model 2, also adjusted for mode of delivery. Tsamantioti. Chorioamnionitis and the risk of long-term neurodevelopmental disorders in offspring. Am J Obstet Gynecol 2022. a Causal effects were adjusted for confounding effects of maternal age at child birth, parity, maternal educational level, country of mother's birth, smoking during pregnancy, maternal height, early-pregnancy BMI, any psychiatric disorders, child's sex, calendar year of birth, and cohabitation with a partner. Tsamantioti. Chorioamnionitis and the risk of long-term neurodevelopmental disorders in offspring. Am J Obstet Gynecol 2022.

Univariable and multivariable analysis
Unadjusted cumulative hazard curves showed a significantly higher cumulative hazard of cerebral palsy, autism, ADHD, and intellectual disability among offspring born to mothers with chorioamnionitis than among those born to mothers without chorioamnionitis (Figure). After adjusting for potential confounders, compared with offspring of mothers without chorioamnionitis, the adjusted HRs were higher for cerebral palsy, autism, ADHD, and intellectual disability in offspring of mothers with chorioamnionitis ( Table 2, Model 1), whereas chorioamnionitis was not significantly associated with epilepsy in offspring. Although slightly attenuated, the same pattern of associations remained after adjusting for major malformations and mode of delivery (Table 2, Model 2 and 3).
We examined the impact of preterm delivery on the association between chorioamnionitis and cerebral palsy, autism, ADHD, and intellectual disability ( Table 3). The HRs for the natural direct and natural indirect (mediated) effects of chorioamnionitis on cerebral palsy were 2.91 (95% CI, 2.01e4.21) and 2.57 (95% CI, 1.93e3.41), respectively. This indicates that 70% of the total effect of chorioamnionitis on cerebral palsy was mediated through preterm delivery, and about 30% of the total effect was explained through other undiscovered pathways (other than preterm birth). Furthermore, 41% of the total effect was jointly mediated through preterm delivery and neonatal infections, and 57% jointly mediated through preterm delivery and RDS. Similar mediation effects of preterm delivery, neonatal infection, and RDS were also observed for autism, ADHD, and intellectual disability ( Table 3).

Sensitivity analyses
In the sensitivity analyses, the E-value for the HR and the lower 95% CI were greater than the observed estimates (Supplemental Table 4). These results suggest that the casual mediation parameters were robust to unmeasured confounding. Analyses addressing the competing risk of death showed similar associations between chorioamnionitis and each composite outcome of death and/or each neurodevelopmental disorder (Supplemental Table 5). Furthermore, restricting the age at diagnosis of the outcomes did not change our results (Supplemental Table 6). The results were also unchanged in supplemental analyses using multiple imputation for missing data (Supplemental Table 7).

Principal findings
In this nationwide cohort study, chorioamnionitis conferred increased risk of cerebral palsy, autism, ADHD, and intellectual disability. The causal mediation analysis suggests that this effect is largely mediated through preterm birth. The controlled direct effects indicate that chorioamnionitis substantially increases the risks of neurologic complications, even among term births. These findings highlight that infection-mediated pathways are implicated in neurologic complications, and the complex associations that shape these risks.

Results in the context of what is known
Consistently with our findings, previous studies have found increased risk of neurologic disorders, specifically cerebral palsy, among offspring exposed to chorioamnionitis. 11,24,25 In addition, elevated risks of other adverse neurodevelopmental outcomes, such as autism, epilepsy, ADHD, cognitive impairment, speech delay, and hearing loss, have been reported in children exposed to maternal infection or chorioamnionitis, 9,26,27 suggesting that exposure to inflammation in utero could alter brain development and function. 28 In contrast, several studies have concluded that chorioamnionitis poses no independent risk on short-term neurodevelopmental outcomes. 16,29e31 However, a direct comparison of our results with those of previous studies is hampered by the absence of a uniform, well-established clinical diagnostic algorithm for chorioamnionitis. 32 Consequently, there is large variation in the diagnosis and definition of clinical chorioamnionitis. 27,33 Clinical chorioamnionitis is frequently used interchangeably with histologically proven chorioamnionitis without being distinguished among clinicians and researchers, further complicating comparisons between individual studies. 33

Clinical and research implications
Our results demonstrated that the association between chorioamnionitis and neurodevelopmental disorders is mainly mediated through associated neonatal conditions including preterm birth, RDS, and neonatal infections. However, about 30% of the total effect was explained through other undiscovered pathways, and we also observed increased risks of cerebral palsy, autism, and intellectual disability among term infants. Overall, preterm infants are at greater risk of developing major shortand long-term neurodevelopmental disorders. 34,35 These risks increase with decreasing gestational age at birth, 36 whereas the inflammatory environment in chorioamnionitis decreases with increasing gestational age. 37 Some studies have previously shown associations between preterm birth and neurodevelopmental conditions, including autism, ADHD, cerebral palsy, and cognitive impairment. 10,38e41 Our results provide further evidence that preterm birth, triggered by infection such as chorioamnionitis, increases the risk of neurodevelopmental adversity. 42,43 Chorioamnionitis is also associated with significant neonatal morbidity, such as neonatal sepsis, intraventricular hemorrhage, and respiratory syndrome, 32,33,44,45 which puts the surviving neonates at higher risk of long-term neurologic adversity. 46e48 The underlying mechanism linking chorioamnionitis with neurodevelopmental outcomes may involve the activation and upregulation of infection and inflammatory processes in the mother and the fetus. 8,9 This activation results in the release of cytokines and chemokines from decidua and fetal membranes, such as interleukin-6, which stimulate the synthesis of prostaglandins and metalloproteases that can ajog.org OBSTETRICS Original Research lead to the ripening of the cervix, rupture of the membranes, and spontaneous labor or induction of labor and delivery. 8 In addition, chorioamnionitis can induce a fetal inflammatory response syndrome, resulting in the release of inflammatory products and reactive oxygen species, which can directly damage the sensitive fetal cerebral cells, predominantly the white matter, resulting in cerebral palsy and other neurodevelopmental disorders. 8,16,49

Strengths and limitations
Our study has several strengths. First, the population-based study design along with high-quality registry data that are prospectively and independently collected minimized the possibility of information bias in our study. Second, we adjusted for a number of important maternal confounders, including maternal BMI, smoking during pregnancy, and maternal psychiatric history. 50,51 We used the Swedish version of ICD-10 diagnostic codes to ascertain chorioamnionitis and adverse outcomes in the offspring. Although these specific codes have not been externally validated, previous studies have shown that in the Swedish National Inpatient Register, positive predictive values of diagnostic codes are between 85% and 95%. 52 In Sweden, the diagnosis of chorioamnionitis is routinely made by the obstetrician, which adds to the clinical accuracy of the diagnosis and its ascertainment by the ICD codes. Lastly, we quantified the effects of chorioamnionitis on the composite outcome, including death and neurodevelopmental conditions, and thereby addressed the competing risk of death. 53,54 Nonetheless, our study might have some limitations. Our outcome ascertainment was based on inpatienteoutpatient clinical data and drug register data. Therefore, it is possible that we included only patients with the most severe cases of the disorders who sought clinical help, whereas milder cases were not captured. However, this misclassification of the outcomes is nondifferential, thus possibly resulting in an underestimation of the true associations. Furthermore, although the diagnosis of chorioamnionitis was based on specific clinical criteria, the clinical presentation varies between patients, which can lead to variation in diagnosis between clinicians. The incidence of chorioamnionitis was 0.26% in our study, which is lower than reported previously. 12,14 This discrepancy can imply that only the most severe cases of chorioamnionitis were included in our cohort, and the clinically silent or subclinical cases were not detected. This potential underascertainment of chorioamnionitis could also lead to an underestimation of the true effects. Furthermore, premature or ill newborns might serve as a stimulus for clinicians to investigate their mothers more actively for an underlying infection than mothers of apparently healthy infants, which makes such newborns more likely to receive a diagnosis of chorioamnionitis. However, in our data, all infants' records for chorioamnionitis coincided with maternal diagnosis, therefore limiting bias. In addition, in our mediation analyses we have assumed that the causal pathways between preterm birth, preterm birtheneonatal infections, and preterm birtheinfectioneRDS are not sequential. Future studies may consider exploring these pathways through a mediation analysis with sequentially ordered causal mediators. Lastly, we cannot exclude that residual confounders by unmeasured or unknown factors could drive the observed association.

Conclusion
This study revealed a significant association between chorioamnionitis and neurodevelopmental disorders. A large driver of this risk was preterm delivery; however, increased risk was also observed among term infants. Efforts for timely identification and appropriate interventions to treat infections during pregnancy will have sustained benefits in reducing the burden of neurologic complications in children at the population level. n

SUPPLEMENTAL FIGURE
Simplified DAG of the relation between chorioamnionitis and neurodeveopmental disorders with preterm delivery as the mediator ADHD, attention deficit hyperactivity disorder; CP, cerebral palsy.
Tsamantioti. Chorioamnionitis and the risk of long-term neurodevelopmental disorders in offspring. Am J Obstet Gynecol 2022. SUPPLEMENTAL

SUPPLEMENTAL TABLE 3
Preterm birth and respiratory distress syndrome and neonatal infections, odds ratios, and 95% confidence intervals for cerebral palsy, epilepsy, autism, attention deficit hyperactivity disorder, and intellectual disability. Liveborn singleton infants born in Sweden from 1998e2019