By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Please be advised that we experienced an unexpected issue that occurred on Saturday and Sunday January 20th and 21st that caused the site to be down for an extended period of time and affected the ability of users to access content on Wiley Online Library. This issue has now been fully resolved. We apologize for any inconvenience this may have caused and are working to ensure that we can alert you immediately of any unplanned periods of downtime or disruption in the future.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Dr LF Nielsen, NANEA, Department of Epidemiology, Institute of Public Health, University of Aarhus, Vennelyst Boulevard 6, 8000 Aarhus C, Denmark. Email: firstname.lastname@example.org
Objective To investigate the association of asphyxia-related conditions (reducing blood flow or blood oxygen levels in the fetus) with spastic cerebral palsy (CP) considering different gestational age groups and the timing of risk.
Design Population-based case–control study.
Setting Danish Cerebral Palsy Register in eastern Denmark and Danish Medical Birth Register.
Population or Sample 271 singletons with spastic CP and 217 singleton controls, frequency matched by gestational age group, born 1982–1990 in eastern Denmark.
Methods Data were abstracted from medical records, and a priori asphyxia-related conditions and other risk factors were selected for analysis. Each factor was classified according to the time at which it was likely to first be present.
Main outcome measures Spastic CP.
Results Placental and cord complications accounted for the majority of asphyxia conditions. In multivariate analysis, placental infarction was significantly associated with a four-fold increased risk for spastic quadriplegia and cord around the neck was significantly associated with a three-fold increased risk for spastic CP overall. The combination of placental infarction and being small for gestational age (SGA) afforded an especially high risk for spastic quadriplegia. Placental and cord complications were present in 21% of cases and 12% of controls.
Conclusions The risk for spastic quadriplegia from placental infarction may be linked in some cases with abnormal fetal growth (17% of all children with spastic quadriplegia and 3% of control children both had an infarction and were SGA)—suggesting an aetiologic pathway that encompasses both factors. The risk for spastic CP from cord around the neck is not accounted for by other prepartum or intrapartum factors we examined. Considering the relative timing of risk factors provides a useful framework for studies of CP aetiology.
Cerebral palsy (CP) is the most common neuromotor developmental disability in childhood, affecting 2–3 per 1000 liveborn children. Its prevalence has remained largely unchanged despite major improvements in clinical care, such as detecting ‘fetal distress’, reducing complications in labour and delivery, and enhancing neonatal support.1–3 Intrapartum conditions leading to birth asphyxia were long believed to be the primary cause of CP, but nonspecific markers, such as low Apgar scores, abnormal fetal heart rate tracing, and meconium in the amniotic fluid, are markers of fetal and newborn difficulties and need not represent asphyxia.4,5
Nelson and Grether6 evaluated the effect of specific birth complications that can limit oxygen supply to the fetus and found that only tight umbilical cord around the neck was associated with spastic CP, specifically for spastic quadriplegia. Their investigation was limited to children weighing ≥2500 g at birth, reflecting the focus on term and near-term infants in investigations of precursors, especially birth asphyxia, for abnormal neonatal neurologic findings (neonatal encephalopathy as the defined clinical syndrome) and such developmental sequelae as CP. Investigation of these associations in preterm neonates is complicated by their high risk for a variety of neurologic problems, although preterm infants account for about half of all children with CP.
Nonspecific intrapartum markers provide poor information on the timing and duration of an asphyxiating insult, and it seems likely that intrapartum events are only one of several steps on the causal pathway to CP.7,8 It has been suggested that 70–80% of cases of CP are caused by prepartum factors such as intrauterine infections, antepartum bleeding, intrauterine growth restriction, preterm birth, and genetic anomalies5,7,9—factors that, in some cases, may also lie on the pathway linking an asphyxia-related condition and CP. Ultimately, understanding the timing and sequence of risk events can provide insight into the development of clinical interventions.
The timing of prepartum, intrapartum, or neonatal events and their relationship to the pathogenesis of CP is complex and incompletely understood, and an intrapartum event is not likely to be the sole cause of CP. It may be only one of the several interacting factors, and/or it may represent only one factor in a sequence. It is therefore important to investigate how a combination or sequence of events (and their timing in the sequence) might contribute to the risk of developing CP. We therefore investigated the association between asphyxia-related conditions, assessed from routine medical record data and spastic CP and taking into account the gestational age and the timing of risk factors. As spastic CP is more likely than nonspastic CP to be the consequence of intrapartum asphyxia, cases with nonspastic CP were not included. Similarly, multiple pregnancies are confounded by a number of obstetric complications, and so, our study was restricted to singletons.
The current case–control study was based upon data collected as part of a pilot study assessing biomarkers and risk factors for CP. For the purposes of the pilot study, the sample size was 100 children with spastic CP and 100 control children in each of three gestational age groups. The case children were identified from the Danish Cerebral Palsy Register (DCPR; established in 1967 and covering eastern Denmark).10 Data for the DCPR are based on reports from contact persons in all paediatric departments in eastern Denmark in combination with information from The Danish National Hospital Register. For the DCPR, CP is defined as ‘a disorder of movement and posture due to a defect or lesion of the immature brain, excluding disorders that are of short duration, due to progressive disease, or due solely to mental deficiency’.11 Children were included in the DCPR if they were born in Denmark, were living there at the age of 5 years, and were believed to have a pre- or perinatal aetiology (events occurring before 28 days of postdelivery age). All diagnoses in the DCPR are validated by a neuropaediatrician on the basis of a review of physical findings in the child’s medical record.11 Case children born from 1982 to 1990 were randomly selected from the DCPR on the basis of diagnosis (spastic CP) and gestational age (three strata: <32, 32–36, and ≥37 weeks).
Control children were randomly selected from the Danish Medical Birth Register (DMBR),12 which collects data on all live births and stillbirths in Denmark. Control children were frequency matched to index children with spastic CP by year of birth, within the three gestational age strata. Control children were then excluded if they had a history of developmental disorders in the Danish National Hospital Register.
We initially identified 471 case and 382 control children for inclusion in the study (Figure 1). We invited all selected individuals to participate, but 119 children and/or their parents refused to take part and were therefore excluded. This left 271 case and 217 control children (Figure 1).
The study was approved by the Danish Data Protection Agency and the Ethical Committee in Vejle and Funen Counties, Denmark (Ref. No. 19990258).
Data collection and variable description
All risk factor data were abstracted from maternal obstetric and neonatal medical records at 24 hospitals in eastern Denmark by trained medical abstractors (midwives). The presence of a specific condition was based on a clinical diagnosis (e.g. placental abruption) or quantitative data (e.g. maternal temperature) as recorded in the medical notes. Gestational age for each birth was based on last menstrual period dates when available (64%); otherwise, it was based on early ultrasound (24%) or newborn (12%) examination.
Based on a review of the abstracted data, we selected for analysis potential asphyxia-related conditions (i.e. conditions that can reduce blood flow or blood oxygen levels in the brain and lead to permanent changes), other potential risk factors for CP prior to delivery, and markers of newborn and neonatal status. We then classified each condition according to the time at which it was thought first to be present (pre-pregnancy, prepartum, intrapartum, and newborn/neonatal period). Details about each selected condition by time period are presented in Table 1.
Table 1. Description of potential asphyxia-related conditions, other potential risk factors for CP prior to delivery, and measures of newborn/neonatal status
Maternal age (years)
<18, 18–40 (referent), >40
Primi-, multiparous (referent)
Previous preterm delivery
Yes, no (referent)
Previous spontaneous abortion/therapeutic abortion/stillbirth/neonatal death
The placenta and cord were examined systematically by the attending midwife at delivery, and the findings were recorded on a standard form as part of routine clinical practice. Data included whether the placenta was intact; if placental size was designated as small (estimated placental weight less than 400 g for a mature child); if any infarcts were present (their approximate size and numbers were not always recorded); if there were any clots on the maternal side; if the membranes were intact; whether the cord had only two vessels, true knots, or a marginal insertion; and if the cord was thin. Cord length measurements were not always recorded by the midwife. As described previously, all the other potential asphyxia-related conditions, such as abruption, placenta praevia, cord around the neck (including tight cord), prolapsed cord, shoulder dystocia, uterine rupture, massive bleeding at delivery, and maternal hypotension, were based on a clinical diagnosis or an entry in the medical notes.
The analysis was conducted in several steps as follows.
Univariate analyses (chi-square test) were performed on all variables to determine their association with CP. Because of the original sampling scheme, all univariate analyses included the gestational age strata in a basic logistic regression model.13 For each of the asphyxia-related conditions, we also performed stratified analyses by gestational age strata and by spastic CP subtypes (diplegia, hemiplegia, and quadriplegia). The measures of association with CP were odds ratios, reported with 95% CI. Since a main objective was to consider the timing of risk events in pregnancy to better understand the pathways leading to CP, we used the multivariate analyses to segregate the analysis by time period. For variables categorised within the pre-pregnancy, prepartum, and intrapartum periods, variables that had a statistically significant association with CP in the basic analyses were added—individually and all combined—to a logistic model consisting of the gestational age strata variables plus an asphyxia-related condition. The results of these analytic steps were designed to indicate 1) whether the asphyxia-related variables were associated with CP (and when the risk arose), independent of other potential risk factors, and 2) which of the other potential risk factors arising at different times in pregnancy influenced the association between asphyxia-related factors and CP. These steps were performed for each statistically significant asphyxia-related condition—for all cases with spastic CP combined as well as stratified on each of the spastic subtypes (diplegia, hemiplegia, and quadriplegia).
The final analytic step was to assess the association between each of the statistically significant asphyxia-related conditions and the markers of fetal status (meconium staining of the amniotic fluid) and newborn/neonatal status (Apgar score <7 at 5 minutes, resuscitation at delivery, and neonatal seizures). The purpose of these final analyses was to assess whether children with asphyxia-related conditions associated with CP were more likely to display signs of fetal distress, depression at delivery, or neonatal neurologic problems. In other words, since these markers of fetal or newborn/neonatal status are not specific to asphyxia, we did not consider them as adverse exposures, per se, but as potential sequelae of interest. We assessed the association in children with spastic CP and controls separately. For our next analyses (stratified on cases or controls), the logistic regression models included both the gestational age strata variable and the significant asphyxia-related condition as independent variables, with each marker of fetal and newborn/neonatal status serving as the dependent outcome variable.
All analyses were performed in Stata version 9.0.
The distribution of cases and controls by gestational age strata is shown in Table 2. Due to some differences in exclusions between case and control children (Figure 1), there was a slightly higher proportion of late preterm infants (32- to 36-week gestation) among control children and a slightly higher proportion of early preterm infants (<32-week gestation) among case children. These differences did not reach statistical significance (P = 0.29). Among children with CP, 176 (65%) had diplegia, 57 (21%) had hemiplegia, 29 (11%) had quadriplegia, and 9 (3%) had spastic ataxia.
Table 2. Distribution of subjects by gestational age strata and spastic CP diagnosis
Basic analyses: pre-pregnancy and prepartum periods
Although there was a tendency for mothers of children with CP to be older and to have a lower pre-pregnancy body mass index, none of the maternal pre-pregnancy characteristics was associated with CP. Prepartum characteristics significantly associated with CP included maternal urinary tract infection confirmed by culture and the baby being small for gestational age (SGA) (Table 3).
Table 3. Univariate analyses of the association between CP and other risk factors by time period prior to delivery and measures of newborn/neonatal status
BMI, body mass index; ROM, rupture of membranes; UTI, urinary tract infection.
Due to missing data, all percentages reflect the number of individuals with data for each factor.
Includes gestational age strata in the basic univariate logistic regression model.
Maternal age at delivery <18 years
Maternal age at delivery >40 years
Previous abort/stillbirth/neonatal death
Previous preterm delivery
Pre-existing maternal disease
Other maternal disease in pregnancy
UTI culture confirmed
Decreased fetal movement
Large for gestational age
Augmentation during labour
Labour length <3 hours
Labour length >20 hours
Second stage >1 hour
ROM <12 hours
24 ≤ ROM ≤ 72 hours
ROM >72 hours
Fever during delivery >37.7°C
Meconium in amniotic fluid
Emergency caesarean delivery
Status of Newborn and Neonate
Apgar score <7 after 5 minutes
Resuscitation at delivery
Cord gas pH <7.00
Seizures within the first 28 days
In relation to potentially asphyxia-related conditions, prepartum cord complications (Table 4) were rare and were not significantly associated with spastic CP. Placental infarcts (based on macroscopic examination of the placenta at delivery by the attending midwife) accounted for the great majority of the selected prepartum placental complications, being reported in about 11% of cases and 6% of controls (Table 4). In the basic analysis, placental infarcts were associated with a significant two-fold increased risk for spastic CP (Table 4) independent of gestational age. There were no significant differences in the magnitude of risk for spastic CP from infarcts by gestational age. Notably, placental infarcts were associated with an almost seven-fold increased risk for spastic quadriplegia, whereas the odds ratios for spastic diplegia or hemiplegia were both about 1.7 and not statistically significant (Table 5).
Table 4. Univariate analyses of the association between CP and potential asphyxia-related conditions by time period
Includes the gestational age strata variable, prepartum urinary tract infection confirmed by culture, SGA.
Includes the gestational age strata variable, labour induction, duration of labour >20 hours, prolonged second stage of labour >1 hour, acute caesarean delivery, and breech presentation.
All spastic CP
Cord around neck
All spastic CP
Basic analyses: intrapartum period
Markers of abnormal labour significantly associated with spastic CP in the basic analyses were total length of labour >20 hours and prolonged second stage of labour (Table 3). Labour induction and duration of labour <3 hours also increased the risk for spastic CP, but only the effect of labour induction was statistically significant. Other intrapartum risk factors associated with spastic CP included breech presentation and emergency caesarean delivery. A marker of fetal status in the intrapartum period, meconium in the amniotic fluid, was also associated with spastic CP (Table 3).
Concerning the asphyxia-related placental and cord complications in the intrapartum period (Table 4), one or more of the selected cord complications was present in about 21% of the cases and in about 12% of the controls (P = 0.007). The most common complication, cord around the neck, was associated with about a two-fold increased risk for spastic CP, specifically for spastic diplegia. The risk did not vary significantly by gestational age (Table 5). None of the other cord complications arising in the intrapartum period was associated with an increased risk of spastic CP. Although a long cord had a four-fold increased risk for CP, it was not statistically significant due to the small numbers. Notably, there was little overlap between long cord and cord around the neck: only two of the ten cases and one of the two controls with long cord also had cord around the neck. Intrapartum placental complications were observed in about 11% of both case and control children, and neither abruption (the most common intrapartum placental complication) nor placenta praevia was associated with an increased risk for spastic CP.
The category of other intrapartum asphyxia-related conditions was accounted for almost completely by massive bleeding although numbers were small (4% of cases and 2% of controls).
Basic analyses: status of newborn and neonate
We examined fetal depression at birth, including Apgar score <7 at 5 minutes and resuscitation measures in the delivery room, both of which were associated with spastic CP overall (Table 3). Cord blood pH was measured on less than half of all births (cases, n = 128, 64% and controls, n = 72, 36%), but only children with CP had an umbilical cord pH <7.00 (12 of 128 children with CP and a cord gas measurement, 9.4%). Seizures in the first 28 days of life were also associated with spastic CP overall (Table 3).
In multivariate analyses allowing for the influence of other variables, the association between placental infarction and spastic CP overall was no longer statistically significant. Furthermore, there was no association between placental infarction and either spastic diplegia or hemiplegia (Table 5). The strength of the association between placental infarction and spastic quadriplegia was reduced but remained significant with a near four-fold increased risk after controlling for prepartum risk factors and a 13-fold increased risk after controlling for intrapartum factors (Table 5).
Notably, the magnitude of the association between placental infarction and spastic quadriplegia was reduced after considering the effect of maternal urinary tract infection on the association (adjusted OR = 6.1) and to an even greater extent after considering the effect of SGA (adjusted OR = 4.9). The latter reflects the fact that overall, 5 of 29 (17.2%) of all children with spastic quadriplegia both had an infarction and were SGA compared with 7 of 217 (3.2%) of all control children. Furthermore, we observed in stratified analyses that among children who had a placental infarction, there was a 16-fold increased risk of spastic quadriplegia for those who were SGA (adjusted OR = 16.0, 95% CI = 1.42–180.1) compared with a 1.6 times higher risk for non-SGA children (adjusted OR = 1.6, 95% CI = 0.17–14.67). Although these results suggest that placental infarction associated with SGA leads to spastic quadriplegia, the numbers in the different substrata were quite small and the corresponding confidence limits for each term in the multivariate models were broad, indicating that the magnitude of the reported risk estimates would need evaluation in larger studies.
With the exception of emergency caesarean delivery, the inclusion of each of the other intrapartum risk factors added individually in multivariate analyses increased the association between spastic quadriplegia and placental infarcts, although the confidence limits were very broad. The distributions among the different risk categories were sparse such that there were either no observations or very few for case and control children in the combined categories of risk (e.g. no children had a history of placental infarcts and long second stage of labour).
Unlike placental infarcts, the association between cord around the neck and spastic CP remained significant—and almost unchanged—in multivariate analyses (Table 5). The association between cord around the neck and spastic CP overall also remained after adding the intrapartum risk factors, although the risk estimate increased to nearly three-fold (adjusted OR = 2.8) (Table 5). The risk estimates between cord around the neck and both hemiplegia and quadriplegia also increased and approached significance after adjusting for the intrapartum risk factors. When considered individually, the specific intrapartum risk factors with the greatest effect on the adjusted risk estimates from cord around the neck were the measures of abnormal labour length: total labour >20 hours or second stage of labour >1 hour. For these two variables, however, there were typically few or no observations for case and control children in the various combined categories of risk (e.g. having both cord around the neck and labour >20 hours). Thus, children with CP (17.4%) were more likely than control children (8.7%) to have cord around the neck and normal labour length (defined as having both total labour 3–19.9 hours plus second stage of labour <1 hour) (P = 0.04).
Associations between placental infarction and cord around the neck with indicators of fetal/newborn/neonatal problems
We examined whether children with placental infarction or cord around the neck were more likely to have meconium in the amniotic fluid, depression at delivery (5-minute Apgar score <7 or resuscitation in the delivery room), or neonatal neurologic problems (seizures in the first 28 days of life). We assessed the association in control children and children with spastic CP separately (Table 6). In control children, there were no increased risks for meconium in the amniotic fluid, 5-minute Apgar score <7, or resuscitation in the delivery room associated with a history of either placental infarction or cord around the neck. There was, however, a seven-fold, statistically significant increased risk for neonatal seizures associated with placental infarction and a four-fold, nonsignificant increased risk for neonatal seizures associated with cord around the neck in control children. In contrast, among children with spastic CP overall (or by subtype), none of the markers of fetal or neonatal problems was significantly associated with a history of placental infarction or cord around the neck.
Table 6. The risk for indicators of fetal, newborn, or neonatal problems from a history of placental infarction or cord around the neck
Odds ratio from basic logistic regression model including gestational age strata variable.
(n = 12), n (%)
(n = 205), n (%)
(n = 22), n (%)
(n = 195), n (%)
5-minute Apgar score <7
Resuscitation at delivery
(n = 29)
(n = 242)
(n = 47)
(n = 224)
5-minute Apgar score <7
Resuscitation at delivery
In this study, we selected a priori a number of specific conditions based on routine clinical notations found in medical records during the study period—arising at different times in pregnancy—that could produce asphyxia conditions in the fetus by reduced blood flow or blood oxygen levels and leading to brain damage and CP. We examined the effects of the selected conditions on the association with CP after controlling for a number of risk factors that also arise in different time periods during pregnancy to better understand the timing and sequence of risk events. Two conditions were significantly associated with spastic CP in multivariate analysis: placental infarction and cord around the neck.
Placental infarcts were associated with an increased risk for spastic CP overall in basic analyses, which appeared largely attributable to a seven-fold increased risk, specifically for spastic quadriplegia. In the multivariate analyses by time period, only prepartum risk factors accounted markedly for the association of infarcts with spastic quadriplegia, indicating that in this study, the risk was not confounded by adverse events during labour and delivery. Furthermore, the combination of infarcts and SGA was more common among children with spastic quadriplegia than among control children, leading to a marked difference in the risk for spastic quadriplegia from infarcts in SGA versus non-SGA children. SGA was observed in 28% of children with spastic quadriplegia in which 63% (5 of 8) also had placental infarction; only 13% of control children were SGA in which 24% (7 of 29) also had placental infarction.
Our measure of placental infarction is based on a macroscopic examination of the placenta routinely performed at delivery by the attending midwife; expert placental pathology reports were not available. For the midwife examination, infarcts are areas of pale, white, dense tissue that are palpable through the entire thickness of the placenta; they can vary in number, size, and distribution on the placenta. It is not clear how these findings would be defined histologically, but we propose that they are ‘old’ lesions (i.e. arising before the intrapartum period), likely due to multiple causes, and contributing to a loss of functional placental tissue. They could have a direct impact on fetal status or indirectly increase fetal vulnerability to further insults, especially in susceptible fetuses. Unfortunately, we did not have more details on placental pathology, such as the proportion of the placenta that was infarcted. In the literature, the placenta has been described as a diary of intrauterine life and a useful tool for understanding the causes of CP and other adverse outcomes.14–16 In a recent review, Redline17 emphasised that expert placental assessment, in combination with data on the presence and timing of other clinical features, constitutes a powerful tool for understanding the various aetiologic scenarios associated with CP.
The high risk for spastic quadriplegia from a combination of macroscopic placental infarction and SGA suggests a common aetiologic pathway, perhaps through the adverse effects on fetal growth and neurologic status from impaired fetal and umbilical blood flow or decreased placental reserve14,18 or a shared underlying aetiology such as thrombophilia. The significant association between SGA and CP has been well documented.18 Our measure of SGA was based on the Marsal’s criteria,19 and we observed a significant association with spastic CP overall, as well as with spastic quadriplegia specifically, which has been reported previously in a Swedish population.20 The Marsal’s criteria are based on only a few factors reflecting fetal growth (gestational age, birthweight, and gender), which may lead to some misclassification of individuals with SGA. Although any misclassification of SGA may be nondifferential with regards to case or control status in this study, more sophisticated approaches to modelling and measuring impaired fetal growth may provide additional insights into its role as a risk for spastic quadriplegia, especially in combination with a history of placental infarction.
The second condition significantly associated with spastic CP in the multivariate analysis was cord around the neck. Adjusting for risk factors from the prepartum period did not influence its association with spastic CP. Adjusting for other intrapartum risk factors, particularly abnormal labour length, actually increased the magnitude of risk from cord around the neck for spastic CP overall and for each of the subtypes that approached significance. Cord around the neck may cause asphyxia through obstruction of blood flow during labour and delivery or even periodically before delivery. Controlling for other intrapartum risk factors that might impact that process, such as the length of labour, revealed that while cord around the neck may occur in some instances in combination with these other risk factors, in the majority of instances it did not. The special instance of a tight cord around the neck – which arose in one-quarter of cases and one-third of controls with cord around the neck – was not significantly associated with spastic CP in our data, unlike the findings of Nelson and Grether.6 Our understanding of the clinical significance of these findings related to CP, however, is limited by our source of data on cord complications (routine clinical records) and thereby the potential for exposure misclassification, especially since we have no way of confirming if asphyxia was ever present. Nevertheless, pathologic umbilical cord irregularities, such as abnormal cord length, cord entanglement, and abnormal cord coiling, have been associated with adverse perinatal outcomes21,22 and warrant further investigation in studies of CP.
In our data, signs of fetal, newborn, or neonatal problems (all of which were more common among children with CP than among control children) were not reliable flags for a subsequent diagnosis of CP in infants with a history of placental infarction or cord around the neck.
Our sample included individuals from different gestational age groups, and the associations between the asphyxia-related conditions and the spastic CP by gestational age group were examined in both stratified and multivariate analyses. There were no differences by gestational age group in the association with spastic CP and any of the selected asphyxia-related conditions, suggesting that the effects of these conditions on the risk for spastic CP are not limited to either term or preterm infants. It should be remembered that because our sample selection process was stratified on gestational age, the proportions of specific conditions reported in this study for case children overall may not be representative of the corresponding proportions observed in a randomly selected sample of individuals with CP in the population.
The strength of this study is the large number of cases with spastic CP (n = 271) drawn from a population-based registry and based on well-defined diagnostic criteria. Because of our sampling design, it was possible to look across gestational ages for the risk factor associations with CP overall and with different CP subtypes. We based our analysis on data found in medical records. As a primary data source, medical records reflect the kinds of data available in routine clinical practice and recorded before the CP diagnosis, but the quality of recording may vary. Any misclassification due to faulty recording—especially of conditions arising before birth—would likely be nondifferential between case and control children as data collection is independent of CP status. Although we hypothesised that the selected conditions associated with CP may potentially provoke asphyxia in the fetus, a study weakness is that we have no way of confirming that asphyxia was ever present. Thus, the biologic mechanism linking placental infarction or cord around the neck with spastic CP remains hypothetical and in the case of placental infarction may be linked to a different mechanism entirely, such as inflammation and infection (to be examined in a separate analysis). Because this study was based upon pilot study data, the numbers of observations in the different strata of risk were often quite small and the study lacks power to detect moderately sized effects, thereby risking type II errors. The findings need to be confirmed in larger studies.
In routine clinical data documenting conditions that may provoke asphyxia in the fetus, we were able to identify conditions that increase the risk for spastic CP or its diagnostic subtypes. In our data, placental and cord complications accounted for the majority of the identified conditions and, collectively, were observed in a substantial minority of our cases (over 20%). Although the opportunities to prevent the asphyxia-related conditions observed in this study are limited, by considering time period in the analyses, we can have a better understanding of when and where to look, or not to look, in future studies (e.g. risk factors for placental infarction and the pattern and type of abnormal fetal growth). A further step would be to evaluate in larger studies the predictive value of conditions such as fetal growth failure following a history of placental infarction for the diagnosis of spastic CP.
Disclosure of interests
The authors have no conflicts of interest.
Contribution to authorship
L.F.N., P.T., and D.S. conceived and designed the study and analysed and wrote the paper.
J.G., D.H., and T.J. analysed the data. B.J., M.V., and P.U. contributed with interpretation of data and critical revision of the manuscript.
Details of ethics approval
The study was approved by the Danish Data Protection Agency and the Ethical Committee in Vejle and Funen Counties, Denmark (Ref. No. 19990258).
The study was supported by the Elsass Foundation, Centers for Disease Control and Prevention, Dagmar Marshall Foundation, and Hede Nielsen Foundation.
We thank The National Cerebral Palsy Registry and the Paediatric Departments in Eastern Denmark for access to medical records.