Objective The incidence of perinatal arterial ischaemic stroke (PAIS) is about 1 in 2300 live births. Evidence about the aetiology is still lacking. The aim of this study was to identify maternal, perinatal and neonatal risk factors for symptomatic PAIS in full-term infants.
Methods Each full-term infant with PAIS was matched to three healthy controls for gestational age, date of birth and hospital of birth. Antenatal and perinatal risk factors were studied using univariate and multivariate conditional logistic regression analysis.
Results Fifty-two infants were diagnosed with PAIS. Significant risk factors in the univariate analysis (p<0.05) were nulliparity (64% vs 47%), maternal fever (>38°C) during delivery (10% vs 1%), fetal heart rate decelerations (63% vs 16%), meconium-stained amniotic fluid (44% vs 17%), emergency caesarean section (35 vs 2%), Apgar score (1 min) ≤3 (29% vs 1%), Apgar score (5 min) <7 (25% vs 1%), umbilical artery pH <7.10 (56% vs 10%), hypoglycaemia <2.0 mmol/l (29% vs 3%) and early-onset sepsis/meningitis (14% vs 2%). In the multivariate analysis, maternal fever (OR 10.2; 95% CI 1.3 to 78.5), Apgar score (5 min) <7 (OR 18.1; 95% CI 3.4 to 96.8), hypoglycaemia <2.0 mmol/l (OR 13.0; 95% CI 3.2 to 52.6) and early-onset sepsis/meningitis (OR 5.8; 95% CI 1.1 to 31.9) were significantly associated with PAIS.
Conclusions Maternal fever during delivery and early-onset sepsis/meningitis were found to be involved with PAIS as was previously noted. Apgar score (5 min) <7 and hypoglycaemia were found to be important risk factors in term PAIS.
Statistics from Altmetric.com
During the perinatal period, there is a high risk of developing arterial or venous territory stroke. This has been increasingly recognised with the more widespread use of sophisticated neuro-imaging techniques. Perinatal ischaemic stroke is defined as a heterogeneous condition in which there is a focal disruption of cerebral blood flow (CBF) secondary to arterial or venous thrombosis.1 Perinatal arterial ischaemic stroke (PAIS) is the most common form of cerebral infarction in children with an incidence of 1:2300 live births and an important cause of life-long neurodevelopmental disabilities1 Half of the newborn infants with PAIS present within 12–36 h after birth with (hemi)convulsions, apnoeas or hypotonia.1 Patterns of MR signal intensity (SI) changes on conventional and diffusion-weighted imaging (DWI) show that PAIS occurs within a limited time-frame around birth.2
What is already known on this topic
The pathogenesis of perinatal arterial ischaemic stroke (PAIS) is multifactorial, but the different pathways are still not elucidated.
Maternal infection, as well as early-onset sepsis/meningitis, appears to be important in the genesis of PAIS.
What this study adds
An Apgar score <7 at 5 min, suggestive of fetal distress, is more commonly associated with PAIS than previously reported.
Hypoglycaemia is associated with preterm PAIS and also a risk factor for PAIS in the full-term infant.
Although PAIS is a clinically relevant type of brain injury, aetiology is still a matter of debate.3 Thrombosis or embolism of a major cerebral artery is probably the main cause. Several factors appear to be related to this pathogenesis reflecting problems at different times during late pregnancy, delivery or the immediate neonatal period such as oligohydramnios, chorioamnionitis, premature rupture of membranes, pre-eclampsia and low Apgar score at 1 min.4,–,7 Thrombophilic abnormalities are also likely to increase the risk of PAIS.8,–,10
Previous studies, identifying risk factors of PAIS, were negatively affected by the heterogeneity of the study population. Preterm and full-term infants were taken together as well as lesions with a variable underlying pathophysiology or lesions occurring at different points in time.6 The study groups were often rather small or there was no adequate comparison group, except for one preterm study.4 ,5 ,11,–,14 The aim of this study was to identify maternal, perinatal and neonatal risk factors in full-term infants with symptomatic PAIS in a comparative study.
Patients and methods
In a retrospective analysis, infants (>37 weeks gestational age (GA)) with a PAIS were studied. All were admitted to the Neonatal Intensive Care Unit (NICU) of the Wilhelmina Children's Hospital, Utrecht, between 2000 and 2010. PAIS cases were established using the neonatal cranial ultrasound (cUS) and MRI database. In infants with clinical symptoms, PAIS was most often primarily suggested using cUS and subsequently confirmed by MRI. For every case, three controls were selected, matched for GA (≤7 days), date of birth (≤7 days) and hospital of birth, using the birth registry database. These births were under supervision of an obstetrician. Control infants for PAIS cases born at home or in the outpatient clinic under supervision of an independent midwife were collected from the University Obstetric Centre for primary obstetric care. When correct matching, based on above-mentioned criteria, was not possible, cases were excluded. Infants with stroke who also had a congenital anomaly, chromosomal abnormality or cardio-embolic stroke were excluded.
For this type of study, no informed consent or permission from the Internal Review Board is required in our hospital.
All records were retrieved for infants with a diagnosis of middle cerebral artery (MCA), posterior cerebral artery (PCA) or anterior cerebral artery (ACA) infarction. PAIS on cUS was diagnosed according to the following criteria: (1) presence of a wedge-shaped area of echogenicity in the region supplied by the MCA, ACA or PCA with a linear demarcation line and (2) cystic evolution of this area of increased echogenicity after 2 to 4 weeks.
Diagnosis of PAIS on MRI was based on decreased SI on T1 and increased SI on T2 within the affected area (figure 1 A–C). In those infants who had their MRI during the first week after birth (n=47), DWI showed high SI within the region of infarction (figure 1 D–F). The infarcts were classified according to the vessel, branch and site involved. MRI was conducted using 1.5-T and 3-T Philips system (Intera or Achieva 1.5 T; Philips, Healthcare, Best, The Netherlands).
Antenatal data from late pregnancy as well as data regarding the process of birth and the early neonatal period were collected from cases and controls from the patient notes. These included maternal-related and pregnancy-related variables: maternal age, parity, previous spontaneous abortion, history of infertility and pre-eclampsia, defined as either pre-eclampsia or pregnancy-induced hypertension as diagnosed by the obstetrician. Intrapartum variables included induction or augmentation of labour, prolonged rupture of membranes >24 h and maternal fever (≥38°C) often combined with intrapartum use of antibiotics. Fetal heart rate abnormalities were considered present when the attending physician noted repetitive or prolonged late decelerations, fetal bradycardia, non-reassuring fetal heart tracing or fetal distress by electronic fetal heart rate monitoring. A prolonged second stage of labour (>2 h), vacuum or forceps-assisted delivery and emergency caesarean section due to fetal distress, were also noted. GA, birth weight (BW), BW <3rd percentile or >97th (http://www.perinatreg.nl/registratiecurven), Apgar scores, umbilical cord artery pH, blood glucose <2 mmol/l and early-onset sepsis or meningitis were noted as infant characteristics. Early-onset sepsis was determined when the child developed a fever (>38°C) or an elevated C reactive protein >30 mg/l with a positive surface, blood or cerebrospinal fluid culture or both.
Because infants with PAIS were matched for several factors to controls, conditional logistic regression was used. The presence of PAIS was the outcome variable and maternal, intrapartum and neonatal characteristics were the independent variables in this analysis.
In the univariable analysis, the association between the individual possible risk factors and PAIS was studied. To determine whether risk factors were independently associated with PAIS, a multivariable conditional logistic regression analysis was performed, using a backward stepwise method. In the initial model, factors from the univariable analysis that were statistically significant and important factors from previous studies were included. Thereafter, variables were removed from the model based on the statistical significance (p>0.1) of the association with PAIS and changes in the regression coefficient of other variables. Results were expressed as ORs with 95% CI. p Values <0.05 were considered statistically significant. SPSS for Windows version 16.0 (SPSS Chicago, Illinois, USA) was used for statistical analyses.
Between January 2000 and December 2010, a total of 6126 infants were admitted to our NICU. Fifty-two full-term infants with clinical symptoms had a PAIS (figure 2) with a median GA of 40 5/7 weeks (range 37 1/7–42 5/7) and median BW 3440 g (range 2145–5230 g). Thirty-nine infants (75%) had been born under the supervision of an obstetrician in 13 different hospitals, 36 (69%) in a level-two hospital and 3 (6%) in our level-three unit. Thirteen (25%) infants were born under the supervision of an independent midwife, either at home or in the hospital.
Clinical presentation of PAIS
Twenty-nine infants (56%) had seizures (18 with hemiconvulsions) in the first 48 h after birth as their first symptom. Seventeen infants (33%) were initially admitted because of perinatal asphyxia and all, but two, developed seizures. Two of them underwent therapeutic hypothermia. Three infants were admitted with Group B Streptococcus meningitis and seizures. Three infants had symptomatic hypoglycaemia and seizures. Of all PAIS infants, 96% developed seizures within the first 48 h after birth.
In 40 infants, the diagnosis was made using cUS and confirmed by MRI at a median age of 5 days (range 2–14 days). In 12 infants with a small cortical or PCA infarct, the diagnosis was missed with cUS, and MRI was performed because of neonatal seizures. In one infant, only a CT scan was done. The majority of strokes involved the MCA (n=39, 75%). Of these infants with MCA distribution, the main branch was involved in 10 (26%), the cortical branch in 23 (59%) and one or more of the lenticulostriate branches in six infants (15%). Eleven infants (21%) had PCA infarction and in one of them, there was also bilateral ACA involvement. In two (4%) infants, an ACA distribution was seen. Left-sided stroke was more common (58%).
MR angiography was performed in 36 infants (69%) and CT angiography in 1 infant. These revealed loss of flow across the vessel involved in four infants. Cervical Doppler ultrasound was performed in 17 infants, but no dissection of the carotid artery was demonstrated.
An echocardiography was performed in 16 infants (31%) and showed a persistent foramen ovale in three, a pulmonary valve stenosis in one infant and a mid-muscular ventricular septal defect in one infant.
Table 1 shows the results obtained with the univariable analysis. Nulliparity, maternal fever combined with intrapartum use of antibiotics, meconium-stained amniotic fluid, fetal heart rate abnormalities, emergency caesarean section, Apgar score ≤3 after 1 min and <7 after 5 min, umbilical arterial pH less than 7.10, hypoglycaemia (≤2.0 mmol/l) and early-onset Group B Streptococcus sepsis/meningitis were significantly associated with PAIS.
After exclusion of PAIS infants initially diagnosed with perinatal asphyxia (n=17) and their controls, only maternal fever and Apgar score (5 min) <7 were no longer significant.
The multivariable analysis identified maternal fever during delivery (OR 10.2; 95% CI 1.3 to 78.5), Apgar score (5 min) <7 (OR 18.1; 95% CI 3.4 to 96.8), hypoglycaemia (OR 13.0; 95% CI 3.2 to 52.6) and early-onset sepsis/meningitis (OR 5.8; 95% CI 1.1 to 31.9) (table 2) to be independently associated with PAIS.
This is the first case-controlled study that has examined risk factors for PAIS in symptomatic full-term infants also taking the place of birth into account. In previous studies, lesions with a variable underlying pathophysiology or lesions occurring at different points in time were taken together.6 Maternal fever during delivery, Apgar score (5 min) <7, early-onset sepsis/meningitis and hypoglycaemia were identified as independent risk factors for PAIS.
A 5 min Apgar score of <7 likely reflects adverse events during delivery and suggests an important role for fetal distress and subsequent hypoxia-ischaemia in the pathogenic pathway of PAIS. Hypoxic-ischaemic encephalopathy (HIE) is an established cause of diffuse brain damage in the full-term newborn, but was not considered to be related to focal infarction.4 ,15 Only 6/126 PAIS infants were diagnosed with HIE in a study by Ramaswamy et al,16 and 15/315 in a study by Harbert et al,17 but recent studies suggested a role of perinatal asphyxia in the development of PAIS.7 ,18 Among our cases, 17 infants (32%) were initially admitted to the NICU because of perinatal asphyxia and were subsequently diagnosed to have PAIS. This is in contrast to a recent study with strict entry criteria for therapeutic hypothermia after HIE indicative of moderate-severe perinatal asphyxia, where no infants with PAIS were found.19 However, intrapartum complications suggestive of hypoxic-ischaemic events including fetal heart rate abnormalities, meconium-stained amniotic fluid and low Apgar scores have frequently been reported in infants with PAIS.3 ,5 ,6 ,11 ,13 ,14 These studies imply that PAIS is more often associated with a complicated delivery, resulting in a milder degree of perinatal asphyxia. Neuro-imaging will only be performed in those infants with symptoms, most likely (hemi)convulsions. In the absence of seizures or in centres where MRI is not routinely performed in infants with neonatal seizures, coexistence of focal cerebral pathology is likely to remain undiagnosed.18 Rafay et al,20 comparing neonates with seizures caused by stroke to seizures caused by HIE, found that seizure onset >12 h and clinically observed focal seizures are more often associated with PAIS.
Hypoxia-ischaemia might play a role in activation of thrombogenesis, as it decreases physiologically inhibiting anticoagulants, including antithrombin III, proteins C and S resulting in hypercoagulability.21 ,22 Adhami et al23 found that ischaemia-hypoxia is a rapid and powerful stimulus for spontaneous coagulation in mice that underwent the Levine/Vannucci procedure, mimicking perinatal hypoxic-ischaemic brain injury.24 However, our study does not answer the question whether a 5 min Apgar score <7 indicates a hypoxic-ischaemic event which might be involved in the causal pathway of PAIS or whether the low Apgar score is the consequence of PAIS already present at birth.3 7 25 The latter is, however, unlikely in view of the changes seen using diffusion-weighted imaging during the first week after birth.
Maternal fever during delivery and early-onset sepsis/meningitis were involved in the genesis of PAIS, which is in agreement with previous studies, identifying prolonged rupture of membranes and chorioamnionitis as risk factors.6 ,26 ,27 Infection shifts the coagulation system towards a prothrombotic state due to release of inflammatory cytokines and direct endothelial injury. Chorioamnionitis may also cause direct thromboembolism in the placenta increasing the risk for emboli to access the fetal brain. In this cohort, three infants (6%) had Group B Streptococcus meningitis. This central nervous system infection can lead to a local inflammatory response including thrombosis and is an established cause of PAIS in some cases.28 ,29
As in the preterm population, hypoglycaemia was identified as a significant neonatal risk factor.12 In full-term infants, an association between symptomatic hypoglycaemia and bilateral occipital infarction has been reported.30 Burns et al31 found stroke among diverse patterns of brain injury in infants with symptomatic neonatal hypoglycaemia without evidence for HIE. Sixteen of our cases (31%) developed hypoglycaemia <2.0 mmol/l within a few hours of birth preceding seizure activity and subsequent diagnosis of stroke. It is of interest that 7 of 11 (64%) infants with involvement of PCA had hypoglycaemia, being symptomatic in three of them. Nine of sixteen infants (56%) with hypoglycaemia had evidence of perinatal asphyxia. Because asphyxia is the most common cause of (asymptomatic) hypoglycaemia as was demonstrated in different animal models, it is difficult to distinguish whether hypoglycaemia is merely the consequence of asphyxia or might be causally related to PAIS.32 Anwar et al33 observed a positive correlation between mean arterial blood pressure and CBF in hypoglycaemic newborn dogs, suggesting loss of CBF autoregulation. Areas displaying the lowest blood flow are known to be most vulnerable.34 Recently, attention has been paid to hypoglycaemia as a cardiovascular risk factor in adults. In a study by Gogitidze Joy et al,35 acute hypoglycaemia led to an increase in inflammatory and pro-atherothrombotic biomarkers, suggesting an important role in the induction of inflammatory and prothrombotic responses.
The relative risk of genetic thrombophilic factors in the development of PAIS is still a matter of debate. A recent meta-analysis has shown that thrombophilia was a risk factor in children with first-episode stroke, including PAIS.8 However, a recent study regarding neonatal stroke did not support this.36 Further case-control studies are essential to investigate whether (extensive) testing is required in all infants with PAIS.
In contrast with previous studies, we did not find pre-eclampsia to be related to PAIS in this case-control study.3 ,6 However, in these studies, presumed perinatal stroke cases were also included, suggesting other risk factors may be important in this group. A history of infertility was not more common either.6 Cervico-cephalic injuries due to traumatic forces during the process of birth has sporadically been described as a pathogenic mechanism.14 In our study, MR angiography or ultrasonographic cervical imaging was performed (in 69% and 33%, respectively) without evidence for dissection of the cervical vessels.
In only three of our PAIS infants, placental histopathology was available; in two of them, there was evidence of decreased placental reserve. In a small cohort (n=5) of PAIS cases, placental pathology was present in all, showing mainly decreased placental reserve and thromboinflammatory processes.37
In a preterm case-control study, fetal heart rate abnormalities, twin-to-twin transfusion syndrome and hypoglycaemia were found to be independent factors associated with early preterm stroke.12 A role for complicated gestational course and delivery as well as sepsis/meningitis was recognised in another preterm case study.38 Although there is a difference in vascular anatomy and brain vulnerability between (early) preterm infants and (near)-full-term infants, risk factors for arterial stroke share a common background.
This study has several limitations. Since it recruited infants over 10 years, obstetric policies, MR imaging protocols and neonatal procedures may have changed. PAIS in asymptomatic neonates or in infants with a mild clinical presentation admitted to a community hospital may remain undiagnosed; therefore, this single-tertiary centre study may represent a selected group. Information was obtained from patient charts, a risk for bias in reporting as was the difference between hospitals in the completeness of the patient data. Because measurement of umbilical cord pH is not part of routine care in Dutch uncomplicated obstetrics, we only had these data in 48% of control infants, not dissimilar from the 62% of the cases. Information of the control infants was not available after discharge from the hospital (usually within a few hours to 3 days after birth), hence asymptomatic hypoglycaemia or infections without admission to the hospital of birth may have occurred.
Maternal fever, a 5 min Apgar score of <7, neonatal infection and hypoglycaemia are associated with symptomatic PAIS in full-term infants. The importance of infection has previously been reported. A role for fetal distress has only recently been recognised and the role of hypoglycaemia appears not just restricted to the preterm population.
The authors thank all obstetricians and paediatricians from Flevo Ziekenhuis, Almere; Meander Medisch Centrum, Amersfoort; Gelre Ziekenhuizen, Apeldoorn; Rijnstate Ziekenhuis, Arnhem; Deventer Ziekenhuis, Deventer; Tergooi Ziekenhuizen, Hilversum; Sint Antonius Ziekenhuis, Nieuwegein; Rivierenland Ziekenhuis, Tiel; Sint Elizabeth Ziekenhuis, Tilburg; TweeSteden Ziekenhuis, Tilburg; Diakonessenhuis, Utrecht, University Obstetric Centre, Utrecht and Hofpoort Ziekenhuis, Woerden for their collaboration. The authors are grateful for the suggestions of Dr F M Cowan to improve this manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.