Perinatal stroke in term infants with neonatal encephalopathy

Methods.

Cases were identified in a prospective cohort study of MRI predictors of outcome that enrolled 124 term neonates born in or transferred to our institution’s intensive care nursery from 1994 to 2000.7 The inclusion criteria for this cohort of neonatal encephalopathy were 1) umbilical artery pH < 7.1, 2) umbilical artery base deficit > 10, 3) 5-minute APGAR score ≤ 5, or 4) encephalopathy. Clinical encephalopathy was defined as abnormal tone, feeding, alertness, respiratory status, or reflexes. Those with congenital infection, malformation, or metabolic disease were excluded. Newborns were studied with MRI at a median of 7 days (range: 1 to 18), including 4 mm (1 mm gap) sagittal spin-echo (SE) (500/11/2 [repetition time/echo time/excitations]), 4 mm (1 mm gap) axial SE (500/11/2) images, and 4 mm (2 mm gap) axial SE (3000/60,120/1) images through the entire brain.7 Strokes were defined as focal parenchymal infarcts, either in an arterial or venous distribution, and were identified by a neuroradiologist blinded to the clinical condition. Outcome was determined at 30 months by the Bayley Scales of Infant Development Mental Index (MDI) and a neurologic examination by a child neurologist blinded to the neonatal condition and imaging studies (tables 1 and 2⇓).

Results.

Six newborns in the cohort were identified with perinatal strokes (4.8%; prevalence 95% CI: 2 to 10%, see table 1). The infarcts were arterial in five (three left and two right middle cerebral artery, all involving less than 2/3 of the middle cerebral artery territory) and venous in one (superior sagittal sinus thrombosis) (figure). Two subjects with arterial infarcts also had injury bilaterally in the intervascular watershed zone. The characteristics of the newborns with stroke are summarized in table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Case 3: right middle cerebral artery infarct with bilateral watershed injury. (A) Axial T1-weighted image (repetition time [TR]/echo time [TE]: 500/11) demonstrating hypointense signal of the cortex (arrowheads), caudate and anterior putamen in the distribution of the right middle cerebral artery. (B) T2-weighted image (TR/TE: 3000/120) of the same newborn demonstrating hyperintense signal and consequent blurring of the normal cortical ribbon in the same distribution as (A) (arrowhead) and coexisting hyperintense signal of the posterior intravascular watershed zones (arrows). Case 4: left middle cerebral artery infarct. (C) Axial T1-weighted image (TR/TE: 500/11) demonstrating hyperintense signal in the distribution of the left middle cerebral artery (arrowhead). (D) Axial T2-weighted image (TR/TE: 3000/120) demonstrating hyperintense signal and consequent blurring of the normal cortical ribbon in the same distribution as (C) (arrowhead).

Of the six newborns with stroke, the median birthweight was 3,518 g (range: 3,000 to 4,500 g) and five were male (83%). All six newborns required significant resuscitation with positive pressure ventilation at birth (bag mask in three and endotracheal intubation in three). The median 5-minute APGAR score was 5 (range 2 to 9). Cases 5 and 6 had meconium aspiration syndrome. Clinical seizures on the first day of life was a presenting feature in all six subjects with stroke including two newborns with status epilepticus in the first few hours of life (Cases 2 and 3). All seizures were treated with IV phenobarbital with the exception of Case 6 who only had a single seizure. The incidence of seizures across the entire cohort was 37.8% (p = 0.004), including 14.7% for those infants without brain injury and 33.3% with only watershed injury.

Pregnancy complications were seen in mothers of individual newborns including gestational diabetes (1), chorioamnionitis (1), Graves disease (1), and premature rupture of membranes > 18 hours (2). The incidence of these disorders in our cohort was similar in newborns with and without stroke (all p > 0.4). None of the pregnancies were complicated by preeclampsia, intrauterine growth restriction, smoking, drug abuse, or alcohol consumption. None of the newborns had a nuchal cord or a defined placental insult. The occurrence of a complicated delivery (failure to progress, failed vacuum) was seen in 4 of the 6 newborns with perinatal stroke and in only 17 of the 118 other newborns in the remainder of the cohort (14.4%) (p = 0.001, Mann-Whitney U). Blood clotting disorders thought to predispose infants to stroke (polycythemia, factor V Leiden, hemoglobin disorders, MTHFR [methylene tetrahydrofolate reductase, L677C], and prothrombin mutations) were not identified in the four cases tested (Cases 3 through 6); these tests were not routinely available to Cases 1 and 2 at the time of birth.

All six patients presenting with stroke had some degree of abnormality on neurologic examination at 30 months of age. Two newborns had cerebral palsy at follow-up (see table 1). The other newborns had abnormalities of muscle tone and/or stretch reflexes without functional weakness. Five of six infants had an MDI score greater than two standard deviations below the normal mean (see table 2). In order to determine the contribution of the infarct to cognitive outcome measured by the MDI score and the occurrence of cerebral palsy, newborns with infarcts were compared to the other newborns in the cohort with similar patterns of brain injury without the infarct (no lesion, watershed injury). Infarcts appeared to be associated with worse neurodevelopmental outcome, in particular lower cognitive test scores (see table 2).

Discussion.

In a prospective cohort of neonatal encephalopathy, perinatal strokes were an uncommon but serious occurrence with a distinct clinical presentation. All six term newborns with stroke presented with seizures in the context of severe encephalopathy on the first day of life and all subsequently had an abnormal neurodevelopmental outcome, with more cognitive than motor abnormalities. The occurrence of stroke in our cohort was substantially more frequent than population-based estimates, even though newborns with congenital infection, malformations, or metabolic disease were excluded from this cohort. While some of this increase may relate to referral bias of sicker patients to our tertiary care center, our cohort nonetheless included the entire spectrum of severity of neonatal encephalopathy from very mild to severe. This suggests that neonatal encephalopathy, particularly that related to hypoxia-ischemia, should remain a risk factor for perinatal stroke.

Consistent with observations of childhood stroke, our newborns had multiple risk factors for brain injury,8 particularly intrapartum complications. Intrapartum complications were prominent in this series of stroke patients: complicated delivery in four and fetal distress in two. The occurrence of delivery complications is consistent with previous reports suggesting that nonpenetrating head injury during the birthing process can result in thrombosis by occlusion of the internal carotid and middle cerebral arteries.6,9⇓ While pregnancy and maternal complications previously associated with neonatal encephalopathy10 were observed in this cohort, their occurrence was similar in newborns with and without stroke.

Previous reports have demonstrated that stroke in term newborns without significant neonatal encephalopathy has a favorable neurodevelopmental outcome in over 90% of infants.4,5⇓ The majority of patients in these two previous studies presented with seizures (in the absence of birth asphyxia) and isolated unilateral focal infarctions. This is in contrast to the newborns in our series with encephalopathy and seizures, who all had abnormal neurodevelopmental outcome with motor and cognitive deficits at 30 months of age, despite infarcts involving less than 2/3 of the middle cerebral artery territory. Moreover, the neurodevelopmental outcome in our series of infants with perinatal stroke was significantly worse than the remainder of our cohort of encephalopathic term neonates, although a larger sample size would have been desirable. In particular, the newborns with perinatal stroke had lower Mental Development Indices than newborns with similar patterns of brain injury in the absence of infarcts (i.e., watershed injury without focal stroke or encephalopathy with normal MRI studies). Not all newborns with stroke and abnormal cognitive scores had cerebral palsy, indicating that clinicians should be alert to cognitive abnormalities in the absence of significant motor deficits. Overall, the occurrence of perinatal stroke in term infants within the context of neonatal encephalopathy results in poor neurodevelopmental outcome, particularly cognitive abnormalities, compared to either event occurring in isolation.

The recent observation that brain imaging demonstrates evidence of an acute insult in 80% of newborns with neonatal encephalopathy suggests that events in the immediate perinatal period are important in the origin of neonatal brain injury.3 These data, together with the high incidence of intrapartum complications in newborns with stroke and encephalopathy, suggests that the role of intrapartum events in the pathogenesis of perinatal stroke should be re-examined.