Background and Purpose. The aim of this study was to evaluate the occurrence of prothrombotic disorders in a well-characterized cohort of infants with neonatal stroke and to document any association of prothrombotic disorders with the type of infarct seen on magnetic resonance imaging (MRI) and clinical outcome.
Methods. Twenty-four infants with perinatal cerebral infarction confirmed by neonatal MRI were enrolled in the study. All the infants and, when possible, both parents were tested to identify inherited and acquired prothrombotic disorders.
Results. None of the infants had a significant bleeding diathesis, but 10 (42%) had at least 1 prothrombotic risk factor. Five children showed heterozygosity for factor V Leiden, and 6 had high factor VIIIc concentrations. There was a striking association between the occurrence of these abnormalities and both the presence of cerebral hemorrhage on MRI and poor neurologic outcome. Eight of the 11 patients (73%) with hemiplegia or global developmental delay had factor V Leiden and/or raised factor VIIIc, whereas only 1 of the 13 patients (8%) with normal outcome had any prothrombotic risk factors. In particular, all 5 infants with factor V Leiden had hemiplegia, compared with only 4 of the 19 infants without factor V Leiden (21%).
Conclusions. These data suggest that the presence of prothrombotic risk factors and, in particular, of the factor V Leiden mutation, is significantly associated with poor outcome after perinatal cerebral infarction.
As a consequence of the wider availability of brain imaging in the newborn, neonatal cerebral infarction has become an increasingly recognized entity, and recent population-based data report an incidence of 1 in 4000 term infants.1,,2 In the past, these lesions were mainly diagnosed on postmortem examination and were always associated with severe conditions, such as sepsis, congenital heart disease, disseminated intravascular coagulation, or severe asphyxia. In the last decade, however, several studies using early and serial imaging have demonstrated that infarcts can also be observed in the absence of severe antenatal and/or perinatal events.3,,4
Although neonatal imaging has dramatically improved the detection of infarcts and may also provide information on the timing of the insult,5 the cause of these lesions is usually unknown. By contrast, there is increasing evidence that inherited or acquired prothrombotic disorders may play a significant role in patients who have suffered a stroke. In particular, the incidence of factor V Leiden mutation and/or increased factor VIII have been found to be significantly associated with increased risk for thrombosis.6–13 The majority of these studies, however, reported strokes occurring in childhood or adulthood6–10and only a small number of newborns with infarction have been studied.11–14
The aim of this study was to assess the thrombophilic status in a cohort of infants who showed cerebral infarction on neonatal magnetic resonance imaging (MRI) and to establish: 1) the prevalence of factor V Leiden mutation and of increased factor VIII in these infants, and 2) whether the presence of these risk factors is related to the type and extent of lesions on MRI scans or to outcome.
The study has been approved by the Hammersmith Hospital Trust Research Ethics Committee. Twenty-four infants who showed evidence of cerebral infarction on neonatal MRI were enrolled in this prospective study. In 23 of the 24 infants, the infarct was detected after convulsions in the first days after birth; in 1 infant, the lesion was an incidental finding when the infant was scanned on day 2 as part of a normal control group. All infants had normal Apgar scores at 5 minutes (≥8) and normal cord pH (>7.2). Twenty-one of the 24 were white and 3 were Asian.
The infants were imaged on a 1.0 Tesla Picker HPQ system using conventional T1-weighted spin echo (860/20 ms), inversion recovery (3800/30/950), and T2-weighted spin echo (3000/120 ms) sequences. Although infants have been serially studied, in this study, we only assessed the scans performed between 1 and 4 weeks after birth, when perinatally acquired lesions are at their most obvious. The scans were assessed by an independent observer (M.R.) who was blind to the hematologic data.
The infarcts were classified according to the arterial distribution of the lesions. This was based on the location, extent, and shape of the lesions.4 Infarcts were characterized as being in the territory of the main arteries or in a border zone distribution. The infarcts in the territory of the main arteries were further subdivided according to whether they occurred in the main branch or in 1 of the cortical branches of the artery.
The lesions were also classified according to whether the ischemic lesion was associated with hemorrhage, either within the ischemic lesion or in other sites.
Follow-up assessments were performed at 6, 12, 18, and 24 months of age, and then at least yearly after that. All the infants were assessed with a structured neurologic examination15 and on the developmental scales of Griffiths.16
Coagulation and Thrombophilia Profile
The coagulation profile included prothrombin time, activated partial thromboplastin time, thrombin time, platelet count, fibrinogen, and von Willebrand factor antigen. The prothrombotic screen included measurement of factor VIIIc, protein C, protein S, and antithrombin. In addition, the presence of the factor V Leiden mutation and the G → A transition at position 20210 of the prothrombin gene were screened using polymerase chain reaction-based mutation analysis. Blood groups were determined as factor VIII levels are related to the blood group systems of A, AB, B, and O.
Blood samples from children and parents were drawn into 0.105 M trisodium citrate bottles for coagulation tests and ethylenediaminetetraacetic acid bottles for full blood counts, blood film, and blood grouping. Blood samples were immediately centrifuged at 4°C for 10 minutes at 3000 rpm; platelet-poor plasma was separated and either tested immediately or stored in aliquots at −80°C for testing. Ethylenediaminetetraacetic acid samples were frozen at −40°C for DNA analysis. Coagulation screens were performed on fresh platelet poor plasma on the ACL 300 R (Instrumentation Laboratory, Lexington, MA) and KC10 (Brownes Ltd, Reading, UK) instruments before January 1997 and on Sysmex CA 6000 (Sysmex UK Ltd, Milton Keynes, UK) after January 1997. All the children born after May 1996 were tested within the first 2 weeks after birth (n = 13). The children born before that date (n = 11) were tested at 1 of their follow-ups.
Data were analyzed using χ2 tests with Fisher's exact test, where appropriate. Testing the level of significance was set at 0.01 because of multiple significances.
Twenty-two of the 24 children had arterial infarction in the distribution of the middle cerebral artery and 2 had border zone lesions. Of the 22 children with infarction of the middle cerebral artery, 5 had an infarction in the main branch and 17 in 1 of the cortical branches. Eleven of the 24 children also showed signal changes on T1 and transverse relaxation time images consistent with hemorrhage, which was always in other sites (Figs 1and 2). Details of the site and size of the lesions are shown in Table 1.
All children were at least 24 months old at the time of their final assessment (range: 2–7 years). Thirteen children had a normal outcome, 9 had a hemiplegia, and 2 had global developmental delay but no signs of hemiplegia. In all 9 children with hemiplegia, the clinical signs suggestive of hemiplegia were already present before the age of 2 years.
Coagulation and Thrombophilia Profile
None of the 24 children studied had thrombocytopenia or any significant abnormality in the coagulation profile, and there was no evidence of disseminated intravascular coagulation.
Thrombophilia screening was normal in 14/24 patients (58%). In the 10 patients with abnormal results, the abnormalities were increased factor VIIIc (n = 6) and heterozygous factor V Leiden (n = 5). One patient had both increased factor VIIIc and heterozygous factor V Leiden. Table 1 shows the details of the individual findings. Parental samples were available for 4 of the 5 children with factor V Leiden heterozygosity. In all 4, 1 of the parents (the father in 2 and the mother in the remaining 2), was also heterozygous for factor V Leiden. Of the 6 children with increased factor VIIIc, parental samples were available for 11 parents and were normal in 10 (1 mother had a raised factor VIII when tested 11 months postnatally). In 4 of the 6 children, the increased factor VIIIC was detected at follow-up. Antithrombin, protein C, and protein S were normal in all of the infants tested and in their parents. There was no evidence of the G20210A mutation of the prothrombin gene in any of the children tested.
Correlation Between Hematologic Factors and MRI Findings
Hematologic Factors and Extent and Site of Lesion
There was no association between the extent of the lesions and abnormal hematologic factors. Both normal and abnormal factor VIIIc and factor V Leiden heterozygosities were found in lesions in the territory of cortical and main branch arteries. Both children with border zone lesions had increased factor VIIIc.
Hematologic Factors and Hemorrhage
Factor VIIIc was increased both in children with purely ischemic lesions and in those with additional hemorrhage (4/13 and 2/11). Factor V Leiden heterozygosity was found in 5/11 children with a hemorrhage and in none of the infants with purely ischemic lesions (χ2, P = .006; Fisher's exact test, P = .01).
Outcome and Hematologic Factors
Eight of the 11 children (73%) with hemiplegia or developmental delay had at least 1 abnormality of their thrombophilic profile, whereas only 2 of the 13 (8%) with normal outcome had abnormal thrombophilic profile (χ2, P = .005). Factor V Leiden heterozygosity was significantly associated with the presence of a hemiplegia (Fisher's exact test, P = .003). All 5 children with factor V Leiden developed a hemiplegia, compared with only 4 of the 20 children without this mutation.
Of the 6 patients with increased Factor VIIIc, 3 had hemiplegia and 1 had global delay, whereas 2 had a normal outcome. The difference was not statistically significant. Table 1 shows individual details of outcome and of the coagulation and thrombophilia profile.
The results of this study showed that ∼40% of infants with neonatal cerebral infarction had prothrombotic risk factors. Specifically, 5 infants were heterozygous for factor V Leiden, 1 of whom also had a raised factor VIIIc, and 5 infants had raised factor VIIIc as an isolated finding.
The prevalence of factor V Leiden mutation in our cohort (24%) was substantially higher than that reported in the normal population, which has been estimated between 2.7% and 10% in Europe and North America. In the United Kingdom, the frequency of factor V Leiden mutation is 3.4%.19 Factor V Leiden has previously been reported in childhood and adult stroke9 and is well-known as a risk factor for venous thrombosis.20–22 Recently, it has also been found in neonatal porencephalopathy and in other newborn infants with arterial thrombosis.11,,23
This study is the first to demonstrate a relationship between the presence of factor V Leiden and both the nature and outcome of a major vascular occlusive event. All 5 infants in this study with factor V Leiden had residual hemiplegia (100%), whereas only 4 of the 19 infants without factor V Leiden (21%) had hemiplegia.
The reasons for this observation are not clear. In our cohort, factor V Leiden was not specifically associated with the size of the infarct because it was found in both main branch and cortical branch distributions. All of the infants with factor V Leiden, however, showed additional hemorrhagic lesions, which were found in only 30% of infants without factor V Leiden. We have previously reported that the presence of MRI abnormalities within the basal ganglia, the cerebral hemisphere, and in the posterior limb of the internal capsule predicted the development of hemiplegia.4 All the infants with factor V Leiden in the present study had MRI abnormalities in these 3 areas. It is, therefore, possible that although the presence of prothrombotic risk factors is not associated with the absolute size of an infarct, they may influence both its nature and the combination of sites involved, thereby leading to a high risk of hemiplegia. Factor V Leiden might equally modulate the chance of reperfusion after ischemia or the ability of neural tissue to repair after the insult.
The association between factor V Leiden and hemiplegia support recent findings by Debus et al23 and Nelson et al.24Both studies looked retrospectively at children with cerebral palsy and/or neurodevelopmental delay and found a high incidence of prothrombotic factors. Debus et al23 also found that >25% of the infants in their cohort had deficiencies of proteins C and S. These were normal in our cohort. This may reflect both sampling error and different gene pools among the study populations.
The other primary finding in the present study was the presence of a raised factor VIIIc level in 6 of the 22 infants in whom this was measured (27%). This is similar to the prevalence of raised factor VIIIc (>1.5 IU/mL) among adults with venous thrombosis in the Leiden Thrombophilia Study (25%)25 and in patients with ischemic cerebrovascular disease and compares with an expected prevalence of raised factor VIIIc in the normal healthy population of 11%. Unlike these studies, we did not demonstrate an increased level of von Willebrand factor, which has also been associated with an increased risk of cerebral infarction.25
Family studies were available for 11 of the parents of the 6 infants with a raised factor VIIIc but only 1 mother had a moderately raised factor VIIIc, which could possibly suggest a genetic basis for the raised factor VIIIc. Factor VIII is well known to behave as an acute phase reactant and so it is possible that increased factor VIII may, in some cases, be a result of rather than a cause of the stroke. However, of the 6 children with raised factor VIIIc, 4 maintained raised factor VIIIc levels beyond the neonatal period, 1 returned to normal at 5 months of age, and 1 was not retested. Normal levels of antithrombin, protein C, and protein S in all the parents tested make it seem unlikely that an anticoagulant deficiency has been missed as a result of the difficulty in applying normal ranges to young children.
These data indicate that the role of elevated factor VIIIc in neonatal stroke is worthy of additional investigation. Three of the 6 infants who had abnormal outcome but did not have the factor V Leiden mutation had increased factor VIIIc (>1.5 IU/mL). Taken together, our results show that infants with a neonatal stroke but a normal factor VIIIc and factor V Leiden seem more likely to have a favorable long-term outcome; only 3/14 (21%) having hemiplegia or developmental delay. In contrast, the presence of factor V Leiden and/or a raised Factor VIIIc in infants with neonatal stroke is associated with a high risk (80%) of poor neurologic outcome.
The high prevalence of factor V Leiden in our cohort also suggests that genetically determined abnormalities of thrombophilia play a role in the cause of cerebral infarction. These lesions are probably attributable to a multifactorial cause rather than to a single factor. Although by using early and serial imaging we were able to time the insult to the perinatal period, none of our children had severe acute perinatal events that could be clearly implicated as responsible for the lesion. It is, therefore, likely that genetic factors increase the risk for thrombosis, which can then be triggered even by otherwise minor adverse perinatal events.
This study highlights the need to assess all infants with focal hypoxic-ischemic and/or hemorrhagic brain injury for prothrombotic risk factors. The results have implications for prognosis and likely development of cerebral palsy, as well as for the possibility of future thrombotic events. Preventive measures and counseling of families may be indicated.
- Received June 26, 2000.
- Accepted December 4, 2000.
Reprint requests to (E.M.) Department of Paediatrics, Hammersmith Hospital, Du Cane Road, London, W12 OHN, United Kingdom. E-mail:
- MRI =
- magnetic resonance imaging •
- T1 =
- longitudinal relaxation time
- Mercuri E,
- Rutherford M,
- Cowan F,
- et al.
- Cowan FM,
- Pennock JM,
- Harahan JD,
- Manji KP,
- Edwards AD
- Ganesan V,
- McShane MA,
- Liesner R,
- Cookson J,
- Hann I,
- Kirkham FJ
- ↵Dubowitz L, Dubowitz V, Mercuri E. The Neurological Assessment of the Preterm and Full-Term Infant. London, United Kingdom: McKeith Press; 1999
- ↵Griffiths R. The Abilities of Babies: An Association for Research in Infant and Child Development. Piscataway, NJ: Amersham Pharmacia Biotech; 1976
- Andrew M,
- Paes B,
- Milner R,
- et al.
- Andrew M,
- Vegh P,
- Johnston M,
- Bowker J,
- Ofosu F,
- Mitchell L
- Copyright © 2001 American Academy of Pediatrics