Editor—Two mutations have been identified in recent years that predispose a heterozygous carrier to venous thrombosis. One is a mutation localised to the factor V gene, Arg 506 to Gln (factor V Leiden mutation, FVL), which has been shown to be the most common cause of familial thrombosis1 2 through resistance to activated protein C (APC), which is an inhibitor of activated factors V and VIII.3 The second is in the gene for the coagulation factor prothrombin (factor II), the G20210A mutation.4

There are two published estimates of prevalence of theFVL mutation in the Australian population, the first in a study of recurrent miscarriage,5 where 3.5% of the controls had FVL. The other is a recent study of blood donors where 3.6% were found to be heterozygous.6

Fewer studies have been done on the prothrombin (PT) mutation which is not as prevalent, thought to occur in approximately 2% of healthy, normal controls.1 4

Of particular relevance to this study are the findings of children with ischaemic strokes7-9 or thromboembolism,10 11 who have been reported as having a high prevalence of the FVL mutation. A more recent study12 suggests that neitherFVL nor PT is a risk factor for childhood stroke and that “a large prospective multicentre study is required” to investigate this further. There has also been a report of three babies with hemiplegic cerebral palsy (CP) who were heterozygous for the FVLmutation.13 In the three cases, there was a suggestion that placental infarction/thrombosis or neonatal stroke may have occurred and resulted in the hemiplegia. Other relevant studies have also been done, suggesting that placental infarction and late fetal loss14 15 may occur more frequently in women who are carriers of FVL.

Cerebral palsy (CP) is a common physical disability in childhood occurring in 2.0-2.5 per 1000 live births. In a significant proportion, no cause can be established. Current research suggests that between 6 and 8% of cases are the result of birth asphyxia16-18and postnatal problems account for a further 10%.19 20Prenatal events are now thought to be responsible for approximately 75% of all cases of cerebral palsy, although it is usually impossible to determine the nature and exact timing of the damaging event. Given the lifetime impact that cerebral palsy has on the person and the family, a search for further causative factors is essential, as new information regarding aetiology may provide the first step to instituting preventive strategies.

In Victoria, there is a CP Register containing demographic and clinical details of 2093 babies born between 1970 and 1995. A letter was sent to the parents of 69 selected patients requesting permission to access the blood spot collected routinely at birth (Guthrie card) for newborn screening purposes. These 69 were chosen because they were private patients of one of the authors (DR) and could therefore be approached directly. The identifiable blood spots are held at The Murdoch Children's Research Institute and are not available for research or other purposes unless permission is obtained from the parents. All had a diagnosis of cerebral palsy with a specific motor diagnosis of spastic hemiplegia.

We estimated that a sample size of 50 would detect a significant increase in prevalence of the FVL mutation from a background prevalence of 4% to one of 15%, with a power of 80% and alpha of 0.05.

Two 3 mm² sections of Guthrie blood spots were cut out and placed in 50 μl of PCR buffer. They were initially incubated at 95°C for 30 minutes, then kept at 37°C overnight, after which they were pulse centrifuged. The supernatant was then transferred to a sterile 500 μl tube and kept at 4°C until the day of analysis.

DNA encompassing the FVL codon 506 mutation was amplified by PCR. The resulting product was checked on a 1% agarose gel stained with ethidium bromide before restriction enzyme digestion. DNA was digested overnight at 37°C using the enzymeMnlI. Fragments were then separated on a 6% polyacrylamide gel and subsequently stained with ethidium bromide. Both normal and heterozygote samples were run as controls in each assay.

The same methodology was applied for detection of the prothrombin 20 210 polymorphism, but with HindIII restriction enzyme used, and with digest fragments being separated on a 3% agarose gel.

Power calculations were done in SPSS for Windows, version 8.0, SAMPLE POWER, and the binomial comparison of the sample prevalence with the estimated population prevalence was also done in SPSS version 8.0.

Of the 69 parents contacted, 58 responded (84%). All gave permission for use of the blood spot, except for one where the child was in the care of foster parents who did not feel that it was appropriate to give consent. Two Guthrie cards were unable to be found leaving a sample of 55 cases.

Of the 55 subjects recruited, 54 and 52 could be amplified successfully for the FVL andPT mutations, respectively. Of the 54 samples screened for FVL, one was found to be homozygous while three were found to be heterozygous. Therefore, the frequency of cases with at least one mutation ofFVL was 7.4% which does not differ significantly from the average Australian population heterozygote prevalence of 3.6% (binomial test, one tailed p value=0.13). Only one of the 52 samples analysed for PT was found to be heterozygous (1.9%).

The clinical features of the five subjects with positive findings are as follows.

(1) Homozygous for FVL mutation. This child was born by caesarean section at 32 weeks following a pregnancy complicated by pre-eclamptic toxaemia. His birth weight was 1815 g. Hemiplegic cerebral palsy was diagnosed at the age of 4 months. A CT brain scan showed appearances consistent with infarction. There was no family history of note but testing showed that his mother is a heterozygote for FVL.

(2) Heterozygous for FVL mutation. These three children were all normal birth weight, term infants. One child required some resuscitation at birth and meconium had been passed before delivery. Her MRI showed middle and anterior cerebral atrophy infarction. The family history was unremarkable apart from the unexpected death of the paternal grandfather, from a myocardial infarct, at 50 years of age. The other two children had uneventful perinatal periods and hemiplegia became apparent during the first year of life. Both had CT scans which showed loss of cerebral hemisphere substance in both grey and white matter with enlarged ventricles, not necessarily typical of major vessel infarction. While one child had a negative family history, the other child's grandmother had sustained a deep venous thrombosis following a hysterectomy at the age of 50 years.

(3) Heterozygous for the prothrombin mutation. This child had an uneventful perinatal period. A diagnosis of hemiplegic cerebral palsy was made at the age of 8 months. A CT brain scan showed signs of infarction with almost complete absence of the left frontal temporal and parietal lobes, with appearances typical of congenital occlusion of the left anterior and middle cerebral arteries and subsequent porencephaly. There was no family history of note.

All available radiological findings were assessed and are summarised in table 1.

If only cases with radiological evidence of ischaemia (14 cases) are used in the calculations, the frequency of either of the thrombophilia mutations (three cases) is 21%. This is significantly higher than the population prevalence (binomial test, one tailed p value=0.013).

In summary, evidence of the influence of theFVL mutation on thrombosis led us to consider whether a significant attributable factor to CP may be the presence of this mutation, which could cause an adverse vascular event such as placental infarcts or stroke in utero or early in postnatal life. To test this hypothesis, we obtained the frequency of the mutation in blood spots from a sample of children, all of whom had hemiplegic CP. This group was chosen because it was thought to be relatively homogeneous and it seemed that if a vascular event were causative, then spastic hemiplegia could be the most likely outcome. However, an examination of their radiological findings, after the mutation detection had been undertaken, showed quite marked heterogeneity. It was among the 14 cases with a known ischaemic event that three mutations were found, giving a significantly higher prevalence of 21%.

From these results, we conclude that there may be a relationship between carrier status for mutations predisposing an infant to thrombophilia and cerebral vascular events in utero (or neonatally) that lead to CP. To confirm such a relationship it will be necessary to study a larger sample of CP patients with a vascular basis, such as venous or arterial occlusion, and compare mutation frequencies with children with CP who do not have a vascular basis for CP.

In addition, investigation of maternal mutation status would be useful, particularly because the frequency of maternal pre-eclampsia in this sample was higher than expected (data not shown), 11% compared with the reported population frequency of approximately 5%21 22 and it is not known whether these mothers hadFVL or PTmutations.

We report our findings at this stage to illustrate how DNA testing for common polymorphisms, which are of unknown importance in the general population, may be of importance to subjects within subgroups of the population. Primarily this would be for clinical management, but perhaps when potentially interactive aetiological factors, such as smoking in pregnancy, are better defined, this genetic information could be used for preventive measures.