American Journal of Hematology 73:81–86 (2003)
Factor V Leiden and Prothrombin Gene G20210A Mutation in Children With Cerebral Thromboembolism Mariana Bonduel,* Gabriela Sciuccati, Mirta Hepner, Graciela Pieroni, Aurora Feliu´ Torres, Claudia Mardaraz, and Juan Pablo Frontroth Servicio de Hematologı´a-Oncologı´a, Hospital de Pediatrı´a “Prof. Dr. Juan P. Garrahan”, Buenos Aires, Argentina
We investigated whether there is an association between factor V Leiden (FVL) and/or prothrombin gene G20210A mutation (PT20210A) and cerebral thromboembolism in a pediatric Argentinean population. From May 1992 to January 2002, 44 consecutive children with arterial ischemic stroke (AIS) and 23 children with cerebral sinovenous thrombosis (SVT) were prospectively studied at a single center. The prevalence of both mutations was compared with a 102 age-matched controls. In children with AIS, the frequencies (patients vs. controls), odds ratio (OR), and 95% confidence interval (95% CI) for the presence of FVL were as follows: 2.3% vs. 2%, OR/95% CI, 1.16/0.2 to 13.2; P value = 0.99. No cases of PT20210A were found in this group. In children with SVT, the frequencies (patients vs. controls), OR, and 95% CI were as follows: FVL (4.3% vs. 2%, OR/95% CI, 2.27/0.22 to 6.2; P value = 0.99) and PT20210A (4.3% vs. 1%; OR/95% CI, 4.6/0.3 to 76.3; P value = 0.3354). One child with PT20210A also had an inherited protein C deficiency. In 12 (18%) out of the 67 children with cerebral thromboembolism, without the aforementioned mutations, other prothrombotic disorders were detected. Although a multi-center prospective study with a large number of Argentinean pediatric patients is needed to obtain considerable evidence, no association between factor V Leiden and/or prothrombin gene G20210A mutation and cerebral thromboembolism was found in this © 2003 Wiley-Liss, Inc. pediatric series. Am. J. Hematol. 73:81–86, 2003. Key words: factor V Leiden; prothrombin gene G20210A mutation; stroke; children
INTRODUCTION
Cerebral thromboembolism (CTE) is a rare entity with an incidence estimated at 0.63–1.2 per 100,000 children per year and a ratio of arterial ischemic stroke (AIS) to sinovenous thrombosis (SVT) of 3:1 [1–3]. Clinical awareness and more sensitive neuroimaging techniques have increased the diagnosis of these entities, which were probably underdiagnosed in the past [3]. Resistance to activated protein C (APC), in the majority of cases due to arginine in residue 506 to glutamine, results in the production of a mutant factor V protein resistant to inactivation by APC [4]. Factor V Leiden (FVL) has been reported to have a prevalence of approximately 3–11% in the general Caucasian population [5] and has emerged as the most common association of venous thrombosis in adults [6]. The recently described prothrombin gene G20210A mutation (PT20210A) within the 3⬘-untranslated region of the prothrombin gene has been described as common but probably mild risk factor of venous thrombosis in adults [7]. This mutation is present in © 2003 Wiley-Liss, Inc.
about 2% of the Caucasian population [8]. PT20210A possibly has a similar distinctive ethnic and geographical distribution, as has been described for FVL. Most of the studies in adults did not find FVL and PT20210A as risk factors for AIS [9–12]. FVL and PT20210A were associated with cerebral venous thrombosis in Italian adults [13,14]. However, there is controversial evidence between the association of these mutations and CTE in children depending on their ethnic background and geographical distribution [15–21].
*Correspondence to: Mariana Bonduel, M.D., Servicio de Hematologı´a-Oncologı´a, Hospital de Pediatrı´a “Prof. Dr. Juan P. Garrahan”, Combate de los Pozos 1881, Buenos Aires, C1245AAM, Argentina. E-mail:
[email protected],
[email protected] Received for publication 28 May 2002; Accepted 15 February 2003 Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ajh.10326
82
Bonduel et al.
The main purpose of this study is to determine whether there is an association between these mutations and cerebral thromboembolism in a pediatric Argentinean population. The frequency of other inherited and acquired prothrombotic disorders was also reported. PATIENTS AND METHODS
From May 1992 to January 2002, 44 consecutive children with AIS (14 F/30 M) and 23 children with SVT (3 F/20 M) receiving care in the Hospital de Pediatrı´a “Prof. Dr. Juan P. Garrahan”, Buenos Aires, Argentina, were studied. The inclusion criteria encompassed infants and children aged from 2 months to 16 years with first onset of cerebral thromboembolism. Newborns and children with congenital heart diseases and vascular abnormalities were excluded from this study for additional evaluation. The thrombotic events were confirmed by computed tomography, magnetic resonance (MR) imaging, and MR angiography or arteriography. All data were collected prospectively. Each patient was evaluated by the hematologists involved in this study and a neurologist at the initiation of the study and a minimum of once yearly. A standardized history including clinical presentation, personal and family history of thrombosis, underlying disorders and/or circumstantial risk factors, radiological studies, laboratory findings, treatment, and outcome was updated at each visit. With informed parental consent, a total of 102 healthy children (42 F/60 M; median age 7.1 years, range 0.2– 15.9 years) admitted to the hospital for elective minor surgery or as potential bone marrow donors were eligible for reference ranges and as control group for DNA analysis. The children were gathered in groups of ages from 0.2 to 1 to 5 (n ⳱ 28), 6–10 (n ⳱ 30), and 11–16 years (n ⳱ 31). The laboratory evaluation began after obtaining informed parental consent. Blood samples were collected into 0.11 M sodium citrate at a ratio of 9:1 by clean venipuncture. Blood was centrifuged at 2,500g for 10 min to obtain platelet-poor plasma (PPP). The PPP was then centrifuged for an additional 5 min at 2,500g, and plasma was immediately processed or stored in 1 mL aliquots at −70°C. For DNA analysis blood was collected into glass Vacutainer tubes containing EDTA (Becton Dickinson Vacutainer威 Systems, Franklin Lakes, NJ). DNA was extracted from white cells fraction using DNAzol威 reagent (Life Technologies, Grand Island, NY). To determine FVL and PT20210A simultaneously, we performed a multiplex allele specific amplification PCR as previously described in patients and controls [22]. Four children, who were admitted between 1992 and 1994, were recalled later in order to obtain DNA samples for studying FVL and other mutations that could be dis-
covered over the following years, as has been the case of PT20210A. All the patients, including the four aforementioned children, were evaluated for prothrombotic disorders. Evaluation for prothrombotic disorders included the following assays: prothrombin time, activated partial thromboplastin time, thrombin time, reptilase time, and fibrinogen and factor V activity by standard procedures. Functional activities of protein C (PC), antithrombin (AT), and plasminogen (PLG) were measured by amidolytic assays (Chromogenix AB, Mo¨lndal, Sweden) and PC and protein S (PS) by clotting assays (Diagnostica Stago, Asnie`res, France) using a Model ST4 coagulometer (Diagnostica Stago). PC antigen was measured by a commercial ELISA kit (Diagnostica Stago). Immunologic measurements of PS (total and free), AT, and PLG were made by Laurell’s technique [23] using polyclonal rabbit antibodies against the respective antigens (PS and AT, Dakopatts, Glostrup, Denmark; ASSERA PLG, Diagnostica Stago). Free PS was assayed by precipitating bound protein with polyethylene glycol 8000 (3.75%) according to the method by Comp et al. [24]. The presence of activated PC resistance (APCR) was tested using a commercial kit with factor V-deficient plasma as a prediluent for plasma samples (Chromogenix AB). The presence of lupus anticoagulant (LA) was determined on the bases of abnormal screening assays, of mixing studies, and of confirmatory assays based on the method that initially gave the abnormal screening test results [25]. The tests used were activated partial thromboplastin time (PTT-LA, Diagnostica Stago) and diluted Russell viper venom time (Sigma-Aldrich, St. Louis, MO) with confirmation using a platelet neutralization procedure [26], as recommended by the International Society on Thrombosis and Haemostasis [25]. Sera for anticardiolipin antibodies (ACA) were assayed for IgG and IgM isotypes using an in-house ELISA previously described [27], calibrated against Dr. Harris’ standards from the University of Louisville, Kentucky. According to proposed criteria for the antiphospholipid syndrome (APS), only those patients (pts) with AIS or SVT and positive LA test results and/or positive IgG and/or IgM ACA at moderate/high levels (>20 GPL or MPL units) in two determinations performed more than 6 weeks apart were considered positive for APS [28]. LA and ACA were tested within 1 month after the thrombotic event in all cases. Their parents were also studied to discard familial APS. In five children with cerebral thromboembolism and acute lymphoblastic leukemia (ALL), the assay for detection of acquired AT deficiency was performed immediately after the thrombotic episode was radiologically confirmed. At least 3–6 months after the thrombotic event, evaluation for inherited prothrombotic disorders was carried
FVL and PT20210A in Children With Stroke
out. The final diagnosis was made when repeatedly measured plasma concentrations of the coagulation proteins investigated were outside the age-appropriate reference ranges. The finding of the same abnormal laboratory results in the parents confirmed the suspected inherited coagulation defect. In children who were under oral anticoagulation plasma samples were obtained 15–30 days after withdrawal of the therapy. Two hundred twelve subjects were recruited from healthy professionals and technicians of our hospital. These adult samples were used to establish the reference values we required to evaluate the results obtained from patients’ parents. In this way we were able to determine the hereditary pattern of the prothrombotic disorder in this pediatric population. We applied separate local laboratory reference ranges for women and men in the assessment of PS levels. A positive family history of venous and/or arterial thrombosis (including unusual sites) required a first- and/ or second-degree relative with thrombosis at an age younger than 45 years. Statistical Analysis
For comparison, Fisher’s exact test was used. Odds ratios and 95% confidence intervals were calculated for FVL and PT20210A. Statistical significance was taken as a P value ⱕ0.05, for two extremes. The software used was CSS/Statistica, 5.1 (StatSoft Corp., Tulsa, OK) [29,30]. RESULTS
Data of baseline demographics, clinical data, underlying diseases and/or circumstantial risk factors, location of the thrombotic event, and family history of thrombosis in our cohort are shown in Table I. The median ages at the first thrombotic onset in children with AIS and SVT were 7.7 years and 7.2 years, respectively. There was a male predominance in both groups. Risk factors were detected in 9 children (20.5%) with AIS and in all but 3 children (87%) with SVT. Eight children (11.9%) had a positive family history of thrombosis. The frequencies (patients vs. controls), odds ratio (OR), 95% confidence interval (95% CI), and P value obtained from the comparison between children with CTE and FVL and PT20210A versus the control group are summarized in Table II. One child with the PT 20210A also had an inherited PC deficiency. This patient was a 15-year-old boy who presented a superior sagittal sinus thrombosis and deep left femoral vein thrombosis. He developed a pulmonary embolism while receiving low molecular weight heparin therapy. In patients without the aforementioned mutations we
83
found abnormalities in children with AIS (inherited PC deficiency, 1 pt; inherited PS deficiency, 1 pt; acquired AT deficiency, 1 pt; LA, 1 pt; ACA, 1 pt; and inherited PLG deficiency combined with LA, 1 pt) and in children with SVT (inherited AT deficiency, 1pt; acquired AT deficiency, 4 pts and LA, 1 pt). No cases of familial APS were found in this series. DISCUSSION
Within the past decade, several publications described inherited and acquired prothrombotic disorders as possible thrombophilic factors in a variable percentage of children with cerebral thromboembolism [15–21,31–35]. The prevalence of FVL and PT20210A mutation are 1.6– 2.4% and 2–2.4%, respectively, in the Argentinean population [36,37]. These observations need to be kept in mind for prediction of the risk of thrombosis emanating in different populations, for FVL, PT20210A, or their co-inheritance. The evidence for these mutations as risk factors for arterial disease in adults appears conflicting, as most studies showed a lack of association between FVL and/or PT20210A and arterial ischemic stroke [9–12]. Whether being a carrier of FVL and PT20210A is associated with arterial ischemic stroke in children remains controversial considering the different ethnic background and geographical distribution of the pediatric series reported [15–21]. Previous studies from several countries, which considered different inclusion criteria than ours, examined the prevalence of FVL and PT20210A mutation in children of some ethnic groups with AIS [15–21]. Although studies from Austria, Germany, Israel, and Turkey found FVL as risk factor for AIS, two studies from UK and our study from Argentina did not find this association [15,16,18–21]. Concerning PT20210A, whereas studies from Germany and Turkey found this mutation as a risk factor for AIS, studies from Austria, Israel, UK, and this present study did not find an association between PT20210A and AIS in children [15,18–21]. In Table III, the findings of our study have been expanded to include geographical distribution of the study populations, sample sizes, inclusion criteria, and control group selection. In this way the findings of our study can be directly compared and contrasted to the findings of the other aforementioned studies. Due to the fact that AIS in children has been underdiagnosed in most countries, future trials with larger sample sizes would give greater power to the studies required to establish significant differences in the prevalence of these mutations between cases and controls. FVL and PT20210A were associated with cerebral venous thrombosis in Italian adults [13,14]. Two previous studies from Germany and Canada reported the prevalence of these mutations in children with SVT [38,39]. In
84
Bonduel et al. TABLE I. Patients’ Characteristics Patients with arterial ischemic stroke (n ⳱ 44)
Patients with sinovenous thrombosis (n ⳱ 23)
6 13 13 12 7.7 (0.2–15.2) 14/30
5 4 8 6 7.2 (0.7–15.4) 3/20
5 — 2 1 1 1 — 1 — — 1 35
10 3 2 4 — 1 1 — 1 2 — 3
40 22 13 12 7 7
11 16 10 11 1 —
29 6 3 2 2 2 — — — — — 4
— — — — — — 11 5 3 3 1 4
Age >2 months to 1 year to 5 years >6 years to 10 years >11 years Age median (range), years F/M Underlying disease and/or circumstantial risk factor Infection Dehydration Head trauma ALL with ASP therapy Non-Hodgkin’s lymphoma Non-Hodgkin’s lymphoma with ASP therapy Hodgkin’s disease Biliary atresia Surgery Brain tumour Hyperlipidemia type IV None Symptoms Hemiparesis Headache Seizures Change in mental status Aphasia Ataxia Radiologic findings Middle cerebral artery occlusion Posterior cerebral artery occlusion Carotid artery occlusion Vertebral artery occlusion MCA and PCA occlusion Anterior cerebral artery and MCA occlusion Superior saggital sinus thrombosis Lateral sinuses thrombosis Saggital and straight sinuses thrombosis Saggital and lateral sinuses thrombosis Internal cerebral vein Family history of thrombosis
*ALL, acute lymphoblastic leukemia; ASP, L-asparaginase; MCA, middle cerebral artery; PCA, posterior cerebral artery. TABLE II. Prevalence of Factor V Leiden and Prothrombin Gene G20210A Mutation in Cases and Controls FVL
AISa SVTa
PT20210A
No. of cases (prevalence)
No. of controls (prevalence)
OR
95% CI
P value
No. of cases (prevalence)
No. of controls (prevalence)
OR
95% CI
P value
44 (2.3%) 23 (4.3%)
102 (2%) 102 (2%)
1.16 2.27
0.2–13.2 0.22–6.2
0.99 0.99
44 (0%)* 23 (4.3%)
102 (1%) 102 (1%)
— 4.6
— 0.3–76.3
— 0.3354
a AIS, arterial ischemic stroke; SVT, cerebral sinovenous thrombosis. *OR, CI 95%, and P value of the association between PT20210A and AIS could not be calculated.
our series, we found a child with FVL and another one with PT20210A among the 23 cases with SVT. Given the fact that the control group prevalences of our study were 2% and 1%, respectively, this may be indicative that
these mutations are more frequent in patients with SVT than in the healthy children evaluated. Therefore, a large number of Argentinean pediatric patients is needed to obtain considerable evidence for this conclusion.
FVL and PT20210A in Children With Stroke
85
TABLE III. Factor V Leiden and Prothrombin Gene G20210A Mutation in Children With Arterial Ischemic Stroke Study
FVL
Reference country/year
Age
Design
Zenz et al. Austria/1998 Ganesan et al. UK/1998
Newborns– 4 years 2 months– 15 years
Hagstrom et al. USA/1998
2 months– 20 years
Retrospective/ multi-center Retrospective and prospective/single institution Retrospective/single institution
Akar et al. 2 months– Turkey/1999 13 years McColl et al. Newborns– UK/1999 15 years Nowak-Go¨ttl et al. 6 months– Germany/1999 16 years Kenet et al. Israel/2000 Bonduel et al. Argentina/2002
NA Retrospective/ multi-center Prospective/ multi-center
1.5 months– Retrospective/single 18 years institution 2 months– Prospective/single 16 years institution
Inclusion criteriaa
Control group
No. of No. of cases controls (prevalence) (prevalence)
OR
95% CI
P value
AIS
General population
33 (18%)
152 (4.6%)
4.6*
1.4–14.7*
0.01
AIS MMS TIA AIS TIA SCD SVT* AIS
Hospitalized children
50 (12%)
77 (5.2%)
2.5*
0.7–9.3*
0.19
Healthy adults
18 (11%)
65 (3.1%)
3.9* 0.51–30.1*
General population
32 (28.1%) 189 (10.6%) 3.3*
1.3–8.13*
0.0115*
Cord blood samples Age-/sex-matched healthy children
37 (5.4%)
0.5–13.5%
0.2597*
AIS Spontaneous AIS AIS AIS
224 (2.2%)
148 (20.2%) 296 (4%)
2.5 6.0
2.97–12.1
0.2037*
