This study evaluated 56 children with a family history of early thromboembolism but no personal history. Comprehensive thrombophilia testing found at least one abnormality in most children, with 34% having two or more traits. Testing led to recommendations for transient anticoagulation in 71% during periods of increased risk. No thromboembolisms occurred during follow-up, though exposures were infrequent. Risk-stratified thrombophilia screening in children may help identify those at heightened risk for targeted prevention.

1. High Prevalence of Thrombophilic Traits in Children with
Family History of Thromboembolism
Meghan J. Calhoon, MS, Cassie N. Ross, Elizabeth Pounder, PA-C, Danielle Cassidy, PharmD, Marilyn J. Manco-Johnson, MD,
and Neil A. Goldenberg, MD, PhD
Objectives To determine a proximate family history of venous thromboembolism (VTE) in (1) the prevalence of
thrombophilia; (2) the frequency of recommended changes in management resulting from thrombophilia evaluation;
and (3) outcomes in longitudinal follow-up.
Study design Laboratory thrombophilia investigation was performed in 56 children with first- or second-degree
family history of thromboembolism before age 55 years, but without personal history of thromboembolism, who
were enrolled in a prospective inception cohort. VTE risk factors, family history, thrombophilia findings, and man-
agement recommendations were systematically collected, along with thromboembolism risk episodes/exposures,
prophylactic anticoagulation, major bleeds, and thromboembolism events during follow-up.
Results The frequencies of all thrombophilia traits were higher than the general population. Among 32 children
who underwent complete laboratory evaluation, 34% had $2 traits. Thrombophilia testing led to recommendations
for risk-based transient antithrombotic prophylaxis in 71% of subjects. No thromboembolism episodes developed
during more than 900 patient-months of follow-up, although at-risk exposures were infrequent.
Conclusion Risk-stratified approaches to primary prevention of pediatric VTE should be further evaluated in co-
operative prospective studies. (J Pediatr 2010;157:485-9).
T
hrombophilia can potentially be caused by any alteration in hemostasis whose net effect increases thrombin produc-
tion, enhances platelet activation/aggregation, induces endothelial activation/damage, or reduces fibrinolytic activity.
Mild inherited thrombophilia traits are common; for example, the factor V Leiden polymorphism is present in ap-
proximately 5% of Caucasians and increases the risk of venous thromboembolism (VTE) 2- to 7-fold.1 Severe thrombo-
philia traits are rare; the prevalence of homozygous protein C deficiency, for instance, is estimated at 1 in 4 million
births. Nearly all reported individuals were identified on the basis of purpura fulminans or thromboembolism (typically
life-threatening) in early childhood.2 Furthermore, numerous studies have indicated that VTE risk is also substantially
increased among individuals who possess 2 or more otherwise-mild inherited thrombophilia traits, among both children3
and adults.4-7
Which symptom-free individuals should be investigated for thrombophilia, the extent of testing, and the decisions to be
made on the basis of the findings are debated.8-12 Furthermore, present consensus-based guidelines rely upon mostly on
low-grade clinical evidence.13 Observational studies in adults have indicated that, while the annual incidence of VTE in
symptom-free carriers of factor V Leiden is <1%,14 18% of periods of clinical prothrombotic risk in such individuals result
in VTE, in the absence of antithrombotic prophylaxis.15 Still, thrombophilia assessment (whether limited to factor V Leiden
or more expansive) is most likely to have clinical utility among individuals with a personal or close family history of a throm-
botic event before the age of 55 years. Particularly in children and young adults, wherein the incidence of VTE is low, efforts to
identify individuals with clinical risk factors for VTE who have meaningful underlying thrombophilia may provide a rational
approach to VTE prevention by targeting a subpopulation at heightened risk.
The principal aims of this study were to determine thrombophilia findings,
recommendations for targeted antithrombotic primary prevention, and out-
comes in longitudinal follow-up of symptom-free children with a proximate From the Section of Hematology, Oncology, and Bone
family history of early thromboembolism. We hypothesized that, among such Marrow Transplantation, Department of Pediatrics, and
the Mountain States Regional Hemophilia and
children, comprehensive thrombophilia testing (1) frequently reveals underlying Thrombosis Center, University of Colorado Denver and
The Children’s Hospital, Aurora, CO (M.C., C.R., E.P.,
thrombophilia (of particular interest, severe or multitrait thrombophilia) and (2) M.M-J., N.G.); the Clinical and Investigational Pharmacy
Units, The Children’s Hospital, Aurora, CO (D.C.);
Division of Hematology/Oncology, Department of
Internal Medicine, University of Colorado Denver, Aurora,
APA Antiphospholipid antibody CO (N.G.); Venous Thromboembolism Trials Unit,
Colorado Prevention Center, Denver, CO (N.G.)
ELISA Enzyme-linked immunosorbent assay
Supported by the Centers for Disease Control and Pre-
FVIII Factor VIII vention (UR6/CCU820552 [M.M-J.] and the National In-
IgG Immunoglobulin G stitutes of Health National Heart Lung and Blood Institute
IgM Immunoglobulin M (1K23HL084055-01A1 [N.G.]). The authors declare no
conflicts of interest.
OCP Oral contraceptive pills
VTE Venous thromboembolism 0022-3476/$ - see front matter. Copyright Ó 2010 Mosby Inc.
All rights reserved. 10.1016/j.jpeds.2010.03.031
485

2. THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 157, No. 3
often leads to a recommendation for use of transient antico-
agulation as primary prophylaxis during periods of height- Table I. Prevalence of thrombophilia traits, by type, in
ened clinical prothrombotic risk. the study population of symptom-free children with
a proximate family history of early thromboembolism
Factor V Leiden
Heterozygous 34% (17/50)
Methods Homozygous 0% (0/50)
FII G20210A
With the approval of the Colorado Multiple Institutional Re- Heterozygous 17% (8/47)
Homozygous 0% (0/47)
view Board (no. 05-0339), and after informed consent was Protein C deficiency
obtained, children with a proximate family history (ie, first- Mild 7% (3/46)
or second-degree relative with) of a venous or arterial throm- Severe 0% (0/46)
Protein S deficiency
boembolism that occurred before the age of 55 years (with or Mild 0% (0/45)
without known thrombophilia), but without a personal his- evere 0% (0/45)
tory of thromboembolism, were consecutively enrolled in an Antithrombin deficiency
Mild 2% (1/47)
institution-based prospective inception cohort study be- Severe 0% (0/47)
tween March 1, 2006, and June 1, 2009. Subjects were ex- Elevated lipoprotein(a) 28% (13/46)
cluded if no thrombophilia evaluation was performed. Elevated homocysteine 0% (0/49)
Elevated factor VIII 17% (7/41)
Demographics, clinical risk factors for VTE, family history Antiphospholipid antibody
of thromboembolism, extent of thrombophilia testing, Lupus anticoagulant 8% (4/48)
thrombophilia findings, and management recommendations Anticardiolipin IgG, IgM 7% (3/46), 0% (0/46)
Beta-2-glycoprotein-I IgG, IgM 4% (2/45), 0% (0/45)
(including targeted antithrombotic prevention, as well as sys- Any 20% (9/45)
temic estrogen restriction) were systematically collected at
enrollment. In longitudinal follow-up, data on type and du- For definitions and thresholds used, see Methods.
ration of episodes/exposures involving heightened risk of
thromboembolism, observed thromboembolism events, ma-
jor bleeding complications, and duration of follow-up were 1.22. The upper limit of normal values for anticardiolipin
uniformly recorded. and beta-2-glycoprotein-I IgG and IgM was defined as less
Comprehensive thrombophilia testing consisted of the fol- than 10 U/mL. Persistent antiphospholipid antibodies
lowing panel: plasma protein C and antithrombin activities (APA) were defined by testing for the lupus anticoagulant,
(chromogenic assay); plasma free protein S antigen concen- anticardiolipin, or anti-beta-2-glycoprotein-I that was posi-
tration (enzyme-linked immunosorbent assay [ELISA]); tive at initial evaluation and again positive on follow-up as-
plasma homocysteine concentration (gas chromatography sessment at least 12 weeks thereafter. Cutoffs for abnormal
and mass spectrometry); factor V Leiden polymorphism laboratory values used locally-established pediatric reference
(polymerase chain reaction [PCR]); factor II G20210A poly- ranges in the case of FVIII activity, homocysteine, and
morphism (PCR); lupus anticoagulant in plasma (dilute Rus- dRVVT. For all other testing, reference ranges used published
sell viper venom time [dRVVT]); serum immunoglobulin G pediatric data or manufacturer’s recommendations. Labora-
and M levels of anticardiolipin anti-beta-2-glycoprotein-I tory evaluation was described as ‘‘complete’’ if all of the
(ELISA); plasma factor VIII (FVIII) activity (1-stage, acti- aforementioned testing was performed. Erythrocyte sedi-
vated partial thromboplastin time-based clotting assay); mentation rate and C-reactive protein were not routinely
and serum lipoprotein(a) concentration (ELISA). Mild in- measured.
herited protein C or S deficiency was defined as activity or With regard to the data analyses, descriptive statistics in-
free antigen levels (respectively) between 20% and 40% of volved calculation of frequencies (ie, proportions), medians,
the value for the pooled normal adult plasma standard on and observed ranges. Frequencies were compared between
two separate venipuncture specimens at least 6 weeks apart groups by c2 or Fisher exact test, as appropriate. For these
after age 6 months (12 months for protein C), in the absence hypothesis tests, alpha was set at 0.05. All statistical analyses
of clinical evidence of underlying infection or inflammatory utilized SAS version 9.1 (SAS Institute, Cary, North
condition, and often in the presence of a family history of de- Carolina).
ficiency. Similar criteria, but with a range of 30% to 60%,
were used for mild antithrombin deficiency. Persistently ele-
vated FVIII was defined as an activity level greater than 191 Results
U/dL on 2 separate venipuncture specimens obtained at least
6 weeks apart; similar criteria were used for persistently ele- Fifty-eight children met eligibility criteria and consented to
vated lipoprotein(a), with a concentration of greater than participate. Two enrolled subjects were excluded from anal-
30 mg/dL. ysis on the basis of lack of available thrombophilia testing re-
Elevated homocysteine was defined as a concentration of sults (ie, declined testing). Among the 56 children in the
greater than 11.9 mmol/L. Presence of the lupus anticoagulant analysis, median age was 10 years (range 6 months to 18
was defined by a dRVVT screen/confirm ratio of greater than years), with 50% females. Eighty-four percent of children
486 Calhoon et al

3. September 2010 ORIGINAL ARTICLES
of early thromboembolism in all subjects periprocedural
anticoagulant primary prevention for surgical procedures
expected to cause reduced postoperative mobility for a pe-
riod of at least 48 hours, among those children in whom
thrombophilia was disclosed (ie, irrespective of age). Addi-
tionally, restriction from use of estrogen-containing oral
contraceptive pills (OCP) was advised in 32% of females
on the basis of factor V Leiden testing alone, as compared
with 64% in whom additional laboratory evaluation was per-
formed (P = .02).
Table II reports outcomes in longitudinal follow-up.
Complete follow-up data were available in 36 children of
56 (64%) in this cohort of asymptomatic children. Over
Figure. Study population distribution by number of throm- a cumulative follow-up duration of 923 patient-months
bophilia traits among all children in whom comprehensive (median [range] of 22 months [8-44 months] per subject),
evaluation was performed (n = 32).
there were 6 episodes involving heightened thromboe-
mbolism risk (not including estrogen-containing OCP use
in 2 patients), and no thromboembolism events. No
were of Caucasian race, and 9% were African-American; eth- anticoagulant-associated, clinically significant bleeding
nicity was Hispanic for 5% of the study population. Seventy- occurred in the one case in which transient antithrombotic
nine percent of children underwent thrombophilia evalua- prophylaxis was used.
tion for a proximate family history of early VTE, and 21%
for arterial thromboembolism. Discussion
Table I gives the prevalence of thrombophilia traits, by
type, in the study population. Factor V Leiden, elevated In 2008, the U.S. Surgeon General issued a Call-to-Action
factor VIII activity, elevated lipoprotein(a) concentration, aimed at increasing awareness of VTE risks and reducing
and antiphospholipid antibody positivity were most the incidence, morbidity, and mortality rates of this disease.
common. At least one APA test result was positive on initial A Surgeon General’s Working Group on DVT/PE Preven-
thrombophilia testing in 11 (22%) of 50 subjects tested for tion, cosponsored by the Centers for Disease Control and
APA. Among 7 subjects who underwent repeat APA testing Prevention and the National Heart, Lung and Blood Insti-
after initial positive APA results, 3 had persistently positive tute, recognized that the morbidity of VTE impacts even chil-
APA results. Accounting for the 39 children who were APA- dren and identified the study of strategies for VTE prevention
negative on initial testing, the overall frequency of APA as one of several key pediatric goals.16
persistence was therefore 7% of 46 evaluable subjects. Given the low incidence of VTE in the general pediatric
Inherited deficiencies of protein C and antithrombin were population, widespread, unselected thrombophilia screen-
found in 7% and 2% of children, respectively. Other than ing would neither be ethical nor cost-effective. Therefore
for protein S deficiency (for which no cases were disclosed), we have advocated a risk-stratified approach to thrombo-
the frequencies of all thrombophilia traits were higher than philia evaluation in asymptomatic children, targeting indi-
expected. viduals with familial early TE. Even though prior
In the Figure, the population distribution is shown by prospective investigation in children by Tormene et al17
number of thrombophilia traits, among 32 symptom-free
children with a first- or second-degree family history of
early VTE in whom complete laboratory evaluation was
Table II. Outcomes in the study cohort in longitudinal
performed. For this analysis, dual IgG and IgM positivity to
follow-up
cardiolipin was considered as a single trait, and similarly
Number (%) study subjects in whom prospective 36 (64%)
for beta-2-glycoprotein-I. Two or more thrombophilia follow-up data were available
traits were found in 34% of these children. The most Median (range) duration of follow-up (patient-months) 22 (8-44)
commonly observed combinations of thrombophilia traits Total cohort follow-up duration (patient-months) 923
Number of major episodes at risk for thromboembolism* 6
were heterozygous factor V Leiden or factor II G20210A Number (%) of at-risk episodes in which transient 1 (17%)
with elevated levels of lipoprotein(a) or FVIII. When the anticoagulant prophylaxis was administered
data were restricted to children with a first-degree family Number of clinically significant bleeding episodes† 0
Number of TE episodes 0
history only, the proportion with multitrait thrombophilia
was not appreciably changed (data not shown). *Included major surgery, hospitalization for severe infection/inflammatory condition, immobility
for $48 hours; excludes 2 exposures to systemic estrogen (ie, use of oral contraceptive pills in
Thrombophilia testing led to a change in recommended 2 subjects).
management in 70% of subjects, consisting principally of †Defined as overt bleeding characterized by any one of the following: central nervous system or
retroperitoneum involved; hemoglobin concentration declined by $2 g/dL in a 24-hour period;
transient anticoagulant prophylaxis during periods of surgical intervention performed in an operating suite to achieve hemostasis; urgent medical
heightened risk for VTE. Given the proximate family history attention sought.
High Prevalence of Thrombophilic Traits in Children with Family History of Thromboembolism 487

4. THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 157, No. 3
found that VTE was rare even during periods of increased for such decision to be made). Nevertheless, prior retro-
clinical risk in symptom-free children with inherited throm- spective work in adults with native anticoagulant deficiency
bophilia who were identified on the basis of an affected suggests that a substantial decrease in VTE incidence can be
family member with a history of VTE, this observation achieved via recommendation of transient antithrombotic
must be considered in the context that (1) thrombophilia prophylaxis in asymptomatic individuals, particularly those
was generally single-trait; (2) familial VTE history was not under 40 years of age.18 Last, and most importantly, at-risk
necessarily before age 55 years; and (3) clinical prothrom- episodes for thromboembolism were uncommon during
botic risk situations encountered did not include any in- follow-up of the cohort, and antithrombotic prophylaxis
stances of central venous catheterization (a prevalent risk was used in only 1 of 6 instances for which it had been rec-
factor in pediatric VTE). ommended. Larger cooperative, controlled studies will be
These findings have important potential implications for necessary to definitively evaluate the safety and efficacy of
future evaluation and management of symptom-free chil- risk-stratified antithrombotic primary prevention for pedi-
dren with a proximate family history of early thromboembo- atric VTE. n
lism. The high likelihood of multitrait thrombophilia as
determined here in children with a family history of early Submitted for publication Aug 28, 2009; last revision received Jan 6, 2010;
thromboembolism, taken together with prior evidence that accepted Mar 26, 2010.
combined thrombophilia traits pose a further substantial in- Reprint requests: Neil Goldenberg, MD, PhD, Mountain States Regional
Hemophilia and Thrombosis Center, P.O. Box 6507, Aurora, CO 80045. E-mail:
crease in risk for VTE, identifies this pediatric subpopulation neil.goldenberg@ucdenver.edu.
as an important target for comprehensive thrombophilia
evaluation and potential antithrombotic prophylaxis in fu-
ture cooperative studies investigating the safety and efficacy References
of in-hospital anticoagulant prophylaxis in children. Addi-
tionally, with regard to precontraceptive thrombophilia test- 1. Rosendaal F, Koster T, Vandenbrouk J, Reitsma P. High risk of throm-
bosis in patients homozygous for factor V Leiden (activated protein C
ing, the frequency with which thrombophilia is disclosed in
resistance). Blood 1995;85:1504-8.
females with familial thromboembolism risk suggests that 2. Goldenberg NA, Manco-Johnson MJ. Protein C deficiency. Haemophilia
a strategy of focusing thrombophilia testing on this segment 2008;14:1214-21.
of the population would be more rational, and seemingly 3. Young G, Albisetti M, Bonduel M, Brandao L, Chan A, Friedrichs F,
more cost-effective, than the mass-screening of prospective Goldenberg NA, et al. Impact of inherited thrombophilia on venous
thromboembolism in children: a systematic review and meta-analysis
users of estrogen-containing OCP that has become increas-
of observational studies. Circulation 2008;118:1373-82.
ingly common in many countries. 4. Almawi WY, Tamim H, Kreidy R, Timson G, Rahal E, Nabulsi M, et al.
A few limitations of this work are noteworthy. First, the A case control study on the contribution of factor V-Leiden, pro-
prevalence of individual thrombophilia traits observed in thrombin G20210A, and MTHFR C677T mutations to the genetic sus-
this study may be an overestimate of that in symptom- ceptibility of deep venous thrombosis. Thromb Thrombolysis 2005;19:
189-96.
free children with familial thromboembolism in the general
5. Brouwer JL, Bijl M, Veeger NJ, Kluin-Nelemans HC, van der Meer J. The
population, because of the potential for referral bias of pa- contribution of inherited and acquired thrombophilic defects, alone or
tients most likely to have underlying thrombophilia. At the combined with antiphospholipid antibodies, to venous and arterial
same time, the potential for this bias is minimized by the thromboembolism in patients with systemic lupus erythematosus. Blood
fact that the indication (albeit consensus-based) for screen- 2004;104:143-8.
6. Tirado I, Mateo J, Soria JM, Oliver A, Borrell M, Coll I, et al. Contribu-
ing in all children was the presence of a first- or second-
tion of prothrombin 20210A allele and factor V Leiden mutation to
degree family history of early thromboembolism, rather thrombosis risk in thrombophilic families with other hemostatic defi-
than heterogeneous reasons for screening. Secondly, the ciencies. Haematologica 2001;86:1200-8.
study population consisted mostly of non-Hispanic Cauca- 7. Emmerich J, Rosendaal FR, Cattaneo M, Margaglione M, De Stefano V,
sians, and its findings therefore may not apply to other Cumming T, et al. Combined effect of factor V Leiden and prothrombin
20210A on the risk of venous thromboembolism–pooled analysis of 8
races and ethnicities. For example, the factor V Leiden
case-control studies including 2310 cases and 3204 controls. Study
and factor II G20210A polymorphisms are known to be Group for Pooled-Analysis in Venous Thromboembolism. Thromb
rare among African American and Asian individuals. Third, Haemost 2001;86:809-16.
because thrombophilia testing in routine clinical practice 8. Martinelli I. Pros and cons of thrombophilia testing: pros. J Thromb
among adults with VTE is nonuniform and typically quite Haemost 2003;1:410-1.
9. Machin SJ. Pros and cons of thrombophilia testing: cons. J Thromb Hae-
limited, thrombophilia findings in relatives affected by
most 2003;1:412-3.
VTE were not systematically collected as part of the present 10. Martinelli I. Rebuttal to: Pros and cons of thrombophilia testing—cons.
study. Therefore such data could not be compared with J Thromb Haemost 2003;1:1311-2.
thrombophilia findings in pediatric subjects evaluated in 11. Baglin T. Management of thrombophilia: who to screen? Pathophysiol
the study. Additionally, the frequency of antithrombotic Haemost Thromb 2003;33:401-4.
12. Thornburg CD, Dixon N, Paulyson-Nunez K, Oretel T. Thrombophilia
prophylaxis recommendations (as well as contraceptive pre-
screening in asymptomatic children. Thromb Res 2008;121:597-604.
cautions) reported here is based on clinical decisions made 13. British Committee for Standards in Haematology. Investigation and
in these subjects, in the absence of strong supportive clinical management of heritable thrombophilia. Brit J Haematol 2001;114:
evidence (and yet in the context of a practical clinical need 512-28.
488 Calhoon et al

5. September 2010 ORIGINAL ARTICLES
14. Middeldorp S, Meinardi JR, Koopman MM, van Rampus EC, Health and Human Services. http://www.surgeongeneral.gov/topics/
Hamulyak K, an Der Meer J, et al. A prospective study of asymptomatic deepvein/index.html.
carriers of the factor V Leiden mutation to determine the incidence of 17. Tormene D, Simioni P, Prandoni P, Franz F, Zerbinati P, Tognin G,
venous thromboembolism. Ann Intern Med 2001;135:322-7. Girolami A. The incidence of venous thromboembolism in thrombo-
15. Simioni P, Tormene D, Prandoni P, Zerbinati P, Gavasso S, Cefalo P, philic children: a prospective cohort study. Blood 2002;100:2403-5.
Girolami A. Incidence of venous thromboembolism in asymptomatic 18. De Stefano V, Leone G, Mastrangelo S, Tripodi A, Rodeghiero F,
family members who are carriers of factor V Leiden: a prospective cohort Castaman G, et al. Clinical manifestations and management of inherited
study. Blood 2002;99:1938-42. thrombophilia: analysis and follow-up after diagnosis of 238 patients
16. The Surgeon General’s Call To Action to Prevention Deep Vein with congenital deficiency of antithrombin III, protein C, protein S.
Thrombosis and Pulmonary Embolism. 2008. U.S. Department of Thromb Haemost 1994;72:352-8.
High Prevalence of Thrombophilic Traits in Children with Family History of Thromboembolism 489