Journal of Thrombosis and Haemostasis, 5: 19301935

ORIGINAL ARTICLE

A genetic basis for the interrelation of coagulation factors

C. Y. VOSSEN,* P. W. CALLAS, S. J. HASSTEDT, G. L. LONG, F. R. ROSENDAAL*,** and E . G . B O V I L L
*Department of Clinical Epidemiology, Leiden University Medical Centre, Leiden, The Netherlands; Department of Pathology, University of Vermont, Burlington, VT; Department of Biostatistics, University of Vermont, Burlington, VT; Department of Human Genetics, University of Utah, Salt Lake City, UT; Department of Biochemistry, University of Vermont, Burlington, VT, USA; and **Haematology, Leiden University Medical Centre, Leiden, The Netherlands

To cite this article: Vossen CY, Callas PW, Hasstedt SJ, Long GL, Rosendaal FR, Bovill EG. A genetic basis for the interrelation of coagulation factors. J Thromb Haemost 2007; 5: 19305.

Summary. Background: Evidence found in the literature for a strong correlation between coagulation factors suggests that single genes might inuence the plasma concentrations of multiple coagulation factors (i.e. pleiotropically acting genes). Objective: To determine whether there is a genetic basis for the correlation among coagulation factors by assessing the heritability of interrelated coagulation factors. Patients/ methods: We performed principal components analysis, and subsequently variance components analysis, to estimate the heritability of principal components of coagulation factors in family members of a large French-Canadian kindred. Results: Four clusters were identied by principal components analysis in 200 family members who did not carry the protein C 3363C mutation. Cluster 1 consisted of prothrombin, factor VII (FVII), FIX, FX and protein S; cluster 2 consisted of FV, FIX, protein C and tissue factor pathway inhibitor; cluster 3 consisted of FVIII and von Willebrand factor; and cluster 4 consisted of antithrombin, protein C and FVII. The heritability of the principal components estimated by variance components analysis was, respectively, 37%, 100%, 37%, and 37%. Conclusion: Our ndings support the hypothesis that genes can inuence plasma levels of interrelated coagulation factors. Keywords: coagulation, heritability, principal components analysis. Introduction Thrombin plays a central role in the blood coagulation process and is formed from prothrombin after a series of
Correspondence: Prof. Edwin G. Bovill, Department of Pathology, University of Vermont, Given Building #E208, 89 Beaumont Avenue, Burlington VT 05405, USA. Tel.: +1 802 656 0359; fax: +1 802 656 8892; e-mail: edwin.bovill@ uvm.edu Received 7 August 2006, accepted 24 June 2007

enzymatic reactions involving the procoagulant coagulation factors V (FV), VII, VIII, IX and X [1]. Regulation of the formation of thrombin occurs through tissue factor pathway inhibitor (TFPI), antithrombin, the protein C pathway, and the protein Zprotein Z inhibitor (ZPI) system. Defects, either genetic or environmental, in the hemostatic balance maintained by these procoagulant and anticoagulant systems can produce venous thrombotic disease. In several studies, genes have been shown to account for a large proportion of the variation in plasma levels of hemostatic factors [25]. A high phenotypic correlation has been found between various coagulation factors, and this seems to have a genetic basis as well [6]. This indicates that single genes might be able to pleiotropically inuence multiple factors involved in coagulation, which is supported by ndings from the Leiden Thrombophilia Study (LETS), which revealed clusters of interrelated procoagulant and anticoagulant factors [7]. The aim of the present study was to estimate the heritability of interrelated hemostatic factors in a large French-Canadian kindred. We included in the analysis the following coagulation factors: prothrombin, FV, FVII, FVIII, FIX, FX, brinogen, von Willebrand factor (VWF), protein C, protein S, antithrombin and TFPI. Methods
Study population

Blood samples were collected from a large kindred of French-Canadian origin with type I protein C deciency. The ascertainment and evaluation of the family members has been described previously [8]. All participating subjects or their legal guardians gave informed consent. The Human Experimentation Committees of the University of Vermont College of Medicine, Burlington (VT) and the Beth Israel Hospital, Boston (MA) approved the study. The pedigree of this family spans seven generations, with data available on the most recent ve. Although we have obtained blood samples from more than 400 family members over the past
2007 International Society on Thrombosis and Haemostasis

Heritability of interrelated coagulation factors 1931

20 years, in the current analysis we only included those with all analytes measured in samples from the same blood draw to ensure correlation between coagulation factor levels and pedigree data validated after genotyping (n = 299). In addition, we excluded family members who were on oral anticoagulant treatment at the time of the blood draw (n = 33), those who had levels of VWF under 25 U dL)1 (n = 8), those for whom we did not have information regarding their age (n = 3), and those who were protein C decient, to remove the effect of the protein C mutation on the correlations between factors (n = 49).
Blood collection and processing

Blood was collected into siliconized glass Vacutainer tubes containing two different anticoagulants: 3.8% buffered citrate solution (Becton Dickinson, Franklin, NJ, USA), and SCAT-1 (25 lM PPACK, 200 KIU mL)1 aprotinin, 4.5 mM EDTA; Haematologic Technologies, Essex, VT, USA). Platelet-poor plasma was produced within 1 h by centrifugation at 3000 g for 10 min at room temperature. Samples were stored at )70 C and thawed at 37 C just before assay performance.
Assay methodology

Assays were performed in the investigators laboratories. Antigen plasma concentrations of FV [in-house assay; interassay coefcient of variation (CV) 7%] [9], FVII (CV 7%), FVIII (CV 8%), FIX (CV 10%), FX (CV 7%), VWF [10] (CV 3%) and total TFPI (CV 6%) were measured by enzyme-linked immunosorbent assay, and antigen levels of prothrombin [11] (CV 6%), antithrombin [12,13] (CV 5%), and free protein S [14] (CV 10%) were determined by radioimmunoassay. The Clauss method was used for determining the plasma concentration of brinogen [15,16] (CV 2%), using the ST4 instrument (Diagnostica Stago, Parsipanny, NJ, USA). We measured protein C activity levels (CV 6%) by performing a clot-based functional assay [17]. Kits for measuring FVII, FVIII, FIX, FX, protein C, VWF and TFPI were provided by Diagnostica Stago, and the assays were performed following the manufacturers instructions.
Statistical analysis

We used the principal components method with orthogonal Varimax rotation to reveal clusters of interrelated coagulation factors [18]. The sum of all squared (unrotated) factor loadings for each principal component (i.e. the eigenvalue) indicates the amount of variance attributable to this component. We only considered principal components with an eigenvalue greater than 1, which indicates that the components account for more of the total variance than the original variables. The inclusion of coagulation factors in a cluster depends on their factor loading, i.e. the correlation between a variable and a principal component. We used a factor loading of > 0.40 as a marginal value to include coagulation factors, which indicates that at
2007 International Society on Thrombosis and Haemostasis

least 16% of the variance of that variable can be explained by a particular principal component. To adjust for age and sex, we used the regression coefcients (betas) for each variable derived via linear regression. Adjusted values were calculated by subtracting the effect of age and sex from the original value of a variable for each family member. The adjusted values did not lead to markedly different results from an analysis without these adjustments. Using the variance component method in Sequential Oligogenic Linkage Analysis Routines [19], we estimated the proportion of the phenotypic variance attributed to polygenes (heritability) and the proportion of the variance attributed to environmental factors shared within a household (common household effect), using the specic value of a principal component for each subject (subjects factor scores). The distribution of each score was assumed to be multivariate normal with a variancecovariance matrix, following the formula: covariance (one person to another person) = h22 K + c2H + e2I, with K from the kinship matrix, H from the household matrix, and I from the identity matrix. The parameters for heritability (h2) and household effect (c2), as well as the effects of covariates, were estimated simultaneously using maximum likelihood analysis. In addition, to determine whether the principal components, which do not take family structure into account, can be corroborated by correlation data taking family structure into account, we determined the phenotypic correlation between pairs of factors using Pedigree Analysis Package 6 (PAP), including the original (nonadjusted) levels and adjusting for age and sex as part of the correlation analysis [20]. To optimize the principal components analysis and heritability analysis, which assume a normal distribution for the included variables, levels of coagulation factors were transformed to obtain normal distributions of the single levels and thereby of the factor scores. An additional six family members were removed to normalize the distribution of FVII and antithrombin levels. The transformations did not lead to markedly different results from an analysis without these transformations. To determine whether the variables that showed a correlation above 0.4 with a principal component were responsible for the heritability of that principal component, we put each variable from a cluster in the analysis one at a time as a covariate when estimating the heritability of a cluster. A decrease in heritability after putting in a coagulation factor as a covariate indicates that (a part of) the heritability can be explained by that specic coagulation factor. Results In total, 200 family members (91 men, 109 women) from 132 households (range of individuals per household 14) were included in the analysis. Their mean age at the blood draw was 31 years (range: 174); 24 subjects (12%; one not tested) carried the prothrombin G20210A variant, and three family members had a validated history of venous thrombosis. None of the family members carried the FV Leiden

1932 C. Y. Vossen et al
Table 1 Mean plasma concentration, standard deviation (SD) and range of all analytes in 200 family members without protein C deciency Analytes (units) n = 200 Prothrombin (U dL ) Factor V (FV) (U dL)1) FVII (% of normal) FVIII antigen (U dL)1) FIX antigen (U dL)1) FX (% of normal) Fibrinogen (g L)1) VWF (U dL)1) Antithrombin (% of normal) Protein C (% of normal) Protein S free (% of normal) TFPI total (ng mL)1)
)1

Mean 105.1 104.6 100.2 110.8 104.2 107.8 2.7 105.0 97.0 111.3 95.8 78.6

SD 21.6 33.1 22.4 37.6 32.8 24.1 0.6 48.6 15.0 32.4 24.6 19.7

Range 39164 39293 40168 34246 53250 51164 1.44.8 27300 49142 53204 37192 41161

Table 2 Factor loadings, heritability, and household eect per principal component for procoagulant and anticoagulant factors after adjusting for age and sex Components 1 Prothrombin 0.665* Factor V (FV) 0.041 FVII 0.589* FVIII 0.027 FIX 0.419* FX 0.777* Fibrinogen 0.349 VWF 0.042 Antithrombin )0.087 Protein C 0.307 Protein S 0.682* TFPI total 0.332 Variance 20.8 explained by component (%) Heritability 37.4 (077) [% (95% CI)] Household 5.5 (031) [% (95% CI)] 2 0.039 0.746* )0.209 0.090 )0.526* 0.167 0.391 0.003 )0.014 )0.648* )0.008 0.722* 18.3 3 0.100 0.044 0.028 0.830* )0.124 )0.144 0.274 0.830* )0.031 )0.084 0.116 )0.073 11.8 4 )0.028 )0.065 0.402* )0.075 )0.020 0.030 0.265 0.037 0.916* 0.456* )0.054 0.047 9.1

VWF, von Willebrand factor; TFPI, tissue factor pathway inhibitor.

mutation. The mean plasma concentrations are shown in Table 1.

Principal components

99.6 (75100) 36.8 (076) 37.0 (965) ** 8.6 (039) **

The principal components analysis identied four components with an eigenvalue above 1.0, explaining 60.0% of the total phenotypic variance in factor levels. When a factor loading of 0.40 was used as a cut-off point to identify clusters of interrelated coagulation factors, the rst cluster consisted of the factors prothrombin, FVII, FIX, FX, protein S, the second cluster consisted of FV, FIX, protein C and TFPI, the third cluster consisted of VWF and FVIII, and the fourth cluster consisted of antithrombin, protein C and FVII (Table 2). Table 3 shows the phenotypic correlation of pairs of coagulation factors as derived from PAP, taking the family structure into account. The correlations in Table 3 corroborate the identied principal components. The vitamin K-dependent coagulation factors almost exclusively show high correlation with each other in Table 3, explaining the constellation of cluster 1 of the principal components analysis (Table 2). Protein C is an exception to this rule, and shows, in addition, high correlation with FV and antithrombin, which is reected by clusters 2 and 4.

VWF, von Willebrand factor; TFPI, tissue factor pathway inhibitor; CI, condence interval. *Factor loading > 0.40 after orthogonal varimax rotation. **Removed from the model. Dened as the sum of squares of the unrotated factor loadings. Calculated as heritability 1.96 SE.

The appearance of FIX in cluster 2 and of FVII in cluster 4 seem to be related to their high correlation with protein C and not to the other factors in these clusters. FV shows a high correlation with protein C and TPFI (Table 3), underlining cluster 2 (Table 2). FVIII and VWF show a high correlation in Table 3, but not with other factors, which explains why they form a separate cluster (Table 2). The correlation between TFPI and FX, FV and brinogen, and the correlation between brinogen and protein S, are not reected in the clusters derived from the principal components analysis (Table 2). However, these

Table 3 Phenotypic correlation (%) between pairs of coagulation factors taking family structure into account Factor V (FV) PT FV FVII FVIII FIX FX Fib VWF AT PC PS 8.6 FVII 23.0 )8.6 FVIII 6.4 15.6 8.8 FIX 13.1 )6.8 32.0 )5.9 FX 41.1 )3.3 46.9 )0.8 26.2 Fib 17.6 22.9 17.1 19.6 10.1 16.8 VWF 14.0 5.9 10.3 43.7 )9.0 2.9 15.3 AT 8.1 )2.2 19.0 )1.7 )0.9 5.7 6.5 )0.3 PC 26.6 )27.3 43.7 )12.3 39.5 25.2 0.3 )4.5 39.7 PS 35.7 )3.8 28.6 12.8 17.7 39.4 26.8 9.2 )0.27 18.0 TFPI 19.6 26.0 4.3 1.9 1.6 25.5 22.4 8.5 7.4 )17.8 9.3

PT, prothrombin; Fib, brinogen; AT, antithrombin; PC, protein C; PS, protein S; VWF, von Willebrand factor; TFPI, tissue factor pathway inhibitor. 2007 International Society on Thrombosis and Haemostasis

Heritability of interrelated coagulation factors 1933

A
0.75 0.50
Fib FV TFPI

Component 2

0.25
FX

0.00 0.25 0.50

AT

FVIII vWF

PT PS FVII

FIX PC

0.75 0.2 0.0 0.2 0.4 0.6 0.8

correlations might be overshadowed by correlations of these factors with other principal components. Figure 1AC presents the original variables and their relation to the components after Varimax rotation. Figure 1A shows the clustering of the vitamin K-dependent proteins and clustering of FV and TFPI. The latter factors indeed show the highest correlation with component 2 (Table 2) and the highest positive correlation in Table 3, whereas the correlation with FIX and protein C of these factors is either negative or negligible (Table 3). The clustering of FVIII and VWF on component 3 is apparent in Fig. 1B. Antithrombin is separate from all other variables (Fig. 1C), which can also be concluded from the fact that antithrombin shows by far the highest correlation with component 4 (Table 2). Fibrinogen shows a similar correlation with each component (around 0.3) (Table 3).
Heritability and household effect estimates of principal components

Component 1

B
1.0 0.8 0.6 0.4
Fib FVIII vWF

0.2 0.0
FV AT PC TFPI FIX FX PT PS FVII

0.2 0.2 0.0 0.2 0.4 0.6

0.8

Component 1

C
1.0
AT

We estimated an intermediate heritability of 37% for, respectively, the rst, third and fourth principal component, and a high heritability of nearly 100% for the second component (Table 2). No household effect was found for the principal components (Table 2). We put in each variable, one at a time, as a covariate in the heritability analysis to nd which variables decreased the heritability and explained a portion of the heritability. For components 1 and 4, none of the variables decreased the heritability below 37%. For component 2, the heritability decreased but remained considerably high when FV (72%), FIX (72%), protein C (90%) or TFPI (78%) were added (separately) to the model. For component 3, heritability with ABO blood group, VWF or FVIII included in the analysis as a covariate, was, respectively, 36%, 0% and 38%. So, of all factors showing a correlation above 0.4 with a principal component, only VWF seems to explain the heritability of the principal component. Discussion

Component 3 Component 4

0.8 0.6 0.4 0.2 0.0 0.2 0.2 0.0 0.2 0.4 0.6 0.8
vWF FV FVIII TFPI FIX PT PS FX PC FVII Fib

Component 1

Fig. 1. The original variables and their relation to the components after Varimax rotation in the principal components analysis (AC). PT, prothrombin, F, factor, Fib, brinogen, AT, antithrombin, PC, protein C, PS, protein S.

In 200 members of a large French-Canadian kindred, we estimated the heritability of principal components of coagulation factors to nd evidence for the hypothesis that plasma concentrations of interrelated coagulation factors may be regulated by genes (pleiotropically acting genes). Principal components analysis, with adjustment of the levels for age and sex, revealed four principal components in this family, of which one showed high heritability and three showed intermediate heritability. From these components, we could derive four clusters when taking a factor loading of > 0.4 as the cut-off point: the rst cluster contained vitamin K-dependent factors only, the second cluster included FV, FIX, protein C and TFPI, the third cluster included FVIII and its carrier VWF, and the fourth cluster included antithrombin, protein C and FVII. Clustering of vitamin K-dependent proteins has also been observed in other studies [7,21]. A paper by Souto et al. [6]

2007 International Society on Thrombosis and Haemostasis

1934 C. Y. Vossen et al

showed substantial phenotypic correlations between different pairs of vitamin K-dependent proteins, and showed that most of the pairs exhibited signicant genetic correlation. In the literature, several genes have been reported to inuence the plasma concentrations of vitamin K-dependent coagulation factors. For example mutations in the c-glutamyl carboxylase gene at chromosome 2p12 and the vitamin K epoxide reductase (VKORC1) gene on chromosome 16p11.2 can cause a combined deciency of vitamin K-dependent coagulation factors [2224]. In addition, calumenin on chromosome 7q32 can inhibit the vitamin K-dependent c-carboxylation system [25], and the transcription factor hepatocyte factor 4 (HNF-4) on chromosome 20q12q13.1 is essential for the expression of prothrombin, FVII, FIX and FX [2629]. A possible explanation for the clustering of FV and FIX could be a common posttranslational modication; tyrosine sulfation [30]. As FVIII also undergoes tyrosine sulfation [30], this could explain the inclusion of FVIII in this cluster in the LETS. No data are available in the literature on tyrosine sulfation of protein C and TFPI. The clustering of VWF and FVIII levels in our study seems logical, as VWF regulates FVIII levels by acting as a carrier protein. Adding VWF levels as a covariate to the heritability analysis of the principal component, to which FVIII and VWF belong, reduced the heritability substantially, suggesting a pivotal role for VWF. A recent linkage analysis in families with type 1 von Willebrand disease (VWD) showed that 48% of the variation in VWF antigen levels could be attributed to the VWF locus [31], although the heritability attributed to the VWF locus might be different for individuals with normal VWF levels as compared to individuals with type 1 VWD. We have no explanation for the clustering of antithrombin and protein C (FVII only shows a strong correlation with protein C) other than that both are inhibitors of the coagulation system. Although the LETS study also found clustering of vitamin K-dependent proteins, of FV and FIX, and of protein C and antithrombin, the clusters observed in this study differ significantly from those observed in the LETS. Differences between the LETS and our analysis could be explained by the different sets of coagulation factors used in the analyses. For example, the strong correlation between FVIII and VWF in our study (VWF was not included in the LETS analysis) could have led to the exclusion of FVIII from another cluster (FVIII was included in a cluster with FV, FIX and brinogen in the LETS). In addition, there are differences in study design: the LETS is a population-based study, and our study is a familybased study. Principal components analysis does not take family structure into account, but we found that the clusters identied by this method could be explained by the phenotypic correlations between the pairs of coagulation factors, which we calculated taking family structure into account. The population-based study will have more genetic and environmental variation among the study subjects than the members from the large protein C family; family members will share genes, but also households and thereby environmental factors. Genetic variations inuencing plasma levels in family members could be

family-specic and of less inuence in the general population. In addition, family members are French-Canadian; the FrenchCanadian population has been more isolated than the Dutch population. The high and intermediate heritabilities found for the principal components seem to provide evidence for a genetic basis for the interrelationship between the variables within these principal components. However, the 95% CI of the intermediate heritabilities were wide. In addition, it might be possible that the correlations within principal components (at least in part) are due to the same environmental factors in an individual affecting all the variables in the cluster. However, when these environmental factors are shared within families, they should contribute to the shared environmental effect (household effect), which was non-signicant in our analyses, and not to heritability. Nevertheless, as the principal components method does not incorporate biological effects, one has to keep in mind that other non-genetic factors could explain the high heritability. In conclusion, our ndings support the hypothesis that genes can inuence plasma levels of interrelated coagulation factors. Knowledge of normal genetic variation in single genes leading to changes in the levels of interrelated coagulation factors can ultimately increase our understanding of the pathways involved in venous thrombosis. Acknowledgements We would like to thank J. E. Valliere for performing the assays needed for this study. Disclosure of Conict of Interests This study was supported by NHLBI grant PHS HL46703 and a Transatlantic Network for Excellence in Cardiovascular Research grant from Fondation Leducq, Paris, France. References
1 Mann KG, Butenas S, Brummel K. The dynamics of thrombin formation. Arterioscler Thromb Vasc Biol 2003; 23: 1725. 2 Souto JC, Almasy L, Borrell M, Gar M, Mart nez E, Mateo J, Stone WH, Blangero J, Fontcuberta J. Genetic determinants of hemostasis phenotypes in Spanish families. Circulation 2000; 101: 154651. 3 de Lange M, Snieder H, Arie ns RA, Spector TD, Grant PJ. The genetics of haemostasis: a twin study. Lancet 2001; 357: 1015. 4 Arie ns RA, de Lange M, Snieder H, Boothby M, Spector TD, Grant PJ. Activation markers of coagulation and brinolysis in twins: heritability of the prethrombotic state. Lancet 2002; 359: 66771. 5 Vossen CY, Hasstedt SJ, Rosendaal FR, Callas PW, Bauer KA, Broze GJ, Hoogendoorn H, Long GL, Scott BT, Bovill EG. Heritability of plasma concentrations of clotting factors and measures of a prethrombotic state in a protein C decient family. J Thromb Haemost 2004; 2: 16. 6 Souto JC, Almasy L, Blangero J, Stone W, Borrell M, Urrutia T, Mateo J, Fontcuberta J. Genetic regulation of plasma levels of vitamin K-dependent proteins involved in hematostatis: results from the GAIT Project. Genetic Analysis of Idiopathic Thrombophilia. Thromb Haemost 2001; 85: 8892.

2007 International Society on Thrombosis and Haemostasis

Heritability of interrelated coagulation factors 1935

7 van Hylckama Vlieg A, Callas PW, Cushman M., Bertina RM, Rosendaal FR. Interrelation of coagulation factors and D-dimer levels in healthy individuals. J Thromb Haemost 2003; 1: 51622. 8 Bovill EG, Bauer KA, Dickerman JD, Callas P, West B. The clinical spectrum of heterozygous protein C deciency in a large New England kindred. Blood 1989; 73: 7127. 9 Folsom AR, Cushman M, Tsai MY, Aleksic N, Heckbert SR, Boland LL, Tsai AW, Yanez ND, Rosamond WD. A prospective study of venous thromboembolism in relation to factor V Leiden and related factors. Blood 2002; 99: 27205. 10 Cushman M, Legault C, Barrett-Connor E, Stefanick ML, Kessler C, Judd HL, Sakkinen PA, Tracy RP. Eect of postmenopausal hormones on inammation-sensitive proteins: the Postmenopausal Estrogen/Progestin Interventions (PEPI) Study. Circulation 1999; 100: 71722. 11 Church WR, Bhushan FH, Mann KG, Bovill EG. Discrimination of normal and abnormal prothrombin and protein C in plasma using a calcium ion-inhibited monoclonal antibody to a common epitope on several vitamin K-dependent proteins. Blood 1989; 74: 241825. 12 Bauer KA, Ashenhurst JB, Chediak J, Rosenberg RD. Antithrombin Chicago: a functionally abnormal molecule with increased heparin anity causing familial thrombophilia. Blood 1983; 62: 124250. 13 Bauer KA, Broekmans AW, Bertina RM, Conard J, Horellou MH, Samama MM, Rosenberg RD. Hemostatic enzyme generation in the blood of patients with hereditary protein C deciency. Blood 1988; 71: 141826. 14 Bovill EG, Landesman MM, Busch SA, Fregeau GR, Mann KG, Tracy RP. Studies on the measurement of protein S in plasma. Clin Chem 1991; 37: 170814. 15 Clauss A. Gerrinnungs-physiologische Schnell-methode zur Bestimmung des Fibrinogens. Acta Haematol 1957; 17: 23746. 16 Geken DF, Keating FG, Kennedy MH, Cornell ES, Bovill EG, Tracy RP. The measurement of brinogen in population-based research. Studies on instrumentation and methodology. Arch Pathol Lab Med 1994; 118: 11069. 17 Sala N, Owen WG, Collen D. A functional assay of protein C in human plasma. Blood 1984; 63: 6715. 18 Kline P. An Easy Guide to Factor Analysis. New York: Routledge, 1994. 19 Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 1998; 62: 1198211. 20 Hasstedt SJ. jPAP: document-driven software for genetic analysis. Genet Epidemiol 2005; 29: 255. 21 Sakkinen PA, Wahl P, Cushman M, Lewis MR, Tracy RP. Clustering of procoagulation, inammation, and brinolysis variables with metabolic factors in insulin resistance syndrome. Am J Epidemiol 2000; 152: 897907. nchez-Vega B, Wu SM, Lanir N, Staord DW, Solera J. Brenner B, Sa A missense mutation in gamma-glutamyl carboxylase gene causes combined deciency of all vitamin K-dependent blood coagulation factors. Blood 1998; 92: 45549. Spronk HM, Farah RA, Buchanan GR, Vermeer C, Soute BA. Novel mutation in the gamma-glutamyl carboxylase gene resulting in congenital combined deciency of all vitamin K-dependent blood coagulation factors. Blood 2000; 96: 36502. Rost S, Fregin A, Ivaskevicius V, Conzelmann E, Ho rtnagel K, Pelz HJ, Lappegard K, Seifried E, Scharrer I, Tuddenham EGD, Mu ller CR, Strom TM, Oldenburg J. Mutations in VKORC1 cause warfarin resistance and multiple coagulation factor deciency type 2. Nature 2004; 427: 53741. Wajih N, Sane DC, Hutson SM, Wallin R. The inhibitory eect of calumenin on the vitamin K-dependent gamma-carboxylation system. Characterization of the system in normal and warfarin-resistant rats. J Biol Chem 2004; 279: 2527683. Ceelie H, Spaargaren-Van Riel CC, de Jong M, Bertina RM, Vos HL. Functional characterization of transcription factor binding sites for HNF1-alpha, HNF3-beta (FOXA2), HNF4-alpha, Sp1 and Sp3 in the human prothrombin gene enhancer. J Thromb Haemost 2003; 1: 1688 98. Erdmann D, Heim J. Orphan nuclear receptor HNF-4 binds to the human coagulation factor VII promoter. J Biol Chem 1995; 270: 2298896. Reijnen MJ, Peerlinck K, Maasdam D, Bertina RM, Reitsma PH. Hemophilia B Leyden: substitution of thymine for guanine at position21 results in a disruption of a hepatocyte nuclear factor 4 binding site in the factor IX promoter. Blood 1993; 82: 1518. Hung HL, High KA. Liver-enriched transcription factor HNF-4 and ubiquitous factor NF-Y are critical for expression of blood coagulation factor X. J Biol Chem 1996; 271: 232331. Kaufman RJ. Post-translational modications required for coagulation factor secretion and function. Thromb Haemost 1998; 79: 106879. Eikenboom J, van Marion V, Putter H, Goodeve A, Rodeghiero F, Castaman G, Federici AB, Batlle J, Meyer D, Mazurier C, Goudemand J, Schneppenheim R, Budde U, Ingerslev J, Vorlova Z, Habart D, Holmberg L, Lethagen S, Pasi J, Hill F, et al. Linkage analysis in families diagnosed with type 1 von Willebrand disease in the European study, molecular and clinical markers for the diagnosis and management of type 1 VWD. J Thromb Haemost 2006; 4: 77482.

22

23

24

25

26

27

28

29

30

31

2007 International Society on Thrombosis and Haemostasis