Heritability of elevated factor VIII antigen levels in factor V Leiden families with thrombophilia
Professor F. R. Rosendaal, Clinical Epidemiology, C0-P, Leiden University Medical Centre, PO Box 9600, 2300 RC Leiden, The Netherlands . E-mail: email@example.com
Factor VIII activity (factor VIII:C) and factor VIII antigen (factor VIII:Ag) levels above150 IU/dl are associated with a five- to sixfold increased risk of venous thrombosis compared with levels < 100 IU/dl. These high levels are present in 25% of patients with a first episode of deep-vein thrombosis and in 11% of healthy controls. von Willebrand factor (VWF) and blood group are important determinants of the factor VIII level in plasma and therefore contribute to thrombotic risk, while factor VIII appears to be the final effector. Previously, we found familial clustering of factor VIII:C levels in women, which remained after adjustment for VWF and blood group. In the present study, we analysed the familial influence on factor VIII:Ag levels exceeding 150 IU/dl in 12 large families with thrombophilia in which high factor VIII:Ag levels contribute to thrombotic risk. As expected, blood group was a main determinant of the plasma factor VIII level: 58 relatives (32%) had factor VIII levels above 150 IU/dl and 50 (86%) of these had blood group non-O. After adjustment for blood group and age, we found an association between factor VIII:Ag levels in sister pairs (0·35, P = 0·003), brother pairs (0·35, P = 0·003), brother–sister pairs (0·35, P < 0·001) and in mother–son pairs (0·45, P = 0·02), but not in father–daughter or father–son pairs. The familial aggregation test was strongly positive for factor VIII:Ag levels (P < 0·001) and remained so after adjustment for the influence of blood group. We conclude that high factor VIII:Ag levels are a highly prevalent risk factor for venous thrombosis and contribute to risk in families with thrombophilia, and that these high levels are likely to be genetically determined by factors other than just blood group.
Elevated factor VIII activity (factor VIII:C) levels are associated with an increased risk of venous thrombosis ( Koster et al, 1995 ). Factor VIII:C levels higher than 150 IU/dl increase the thrombosis risk five- to sixfold compared with levels below 100 IU/dl ( Koster et al, 1995 ). von Willebrand factor (VWF) and blood group are well-known determinants of the factor VIII level in plasma and so contribute to thrombotic risk, whereas factor VIII itself appears to be the final effector in promoting thrombosis ( Koster et al, 1995 ). Elevated factor VIII:C levels are highly correlated with factor VIII antigen (factor VIII:Ag) levels ( Kamphuisen et al, 1997; O'Donnell et al, 1997 ). This suggests that the observed elevation of factor VIII:C levels reflects a true increase in factor VIII protein and is not the result of activation of the coagulation system during the blood collection procedure ( Kamphuisen et al, 1997; O'Donnell et al, 1997 ). Factor VIII:Ag levels above 150 IU/dl were, like factor VIII:C levels, also associated with a fivefold increased risk of venous thrombosis ( Kamphuisen et al, 1997 ). We have recently shown that elevated factor VIII levels in thrombosis patients are not the result of acute phase reactions because elevated factor VIII levels remained associated with a sixfold increased risk after adjustment for C-reactive protein (CRP), a sensitive marker for acute phase processes ( Kamphuisen et al, 1999 ). These observations lend further support to a causal relationship between high factor VIII levels and venous thrombosis.
High factor VIII levels are common; in our study, 25% of the patients with a first episode of venous thrombosis and 11% of the healthy controls had factor VIII levels above 150 IU/dl ( Koster et al, 1995 ). Considering the sixfold increased thrombosis risk and the high prevalence in the population, factor VIII levels exceeding 150 IU/dl contribute importantly to deep-vein thrombosis.
We previously reported that factor VIII:C levels show a familial clustering, which remains after correction for VWF and blood group ( Kamphuisen et al, 1998 ). That study was performed in female relatives of probands with documented haemophilia A who came to the Leiden haemophilia centre for carriership testing. We designed a study that allowed us to incorporate a large number of families that were seen over a period of more than 10 years for carrier testing. However, the design of this study also had several drawbacks. First, factor VIII:C levels were measured over a 10-year period, which might have led to extra variation in the plasma factor VIII level. Furthermore, only women were tested, which provides less information on familial clustering than when both women and men are tested. And, finally, in this group of women, high levels of factor VIII were not as common as among patients with thrombosis, and the association between levels and thrombosis was not studied nor was the heritability of high levels (above 150 IU/dl) itself.
In the present study, we have analysed the familial aggregation of factor VIII levels in men and women of 12 large families with thrombophilia who had a proband with both venous thrombosis and factor V Leiden. In these families, factor VIII:Ag levels above 150 IU/dl contribute to the thrombosis risk of factor V Leiden carriers ( Lensen et al, 1999 ). We tested whether clustering of factor VIII:Ag levels higher than 150 IU/dl occurred and the effect of blood group on these high levels.
SUBJECTS and METHODS
Family data The 12 probands originated from a larger panel of 28 patients who were referred to our centre for diagnostic work-up for venous thrombophilia, i.e. patients with a positive family history of venous thrombosis (in addition to the proband, at least two symptomatic relatives), and who did not have deficiencies of protein C, protein S or antithrombin. These 28 patients were screened for the presence of the factor V Leiden mutation that was detected in 12 patients. We invited the siblings, parents and children of these 12 probands as well as uncles and aunts of the affected parental side and, if these were carriers, their children (first cousins of the proband). Family members under 15 years of age were excluded for practical purposes. Of the 12 probands, 182 family members (93%) participated in the study while 12 did not, three because they lived abroad and nine for reasons unknown.
The probands and family members were not selected on the basis of factor VIII:Ag levels because these levels were not known at that time.
Laboratory assays Blood was collected from the antecubital vein in 0·106 m trisodium citrate. ABO phenotypes were deduced from the reactions of plasma isoagglutinins to A1, B, A1B or (as negative control) O test blood cell suspensions (3% v/v) in a standard spin-tube agglutination technique carried out at room temperature ( Walker, 1990). All the subjects were considered to be immune competent and free from malignancy or infection. The chance of an ABO misphenotyping because of the absence of the appropriate isoagglutinin, anomalous occurrence of anti-A or -B or acquisition of a B-antigen was considered to be unlikely. Factor VIII antigen was measured by a sandwich type enzyme-linked immunosorbent assay (ELISA) using two monoclonal antibodies directed against the light chain of factor VIII ( Kamphuisen et al, 1997 ). Pooled normal plasma, calibrated against the WHO standard (91/666) for factor VIII:Ag, was used as a reference.
Statistical analysis The distribution of factor VIII:Ag levels was skewed and was logarithmically transformed for all statistical analyses. The analysis was performed using linear regression, with age entered as a continuous variable (in years), and blood group dichotomized into two groups (0 for blood group O, 1 for non-O).
To investigate genetic effects on factor VIII:Ag levels, the residuals of the multiple regression models were used. Residuals are the differences between the observed level Yi for person i and the predicted value µi obtained from the multiple regression model. As a first test for familial effects, correlations between the residuals (obtained from the multiple regression model) of pairs of relatives were calculated. Familial aggregation of factor VIII:Ag levels was studied using a recently developed method ( Houwing-Duistermaat et al, 1995 ) that tests the null hypothesis of no correlation within randomly chosen pedigrees. The correlation of the genetic effects is tested by the weighted sum of correlations between pairs of relatives within a pedigree Q. The test for familial aggregation is positive when the calculated Q is larger than the expected Q-value under the null hypothesis of no aggregation (for more details, see Houwing-Duistermaat et al, 1995; Kamphuisen et al, 1998 ). The familial clustering of factor VIII levels exceeding 150 IU/dl was tested with the individual factor VIII level dichotomized into two groups (Yi = 0 for factor VIII:Ag levels < 150 IU/dl, Yi = 1 for factor VIII:Ag levels ≥ 150 IU/dl). In this way, the value of Q will be determined mainly by pairs of relatives who both have factor VIII levels higher than 150 IU/dl.
We studied 182 relatives of 12 probands with factor V Leiden and a positive family history of venous thrombosis. The mean size of the families was 19 members, with a range from 3 to 29 members. The mean age at the time of the study was 40 years (range 15–88 years); 92 (50%) were men and 90 were women. Of these 182 relatives, 91 were heterozygous for factor V Leiden, one relative was homozygous and 90 were non-carriers.
Mean factor VIII:Ag level was 137 IU/dl with a range between 54 and 339 IU/dl. Plasma factor VIII:Ag levels were not clearly different in factor V Leiden carriers [142 IU/dl; 95% confidence interval (CI) 132–151 U/dl] and non-carriers (132 IU/dl; 95% CI 121–142 U/dl). Factor VIII:Ag levels were higher in the 126 subjects (69%) with blood group non-O than in the 56 subjects (31%) with blood group O (150 vs. 107 IU/dl) (mean difference 43 IU/dl; 95% CI 29–57 IU/dl). Factor VIII levels were above 150 IU/dl in 58 individuals (32%). Of these individuals, 50 subjects had blood group non-O (86%) and eight (14%) had blood group O.
We tested the correlations of factor VIII:Ag levels between all first-degree relative pairs. After adjustment for blood group and age, the factor VIII levels in siblings showed a strong association (Table I). In sister pairs, the correlation of factor VIII levels was 0·35 (P = 0·003) and in brother pairs it was 0·35 (P = 0·004), whereas in brother–sister pairs this correlation was 0·31 (P < 0·001). In mother–son pairs, factor VIII:Ag levels were also highly correlated (r = 0·45, P = 0·02). The other first-degree relationships showed no clear correlations (Table I).
We assessed the familial clustering of factor VIII:Ag levels with the familial aggregation test. Table II shows that when factor VIII:Ag levels are tested as continuous variables the familial aggregation statistic Q was clearly higher than the expected Q under the null hypothesis of no correlation, also after adjustment for age and blood group. Unadjusted, Q was 308, whereas the expected value of no correlation was 175 (P < 0·001). Adjusted, Q was 301 with an expected Q of 172 (P < 0·001). Table III shows the familial clustering of factor VIII levels above 150 IU/dl as a dichotomized variable. The familial aggregation test was strongly positive again. In this test, Q was 1·65 with an expected value of no correlation of 0·97 (P < 0·001). Adjustment for blood group and age did not essentially change Q (Table III).
Our study showed that within thrombophilia families factor VIII:Ag levels are highly aggregated. The families we studied had thrombophilia, in which high factor VIII levels contribute to the risk of factor V Leiden carriers ( Lensen et al, 1999 ). Familial clustering of factor VIII:Ag levels higher than 150 IU/dl was clearly demonstrated by the familial aggregation test, and this remained after adjustment for the effects of blood group and age. This suggests a genetic influence on high factor VIII levels other than just blood group.
Familial aggregation was prominent, indicating that factor VIII levels were strongly influenced by familial factors. The support for familial aggregation of factor VIII levels was much stronger for the families studied here than that obtained with the female relatives of haemophilia A patients tested in our previous study ( Kamphuisen et al, 1998 ). This might have been the result of the larger size of the families which positively influenced the statistic Q of familial aggregation. The inclusion of men and women tested in the present study could also affect the outcome because the familial aggregation test uses all possible combinations within pedigrees. Further, factor VIII levels were measured over a much shorter period than in our previous study (< 4 weeks), which will reduce interassay variations in the factor VIII level.
These are families with thrombophilia who will probably have several risk factors for thrombosis. As high factor VIII levels are associated with venous thrombosis, we may have selected for high factor VIII levels in choosing these families. This is reflected by the higher number of factor VIII levels above 150 IU/dl in the thrombophilia families (32%) than in the thrombosis patients in the LETS study (25%) ( Koster et al, 1995 ). The calculated Q of familial aggregation for high factor VIII levels is mainly the result of concordant pairs with high factor VIII levels. The probability that siblings have both elevated factor VIII levels is more likely in families that have a high prevalence of elevated factor VIII levels than in families in which low factor VIII levels are more common. This means that a selection for high factor VIII levels in families will underestimate the heritability of high factor VIII levels.
In this study, correlations of factor VIII:Ag levels between first-degree relatives only partially confirm the theory that variation of plasma factor VIII levels is under control of X-linked alleles ( Filippi et al, 1984 ). After adjustment for the effect of blood group and age, factor VIII levels correlated in siblings and in mother–son pairs, but not in mother–daughter, father–daughter or father–son pairs. Especially the finding that factor VIII levels were not correlated in father–daughter pairs, who share the same X-chromosome, does not support an X-linked determinant of factor VIII levels. We have to consider the possibility that the correlation of factor VIII between fathers and daughters is lowered by the maternal X-chromosome of the daughter. A daughter who has a father with high factor VIII and a mother with low factor VIII may have an intermediate factor VIII level. This will negatively influence the correlation between father and daughter. Very recently, Mansvelt et al (1998) investigated the promoter and 3′ terminus of the factor VIII gene for variations associated with high factor VIII:C levels, but found none. It remains possible that a part of the variation of factor VIII is determined by genetic factors located on the X-chromosome, outside the factor VIII gene. We can also not rule out the possibility that environmental factors and clustering within families also contribute to the familial aggregation of high factor VIII.
The mean difference in factor VIII:Ag level between blood group O and non-O was 43 IU/dl. Among subjects with factor VIII levels above 150 IU/dl, 86% had blood group non-O, indicating that a substantial part of the elevation in factor VIII is attributable to blood group. Most of the effect of blood group on the factor VIII level is mediated through VWF ( Koster et al, 1995 ; Kamphuisen et al, 1998 ), but the exact mechanism of how ABO blood group influences VWF is unclear. Blood group A, B and H(O) oligosaccharide structures have been identified on VWF ( Sodetz et al, 1979; Matsui et al, 1992 ). As modification of carbohydrates has been shown to influence the half-life of VWF in the circulation in animal models ( Stoddart et al, 1996; Sodetz et al, 1977 ), it is possible that ABO blood group determinants affect the clearance of VWF in plasma. VWF is an important determinant of the factor VIII level in plasma, which is explained by VWF being the carrier protein for factor VIII ( Tuddenham et al, 1982; Brinkhous et al, 1985 ). Whether the different types of ABO blood group also influence the factor VIII survival in plasma remains to be determined.
We conclude that there is strong support that factor VIII levels exceeding 150 IU/dl aggregate in families, also after adjustment for blood group and age. This is further evidence that the high plasma factor VIII levels that previously were found to be associated with a thrombotic risk in case–control and family studies are determined by genetic factors other than just blood group.
We would like to thank H. de Ronde for determination of the factor VIII antigen levels. This study was supported by a grant (no. 950-10-629) from The Netherlands Organization for Scientific Research (NWO) and by a grant (no. 95.026) from The Netherlands Heart Foundation (NHS).