In a recently published paper in Blood, Hoppe et al1 showed in 213 patients and 213 controls that carriers of the Marburg I polymorphism of the factor VII-activating protease (FSAP)2 had an increased risk of venous thrombosis (odds ratio [OR], 3.5; 95% confidence interval [CI], 1.2-10.0).
FSAP is a serine protease that has 2 functions in hemostasis. It activates factor VII, thereby promoting coagulation,3 but it can also activate single-chain plasminogen activators, thus promoting fibrinolysis.4 Recently, a single nucleotide polymorphism (1601G>A) was discovered in the gene coding for FSAP, which results in the substitution of glycine 511 by glutamic acid (FSAP Marburg I) and which is present in 2% to 9% of the white population.1,5,6 The Marburg I variant has an impaired prourokinase activating potency, whereas it can still activate factor VII normally.7
If true, the finding of Hoppe et al1 would support the hypothesis that reduced fibrinolysis contributes to the risk of venous thrombosis. However, the authors indicated that their control group might be biased due to the exclusive inclusion of healthy blood donors and that larger studies were needed to validate their results. Therefore, we determined the Marburg I polymorphism in 471 consecutive patients with a first episode of deep venous thrombosis (DVT) and 471 sex- and age-matched healthy controls of the Leiden Thrombophilia Study (LETS), a case-control study on the causes of venous thrombosis.8
We determined the Marburg I 1601G>A polymorphism with a 5′ nuclease/TaqMan assay (Assay by design; Applied Biosystems, Foster City, CA).9 Nucleotide sequences of primers and probes are available on request. An odds ratio with 95% CI was calculated as a measure of the relative risk of thrombosis for carriers of the Marburg I allele (homozygous or heterozygous) compared with homozygous wild-type allele carriers. Factor VII activity was measured previously using Thromborel S reagent (Behringwerke, Warburg, Germany) and factor VII-deficient plasma.10 In Table 1, the risk of venous thrombosis is shown for the Marburg I polymorphism. Marburg I was found in 30 controls (allele frequency 0.034) and 27 cases (allele frequency 0.030). No association between Marburg I and venous thrombosis was found. Similar results were obtained when the analysis was stratified by sex or age (individuals < 45 years of age versus individuals ≥ 45 years of age). Factor VII activity was not influenced by the presence of Marburg I.
Our results indicate, in contrast with the finding of Hoppe et al,1 that the Marburg I allele of FSAP is not a risk factor for venous thrombosis. This difference may be explained by the frequency of the Marburg I allele in the control population studied by Hoppe et al,1 which was considerably lower (0.012) than that reported in other studies, including ours (0.023-0.043).5,6 In thrombosis patients the frequency of the Marburg I allele was similar in the study of Hoppe et al1 (0.039) and ours (0.030), so the low prevalence in the control group (5/426) seems to be the explanation for the findings by Hoppe et al, either because these were blood donors, in which case the mutant allele would be infrequent in a group selected on health, or because the control group was relatively small. On the other hand we cannot exclude small geographic differences in the Marburg I frequency.
Association of Marburg I polymorphism of factor VII-activating protease with venous thromboembolism is limited to idiopathic events
In this issue, van Minkelen et al report on venous thrombosis and the Marburg I variant of factor VII-activating protease (FSAP), commenting on a study published recently by our group in the journal.1 They essentially describe 2 findings that appear to be in contrast to our study: (1) The frequency of carriers of FSAP Marburg I in the controls of van Minkelen et al was considerably higher (6.4%) than that of our control group (2.3%); and (2) an association of FSAP Marburg I with deep vein thrombosis (DVT) could not be confirmed.
In reply to these comments we want to remark the following. (1) Indeed, the frequency of carriers of FSAP Marburg I in our controls (5/213; 2.3%)1 comprising healthy blood donors was lower than that of the Bruneck study cohort (37/810; 4.6%),2 the controls of van Minkelen et al (30/471; 6.4%), and the Northwick Park Heart Study cohort (176/2066; 8.5%).3 Different frequencies in the described study populations could be due to low sample size, selection bias, and different geographic distributions. To exclude the possibility that our controls might be biased by negative selection of FSAP Marburg I due to its association with atherosclerosis,2 we analyzed a larger, independent control group consisting of consecutive patients admitted to our institution for reasons other than venous thromboembolism (non-VTE patients). Main diagnoses of these patients included transient ischemic attack/stroke, family testing for hereditary thrombophilia, pregnancy complications, and myocardial infarction. Considering the positive association of FSAP Marburg I with progression of atherosclerosis2 and assuming at least no negative association of this variant with VTE (Hoppe et al1 and van Minkelen et al), the frequency of FSAP Marburg I carriers in these non-VTE patients (15/327; 4.6%) will presumably overestimate the normal frequency of this variant in our geographic region.
(2i) The data presented by van Minkelen et al included patients with DVT, making no distinction between idiopathic and secondary events. As we described in the study,1 the association of FSAP Marburg I with VTE (including DVT and/or pulmonary embolism) was nearly exclusively attributable to idiopathic events (Table 1; “Comparison vs controls I”). Secondary VTE due to acquired risk factors (immobilization, surgery, trauma, pregnancy, puerperium, malignancy) was not associated with this variant. (ii) To avoid possible biases by the control group (blood donors), we performed a statistical analysis based on the second, independent control group (non-VTE patients; Table 1; “Comparison vs controls II”). Again, FSAP Marburg I was significantly associated with idiopathic but not secondary VTE (odds ratio, 2.7; 95% confidence interval [CI], 1.2-6.1).
The FSAP genotypes have no influence on its capacity to activate factor VII in vitro.2 Moreover, the physiologic role of FSAP and the site where its action takes place remains unclear. Thus, the absence of a detectable difference in circulating factor VII activities with both FSAP genotypes as described by van Minkelen et al is not surprising.
Based on the additional data described here and as stated in our initial report,1 FSAP Marburg I is associated with idiopathic VTE.
Correspondence: Berthold Hoppe, Institute of Transfusion Medicine, Campus Virchow-Klinikum, Charité-Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin, Germany; e-mail: berthold.hoppe@charite.de.
Supported by grant 912-02-036 from the Netherlands Organization for Scientific Research (NWO). The LETS study was supported by grant 89-063 from the Netherlands Heart Foundation.
