Paul A. Kyrle MD, Allgemeines Krankenhaus Wien, Klinik für Innere Medizin I, Abteilung für Hämatologie und Hämostaseologie, Währinger Gürtel 18–20, A-1090 Wien, Austria. Tel.: + 43 14 0400 4410; fax: + 43 1402 6930; e-mail: email@example.com
Summary. High factor IX (FIX) is a risk factor of deep vein thrombosis. The impact of high FIX on the risk of recurrent venous thrombosis is unknown. We prospectively followed 546 patients after anticoagulation for a first spontaneous venous thromboembolism. Patients with a natural coagulation inhibitor deficiency, lupus anticoagulant or cancer were excluded. At 3 years, the likelihood of recurrence was 23% among patients with high FIX (exceeding the 75th percentile) compared with 11% among patients with lower levels. Among patients with high FIX, the relative risk of recurrence was 2.2 (95% CI: 1.3–3.6) before and was 1.6 (95% CI: 1.0–2.8) after adjustment for age, gender, duration of anticoagulation, FV Leiden, FII G20210A, high FVIII and hyperhomocysteinemia. Compared with patients with low factor IX (< 138 IU dL−1) and low FVIII (≤ 234 IU dL−1), the relative risk of recurrence was 1.5 among patients with high FIX and low FVIII, 2.7 among patients with low FIX and high FVIII and 6.6 among patients with high FIX and high FVIII. High levels of FIX confer an increased risk of recurrent venous thromboembolism and enhance the risk of recurrence among patients with high FVIII.
Deep vein thrombosis (DVT) is a frequently occurring disease with an incidence of 1–2 per 1000 persons per year [1–3]. Various risk factors for venous thrombosis have been established [4, 5]. Among the environmental risk factors, surgery, malignancy, pregnancy and contraceptive pill use are the most important. A large proportion of thrombosis patients carries at least one congenital or acquired abnormality in the coagulation system, such as a natural coagulation inhibitor deficiency (antithrombin, protein C, protein S), a mutation in a gene encoding for a coagulation factor (FV G1691A, FII G20210A), lupus anticoagulant or hyperhomocysteinemia. There is now strong evidence that high plasma levels of some coagulation factors are independent risk factors of a first episode of DVT. Both high levels of factor VIII (FVIII) and factor XI are independent risk factors for a first DVT [6–9]. Likewise, a high level of factor IX (FIX) has recently been reported as an independent risk factor of venous thrombosis . In the Leiden Thrombophilia Study, individuals with a FIX level exceeding the 90th percentile (129 IU dL−1) had a two- to three-fold increased risk of a first episode of DVT. The risk of recurrence among patients with high plasma levels of FIX is, however, unknown. Guidelines regarding the optimal duration of secondary thromboprophylaxis for this patient group are therefore lacking. We prospectively followed 546 patients with a first episode of spontaneous venous thromboembolism and investigated the effects of high levels of FIX on the risk of recurrent venous thromboembolism.
Patients and study design
The Austrian Study on Recurrent Venous Thromboembolism (AUREC) is an ongoing, prospective study involving patients from four thrombosis centers in Vienna, Austria. Patients were eligible if they were older than 18 years and had been treated with oral anticoagulants for at least 3 months after an episode of venous thromboembolism. All the patients had received standard heparin or low-molecular-weight heparin at therapeutic dosages. Patients were excluded for the following reasons: venous thromboembolism before the recent episode, surgery, trauma, or pregnancy within the previous 3 months, deficiency of antithrombin, protein C, or protein S, the presence of lupus anticoagulant, cancer, or a requirement for long-term treatment with antithrombotic drugs for reasons other than venous thrombosis.
The day of discontinuation of oral anticoagulant therapy was defined as the day of enrollment in the study. Blood for measurement of the FIX level was obtained three weeks after enrollment. At that time the prothrombin time was normal in all patients. Patients were seen at 3-month intervals during the first year after enrollment and every 6 months thereafter. They were given detailed written information on the symptoms of venous thromboembolism and were instructed to report immediately to one of the thrombosis centers if such symptoms appeared. At each visit, the medical history was obtained and a physical examination was performed.
The study was approved by the Ethics Committee of the University of Vienna and patients had to provide informed consent prior inclusion.
Diagnosis of venous thromboembolism
The diagnosis of DVT was based on a positive finding on venography or colour duplex sonography (in the case of proximal DVT). To be considered positive, the venograms had to meet at least one of the following direct or indirect criteria: a constant filling defect seen on two views; an abrupt discontinuation of the contrast filled vessel at a constant level of the vein; and the absence of filling in the entire deep vein system (without a compression), with or without venous flow through collateral veins. With colour duplex sonography, at least one of the two following criteria for DVT had to be met: visualization of an intraluminal thrombus in a deep vein and incomplete compressibility or absence of compressibility. The diagnosis of pulmonary embolism was established either by a positive finding on ventilation–perfusion lung scanning according to the criteria of the Prospective Investigation of Pulmonary Embolism Diagnosis  or by spiral computed tomography revealing one or several low-attenuation areas that partly or completely filled the lumen of an opacified vessel. Patients with both DVT and pulmonary embolism were classified as pulmonary embolism.
The end point of the study was recurrence of symptomatic venous thromboembolism, confirmed by venography, ventilation-perfusion lung scanning or spiral computed tomography of the lung according to the same diagnostic criteria. DVT was considered to have recurred if the patient had a thrombus in the leg or arm other than that affected by the previous thromboembolic event; a thrombus in another deep vein in the same leg or arm as the previous event; or a thrombus in the same venous system as the previous event, with proximal extension of the thrombus (if the upper limit of the original thrombus had been visible) or with a constant filling defect surrounded by contrast medium (if the original thrombus had not been visible).
After overnight fasting, venous blood was collected in a 1 : 10 dilution of 0.11 mol L−1 trisodium citrate. A portion of the collected blood was centrifuged for 20 min at 2000 × g, and the plasma was stored at −80 °C. For measurement of homocysteine, another portion of the collected blood was immediately centrifuged at 1600 × g for 20 min at 4 °C. The plasma was snap-frozen and stored at −80 °C. Genomic DNA was isolated by standard methods. FVIII was measured by a one-step clotting assay as recently described . FIX was measured by a one-step clotting assay with use of a Factor IX-deficient plasma obtained from Immuno Baxter (Baxter Healthcare, Vienna, Austria), a fully automated coagulation analyser (CA 6000, Sysmex, Kobe, Japan) and commercially available pooled normal plasma (Coag Cal N, Dade Diagnostics, Duedingen, Switzerland) calibrated against World Health Organization standard 91/666. Intra- and interassay variabilities of the FIX assay system were below 5%, respectively.
Antithrombin, protein C, protein S and total plasma homocysteine were determined as previously reported [13, 14]. Hyperhomocysteinemia was diagnosed when the homocysteine level was above the 95th percentiles (8.8 µmol per litre in women and 11.6 µmol per litre in men) of the levels measured in 73 healthy control subjects who were similar to the study patients with regard to age and gender distribution. Screening for FV Leiden and for FII G20210A was carried out as described [15, 16]. The presence of lupus anticoagulant was assessed according to the criteria of the International Society on Thrombosis and Haemostasis .
For numerical operations, SPSS software (SPSS Inc. Headquarters, Chicago, Illinois, USA) was used. Values are given as means ± SD. For comparison of categorical and numerical data the chi-squared test and the Mann–Whitney U-test, respectively, were applied. All P-values are two-tailed. Times to recurrence (uncensored observations) or follow-up times in patients without recurrence (censored observations) were analysed according to survival-time methods . The probability of recurrence was estimated according to the method of Kaplan and Meier . Data of patients who left the study because of requirement of antithrombotic treatment (for causes other than venous thrombosis), diagnosis of cancer, lost to follow-up or death were censored at the time of withdrawal. To test for homogeneity among the various groups of patients, we used the log-rank test and the generalized Wilcoxon test. The plasma level of factor IX was analyzed in Cox proportional-hazards models as a continuous variable and after stratification into several groups according their FIX levels (in a separate analysis) to compare the relative risks and the 95% confidence intervals (CI) of recurrent venous thromboembolism associated with different FIX levels. The data were adjusted for age, gender, duration of oral anticoagulation, the presence or absence of FV Leiden and of FII G20210A, homocysteine (dichotomized at the 95th percentile of normal) and FVIII (dichotomized at 234 IU dL−1, which is the threshold level for an increased risk of recurrent venous thromboembolism in our study population) .
The baseline characteristics of the 546 patients are given in Table 1. The mean age of the patients was 48 years and 54% were female. The patients had received oral anticoagulants for an average of 8 months. Carriers of the FV Leiden mutation numbered 170 patients (31%), and 44 patients (8%) had FII G20210A. Hyperhomocysteinemia was found in 102 patients (20%), and 62 patients (11%) had FVIII levels exceeding 234 IU dL−1. After discontinuation of treatment with oral anticoagulants, the patients were followed for an average of 31 months.
Table 1. Baseline characteristics of the 546 patients
Recurrent venous thromboembolism occurred in 66 patients (12%). None of these events was fatal. At recurrence, 44 patients (67%) had DVT only and 22 (33%) patients had pulmonary embolism with or without DVT. In 21 patients DVT occurred in the ipsilateral leg and in 26 patients in the contralateral leg. Recurrence was seen in 46 men and in 20 women. There was no statistically significant difference between patients with and without recurrence with regard to age (52 vs 48 years, respectively), the presence of factor V Leiden (29% and 32%) or FII G20210A (14% and 8%), or hyperhomocysteinemia (23% and 20%). The proportion of patients with high FVIII levels was higher among patients with recurrence than among patients without recurrence (24% and 10%, P = 0.001).
Recurrent venous thromboembolism and factor IX levels
Patients with recurrent venous thromboembolism had significantly higher levels of FIX than those without recurrence (133 ± 24 IU dL−1vs 123 ± 26 IU dL−1, P = 0.001). When FIX was analysed as a continuous variable in a Cox proportional-hazard model, the relative risk of recurrent venous thromboembolism was 1.16 (CI: 1.07–1.25) for each FIX increase of 10 IU dL−1 in the univariate analysis and was 1.08 (CI: 0.99–1.18) after adjustment for age, gender, duration of anticoagulant treatment, FV Leiden, FII G20210A, homocysteine and FVIII.
To determine whether there is a dose–response relation between FIX and the risk of recurrent venous thromboembolism, stratification of patients according to their FIX level was performed. Compared with the reference group (patients with FIX levels ≤ 100 IU dL−1), the relative risk of recurrence was 3.0 (CI 1.0–8.7) among patients with FIX levels between 101 and 124 IU dL−1, 3.9 (CI 1.3–11.4) among patients with FIX levels between 125 and 149 IU dL−1 and 5.2 (CI 1.8–15.6) among patients with FIX levels > 150 IU dL−1.
According to the Kaplan–Meier analysis, patients with FIX levels exceeding the 75th percentile (≥ 138 IU dL−1) had a higher likelihood of recurrence than patients with FIX levels below 138 IU dL−1 (P < 0.001 by the Wilcoxon test and P = 0.001 by the log rank test) (Figure 1). At 3 years, the probability of recurrence was 23% (CI: 14–31%) among patients with high FIX levels as compared with 11% (CI: 7–15%) among patients with lower levels. FIX levels above the 75th percentile conferred a relative risk of recurrence of 2.2 (CI: 1.3–3.6). After adjustment for age, gender, duration of anticoagulation, FV Leiden, FII G20210A, homocysteine and FVIII, the relative risk of recurrence was 1.6 (CI: 1.0–2.8).
FVIII levels exceeding 234 IU dL−1 confer a high risk of recurrent venous thromboembolism . We therefore investigated the relationship between FIX and FVIII with regard to the risk of recurrence. Compared with the reference group (patients with FIX < 138 IU dL−1 and FVIII < 234 IU dL−1, patients with high FVIII (> 234 IU/dL) and low FIX (< 138 IU dL−1) had a higher risk of recurrence than patients with low FVIII (≤ 234 IU dL−1) and high FIX (≥ 138 IU dL−1). The risk of recurrence was highest (6.6-fold) among patients with both high FVIII (> 234 IU dL−1) and high FIX (≥ 138 IU dL−1) (Table 2). The presence of the FV Leiden mutation or high plasma levels of homocysteine did not affect the risk of recurrent venous thromboembolism among patients with high FIX (data not shown).
Table 2. Relative risk of recurrent venous thromboembolism according to plasma levels of FVIII and FIX
In the Leiden thrombophilia study a high level of FIX emerged as an independent risk factor of a first episode of DVT . The risk of recurrence of venous thromboembolism among patients with high plasma levels of FIX is, however, unknown. Thus, it is currently unclear whether patients with high FIX may benefit from extended anticoagulant therapy. The principal finding of our study was that a high level of FIX confers an increased risk of recurrent venous thromboembolism. At three years, the likelihood of recurrence was 23% among patients with FIX level exceeding the 75th percentile, compared with 11% among patients with lower levels. Patients with high FIX (> 75th percentile) levels had a more than two-fold increased risk of recurrence compared with patients with lower levels. In the multivariate analysis, the relative risk reduced to 1.6 (CI 1.0–2.8), indicating that among patients with high FIX the risk of recurrence is partially mediated by other confounders. Stratification of patients according to their FIX levels revealed a gradual relationship between increasing FIX levels and the risk of recurrent venous thromboembolism. This is in agreement with the Leiden thrombophilia study, which also revealed a linear relationship between FIX and the risk for a first episode of DVT .
What are the implications of our findings with regard to the management of thrombosis patients? The optimal duration of secondary thromboprophylaxis entails balancing the risk of recurrence against the risk of hemorrhage as a consequence of anticoagulation. The annual incidence of severe bleeding among patients receiving anticoagulant therapy ranges from 1 to 2% and approximately 25% of these bleeds are fatal [20–22]. On the other hand, there are several potent risk factors of recurrent venous thrombosis including antithrombin deficiency, lupus anticoagulant, malignancy and recurrent venous thrombosis [23–26]. In these patients, the risk of recurrence most probably outweighs the risk of bleeding during antithrombotic treatment. Therefore, patients harbouring one of these risk factors usually receive long-term secondary thromboprophylaxis. Compared with the severity of the aforementioned risk factors, the risk of recurrent venous thromboembolism conferred by high FIX levels is considerably smaller. We thus believe that patients with high factor IX, although at a somewhat increased risk of recurrence, probably will not benefit from extended anticoagulation.
The risk of venous thrombosis is particularly high in individuals with more than one acquired and/or congenital thrombotic risk factor. For instance, Ridker et al. reported a higher risk of venous thromboembolism among apparently healthy men with both FV Leiden and hyperhomocysteinemia as compared with individuals with only one of the risk factors . There is evidence from family studies for a synergistic effect of FV Leiden and protein S or protein C deficiency with regard to the risk of a first episode of venous thrombosis [28, 29]. Likewise, elevated FIX levels have been shown to further enhance the risk of a first DVT in subjects with high FVIII . Only limited data are available regarding the risk of recurrence of venous thromboembolism in patients with more than one thrombotic risk factor. According to two large retrospective studies from Italy, the coexistence of FV Leiden and the FII G20210A is associated with a 3–5 fold increased risk of recurrent thrombosis [30, 31]. We therefore investigated the impact of high levels of FIX on the risk of recurrence among patients with FV Leiden, hyperhomocysteinemia or high FVIII levels. We and others have shown, that FV Leiden does not confer an increased risk of recurrent venous thromboembolism, whereas elevated total homocysteine is an independent risk factor of recurrence [14, 30–34]. From the present study it is evident, that the coexistence of one of the two aforementioned risk factors with high FIX levels does not result in a higher risk of thrombosis than that conferred by FV Leiden or hyperhomocysteinemia themselves. We have recently shown that F VIII levels exceeding 234 IU dL−1 confer a very high risk of recurrent venous thromboembolism . Accordingly, in the present analysis patients with high FVIII had a substantially higher risk of recurrence than patients with lower FVIII regardless of the FIX level. The risk of recurrence, however, was highest (six-fold) among patients with both high FVIII and high FIX. Our observation of a synergistic effect between high levels of factor VIII and factor IX again supports the concept that the pathogenesis of (recurrent) venous thrombosis is multicausal.
This work was supported by a grant of the Jubiläumsfonds of the Oesterreichische Nationalbank (Nr. 7529).