Thrombophilia and outcomes of venous thromboembolism in older patients

Background Limited data exist on thrombophilic risk factors and clinical outcomes in the elderly with venous thromboembolism (VTE). Objectives To describe the prevalence of laboratory thrombophilic risk factors and their association with VTE recurrence or death in a cohort of elderly people with VTE. Methods In 240 patients aged ≥65 years with acute VTE without active cancer or an indication for extended anticoagulation, we performed laboratory thrombophilia testing 1 year after the index VTE. Recurrence or death was assessed during the 2-year follow-up. Results A total of 78% of patients had ≥1 laboratory thrombophilic risk factor(s). Elevated levels of von Willebrand factor, homocysteine, coagulant activity of factor VIII (FVIII:C), fibrinogen, FIX:C, and low antithrombin activity were the most prevalent risk factors (43%, 30%, 15%, 14%, 13%, and 11%, respectively). Additionally, 16.2% of patients experienced VTE recurrence and 5.8% of patients died. Patients with a von Willebrand factor of >182%, FVIII:C level >200%, homocysteine level >15μmol/L, or lupus anticoagulant had a significantly higher rate of recurrence than those without these risk factors (15.0 vs. 6.1 [P = .006], 23.5 vs. 8.2 [P = .01], 17.0 vs. 6.8 [P = .006], and 89.5 vs. 9.2 [P = .02] events per 100 patient-years, respectively). Furthermore, patients with a high fibrinogen level or hyperhomocysteinemia with a homocysteine level ≥30 μmol/L had significantly higher mortality than patients with normal levels (18.5 vs. 2.8 [P = .049] and 13.6 vs. 2 [P = .002] deaths per 100 patient-years, respectively). After adjustments for relevant confounders, these associations remained unchanged. Conclusion Laboratory thrombophilic risk factors are common in elderly people with VTE and allow for the identification of a population at the risk of worse clinical outcomes.


| I N T R O D U C T I O N
Risk factors for venous thromboembolism (VTE) include clinical conditions and laboratory-based blood parameters. [1] Age is a strong driver of index VTE risk, likely owing, at least in part, to comorbidities, medications, immobilization, changes in the vessel wall, and stasis as part of the Virchow triad. [1] Whether laboratory-based blood parameters influence VTE recurrence and outcomes in elderly patients is unknown. Elderly patients are often excluded from studies owing to comorbidities, short life expectancy, and logistical difficulties concerning data collection. [2] In the case-control Age and Thrombosis, Acquired and Genetic risk Factors in the Elderly Study (AT-AGE) study, a subgroup of 312 outpatients aged ≥70 years with a first episode of VTE showed an association between high levels of coagulation factors VIII (FVIII) and XI (FXI) and the occurrence of VTE, and another subgroup of 163 patients showed an association between a high level of factor IX (FIX) with first VTE; the associations were calculated by comparing the highest and lowest quartiles (odds ratios of 4.5, 1.7, and 2.4 for FVIII, FXI, and FIX, respectively). [2] People with active malignancy, a history of VTE, or severe cognitive impairment were excluded from that study. No outcome data were reported. In another subgroup of 394 patients from the AT-AGE study, genetic analyses showed a prevalence of 8.6% for factor (F)V Leiden and 2.3% for the prothrombin G20210A polymorphism, and the presence of the FV Leiden variant was associated with a 2.4-fold increased risk of a first unprovoked VTE, but not the prothrombin G20210A polymorphism. [3] Analysis of 65 patients aged ≥65 years in the Cardiovascular Health Study showed a statistically significant association between levels of FVIII >124% and incident cases of first VTE, whereas no such association was found for FVII or von Willebrand factor (VWF). [4] In a separate analysis of this study population using a nested case-control study design, an FXI level >157% was independently associated with the increased risk of a future first VTE, whereas elevated levels of FIX, factors X (FX), XII (FXII), and XIII (FXIII) were not. [5] Further analyses of this study population showed an association between the FV Leiden variant or the prothrombin G20210A polymorphism and VTE, but not for levels of FII, FV, or homocysteine or for the methylene tetrahydrofolate reductase C677T gene polymorphism. [6][7][8] In 2 retrospective studies, the risk of a first VTE increased linearly with increasing levels of FVIII in patients aged >70 years, reaching a 2.4fold increased risk in people with FVIII >225%, [9] and VTE occurrence was associated with FV Leiden in patients aged >60 years. [10] In addition, we have previously shown in 354 elderly in-and outpatients with a first unprovoked VTE and a median follow-up Essentials • Blood tests often show a risk of clotting in older patients who have already had a blood clot.
• We performed these tests in 240 patients aged ≥65 years, 1 year after a blood clot.
• Specific blood tests showed a risk of a further blood clot after stopping a blood-thinning therapy.
• The risk of a recurrent blood clot increased with the number of laboratory risk factors present.
2 of 11period of 30 months that there was no association between FV Leiden and the prothrombin G20210A polymorphism with recurrence [11] and that prohemostatic GAS6 plasma concentrations were independently associated with VTE recurrence and death. [12] Despite limited data on laboratory risk factors for first VTE in the elderly, it stands to reason that laboratory thrombophilic risk factors may also play a role in the elderly regarding outcomes such as VTE recurrence and death. Therefore, we explored the prevalence of elevated levels of procoagulant factors, such as fibrinogen, FVIII, VWF, FIX, FXI, and homocysteine; the presence of anti-phospholipid (aPL) antibodies including anticardiolipin (aCL), anti-β2-glycoprotein I IgG and IgM antibodies, and lupus anticoagulant (LA); FV Leiden and prothrombin G20210A polymorphism; and low levels of inhibitors of coagulation antithrombin (AT), protein C (PC), and protein S (PS) and the association of these factors with VTE recurrence and death in a prospective cohort of ≥65-year-old people with VTE.

| Study design, setting, and participants
This study was conducted between September 2, 2009, and December 6, 2013, as part of the Swiss Thromboembolism Cohort (SWITCO65+) study, a prospective multicenter cohort study that assessed long-term clinical outcomes in elderly people with acute VTE from 5 university and 4 high-volume nonuniversity public academic teaching hospitals in Switzerland. The study rationale and full design have been published previously. [11][12][13][14] The trial has been registered at http://clinicaltrials.gov (NCT00973596). Briefly, consecutive patients aged ≥65 years with an acute, objectively confirmed symptomatic VTE were prospectively identified in the in-and outpatient services of all participating study sites. Symptomatic pulmonary embolism (PE) was defined as a positive spiral computed tomography or pulmonary angiography, a high probability ventilation-perfusion scan, or proximal deep venous thrombosis (DVT) documented using compression ultrasonography or contrast venography in people with acute chest pain, new or worsening dyspnea, hemoptysis, or syncope.
Symptomatic DVT was defined as an acute onset of leg pain or swelling plus incomplete compressibility of a venous segment on ultrasonography or an intraluminal filling defect on contrast venography. Vitamin-K antagonist therapy was the long-term anticoagulant of choice for 95% of patients. The study was approved by the Institutional Review Board at each participating center.
In the present study, laboratory blood parameters, including fibrinogen, factor VIII coagulant activity (FVIII:C), VWF antigen (VWF:Ag), FIX:C, FXI:C, fasting homocysteine, aPL antibodies, AT activity, PC activity, and free PS antigen, were analyzed in 240 people with VTE. Similar to prior studies, [2,3,[14][15][16] blood sampling for these tests was performed after the completion of therapeutic anticoagulation at a follow-up visit 1 year after the index VTE event, and persons with ongoing anticoagulation or active malignancy were excluded. The flow chart of the study population is summarized in Supplementary Figure 1.
The total follow-up time was 2 years after thrombophilia testing (median, 718 days; IQR, 521-896 days). To avoid a bias in the choice or duration of therapy during the follow-up, thrombophilia tests were performed independently of the clinical management, and the results were not communicated to the patients or their managing physicians.

| Laboratory measurements
Details regarding blood sampling/processing and the methods of measuring thrombophilic factors were described elsewhere. [14] Citrated platelet-poor plasma was used for the following assays: i) FVIII:C, FIX:C, and FXI:C using one-stage coagulation assays using  as activity <70%, and free PS antigen deficiency as <55% in women and <60% in men; these latter cut-offs were the lower limits of the reference range established in the central laboratory. For aPL antibodies, we used the revised cut-off values as published earlier. [18] We defined positive aPL as aCL IgG level >13.6 U/mL, aCL IgM level >18.7 U/mL, aβ2GPI IgG level >17.4 U/mL, aβ2GPI IgM level >12.0 U/mL, or a positive LA.

| Clinical outcomes
The primary outcome was the recurrence of an objectively confirmed,

| Statistical analyses
Baseline characteristics are presented as median and IQR or number and percentage as appropriate. We compared baseline characteristics of patients with and without the presence of any thrombophilia. We defined unprovoked VTE as the absence of major surgery within the previous 3 months, estrogen therapy, and immobilization (fracture or cast of the lower extremity, bed rest >72 hours, or voyage in sitting position for >6 hours) during the last 3 months.
We calculated incidence rates of recurrent VTE and overall mortality for patients with or without a given thrombophilic risk factor using mid two-sided P values to test for incidence difference. We compared cumulative incidences using Kaplan-Meier curves and the log-rank test starting from the observation period at the time of blood sampling. We examined associations between laboratory thrombophilic risk factors and time-to-VTE recurrence using competing risk regression accounting for non-VTE-related death as a competing event and the association with death using Cox proportional hazards models. We examined the strength of these associations in further models adjusting for prior VTE and unprovoked VTE in the VTE recurrence model, and additionally for C-reactive protein when calculating the subhazard ratios (SHR) for FVIII and VWF, and for age in the mortality model.
Additional adjustment for inflammation using C-reactive protein was performed for the association between elevated fibrinogen levels and overall mortality. Adjustments for FVIII, VWF, and fibrinogen were made using C-reactive protein as a surrogate marker for inflammation as they act as positive acute phase proteins. We examined the interactions of VWF and FVIII in a joint model, with additional adjustment for unprovoked index VTE and prior VTE. We assumed missing and noninterpretable laboratory values to be normal. All analyses were performed using Stata 17 (Stata Corporation).

| R E S U L T S
As indicated in Supplementary Figure 1 Table 4).
The Kaplan-Meier estimates demonstrated that patients with a high level of FVIII had a 2-fold higher 2-year cumulative incidence of recurrent VTE than patients without this risk factor ( Figure 1A). The risk increased over time. Similar findings were noted with the presence of elevated VWF or high homocysteine (>15 μmol/L; Figure 1B, C). After adjustment for prior VTE and unprovoked index VTE, elevated FVIII, elevated VWF, and high homocysteine level >15 μmol/L all remained statistically significantly associated with VTE recurrence ( Kaplan-Meier estimates in Figure 2) as well as when adjusted for age (  thrombophilic risk factors. In this cohort, some laboratory parameters identified a population at risk of VTE recurrence and death. Our data show that the risk of VTE recurrence more than doubled in the presence of elevated levels of FVIII, VWF, or homocysteine and that the risk of death was more than 6-fold higher in the presence of elevated levels of fibrinogen. Moreover, despite a very low prevalence, LA was also associated with VTE recurrence and death.

| Prevalence of laboratory indicators of hypercoagulability
To our knowledge, only few prior studies addressed the prevalence of thrombophilic risk factors among persons aged ≥65 years with VTE, mostly in subgroups only. [2,4,[9][10][11]19] Most studies focused on isolated or selected thrombophilic risk factors in relation to a first VTE.
6 of 11surrogate marker, the elevated FVIII levels were independent of any inflammation. [22] We made the same observation in our elderly cohort. Although we excluded people with known active cancer from the analysis, we cannot exclude the possibility of an underlying occult malignancy. However, as only 2 patients died from cancer during the 2-year follow-up, a high prevalence of undiagnosed neoplastic disorders is unlikely.
We found elevated FIX and FXI in 13% and 7% of patients, respectively. We did not find comparable data for the elderly in the literature. Furthermore, a combined prevalence of endogenous anticoagulant (PS, PC, and AT) deficiencies of 16.25% was found in our study, with the highest prevalence for decreased AT (11%). This is surprising, as it is clearly higher than in a smaller subgroup analysis of the MAISTHRO registry in the patients aged ≥70 years, in which the findings for AT, PC, and PS deficiency were 1.4%, 0%, and 1%, respectively, [19] and it is also higher than that in younger patients with a mean age of 45 years (5%). [15] In the general population, the prevalence is <0.5%. [23] A detailed analysis of the low AT activity found in our study population showed a high median of 78.5% with an IQR between 74% and 81% ( Table 2). Three patients only had AT levels ≤70%, with the lowest level being 62%. A likely explanation for the antithrombin activities mostly moderately below the normal range in our elderly population is consumption or loss of the anticoagulant factor by renal (nephrotic syndrome) or gastrointestinal (proteinlosing enteropathy) disease.

| Clinical outcomes
In our analysis, among elderly people with VTE, elevated FVIII clotting activity, elevated VWF, increased homocysteine, and presence of T A B L E 3 Incidence rates of VTE recurrence in 240 patients followed up for 2 years after thrombophilia testing. a  anticoagulation. As in those other studies, [2,15,16]  In conclusion, laboratory thrombophilic risk factors are frequent in elderly people with VTE. They potentially allow identification of a population at risk of worse clinical outcome. Whether patients with high levels of VWF, FVIII, and homocysteine or with an LA might especially benefit from prolonged secondary prophylactic anticoagulation will need further prospective randomized trials.