Venous thromboembolism (VTE) is estimated to affect 300,000-600,000 people in the United States each year.¹ VTE is the third leading cause of cardiovascular death², and it disproportionately affects Blacks.³ Because current diagnostic testing for VTE fails to identify underlying pro-thrombotic tendency in about 50% of patients, identification of novel risk factors for VTE is essential.⁴ Furthermore, several risk factors known to be associated with risk of VTE in Whites have been shown to have little impact on VTE risk in Blacks.^5-7 Identification of risk factors that may explain these racial differences may prove important in preventing VTE and reducing associated health disparities.

Several reports have indicated that high levels of pro-coagulant proteins may be independent risk factors for VTE.^8-12 Factor VIII (FVIII) circulates in plasma bound to von Willebrand factor (VWF) and is proteolytically cleaved during clot formation to yield activated FVIII which serves as a cofactor for the activation of Factor X (FX). Subsequently, activated FX serves as a cofactor for the conversion of prothrombin to thrombin, which acts on fibrinogen to form a fibrin clot. VWF stabilizes FVIII and provides an adhesive linkage between platelets and the subendothelium at sites of vascular injury. Elevated levels of FVIII have consistently been shown to be associated with risk of VTE, ^8-10 while elevated levels of VWF and fibrinogen have not been consistently associated with an increased risk of VTE.^8,11-14 Furthermore, ethnic differences in mean steady-state levels of these proteins have been reported, with Blacks having higher average levels of both FVIII and VWF.^15,16 FVIII, VWF, and fibrinogen, however, are acute phase reactants and are elevated in some conditions known to be risk factors for VTE. This study examines the relationship between pro-coagulant protein levels measured after VTE events in a group of VTE cases compared to protein levels measured in a group of control patients and risk of VTE in both Blacks and Whites after adjustment for covariates.

Out of 1145 enrolled cases, 256 were eligible for this study (51 reported being a race other than Black or White and 889 did not return to the CDC laboratory for blood specimen collection after completion of anticoagulant therapy). Of the 256 eligible VTE cases, 152 cases were considered provoked, and 104 cases were considered idiopathic. The mean time between VTE event and return to CDC laboratory for testing was 10.5 (± 6.0) months. Out of the 1309 enrolled controls, 1242 were eligible for this study (45 reported being a race other than Black or White and 67 did not have a blood specimen). There were 667 White and 575 Black controls and 140 White and 116 Black cases eligible for the study. The mean FVIII, VWF, and fibrinogen levels for White controls were 142.3 IU/dl, 130.9 IU/dl, and 3.4 g/L, respectively. Whereas the mean FVIII, VWF, and fibrinogen levels for Black controls were 168.3 IU/dl, 149.0 IU/dl, and 3.7 g/L, respectively. The mean FVIII, VWF, and fibrinogen levels for White cases were 179.7 IU/dl, 162.5 IU/dl, and 3.8 g/L, respectively. The mean FVIII, VWF, and fibrinogen levels for Black cases were 209.4 IU/dl, 195.8 IU/dl, and 3.9 g/L, respectively. There was no statistically-significant difference in the mean FVIII (183.6 IU/dl vs 177.5 IU/dl, p=0.61), VWF (163.4 IU/dl vs 162.0 IU/dl, p=0.92), or fibrinogen (3.79 g/L vs 3.78 IU/dl, p=0.93) levels measured in White idiopathic VTE cases compared to White provoked VTE cases. Similarly, there was no statistically-significant different in the mean FVIII (213.4 IU/dl vs 205.8 IU/dl, p=0.56), VWF (193.4 IU/dl vs 198.0 IU/dl, p=0.78), or fibrinogen (3.79 g/L vs 3.97 g/L, p=0.28) levels measured in Black idiopathic cases compared to Black provoked cases.

Covariates that were found to be independent predictors of VTE are shown in Table 1. The strongest independent predictors of VTE risk included cancer diagnosis, renal disease diagnosis, and recent surgery. The variables shown in Table 1 were used in the full model to estimate adjusted odds ratios for risk of VTE.

Crude odds ratios for odds of VTE in patients with high protein levels compared to those with low protein levels are shown in Table 2. Before adjustment for independent predictors of VTE risk, high levels of FVIII, VWF, and fibrinogen were significantly associated with increased risk of VTE in both Blacks and Whites. Furthermore, combinations of high protein levels tended to be associated with greater increased risk of VTE, particularly for Whites. However, after adjustment for ABO type, FVII levels, hypertension, kidney disease, recent surgery, diabetes, annual household income, alcohol use, and other proteins of interest (FVIII, VWF, and/or fibrinogen) (Table 2) several associations were no longer significant. High FVIII and VWF were associated with increased risk of VTE in Blacks (OR: 1.97 (1.01-3.84) and 3.39 (1.58-7.27), respectively), with high levels of both proteins conferring even greater risk (OR: 4.20 (2.24-7.89)). High fibrinogen was no longer associated with significantly increased risk of VTE in Blacks (OR: 1.38 (0.75-2.54). High FVIII was the only protein level significantly associated with risk of VTE in Whites (OR: 2.35 (1.16-4.75)).

The objective of this study was to assess the association of FVIII, VWF, and fibrinogen levels and risk of VTE and to assess any differences in risk by race. Previous studies have implicated high FVIII, VWF, and fibrinogen levels as risk factors for VTE, with high FVIII levels being more consistently associated with increased risk.^8-12,14,19 However, many of these studies only adjusted for age and sex and failed to adjust for other covariates that impact both factor level and risk of VTE. We have used logistic regression models to control for such confounders and in doing so have eliminated high fibrinogen and VWF levels as risk factors for VTE in Whites. Both FVIII and VWF, however, remain independent predictors of risk for VTE in Blacks. While FVIII, VWF, and fibrinogen levels are related, controlling for levels of the proteins when assessing the effect of high levels of the other proteins allowed the assessment of the independent contribution of each protein to risk of VTE.

Recent publications have suggested the mechanism conferring VTE risk for high FVIII levels relates to constitutively high levels as opposed to spikes in levels, as is the case with the acute-phase response.^20,21 Our findings agree with these results. After controlling for variables likely implicated in acute-phase response (e.g., diabetes diagnosis and recent surgery), we find high levels of FVIII remain an independent risk factor for VTE in both Blacks and Whites. Furthermore, after controlling for FVIII and VWF levels as well as other variables implicated in the acute-phase response, fibrinogen (a marker of the acute-phase response) is no longer associated with risk of VTE.

This report is the first to suggest that high VWF levels are associated with risk of VTE in Blacks only. Reports have suggested certain polymorphisms in the gene coding VWF are related to levels of VWF and that the distribution of these polymorphisms differs by race.²² Perhaps these polymorphisms also confer increased risk for VTE and help explain the differing results between Blacks and Whites in our population. Future studies are needed to assess this relationship. However, if these findings hold, high levels of VWF as a risk factor for VTE in Blacks may help explain some of the racial disparity in risk of VTE and could prove an important risk factor to assess in the clinical setting.

Because traditional diagnostic techniques for VTE fail to implicate an underlying inherited or acquired prothrombotic tendency in up to 50% of patients⁴, identification of novel risk factors such as FVIII, VWF, or fibrinogen levels could aid in the identification of patients at-risk for developing VTE before the event occurs. The findings in this report suggest measurement of FVIII activity could be considered when assessing VTE risk clinically in Whites and that both FVIII and VWF levels could be considered when assessing risk in Blacks. However, more work regarding the predictive value of these measurements is necessary before recommending the tests be used routinely in clinical settings.

The strengths of this study include its relatively large sample of VTE cases and controls and its ability to control for likely confounders of the association between FVIII, VWF, and fibrinogen levels and risk of VTE. A weakness of this study is the high rate of lost to follow-up for cases. A number of cases did not return to the CDC laboratory after completion of anticoagulant therapy for blood sample collection, possibly resulting in a biased case group. However, comparison of key indicators of health collected in the study questionnaire and completed by all cases indicates those who returned for blood sample collection were not significantly different than those cases who did return (Table 3). Another limitation of this study is the timing of procoagulant protein measurement. VTE cases were identified prior to enrollment in the study, and procoagulant proteins were measured subsequent to enrollment. In order to prevent measurement bias, we required all cases to complete anticoagulant therapy before blood sample collection. However, it is possible levels of FVIII, VWF, or fibrinogen may differ before and after a biologically stressful event such as VTE. This report assumes ample time between event and measurement allowed for the return to of these proteins basal levels.

The results of this study indicate that high levels of FVIII and VWF are independent risk factors for VTE in Blacks and high levels of FVIII are a risk factor for VTE in Whites. After accounting for likely confounders, high fibrinogen levels were not a risk factor for VTE in this population. Future research into the inclusion of FVIII and VWF levels in risk models for VTE could help define those at highest risk for an event and could help explain some of the racial disparity in risk of VTE between Blacks and Whites.