Deep Vein Thrombosis and Pulmonary Embolism

A Single Disease Entity With Different Risk Factors?

In recent years, deep vein thrombosis and pulmonary embolism, its major complication, have been increasingly considered as a single disease entity, namely venous thromboembolism. Indeed, both autopsic¹ and clinical studies² have shown that approximately 90% of pulmonary emboli arise from the deep veins of the lower limbs. Moreover, an asymptomatic pulmonary embolism can be found in about half the patients presenting with a symptomatic proximal deep vein thrombosis.³ Finally, deep vein thrombosis and pulmonary embolism share many risk factors, such as age, major surgery or trauma, cancer, immobilization, pregnancy, oral contraceptives, and hormone replacement therapy.

In 1993, in a landmark study, Dahlbäck et al⁴ identified a functional anomaly in certain families that predisposed to deep venous thrombosis, which they named resistance to activated protein C. Shortly thereafter, it was shown that this anomaly was due to a mutation of factor V, named factor V Leiden, that reduced its normal capacity to be inactivated by activated protein C, thus creating a hemostatic imbalance in favor of thrombosis.⁵ More recently, a mutation of prothrombin, prothrombin G20210A, was also shown to be more frequent in patients with deep vein thrombosis than in the general population.⁶ Hence, it seemed logical that these inherited thrombophilic defects would also predispose to pulmonary embolism. Intriguingly, a study performed in consecutive nonselected patients admitted for suspected pulmonary embolism⁷ found no difference between the prevalence of factor V Leiden in patients in whom pulmonary embolism was ruled in or out. This series was criticized because factor V Leiden was diagnosed by a functional test and not by DNA analysis. However, in the past few years, this finding has been repeatedly confirmed^8910111213; the study by Margaglione et al, published in this issue of CHEST (see page 1405), adds to this growing body of data. Moreover, these authors show similar observations for factor II G20210A, but their results are not confirmed by another recent series,⁹ in which the prevalence of the mutation was similar in patients with or without venous thromboembolism (Table 1 ).

Admittedly, all the studies on this topic suffer from methodologic flaws. Table 1 summarizes only the series in which a deep vein thrombosis was systematically searched in patients with confirmed pulmonary embolism by lower-limb venous compression ultrasonography or venography. However, none of these studies searched for an asymptomatic pulmonary embolism in patients with so-called isolated deep vein thrombosis. Nonetheless, taking into account asymptomatic pulmonary embolism in deep vein thrombosis patients would only increase the observed risk of deep vein thrombosis conferred by factor V Leiden, and would not change the conclusion that this defect is a much weaker risk factor for pulmonary embolism. Second, some of these studies, such as that by Margaglione et al, were performed in highly selected patients, as reflected by the mean age of patients with venous thromboembolism (43 to 48 years). Indeed, that series included patients referred for a thrombophilic workup, a group in which the prevalence of all inherited thrombophilic anomalies is expected to be greater than in nonselected patients. But again, the fact that the results show the same pattern for factor V Leiden as studies performed in nonselected populations only strengthens its conclusion. However, the selection bias in the series by Margaglione et al could account for the lack of association between the prothrombin mutation and venous thromboembolism observed in nonselected patient populations.⁹¹³ Finally, autopsic studies have shown that the prevalence of factor V Leiden among patients with fatal pulmonary embolism was not different from that of the general population,¹⁴ although similar data are lacking for the prothrombin mutation. Hence, the conclusion is inescapable: factor V Leiden is a strong risk factor for deep vein thrombosis, but not as strong a risk factor for pulmonary embolism. Regarding factor II G20210A, the jury is still out.

How then can this observation on factor V Leiden be explained? Two main hypotheses may be proposed. First, the intimate structure of the thrombus might be different in patients with a thrombophilic defect and render it more adherent to the venous wall in patients with the factor V Leiden or the prothrombin G20210A mutations, and, therefore less likely to embolize. Although this field is still in its infancy, some in vitro studies lend support to this hypothesis. Thrombin generation appears to be enhanced in the presence of factor V Leiden compared to normal factor V.¹⁵ Whether this might generate a more potent local inflammatory reaction in vivo and, thus, enhanced thrombus adherence remains to be shown. Moreover, increased thrombin formation may impair fibrinolytic response through activation of thrombin activable fibrinolysis inhibitor, a mechanism shown to exist in vitro in the presence of factor V Leiden.¹⁶ Second, an interesting study by Björgell et al,¹⁷ performed in patients with venographically confirmed deep vein thrombosis, showed that patients with factor V Leiden had a much lower prevalence of iliofemoral deep vein thrombosis than those with a normal factor V (2.8% vs 23.2%). Since it is widely admitted that calf deep vein thrombosis has a low potential for embolization unless proximal extension of the clot occurs, that observation might provide an elegant explanation for the weak association between factor V Leiden and pulmonary embolism. However, the link between that phenomenon and the in vitro anomalies described previously is elusive, and the reason why patients with the Leiden mutation tend to have thrombi limited to the calf veins is unknown.

Could these observations have any clinical relevance? Considering the ongoing debate whether anticoagulant treatment should be prolonged or even administered indefinitely in patients with a first episode of deep vein thrombosis and factor V Leiden, these data are certainly important. Indeed, pulmonary embolism is the only potentially fatal complication for deep vein thrombosis, and the potential of anticoagulant treatment to prevent the postthrombotic syndrome is at best unclear. Moreover, postthrombotic syndrome is rare in patients with calf deep vein thrombosis. Hence, this raises the challenging thought that patients with the factor V Leiden mutation might paradoxically be treated even less rather than more aggressively than those with a normal factor V. Of course, this remains to be demonstrated by well-designed clinical studies.

In summary, the weak association between the factor V Leiden and pulmonary embolism is in sharp contrast with the observations regarding deep vein thrombosis. This is certainly insufficient to make us entirely question the concept of venous thromboembolism as a single entity. However, it highlights that venous thromboembolism is a disease with a wide spectrum of clinical manifestations, the severity of which may be determined at least in part by genetic factors.