Cover Page The handle http://hdl.handle.net/1887/19768 holds various files of this Leiden University dissertation. Author: Langevelde, Kirsten van Title: Are pulmonary embolism and deep-vein thrombosis always one disease? Date: 2012-09-11
C h a p t e r 1 0 Summar y and general discussion
Summary and general discussion 123 The aim of this thesis was to assess whether pulmonary embolism and deep-vein thrombosis are always the same disease or not, i.e., that one always occurs as a consequence of the other, and to answer this question with regard to etiology (risk factors) and anatomical distribution of thrombi. We studied this question in two populations: the PEDLAR study and the MEGA case-control study. In addition, we aimed to investigate the effect of aging on venous valve thickness. This was performed in an ultrasonography study, with participants from 20 to 80 years old (the aging venous valves study). The thesis starts with two review articles of the literature. In Chapter 2 common acquired and genetic risk factors for venous thrombosis were described. We focused on their effect size for deep-vein thrombosis versus pulmonary embolism separately. Chapter 3 gave an overview of the state-of-the-art techniques of imaging venous thrombosis, more specifically focused on MRI techniques. Chapter 4 described the aging effect on valve thickness in a group of healthy volunteers. In Chapter 5 we presented total body MRI findings in 99 patients with pulmonary embolism. We found no thrombus on MRI in 55 patients. Of the 44 pulmonary embolism patients with a thrombus on MRI, 12 patients had isolated calf vein thrombosis (27%) and 6 patients had a superficial vein thrombosis (14%) as the only peripheral thrombus. In Chapter 6 we assessed thrombus extent in the legs with MRI, and related that to the degree of pulmonary obstruction on CT pulmonary angiography in 56 patients with pulmonary embolism. Thrombus extent was defined as the length of the total clot and as the number of occluded veins. The Qanadli pulmonary obstruction score and the right ventricle/left ventricle diameter ratio were used to assess the severity of pulmonary embolism. Chapter 7 described the anatomical location of peripheral thrombi in patients with a deep-vein thrombosis alone, versus patients with a pulmonary embolism and a concomitant deep-vein thrombosis (PEDLAR study). In addition, we compared the prevalence of acquired and genetic risk factors between patient groups. Chapter 8 explored the relation between superficial vein thrombosis and deep-vein thrombosis or pulmonary embolism in the MEGA study. We found that the overall risk of venous thrombosis was 5-fold increased when somebody had a history of clinically diagnosed superficial vein thrombosis, compared with a negative superficial vein thrombosis status. In Chapter 9 we described CT pulmonary angiography yield in the Leiden University Medical Center. We found one out of five CT pulmonary angiographies to be positive for pulmonary embolism (19%). CT pulmonary angiography yield, or indirectly patient referral, was suboptimal in outpatients, women, and patients below 40 years of age.
124 Chapter 10 We discuss the main findings of this thesis in five themes: the role of venous valves in thrombus formation, the differences between pulmonary embolism and deep-vein thrombosis regarding risk factors and distribution of peripheral thrombi, the role of superficial vein thrombosis, imaging of venous thrombosis, and future directions. T h e r o l e o f v e n o u s v a l v e s i n t h r o m b u s f o r m a t i o n In the study on the aging of venous valves, we found that valve thickness slightly increased with age when studying the relation linearly. However, when we compared people above 50 years of age with people below 50 years, we found a 5-fold increased risk of having thicker valves in the older group. It could therefore well be that the increased risk of venous thrombosis with age is in part explained by thicker valves in the elderly. (Chapter 4) On the one hand, venous valves have been shown to be the starting point of thrombus formation, due to the low blood flow velocity and hypoxia in the valve pockets. On the other hand, the valves have antithrombotic properties, as their function is to maintain the return of venous blood to the heart to ensure continuous cardiac output. Venous valves are important to prevent stasis of blood and consequently in preventing venous thrombosis. In a histology study with immunofluorescence the microenvironment around valves from the great saphenous veins was shown to be antithrombotic. The expression of Von Willebrand factor was lower, whereas thrombomodulin and endothelial protein C receptor expression were higher than in the non-valvular lumenal endothelium. 169,269 The authors suggested that variation in valvular sinus thromboresistance may be an important factor in venous thrombogenesis. The valvular thromboresistant state may become less so with aging. D i f f e r e n c e s b e t w e e n p u l m o n a r y e m b o l i s m a n d d e e p - v e i n t h r o m b o s i s r e g a r d i n g r i s k f a c t o r s a n d d i s t r i b u t i o n o f p e r i p h e r a l t h r o m b i The results of this thesis offer new insights in the etiology of pulmonary embolism. The classic idea of a deep-vein thrombosis that becomes detached from the vessel wall and embolizes to the pulmonary arteries, seems not tenable at all times. Risk factors In Chapter 2 we reviewed the literature to study common acquired and genetic risk factors for venous thrombosis, focusing on their effect size for deep-vein thrombosis versus pulmonary embolism separately. We found several risk factors that gave a higher risk of deep-vein thrombosis than of pulmonary embolism, similar to factor V Leiden. These risk factors encompassed use of an oral contraceptive pill, 18,49 pregnancy and puerperium, 50-52
Summary and general discussion 125 and obesity. 43,44 Not only were they in line with the factor V Leiden paradox, they were also based on the mechanism of APC resistance, like factor V Leiden is. In addition, we found minor leg injury to be a stronger risk factor for deep-vein thrombosis than for pulmonary embolism. 48 We hypothesized that in minor leg injury only a local thrombosis is triggered instead of systemic activation of the clotting system. For example, knee ligament ruptures and ankle sprains may cause local swelling, aching and pain and possibly even mechanical damage to adjacent veins. We found that chronic obstructive pulmonary disease (COPD) and sickle cell disease posed a stronger risk of pulmonary embolism than of deep-vein thrombosis. 30,188 A localized process of inflammation in the lungs may be a provocative explanatory mechanism. Another example is pneumonia, an acute inflammatory process taking place in the lungs, which increased the risk of pulmonary embolism 8-fold in the MEGA study. For deep-vein thrombosis alone, the risk was only 3-fold increased. (Ribeiro et al., submitted, MEGA study). In Chapter 7 we compared risk factors for deep-vein thrombosis and pulmonary embolism in the PEDLAR study. The distribution of acquired risk factors was very similar, except for cancer. The prevalence of cancer was higher in pulmonary embolism patients (33%) than in deep-vein thrombosis patients (15%). The factor V Leiden mutation was present in 12% of deep-vein thrombosis patients, and in 14% of pulmonary embolism patients with a concomitant deep-vein thrombosis. 7% of patients with an isolated pulmonary embolism had the factor V Leiden mutation. Again, this is in line with the factor V Leiden paradox, as patients with both pulmonary embolism and a thrombosis of the legs tend to have a prevalence of the mutation similar to patients with an isolated deep-vein thrombosis, as opposed to patients with isolated pulmonary embolism. 17 Distribution of thrombi in patients with pulmonary embolism versus in patients with deep-vein thrombosis Pulmonary artery obstruction In Chapter 6 we assessed peripheral thrombus extent in the legs with MRI, and related that to the degree of pulmonary obstruction as measured on CT pulmonary angiography in 56 patients with pulmonary embolism. We found that patients with pulmonary embolism and a concomitant deep-vein thrombosis had more severe obstruction of the pulmonary arteries than those without deep-vein thrombosis. Besides, both thrombus length and the number of occluded veins were positively associated with the percentage of obstruction in the pulmonary arteries.
126 Chapter 10 Peripheral thrombus load In Chapter 7 we described the differences between patients with an isolated deep-vein thrombosis and those with a pulmonary embolism and a concomitant deep-vein thrombosis, regarding the anatomical location of thrombi in the legs. We compared MR images in both patient groups to measure thrombus length and the number of occluded veins. Patients with isolated deep-vein thrombosis had a higher number of occluded veins and longer thrombi than patients with pulmonary embolism and concomitant deepvein thrombosis. No difference was seen regarding proximal extent of a thrombus, i.e., in both patient groups 9% of thrombi were present in the abdominal or pelvic veins. We therefore concluded that these results do not confirm the idea that patients with pulmonary embolism have more extensive and more proximal deep-vein thrombi than patients with deep-vein thrombosis alone. It rather seems that patients with pulmonary embolism and a concomitant deep-vein thrombosis have less extensive clots in the legs, as the clots have already embolized to the lungs. When we seek to combine the results from the Chapters 6 and 7, we see a dose response relationship : patients with isolated deep-vein thrombosis had the most extensive thrombi, followed by patients with pulmonary embolism plus deep-vein thrombosis. The latter group had more severe obstruction of the pulmonary arteries than patients with isolated pulmonary embolism. This supports a different etiology for pulmonary embolism without peripheral deep-vein thrombosis possibly with a local origin. (see Figure) Deep-vein thrombosis Extensive thrombi regarding number of occluded veins and thrombus length Pulmonary embolism with deep-vein thrombosis Less extensive peripheral thrombi (however, dose response effect: with increasing thrombus length and number of occluded veins, more severe pulmonary obstruction) Extensive pulmonary obstruction Pulmonary embolism Less extensive obstruction of the pulmonary arteries Figure 1. Schematic representation of the results from chapter 6 and 7.
Summary and general discussion 127 Absence of peripheral thrombi in patients with pulmonary embolism In the PEDLAR study we found that in more than half of the patients with pulmonary embolism (55 out of 99 patients), no thrombi were present in the legs or abdomen. (Chapter 5) We postulated alternative hypotheses to explain the origin of pulmonary embolism without a peripheral thrombus. One of the hypotheses is that emboli originate from the right heart, e.g. in patients with atrial fibrillation. Patients with various heart diseases have been shown to be at higher risk of isolated pulmonary embolism than of deepvein thrombosis with or without pulmonary embolism. 196 Right-sided valvular disease gave a higher risk of isolated pulmonary embolism (OR 74.6; 95% CI 28.4-195.8) than left-sided valvular heart disease (OR 13.5; 95% CI 11.3-16.1), further strengthening the hypothesis of a cardiac origin of pulmonary embolism. Second, local thrombus formation in the lungs could be the case. This hypothesis was proposed previously in a study on trauma patients. 193 A local origin of pulmonary embolism is supported by the finding that patients with lung disease, like COPD or asthma, 188 and patients with pneumonia have a higher risk of pulmonary embolism alone, than of deep-vein thrombosis with or without pulmonary embolism. This difference may be attributed to inflammation that occurs locally in the pulmonary arteries, triggering coagulation. Inflammation affects coagulation via two mechanisms: first, via tissue factor mediated thrombin generation, and secondly, via the impairment of anticoagulant pathways, the protein C system in particular. 200 In addition, coagulation can be further enhanced by the complement system that is activated during inflammation. One might speculate on local inflammation of the vessel wall leading to thrombus formation. T h e r o l e o f s u p e r f i c i a l v e i n t h r o m b o s i s Superficial vein thrombosis was studied in two chapters of this thesis: in the PEDLAR study (Chapter 5) and in the MEGA study (Chapter 8). Superficial vein thrombosis is commonly regarded as a benign and self-limiting disease. Recently, that idea has been challenged. 223 In Chapter 5, we assessed the presence of peripheral thrombi in patients with pulmonary embolism. We found that in 14% of patients with a thrombus on MRI, superficial vein thrombosis was the only peripheral thrombus (6/44). It should be noted that we could not verify whether a thrombus in a more proximal deep vein had been present which had already embolized to the lungs and could not be identified anymore. This study showed an association rather than a causal relation between superficial vein thrombosis and pulmonary embolism. At the time of pulmonary embolism diagnosis, superficial vein thrombosis may be the only thrombus left to be detected by imaging techniques, whereas the embolism may have been caused by a proximal deep venous extension of the clot. Therefore, more reliable information on the frequency of progression from superficial veins to proximal deep veins is required.
128 Chapter 10 Results from the MEGA study showed that a history of clinical superficial vein thrombosis was associated with a 6-fold increased risk of deep-vein thrombosis and a 4-fold increased risk of pulmonary embolism. Blood group non-o and factor V Leiden hardly affected the risk of superficial vein thrombosis in controls. We concluded that clinically diagnosed superficial vein thrombosis was a moderate risk factor for venous thrombosis. Given that thrombophilia was only weakly associated with superficial vein thrombosis, it is likely that other factors (such as varicose veins, obesity, stasis) also play a role in its etiology. (Chapter 8) Regarding venous anatomy, deep and superficial veins are approached differently. A blood clot that is formed in the venous side of the circulation is a deep-vein thrombosis when formed in any of the deep veins, no matter their size or location in the body. Interestingly, if clot formation takes place in a vein that falls in the category of superficial veins, the disease process is regarded differently and often not treated as thrombosis in the deep venous system would have been. Even patients with thrombosis in a large superficial vein, such as the greater saphenous vein that starts from the medial malleolus and continues up to the groin where it joins the common femoral vein, are not treated with anticoagulants. Superficial veins contain venous valves, just like deep veins. As these valves are seen as the places where thrombus formation starts, it seems ambiguous to let the anatomical location of veins, whether positioned deep or superficially in the body, determine whether a blood clot is of importance and needs to be treated with anticoagulants or not. I m a g i n g o f v e n o u s t h r o m b o s i s In Chapter 3, we reviewed the literature on present imaging techniques for deep-vein thrombosis and pulmonary embolism. For pulmonary embolism, CT pulmonary angiography continues to be the preferred modality in most clinical scenarios. High resolution and velocity of image acquisition as well as a high percentage of technically adequate images offer significant benefits over MR angiography. Imaging of pulmonary embolism with MR molecular imaging such as fibrin labeling seems a promising technique. Diagnostic evaluation of deep-vein thrombosis is mainly done by ultrasonography, due to its high diagnostic yield and wide availability. CT venography can be useful in selected patient groups, e.g. in patients with morbid obesity, casts, and in case of a suspected thrombosis in the iliac or inferior caval vein. Radiation exposure is an important disadvantage of CT venography, compared with ultrasonography of the legs. In a population of patients with a suspected pulmonary embolism from the Leiden area, we showed that CT pulmonary angiography was positive in 19% of cases. (Chapter 9) This means that in the majority of patients, radiation and iodinated contrast were given while the diagnosis of pulmonary embolism was not made. However, we should take into consideration that a certain number of CT pulmonary angiographies is re-
Summary and general discussion 129 quested by clinicians to exclude the diagnosis, rather than to confirm it. In addition, other diagnoses can be made on CT pulmonary angiography that explain the clinical complaints. We pointed out three groups with a suboptimal CT yield: outpatients compared with clinical patients or emergency room patients, women, and patients below 40 years of age. As the sensitivity and specificity of CT pulmonary angiography have been shown to be similar for different strata of age and sex, 256 differences in patient referral could be an explanation for our findings. F u t u r e d i r e c t i o n s The results presented in this thesis concern etiological questions, i.e., they are related to the development and risk factors of pulmonary embolism and deep-vein thrombosis. Further studies could be valuable to confirm our results, e.g. in relation to superficial vein thrombosis as a risk factor for deep-vein thrombosis. Before changing therapeutic strategies and treating all patients who present with superficial vein thrombosis, we need to know the incidence of proximal extension of these superficial thrombi more precisely. Bleeding, the most important disadvantage of anticoagulant medication, calls for careful patient selection before starting treatment. Regarding alternative origins of pulmonary embolism, more studies could be done on diagnosing (right) cardiac thrombi. Fundamental research could focus on coagulation as a local process taking place in the pulmonary arteries, when inflammatory diseases such as COPD or asthma are present.