Type and location of venous thromboembolism in patients with factor V Leiden or prothrombin G20210A and in those with no thrombophilia
Ida Martinelli, Hemophilia and Thrombosis Center, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, University of Milan, Via Pace 9, 20122 Milan, Italy.
Tel.: +39 02 55035468; fax: +39 02 50320723; e-mail:firstname.lastname@example.org
Summary. Background: Patients with factor (F) V Leiden or the prothrombin G20210A polymorphism are at increased risk of developing deep vein thrombosis (DVT). On the other hand, the risk of developing pulmonary embolism (PE) appears to be low in carriers of FV Leiden, perhaps because of a lower tendency to develop iliofemoral DVT than non-carriers. For prothrombin G20210A, data are scanty and controversial. Methods: The clinical manifestations (isolated DVT, DVT and PE, and isolated PE), the extension of DVT, and the presence of transient risk factors were retrospectively investigated in 115 patients with heterozygous FV Leiden, 87 with prothrombin G20210A and 200 with no thrombophilia marker. Results: Isolated symptomatic PE was less prevalent in patients with FV Leiden (6%) than in those with prothrombin G20210A (21%) and no thrombophilia (23%) (P > 0.0001). The rate of distal DVT was higher in patients with no thrombophilia (16% vs. 7% for FV Leiden and 6% for prothrombin G20210A) (P = 0.02). No difference in the incidence of PE from distal and proximal DVT, the extension of proximal DVT and the type of transient risk factors for venous thromboembolism (VTE) was found in the three groups. Patients with prothrombin G20210A had a younger age at their first VTE (24 years, P < 0.0001) and a higher rate of DVT accompanying PE (P = 0.04) than those with FV Leiden or no thrombophilia. Conclusions: Carriers of prothrombin G20210A, unlike those of FV Leiden, have an increased risk of developing isolated PE. This difference was not explained by a different rate of distal DVT, extension of proximal DVT, or distribution of transient risk factors in the two groups. Patients with prothrombin G20210A have more severe clinical manifestations than those with FV Leiden or no thrombophilia.
The guanine to adenine mutation at position 1691 of coagulation factor (F) V gene (factor V Leiden) and the guanine to adenine mutation at position 20210 of the 3′-untranslated region of the prothrombin gene (prothrombin G20210A) are the most common genetic causes of thrombophilia predisposing to venous thromboembolism (VTE) . The former mutation renders activated FV partially resistant to the inactivation by its naturally occurring anticoagulant protein C , and the latter leads to an approximately 30% increased concentration of prothrombin in plasma . The hypercoagulable state associated with these two gain-of-function mutations leads to different clinical manifestations of VTE. For instance, prothrombin G20210A is associated with an increased risk of portal vein thrombosis whereas FV Leiden is not . Several studies consistently showed a lower incidence of FV Leiden in patients with symptomatic isolated pulmonary embolism (PE), for example that occurring without a concomitant deep vein thrombosis (DVT), than in patients with DVT with or without PE [(5–14]. As a possible explanation of the weak association between FV Leiden and isolated PE, the observation was offered that carriers of the mutation develop iliofemoral DVT less frequently than non-carriers , supporting the much stronger association between proximal than distal (i.e. in the veins of the calf) DVT and PE [15–17]. However, this finding is controversial, because similar locations of DVT in carriers and non-carriers of FV Leiden have been reported . Previous studies evaluating the relationship between FV Leiden and the extension of DVT have included among ‘non-carriers’ patients that may carry prothrombin G20210A or other causes of thrombophilia [5–9]. Pertaining to prothrombin G20210A, data on its incidence in patients with isolated PE and PE associated with DVT are more scanty and controversial, as a lower [12,14], similar  or higher  incidence was reported in isolated PE.
To investigate whether the clinical manifestations of VTE (isolated PE, PE with DVT, and isolated DVT), as well as the extension of DVT and the presence of transient risk factors at the time of VTE, were different among patients with FV Leiden, prothrombin G20210A or no evident thrombophilia, a large retrospective cohort of patients was investigated.
Nine hundred and sixteen patients were consecutively referred to our Thrombosis Center from June 1996 to December 2004 to be investigated for thrombophilia screening after a first episode of symptomatic VTE (i.e. DVT of the lower limbs with or without symptomatic PE and isolated PE). Five hundred and fourteen patients were excluded from the study because they had antithrombin, protein C, protein S deficiency, homozygous FV Leiden or prothrombin G20210A, antiphospholipid antibodies or combined abnormalities (n = 308) or because diagnosis of VTE was not objectively confirmed (n = 206). Hence, 402 patients were considered for the study. DVT was objectively confirmed by bilateral B-mode compression ultrasound (n = 286) or contrast venography (n = 44) and PE by (ventilation)/perfusion lung scan (n = 83), computed tomography (n = 56) or angiography (n = 2). Patients presenting with symptomatic DVT underwent objective investigation for PE only if they had pulmonary symptoms, whereas patients presenting with symptomatic PE underwent objective investigation (compression ultrasound) for lower-limb DVT independently on the presence of specific symptoms. None of the 72 patients with isolated PE reported symptoms of DVT or superficial vein thrombosis. PE/DVT was considered isolated when no site of DVT/PE could be objectively detected. DVT of the veins of the calf were considered distal and those involving the remaining vein segments (popliteal, superficial femoral, common femoral, external iliac, internal iliac, common iliac, or inferior cava) were considered proximal.
The presence of transient risk factors in the month preceding thrombosis was recorded; these were surgery, trauma, leg cast, prolonged immobilization (> 10 days), oral contraceptive or hormone replacement therapy use, pregnancy and puerperium. In the absence of the aforementioned conditions, thrombosis was considered unprovoked.
All patients were of Caucasian descent and free from overt autoimmune or neoplastic diseases. All of them gave their written informed consent and the study was approved by the Institutional Review Boards.
DNA analyses for the 1691 G > A substitution in FV and for the 20210 G > A substitution of the prothrombin genes were carried according to previously described methods [2,3]. Plasma functional and (when these were abnormal) antigen levels of antithrombin, protein C and protein S; antiphospholipid antibodies (lupus anticoagulant and anticardiolipin antibodies) were tested as previously described [19,20].
The non-parametric Kruskal–Wallis test was used to detect group differences in continuous variables. The chi-squared test or Fisher's exact test were used to detect group differences in the distribution of categorical variables (e.g. type of event, risk factors, and male to female ratio). All statistical tests were two-sided with an α level of 0.05. The analyses were performed using the SPSS package version 14.0 (SPSS Inc., Chicago, IL, USA).
Of the 402 patients with VTE, 330 (72%) had had DVT [which in 69 (21%) was complicated by symptomatic PE] and 72 (18%) had isolated PE. The median age at thrombosis was 38 years (range 6–79 years). In 182 patients (45%) VTE was unprovoked. One hundred and fifteen patients had FV Leiden, 87 prothrombin G20210A and 200 no thrombophilia. Of the 330 patients who had DVT with or without PE, 108 (33%) had FV Leiden, 69 (21%) prothrombin G20210A and 153 (46%) no thrombophilia (P < 0.001), and of the 72 who had isolated PE, 7 (10%) had FV Leiden, 18 (25%) prothrombin G20210A and 47 (65%) no thrombophilia (P < 0.001). The rate of bilateral DVT was higher in patients with prothrombin G20210A (13%) than in those with FV Leiden (6%) and no thrombophilia (5%) (P = 0.07), but the rate of PE as a complication of DVT was similar in patients with unilateral and bilateral DVT in the three groups.
Table 1 shows the general characteristics, type of VTE and presence of transient risk factors at the time of thrombosis according to the presence or absence of thrombophilia. Patients with FV Leiden or prothrombin G20210A had had VTE at a significantly younger median age (33 and 24 years, respectively) than those with no thrombophilia (43 years) (P < 0.0001). No difference in the incidence of transient risk factors at the time of VTE was observed among groups of patients with or without thrombophilia. Among patients with DVT, those with prothrombin G20210A had a higher rate of accompanying PE (32%) than those with FV Leiden (19%) or without thrombophilia (17%) (P = 0.04).
Table 1. General characteristics, type of venous thromboembolism (VTE) and transient risk factors at the time of VTE in the study population
|M/F|| 55/60||41/46|| 84/116||0.5|
|Median age (range) at thrombosis (years)|| 33 (13–79)||24 (16–77)|| 43 (6–78)||< 0.0001|
|Isolated DVT, n (%)|| 90 (78)||47 (54)||124 (62)||< 0.0001|
|DVT and PE, n (%)|| 18 (16)||22 (25)|| 29 (15)|| |
|Isolated PE, n (%)|| 7 (6)||18 (21)|| 47 (23)|| |
|Unprovoked thrombosis, n (%)|| 44 (38)||45 (39)|| 99 (45)|| |
|Provoked thrombosis, n (%)|| 71 (62)||48 (55)||101 (51)||0.5|
| Surgery|| 21 (30)||14 (29)|| 29 (29)||1.0|
| Trauma/prolonged immobilization|| 20 (28)||13 (27)|| 25 (25)||0.9|
| Oral contraceptives/HRT*|| 39 (78)||28 (80)|| 49 (63)||0.1|
| Pregnancy/puerperium*|| 7 (14)|| 3 (9)|| 13 (17)||0.5|
Table 2 shows the extension of DVT in patients with or without thrombophilia. Patients with either FV Leiden or prothrombin G20210A had a significantly lower incidence of distal DVT than those without thrombophilia (P = 0.02).
Table 2. Incidence and extension of distal and proximal deep vein thrombosis (DVT) in patients with and without thrombophilia
|Distal (of the calf) DVT, n (%)||8 (7)||4 (6)||25 (16)||0.02|
|Proximal DVT, n (%)||100 (93)||65 (94)||128 (84)|
| Isolated inferior cava||8 (7)||5 (7)||3 (2)|| |
| Isolated iliac vein||2 (2)||1 (1)||1 (1)|| |
| Iliofemoral veins||12 (11)||12 (17)||12 (8)|| |
| Isolated femoral vein||4 (4)||5 (7)||6 (4)|| |
| Popliteal-iliofemoral veins||11 (10)||7 (10)||21 (14)|| |
| Popliteal-femoral veins||29 (27)||16 (23)||32 (21)|| |
| Isolated popliteal vein||34 (32)||19 (28)||53 (35)|| |
The rate of DVT complicated by PE was similar in patients with distal and proximal location of DVT (14% and 22%, P = 0.2), and also when considering the three groups of carriers of FV Leiden, prothrombin G20210A and no thrombophilia separately (data not shown). Four of the five patients with distal DVT complicated by PE had no thrombophilia. When the median number (range) of vein segments was considered, there was no statistically significant difference between patients with, 2 (1–4), or without, 2 (1–7), PE (P = 0.7), and between patients with prothrombin G20210A, 2 (1–7), FV Leiden, 2 (1–5), or no thrombophilia, 1 (1–5) (P = 0.13).
Although VTE has a wide spectrum of clinical manifestations and different risk factors may play a different role in determining the various locations of thrombi, DVT and its major complication, PE, are usually considered as a single disease entity . In recent years, a growing body of evidence showed that FV Leiden, which is the most common genetic risk factor for VTE, was strongly associated with an increased risk of DVT but not of PE [5–14]. The incidence of this mutation appears highest in patients with isolated DVT, intermediate in those with DVT associated with PE, and lowest in those with isolated PE. The low incidence of FV Leiden in patients with PE has been strengthened by the same post-mortem findings in patients with fatal PE . Whether or not the so-called ‘FV Leiden paradox’  is actually true, is at present a matter of debate . On the one hand several studies have consistently confirmed this issue [5–14], and on the other hand it cannot be ruled out that diagnostic, suspicion and selection bias affected many of those studies . Pertaining to the role in DVT and PE of the second most common genetic abnormality predisposing to VTE, the prothrombin G20210A mutation, data are more scanty and controversial; two studies showed similar results as for FV Leiden [12–14], but others did not [10,13].
In a previous study carried out in a small series of patients with isolated PE , we were among those who corroborated the ‘FV Leiden paradox’. In order to address this issue in a larger series of new patients and to investigate the role of prothrombin G20210A and transient risk factors for VTE, we compared the type and extension of VTE, and the presence of transient risk factors for VTE in three groups of patients with either mutation or no evident thrombophilia. This study shows that isolated PE is indeed less frequent in carriers of FV Leiden, but not in carriers of prothrombin G20210A. We also found that PE complicated a lower proportion of distal than proximal DVT, mainly in patients without thrombophilia, but the difference was not statistically significant. Although this finding may support the hypothesis that thrombophilia leads to more stable clots, which are less likely to detach , the rate of PE from proximal DVT was similar in patients with or without thrombophilia and was slightly higher in carriers of prothrombin G20210A, who also more often developed bilateral DVT. Distal DVT, although mainly searched for with B-mode compression ultrasound, which is not a very sensitive method for detecting DVT in such sites , was observed more rarely in patients with thrombophilia than in those without, and, finally, the extension of proximal DVT evaluated as number of involved venous segments was similar in the three groups of patients. Hence, our findings are against the hypothesis that carriers of FV Leiden have a tendency to form small thrombi that adhere to the venous wall and have a low potential to embolize, or to develop mainly distal DVT, and the same can be said for carriers of prothrombin G20210A.
Among the limitations of this study there is the highly selected population of young patients who survived after VTE (patients who died from massive PE were obviously not included in this study) and were referred for a thrombophilia screening, which renders the results not necessarily applicable to thrombosis patients in general. In addition, the study has some methodological flaws. In patients with DVT, PE was objectively searched for only if symptoms were present, and in those with isolated PE, DVT was ruled out only in the legs and mainly with compression ultrasound, whose accuracy in the diagnosis of distal and asymptomatic proximal DVT is lower than in that of symptomatic proximal DVT . However, a misclassification of patients with DVT and PE in the group of isolated PE would have lead to an overestimation of the incidence of thrombophilia, and there is no reason to think that it might have happened only in one of the three groups of patients with or without thrombophilia.
In conclusion, carriers of prothrombin G20210A develop PE either in isolation or as a complication of DVT in 21% and 25% of cases, respectively, whereas the corresponding rates in carriers of FV Leiden were only 6% and 16%. These differences were not explained by a different rate of distal DVT, a different extension of proximal DVT or a different distribution of transient risk factors for VTE between patients with one or the other mutation. To us, the ‘FV Leiden paradox’ remains unexplained and peculiar to this mutation. In terms of age at occurrence of first VTE and type of VTE (PE and bilateral DVT), prothrombin G20210A appears to be associated with more severe clinical manifestations than FV Leiden or no thrombophilia. The potential clinical implications of our findings in terms of tailored duration of anticoagulant therapy require further investigations.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.