A common mutation, a GA transition at nucleotide position 20210 in the prothrombin gene locus, has recently been described [5]. The A allele is associated with higher plasma prothrombin levels, and in one study [5], carriers of this allele had a 2.8-fold increased risk for venous thrombosis compared with persons homozygous for the G allele.

We sought to determine the presence of this mutation in 281 patients with venous thrombosis and investigated whether the presence of the prothrombin A (²⁰²¹⁰) allele was an independent, inherited risk factor for venous thrombosis.

Methods

Top Methods Results Discussion Author & Article Info References

After approval by the local ethics committees, the study was conducted according to the principles of the Declaration of Helsinki. All participants provided informed consent.

Patients

Persons with documented venous thrombosis who had been referred to one of two thrombosis centers (the IRCCS Casa Sollievo della Sofferenza in San Giovanni Rotondo, Italy, and the Ospedale A. Cardarelli in Naples, Italy) were investigated. Deep venous thrombosis was confirmed by phlebography or ultrasonography. Pulmonary embolism was diagnosed by angiography or ventilation-perfusion lung scanning. By using a previously validated questionnaire based on the World Health Organization questionnaire for cardiovascular disease [6], specially trained staff obtained a complete clinical summary from all patients, with emphasis on personal history of circumstantial risk factors for venous thromboembolism (surgery, immobilization, pregnancy, postpartum, trauma, use of oral contraceptives, varicose veins, and cancer) and for arterial thrombosis.

Controls

While patients were being recruited, apparently healthy persons were also randomly selected from a healthy population in southern Italy. None of these persons had been exposed to circumstantial risk factors for 8 weeks before the visit or had a history of venous thromboembolism, as determined by using a structured questionnaire validated for the retrospective diagnosis of venous thrombosis [7]. As was done with patients, trained staff obtained a detailed clinical summary from all controls, with emphasis on personal and family history for angina pectoris, myocardial infarction, ischemic stroke, and peripheral arterial disease [6].

Blood Collection and Coagulation Tests

Blood samples were collected into vacuum tubes that contained 3.8% trisodium citrate as an anticoagulant. Samples were subjected to centrifugation at 2000 g for 15 minutes to obtain platelet-poor plasma, which was frozen and stored at –70°C until assays were performed. Antiphospholipid antibodies (lupus anticoagulant and IgG anticardiolipin antibodies [measured by enzyme-linked immunosorbent assay, Byk Gulden, Italy]); antithrombin III, protein C, and amidolytic and immunologic protein S antigen (Behering, Marburg, Germany); and total and free protein S antigen (measured by enzyme-linked immunosorbent assay [Diagnostica Stago, Asnieres, France]) were measured in all patients [8, 9]. A thromboplastin-based assay on factor II-deficient plasma (Diagnostica Stago) was performed to measure prothrombin activity in 157 controls (72 men and 85 women) who were indistinguishable from other controls in age, sex, and social status (occupation). Results were expressed as the percentage of the amount of prothrombin in pooled normal plasma (arbitrarily designated as 100%). Clotting assays were performed on a KC4 Amelung coagulometer (Amelung, Austria).

Extraction and Analysis of DNA

Standard protocols were used to extract DNA from peripheral blood leukocytes [6]. The presence of factor V Leiden mutation was detected as described elsewhere [10], with some modifications [11]. The presence of the GA mutation of the prothrombin gene was tested according to the method of Poort and colleagues [5].

Statistical Analysis

All analyses were done by using the Statistical Package for Social Science, version 6.1 for Macintosh (SPS, Inc., Chicago, Illinois). Differences in means and proportions were tested by using the Mann-Whitney U-test or the Kruskal-Wallis one-way analysis of variance and chi-square statistic or the Fisher exact test, as appropriate. We performed all analyses after grouping homozygous and heterozygous carriers of the prothrombin mutation and factor V Leiden mutation. Prevalence odds ratios (ORs), considered as the prevalence of existing disease, and 95% CIs were calculated by the normal approximation. If np(1 –p) was small (<5), we used the exact method to calculate CIs (n = total number of participants, p = proportion of participants with a specific condition, and 1 –p = proportion of participants without that condition). Adjusted odds ratios and their 95% CIs were calculated by logistic regression models that controlled for age, sex, presence of arterial thrombotic episodes, and factor V Leiden mutation. For all tests, a P value of 0.05 or less was considered statistically significant.

Results

Top Methods Results Discussion Author & Article Info References

Between May 1996 and December 1997, 348 patients (151 men and 197 women) with documented venous thrombosis were investigated. Sixty-seven patients (19.25%) had had at least one previous thrombotic events. Because these 67 patients are likely to have a greater susceptibility to thromboembolic episodes, we excluded them from the analysis. Thus, we analyzed 281 patients (128 men and 153 women; age range, 3 to 81 years). The median age at the time of the first thrombotic episode was 38.0 years (range, 16 to 81 years) for men and 35.0 years (range, 3 to 77 years) for women. The presenting thrombotic episode was deep venous thrombosis in one leg in 234 patients (106 men and 128 women; age range, 16 to 81 years), thrombosis of the upper extremities in 27 patients (13 men and 14 women; age range, 3 to 72 years), and isolated mesenteric vein thrombosis in 20 patients (age range, 17 to 70 years; 9 men and 11 women). Forty-five patients (17 men and 28 women) who had had deep venous thrombosis had also experienced an episode of pulmonary embolism.

The control group consisted of 850 randomly selected, apparently healthy persons (388 men and 462 women; median age, 36.0 years [range, 22 to 66 years]). All patients and controls were white and were from the same region. The clinical characteristics of patients and controls are shown in Table 1.

Forty patients (14.23% [95% CI, 10.14% to 18.32%]) carried the prothrombin GA²⁰²¹⁰ mutation; 35 were heterozygotes and 5 were homozygotes. Thirty-nine controls (4.59% [CI, 3.17% to 6.01%]) (P < 0.001) were heterozygotes, and none were homozygotes. The A²⁰²¹⁰ allele was noted in 8.01% of patients (CI, 5.77% to 10.25%) and 2.29% of controls (CI, 1.58% to 3.00%) (P < 0.001). The distribution of genotypes did not significantly differ from distributions predicted from the Hardy-Weinberg equilibrium in patients (P = 0.151) or controls (P > 0.2) and closely resembled those reported elsewhere [5]. Mean (±SD) prothrombin activity was 99.33% ± 16.30% in the 150 patients without the prothrombin A²⁰²¹⁰ allele and 125.20% ± 7.69% in the 7 controls with the allele (Mann-Whitney U-test, P < 0.001). Fifty-one patients carried the factor V Leiden mutation (18.15% [CI, 13.64% to 22.66%]); 50 were heterozygotes and 1 was a homozygote. Forty-three controls carried this mutation (5.06% [CI, 3.59% to 6.53%]; P < 0.001); all were heterozygotes.

Crude odds ratios for venous thrombosis associated with the presence of the prothrombin GA²⁰²¹⁰ or factor V Leiden mutation are reported in Table 1. We analyzed the association between the two gene variants and a history of venous thrombosis by stratifying patients and controls according to the presence of one or two gene variants (Table 2). The increased risk (odds ratio, 2.51) associated with the prothrombin GA²⁰²¹⁰ mutation was clearly unrelated to the overrepresentation of the factor V Leiden mutation; the coexistence of both mutations was seen only in patients (n = 14; 4.98% [CI, 2.44% to 7.52%]). The estimated risk associated with the prothrombin GA²⁰²¹⁰ mutation was similar when we excluded carriers of inherited defects of natural anticoagulants, patients with antiphospholipid antibodies, and patients exposed to circumstantial risk factors for venous thromboembolism. However, the risk for venous thrombosis associated with the prothrombin GA²⁰²¹⁰ mutation approached statistical significance only in the subset of patients with additional risk factors (Table 2). The distribution of the prothrombin GA²⁰²¹⁰ gene variant was similar in patients with and those without additional risk factors (P = 0.171). The prevalence of the factor V Leiden mutation did not differ between patients with and those without risk factors. The median age at the time of the first thrombotic episode was 38.0 years (range, 10 to 67 years) in the 26 patients with the prothrombin GA²⁰²¹⁰ mutation and 39.0 years (range, 16 to 65 years) in the 37 patients carrying the factor V Leiden mutation. The median age was 37.0 years (range, 3 to 81 years) in the 204 patients without either mutation and 31.5 years (range, 16 to 49 years) in the 14 patients carrying both mutations (Kruskal-Wallis test, P = 0.129).

View this table: [in this window] [in a new window]   Table 2. Risk Estimate in All Study Participants and after Stratification by Risk Factors for Venous Thrombosis according to Presence of Prothrombin GA (20210) or Factor V Leiden Mutation*

The estimated risk for having venous thrombosis when carrying the prothrombin GA²⁰²¹⁰ mutation was further analyzed in a logistic model that adjusted for age, sex, presence of arterial thrombotic episodes, and factor V Leiden mutation. The adjusted odds ratios were 3.13 (CI, 1.89 to 5.21) with the prothrombin GA²⁰²¹⁰ mutation and 3.38 (CI, 2.10 to 5.45) with the factor V Leiden mutation. A significant association was also noted between venous thrombosis and age (odds ratio per year, 1.02 [CI, 1.01 to 1.04]). A logistic regression model done by using an interaction term excluded any interaction between the prothrombin GA²⁰²¹⁰ mutation and the factor V Leiden mutation (P > 0.2). When the 27 patients who had thrombosis of the upper extremities were analyzed separately, none was found to exhibit inherited or acquired (that is, antiphospholipid antibodies) risk factors for venous thromboembolism. The prothrombin GA²⁰²¹⁰ mutation was found in 2 patients (7.41% [CI (exact method), 0.91% to 24.29%]), and the frequency of the carriers did not significantly differ from that in the control group (Fisher exact test, P > 0.2) or that in the 234 patients with deep venous thrombosis in one leg (14.96% [CI, 10.39% to 19.53%]; Fisher exact test, P > 0.2). Similarly, 2 patients with thrombosis of the upper extremities carried the factor V Leiden mutation (7.41% [CI (exact method), 0.91% to 24.29%]), and the frequency of mutation was similar to that of controls (Fisher exact test, P > 0.2) and to that of patients with thrombosis of the lower extremities (20.09% [CI, 14.96% to 25.22%]; Fisher exact test, P = 0.094). The overall prevalence of inherited coagulation abnormalities was significantly higher in the patients with thrombosis of the lower extremities (odds ratio, 3.77 [CI, 1.10 to 12.93]).

Discussion

Top Methods Results Discussion Author & Article Info References

Among the inherited factors of thrombophilia, the prothrombin GA²⁰²¹⁰ mutation has recently been emphasized because carriers of the A allele have higher plasma prothrombin levels and increased risk for venous thrombosis [5]. In agreement with the findings of that study, we found that the presence of the GA²⁰²¹⁰ gene variant was strongly associated with the occurrence of venous thrombosis; this association was not affected by the simultaneous presence of the factor V Leiden mutation. The distribution of prothrombin A²⁰²¹⁰ carriers differed slightly between patients with and those without inherited, acquired, or circumstantial risk factors. However, because the prothrombin GA²⁰²¹⁰ mutation was not differently distributed in the two groups of patients, differences in estimated risks may reflect the effect of subgroup analyses.

The adjusted prevalence odds ratio of venous thrombosis (3.13) was similar to that associated with the factor V Leiden mutation (3.38) in our study and to that estimated (2.8) by Poort and colleagues [5]. Reports that higher plasma levels of prothrombin are independently related to venous thrombosis and that carriers of the A²⁰²¹⁰ allele showed higher circulating prothrombin values give our findings a biological plausibility [5]. The significantly increased prothrombin activity that we observed in controls who carried the prothrombin GA²⁰²¹⁰ mutation among a subset of healthy controls further supports this hypothesis. However, the possibility that the prothrombin GA²⁰²¹⁰ polymorphism is a neutral marker closely linked to a functional variant of an undefined sequence cannot be ruled out on the basis of the available data.

A significant subset of thrombophilic patients had several genetic risk factors [10, 12, 13]. The factor V Leiden mutation has been found in up to 15% of white persons in general [14]. When we identified thrombotic events by using the criteria used in other studies [15-19], the results confirm the independent and significant association with the factor V Leiden mutation in our study. Poort and colleagues [5] observed a 2.3% prevalence of the prothrombin A²⁰²¹⁰ allele in controls; we found this gene variant in 4.59% of controls. These estimates are similar (given the ethnic difference between the two settings) and indicate that both mutations may eventually coexist. Moreover, the presence of these carriers only in the patient group is consistent with the concept that the two defects confer a higher risk for venous thrombosis than does either defect alone.

Depending on the extremities involved (upper or lower), the prevalence of inherited coagulation abnormalities, anticoagulant factors, and the factor V Leiden mutation has been reported as statistically significantly different [19]. Our study extends these data to the prothrombin A²⁰²¹⁰ allele and further suggests mechanistic differences in the pathogenesis of thrombotic events of the upper and lower extremities.

Our data confirm previous risk estimates associated with the prothrombin A²⁰²¹⁰ allele and provide strong evidence that this gene variant is a common, inherited, independent entity that predisposes patients to thrombosis, particularly patients with thrombosis of the lower extremities. Future studies should examine whether patients with more than one inherited risk factor require usual or more aggressive prophylaxis before undergoing surgical, medical, or obstetric procedures that carry an increased risk for thrombosis.

Dr. Brancaccio and Mr. Iannaccone: Divisione di Ematologia, Unita di Coagulazione, Ospedale A. Cardarelli, via A. Cardarelli, 9, 80131 Napoli, Italy.

Dr. Di Minno: Istituto di Medicina Interna e Geriatria, Universita di Palermo, viale delle Alpi, 86, 90144 Palermo, Italy.