Age-Specific Incidence Rates of Venous Thromboembolism among Heterozygous Carriers of Factor V Leiden Mutation

Methods

The Physicians' Health Study was a randomized trial of aspirin and β-carotene in 22 071 male physicians in the United States who did not have cardiovascular disease at baseline. These physicians were prospectively followed for the occurrence of vascular disease and cancer [9]. Whole-blood specimens were obtained at baseline from 14 916 of these men. The methods used to recover DNA from these samples and the polymerase chain reaction methods used to assess factor V Leiden mutation status have been described elsewhere [7, 8].

Between 1982 (the year that the study began) and August 1994, 156 venous thromboembolic events were confirmed by review of hospital records and autopsy reports. Deep venous thrombosis was confirmed by venography or ultrasonography. Pulmonary embolism was confirmed by angiography or perfusion scanning that showed at least two segmental defects without ventilation defects. Venous thromboses that were not associated with cancer, surgery, or trauma were defined as primary events.

At year 11 of follow-up, data on morbidity were obtained from 99.1% of participants and data on mortality were obtained from 99.9% of participants. For persons who had venous thromboembolism, exposure time was calculated as the time between study initiation and diagnosis. For persons who did not develop venous thromboembolism, exposure time was calculated as the time between study initiation and August 1994 (or death if this event occurred before August 1994).

Factor V Leiden mutation status was assessed in the 156 study participants who developed venous thromboembolism during study follow-up and in a sample of 2406 randomly selected study participants who remained free of vascular disease. The latter group was used to estimate the prevalence of factor V Leiden mutation for the cohort as a whole.

Venous thromboembolic incidence rates and incidence rate differences between men with and without the factor V Leiden mutation were computed after the cohort was divided into four age groups (40 to 49 years of age, 50 to 59 years of age, 60 to 69 years of age, and 70 years of age or older). Study exposure time was proportionally allocated to patients with and without the factor V Leiden mutation according to the observed prevalence of mutation in the cohort, appropriately aging the cohort over the course of follow-up. To account for the proportional allocation of study exposure time, the variance of an incidence rate was calculated to be the variance (assuming that this proportion was known) plus the variance of the incidence rate conditional on this proportion [10]. The significance of any trend across age groups for computed incidence rate differences was tested using a weighted regression, with weights equal to the inverse of the variance of each difference [11]. Data on incidence rate were further analyzed by using Poisson regression models. In these analyses, formal tests of interaction between age and mutation were of borderline significance; however, graphic display showed a better fit when an interaction term was introduced between age and the presence or absence of the factor V Leiden mutation. Subgroup analyses were done separately for the 77 venous thromboembolic events classified by the end points committee as primary and for the 78 events classified as secondary. One event, classified as indeterminate, was excluded from the subgroup analysis. All P values were two tailed, and all CIs were calculated at the 95% level.

Results

Venous thromboembolism was rare in patients in our cohort who were younger than 60 years of age. However, incidence rates increased markedly with older age in a pattern consistent with that previously reported in community-based studies of venous thromboembolism [12]. Among the 2406 study participants in whom vascular disease did not develop, the prevalence of the factor V Leiden mutation was 5.2%.

Incidence rates and incidence rate differences for any type of venous thromboembolism among study participants with or without factor V Leiden, stratified by age, are shown in Table 1. Incidence rates of venous thromboembolism were very low for men younger than 50 years of age, and the incidence rate difference between men in this age group with and without the factor V Leiden mutation was not significant (P > 0.2). However, incidence rate differences per 1000 person-years of observation between patients with and without the factor V Leiden mutation were 1.23 for men 50 to 59 years of age, 1.61 for men 60 to 69 years of age, and 5.97 for men 70 years of age or older (P for trend across age groups = 0.008). The incidence rate among men with factor V Leiden mutation who were 70 years of age or older was significantly greater than that among men in the same age group who did not have this condition (7.83 events compared with 1.86 events per 1000 person-years; P = 0.028).

Table 1 also shows incidence rates and incidence rate differences for primary venous thromboembolism among study participants with and without the factor V Leiden mutation. In this subgroup, a significant increase in incidence rate differences was found across age groups (P for trend = 0.017), and the incidence rate for men with the mutation who were 70 years of age or older (5.22 per 1000 person-years) was significantly greater than that of men in the same age group without the mutation (0.76 per 1000 person-years) (P = 0.04). In contrast, no evidence showed that the factor V Leiden mutation was associated with an increase in rates of secondary venous thromboembolism, either overall or in any age category. Poisson regression analyses confirmed these observations and are shown in Figure 1.

View larger version:

• In this window
• In a new window

Figure 1. Any venous thromboembolism. Idiopathic venous thromboembolism. Venous thromboembolism associated with cancer or surgery. Estimated age-specific incidence rates for venous thromboembolism among men with (dashed lines) and men without (solid lines) factor V Leiden mutation.Left.Middle.Right.

Discussion

These prospective data indicate that incidence rates of venous thromboembolism in heterozygous male carriers of the factor V Leiden mutation increase with age at a rate significantly greater than that in men who do not have this mutation. The difference in incidence rates of venous thromboembolism between men with and without the factor V Leiden mutation was greatest in men older than 70 years of age, whereas incidence rates in men younger than 50 years of age with and without the factor V Leiden mutation were almost identical. This trend was most pronounced for venous thromboembolism that was not associated with cancer, surgery, or trauma.

Our finding that the incidence rate of venous thromboembolism increases with age among persons with the factor V Leiden mutation may appear counterintuitive because this prothrombotic defect is genetically determined and previous reports suggest that thrombosis in persons with the factor V Leiden mutation tends to occur at younger ages. However, our data were not influenced by such factors as pregnancy and use of oral contraceptives, both of which are independently associated with high risk for venous thromboembolism in young patients [5]. In contrast to retrospective and cross-sectional studies, which may be subject to inadvertent diagnostic, referral, and treatment biases (all of which are likely to result in overrepresentation of young patients [6]), the prospective design of the Physicians' Health Study ensures that genetic testing is done independently of case accrual. Although our study did not include patients younger than 40 years of age, previously obtained population-based data suggest that venous thromboembolism from any cause is very rare in men in this age range [12].

Our finding-that the risk for venous thromboembolism among heterozygous carriers of the factor V Leiden mutation increases with age at a rate significantly greater than that in men without this mutation-strongly suggests that the pathogenesis of venous thromboembolism is multifactorial and requires interactions between both inherited and acquired risk factors. Our data are therefore consistent with those of studies suggesting that heterozygosity for the factor V Leiden mutation alone may be a relatively weak risk factor unless a concomitant second genetic risk factor is present [13] or an acquired factor, such as older age or hyperhomocystinemia [14], also exists. Although homozygosity for the factor V Leiden mutation confers higher risk, this condition is rare.

Our data are likely to become clinically important as the availability of testing for the factor V Leiden mutation increases. In contrast to findings for many other genetic risk factors, our data suggest that the age-specific risk for thromboembolism attributable to the factor V Leiden mutation increases with age, even though the mutation is present from birth. Thus, these data indicate that the decision about whether to determine factor V Leiden mutation status in patients with a first venous thromboembolism should not be limited to young patients. In fact, our data suggest that an approach to screening for factor V Leiden mutation that excludes older persons may overlook many patients who are at high risk for recurrent thromboembolic events.

From Brigham and Womens' Hospital, Harvard Medical School, and Harvard School of Public Health, Boston, Massachusetts; and Washington University School of Medicine, St. Louis, Missouri.

Drs. Glynn, Stampfer, and Hennekens: Department of Medicine, Brigham and Womens' Hospital, 900 Commonwealth Avenue East, Boston, MA 02115.

Dr. Miletich: Division of Laboratory Medicine, Washington University, 660 South Euclid Avenue, St. Louis, MO 63110.