- Top of page
- Patients and methods
Summary. The duration of anticoagulant treatment after a first episode of venous thromboembolism primarily depends on the risk of recurrence. Variability of recurrence rates in factor (F) V Leiden carriers may be due to concomitant thrombophilic disorders. A retrospective study was performed in 329 FV Leiden carriers with a history of venous thromboembolism (262 probands, 67 relatives). The annual rate of first recurrence was estimated in relatives. The contribution of concomitant thrombophilic disorders to the recurrence rate was evaluated in probands and relatives by a nested case–control analysis in 105 matched pairs of carriers either with or without recurrence. The overall annual recurrence rate was 2·3 per 100 patient–years. The adjusted risk of recurrence for concomitant thrombophilic disorders was: 9·1 (1·3–62·8) for the FII mutation; 1·0 (0·2–4·9) for homozygosity for FV Leiden; 1·5 (0·2–9·5) for inherited deficiencies of protein C or S; 1·8 (0·7–4·9) for FVIII coagulant activity (FVIII:C) levels > 122%; 5·4 (1·6–18·6) for fasting homocysteine levels > 15·2 µmol/l; and 4·4 (1·0–18·7) for loading homocysteine levels > 45·8 µmol/l. Of these disorders, only the FII mutation and hyperhomocysteinaemia significantly increased the risk of recurrence in FV Leiden carriers. The estimated recurrence rate ranged from 0·45 per 100 patient–years after a secondary first event in the absence of concomitant disorders to 4·8 per 100 patient–years when a spontaneous first event was combined with concomitant disorders. Our study provides supportive evidence that the incidence of recurrent venous thromboembolism in heterozygous FV Leiden carriers depends on the concomitance of other thrombophilic disorders, in addition to whether the first thrombotic event occurred spontaneously.
The recurrence rate of venous thromboembolism ranges from 10% to 18% within 2 years after the first episode and increases to 30% after 8–12 years (Schulman et al, 1995; Franzeck et al, 1996; Prandoni et al, 1996). It is higher in patients with persistent risk factors than in patients who are exposed to transient risk factors for venous thromboembolism (Schulman et al, 1995; Prandoni et al, 1996). It remains to be determined whether it is worthwhile to identify thrombophilic disorders in order to adjust the duration of secondary thromboprophylaxis.
The factor (F) V:Q506 mutation or FV Leiden is the most common genetic defect associated with venous thromboembolism. Its prevalence in Caucasians is approximately 5%, while it has been demonstrated in 20–50% of patients with venous thromboembolism (Koster et al, 1993; Bertina et al, 1994). Compared with non-carriers, heterozygous carriers of this mutation showed a three- to sevenfold higher risk of venous thromboembolism; in homozygotes this risk was considered to be as much as 80-fold higher (Koster et al, 1993; Ridker et al, 1995a; Rosendaal et al, 1995; Middeldorp et al, 1998). Whether FV Leiden influences the recurrence rate is still controversial. Of five prospective studies on this subject, one reported a twofold and one a fourfold increased risk of recurrence (Ridker et al, 1995b; Simioni et al, 1997), while the remaining three studies did not show such an association (Eichinger et al, 1997; Kearon et al, 1999; Lindmarker et al, 1999). These discrepant findings might partly be due to differences in the distribution of concomitant thrombophilic defects between the studies, as these were not consistently evaluated. We performed a study to assess the contribution of concomitant thrombotic risk factors, either thrombophilic disorders or exogenous conditions, to the recurrence of venous thromboembolism in a large cohort of FV Leiden carriers.
- Top of page
- Patients and methods
Of 793 FV Leiden carriers who enrolled the original study, 337 had a history of venous thromboembolism. Eight patients (2%) were excluded because their first thrombotic event was located at an unusual site: retinal vein (four), caval vein (two), mesenteric vein (one) and cerebral sinus (one). The remaining 329 carriers (262 probands and 67 relatives) were analysed in the present study; of these, 115 (35%) had recurrent deep vein thrombosis or pulmonary embolism. Their characteristics are summarized in Table I. Probands and relatives showed a similar distribution of characteristics among groups with or without recurrences. Of 329 first thrombotic events, 80% presented as deep vein thrombosis and 45% occurred spontaneously. A majority (72%) of these events was diagnosed by objective tests. In women with recurrent venous thromboembolism, 21% of the first episodes were associated with the use of oral contraceptives compared with 47% in women without recurrences. An association of the first episode with pregnancy or puerperium was observed in 32% and 15% of women respectively. There were no differences in duration of anticoagulant therapy after the first event. Of 115 first recurrences, 71% occurred spontaneously and 83% were located in the limbs. Of these recurrences, 70% were demonstrated by objective tests. Of 80 patients in whom both first and recurrent events were located in the limbs, 44% had a contralateral recurrence. In one case, initial leg vein thrombosis recurred in the caval vein. The median (range) interval between the first event and recurrence was 4·2 (0·1–35·1) years in probands and 9·3 (0·4–42·8) in relatives (P = 0·01).
Table I. Characteristics of 329 FV Leiden carriers with a history of venous thromboembolism.
|Recurrence (n = 91)||No recurrence (n = 171)||Recurrence (n = 24)||No recurrence (n = 43)|
|Median age at enrolment (range), years||49 (19–81)||36 (8–88)||< 0·001||57 (31–80)||53 (20–79)||0·42|
| || || || || || |
|Median age at first event (range), years||36 (12–77)||31 (8–88)||0·27||31 (16–63)||38 (15–79)||0·02|
| || || || || || |
| Deep vein thrombosis, %||81||77||0·53||83||83||1·0|
| Surgery, trauma||20||15||0·39||21||19||1·0|
| or immobilization|| || || || || || |
| puerperium|| || || || || || |
| Oral contraceptives||11||36||< 0·001||13||19||0·73|
|Anticoagulant therapy, years||0·28 (0–9)||0·33 (0–18)||0·22||0·50 (0–32)||0·26 (0–10)||0·39|
The annual recurrence rate in relatives was 2·3 per 100 patient–years (24 first recurrences/1062 observation years). The recurrence-free survival is shown in Fig 1. The 1-, 5-, 10- and 15-year cumulative recurrence rates were 5%, 10%, 28% and 38% respectively.
The influence of concomitant thrombophilic disorders on the recurrence rate was analysed in 105 pairs of cases (FV Leiden carriers with recurrences) and controls (carriers without recurrences) (Table II). For 10 out of 115 FV Leiden carriers with recurrences no control was available that fulfilled all matching criteria. Any concomitant thrombophilic disorder was observed in 75% of cases compared with 63% of controls (P = 0·05). Of these, heterozygous FII mutation, homozygosity for FV Leiden and inherited deficiencies of protein C or protein S (antithrombin deficiency was not observed) were demonstrated approximately twice as often in cases as in controls. The difference was statistically significant only for the combination of FV Leiden and FII mutation (P = 0·0002). Elevated FVIII:C levels (> 150%) were found in 55% of cases compared with 40% of controls (P = 0·06). Median levels were 158% compared with 140% (P = 0·008). Hyperhomocysteinaemia (fasting level > 18·5 µmol/l and/or loading level > 58·8 µmol/l) was more common in cases than in controls (19% versus 12%, P = 0·22). Of the cases, 11% had elevated fasting levels, 15% had elevated loading levels, and in 7% both levels were elevated, compared with 7%, 9% and 3% of controls respectively. Median homocysteine levels were higher in cases than in controls: fasting 12 versus 11 µmol/l (P = 0·008) and after loading 40 versus 36 µmol/l (P = 0·09). The results of the multivariate analysis are shown in Table III. Of the established genetic defects, the FII mutation was revealed as an independent contributor to recurrent venous thromboembolism in FV Leiden carriers (adjusted OR 9·1; 95% CI 1·3–62·8), in contrast to a homozygous FV Leiden carrier state (1·0; 0·2–4·9) and deficiencies of protein C or protein S (1·5; 0·2–9·5). The risk of recurrence was higher at FVIII:C levels between 122% and 144% (statistically not significant), but did not increase further at levels that were higher still. Overall, the risk of recurrence at FVIII:C levels > 122% was approximately twofold higher than that at lower levels (adjusted OR 1·8; 95% CI 0·7–4·9). The homocysteine-related risk showed an increase at fasting levels > 11·2 µmol/l (5·4; 1·6–18·6 for levels > 15·2 µmol/l). A significant increase was demonstrated at loading levels > 45·8 µmol/l (4·4; 1·0–18·7).
Table II. Results of univariate analysis of concomitant thrombophilic disorders in 105 matched pairs of heterozygous FV Leiden carriers either with (cases) or without (controls) recurrent venous thromboembolism.
|Variable||Cases (n = 105)||Controls (n = 105)||P|
|Women, n (%)||61 (58)||62 (59)||1·0|
|Age at enrolment, median (range), years||48 (19–80)||47 (16–81)||0·74|
|Observation time, median (range), years||4 (0–42)||4 (1–47)||0·56|
|Spontaneous first thrombotic event, n (%)||52 (50)||49 (47)||0·78|
|Concomitant thrombophilic disorders, n (%)|
| FII mutation||10 (10)||4 (4)||0·0002|
| Homozygosity for FV Leiden||9 (9)||5 (5)||0·41|
| Deficiencies of antithrombin, protein C or protein S*||6 (6)||3 (3)||0·50|
| FVIII:C > 150%||48 (55)||37 (40)||0·06|
| Hyperhomocysteinaemia†||16 (19)||11 (12)||0·22|
Table III. Influence of concomitant thrombophilic disorders on the risk of recurrent venous thromboembolism in heterozygous FV Leiden carriers.
|Variable|| || ||Adjusted odds ratio* (95% CI)|
|FII mutation|| || ||9·1 (1·3–62·8)|
|Homozygosity for factor Leiden|| || ||1·0 (0·2–4·9)|
|Deficiencies of protein C or protein S|| || ||1·5 (0·2–9·5)|
| 1st quartile||< 122|| ||1 (reference)|
| 2nd quartile||122–144|| ||2·0 (0·6–6·6)|
| 3rd quartile||145–179|| ||1·6 (0·5–5·4)|
| 4th quartile||> 179|| ||1·8 (0·5–6·3)|
|Homocysteine, µmol/l||Fasting level||Loading level||Fasting level|
| 1st quartile||< 9·4||< 28·8||1 (reference)|
| 2nd quartile||9·5–11·2||29–36·5||1·0 (0·3–3·5)|
| 3rd quartile||11·3–15·2||37–45·8||3·8 (1·1–13·0)|
| 4th quartile||> 15·2||> 45·8||5·4 (1·6–18·6)|
To estimate the effect of interactions between endogenous and exogenous thrombotic risk factors on the recurrence rate, we stratified cases and controls according to the presence or absence of concomitant thrombotic risk factors (Table IV). In this analysis, fasting homocysteine levels > 11·2 µmol/l were considered to be a risk factor for recurrence, as demonstrated by multivariate analysis. The lowest recurrence rate was observed in heterozygous FV Leiden carriers who had a secondary first event without concomitant thrombophilic disorders. The risk of recurrence was approximately fourfold higher after either a spontaneous first event without a concomitant thrombophilic disorder or a secondary first event in the presence of any of these disorders. It was almost 11-fold higher when a spontaneous first event was associated with a concomitant disorder. Interactions between endogenous and exogenous risk factors could not be attributed to the FII mutation exclusively. Estimated annual recurrence rates ranged from 0·45 to 4·8 per 100 patient–years.
Table IV. Effects of interactions between endogenous and exogenous thrombotic risk factors on the rates of first recurrence of venous thromboembolism in heterozygous carriers of FV Leiden.
|Concomitant thrombophilic disorder*||Spontaneous first event||Adjusted odds ratio† (95% CI)||Annual recurrence rate (per 100 patient–years)|
- Top of page
- Patients and methods
This study shows that concomitance of other thrombophilic disorders and a spontaneous first episode of venous thromboembolism are determinants of the risk of recurrence in heterozygous carriers of FV Leiden. The estimated annual recurrence rate ranged from 0·45 per 100 patient–years in heterozygous FV Leiden carriers who experienced a secondary first thromboembolic episode in the absence of other thrombophilic disorders to 4·8 per 100 patient–years in those with a spontaneous first event in combination with a concomitant disorder.
The FII mutation increased the risk of recurrence ninefold. Considering that previous studies did not demonstrate an increased risk of recurrence in single carriers of the FII mutation (Eichinger et al, 1999; Kearon et al, 1999; Lindmarker et al, 1999), the genetic interaction between FV Leiden and the FII mutation apparently results in a synergistically increased risk. Our finding is in agreement with two recent reports that showed a two- to threefold higher risk of recurrence in double-heterozygous carriers of FV Leiden and the FII mutation than in single carriers of FV Leiden (De Stefano et al, 1999; Margaglione et al, 1999). Of established genetic defects, deficiencies of protein C or protein S and homozygosity for FV Leiden were not identified as independent predictors of recurrence in our study, although these were found more frequently in the recurrence group. This finding is not plausible and can be explained by the small numbers of patients with these combined defects. Elevated levels of FVIII:C and hyperhomocysteinaemia have been recognized more recently as possible inherited risk factors for (recurrent) venous thromboembolism (Den Heijer et al, 1995, 1996; Koster et al, 1995; Eichinger et al, 1998; Kraaijenhagen et al, 2000). Both disorders were frequently observed in our patients and did increase the risk of recurrence. Actually, the effects of FVIII:C and homocysteine were already demonstrated at levels within the defined normal ranges, suggesting an interaction with FV Leiden. Levels of FVIII:C > 122% and fasting homocysteine > 11·2 µmol/l were associated with an approximately two- and fivefold increased risk of recurrence respectively. Although these risk estimates are modest and not statistically significant for FVIII:C, it should be noted that these levels were found in 75% and 50% of our patients, respectively, and may consequently account for a substantial proportion of recurrences in FV Leiden carriers. In the absence of any concomitant thrombophilic disorder, the risk of recurrence was still four times higher in FV Leiden carriers who experienced a spontaneous first thrombotic event than in those who experienced a secondary first event. This finding suggests the presence of other, thus far unrecognized, thrombophilic disorder(s). On the other hand, the lower risk of recurrence in patients with a secondary first event may be partly attributed to discontinuation of oral contraceptives after the first event and thromboprophylaxis at renewed exposure to exogenous risk factors. It might explain why women had fewer recurrences than men, as well as the large proportion of spontaneous recurrences.
The survival curve suggests a steadily continuing risk of recurrence over a period of 35 years. Beyond a period of 8 years, this finding may be flawed because fewer carriers without recurrences had a follow-up longer than 8 years. In addition, the diagnosis of the first event in patients with a recurrence after more than 20 years was made before the introduction of objective tests, and hence may be less reliable than in patients who had their first event fewer than 20 years ago.
Two out of five prospective studies have reported an increased recurrence risk in FV Leiden carriers (Ridker et al, 1995b; Eichinger et al, 1997; Simioni et al, 1997; Kearon et al, 1999; Lindmarker et al, 1999). Notably, one of these studies considered men and spontaneous first events (Ridker et al, 1995b), clinical variables that were related to the risk of recurrence in the present and previous studies (Schulman et al, 1995; Prandoni et al, 1996; Lindmarker et al, 1999). Moreover, recurrence rates in prior studies were not adjusted for the influence of the FII mutation and elevated FVIII:C or homocysteine levels.
This retrospective study has limitations. First, referral bias might have occurred easily, as it is highly likely that patients with recurrences will be referred. This is supported by the longer observation time in our patients with recurrences. However, the proportion of patients with recurrences did not differ between probands and relatives, whereas the annual incidence rate was estimated in relatives. Second, selection bias might have been introduced since the study population was not an inception cohort. This is less likely because an excess of mortality due to venous thromboembolism was not observed. Third, the results might have been influenced by thromboembolic episodes that were not confirmed by objective testing. However, the prevalence of concomitant thrombophilic defects in patients with or without recurrences remained similar if these events were excluded (FII mutation 13% versus 4%, homozygosity for FV Leiden 8% versus 5%, and deficiencies of protein C or protein S 6% versus 1%). Fourth, post-phlebitic signs and symptoms might have been wrongly interpreted as recurrence. We found an ipsilateral recurrence in about half of the patients with deep vein thrombosis, similar to previously reported rates in venographically controlled studies (Schulman et al, 1995; Prandoni et al, 1996). Moreover, post-phlebitic syndrome as a reason for misclassification can be ruled out in the remaining patients who experienced contralateral recurrences. Finally, thrombophilic disorders might be a consequence of recurrence rather than its cause, as all assays were performed in patients who had already experienced recurrence. For most thrombophilic defects (FV Leiden, FII mutation, deficiencies of antithrombin, protein C and protein S), acquired abnormalities can be ruled out by the use of DNA tests and strict criteria for inherited deficiencies.
A possible clinical implication of our findings is the need for risk stratification in heterozygous FV Leiden carriers who experience a first episode of venous thromboembolism since their risk of recurrence may range from 0·45 to 4·8 per 100 patient–years. It should be noted, however, that although concomitant disorders were demonstrated in a majority (75%) of FV Leiden carriers with recurrent venous thromboembolism, these were also frequently (63%) found in carriers who had experienced only a single episode of venous thromboembolism. The optimal duration of anticoagulant treatment after a first event primarily depends on the risk of recurrence over time, as well as the risk of major bleeding. Considering an individual risk assessment, especially heterozygous carriers of FV Leiden with a first spontaneous episode of venous thromboembolism may be tested for all known thrombophilic disorders.
In conclusion, our study provides supportive evidence that the risk of recurrent venous thromboembolism in heterozygous FV Leiden carriers depends on the concomitance of other thrombophilic disorders, as well as on whether the first thrombotic event occurred spontaneously.