Journal of Thrombosis and Haemostasis, 12: 306–312
DOI: 10.1111/jth.12499
ORIGINAL ARTICLE
Family history of venous thromboembolism (VTE) and risk of recurrent hospitalization for VTE: a nationwide family study in Sweden € L E R , * H . O H L S S O N , * J . S U N D Q U I S T * † and K . S U N D Q U I S T * † B . Z OL
€, Sweden; and †Stanford Prevention Research Center, *Center for Primary Health Care Research, Region Sk ane/Lund University, Malmo Stanford University School of Medicine, Stanford, CA, USA
€ ller B, Ohlsson H, Sundquist J, Sundquist K. Family history of venous thromboembolism (VTE) and risk of recurrent hospiTo cite this article: Zo talization for VTE: a nationwide family study in Sweden. J Thromb Haemost 2014; 12: 306–12.
Summary. Background: Data concerning the importance of a family history of venous thromboembolism (VTE) for the risk of recurrent VTE are sparse. The aim of this nationwide study was to determine whether a family history of VTE is a risk factor for recurrent hospitalization for unprovoked VTE (deep vein thrombosis of the lower extremities or pulmonary embolism). Methods: We linked Multigeneration Register data on individuals aged 0–77 years to the Swedish nationwide Hospital Discharge Register data for the period 1987–2009 to compare the risk of hospitalization for unprovoked recurrent VTE among individuals with and without a parental or sibling history of VTE. We calculated hazard ratios (HRs) to determine the familial HR for recurrent hospitalization for VTE. Results and Conclusions: The risk of recurrent VTE hospitalization was 1.20 (95% confidence interval [CI] 1.10–1.32) for individuals with affected parents, and 1.30 (95% CI 1.14–1.49) for those with affected siblings. The risk of recurrent VTE hospitalization in individuals with two affected parents was 1.92 (95% CI 1.44– 2.58). There was an interaction between age at diagnosis of VTE and a family history of VTE, with a family history having a stronger effect on VTE risk in younger patients. We conclude that a family history of VTE is a modest risk factor for recurrent VTE hospitalization in Sweden. Keywords: genetics; heredity; pulmonary embolism; risk factors; venous thromboembolism; venous thrombosis.
Correspondence: Bengt Z€ oller, Center for Primary Health Care Research, CRC, Building 28, Floor 11, Jan Waldenstr€ oms gata 35, Sk ane University Hospital, S-205 02 Malm€ o, Sweden. Tel.: +46 70 6691476; fax: +46 40 391370. E-mail: [email protected] Received 17 July 2013 Manuscript handled by: F. R. Rosendaal Final decision: F. R. Rosendaal, 18 December 2013
Background Venous thromboembolism (VTE), which comprises deep vein thrombosis (DVT) of the lower extremities and pulmonary embolism (PE), is often a chronic disease, with a risk of recurrence within 5 years of ~ 25% [1–3]. VTE recurrence can be prevented with anticoagulants, but their use increases the risk of bleeding [4–6]. Risk assessment is therefore important to balance the risks and benefits of anticoagulant therapy [4–6]. However, the risk of recurrent thrombosis after stopping treatment is not easily predicted, although a number of risk factors for recurrent VTE have been identified [1–7]. The risk of recurrent VTE is especially high in patients with unprovoked VTE (i.e. VTE in the absence of transitory risk factors such as surgery, trauma, pregnancy, and oral contraceptive use) [1–7]. Other risk factors for recurrent VTE include PE (as compared with DVT), proximal DVT (as compared with distal DVT), male sex, cancer, obesity, residual venous thrombosis, and elevated D-dimer levels [1–7]. Several genetic defects, such as deficiencies of antithrombin, protein C, and protein S, resistance to activated protein C (APC) (APC resistance or the presence of factor V Leiden, rs6025), prothrombin mutations (rs1799963), and non-O blood group, have been associated with an increased risk of a first VTE event [8,9]. However, the clinical importance of these inherited genetic defects in recurrent VTE risk is controversial [1–3,7]. Moreover, the predictive value of the family history for any of these genetic defects is low [10], suggesting that many unknown genetic defects remain to be identified. A family history of VTE could therefore be a better and more comprehensive predictor of VTE than the present established genetic variants associated with VTE. Although a number of studies have shown that a family history is a risk factor for a first VTE event [10–19], only two studies have investigated whether a family history of VTE is associated with an increased risk of recurrent VTE [20,21]. These two studies did not find an association between a family his© 2013 International Society on Thrombosis and Haemostasis
Familial recurrent venous thromboembolism 307
tory of VTE and the risk of recurrent VTE. However, no published large-scale nationwide study has investigated a family history of VTE as a predictor of recurrent VTE after a first VTE event. In this nationwide study, we investigated the relationship between a family history of VTE and the risk of recurrent hospitalization for VTE in patients with unprovoked DVT/PE in Sweden. Methods To assess VTE among individuals in Sweden, comprehensive nationwide data, including healthcare data, were linked [22–25]. This linking was based on the unique 10digit personal ID numbers assigned at birth or immigration to all Swedish residents for life, information on which is nearly 100% complete. These numbers were replaced with serial numbers to preserve anonymity. Our database contains data from four sources: 1 The Swedish Multigeneration Register, which contains information on family relationships (siblings and parent–offspring). The register contains information on index persons registered in Sweden between 1 January 1961 and 31 December 2009 and born between 1 January 1932 and 31 December 2009. 2 The Swedish Hospital Discharge Register, which contains information on all hospital diagnoses for all people in Sweden. Each record includes the main discharge diagnosis. 3 The Swedish Cause of Death Register, which contains data on date of death. 4 The Total Population Register, which contains data on year of birth, sex, country of birth, and education. This study was approved by the Ethics Committee of Lund University, Sweden. Definition of a family history of VTE
Individuals diagnosed with VTE according to the seventh (1964–1968), eighth (1969–1986), ninth (1987–1996) and tenth (1997–2009) versions of the International Classification of Diseases (ICD) were identified from hospital discharge data. We used only main hospital diagnoses of VTE, to guarantee high validity [25,26]. A family history of VTE was defined as hospitalization of full siblings or parents for VTE manifestations between 1964 and 2009 (Table S1) [13,16–18]. Sample
The dataset for the analyses was created by identifying all individuals diagnosed with VTE from the Swedish Hospital Discharge Register. The study period was 1987– 2009, as the Swedish Hospital Discharge Register has had nationwide coverage only since 1987. We selected the © 2013 International Society on Thrombosis and Haemostasis
first occasion on which an individual was found in the register with a VTE diagnosis to ensure that only recurrence risk in individuals with a first VTE event was studied. For our main analysis, we selected only individuals diagnosed with DVT (ICD-9, 451 with exclusion of 451A; ICD-10, I80 with exclusion of I800) or PE (ICD-9, 415B and 416W; ICD-10, I26) during 1987–2009 and not previously registered with any previous VTE manifestation between 1964 and 2009, i.e. only first-time hospitalizations for DVT/PE. Superficial thrombophlebitis (ICD9, 451A; ICD-10, I800) and forms of VTE other than PE and DVT of the lower extremities were not included. Patients registered with a diagnosis of pregnancy-related or abortion-related VTE were not included (Table S1). We also excluded individuals with pregnancy/delivery or abortion 90 days before the diagnosis of DVT/PE (ICD9, 630–676; ICD-10, O00–O99) and individuals with other major circumstantial risk factors, i.e. hospitalization for malignancy 5 years before or 1 year after the first DVT/ PE event, hospitalization for any type of surgery 90 days before the first VTE event (hospitalization with any surgery code AA-ZZ or 0000–9999), or hospitalization for any type of fracture or trauma within 90 days before the DVT/PE event (ICD-9, 800–929; ICD-10, S00-S99 and T00-T14). As we had no data on anticoagulation treatment, we started the follow-up 180 days after the first PE/DVT event, as most PE/DVT cases are treated for 90–180 days. We also performed analysis with follow-up from 90 days, in order to investigate whether this would affect the results. In total, 34 006 individuals fulfilling the above inclusion and exclusion criteria with a first DVT/PE event were initially identified. However, 1080 patients died before follow-up from 180 days, and 1337 were lost to follow-up, owing to inclusion in the last part of 2009. Moreover, patients with recurrence of any type of VTE (excluding superficial thrombophlebitis; Table S2) before 180 days (2778) were also excluded. Hence, the analyses were based on a dataset containing information on 28 811 individuals with a first event of unprovoked DVT/PE (mean age of 51.5 years [standard deviation 13], 52% men, VTE recurrence rate of 11.8%). Of these 28 811 individuals, 3991 had at least one parent diagnosed with VTE, and 1355 had at least one full sibling diagnosed with VTE. In the models, we included patient sex and age at VTE diagnosis. For all individuals, we also retrieved information from the registers on level of education. Education was divided into three categories: (i) 0–9 years (low education); (ii) 10–11 years (intermediate education); and (iii) > 12 years (high education). High education was used as the reference in the analysis. Statistical methods
Similarly to a previous study of a family history of atrial fibrillation (AF) and recurrent AF risk in patients with
308 B. Z€oller et al
lone AF [27,28], we used Cox proportional hazards models to investigate the recurrence of VTE. We started the follow-up 180 days after the first diagnosis of unprovoked DVT/PE, and followed the subjects until VTE recurrence (hospitalization for any type of VTE except for superficial thrombophlebitis; Table S2), death, emigration, or the end of the follow-up period (31 December 2009) (whichever came first). The analysis was repeated with follow-up from 90 days after the first diagnosis of unprovoked DVT/PE. In model A, parental history of VTE (yes/no) was included as a covariate. Model B also included sex, age at diagnosis of DVT/PE (centered at the mean value), and education. In model C, parental history was categorized as no parental history of VTE (reference in the model), one parent with VTE, and two parents with VTE. In model D, parental history was categorized as no parental history of VTE (reference in the model), mother with VTE, and father with VTE. Model E was similar to model B, except that it also included an interaction term for parental history of VTE and age at diagnosis of DVT/PE. Model F only included VTE in a sibling as a covariate, whereas model G also included sex, age at diagnosis of DVT/PE (centered at the mean value), and education, as well as an interaction term for VTE in a sibling and age at diagnosis of DVT/PE. In model H, sibling history of VTE was categorized into three groups: no sibling history of VTE (reference in the model), one sibling with VTE, and two or more siblings with VTE. Model I included both parental history of VTE and sibling history of VTE. In the final model, parental history and sibling history were combined into one variable, and included in the model as a single variable. All models except for models A and F included sex, age at diagnosis of DVT/PE, and education. In order to further evaluate the results, we also investigated DVT and PE separately. For both of these types of VTE, we created three separate models with three different outcomes. For all three models, we started the follow-up 180 days after the first diagnosis of DVT/PE, and followed the subjects until the diagnosis of DVT, PE, and all other types of VTE except for DVT and PE (OVTE) (Table S2), death, emigration, or the end of the follow-up period (31 December 2009) (whichever came first). In all models, we used a family history of VTE (in parents and/or sibling) as a combined variable. All models were adjusted for sex, age at first DVT/PE diagnosis, and education. To take into account the non-independence of observations from the same family, we used a robust sandwich estimator in all models. We present hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs). The proportional hazards assumption was fulfilled for the variables of interest. All calculations were performed with SAS version 9.3 (SAS Institute, Cary, NC, USA).
Results In total, 28 811 individuals with unprovoked DVT/PE were included in the study (Table 1), of whom 3404 (11.8%) suffered recurrent hospitalization for VTE. Five hundred and sixty individuals (16.5%) with recurrence and 3276 (13.5%) individuals without recurrence had a parent with VTE. Two hundred and fifteen (6.3%) individuals with recurrence and 1140 (4.5%) individuals without recurrence had a sibling history of VTE. Model A in Table 2 shows that proband cases with parents diagnosed with VTE had a 1.20-fold higher HR for recurrent VTE hospitalization than proband cases without a parental history of VTE. Follow-up was started 180 days after the first unprovoked DVT/PE event. The crude HR was not attenuated when age, sex and education level were included in the model (Table 2, model B). There was an interaction between age at diagnosis and parental history of VTE (HR 0.987) (model E), indicating a decreased association between parental VTE and recurrent VTE in older cases. Model C indicates that the HR was higher for proband cases with two affected parents (HR 1.92) than for proband cases with only one affected parent (HR 1.18). Model D shows that proband cases with mothers diagnosed with VTE had an HR of 1.16 for recurrent VTE hospitalization, and probands with affected fathers had an HR of 1.31. Figure 1 shows that, at the end of the follow-up period, ~ 21% of proband cases with a parental history of VTE had recurrent hospitalization for VTE; the corresponding value for proband cases without a parental history was 18%. Table S3 shows models A and B with follow-up from 90 days after the first unprovoked DVT/PE event. HRs Table 1 Descriptive statistics for individuals with unprovoked deep vein thrombosis (DVT)/pulmonary embolism (PE) in the Swedish population (1987–2009) DVT/PE
No recurrence
Recurrence*
N VTE in parents VTE in one parent VTE in two parents VTE in mother VTE in father VTE in one sibling VTE in two or more siblings Age (years), mean (SD) Males Females Low education Intermediate education High education Information on education
25 407 3431 3276 155 1961 1625 1140 135 51.5 12 977 12 430 8593 7640 9071 103
3404 560 (16.5) 516 (15.2) 44 (1.3) 314 (9.2) 290 (8.5) 215 (6.3) 19 (0.6) 51.6 (12) 2019 (59.3) 1385 (40.7) 1264 (37.1) 1030 (30.3) 1105 (32.5) 5 (0.1)
(13.5) (12.9) (0.6) (7.7) (6.4) (4.5) (0.5) (14) (51.1) (48.9) (33.8) (31.1) (35.7) (0.4)
SD, standard deviation. Except where otherwise stated, data are given as no. (%). *Recurrence of any type of VTE, excluding superficial thrombophlebitis. © 2013 International Society on Thrombosis and Haemostasis
Familial recurrent venous thromboembolism 309 Table 2 Results from Cox proportional hazards models for individuals with deep vein thrombosis (DVT)/pulmonary embolismPE in the Swedish population (1987–2009): parent–offspring analysis with follow-up from 180 days after the first venous thromboembolism (VTE) event 180 days DVT/PE
Model A HR (95% CI)
Model B HR (95% CI)
Model C HR (95% CI)
Model D HR (95% CI)
Model E HR (95% CI)
VTE in parents VTE in one parent VTE in two parents VTE in mother VTE in father Age at diagnosis of VTE (centered) Male vs. female Low education Intermediate education High education VTE in parents, age at diagnosis of VTE*
1.20 (1.10–1.32) – – – – – – – – – –
1.21 (1.11–1.33) – – – – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
– 1.18 (1.07–1.29) 1.92 (1.44–2.58) – – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
– – – 1.16 (1.03–1.30) 1.31 (1.16–1.48) 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
1.18 (1.07–1.30) – – – – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference 0.987 (0.980–0.995)*
CI, confidence interval; HR, hazard ratio. *The HR for VTE in a 52-year-old individual with affected parents is 1.18; the HR for VTE in a 42-year-old individual with affected parents is 1.34; and the HR for VTE in a 62-year-old individual with affected parents is 1.04. Model A included parental history of VTE. Model B also included sex (reference: female), age at diagnosis of DVT/PE (centered at the mean value), and education. In model C, parental history was categorized as no parental history of VTE (reference in the model), one parent with VTE, and two parents with VTE. In model D, parental history was categorized as no parental history of VTE (reference in the model), mother with VTE, and father with VTE. Model E contained a term for the interaction between parental history of VTE and age at diagnosis. An interaction term was only included in the model if the P-value for the interaction term was < 0.05.
© 2013 International Society on Thrombosis and Haemostasis
1.00 0.99 0.98 0.97 0.96 0.95
No family history of VTE
0.94
Family history of VTE
0.93 0.92 Survivorship function
were similar to those for follow-up from 180 days (Table 2). We also calculated HRs without exclusion of 2778 cases who were rehospitalized for VTE before the start of the follow-up at 180 days after the first VTE event (data not shown in tables). Thus, 31 589 cases were included in the follow-up from 180 days. The HR for those with affected parents was 1.19, which was similar to the HR obtained when those who were rehospitalized for VTE before the start of the follow-up from 180 days were excluded (HR 1.20, Table 2). Thus, the exclusion of 2778 cases with early recurrence before 180 days did not affect the familial risks of recurrent VTE. Model F in Table 3 shows that proband cases with at least one sibling diagnosed with VTE had a 1.30-fold higher HR for recurrent VTE hospitalization than proband cases without a sibling history of VTE. In the multivariate model B, which shows the risk of recurrent VTE associated with a sibling history of VTE, there seemed to be an interaction between age at diagnosis and a sibling history of VTE (HR 0.977) (model E), indicating a decreased association between a sibling history of VTE and recurrent VTE among older cases. Surprisingly, model H shows no significant effect on recurrent VTE risk in individuals with two affected siblings. In model I, we see that parental and sibling histories of VTE were independent risk factors for recurrent VTE. In model J, a family history (in a sibling or parent) was significantly associated with an increased risk of recurrent VTE. In Table 4, subanalysis of the familial risks (parent or sibling) of recurrent DVT, PE and OVTE among patients with unprovoked DVT or PE are presented. A family history was a risk factor for recurrent DVT, PE and OVTE among patients with unprovoked DVT or unprovoked
0.91 0.90 0.89 0.88 0.87 0.86 0.85 0.84 0.83 0.82 0.81 0.80 0.79 0.78 0
1000
2000
3000
4000 Days
5000
6000
7000
8000
Fig. 1. The survivorship function from model A. After 8000 days, ~ 21% of individuals with a parent with venous thromboembolism (VTE) had recurrence, as compared with only 18% of individuals without a parental history of VTE.
PE. A family history of VTE was a significant risk factor for all types of recurrent event in patients with a first DVT. Among patients with a first PE, a family history of VTE was a significant risk factor only for recurrence of PE, but the number of recurrent DVT and OVTE events was low. We also determined familial recurrent risk after exclusion of patients with malignancies > 1 year after the first diagnosis of VTE (similar to model A in Table 2). The familial risk after exclusion (1.22, 95% CI 1.11–1.34) was similar to that in model A in Table 2 (1.20, 95% CI 1.10–1.32). Moreover, exclusion of cases with a family history of coronary heart disease (ICD-10, I20– I25; ICD-9, 410–414) did not change the familial risks to
310 B. Z€oller et al
a major degree (1.19, 95% CI 1.09–1.31). Moreover, only 29 cases had a recurrent episode associated with pregnancy or abortion, which could not have affected the results to any major degree. Discussion The present study is, to our knowledge, the first nationwide study to estimate the familial risk of recurrent hospitalization for unprovoked VTE. Our findings suggest that a family history of VTE is a novel but modest risk factor for recurrent hospitalization for unprovoked VTE. The higher risk of recurrent unprovoked VTE hospitalization among probands with two affected parents may be genetic, although non-genetic familial factors cannot be ruled out. There was an interaction between age at diagnosis and a family history of VTE, with a family history having a stronger effect on VTE hospitalization risk in younger age groups. This further suggests a genetic contribution. Genetic risk factors for VTE may have prognostic value regarding recurrence risk. The mechanism behind the observed familial risk of recurrent unprovoked VTE hospitalization may also include non-genetic components. For instance, individuals with a family history of unprovoked VTE may have greater VTE awareness and a lower threshold for seeking healthcare. However, several genetic variants have been linked to the risk of a first VTE event [9]. It is possible that these variants may also predispose individuals to an increased risk of recurrent hospitalization for unprovoked VTE. Most previous studies have only investigated the familial risk of a first VTE event [10– 18]. Only two studies have investigated whether a family history is associated with an increased risk of recurrent unprovoked VTE [20,21]. The familial risk of recurrent unprovoked VTE hospitalization was lower than that
reported previously for the risk of first-time VTE hospitalization [10–18], suggesting that familial, and possibly genetic, factors may be more important for first-time VTE hospitalization than for recurrent unprovoked VTE hospitalization. Our results fit with studies showing that genetic variants are less important for recurrent VTE than for a first VTE event [1–3,7,9]. The lower risk of recurrent VTE than for a first VTE event among individuals with a family history of VTE could explain why the two previously published studies did not find an association between a family history and recurrent VTE risk [20,21]. The present study has a number of strengths. These include complete nationwide coverage in a country with high medical standards, and medical diagnosis of patients by specialists during examinations in hospitals. In addition, the results were not affected by recall bias, because both the probands and their relatives were medically diagnosed. Importantly, the Multigeneration Register is a validated source that has been proved to be reliable in the study of many familial diseases [22–25]. The present study also has a number of limitations. One limitation is the lack of risk factors for recurrent VTE, such as body mass index, D-dimer, and residual thrombosis [2,3]. The most important limitation is that we do not know for how long patients were treated with anticoagulants after the first unprovoked VTE event. We therefore started the follow-up analysis at 180 days after the first event. Moreover, we do not know whether recurrent VTE hospitalization was attributable to a new episode of VTE or to the same persistent episode of VTE. However, the latter alternative is unlikely, as we started survival analysis after 180 days, and progression of existing VTE is likely to occur soon after treatment is started. The risk of recurrent hospitalization was increased during the whole follow-up period, which sug-
Table 3 Results from Cox proportional hazards models for individuals with deep vein thrombosis (DVT)/pulmonary embolism (PE) in the Swedish population (1987–2009): sibling analysis with follow-up from 180 days after the first venous thromboembolism (VTE) event 180 days
Model F
Model G
Model H
Model I
Model J
DVT/PE VTE in parents VTE in siblings VTE in one sibling VTE in two or more siblings VTE in relatives (sibling or parent) Age at VTE (centered) Male vs. female Low education Intermediate education High education VTE in siblings, age at diagnosis of VTE*
HR (95% CI) – 1.30 (1.14–1.49) – – – – – – – – –
HR (95% CI) – 1.23 (1.08–1.41) – – – 1.01 (1.01–1.02) 1.44 (1.34–1.54) 1.11 (1.02–1.20) 1.07 (1.02–1.20) Reference 0.977 (0.966–0.988)
HR (95% CI) – – 1.30 (1.14–1.49) 0.92 (0.59–1.45) – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
HR (95% CI) 1.20 (1.09–1.31) 1.22 (1.07–1.40) – – – 1.01 (1.01–1.01) 1.43 (1.34–1.53) 1.11 (1.02–1.21) 1.07 (0.98–1.17) Reference –
HR (95% CI) – – – – 1.19 (1.10–1.30) 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
CI, confidence interval; HR, hazard ratio. *The HR for VTE in a 52-year-old individual with affected siblings is 1.23; the HR for VTE in a 42-year-old individual with affected siblings is 1.55; and the HR for VTE in a 62-year-old individual with affected siblings is 0.91. Model F: crude HR for sibling history of VTE. Model G also included sex, age at diagnosis of DVT/PE (centered at the mean value), education, and an interaction term between VTE in siblings and age at diagnosis of DVT/PE. In model H, sibling history of VTE was categorized as no sibling history of VTE (reference in the model), one sibling with VTE, and two or more siblings with VTE. Model I included both parental history of VTE and sibling history of VTE. In the model J, parental history and sibling history were combined to give a single variable. © 2013 International Society on Thrombosis and Haemostasis
Familial recurrent venous thromboembolism 311 Table 4 Subanalysis of the importance of family history of venous thromboembolism (VTE) for patients with a first deep vein thrombosis (DVT) or pulmonary embolism (PE) event for recurrence of DVT, PE and all other types of VTE except for PE and DVT of the lower extremities (OVTE) in the Swedish population (1987–2009): follow-up from 180 days after the first VTE event
First event (N)
Type of recurrent event
HR (95% CI)*
Recurrent events (N)
Recurrent events in patients with FH, N (%)
Recurrent events in patients without FH, N (%)
DVT (14 404)† DVT (15 216)† DVT (15 196)† PE (16 146)† PE (14 505)† PE (16 222)†
DVT PE OVTE DVT PE OVTE
1.21 1.31 1.21 1.23 1.14 1.25
1368 763 360 372 1581 128
297 (11.3) 193 (7.0) 84 (3.0) 77 (2.9) 298 (12.3) 27 (1.0)
1071 (9.1) 570 (4.6) 276 (2.2) 295 (2.2) 1283 (10.6) 101 (0.8)
(1.06–1.37) (1.03–1.67) (1.06–1.37) (0.96–1.59) (1.01–1.30) (0.82–1.91)
CI, confidence interval; FH, family history of VTE in a sibling or parent; HR, hazard ratio. *Models controlled for sex, age at diagnosis of VTE, and education. †The numbers of patients included in the different analyses are different, because patients with the same early recurrent event as the respective outcome variable (DVT, PE, or OVTE) before the start of follow-up at 180 days from first-time DVT or PE were excluded.
gests that our results mainly reflect an increased risk of new VTE episodes. In addition, hospitalization for postthrombotic syndrome is unlikely to be a problem, as this diagnosis should have a main diagnosis of I870 (ICD.10) or 459B (ICD-9), and these ICD codes were excluded. Moreover, Hansson et al. [28] have previously determined the risk of recurrent hospitalization for VTE and reviewed the validity of medical records. The absence of an increased risk of recurrent VTE among those with two or more affected siblings might be attributable to the fact that these individuals have very high risks, possibly owing to multiple genetic defects, and were possibly treated indefinitely with anticoagulants after the first event [18]. However, we have no information about this. We only included a main diagnosis of VTE in hospitalized patients, in order to increase the probability that the VTE events were new thrombotic events. The validity of VTE as a primary diagnosis is 95% [26]. The lack of the less severe outpatients with VTE would be expected to diminish the risk of recurrent VTE, which is a limitation of the study. However, familial cases are probably not more likely to be hospitalized than non-familial cases, suggesting that the familial risks are not affected. Moreover, almost all cases of PE in Sweden are treated in hospitals [29], but hospital treatment of DVT in Sweden fell between 1987 and 1998, such that, in 1998, 50% of DVT patients in Sweden were treated directly as outpatients [29]. However, the increased recurrence risk for familial cases was observed during the whole follow-up period (Fig. 1). The Swedish Hospital Discharge Register only contains complete data for the period since 1987. However, the lack of data for the period before 1987 will have been similar for both cases and controls, and will not have affected the estimates of familial aggregation (i.e. the HRs). It was most likely a source of non-differential bias regarding familial risk estimates. Another potential limitation is that we do not have access to the methods used for objective diagnosis. However, the Swedish Hospital Discharge Register has high © 2013 International Society on Thrombosis and Haemostasis
validity, especially for cardiovascular disorders such as VTE (95%) [25,26]. Conclusions The present study suggests that a family history of VTE is a modest risk factor for recurrent hospitalization for unprovoked VTE in Swedish patients. The risk of recurrent hospitalization for unprovoked VTE was higher in individuals with two affected parents and in younger individuals, which suggests a genetic contribution. Addendum B. Z€ oller, H. Ohlsson, J. Sundquist, and K. Sundquist contributed to the conception and design of the study, and to the analysis and interpretation of data. J. Sundquist and K. Sundquist contributed to the acquisition of data. B. Z€ oller drafted the manuscript. All authors critically revised the paper and approved the final version. All authors had full access to all of the data (including statistical reports and tables), and take responsibility for the integrity of the data and the accuracy of their analysis. Acknowledgements The authors wish to thank the CPF’s Science Editor, Stephen Gilliver, for his useful comments on the text. The registers used in the present study are maintained by Statistics Sweden and the National Board of Health and Welfare. This work was supported by grants to B. Z€ oller from the Swedish Heart-Lung Foundation and Region Sk ane (REGSKANE-124611), to K. Sundquist from the Swedish Research Council (K2009-70X-15428-05-3 and K2012-70X-15428-08-3), and to J. Sundquist from the Swedish Council for Working Life and Social Research (2007-1754) and the Swedish Freemasons Foundation, as well as by ALF funding from Region Sk ane awarded to B. Z€ oller, J. Sundquist, and K. Sundquist.
312 B. Z€oller et al
Disclosure of Conflict of Interests The authors state that they have no conflict of interest. Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Family history of VTE was defined by hospitalization of full siblings or parents for any of the following VTE manifestations between 1964 and 2009 [13,16–18]. Table S2. VTE endpoints. Table S3. Results from Cox proportional hazard models for individuals with DVT/PE in the Swedish population (1987–2009): parent–offspring analysis with follow-up from 90 days after the first VTE event. References 1 Kyrle PA, Rosendaal FR, Eichinger S. Risk assessment for recurrent venous thrombosis. Lancet 2010; 376: 2032–9. 2 Labropoulos N, Spentzouris G, Gasparis AP, Meissner M. Impact and clinical significance of recurrent venous thromboembolism. Br J Surg 2010; 97: 989–99. 3 Ribeiro DD, Lijfering WM, Barreto SM, Rosendaal FR, Rezende SM. Epidemiology of recurrent venous thrombosis. Braz J Med Biol Res 2012; 45: 1–7. 4 Goldhaber SZ, Piazza G. Optimal duration of anticoagulation after venous thromboembolism. Circulation 2011; 123: 664–7. 5 de Jong PG, Coppens M, Middeldorp S. Duration of anticoagulant therapy for venous thromboembolism: balancing benefits and harms on the long term. Br J Haematol 2012; 158: 433–41. 6 Schulman S. Optimal duration of anticoagulant therapy. Semin Thromb Hemost 2013; 39: 141–6. 7 Zhu T, Martinez I, Emmerich J. Venous thromboembolism: risk factors for recurrence. Arterioscler Thromb Vasc Biol 2009; 29: 298–310. 8 Morange PE, Tregouet DA. Lessons from genome-wide association studies in venous thrombosis. J Thromb Haemost 2011; 9 (Suppl. 1): 258–64. 9 Rosendaal FR, Reitsma PH. Genetics of venous thrombosis. J Thromb Haemost 2009; 7(Suppl. 1): 301–4. 10 Bezemer ID, van der Meer FJ, Eikenboom JC, Rosendaal FR, Doggen CJ. The value of family history as a risk indicator for venous thrombosis. Arch Intern Med 2009; 169: 610–15. 11 Dowling NF, Austin H, Dilley A, Whitsett C, Evatt BL, Hooper WC. The epidemiology of venous thromboembolism in Caucasians and African-Americans: the GATE Study. J Thromb Haemost 2003; 1: 80–7. 12 Noboa S, Le Gal G, Lacut K, Mercier B, Leroyer C, Nowak E, Mottier D, Oger E. EDITH Collaborative Study Group. Family history as a risk factor for venous thromboembolism. Thromb Res 2008; 122: 624–9. 13 Z€ oller B, Li X, Sundquist J, Sundquist K. Parental history and venous thromboembolism: a nationwide study of age and sex-specific familial risks in Sweden. J Thromb Haemost 2011; 9: 64–70.
14 Couturaud F, Leroyer C, Julian JA, Kahn SR, Ginsberg JS, Wells PS, Douketis JD, Mottier D, Kearon C. Factors that predict risk of thrombosis in relatives of patients with unprovoked venous thromboembolism. Chest 2009; 136: 1537–45. 15 Sørensen HT, Riis AH, Diaz LJ, Andersen EW, Baron JA, Andersen PK. Familial risk of venous thromboembolism: a nationwide cohort study. J Thromb Haemost 2011; 9: 320–4. 16 Z€ oller B, Li X, Sundquist J, Sundquist K. Familial risks for unusual forms of venous thrombosis: a nationwide epidemiological study in Sweden. J Intern Med 2010; 270: 158–65. 17 Z€ oller B, Li X, Sundquist J, Sundquist K. Determination of agespecific and sex-specific familial risks for the different manifestations of venous thromboembolism: a nationwide family study in Sweden. Thromb Haemost 2011; 106: 102–12. 18 Z€ oller B, Li X, Sundquist J, Sundquist K. Age- and gender-specific familial risks for venous thromboembolism: a nationwide epidemiological study based on hospitalizations in Sweden. Circulation 2011; 124: 1012–20. 19 Prandoni P, Prins MH, Ghirarduzzi A, Pengo V, Sartori MT, Ugolotti MC, Bracco A, Veropalumbo MR, Noventa F, Lensing AW. Family history of venous thrombosis or sudden death as a risk factor for venous thromboembolism. Thromb Haemost 2012; 107: 1191–2. 20 Hron G, Eichinger S, Weltermann A, Minar E, Bialonczyk C, Hirschl M, Stain M, Gartner V, Kyrle PA. Family history for venous thromboembolism and the risk for recurrence. Am J Med 2006; 119: 50–3. 21 Gauthier K, Kovacs MJ, Wells PS, le Gal G, Rodger M; REVERSE investigators. Family history of venous thromboembolism (VTE) as a predictor for recurrent VTE in unprovoked VTE patients. J Thromb Haemost 2013; 11: 200–3. 22 Statistics Sweden. The Swedish Multigeneration Register (1960– 1990): Registret o€ver totalbefolkningen/RTB. In Swedish. Stockholm: Statistics Sweden; 2005. 23 The National Board of Health and Welfare. Validity of the Diagnoses from the Swedish In-Care Register 1987 and 1995. In Swedish. Stockholm: Epidemiologiskt Centrum, Socialstyrelsen; 2000. 24 Rosen M, Hakulinen T. Use of disease registers. In: Handbook of Epidemiology. Berlin: Springer-Verlag, 2005: 231–51. 25 Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, Heurgren M, Olausson PO. External review and validation of the Swedish national inpatient register. BMC Public Health 2011; 11: 450. [online]. 26 Rosengren A, Freden M, Hansson PO, Wilhelmsen L, Wedel H, Eriksson H. Psychosocial factors and venous thromboembolism: a long-term follow-up study of Swedish men. J Thromb Haemost 2008; 6: 558–64. 27 Z€ oller B, Ohlsson H, Sundquist J, Sundquist K. Family history as a risk factor for recurrent hospitalization for lone atrial fibrillation: a nationwide family study in Sweden. BMC Cardiovasc Disord 2012; 12: 121. [online]. 28 Hansson PO, S€ orbo J, Eriksson H. Recurrent venous thromboembolism after deep vein thrombosis: incidence and risk factors. Arch Intern Med 2000; 160: 769–74. 29 Statens beredning f€ or medicinsk utv€ardering (SBU). Blodpropp – F€ orebyggande, Diagnostik Och Behandling av Ven€ os Trombembolism SBU-Rapport 158. In Swedish. Stockholm: The Swedish Council on Technology Assessment in Health Care; 2002 I–III.
© 2013 International Society on Thrombosis and Haemostasis
DOI: 10.1111/jth.12499
ORIGINAL ARTICLE
Family history of venous thromboembolism (VTE) and risk of recurrent hospitalization for VTE: a nationwide family study in Sweden € L E R , * H . O H L S S O N , * J . S U N D Q U I S T * † and K . S U N D Q U I S T * † B . Z OL
€, Sweden; and †Stanford Prevention Research Center, *Center for Primary Health Care Research, Region Sk ane/Lund University, Malmo Stanford University School of Medicine, Stanford, CA, USA
€ ller B, Ohlsson H, Sundquist J, Sundquist K. Family history of venous thromboembolism (VTE) and risk of recurrent hospiTo cite this article: Zo talization for VTE: a nationwide family study in Sweden. J Thromb Haemost 2014; 12: 306–12.
Summary. Background: Data concerning the importance of a family history of venous thromboembolism (VTE) for the risk of recurrent VTE are sparse. The aim of this nationwide study was to determine whether a family history of VTE is a risk factor for recurrent hospitalization for unprovoked VTE (deep vein thrombosis of the lower extremities or pulmonary embolism). Methods: We linked Multigeneration Register data on individuals aged 0–77 years to the Swedish nationwide Hospital Discharge Register data for the period 1987–2009 to compare the risk of hospitalization for unprovoked recurrent VTE among individuals with and without a parental or sibling history of VTE. We calculated hazard ratios (HRs) to determine the familial HR for recurrent hospitalization for VTE. Results and Conclusions: The risk of recurrent VTE hospitalization was 1.20 (95% confidence interval [CI] 1.10–1.32) for individuals with affected parents, and 1.30 (95% CI 1.14–1.49) for those with affected siblings. The risk of recurrent VTE hospitalization in individuals with two affected parents was 1.92 (95% CI 1.44– 2.58). There was an interaction between age at diagnosis of VTE and a family history of VTE, with a family history having a stronger effect on VTE risk in younger patients. We conclude that a family history of VTE is a modest risk factor for recurrent VTE hospitalization in Sweden. Keywords: genetics; heredity; pulmonary embolism; risk factors; venous thromboembolism; venous thrombosis.
Correspondence: Bengt Z€ oller, Center for Primary Health Care Research, CRC, Building 28, Floor 11, Jan Waldenstr€ oms gata 35, Sk ane University Hospital, S-205 02 Malm€ o, Sweden. Tel.: +46 70 6691476; fax: +46 40 391370. E-mail: [email protected] Received 17 July 2013 Manuscript handled by: F. R. Rosendaal Final decision: F. R. Rosendaal, 18 December 2013
Background Venous thromboembolism (VTE), which comprises deep vein thrombosis (DVT) of the lower extremities and pulmonary embolism (PE), is often a chronic disease, with a risk of recurrence within 5 years of ~ 25% [1–3]. VTE recurrence can be prevented with anticoagulants, but their use increases the risk of bleeding [4–6]. Risk assessment is therefore important to balance the risks and benefits of anticoagulant therapy [4–6]. However, the risk of recurrent thrombosis after stopping treatment is not easily predicted, although a number of risk factors for recurrent VTE have been identified [1–7]. The risk of recurrent VTE is especially high in patients with unprovoked VTE (i.e. VTE in the absence of transitory risk factors such as surgery, trauma, pregnancy, and oral contraceptive use) [1–7]. Other risk factors for recurrent VTE include PE (as compared with DVT), proximal DVT (as compared with distal DVT), male sex, cancer, obesity, residual venous thrombosis, and elevated D-dimer levels [1–7]. Several genetic defects, such as deficiencies of antithrombin, protein C, and protein S, resistance to activated protein C (APC) (APC resistance or the presence of factor V Leiden, rs6025), prothrombin mutations (rs1799963), and non-O blood group, have been associated with an increased risk of a first VTE event [8,9]. However, the clinical importance of these inherited genetic defects in recurrent VTE risk is controversial [1–3,7]. Moreover, the predictive value of the family history for any of these genetic defects is low [10], suggesting that many unknown genetic defects remain to be identified. A family history of VTE could therefore be a better and more comprehensive predictor of VTE than the present established genetic variants associated with VTE. Although a number of studies have shown that a family history is a risk factor for a first VTE event [10–19], only two studies have investigated whether a family history of VTE is associated with an increased risk of recurrent VTE [20,21]. These two studies did not find an association between a family his© 2013 International Society on Thrombosis and Haemostasis
Familial recurrent venous thromboembolism 307
tory of VTE and the risk of recurrent VTE. However, no published large-scale nationwide study has investigated a family history of VTE as a predictor of recurrent VTE after a first VTE event. In this nationwide study, we investigated the relationship between a family history of VTE and the risk of recurrent hospitalization for VTE in patients with unprovoked DVT/PE in Sweden. Methods To assess VTE among individuals in Sweden, comprehensive nationwide data, including healthcare data, were linked [22–25]. This linking was based on the unique 10digit personal ID numbers assigned at birth or immigration to all Swedish residents for life, information on which is nearly 100% complete. These numbers were replaced with serial numbers to preserve anonymity. Our database contains data from four sources: 1 The Swedish Multigeneration Register, which contains information on family relationships (siblings and parent–offspring). The register contains information on index persons registered in Sweden between 1 January 1961 and 31 December 2009 and born between 1 January 1932 and 31 December 2009. 2 The Swedish Hospital Discharge Register, which contains information on all hospital diagnoses for all people in Sweden. Each record includes the main discharge diagnosis. 3 The Swedish Cause of Death Register, which contains data on date of death. 4 The Total Population Register, which contains data on year of birth, sex, country of birth, and education. This study was approved by the Ethics Committee of Lund University, Sweden. Definition of a family history of VTE
Individuals diagnosed with VTE according to the seventh (1964–1968), eighth (1969–1986), ninth (1987–1996) and tenth (1997–2009) versions of the International Classification of Diseases (ICD) were identified from hospital discharge data. We used only main hospital diagnoses of VTE, to guarantee high validity [25,26]. A family history of VTE was defined as hospitalization of full siblings or parents for VTE manifestations between 1964 and 2009 (Table S1) [13,16–18]. Sample
The dataset for the analyses was created by identifying all individuals diagnosed with VTE from the Swedish Hospital Discharge Register. The study period was 1987– 2009, as the Swedish Hospital Discharge Register has had nationwide coverage only since 1987. We selected the © 2013 International Society on Thrombosis and Haemostasis
first occasion on which an individual was found in the register with a VTE diagnosis to ensure that only recurrence risk in individuals with a first VTE event was studied. For our main analysis, we selected only individuals diagnosed with DVT (ICD-9, 451 with exclusion of 451A; ICD-10, I80 with exclusion of I800) or PE (ICD-9, 415B and 416W; ICD-10, I26) during 1987–2009 and not previously registered with any previous VTE manifestation between 1964 and 2009, i.e. only first-time hospitalizations for DVT/PE. Superficial thrombophlebitis (ICD9, 451A; ICD-10, I800) and forms of VTE other than PE and DVT of the lower extremities were not included. Patients registered with a diagnosis of pregnancy-related or abortion-related VTE were not included (Table S1). We also excluded individuals with pregnancy/delivery or abortion 90 days before the diagnosis of DVT/PE (ICD9, 630–676; ICD-10, O00–O99) and individuals with other major circumstantial risk factors, i.e. hospitalization for malignancy 5 years before or 1 year after the first DVT/ PE event, hospitalization for any type of surgery 90 days before the first VTE event (hospitalization with any surgery code AA-ZZ or 0000–9999), or hospitalization for any type of fracture or trauma within 90 days before the DVT/PE event (ICD-9, 800–929; ICD-10, S00-S99 and T00-T14). As we had no data on anticoagulation treatment, we started the follow-up 180 days after the first PE/DVT event, as most PE/DVT cases are treated for 90–180 days. We also performed analysis with follow-up from 90 days, in order to investigate whether this would affect the results. In total, 34 006 individuals fulfilling the above inclusion and exclusion criteria with a first DVT/PE event were initially identified. However, 1080 patients died before follow-up from 180 days, and 1337 were lost to follow-up, owing to inclusion in the last part of 2009. Moreover, patients with recurrence of any type of VTE (excluding superficial thrombophlebitis; Table S2) before 180 days (2778) were also excluded. Hence, the analyses were based on a dataset containing information on 28 811 individuals with a first event of unprovoked DVT/PE (mean age of 51.5 years [standard deviation 13], 52% men, VTE recurrence rate of 11.8%). Of these 28 811 individuals, 3991 had at least one parent diagnosed with VTE, and 1355 had at least one full sibling diagnosed with VTE. In the models, we included patient sex and age at VTE diagnosis. For all individuals, we also retrieved information from the registers on level of education. Education was divided into three categories: (i) 0–9 years (low education); (ii) 10–11 years (intermediate education); and (iii) > 12 years (high education). High education was used as the reference in the analysis. Statistical methods
Similarly to a previous study of a family history of atrial fibrillation (AF) and recurrent AF risk in patients with
308 B. Z€oller et al
lone AF [27,28], we used Cox proportional hazards models to investigate the recurrence of VTE. We started the follow-up 180 days after the first diagnosis of unprovoked DVT/PE, and followed the subjects until VTE recurrence (hospitalization for any type of VTE except for superficial thrombophlebitis; Table S2), death, emigration, or the end of the follow-up period (31 December 2009) (whichever came first). The analysis was repeated with follow-up from 90 days after the first diagnosis of unprovoked DVT/PE. In model A, parental history of VTE (yes/no) was included as a covariate. Model B also included sex, age at diagnosis of DVT/PE (centered at the mean value), and education. In model C, parental history was categorized as no parental history of VTE (reference in the model), one parent with VTE, and two parents with VTE. In model D, parental history was categorized as no parental history of VTE (reference in the model), mother with VTE, and father with VTE. Model E was similar to model B, except that it also included an interaction term for parental history of VTE and age at diagnosis of DVT/PE. Model F only included VTE in a sibling as a covariate, whereas model G also included sex, age at diagnosis of DVT/PE (centered at the mean value), and education, as well as an interaction term for VTE in a sibling and age at diagnosis of DVT/PE. In model H, sibling history of VTE was categorized into three groups: no sibling history of VTE (reference in the model), one sibling with VTE, and two or more siblings with VTE. Model I included both parental history of VTE and sibling history of VTE. In the final model, parental history and sibling history were combined into one variable, and included in the model as a single variable. All models except for models A and F included sex, age at diagnosis of DVT/PE, and education. In order to further evaluate the results, we also investigated DVT and PE separately. For both of these types of VTE, we created three separate models with three different outcomes. For all three models, we started the follow-up 180 days after the first diagnosis of DVT/PE, and followed the subjects until the diagnosis of DVT, PE, and all other types of VTE except for DVT and PE (OVTE) (Table S2), death, emigration, or the end of the follow-up period (31 December 2009) (whichever came first). In all models, we used a family history of VTE (in parents and/or sibling) as a combined variable. All models were adjusted for sex, age at first DVT/PE diagnosis, and education. To take into account the non-independence of observations from the same family, we used a robust sandwich estimator in all models. We present hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs). The proportional hazards assumption was fulfilled for the variables of interest. All calculations were performed with SAS version 9.3 (SAS Institute, Cary, NC, USA).
Results In total, 28 811 individuals with unprovoked DVT/PE were included in the study (Table 1), of whom 3404 (11.8%) suffered recurrent hospitalization for VTE. Five hundred and sixty individuals (16.5%) with recurrence and 3276 (13.5%) individuals without recurrence had a parent with VTE. Two hundred and fifteen (6.3%) individuals with recurrence and 1140 (4.5%) individuals without recurrence had a sibling history of VTE. Model A in Table 2 shows that proband cases with parents diagnosed with VTE had a 1.20-fold higher HR for recurrent VTE hospitalization than proband cases without a parental history of VTE. Follow-up was started 180 days after the first unprovoked DVT/PE event. The crude HR was not attenuated when age, sex and education level were included in the model (Table 2, model B). There was an interaction between age at diagnosis and parental history of VTE (HR 0.987) (model E), indicating a decreased association between parental VTE and recurrent VTE in older cases. Model C indicates that the HR was higher for proband cases with two affected parents (HR 1.92) than for proband cases with only one affected parent (HR 1.18). Model D shows that proband cases with mothers diagnosed with VTE had an HR of 1.16 for recurrent VTE hospitalization, and probands with affected fathers had an HR of 1.31. Figure 1 shows that, at the end of the follow-up period, ~ 21% of proband cases with a parental history of VTE had recurrent hospitalization for VTE; the corresponding value for proband cases without a parental history was 18%. Table S3 shows models A and B with follow-up from 90 days after the first unprovoked DVT/PE event. HRs Table 1 Descriptive statistics for individuals with unprovoked deep vein thrombosis (DVT)/pulmonary embolism (PE) in the Swedish population (1987–2009) DVT/PE
No recurrence
Recurrence*
N VTE in parents VTE in one parent VTE in two parents VTE in mother VTE in father VTE in one sibling VTE in two or more siblings Age (years), mean (SD) Males Females Low education Intermediate education High education Information on education
25 407 3431 3276 155 1961 1625 1140 135 51.5 12 977 12 430 8593 7640 9071 103
3404 560 (16.5) 516 (15.2) 44 (1.3) 314 (9.2) 290 (8.5) 215 (6.3) 19 (0.6) 51.6 (12) 2019 (59.3) 1385 (40.7) 1264 (37.1) 1030 (30.3) 1105 (32.5) 5 (0.1)
(13.5) (12.9) (0.6) (7.7) (6.4) (4.5) (0.5) (14) (51.1) (48.9) (33.8) (31.1) (35.7) (0.4)
SD, standard deviation. Except where otherwise stated, data are given as no. (%). *Recurrence of any type of VTE, excluding superficial thrombophlebitis. © 2013 International Society on Thrombosis and Haemostasis
Familial recurrent venous thromboembolism 309 Table 2 Results from Cox proportional hazards models for individuals with deep vein thrombosis (DVT)/pulmonary embolismPE in the Swedish population (1987–2009): parent–offspring analysis with follow-up from 180 days after the first venous thromboembolism (VTE) event 180 days DVT/PE
Model A HR (95% CI)
Model B HR (95% CI)
Model C HR (95% CI)
Model D HR (95% CI)
Model E HR (95% CI)
VTE in parents VTE in one parent VTE in two parents VTE in mother VTE in father Age at diagnosis of VTE (centered) Male vs. female Low education Intermediate education High education VTE in parents, age at diagnosis of VTE*
1.20 (1.10–1.32) – – – – – – – – – –
1.21 (1.11–1.33) – – – – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
– 1.18 (1.07–1.29) 1.92 (1.44–2.58) – – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
– – – 1.16 (1.03–1.30) 1.31 (1.16–1.48) 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
1.18 (1.07–1.30) – – – – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference 0.987 (0.980–0.995)*
CI, confidence interval; HR, hazard ratio. *The HR for VTE in a 52-year-old individual with affected parents is 1.18; the HR for VTE in a 42-year-old individual with affected parents is 1.34; and the HR for VTE in a 62-year-old individual with affected parents is 1.04. Model A included parental history of VTE. Model B also included sex (reference: female), age at diagnosis of DVT/PE (centered at the mean value), and education. In model C, parental history was categorized as no parental history of VTE (reference in the model), one parent with VTE, and two parents with VTE. In model D, parental history was categorized as no parental history of VTE (reference in the model), mother with VTE, and father with VTE. Model E contained a term for the interaction between parental history of VTE and age at diagnosis. An interaction term was only included in the model if the P-value for the interaction term was < 0.05.
© 2013 International Society on Thrombosis and Haemostasis
1.00 0.99 0.98 0.97 0.96 0.95
No family history of VTE
0.94
Family history of VTE
0.93 0.92 Survivorship function
were similar to those for follow-up from 180 days (Table 2). We also calculated HRs without exclusion of 2778 cases who were rehospitalized for VTE before the start of the follow-up at 180 days after the first VTE event (data not shown in tables). Thus, 31 589 cases were included in the follow-up from 180 days. The HR for those with affected parents was 1.19, which was similar to the HR obtained when those who were rehospitalized for VTE before the start of the follow-up from 180 days were excluded (HR 1.20, Table 2). Thus, the exclusion of 2778 cases with early recurrence before 180 days did not affect the familial risks of recurrent VTE. Model F in Table 3 shows that proband cases with at least one sibling diagnosed with VTE had a 1.30-fold higher HR for recurrent VTE hospitalization than proband cases without a sibling history of VTE. In the multivariate model B, which shows the risk of recurrent VTE associated with a sibling history of VTE, there seemed to be an interaction between age at diagnosis and a sibling history of VTE (HR 0.977) (model E), indicating a decreased association between a sibling history of VTE and recurrent VTE among older cases. Surprisingly, model H shows no significant effect on recurrent VTE risk in individuals with two affected siblings. In model I, we see that parental and sibling histories of VTE were independent risk factors for recurrent VTE. In model J, a family history (in a sibling or parent) was significantly associated with an increased risk of recurrent VTE. In Table 4, subanalysis of the familial risks (parent or sibling) of recurrent DVT, PE and OVTE among patients with unprovoked DVT or PE are presented. A family history was a risk factor for recurrent DVT, PE and OVTE among patients with unprovoked DVT or unprovoked
0.91 0.90 0.89 0.88 0.87 0.86 0.85 0.84 0.83 0.82 0.81 0.80 0.79 0.78 0
1000
2000
3000
4000 Days
5000
6000
7000
8000
Fig. 1. The survivorship function from model A. After 8000 days, ~ 21% of individuals with a parent with venous thromboembolism (VTE) had recurrence, as compared with only 18% of individuals without a parental history of VTE.
PE. A family history of VTE was a significant risk factor for all types of recurrent event in patients with a first DVT. Among patients with a first PE, a family history of VTE was a significant risk factor only for recurrence of PE, but the number of recurrent DVT and OVTE events was low. We also determined familial recurrent risk after exclusion of patients with malignancies > 1 year after the first diagnosis of VTE (similar to model A in Table 2). The familial risk after exclusion (1.22, 95% CI 1.11–1.34) was similar to that in model A in Table 2 (1.20, 95% CI 1.10–1.32). Moreover, exclusion of cases with a family history of coronary heart disease (ICD-10, I20– I25; ICD-9, 410–414) did not change the familial risks to
310 B. Z€oller et al
a major degree (1.19, 95% CI 1.09–1.31). Moreover, only 29 cases had a recurrent episode associated with pregnancy or abortion, which could not have affected the results to any major degree. Discussion The present study is, to our knowledge, the first nationwide study to estimate the familial risk of recurrent hospitalization for unprovoked VTE. Our findings suggest that a family history of VTE is a novel but modest risk factor for recurrent hospitalization for unprovoked VTE. The higher risk of recurrent unprovoked VTE hospitalization among probands with two affected parents may be genetic, although non-genetic familial factors cannot be ruled out. There was an interaction between age at diagnosis and a family history of VTE, with a family history having a stronger effect on VTE hospitalization risk in younger age groups. This further suggests a genetic contribution. Genetic risk factors for VTE may have prognostic value regarding recurrence risk. The mechanism behind the observed familial risk of recurrent unprovoked VTE hospitalization may also include non-genetic components. For instance, individuals with a family history of unprovoked VTE may have greater VTE awareness and a lower threshold for seeking healthcare. However, several genetic variants have been linked to the risk of a first VTE event [9]. It is possible that these variants may also predispose individuals to an increased risk of recurrent hospitalization for unprovoked VTE. Most previous studies have only investigated the familial risk of a first VTE event [10– 18]. Only two studies have investigated whether a family history is associated with an increased risk of recurrent unprovoked VTE [20,21]. The familial risk of recurrent unprovoked VTE hospitalization was lower than that
reported previously for the risk of first-time VTE hospitalization [10–18], suggesting that familial, and possibly genetic, factors may be more important for first-time VTE hospitalization than for recurrent unprovoked VTE hospitalization. Our results fit with studies showing that genetic variants are less important for recurrent VTE than for a first VTE event [1–3,7,9]. The lower risk of recurrent VTE than for a first VTE event among individuals with a family history of VTE could explain why the two previously published studies did not find an association between a family history and recurrent VTE risk [20,21]. The present study has a number of strengths. These include complete nationwide coverage in a country with high medical standards, and medical diagnosis of patients by specialists during examinations in hospitals. In addition, the results were not affected by recall bias, because both the probands and their relatives were medically diagnosed. Importantly, the Multigeneration Register is a validated source that has been proved to be reliable in the study of many familial diseases [22–25]. The present study also has a number of limitations. One limitation is the lack of risk factors for recurrent VTE, such as body mass index, D-dimer, and residual thrombosis [2,3]. The most important limitation is that we do not know for how long patients were treated with anticoagulants after the first unprovoked VTE event. We therefore started the follow-up analysis at 180 days after the first event. Moreover, we do not know whether recurrent VTE hospitalization was attributable to a new episode of VTE or to the same persistent episode of VTE. However, the latter alternative is unlikely, as we started survival analysis after 180 days, and progression of existing VTE is likely to occur soon after treatment is started. The risk of recurrent hospitalization was increased during the whole follow-up period, which sug-
Table 3 Results from Cox proportional hazards models for individuals with deep vein thrombosis (DVT)/pulmonary embolism (PE) in the Swedish population (1987–2009): sibling analysis with follow-up from 180 days after the first venous thromboembolism (VTE) event 180 days
Model F
Model G
Model H
Model I
Model J
DVT/PE VTE in parents VTE in siblings VTE in one sibling VTE in two or more siblings VTE in relatives (sibling or parent) Age at VTE (centered) Male vs. female Low education Intermediate education High education VTE in siblings, age at diagnosis of VTE*
HR (95% CI) – 1.30 (1.14–1.49) – – – – – – – – –
HR (95% CI) – 1.23 (1.08–1.41) – – – 1.01 (1.01–1.02) 1.44 (1.34–1.54) 1.11 (1.02–1.20) 1.07 (1.02–1.20) Reference 0.977 (0.966–0.988)
HR (95% CI) – – 1.30 (1.14–1.49) 0.92 (0.59–1.45) – 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
HR (95% CI) 1.20 (1.09–1.31) 1.22 (1.07–1.40) – – – 1.01 (1.01–1.01) 1.43 (1.34–1.53) 1.11 (1.02–1.21) 1.07 (0.98–1.17) Reference –
HR (95% CI) – – – – 1.19 (1.10–1.30) 1.01 (1.01–1.01) 1.43 (1.34–1.54) 1.12 (1.03–1.21) 1.07 (0.99–1.17) Reference –
CI, confidence interval; HR, hazard ratio. *The HR for VTE in a 52-year-old individual with affected siblings is 1.23; the HR for VTE in a 42-year-old individual with affected siblings is 1.55; and the HR for VTE in a 62-year-old individual with affected siblings is 0.91. Model F: crude HR for sibling history of VTE. Model G also included sex, age at diagnosis of DVT/PE (centered at the mean value), education, and an interaction term between VTE in siblings and age at diagnosis of DVT/PE. In model H, sibling history of VTE was categorized as no sibling history of VTE (reference in the model), one sibling with VTE, and two or more siblings with VTE. Model I included both parental history of VTE and sibling history of VTE. In the model J, parental history and sibling history were combined to give a single variable. © 2013 International Society on Thrombosis and Haemostasis
Familial recurrent venous thromboembolism 311 Table 4 Subanalysis of the importance of family history of venous thromboembolism (VTE) for patients with a first deep vein thrombosis (DVT) or pulmonary embolism (PE) event for recurrence of DVT, PE and all other types of VTE except for PE and DVT of the lower extremities (OVTE) in the Swedish population (1987–2009): follow-up from 180 days after the first VTE event
First event (N)
Type of recurrent event
HR (95% CI)*
Recurrent events (N)
Recurrent events in patients with FH, N (%)
Recurrent events in patients without FH, N (%)
DVT (14 404)† DVT (15 216)† DVT (15 196)† PE (16 146)† PE (14 505)† PE (16 222)†
DVT PE OVTE DVT PE OVTE
1.21 1.31 1.21 1.23 1.14 1.25
1368 763 360 372 1581 128
297 (11.3) 193 (7.0) 84 (3.0) 77 (2.9) 298 (12.3) 27 (1.0)
1071 (9.1) 570 (4.6) 276 (2.2) 295 (2.2) 1283 (10.6) 101 (0.8)
(1.06–1.37) (1.03–1.67) (1.06–1.37) (0.96–1.59) (1.01–1.30) (0.82–1.91)
CI, confidence interval; FH, family history of VTE in a sibling or parent; HR, hazard ratio. *Models controlled for sex, age at diagnosis of VTE, and education. †The numbers of patients included in the different analyses are different, because patients with the same early recurrent event as the respective outcome variable (DVT, PE, or OVTE) before the start of follow-up at 180 days from first-time DVT or PE were excluded.
gests that our results mainly reflect an increased risk of new VTE episodes. In addition, hospitalization for postthrombotic syndrome is unlikely to be a problem, as this diagnosis should have a main diagnosis of I870 (ICD.10) or 459B (ICD-9), and these ICD codes were excluded. Moreover, Hansson et al. [28] have previously determined the risk of recurrent hospitalization for VTE and reviewed the validity of medical records. The absence of an increased risk of recurrent VTE among those with two or more affected siblings might be attributable to the fact that these individuals have very high risks, possibly owing to multiple genetic defects, and were possibly treated indefinitely with anticoagulants after the first event [18]. However, we have no information about this. We only included a main diagnosis of VTE in hospitalized patients, in order to increase the probability that the VTE events were new thrombotic events. The validity of VTE as a primary diagnosis is 95% [26]. The lack of the less severe outpatients with VTE would be expected to diminish the risk of recurrent VTE, which is a limitation of the study. However, familial cases are probably not more likely to be hospitalized than non-familial cases, suggesting that the familial risks are not affected. Moreover, almost all cases of PE in Sweden are treated in hospitals [29], but hospital treatment of DVT in Sweden fell between 1987 and 1998, such that, in 1998, 50% of DVT patients in Sweden were treated directly as outpatients [29]. However, the increased recurrence risk for familial cases was observed during the whole follow-up period (Fig. 1). The Swedish Hospital Discharge Register only contains complete data for the period since 1987. However, the lack of data for the period before 1987 will have been similar for both cases and controls, and will not have affected the estimates of familial aggregation (i.e. the HRs). It was most likely a source of non-differential bias regarding familial risk estimates. Another potential limitation is that we do not have access to the methods used for objective diagnosis. However, the Swedish Hospital Discharge Register has high © 2013 International Society on Thrombosis and Haemostasis
validity, especially for cardiovascular disorders such as VTE (95%) [25,26]. Conclusions The present study suggests that a family history of VTE is a modest risk factor for recurrent hospitalization for unprovoked VTE in Swedish patients. The risk of recurrent hospitalization for unprovoked VTE was higher in individuals with two affected parents and in younger individuals, which suggests a genetic contribution. Addendum B. Z€ oller, H. Ohlsson, J. Sundquist, and K. Sundquist contributed to the conception and design of the study, and to the analysis and interpretation of data. J. Sundquist and K. Sundquist contributed to the acquisition of data. B. Z€ oller drafted the manuscript. All authors critically revised the paper and approved the final version. All authors had full access to all of the data (including statistical reports and tables), and take responsibility for the integrity of the data and the accuracy of their analysis. Acknowledgements The authors wish to thank the CPF’s Science Editor, Stephen Gilliver, for his useful comments on the text. The registers used in the present study are maintained by Statistics Sweden and the National Board of Health and Welfare. This work was supported by grants to B. Z€ oller from the Swedish Heart-Lung Foundation and Region Sk ane (REGSKANE-124611), to K. Sundquist from the Swedish Research Council (K2009-70X-15428-05-3 and K2012-70X-15428-08-3), and to J. Sundquist from the Swedish Council for Working Life and Social Research (2007-1754) and the Swedish Freemasons Foundation, as well as by ALF funding from Region Sk ane awarded to B. Z€ oller, J. Sundquist, and K. Sundquist.
312 B. Z€oller et al
Disclosure of Conflict of Interests The authors state that they have no conflict of interest. Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Family history of VTE was defined by hospitalization of full siblings or parents for any of the following VTE manifestations between 1964 and 2009 [13,16–18]. Table S2. VTE endpoints. Table S3. Results from Cox proportional hazard models for individuals with DVT/PE in the Swedish population (1987–2009): parent–offspring analysis with follow-up from 90 days after the first VTE event. References 1 Kyrle PA, Rosendaal FR, Eichinger S. Risk assessment for recurrent venous thrombosis. Lancet 2010; 376: 2032–9. 2 Labropoulos N, Spentzouris G, Gasparis AP, Meissner M. Impact and clinical significance of recurrent venous thromboembolism. Br J Surg 2010; 97: 989–99. 3 Ribeiro DD, Lijfering WM, Barreto SM, Rosendaal FR, Rezende SM. Epidemiology of recurrent venous thrombosis. Braz J Med Biol Res 2012; 45: 1–7. 4 Goldhaber SZ, Piazza G. Optimal duration of anticoagulation after venous thromboembolism. Circulation 2011; 123: 664–7. 5 de Jong PG, Coppens M, Middeldorp S. Duration of anticoagulant therapy for venous thromboembolism: balancing benefits and harms on the long term. Br J Haematol 2012; 158: 433–41. 6 Schulman S. Optimal duration of anticoagulant therapy. Semin Thromb Hemost 2013; 39: 141–6. 7 Zhu T, Martinez I, Emmerich J. Venous thromboembolism: risk factors for recurrence. Arterioscler Thromb Vasc Biol 2009; 29: 298–310. 8 Morange PE, Tregouet DA. Lessons from genome-wide association studies in venous thrombosis. J Thromb Haemost 2011; 9 (Suppl. 1): 258–64. 9 Rosendaal FR, Reitsma PH. Genetics of venous thrombosis. J Thromb Haemost 2009; 7(Suppl. 1): 301–4. 10 Bezemer ID, van der Meer FJ, Eikenboom JC, Rosendaal FR, Doggen CJ. The value of family history as a risk indicator for venous thrombosis. Arch Intern Med 2009; 169: 610–15. 11 Dowling NF, Austin H, Dilley A, Whitsett C, Evatt BL, Hooper WC. The epidemiology of venous thromboembolism in Caucasians and African-Americans: the GATE Study. J Thromb Haemost 2003; 1: 80–7. 12 Noboa S, Le Gal G, Lacut K, Mercier B, Leroyer C, Nowak E, Mottier D, Oger E. EDITH Collaborative Study Group. Family history as a risk factor for venous thromboembolism. Thromb Res 2008; 122: 624–9. 13 Z€ oller B, Li X, Sundquist J, Sundquist K. Parental history and venous thromboembolism: a nationwide study of age and sex-specific familial risks in Sweden. J Thromb Haemost 2011; 9: 64–70.
14 Couturaud F, Leroyer C, Julian JA, Kahn SR, Ginsberg JS, Wells PS, Douketis JD, Mottier D, Kearon C. Factors that predict risk of thrombosis in relatives of patients with unprovoked venous thromboembolism. Chest 2009; 136: 1537–45. 15 Sørensen HT, Riis AH, Diaz LJ, Andersen EW, Baron JA, Andersen PK. Familial risk of venous thromboembolism: a nationwide cohort study. J Thromb Haemost 2011; 9: 320–4. 16 Z€ oller B, Li X, Sundquist J, Sundquist K. Familial risks for unusual forms of venous thrombosis: a nationwide epidemiological study in Sweden. J Intern Med 2010; 270: 158–65. 17 Z€ oller B, Li X, Sundquist J, Sundquist K. Determination of agespecific and sex-specific familial risks for the different manifestations of venous thromboembolism: a nationwide family study in Sweden. Thromb Haemost 2011; 106: 102–12. 18 Z€ oller B, Li X, Sundquist J, Sundquist K. Age- and gender-specific familial risks for venous thromboembolism: a nationwide epidemiological study based on hospitalizations in Sweden. Circulation 2011; 124: 1012–20. 19 Prandoni P, Prins MH, Ghirarduzzi A, Pengo V, Sartori MT, Ugolotti MC, Bracco A, Veropalumbo MR, Noventa F, Lensing AW. Family history of venous thrombosis or sudden death as a risk factor for venous thromboembolism. Thromb Haemost 2012; 107: 1191–2. 20 Hron G, Eichinger S, Weltermann A, Minar E, Bialonczyk C, Hirschl M, Stain M, Gartner V, Kyrle PA. Family history for venous thromboembolism and the risk for recurrence. Am J Med 2006; 119: 50–3. 21 Gauthier K, Kovacs MJ, Wells PS, le Gal G, Rodger M; REVERSE investigators. Family history of venous thromboembolism (VTE) as a predictor for recurrent VTE in unprovoked VTE patients. J Thromb Haemost 2013; 11: 200–3. 22 Statistics Sweden. The Swedish Multigeneration Register (1960– 1990): Registret o€ver totalbefolkningen/RTB. In Swedish. Stockholm: Statistics Sweden; 2005. 23 The National Board of Health and Welfare. Validity of the Diagnoses from the Swedish In-Care Register 1987 and 1995. In Swedish. Stockholm: Epidemiologiskt Centrum, Socialstyrelsen; 2000. 24 Rosen M, Hakulinen T. Use of disease registers. In: Handbook of Epidemiology. Berlin: Springer-Verlag, 2005: 231–51. 25 Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, Heurgren M, Olausson PO. External review and validation of the Swedish national inpatient register. BMC Public Health 2011; 11: 450. [online]. 26 Rosengren A, Freden M, Hansson PO, Wilhelmsen L, Wedel H, Eriksson H. Psychosocial factors and venous thromboembolism: a long-term follow-up study of Swedish men. J Thromb Haemost 2008; 6: 558–64. 27 Z€ oller B, Ohlsson H, Sundquist J, Sundquist K. Family history as a risk factor for recurrent hospitalization for lone atrial fibrillation: a nationwide family study in Sweden. BMC Cardiovasc Disord 2012; 12: 121. [online]. 28 Hansson PO, S€ orbo J, Eriksson H. Recurrent venous thromboembolism after deep vein thrombosis: incidence and risk factors. Arch Intern Med 2000; 160: 769–74. 29 Statens beredning f€ or medicinsk utv€ardering (SBU). Blodpropp – F€ orebyggande, Diagnostik Och Behandling av Ven€ os Trombembolism SBU-Rapport 158. In Swedish. Stockholm: The Swedish Council on Technology Assessment in Health Care; 2002 I–III.
© 2013 International Society on Thrombosis and Haemostasis
