Zöller B, Li X, Sundquist J, Sundquist K. (Center for Primary Health Care Research, Lund University, Malmö, Sweden; Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, CA, USA). Familial risks of unusual forms of venous thrombosis: a nationwide epidemiological study in Sweden. J Intern Med 2011; 270: 158–165.
Objective. This is the first nationwide study to determine familial risks of unusual forms of venous thrombosis amongst offspring of affected parents and amongst siblings.
Design and settings. The Swedish Multigeneration Register of 0- to 75-year-old subjects was linked to the Hospital Discharge Register for the period 1987–2007. Standardized incidence ratios (SIRs) were calculated for individuals whose relatives were hospitalized for venous thromboembolism (VTE), as determined by the International Classification of Diseases, compared to those whose relatives were not affected by VTE.
Results. The total number of hospitalized patients with VTE was 45 362, of which 1824 (4.0%) were affected by a rare thrombotic condition. The familial SIRs in cases with a history of VTE in parents or siblings were significantly increased for migrating thrombophlebitis (1.81; 95% confidence interval (CI) 1.40–2.31), portal vein thrombosis (2.35; 95% CI 1.77–3.06), vena cava thrombosis (1.96; 95% CI 1.42–2.64) and cerebral venous thrombosis (1.74; 95% CI 1.30–2.28). Budd–Chiari syndrome (SIR, 0.92; 95% CI 0.24–2.38) and renal vein thrombosis (SIR, 1.72; 95% CI 0.62–3.77) were not significantly associated with parental or sibling history of VTE; however, these two conditions were very rare, and therefore, we cannot draw any definite conclusions from this finding.
Conclusions. Family history is an important risk factor for most unusual forms of VTE. Moreover, even the paraneoplastic phenomenon, migrating thrombophlebitis (Trousseau’s syndrome), is associated with a family history of VTE. Thus, our data suggest that most rare forms of VTE have a familial background.
Venous thromboembolism (VTE) is a multicausal disorder that results from multiple interactions between acquired and inherited risk factors . The most common manifestations are superficial thrombophlebitis, deep venous thrombosis (DVT) and pulmonary embolism (PE) [1–3]. The most important known genetic risk factors for venous thrombosis are deficiencies of the natural anticoagulant proteins antithrombin, protein C and protein S, activated protein C resistance owing to a point mutation in the factor V gene (factor V G1691A or factor V Leiden) and the prothrombin G20210A mutation, although many weaker genetic risk factors have been identified [1, 4]. The predictive value of family history for finding any of these genetic risk factors is, however, low [5, 6], suggesting that other genetic or non-genetic familial factors might be important.
Many rare forms of VTE have been described, including migrating thrombophlebitis (Trousseau’s syndrome when associated with visceral cancer) [7, 8], Budd–Chiari syndrome (hepatic vein occlusions) and thromboses of the vena cava and cerebral, portal, mesenteric, renal and retinal veins [9–14]. Many of these disorders are associated with specific, acquired risk factors such as nephrotic syndrome in renal vein thrombosis , myeloproliferative disorders in Budd–Chiari syndrome , liver cirrhosis or hepatocarcinoma in portal vein thrombosis , cancer and central venous catheters in vena cava thrombosis  and head injury, mastoiditis or otitis in cerebral vein thrombosis . The most well-known example of a rare form of VTE is Trousseau’s syndrome. In 1865, Armand Trousseau noted that migratory thrombophlebitis could be a forewarning of an occult visceral malignancy [7, 8]. Deficiencies of antithrombin, protein C and protein S have been described in some patients with rare thrombotic disorders . In some rare thrombotic conditions, an association has been made in case–control studies with factor V Leiden and prothrombin G20210A mutations reviewed by Martinelli et al. . These two mutations have been associated with an increased risk of cerebral vein  and portal vein thrombosis . Budd–Chiari syndrome has been associated with factor V Leiden . The role of genetic risk factors such as factor V Leiden and prothrombin G20210A mutations in migrating thrombophlebitis  and thrombosis of the vena cava , renal vein  and retinal vein [9, 14] has not been established.
The importance of family history as a risk factor for unusual forms of VTE has not been determined. We have therefore estimated the relative importance of family history for different rare manifestations of VTE. Another novel aspect of this study is its design: it was based on a nationwide register of all hospitalizations in Sweden between 1987 and 2007. Use of hospitalizations eliminated potential self-report and recall bias. The Swedish family data set, the Multigeneration Register, is a validated source that has been proven to be reliable in the study of many familial diseases [25–29]. In this nationwide study, we have analysed, for the first time, the familial risks of hospitalization for unusual manifestations of VTE amongst offspring of affected parents and amongst siblings.
Material and methods
This study was approved by the Ethics Committee of Lund University, Sweden. The VTE research database used for this study was constructed by linking several national Swedish registers, based on the MigMed2 data sets at the Center for Primary Health Care Research, Lund University, Malmö, Sweden. Statistics Sweden, the Swedish government-owned statistics bureau, provided data for the Multigeneration Register in which individuals (second generation) born in Sweden from 1932 are registered shortly after birth and linked to their parents (first generation). Families could be defined by linking all siblings to their parents. The second generation represented the present study population. Linkages were made to national census data to obtain information about socio-economic status. The final link was made by including data from the Swedish Hospital Discharge Register, which records complete information on all discharges with dates of hospitalization and diagnoses since 1986. The centralized Swedish Patient Registry has a very high coverage , and its scientific value has been confirmed in a number of studies, including those of VTE [31–33]. All linkages were made by using the unique individual national identification number assigned to each person in Sweden. This number was replaced by a serial number for each person to provide anonymity. The serial number was used to check that each individual was only entered once on first appearance with a VTE diagnosis. More than 11.8 million individuals in 3.9 million families were included in this database with 8.9 million belonging to the second generation in which the oldest, who were born in 1932, had reached the age of 75 at the end of the 20-year follow-up in 2007 [25–29]. Data in the MigMed2 database are almost complete; in 2001, personal identification numbers were missing in only 0.4% of hospitalizations and the main diagnosis in 0.9% [25–29]. Information on occupational status, retrieved from the national census records in the database, was 99.2% complete [25–29].
Patients with VTE were identified from hospital discharge data, reported according to the ninth (1987–1996) and 10th (1997–2007) versions of the International Classification of Diseases (ICD-9 and ICD-10, respectively). In agreement with Souto et al. , VTE was defined as not only DVT and PE but also superficial venous thrombosis (SVT) and other forms of venous thrombosis. This definition of VTE is common in studies of familial thrombotic risk of defined genetic defects [35–37]. Moreover, this definition of VTE avoided the problem of some ICD-9 and ICD-10 codes being nonspecific and not always allowing for localization of VTE or differentiation between SVT and DVT. VTE was defined as the following ICD-9 and ICD-10 codes: PE (ICD-9: 415B, 416W and ICD-10: I26), superficial or deep phlebitis or thrombophlebitis (ICD-9: 451 and ICD-10: I80), portal vein thrombosis (ICD-9: 452 and ICD-10: I81), other venous embolism or thrombosis (ICD-9: 453 or ICD-10: I82), cerebral vein thrombosis and cerebral infarction owing to cerebral vein thrombosis (ICD-9: 437G and ICD-10: I636, I676), pregnancy-related VTE (ICD-9: 671C, 671D, 671E, 671F, 671X, 673C and ICD-10: O222, O223, O225, O228, O229, O870, O871, O873, O879, O882), or abortion-related VTE (ICD-9: 639G and ICD-10: O082, O087). There is no official translation from ICD-9 to ICD-10. For this reason, DVT before delivery (code 671C in ICD-9) was defined by a grouping of ICD-10 codes (O223, O228, O871), chosen to correspond as closely as possible to the previous code 671C. This is in agreement with the Swedish National Board of Health and Welfare ICD-9 to ICD-10 translator. A total of 45 362 patients were identified for inclusion in the study, based on their first discharge recorded in the Hospital Discharge Register.
To analyse the familial risk of different rare manifestations of VTE, the following specific subtypes were identified: migrating thrombophlebitis (ICD-9: 453B and ICD-10: I821), cerebral venous thrombosis (ICD-9: 437G, 671F and ICD-10: I636, I676, O225, O873), portal vein thrombosis (ICD-9: 452 and ICD-10: I81), vena cava thrombosis (ICD-9: 453C and ICD-10: I822), renal vein thrombosis (ICD-9: 453D and ICD-10: I823) and Budd–Chiari syndrome (ICD-9 453A and ICD-10 I820). Only the main diagnosis in the Hospital Discharge Register was considered.
Individual variables included in the analysis
Data were analysed according to sex. Socio-economic status for both men and women was divided into six groups according to occupation: (i) farmers, (ii) unskilled/skilled workers, (iii) white-collar workers, (iv) professionals, (v) self-employed individuals and (vi) all others [25–28]. Geographical region was divided into three groups: (i) large cities (i.e. Stockholm, Gothenburg and Malmo), (2) southern Sweden and (3) northern Sweden, enabling adjustments for regional differences in hospitalization [25–29].
Person-years were calculated from the start of follow-up on 1 January 1987 until hospitalization for VTE, death, emigration or study end (31 December 2007). Age-adjusted incidence rates were calculated for the complete duration of follow-up and divided into 5-year periods. Standardized incidence ratios (SIRs) were calculated as the ratio of the observed (O) to the expected (E) number of cases . The expected number of cases was calculated according to sex-, period (5-year groups)-, region- and socio-economic status-specific standardized incidence rates. Familial risks in offspring of affected parents and in affected siblings were calculated for men and women, compared with those whose parents or siblings were not affected by these conditions, using the cohort method as previously described . Confidence intervals (95% CI) were calculated assuming a Poisson distribution . Age-adjusted incidence rates were calculated based on the European standard population .
We analysed the familial risks of rare thrombotic events in offspring and siblings (aged 0–75) of individuals hospitalized for VTE in Sweden between 1987 and 2007. The total follow-up period for all individuals was 132 222 639 person-years. Amongst men, the follow-up periods were 64 785 879 and 2 873 433 person-years for those without and with a parental history of VTE, respectively. Amongst women, the follow-up periods were 61 815 359 and 2 747 968 person-years for those without and with a parental history of VTE, respectively. The follow-up period was 66 868 412 person-years for men without a sibling history of VTE and 790 900 person-years for men with a sibling history of VTE. The corresponding follow-up periods were 63 790 722 and 772 605 person-years for women without and with a sibling history of VTE, respectively. Amongst a total of 45 362 hospitalized offspring of parents with VTE, 1824 (4.0%) were affected by the following rare thrombotic manifestations: migrating thrombophlebitis, portal vein thrombosis, vena cava thrombosis, cerebral venous thrombosis, Budd–Chiari syndrome and renal vein thrombosis. Table 1 shows the sex-specific hospitalized incidence rates for men and women. All incidence rates were lower than one per 100 000 person-years for both women and men. No sex differences in incidence rates were observed except for cerebral vein thrombosis, which was significantly less frequent amongst men than among women. Both Budd–Chiari syndrome and renal vein thrombosis were very rare at ≤0.1 per 100 000 person-years.
|ICD-9||ICD-10||n||IR (per 100 000)||95% CI||n||IR (per 100 000)||95% CI|
|Cerebral venous thrombosis||437G, 671F||I636, I676, O225, O873||199||0.3||0.3||0.3||335||0.5||0.5||0.6|
|Portal vein thrombosis||452||I81||205||0.3||0.3||0.3||160||0.2||0.2||0.3|
|Vena cava thrombosis||453C||I822||153||0.2||0.2||0.3||176||0.3||0.2||0.3|
|Renal vein thrombosis||453D||I823||35||0.1||0.0||0.1||26||0.0||0.0||0.1|
Amongst a total of 1824 hospitalized cases with unusual forms of venous thrombosis, 169 (9.3%) had a parental history of VTE. Table 2 shows the familial SIRs for offspring with a parent affected by VTE. No significant differences were observed between men and women. Familial SIRs for offspring with a parent affected by VTE were significantly increased for all men and women, for all studied conditions except renal vein thrombosis and Budd–Chiari syndrome (Table 2). These two thrombotic disorders were, however, very rare. The highest familial SIR was noted for portal vein thrombosis. However, the differences in familial SIRs between the different conditions were not significant.
|Subtype of VTE||Men||Women||All|
|O||SIR||95% CI||O||SIR||95% CI||O||SIR||95% CI|
|Cerebral venous thrombosis||14||1.41||0.77||2.36||26||1.86||1.21||2.73||40||1.67||1.19||2.28|
|Portal vein thrombosis||24||2.32||1.49||3.46||17||2.13||1.24||3.42||41||2.24||1.61||3.04|
|Vena cava thrombosis||14||1.62||0.89||2.73||20||2.34||1.42||3.61||34||1.98||1.37||2.77|
|Renal vein thrombosis||2||1.30||0.12||4.79||2||1.66||0.16||6.11||4||1.46||0.38||3.77|
Amongst the 1824 hospitalized cases with unusual forms of venous thrombosis, 66 (3.6%) had a sibling with a history of VTE. Table 3 shows the familial SIRs for the different subtypes of VTE following diagnosis of a sibling with any type of VTE. Although there was a tendency towards similar increases in familial SIRs amongst siblings and offspring of affected individuals, significantly increased risks were observed only for portal vein thrombosis for both men and women, and for renal vein thrombosis amongst women (Table 3). However, only three familial cases of renal vein thrombosis were observed.
|Subtype of VTE||Men||Women||All|
|O||SIR||95% CI||O||SIR||95% CI||O||SIR||95% CI|
|Cerebral venous thrombosis||5||1.55||0.34||5.14||8||1.99||0.60||5.58||13||1.79||0.67||4.35|
|Portal vein thrombosis||10||2.70||0.91||7.06||7||2.47||0.69||7.25||17||2.60||1.07||5.91|
|Vena cava thrombosis||6||2.13||0.54||6.59||4||1.30||0.24||4.74||10||1.69||0.57||4.42|
|Renal vein thrombosis||0||3||8.02||1.07||33.59||3||3.21||0.43||13.45|
Table 4 shows the familial risk following diagnosis of a parent or sibling with any type of VTE. Amongst the 1824 hospitalized cases in this study, 225 (12.3%) had parental or sibling histories of VTE. All unusual subtypes were associated with a significantly increased familial risk except for the very rare Budd–Chiari syndrome and renal vein thrombosis. Increased risks of migrating thrombophlebitis, portal vein thrombosis and vena cava thrombosis in both men and women were observed. The familial risk of cerebral vein thrombosis was significantly increased only in women.
|Subtype of VTE||Men||Women||All|
|O||SIR||95% CI||O||SIR||95% CI||O||SIR||95% CI|
|Cerebral venous thrombosis||19||1.50||0.90||2.35||33||1.91||1.31||2.69||52||1.74||1.30||2.28|
|Portal vein thrombosis||32||2.41||1.64||3.40||23||2.28||1.44||3.43||55||2.35||1.77||3.06|
|Vena cava thrombosis||20||1.86||1.13||2.87||23||2.05||1.30||3.09||43||1.96||1.42||2.64|
|Renal vein thrombosis||2||0.96||0.09||3.53||4||2.84||0.74||7.35||6||1.72||0.62||3.77|
The results of this study demonstrate, for the first time, an association between family history of VTE and several rare forms of venous thrombosis. Previously, family history has been associated with such common manifestations of VTE as DVT in the lower limbs and PE [6, 41, 42]. Our data suggest that most rare forms of venous thrombosis (i.e. migrating thrombophlebitis and thrombosis of the vena cava and cerebral and portal veins) have a familial background too. We cannot draw any firm conclusions about the familial risks of Budd–Chiari syndrome and renal vein thrombosis because of their rareness and the small numbers of individuals with these conditions in this study; however, an association with family history amongst female siblings was observed for renal vein thrombosis.
Our finding that familial factors are important is in agreement with previous studies that have found an association between factor V Leiden and cerebral vein thrombosis, portal vein thrombosis and Budd–Chiari syndrome [13, 20–22]. Other studies have found an association between the prothrombin G2010A mutation and thrombosis of the cerebral and portal veins [13, 20, 21]. A genetic or familial contribution to migrating thrombophlebitis (Trousseau’s syndrome when associated with malignancy) has not been demonstrated previously [7, 8]. Migrating thrombophlebitis is intimately associated with malignant disease. However, our data show that familial and probably genetic factors are also predisposing risk factors in this paraneoplastic manifestation. This is in agreement with a report suggesting that the factor V Leiden and the prothrombin G2010A mutations are additional thrombotic risk factors in patients with cancer . The same is true for vena cava thrombosis, which is often associated with cancer . Our study is the first to demonstrate a familial and possible genetic contribution to this thrombotic manifestation, although genetic prothrombotic risk factors have been described in a small study of childhood vena cava thrombosis . Thus, our data reinforce the concept that venous thrombosis is a multicausal disease involving multiple interactions of acquired and genetic risk factors .
The strengths of the study include a complete nationwide coverage in a country with a high standard of medical diagnosis of patients, usually by a specialist during extensive examinations in the clinic. An important strength of this nationwide register-based study is that the results are not affected by recall bias because both the probands and cases were medically diagnosed. Also, the Swedish family data set (i.e. the Multigeneration Register) is a validated source that has been proven to be reliable in the study of many familial diseases [25–29]. Data in the MigMed2 database are almost complete.
There are some potential limitations concerning the present study. However, most of the limitations involve a nondifferential bias, and the design of this study has been used successfully in a large number of other studies to determine familial risks of many complex diseases [25–29]. The Swedish Hospital Discharge Register only contains complete data since 1987. Thus, the present study covered a period of 20 years, which implies that pedigree charts would not be very informative. This is compensated for by a total follow-up period of 132 222 639 person-years. If familial and nonfamilial VTE cases before 1987 are lost in proportion to familial risks, which must be assumed to be the case, this is a nondifferential bias. Moreover, published studies estimating familial risks based on a history of VTE amongst relatives will probably suffer from recall bias regarding events occurring many years ago, but our approach is not biased by this potential problem. In a study by Couturaud et al., VTE was observed in a first-degree relative in 5.3% of patients with venous thrombosis, which is much lower than in the present study in which 12.3% of cases of unusual venous thrombosis were familial. This difference might be attributable to recall or inclusion bias in the study by Couturaud et al. and emphasizes the advantages of our study design. Our estimates of familial SIRs for the rare manifestations of VTE are not very different from previously reported odds ratios of between 2.2 and 2.7 for the risk of venous thrombosis owing to a family history of VTE [6, 41, 42], suggesting that our results give a valid estimate of the familial risks.
Another potential limitation is that the Swedish Hospitalization Discharge Register does not contain information on diagnostic procedures. However, the overall diagnostic validity of the Swedish Inpatient Register is close to 90% . Moreover, the diagnosis of VTE in the Swedish Inpatient Register has been validated and found to be correct in 95% of VTE diagnoses . In addition, in 91% of cases, VTE is diagnosed by an objective method . This latter finding is in agreement with three recent Swedish studies demonstrating that VTE is seldom diagnosed without objective investigations [47–49]. As it is possible that the diagnostic accuracy could have varied between geographical regions, the analysis was adjusted to minimize this possible bias. The high validity of the Swedish Inpatient Register for VTE is in agreement with previous estimates of a validity of about 95% for stroke and myocardial infarction [50–52]. The concordance between the hospital discharge diagnosis and the underlying causes of death of those who were hospitalized and later died under dramatic conditions was 89% . Only a main diagnosis of VTE was used in the present study to further increase the accuracy.
Our data do not include major risk factors. In an attempt to compensate for this, we adjusted for socio-economic status (occupation) in the models. We therefore cannot compare the importance of familial history of VTE in provoked and unprovoked cases. However, according to a recent study, family history is important in both provoked and unprovoked VTE . This is in agreement with familial studies of defined genetic defects, in which approximately 50% of cases involving a first thrombotic event are unprovoked [35–37]. Acquired risk factors are usually not taken into account when considering the overall importance of a family history of VTE [6, 41, 42], although familial risks have been estimated in unprovoked VTE cases by Couturaud et al. .
Another possible limitation is that only hospitalized patients were included in the present study, as outpatient data were not available. However, most unusual manifestations of VTE would be treated in hospital. Moreover, our estimates of incidence rates for portal vein thrombosis, Budd–Chiari syndrome, cerebral vein thrombosis and vena cava thrombosis are in agreement with previously estimated values (Table 1) [13, 18, 54]. Also in agreement with others , we confirmed a significantly higher incidence rate amongst women than amongst men for cerebral vein thrombosis. The incidence of migrating thrombophlebitis has not been reported previously. Our estimated incidence rate for this condition could be too low as we used only the main diagnosis (to increase the accuracy), and patients with this diagnosis will sometimes also be treated as outpatients. Another possible source of bias is changing incidences and hospitalization rates of VTE over time. However, incidence rates were calculated for the total follow-up duration, divided into five 5-year periods, and adjustments were made for the possibility of changes in the incidence rate over time.
In conclusion, familial history of VTE is an important risk factor for most unusual forms of venous thrombosis. Family history is a potentially useful risk indicator for the development of unusual forms of thrombosis in certain clinical conditions known to predispose to a specific thrombotic disorder. In addition, thrombotic disorders that are intimately associated with malignancies, such as portal vein thrombosis, vena cava thrombosis and migrating thrombophlebitis, are associated with a family history of VTE. Thus, most rare forms of venous thrombosis have a familial background.
The registers used in this study are maintained by Statistics Sweden and the Swedish National Board of Health and Welfare. This work was supported by grants to Drs Kristina and Jan Sundquist from the Swedish Research Council (2008-3110 and 2008-2638), the Swedish Council for Working Life and Social Research (2006-0386, 2007-1754 and 2007-1962) and the Swedish Research Council Formas (2006-4255-6596-99 and 2007-1352).
All authors contributed to the conception and design of the study, JS and KS contributed to the acquisition of data, all authors contributed to the analysis and interpretation of data, BZ drafted the manuscript and all authors revised it critically and approved the final version.
Disclosure of conflicts of interest
We declare that we have no conflicts of interest.