Postoperative Joint Replacement Complications in Swedish Patients With a Family History of Venous Thromboembolism

Key Points Question Is a family history of venous thromboembolism associated with postoperative venous thromboembolism and major bleeding in patients with primary hip or knee replacement surgical procedures? Findings In this cohort study of 69 505 Swedish patients, a family history of venous thromboembolism was associated with statistically significant increased venous thromboembolism risk and reduced bleeding risk after hip and knee replacement surgical procedures. The heritability (SE) for postoperative venous thromboembolism was 20% (6%). Meaning Familial and most likely genetic factors appear to affect venous thromboembolism and major bleeding risk following hip and knee replacement surgical procedures.


Introduction
Hip and knee replacement operations are generally safe procedures but do carry a small risk of serious complications. 1 The risk of venous thromboembolism (VTE) and major bleeding with pharmacologic thromboprophylaxis has been studied extensively. 2 VTE risk has been estimated to be 1.15% within 90 days of postoperative care in contemporary trials since 2003. 2 After 90 days, the VTE risk returns to the baseline risk before surgery. 2 Major bleeding rates have been reported to be between 0.1% and 3.1% in prevention trials of patients undergoing hip arthroplasty and between 0.2% to 1.4% for patients undergoing knee arthroplasty. 3 A large population-based study found a total VTE risk of 1.3% and a major bleeding risk of 0.6%. 4 Although studies have examined predictors associated with VTE, data on predictors associated with major bleeding are sparse. [4][5][6] The involvement of genetic factors in postoperative thrombosis is not clear. A study by Svensson et al 7 found an association between factor V Leiden (FVL), single-nucleotide polymorphism rs6025, and VTE risk among patients undergoing hip arthroplasty with short-term low-molecular-weight-heparin (LMWH) prophylaxis during hospitalization but not in those patients with prolonged LMWH prophylaxis (exactly 3 weeks). A study by Wåhlander et al 8 of patients with a hip or knee replacement surgical procedure found an increased risk for VTE in patients with the prothrombin gene G20210A variant (singlenucleotide polymorphism rs1799963) but not in patients with FVL. However, a study by Ryan et al 9 found no association between FVL and VTE in patients after joint replacement surgical procedures.
Thus, the results of genetic studies for postoperative VTE are divergent.
Family history of VTE (FH-VTE) is a risk factor in first-degree relatives and is associated with 2 to 3 times the increased familial relative risk. 10,11 However, to our knowledge, FH-VTE has only been investigated in joint replacement surgical procedures in 1 small study (13 VTE cases). 12 Familial aggregation represents the sum of shared family environmental and genetic factors. 10,11 However, previous studies indicate a weak involvement of shared environmental factors to the familial aggregation of VTE. 10,11 The FH-VTE is therefore an important research tool and an important risk factor for VTE reflecting a potential genetic predisposition. 10,11 In this nationwide Swedish cohort study, we examined the risk of VTE and major bleeding associated with an FH-VTE within 90 days following primary hip and knee replacement surgical procedures.

Family History of VTE
The FH-VTE predictor variable was defined as the diagnosis of VTE in a parent or full sibling between 1964 and the surgical procedure index date. The IPR, OPR, and Cause of Death Register were used to define VTE in parents and full siblings. The ICD diagnosis codes used to define FH-VTE are listed in eTable 1 in the Supplement.

Adjusting Variables
Patient age at the discharge date was treated as a continuous variable. Augmented Charlson Comorbidity Index (aCCI) was defined by ICD-10 codes within 4.5 years before the surgical procedure index date (eTable 3 in the Supplement). 24,25 Patients with aCCI = 0 have no disease, and patients with aCCI = 1 have only 1 disease with the weight = 1. Patients with aCCI = 2 have 2 or more points (ie, between 2 and 37 points). Educational level was dichotomized as fewer than 12 years of education vs 12 or more years of education. Among patients, there were 150 of 69 505 (0.2%) individuals who had not been registered with any form of education, and they were regarded as having the lower educational level.

Statistical Analysis
Cox proportional hazards regression was used to determine the hazard ratio (HR) of VTE and major bleeding in individuals with FH-VTE compared with those with no family history (NFH) of VTE within Heritability is defined as a ratio of variances (ie, the proportion of total variance because of variation in additive genetic factors). 27 The heritability of a binary trait could be estimated using

JAMA Network Open | Cardiology
Falconer regression by presuming a liability threshold model of the disease (ie, whereby everyone has a liability to develop the disease, but only individuals above a threshold value do so). [27][28][29] To evaluate heritability for postoperative VTE, Falconer regression was used. 28,29 Using the prevalence rate of the relatives of the first-degree probands (ie, first-degree relatives to affected patients) and the controls (ie, first-degree relatives to unaffected patients) from the case-control study, the heritability (SE) was calculated with the assumption that the contribution of shared environment factors is negligible. 10,11,22,29 To estimate heritability based on the Falconer method, we used a case-control exact matching method (1:5) by drawing a sample of affected patients as cases with matched control groups of unaffected patients. 30 The control groups were matched based on sex, birth year, number of full siblings in each family, and educational level. In the case-control study, both groups were linked to their first-degree relatives (parents and full siblings) that were used to calculate postoperative heritability. 29 Falconer describes the calculation of heritability in detail. 29

Heritability
A matched (1:5) case-control study was performed to determine the heritability for postoperative VTE (782 cases and 3910 controls) and major bleeding (1245 cases and 6225 controls). Postoperative VTE was associated with VTE in the first-degree relatives (parent and/or full sibling), with an odds ratio of 1.34 (95% CI, 1.12 to 1.59) in patients with at least 1 affected first-degree relative.
Postoperative bleeding was significantly associated with at least 1 affected first-degree relative, with an odds ratio of 0.83 (95% CI, 0.71-0.97). By using the Falconer method, heritability was estimated based on VTE in the first-degree relatives. The heritability (SE) for VTE calculated from the casecontrol study was 20% (6%).

Discussion
The present study linked FH-VTE to an increased incidence of VTE and decreased risk of major bleeding following the surgical procedures. Hip or knee replacement surgical procedures increased the risk for VTE, but LWMH treatment postoperatively for 7 to 10 days had considerably reduced the postoperative VTE risk. Previously, there have been contradictory data about whether genetic risk factors for VTE (ie, rs6025, rs1799963, and ABO blood type variants) are also risk factors for postoperative VTE. [7][8][9] The present study suggested that familial and probably genetic factors including common risk variants (rs6025 and rs1799963) were risk factors for postoperative VTE. The present study explained why a previous study did not find an association with FVL and venography-diagnosed VTE after 1 week. 9 The findings were that FH-VTE, and presumably genetic factors, affected the risk of VTE to a larger degree only after more than 7 days after discharge. This conclusion was similar to a study by Svensson et al 7 who found that FVL was associated with VTE risk only in patients treated for 1 week with LMWH and not in those patients treated for 3 weeks with LMWH. In Sweden, LMWH prophylactic treatment was provided for 7 to 10 days and not extended (Ն3 weeks). FVL (rs6025) and the prothrombin variant (rs1799963) are the most common known strong genetic risk factors for VTE in Sweden that are linked to FH-VTE. Especially FVL might have contributed to our findings because of its high frequency in Sweden. 7,10,11 The importance of ABO blood type is less well studied in Sweden, but non-O blood type is associated with an approximately 2-fold increased risk of VTE. 11 It possibly could have been worthwhile to screen for the rs6025, rs1799963, and ABO risk variants and to give those patients who were carriers extended LMWH prophylaxis treatment for 3 weeks. Missense PROS1 variants translating into low protein S levels and/or low protein S activity were more common in the general Swedish population than previously anticipated and might have contributed to the situation. 31 However, it was also possible to extend prophylaxis in all patients with FH-VTE without genetic analysis because known variants only explain 30% of FH-VTE. 32 Another finding in the present study was that FH-VTE protected against bleeding. Most bleeding events occurred prior to discharge; FH-VTE was not protective against bleeding after 1 week from discharge (ie, not after LMWH treatment was stopped). The present study was the first study, to our knowledge, to show an evolutionary advantage with FH-VTE. Previously, FVL had been shown to protect against intrapartum and menstrual blood loss. 33,34 Our findings suggest that the genes collectively associated with FH-VTE conferred a procoagulant and evolutionary advantage in situations with trauma to the body. Speculatively, an evolutionary advantage of FH-VTE with less bleeding may have occurred after an attack from a wild animal or a fellow human or after accidental trauma, with a higher chance to survive and achieve reproduction.
A major part of the association of FH-VTE was because of genetic and not shared familial environmental factors. 10,11 Thus, genes were important even in thrombogenic situations, such as joint replacement surgical procedures, with a heritability of 20%. This result was lower than estimated in young individuals and in selected families with inherited VTE. 10,11,22 It is therefore possible that the heritability for postoperative VTE was higher in younger individuals and in other less thrombogenic types of operations than joint replacement surgical procedures.

Strengths and Limitations
The strength of our study was the large study size using validated nationwide registers provided by A limitation was the lack of biological data such as obesity, patient height, and smoking status.
However, we adjusted for educational level that was related to lifestyle factors. 35  Another potential limitation was use of the aCCI, which may not always adequately adjust for all risk factors for VTE and major bleeding. However, the aCCI allowed adjustment for the sum of a large number of comorbidities that would not have been possible otherwise. There was a tendency for fewer comorbidities and lower educational attainments among those with FH-VTE ( and families was more powerful. However, we knew from previous studies that genetic factors made a strong contribution to the familial transmission of VTE. 10,11,22,[36][37][38][39] In the present study, family history was used as a proxy for genetic defects, but the question remained as to which defects were responsible for the outcomes. For instance, FVL was common in Sweden. 40

Conclusions
Family history of VTE appears to have predictive associations with postoperative VTE, and FH-VTE might also have provided protection against major bleeding after joint replacement surgical procedures. This study suggests that extended LMWH treatment might be beneficial in genetically predisposed individuals. We also hypothesize a possible evolutionary advantage of prothrombotic genes protecting against traumatic bleeding.