The Risk of Venous Thromboembolism in Surgically Treated Hip Fracture: A retrospective cohort study of 5184 patients.

ABSTRACT Introduction: Elderly patients undergoing surgery for a hip fracture are at risk of venous thromboembolism (VTE). The known risk of VTE is low due to thromboprophylaxis (1.8-2.5% at three months). Most previous studies have ignored the high mortality in that population when calculating the risk of VTE. Hip fracture treatment has also evolved over the last decade. The real risk of VTE today is unknown. We asked what was the risk of symptomatic VTE following surgery for a hip fracture in an elderly population. We also asked what kind of bleeding events were associated with thromboprophylaxis. Hypothesis: The risk of VTE will be lower than previously described due to modern postoperative care. Patients and Methods: Retrospective cohort study of all patients ≥ 65 years old undergoing surgery for a hip fracture in two Canadian academic centers, between January 1, 2008, and January 1, 2019. Symptomatic VTE (pulmonary embolism or deep venous thrombosis) confirmed by imagery were assessed. The follow-up was fixed at 3 months. The cumulated risks of VTE and bleeding events were calculated using the Kaplan-Meier estimator and a logistic regression model was used to determine risk factors. Results: The cohort included 5184 patients. The mean age was 83 ± 8 years old and 76% of patients were female. 98.8% of this cohort received postoperative thromboprophylaxis. Low-molecular-weight heparin was given for 35 days in 87% of this cohort. The risk of venous thromboembolism was 4.7 ± 0.5% at 3 months (N=144). Chronic obstructive pulmonary disease (odds ratio 1.6 [1.0-2.4]) and the use of warfarin as extended thromboprophylaxis (odds ratio 2.1 [1.3-3.6]) were associated with venous thromboembolism. The risk of bleeding was 5.9% (N=179) at 3 months. 78% (N=141) of bleeding events were hematomas, of which only 12% (N=16) needed reoperation. The use of direct oral anticoagulant was associated with an increased risk of bleeding events (odds ratio, 2.8 [1.5-5.0]). Mortality at 3 months was 8.4%. Discussion: The risk of venous thromboembolism is higher than expected in a population treated for this condition (4.7% vs 1.8-2.5% at 3 months as previously described). Bleeding events were mostly hematomas and few needed reoperations for wound complication. Future research should focus on the management of thromboprophylaxis in that population. Level of evidence: III; retrospective cohort study