• arthroplasty, replacement, hip;
  • arthroplasty, replacement, knee;
  • pharmacoepidemiology;
  • pulmonary embolism;
  • venous thromboembolism;
  • venous thrombosis



Long-term risk of venous thromboembolism (VTE) following total hip or knee replacement (THR/TKR) compared with controls has not been studied extensively, and the long-term influence of outpatient anticoagulant use on VTE risk remains unknown. The objectives were to evaluate long-term VTE risk following THR/TKR compared with matched controls, and to investigate effect modification by prolonged outpatient vitamin K antagonist use.


A Danish retrospective nationwide cohort study was conducted. All patients undergoing primary THR/TKR (n = 95,227) between 1998 and 2007 were selected, each matched by age, sex and region with three controls (no THR/TKR). Patients were stratified by prolonged outpatient vitamin K antagonist use in the previous 3 months (in a time-dependent manner). All subjects were followed for VTE, and Cox models were used to calculate disease and medication history adjusted hazard ratios (HRs).


Within 6 weeks following surgery, a 13-fold increased risk of VTE was found for THR (adj. HR 12.9; 95% CI 11.2–14.7), and a 14-fold elevated risk for TKR (adj. HR 13.6; 95% CI 11.0–16.7), compared with matched controls. The risk remained substantially increased for at least 4 months following THR/TKR. Within this period, prolonged outpatient vitamin K antagonist use reduced the increase in VTE risk by 69% for THR and 54% for TKR.


The risk of VTE remains substantially elevated for at least 4 months following THR/TKR; this is well beyond the recommended duration of anticoagulant use. The increase in VTE risk is less pronounced in prolonged outpatient vitamin K antagonist users.