• With the use of an estimated annual incidence of the primary outcome of 4.5% in the BMI ≥50 kg/m2 cohort
    and 2.0% in the BMI <30 kg/m2 cohort in patients receiving anticoagulation, a sample size of 1288 patients
    (644 patients in each cohort) was calculated using 80% power and a significance level set at 0.05
  • Multivariable logistic regression analysis was performed to investigate the association between outcomes of
    interest and the 2 BMI groups, controlling for confounding effects from age, renal function, and apixaban or
    rivaroxaban therapy
    • This included the non-linear relationship between age, renal function, and outcomes of interest;
      the interaction between BMI and apixaban or rivaroxaban therapy; and the interaction between BMI and
      the linear component of age and renal function