Gastrointestinal Surgery Status and the Risks of Ischemic Stroke and Safety Outcomes With Direct Oral Anticoagulants Versus Warfarin in Atrial Fibrillation: A Nationwide Population-Based Cohort Study

Figures

• 

  Figure 1.. Study design and flowchart of study population selection. The design of the study (A) and flowchart of study population selection (B). Exclusions were performed stepwise in the order presented. Patients were censored at earliest of the following: the occurrence of the outcome of interest, death, treatment discontinuation (defined as the absence of OAC claims for at least 60 consecutive days), or the end of the study period (December 31, 2021). OAC, oral anticoagulant; AF, atrial fibrillation; ICD, International Classification of Diseases; DOAC, direct oral anticoagulants.

• 

  Figure 2.. Weighted cumulative incidence curves for outcomes by OAC type and GI surgery status. Weighted cumulative incidence curves for ischemic stroke (A), major bleeding (B), all-cause death (C), and composite outcome (D) by OAC type and GI surgery status are shown. WFR/GIS–, warfarin user without prior GI surgery; DOAC/GIS–, direct oral anticoagulant user without prior GI surgery; WFR/GIS+, warfarin user with prior GI surgery; DOAC/GIS+, direct oral anticoagulant user with prior GI surgery; OAC, oral anticoagulant; GI, gastrointestinal.

• 

  Figure 3.. Pairwise comparisons of outcome risks by OAC type and GI surgery status with interaction-term analysis. aHRs for ischemic stroke and major bleeding represent adjusted subdistribution hazard ratios, with all-cause death treated as a competing event. aHRs were derived by adjusting for age, Charlson Comorbidity Index, CHA2DS2-VASc score, sex, prior ischemic stroke, hypertension, diabetes, hyperlipidemia, heart failure, myocardial infarction, peripheral vascular disease, kidney disease, chronic obstructive pulmonary disease, cancer, and prior use of antiplatelet therapy. Pint indicates P for interaction between oral anticoagulant type and GI surgery status. WFR, warfarin; DOAC, direct oral anticoagulant; GIS+, with prior GI surgery; GIS–, without prior GI surgery; aHR, adjusted hazard ratio; CI, confidence interval; OAC, oral anticoagulant; GI, gastrointestinal.

• 

  Figure 4.. Pairwise comparisons of ischemic stroke and major bleeding risks in upper and lower GI surgery strata. aHRs for ischemic stroke and major bleeding represent adjusted subdistribution hazard ratios, with all-cause death treated as a competing event. aHRs were derived by adjusting for age, Charlson Comorbidity Index, CHA2DS2-VASc score, sex, prior ischemic stroke, hypertension, diabetes, hyperlipidemia, heart failure, myocardial infarction, peripheral vascular disease, kidney disease, chronic obstructive pulmonary disease, cancer, and prior use of antiplatelet therapy. Pint indicates P for interaction between oral anticoagulant type and GI surgery status. WFR, warfarin; DOAC, direct oral anticoagulant; GIS+, with prior GI surgery; GIS–, without prior GI surgery; aHR, adjusted hazard ratio; CI, confidence interval; GI, gastrointestinal.