Risk of Recurrent Bleeding Events in Nonvalvular Atrial Fibrillation Treated with Vitamin K Antagonists: A Clinical Practice Research Datalink Study

Figures

• 

  Fig. 1. Schematic description of study design. Individuals are followed up from their first VKA prescription until discontinuation of VKA treatment (unless the study period ends, the individual leaves the database, or the individual dies before discontinuing treatment; not shown). In the analysis the risk of primary and secondary outcomes within each of exposure periods 1 to 3 (defined by the number of bleeds that has occurred; 1, 2, or ≥3) was compared with the risk in the reference period (the period prior to the first bleed under VKA treatment). Adjusted models include covariate information extracted from the baseline characteristic period (12-month period prior to the first VKA prescription). VKA, vitamin K antagonist.

• 

  Fig. 2. Results of sensitivity analysis investigating the impact of treating records of bleeds as duplicates if they are recorded within certain time periods of each other. The main analysis treats all records of bleeds as separate events of bleeds. Hazard ratios adjusted for all characteristics listed in Table 1 .

• 

  Fig. 3. Forest plot describing the hazard of primary and secondary bleed outcomes during VKA treatment in the period following the first, second, and third bleeds compared with the hazard in the period before any bleeds have occurred. Hazard ratios adjusted for all characteristics are listed in Table 1 . VKA, vitamin K antagonist.

• 

  Fig. 4. Forest plot describing the hazard of secondary clinical outcomes during VKA treatment in the period following the first, second, and third bleeds compared with the hazard in the period before any bleeds have occurred. Hazard ratios adjusted for all characteristics listed in Table 1 . VKA, vitamin K antagonist.

• 

  Fig. 5. Forest plot describing the incidence rate of GP visits, GP prescriptions, and hospital admissions during VKA treatment in the period following the first, second, and third bleeds compared with the incidence in the period before any bleeds have occurred. Incidence rate ratios adjusted for all characteristics listed in Table 1 . GP, general practitioner; VKA, vitamin K antagonist.