TARGET-EU: Rivaroxaban and risk of major gastrointestinal bleeding in elderly patients with non-valvular atrial fibrillation

Study type

• Study topic:
  Disease /health condition

  Human medicinal product
• Study type:
  Non-interventional study
• Scope of the study:
  Safety study (incl. comparative)
• Data collection methods:
  Secondary use of data

Non-interventional study

• Non-interventional study design:
  Cohort

Study drug and medical condition

• Study drug International non-proprietary name (INN) or common name:
  RIVAROXABAN

  APIXABAN
• Anatomical Therapeutic Chemical (ATC) code:
  (B01AF01) rivaroxaban

  rivaroxaban

  (B01AF02) apixaban

  apixaban
• Medical condition to be studied:
  Atrial fibrillation

Population studied

• Short description of the study population:
  The study population will consist of adults aged 75 years or older with non-valvular atrial fibrillation (NVAF) who initiate treatment with either rivaroxaban or apixaban during the study period. Cohort entry (index date) is defined as the date of the first prescription of either treatment.; To be eligible, individuals must have a diagnosis of atrial fibrillation within a 30-day window around the index date and be aged 75 years or older at cohort entry.; Several exclusion criteria will be applied: individuals must have at least 365 days of recorded medical history prior to the index date; those with a documented diagnosis of valvular atrial fibrillation at any time before the index date will be excluded; and individuals with any prior use of a DOAC or vitamin K antagonist in the year preceding cohort entry will be excluded to implement a new-user design.
• Age groups:
  • Adults (75 to < 85 years)
  • Adults (85 years and over)

Study design details

• Study design:
  This study will emulate a hypothetical target trial using a retrospective, active-comparator, new-user cohort design based on routinely collected electronic health records.
• Main study objective:
  The overall aim is to assess whether rivaroxaban is associated with a different risk of major gastrointestinal bleeding compared with apixaban in adults aged 75 years or older with non-valvular atrial fibrillation.
• Setting:
  This study is conducted using routinely collected electronic health records from 2013 to 2023, reflecting the period of routine clinical use of direct oral anticoagulants in older adults. The study is set in hospital and outpatient settings (Danish registers) and primary care with hospital linkage (SIDIAP), providing population-based coverage of real-world clinical care in Denmark and Spain.
• Comparators:
  The comparator group consists of patients who initiate apixaban at any dose, formulation, or regimen. The use of an active comparator is the most appropriate approach as it aligns with real-world clinical practice, addresses confounding by indication, and allows the safety of rivaroxaban to be benchmarked against a commonly prescribed alternative for NVAF. Both treatments are used for the same indication, and treatment episodes may last for up to two years.
• Outcomes:
  The primary endpoint is time to first occurrence of major gastrointestinal bleeding, identified through hospital admissions, primary care records, or death records using ICD-10 diagnostic codes. The definition of major bleeding follows the International Society on Thrombosis and Haemostasis (ISTH) criteria, operationalised in real-world data using relevant diagnostic codes.
• Data analysis plan:
  Analyses are conducted within a target trial emulation framework to estimate the comparative effect of rivaroxaban versus apixaban on risk of major gastrointestinal bleeding.; For Estimand 1 (primary), the causal effect summary measure is the hazard ratio for time to first major GI bleeding, estimated using an inverse probability of treatment weighted (IPTW) Cox proportional hazards model. The Cox model will be fitted separately within each data source (Danish registers and SIDIAP), and the resulting hazard ratios pooled using a random-effects meta-analysis; potential sources of heterogeneity will be described qualitatively.; Two supplemental estimands are also defined: Estimand 2 applies a while-on-treatment strategy using restricted mean survival time (RMST) at 1 and 2 years; Estimand 3 applies a hypothetical strategy for treatment discontinuation and switching. Sensitivity analyses include inverse probability of censoring weighting (IPCW), tipping point analysis, and probabilistic bias analysis for outcome misclassification