The purpose of the study is to estimate the impact of bleeding events in patients in the UK with non-valvular atrial fibrillation (NVAF) treated with vitamin K antagonists (VKA).

Inclusion Criteria:-

For the CPRD-HES cohort, patients will be included if they meet these criteria:

• Have CPRD data linked to HES

• Have an AF diagnosis (based on Read codes from CPRD or ICD-10 codes from HES) and have been treated with VKA during the study period

  -≥18 years old at index date

• CPRD acceptability quality criteria are present.

For the CPRD cohort, patients will be included if they meet the following criteria:

• Have an Atrial Fibrillation (AF) diagnosis (based on Read codes from CPRD) and have been treated with VKAs during the study period
• ≥18 years old at index date
• CPRD acceptability quality criteria are present.

Exclusion Criteria:

For the two cohorts of interest, patients will be excluded if they meet the following criteria:

• Fewer than 12 months of computerised data available prior to the index date (in order to provide a minimum amount of computerised history and to avoid issues associated with updating records when a patient joins a practice). The date of start of computerized records will be the latter of the patient's date of registration with the practice or the practice's Up-to-Standard (UTS) date (when the practice is deemed to be contributing data of adequate quality)
• The practice is not UTS, or the patient is deemed "unacceptable" according to CPRD's quality criteria
• There is evidence of valvular disease, repair, or replacement before (ie, up to 12 months prior to the index date) or during the study period
• There are records of Novel anticoagulant (NOAC)

Retrospective cohort with study period from January 1, 2003 to December 31, 2013 using the Clinical Practice Research Datalink (CPRD). The study will identify AF patients treated with VKA during the study period and will follow them from the first record of VKA use (index date). until the first of end of study period, leaving the database, VKA discontinuation/switch, or death. It will use person-time denominators to handle the varying length of follow-up of patients