This study was designed to assess the effectiveness of oral anticoagulants in elderly patients (>=75 years) with atrial fibrillation. All patients discharged from one large hospital in Stockholm, Danderyds Hospital, between november 1st 2010 and december 31st 2017 with atrial fibrillation as main diagnosis have been included (n=2943). Adverse events (cardiovascular death, fatal/non fatal ischemic stroke, peripheral thromboembolism, fatal/non fatal intracerebral hemorrhage and extracranial bleeding) are recorded through linkage to mandatory and qualitative national registries and review of the medical records until december 31 2018.

The investigators established a prospective cohort of 2943 of elderly (>=75 years old) discharged with atrial fibrillation as main diagnosis from November 1st 2010 to December 31st 2017 from a large hospital in the Northern part of Stockholm in Sweden, Danderyds Hospital. Atrial fibrillation diagnosis is defined by the (ICD-10) codes: I48.0-I48.9 (atrial fibrillation and atrial flutter) according to the International Classification of Diseases, 10th revision. The first admission to the hospital during the study period is recorded as the index admission.

Clinical and anamnestic data, anthropometric measures, biochemical data, treatment at admission and at discharge have been recorded for each patient through the screening of medical journals.

Inclusion Criteria:

• All consecutive patients discharged with atrial fibrillation as main diagnosis from Danderyds Hospital from November 1st 2011 to December 31 2017 aged 75 years or more.

Exclusion Criteria:

• Patients who have not given their consensus to the use of medical journals for register based research.
• Patients who were discharged from the hospital with atrial fibrillation as main diagnosis but were admitted to the hospital for medical reasons unrelated to atrial fibrillation (as an example change of pacemakers batteries).
• Patients whose medical records are not consistent with the main diagnosis (as an example wrong ICD code).

Prospective cardiovascular cohort

Data will be shared according to the rules on privacy at Karolinska Institutet and according to General Data Protection Regulations in Europe.

Individual data will not be shared, but summary statistics of the variable of interest will be provided upon request.

from January 2020