This is a retrospective multi-center patient cohort study investigating patients with persistent atrial fibrillation (AF) and elective cardioversion (ECV). The aim is to identify clinical predictors for ineffective ECV to improve management quality of persistent AF.

Data was gathered from patient registries of two finnish university hospitals and two regional hospitals in a time period of 2003-2015. A data search was conducted using the ICD-10 code for atrial fibrillation and NCSP code for cardioversion. All patients with AF and performed cardioversion were eligible for the study.

Initially, 2373 patients were captured after which all ECV cases were identified and included manually in the study using a structured electronic case report form. The study cohort comprised 1998 ECVs and 1342 patients. Follow-up data was gathered for 30 days after ECV and all strokes, transient ischemic attacks, arrhythmic complications, AF recurrences and failed cardioversions were recorded and subjected to careful exploration. IBM SPSS Statistics software version 22.0 was used to perform all analyses.

Inclusion Criteria:

• All patients with persistent (>48h) atrial fibrillation and elective electrical cardioversion

Exclusion Criteria:

• Cardioversions for patients with acute (<48h) atrial fibrillation
• Pharmacological cardioversions

1342 patients with 1998 elective cardioversions