New-onset atrial fibrillation in critically ill patients: risk factors, treatments, and outcomes

Thesis

New-onset atrial fibrillation (NOAF) is defined as atrial fibrillation (AF) occurring in a patient with no history of chronic or paroxysmal AF. It is a common arrhythmia in critically ill patients. Existing tools to estimate the risk of a patient developing NOAF in a general intensive care unit (ICU) have significant limitations. We know little about the haemodynamic changes associated with NOAF in patients in an ICU. The best treatments to be compared in randomised trials are unknown. Studies investigating the survival impact of NOAF in critically ill patients are limited by size or methodology. In this thesis, an international Delphi study examining the risk factors for NOAF in critically ill patients is undertaken. Using these systematically-identified variables, a NOAF prediction model is developed using international data. This model outperforms a comparable model in an assessment of external validity using multi-centre UK data. A user-friendly tool is built to estimate NOAF risk using 10 readily available clinical variables. Outcomes associated with first-line treatments for ICU-acquired NOAF using a propensity-weighted analysis are then investigated. Digoxin therapy is found to be associated with inferior rate control compared with amiodarone, with calcium channel blocker therapy being associated with inferior rhythm control compared with amiodarone. Apparent mortality differences between beta blockers and amiodarone in unadjusted analyses are no longer evident when cardiovascular stability at the time of AF onset is taken into account. This finding has now been included in American national guidelines. The immediate haemodynamic impact of NOAF in the ICU is subsequently explored. This work demonstrates that NOAF is associated with a significant increase in heart rate and reduction in systolic blood pressure. Finally, a competing risks analysis of the association of NOAF with hospital mortality is presented. This study reveals that NOAF is associated with both an increased duration of hospital stay and an increased rate of in-hospital death. NOAF lasting over 30 minutes was associated with increased hospital mortality. This thesis argues that ICU-acquired NOAF is associated with poor outcomes. It highlights optimal therapies that may be compared in randomised trials. Lastly, it suggests that its prediction is possible. Targeted prevention strategies in high-risk patients combined with optimal treatment has the potential to improve outcomes for a vulnerable cohort.

Authors: Bedford, JP

• DOI: 10.5287/ora-vynkbwwd5
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: critical care; atrial fibrillation; arrhythmia; machine learning; critical illness; prognostic modelling
• Subjects: Critical care medicine