Novel risk model for predicting incident atrial fibrillation in patients taking ω-3 fatty acid: sub-analysis of the STRENGTH randomized trial

Fish oils are associated with higher risk of incident atrial fibrillation (AF) in randomized trials, risk factors for this remain unknown. We analyzed associated factors and developed a novel risk score for predicting AF in patients at elevated cardiovascular risk in this secondary analysis of the STRENGTH trial. The STRENGTH trial randomized 13,078 patients at high cardiovascular risk at 1:1 ratio to receive either ω-3 carboxylic acid (CA) or corn oil placebo at 675 centers during 10/30/2014-6/14/2017. Multivariable Cox proportional hazards regression was performed to develop a risk score to help identify new AF during follow-up in the ω-3 CA group without prior history of AF. Amongst 6,000 ω-3 CA group and 6,012 placebo group patients without history of AF, 126 (2.1%) and 75 (1.2%) developed AF, respectively, during mean follow-up of 3.5±0.8 years. In the ω-3 CA arm, older age, males, higher body mass index, history of heart failure, insulin-dependent diabetes, and elevated high-sensitivity C-reactive protein were independently associated with incident AF during follow-up. A novel risk score based on these six parameters predicted incident AF with c-index 0.72 in the ω-3 CA group, and 0.73 when applied to the placebo group. A novel risk score for predicting AF development in patients at elevated cardiovascular risk receiving ω-3 CA had moderate discriminative ability in both patients receiving ω-3 CA and placebo. Further research is necessary to externally validate our risk score and determine how it may guide AF management and prevention and improve clinical outcomes. Novel risk score for predicting incident atrial fibrillation in patients at elevated cardiovascular risk receiving ω-3 CA (top left), receiver-operating characteristic curve for predicting incident atrial fibrillation (top right), and atrial fibrillation event rates during follow-up by risk score categories (bottom).