Background: Atrial fibrillation (AF) often co-exists with left ventricular systolic dysfunction (LVSD), with LVSD being either the trigger or consequence of AF. In the CASTLE-AF trial, patients who suffered from both AF and LVSD had an improved outcome when treated with catheter ablation compared to medical therapy. A subanalysis of CASTLE-AF established the amount of AF burden at 6 months after catheter ablation as an important predictor defining the outcome after rhythm control therapy. Thus, we evaluated if an AF burden-dependent effect on cardiac function could be observed in vitro.

Methods: Experiments were conducted using human induced pluripotent stem cell cardiomyocytes (iPSC-CM). To investigate the effect of different AF burden, an AF simulation was created. IPSC-CM were divided into four groups and stimulated for 48 h and 7 days with a frequency of 90 bpm. Apart from a control group, which was stimulated regularly the whole time, three additional iPSC-CM groups with an arrhythmia burden of 30%, 60% and 90% were investigated to simulate AF. During times of irregular stimulation, the beat-to-beat variability was 40%. Epifluorescence microscopy for measurements of Ca²⁺ homeostasis was performed using the ratiometric Ca²⁺ dye Fura-2 AM. Action potential measurements were carried out using ruptured-patch whole-cell current clamp technique.

Results: After 7 days, Ca²⁺ transient amplitude decreased significantly and action potential duration (APD) 90 increased significantly between groups with low and high AF burden, with a cut off being detected between Burden 30% and Burden 60%. Thus, our AF simulation had a negative impact on cellular ventricular function when AF burden was 60% or higher which matches the clinical findings of CASTLE-AF (figure 1, 30 bpm). After 48 h, there were no significant effects of higher AF burden on Ca²⁺ transient amplitude, but a significant APD 90 prolongation could be seen, which was, however, less pronounced than after 7 days. Diastolic Ca²⁺levels, decay kinetic of Ca²⁺ transients, resting membrane potential, action potential amplitudes and upstroke velocity did not display significant changes.

Fig. 1: Effects of AF simulation on Ca²⁺ transients and APD90 at 30 bpm after 7 days. The total number of analyzed cells is provided in each column. Data are presented with SEM, 2way ANOVA were performed, **p<0.01, ***p<0,001, ***p<0,0001).

Conclusion: We could show for the first time that a high arrhythmia burden leads to impaired cellular electrophysiological function, while low arrhythmia burden does not have a significant negative impact. Furthermore, we detected a threshold of AF burden of >50% where cellular ventricular function starts to be impaired which clearly relates to the findings of the CASTLE-AF trial. Thus, our results may help to understand these clinical findings in a “back to bench” approach.