Background: Recent studies demonstrated the beneficial effects of sinus rhythm restoration in patients with heart failure (HF) and concomitant atrial fibrillation (AF). Despite frequent coexistence of HF and AF, the adverse mechanisms of AF mediating the aggravation of ventricular dysfunction in pre-existing HF are largely unknown. 

Methods and results: To elucidate the effects of persistent AF on human failing myocardium, we performed contractility experiments using in toto isolated ventricular trabeculae from patients with end-stage HF. Simulation of normofrequent AF by electric field stimulation (60bpm, 40% beat-to-beat-variability vs. 60bpm in control) for 8h progressively compromised systolic as well as diastolic contractile function. Peak force and relaxation velocity were significantly decreased after 8h of simulated AF, while diastolic dysfunction was increased compared to control (n=8 trabeculae/8 patients AF vs. 8/8 control). For healthy control, we performed respective experiments in human non-failing ventricular trabeculae. In contrast to failing myocardium, 8h of arrhythmic stimulation had no adverse effects on any contractile parameters in non-failing trabeculae (n=3/3 AF vs. 4/4 control). 

To investigate underlying cellular mechanisms, we utilized human induced pluripotent stem-cell cardiomyocytes (iPSC-CM) from 2 patients with familial dilated cardiomyopathy (DCM) and 1 healthy donor for 48h AF simulation (90bpm, 30% beat-to-beat-variability) vs. rhythmic stimulation (60bpm) as control in culture. 

Ca-measurements (Fura-2 AM) were performed to elucidate potential mechanisms mediating the impaired contractile function. In healthy iPSC-CM 48h of AF-like stimulation had overall no significant effect on Ca-transient parameters (n=39 cells/2 differentiations AF vs. 29/2 control). In both DCM iPSC-CM cell lines however, 48h of AF simulation resulted in a substantial decrease of systolic Ca-transient amplitude, accompanied by a slowed Ca-elimination (RT80, decay time constant tau) and a reduced cytosolic Ca-concentration (n=86/8 AF vs. 116/8 control) compared to rhythmically stimulated cells. Evaluation of SR Ca-handling by Caffeine application showed a significantly reduced systolic fractional Ca-release a trend towards a reduction of SERCA2a activity (K_sys-K_Caff), which might explain the observed Ca-handling alterations (14/8 AF vs. 16/8 control).

Whole-cell current clamp experiments after 48h of AF-like stimulation revealed a significant prolongation of action potential duration compared to control (60bpm) in iPSC-CM from both DCM patients. Resting membrane potential, action potential amplitude and upstroke velocity were not altered (n=35/7 AF vs. 38/7 control). In healthy iPSC-CM, no effect of 48h AF on any action potential parameters could be observed compared to control (n=10/2 AF vs. 11/2 control).

Conclusion: Arrhythmic stimulation severely compromised contractile function in human ventricular myocardium with pre-existing systolic HF. In DCM iPSC-CM 48h of AF-like stimulation adversely affected cellular Ca-handling and electrophysiology, but did not induce alterations in healthy cardiomyocytes. Thus, this study shed light on the detrimental contribution of AF to arrhythmia-mediated aggravation of pre-existing heart failure and gives rise to clinical investigation.