Funding Acknowledgements
Type of funding sources: Public Institution(s). Main funding source(s): National Heart, Lung, and Blood Institute
Background
Many patients with heart failure who are considered for cardiac resynchronization therapy (CRT) have a history of (h/o) atrial fibrillation (AF) but there are doubts about the efficacy of CRT in patients with AF.
Purpose
To investigate the association of CRT on morbidity and mortality among patients with and without a h/o AF.
Methods
Original, patient-level data from five clinical trials of CRT that permitted enrolment of patients with a h/o AF were included: COMPANION, MADIT-CRT, BLOCK HF, REVERSE, and MIRACLE trial. Patients with permanent or persistent AF were excluded from these trials, and therefore from this analysis. The outcomes of interest were the composite endpoint of time to heart failure hospitalization (HFH) or all-cause mortality or all-cause mortality alone. The association of CRT (versus no CRT) with outcomes for patients with and without a h/o AF was assessed using a Bayesian-Weibull survival regression model with random terms for the trial-specific treatment effects and the trial-specific baseline hazard functions including an interaction between history of paroxysmal AF and CRT. All results are presented as hazard ratios (HRs) with 95% posterior credible intervals (CIs) and posterior probabilities of no association, adjusting for baseline characteristics.
Results
A total of 4062 patients were included, 661 (16.3%) of whom had a h/o AF. Patients with a h/o AF were older (mean [SD] age 68 [10] years versus 64 [11] years) and had a higher proportion of ischemic cardiomyopathy (67% versus 53%, p<0.001), a higher baseline serum creatinine (1.3 mg/dl versus 1.2 mg/dl, p<0.001), and a lower left ventricular ejection fraction (25% versus 26%, p<0.001). The HRs for all outcomes and the interaction term are shown in Table 1. For the overall population, CRT delayed the time to HFH or all-cause mortality (HR: 0.74, 95% CI: 0.62 – 0.87, p=0.005); for patients with a h/o AF, it did not (HR: 0.87, 95% CI: 0.64 to 1.19, p=0.37). In this patient-level meta-analysis, CRT was not associated with a reduction in mortality, overall or by h/o AF. Howevber, the interaction (estimate shown as a ratio of HRs) between those with or without a h/o AF and the effects of CRT was not significant for either outcome (Table 1).
Conclusion
In the largest post hoc analysis to date, we confirm the benefits of CRT in patients without a h/o AF in reducing HFH or mortality. There was no statistically significant interaction between CRT and h/o AF for any analysed outcome.