Secondary outcomes
Mortality
Three patients died during follow-up, all in the HF group. There was only one cardiovascular death resulting from cardiogenic shock during pulmonary sepsis. The two remaining patients died from a biliary tract cancer and from haemorrhagic complications of liver cirrhosis. Therefore one-year survival was 98.8% with no difference between groups (p=0.254).
Heart failure hospitalization
As expected, more HF patients (19/135, 14.1%) were hospitalized for worsening of HF during follow-up than controls (2/111, 1.8%), p<0.001 (Figure 2). If AF/AT recurrence tended only to increase the risk of HF hospitalization (HR=2.71, 95% CI 0.90 to 8.20, p=0.078), HF status was the only independent predictive factor in multivariate analysis (HR=10.2, 95% CI 2.29 to 10.43, p=0.002). Conversely, patients referred for redux procedure were significantly less hospitalized for HF worsening, than patients with index procedure (HR=0.32, 95% CI 0.12 to 0.82, p=0.018). Among HF patients, we did not find predictive factor associated with hospitalization.
Functional status
At the end of follow-up, 69/135 HF patients experienced an improvement by one or more NYHA class (51.9%) compared to 36 non-HF patients (34%), p=0.006. Among HF patients, AF/AT recurrence was negatively associated with NYHA improvement (HR=0.42, 95% CI 0.19 to 0.91, p=0.028).
Cardiac function
At one-year after the index procedure, we observed an improvement in LVEF, by 5% or more, in 46.4% of the HF patients and in 5.7% of the control patients (p<0.001). HF patients who experienced AF/AT recurrence were less likely to improve their LVEF, but with no statistical difference (35.7% vs 51.8% respectively, p=0.126). HF patients with paroxysmal AF were less likely to improve their LVEF (HR=0.08, 95% CI 0.01 to 0.79, p=0.030) than HF patients with AT or persistent AF.
HF Subgroup analysis
Patients with HFpEF and HFrEF were comparable in their baseline characteristics, comorbidities, symptoms, previous ablations, medications, and echocardiography findings except LVEF. Baseline characteristics of patients of the subgroup analysis are detailed in Table 3. AF/AT recurrence rates were not different whether HF patients had preserved (37.1%) or reduced (26.4%) LVEF, p=0.196. However, patients in HFpEF subgroup (35.6%) had more antiarrhythmic drug therapy discontinuation than HFrEF patients (15.9%), p=0.014. At one-year, there was a greater proportion of LVEF improvement among patients with HFrEF (40/72, 55.6%) compared to patients with HFpEF (18/62, 29%, p=0.002), whereas NYHA improvement was not different between the subgroups (47.2% versus 54.8% respectively, p=0.49).
Complications
Fourteen complications occurred in 13 procedures with no difference between the groups: 8 procedures (5.3%) in the HF group had complications versus 5 in the non-HF group (4.8%) (p=1.000), with no difference whether LVEF was preserved or reduced. The types of complications were equally distributed between groups: two tamponades in each group; three groin bleedings, two of them in the HF group; two other vascular complications in each group, either femoral arteriovenous fistula or pseudoaneurysms; and three strokes, all in the HF group (p=0.231). One HF patient experienced both femoral pseudoaneurysm and a stroke. There was no hemodynamic complication or congestive heart failure but patients of the HFrEF group were more likely to need diuretics increase after ablation (p=0.013).