Long-term Outcome of Atrial Fibrillation Ablation: Impact and Predictors of Very Late Recurrence

Anita Wokhlu, M.D.; David O. Hodge, M.S.; Kristi H. Monahan, R.N.; Samuel J. Asirvatham, M.D.; Paul A. Friedman, M.D.; Thomas M. Munger, M.D.; Yong-Mei Cha, M.D.; Win-Kuang Shen, M.D.; Peter A. Brady, M.D.; Christine M. Bluhm, R.N.; Janis M. Haroldson, R.N.; Stephen C. Hammill, M.D.; Douglas L. Packer, M.D.

Disclosures

J Cardiovasc Electrophysiol. 2010;21(10):1071-1078. 

In This Article

Abstract and Introduction

Abstract

Long-Term Outcome of AF Ablation.
Introduction: Ablation eliminates atrial fibrillation (AF) in studies with 1 year follow-up, but very late recurrences may compromise long-term efficacy. In a large cohort, we sought to describe the determinants of delayed recurrence after AF ablation.
Methods and Results: Seven hundred and seventy-four patients with AF (428 paroxysmal [PAF, 55%] and 346 persistent or longstanding persistent [PersAF, 45%]) underwent wide area circumferential ablation (WACA, 62%) or pulmonary vein isolation (38%). Over 3.0 ± 1.9 years, there were 135 recurrences in PAF patients and 142 in PersAF patients. AF elimination was achieved in 61% of patients with PersAF at 2 years after last ablation and in 71% of patients with PAF (P = 0.04). This finding was related to a higher initial rate of very late recurrence in PersAF. From 1.0 to 2.5 years, the recurrence increased by 20% (from 37% to 57%) in PersAF patients versus only 12% (from 27% to 39%) in PAF patients. Independent predictors of overall recurrence included diabetes (HR 1.9 [1.3–2.9], P = 0.002) and PersAF (HR 1.6 [1.2–2.0], P < 0.001). Independent predictors of very late recurrence included PersAF (HR 1.7 [1.1–2.7], P = 0.018) and WACA (HR 1.8 [1.1–2.7], P = 0.018), while diabetes came close to significance. In PAF patients, left atrial size >45 mm was identified as an AF-type specific predictor (HR 2.4 [1.3–4.7], P = 0.009), whereas in PersAF patients, no unique predictors were identified.
Conclusion: Late recurrences reduced the long-term efficacy of AF ablation, particularly in patients with PersAF and underlying cardiovascular diseases.

Introduction

Radiofrequency (RF) ablation of atrial fibrillation (AF) has variable efficacy with elimination rates ranging from 30% to 95% in studies with less than 12 months of follow-up.[1–7] This wide range is based on data from observational studies and a limited number of small, randomized clinical trials, predominantly in patients with paroxysmal atrial fibrillation (PAF) and no underlying heart disease.[1–5] Extrapolation of these data to prognosticate for more complex patient populations has been further complicated by nonstandardized definitions for ablative success and underreporting of single procedure outcomes and required drug therapy.[6–7]

Ablation in patients with persistent AF is estimated to be 10–20% less effective than in those with paroxysmal AF.[6] However, limited studies suggest that with a substrate-modification approach or multiple procedures, outcomes would be comparable among patients with paroxysmal versus persistent disease.[4,8–10]

Nevertheless, due to limited follow-up within these studies, the long-term outcomes of ablation remain ambiguous. In a worldwide survey of 181 centers of AF ablation, recurrence rates without antiarrhythmic drug (AAD) use are as high as 34% in patients at 12 months follow-up, but increase to 60% in patients with ≥18 months follow-up.[11] Conversely, in other studies with relatively smaller patient numbers, very late recurrence (beyond12 months) has been described as an uncommon phenomenon with reported rates of 3–10%.[8,12–15] Electrophysiological characteristics of patients with very late recurrences have been described, but clinical predictors critical to treatment and relevant to the planning of large clinical trials have yet to be identified.[14–16]

We sought to determine the long-term prognosis after AF ablation in a large, longitudinal cohort of patients with symptomatic AF at presentation and to clarify the predictors of late recurrence.

processing....