Discussion
The CMP-IV remains the most effective treatment for AF, and has been shown to be effective for both paroxysmal and non-paroxysmal AF.[13–15,18] There have been few published reports on the late results of the CMP-IV in patients with longstanding persistent AF, who composed the most common subgroup of patients referred for surgical ablation.[12]
Our results showed excellent and durable success with the CMP-IV for longstanding persistent AF from early through late follow-up amongst 174 consecutive patients. Freedom from ATAs and freedom from both ATAs and AADs was 88% and 68% respectively at 7 years postoperatively. This is particularly impressive in this group of difficult-to-treat patients, with a median preoperative duration of AF of 7.8 years and a mean LA diameter of 4.8cm, the majority of whom (71%, 124/174) had failed one or more catheter ablations. Our findings support the use of surgical ablation as a treatment of symptomatic AF refractory to AADs in patients who have failed or are poor candidates for catheter-based interventions, as recommended by the HRS, STS, and European Cardiac Arrhythmia Society.[2,10,30]
The late results of the CMP-IV in our population were superior to previous studies published on catheter ablation in longstanding persistent AF.[7,8] Within our cohort, ATA-free survival at 5 and 7 years was 83% and 88%, compared with 24% AF-free survival at 10 years in persistent AF patients in one study [7], and just 17% ATAs recurrence-free survival at 5 years after a single catheter ablation in persistent AF patients in another[8]. This suggests that the CMP-IV should be considered in lieu of catheter ablation in selected patients at high risk for failure (i.e. large left atrial volume, long duration of preoperative AF).
The late results of the CMP-IV in our population were also superior to other surgical ablation techniques.[31,32] At 5 years follow-up, our cohort had a 76% ATA-free survival, compared to one study’s 29% AF-free survival in longstanding persistent AF patients after LAA exclusion, pulmonary vein isolation, and ganglionated plexi ablation [32] at the same time point. This supports the use of the full CMP-IV lesion set for surgical ablation in patients with long-standing persistent AF who have failed catheter ablation, especially in light of our low complication rates. There was no significant relationship between the type of ablation device used and freedom from ATA recurrence, though there was a trend toward better late outcomes with dry bipolar radiofrequency clamps.
The stand-alone CMP-IV was performed with minimal morbidity and no 30-day mortality, regardless of surgical approach. There was only one late CVA (0.6%). Postoperative pacemaker placement (7.5%, 13/174) was comparable to previously published studies, with 85% of those pacemakers (11/13) placed for sick sinus syndrome, and 15% (2/13) placed for complete heart block.[12,22] Patients with ATA recurrence had significantly longer median postoperative hospital length of stay, ICU length of stay, and mechanical ventilation time, compared to those without recurrence. This may have been secondary to differences in patients’ underlying baseline health, but could also indicate that greater stress in the postoperative period contributed to eventual recurrence.
On Fine-Gray multivariable regression, the only two factors predictive of late ATAs recurrence were preoperative length of time in AF and early postoperative ATAs. This is similar to previously published results from both our group and others.[15,33,34]
Although this study is one of the largest in the literature to report late rhythm outcomes for stand-alone surgical ablation in patients with longstanding persistent AF, it is not without limitations. This study was retrospective and nonrandomized, thus subject to inherent selection bias. All operations were performed at a single institution, with most completed by a single, highly-experienced surgeon (n=164). This may limit generalizability to other centers. Referral patterns and patient selection may have had an impact on our results. Moreover, sample size may have contributed to the variability of these results and increased risk of type II error. An inherent limitation in the study design was the lack of continuous monitoring on all patients, which can lead to interval censoring and underestimation of ATAs recurrence. However, the majority of patients underwent prolonged monitoring during follow-up, and Fine-Gray regression does not assume non-informative censoring. Moreover, the number of patients both followed longer than 5 years and who had prolonged monitoring is higher than most previously published studies. While the majority of patients underwent prolonged monitoring, AF burden was not recorded in the available medical record for most of these patients, precluding a burden-based analysis. Burden calculations are only possible in patients with pacemakers or implantable loop recorders (ILRs), which were present in only a small minority of our patients. We have previously studied the utility of ILR use following CMP-IV, and found that ILR use did not increase ATA detection compared to Holter monitoring or ECG.[35] Given this, it is unlikely that increasing available burden data via more liberal use of ILRs would significantly alter our findings.