Trends in mortality over time in patients with oesophageal fistula post AF ablation: results from a worldwide FDA database - a subanalysis of the POTTER-AF 2 study

Oesophageal fistula (OF) is one of the deadlines complications following atrial fibrillation (AF) ablation.(1) Due to its reduced incidence, data regarding this complication are scarce. The POTTER-AF 2 study is the second largest study to date to report on OF complicating AF ablation.(2)

To analyse the changes in reporting, management and outcome of OF in the POTTER-AF 2 study over time.

The POTTER-AF 2 Study analysed the OF and oesophageal injury cases complicating AF ablation reported in the mandatory Manufacturer and User Facility Device Experience (MAUDE) database of the Food and Drug Administration (FDA) between August 2009 and August 2019. Herein we perform an analysis of OF cases based on the time of event to investigate the reporting behaviour, the treatment and whether the increased knowledge over time could improve the outcome of these patients. A total of 47 OF were reported in the previously specified time. The date of event was available for 44 patients. We divided this population in two groups (half of the period): patients exhibiting OE until August 15, 2014 (Group 1) and after August 15, 2014 (Group 2).

A total of 19 (43.2%) patients suffered OF until August 15, 2014, and 25 (56.8%) after this date (Table 1). Regarding the year of report, 36.4% of cases were reported in the first time frame and 63.6% in the second one. In Group 1, data on the location of report were available in 47.4% of cases, while in Group 2 in 84.0%, with most patients reported in the United States. The complications were reported after finding them in the literature and not directly to FDA in 36.8% of patients in Group 1 and in 20.0% in Group 2 (p=0.308). The energy source was laser energy, cryoenergy and radiofrequency in 0.0%, 36.8% and 63.2% of patients in Group 1 and 4.0%, 32.0% and 64.0% in Group 2 (p=0.659). The symptoms were reported in 78.9% in Group 1 and 56.0% in Group 2 (p=0.199). The most common symptoms at presentation were neurological events in 66.7% and 64.3% of patients in the two groups (p=1). The diagnostic method was mentioned in 36.8% in the first group and in 40.0% in the second one (p=1), the most common being computed tomography. The treatment was reported in 73.7% patients in Group 1 and 48.0% in Group 2 (p=0.124). The treatment was surgical, only endoscopic or conservative in 42.9%, 42.9% and 14.3% of patients in the first group and in 75.0%, 25.0% and 0.0% of them in Group 2 (p=0.177). Data on survival were available in all patients. The mortality was significantly higher in Group 1 compared to Group 2 (94.7% vs. 64.0%; p=0.027) (Figure 1).

More OF cases were reported in the latter years. The quality of data showed no improvement over time. The distribution across energy sources remained consistent, approximately one-third of patients developing OF following cryoablation. The treatment in Group 2 was more often surgical. Mortality rates significantly declined over the years.Table 1.

Characteristics of the groups