Comparative efficacy of esophageal protection strategies during radiofrequency atrial fibrillation ablation: a systematic review and network meta-analysis of randomized trials

Background Esophageal thermal injury remains a clinically significant complication of radiofrequency catheter ablation for atrial fibrillation (AF), with endoscopically detected esophageal lesions (EDELs) occurring in up to 10–15% of cases and atrio-esophageal fistula carrying high mortality. Multiple protective strategies have been proposed; however, their comparative efficacy remains uncertain. Methods We conducted a systematic review and network meta-analysis of randomized controlled trials evaluating esophageal protection strategies during AF ablation. PubMed, Web of Science, Scopus, and ClinicalTrials.gov were searched from inception through August 7, 2026. Parallel-group randomized trials in adult patients undergoing radiofrequency ablation were included. The primary outcome was incidence of EDEL. A frequentist random-effects network meta-analysis was performed, with treatment hierarchy assessed using P-scores. Risk of bias and certainty of evidence were evaluated using the Cochrane RoB 2 and CINeMA frameworks, respectively. Results Twenty-five randomized controlled trials involving approximately 2,500 patients were included in the systematic review. Twelve trials involving 1,077 patients contributed to the connected network meta-analysis. Mechanical esophageal deviation combined with luminal esophageal temperature monitoring (LET) had the most favorable point estimate compared with no dedicated protection (OR, 0.27; 95% CI, 0.02–3.05) and achieved the highest P-score (0.861). The rankogram estimated a 58% probability of ranking first. Visualization-guided strategies also showed a favorable point estimate (OR, 0.48; 95% CI, 0.08–3.09), whereas LET alone and active cooling did not demonstrate significant reductions in EDEL. Directly, mechanical deviation plus LET significantly reduced the odds of EDEL compared with single-sensor LET (OR, 0.13; 95% CI, 0.04–0.41). Seven randomized trials involving 990 patients showed no significant association between higher- versus standard-power ablation and EDEL (OR, 1.19; 95% CI, 0.56–2.54), with no observed heterogeneity (I² = 0%). Network heterogeneity was moderate (I² = 46.9%), with no significant evidence of network inconsistency or publication bias. Conclusions No single esophageal protection strategy completely eliminates the risk of thermal injury during AF ablation. Mechanical esophageal deviation combined with LET demonstrated the most favorable overall ranking, while visualization-guided strategies also showed a potentially beneficial signal. In contrast, LET or cooling alone did not reliably reduce EDEL incidence. However, overall confidence in the evidence is constrained by small trial samples and reliance on surrogate outcomes. Larger, contemporary trials are needed to confirm clinical benefits and assess the impact on rare but catastrophic complications. Clinical trial number Not Applicable.