BackgroundBalloon-based catheter ablation (CA) technologies, including hot balloon ablation (HBA), laser balloon ablation (LBA) and cryoballoon ablation (CBA) have been introduced in recent years as alternatives to conventional radiofrequency ablation therapy for atrial fibrillation (AF). However, the results remain controversial concerning the optimal approach. Thus, we conducted a network meta-analysis (NMA) to comprehensively evaluate the efficacy and safety of HBA, LBA and CBA.MethodsClinical trials comparing the efficacy and safety of HBA, LBA and CBA were identified through a systematic search up to October 2022. The primary outcomes of interest were the recurrence of AF and procedure-related complications.ResultsTwenty clinical trials with a total of 1,995 patients were included in the meta-analysis. The NMA results demonstrated that HBA, LBA and CBA had comparable AF recurrence rates (HBA vs. CBA: odds ratio OR = 0.88, 95% credible interval CrI: 0.56–1.4; LBA vs. CBA: OR = 1.1, 95% CrI: 0.75–1.5; LBA vs. HBA: OR = 1.2, 95% CrI: 0.70–2.0) and procedure-related complications (HBA vs. CBA: OR = 0.93, 95% CrI: 0.46–2.3; LBA vs. CBA: OR = 1.1, 95% CrI: 0.63–2.1; LBA vs. HBA: OR = 1.2, 95% CrI: 0.44–2.8). The surface under the cumulative ranking curve (SUCRA) suggested that HBA may be the optimal approach concerning the primary outcomes (SUCRA = 74.4%; 61.1%, respectively). However, HBA (40.1%) had a significantly higher incidence of touch-up ablation (TUA) than LBA (8.5%, OR = 2.8, 95% CrI: 1.1–7.1) and CBA (11.9%, OR = 3.7, 95% CrI: 1.9–7.5). LBA required more procedure time than CBA [mean difference (MD = 32.0 min, 95% CrI: 19.0–45.0 min)] and HBA (MD = 26.0 min, 95% CrI: 5.6–45.0 min), but less fluoroscopy time than HBA (MD = −9.4 min, 95% CrI: −17.0–−2.4 min).ConclusionsHBA, LBA and CBA had comparable efficacy and safety as initial treatments for AF. HBA ranked highest in the primary outcomes, but at the cost of a higher incidence of TUA and longer fluoroscopy time.Systematic Review Registrationwww.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022381954, identifier: CRD42022381954.