Obstructive Sleep Apnea Treatment Modifies the Effectiveness of Catheter Ablation for Atrial Fibrillation: Post Hoc Analysis of the CABANA Trial Authors: Matthew W Segar, Kaleb D Lambeth, Brett Doliner, Joshua R Hirsch, Abdi Rasekh, Mehdi Razavi Published: 2026 Obstructive sleep apnea (OSA) is highly prevalent affecting ∼50% of patients with atrial fibrillation (AF). 1 Although observational studies suggest that untreated OSA may attenuate the effectiveness of catheter ablation for AF, the impact of OSA treatment on ablation outcomes remains unclear. 2 Although OSA was prespecified as a differential treatment effect modifier in the CABANA (Catheter Ablation vs Anti‐arrhythmic Drug Therapy for Atrial Fibrillation Trial) protocol, its impact on AF recurrence outcomes has not been previously reported. Thus, we performed a post hoc analysis of the CABANA trial to evaluate whether OSA treatment status modifies the effectiveness of catheter ablation on AF recurrence compared with drug therapy. Deidentified data were acquired from and are available to qualified researchers through the National Institutes of Health's Biologic Specimen and Data Repository Information Coordinating Center. Data are available from the corresponding author upon reasonable request. The study was exempt from institutional review board review by CommonSpirit Health Research Institute (IRB00009715) and all CABANA participants provided written consent. The CABANA trial randomized 2204 patients with AF to catheter ablation versus drug therapy. 3 We excluded 1 participant with missing OSA status resulting in an analytic cohort of 2203 patients. OSA assessment was not protocolized in CABANA, thus OSA status was ascertained at baseline from medical history and medical records. Patients were classified as having OSA based on documented clinical diagnosis. Among patients with OSA, treatment status (continuous positive airway pressure or oral appliance) was determined by patient report and medical record review. The primary outcome was AF recurrence, defined as any documented atrial tachyarrhythmia (AF, atrial flutter, or atrial tachycardia) lasting ≥30 seconds after a 90‐day blanking period. AF was detected using a proprietary ECG monitoring system that included symptom‐activated transtelephonic recordings, monthly 24‐hour autodetect loop recordings, and 96‐hour Holter recordings obtained biannually. 3 Cox proportional hazards models adjusted for age, sex, race, arrhythmia type (paroxysmal versus persistent), body mass index, CHA2DS2‐VASc score, and randomized treatment arm were used to evaluate associations. The proportional hazards assumption was verified for all models using Schoenfeld residuals (global test P>0.10) and all CIs were constructed using robust SEs. Interaction terms were included to assess treatment effect modification by OSA status and OSA treatment and were tested using a significance level of α=0.025 and α=0.017, respectively, to account for multiple comparisons. For descriptive subgroup hazard ratios (HRs) stratified by OSA treatment status, Bonferroni correction for 3 pairwise comparisons were applied (threshold P<0.017). The magnitude of treatment effect modification was quantified using interaction HRs (ratio‐of‐HRs), representing the relative attenuation of catheter ablation benefit across subgroups with OSA. Of 2203 participants, 508 (23.1%) had OSA. Patients with OSA were more likely to be male (73.2% versus 59.8%, P<0.001), have higher body mass index (34.3±6.4 versus 30.0±5.7 kg/m2, P<0.001), and more likely to drink alcohol (41.0% versus 30.1%, P<0.001) compared with those without OSA. Among patients with OSA, 335 (65.9%) were receiving treatment. Over a median follow‐up of 48 months, AF recurred in 51.4% of patients with OSA versus 46.0% without OSA. In adjusted analyses, the presence of OSA alone was not significantly associated with AF recurrence (HR, 1.07 [95% CI, 0.92–1.25], P=0.40) or when stratifying by treatment status (treated OSA: HR, 1.01 [95% CI, 0.84–1.21], P=0.95; untreated OSA: HR, 1.19 [95% CI, 0.95–1.48], P=0.14). Notably, we observed significant effect modification of catheter ablation effectiveness by OSA status (P‐interaction=0.008). Among patients without OSA, catheter ablation was associated with a 51% relative reduction in AF recurrence compared with drug therapy (HR, 0.49 [95% CI, 0.42–0.57], P<0.001). This benefit was attenuated in patients with OSA (HR, 0.73 [95% CI, 0.57–0.94], P=0.008). When stratifying by OSA treatment status, the interaction remained significant (P‐interaction=0.017). The interaction HR for untreated OSA versus no OSA was 1.72 (95% CI, 1.11–2.66), indicating significantly greater attenuation of catheter ablation benefit in untreated OSA relative to no OSA. Specifically, catheter ablation improved outcomes in patients with treated OSA (HR, 0.68 [95% CI, 0.50–0.93], P=0.016) but was not significant in those with untreated OSA (HR, 0.83 [95% CI, 0.54–1.27], P=0.40) (Figure 1). Cumulative incidence of atrial fibrillation recurrence and treatment effect modification by obstructive sleep apnea status. A, Kaplan–Meier curves for AF recurrence stratified by randomized treatment arm (catheter ablation vs drug therapy) and OSA status (OSA vs no OSA) and B, forest plot of hazard ratios and 95% CIs for the effect of catheter ablation vs drug therapy on AF recurrence. The upper section (Overall OSA Effect) presents results stratified by OSA status; the lower section (OSA Treatment‐Stratified Effect) presents results stratified by OSA treatment status. Hazard ratios are from multivariable Cox models adjusted for age, sex, race, arrhythmia type, body mass index, and CHA2DS2‐VASc score, with robust SEs. P‐interaction indicates the P value for interaction between treatment arm and OSA status. AF indicates atrial fibrillation; HR, hazard ratio; and OSA, obstructive sleep apnea. Our findings suggest that OSA treatment status importantly modifies the effectiveness of catheter ablation for AF. The mechanistic basis for this interaction likely relates to multiple pathophysiologic effects of untreated OSA that promote AF, including repetitive hypoxemia, autonomic dysregulation, systemic inflammation, and atrial structural remodeling. These ongoing triggers may overwhelm the benefits of substrate modification achieved through catheter ablation, whereas OSA treatment appears to partially mitigate these adverse effects. 2 , 4 However, although the mechanistic basis for this interaction likely relates to pathophysiologic effects of OSA, we cannot exclude that treatment‐seeking behavior, adherence patterns, or other unmeasured confounders contribute to observed associations. The clinical implications of these findings are important. First, they support routine screening for OSA in patients being considered for catheter ablation, as OSA prevalence exceeds 20% in this analysis and other studies suggest even higher rates when formal polysomnography is performed. Second, they suggest that optimization of OSA treatment before catheter ablation may improve procedural outcomes. Third, they provide a potential explanation for heterogeneity in catheter ablation outcomes across studies and populations that may differ in OSA prevalence and treatment rates. Several limitations warrant consideration. First, although OSA was a prespecified protocol variable, assessment methods were not standardized across sites. We cannot verify diagnostic criteria uniformity, OSA severity (apnea‐hypopnea index), or objective treatment adherence. The observational nature of OSA treatment comparisons precludes causal inference despite multivariable adjustment. Patients receiving OSA treatment likely differ from untreated patients in unmeasured confounders (adherence, health literacy, health‐seeking behaviors) that cannot be addressed through statistical adjustment alone. Second, adherence to OSA therapy was not systematically assessed and incomplete adherence, suboptimal treatment efficacy, or intermittent noncompliance could result in misclassification of treatment status and attenuation of observed benefits. 5 The “treated” designation reflects prescribed therapy but does not confirm effective ongoing use, which may explain why treated OSA showed intermediate rather than complete restoration of ablation effectiveness. Third, the binary end point of AF recurrence (≥30 seconds) does not capture AF burden or symptom severity. However, this end point remains the guideline‐recommended standard for catheter ablation studies and facilitates comparison with prior literature. Finally, the relatively small number of patients with untreated OSA limited power for pairwise comparisons within OSA subgroups, though the overall treatment × OSA interaction remained statistically significant after Bonferroni correction. In conclusion, OSA treatment status significantly modifies the effectiveness of catheter ablation compared with drug therapy for AF recurrence. Although catheter ablation provides substantial benefit in patients without OSA and intermediate benefit in those with treated OSA, its effectiveness is substantially attenuated in patients with untreated OSA. These findings support treatment optimization in patients undergoing catheter ablation for AF and suggest that addressing OSA may improve procedural outcomes. Sources of Funding None. Disclosures Dr Segar reports honoraria from Idorsia and Otsuka, nonfinancial support from Merck, is an advisor for AccurKardia, is on the advisory board of descendantsDNA, and is the founder of ReCODE Medical. The remaining authors have no disclosures to report. This article was sent to Peng Li, PhD, Associate Editor, for review by expert referees, editorial decision, and final disposition. For Sources of Funding and Disclosures, see page 3. REFERENCES