TABLE VIII.B. 4.b.

Evidence for use of PAP and atrial fibrillation control

Study	Year	LOE	Study design	Study groups	Clinical endpoints	Conclusion
Kanagala et al.1216	2003	4	Clinic-based cohort study	Patients referred for DC cardioversion; OSA diagnosis based on prior documented polysomnography; CPAP use based on self-report 1. 12 patients with treated OSA 2. 27 patients with untreated (or inadequately treated) OSA 3. 79 patients without known OSA	Recurrence of AF at 12 months	AF recurred in 82% of untreated OSA patients, which was significantly high than the 42% of treated OSA patients (p = 0.01) and 53% in the patients without known OSA (p = 0.009).
Jongnarangsin et al.728	2008	4	Clinic-based cohort study	324 patients referred for radiofrequency catheter ablation of AF (pulmonary vein isolation, other atrial regions if needed; OSA diagnosis based on prior documented polysomnography 1. 18 patients with treated OSA 2. 14 patients with untreated OSA 3. 292 patients without known OSA	Recurrence of AF over a mean follow-up of 7 (SD 4) months	AF recurred in 37% of patients without known OSA and 59% of patients with OSA (p = 0.02). The recurrence rate was 50% in the CPAP-treated OSA patients and 71% in the untreated OSA patients (p = 0.29). In multivariate analysis, the adjusted OR for AF recurrence was 3.04 (95% CI 1.11–8.32) in those with OSA compared to those without known OSA.
Patel et al.730	2010	2b	Clinic-based, multi-center cohort study	3000 consecutive patients undergoing pulmonary vein and posterior wall isolation procedure for AF; OSA diagnosis based on prior documented polysomnography with AHI >15 and >80% of events obstructive; CPAP use based on self-report 1. 315 OSA patients with regular CPAP use 2. 325 OSA patients with no or erratic CPAP use 3. 2360 patients without known OSA	Recurrence of AF over a mean follow-up of 32 (SD 14) months	AF recurred in 22% of patients without known OSA and 27% of patients with OSA. The adjusted OR for AF recurrence was 1.59 (95% CI 1.25–2.08) for OSA versus no known OSA. Among those with OSA, AF recurred in 21% of those treated with CPAP and 32% of those not using CPAP (p = 0.003). The adjusted OR for AF recurrence was 0.16 (95% CI 0.07–0.35) for CPAP use versus no or erratic CPAP use.
Fein et al.1215	2013	4	Clinic-based cohort study	92 patients undergoing pulmonary vein isolation for the treatment of AF; OSA diagnosis based on prior documented polysomnography with AHI >15 and ≥80% of events obstructive; method of CPAP use determination not stated 1. 32 patients with treated OSA 2. 30 patients with untreated OSA 3. 30 patients without known OSA	AF-free survival	The 1-year AF-free survival was 71.9% in those with CPAP-treated OSA, significantly higher than the 33.3% in those with untreated OSA (p = 0.02) but not significantly different from the 66.7% in those without OSA. Compared to those without OSA, the adjusted HR for AF recurrence was 2.15 (95% CI 1.10–5.44) in those with OSA not using CPAP and 0.7 (95% CI 0.3–1.59) in those with OSA using CPAP.
Naruse et al.708	2013	2b	Clinic-based cohort study	153 patients undergoing pulmonary vein isolation for the treatment of AF; all patients underwent polysomnography approximately 1 week after the procedure, with OSA defined as AHI ≥ 5 and ≥50% of events obstructive; objective CPAP data 1. 82 patients with CPAP-treated OSA (60 with use ≥4 h/night, 22 with use <4 h/night) 2. 34 patients with untreated OSA 3. 37 patients without OSA	AF recurrence over mean follow-up of 19 (SD 10) months	AF recurred in 53% of those with untreated OSA, 30% in those treated with CPAP, and 22% of those without OSA. The multivariate adjusted HR for AF recurrence was 2.61 (95% CI 1.12–6.09) for presence of OSA and 0.41 (95% CI 0.22–0.76) for CPAP use.
Neilan et al.729	2013	2b	Clinic-based cohort study	720 consecutive patients referred for pulmonary vein isolation for AF; OSA diagnosis based on prior documented polysomnography; CPAP usage by self-report 1. 71 patients with treated OSA (CPAP use >4 h/night) 2. 71 patients with untreated OSA (or CPAP use <4 h/night) 3. 578 patients without known OSA	Late AF recurrence (AF occurring >3 months after PVI) over median follow-up of 42 months	The cumulative incidence of late AF was 68% in patients with untreated OSA, 35% in treated OSA, and 30% in those without OSA. Compared to those without OSA, the multivariate adjusted HR for AF recurrence was 2.79 (95% CI 1.97–3.94) in untreated OSA and 1.14 (95% CI 0.741.76) in treated OSA.
McMillan et al.1219	2014	2b	Randomized multi-center clinical trial	278 patients with 4% oxygen desaturation index >7.5 events/h plus Epworth Sleepiness Scale score of ≥9 1. CPAP plus “best supportive care” 2. “Best supportive care” alone	Incident AF was reported as part of a secondary composite endpoint and was not itself a pre-specified endpoint	Supplementary table S13b includes baseline prevalence and 12-month incidence of AF. In the Best Supportive Care group, incident AF occurred in 14 of 85 patients free of AF at baseline, versus seven of 92 in the CPAP group (OR 0.42 for CPAP group, p = 0.07).
Caples et al.1218	2019	2b	Randomized clinical trial	25 patients with recent successful electrocardioversion for AF screened for enrollment; of 1757 patients screened, 34 were enrolled and underwent in-lab polysomnography; 25 with AHI ≥ 5 and predominantly obstructive events were randomized to: 1. CPAP plus usual	Recurrence of AF	36% of participants withdrew or were lost to follow-up prior to 1-year follow-up; three patients in each group had recurrence of AF (p = 0.98). This study was markedly underpowered to observe a treatment effect.
Hojo et al.712	2019	4	Clinic-based cohort study	100 patients undergoing pulmonary vein isolation for AF; all had cardiorespiratory polygraphy at baseline, with OSA defined as AHI ≥ 15 and ≥50% of events obstructive; all patients had a second PVI procedure at 6 months, after which CPAP was recommended based on local guidelines 1. 11 patients with treated OSA 2. 23 patients with untreated OSA 3. 66 patients without OSA	Recurrence of AF or other atrial tachyarrhythmia following the second PVI procedure	No significant difference among groups was noted in rate of AF recurrence (12.1% in no OSA group, 9.1% un treated OSA group, 8.7% in untreated OSA group).
Srivali et al.1217	2019	2b	Retrospective clinic-based cohort study	429 patients with a diagnosis of sleep apnea by in-lab PSG (351 OSA, 21 CSA, 57 mixed sleep apnea), with a subsequent diagnosis of AF, treated with PVI or with electrical or chemical cardioversion 1. 269 PAP-adherent patients, defined as ≥4 h/night at least 70% of nights 2. 160 PAP non-adherent patients	Recurrence of AF over a median follow-up of 4.6 months; analyzed for all sleep apnea and separately by type of sleep apnea	Time to recurrence of AF did not differ between PAP-adherent and non-adherent patients overall or for any sleep apnea type. In OSA patients, median recurrence-free survival was 9.4 months in the adherent and 9.5 months in the non-adherent group. PVI (vs. cardioversion) was the strongest predictor of recurrence-free survival, and was more common in PAP non-adherent (33%) than PAP adherent (20%) patients; however, the authors state that in multivariate models adjusting for type of intervention, there was no significant difference between PAP adherent and non-adherent patients in hazard of AF recurrence.
Traaen et al.710	2021	1b	Randomized clinical trial	Patients with paroxysmal atrial fibrillation, screened for sleep apnea with two nights of home respiratory polygraphy with a mean AHI ≥ 15, and use of CPAP ≥4 h on each night of a 7-night run-in period 1. Usual care alone (n = 54, no dropouts), baseline AF burden 5.0% 2. Usual care plus CPAP (n = 55, one dropout), baseline AF burden 5.6%; mean CPAP use 4.4 (SD 1.9) h/night	Primary outcome: between-group difference in change in AF burden (percent of time in AF) from the 1-month period prior to treatment to the final 3 months of the 5-month treatment period Secondary outcome: percent of patients with a ≥25% reduction in AF burden Post-hoc analyses: (1) frequency of AF episodes; (2) duration of AF episodes	AF burden decreased from 5.0% to 4.3% in the control group and from 5.6% to 4.1% in the CPAP group (adjusted mean difference −0.6%, 95% CI −2.6 to +1.3%, p = 0.52). No significant difference in “per protocol” analysis including only those using CPAP ≥4 h/night. The percent of patients with a ≥25% reduction in AF burden was 41% in the CPAP group and 31% in the usual care group (adjusted mean difference −9.3%, 95% CI −26.4 to +8.7%, p = 0.33). No significant difference in frequency or duration of AF events. Serious adverse events occurred in 13% of CPAP and 4% of control patients.