Sleep apnea severity in patients undergoing atrial fibrillation ablation: Home sleep apnea‐test and polysomnography comparison

Abstract Background Sleep apnea (SA) is highly prevalent and should be treated in patients referred for catheter ablation (CA) of atrial fibrillation (AF). Watch‐type peripheral arterial tonometry (WP) for home SA testing has demonstrated a high correlation of the apnea‐hypopnea index (AHI) with Polysomnography (PSG), but the evidence of its accuracy in AF patients is not adequate. Methods This study was conducted under a retrospective, single‐center, observational design. We included 464 consecutive AF patients (age 65 ± 11 years, 76.5% male, 45.0% paroxysmal‐AF) who received both WP and PSG during the periprocedural period of the CA. We compared the AHI using the WP (WP‐AHI) to that using PSG (PSG‐AHI). Results The WP‐AHI was 25.9 ± 12.7 and PSG‐AHI 31.4 ± 18.9 (r = .48). Among 325 patients with a WP‐AHI < 30, 116 (35.7%) exhibited a PSG‐AHI ≥ 30. Only 12.5% of the patients were indicated for continuous positive airway pressure (CPAP) treatment only by the WP‐AHI, while 70.9% were indicated for CPAP by the PSG‐AHI according to the Japanese health insurance system. The best cut‐off value of the WP‐AHI was 18.1 to predict a PSG‐AHI ≥ 20 with an area under the curve of 0.72 (95% confidence interval, 0.67–0.76). Conclusions The WP‐AHI and PSG‐AHI were weakly correlated in AF patients receiving CA. About one‐third of the patients with moderate SA using the WP was diagnosed with severe SA evaluated by PSG. The majority required PSG for the CPAP indication.


| INTRODUC TI ON
Atrial fibrillation (AF) and sleep apnea (SA) are common, and both have a risk of cardiovascular morbidity and mortality. 1,2 SA is highly prevalent in AF patients because of the linking mechanisms between SA and AF including activation of the sympathetic nervous system during intermittent hypoxia and structural remodeling of the left atrium through the negative intrathoracic pressure during an obstructive apnea episode. 3,4 Experimental, epidemiological, and clinical studies have consistently shown that SA is a potential risk factor for the incidence and recurrence of AF. 5 A meta-analysis reported that obstructive SA (OSA) is associated with a 31% greater risk of AF recurrence after catheter ablation (CA) of AF. 6 Treating OSA with continuous positive airway pressure (CPAP) reduces the AF recurrence after pulmonary vein isolation. 7,8 Nevertheless, OSA is frequently undiagnosed and consequently undertreated in AF patients receiving CA.
We previously reported that patients undergoing CA of AF had a high prevalence of SA using a watch-type peripheral arterial tonometry (WP) device for the home sleep apnea test (HSAT) even if they did not have sleepiness or risk factors such as obesity and hypertension. 9 Although polysomnography (PSG) is currently considered the gold standard diagnostic test for SA, carrying out PSG in all AF patients is difficult in clinical practice due to the limited access and cost issues. 10 HSAT using WP is a simplified alternative screening tool, which can be performed in all AF patients. 11 Although a high degree of correlation between the apnea-hypopnea index (AHI) derived from WP (WP-AHI) and PSG (PSG-AHI) has been demonstrated in patients suspected of SA, the evidence of the accuracy in AF patients is not adequate.

| Patient population and study design
This was a retrospective, single-center, and observational designed study to compare an automatically scored WP-AHI with the PSG-AHI in patients that underwent CA of AF. Out of 1435 patients who underwent CA of AF from March 2018 to December 2020 in our institute, we enrolled a total of 464 patients who received sleep studies both at home and during hospitalization including 2 patients that received alpha blockers. The HSAT was conducted using a Watch-PAT200U (WP) (Watch-PAT; Itamar Medical Ltd.) before the CA of AF. An in-hospital examination was conducted using PSG after the CA of AF. We sent the WP to the patients' homes after we decided in the outpatient clinic to perform a CA of AF. The patients self-administered the WP in their homes and returned both the WP and screening questionnaire to us before their admission for CA.
We performed the PSG in a post-operative stable condition at least 1 day after the CA of AF. During the CA procedures, most patients underwent conscious sedation with a combination of a bolus of thiamylal sodium, pentazocine, and continuous dexmedetomidine hydrochloride, except for 10 patients who underwent general anesthesia. 12 Almost all patients received the PSG within 1 week after the CA. The mean duration between the CA and PSG was 9.3 days (median 1 day). All patients gave their informed consent for both the ablation procedures and the use of their clinical data in a retrospective study. This study complied with the Declaration of Helsinki and was performed according to the institutional ethics committee's approval.

| Watch PAT system
The apnea-hypopnea index measurements using the WP (WP-AHI) consisted of the peripheral arterial tonometry (PAT) signal, oxygen saturation, heart rate, wrist activity (actigraphy), snoring, and body position. The WP indirectly detected any apnea-hypopnea events by selectively measuring the peripheral arterial volume changes using a finger-mounted plethysmograph. This information was collated with the pulse oximetry in conjunction with the heart rate and was further analyzed using a predeveloped automated computer program.
The WP provided an algorithm able to differentiate between sleep and awake states every 30 s and to calculate both the total sleep time and total recording time. PAT-based device determinations of the AHI (i.e., WP-AHI) were computed automatically.

| Polysomnography
All patients underwent overnight PSG during hospitalization using a digital polygraph system (Alice; Philips Respironics). The apnea and hypopnea events were quantified, and the SDB severity was assessed using the frequency of the apnea and hypopnea events per hour of sleep (i.e., PSG-AHI).

| Definitions and endpoints
The diagnosis of sleep apnea was determined according to the Adult Obstructive Sleep Apnea Task Force of the American Academy of Sleep Medicine. A normal sleep study was defined as an AHI < 5, mild sleep apnea as an AHI range ≥ 5 and <15, moderate sleep apnea as an AHI range ≥ 15 and <30, and severe sleep apnea as an AHI ≥ 30. The CPAP indications were defined as an AHI ≥ 20 assessed by PSG or an AHI ≥ 40 assessed by HSAT according to the current Japanese health insurance system. The Epworth Sleepiness Scale (ESS) is a commonly used 8-item questionnaire to assess subjective daytime sleepiness. 13 The ESS is unidimensional and is closely related to the frequency of apnea events during sleep apnea. The ESS scores were collected by sending a paper questionnaire to the patients' homes together with the WP before the CA. Paroxysmal AF was defined as an AF episode that terminated spontaneously or following the administration of antiarrhythmic drugs within 7 days of the onset. Persistent AF was defined as an AF episode lasting for more than 7 days and up to 1 year. Long-lasting AF was defined as that persisting for more than 1 year.
The aim of this study was to compare the automatically scored WP-AHI with the PSG-AHI and to elucidate the risk of a CPAP indication assessed using PSG. Consequently, the primary endpoints of the study were (1) the distribution of the WP-AHI and PSG-AHI among the patients and (2) the predictors of a CPAP indication (PSG-AHI ≥ 20) among them.

| Statistical analysis
The descriptive statistics are reported as the mean ± SD for continuous variables and as absolute frequencies and percentages for categorical variables. The parametric data were compared using a Student's t-test or paired Student's t-test, and the non-parametric data were compared using the Mann-Whitney U-test, Wilcoxon signed rank test, chi-squared test, or Fisher's exact test, as appropriate. A receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut-off value of the WP-AHI to predict a CPAP indication assessed using PSG (AHI ≥ 20). Those cutoff values were then used to stratify the patients. A multivariable logistic regression analysis was performed to determine the risk factors for a CPAP indication by PSG using the following variables:

| Overall description of the sample
Overall, 464 patients were enrolled. The baseline characteristics are listed in Table 1. The mean age was 64.9 ± 10.6 years, mean BMI 24.6 ± 3.5 kg/m 2 , and 76.5% of the patients were men. The prevalence of paroxysmal AF was 45%. The mean CHADS 2 score and CHA 2 DS 2 -VASc score were 1.17 and 2.00, respectively. The mean ESS scale was 6.8 ± 4.5.

| Distribution of AHI
The distribution of the WP-AHI and PSG-AHI is shown in Figure 1

| Correlation between the WP-AHI and PSG-AHI
There was a correlation between the WP-AHI and PSG-AHI (r = .48,    (Table 2). However, the prevalence of a CPAP indication in patients with a female sex, without obesity, and without sleepiness, was 57.8%, 65.8%, and 68.1%, respectively ( Figure 6).

| Main findings
The present study to compare the WP-AHI with the PSG-AHI for un-

| Sleep apnea in AF patients receiving CA
Sleep apnea is a risk factor for AF recurrence after CA. Several meta-analyses in some nonrandomized studies have established that CPAP treatment for severe OSA is associated with higher AF-free survival rates in AF patients undergoing CA. 6 The WP is a type of HSAT that allows an automatic calculation of the scoring. Previous studies showed a strong correlation between the AHI using the WP and the simultaneously recorded AHI using PSG in patients with cardiovascular diseases including AF. 14

| Weak correlation between the WP-AHI and PSG-AHI
OSA screening using HSAT before the CA is more cost-effective than when using PSG in AF patients. 17 Many studies using the WP have excluded patients with arrhythmias due to the potential effect of the F I G U R E 4 Indication for CPAP (PSG-AHI ≥ 20) using PSG according to the severity assessed using the Watch PAT (WP-AHI).

F I G U R E 5 Receiver operating characteristic curves of the WP-AHI for a PSG-AHI ≥ 20 (CPAP indication).
arrhythmias on the peripheral arterial tonometry amplitude and rate changes. Tauman R et al. showed that a strong correlation was found between the PSG-AHI and WP-AHI (r = .80, p < .0001), and the presence of AF during the night did not decrease the precision of the WP-AHI when full in-lab PSG and WP were conducted simultaneously in patients with AF. 18 We revealed that the AHI obtained by PSG in our institution after the CA was higher than that of the WP at home before the CA. More than half of the mild sleep apnea patients assessed by the WP were diagnosed with moderate or severe sleep apnea by PSG, and 42% of the moderate sleep apnea patients assessed by the WP were diagnosed with severe sleep apnea by PSG.
There were 3 possible reasons why a weak correlation was found between the WP-AHI and PSG-AHI in the current study which are as follows: (1) the difference in the tools used, (2) rhythm differences, and (3) the night-to-night variability.
1. In the aforementioned trial, WP-AHI was well correlated with the PSG-AHI when performing the WP and PSG simultaneously in AF patients. 18 The number of study subjects was relatively small (n = 101) and did not have a perfect correlation (r = .80).
2. Some studies showed that conversion from AF to sinus rhythm reduces the AHI following CA or cardioversion. Naruse Y et al. 3. There is a considerable number of studies assessing the night-tonight variability in the AHI. 20

| Study limitations
The main limitations of this study were associated with the retrospective, observational design. We believe, however, the findings obtained in our clinical practice compare favorably with those of a well-organized prospective study because we collected sequential data on unscreened sleep apnea patients with AF who underwent the WP and PSG during a relatively short period, and the sample size was large (n = 464). Second, CA might have influenced our results because the WP was performed before the CA, and the patients received PSG after the CA. Third, the WP demonstrated a high degree of correlation of the AHI with the PSG, but the evidence of the accuracy in patients with AF or in those who have central sleep apnea is not adequate. Fourth, this study was based on a single-center experience. To ensure that AF patients should receive PSG after the HSAT, our results should be confirmed by a multicenter study.

| CON CLUS IONS
In AF patients receiving catheter ablation, there was a significant variation between the WP-AHI and PSG-AHI. About one-third of the patients with mild or moderate sleep apnea using WP was diagnosed with severe sleep apnea evaluated by PSG. Even if routine sleep apnea screening using HSAT before CA results in only mild sleep apnea, PSG should be considered especially in patients that are male or have daytime sleepiness, in order to determine an indication for CPAP treatment.

ACK N OWLED G M ENTS
We thank the nursing staff, clinical engineers, and office administrators of Sakurabashi-Watanabe Hospital for their invaluable help in conducting this study. We also thank Mr. John Martin for his linguistic assistance with the manuscript.

CO N FLI C T O F I NTER E S T S TATEM ENT
N. Tanaka has received honoraria from Johnson and Johnson KK, Bayer, and Boehringer Ingelheim. The other authors declare no conflicts of interest.

E TH I C S A PPROVA L S TATEM ENT
The present study was approved by Sakurabashi Watanabe Hospital (Reference number. 21-89).

PATI ENT CO N S ENT S TATEM ENT
Informed consent was obtained in the form of opt-out on the website. Those who rejected were excluded.