PFA Under Deep Sedation and Non‐Invasive Monitoring in Atrial Fibrillation Patients With Obstructive Sleep Apnea

ABSTRACT

Background

Clinical data about pulsed field ablation (PFA) under deep sedation among patients with obstructive sleep apnea (OSA) are limited, while the optimal periprocedural monitoring is not established. We report our experience in PFA among patients with atrial fibrillation (AF) and OSA, using a continuous non‐invasive respiratory and hemodynamic monitoring protocol.

Methods

A total of 32 patients with AF suffering from OSA (mean age 68.5 ±9 years, 31% female) were compared with 60 consecutive patients without OSA. The monitoring protocol comprised continuous beat‐to‐beat blood pressure recording via an infrared light system and near‐infrared spectroscopy for both cerebral and peripheral tissue oximetry. The Acumen Hypotension Prediction Index (HPI) that detects hypotensive trending was also assessed.

Results

Three of the patients with OSA (9%) and four of the controls (6%) developed transient decrease of cerebral saturation during induction to deep sedation, early detected and effectively managed with airway manoeuvres and supplemental oxygen up‐titration. In nine patients with OSA (28%) and twenty controls (33%), increased HPI values followed by transient decrease in blood pressure were detected after induction to sedation and/or during PFA, responsive to intravenous fluids. The average time of hypotension was comparable between patients with and without OSA (2.69% vs. 3.23% of the total procedure time, respectively). No serious complications occurred.

Conclusion

PFA under deep sedation in patients with OSA and AF was not associated with higher incidence of adverse events compared to those without OSA. This is confirmed by a sensitive continuous non‐invasive monitoring protocol, which could be further evaluated by dedicated multicenter studies.

1. Introduction

The single‐shot irreversible electroporation by pulsed field ablation (PFA) for pulmonary vein isolation (PVI) seems to gain ground against other PVI techniques in patients with paroxysmal atrial fibrillation (PAF) [1]. PFA under deep sedation has become a routine in progressively more electrophysiology laboratories but requires intensive monitoring, especially for patients with high body mass index (BMI) and obstructive sleep apnea (OSA) [2, 3, 4]. Moreover, a uniform protocol of monitoring and sedation during PFA for PAF has not been established for the time being.

OSA is associated with the presence of atrial fibrillation (AF) and increases the rate of AF recurrence following catheter ablation [2, 5, 6, 7]. Besides, during AF ablation under deep sedation, a risk for obstructive respiratory events due to collapse of the upper airways has been reported [8, 9]. On top of this, a considerable number of patients with OSA remain undiagnosed due to mild or underestimated symptoms, and this may increase the risk during ablation [10, 11, 12]. For all these reasons, patients with OSA are underrepresented in clinical studies evaluating ablation technologies [13, 14, 15].

The purpose of this study is to test whether our experience supports the hypothesis that patients with OSA might be more vulnerable to adverse events during deep sedation than those without. Therefore, we compare the incidence of potential adverse respiratory and hemodynamic events during PVI with PFA under deep sedation between a specific population of patients with AF and OSA and patients without OSA, using continuous non‐invasive respiratory and hemodynamic monitoring during the procedure.

1.1. Materials & Methods

Between October 2023 and October 2025, adult patients (18–80 years of age) with symptomatic AF resistant to antiarrhythmic drugs, who reported concomitant OSA at their admission, diagnosed by experts on the basis of symptoms and diagnostic testing [16, 17], were prospectively evaluated for inclusion in our study. They had to be admitted for their first PVI procedure and had their antiarrhythmic drugs interrupted before the ablation. During the same period, consecutive patients with drug‐resistant AF without OSA served as the control group. Oral anticoagulants were discontinued the day before admission. Patients with severe heart failure (ejection fraction <30% or NYHA class >II), recent acute coronary syndrome (≤3 months), severe renal impairment (estimated glomerular filtration rate <30 mL/min/1.73m²), moderate or severe chronic obstructive pulmonary disease (COPD) as defined by the Global Initiative for Chronic Obstructive Lung Disease or recent respiratory infection were excluded [18, 19]. The study was approved by the Hospital's Ethics Committee (Clinical Study Registration Number: 650/18‐11‐2024) and was conducted according to the principles of the Declaration of Helsinki, while study participants provided written consent prior to the procedure.

1.2. Monitoring Protocol

As soon as patients entered the electrophysiology laboratory, continuous noninvasive monitoring was established using the Acumen IQ cuff (Edwards Lifesciences, Irvine, CA, United States). Therefore, a finger cuff combining i) an infrared light system, ii) an inflatable bladder to accurately measure continuous beat‐to‐beat blood pressure and iii) a light sensor, was attached to their finger and connected with a pressure controller component at the wrist and the heart reference sensor at the heart level. Two tissue oximetry sensors (ForeSight) were attached to the patient: one at the forehead to monitor cerebral tissue oxygen saturation (StO₂ A₁) and another at the forearm for peripheral tissue oximetry (StO₂ A₂) monitoring. All the assessed hemodynamic and respiratory parameters were shown on a bedside screen with the option of jump‐overs to different screen setups. Hemodynamic parameters included pulse rate (PR), systolic and diastolic arterial blood pressure (SBP and DBP, respectively), mean arterial pressure (MAP), cardiac output (CO), cardiac index (CI), stroke volume (SV), stroke volume index (SVI), systemic vascular resistance (SVR), systemic vascular resistance index (SVRI), dP/dt, and dynamic arterial elastance (Ea_dyn) defined as the ratio between pulse pressure variation and SV variation. Additionally, the Acumen Hypotension Prediction Index (HPI) that detects potential hypotensive trending of a patient's MAP was assessed and updated every 20s [18].

1.3. Deep Sedation Protocol

Deep sedation was initiated before femoral vein puncture with intravenous administration of midazolam (1–2 mg bolus) and continuous propofol administration at a rate of 50–150 mg/hr. Additionally, an initial intravenous bolus of 0.025 mg fentanyl was administered, which was repeated before the transeptal puncture (TSP). In all patients a nasopharyngeal airway was inserted, and in those who were under treatment with a continuous positive airway pressure (CPAP) device during sleep at home, the device was used throughout the procedure at the previously adjusted settings. The remaining patients received an oxygen supply via a Venturi mask (8–10 L/min, FiO₂ 35%) until the end of the ablation procedure. Before the ablation, an additional dose of fentanyl 0.025 mg was administered, followed by equal bolus doses during the ablation, if necessary. The rate of continuous propofol infusion was titrated up to 200 mg/hr, according to the level of sedation and the indications obtained by the ClearSight system of continuous noninvasive monitoring. Additional bolus doses of midazolam 1–2 mg were administered, accordingly. The depth of sedation was recorded using a Bispectal Index (BIS) monitor (BIS LoC 2 Channel, Covidien, Singapore) with a target‐range between 50 and 60. [20, 21]All electrophysiologists and electrophysiology (EP) nurses were trained in management of sedation, had intensive care experience, and advanced life support training. Besides, if intubation was required, a resuscitation team was consistently available in the hospital.

1.4. Ablation Protocol

A decapolar diagnostic catheter was placed in the coronary sinus as a guide for the TSP and for backup pacing in case of bradycardia during ablation. The TSP procedure was performed under fluoroscopic guidance using a Brockenbrough needle via a SL1 or SL0 sheath (Abbott, Swartz) and a guidewire (CHOICE Floppy Guidewire, Boston Scientific) was used to confirm that the needle was advanced in the left atrium (LA). After TSP, heparin boluses were administered targeting an activated clotting time >300s.

Subsequently, the transeptal sheath was withdrawn, and the 13‐F FARADRIVE steerable sheath was inserted in the LA over the wire, and the 12‐F over‐the‐wire PFA catheter (FARAWAVE, FARAPULSE, Boston Scientific Inc.) was introduced such that the splines achieve circumferential contact/proximity with the pulmonary vein (PV) antra. The catheter was deployed in a basket and flower configuration and rotated between applications to ensure circumferential PV ostial and antral coverage without gaps. Eight bipolar and biphasic pulse trains 2 kV/2.5 sec (4 in a basket and 4 in a flower configuration) were delivered to each PV, starting with the left‐sided veins. Additional pulse trains were optionally delivered, according to the local electrograms. Prior to the first pulse delivery, atropine 0.04 mg/kg was given intravenously to avoid vagal reactions such as sinus arrest or intermittent atrioventricular block. PVI was considered successful if the absence of the local PV electrograms obtained by the FARAWAVE catheter was observed, along with low voltage areas < 0.1 mV around the PV ostia on the voltage map created by the EnSite Precision Cardiac Mapping System (Abbott Medical, St. Paul, MN, USA) and bidirectional conduction block during pacing in sinus rhythm.

1.5. Events and Clinical Outcomes

During ablation, StO2 A₁ and StO2 A₂ were carefully attended, and a decrease below normal range (60%–80%) or more than 10% was considered the alarm limit for a potent apneic event. Additionally, HPI values above 85 for two consecutive measurements were interpreted as a sign of an imminent hypotensive event, and MAP values were estimated accordingly. Hypotensive events were defined as MAP< 65 mmHg [22, 23, 24]. The primary endpoint was to assess the incidence of actionable cerebral tissue desaturation during PVI. The secondary endpoint was to detect and evaluate the number of hypotensive episodes, as stated above, or other remarkable hemodynamic adverse events during ablation.

1.6. Statistical Analysis

Statistical analysis was performed using the SPSS software to compare baseline characteristics of the patients between groups. Mean values and standard deviations were used to describe continuous variables. Continuous variables were assessed for normality and were found to follow a normal distribution. Therefore, comparisons between groups were conducted using the independent t‐test for continuous variables and the chi‐square (X ²) test for categorical variables. The significance level of p < 0.05 was considered statistically significant. Data derived from the non‐invasive monitoring system were retrospectively analyzed using the Acumen Analytics software from the HemoSphere advanced monitoring platform.

2. Results

During the study period, thirty‐four patients with AF and OSA were enrolled in the study. Two of them were excluded due to low quality of the respiratory and hemodynamic recordings during the ablation. All patients included in the OSA group were classified as having moderate or severe OSA (apnea‐hypoxia index>15/h), diagnosed by polysomnography, and treated with continuous positive airway pressure for at least the last 3 months, according to the experts’ indications [25, 26, 27, 28]. During the same time period, sixty consecutive patients with AF and no history of OSA or OSA‐related symptoms reported served as the control group (control group). Patient characteristics are presented in Table 1.

TABLE 1.

Patients’ demographics.

Variables	OSA group (N = 32)	Control group (N = 60)	p value
Age	68.5±9*	66.8±8*	NS
Body mass index	27.1±2*	28.3±3*	NS
Male gender	20 (62.5%)	36 (60%)	NS
Paroxysmal AF	22 (69%)	42 (70%)	NS
Hypertension	22 (69%)	45 (75%)	NS
Diabetes mellitus	3 (9%)	7 (12%)	NS
Coronary artery disease	4 (12.5%)	9 (15%)	NS
Dyslipidemia	16 (50%)	29 (48%)	NS
HFrEF	1 (3%)	3 (5%)	NS
Mitral regurgitation	1 (3%)	3 (5%)	NS
Stroke	2 (6%)	3 (5%)	NS
Smoking	5 (16%)	12 (20%)	NS
Hypothyroidism	3 (9%)	5 (8%)	NS
Hyperthyroidism	2 (6%)	5 (8%)	NS

Abbreviations: AF, atrial fibrillation; HFrEF, heart failure with reduced ejection fraction; N, number; NS, statistically non‐significant *mean value ± standard deviation.

The PVI ablation procedure was successful in all patients. No serious procedure‐related complications were observed, except two cases of femoral artery injury, requiring transcutaneous embolism by interventional radiologists. The total procedure time was 103±27 min for patients with OSA and 95±22 min for controls (p = 0.19). In all patients, mild cough and skeletal muscle recruitment were observed during some PFA applications, without a map shift or dislodgement of the catheters affecting the smooth procedure workflow. The average number of PFA applications was 66±17 in the OSA group and 68±15 in the controls. In the 24‐hour post‐ablation follow‐up, the patients did not report any memories of painful experiences or chest discomfort, either during the ablation or immediately following the procedure until discharge.

Throughout the ablation, instantaneous values and continuous curves of the monitored parameters were obtained by the non‐invasive monitoring system (Figure 1). Changes in recorded parameters beyond the alarm threshold are presented in Table 2. In three cases of the OSA group and in four patients among controls, a decrease in the cerebral StO2Α₁ values beyond the alarm limits was observed during the procedure (Figure 2). Guided by the doctors of the EP laboratory and implemented by the trained EP nurses, a temporary decrease in propofol administration along with airway management maneuvers (head tilt‐chin lift and jaw‐thrust) and an increase in oxygen supplementation reverted the levels of oxygen saturation to their initial levels within the next few minutes.

FIGURE 1.

Parameters derived from the ClearSight system: indicatively, the central and peripheral tissue oxygen saturation (StO₂ A₁ and StO₂ A₂, respectively), the Hypotension Prediction Index (HPI), the mean arterial pressure (MAP), the pulse rate (PR), the stroke volume (SV) and the dP/dt ratio are presented. In this patient, one can notice that induction to deep sedation (DS) was associated with increased HPI, along with low MAP. During the procedure, AF was induced (asterisk) followed by a decrease in SV and dP/dt. During PFA, an increase in HPI and decrease in MAP was recorded, which was restored after PFA and sinus rhythm restoration (arrow).

TABLE 2.

Procedure time and non‐invasive respiratory and hemodynamic parameters during the procedure.

	OSA group (n = 32)		Control group (n = 60)
Parameters	N	% of total	N	% of total	p value
StO2 drop	3	9	4	7	NS
Hypotension	9	28	20	33	NS
After sedation	5 (55%)	16	10 (50%)	17	NS
During PFA	7 (78%)	22	16 (80%)	27	NS

Abbreviations:HPI, hypotension prediction index; N, number; NS, not significant; PFA, pulsed field ablation; StO2, tissue oxygen saturation.

FIGURE 2.

Before PFA, a significant decrease in cerebral tissue oxygenation (StO₂ A₁—green line) was observed following continuous propofol and bolus fentanyl administration for deep sedation (arrow). It was soon restored after head tilting and the temporary interruption of propofol intravenous injection. No significant decrease in peripheral oxygenation was observed (StO₂ A₂).

In nine patients with OSA (28%), temporary hypotension preceded by an increase in HPI levels reaching the alarm threshold of 85 was observed (Table 2) right after the induction to deep sedation or/and abruptly during PFA application (Figure 3). These rates of HPI increase were comparable with those in the control group, as presented in Table 2. Administration of 0.9% normal saline solution resulted in the restoration of normal HPI and MAP values. The average time of hypotension during the ablation procedure (2.8±3.1 min) represented only 2.69% of the total procedure time and did not differ from controls (3.23%, 3.1±3.0 min, p = 0.55) (Figure 4). In the cases of hypotension after the induction to sedation, the HPI increase was simultaneous with, or preceded the MAP fall by 4.9 ± 2.7 min. In patients with short and abrupt drops in MAP during PFA applications, HPI increase was simultaneous with hypotension, and the next application of energy was postponed for a few minutes until blood pressure returned to baseline.

FIGURE 3.

During the ablation procedure, an increase in HPI values above the limit of 85 was observed among some patients from both OSA group and controls. The three different patterns of increase, along with the MAP curve, are presented in this figure: (A) in some patients HPI increased following the induction to sedation, (B) in others HPI abruptly increased during PFA, and (C) in some cases, HPI increased after the induction to deep sedation, but also during PFA, with a quick recovery in both circumstances. HPI: hypotension prediction index, MAP: mean arterial pressure, PFA: pulsed field ablation.

FIGURE 4.

The average procedure time, the average duration of hypotensive events and the average percentage of time in hypotension per patient are presented in the data summary (Acumen Analytics software).

OSA: obstructive sleep apnoea, MAP: mean arterial pressure, min: minutes.

3. Discussion

In this observational study, we present our experience with PFA for PVI under deep sedation in the specific group of patients with AF and OSA in comparison to counterparts with AF but not OSA, using a continuous non‐invasive respiratory and hemodynamic monitoring protocol to detect possible adverse events and evaluate the safety of the procedure. Our observations suggest that (1) PFA under deep sedation in this group of patients is as safe as in patients without OSA, and its acute efficacy is similarly high; (2) non‐invasive continuous monitoring provides reliable information regarding the respiratory and hemodynamic condition of the patients and could be used as an additional tool to ensure safety in the procedure.

Data from previous observational studies suggest that catheter ablation for AF is safe in patients with OSA and its long‐term efficacy, especially when combined with CPAP treatment, is comparable to that in patients without OSA [25]. However, these data derive from various ablation procedures, while detailed information regarding the use of general anesthesia or deep sedation is not provided. Additionally, the experience in the use of PFA and analgesia in such patients has not been reported yet. Although clinical data show that deep sedation with midazolam, propofol, and fentanyl administered by the electrophysiology laboratory personnel during left atrial ablation procedures is safe, these respiratory depressants could cause collapse of the upper airways, and the risk of acute hypercapnia and respiratory acidosis in patients with OSA still exists [8, 9, 26]. This raises the question about the need of real‐time monitoring for the early detection of such events and the subsequent proper respiratory management of these vulnerable patients [29].

While cerebral tissue hypoxia typically follows peripheral oxygen saturation fall during apneic events (possibly due to effective central autoregulation), the reverse sequence has been described in various hemodynamic situations, like vasovagal syncope [30, 31]. Therefore, cerebral tissue oximetry has been proposed for patients’ monitoring in case of respiratory emergencies and pulseless situations. The ForeSight system uses near‐infrared spectroscopy (NIRS) for non‐invasive monitoring of cerebral oxygenation, and it has been proved more sensitive than standard monitoring in predicting changes in cerebral oxygenation during arrhythmias and circulatory arrest, as well as in identifying incomplete cerebral recovery, despite seemingly restored hemodynamics [32, 33, 34]. Therefore, it has been suggested as a continuous non‐invasive monitoring method for perioperative assessment during cardiac or non‐cardiac surgery [35, 36]. Additionally, it has been previously validated in sedated patients undergoing interventional electrophysiology procedures and transcatheter aortic valve replacement [37, 38, 39].

In our study, major respiratory events did not occur with the sedation protocol described. The minor decrease in cerebral tissue oxygenation levels was comparable between the OSA group and controls and could be attributed to the cumulative effect of sedatives. Furthermore, it served as a warning for an imminent respiratory event that led to prompt and successful airway management along with titration of drug administration and hence rapid recovery of the respiratory parameters. As previously suggested, intraoperative cerebral oxygen desaturation (StO₂ A₁< 60%) or a decrease by 10% should be taken into consideration, and factors that influence oxygen saturation should be investigated [18, 33]. Our experience shows that in the case of StO₂ reduction, the decrease in propofol infusion rate and the increase of supplemented oxygen concentration can be effective in restoring normal breathing. Non‐invasive early indications regarding cerebral tissue oximetry might be a useful tool for avoiding serious obstructive respiratory events during a PFA procedure.

Although PFA demonstrates a favorable safety profile, complications with potentially serious hemodynamic impact have been reported, such as cardiac tamponade, coronary artery spasm, and vascular complications, all manifested in blood pressure hemodynamics [40]. Therefore, apart from respiratory monitoring, continuous and reliable monitoring of hemodynamic parameters is mandatory during the ablation procedure. The non‐invasive monitoring system that we used has been proposed as a reliable tool for hemodynamic monitoring during induction to anesthesia for cardiac surgery, while the HPI has been reported to provide accurate predictions of impending intraoperative hypotension in patients undergoing surgery under general anesthesia [37, 41, 42, 43]. In both patients with and without OSA, the blood pressure fall, when observed, represented a small proportion of the total ablation time and was attributed to the induction of sedation and vasovagal reactions during the application of energy. In patients with hypotension after deep sedation because of the gradual and cumulative vasodepressive effect of the sedatives, the earlier HPI elevation served as a warning sign of the impending hypotension, while none of our patients developed periprocedural hypotension with normal HPI levels [44]. In the cases of simultaneous HPI increase and transient hypotension during ablation, the acute vagal stimulation could explain this timing.

3.1. Study Limitations

We present the preliminary experience of a single center on PFA under deep sedation with a limited number of patients with AF and OSA. The relatively small number of enrolees combined with the low incidence of adverse events might have limited the statistical power for this analysis. Larger studies would enrich our experience in the application of this ablation technique among these patients.

Major respiratory and hemodynamic events did not occur in our patients during the PFA procedure. The relatively low number of cases developing desaturation might be due to the usage of CPAP during ablation and the exclusion of patients with severe COPD or morbid obesity. Additionally, hemodynamic instability was promptly addressed in accordance with the HPI values, as ethically required.

Although patients in the control group did not report symptoms or a history of OSA, the inclusion of individuals with at least mild apnea cannot be ruled out.

4. Conclusion

Our experience on PFA under deep sedation combined with continuous non‐invasive respiratory and hemodynamic monitoring is not aligned with the assumption that OSA might be associated with more respiratory or hemodynamic adverse events in patients suffering from OSA and AF than in those without OSA. In addition, this report demonstrates that non‐invasive monitoring is sensitive, reliable and beneficial for patients at elevated risk of respiratory and hemodynamic complications, such as those with OSA, thereby enhancing the safety of ablation procedures conducted under deep sedation.

Author Contributions

Conceptualization: Dionyssios Leftheriotis and Panagiota Flevari. Methodology: Dionyssios Leftheriotis, Panagiota Flevari, and Konstantinos A. Papathanasiou. Investigation: Dionyssios Leftheriotis, Panagiota Flevari, Konstantinos A. Papathanasiou, Areti Vlachodiamanti, Evangelia Stamatopoulou, and Maria Polikandrioti. Software: Dionyssios Leftheriotis, Konstantinos A. Papathanasiou, and Efstratios Karamanolis. Data curation: Dionyssios Leftheriotis, Konstantinos A. Papathanasiou, and Efstratios Karamanolis. Formal analysis: Dionyssios Leftheriotis and Efstratios Karamanolis. The first draft of the manuscript was written by Dionyssios Leftheriotis and Konstantinos A. Papathanasiou, and all authors commented on previous versions of the manuscript. Visualization: Dionyssios Leftheriotis, Panagiota Flevari, Maria Polikandrioti, and Konstantinos A. Papathanasiou. Supervision: Dionyssios Leftheriotis, Panagiota Flevari, and Maria Polikandrioti. All authors have read and agreed to the published version of the manuscript.

Funding

The authors have nothing to report.

5.

The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of, Attikon University Hospital, National and Kapodistrian University of Athens, Athens, Greece.

Consent

Informed consent was obtained from all subjects involved in the study.

Conflicts of Interest

The authors declare no conflicts of Interest.

Data Availability Statement

Data available upon reasonable request.