Sex differences in characteristics of atrial fibrillation recurrence post surgical pulmonary vein isolation

Graphical abstract


Introduction
The age-related prevalence of atrial fibrillation (AF) is higher in males than in females.[2][3][4] Female patients with AF have, however, more symptoms [4] and their AF episodes are longer in duration.[5] Consequently, females with AF experience a poorer quality of life.[4][5][6][7] Female patients with AF are less often referred for pulmonary vein isolation (PVI) [4,6] and also experience more recurrences after catheter based PVI.[8][9][10] These observations indicate that the AF-related arrhythmogenic substrate may be more complex in females.Indeed, electro-anatomical mapping prior to endovascular PVI revealed more slowing of conduction, a higher degree of fractionation and lower voltages of left atrial bipolar electrograms in females.[11] Data on sex differences in outcome of surgical PVI are sparse and limited to open chest surgery.AF recurrences assessed by 24-hour continuous rhythm monitoring at 6 and 12 months after a Cox-Maze procedure performed in 540 patients did not differ between males (N = 356, 66 %) and females (N = 184, 34%).[12] Likewise, AF Abbreviations: AF, Atrial Fibrillation; BMI, body mass index; CRM, continuous rhythm monitoring; CV, conduction velocity; OAC, oral anticoagulation; RF, radio frequency; SR, sinus rhythm; VATS PVI, video assisted thoracoscopic pulmonary vein isolation.recurrences were also not different between 251 matched male and female patients after concomitant arrhythmia surgery.Male and female patients had similar AF recurrence rates (28.4% in male patients versus 25.4% in female patients).Additionally, when having AF recurrences, male and female patients did not differ from each other in receiving therapy (cardioversion or catheter ablation) to restore sinus rhythm.[13] Surgical video assisted thoracoscopic (VATS) PVI is nowadays widely adopted and has a higher efficacy compared to catheter based PVI, as reported in a systematic review in which surgical PVI was compared with catheter based PVI.[14] However, it is not known whether there are sex differences in outcomes of VATS PVI.The goal of the present study was therefore to investigate whether there are sex differences in VATS PVI outcomes by comparing not only the occurrence but also the characteristics of AF episodes between matched male and female patients preoperatively-and one-year post-VATS PVI.

Patient inclusion and data collection
This is a retrospective single center study performed at the OLVG Amsterdam, The Netherlands.
All patients included who did not receive a prior endovascular PVI or an atrial pacing lead have been selected from an existing database from a prior conducted study by the cardiothoracic department of the OLVG hospital [1] containing patients who underwent a VATS PVI between 2012 and 2017.Clinical data were collected from an electronic patient information system and male and female patients were subsequently matched on age and body mass index (weight in kilograms divided by height in metres (m) squared).
Approval was granted by the local ethics committee (WO-19.109).The project adhered to the Declaration of Helsinki principles and written consent was obtained from participating patients before the surgical intervention.

Surgical procedure
Surgery was performed thoracoscopically and bilaterally under general anaesthesia with a double-lumen endotracheal tube to facilitate single-lung ventilation of the patient in supine position with both arms alongside and a few centimetres dorsal of the body.A transoesophageal echo was inserted to evaluate thrombus formation in the left atrial appendage prior to the incision and after that, retracted from the patient to prevent damage to the oesophagus during the ablation procedure.
On both sides, two 12 mm and one 5 mm thoracoscopic ports were introduced.Surgery was performed under CO2 insufflation to optimize visualization during the procedure.After opening the right hemithorax, with selective left lung ventilation, the phrenic nerve was identified, and the pericardium on the right side was opened 2 cm above it and retracted with two stay sutures.Blunt dissection of the oblique and transverse sinuses was performed and guiding wires were directed through these sinuses.Ganglionated plexi were identified by focal electrical stimulation and, if present, immediately ablated with a Medtronic monopolar ablation pen.(Cardioblate ablation pen (standard) Medtronic, Minneapolis, MN) A decrease in R-R interval of>2.5 s was considered positive for the presence of ganglionated plexi.Right lung ventilation was reinitiated and the thoracoscope was retracted from the right hemithorax.A similar procedure was performed on the left side, where the pericardium was opened 2 cm below the phrenic nerve.The guiding wires through the sinuses were identified and guided outside the left thorax.Again, the ganglionated plexi were identified, and ablated immediately if present.
Subsequently, the Medtronic Gemini S clamp (Medtronic, Minneapolis, MN) was introduced, following the guiding wires.Ablation with bipolar radiofrequency was performed under continuous irrigation, five times with the clamp's curvature upwards and five times with the curvature downwards.The clamp was retracted from the body, as well as the thoracoscope and the same procedure was performed on the right side.After these ten ablations on each side, if patients had not already converted into SR spontaneously, an electrical cardioversion was performed.Then, bidirectional conduction block across the circular lesions was confirmed by fixed rate pacing from within the pulmonary veins demonstrating exit block when there was no atrial capture outside the lesions.If complete electrical isolation was not confirmed, bipolar ablation was repeated until bidirectional block was achieved, after which a chest tube was inserted on the right side, and the wounds were closed.The guiding wires were then retracted from the body and exclusion of the left atrial appendage was performed using the Powered Echelon® (Johnson and Johnson, USA) and haemostasis was confirmed.A chest tube was inserted on the left side and the wounds were closed.The double-lumen tube was replaced by a single lumen tube if patients were not ready to be extubated in the operating room.
Postoperatively, all patients were transferred to the intensive care unit for intensive monitoring because of the lack of a postoperative anaesthetic care unit or medium care in our hospital.Most patients were extubated directly after or in the first hours after surgery.The next day, the chest tubes were removed, and the patients were transferred to the Cardiothoracic Ward.After 3-5 days, the patients were discharged.

Pharmacological therapy
Patients received 30 prednisone mg daily during the first 5 perioperative days.For the first 3 postoperative months, patients were treated with 200 mg amiodarone daily if no contraindications were present.All (other) preoperative anti-arrhythmic drugs were ceased.After 3 months, amiodarone was ceased in all patients.With regard to oral anticoagulation (OAC), preoperative OAC were restarted the first postoperative day.After 3 months, according to guidelines, OAC were managed according to the thromboembolic risk profile.The final decision for discontinuation of OAC was left to the discretion of the referring cardiologist.

Follow-up
Cardiac rhythm was continuously monitored by using an implanted loop recorder (Reveal XT® or Linq®; Medtronic, Minneapolis, MN) implanted 4 weeks before surgery and programmed with nominal AF detection settings.These devices are equipped with an AF detection algorithm based on beat-to-beat variability, resulting in an accurate detection of AF episodes.[15][16][17][18] As demonstrated in Fig. 1, rhythm recordings were analyzed during 3 different time periods: 0-3 months, > 3-6 months and > 6-12 months after VATS PVI.
For every patient, the total number, median-and longest duration of AF episodes and the total AF burden were quantified during the 3 different time intervals after VATS PVI.The AF burden was defined as the ratio between the total duration of all AF episodes and total recording time (AF Burden (%) = total time in AF in minutes/ recording time in minutes * 100).Additionally, inter-episode intervals and the moment of AF occurrence (day: between 06:00 and 18:00 hrs.and night: between 18:00 and 06:00 hrs.) were evaluated.

Statistical analysis
IBM SPSS Statistics 24 software was used to perform statistical analysis.A propensity score matching analysis was performed, using logistic regression which was based on age and body mass index (BMI).The (nearest) neighboring propensity score with a match tolerance of 0.05 determined the random assignment of cases to controls.Both normally distributed data as well as skewed data were depicted as median [interquartile range].For related non-normally distributed variables the Wilcoxon Test was used, and dichotomous variables were tested using the McNemar test.A two-sided P-value of < 0.05 was considered statistically significant.

AF recurrences after surgical PVI
Fig. 1 summarizes the number of patients with AF episodes (upper left panel), the total number of AF episodes in the female and male group separately (upper right panel) and the number of AF episodes for each individual male and female patient (lower panels).The number of patients who experienced AF episodes during the first 3 months after VATS PVI, did not differ between males: N = 24 (60%) and females: N = 28 (70%); P > 0.05).However, AF episodes occurred significantly more often in females than in males (total number of AF episodes: females: 2324 versus males: 821) and the median number of AF episodes in the female group was higher (females: 3 [0-69]) versus males: 1.5 [0-25]; P = 0.01).
In the period of > 3-6 months after VATS PVI, the number of patients with AF episodes decreased (males: N = 17 (43%) versus females: N = 14 (35%; P > 0.05).The total number of AF episodes also decreased but remained higher in female patients (females: 1298 versus males: 451 episodes).Likewise, the average number of AF episodes was lower compared to the first 3 months but was still higher in the female group (females: 1 [0-4] versus males: 0 [0-4]; P = 0.034).
In the > 6-12 month post-VATS PVI period, there was a rise in the total number of patients with AF episodes (males: N = 19 (48%) and females: N = 20 (50%) patients; P > 0.05).Female and male patients had a comparable number of AF episodes (females: 1935 (0 [0-13] and males:1225 (1.5 [0-26]; P > 0.05).However, when comparing the number of AF episodes within the > 3-6 month period, there was a significant increase in the total amount of AF episodes among males     (>3-6 months: 451 episodes versus > 6-12 months: 1225 episodes; P = 0.026).Despite the increase in number of patients with AF recurrences, they were still lower compared to the first 3 months.

Total AF episode duration
The total time in AF for both the male and female groups during the 3 different follow-up intervals are demonstrated in the upper panel of Fig. 2. The middle and lower panels show the total duration of all AF episodes for each individual female and male patient.
During the 0-3 months after VATS PVI, the total AF duration was significantly longer in female patients compared to male patients (1552 hrs.versus 340 hrs.; P = 0.01).The total AF duration decreased during the > 3-6 months after VATS PVI, though it was still longer in female patients (442 hrs.versus 96 hrs.; P = 0.01).During the 6-12 months post-operative period, the total AF duration increased, and again time in AF was longer in female patients (females: 610 hrs.versus males: 364 hrs.; P = 0.01).The median AF episode duration between male and female patients was similar during all follow up period (0-3 months: males: 4

Longest AF episodes
Fig. 3 demonstrates longest AF episode for both male and female patients separately.During the first 3 months after VATS PVI, the longest AF episode in female and male patients ranged from respectively 6 to 1316 minutes and from 2 to 62,640 minutes and these longest AF episodes were significantly longer in females (females: 221 [44-608] minutes versus males: 108  minutes; P = 0.009).
During the > 6-12 months post VATS PVI period, longest AF episodes durations were comparable to the > 3-6 months period.In female patients, the longest episode duration ranged from 2 to 658 minutes compared to 2-780 min in male patients and male patients now had the longest AF episodes (males: 18 [5.5-228]minutes versus females: 16  minutes; P = 0.001).

Key findings
To our knowledge, this is the first study reporting on significant sex differences in characteristics of recurrent AF episodes after VATS PVI.Fig. 3.The left, middle and right panels show the longest AF episodes for both male and female patients, for every period separately.Female patients had longer AF episodes at the 0-3 months period in contrast to the > 6-12 month period in which males had longer AF episodes.(AF = atrial fibrillation).Fig. 4. The left panel shows the inter-episode intervals for both male and female patients during the first 3 months.During this period, female patients had significantly shorter intervals compared to male patients (P = 0.001).Also during the > 3-6 months period, female patients had shorter inter-episode intervals which is shown in the middle panel (P = 0.001).The right panel shows again a significantly shorter inter-episode interval in female patients compared to male patients during the > 6-12 months period (P = 0.04).
During the first 6 months after the procedure, AF burden in females is higher as a result of both a larger number of AF episodes of longer durations and consequently shorter inter-episode intervals.Yet, 6 months after VATS PVI, females and males have a comparable number of AF episodes, but the duration is still longer in F, resulting in a higher AF burden.In male patients, there was a significant increase in the total amount of AF episodes 6 months after VATS PVI compared to the > 3-6 months period.After 3 months, AF recurrences during the night were more frequently observed in female patients.

AF recurrences after PVI
In contrast to our study, previous studies reporting on outcomes of ablative therapy focused on the influence of sex on only early or late AF recurrence rates.[9,10,15] Comparable to our findings during the initial post-procedural 6 months higher recurrences rates were found in female patients in all studies.[9,10,16] Among 550 males and only 183 females with paroxysmal or (longstanding) persistent AF who underwent cryo PVI, AF recurrences occurred more frequently in females compared to males (41% versus 27%).[10] In the Fire and Ice study, in which outcomes of cryoballoon PVI was compared to radiofrequency PVI, female sex was also associated with AF recurrences.[9] After an average follow up period of 1.5 years, female patients (N = 293) had a 40% higher AF recurrence rate compared to male patients (N = 457).[9] In contrast, in 163 patients (77% males) who underwent cryo balloon PVI, male sex was correlated with early arrhythmia recurrence during the first 3 months follow-up.[15] However, arrhythmia screening in the above mentioned studies was performed using ECG, 24 hrs.holter monitoring or weekly transtelephonic monitoring instead of using continuous rhythm recordings obtained with implantable loop recorders.[9,10,15] It, therefore, remains unclear at what moment AF episodes recurred and neither the AF burden nor inter-episode intervals and circadian rhythms were described.[9,10,15] Even though the prevalence of AF is in general higher in male patients compared to female patients, females have more and severer symptoms.[4,17] This observation suggests that female patients may have a more advanced arrhythmogenic substrate which is in line with our findings that females in our population have more AF episodes with a longer duration and shorter inter-episode intervals.
A more extensive arrhythmogenic substrate can be explained by a previously reported larger referral delay for PVI for female compared to male patients, suggesting that females experience AF episodes for a longer time, hence resulting in more extensive electrical and structural remodeling of atrial tissue.[18,19] In The East -AFNET 4 trial [21], females more often had symptomatic AF and the CABANA trial [22] reported a lower quality of life in females with AF compared to males.In our study, the number of male and female patients experiencing recurrent AF episodes after VATS PVI were comparable, though the number, duration, inter-episode intervals and burden of AF episodes were higher in females.These findings may explain why females are more symptomatic and subsequently have a lower quality of life.

Type of rhythm control and timing line
The study of Kirchhof et al. [20] showed that patients who received early rhythm control (median of 36 days after the initial diagnosis) consisting of antiarrhythmic drugs or PVI) more often remained in SR compared to patients assigned to usual care.However, in our study population, the time between the initial diagnosis and VATS PVI was more than 36 days.In both males as well as in females, most AF recurrences occurend when a VATS PVI was performed more than one year after the initial AF diagnosis.This finding is in contrast to the observation of Kalman who reported that delayed ablation of one year did not affect freedom from recurrent arrhythmia.[23]

Sex differences in atrial electrophysiology
The observed higher AF burden as indicator of enhanced vulnerability for AF in females, may also be explained by differences in atrial electrophysiology.A shorter P-wave duration observed in (post-menopausal) females has been associated with an increased risk for recurrent AF episodes after PVI, which has been attributed to reduction in action potential duration.[24] However, females may also have shorter P-wave durations as a result of smaller left atrial dimensions compared to males of the same age.[25] Several studies have indicated that the AF related arrhythmogenic substrate may also differ between male and female patients.Wong et al. performed 3D electro-anatomical mapping in 116 patients (86% males) with episodes of paroxysmal or persistent AF. [11] Female patients had more intra-atrial slowing of conduction, a larger proportion of complex fractionated potentials and lower bipolar potential voltages compared to male patients (N = 74).[11] This study clearly shows that females have a more advanced substrate and therefore also serves as an explanation for the high number and burden of recurring AF episodes in the female patients observed in our study.Female patients also have more non-PV triggers compared to male patients [16,26] which may be related to the higher AF burden observed in our female population.
The higher recurrence rate after six months in male patients is consistent with reports on a higher recurrence rate of AF episodes in male patients after PVI caused by reconnection of the PV.[27]

Sex differences in structural remodeling
Sex differences in structural remodeling may also explain the higher AF burden and shorter inter-episode interval in female patients detected in our study.
Female patients (without AF) have a higher expression of proinflammatory genes (C-reactive proteins), resulting in higher circulating inflammatory cytokines compared to male patients.[8] As a consequence, females have a higher degree of myocardial necrosis and fibrosis replacement.[8] In patients with AF, female sex was also predictive for increased LA fibrosis.A higher degree of fibrosis enhances conduction heterogeneity and which could make female patients more prone to the development of AF. [8,28,29]

Sex differences in circadian rhythms
Most recurrent AF episodes in female patients occurred during nighttime whereas male patients experienced most AF episodes during daytime.Data on sex differences in circadian rhythms is sparse and limited to only the moment at which the majority of AF episodes occur.[30] In 44,000 patients (56% males), new onset, paroxysmal AF episodes (N = 9989) occurred mainly in the morning (nadir at 11am) or in the evening.[31] The prefential nighttime occurrence of AF episodes in our female patients could be explained by increased sympathetic tone.In a middle aged population, 24 h CRM demonstrated that females had a significantly higher heart rate during the night compared to males caused by a higher sympathetic tone.[32] A higher nightly sympathetic tone increases release of Ca 2+ ions inducing after depolarizations which may trigger AF episodes.[17]

Conclusions
This is the first study reporting on significant sex differences in characteristics of recurrent AF episodes after VATS PVI.After one year follow up, the AF recurrence rate was considerably high (48% among males and 50% among females).During the first post-procedural 6 months, AF burden in females was higher as a result of both a larger number of AF episodes of longer durations and consequently shorter inter-episode intervals.Yet, 6 months after VATS PVI, females and males had a comparable number of AF episodes, but the duration was still longer in females, resulting in a higher AF burden.In male patients, there was a significant increase in the total amount of AF episodes 6 months after VATS PVI compared to the > 3-6 month period.After 3 months, night time AF recurrences were more frequently observed in female patients.Hence, the observed differences in AF burden justify frequent rhythm monitoring after VATS PVI in both male and female patients and a more aggressive treatment of recurrent AF episodes in females.

Strenghts and limitations
During the first 3 months after VATS PVI, all patients used amiodarone.This may have affected (characteristics of) AF recurrences as the efficacy of amiodarone may differ between the two sexes as females have a smaller volume distribution of the hydrophilic amiodarone compared to males.[29,30] Due to the large variation and small number of patients, our study may be underpowered on the level of the individual patient.However, usage of implantable loop recorders enabling continuous monitoring of the heart rhythm prior to and after the VATS PVI is a strength of this study.Continuous rhythm monitoring provides detailed insights into AF characteristics that would not have been possible with single ECGs or standard Holter registrations.

Fig. 1 .
Fig. 1.The left upper panel shows the 3 analyzed intervals: directly after ablation up to three months after ablation (0-3 months), three to six months after ablation (>3-6 months), six to twelve months after ablation (>6-12 months).The left middle panel shows the percentage of patients in who AF occurred during the 0-3, >3-6 and > 6-12 month period, for male and female patients separately (solid color).The right middle panel shows the total number of AF episodes for the 0-3, >3-6 and > 6-12 month period, in both male and female patients separately.(AF = atrial fibrillation) As can be seen in the left lower panel, the individual number of AF episodes are shown for each individual male patient separately, for the 0-3, >3-6 and > 6-12 month period.The right lower panel also shows the individual number of AF episodes however for individual female patient separately, for the 0-3, >3-6 and > 6-12 month period.(AF = atrial fibrillation).

Fig. 2 .
Fig. 2. The upper panel shows the total duration of AF (in hours) is shown for the 0-3, >3-6 and > 6-12 month period, in both male and female patients separately.The middle panel shows the female individual AF episode duration whereas the lower panel of shows the male individual AF episode duration in minutes (outliers excluded).(AF = atrial fibrillation).

Table 1
Patient Characteristics.