Atrial fibrillation incidence after coronary artery bypass graft surgery and percutaneous coronary intervention: the prospective AFAF cohort study

Abstract

Objectives. Atrial fibrillation is a common arrhythmia in patients with ischemic heart disease. This study aimed to determine the cumulative incidence of new-onset atrial fibrillation after percutaneous coronary intervention or coronary artery bypass grafting surgery during 30 days of follow-up. Design. This was a prospective multi-center cohort study on atrial fibrillation incidence following percutaneous coronary intervention or coronary artery bypass grafting for stable angina or non-ST-elevation acute coronary syndrome. Heart rhythm was monitored for 30 days postoperatively by in-hospital telemetry and handheld thumb ECG recordings after discharge were performed. The primary endpoint was the cumulative incidence of atrial fibrillation 30 days after the index procedure. Results. In-hospital atrial fibrillation occurred in 60/123 (49%) coronary artery bypass graft and 0/123 percutaneous coronary intervention patients (p < .001). The cumulative incidence of atrial fibrillation after 30 days was 56% (69/123) of patients undergoing coronary artery bypass grafting and 2% (3/123) of patients undergoing percutaneous coronary intervention (p < .001). CABG was a strong predictor for atrial fibrillation compared to PCI (OR 80.2, 95% CI 18.1–354.9, p < .001). Thromboembolic stroke occurred in-hospital in one coronary artery bypass graft patient unrelated to atrial fibrillation, and at 30 days in two additional patients, one in each group. There was no mortality. Conclusion. New-onset atrial fibrillation during 30 days of follow-up was rare after percutaneous coronary intervention but common after coronary artery bypass grafting. A prolonged uninterrupted heart rhythm monitoring strategy identified additional patients in both groups with new-onset atrial fibrillation after discharge.

Keywords:

• New-onset atrial fibrillation
• coronary artery bypass graft surgery
• percutaneous coronary intervention
• postoperative atrial fibrillation incidence
• silent atrial fibrillation

Introduction

Atrial fibrillation (AF) is the most common arrhythmia seen in clinical practice and is associated with an increased risk of stroke, heart failure, and death [1,2]. Postoperative AF after coronary artery bypass graft surgery (CABG) is associated with a 6-fold increased risk of future AF, reduced long-term survival, and increased risk of cardiovascular and cerebral mortality compared with CABG patients with no postoperative AF [3]. In retrospective studies, patients with postoperative AF have a doubled long-term cardiovascular mortality compared with those who remain in sinus rhythm (SR), mainly explained by an increased risk of cardiac death and ischemic stroke [3–7].

Over the last 30 years, the proportion of CABG for cardiac revascularization has decreased in favor of percutaneous coronary intervention (PCI), currently the most common revascularization strategy for coronary artery disease [8]. In a post hoc analysis of the EXCEL trial, where CABG was compared to PCI for left main stem coronary disease, a post-procedural episode of AF was an independent predictor of stroke and death after three years [9], importantly, AF after PCI was rare. However, systematic detection of post-procedure AF was not performed, and patients with preoperative AF were not systematically detected [9]. Studies where the incidence of AF after CABG and PCI are compared, consist of non-systematic long-term postoperative AF detection [10], or post-hoc analysis [9,11]. The Atrial Fibrillation after CABG and PCI (AFAF) study is a prospective observational study investigating the incidence of new-onset AF in patients undergoing either isolated CABG or PCI for stable/unstable angina or non-ST-elevation myocardial infarction (NSTEMI) during a 2-year follow-up period. Here we present a post-hoc interim analysis from the first 30-days of follow-up where we aimed to determine the cumulative incidence of new-onset post-operative/intervention AF in each group independently using an uninterrupted systematic post-procedure AF detection strategy.

Methods

Patients

Patients scheduled for isolated CABG or PCI were eligible for participation. Inclusion criteria were:

• coronary intervention due to stable/unstable angina or NSTEMI

Exclusion criteria were:

• history of AF

• pacemaker or non-SR before coronary intervention

• bleeding disorder contraindicating oral anticoagulation (OAC) therapy

• cognitive impairment or communication problems leading to the inability to follow instructions and give informed consent

• life expectancy <1 year

• ongoing OAC therapy for any reason before the coronary intervention

• for PCI patients, CABG surgery <3 months before PCI

• for CABG patients, PCI <3 months before CABG surgery.

The study was conducted at three Swedish centers: Örebro University Hospital, Linköping University Hospital, and Danderyd Hospital, Stockholm. Patient enrollment started in May 2016 and ended in April 2019. The study followed the principles of the Declaration of Helsinki [12]; approved by the regional ethics committee in Uppsala, Sweden (2015/413) and registered at ClinicalTrials.gov, NCT04307225. Signed informed consent was obtained from all participants. An external medical committee was appointed to evaluate adverse events and changes to the protocol. Eligible patients were consecutively asked to participate during their hospital stay for the coronary intervention with a span of 1 day before and 5 days after the index procedure (Figure 1). The total follow-up period was 24 months. In this paper, we present results from the initial 30-days.

Figure 1. Study flow-chart. LOS: length of stay; mo: months.

Arrhythmia monitoring and follow-up

New-onset AF was defined as an episode of >30 s of irregular R-R intervals and the absence of P waves. All CABG patients were monitored by telemetry for at least 4 days postoperatively and once terminated, a Zenicor II (Zenicor Medical Systems AB, Stockholm, Sweden) handheld thumb ECG device, was three times daily. PCI patients were monitored by telemetry during their hospital stay and once discharged with the Zenicor II device three times daily for the remainder of the 30-days (Figure 1). The Zenicor II device provides one-lead 30-s ECG strip data to a web-based, password-protected, study-specific database. The device has been thoroughly validated as a reliable AF screening device [13–15]. All recordings were monitored by study coordinators and, in cases of uncertain arrhythmia, by study cardiologists (EF, JE, TK, and JA). Any arrhythmia requiring intervention was reported to the patient and their local health care professional for further care. Thirty days after the index procedure, follow-up was conducted by on-site visit or telephone.

Antithrombotic therapy

All patients received antithrombotic therapy according to ESC guidelines [16,17]. After CABG, antiplatelet therapy with aspirin started on the day of surgery. In addition, patients were routinely anticoagulated with low-molecular-weight heparin (LMWH) for at least 5 days postoperatively. In the presence of acute coronary syndrome or drug-eluting stents (DES), dual antiplatelet therapy (DAPT) (i.e. aspirin and clopidogrel/prasugrel/ticagrelor) was optional postoperatively. After PCI, patients were routinely administered DAPT. For both groups, if the patient developed AF postoperatively, OAC with warfarin or DOAC (direct oral anticoagulants) was considered after estimating stroke risk using the CHA2DS2-VASc score, and individual bleeding risk assessment; With a CHA2DS2-VASc score ≥2, OAC was initiated unless contraindicated.

Statistical methods

Sample size was calculated by non-probability for a total of 2 years of follow-up, and as follows: Based on retrospective data and the assumption of an increased incidence by detection of silent AF, we estimated a cumulative incidence of AF of 35% in the CABG group and 15% in the PCI group. To obtain an estimated power of 90% and a significance level of .05, the total required number of participants was 194. To compensate for missing data and patient dropout during follow-up of 15%, the final required number of participants was 250. The 30-day results presented herein thus constitute a post-hoc analysis.

Categorical variables were analyzed using Fisher’s exact test, and the Mann–Whitney U test was used to analyze non-normally distributed variables. Probability values <0.05 were regarded as statistically significant. Data are mean ± standard deviation, % (n), or median (interquartile range). Predictors of AF were analyzed by univariate and multivariate logistic regression, with results presented as odds ratios (OR) with 95% confidence intervals (CIs). Variables selected were those of most clinical interest. All statistical analyses were performed using SPSS version 22 (IBM Corp, Armonk, NY, USA).

Results

A total of 257 patients were included, 126 patients in the PCI group and 131 patients in the CABG group. At 30 days post-intervention, three patients were excluded from the PCI group and eight patients from the CABG group, providing 123 patients in each group as the final study population (Figure 2). Missing data were limited to preoperative patient characteristic variables in the CABG group (Smoking status, alcohol consumption, congestive heart failure, previous MI, chronic obstructive pulmonary disease, obstructive sleep apnea, hypertension, diabetes mellitus, previous stroke, peripheral vascular disease) with n = 4 missing at the most, and preoperative alcohol consumption in the PCI group (n = 4).

Figure 2. Study design.

Baseline characteristics

Patient characteristics are presented in Table 1. The gender distribution was 29% female in the PCI group and 9% female in the CABG group. There was no significant age difference. Diabetes mellitus (34 vs. 17%, p = .003) and heart failure (13 vs. 2%, p = .002) were more common in patients undergoing CABG than PCI, and CABG patients were more often affected by three-vessel disease (81 vs. 3%, p < .001). Chronic obstructive pulmonary disease (COPD) was more common in the PCI group and peripheral vascular disease was more common in the CABG group. Nearly all patients in the PCI group were discharged with DAPT (98%), whereas in the CABG group, 28/60 postoperative AF patients (46%) were discharged with OAC + aspirin (p < .001). At 30-days follow-up, 58 patients in the CABG group (84% of all with postoperative AF) were prescribed OAC.

Table 1. Preoperative patient characteristics.

	PCI group (n = 123)	CABG group (n = 123)	p-Value
Age	67 ± 9	67 ± 8	.903
Female sex (%)	29% (35/123)	9% (11/123)	<.001
BMI (kg/m^2)	28 ± 4	29 ± 5	.009
Smoking status
Current	11% (14/123)	9% (10/119)
Former	42% (51/123)	41% (49/119)
Non-smoker	47% (59/123)	50% (60/119)	.714
Alcohol consumption/week
>8 glasses	6% (7/119)	6% (7/118)
<8 glasses	86% (102/119)	86% (101/118)
None	8% (10/119)	8% (10/118)	1.000
Congestive heart failure	2% (3/123)	13% (16/120)	.002
Previous MI	40% (49/123)	43% (41/120)	.673
COPD	4% (5/123)	1% (1/120)	.213
OSA	3% (4/123)	5% (6/120)	.493
Hypertension	67% (82/123)	68% (82/120)	.782
Diabetes mellitus	17% (21/122)	34% (40/119)	.003
Previous stroke/TIA	5% (6/123)	6% (7/119)	.557
Peripheral vascular disease	2% (3/123)	6% (7/120)	.183
CHADS-VASc median (IQR)	3 (2)	3 (2)	.334
HAS-BLED median (IQR)	2 (1)	1 (1)	.160
Indication for PCI/CABG
Unstable	14% (17/123)	21% (26/123)
Stable	54% (66/123)	50% (62/123)
NSTEMI	32% (40/123)	29% (35/123)	.310
No. of diseased vessels
1	62% (76/123)	3% (4/123)
2	35% (43/123)	16% (19/123)
3	3% (4/123)	81% (100/123)	<.001
LA Area (cm^2)	21 ± 5	21 ± 5	.887
LVEF (%)	57 ± 8	55 ± 11	.135
CABG
No. of peripheral		3 ± 1
LIMA		90% (110/123)
PCI
No. of treated segments	2 ± 1
DES	97% (119/123)
BMS	1% (1/123)
Unknown	2% (3/123)
Medication
Aspirin	87% (107/123)	90% (111/123)	.422
Clopidogrel	32% (39/122)	2% (2/123)	<.001
Ticagrelor	37% (46/123)	4% (5/123)	<.001
Beta blockers	50% (61/123)	74% (91/123)	<.001
ACE-inh/ARB	55% (68/123)	66% (81/123)	.090
MRA	1% (1/122)	7% (8/123)	.036
Statins	73% (90/123)	85% (105/123)	.018
DOAC	0	0
Warfarin	0	0
LMWH	2% (2/123)	9% (11/123)	.010

BMI: body mass index; MI: myocardial infarction; COPD: chronic obstructive pulmonary disease; DOAC: direct oral anticoagulants; OSA: obstructive sleep apnea; TIA: transient ischemic attack; NSTEMI: non-ST-elevation myocardial infarction; LA: left atrium; LVEF: left ventricular ejection fraction; CABG: coronary artery bypass grafting; PCI: percutaneous coronary intervention; DES: drug-eluting stent; BMS: bare metal stent; ACE-inh/ARB: angiotensin converting enzyme inhibitor/angiotensin receptor blocker; MRA: mineralocorticoid receptor antagonist; LMWH: low-molecular-weight heparin.

Data are mean ± standard deviation, % (n), or median (interquartile range).

New-onset AF

No patient in the PCI group experienced AF during hospitalization, and in the CABG group, 60/123 of patients (49%) had at least one episode of AF (p < .001). Most of the AF episodes were paroxysmal, and 6/123 patients (5%) were discharged with ongoing AF in the CABG group vs. 0/123 in the PCI group (p = .013) (Table 2). At 30-days, an additional, three patients in the PCI group and nine patients in the CABG group were diagnosed with new-onset AF, resulting in a cumulative incidence of AF of 3/123 patients (2%) after PCI and 69/123 (56%) after CABG (p < .001) (Figure 3). CABG was a strong predictor for new-onset AF in both univariate (OR 51.1, 15.4–169.6, p < .001) and multivariate regression analysis (OR 80.2, 95% CI 18.1–354.9, p < .001) (Table 3).

Figure 3. Cumulative new-onset AF incidence. CABG: coronary artery bypass graft; PCI: percutaneous coronary intervention.

Table 2. In-hospital and discharge data.

	PCI group (n = 123)	CABG group (n = 123)	p-Value
AF >30 s	0% (0/123)	49% (60/123)	<.001
Asymptomatic		50% (30/60)
Paroxysmal		91% (55/60)
Persistent		2% (1/60)
Permanent		0
Not classified		7% (4/60)
Diagnosis
Telemetry		98% (59/60)
Zenicor II		2% (1/60)
Median day of onset		2
Median duration (days)		2 (1–4)
AF at discharge	0% (0/123)	5% (6/123)	.013
SR at discharge	100% (123/123)	95% (117/123)	.013
Hospital mortality	0	0
Thromboembolic events	0	1% (1/123)	1.0
Pericardial exudate	1% (1/123)	3% (4/123)	.370
Cerebral hemorrhage	0	0
Cardiac ischemic events	0	2% (2/123)	.498
Heart failure	0	0
Bleeding requiring intervention	2% (3/123)	8% (10/123)	.046
Medication
Aspirin	99% (122/123)	98% (120/123)	.313
Clopidogrel	49% (60/123)	3% (4/123)	<.001
Ticagrelor	50% (61/123)	15% (18/123)	<.001
Beta blockers	70% (86/123)	95% (117/123)	<.001
ACE inh/ARB	75% (92/123)	58% (71/123)	.005
MRA	2% (2/123)	6% (7/123)	.172
Statins	98% (121/123)	94% (116/123)	.172
OAC/DOAC	0	23% (28/123)	<.001
Dabigatran	0	0
Rivaroxaban	0	0
Apixaban	0	6% (7/123)	.014
Edoxaban	0	1% (1/123)	1.00
Warfarin	0	16% (20/123)	<.001
LMWH	1% (1/123)	2% (2/123)	1.00

AF: atrial fibrillation; SR: sinus rhythm; ECG: electrocardiography; ACE-inh/ARB: angiotensin converting enzyme inhibitor/angiotensine receptor blocker; MRA: mineralocorticoid receptor antagonist; OAC: oral anticoagulation; DOAC: direct oral anticoagulants; LMWH: low-molecular-weight heparin.

Data are presented as % (n), n, or median (interquartile range). Thrombomembolic events was defined as any venous or arterial symptomatic and clinically diagnosed embolic episode, such as deep vein thrombosis, pulmonary embolism, arterial embolic ischemia.

Table 3. Predictors of new-onset atrial fibrillation 30 days postoperatively.

	Univariate analysis OR (95% CI)	p-Value	Multivariate analysis OR (95% CI)^a	p-Value
CABG vs. PCI	51.1 (15.4–169.6)	<.001	80.2 (18.1–354.9)	<.001
Age	1.05 (1.01–1.08)	.009	1.08 (1.03–1.13)	.002
Congestive heart failure	4.86 (1.83–13.0)	.002	2.65 (0.78–9.0)	.118
Diabetes mellitus	1.49 (0.8–2.77)	.214	0.77 (0.35–1.72)	.526

CABG: coronary artery bypass grafting; CI: confidence interval; OR: odds ratio; PCI: percutaneous coronary intervention.

^aWith all variables in the table.

Complications

In-hospital rate of hemorrhagic incidents requiring transfusions, adjustment of medication, or surgical intervention was 2% in the PCI group and 8% in the CABG group (p = .046). At 30-days, bleeding events occurred in 1 CABG patient, who developed a hemorrhagic gastrointestinal ulcer during DAPT and NSAID treatment and was successfully managed non-surgically. No PCI patients experienced a bleeding event once discharged. Four patients in the CABG group developed clinically significant pericardial effusion, one of whom required surgical revision for late non-hemorrhagic cardiac tamponade.

Thromboembolic complications occurred in the hospital in one CABG patient (i.e. stroke) which was unrelated to AF, and none in the PCI group. At 30-day follow-up, one additional CABG patient who had developed AF and not yet initiated OAC, and one PCI patient who did not develop AF suffered thromboembolic stroke; There were no deaths (Table 4).

Table 4. 30-Day follow-up.

	PCI group (n = 123)	CABG group (n = 123)	p-Value
Follow-up by telephone	95% (117/123)	93% (114/123)	.313
AF >30 s	2% (3/123)	56% (69/123)	<.001
Asymptomatic	25% (31/123)	38% (47/123)	.001
Dyspnea	1% (1/123)	11% (13/123)	<.001
Palpitations	0	12% (15/123)	<.001
Fatigue	0	10% (12/123)	<.001
Paroxysmal	2% (3/123)	33% (40/123)	<.001
Persistent	0	1% (1/123)	<.001
Permanent	0	0
Unknown	0	23% (28/123)
Mortality	0	0
Thromboembolic events	1% (1/123)	1% (1/123)	1.00
Pericardial effusion	0	2% (3/123)	.247
Cerebral hemorrhage	0	0
Cardiac ischemic events	1% (1/123)	1% (1/123)	1.00
Heart failure	0	1% (1/123)	1.00
Bleeding requiring intervention	0	1% (1/123)	.122
Medication
Aspirin	97% (119/123)	89% (109/123)	.014
Clopidogrel	48% (59/123)	3% (4/123)	<.001
Ticagrelor	46% (57/123)	14% (17/123)	<.001
Beta blockers	68% (84/123)	96% (118/123)	<.001
ACE inh/ARB	76% (94/123)	70% (86/123)	.250
MRA	1% (1/123)	5% (6/123)	.120
Statins	98% (120/123)	94% (116/123)	.197
Dabigatran	1% (1/123)	2% (3/123)	.472
Rivaroxaban	0% (0/123)	1% (1/123)	.316
Apixaban	1% (1/123)	17% (21/123)	<.001
Edoxaban	0% (0/123)	1% (1/123)	1.00
Warfarin	1% (1/123)	27% (34/123)	<.001
LMWH	1% (1/123)	2% (2/123)	1.00
OAC/DOAC	2% (3/123)	48% (60/123)	<.001
DAPT	90% (111/123)	12% (15/123)
DAT	2% (2/123)	37% (45/123)
TAT	1% (1/123)	24% (3/123)

AF: atrial fibrillation; ACE-inh/ARB: angiotensin converting enzyme inhibitor/angiotensine receptor blocker; MRA: mineralocorticoid receptor antagonist; OAC: oral anticoagulation; DOAC: direct oral anticoagulants; DAPT: dual antiplatelet therapy; DAT: dual antithrombotic therapy; LMWH: low-molecular-weight heparin; SR: sinus rhythm; TAT: triple antithrombotic therapy.

Data are presented as % (n), n, or median (interquartile range).

Discussion

In this 30-day follow-up of the AFAF study assessing the incidence of new-onset AF in patients undergoing PCI or CABG, the main findings were that AF is rare after PCI (2%) and common after CABG (56%), where CABG was a strong predictor for new-onset AF compared to PCI. In addition, new-onset AF occurred not only in conjunction with the index procedure but also following discharge, giving a cumulative increase in the incidence of new-onset AF over time. To our knowledge, this is the first prospective cohort trial comparing the incidence of new-onset AF after PCI and CABG.

New-onset AF has been reported to occur at a rate of 18–31% after CABG [3,4,10] and 0.1–6% after PCI [9,18–21]. When interpreting these data, one must take into account several caveats. First, all studies reported the incidence of new-onset postoperative atrial fibrillation (POAF) based on in-hospital 12-lead ECG or telemetry, but information on AF is lacking beyond discharge. Second, the data are mainly retrospective and thus prone to detection bias. In the post-hoc analysis of the randomized EXCEL trial comparing PCI with CABG for left main stem coronary disease [9], new-onset AF occurred in 18% of the CABG population and 0.1% of the PCI population. Notably, AF after CABG was an independent predictor of an increased risk of stroke, MI, and death at 3-year follow-up. Although similar differences between the groups, the incidence was significantly lower than in our study, which could be explained by several factors. The AF incidence was analyzed post hoc and therefore prone to confounding bias. CABG patients with no prior history of AF were to some extent administered preoperative amiodarone and/or beta blocker as AF prophylaxis. Moreover, follow-up data on arrhythmia were not systematically collected.

Other trials comparing PCI and CABG have found a consistent difference in the postoperative stroke rate. In the SYNTAX trial, comparing CABG with PCI for three-vessel disease and/or left main stem coronary artery disease, the stroke incidence at 1-year follow-up was 2.2% in the CABG group vs. 0.6% in the PCI group [22]. In the Freedom trial, comparing CABG with PCI in diabetic patients with multi-vessel coronary artery disease, the stroke incidence at 5 years was 5.2% in the CABG group and 2.4% in the PCI group [23]. Notably, differences in the incidence of postprocedural AF or in treatment with OAC were not recorded in these trials. About one-third of all AF is asymptomatic and often paroxysmal [24], and the risk of stroke seems to be the same in silent and symptomatic AF [25]. Overall, this implies that monitoring for AF beyond discharge is important for CABG patients.

Screening for silent AF using handheld ambulatory ECG recorders has proven to be an effective, noninvasive way of capturing episodes of AF and superior to 24-h Holter [13,26,27]. Alternatively, continuous rhythm surveillance with implantable loop recorders can be performed, but this requires an additional invasive procedure. In the CABG population, this can be done safely as part of their surgery and provides rhythm and AF burden data [28].

Data on the impact of new-onset AF after PCI are limited to retrospective data or post-hoc analyses; furthermore, they mainly investigate the in-hospital incidence of AF in an ST-elevation myocardial infarction (STEMI) population. New-onset AF incidence was 4.5–7%, and AF was a predictor of a worse cardiovascular outcome [11,18–21,29].

The present results may shed some light on the etiology of postoperative AF. Both groups suffer from ischaemic heart disease requiring revascularization but with profoundly different results in the incidence of subsequent AF. Also, most participants were treated for stable angina and thus had no recent ischemic episode. Ischemia per se therefore seems less likely to be the inductive mechanism. However, the inflammatory responses after PCI and CABG differ [30,31], and inflammation is known to induce AF [32–34]. Although attempts to reduce AF after cardiac surgery have been made by administering anti-inflammatory drugs, the effects have been moderate [35]. The present study may indicate that further studies of inflammatory response after CABG and PCI are warranted.

Although postoperative AF has been shown to have a strong negative association with long-term cardiovascular survival after CABG, current guidelines recommend postoperative OAC treatment with a class IIa/b LOE B recommendation [17,36,37]. As ischemic stroke constitutes a significant proportion of such events, underutilization of OAC treatment could theoretically explain the higher frequency of thromboembolic stroke after CABG. Retrospective data show conflicting results regarding OAC treatment in patients with postoperative AF [38], and a current small scale RCT is expected to provide additional data for future recommendations [39]. In all OAC treatment reduces the frequency of thromboembolic events in post-op AF CABG patients and seems safe in terms of bleeding complications. This mirrors our results where in-hospital bleeding events were rare in both groups and at 30-days almost non-existent. However, recommending or deterring routine use of OAC in these patients requires results from long term, large scale RCTs.

The major limitation of the present study is the non-randomized design leading to baseline differences and the absence of STEMI patients, The gender distribution is somewhat skewed in the CABG group, with 9% female participation, vs. 19% in the overall Swedish CABG population [40]. The baseline differences reflect the guidelines for choice of revascularization method, with three-vessel disease and diabetes mellitus being more common among CABG patients. Thus any inferred causality between the groups cannot be made. Further, as the time of telemetry surveillance is generally shorter for the PCI group, concerns on whether early new-onset AF detection in the PCI group would be lacking, and non-comparable, could arise. However, previous data show that peak incidence and duration of new-onset AF is on postoperative day 2 and lasting 48 h [3], thus it is unlikely that any episode of AF would not be captured by the thrice daily intermittent monitoring after discharge. Rather, a major strength of this study is the uninterrupted intermittent rhythm surveillance, yielding a cumulative increase in postoperative AF over time in both groups, which could explain the relatively higher prevalence of new-onset AF after CABG than previously reported.

Conclusion

Postoperative AF at 30-day follow-up occurs at a higher rate than previously reported after CABG and at a low rate after PCI. A prolonged uninterrupted heart rhythm monitoring strategy identified additional patients in both groups with new-onset AF after discharge from in-hospital care. Long-term follow-up with matched populations is needed to determine the prognostic impact of new-onset AF on cardiovascular morbidity and mortality as well as the predictive markers in each population.

Author contributions

All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.

Acknowledgements

Thanks to the study nurses at all study sites for their valuable and arduous work in including patients, conducting follow-up, study logistics, and thumb-ECG interpretation:

Helen Hildingsson, Anna Mattson, Karin Johansson, and Margit Quist, Örebro University Hospital.

Liselotte Persson, Danderyd Hospital.

Maria Eriksson, Linköping University Hospital.

Thanks to Mikolaj Skibniewski M.D, and Emmanouil Charitakis M.D, Linköping University Hospital for their valuable assistance in patient follow-up and data management.

Thanks to Aron Sztaniszlav M.D Örebro University Hospital for aiding with the regression analysis.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data will be made available upon request, please contact the corresponding author.

Additional information

Funding

This work was supported by the Research Committee of Örebro University Hospital, under grant numbers: OLL-597581, OLL-769421, OLL-685871, OLL-575671, OLL-785021, OLL-838331, and OLL-889661.