Eicosapentaenoic Acid and the Outcomes in Older Patients Undergoing Atrial Fibrillation Ablation

Abstract

Background

A lower serum eicosapentaenoic acid (EPA) to arachidonic acid (AA) ratio (EPA/AA) level correlates with cardiovascular events. Nevertheless, elevated serum EPA levels increase the risk of new‐onset atrial fibrillation (AF) in older patients. The relationship between the EPA/AA and outcomes post‐AF ablation remains unclear. This study investigated the impact of the EPA/AA on AF recurrence and cardiovascular events after AF ablation in older patients.

Methods and Results

This retrospective cohort study examined consecutive patients with AF aged ≥65 years who underwent a first‐time AF ablation. We compared the 3‐year AF recurrence and 5‐year major adverse cardiovascular event (MACE) rates between patients divided into high and low EPA/AA levels defined as above and below the median EPA/AA value before ablation. MACE was defined as heart failure hospitalizations, strokes, coronary artery disease, major bleeding, and cardiovascular death. Among the 673 included patients, the median EPA/AA value was 0.35. Compared with the low EPA/AA group, the high EPA/AA group had a significantly higher cumulative incidence of AF recurrence (39.3% versus 27.6%; log‐rank P=0.004) and lower cumulative incidence of MACE (13.8% versus 25.5%, log‐rank P=0.021). A high EPA/AA level was determined as an independent predictor of AF recurrence (hazard ratio [HR], 1.75 95% CI, 1.24–2.49; P=0.002) and MACE (HR, 0.60 [95% CI, 0.36–0.99]; P=0.046).

Conclusions

The EPA/AA was associated with AF recurrence and MACE after ablation in patients with AF aged ≥65 years.

Nonstandard Abbreviations and Acronyms

AA

arachidonic acid

AF

atrial fibrillation

DHA

docosahexaenoic acid

EPA

eicosapentaenoic acid

EPA/AA

eicosapentaenoic acid to arachidonic acid ratio

n‐3 PUFA

omega‐3 polyunsaturated fatty acid

Clinical Perspective.

What Is New?

• Among older patients with atrial fibrillation (AF) aged ≥65 years, a high serum eicosapentaenoic acid (EPA) to arachidonic acid (AA) ratio (EPA/AA) level before AF ablation was associated with a higher incidence of a 3‐year AF recurrence, including a persistent form of AF recurrence.

• Conversely, it was associated with a lower incidence of 5‐year major adverse cardiovascular events after ablation.

• These results were consistent with those using other fatty acid indices, such as EPA alone, EPA + docosahexaenoic acid, and EPA/docosahexaenoic acid, and the hazard ratio for AF recurrence using the EPA/AA was the largest among those using 4 fatty acid indices.

What Are the Clinical Implications?

• Our findings highlight a group of patients with a higher serum EPA/AA level before AF ablation, suggesting a more favorable prognosis, yet a potential risk of AF recurrence; the assessment of the serum EPA/AA level before ablation would be beneficial to guide an appropriate postablation management.

• Evaluation of the serum EPA/AA level may be recommended in patients with AF who are aged ≥65 years and scheduled for AF ablation, especially those aged ≥75 years, with heart failure, or with a CHA₂DS₂‐VASc score ≥ 2 (male) or 3 (female).

Atrial fibrillation (AF) is known to become more common with age ¹ and is associated with various cardiovascular events, including strokes, heart failure, and cardiovascular death. ² , ³ , ⁴ Catheter ablation has become an established standard therapy ⁵ for AF, aiming to enhance both the quality of life and prognosis. However, older patients often experience cardiovascular and major bleeding events even after AF ablation, particularly in light of the need for anticoagulation therapy among those aged ≥65 years, which comes with its own bleeding risks. ⁶ , ⁷ Therefore, there is a need for prognostic indicators following AF ablation in this older patient population.

Omega‐3 polyunsaturated fatty acids (n‐3 PUFAs) have multifaceted effects and have been linked to an increased risk of AF and a decreased risk of cardiovascular events. ⁸ Eicosapentaenoic acid (EPA), a type of n‐3 PUFA, exhibits various beneficial actions, including anti‐inflammatory, antithrombotic, lipid‐lowering, and plaque‐stabilizing effects. ⁹ A lower serum EPA to arachidonic acid (AA) ratio (EPA/AA) level has been identified as a predictor of cardiovascular events in high‐risk patients, such as those with heart failure ¹⁰ or coronary artery disease. ¹¹ However, prior reports indicated a nonpreventive effect of n‐3 PUFA intake on AF recurrence after electrical cardioversion. ¹² Furthermore, recent studies have raised concerns regarding the association between elevated serum EPA levels resulting from n‐3 PUFA intake and the increased risk of new‐onset AF. ⁸ , ¹³ , ¹⁴ , ¹⁵ Additionally, patients with AF who take n‐3 PUFAs have exhibited a higher incidence of AF‐related hospitalizations than those not taking n‐3 PUFAs. ¹⁶

In older patients, we deemed it essential to evaluate the dual impact of the serum EPA/AA levels concerning both AF recurrence and cardiovascular events after ablation. To achieve this, we investigated the patients aged ≥65 years who underwent AF ablation, by comparing the rates of AF recurrence and cardiovascular events, including heart failure hospitalizations, strokes, coronary artery disease, major bleeding, and cardiovascular death, between the 2 groups categorized by the median EPA/AA value before ablation.

Methods

Study Design and Patient Population

The data that support the findings of this study are available from the corresponding author upon reasonable request. This retrospective cohort study included patients aged ≥65 years with AF who underwent their first AF ablation procedure at Kagawa Prefectural Central Hospital between April 2014 and July 2020. Exclusion criteria encompassed patients (1) without preablation serum measurements of the EPA or AA values and (2) lost to follow‐up after ablation. Notably, patients taking n‐3 PUFAs before ablation were not excluded, as this study primarily focused on the serum EPA/AA level rather than the intake of EPA itself. The sample size was based on the registered patients during the study period, which was determined in advance considering the feasibility of the study.

The Institutional Review Board at Kagawa Prefectural Central Hospital granted approval for this study, and patient informed consent was obtained using opt‐out methods, with the study details made available on the hospital website.

Measurement of the Serum EPA/AA Value

Fasting blood samples taken on the day of admission before ablation were used to measure the serum values of fatty acids, for example, EPA, docosahexaenoic acid (DHA), AA, and dihomo‐γ‐linolenic acid. This analysis was conducted by the SRL Co., Ltd. (Tokyo, Japan), using a gas chromatography–flame ionization detector system (GC2010; Shimadzu, Kyoto, Japan). The EPA/AA value was computed, and the patients were classified into 2 groups on the basis of whether their EPA/AA value fell below or above the median value. It is worth noting that no new prescriptions of n‐3 PUFAs were issued following ablation, and such decisions were not influenced by the EPA/AA values observed before ablation.

Data Collection and Variables

We investigated the variables and outcomes from the medical records of Kagawa Prefectural Central Hospital. Before the ablation procedure, we collected comprehensive patient data, including physical examinations, medical histories, medication status, laboratory results, and electrocardiographic and echocardiographic findings. Echocardiography was performed to measure the dimensions of the left atrium (LA) and assess the left ventricular ejection fraction (LVEF), with measurements obtained through the disk summation method and presented as an average over 5 cardiac cycles when AF persisted.

All patients received instructions for self‐pulse checks to identify asymptomatic AF episodes before discharge. Following ablation, 12‐lead ECGs were conducted every 3 months. Furthermore, a 1‐week ECG monitoring was performed using an autotriggered external loop recorder or 24‐hour Holter ECG 6 months after ablation in the patients who had no AF recurrence until 6 months after ablation. In cases in which patients reported symptoms or irregular pulses, a 24‐hour Holter ECG or 1‐week ECG monitoring was arranged. Furthermore, for patients who had completed their clinical visits to our hospital, we collected information on their outcomes by communicating with their primary care physicians through mail and telephone.

AF Ablation

Anticoagulation therapy was initiated at least 1 month before ablation. Antiarrhythmic drugs were discontinued the day before ablation. Following heparin administration to maintain anticoagulation, we performed a pulmonary vein isolation using either radiofrequency ablation or cryoballoon ablation.

For radiofrequency ablation, we used a contact force sensing catheter (Thermocool; Biosense Webster, CA) and a 3‐dimensional mapping system (CARTO; Biosense Webster). Radiofrequency energy was applied with power settings ranging from 25 W to 40 W. The decision to perform a cavotricuspid isthmus ablation was made by the operator, and a superior vena cava isolation was performed when superior vena cava potentials were detected. Additionally, a posterior wall isolation was performed in cases where the duration of AF persistence exceeded 36 months or when AF was not terminated by electrical cardioversion following the pulmonary vein isolation.

We employed a second‐generation 28‐mm cryoballoon (Arctic Front Advance Cardiac Cryoablation Catheter; Medtronic, Minneapolis, MN) exclusively for patients with paroxysmal AF during cryoballoon ablation. In cases in which electrical isolation of the pulmonary veins was not achieved or focal triggers other than the pulmonary veins were identified, we opted for radiofrequency ablation.

The ablation procedure aimed for a bidirectional block evaluated 20 minutes after the initial pulmonary vein isolation with an isoproterenol administration. ¹⁷ Antiarrhythmic drugs were reintroduced on the day after ablation and continued for at least 3 months after ablation. All patients underwent ECG monitoring for at least 3 days after ablation. Anticoagulation therapy was discontinued 6 months after ablation for patients with a CHA₂DS₂‐VASc score <2 (men) or 3 (women), while patients with a CHA₂DS₂‐VASc score ≥2 (men) or 3 (women) continued anticoagulation therapy unless they experienced adverse events related to anticoagulation.

Outcomes

The primary outcome of this study was defined as the initial recurrence of AF within 3 years following the 3‐month blanking period after ablation. AF recurrence encompassed the presence of atrial tachyarrhythmias, including AF and atrial tachycardia, detected on a 12‐lead ECG or documented as lasting at least 30 seconds on a 24‐hour Holter ECG or a 1‐week external loop recorder. A persistent AF recurrence was defined as an AF recurrence lasting at least 7 days or requiring pharmacological or electrical cardioversion and were more clinically problematic and accurately assessable recurrences. ¹⁸ The secondary outcome was defined as initial major adverse cardiovascular events (MACEs), including heart failure hospitalizations, strokes, coronary artery disease, major bleeding, and cardiovascular death. ² , ³ , ⁴ , ⁷ , ¹⁹ The 3‐year follow‐up period was chosen for the primary outcome due to the presumption that the impact of the ablation would not extend beyond 3 years. ²⁰ The 5‐year follow‐up period for the secondary outcome was established due to the limited occurrence of events. In this study, a “heart failure hospitalization” was identified as worsening heart failure requiring treatment in the hospital. “Strokes” encompassed cerebrovascular infarctions, transient ischemic attacks, and intracranial hemorrhages. “Coronary artery disease” included nonfatal myocardial infarctions, unstable angina, and coronary revascularization. A myocardial infarction was defined as a type 1 myocardial infarction, which involved acute myocardial injury with an elevated troponin level caused by a coronary thrombosis. ²¹ Unstable angina was defined as myocardial ischemia at rest or minimal exertion in the absence of an acute cardiomyocyte injury. Coronary revascularization included percutaneous coronary intervention and coronary artery bypass grafting. Coronary revascularization for the lesion detected at the time of the AF ablation was excluded. “Major bleeding” was defined as a reduction in the hemoglobin level of at least 20 g/L, the need for a transfusion of at least 2 units of blood, or the occurrence of intracranial bleeding. “Cardiovascular death” was characterized as death resulting from heart failure, a stroke, acute coronary syndrome, or sudden cardiac death.

Statistical Analysis

The patients were categorized into 2 groups, namely, the high EPA/AA group and the low EPA/AA group, using the median EPA/AA value measured before ablation. Categorical variables were presented as numbers and percentages, with between‐group comparisons performed using the χ² test. Continuous variables were presented as either the mean±SD or median and interquartile range (IQR), depending on their distribution, and compared using either a Student's t test or Wilcoxon rank‐sum test.

We conducted a comparison of the incidence rates for both primary and secondary outcomes between the high and low EPA/AA groups using a Kaplan–Meier analysis and the log‐rank test. Proportional hazard assumptions for the risk‐adjusting variables were assessed on the plots of the log (time) versus log (−log [survival]) stratified by the variable and were verified to be acceptable. To evaluate the impact of the EPA/AA level on AF recurrence, we built a multivariable Cox proportional hazards model. This model was adjusted for the clinically relevant risk factors, including the sex, persistent AF, hypertension, diabetes, chronic kidney disease, a history of heart failure, a reduced LVEF (<50%), and LA dilatation (≥40 mm). ¹⁷ , ²² , ²³ , ²⁴ , ²⁵ We conducted subgroup analyses to assess the impact of the EPA/AA level on AF recurrence on the basis of various factors, including the sex, age, body mass index, hypertension, diabetes, heart failure, chronic kidney disease, CHA₂DS₂‐VASc score, AF type, LVEF, and LA diameter. We compared the incidence of a persistent AF recurrence with the same methods as for the primary outcome.

To evaluate the impact of the EPA/AA level on the MACE, we constructed a multivariable Cox proportional hazard model, adjusting for the confounding factors such as the age (≥75 years), hypertension, diabetes, history of heart failure, chronic kidney disease, and persistent AF. ⁴ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ Furthermore, we conducted a detailed examination of the MACE, specifically heart failure hospitalizations, coronary artery disease, major bleeding, and cardiovascular death using univariable Cox proportional hazard models.

For sensitivity analyses, we evaluated both the primary and secondary outcomes by comparing the high and low EPA/AA groups, excluding patients taking n‐3 PUFAs. We also evaluated both the primary and secondary outcomes using the EPA alone, EPA + DHA, and EPA/DHA to validate the EPA/AA, an optimal fatty acid index, dividing the patients into 2 groups on the basis of each median value. Furthermore, to explore the impact of the fatty acid indices on the autonomic nervous system, we compared the heart rate at 6 months after the ablation between the high and low groups of the EPA/AA, EPA alone, EPA + DHA, and EPA/DHA, respectively. In addition, to evaluate the impact of AF recurrence on heart failure worsening according to the EPA/AA level, we compared the incidence of heart failure hospitalizations occurring from 3 years to 5 years after ablation between the patients with and without AF recurrence within 3 years dividing the high and low EPA/AA groups using the χ² test due to the small number of events.

Missing data were not imputed and were eliminated from the corresponding analyses. All statistical analyses were performed using statistical software (SPSS version 28, IBM, NY). We considered a 2‐tailed P < 0.05 as statistically significant.

Results

Baseline Characteristics

Among 1022 consecutive patients aged ≥65 years who underwent a first‐time catheter ablation of AF, our analysis included 673 patients after excluding those who met the specified exclusion criteria. All exclusions were caused by not measuring the EPA/AA values before the ablation and were treated as missing data. The major reasons for not evaluating the EPA/AA values were patients declining it or being in a nonfasting state (Figure 1). The included patient characteristics were as follows: the median EPA/AA value was 0.35, the median age was 72 (IQR, 68–77) years, 40% were women, 53% had persistent AF, 66% had hypertension, 42% had heart failure, 39% had chronic kidney disease, the median LA diameter was 39 (IQR, 33–42) mm, and the median LVEF was 63% (IQR, 57%–68%). Additionally, 31 patients (4.6%) were taking n‐3 PUFAs internally (Table S1). The 349 excluded patient characteristics are also shown in Table S1, and they had a significantly higher age, rate of chronic kidney disease, and LVEF, and had a significantly lower body mass index, prevalence rate of persistent AF, rate of a history of heart failure, brain natriuretic peptide levels, and LA diameter than the included patients.

Figure 1. Study flowchart.

AF indicates atrial fibrillation; and EPA/AA, eicosapentaenoic acid to arachidonic acid ratio.

All patients were categorized into 2 groups on the basis of the median EPA/AA value before the ablation. Specifically, 345 patients had an EPA/AA value of ≥0.35 (high EPA/AA group), while 328 patients had an EPA/AA value of <0.35 (low EPA/AA group). The distribution of the EPA/AA values is shown in Figure S1.

In the high EPA/AA group, there were significantly lower prevalence rates of persistent AF, a history of heart failure, chronic kidney disease, brain natriuretic peptide levels, and statin use than in the low EPA/AA group (Table 1). However, there were no significant differences observed between the 2 groups in terms of the age; sex; body mass index; hypertension; diabetes; C‐reactive protein level; LA diameter; LVEF; and use of angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, and β blockers.

Table 1.

Baseline Patient Characteristics

	High EPA/AA group (n=345)	Low EPA/AA group (n=328)	P value
Age, y, median (IQR)	73 (69–77)	72 (68–76)	0.23
Female sex, n (%)	128 (37)	138 (42)	0.21
BMI, kg/m2, mean±SD	24±3.2	24±3.5	0.54
Persistent AF, n (%)	162 (47)	192 (59)	0.003
Hypertension, n (%)	228 (67)	213 (65)	0.69
Diabetes, n (%)	68 (20)	69 (21)	0.70
Heart failure, n (%)	128 (37)	156 (48)	0.006
Chronic kidney disease, n (%)	119 (35)	145 (44)	0.011
CHA2DS2VASc score, median (IQR)	3 (2, 4)	3 (2, 4)	0.99
Blood exam
hs‐CRP, mg/dL, median (IQR)	0.08 (0.04–0.17)	0.08 (0.04–0.19)	0.35
BNP, pg/mL, median (IQR)	113 (53, 206)	129 (60, 239)	0.041
eGFR, mL/min per 1.73 m2, median (IQR)	66 (56–76)	62 (51–73)	<0.001
EPA, μg/mL, median (IQR)	95 (74–129)	41 (32–52)	<0.001
DHA, μg/mL, median (IQR)	162 (131–194)	113 (93–138)	<0.001
AA, μg/mL, median (IQR)	173 (147–205)	187 (156–223)	<0.001
DGLA, μg/mL, median (IQR)	32 (27–42)	40 (33–49)	<0.001
EPA/AA, median (IQR)	0.52 (0.43–0.74)	0.23 (0.17–0.29)	<0.001
Medicine
n‐3 PUFA, n (%)	29 (8)	2 (0.6)	<0.001
ACE‐I/ARB, n (%)	152 (44)	163 (50)	0.16
β Blocker, n (%)	182 (53)	169 (52)	0.76
Statin, n (%)	79 (23)	103 (31)	0.015
Echocardiogram
LA diameter, mm, median (IQR)	39 (33–43)	39 (34–42)	0.84
LVEF, %, median (IQR)	64 (59–68)	63 (56–68)	0.35

The data are expressed as the mean±SD or median (IQR) based on the distribution and number of patients (%).

AA indicates arachidonic acid; ACE‐I, angiotensin‐converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin II receptor blocker; BMI, body mass index; BNP, brain natriuretic peptide; DHA, docosahexaenoic acid; DGLA, dihomo‐γ‐linolenic acid; eGFR, estimated glomerular filtration rate; EPA, eicosapentaenoic acid; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; hs‐CRP, high‐sensitivity C‐reactive protein; IQR, interquartile range; LA, left atrial; LVEF, left ventricular ejection fraction; and n‐3 PUFA, omega‐3 polyunsaturated fatty acid.

Outcomes

Of the 673 patients included in the study, 140 experienced AF recurrences during the 3‐year follow‐up period (median, 30 [IQR, 13–36] months). In terms of the primary outcome, the cumulative incidence of an AF recurrence was significantly higher in the high EPA/AA group than in the low EPA/AA group (39.3% versus 27.6%; log‐rank test, P=0.004; Figure 2). Furthermore, an analysis using a multivariable Cox proportional hazard model revealed that a high EPA/AA level had a statistically significant association with AF recurrence (adjusted HR, 1.75 [95% CI, 1.24–2.49]; P=0.002; Table 2). A subgroup analysis of AF recurrence, considering the EPA/AA level, demonstrated a higher frequency of AF recurrence with a high EPA/AA level among those patients aged ≥75 years, with a history of heart failure and a CHA₂DS₂VASc score ≥ 2 (men) or 3 (women) (Figure 3). A persistent AF recurrence occurred in 70 patients during the 3‐year follow‐up period. The cumulative incidence of a persistent AF recurrence was significantly higher in the high EPA/AA group than in the low EPA/AA group (17.9% versus 12.8%; log‐rank, P=0.038; Figure 4). An analysis using a multivariable Cox proportional hazard model revealed that a high EPA/AA level had a statistically significant association with a persistent AF recurrence (adjusted HR, 1.94 [95% CI, 1.18–3.19]; P=0.009), consistent with AF recurrence as the primary outcome.

Figure 2. 3‐year incidence of AF recurrence after ablation.

The 3‐year incidence of AF recurrence was significantly higher in the high EPA/AA group than in the low EPA/AA group. AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; and HR, hazard ratio.

Table 2.

Cox Proportional Hazard Model of AF Recurrence After Ablation

	Crude HR	95% CI	P value	Adjusted HR	95% CI	P value
EPA/AA ≥0.35	1.65	1.17–2.33	0.004	1.81	1.27–2.57	<0.001
Female sex	1.46	1.05–2.04	0.025	1.49	1.06–2.09	0.022
Diabetes	1.38	0.94–2.02	0.11	1.41	0.95–2.10	0.09
Chronic kidney disease	1.30	0.93–1.81	0.12	1.36	0.96–1.92	0.09
Heart failure	1.33	0.95–1.85	0.095	1.24	0.85–1.81	0.26
Persistent AF	1.26	0.90–1.77	0.18	1.19	0.82–1.73	0.37
LA diameter ≥40 mm	1.07	0.77–1.49	0.70	0.90	0.62–1.29	0.55
Hypertension	0.98	0.69–1.40	0.92	0.87	0.60–1.25	0.45
LVEF <50%	0.99	0.52–1.87	0.96	0.86	0.44–1.69	0.67

AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; HR, hazard ratio; LA, left atrial; and LVEF, left ventricular ejection fraction.

Figure 3. Subgroup analysis of AF recurrence based on the EPA/AA.

AF recurrences were more frequent in those with a high EPA/AA level in the patients aged ≥75 y, with a history of heart failure, and with a CHA₂DS₂‐VASc score ≥2 (men) or 3 (women). AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; HR, hazard ratio; LA, left atrial; and LVEF, left ventricular ejection fraction.

Figure 4. 3‐year incidence of persistent AF recurrence after ablation.

The 3‐year incidence of persistent AF recurrence after ablation was significantly higher in the high EPA/AA group than in the low EPA/AA group. AA arachidonic acid; AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; and HR, hazard ratio.

Regarding the secondary outcomes, there were 65 reported cases of MACE during the 5‐year follow‐up period (median, 30 [IQR, 13–52] months). The cumulative incidence of the MACE was significantly lower in the high EPA/AA group than in the low EPA/AA group (13.8% versus 25.5%; log‐rank test, P=0.021; Figure 5). According to the multivariable Cox proportional hazard model, a high EPA/AA level had a statistically significant association with a lower incidence of MACE (adjusted HR, 0.60 [95% CI, 0.36–0.99]; P=0.046; Table 3). Within the MACE, it is worth noting that heart failure hospitalizations were significantly less frequent in the high EPA/AA group (crude HR, 0.41 [95% CI, 0.17–0.96]; P=0.041; Table 4).

Figure 5. 5‐year incidence of major adverse cardiovascular events after ablation.

The 5‐year incidence of major adverse cardiovascular events was significantly lower in the high EPA/AA group than in the low EPA/AA group. AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; and HR, hazard ratio.

Table 3.

Cox Proportional Hazard Model of Major Adverse Cardiovascular Events After Ablation

	Crude HR	95% CI	P value	Adjusted HR	95% CI	P value
Diabetes	2.59	1.58–4.25	<0.001	2.13	1.28–3.55	0.004
Heart failure	2.49	1.51–4.11	<0.001	1.97	1.14–3.39	0.015
Hypertension	1.88	1.03–3.46	0.041	1.51	0.81–2.80	0.20
Age ≥75 y	1.55	0.95–2.51	0.08	1.33	0.81–2.20	0.26
Chronic kidney disease	1.85	1.13–3.01	0.014	1.32	0.80–2.20	0.28
Persistent AF	1.50	0.91–2.46	0.11	1.05	0.62–1.78	0.87
EPA/AA ≥0.35	0.56	0.34–0.92	0.023	0.60	0.36–0.99	0.046

AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; and HR, hazard ratio.

Table 4.

Details of Major Adverse Cardiovascular Events After Ablation

	High EPA/AA group (n=345)	Low EPA/AA group (n=328)	Crude HR	95% CI	P value
Major adverse cardiovascular events, n (%)	27 (7.8)	38 (11.6)	0.56	0.34–0.92	0.023
Heart failure hospitalization, n (%)	8 (2.3)	15 (4.6)	0.41	0.17–0.96	0.041
Stroke, n (%)	11 (3.2)	15 (4.6)	0.58	0.27–1.27	0.17
Ischemic stroke, n (%)	7 (2.0)	9 (2.7)	0.63	0.23–1.69	0.36
Hemorrhagic stroke, n (%)	4 (1.2)	6 (1.8)	0.51	0.14–1.81	0.30
Coronary artery disease, n (%)	3 (0.9)	2 (0.6)	1.25	0.21–7.28	0.81
Nonfatal myocardial infarction, n (%)	2 (0.6)	0 (0)	‐	‐	‐
Unstable angina, n (%)	1 (0.3)	0 (0)	‐	‐	‐
Coronary revascularization, n (%)	3 (0.9)	2 (0.6)	1.25	0.21–7.48	0.81
Major bleeding, n (%)	9 (2.6)	11 (3.4)	0.65	0.27–1.57	0.34
Cardiovascular death, n (%)	7 (2.0)	10 (3.0)	0.56	0.21–1.48	0.25

AF indicates atrial fibrillation; EPA/AA, eicosapentaenoic acid to arachidonic acid ratio; and HR, hazard ratio.

In the sensitivity analyses, when patients taking n‐3 PUFAs were excluded, the 3‐year cumulative incidence of AF recurrence was found to be significantly higher in the high EPA/AA group than in the low EPA/AA group (41.4% versus 27.6%; log‐rank test, P < 0.001; Figure S2). Similarly, the 5‐year cumulative incidence of MACE was significantly lower in the high EPA/AA group than in the low EPA/AA group (13.0% versus 25.5%; log‐rank test, P=0.021; Figure S3).

In the other sensitivity analyses using the EPA alone, EPA + DHA, and EPA/DHA, the results of both the AF recurrence and MACE for each fatty acid index were consistent with those using the EPA/AA (Figure S4A–C, and S5A–C). The hazard ratio for AF recurrence of the EPA/AA was the largest among the 4 fatty acid indices.

Among the included patients, in 661 of 673 (98%) the heart rate during sinus rhythm was evaluated at 6 months after the ablation. The heart rate was numerically lower in the high EPA/AA group than in the low EPA/AA group (70 [IQR, 64–79] bpm versus 72 [IQR, 65–81] bpm; P=0.09). The results using other fatty acid indices were consistent with that using the EPA/AA (EPA alone, 70 [IQR, 64–79] bpm versus 72 [IQR, 65–81] bpm; P=0.06; EPA + DHA, 70 [IQR, 64–79] bpm versus 73 [IQR, 65–81] bpm; P=0.05; EPA/DHA, 73 [IQR, 65–81] bpm versus 70 [IQR, 63–79] bpm; P=0.011 [Figure S6]).

The patients with AF recurrence within 3 years had a numerically higher incidence of heart failure hospitalizations from 3 years to 5 years after the ablation than those without (4/140 [2.9%] versus 7/533 [1.3%]; P=0.36). The patients with persistent AF recurrences had a significantly higher incidence of heart failure hospitalizations than those without (4/70 [5.7%] and 7/603 [1.2%]; P=0.019). In the high EPA/AA group, the incidence of heart failure hospitalizations was similar between the patients with and without AF recurrences (1/89 [1.1%] versus 3/256 [1.2%]; P=1.0), and those with and without persistent AF recurrences (1/45 [2.2%] versus 3/300 [1.0%]; P=1.0), respectively. In contrast, in the low EPA/AA group, the incidence of heart failure hospitalizations was numerically higher in the patients with AF recurrences than in those without (3/51 [5.9%] versus 4/277 [1.4%]; P=0.14) and was significantly higher in the patients with persistent AF recurrences than in those without (3/25 [12.0%] versus 4/303 [1.3%]; P=0.005), respectively.

Discussion

This study found that a high serum EPA/AA level before AF ablation was associated with a higher incidence of a 3‐year AF recurrence, including the persistent form of an AF recurrence. Conversely, it was associated with a lower incidence of a 5‐year MACE after ablation. The median EPA/AA value emerged as an independent predictor for both of those outcomes. Those results remained consistent even when considering patients who were not taking n‐3 PUFAs. Those results were also consistent with those using other fatty acid indices, such as the EPA alone, EPA + DHA, and EPA/DHA, and the hazard ratio for AF recurrence using the EPA/AA was the largest among those 4 fatty acid indices.

Recent research has explored the relationship between AF and the serum EPA levels or the intake of n‐3 PUFAs. A meta‐analysis has highlighted an increased risk of AF in individuals with an n‐3 PUFA intake. ³² Furthermore, a greater increase in the serum EPA levels due to the n‐3 PUFA intake was associated with a higher risk of new‐onset AF while also being linked to a lower risk of cardiovascular events. ¹⁵ Patients with AF were found to have higher serum EPA levels than those without. ³³ Moreover, patients with AF taking n‐3 PUFAs exhibited a higher incidence of hospitalizations for AF than those not taking n‐3 PUFAs.¹⁶ Despite the widespread use of AF ablation, there has been limited research evaluating the relationship between the serum EPA/AA level and post‐AF ablation outcomes. A prior study reported that a low serum EPA/AA level was associated with a higher incidence of AF recurrence following ablation in patients with persistent AF, however, that correlation was not observed in those with paroxysmal AF. ³⁴ In our study, we did not observe an interaction between the AF type and recurrence. This lack of interaction may have resulted from the differences in the patient characteristics, particularly because our study focused on an older population. This study consistently demonstrated an increased risk of AF occurrence with higher serum EPA levels, even after AF ablation, aligning with recent findings from large‐scale trials. ⁸ , ¹³

The mechanisms underlying the proarrhythmic effects of n‐3 PUFAs and higher serum EPA levels remain unclear. Several experimental studies have proposed both anti‐ and proarrhythmic effects of n‐3 PUFAs. These effects include reducing the asynchronous contractile activity of atrial myocytes, ³⁵ prolonging the atrial conduction time, suppressing AF inducibility, and enhancing atrial flutter inducibility. ³⁶ The higher incidence of AF associated with higher serum EPA/AA levels, even after AF ablation, suggests that serum EPA may influence proarrhythmic factors within the atrium beyond the ablation target areas, particularly the pulmonary veins. We considered that the serum EPA level was associated with the degree of parasympathetic nervous tone because previous studies have reported that a high n‐3 PUFA level increases the parasympathetic nervous tone ³⁷ and decreases the heart rate. ³⁸ In fact, we confirmed a lower heart rate during sinus rhythm after ablation among the high EPA/AA, EPA alone, EPA + DHA, and EPA/DHA groups. Regarding the higher risk of EPA on AF recurrence, a study reported that AF could be induced by simultaneous sympathetic and parasympathetic nervous stimulation. ³⁹ In general, older people have an increased sympathetic nervous tone. ⁴⁰ , ⁴¹ If a high serum EPA level could enhance the parasympathetic nervous tone, AF could occur even after AF ablation, especially among older patients. This consideration could be supported by the result of the subgroup analysis for the age. Regarding the lower risk of EPA on MACE, especially regarding heart failure hospitalizations, a study reported that the parasympathetic nervous tone could decrease the ventricular rate during AF leading to protection against heart failure worsening. ⁴² The association between the serum EPA/AA level and heart failure hospitalizations has been reported among an older population. ⁴³ In this study, the most robust association with the serum EPA/AA level was observed with hospitalizations for heart failure as part of MACE, and heart failure worsened less even though AF recurred in the patients in the high EPA/AA group. Furthermore, EPA is known to possess antithrombotic effects, and while there may be concerns regarding the bleeding risk, high serum EPA levels are actually associated with a lower incidence of bleeding events. ⁴⁴ The reduced risk of bleeding associated with EPA may also contribute to the lower occurrence of MACE in this study.

To our knowledge, this study represents the first attempt to investigate the impact of the serum EPA/AA level on both AF recurrence and MACEs in older patients, even after AF ablation. Our findings highlight a group of patients with a higher EPA/AA level before AF ablation, suggesting a more favorable prognosis, yet a potential risk of AF recurrence.

This study had several limitations. First, our analysis was based solely on the EPA/AA level before ablation, which can vary with n‐3 PUFA intake. Therefore, monitoring the EPA/AA level after ablation may be necessary. Second, we were unable to detect all instances of asymptomatic AF recurrence. While an implantable cardiac monitor can identify all AF recurrences, it is not covered by Japanese insurance for this purpose due to economic constraints. Nevertheless, all patients were instructed to regularly self‐check their pulse and underwent routine monitoring with a 1‐week external loop recorder or 24‐hour Holter ECG 6 months after ablation. Hence, we believed that the recurrence of asymptomatic AF was likely rare if it occurred at all. Third, the study had a limited number of patients and relatively few secondary outcomes, attributed to its single‐center design. Fourth, the observation period for AF recurrence was shorter than that for the MACE. However, a study reported that AF recurrences 3 years after ablation occurred less often from the pulmonary vein reconnections and were more common to occur from extrapulmonary vein triggers. ²⁰ Another study reported that the AF recurrence rate 2 years after ablation was constantly at 2%/year regardless of the AF type. ⁴⁵ Those findings suggested that the pathophysiology of the AF occurrence 2 or 3 years after the ablation was associated with causes other than the initial ablated sites. Therefore, we concluded that the 3‐year observation period was rational and sufficient to evaluate AF recurrences after the ablation. Fourth, the potential for residual confounding should be considered due to the significant differences in the baseline characteristics between the high and low EPA/AA groups despite adjusting for the factors in the multivariable Cox hazard models. Fifth, a large selection bias might have been inevitable because the measurement of the EPA/AA value was at the discretion of the physician in charge. Indeed, there were some differences in the baseline characteristics between the included and excluded patients. Finally, the applicability of these correlations to patients aged <65 years remains uncertain due to the lack of evaluable data among the younger population. Therefore, a larger multicenter study is needed to clarify the impact of the serum EPA/AA levels on AF recurrence and cardiovascular events after ablation.

Conclusions

In older patients, a higher serum EPA/AA level before AF ablation was associated with a higher risk of AF recurrence but a lower risk of MACE after ablation. Further research is needed to better understand the mechanism by which EPA influences these dual outcomes after ablation.

Disclosures

None.

Source of Funding

None.

Supporting information

Table S1

Figures S1–S6