Circulating metabolomic profile and its association with atrial fibrillation and systemic inflammation

Abstract

Background

Atrial fibrillation (AF) is a prevalent arrhythmia associated with chronic inflammation and metabolic dysregulation. Metabolic and inflammatory pathways may influence atrial electrophysiology and structural remodeling, contributing to AF onset and persistence, yet key circulating signatures remain poorly defined.

Objective

The purpose of this study was to characterize systemic inflammatory markers and circulating metabolomic profiles in patients with paroxysmal and persistent AF compared with non-AF controls.

Methods

In this prospective study, 101 patients were enrolled (mean age 59.9 ± 14.7 years; 73/101 [72%] male), including 41 with paroxysmal AF, 30 with persistent AF, and 30 controls. Fasting plasma samples were analyzed using targeted gas chromatography–mass spectrometry to quantify 156 metabolites and multiplex flow cytometry to profile 13 inflammatory cytokines.

Results

Interleukin-18 (IL-18) concentrations were significantly elevated in patients with AF compared with controls (P = .009), with highest levels observed in those with persistent AF. No significant differences were observed in IL-6 or high-sensitivity C-reactive protein. 9 metabolites showed significant differential abundance between patients with AF and controls. Metabolites including pyridoxine, N-acetylglutamine, 3-dehydroquinate, d-glucose, glucosamine, ascorbic acid, and galacturonic acid were reduced in AF, while capric acid and caprylic acid were elevated. β-Alanine was the only metabolite increased in persistent AF relative to both patients with paroxysmal AF and controls. Trimethylamine-N-oxide concentrations did not differ significantly between groups.

Conclusion

This study identifies distinct metabolic and inflammatory signatures in AF, with IL-18 and specific energy-related metabolites emerging as potential markers of disease phenotype. These hypothesis-generating findings warrant validation in larger cohorts.

Key Findings.

• ▪We performed targeted metabolomic and inflammatory profiling in a total of 101 patients with atrial fibrillation (AF) and controls with supraventricular tachycardia, with analyses adjusted for age, sex, and body mass index.
• ▪Interleukin-18 (IL-18) levels were significantly higher in patients with AF than in controls, with the highest concentrations seen in persistent AF, and remained independently associated with AF after adjustment for clinical covariates.
• ▪9 circulating metabolites differed in abundance in AF, including lower pyridoxine, ascorbic acid, and D-glucose, alongside higher medium-chain fatty acids such as capric acid and caprylic acid.
• ▪β-Alanine was uniquely elevated in persistent AF groups compared with both paroxysmal AF and control groups, suggesting a potential association with more advanced disease.
• ▪Trimethylamine-N-oxide levels did not differ between groups, indicating that its value as a biomarker may be context dependent and influenced by underlying vascular disease.

Introduction

Atrial fibrillation (AF), the most common sustained heart rhythm disorder in humans, affects 33.5 million people globally.¹ Both pulmonary vein and atrial substrate play an important role in the initiation and perpetuation of AF. There is extensive evidence of atrial and pulmonary vein remodeling in different population types with AF.^2,3 However, the pathophysiology of such remodeling remains a point of considerable ongoing debate. It is proposed that inflammasome activation has a causal role in the etiology of AF.⁴ The inflammasome, an intracellular multiprotein complex that functions as a molecular platform to activate the cysteine protease caspase 1, controls the production of pro-inflammatory cytokines such as interleukin-1β (IL-1β) and interleukin-18 (IL-18).⁵ The NLRP3 inflammasome (nucleotide-binding oligomerisation domain [NACHT], leucine-rich repeat, and pyrin domain–containing protein 3) is the main inflammasome linked to the pathology and progression of cardiovascular diseases.⁶ NLRP3 inflammasome activity is increased in atrial cardiomyocytes of patients with AF.⁷ This inflammasome activation is associated with atrial hypertrophy, fibrosis, shortening of refractory periods, and AF susceptibility.⁸ In addition to the role of inflammation, changes in autonomic nervous system activity can create a substrate that maintains AF.⁹ Patients with AF often share the same modifiable risk factors for coronary artery disease, including obesity, hypertension, and diabetes, all of which are substantially linked to dietary habits. Given the current knowledge on the role of the gut microbiota on hypertension, obesity, and atherosclerosis,¹⁰ its effect on the pathophysiology of AF is becoming a focus for research. Recent reviews have highlighted the growing role of integrative biological profiling approaches in AF, particularly in elucidating interactions between inflammation, metabolism, and arrhythmogenic processes.^11,12 In this prospective study, we aim to characterize the inflammatory markers and metabolomic profile in a cohort of patients with AF and compare them with a cohort of patients without a history of AF.

Methods

Study design and patient selection

This was an investigator-initiated and prospective clinical study. The study was conducted in accordance with the Declaration of Helsinki. The study protocol was approved by the Melbourne Health Research Ethics Committee, and written informed consent was obtained from all patients. The study included a total of 101 patients who were divided into 2 groups: (1) 71 patients with drug refractory paroxysmal or persistent AF and (2) 30 control patients with supraventricular tachycardia (SVT) with no documented history of AF and no AF on cardiac rhythm monitoring. Consecutive patients presenting for first time AF ablation of drug refractory paroxysmal or persistent AF were invited to participate in this study. Patients with a history suggestive of or who had documented SVT with no history of AF were also recruited. Exclusion criteria included (1) prior catheter ablation; (2) underlying acute or chronic inflammatory conditions such as inflammatory bowel disease, connective tissue disease, and active malignant neoplasm; (3) severe renal dysfunction with an estimated glomerular filtration rate (eGFR; <15 mL/min per 1.73 m²); and (4) age <18 years or inability to provide consent.

Metabolomic profiling and inflammatory marker assessment

Sample collection

From every patient, blood samples were obtained with the same tube type (BD plasma collection tubes; BD, Franklin Lakes, NJ, USA) after an overnight fast (>8 hours). The samples were collected from the right femoral vein access at the beginning of their planned cardiac procedure. Plasma samples were prepared by centrifugation at 1500g at −4°C for 12 minutes. This procedure was repeated twice. 25 μl of plasma aliquots was prepared and stored at −80°C until assayed as follows:

• 1.Gas chromatography–mass spectrometry comprehensive targeted metabolite profiling: Plasma samples were chemically derivatized and analyzed using a Shimadzu GCMS-TQ-8050 NX (Shimadzu Corporation, Kyoto, Japan) gas chromatograph, as described previously.¹³ The targeted metabolite panel comprised 156 compounds measured using a commercially available, validated gas chromatography–mass spectrometry platform, capturing metabolites across central carbon, amino acid, and lipid-related metabolic pathways. 2 internal standards were included: l-Valine-¹³C₅, ¹⁵N₁ and ¹³C₆ Sorbitol. Sample loading order was randomized, with quality control pooled samples run after every 5 samples and blank injections run after every 10 samples.
• 2.Inflammatory marker assays: Plasma samples were analyzed using the LEGENDplex Human Inflammation Panel 1 kit (BioLegend, San Diego, CA), measuring the plasma concentrations of 13 cytokines—IL-1β, interferon-α2 [IFN-α2], IFN-γ, tumor necrosis factor alpha, monocyte chemoattractant protein-1 (MCP-1), IL-6, IL-8, IL-10, IL-12p70, IL-17A, IL-18, IL-23, and IL-33—on a BD LSRFortessa X-20 flow cytometer (BD Biosciences, San Jose, CA). Similarly, high-sensitivity C-reactive protein (hsCRP) assay was performed with a LEGENDplex kit (BioLegend).
• 3.Trimethylamine-N-Oxide (TMAO) assay: The plasma concentration of circulating human TMAO was measured using enzyme-linked immunosorbent serologic assay kits (catalog no. EK715704, AFG Bioscience).

Ablation procedure

Procedural study protocol for the AF cohort

Anticoagulation was withheld on the morning of the procedure. All procedures were performed under general anesthesia with periprocedural transesophageal echocardiography to exclude left atrial thrombus. Double transseptal puncture was performed after heparinization (target activated clotting time 300–350 seconds). Endocardial 3-dimensional mapping with multipolar mapping catheters was performed using CARTO (Biosense Webster, Irvine, CA) or EnSite Precision/Velocity (Abbott Medical, St. Paul, MN) mapping systems. Irrigated contact force–sensing ablation catheters (ThermoCool SmartTouch ST, Biosense Webster, Irvine, CA or TactiCath SE, Abbott Medical, St. Paul, MN) were used in all cases. The goal of AF ablation was persistent isolation of all pulmonary veins. Additional ablation was performed at the physician’s discretion.

Procedural study protocol for the SVT cohort

Patients with SVT underwent catheter ablation targeting the specific underlying arrhythmia.

Statistical analysis

Data are expressed as frequency and percentage for categorical variables, means ± standard deviation for normally distributed continuous variables, and median (interquartile range) for nonnormally distributed continuous variables. Normality of data distribution was assessed through normality tests as required. Continuous variables were compared between study groups using the unpaired Student t test (normal distribution) or the Mann-Whitney U test (nonnormal distribution). Categorical variables were compared using the χ² test or the Fisher exact test. Associations were examined using univariable and multivariable analysis with SPSS version 26 (IBM Corporation, Armonk, NY). A probability value of P < .05 indicates statistical significance (∗); P < .01 (∗∗); P < .001 (∗∗∗); P < .0001 (∗∗∗∗). The Bonferroni correction method was applied to adjust P-value significance when testing for multiple metabolites and inflammatory markers; given the number of analytes assessed, this conservative approach was adopted to minimize false-positive discovery and findings are interpreted as hypothesis generating. For the metabolomic analyses, tests for differential metabolite abundance were conducted using the limma R package ¹⁴ on median peak–normalized and log-transformed data, with correction for multiple comparisons using the Benjamini-Hochberg method; age, sex, and body mass index (BMI) were included as covariates. Discriminating feature selection and multivariate statistical methods were applied using the mixOmics R package (supervised partial least squares discriminant analysis).¹⁵ Visualizations of metabolite z scores were produced using the ComplexHeatmap R package.¹⁶

Results

A total of 101 patients were enrolled for the study between September 2020 and September 2021. 41 (41%) patients had paroxysmal AF; 30 (30%) had persistent AF; and 30 (30%) had SVT. Baseline characteristics for the study groups are presented in Table 1. The control group had fewer conventional cardiac risk factors, were younger, had more female patients, and had a lower prevalence of hypertension, obstructive sleep apnea, and use of statin therapy. Certain inflammatory markers and metabolites were significantly different in the AF group compared with the control group (Figure 1).

Table 1.

Baseline characteristics of the AF cohort compared with the control cohort

Characteristic	AF	Control/SVT	P
No. of patients	71	30	<.001
Age (y)	62.3 ± 11	45 ± 15.5
Female gender	8 (11)	20 (67)	<.001
Persistent AF	30 (42)	–
CHA2DS2-VASc score	1.3 ± 1.2	0.9 ± 0.8	.2
BMI (kg/m2)	28.8 ± 4.9	27.9 ± 6.3	.4
Hypertension	27 (38)	3 (10)	.003
Diabetes	2 (3)	1 (3)	.9
Heart failure	8 (11)	2 (7)	.3
Stroke/TIA	2 (3)	0 (0)	.6
Vascular disease	6 (8)	2 (7)	.7
OSA	9 (13)	0 (0)	.04
Regular alcohol intake	27 (38)	9 (30)	.7
Smoking	5 (7)	5 (17)	.09
eGFR (mL/min per 1.73 m2)	79 ± 13	87 ± 5	.008
Total cholesterol level (mmol/L)	4.7 ± 0.9	4.8 ± 0.2	.9
Total bilirubin level (μmol/L)	12.3 ± 5	9.2 ± 4	.1
ALT level (U/L)	30.5 ± 17	27 ± 11	.5
ALP level (U/L)	71 ± 21	86 ± 36	.2
Anticoagulation	53 (75)	1 (3)	<.001
β-Blockers	30 (42)	14 (47)	.7
CCB	7 (12)	1 (3)	.1
Digoxin	7 (12)	0 (0)	.09
Amiodarone	9 (13)	1 (3)	.2
Flecainide	13 (18)	2 (7)	.2
Sotalol	24 (34)	0 (0)	<.001
Statin	17 (24)	1 (3)	.01
ACEI/ARB	19 (27)	3 (10)	.06
LV ejection fraction (%)	55.4 ± 8	63 ± 6.3	<.001
E/eʹ	8.8 ± 2.9	7.7 ± 2.3	.2
LA volume indexed (mL/m²)	44.3 ± 23	28.3 ± 18	.01

Values are presented as mean ± standard deviation or n (%).

ACEI = angiotensin converting-enzyme inhibitor; AF = atrial fibrillation; ALP = alkaline phosphatase; ALT = alanine aminotransferase; ARB = angiotensin receptor blocker; BMI = body mass index; CCB = calcium-channel blocker; CHA₂DS₂-VASc = congestive heart failure, hypertension, age ≥75 y, diabetes mellitus, stroke/transient ischemic attack, vascular disease, age 65–74 y, sex category; E/eʹ = ratio of early mitral inflow velocity to early diastolic mitral annular velocity; eGFR = estimated glomerular filtration rate; LA = left atrial; LV = left ventricular; OSA = obstructive sleep apnea; SVT = supraventricular tachycardia; TIA = transient ischemic attack.

Figure 1.

Link between gut dysbiosis and atrial fibrillation (AF) susceptibility through inflammatory and metabolic pathways. IL-18 = interleukin-18.

Inflammatory marker assessment

A total of 13 inflammatory cytokines were assessed in all patients. Overall, low levels of inflammatory cytokines in the study cohort were seen, except for IL-18. Serum concentrations of IL-18 were significantly higher in patients with AF than in controls (356 ± 196 pg/mL vs 264 ± 114 pg/mL; P = .009). Moreover, IL-18 levels were significantly higher in patients with persistent AF than in those with paroxysmal AF (417 ± 231 pg/mL vs 311 ± 153 pg/mL; P = .01) (Figure 2). In univariate analysis, only BMI and presence of AF were univariate predictors of elevated serum IL-18 concentration. In a multivariate model incorporating age, BMI, and presence of AF, both BMI and AF remained independent predictors of elevated serum IL-18 levels (Table 2). There was no significant difference in hsCRP between the control and AF groups (197 ± 393 vs 178 ± 272; P = .79). These findings suggest that IL-18 may be associated with the inflammatory profile of AF, particularly persistent AF, and this association appears independent of traditional cardiovascular risk markers such as hsCRP.

Figure 2.

Interleukin-18 (IL-18) levels are significantly higher in patients with atrial fibrillation (AF) than in controls. Furthermore, they are higher in patients with persistent AF (PeAF) than in those with paroxysmal AF (PAF).

Table 2.

Univariate and multivariate predictors of elevated serum IL-18 concentration

Variable	Univariate		Multivariate
	Standardized coefficient (95% CI)	P	Standardized coefficient (95% CI)	P
Age	0.17 (−0.3 to 4.5)	.08
Gender	−0.14 (−138 to 19)	.1
AF	0.23 (15.5–167)	.01	0.2 (9–154)	.02
BMI	0.32 (4.5–17.1)	<.001	0.3 (4–16)	.001
CHA2DS2-VASc score	0.1 (−17.1 to 48)	.3
Hypertension	0.18 (−8 to 145)	.08
Alcohol	0.07 (−51 to 103)	.5
Smoking	−0.03 (−136 to 100)	.7

AF = atrial fibrillation; BMI = body mass index; CHA₂DS₂-VASc = congestive heart failure, hypertension, age ≥75 y, diabetes mellitus, stroke/transient ischemic attack, vascular disease, age 65–74 y, sex category; CI = confidence interval; IL-18 = interleukin-18.

Given the significant sex imbalance between groups, univariable regression analyses were performed to assess associations between sex and inflammatory markers. IL-18 was not significantly associated with sex (coefficient −59.8; 95% confidence interval −138.9 to 19.2; P = .14). Similarly, no significant sex associations were observed for TMAO (P = .19), IL-6 (P = .50), tumor necrosis factor alpha (P = .23), and hsCRP (P = .85). Only MCP-1 demonstrated a statistically significant univariable association with sex (P = .03), but in multivariable analysis including AF status, age, BMI, and sex, MCP-1 was associated with sex and BMI but not with AF status. This suggests MCP-1 reflects demographic and anthropometric factors rather than disease phenotype. The primary inflammatory signal (IL-18) and metabolomic findings do not appear to be driven by sex differences within this cohort.

Serum TMAO assessment

The median plasma TMAO level for the study cohort was 131.9 ng/mL (interquartile range 116.7–144.2 ng/mL). In linear regression analysis, plasma TMAO levels did not correlate with age, gender, BMI, or any of the AF traditional risk factors (hypertension, vascular disease, and alcohol intake) (Table 3). In addition, there was no significant difference in plasma TMAO levels between the AF and control groups (130.4 [95% CI: 115.5–142.7] vs 134.4 [95% CI: 120.6–160.6]; P = .3). There was no significant difference in plasma TMAO levels between the paroxysmal AF and persistent AF groups (130.4 [95% CI: 116.4–143] vs 127 [95% CI: 115.5–141.1]; P = .6) (Figure 3). Moreover, the prevalence of vascular disease was similar between AF and control groups (8% vs 7%; P = .7), and no significant differences were observed in TMAO concentrations between these groups.

Table 3.

Univariate analysis to assess predictors of elevated serum TMAO levels

Variable	Univariate
	Standardized coefficient (95% CI)	P
Age	−0.06 (−1.6 to 0.82)	.5
Gender	0.13 (−13.3 to 65)	.2
AF	−0.1 (−60 to 17)	.3
BMI	−0.02 (−3.6 to 2.9)	.8
CHA2DS2-VASc score	0.1 (−17.1 to 48)	.3
Hypertension	−0.7 (−56 to 27)	.5
Alcohol	−0.07 (−48 to 23)	.5
Smoking	0.002 (−54 to 54)	.9
Vascular disease	−0.08 (−97 to 43)	.4

AF = atrial fibrillation; BMI = body mass index; CHA₂DS₂-VASc = congestive heart failure, hypertension, age ≥75 y, diabetes mellitus, stroke/transient ischemic attack, vascular disease, age 65–74 y, sex category; CI = confidence interval; TMAO = trimethylamine-N-oxide.

Figure 3.

Trimethylamine-N-oxide (TMAO) levels were not significantly different between the persistent atrial fibrillation (PeAF), paroxysmal atrial fibrillation (PAF), and control groups. AF = atrial fibrillation.

Serum metabolomic alterations between patients with AF and controls

A total of 156 serum metabolites were analyzed for the study cohort. Based on the Bonferroni correction, P < .0003 would be required to be considered significant. Given the age and sex differences in the study cohort baseline characteristics and the number of metabolites compared, the analysis was performed after adjusting for age, sex, BMI, and the number of comparisons to avoid false-positive discovery. When comparing patients with AF and controls, 9 metabolites had significantly different concentrations between the study cohorts (Figure 4). 7 metabolites: pyridoxine, N-acetylglutamine, 3-dehydroquinate, d-glucose, glucosamine, ascorbic acid, and galacturonic acid—were more abundant in controls than in patients with AF. 2 metabolites: capric acid and caprylic acid—were more abundant in patients with AF than in controls. β-Alanine was the only metabolite that was elevated in patients with persistent AF compared with patients with paroxysmal AF and controls (Figure 5). A heatmap of significantly altered metabolites, identified using limma analysis adjusted for age, sex, and BMI, demonstrated distinct clustering patterns between controls, participants with paroxysmal AF, and participants with persistent AF, with several metabolites showing reduced abundance in AF groups (Figure 6). Hierarchical clustering further suggested that several of the differential metabolites clustered within related metabolic pathways. Formal intermetabolite correlation analyses were beyond the scope of the present study and warrant evaluation in larger cohorts.

Figure 4.

Scatterplots depicting the metabolites that were significantly different between the atrial fibrillation (AF) and control groups. F = female; M = male.

Figure 5.

Scatterplots of β-alanine levels across the 3 groups; these levels were significantly higher in the persistent atrial fibrillation (AF) group than in other groups.

Figure 6.

Heatmap of significantly altered metabolites identified using limma analysis adjusted for age, sex, and body mass index (BMI), demonstrating distinct clustering patterns between the 3 groups. AF = atrial fibrillation; F = female; M = male.

Discussion

AF is increasingly recognized as a condition driven by complex interactions between metabolic, inflammatory, and electrophysiological pathways. In this prospective study, we identified a distinct serum metabolomic and cytokine profile in patients with AF compared with controls, characterized by elevated IL-18 levels and alterations in energy-related metabolites. These findings support the concept of AF as a systemic disorder, in which immune and metabolic dysregulation is associated with AF pathogenesis and persistence. The metabolite shifts observed are consistent with a prooxidative and energy-stressed environment that may contribute to atrial remodeling and substrate formation.

Metabolic dysregulation and AF

The identification of altered metabolic pathways in patients with AF suggests a shift in energy metabolism. The observed reduction in d-glucose and glucosamine levels in patients with AF may reflect impaired glycolysis and increased reliance on alternative energy sources, such as fatty acid oxidation or ketone bodies. This is supported by imaging studies using ¹⁸F-fluorodeoxyglucose positron emission tomography, which have shown increased glucose uptake in the atria of patients with AF, particularly those with persistent AF, suggesting heightened metabolic activity and stress within atrial tissue.¹⁷ Concurrent reductions in ascorbic acid levels point toward a prooxidative environment, which may contribute to the development of a fibrotic and proarrhythmic substrate. These observations align with prior studies exploring antioxidant strategies, including perioperative ascorbic acid supplementation, as a potential means to mitigate AF risk.¹⁸

β-Alanine, a nonessential amino acid involved in histidine metabolism and carnosine synthesis, was elevated in patients with persistent AF in our cohort. Carnosine functions as an intracellular pH buffer and antioxidant, particularly in excitable tissues such as the heart.¹⁹ The observed increase in β-alanine may reflect a metabolic adaptation to sustained atrial stress, suggesting its involvement in long-term remodeling processes in persistent AF. While direct evidence linking β-alanine to AF is limited, its role in myocardial metabolism and redox balance supports further investigation in the context of chronic arrhythmia.^20,21

Fatty acid metabolism

Elevated levels of medium-chain fatty acids (MCFAs), specifically capric acid and caprylic acid, in patients with AF suggest a potential link between gut microbiota composition and AF. These MCFAs can be derived from the diet, but they are also produced by gut microbial metabolism. Their increased abundance may reflect underlying gut dysbiosis, a state increasingly associated with systemic inflammation and cardiovascular disease.²² Normally, MCFAs are rapidly oxidized in the liver for energy via β-oxidation; however, elevated levels in patients with AF could suggest impaired oxidative metabolism or a shift in energy substrate preference toward fatty acid oxidation.²³ This alteration may contribute to metabolic stress in cardiomyocytes, potentially promoting atrial fibrosis and electrical remodeling.²⁴ In addition, caprylic acid plays a role in the acylation of ghrelin, a hormone involved in energy balance and hunger signaling, and disruptions in this process could further affect cardiac metabolic homeostasis.²⁵

TMAO and gut-heart axis: A conditional association?

While several inflammatory and metabolic markers differed between patients with AF and controls, TMAO levels did not differ significantly between groups in our cohort. TMAO, a gut microbiota–derived metabolite, has been implicated in cardiovascular disease through pro-inflammatory and prothrombotic mechanisms, and elevated levels have previously been linked to worse outcomes in patients with AF, including increased stroke risk and cardiovascular mortality.26, 27, 28 However, these associations appear to be more robust in populations with established vascular disease. In our study, the overall prevalence of vascular comorbidities was low, which may partly explain the lack of association observed. This supports the hypothesis that TMAO is likely a marker of underlying vascular pathology rather than a direct mediator of AF.²⁹ Therefore, its utility as a biomarker in AF may depend heavily on the clinical context and comorbidity burden of the population being studied.³⁰

Inflammatory markers in AF: The role of IL-18

The inflammatory response plays a pivotal role in AF and has been frequently implicated in the development and chronicity of AF.^31,32 Among the cytokines assessed, IL-18 was the only marker that differed significantly between patients with AF and controls, with the highest levels observed in those with persistent AF. IL-18 is a pro-inflammatory cytokine known to stimulate IFN-γ production, thereby activating a cascade of immune responses.³³ In the context of AF, IL-18 has been associated with structural remodeling of the atria, particularly through its potential involvement in atrial fibrosis and endothelial dysfunction.³⁴ Elevated IL-18 levels have been observed in patients with AF, with higher concentrations in those with persistent AF, as noted in both our cohort and prior clinical studies.³⁴ These findings suggest that IL-18 may reflect a state of atrial-specific inflammation more closely than general markers such as hsCRP or IL-6, which were not significantly altered in our study.

The absence of IL-6 elevation in our cohort contrasts with some prior studies. Li et al³⁵ reported elevated IL-6 concentrations in patients with AF compared with nonarrhythmia controls, with a complex pattern across AF subtypes. More recently, Valera Soria et al³⁶ found IL-6 elevated in persistent AF compared with paroxysmal AF in femoral vein samples, though not in coronary sinus samples, and notably their study did not include non-AF controls. Our use of patients with SVT as controls, who may themselves have some degree of baseline inflammation related to arrhythmia or autonomic activation, and the cross-sectional design may have limited sensitivity to detect IL-6 differences. Further prospective studies with nonarrhythmia control populations may help clarify these relationships.

IL-18 and metabolism: A possible intersection

Emerging literature suggests that IL-18 may be modulated by metabolic factors, including dietary lipids and potentially gut microbiota activity. Human studies have shown that plasma IL-18 levels vary with dietary fat quality; higher intake of saturated fatty acids is associated with elevated IL-18, whereas omega-3 polyunsaturated fats are linked to reduced levels.³⁷ In parallel, IL-18 plays a key role in maintaining intestinal immune balance and its expression is influenced by gut microbiota composition and function.³⁸ In our study, the observed elevations in IL-18 alongside altered MCFAs raise the possibility of an interaction between inflammation and metabolism. However, as dietary intake and gut microbiota composition were not directly assessed, the mechanistic basis for this association cannot be established from the present data.

IL-18 has been implicated in acute kidney injury and chronic kidney disease, where it may function as both a biomarker and a mediator of renal inflammation.³⁹ In our cohort, patients with AF had modestly lower eGFR than did controls (79 mL/min per 1.73 m² vs 87 mL/min per 1.73 m²). However, both groups had values within the normal to mildly reduced range, and univariable analysis demonstrated no association between IL-18 and eGFR (coefficient −0.14; 95% confidence interval −3.20 to 2.92; P = .93). This suggests that the observed IL-18 elevation in patients with AF is unlikely to be explained by differences in renal function within this cohort.

Clinical implications

In this study, IL-18 levels were independently associated with AF after adjustment for clinical covariates, consistent with prior literature linking inflammation to AF pathogenesis. Several circulating metabolites also showed significant associations with AF phenotype, suggesting broader metabolic differences across disease states. However, the clinical application of these markers for risk stratification or therapeutic targeting cannot be inferred from the present data and will require validation in larger prospectively characterized cohorts. Overall, these findings support the concept of AF as a systemic condition with measurable inflammatory and metabolic correlates.

Study limitations

This study has several limitations, including its modest sample size and lack of functional mechanistic data. While the design was prospective, it was not powered to assess longitudinal outcomes, and thus causal inferences cannot be firmly established. Quantification of AF burden (eg, time in AF before sampling) was not performed, which may influence circulating metabolite and cytokine concentrations. The use of patients with SVT as controls, while excluding patients with overt inflammatory or cardiac disease, may not fully represent a healthy population, and residual confounding from differences in autonomic tone or subclinical conditions cannot be excluded. This control group selection may also have reduced sensitivity for detecting differences in inflammatory markers such as IL-6, which have been reported in studies using nonarrhythmia controls.

Statistical analyses were adjusted for key confounders such as age, sex, and BMI; however, the possibility of residual confounding remains. An especially marked sex imbalance exists between study groups. Although formal testing did not demonstrate sex-driven effects on the primary inflammatory and metabolomic findings, this imbalance may reduce precision and limit the ability to fully separate disease effects from demographic differences. Finally, dietary intake and gut microbiota composition were not assessed, limiting interpretation of the origin of circulating metabolites. These findings should be considered hypothesis generating and warrant validation in larger prospective cohorts with longitudinal follow-up and mechanistic end points.

Conclusion

This study highlights distinct metabolic and inflammatory alterations in patients with AF, with IL-18 emerging as an inflammatory marker associated with AF persistence. The observed differences in oxidative stress markers, energy-related metabolites, and extracellular matrix components may reflect broader pathophysiological changes in AF. These hypothesis-generating findings warrant validation in larger cohorts with complementary dietary and microbiome profiling and in functional studies to clarify the causal relevance of these pathways.

Disclosures

Dr Kalman has received research support from Biosense Webster and Medtronic. Dr Kistler has received speaker fees from Abbott Medical. The rest of the authors have no conflicts of interest.