Prevalence and risk factors of atrial fibrillation in type 2 diabetes mellitus: a hospital-based cross-sectional study Authors: Sukrisd Koowattanatianchai, Vimonsri Rangsrisaeneepitak, Boonyaporn Pheugphoolphol, Kiraphol Kaladee, Chatchai Kreepala Published: 2026 ABSTRACT AbstractBackgroundAtrial fibrillation (AF) is the most common sustained arrhythmia and is associated with adverse cardiovascular outcomes. In type 2 diabetes mellitus (T2DM), AF further increases the risk of heart failure, stroke, and premature death. However, epidemiological data in Asian populations remain limited. This study aimed to clarify the prevalence and identify risk factors of AF among patients with T2DM.MethodsWe conducted a hospital-based cross-sectional study using retrospectively collected medical records. AF was identified from physician documentation with electrocardiographic confirmation, including paroxysmal, persistent, and permanent types. Demographic, clinical, and laboratory variables were compared between AF and non-AF groups, and independent risk factors were assessed using multiple logistic regression.ResultsAmong 427 T2DM patients (mean age 63.6 ± 11.6 years), AF prevalence was 10.5%. Independent predictors of AF were older age (OR 1.06, 95% CI 1.02–1.11; p = 0.001), higher BMI (OR 1.07, 95% CI 1.01–1.13; p = 0.040), and heart failure (OR 4.39, 95% CI 1.56–12.39; p = 0.005). The predictive model incorporating these factors demonstrated fair discriminative ability (AUC 0.75, 95% CI 0.68–0.83).ConclusionsAF was present in one in ten T2DM patients. Older age, elevated BMI, and heart failure were independent predictors, and together provided moderate accuracy for AF risk prediction. These findings support targeted cardiovascular risk assessment and opportunistic AF screening in high-risk patients.Keywords: Atrial fibrillation, Type 2 diabetes mellitus, Prevalence, Risk factors, Cross-sectional study Introduction Type 2 diabetes mellitus (T2DM) is a major non-communicable chronic disease with a rising global prevalence, projected to reach 7079 per 100,000 population by 2030 [1]. It is a leading cause of microvascular and macrovascular complications, including stroke [2]. Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia and a well-established cause of thromboembolic events, stroke, and heart failure (HF), leading to substantial morbidity and mortality [3–5]. The coexistence of AF and T2DM represents a significant clinical challenge. T2DM is incorporated into stroke risk stratification schemes for AF [6] and is recognized as an independent risk factor for AF onset [7, 8]. Meta-analyses report a 28%–40% higher AF risk in individuals with T2DM compared with non-diabetic controls [9, 10]. Pathophysiological links include atrial structural remodeling, autonomic dysfunction, and systemic inflammation—a mechanism also implicated in other diabetic complications and identifiable through molecular and biochemical markers (e.g., inflammatory gene signatures [11], urine- or blood-based biomarkers [12–14]. From a patient care perspective, detecting AF in individuals with T2DM has important implications. AF in these patients necessitates timely anticoagulation to prevent stroke, optimization of cardiovascular risk factors, and careful selection of antidiabetic regimens to avoid adverse drug–drug interactions. In addition to the clinical burden, comorbidity with AF may further increase hospitalization rates and associated healthcare costs compared with T2DM alone, as inpatient care is a major driver of expenditure in chronic disease management [15]. Early identification of modifiable risk factors could therefore help prevent avoidable admissions and reduce both clinical and economic impact. Furthermore, AF may exacerbate glycemic instability via reduced cardiac output and physical activity, whereas effective rhythm and rate control may improve metabolic profiles. Advances in biomarker technology—such as volatile organic compound analysis [12, 13] or genetic risk profiling [14]—may in the future support earlier AF detection and tailored preventive strategies in T2DM. Previous studies have explored possible risk factors for AF in T2DM—including diabetes duration, glycemic control, and renal complications—but findings remain inconsistent [9, 16]. In Thailand, data on the prevalence and determinants of AF among patients with T2DM are scarce. This study aimed to determine the prevalence of AF in patients with T2DM attending a tertiary hospital and to identify clinical factors associated with AF in this population. Methods Study design and participants This cross-sectional study was conducted at Burapha University Hospital, Chonburi, Thailand. Between June 2023 and June 2024, 740 outpatients with T2DM attending the medical clinic were screened. T2DM was diagnosed according to the American Diabetes Association (ADA) criteria [17]. Patients without documented urinary albumin measurements or without 12-lead electrocardiogram (ECG) records from their most recent visit were excluded (n = 313), leaving 427 eligible participants for analysis. Ethics approval The study protocol was approved by the Burapha University Ethics Institutional Review Board (HS 114/2565) and conducted in accordance with the principles of the Declaration of Helsinki. As this study involved retrospective review of existing medical records with minimal risk to participants, the requirement for written informed consent was waived by the Burapha University Ethics Institutional Review Board. Consent for publication Not applicable. AF diagnosis AF, including paroxysmal, persistent, and permanent forms, was diagnosed by one cardiologist and independently confirmed by another. Diagnosis was based on 12-lead ECG criteria—namely, an irregularly irregular R–R interval (with intact atrioventricular conduction), absence of distinct repeating P waves, and irregular atrial activation—consistent with established definitions [18]. Patient ECGs and Holter monitoring, when available, were reviewed on arrival at the clinic and at any time during the study period. Clinical and anthropometric data Baseline characteristics collected included age, sex, body mass index (BMI), comorbidities, T2DM duration, medication use, and diabetes-related complications. Blood pressure and heart rate were measured at the time of the clinic visit during routine vital sign assessment. Hypertension was defined as the use of antihypertensive medication and/or systolic blood pressure > 130 mmHg and/or diastolic blood pressure > 80 mmHg, according to ADA guidelines [17]. Obesity was defined as BMI ≥ 25 kg/m², based on World Health Organization criteria for Asian populations [19]. Laboratory measurements Laboratory data were obtained within 12 months of the most recent clinic visit, acknowledging that some tests were performed annually. Parameters included glycated hemoglobin A1c (HbA1c), lipid profile, serum creatinine, and microalbumin urine. HbA1c was measured by high-performance liquid chromatography and reported as National Glycohemoglobin Standardization Program–equivalent percentages. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [20]. Microalbuminuria was defined as a urinary albumin-to-creatinine ratio (UACR) of 30–300 mg/g, and macroalbuminuria as UACR > 300 mg/g or proteinuria on urinalysis [21]. Sample adequacy assessment Based on a previously reported AF prevalence of 5.9% among patients with T2DM [22], approximately 398 participants would provide a 95% confidence level with a 2.2% margin of error. The final study population included 427 eligible participants, exceeding this estimated number and supporting the adequacy of the available sample. Statistical analysis Continuous variables were summarized as mean ± standard deviation and categorical variables as frequencies and percentages. Comparisons between patients with and without AF were performed using the independent t test for continuous variables (with Welch’s correction applied when variances were unequal) and Fisher’s exact test for categorical variables. Multivariable logistic regression with backward elimination was used to identify independent factors associated with AF, retaining variables with p < 0.20 in the model. Model performance was evaluated by the Hosmer–Lemeshow goodness-of-fit test and by receiver operating characteristic (ROC) analysis, with the area under the curve (AUC) and 95% confidence intervals used to assess the model’s ability to discriminate between patients with and without AF. Statistical significance was defined as p < 0.05. All analyses were performed using STATA version 16 (Stata Corp, College Station, TX, USA). Results Patients’ baseline characteristics The patients’ baseline characteristics are summarized in Table 1. In the cohort of 427 T2DM patients, 45 (10.5%) had documented AF, including both paroxysmal and permanent forms. All patients were aged between 26 and 92 years, with a mean age of 63.58 ± 11.62 years. The mean age of AF patients was significantly higher than that of patients without AF (70.23 ± 10.51 vs. 62.79 ± 11.51 years; p < 0.001). The mean BMI was 27.00 ± 5.29 kg/m². A substantial proportion of patients also presented with comorbidities, notably dyslipidemia (92.7%) and hypertension (74.9%). Chronic kidney disease (CKD), defined as an eGFR of < 60 mL/min/1.73 m², was observed in 33.5% of cases. Additionally, 9.6% had documented coronary artery disease, whereas 6.8% had HF. A history of stroke or transient ischemic attack (TIA) was present in 7.0% of patients. Compared with non-AF patients, those with AF had a higher prevalence of CKD, thyroid disease, and HF (48.9% vs. 31.7%, p = 0.029; 13.6% vs. 5.0%, p = 0.034; and 24.4% vs. 4.7%, p < 0.001, respectively). The duration of T2DM ranged from 1 year to > 10 years, with a mean duration of 8.29 ± 5.81 years. The duration of T2DM did not differ significantly between the two groups. Table 1Patients’ baseline characteristicsCharacteristicsAll patientsN = 427AF (+)N = 45AF (−)N = 382p-valueMale sex, n (%)223 (52.2)26 (57.8)197 (51.6)0.528Age (year), (mean ± SD)63.58 ± 11.6270.23 ± 10.5162.79 ± 11.51< 0.001BMI (kg/m2), (mean ± SD)27.00 ± 5.2928.20 ± 6.2026.86 ± 5.170.107Hypertension, n (%)320 (74.9)39 (88.7)281 (73.6)0.068Dyslipidemia, n (%)396 (92.7)43 (95.6)353 (92.4)0.759Hypertriglyceridemia, n (%)36 (8.4)2 (4.4)34 (8.9)0.406Stroke/TIA, n (%)30 (7.0)5 (11.1)25 (6.5)0.229Chronic kidney disease, n (%)143 (33.5)22 (48.9)121 (31.7)0.029Coronary artery disease, n (%)41 (9.6)4 (8.9)37 (9.7)0.864Gout, n (%)20 (4.7)3 (6.7)17 (4.5)0.456Cirrhosis, n (%)4 (0.9)1 (2.2)3 (0.8)0.361Heart failure, n (%)29 (6.8)11 (24.4)18 (4.7)< 0.001Thyroid disease, n (%)25 (5.9)6 (13.6)19 (5.0)0.034NASH, n (%)3 (0.7)0 (0.0)3 (0.8)0.715Diabetes duration (year), (mean ± SD)8.29 ± 5.818.38 ± 7.038.28 ± 5.660.918SBP (mmHg), (mean ± SD)134.00 ± 15.63130.49 ± 17.95134.42 ± 15.310.111DBP (mmHg), (mean ± SD)71.48 ± 12.0072.09 ± 13.5971.41 ± 11.820.721Heart rate (beats/min), (mean ± SD)81.76 ± 15.1577.27 ± 13.7182.89 ± 15.240.035Abbreviations: AF atrial fibrillation, BMI body mass index, TIA transient ischemic attack, NASH non-alcoholic steatohepatitis, SBP systolic blood pressure, DBP diastolic blood pressure, SD standard deviation Patients’ baseline characteristics Abbreviations: AF atrial fibrillation, BMI body mass index, TIA transient ischemic attack, NASH non-alcoholic steatohepatitis, SBP systolic blood pressure, DBP diastolic blood pressure, SD standard deviation Table 2 presents the laboratory values. The median HbA1c level was 7.53% ± 1.72%. The mean eGFR was significantly lower in AF patients than in non-AF patients (59.07 ± 25.26 vs. 72.36 ± 35.63 mL/min/1.73 m²; p = 0.017). Half of the patients exhibited albuminuria, with 26.1% classified as having microalbuminuria and 22.0% as having macroalbuminuria. The prevalence of microalbuminuria and macroalbuminuria did not differ significantly between AF and non-AF patients. Regarding antidiabetic drug usage among all patients, insulin was used in 30% of cases, and metformin—the most commonly prescribed antidiabetic drug—was used in 75.2% of all T2DM patients. Other antidiabetic drugs included sulfonylureas (43.3%), thiazolidinediones (32.1%), dipeptidyl peptidase-4 (DPP-4) inhibitors (30.4%), sodium–glucose cotransporter-2 (SGLT2) inhibitors (22.7%), and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) (7.7%). Table 2Patients’ laboratory valuesParametersMean ± SDp-valueAll patientsN = 427AF (+)N = 45AF (−)N = 382HbA1c, (%)7.53 ± 1.727.08 ± 1.107.58 ± 1.780.065Creatinine (mg/dL)1.45 ± 2.411.56 ± 1.881.43 ± 2.460.729eGFR (mL/min/1.73 m2)70.85 ± 34.9059.07 ± 25.2672.36 ± 35.630.017TC (mg/dL)161.59 ± 39.97155.48 ± 45.88162.31 ± 39.230.284HDLC (mg/dL)50.96 ± 14.7751.84 ± 12.3450.85 ± 15.050.689TG (mg/dL)132.16 ± 93.21117.55 ± 67.45133.87 ± 95.710.272LDLC (mg/dL)85.59 ± 33.2182.20 ± 40.3785.99 ± 82.420.475Presence of albuminuria, n (%)0.929 Normal217 (51.9)21 (50.0)197 (52.1) Microalbuminuria109 (26.1)12 (28.6)97 (25.8) Macroalbuminuria92 (22.0)9 (21.4)83 (22.1)Abbreviations: AF atrial fibrillation, HbA1c glycated hemoglobin A1c, eGFR estimated glomerular filtration rate, TC total cholesterol, HDLC high-density lipoprotein cholesterol, TG triglyceride, LDLC low-density lipoprotein cholesterol, SD standard deviation Patients’ laboratory values Abbreviations: AF atrial fibrillation, HbA1c glycated hemoglobin A1c, eGFR estimated glomerular filtration rate, TC total cholesterol, HDLC high-density lipoprotein cholesterol, TG triglyceride, LDLC low-density lipoprotein cholesterol, SD standard deviation Furthermore, more than half of the patients used angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) for hypertension control (51.1%), followed by calcium channel blockers (42.6%), beta-blockers (26.9%), and diuretics (17.8%). The majority of patients were prescribed statins (86%). The use of beta-blockers and diuretics was significantly higher in AF patients than in non-AF patients (66.7% vs. 22.3%, p < 0.001, and 35.6% vs. 15.7%, p = 0.003, respectively). Nearly all AF patients received anticoagulants, including both vitamin K antagonists and non-vitamin K oral anticoagulants (NOACs) (Table 3). Table 3Pharmacological treatments in the study populationTreatmentsAll patientsN = 427AF (+)N = 45AF (−)N = 382p-valueDiabetes drugs, n (%) Metformin321 (75.2)31 (68.9)290 (75.9)0.361 Sulfonylurea185 (43.3)15 (33.3)170 (44.5)0.203 DPP-4 inhibitor130 (30.4)12 (26.7)118 (30.9)0.612 Insulin128 (30.0)9 (20.0)119 (31.2)0.168 SGLT2 inhibitors97 (22.7)11 (24.4)86 (22.5)0.851 Thiazolidinedione137 (32.1)9 (20.0)128 (33.5)0.090 GLP1-RA33 (7.7)2 (4.4)31 (8.1)0.558Anti-hypertension drug, n (%) ACE/ARB218 (51.1)25 (56.6)193 (50.5)0.533 Beta-blocker115 (26.9)30 (66.7)85 (22.3)< 0.001 Calcium-channel blocker182 (42.6)21 (46.7)161 (42.2)0.633 Diuretic drug76 (17.8)16 (35.6)60 (15.7)0.003Statin drug, n (%)0.021 None58 (13.6)6 (13.3)52 (13.6) Statin used369 (86.4)39 (86.7)330 (86.4) Fibrate, n (%)30 (7.0)4 (8.9)26 (6.8)0.543 Ezetrol, n (%)13 (3.0)1 (2.2)12 (3.1)0.999Anti-thrombotic, n (%)< 0.001 None238 (55.7)5 (11.1)232 (60.6) Aspirin129 (30.2)7 (15.6)122 (31.9) Clopidogrel12 (2.8)3 (6.7)9 (2.4) Warfarin12 (2.8)12 (27.7)0 (0.0) NOAC18 (4.2)18 (40.0)0 (0) Aspirin + Clopidogrel18 (4.2)0 (0)18 (4.7)Abbreviations: AF atrial fibrillation, DPP-4 dipeptidyl peptidase-4, SGLT2 sodium–glucose cotransporter 2, GLP1-RA glucagon-like peptide-1 receptor agonist, ACE angiotensin-converting enzyme, ARB angiotensin receptor blocker, NOAC non-vitamin K oral anticoagulant Pharmacological treatments in the study population Abbreviations: AF atrial fibrillation, DPP-4 dipeptidyl peptidase-4, SGLT2 sodium–glucose cotransporter 2, GLP1-RA glucagon-like peptide-1 receptor agonist, ACE angiotensin-converting enzyme, ARB angiotensin receptor blocker, NOAC non-vitamin K oral anticoagulant Factors associated with AF Table 4 presents the results of a multivariable logistic regression analysis. The independent variables considered included age, hypertension, HF, thyroid disease, BMI, eGFR, HbA1c, and the use of diuretic medications. The analysis revealed that increasing age (per year), higher BMI (per 1 kg/m²), and the presence of HF were independently associated with AF among patients with T2D. The odds ratios (ORs) for age, BMI, and HF were 1.06 (95% CI, 1.02–1.11; p = 0.001), 1.07 (95% CI, 1.01–1.13; p = 0.040), and 4.39 (95% CI, 1.56–12.39; p = 0.005), respectively. These three variables predicted the risk of AF in patients with T2D with fair discriminative ability (AUC of ROC: 0.75; 95% CI, 0.68–0.83). The model’s goodness-of-fit was confirmed using the Hosmer–Lemeshow test, which indicated statistical appropriateness (p = 0.475) (Fig. 1). Table 4Prognostic factors associated with atrial fibrillation among patients with type 2 diabetes mellitus *Prognostic factorsCrude OR(95% CI)Adjusted OR(95% CI)p-valueAge (per 1 year increase)1.065(1.032–1.099)1.061(1.023–1.100)0.001Hypertension2.336(0.960–5.684)0.863(0.318–2.345)0.773Heart failure6.542(2.858–14.979)4.389(1.556–12.385)0.005Thyroid disease3.02(1.136–8.012)2.455(0.828–7.278)0.105BMI (per 1 kg/m2 increase)1.045(0.990–1.104)1.067(1.003–1.134)0.040eGFR < 60 (ml/min/1.73 m2)2.063(1.107–3.847)1.087(0.509–2.321)0.830HbA1c (per 1% increase)0.802(0.633–1.016)0.801(0.609–1.055)0.114Use of diuretic drug2.96(1.516–5.785)1.334(0.555–3.205)0.519Abbreviations: AF atrial fibrillation, T2DM type 2 diabetes mellitus, OR odds ratio, CI confidence interval, BMI body mass index, eGFR estimated glomerular filtration rate, HbA1c glycated hemoglobin* Multivariable logistic regression analysis adjusted for age, hypertension, heart failure, thyroid disease, BMI, eGFR < 60 mL/min/1.73 m², HbA1c, and use of diuretic drugs. No significant multicollinearity was identified among predictors included in the multivariable model, with all variance inflation factor (VIF) values < 2 Prognostic factors associated with atrial fibrillation among patients with type 2 diabetes mellitus * 1.065 (1.032–1.099) 1.061 (1.023–1.100) 2.336 (0.960–5.684) 0.863 (0.318–2.345) 6.542 (2.858–14.979) 4.389 (1.556–12.385) 3.02 (1.136–8.012) 2.455 (0.828–7.278) 1.045 (0.990–1.104) 1.067 (1.003–1.134) 2.063 (1.107–3.847) 1.087 (0.509–2.321) 0.802 (0.633–1.016) 0.801 (0.609–1.055) 2.96 (1.516–5.785) 1.334 (0.555–3.205) Abbreviations: AF atrial fibrillation, T2DM type 2 diabetes mellitus, OR odds ratio, CI confidence interval, BMI body mass index, eGFR estimated glomerular filtration rate, HbA1c glycated hemoglobin * Multivariable logistic regression analysis adjusted for age, hypertension, heart failure, thyroid disease, BMI, eGFR < 60 mL/min/1.73 m², HbA1c, and use of diuretic drugs. No significant multicollinearity was identified among predictors included in the multivariable model, with all variance inflation factor (VIF) values < 2 Receiver operating characteristic (ROC) curve for the multivariable logistic regression model predicting atrial fibrillation (AF) among patients with type 2 diabetes mellitus (T2DM). The model, incorporating age, body mass index (BMI), and presence of heart failure (HF), demonstrated fair discriminative ability with an area under the ROC curve (AUC) of 0.75 (95% CI, 0.68–0.83). The Hosmer–Lemeshow test confirmed good model fit (p = 0.475) Discussion In the present study, we examined the prevalence of AF among Thai patients with T2DM and identified factors associated with its occurrence. The observed prevalence of 10.5% falls toward the higher end of the range reported for Asian populations. For instance, Gherasim et al. [23] reported an AF prevalence of approximately 15% among patients with diabetes, while Swedish and Danish cohort studies documented rates of 8.9% and 19%, respectively [16, 24]. The Non-invasive Monitoring for Early Detection of Atrial Fibrillation (NOMED-AF) study found AF in one-quarter of patients with concomitant T2DM—substantially higher than in those without diabetes [25]. In contrast, lower rates have been documented in Asia, including 6.1% in a hospital-based study [26] and 4.4% in Japan [27]. The comparatively higher prevalence in our cohort may be explained by methodological differences, including the inclusion of paroxysmal AF, as well as the increasing global burden of AF [28]. Previous studies have linked AF risk in T2DM to advanced age, poor glycemic control, glycemic variability, longer diabetes duration, hypertension, and albuminuria [24]. A meta-analysis by Aune et al. [9] demonstrated a dose–response relationship between glycemia and AF risk, with each 1% increase in HbA1c conferring an odds ratio of 1.14 [29]. In our multivariable analysis, advancing age, higher BMI, and the presence of HF emerged as the strongest independent predictors of AF. Importantly, the predictive model demonstrated moderate discriminative ability, with an AUC of 0.75 (95% CI 0.68–0.83). The discriminative performance of the model, illustrated in Fig. 1, further supports the robustness of these predictors in identifying AF risk among patients with T2DM. This suggests potential clinical utility in incorporating easily obtainable parameters—age, BMI, and HF status—into opportunistic AF screening strategies. Pathophysiologically, T2DM-related AF may be mediated by glycemic fluctuations, oxidative stress, and inflammation, which collectively promote both structural and electrical remodeling of the atria [30]. Aging remains a well-established determinant of AF risk [31], and elderly patients with T2DM are particularly susceptible. HF demonstrated the strongest association with AF in our cohort. T2DM can promote HF through multiple mechanisms, including the accumulation of advanced glycation end products, oxidative stress, impaired inflammatory responses, dysregulated calcium handling, and altered microRNA expression [32]. The relationship between HF and AF is bidirectional—HF predisposes to AF, and AF can exacerbate HF [33]. This underscores the importance of early HF recognition and aggressive management in patients with T2DM to mitigate AF risk. In this context, novel therapies such as sodium–glucose cotransporter-2 (SGLT2) inhibitors may offer additional benefits. Although approximately one in five patients in this cohort received SGLT2 inhibitors, the short follow-up duration and limited use across subgroups precluded assessment of their potential protective effect against AF. Recent meta-analyses suggest that SGLT2 inhibitors may prevent AF in patients with T2DM, although this benefit may be attenuated in those with heart failure with preserved ejection fraction (HFpEF) or heart failure with mildly reduced ejection fraction (HFmrEF) [34, 35]. Emerging evidence also suggests that SGLT2 inhibitors may provide additional cardiovascular and anti-arrhythmic benefits beyond glycemic control [36–38]. Future studies with longer follow-up and more widespread access to SGLT2 inhibitors are warranted. Each 1 kg/m² increase in BMI was associated with a 7% increase in AF risk, consistent with prior studies [39, 40]. Obesity promotes systemic inflammation, atrial remodeling, and conduction abnormalities, partly via epicardial adipose tissue. This tissue secretes pro-inflammatory adipokines, including adiponectin, leptin, apelin, chemerin, omentin-1, interleukin-6, and tumor necrosis factor-α [41], which may contribute to interstitial fibrosis and impaired conduction velocity. Pathological remodeling of epicardial fat is characterized by fibro-fatty infiltration, while obesity-related increases in left atrial size and diastolic dysfunction further elevate atrial pressures. Interestingly, neither T2DM duration, HbA1c levels, nor albuminuria were significantly associated with AF in our analysis. This is in line with meta-analytic findings questioning the direct role of sustained hyperglycemia in AF development [42]. Emerging evidence suggests that glycemic variability, inflammation, oxidative stress, and possibly genetic susceptibility may be more relevant drivers [43]. Our data indicate that, in Thai patients with T2DM, HF and age are the predominant predictors of AF, consistent with findings from other Asian populations, whereas BMI appears to be a stronger predictor in Western cohorts [44]. Such variations may reflect genetic predisposition and lifestyle differences. Given the poorer prognosis of patients with both HF and AF, integrated management of these comorbidities should be a priority in this high-risk population. Beyond the variables analyzed in our study, other established risk factors for AF, such as peripheral artery disease (PAD), have also been reported. A recent large registry study demonstrated that PAD was associated with increased adverse cardiovascular outcomes among patients with AF [45]. In our cohort, however, PAD status was not recorded, and we were therefore unable to evaluate this association. Future studies incorporating PAD assessment are warranted to validate and extend these findings in T2DM populations. Study limitations This study has several limitations. First, the relatively small number of T2DM patients with AF may limit the generalizability of the findings. A larger cohort could allow detection of additional AF-related factors, such as hypertension, thyroid disease, HbA1c, T2DM duration, CKD, eGFR, albuminuria, or diuretic use, which may not have reached statistical significance here. Second, nearly half of the initially screened patients were excluded due to missing ECG or urinary albumin data. Although these exclusions were based solely on data availability and not on clinical outcomes, they may have introduced selection bias and limit the generalizability of our findings. Third, the significant predictors identified—age, HF, and BMI—are not unique to T2DM and have also been reported in the general population; future studies incorporating matched non-diabetic controls and adjusted analyses are warranted to clarify T2DM-specific risk profiles. Fourth, AF diagnoses relied on available ECG and Holter records in medical charts. The absence of systematic screening, such as routine Holter monitoring or periodic ECGs, may have led to underestimation of AF prevalence. Finally, our dataset did not include information on peripheral artery disease, an established AF risk factor, which limited our ability to evaluate its contribution in this population. In addition, lifestyle factors such as smoking and alcohol use were excluded from the analysis due to concerns about the accuracy of retrospective self-report and inconsistent documentation, and information on heart failure subtype (HFpEF vs. HFrEF) was not available. These missing data elements may have limited further exploration of risk factors. Conclusions This study is the first to report that the prevalence of AF among Thai patients with T2DM is 10.5%, which lies at the higher end of Asian estimates. Advancing age, higher BMI, and the presence of HF were the strongest independent predictors. Identifying these high-risk subgroups offers an opportunity for targeted AF screening and early intervention. Given the parallel global rise of T2DM and AF, incorporating systematic AF assessment into routine diabetes care may enable earlier detection, optimize management, and reduce cardiovascular morbidity in this vulnerable population. Acknowledgements The authors are sincerely thankful to Burapha University Hospital for providing their facilities. The authors also acknowledge the medical staff and all patients for their generosity and willingness to participate in the study. Institutional review board statement The study design and protocol were performed according to the tenets of the Declaration of Helsinki for research involving human subjects and were approved by the BUU Ethics Institutional Review Board (No. HS114/2565) on March 29, 2023. Informed consent statement The requirement for informed consent was waived by the Burapha University Ethics Institutional Review Board (HS 114/2565) because this study involved retrospective review of existing medical records and posed minimal risk to participants. Abbreviations Atrial fibrillation Type 2 diabetes mellitus Authors’ contributions Conceptualization, S.K., C.K. and V.R.; methodology, S.K., V.R. and K.K.; software, K.K.; validation, S.K.; formal analysis, K.K.; investigation, S.K. and K.K.; resources, B.P.; data curation, V.R., B.P., C.K. and K.K.; writing—original draft preparation, S.K. and V.R.; writing—review and editing, S.K. and V.R.; visualization, S.K., C.K. and K.K.; supervision, S.K.; project administration, S.K. and V.R.; funding acquisition, S.K. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the Faculty of Medicine, Burapha University. Data availability The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Kiraphol Kaladee, Email: kiraphol.kal@stou.ac.th. Chatchai Kreepala, Email: chatchaikree@gmail.com. References Associated Data Data Availability Statement The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.