Advertisement

Burden of Microvascular Disease and Risk of Atrial Fibrillation in Adults with Type 2 Diabetes

      Abstract

      Background

      Epidemiological data on the associations of microvascular disease with atrial fibrillation are scarce. We evaluated the associations of diabetes-related microvascular disease in multiple vascular beds and its burden with incident atrial fibrillation among adults with type 2 diabetes.

      Methods

      A total of 7603 participants with type 2 diabetes and without atrial fibrillation were assessed for diabetic kidney disease, retinopathy, or neuropathy at baseline in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study. Incident atrial fibrillation events were adjudicated using follow-up electrocardiograms. Modified Poisson regression was used to generate risk ratios (RRs) and 95% confidence intervals (CIs) for atrial fibrillation.

      Results

      Of the 7603 participants (mean age 62.5 years, 38.0% women, 63.4% white), 63.3% (n = 4816) had microvascular disease—defined as the presence of ≥1 of: diabetic kidney disease, retinopathy, or neuropathy at baseline. Over a median of 7 years, there were 137 atrial fibrillation events (1.8%). Participants with microvascular disease had a 1.9-fold higher risk of incident atrial fibrillation compared with those without microvascular disease (RR 1.88; 95% CI, 1.20-2.95). Compared with no microvascular disease, the RRs for atrial fibrillation were 1.62 (95% CI, 1.01-2.61) and 2.47 (95% CI, 1.46-4.16) for those with 1 and ≥2 microvascular territories affected, respectively. The RRs for atrial fibrillation by type of microvascular disease were 1.57 (95% CI, 1.09-2.26), 0.95 (95% CI, 0.53-1.70), and 1.67 (95% CI, 1.15-2.44) for neuropathy, retinopathy, and diabetic kidney disease, respectively.

      Conclusions

      In a large cohort of adults with type 2 diabetes, the presence of microvascular disease and its burden were independently associated with higher risk of incident atrial fibrillation.

      Keywords

      Clinical Significance
      • In a large sample of adults with type 2 diabetes, participants with microvascular disease had a higher risk of incident atrial fibrillation compared with those without microvascular disease.
      • Participants with a greater burden of microvascular disease had an increased risk of atrial fibrillation compared with those with a lower burden of microvascular disease.

      Introduction

      Atrial fibrillation is the most prevalent sustained cardiac arrhythmia, and its prevalence is predicted to double by the year 2050.
      • Go AS
      • Hylek EM
      • Phillips KA
      • et al.
      Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.
      The risk of atrial fibrillation is significantly increased in individuals with type 2 diabetes, who have a 40% greater risk of atrial fibrillation compared with those without diabetes.
      • Huxley RR
      • Filion KB
      • Konety S
      • Alonso A
      Meta-analysis of cohort and case-control studies of type 2 diabetes mellitus and risk of atrial fibrillation.
      In patients with diabetes, the risk of atrial fibrillation increases by 3% for each year lived with diabetes.
      • Dublin S
      • Glazer NL
      • Smith NL
      • et al.
      Diabetes mellitus, glycemic control, and risk of atrial fibrillation.
      About 1 in 6 individuals with atrial fibrillation have diabetes;
      • Van Staa TP
      • Setakis E
      • Di Tanna GL
      • Lane DA
      • Lip GYH
      A comparison of risk stratification schemes for stroke in 79,884 atrial fibrillation patients in general practice.
      and diabetes is associated with worse atrial fibrillation symptoms and reduced quality of life, as well as increased morbidity and mortality among atrial fibrillation patients.
      • Echouffo-Tcheugui JB
      • Shrader P
      • Thomas L
      • et al.
      Care patterns and outcomes in atrial fibrillation patients with and without diabetes: ORBIT-AF Registry.
      The exact mechanisms linking diabetes mellitus to a greater risk of atrial fibrillation are not fully understood, but may involve diabetes-mediated maladaptive and profibrillatory structural, electrical, and microangiopathic alterations.
      • Plitt A
      • McGuire DK
      • Giugliano RP
      Atrial fibrillation, type 2 diabetes, and non-vitamin K antagonist oral anticoagulants: a review.
      While animal studies have suggested a possible link between microvascular dysfunction—a hallmark of diabetes mellitus—and atrial fibrillation,
      • Kato T
      • Yamashita T
      • Sekiguchi A
      • et al.
      AGEs-RAGE system mediates atrial structural remodeling in the diabetic rat.
      ,
      • Kume O
      • Teshima Y
      • Abe I
      • et al.
      Role of atrial endothelial cells in the development of atrial fibrosis and fibrillation in response to pressure overload.
      there is a paucity of population-based studies evaluating the effect of diabetes-related microvascular disease—and its burden—on incident atrial fibrillation among patients with type 2 diabetes.
      • Seyed Ahmadi S
      • Svensson A-M
      • Pivodic A
      • Rosengren A
      • Lind M
      Risk of atrial fibrillation in persons with type 2 diabetes and the excess risk in relation to glycaemic control and renal function: a Swedish cohort study.
      Using data from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study, we evaluated the associations of microvascular disease—assessed in several vascular beds—with incident atrial fibrillation in a large cohort of type 2 diabetes individuals. We hypothesized that a greater burden of microvascular disease would be associated with higher risk of incident atrial fibrillation.

      Methods

      Study Design

      This report followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for observational studies.
      • von Elm E
      • Altman DG
      • Egger M
      • Pocock SJ
      • Gøtzsche PC
      • Vandenbroucke JP
      The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.
      The data used for the analyses are publicly available through the National Heart Lung and Blood Institute Biorepository. The details on the design and methods of ACCORD have been published elsewhere.
      • Buse JB
      • Bigger JT
      • Byington RP
      • et al.
      Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.
      The ACCORD study was a double 2-by-2 factorial trial that enrolled 10,251 participants in 2 noncontiguous periods (January 2001 through June 2001, and January 2003 to October 2005) across 77 locations across the United States and Canada. Participants were randomly assigned to receive either an intensive glucose-lowering intervention aiming for a glycated hemoglobin (HbA1C) of <6% or standard treatment targeting an HbA1C of 7.0%-7.9%. Eligible participants met the following criteria at baseline: age 40 to 79 years (with a history of cardiovascular disease) or 55 to 79 years (with significant albuminuria, atherosclerosis, left ventricular hypertrophy, or a minimum of 2 cardiovascular disease risk factors).
      • Buse JB
      • Bigger JT
      • Byington RP
      • et al.
      Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.
      For the current investigation, we excluded participants with atrial fibrillation (n = 126) at baseline, missing baseline atrial fibrillation status (n = 45), inability to determine atrial fibrillation status at follow-up (n = 769), missing data on the presence of kidney disease (n = 435), retinopathy (n = 1070), or neuropathy (n = 203) at baseline. After these exclusions, 7603 participants were included in our analyses.
      We also conducted a separate analysis of a subset of participants (n = 2627) from the ACCORD Eye study,
      • Chew EY
      • Ambrosius WT
      • Howard LT
      • et al.
      Rationale, design, and methods of the Action to Control Cardiovascular Risk in Diabetes Eye Study (ACCORD-EYE).
      which included individuals without a history of photocoagulation or vitrectomy for diabetic retinopathy, who had an eye examination conducted by a study ophthalmologist or optometrist, as well as 7-field stereoscopic color fundus photographs at baseline.
      • Chew EY
      • Ambrosius WT
      • Howard LT
      • et al.
      Rationale, design, and methods of the Action to Control Cardiovascular Risk in Diabetes Eye Study (ACCORD-EYE).
      The research protocol was approved by the Institutional Review Board or ethics committee at each participating center and each participant gave an informed consent.
      • Buse JB
      • Bigger JT
      • Byington RP
      • et al.
      Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.
      ,
      • Chew EY
      • Ambrosius WT
      • Howard LT
      • et al.
      Rationale, design, and methods of the Action to Control Cardiovascular Risk in Diabetes Eye Study (ACCORD-EYE).

      Assessment of Microvascular Disease

      At baseline, blood and urine samples were obtained from each participant and analyzed on the day of the sample receipt. Serum and urine creatinine levels were determined enzymatically on a Roche Double Modular P Analytics automated analyzer. The inter-assay precision coefficients of these assays have been reported to be <1.4% and <2.2% for the high-quality control and low-quality control samples, respectively.
      • Ismail-Beigi F
      • Craven T
      • Banerji MA
      • et al.
      Effect of intensive treatment of hyperglycaemia on microvascular outcomes in type 2 diabetes: an analysis of the ACCORD randomised trial.
      Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration method.
      • Levey AS
      • Stevens LA
      • Schmid CH
      • et al.
      A new equation to estimate glomerular filtration rate.
      Urine microalbumin was assayed on spot urine by immunonephelometry on a Siemens BN11 nephelometer (Siemens Medical Solutions USA, Inc., Malvern, Pa). The sensitivity of this assay was 0.16 mg/dL, with an inter-assay coefficient of variation of 3.0%, 2.6%, and 4.9% for control levels of 0.89 mg/dL, 6.6 mg/dL, and 16.1 mg/dL, respectively. Urinary albumin excretion was estimated as the urine albumin-to-creatinine ratio (UACR) in mg of albumin per g of creatinine.
      We defined diabetic kidney disease as UACR ≥30 mg/g or eGFR <60 mL/min/1.73m2.
      Peripheral neuropathy was assessed using the Michigan Neuropathy Screening Instrument (MNSI) examination performed at baseline.
      • Herman WH
      • Pop-Busui R
      • Braffett BH
      • et al.
      Use of the Michigan Neuropathy Screening Instrument as a measure of distal symmetrical peripheral neuropathy in Type 1 diabetes: results from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications.
      The MNSI examination consists of a lower-extremity examination that includes inspection and assessment of vibratory sensation and ankle reflexes and is scored by assigning points for abnormal findings. The total possible score is 8 points. Neuropathy was defined based on an MNSI score >2, as this cutoff has been shown to have a good performance at diagnosing diabetic peripheral neuropathy.
      • Herman WH
      • Pop-Busui R
      • Braffett BH
      • et al.
      Use of the Michigan Neuropathy Screening Instrument as a measure of distal symmetrical peripheral neuropathy in Type 1 diabetes: results from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications.
      The presence of retinopathy was initially based on a self-report of diagnosis of diabetic retinopathy.
      • Buse JB
      • Bigger JT
      • Byington RP
      • et al.
      Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.
      In the subset of participants with baseline eye examinations (visual acuity measurement, slit lamp examination, and dilated fundus examination),
      • Chew EY
      • Ambrosius WT
      • Howard LT
      • et al.
      Rationale, design, and methods of the Action to Control Cardiovascular Risk in Diabetes Eye Study (ACCORD-EYE).
      the ACCORD Eye study utilized a modified version of the Early Treatment Diabetic Retinopathy Study (ETDRS) scale.
      Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12. Early Treatment Diabetic Retinopathy Study Research Group.
      Using this subset, we defined retinopathy as a baseline ETDRS score of ≥4, corresponding to at least mild non-proliferative diabetic retinopathy.
      We defined microvascular disease as the presence of at least one of the following: diabetic kidney disease, retinopathy, or neuropathy. The burden of microvascular disease was defined as the number of microvascular beds affected by microvascular disease and categorized as 0, 1, or ≥2.

      Ascertainment of Incident Atrial Fibrillation Events

      Baseline and follow-up resting electrocardiograms (ECGs) were obtained in all ACCORD sites.
      • Buse JB
      • Bigger JT
      • Byington RP
      • et al.
      Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.
      Incident atrial fibrillation cases were ascertained from 12-lead ECGs obtained at the biennial follow-up and close-out visits. ECGs were digitally acquired by trained electrocardiographers via standardized protocol using a GE MAC 1200 electrocardiograph (GE, Milwaukee, Wis) at 10 mm/mV calibration and a speed of 25 mm/s. All ECGs were electronically transmitted to a central core laboratory for processing and coding—the Epidemiological Cardiology Research Center, Wake Forest School of Medicine, Winston-Salem, NC. All ECGs were initially inspected visually for technical errors and inadequate quality prior to being automatically processed using GE 12-SL Marquette Version 2001 (GE); atrial fibrillation was defined by any Minnesota code 8.3.
      • Prineas RJ
      • Crow RS
      • Zhang Z-M
      The Minnesota Code Manual of Electrocardiographic Findings.

      Covariates

      The covariates selected a priori based on their relation with microvascular disease and atrial fibrillation were collected at baseline and included: age, sex, race/ethnicity, treatment arm, body mass index (BMI), cigarette smoking status, alcohol intake, systolic blood pressure (BP), use of BP-lowering medications, HbA1C, duration of diabetes, use of thiazolidinediones, ratio of total to high-density lipoprotein (HDL) cholesterol, history of atherosclerotic cardiovascular disease (including prior myocardial infarction, coronary revascularization, carotid or peripheral revascularization, angina, or stroke), or heart failure.
      • Buse JB
      • Bigger JT
      • Byington RP
      • et al.
      Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.

      Statistical Analyses

      We compared the baseline characteristics of participants according to the burden of microvascular disease using the Analysis of Variance, Kruskal–Wallis test, or the χ2 test as appropriate.
      We used multivariable Poisson regression with robust variance estimation to generate relative risks (RRs) and associated 95% confidence intervals (CI) relating microvascular disease to atrial fibrillation, as well as each individual type of microvascular disease (nephropathy, retinopathy, and neuropathy) and atrial fibrillation. Additionally, UACR and eGFR were log transformed due to their skewed distribution and related to the atrial fibrillation outcome as continuous variables.
      We constructed regression models in a hierarchical fashion. Model 1 adjusted for age, sex, race/ethnicity, treatment arm. Model 2 included variables in Model 1 plus current smoking, alcohol drinking, BMI, systolic BP, use of antihypertensive medications, ratio of total to HDL cholesterol, HbA1C, and duration of diabetes. Model 3 included variables in Model 2 plus further adjustment for prevalent atherosclerotic cardiovascular disease and prevalent heart failure.
      We conducted sensitivity analyses, which consisted of repeating the aforementioned analyses using the subset of participants with baseline eye examinations, in whom retinopathy was defined as a baseline ETDRS score of ≥4, corresponding to at least mild non-proliferative diabetic retinopathy.
      Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12. Early Treatment Diabetic Retinopathy Study Research Group.
      A two-sided P-value of < .05 was considered statistically significant for all analyses. All analyses were performed using STATA 14.2 (Stata, Inc, College Station, Texas).

      Results

      Baseline Characteristics of Study Participants

      Table 1 displays the baseline characteristics of participants. Of the 7603 participants (mean age 62.5 [SD 6.6] years, 38.0% women, 63.4% white) included in our study, 63.3% had microvascular disease (n = 4816), with 41.5% having one microvascular territory affected (n = 3157) and 21.8% having ≥2 microvascular territories affected (n = 1659). Regarding the individual microvascular disease types, 36.0% had diabetic kidney disease (n = 2736); 11.3% had retinopathy (n = 860); and 41.8%, neuropathy (n = 3177). Participants with a higher burden of microvascular disease were generally older, more frequently men, white, and had higher BMI, systolic BP, and HbA1C, as well as a longer duration of diabetes (Table 1).

      Incident Atrial Fibrillation by the Presence and Burden of Microvascular Disease

      During a median follow-up of 7 years, 137 participants experienced an atrial fibrillation event.
      After adjusting for relevant confounders, participants with microvascular disease had a 1.9-fold higher risk of incident atrial fibrillation compared with those without microvascular disease (RR 1.88; 95% CI, 1.20-2.95). A greater burden of microvascular disease was associated with increased risks of atrial fibrillation. Compared with those with no microvascular disease, the RRs of atrial fibrillation were 1.62 (95% CI, 1.01-2.61) for those with microvascular disease in one vascular territory and 2.47 (95% CI, 1.46-4.16) for those with microvascular disease in ≥2 vascular beds (Table 2).

      Incident Atrial Fibrillation by Individual Type of Microvascular Disease

      After adjusting for relevant confounders, the RRs for atrial fibrillation were 1.57 (95% CI, 1.09-2.26), 0.95 (95% CI, 0.53-1.70), and 1.67 (95% CI, 1.15-2.44) for neuropathy, retinopathy, and diabetic kidney disease, respectively (Table 3). Additionally, abnormalities of kidney function markers assessed individually were related to greater risks of atrial fibrillation (RR 1.25; 95% CI, 1.02-1.53) per 1-SD decrease in log(eGFR), and 1.32 (95% CI, 1.12-1.56) per 1-SD higher log (UACR).

      Sensitivity Analyses

      We tested the robustness of our findings by conducting sensitivity analyses using the subset of participants (n = 2627) who had retinopathy defined by baseline ETDRS scores. In this subset, 39 participants developed incident atrial fibrillation. Microvascular disease remained positively associated with a higher risk of atrial fibrillation (RR 2.59; 95% CI, 0.98-6.88). Likewise, a greater burden of microvascular disease was related to higher risk of atrial fibrillation (RR 3.38; 95% CI, 1.12-10.14) for ≥2 microvascular diseased territories compared with no microvascular disease (Supplementary Table 1, available online). Consistent with our main results, the risk of atrial fibrillation was higher among participants with individual types of microvascular disease compared with those without (Supplementary Table 2, available online).
      Supplementary Table 1Rates and Relative Risks for Incident Atrial Fibrillation by Burden of Microvascular Disease in the Subset of Participants with Eye Examinations at Baseline (N = 2627)
      Cases/No. at RiskModel 1Model 2Model 3
      RR (95% CI)P ValueRR (95% CI)P ValueRR (95% CI)P Value
      MVD
       Absent5/579 (0.9)1 (Reference)1 (Reference)1 (Reference)
       Present34/ 2048 (1.7)2.39 (0.91-6.25).0762.57 (0.97-6.82).0572.59 (0.98-6.88).056
      Burden of MVD
       05/579 (0.9)1 (Reference)1 (Reference)1 (Reference)
       110/1009 (1.0)2.04 (0.74-5.65).1702.16 (0.78-5.94).1372.20 (0.80-6.05).127
       ≥ 224/1039 (2.3)2.85 (1.02-7.95).0453.42 (1.15-10.16).0273.38 (1.12-10.14).030
      Microvascular disease was defined as ≥ 1 of nephropathy, neuropathy, or retinopathy.
      Model 1 adjusted for age, sex, race and treatment arm; model 2, model 1 variables plus duration of diabetes, hemoglobin A1C, cigarette smoking, alcohol intake; body mass index, total-to-HDL cholesterol, systolic BP, use of BP-lowering medications; model 3, model 2 variables plus history of ASCVD, history of HF, and use of thiazolidinediones.
      ASCVD = atherosclerotic cardiovascular disease; BP = blood pressure; CI = confidence interval; HDL = high-density lipoprotein; HF = heart failure; MVD = microvascular disease; RR = relative risk.
      Supplementary Table 2Rates and Relative Risks for Incident Atrial Fibrillation by Type of Microvascular Disease in the Subset of Participants with Eye Examinations at Baseline (N = 2627)
      Cases/No. at Risk (%)Model 1Model 2Model 3Model 4
      RR (95% CI)P ValueRR (95% CI)P ValueRR (95% CI)P ValueRR (95% CI)P Value
      Neuropathy
       Absent15/1498 (1.0)1 (Reference)1 (Reference)1 (Reference)1 (Reference)
       Present24/1129 (2.1)1.64 (0.84-3.19).1451.59 (0.80-3.18).1881.58 (0.78-3.20).2031.59 (0.79-3.20).194
      Retinopathy
       Absent20/1266 (1.6)1 (Reference)1 (Reference)1 (Reference)1 (Reference)
       Present19/1361 (1.4)0.97 (0.49-1.95).9371.15 (0.53-2.52).7181.16 (0.52-2.56).7191.14 (0.51-2.56).743
      Kidney disease
       Absent17/1767 (1.0)1 (Reference)1 (Reference)RefNA
       Present22/860 (2.6)2.25 (1.15-4.43).0182.49 (1.23-5.03).0112.48 (1.23-4.99).011NA
      eGFR, mL/min/1.73m2
       Normal eGFR34/2442 (1.4)1 (Reference)1 (Reference)1 (Reference)1 (Reference)
       Low eGFR5/185 (2.7)1.19 (0.47-3.05).7141.25 (0.47-3.30).6531.25 (0.48-3.28).644NA
       Per 1-SD lower log(eGFR)39/2627 (1.5)1.25 (0.88-1.78).2171.29 (0.90-1.84).1641.29 (0.91-1.83).157NA
      Albuminuria
       Absent22/1886 (1.2)1 (Reference)1 (Reference)1 (Reference)1 (Reference)
       Present17/741 (2.3)1.85 (0.96-3.56).0662.05 (1.06-3.98).0332.03 (1.05-3.93).0352.01 (1.04-3.89).037
       Per 1-SD higher log(UACR)39/2627 (1.5)1.27 (0.97-1.66).0851.35 (1.01-1.81).0401.34 (1.01-1.78).0471.32 (0.99-1.75).058
      Microvascular disease was defined as ≥ 1 of nephropathy, neuropathy, or retinopathy.
      Model 1 adjusted for age, sex, race and treatment arm; model 2, model 1 variables plus duration of diabetes, hemoglobin A1C, cigarette smoking, alcohol intake; body mass index, total-to-HDL cholesterol, systolic BP, use of BP-lowering medications; model 3, model 2 variables plus history of ASCVD, history of HF, and use of thiazolidinediones; model 4, model 3 variables plus eGFR.
      ASCVD = atherosclerotic cardiovascular disease; BP = blood pressure; CI = confidence interval; eGFR = estimated glomerular filtration rate; HDL = high-density lipoprotein; HF = heart failure; RR = relative risk.
      Table 1Characteristics of Participants by Burden of Microvascular Disease at Baseline
      CharacteristicTotalBurden of Microvascular DiseaseP Value
      01≥2
      n7603278731571659
      Age, years62.5 (6.6)61.3 (6.3)62.7 (6.6)64.2 (6.8)< .001
      Women, %38.042.335.834.9< .001
      Race/ethnicity, %.002
       White63.461.264.864.6
       Black18.118.217.519.0
       Hispanic6.88.16.25.6
       Other11.712.511.510.9
      Body mass index, kg/m232.3 (5.4)31.8 (5.3)32.5 (5.5)32.9 (5.4)< .001
      Current smoking, %13.913.114.414.4.303
      Alcohol drinking, %24.624.624.923.8.689
      Systolic BP, mm Hg135.8 (16.9)133.5 (15.7)135.8 (17.1)139.6 (17.8)< .001
      Diastolic BP, mm Hg74.9 (10.6)75.6 (10.2)74.9 (10.6)73.6 (10.9)< .001
      Use of BP-lowering drug, %84.080.684.289.3< .001
      Use of thiazolidinediones, %22.522.822.322.7.885
      Use of insulin, %34.726.133.451.5< .001
      Hemoglobin A1C, %8.3 (1.0)8.1 (1.0)8.3 (1.0)8.4 (1.0)< .001
      Duration of diabetes, years9.0 (5.0-15.0)7.0 (4.0-12.0)9.0 (5.0-15.0)13.0 (8.0-20.0)< .001
      Prevalent atherosclerotic cardiovascular disease, %34.730.036.040.1< .001
      Prevalent heart failure, %3.92.73.66.6< .001
      Total cholesterol, mg/dL182.5 (41.9)182.5 (40.1)182.9 (43.2)181.6 (42.5).553
      HDL cholesterol, mg/dL41.8 (11.3)42.8 (11.5)41.2 (11.0)41.2 (11.6)< .001
      LDL cholesterol, mg/dL103.6 (33.7)104.9 (33.0)103.5 (34.6)101.9 (33.4).017
      Total/HDL cholesterol ratio4.6 (1.7)4.5 (1.6)4.7 (1.8)4.7 (1.6)< .001
      eGFR, mL/min/1.73m285.4 (16.9)89.3 (13.5)85.3 (16.9)78.8 (19.6)< .001
      UACR, mg/g14.0 (7.0-43.0)8.0 (5.0-14.0)15.0 (7.0-48.0)64.0 (31.0-191.0)< .001
      MNSI score2.0 (0.5-3.0)1.0 (0.0-2.0)2.5 (1.0-4.0)3.5 (3.0-5.0)< .001
      Data are mean (standard deviation), median (interquartile range) or proportion (%) unless otherwise indicated.
      Microvascular disease was defined as ≥1 of nephropathy, neuropathy, or retinopathy.
      BP = blood pressure; eGFR = estimated glomerular filtration rate in mL/min/1.73m2; HDL = high-density lipoprotein; LDL = low-density lipoprotein; MNSI = Michigan Neuropathy Screening Instrument; UACR = urine albumin-to-creatinine ratio in mg/g.
      Table 2Rates and Relative Risks for Incident Atrial Fibrillation by Burden of Microvascular Disease
      Cases/No. at Risk (%)Model 1Model 2Model 3
      RR (95% CI)P ValueRR (95% CI)P ValueRR (95% CI)P Value
      Microvascular disease
       Absent26/2787 (0.9)1 (Reference)1 (Reference)1 (Reference)
       Present111/4816 (2.3)1.99 (1.29-3.06).0021.90 (1.22-2.97).0041.88 (1.20-2.95).005
      Burden of microvascular disease
       026/2787 (0.9)1 (Reference)_1 (Reference)1 (Reference)
       159/3157 (1.9)1.69 (1.06-2.68).0271.62 (1.01-2.59).0461.62 (1.01-2.61).046
       ≥252/1659 (3.1)2.53 (1.57-4.08)< .0012.55 (1.53-4.28)< .0012.47 (1.46-4.16).001
      Microvascular disease was defined as ≥ 1 of nephropathy, neuropathy, or retinopathy.
      Model 1 adjusted for age, sex, race, and treatment arm; model 2, model 1 variables plus duration of diabetes, hemoglobin A1C, cigarette smoking, alcohol intake; body mass index, total-to-HDL cholesterol, systolic BP, use of BP-lowering medications; model 3, model 2 variables plus history of atherosclerotic cardiovascular disease, history of heart failure, and use of thiazolidinediones.
      BP = blood pressure; CI = confidence interval; HDL = high-density lipoprotein; RR = relative risk.
      Table 3Rates and Relative Risks for Incident Atrial Fibrillation by Type of Microvascular Disease
      Cases/No. at Risk (%)Model 1Model 2Model 3Model 4
      RR (95% CI)P ValueRR (95% CI)P ValueRR (95% CI)P ValueRR (95% CI)P Value
      Neuropathy
       Absent56/4426 (1.3)1 (Reference)1 (Reference)1 (Reference)1 (Reference)
       Present81/ 3177 (2.6)1.61 (1.14-2.26).0071.58 (1.10-2.26).0121.58 (1.10-2.27).0131.57 (1.09-2.26).015
      Retinopathy
       Absent123/6743 (1.8)1 (Reference)1 (Reference)1 (Reference)1 (Reference)
       Present14/860 (1.6)0.88 (0.51-1.53).6551.02 (0.57-1.81).9551.00 (0.56-1.79).9970.95 (0.53-1.70).860
      Kidney disease
       Absent63/4867 (1.3)1(Reference)1 (Reference)1 (Reference)NA
       Present74/2736 (2.7)1.82 (1.29-2.57).0011.74 (1.20-2.52).0031.67 (1.15-2.44).007NA
      eGFR, mL/min/1.73m2
       Normal eGFR115/6963 (1.7)1 (Reference)1 (Reference)1 (Reference)NA
       Low eGFR22/640 (3.4)1.43 (0.90-2.27).1311.43 (0.87-2.34).1571.36 (0.83-2.23).224NA
       Per 1-SD lower log(eGFR)137/7603 (1.8)1.25 (1.04-1.50).0171.28 (1.05-1.57).0151.25 (1.02-1.53).028NA
      Albuminuria
       Absent75/5235 (1.4)1 (Reference)1 (Reference)Ref1 (Reference)
       Present62/2368 (2.6)1.75 (1.25-2.47).0011.74 (1.20-2.54).0041.70 (1.17-2.48).0061.68 (1.15-2.44).007
       Per 1-SD higher log(UACR)137/7603 (1.8)1.35 (1.17-1.56)< .0011.37 (1.16-1.63)< .0011.35 (1.14-1.60)< .0011.32 (1.12-1.56).001
      Microvascular disease was defined as ≥ 1 of nephropathy, neuropathy, or retinopathy.
      Model 1 adjusted for age, sex, race and treatment arm; model 2, model 1 variables plus duration of diabetes, hemoglobin A1C, cigarette smoking, alcohol intake; body mass index, total-to-HDL cholesterol, systolic BP, use of BP-lowering medications; model 3, model 2 variables plus history of atherosclerotic cardiovascular disease, history of heart failure, and use of thiazolidinediones; model 4, model 3 variables plus eGFR.
      BP = blood pressure; CI = confidence interval; eGFR = estimated glomerular filtration rate; HDL = high-density lipoprotein; NA = not applicable; RR = relative risk; UACR = urine albumin-to-creatinine ratio.

      Discussion

      In this study, we performed a comprehensive evaluation of the association of microvascular disease, assessed in multiple vascular beds, as well as its burden, with incident atrial fibrillation in a large sample of individuals with type 2 diabetes. We found that the presence of microvascular disease and the extent of microvascular disease burden were each independently associated with an increased risk of incident atrial fibrillation, after accounting for multiple confounders, including the degree of glycemic control, duration of diabetes, blood pressure, and the presence of cardiovascular disease. Individual microvascular disease types were also associated with a higher risk of atrial fibrillation. Our findings point to the relevance of microvascular disease in atrial fibrillation risk assessment within the framework of type 2 diabetes.
      Our study is one of the few to comprehensively evaluate the effect of microvascular disease assessed in multiple vascular beds on the incidence of atrial fibrillation in patients with type 2 diabetes. Previous investigations of the influence of microvascular disease on atrial fibrillation occurrence were limited in many ways, including the assessment of only one microvascular bed (often the kidney vasculature)
      • Seyed Ahmadi S
      • Svensson A-M
      • Pivodic A
      • Rosengren A
      • Lind M
      Risk of atrial fibrillation in persons with type 2 diabetes and the excess risk in relation to glycaemic control and renal function: a Swedish cohort study.
      ,
      • Bansal N
      • Zelnick LR
      • Alonso A
      • et al.
      eGFR and albuminuria in relation to risk of incident atrial fibrillation: a meta-analysis of the Jackson Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Cardiovascular Health Study.
      ,
      • Molnar AO
      • Eddeen AB
      • Ducharme R
      • et al.
      Association of proteinuria and incident atrial fibrillation in patients with intact and reduced kidney function.
      and the use of a retrospective design.
      • Seyed Ahmadi S
      • Svensson A-M
      • Pivodic A
      • Rosengren A
      • Lind M
      Risk of atrial fibrillation in persons with type 2 diabetes and the excess risk in relation to glycaemic control and renal function: a Swedish cohort study.
      ,
      • Molnar AO
      • Eddeen AB
      • Ducharme R
      • et al.
      Association of proteinuria and incident atrial fibrillation in patients with intact and reduced kidney function.
      Additionally, most of the prior studies were not focused on individuals with type 2 diabetes.
      • Bansal N
      • Zelnick LR
      • Alonso A
      • et al.
      eGFR and albuminuria in relation to risk of incident atrial fibrillation: a meta-analysis of the Jackson Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Cardiovascular Health Study.
      ,
      • Molnar AO
      • Eddeen AB
      • Ducharme R
      • et al.
      Association of proteinuria and incident atrial fibrillation in patients with intact and reduced kidney function.
      Our results are in agreement with prior reports of a positive association between microvascular disease and left atrial volume, a predictor of atrial fibrillation.
      • Bonapace S
      • Rossi A
      • Lipari P
      • et al.
      Relationship between increased left atrial volume and microvascular complications in patients with type 2 diabetes.
      ,
      • Kadappu KK
      • Boyd A
      • Eshoo S
      • et al.
      Changes in left atrial volume in diabetes mellitus: more than diastolic dysfunction?.
      Our findings also corroborate those from prior reports observing an increased risk of atrial fibrillation among individuals with kidney disease.
      • Seyed Ahmadi S
      • Svensson A-M
      • Pivodic A
      • Rosengren A
      • Lind M
      Risk of atrial fibrillation in persons with type 2 diabetes and the excess risk in relation to glycaemic control and renal function: a Swedish cohort study.
      ,
      • Bansal N
      • Zelnick LR
      • Alonso A
      • et al.
      eGFR and albuminuria in relation to risk of incident atrial fibrillation: a meta-analysis of the Jackson Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Cardiovascular Health Study.
      ,
      • Molnar AO
      • Eddeen AB
      • Ducharme R
      • et al.
      Association of proteinuria and incident atrial fibrillation in patients with intact and reduced kidney function.
      Likewise, our findings of a positive association between neuropathy and atrial fibrillation is consistent with a previous report from the general population (including a subset of individuals with diabetes).
      • Agarwal SK
      • Norby FL
      • Whitsel EA
      • et al.
      Cardiac autonomic dysfunction and incidence of atrial fibrillation: results from 20 years follow-up.
      The lack of a positive relation between retinopathy and atrial fibrillation in our study could be related to the smaller number of participants with objectively assessed retinal disease and is inconsistent with prior studies reporting a positive association between retinopathy and atrial fibrillation risk.
      • Lee S-R
      • Choi E-K
      • Rhee T-M
      • et al.
      Evaluation of the association between diabetic retinopathy and the incidence of atrial fibrillation: a nationwide population-based study.
      ,
      • Phan K
      • Mitchell P
      • Liew G
      • et al.
      Relationship between macular and retinal diseases with prevalent atrial fibrillation - an analysis of the Australian Heart Eye Study.
      The pathways linking microvascular disease to atrial fibrillation in patients with type 2 diabetes largely remain to be clarified. One pathway could be related to structural microvascular changes leading to atrial remodeling. Indeed, chronic hyperglycemia leads to the formation of advanced glycation end products that deposit in the endothelium and arteriolar walls; this may ultimately result in thickening of capillary basement membranes and microvascular remodeling throughout the myocardium.
      • Falcão-Pires I
      • Leite-Moreira AF
      Diabetic cardiomyopathy: understanding the molecular and cellular basis to progress in diagnosis and treatment.
      ,
      • Low Wang CC
      • Hess CN
      • Hiatt WR
      • Goldfine AB
      Clinical update: cardiovascular disease in diabetes mellitus: atherosclerotic cardiovascular disease and heart failure in type 2 diabetes mellitus - mechanisms, management, and clinical considerations.
      The increased production of advanced glycation end products and its receptors has also been shown to contribute to atrial fibrosis.
      • Kato T
      • Yamashita T
      • Sekiguchi A
      • et al.
      AGEs-RAGE system mediates atrial structural remodeling in the diabetic rat.
      Additionally, endothelial damage and decreased nitric oxide bioavailability leads to endothelial dysfunction, which, in turn, results in a decreased coronary blood flow reserve, with ensuing cardiac hypertrophy, myocardial stiffening, and diastolic dysfunction.
      • Seddon M
      • Looi YH
      • Shah AM
      Oxidative stress and redox signalling in cardiac hypertrophy and heart failure.
      Diastolic dysfunction can lead to left atrial dilation, another substrate of atrial fibrillation.
      • Tiwari S
      • Schirmer H
      • Jacobsen BK
      • et al.
      Association between diastolic dysfunction and future atrial fibrillation in the Tromsø Study from 1994 to 2010.
      Another pathway is related to electromechanical remodeling. Indeed, diabetes-related microvascular disease is associated with atrial fibrosis, which has been linked to prolonged atrial conduction times, resulting in reentry,
      • Schotten U
      • Dobrev D
      • Platonov PG
      • Kottkamp H
      • Hindricks G
      Current controversies in determining the main mechanisms of atrial fibrillation.
      decreases in left atrial emptying volume and emptying fraction, as well as higher interatrial and intra-atrial electromechanical delays.
      • Wang A
      • Green JB
      • Halperin JL
      • Piccini JPS
      Atrial fibrillation and diabetes mellitus: JACC review topic of the week.
      ,
      • Mahajan R
      • Lau DH
      • Brooks AG
      • et al.
      Electrophysiological, electroanatomical, and structural remodeling of the atria as consequences of sustained obesity.
      Another potential mechanism might be linked to autonomic remodeling.
      • Otake H
      • Suzuki H
      • Honda T
      • Maruyama Y
      Influences of autonomic nervous system on atrial arrhythmogenic substrates and the incidence of atrial fibrillation in diabetic heart.
      Cardiac autonomic dysfunction is a known complication of diabetes, characterized by sympathetic denervation, exhaustion of cardiac catecholamine, and impairment in cardiac nerve fibers. The imbalance between parasympathetic and sympathetic may promote the development of atrial fibrillation.
      • Wang A
      • Green JB
      • Halperin JL
      • Piccini JPS
      Atrial fibrillation and diabetes mellitus: JACC review topic of the week.
      Our findings have important implications for individuals with type 2 diabetes. Our study confirms the relevance of microvascular disease in the pathogenesis of diabetes-related atrial fibrillation and points to the potential utility of microvascular disease in atrial fibrillation risk prediction, as illustrated by the inclusion of kidney disease in a recently reported electronic health record atrial fibrillation score that improved the prediction of atrial fibrillation compared with other risk tools that did not include microvascular disease components such as the Cohorts for Heart and Aging Research in Genomic Epidemiology Atrial Fibrillation (CHARGE-AF) score.
      • Khurshid S
      • Kartoun U
      • Ashburner JM
      • et al.
      Performance of atrial fibrillation risk prediction models in over 4 million individuals.
      Further research is needed to investigate the potential pathways linking microvascular disease to atrial fibrillation, and to further explore the incremental predictive value of microvascular disease for the risk of atrial fibrillation.
      This study has several strengths. First, we used a prospective design leveraging a large sample of adults with type 2 diabetes among whom microvascular disease was assessed in multiple vascular territories. Second, neuropathy was assessed using the MNSI, a validated instrument that has been shown to have a good performance at detecting diabetic peripheral neuropathy;
      • Herman WH
      • Pop-Busui R
      • Braffett BH
      • et al.
      Use of the Michigan Neuropathy Screening Instrument as a measure of distal symmetrical peripheral neuropathy in Type 1 diabetes: results from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications.
      and retinopathy assessed using retinal photography in a subset of individuals. Finally, the adjudication of atrial fibrillation events was standardized using central ECG core analysis, and we conducted robust adjustments for relevant cofounders.
      The limitations of our study should be acknowledged. First, although ECGs were performed at prespecified study visits (at enrollment, every 24 months, and study completion), continuous ECG monitoring was not available in ACCORD. Because atrial fibrillation was not an endpoint in ACCORD, and medical records were not reviewed as part of this analysis, it is possible that we missed cases of atrial fibrillation occurring between visits, such as paroxysmal atrial fibrillation, for example. This may have led to nondifferential misclassification of the outcome with a resulting bias toward the null. Additionally, we could not assess the effect of microvascular disease on atrial fibrillation phenotypes such as paroxysmal, persistent, or permanent atrial fibrillation. Second, echocardiography was not performed as part of ACCORD; therefore, we could not account for left atrial diameter or volume index in our analyses. Third, the small number of atrial fibrillation cases, especially in the subset of participants with retinal photography, may have limited our statistical power to detect differences. Fourth, ACCORD did not assess the microvasculature in the coronary bed using, for example, cardiac positron emission tomography, which could have provided more direct evidence of the impact of myocardial microvascular dysfunction on the atrium and thus, atrial fibrillation. Finally, this study used an observational design; hence, the possibility of residual confounding remains.
      In conclusion, in a large cohort of adults with type 2 diabetes, we observed that the presence and burden of microvascular disease, assessed in multiple vascular beds, are associated with a greater risk of incident atrial fibrillation, independently of other atrial fibrillation risk factors. Our findings support the notion that microvascular disease may be involved in the pathogenesis of atrial fibrillation in type 2 diabetes.

      Data Statement

      The data for these analyses were made available to the authors by a third party. As a result, the authors cannot provide the raw data themselves.

      Acknowledgements

      The authors wish to thank the staff and participants of the ACCORD (Action to Control Cardiovascular Risk in Diabetes) study for their valuable contributions. The ACCORD study has been funded by federal funds from the National Heart Lung and Blood Institute (NHLBI). This manuscript was prepared using ACCORD Research Material obtained from the NHLBI Biologic Specimen and Data Repository Information Coordinating Center, and does not necessarily reflect the opinions or views of the ACCORD or the NHLBI.

      Supplementary Data

      Supplementary data to this article can be found online at https://doi.org/10.1016/j.amjmed.2022.04.012.

      References

        • Go AS
        • Hylek EM
        • Phillips KA
        • et al.
        Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.
        JAMA. 2001; 285: 2370-2375
        • Huxley RR
        • Filion KB
        • Konety S
        • Alonso A
        Meta-analysis of cohort and case-control studies of type 2 diabetes mellitus and risk of atrial fibrillation.
        Am J Cardiol. 2011; 108: 56-62
        • Dublin S
        • Glazer NL
        • Smith NL
        • et al.
        Diabetes mellitus, glycemic control, and risk of atrial fibrillation.
        J Gen Intern Med. 2010; 25: 853-858
        • Van Staa TP
        • Setakis E
        • Di Tanna GL
        • Lane DA
        • Lip GYH
        A comparison of risk stratification schemes for stroke in 79,884 atrial fibrillation patients in general practice.
        J Thromb Haemost. 2011; 9: 39-48
        • Echouffo-Tcheugui JB
        • Shrader P
        • Thomas L
        • et al.
        Care patterns and outcomes in atrial fibrillation patients with and without diabetes: ORBIT-AF Registry.
        J Am Coll Cardiol. 2017; 70: 1325-1335
        • Plitt A
        • McGuire DK
        • Giugliano RP
        Atrial fibrillation, type 2 diabetes, and non-vitamin K antagonist oral anticoagulants: a review.
        JAMA Cardiol. 2017; 2: 442-448
        • Kato T
        • Yamashita T
        • Sekiguchi A
        • et al.
        AGEs-RAGE system mediates atrial structural remodeling in the diabetic rat.
        J Cardiovasc Electrophysiol. 2008; 19: 415-420
        • Kume O
        • Teshima Y
        • Abe I
        • et al.
        Role of atrial endothelial cells in the development of atrial fibrosis and fibrillation in response to pressure overload.
        Cardiovasc Pathol. 2017; 27: 18-25
        • Seyed Ahmadi S
        • Svensson A-M
        • Pivodic A
        • Rosengren A
        • Lind M
        Risk of atrial fibrillation in persons with type 2 diabetes and the excess risk in relation to glycaemic control and renal function: a Swedish cohort study.
        Cardiovasc Diabetol. 2020; 19: 9
        • von Elm E
        • Altman DG
        • Egger M
        • Pocock SJ
        • Gøtzsche PC
        • Vandenbroucke JP
        The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.
        Ann Intern Med. 2007; 147: 573-577
        • Buse JB
        • Bigger JT
        • Byington RP
        • et al.
        Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods.
        Am J Cardiol. 2007; 99: 21i-33i
        • Chew EY
        • Ambrosius WT
        • Howard LT
        • et al.
        Rationale, design, and methods of the Action to Control Cardiovascular Risk in Diabetes Eye Study (ACCORD-EYE).
        Am J Cardiol. 2007; 99: 103i-111i
        • Ismail-Beigi F
        • Craven T
        • Banerji MA
        • et al.
        Effect of intensive treatment of hyperglycaemia on microvascular outcomes in type 2 diabetes: an analysis of the ACCORD randomised trial.
        Lancet. 2010; 376: 419-430
        • Levey AS
        • Stevens LA
        • Schmid CH
        • et al.
        A new equation to estimate glomerular filtration rate.
        Ann Intern Med. 2009; 150: 604-612
        • Herman WH
        • Pop-Busui R
        • Braffett BH
        • et al.
        Use of the Michigan Neuropathy Screening Instrument as a measure of distal symmetrical peripheral neuropathy in Type 1 diabetes: results from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications.
        Diabet Med. 2012; 29: 937-944
      1. Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12. Early Treatment Diabetic Retinopathy Study Research Group.
        Ophthalmology. 1991; 98: 823-833
        • Prineas RJ
        • Crow RS
        • Zhang Z-M
        The Minnesota Code Manual of Electrocardiographic Findings.
        Springer Science & Business Media, Heidelberg, Germany2009
        • Bansal N
        • Zelnick LR
        • Alonso A
        • et al.
        eGFR and albuminuria in relation to risk of incident atrial fibrillation: a meta-analysis of the Jackson Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Cardiovascular Health Study.
        Clin J Am Soc Nephrol. 2017; 12: 1386-1398
        • Molnar AO
        • Eddeen AB
        • Ducharme R
        • et al.
        Association of proteinuria and incident atrial fibrillation in patients with intact and reduced kidney function.
        J Am Heart Assoc. 2017; 6e005685
        • Bonapace S
        • Rossi A
        • Lipari P
        • et al.
        Relationship between increased left atrial volume and microvascular complications in patients with type 2 diabetes.
        J Diabetes Complications. 2015; 29: 822-828
        • Kadappu KK
        • Boyd A
        • Eshoo S
        • et al.
        Changes in left atrial volume in diabetes mellitus: more than diastolic dysfunction?.
        Eur Heart J Cardiovasc Imaging. 2012; 13: 1016-1023
        • Agarwal SK
        • Norby FL
        • Whitsel EA
        • et al.
        Cardiac autonomic dysfunction and incidence of atrial fibrillation: results from 20 years follow-up.
        J Am Coll Cardiol. 2017; 69: 291-299
        • Lee S-R
        • Choi E-K
        • Rhee T-M
        • et al.
        Evaluation of the association between diabetic retinopathy and the incidence of atrial fibrillation: a nationwide population-based study.
        Int J Cardiol. 2016; 223: 953-957
        • Phan K
        • Mitchell P
        • Liew G
        • et al.
        Relationship between macular and retinal diseases with prevalent atrial fibrillation - an analysis of the Australian Heart Eye Study.
        Int J Cardiol. 2015; 178: 96-98
        • Falcão-Pires I
        • Leite-Moreira AF
        Diabetic cardiomyopathy: understanding the molecular and cellular basis to progress in diagnosis and treatment.
        Heart Fail Rev. 2012; 17: 325-344
        • Low Wang CC
        • Hess CN
        • Hiatt WR
        • Goldfine AB
        Clinical update: cardiovascular disease in diabetes mellitus: atherosclerotic cardiovascular disease and heart failure in type 2 diabetes mellitus - mechanisms, management, and clinical considerations.
        Circulation. 2016; 133: 2459-2502
        • Seddon M
        • Looi YH
        • Shah AM
        Oxidative stress and redox signalling in cardiac hypertrophy and heart failure.
        Heart. 2007; 93: 903-907
        • Tiwari S
        • Schirmer H
        • Jacobsen BK
        • et al.
        Association between diastolic dysfunction and future atrial fibrillation in the Tromsø Study from 1994 to 2010.
        Heart. 2015; 101: 1302-1308
        • Schotten U
        • Dobrev D
        • Platonov PG
        • Kottkamp H
        • Hindricks G
        Current controversies in determining the main mechanisms of atrial fibrillation.
        J Intern Med. 2016; 279: 428-438
        • Wang A
        • Green JB
        • Halperin JL
        • Piccini JPS
        Atrial fibrillation and diabetes mellitus: JACC review topic of the week.
        J Am Coll Cardiol. 2019; 74: 1107-1115
        • Mahajan R
        • Lau DH
        • Brooks AG
        • et al.
        Electrophysiological, electroanatomical, and structural remodeling of the atria as consequences of sustained obesity.
        J Am Coll Cardiol. 2015; 66: 1-11
        • Otake H
        • Suzuki H
        • Honda T
        • Maruyama Y
        Influences of autonomic nervous system on atrial arrhythmogenic substrates and the incidence of atrial fibrillation in diabetic heart.
        Int Heart J. 2009; 50: 627-641
        • Khurshid S
        • Kartoun U
        • Ashburner JM
        • et al.
        Performance of atrial fibrillation risk prediction models in over 4 million individuals.
        Circ Arrhythm Electrophysiol. 2021; 14e008997