Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis Authors: Shunsuke Kobayashi, Mototsugu Nagao, Izumi Fukuda, Shinichi Oikawa, Masato Iwabu Published: 2026 ABSTRACT AbstractBackground:A prolonged QT interval is a recognized risk factor for life-threatening arrhythmias and sudden cardiac death. In patients with type 2 diabetes, identifying clinical markers associated with myocardial repolarization abnormalities is essential for risk stratification.Objectives:To investigate the association between multiple non-invasive atherosclerosis indicators and corrected QT interval prolongation in patients with type 2 diabetes.Design:Retrospective cross-sectional analysis.Methods:We evaluated 96 patients with type 2 diabetes admitted for glycemic control. The QTc interval was calculated using the Bazett (QTcB), Fridericia (QTcF), and Framingham (QTcFra) formulas. We examined the association between QTc and multiple atherosclerosis indicators, including the ankle-brachial index (ABI), cardio-ankle vascular index (CAVI), carotid intima-media thickness (IMT), and flow-mediated dilatation (FMD).Results:Among all participants, 2 women (7%) and 10 men (15%) had prolonged QTcB. In univariate analyses, FMD was negatively correlated with QTcB (r = −0.28, p = 0.0054), whereas ABI, CAVI, mean, and maximum carotid IMT did not exhibit significant associations. In multivariable linear regression analyses, FMD was independently associated with QTcB (unstandardized coefficient β = −1.95 ms per 1% increase in FMD; 95% confidence interval −3.60 to −0.30; p = 0.021). Subgroup analysis revealed that this association was prominent specifically in the non-obese group (body mass index, <25 kg/m²). Notably, these findings remained consistent regardless of the correction formula used.Conclusion:Vascular endothelial dysfunction, as measured by FMD, is independently associated with QTc prolongation, particularly in non-obese patients with type 2 diabetes. This suggests a significant link that may help identify individuals at higher risk for life-threatening arrhythmias.Keywords: atherosclerosis, corrected QT interval, flow-mediated dilatation, type 2 diabetes mellitus Introduction Type 2 diabetes is associated with an increased risk of death from cardiovascular disease and sudden cardiac death due to pathophysiological alterations in the coronary vasculature, myocardium, and cardiac electrophysiology.1–7 Among them, the QT interval, which defines the period between ventricular depolarization and repolarization, reflects the action potential duration (APD) of cardiomyocytes and serves as a critical window for arrhythmogenic events. Prolongation of the corrected QT interval (QTc) leads to life-threatening arrhythmias, particularly torsades de pointes-type ventricular tachycardia, which increases the risk of sudden cardiac death. 8 This condition is more common in individuals with diabetes than in those without.9–11 Emerging evidence underscores the prognostic significance of prolonged QTc as an independent predictor of both all-cause and cardiovascular mortality in type 2 diabetes.12,13 Hence, it is important to clarify the patient characteristics that influence QTc to reduce mortality in type 2 diabetes. Several studies have suggested that hypoglycemia per se prolongs QTc and is likely to lead to fatal arrhythmias and sudden death in patients with diabetes.14–16 In patients with type 2 diabetes, hyperglycemia,17,18 serum insulin concentration, 19 hypertension,10,20 diabetic microvascular complications, 21 and coronary heart disease22,23 have been reported to be associated with the QTc prolongation. In addition, obesity and insulin resistance have been demonstrated to be associated with cardiac repolarization abnormalities.10,19,24 These factors represent important clinical considerations in the evaluation of the QTc interval in the diabetic population. Since myocardial microischemia has been suggested as a major cause of QTc prolongation in diabetic patients,25,26 the relationship between atherosclerosis indicators and QT interval has been investigated. In clinical practice, ultrasonography, ankle-brachial index (ABI), brachial-ankle pulse-wave velocity (baPWV), cardio-ankle vascular index (CAVI), and flow-mediated dilatation (FMD) are employed as non-invasive assessments of atherosclerosis. 27 Furthermore, cardiac autonomic neuropathy (CAN), a common complication of diabetes, is known to influence both cardiac repolarization and vascular regulation.28–30 Since CAN leads to QTc prolongation via sympathovagal imbalance and exacerbates endothelial dysfunction, 31 it represents a potential shared determinant linking these two pathological states. While baPWV 32 and carotid intima-media thickness (IMT) 33 have each been reported to be correlated with QTc prolongation, few studies have focused on FMD, which detects earlier stages of atherosclerosis. 34 Furthermore, few studies have comprehensively and directly compared these functional and structural atherosclerosis indicators within a single cohort of patients with type 2 diabetes to determine which marker is most closely associated with QTc prolongation. In the present study, we therefore retrospectively analyzed the relationship between QTc and the results of multiple atherosclerosis examinations, including ABI, CAVI, carotid IMTs, and FMD in Japanese patients with type 2 diabetes. We found that, among the atherosclerosis indicators examined, QTc was significantly associated only with FMD. Materials and methods Participants We conducted a retrospective cross-sectional survey of medical records of patients with type 2 diabetes who were admitted to Nippon Medical School Hospital (Tokyo, Japan) for glycemic control from April 2009 to March 2013. We screened patients who underwent electrocardiogram (ECG) and atherosclerosis examinations, including ABI, CAVI, carotid IMT, and FMD. Exclusion criteria were patients with diabetic ketosis, type 1 diabetes, medications that could affect QTc, such as antiarrhythmic agents, antipsychotic agents, tricyclic antidepressants, tetracyclic antidepressants, α-blockers, β-blockers, digoxin, antifungal agents, and antibacterial agents. We also excluded patients with insulin therapy before admission, atrial fibrillation, infectious disease, or malignant tumors under treatment. The study protocol was registered with the University Hospital Medical Information Network (UMIN; No. UMIN000040833), approved by the institutional review board (R1-07-1159), and conformed to the provisions of the Declaration of Helsinki in 1995 (as revised in Edinburgh 2000). Clinical measurements We collected information on physical examinations, including height, body weight, and blood pressure, as well as laboratory data obtained early in the morning fasting state on the day after hospitalization. Obesity was defined as a body mass index (BMI) ⩾25 kg/m2. Fasting plasma glucose was measured using the glucose oxidase method (ADAMS Glucose GA-1170; Arkray, Kyoto, Japan). Glycated hemoglobin (HbA1c) was measured using high-performance liquid chromatography (ADAMS A1c HA-8160; Arkray) and is expressed as the percentage value of the National Glycohemoglobin Standardization Program according to the Japan Diabetes Society guideline. 35 The serum calcium level was corrected for the concentration of albumin using the following formula if the serum albumin level was <4.0 g/dL: corrected Ca (mg/dL) = Ca (mg/dL) − (4.0 − albumin (g/dL)). 36 Estimated glomerular filtration rate (eGFR) was calculated using the following formula: eGFR (mL/min/1.73 m2) = 194 × Cr−1.094 × Age−0.287 (×0.742 if female). 37 Corrected QT interval We analyzed 12-lead ECG that were recorded on the day of admission using an electrocardiograph (ECG-1550; Nihon Kohden Co., Ltd., Tokyo, Japan). We printed the recorded ECG and removed patient information from the recorded ECGs for blinding. QT and RR intervals of three consecutive beats were measured manually on lead II.38,39 The QT intervals were defined as the length from the beginning of the QRS complex to the end of the downslope of the T wave. A QTc was calculated according to the Bazett formula: QTcB = QT/(RR)1/2, 40 Fridericia formula: QTcF = QT/(RR)1/3, 41 Framingham formula: QTcFra = QT + 0.154 (1 − RR). 42 In accordance with the AHA/ACCF/HRS recommendations, QTc prolongation was defined as 460 ms or longer for women and 450 ms or longer for men. 43 ABI and CAVI ABI and CAVI were evaluated using VaSeraVS-1500 (Fukuda Denshi, Co., Ltd., Tokyo, Japan) as previously described. 44 ABI and CAVI were measured bilaterally on the right and left sides, and the lower value for ABI and the higher value for CAVI on both sides were used for analysis. Carotid ultrasonography Carotid IMT was measured with high-resolution B-mode ultrasound systems (SDU-2000; Shimadzu, Kyoto, Japan; iU22 and EnVisor; Philips Healthcare, Andover, MA, USA; LOGIQ 7; GE Healthcare, Little Chalfont, UK) equipped with linear transducers with a frequency of 3–12 MHz as previously described. 45 Patients were examined in the supine position. Mean IMT was defined as the mean of the IMTs at three points: the thickest point on the far wall in the carotid artery, 1 cm on the proximal side, and 1 cm on the distal side from the thickest point on long-axis scans. 46 Max IMT was defined as the thickest IMT in the carotid arteries on short-axis scans. All measurements were performed by well-trained physicians. The higher values for mean IMT and max IMT among both sides were used for analysis. Flow-mediated dilatation FMD was measured in an overnight fasting state using an ultrasonography device with one long-axis probe and two short-axis probes (UNEXEF 38G; UNEX Corporation, Nagoya, Japan). Patients were rested in the supine position for at least 5 min in a temperature-controlled room (22–25°C). Resting blood pressure was measured on the left arm using a sphygmomanometer. An occlusion cuff was placed on the right forearm. The ultrasound probe was placed 5–10 cm proximally to the elbow with a probe holder, and the resting diameter of the right brachial artery was measured. The right brachial artery was occluded for 5 min by pressurizing the cuff 50 mmHg higher than the systolic pressure measured in the left arm. After 5 min, the cuff was decompressed to release occlusion, and the diameter of the right brachial artery was tracked for 2 min. We recorded the maximum diameter of the right brachial artery after 5 min of occlusion as max diameter and calculated %FMD as ((max diameter − resting diameter)/resting diameter) × 100. 34 Assessment of diabetic microvascular complications Neuropathy was screened with criteria for typical diabetic peripheral neuropathy proposed by the Toronto Neuropathy Expert Group. 28 The presence of typical diabetic peripheral neuropathic symptoms (decreased sensation, positive neuropathic sensory symptoms (e.g., numbness, prickling or stabbing, burning, or aching pain) in lower legs) or abnormal (decreased or absent) Achilles tendon reflex was diagnosed as neuropathy (including possible neuropathy). Retinopathy was diagnosed using ophthalmoscopy by an ophthalmologist. 47 Nephropathy was diagnosed by the presence of albuminuria (urinary albumin excretion ⩾30 mg/g·Cr (spot)). 48 Statistical analysis Participants with missing values for any of the required variables were excluded from the analysis. Data distribution of continuous variables was assessed for normality using the Shapiro–Wilk test and visual inspection of Q–Q plots. Continuous variables were described as the mean ± standard deviation for normally distributed variables and median and interquartile range for non-normally distributed variables. Categorical variables were expressed as numbers and percentages. Correlations between QTc and other continuous variables were examined using Pearson’s correlation test for normally distributed variables and Spearman’s rank correlation test for non-normally distributed variables. Differences in QTc between patients with and without microvascular complications were analyzed by Student’s t-test. Multiple regression analyses were performed to determine statistically significant predictive factors for QTc. A p value <0.05 was considered significant. All analyses were performed using JMP Pro 14.10 software (SAS Institute Inc., Cary, NC, USA). Reporting method This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist for cross-sectional studies (Supplemental Material). 49 Results We screened 135 patients who underwent ECG and atherosclerosis examinations. We excluded 6 patients with diabetic ketosis, 2 with type 1 diabetes, 16 with medications affecting the QT interval, 11 with insulin therapy before admission, 1 with atrial fibrillation, 1 with infection, and 2 with missing data. Clinical characteristics at admission of the participants are shown in Table 1. Symptomatic hypoglycemic episodes were not observed in any participants prior to or during the ECG recordings. Two patients had a history of myocardial infarction. The antidiabetic medications that participants received at admission were sulfonylurea in 32, metformin in 16, α-glucosidase inhibitor in 11, dipeptidyl peptidase-4 inhibitor in 10, thiazolidine in 4, and glinide in 3 cases. Patients taking angiotensin receptor blockers, calcium channel blockers, and statins were 22, 17, and 12 cases, respectively. Comparison of serum electrolyte levels between participants with and without diabetic nephropathy revealed no statistically significant differences in sodium (139 ± 3 vs 139 ± 3, p = 0.31), potassium (4.3 ± 0.5 vs 4.2 ± 0.3, p = 0.35), chloride (104 ± 3 vs 103 ± 3, p = 0.56), and corrected calcium levels (9.3 ± 0.6 vs 9.4 ± 0.4, p = 0.36). Regarding the screening for atherosclerosis indicators, there were no participants with ABI greater than 1.30, which is suggestive of advanced arterial calcification. Mean QTcB among all participants was 430 ± 21 ms, with prolongation identified in 2 (7%) women and 10 (15%) men. Using the QTcF and QTcFra formulas, the mean values were 415 ± 19 ms; prolongation was observed in 0 (0%) women and 2 (3%) men for both correction methods. Table 1.Clinical characteristics of the participants.Variablesn = 96Age (years)56 ± 12Female/male (n)28/68Duration of type 2 diabetes (years)5 (1–10)BMI (kg/m2)25.0 ± 4.7Systolic blood pressure (mmHg)129 ± 19Diastolic blood pressure (mmHg)74 ± 12Fasting plasma glucose (mg/dL)188 ± 52Fasting C-peptide (ng/mL)1.9 ± 0.9GA (%)29.5 ± 9.7HbA1c (%)10.3 ± 2.0Serum sodium (mEq/L)139 ± 3Serum potassium (mEq/L)4.2 ± 0.4Serum chloride (mEq/L)103 ± 3Corrected serum calcium (mg/dL)9.3 ± 0.5eGFR (mL/min/1.73 m2)79.8 ± 25.1Microvascular complications Neuropathy (n (%))46 (48) Retinopathy (n (%))24 (25) Nephropathy (n (%))39 (41)Current smoking (n (%))39 (41)ABI1.10 ± 0.09CAVI8.6 ± 1.3Carotid mean IMT (mm)0.79 ± 0.27Carotid max IMT (mm)0.92 ± 0.41FMD (%)4.7 ± 2.3Heart rate (bpm)74 ± 13QTcB (ms)430 ± 21QTcF (ms)415 ± 19QTcFra (ms)415 ± 19Continuous variables are presented as mean ± SD for normally distributed variables and as median (IQR) for non-normally distributed variables.ABI, ankle-brachial index; BMI, body mass index; CAVI, cardio-ankle vascular index; eGFR, estimated glomerular filtration rate; FMD, flow-mediated dilatation; GA, glycated albumin; HbA1c, glycated hemoglobin; IMT, intima-media thickness; IQR, interquartile range; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; QTcF, QTc corrected by Fridericia formula; QTcFra, QTc corrected by Framingham formula; SD, standard deviation. Clinical characteristics of the participants. Continuous variables are presented as mean ± SD for normally distributed variables and as median (IQR) for non-normally distributed variables. ABI, ankle-brachial index; BMI, body mass index; CAVI, cardio-ankle vascular index; eGFR, estimated glomerular filtration rate; FMD, flow-mediated dilatation; GA, glycated albumin; HbA1c, glycated hemoglobin; IMT, intima-media thickness; IQR, interquartile range; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; QTcF, QTc corrected by Fridericia formula; QTcFra, QTc corrected by Framingham formula; SD, standard deviation. Table 2 shows the correlation between QTc and clinical characteristics. Among the atherosclerosis indicators, FMD was negatively correlated with QTcB (r = −0.28, p = 0.0054; Figure 1), but ABI, CAVI, and carotid artery mean, and max IMT were not significantly correlated with QTcB. We have previously reported that QTc prolongation is associated with the severity of microvascular complications. 21 We therefore analyzed the relationships between QTc and microvascular complications (Table 3). Patients with neuropathy (434 ± 18 vs 425 ± 23, p = 0.036) and nephropathy (437 ± 19 vs 425 ± 21, p = 0.0062) had longer QTcB than those without. No statistically significant differences in QTcF and QTcFra were observed between the two groups with and without microvascular complications. Table 2.Correlations with QTc and clinical characteristics.VariablesQTcBQTcFQTcFraCorrelationp ValueCorrelationp ValueCorrelationp ValueAge (years)0.0910.380.0450.660.0440.67Duration of type 2 diabetes (years)0.150.130.100.340.100.32BMI (kg/m2)0.0520.610.0310.760.0160.88Systolic blood pressure (mmHg)0.0950.290.0460.660.0470.65Diastolic blood pressure (mmHg)0.0810.62−0.0360.73−0.0300.77Fasting plasma glucose (mg/dL)0.0720.49−0.0830.42−0.0800.44Fasting C-peptide (ng/mL)0.00750.94−0.0370.72−0.0300.77GA (%)0.0900.390.0610.550.0590.57HbA1c (%)0.0370.72−0.0100.92−0.0170.87Serum sodium (mEq/L)−0.0400.700.0160.88−0.00110.99Serum potassium (mEq/L)−0.110.28−0.0360.73−0.0200.85Serum chloride (mEq/L)−0.0590.57−0.0440.67−0.0420.68Corrected serum calcium (mg/dL)−0.230.026−0.240.019−0.230.027eGFR (mL/min/1.73 m2)0.180.0840.170.0990.150.14ABI−0.180.0770.0890.390.0960.354CAVI0.160.130.0600.560.0480.64Carotid mean IMT (mm)0.110.280.0590.570.0540.60Carotid max IMT (mm)0.200.0560.120.240.110.27FMD (%)−0.280.0054−0.210.038−0.210.045Heart rate (bpm)0.44<0.0001−0.190.070−0.210.045Correlations between continuous variables and QTc were examined by Pearson’s correlation test for normally distributed variables and Spearman’s rank correlation test for non-normally distributed variables.ABI, ankle-brachial index; BMI, body mass index; CAVI, cardio-ankle vascular index; eGFR, estimated glomerular filtration rate; FMD, flow-mediated dilatation; GA, glycated albumin; HbA1c, glycated hemoglobin; IMT, intima-media thickness; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; QTcF, QTc corrected by Fridericia formula; QTcFra, QTc corrected by Framingham formula. Correlations with QTc and clinical characteristics. Correlations between continuous variables and QTc were examined by Pearson’s correlation test for normally distributed variables and Spearman’s rank correlation test for non-normally distributed variables. ABI, ankle-brachial index; BMI, body mass index; CAVI, cardio-ankle vascular index; eGFR, estimated glomerular filtration rate; FMD, flow-mediated dilatation; GA, glycated albumin; HbA1c, glycated hemoglobin; IMT, intima-media thickness; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; QTcF, QTc corrected by Fridericia formula; QTcFra, QTc corrected by Framingham formula. Correlation with atherosclerosis indicators and QTcB. Correlations between QTcB and (a) ABI, (b) CAVI, (c) mean IMT, (d) max IMT, and (e) FMD were examined using Pearson’s correlation test. ABI, ankle-brachial index; CAVI, cardio-ankle vascular index; FMD, flow-mediated dilatation; IMT, intima-media thickness; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula. Table 3.Diabetic microvascular complications and QTc.Microvascular complicationsPresencen (%)QTcB (ms)p ValueQTcF (ms)p ValueQTcFra (ms)p ValueNeuropathy+46 (48)434 ± 18416 ± 17416 ± 16−50 (52)425 ± 230.036414 ± 210.60414 ± 210.58Retinopathy+24 (25)433 ± 18413 ± 16412 ± 15−72 (75)428 ± 220.32416 ± 200.48416 ± 200.42Nephropathy+39 (41)437 ± 19418 ± 17417 ± 16−57 [59]425 ± 210.0062413 ± 210.26414 ± 200.38QTc are expressed as mean ± standard deviation. Differences between the two groups were analyzed by Student’s t-test.QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; QTcF, QTc corrected by Fridericia formula; QTcFra, QTc corrected by Framingham formula. Diabetic microvascular complications and QTc. QTc are expressed as mean ± standard deviation. Differences between the two groups were analyzed by Student’s t-test. QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; QTcF, QTc corrected by Fridericia formula; QTcFra, QTc corrected by Framingham formula. We performed multiple linear regression analyses to identify predictive factors for QTc. In addition to age and gender, corrected serum calcium, neuropathy, nephropathy, heart rate, and FMD, all of which were significant in univariate analyses, were included in the model. Gender (female), corrected serum calcium, heart rate, and FMD were independently associated with QTcB (unstandardized β = −1.95 ms per 1% increase in FMD, 95% confidence interval (CI) −3.60 to −0.30, p = 0.021; Table 4). The association between FMD and QTc remained consistent when analyzed using QTcF and QTcFra (Supplemental Tables 1–1 and 1-2). Table 4.Multivariable linear regression models for QTcB.Variablesβ coefficient95% CIStandardized βVIFp ValueAge−0.086−0.39 to 0.22−0.0501.120.58Gender (female)6.272.11–10.430.271.130.0036Corrected serum calcium−12.7−20.73 to −4.62−0.281.090.0024Neuropathy0.93−2.86 to 4.730.0441.140.63Nephropathy2.35−1.68 to 6.380.111.240.25Heart rate0.570.29–0.860.361.140.0001FMD−1.95−3.60 to −0.30−0.211.160.021Adjusted R2 = 0.32. Data are expressed as β coefficient, 95% CI, Standardized β, VIF, and p Values for the association of QTcB.CI, confidence interval; FMD, flow-mediated dilatation; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; VIF, variance inflation factors. Multivariable linear regression models for QTcB. Adjusted R2 = 0.32. Data are expressed as β coefficient, 95% CI, Standardized β, VIF, and p Values for the association of QTcB. CI, confidence interval; FMD, flow-mediated dilatation; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; VIF, variance inflation factors. To evaluate the effect of obesity and insulin resistance, participants were stratified into non-obese (BMI <25 kg/m2) and obese (BMI ⩾25 kg/m2) groups for subgroup analyses. Compared with participants without obesity, those with obesity exhibited significantly higher BMI, systolic blood pressure, and fasting C-peptide levels. Furthermore, the prevalence of diabetic neuropathy was significantly higher in the obese group than in the non-obese group (Supplemental Table 2). In the non-obese group, FMD demonstrated significant negative correlation with QTcB, QTcF, and QTcFra (Supplemental Tables 3–1–3-3). In the obese group, no statistically significant differences in QTcB were observed between participants with and without microvascular complications, and correlation with FMD. Conversely, in the non-obese group, those with diabetic nephropathy exhibited significantly prolonged QTcB compared to those without (Supplemental Table 4). Regarding QTcF and QTcFra, no statistically significant differences in QTc were found based on the presence of microvascular complications in either the obese or the non-obese groups. Multivariable linear regression analysis confirmed that FMD was an independent predictor of the QTcB, specifically in the non-obese group (Table 5). These significant associations in the non-obese group remained consistent when QTcF and QTcFra were utilized as the dependent variables (Supplemental Tables 5–1 and 5-2). Table 5.Multivariable linear regression models for QTcB in the non-obese group.Variablesβ coefficient95% CIStandardized βVIFp ValueAge−0.089−0.45 to 0.27−0.0551.220.63Gender (female)5.180.85–9.520.251.080.020Corrected serum calcium−14.4−24.3 to −4.57−0.311.080.0050Heart rate0.600.32–0.890.441.04<0.0001FMD−2.90−4.58 to −1.22−0.391.250.0011Adjusted R2 = 0.44. Data are expressed as β coefficient, 95% CI, Standardized β, VIF, and p Values for the association of QTc.CI, confidence interval; FMD, flow-mediated dilatation; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; VIF, variance inflation factors. Multivariable linear regression models for QTcB in the non-obese group. Adjusted R2 = 0.44. Data are expressed as β coefficient, 95% CI, Standardized β, VIF, and p Values for the association of QTc. CI, confidence interval; FMD, flow-mediated dilatation; QTc, corrected QT interval; QTcB, QTc corrected by Bazett formula; VIF, variance inflation factors. Discussion While previous studies have linked QTc prolongation to arterial stiffness markers such as baPWV and carotid IMT, the role of endothelial dysfunction has remained unclear. To the best of our knowledge, this is the first study to evaluate the relationship between QTc and multiple atherosclerosis indicators, including FMD, in patients with type 2 diabetes. In the present study, we found that among the examined indicators, only FMD was correlated with QTc in univariate analyses, and it was independently associated with QTc in multivariable linear regression analysis in patients with type 2 diabetes. FMD is a well-established marker of vascular endothelial function. Increased blood flow after release of brachial artery occlusion induces shear stress on endothelial cells. Shear stress induced by increased blood flow promotes nitric oxide (NO) production via endothelial nitric oxide synthase (eNOS). NO is synthesized in vascular endothelial cells and diffuses to smooth muscle cells on the vascular wall and activates guanylate cyclase, resulting in vasodilatation.50,51 Therefore, %FMD begins to decrease from the early stage of atherosclerosis due to vascular endothelial dysfunction. 52 Since hyperglycemia per se induces oxidative stress and promotes the formation of advanced glycation end products, which cause vascular endothelial dysfunction,53–55 it is common for patients with type 2 diabetes to have decreased %FMD.56,57 The relationship between FMD and QT interval is not clear, but changes in ionic currents in cardiomyocytes induced by NO could be considered a possible mechanism. NO shortens APD, corresponding to the QT interval, by decreasing the inward Na+ current and by increasing the outward K+ current in the myocardium. 58 These changes in ion current are caused by functional modification of ion channels by cGMP-dependent (protein kinase G and cGMP-regulated phosphodiesterases) and cGMP-independent mechanisms (S-nitrosylation and direct effects on G proteins). 58 Epidemiological studies have reported an association between genetic polymorphisms in nitric oxide synthase 1 adaptor protein (NOS1AP) and QT interval.59,60 NOS1AP gene encodes “carboxy-terminal PDZ ligand of a neuronal nitric oxide synthase (nNOS) protein (CAPON),” a nNOS adaptor protein that regulates the subcellular location of nNOS. In cardiomyocytes overexpressing CAPON, decreased inward Ca current, increased outward K current, and shortened APD were observed. 61 In contrast, a prolonged APD was observed in cardiomyocytes isolated from eNOS knockout mice. 62 These previous studies, which suggest a link between NO and APD, support the finding that FMD was independently associated with QTc among the atherosclerosis indicators. The present study showed a statistically significant correlation and an independent association between QTc and FMD, but not with the other atherosclerosis indicators, ABI, CAVI, or carotid artery IMTs in patients with type 2 diabetes. However, a previous study has reported a positive correlation between QTc and carotid mean IMT in patients with type 2 diabetes. 33 Whereas carotid IMTs are indices that reflect morphological changes due to relatively advanced atherosclerotic lesion formation, FMD is an indicator of vascular endothelial dysfunction in the early stages of atherosclerosis. 52 Such differences in the characteristics of these atherosclerosis indicators may have influenced the present findings regarding associations with QTc. In addition, the discrepancy between the previous study and the present study may be due to differences in the distribution of mean IMT. The mean IMT in the present study was thinner than that in the previous study, suggesting that the participants in the present study were limited to those with less advanced carotid atherosclerosis. CAN represents a plausible common determinant of both QTc prolongation and impaired FMD. In patients with type 2 diabetes, autonomic dysfunction—characterized by attenuated parasympathetic activity and relative sympathetic predominance—exerts substantial effects on cardiac electrophysiology and vascular homeostasis. Specifically, CAN has been implicated in delayed ventricular repolarization and, concurrently, in the attenuation of NO-mediated vasodilation. 63 Considering these overlapping mechanisms, CAN warrants consideration as a potential confounding factor in the observed association between FMD and the QTc interval. Furthermore, the Bazett formula—QT correction formula widely used in clinical practice—is sensitive to heart rate and may misestimate the QTc interval in patients with tachycardia,64,65 a common clinical feature of CAN. To address these potential confounders, heart rate was included as a covariate in the multivariable regression models. In addition, the consistent results obtained from analyses using the Fridericia and Framingham formulas suggest that the observed association between FMD and QTc is not merely an artifact of heart rate or the correction formula used. The observed association between impaired FMD and QTc prolongation likely reflects a complex interplay of several shared pathophysiological mechanisms. Chronic hyperglycemia in type 2 diabetes promotes the overproduction of reactive oxygen species and pro-inflammatory cytokines, such as interleukin-6 and tumor necrosis factor-alpha.53–55 These factors decrease NO bioavailability, leading to endothelial dysfunction as measured by FMD. Concurrently, oxidative stress and inflammation are known to modulate the expression and kinetics of cardiac ion channels—specifically IKr and IKs currents—thereby delaying ventricular repolarization and prolonging the QT interval.66–68 Regarding clinical confounders, diabetic nephropathy and associated electrolyte disturbances, particularly corrected serum calcium, are primary determinants of QT interval. In the present study, the inclusion of corrected serum calcium and nephropathy in the multivariable models, along with the consistency of our findings across different QTc correction formulas, suggests that the link between FMD and QTc is robust and independent of these factors. The data in this study were collected between 2009 and 2013, allowing the characterization of type 2 diabetes pathophysiology prior to the widespread clinical use of SGLT2 inhibitors (SGLT2i) and GLP-1 receptor agonists (GLP-1RAs). SGLT2i and GLP-1RAs are recognized not only for their glucose-lowering effects but also for the prevention of cardiovascular disease.69,70 The mechanisms underlying the cardioprotective effects of SGLT2i are thought to include improvements in vascular endothelial function and antiarrhythmic effects.71,72 In fact, improvements in endothelial function measured by FMD following treatment with dapagliflozin have been reported.73–75 Although SGLT2i and GLP-1RAs may further improve endothelial function and ventricular repolarization, potentially attenuating the clinical burden of QTc prolongation, our findings remain valuable in that they elucidate the pathophysiological relationship between vascular endothelial function and QTc in a cohort unaffected by SGLT2i or GLP-1RAs. The clinical relevance of QTc interval prolongation in the diabetic population is supported by recent outcome-based evidence. Diabetes has been identified as an independent risk factor for both clinically significant QTc prolongation and the occurrence of torsades de pointes.76,77 Our findings, which demonstrate a significant association between impaired FMD and QTc prolongation, are consistent with this established electrical vulnerability. These data suggest that functional endothelial impairment may be intrinsically linked to the heightened arrhythmic risk observed in patients with type 2 diabetes. The subgroup analysis by BMI revealed that the association between impaired FMD and QTc prolongation was prominent in non-obese individuals but absent in those with obesity. This discrepancy may be attributed to the multifactorial nature of QTc prolongation in obese subjects. Obesity is often associated with a clustering of metabolic disorders and chronic low-grade inflammation, both of which are independent drivers of cardiac repolarization abnormalities. 78 It is plausible that in obese patients, the predominant influence of insulin resistance and metabolic clustering overshadows the specific contribution of endothelial dysfunction, as measured by FMD. 78 In contrast, in non-obese individuals where such systemic metabolic interference is less pronounced, the intrinsic link between vascular function and the QTc interval appears more evident. These findings suggest that the clinical utility of FMD as a marker for QTc prolongation may be particularly significant in non-obese diabetic populations. Limitations The present study has several limitations. First, this study is a cross-sectional design and therefore cannot establish a causal relationship between vascular endothelial dysfunction assessed by FMD and QTc prolongation. Second, this study included only patients with type 2 diabetes admitted to our hospital for glycemic control, the majority of whom had poor glycemic control with an HbA1c of approximately 10%. Accordingly, it remains unclear whether the present findings are applicable to patients with type 2 diabetes who maintain better glycemic control. Third, the participants in this cohort had a relatively low prevalence of advanced atherosclerosis. Fourth, our exclusion criteria of insulin or medications that influence the QT interval have introduced a selection bias toward individuals with earlier-stage type 2 diabetes and lower cardiovascular risk, resulting in a relatively low prevalence of clinically significant QTc prolongation among study participants. Consequently, the generalizability of these findings to patients with more advanced diabetic complications, cardiovascular disease, or those requiring complex pharmacological interventions may be limited. Further validation studies, including patients with varying degrees of glycemic control and atherosclerotic burden, are necessary to confirm the broader clinical utility of these findings and to clarify the relationship between FMD and the QTc interval in patients with type 2 diabetes. Conclusion Vascular endothelial dysfunction, as measured by FMD, is independently associated with QTc prolongation in patients with type 2 diabetes, whereas no significant associations were observed with ABI, CAVI, or carotid IMT. Notably, this association is particularly prominent in the non-obese group and remains consistent across multiple QT correction formulas, suggesting a significant pathophysiological link. These findings highlight the importance of monitoring the QTc interval in patients with type 2 diabetes who exhibit reduced %FMD to help identify individuals at higher risk for life-threatening arrhythmias and cardiac events. Supplemental Material sj-docx-1-tae-10.1177_20420188261448416 – Supplemental material for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis Supplemental material, sj-docx-1-tae-10.1177_20420188261448416 for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis by Shunsuke Kobayashi, Mototsugu Nagao, Izumi Fukuda, Shinichi Oikawa and Masato Iwabu in Therapeutic Advances in Endocrinology and Metabolism sj-docx-2-tae-10.1177_20420188261448416 – Supplemental material for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis Supplemental material, sj-docx-2-tae-10.1177_20420188261448416 for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis by Shunsuke Kobayashi, Mototsugu Nagao, Izumi Fukuda, Shinichi Oikawa and Masato Iwabu in Therapeutic Advances in Endocrinology and Metabolism Acknowledgments All authors thank all team members of the Department of Endocrinology, Metabolism and Nephrology, Graduate School of Medicine, Nippon Medical School. Footnotes ORCID iDs: Shunsuke Kobayashi https://orcid.org/0000-0002-5674-5998 Mototsugu Nagao https://orcid.org/0000-0001-6880-5161 Izumi Fukuda https://orcid.org/0000-0003-0772-3837 Supplemental material: Supplemental material for this article is available online. Contributor Information Shunsuke Kobayashi, Department of Endocrinology, Metabolism and Nephrology, Graduate School of Medicine, Nippon Medical School, 1-1-5 Sendagi, Bunkyo-ku, Tokyo 113-8603, Japan. Mototsugu Nagao, Department of Endocrinology, Metabolism and Nephrology, Graduate School of Medicine, Nippon Medical School, 1-1-5 Sendagi, Bunkyo-ku, Tokyo 113-8603, Japan. Izumi Fukuda, Department of Endocrinology, Metabolism and Nephrology, Graduate School of Medicine, Nippon Medical School, Tokyo, Japan. Shinichi Oikawa, Department of Endocrinology, Metabolism and Nephrology, Graduate School of Medicine, Nippon Medical School, Tokyo, Japan. Masato Iwabu, Department of Endocrinology, Metabolism and Nephrology, Graduate School of Medicine, Nippon Medical School, Tokyo, Japan. Declarations Ethics approval and consent to participate: This study protocol was registered with the University Hospital Medical Information Network (UMIN; No. UMIN000040833), approved by the Ethics Committee of Nippon Medical School Hospital (R1-07-1159), and conformed to the provisions of the Declaration of Helsinki. The Ethics Committee approved the use of an opt-out consent procedure. Patients were informed about the study and given the opportunity to opt out of the research. Consent for publication: Not applicable. Author contributions: Shunsuke Kobayashi: Data curation; Formal analysis; Investigation; Project administration; Writing – original draft. Mototsugu Nagao: Formal analysis; Investigation; Methodology; Supervision; Writing – review & editing. Izumi Fukuda: Investigation; Writing – review & editing. Shinichi Oikawa: Conceptualization; Writing – review & editing. Masato Iwabu: Conceptualization; Writing – review & editing. Funding: The authors received no financial support for the research, authorship, and/or publication of this article. The authors declare that there is no conflict of interest. Availability of data and materials: The data that support the findings of this study are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials sj-docx-1-tae-10.1177_20420188261448416 – Supplemental material for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis Supplemental material, sj-docx-1-tae-10.1177_20420188261448416 for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis by Shunsuke Kobayashi, Mototsugu Nagao, Izumi Fukuda, Shinichi Oikawa and Masato Iwabu in Therapeutic Advances in Endocrinology and Metabolism sj-docx-2-tae-10.1177_20420188261448416 – Supplemental material for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis Supplemental material, sj-docx-2-tae-10.1177_20420188261448416 for Exploring the link between atherosclerosis indicators and QT interval in type 2 diabetes: a retrospective cross-sectional analysis by Shunsuke Kobayashi, Mototsugu Nagao, Izumi Fukuda, Shinichi Oikawa and Masato Iwabu in Therapeutic Advances in Endocrinology and Metabolism