Cardiac Autonomic Function Is Associated With the Coronary Microcirculatory Function in Patients With Type 2 Diabetes

Cardiovascular autonomic neuropathy (CAN) is an often overlooked and severe complication of type 2 diabetes mellitus (T2DM). CAN results from damage to the autonomic nerve fibers that innervate the heart and blood vessels and results in abnormalities in heart rate control and vascular dynamics (1). Early signs of CAN, characterized by increased sympathetic activity and/or decreased parasympathetic activity at rest, are reflected in detrimental changes in indices of heart rate variability (HRV). Failure in autoregulation in response to physical stimuli (e.g., cardiovascular autonomic reflex tests [CARTs]) is seen in later more severe stages of CAN. CAN is strongly associated with cardiovascular morbidity and mortality (1,2).

CARTs, indices of HRV, and cardiac ¹²³I metaiodobenzylguanidine (¹²³I-MIBG) scintigraphy have all been reported to be valid measures of cardiac autonomic function in patients with diabetes (3). ¹²³I-MIBG scintigraphy allows a direct assessment of the integrity of the adrenergic cardiac innervation in contrast to CAN assessment by HRV and CART analyses, which are indirect measures of nervous dysfunction. Thus, ¹²³I-MIBG scintigraphy may be more reliable to evaluate cardiac autonomic function (4). Cardiac ¹²³I-MIBG scintigraphy may also diagnose CAN in early clinical stages before it can be detected by tests that indirectly show the autonomic function of the heart (5).

Impaired cardiac autonomic function has been suggested to promote cardiovascular disease by inducing ventricular arrhythmias and sudden death and by impairing circadian blood pressure fluctuations (6,7). Coronary flow reserve (CFR) is an important physiological variable in the cardiac circulation that reflects the function of large epicardial arteries and the microcirculation. Impaired CFR has been described as a powerful, independent predictor of cardiac mortality among patients with diabetes (8). The coronary artery calcium (CAC) score is known to be highly correlated with the extent of coronary atherosclerosis and can identify asymptomatic patients who are at higher risk for cardiac events and death. The presence of calcium in the coronary arteries is a specific marker of atherosclerosis, independent of its etiology (9).

We undertook a cross-sectional study of patients with T2DM, with or without albuminuria, and age- and sex-matched healthy control subjects without clinical cardiovascular disease. The aims were to determine cardiac autonomic function and the potential association between different measures of cardiac autonomic function and cardiac vascular function assessed by CFR and coronary atherosclerosis assessed by CAC; both measured by cardiac ⁸²Rb-positron emission tomography/computed tomography (PET/CT).

Further, we determined the correlation between cardiac autonomic function assessed by HRV analyses and CARTs and by cardiac ¹²³I-MIBG scintigraphy. We hypothesized that impaired cardiac autonomic function would be associated with lower CFR and higher CAC.

The study population has previously been described (10). In brief, a cohort of 60 consecutive outpatients with T2DM, defined according to the World Health Organization criteria, was identified at Steno Diabetes Center. Participants were aged between 35 and 80 years and had the ability to understand and give informed consent. Patients were stratified as normoalbuminuric if the urinary albumin excretion rate (UAER) was <30 mg/24 h in two of three consecutive urine collections (two of the samples were collected over 24 h in relation to the current study). A priori we included 30 patients with normoalbuminuria and 30 with persistent elevated albuminuria (UAER ≥30 mg/24 h). In addition, 30 control subjects without diabetes were recruited from a newspaper advertisement and matched for age and sex to the 30 patients with normoalbuminuria. We divided participants in these three groups because one of the aims of the overall study was to examine the prevalence of impaired CFR and elevated CAC in 1) patients with T2DM and albuminuria, 2) patients with T2DM and normoalbuminuria, and 3) healthy control subjects (10). In keeping with the original study design, we considered it most appropriate to present this division also in the present report. Participants were excluded if one of the following characteristics was present:

1. history of coronary artery disease or other cardiovascular disease (including stroke) or heart symptoms, assessed from medical records and patient interviews and questionnaires;

2. asthma or chronic obstructive pulmonary disease requiring treatment;

3. history of kidney disease other than diabetic nephropathy;

4. end-stage renal disease;

5. office blood pressure >200/110 mmHg;

6. second- or third-degree atrioventricular block; or

7. pregnancy or lactating.

For the analyses presented in this report, 83 participants were included, with 7 participants (2 control subjects, 2 normoalbuminuric patients, and 3 albuminuric patients) excluded because of incomplete data on the HRV indices (n = 6) or atrial fibrillation (n = 1).

The first 15 participants randomly enrolled in each group were invited to a cardiac ¹²³I-MIBG scintigraphy. All accepted initially, but 1 patient in the albuminuric group was unable to undergo the examination because of a newly diagnosed breast cancer, and 1 control person declined, thus a total of 43 participants were analyzed. The 43 participants with data for cardiac ¹²³I-MIBG scintigraphy and HRV indices did not differ from the 83 patients in the overall population in relation to age (P = 0.89), sex (P = 0.47), heart rate (P = 0.39), smoking (P = 0.52), HbA_1c (P = 0.75), or any of the CARTs (P ≥ 0.10). Treatment with insulin was similar in participants with (n = 43) and without (n = 40) data on cardiac ¹²³I-MIBG scintigraphy (15 vs. 14 patients, P = 0.86).

Power calculation was performed for the primary study (10) and not for the present substudy. On the basis of existing literature, we anticipated that our sample size was sufficient (11,12).

The study was performed in compliance with the Declaration of Helsinki. All participants gave informed written consent, and the study protocol was approved by the Capital Region of Denmark Research Ethics Committee (ref. no. H-3-2013-015).

HbA_1c was measured by high-performance liquid chromatography and plasma creatinine by an enzymatic method (Hitachi 912; Roche Diagnostics, Mannheim, Germany). The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation (13).

Measurement of 24-h blood pressure was conducted using BPro (HealthSTATS, Singapore), a watch-like device that captures radial pulse wave reflection with tonometry and calculates brachial 24-h blood pressure from the pulse wave after calibration to the brachial blood pressure. The device meets the standards from European Society of Hypertension and the Association for the Advancement of Medical Instrumentation (14). The device was programmed to capture blood pressure every 15 min for 24 h. Mean blood pressure was calculated using all readings over the 24-h period. To evaluate the circadian blood pressure fluctuation, we calculated the night-to-day blood pressure ratio.

UAER was measured in two 24-h urine collections by an enzyme immunoassay and calculated as the geometric mean of the two collections.

A detailed medical history was obtained along with demographic and anthropometric variables, including smoking status. Current smoking was defined as one or more cigarettes, cigars, or pipes per day. Information on medical treatment was obtained from questionnaires and cross checked against medical records at the Steno Diabetes Center.

All participants underwent a dynamic, electrocardiogram-gated cardiac PET scan using a hybrid PET/CT scanner in three-dimensional mode (Biograph mCT 128; Siemens, Munich, Germany) after the administration of ∼1,100 MBq ⁸²Rb (Cardiogen·82; Bracco Diagnostics Inc., Monroe Township, NJ). Myocardial blood flow was calculated automatically with the Siemens Syngo MBF 2.3 (Siemens Medical Solutions, Malvern, PA), using one-compartment tracer kinetic models for ⁸²Rb, including regional uptake and clearance parameters, blood to myocardium spillover and partial volume corrections, and the extraction curve from Lortie et al. (15). Maximal hyperemia was induced with adenosine infused at 140 µg/kg/min for 6 min. The CAC score was quantified using the method described by Agatston et al. (16) and semiautomated Corridor4DM software (INVIA, Ann Arbor, MI). CAC scores specific to the three main coronary arteries were calculated and then summed to provide a total CAC score for each participant.

All electrocardiographic signals used for HRV analyses and CARTs were measured by laboratory technicians using the Vagus (Medicus Engineering, Aarhus, Denmark) device, with a sampling frequency of 1,000 Hz. After the subject rested supine for 5 min, 5-min resting heart rate measures for HRV analyses were obtained. From the 5-min resting heart rate recordings, time-domain HRV indices were derived: the SD of normal-to-normal (SDNN) intervals and the root mean square of successive differences (RMSSD). Frequency-domain HRV indices were calculated using fast Fourier transformation and included low-frequency (LF) power (0.04–0.15 Hz), high-frequency (HF) power (0.15–0.4 Hz), and the total power (≤0.4 Hz). The ratio of LF-to-HF power was also calculated. The SDNN is a measure of combined sympathetic and parasympathetic activity, and RMSSD reflects parasympathetic activity. LF power is a measure of sympathetic and parasympathetic tone, and HF power reflects parasympathetic activity. Total power reflects parasympathetic and sympathetic tone.

The CARTs included: 1) response to standing (30-to-15 ratio), as the ratio between the shortest R-R interval (around the 15th heartbeat after the rise) and the longest R-R interval (around the 30th heartbeat); 2) deep breathing (E-to-I ratio) while sitting, expressed as the mean of the longest R-R interval during inhalation divided by the mean of the shortest R-R interval during exhalation over a 1-min deep breathing test; and 3) Valsalva test, as the ratio of the maximum and minimum R-R interval in the 45-s period after a 15-s exhalation while sitting through a 40 mmHg resistance mouthpiece. The E-to-I ratio is a measure of parasympathetic function. The 30-to-15 ratio is largely a measure of parasympathetic function. The Valsalva test is a measure of parasympathetic and sympathetic function (17). The Valsalva test was not performed in patients with laser-treated retinopathy (n = 6).

The CARTs were evaluated according to age-related reference intervals (18), and CAN was defined using the American Diabetes Association criteria (19) as no CAN: no pathological CARTs; borderline CAN: one abnormal CART; definite CAN: two or three abnormal CARTs.

All tests were performed between 8:00 a.m. and 2:00 p.m. in a quiet examination room. A standard protocol was applied in accordance with recommendations (20). Participants were advised to abstain from hard physical activity 24 h before the examination.

On a separate day, within 3 weeks after the cardiac PET/CT scan, a planar cardiac ¹²³I-MIBG scintigraphy was performed. Patients were given 130 mg potassium iodine 1 h before tracer injection and 20 mg potassium iodine 24 h after tracer injection to block thyroid iodine uptake. Approximately 200 MBq of ¹²³I-MIBG was injected intravenously, and planar anterior-posterior images of the chest were obtained 15 min (early) and 240 min (late) after the tracer injection using a Philips SKYLight gamma camera with JETStream software (Philips Medical Systems, Best, the Netherlands), with medium energy collimator, 256 × 256 matrix, acquisition time of 600 s. ¹²³I was imaged with a 15% energy window set symmetrically over the 159-keV photo peak. Image interpretation was done using Extended Brilliance Workspace NM Application Suite V4.5.3.40140 (Philips Medical Systems). One experienced observer assessed the images. A region of interest (ROI) was drawn above the heart, following the epicardial contour, and a rectangular ROI was drawn above the mediastinum on early and late anterior images in accordance with recently published guidelines (21). The mean count within each ROI was reported. The myocardial washout rate from early to late images was calculated according to guidelines of the European Association of Nuclear Medicine Cardiovascular Committee and the European Council of Nuclear Cardiology (21).

Evidence supports the use of the late heart-to-mediastinum ratio for assessment of symptomatic CAN (22,23). An abnormal late heart-to-mediastinum ratio was defined as <1.6 (24).

The distribution of the time- and frequency-domain HRV indices, CAC, UAER, and known duration of diabetes was skewed, and these variables were log2-transformed (CAC+1) in all analyses. These variables are given as medians with the interquartile range. All other continuous variables are given as means ± SD, and the categorical variables are given as total numbers with corresponding percentages. When analyzing differences between two groups (e.g., control subjects vs. normoalbuminuric patients, and normoalbuminuric vs. albuminuric patients), we used the independent-samples t test for continuous and the χ² test for categorical variables. Analysis of covariance was applied when levels of the continuous variables among three groups were compared. We ascertained that the four principal assumptions of linear regression were fulfilled. The proportion of the variability in the dependent variable explained by the model is presented as the R².

All patients were pooled in the linear regression analyses, where we applied stepwise adjustment. First, we used unadjusted models (model 1) to determine whether any association existed among the measures of cardiac autonomic function and CFR as well as CAC. The subsequent adjustments included age (model 2), age and heart rate (model 3), and age, heart rate, and risk factors based on prior evidence (sex, 24-h systolic blood pressure, HbA_1c, UAER, and smoking; model 4). For the measures of CARTs, model 3 was omitted, and model 4 did not include heart rate, because these tests are slightly influenced by the resting heart rate (20).

Owing to bias by indication, we did not include variables for medical treatment. Moreover, total cholesterol was not included because patients had lower levels than control subjects, likely due to lipid-lowering treatment. Standardized regression coefficients were reported. A two-tailed P < 0.05 was considered statistically significant. Statistical analyses were performed using SAS 9.3 software (SAS Institute, Inc., Cary, NC).

The total cohort (n = 83) comprised 35% women, and the mean age was 62.1 ± 9.3 years. The characteristics of the participants in the three groups are reported in Table 1. The normoalbuminuric patients had higher mean heart rate, were more frequently receiving renin-angiotensin-aldosterone (RAAS) system inhibition and lipid-lowering treatment, and had lower total cholesterol compared with the control subjects (P ≤ 0.003). All time- and frequency-domain HRV indices, except the LF-to-HF ratio, were lower in normoalbuminuric patients than in control subjects (P ≤ 0.005). The E-to-I ratio and the late heart-to-mediastinum ratio were also lower in normoalbuminuric patients than in control subjects (P ≤ 0.004). Patients with albuminuria had a lower eGFR, 30-to-15 ratio, and Valsalva test than the normoalbuminuric patients (P ≤ 0.004).

All patients were treated with oral glucose-lowering medication, 49% received insulin, and none received glucagon-like peptide 1 receptor agonists. Most patients received lipid-lowering (93%) and RAAS-blocking (89%) treatment. Treatment with calcium channel blockers was prescribed in 57%, diuretics in 49%, and β-blockers in 11% of the patients; none were treated with α-blockers. Three patients were treated with allopurinol and one with levothyroxine, and no other medications were prescribed.

The late heart-to-mediastinum ratio in the total population was 2.6 ± 0.5. Two participants (5%) had CAN according to a late heart-to-mediastinum ratio of <1.6. On the basis of the CARTs, 6 participants (8%) had definitive CAN, 24 (31%) had borderline CAN, and 47 (61%) had no signs of CAN. One of the two participants with CAN according to late heart-to-mediastinum ratio had CAN based on the CARTs.

In unadjusted analyses (model 1), all measures of cardiac autonomic function, except the LF-to-HF ratio, correlated positively with CFR (P ≤ 0.005) (Table 2). In age-adjusted analyses (model 2), all measures, except the LF-to-HF ratio and the 30-to-15 ratio, remained positively associated with CFR (P ≤ 0.04). In model 3 (adjusted for age and heart rate), the late heart-to-mediastinum ratio and LF power were positively associated with CFR (P ≤ 0.01). After adjustment for additional risk factors (model 4), the late heart-to-mediastinum ratio remained positively associated with CFR (β per 1 SD = 0.53 increase: 0.43; P < 0.001).

The unadjusted correlation of CFR to the late heart-to-mediastinum ratio is illustrated in Fig 1A, total power in Fig. 1B, and Valsalva test in Fig. 1C. The levels of CFR, according to tertiles of the late heart-to-mediastinum ratio, are illustrated in Fig. 2A and absence of CAN, borderline CAN, or definitive CAN, based on the CARTs, is illustrated in Fig. 2B.

In unadjusted analyses (model 1), all measures of cardiac autonomic function, except the late heart-to-mediastinum ratio, were negatively associated with CAC (P ≤ 0.043) (Table 3). In model 2 (age adjusted) all measures, except the late heart-to-mediastinum ratio, HF power, 30-to-15 ratio, and the Valsalva test, were negatively associated with CAC (P ≤ 0.027). Further adjustment for heart rate (model 3) of the time- and frequency-domain HRV indices did not alter significance. However, after adjustment for additional risk factors (model 4), none of the measures of cardiac autonomic function were associated with CAC (P ≥ 0.072).

In unadjusted analyses, the late heart-to-mediastinum ratio correlated positively with the time- and frequency-domain HRV indices (P ≤ 0.004), except for the LF-to-HF ratio (P = 0.17). For the CARTs, the late heart-to-mediastinum ratio correlated positively with the 30-to-15 ratio (P = 0.04) but not with the E-to-I ratio or the Valsalva test (P ≥ 0.43). After adjustment for age, heart rate (only for the time- and frequency-domain HRV indices), sex, 24-h systolic blood pressure, HbA_1c, UAER, and smoking, the late heart-to-mediastinum ratio correlated positively with LF power (P = 0.002), SDNN (P = 0.049), and RMSSD (P = 0.037), but not with any of the other HRV measures or the CARTs (P ≥ 0.07).

Analyses only including the patients with diabetes (HRV: n = 55; ¹²³I-MIBG scintigraphy: n = 29) revealed confirmatory results: CFR was positively correlated with the late heart-to-mediastinum ratio (R² = 0.19; P = 0.017), total power (R² = 0.14; P = 0.01), and the Valsalva test (R² = 0.26; P < 0.001). In adjusted analyses (model 4), the late heart-to-mediastinum ratio correlated positively with CFR (P < 0.001), but the correlation between CFR and the other measures of cardiac autonomic function lost significance (P ≥ 0.13).

The washout rate was not correlated with CFR in unadjusted (P = 0.92) or adjusted analyses (P = 0.29). A lower late heart-to-mediastinum ratio correlated significantly with lower early heart-to-mediastinum ratio (P < 0.001) but not with a higher washout rate (P = 0.06). The systolic night-to-day blood pressure ratio was not correlated with CFR in unadjusted (P = 0.08) or adjusted analyses (P = 0.69).

To avoid the potential confounding effect of treatment with β-blocker medications on measures of cardiac autonomic function, we performed a sensitivity analysis including only the participants without β-blocker treatment (HRV: n = 77; ¹²³I-MIBG: n = 40). Results were confirmatory: CFR was positively correlated with the late heart-to-mediastinum ratio (R² = 0.33; P < 0.001), total power (R² = 0.24; P < 0.001), and the Valsalva test (R² = 0.10; P = 0.013). In adjusted analyses (model 4), the late heart-to-mediastinum ratio correlated positively with CFR (P = 0.0002), but the correlation between CFR and the other measures of cardiac autonomic function lost significance (P ≥ 0.22).

In this cross-sectional study of patients with T2DM without clinical cardiovascular disease, we demonstrate that impaired function of the cardiac autonomic system correlated with lower CFR measured with cardiac ⁸²Rb-PET/CT. Especially, the late heart-to-mediastinum ratio, assessed by cardiac ¹²³I-MIBG scintigraphy, and LF power were associated with CFR. We further found agreement between the measures of cardiac autonomic function. In this cohort of asymptomatic patients with T2DM, CAN was present in 7%, defined by a late heart-to-mediastinum ratio <1.6 (24) and in 11% according to the American Diabetes Association criteria based on CARTs (19). We demonstrated cardiac autonomic function, assessed by cardiac ¹²³I-MIBG scintigraphy (the late heart-to-mediastinum ratio) and by HRV indices, was lower in patients with T2DM compared with control subjects without diabetes. However, only the 30-to-15 ratio and the Valsalva test were lower in albuminuric patients compared with normoalbuminuric patients.

CAN is an overlooked and serious complication associated with increased risk of cardiovascular morbidity and mortality, including cardiac arrhythmias and sudden death. In the Detection of Ischemia in Asymptomatic Diabetics study of 1,123 patients with T2DM, CAN, assessed by the 30-to-15 ratio, was a strong predictor of silent ischemia (evaluated with adenosine-stress myocardial perfusion imaging) and cardiovascular events (25).

The CARTs are validated and recommended by American Diabetes Association for the diagnosis of CAN. These tests are simple and can be performed in the practitioner’s office. In this study, we additionally included time and frequency HRV indices to acquire more comprehensive information on the tone of the autonomic nervous system (26). The Atherosclerosis Risk in Communities Study found lower HRV was associated with an increased risk of incident coronary heart disease during an average follow-up of >8 years among patients with T2DM but not in individuals without diabetes at baseline (27).

Cardiac radionuclide imaging enables a direct quantification of cardiac sympathetic innervation in various diseases, including CAN (6). Cardiac ⁸²Rb PET/CT is a promising technique providing CFR, a quantitative measure of the coronary microcirculation and the function of the large epicardial arteries, and CAC, quantifying the overall atherosclerotic burden of the heart. In individuals without epicardial coronary stenosis, reduced CFR cannot be entirely attributed to structural microvascular disease but can be functional and reversible. PET imaging is considered the gold standard for quantification of myocardial blood flow and CFR (28). Lower CFR is strongly associated with the risk of cardiovascular disease and death in patients with diabetes (8).

We are, to the best of our knowledge, the first to investigate the association between a comprehensive panel of cardiac autonomic function measures and CFR in asymptomatic patients with T2DM. We demonstrate a positive correlation between all measures of cardiac autonomic function (except the washout rate) and CFR; however, of particular interest, the late heart-to-mediastinum ratio, assessed by cardiac ¹²³I-MIBG scintigraphy, was strongly associated with CFR, even after adjustment for appropriate risk factors. In age- and heart rate–adjusted models, LF power was also associated with CFR. However, our limited sample size implies a higher sampling variability increasing the risk of a type II error.

Few studies have investigated similar associations in patients with diabetes. In a study including 28 patients with type 1 or T2DM, patients with evidence of sympathetic nerve dysfunction, as assessed by the norepinephrine analog ¹¹C-hydroxyephedrine, had impaired sympathetically mediated dilation of coronary resistance vessels (11). A study in 28 patients with type 1 diabetes concluded that subjects with preclinical microangiopathy had wide-ranging abnormalities of cardiac sympathetic innervation and blood flow regulation (12).

An impaired late heart-to-mediastinum ratio might reflect damage in adrenergic receptors and also be a result of enhanced washout. The washout rate in our study population was not related to CFR or to the late heart-to-mediastinum ratio, indicating that damage in adrenergic receptors might be the impelling cause of the impaired coronary microcirculation. Our novel findings may be useful in further investigation of the elevated risk of cardiovascular disease in asymptomatic patients with T2DM. We reveal that the tests primarily reflecting sympathetic autonomous control and activity had the strongest association with the CFR. Elevated cardiac sympathetic tone and damaged adrenergic receptors may play an important pathogenetic role in the development of myocardial injury and cardiac events in T2DM. It has been hypothesized that increased cardiac sympathetic tone may decrease myocardial vascularity, increase mitochondrial reactive oxygen species production, precipitate myocardial apoptosis, and promote myocardial remodelling (12), leading to impairment of the vascular performance of the heart and reduction of the coronary blood flow (29).

The association between cardiac autonomic function and CAC was investigated in a cross-sectional study of 160 patients with type 1 diabetes and 163 control subjects without diabetes (30). Reduced HRV (evaluated as total power) was associated with increased coronary calcification; however, the association lost significance after adjustment for systolic blood pressure (30). A cross-sectional study including patients with T2DM showed increased CAC was associated with lower HRV; however, this relationship became insignificance after adjustment for diabetes and other conventional risk factors (31). We confirm the association between reduced HRV and increased CAC in our cohort and, likewise, that this association lost significance after adjustment for conventional risk factors, including systolic blood pressure, which could reflect shared risk factors.

The agreement between cardiac autonomic function measured by HRV and cardiac ¹²³I-MIBG scintigraphy has been investigated in previous studies in T2DM. Murata et al. (4) observed a significant correlation, whereas Scholte et al. (32) observed disagreement between HRV and cardiac ¹²³I-MIBG scintigraphy for the assessment of CAN. We demonstrate positive correlations between the late heart-to-mediastinum ratio and the time- and frequency-domain HRV indices. In unadjusted analyses, the late heart-to-mediastinum ratio correlated with the 30-to-15 ratio, an index of parasympathetic function, but not with the Valsalva test, which is a measure of parasympathetic and sympathetic function. This was unexpected and might be a chance finding resulting from the limited sample size. After comprehensive adjustment, no significant agreement was found between the late heart-to-mediastinum ratio and any of the CARTs. These findings could be because of differences in measuring modalities or because the sample size limits its usability for complex statistical analyses. Outcomes of CARTs do not yield specific information about sympathetic function, whereas the late heart-to-mediastinum ratio is a measure reflecting the function of adrenergic receptors and sympathetic activity.

The strength of this study is that, to our knowledge, it is the first to evaluate the association between a comprehensive panel of cardiac autonomic function measures, including cardiac ¹²³I-MIBG scintigraphy and CFR, as assessed by cardiac ⁸²Rb-PET/CT, in patients with T2DM. Limitations of the study include the small sample size increasing the likelihood of a type II error. Importantly, the sample size is rather limited for complex statistical analyses, and the results from our multivariate analyses have ideally to be confirmed in larger studies. Finally, the cross-sectional nature of this study makes it impossible to assess cause-and-effect associations and predictive value.

In patients with T2DM without clinical cardiovascular disease, we demonstrate a positive association between the late heart-to-mediastinum ratio, a measure reflecting the function of adrenergic receptors and sympathetic activity, and CFR, a measure of the coronary microcirculatory function. A reduced cardiac autonomic function and damage to adrenergic receptors may contribute to the development of cardiac microvascular dysfunction.

Acknowledgments. The authors thank all participants and acknowledge the work of study nurse L. Jelstrup and laboratory technicians A.G. Lundgaard, B.R. Jensen, T.R. Juhl, and J.A. Hermann (Steno Diabetes Center, Gentofte, Denmark).

Funding. Internal funding was provided by Steno Diabetes Center, Gentofte, Denmark, and Rigshospitalet, Copenhagen, Denmark.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. B.J.v.S. conceived and designed the research, acquired the data, performed statistical analysis, and drafted the manuscript. C.S.H. acquired the data and critically revised the manuscript for key intellectual content. P.H., A.K., and P.R. conceived and designed the research, acquired the data, handled funding and supervision, and critically revised the manuscript for key intellectual content. T.W.H. conceived and designed the research, acquired the data, performed statistical analysis, and critically revised the manuscript for key intellectual content. B.J.v.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.