Myocardial Dysfunction and Adrenergic Cardiac Innervation in Patients With Insulin-Dependent Diabetes Mellitus

1.1 Study Patients

We studied 14 consecutive asymptomatic patients with IDDM (10 men, 4 women; mean [±SD] age 28.5 ± 6 years) who fulfilled the following inclusion criteria: 1) no overt long-term diabetic complications; 2) no overt heart disease; 3) blood pressure <150 mm Hg systolic and <90 mm Hg diastolic; 4) no major additional risk factors for coronary artery disease (i.e., blood cholesterol concentration ≤5.2 mmol/liter, no smoking); 5) absence of electrocardiographic (ECG) signs and clinical history of myocardial infarction; 6) no angina pectoris; 7) absence of ischemic ST segment changes induced by isotonic and isometric exercise; 8) no LV wall motion abnormalities induced by isometric exercise; 9) no evidence of significant obstruction in coronary vessels by coronary angiography; 10) serum creatinine <1.2 mg/dl; and 10) no assumption of any drugs known to affect cardiac function or to interfere with myocardial MIBG uptake.

In addition, 10 healthy control subjects matched for gender (7 men, 3 women), body mass index (23.6 ± 1.4 vs. 23.1 ± 1.3 g/m², p = NS) and age (28.5 ± 3 years) were recruited to perform PESP isometric exercise and myocardial MIBG scintigraphy.

1.2 Study Protocol

The study protocol was approved by the local ethics committee, and informed consent was obtained from all participants. Assessment of the study cohort included 1) assessment of retinal and renal long-term diabetic complications; 2) determination of the degree of metabolic control; 3) coronary angiography; 4) LV function study at rest and during isometric exercise by quantitative two-dimensional echocardiography; 5) assessment of LV contractile function at rest by dobutamine infusion and PESP; 6) evaluation of LV contractile function during isometric exercise by PESP; 7) evaluation of autonomic cardiac innervation by MIBG scintigraphy.

1.3 Assessment of Degree of Metabolic Control

All participants were asked to fast for at least 12 h before the blood sampling. Blood samples were taken without cuffing to determine plasma glucose, hemoglobin A1c, electrolytes and creatinine levels and lipid profile. Patients with IDDM were all able to self-monitor their plasma glucose levels at least three times a day, both in the fasting condition and 2 h after principal meals. A complete lifestyle questionnaire to obtain medical histories, parental history of cardiovascular disease, information on smoking habits and physical activity was completed by each patient, and answers were discussed with one of the physicians (A.A.). The metabolic variables are reported in Table 1.

1.4 Assessment of Retinal and Renal Long-Term Diabetic Complications

The fundus oculi was examined by a trained ophthalmologist, and each patient was classified according to the presence or absence of diabetic retinopathy; 24-h microalbuminuria was measured for 3 days.

1.5 MIBG Scintigraphy

Myocardial MIBG scintigraphy was performed in diabetic patients and control subjects. Thyroid blockade was achieved with potassium iodide (Lugol’s solution) for several days before the study. Participants were instructed to avoid tea, coffee and alcoholic beverages on the day of the study. Otherwise, diabetic patients consumed their regular diets, and insulin was administered as usual. After a 30-min rest period, a dose of 100 to 185 MBq of I-123 MIBG was slowly infused intravenously. A dynamic acquisition was performed 1 h after injection in the left anterior oblique position (40°) for a better visualization of the left ventricle by means of a large field of view gamma camera (Siemens Medical System) equipped with a high resolution low energy collimator. The set of scintigraphic images consisted of 10 consecutive frames, each of 6 min, for a duration of 1 h, recorded and processed on a dedicated computer (Max Delta, Microvax Digital). Equivalent regions of interest (ROIs) of the left ventricle and lung were established to calculate both the heart/lung ratio and myocardial I-123 MIBG washout and to assess adrenergic cardiac innervation. The following variables were measured: 1) I-123 MIBG myocardial half-life (washout), expressed in minutes and determined from heart and lung curves; 2) heart/lung ratio, measured from equivalent ROIs referred to LV and lung tissue uptake; 3) sympathetic myocardial innervation, evaluated through the calculation of cpm/administered dose, normalized for body weight (expressed as cpm/MBq per kg) or for LV myocardial mass (expressed as cpm/MBq per g) calculated by two-dimensional echocardiography.

1.6 Echocardiographic Analysis

Echocardiographic examinations were performed with a Hewlett-Packard 77030A phase array ultrasonoscope and a 2.5- or 3.5-MHz transducer. Echocardiographic studies were coded and read by two independent observers (R.S., G.F.) in blinded manner, with no knowledge of to the patient’s identity and experimental condition. Echocardiographic analysis was performed using the digitized cine loop method (Prevue III System, Nova Microsonics, Inc.). Wall motion and myocardial thickening were detected by examining echocardiographic images of the left ventricle obtained in the apical four- and two-chamber views and in the parasternal long- and short-axis views. LV wall motion was analyzed by repeated viewing. Agreement of interobserver analysis for segmental asynergy was seen in 98% of the segments visualized; discrepancies were resolved by consensus. LV volumes were calculated by an ellipsoid biplane area–length method [10]. LVEF was derived as end-diastolic volume minus end-systolic volume divided by end-diastolic volume. LV endocardial echocardiograms in the apical four- and two-chamber views, for a minimum of two to four cardiac cycles, were digitized at end-diastole (R wave peak) and end-systole (time of smallest cavity area) by two independent observers. A discrepancy >10 ml for LV volume required the analysis of the echocardiographic tracing by a third observer; agreement was achieved by consensus. However, interobserver and intraobserver variability for LV area (r = 0.94 and r = 0.98, respectively) and for LV length (r = 0.95 and r = 0.96, respectively) was acceptable. LV mass was calculated by multiplying LV myocardial volume by the specific weight of the myocardium (1.05) [11]. Calibration of the carotid pulse tracings was performed at rest and during isometric exercise, with assignment of systolic blood pressure to the peak and diastolic blood pressure to the nadir of the tracing [12]. Linear interpolation to the level of the incisura was then performed to estimate end-systolic pressure. LV end-systolic circumferential wall stress (kdynes/cm²) was calculated by S= (1.332PD/2h)/(1 − h/D− D2/2L2), where P= end-systolic pressure(simultaneous with LV echocardiographic measurements); D= LV end-systolic short-axis diameter(in parasternal short-axis view); h= LVwall thickness; and1.332 is the factor to convert from mm Hg to kdynes/cm²[13]. Wall stress was calculated at rest, during handgrip testing and during dobutamine infusion.

1.7 Isometric Exercise

Handgrip testing was performed with the patient supine. Arterial pressure was measured every 30 s by an oscillometric method (Nippon Colin Co. Ltd.), and LV function was continuously monitored by two-dimensional echocardiography. Maximal voluntary contraction was determined by means of a handgrip dynamometer. Isometric exercise was performed for 3 min at 40% maximal voluntary contraction. Patients were instructed to avoid performing the Valsalva maneuver during handgrip. Left ventricular function was assessed before and every 30 s during exercise. Heart rate and electrocardiographic tracings were continuously monitored.

1.8 Post-Extrasystolic Potentiation

A quadripolar electrode catheter for transesophageal cardiac stimulation (interelectrode distance 2 to 3 cm) was passed through the nares into the distal esophagus. The lead was secured where the unipolar atrial electrogram, recorded from the middle electrodes, exhibited the greatest amplitude and the most rapid deflection. In this position, consistent and constant atrial capture was achieved by means of a programmable stimulator with a pulse width of 10 ms and an amplitude of 15 mA. A bipolar single atrial extrastimulus was delivered from the two middle electrodes every seventh sensed, spontaneous sinus beat. At the baseline assessment, the coupling interval between the spontaneous sinus beat and the induced extrasystole was progressively decreased by 10-ms increments from 500 to 300 ms. The post-extrasystolic beat was then allowed to occur spontaneously according to the patient’s intrinsic rhythm. During the procedure, LV volumes were monitored by echocardiography, and only beats without significant changes in LV end-diastolic volume (increase <10% from control value) were considered for analysis. PESP was repeated at peak handgrip using a coupling interval ranging from 400 to 300 ms. The procedures produced no complications or significant patient discomfort.

1.9 Dobutamine Echocardiography

Dobutamine echocardiography was performed 1 h after the PESP procedure. A quad screen, continuous loop was used to simultaneously compare rest images with dobutamine and recovery images. Dobutamine was infused intravenously by an infusion pump at 10 μg/kg for 5 min. Patients underwent continuous ECG monitoring, and blood pressure was recorded every 30 s. LV wall motion, LV volumes and LVEF were assessed before and at the fifth minute of infusion.

1.10 Coronary Angiography

Coronary cineangiography was performed in all study patients using the Judkins technique; all examinations were evaluated in blinded manner.

1.11 Tests for Cardiovascular Autonomic Nervous Function

Assessment was performed at least 3 h after the last insulin injection and included standard autonomic function tests by measuring heart rate variability during three standardized Valsalva maneuvres (blowing against 40 mm Hg for 15 s), heart rate variability on deep breathing and heart rate variability and blood pressure variability on lying to standing [14].

1.12 Laboratory Analysis

Plasma glucose was assayed using a glucose analyzer (Beckman). Hemoglobin A_1cwas determined with a high performance liquid chromatographic method [15]. Microalbuminuria was determined by the method of Keen and Chlouverakis [16]. Sodium and potassium were measured with an automated assay.

1.13 Statistical Analysis

Results are expressed as mean value ± SD. Multiple comparisons were performed by two-way repeated measures analysis of variance, followed by the Fisher protected least significant difference test. Correlations and regressions were determined by a linear least squares method. A p value ≤0.05 by the two-tailed test was considered to indicate statistical significance.

2.1 Baseline LV Function

Baseline characteristics of normal subjects and study patients are shown in Table 2. Measurements of LV end-diastolic volume (64 ± 9 ml/m², range 54 to 74) and LVEF (60 ± 9%, range 57% to 70%) in the control group were used to determine the mean and the 95% confidence limits of normal values. In diabetic patients, the baseline mean values of LV end-diastolic volume and LVEF did not differ compared with those of control subjects. Three patients (two in group A, one in group B) had an enlarged left ventricle, as indicated by an LV end-diastolic volume index greater than the upper normal limit. All patients showed a normal LVEF. Heart rate and systolic and diastolic blood pressures did not differ between diabetic patients and control subjects.

2.2 LV Response to Handgrip (Table 2)

Myocardial ischemia was not induced by isometric exercise as assessed by the appearance of wall motion abnormalities or ST segment changes. Heart rate and systolic blood pressure increased at peak exercise in both groups of diabetic patients and in normal subjects without significant differences. LVEF response to handgrip in the control group was characterized by a >0.05-U increase in all but two subjects, who showed no change during isometric exercise (mean baseline value 60 ± 9% vs. mean peak value 78 ± 8%, p < 0.005). Seven of the diabetic patients had a >0.05-U decline in LVEF (group A); peak LVEF was significantly lower than the baseline value (42 ± 7% vs. 60 ± 6%, p < 0.001). The remaining seven diabetic patients had a normal LV response to handgrip (five had a significant increase, whereas two showed no change, in LVEF [group B]). At peak handgrip, LVEF was greater in group B than in group A patients (72 ± 8% vs. 42 ± 7%, p < 0.001), whereas LV end-diastolic volume, heart rate and systolic and diastolic blood pressures did not differ. LVEF values at rest did not differ in patients with an abnormal response to isometric exercise, in patients with a normal response and in normal subjects (60 ± 6%, 61 ± 7% and 60 ± 9% respectively).

2.3 Circumferential Wall Stress–LVEF Relation

When control points obtained at rest and during handgrip were considered, the relation between LV circumferential wall stress and LVEF during exercise was linear for normal subjects, with a correlation coefficient of 0.82. As shown in Fig. 1, values for group A patients were downwordly displaced and fell below the 95% confidence interval, showing an impairment of ejection performance that was disproportionate to the degree of afterload during handgrip.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1

Relation between LVEF and LV circumferential end-systolic wall stress at rest and at peak handgrip. Open symbols= baseline values in normal subjects (squares),group A patients (triangles)and group B patients(circles); solid symbols= values at peak exercise in normal subjects (squares),group A patients (triangles)and group B patients (circles); dashed lines= 95% confidence limit of relation in normal subjects.

2.4 PESP Potentiation (Table 3)

PESP significantly increased LVEF both in group A (60 ± 6% vs. 74 ± 6, p < 0.001) and in group B patients (61 ± 7% vs. 72 ± 5, p < 0.001). To further clarify the etiology of LV dysfunction by exercise in diabetic patients, PESP was performed at peak handgrip. PESP significantly increased LVEF at peak exercise in group A (72 ± 5% vs. 42 ± 7%, p < 0.001), whereas no significant changes occurred in group B (75 ± 4% vs. 72 ± 5%, p = NS). In group A patients, potentiated peak handgrip LVEF did not differ from potentiated values at rest (75 ± 4% vs. 72 ± 5%, p = NS).

2.5 Dobutamine Echocardiography

On dobutamine administration, mean LVEF improved significantly (74 ± 5% vs. 61 ± 7%, p = 0.001) in group B patients, but no significant changes occurred in group A patients (64 ± 10% vs. 60 ± 6%, p = NS) (Fig. 2). LV asynergies did not appear during dobutamine infusion in either group of patients.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2

Changes in LVEF induced by dobutamine inotropic stimulation in study patients. Squares= Group A; circles= Group B.

2.6 Cardiovascular Autonomic Nervous Function

Traditional autonomic function tests did not reveal any differences between the two groups of diabetic patients or between patients and normal subjects.

2.7 MIBG Scintigraphy (Fig. 3)

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3

Scintigraphic images obtained for three individual patients: (A)normal cardiac MIBG uptake; (B)heterogeneity in MIBG distribution; and (C)severe reduction in cardiac MIBG uptake. Numbers= ROIs of LV myocardial wall (area 1) and lung background (areas 2 and 5).

MIBG myocardial half-life was reduced in diabetic patients compared with that in control subjects (332 ± 68 vs. 472 ± 40 min, p < 0.05), with groups A and B exhibiting similar values (330 ± 44 vs. 334 ± 67 min, p = NS). Calculation of cpm/administered dose did not reveal differences between normal subjects and group B diabetic patients when values were normalized for body weight (4.97 ± 0.92 vs. 5.09 ± 0.85 cpm/kg, p = NS) or for LV myocardial mass (2.98 ± 0.82 vs. 2.79 ± 0.94 cpm/g, p = NS). MIBG uptake normalized for body weight or for LV myocardial mass was significantly lower in group A diabetic patients with an abnormal LV response to handgrip than in both normal subjects (3.82 ± 1.03 vs. 4.97 ± 0.92 cpm/MBq per kg, p < 0.05; 1.69 ± 0.30 vs. 2.98 ± 0.82 cpm/MBq per g, p = 0.01) and group B patients with a normal LV response to exercise (3.82 ± 1.03 vs. 5.09 ± 0.85 cpm/MBq per kg, p < 0.05; 1.69 ± 0.30 vs. 2.79 ± 0.94 cpm/MBq per g, p = 0.01). Finally, the relation between LVEF at peak handgrip and myocardial uptake normalized for LV mass exhibited a significant negative correlation (y = 10x + 31, r = 0.68, p = 0.0063, SD 9.86) (Fig. 4).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4

Linear relation between peak handgrip LVEF (EF) and myocardial MIBG uptake normalized for left ventricular mass (LVM).

2.8 Coronary Angiography

Significant (>50%) coronary stenosis could not be detected in any patient by selective coronary angiography.

3.1 Myocardial Contractility Assessment by PESP

In the present study, cardiac inotropism was stimulated by two interventions (PESP and dobutamine) acting with different mechanisms. PESP, as part of the force–interval relation represents an inotropic stimulus clearly independent of variations in diastolic ventricular filling because it has been demonstrated in the isovolumetrically contracting heart [17]. In the ejecting ventricle, when the premature beat is followed by a compensatory pause, LV end-diastolic volume may be augmented, and this increased preload contributes to the enhanced performance. To minimize these changes in preload, only potentiated beats without significant changes in LV end-diastolic volume were evaluated in the present clinical study. Moreover, the maximal rate of ventricular pressure can be affected if afterload falls sufficiently. However, it does not appear likely that this would occur in the intact patient, given the relatively short test interval used to evaluate the force–interval relation. This assumption is supported by cardiac catheterization studies in which the plateau of restitution was sustained for >1,000 ms. During this period, aortic diastolic pressure would have fallen significantly, yet the maximal rate of rise of pressure appears to have been unaffected [18]. Finally, calcium availability at the level of the myofibrills is a critical determinant of PESP, but the interaction at the beta-adrenoceptor and increases in intracellular cyclic adenosine monophosphate (cAMP) concentration play little direct role in PESP. Thus, PESP is a potent contractile stimulus dependent on the sarcoplasmic reticulum function and on the calcium sensitivity of contractile proteins [19]. In contrast, the inotropic stimulus induced by dobutamine is mediated by the interaction with myocyte adrenoceptor and by the secondary increase in intracellular cAMP.

3.2 Muscle and Sympathetic Fiber Function in Diabetic Heart

Several previous studies have demonstrated an abnormal response of LVEF to exercise in a large subset of young adult patients with IDDM [20–23], although baseline myocardial contractility was normal [8, 9]. In our study patients, contractile reserve, defined as the capability to improve LV pump function by PESP, was normal in all study patients regardless of LVEF response to exercise. However, dobutamine infusion increased LVEF in diabetic patients with a normal response to exercise but failed to improve LVEF in diabetic patients with LV dysfunction induced by exercise. The afterload–LVEF relation has been used to examine myocardial contractility in isolated papillary muscle and in clinical investigations [24–26]. The response of the left ventricle to a sustained increase in afterload during exercise in normal subjects is described by a linear correlation: The increase in LV afterload is matched by an adequate increase in myocardial contractility. Sustained static muscular contraction increases heart rate and blood pressure, imposing a pressure load on the left ventricle. Activation of fast-twitch fibres during static contraction stimulates unmyelinized free nerve endings in the contracted muscle, inducing an arterial pressure reflex, which results in an important increase in arterial blood pressure and LV afterload [27, 28]. Moreover, a twofold increase in plasma norepinephrine and a fivefold increase in plasma epinephrine were demonstrated during handgrip exercise. These changes in plasma catecholamine levels may represent the basis of chronotropic and inotropic response [29]. LVEF values for Group A patients with IDDM were downwardly displaced and fell below the 95% confidence interval described by the control group values, indicating an inadequate increase in myocardial contractility at each comparable level of afterload (Fig. 1). Nevertheless, these diabetic patients with an abnormal LVEF response to exercise were able to normalize their response by PESP inotropic stimulation performed at peak handgrip. Thus, a defective recruitment of contractility seems to be a major cause of LV dysfunction during exercise in these patients. Moreover, the strong relation between the value of LVEF at peak handgrip and MIBG myocardial uptake (Fig. 4) demonstrates that the inadequate contractile response during isometric exercise may be related to impaired function of cardiac sympathetic nerve fibers. The blunted response to inotropic dobutamine stimulation in these patients with IDDM offers additional evidence supporting this hypothesis. PESP can normalize LVEF at peak handgrip because this inotropic stimulus is largely independent of the integrity of adrenergic receptors [19, 30]. The altered adrenergic innervation is also suggested by the rapid washout of MIBG from the myocardium in patients with IDDM with an abnormal LV response to exercise compared with both to normal subjects and patients with IDDM with a normal LV response to exercise. These data in humans are in agreement with the defective inotropic response to catecholamines demonstrated in animal studies [31–33]and extend the findings of Kreiner et al. [34], who, by using MIBG scintigraphy, recently showed that sympathetic myocardial dysinnervation is much more common in patients with IDDM than previously thought and that subclinical LV dysfunction is related to disturbances in adrenergic cardiac innervation.

As shown in animal models [32, 33], the sympathetic nervous system appears to be activated during the early stages of diabetes on the basis of reports of elevated plasma catecholamine levels. Prolonged exposure to catecholamines in humans induces a downward regulation of adrenergic receptors and alterations in adrenergic nervous fibers in the myocardium [35]. A decrease in serum T3 levels, hyperglycemia and insulin deficiency are additional possible factors contributing to the abnormalities in cardiac innervation in diabetes [36].

As demonstrated by coronary angiography, myocardial ischemia was ruled out as a major cause of exercise LV dysfunction in our study patients. In fact, LV dysfunction induced by exercise was always of the global, diffuse form, without regional asynergy. This finding is also strengthened by the lack of angina and ST segment changes during exercise, in accordance with a previous study [37]in which thallium imaging revealed no perfusion defects associated with LV dysfunction in IDDM. Finally, the normal increase in LV end-diastolic volume index at peak handgrip indicates that venous return was not a limiting factor in abolishing the normal LVEF response during exercise.

Traditionally, detection of autonomic dysfunction has been based on abnormal responses in heart rate or blood pressure to various bedside maneuvers. This approach has significant limitations in that it does not directly assess sympathetic cardiac innervation, lacks the ability to examine specific anatomic regions and is nonquantitative and thereby subject to large measurement variability. Our study demonstrates that there is an important difference in sensitivity between the traditional approaches, which were unable to demonstrate cardiac autonomic dysfunction, and the findings obtained both by myocardial MIBG uptake, which indicated a significant cardiac autonomic nervous impairment in patients with IDDM with an abnormal LV response to exercise.

3.3 Conclusions

The present study showed that a subset of uncomplicated young patients with IDDM exhibit an abnormal LV response to exercise despite normal contractility and that defective inotropic recruitment plays an important role in determining exercise LV dysfunction. Results of MIBG scintigraphy and dobutamine infusion revealed functional impairment of cardiac sympathetic nerve fibers as a major cause of this blunted contractile response during exercise. Finally, these results suggest that abnormalities in the cardiac beta-adrenergic system in diabetes may contribute to LV dysfunction before the appearance of irreversible damage to contractile machinery and overt heart failure.