α-Glucosidase inhibitor miglitol attenuates glucose fluctuation, heart rate variability and sympathetic activity in patients with type 2 diabetes and acute coronary syndrome: a multicenter randomized controlled (MACS) study

It is well established that hyperglycemia on admission to hospital is associated with increased mortality in patients with acute myocardial infarction (AMI) [1, 2]. Recent studies have shown that hypoglycemia is also associated with increased short- and long-term mortality in AMI patients [3‐5]. Lee et al. [6] compared the prognostic significance of hypoglycemia and hyperglycemia in 34,943 AMI patients with or without type 2 diabetes mellitus (T2DM) from Korean registries. They found that admission hypoglycemia in patients with poorly controlled T2DM was associated with a markedly decreased 30-day survival rate. Nevertheless, HbA1c level was not associated with the risk of 30-day mortality.

Although the mechanism by which hypoglycemia leads to increased mortality is not clear, an abnormal QT prolongation during severe hypoglycemia may be associated with increased 30-day mortality [7]. In experimentally-induced hypoglycemia, abnormal QT prolongation, sinus bradycardia, and ventricular arrhythmias occurred that were potentially fatal [8, 9]. A clinical study has described bradycardia during severe hypoglycemia in both diabetic and non-diabetic patients, which may reflect imbalance of the sympathetic nervous system (SNS) [7, 8]. However, associations between hypoglycemia, heart rate variability (HRV), and the activity of the SNS have not been studied in patients with AMI/acute coronary syndrome (ACS).

Alpha-glucosidase inhibitors (αGIs) are a class of oral anti-diabetic drugs that act by inhibiting the hydrolysis of complex carbohydrates into glucose. Although they lower the degree of postprandial hyperglycemia, they also reduce the degree of hypoglycemia, probably by diminishing postprandial hyperinsulinemia [10]. Therefore, we hypothesized that αGI drugs stabilize glucose fluctuations, and thereby decrease the number of hypoglycemic episodes, in consequence lessening variation(s) in heart rate (HR) and/or imbalance of the SNS.

In this multicenter randomized controlled trial, we evaluated the effects of glucose fluctuations on HRV and autonomic nervous system activity in T2DM patients with ACS. We also investigated the effects of the αGI miglitol on glucose fluctuations and the other variables.

The new findings of this pilot study are: (1) T2DM patients with recent ACS manifested a substantial number of episodes of hypoglycemia (blood glucose ≤4.44 mmo/L) during the first 24-h CGM (Day 1). However, in the miglitol group, MAGE was significantly decreased on Day 2, and the min glucose level was increased and the episodes of hypoglycemia tended to be fewer after miglitol on Day 2; (2) mean and max HR during 24-h were increased on Day 2 in the control group, but the increases on Day 2 were mitigated in the miglitol group. Simultaneously, the increase in mean LF/HF during 24-h and day-time in the control group were also suppressed in the miglitol group. This is the first study to evaluate the associations between glucose fluctuations including hypoglycemia, HR and SNS activity in T2DM patients who had recently developed ACS. The study found that there is a substantial number of subclinical episodes of hypoglycemia in ACS and that they can be attenuated by diminishing the glucose fluctuations with a αGI, accompanied by concurrent reductions of HRV and autonomic nervous system activity.

There have been no published studies of 24-h continuous glucose monitoring in ACS patients. However, previous reports have suggested that there may be a substantial number of subclinical episodes of hypoglycemia in ACS patients. Lee et al. [6] reported that AMI patients had admission glucose levels <3.89 mmol/L in 1.20% (255/21,096) of diabetic patients and 0.73% (154/13,156) of non-diabetic patients. On Day 1 in the current study, diabetic patients with ACS had blood glucose levels ≤4.44 mmol/L in 35% (6/17) of participants in the control group and 31% (4/13) in the miglitol group. Although there are several differences in the protocols between our study and that of Lee et al. (admission vs post-admission, one-point vs continuous, and plasma glucose level vs interstitial blood glucose level estimated by CGMS), the high frequency of subclinical hypoglycemia in our study is not so questionable. Hermanns et al. [14] reported that if the POC measurement of hypoglycemic plasma showed a glucose value of ≤3.89 mmol/L, the CGM system gave a glucose level ≤4.44 mmol/L in interstitial fluid. Accordingly, we adopted ≤4.44 mmol/L in the CGM as a hypoglycemic alert. The value for the min glucose was increased and the frequency of episodes of blood glucose ≤4.44 mmol/L tended to be lower in the miglitol group on Day 2. It has been reported that αGI not only lowers the degree of postprandial hyperglycemia, but also decreases the degree of hypoglycemia, probably by diminishing postprandial hyperinsulinemia [10]. It is conceivable that αGI stabilizes glucose fluctuations, thereby decreasing the frequency of hypoglycemic episodes.

In the control group, the average of 24-h min HR was comparable between Day 1 and Day 2, but mean and max HR were increased on Day 2. Jabre et al. [16] reported that an increase in resting HR after the onset of ACS was a strong predictor of cardiovascular (CV) mortality. Although it has been suggested that elevated resting HR is an intrinsic risk factor for poor outcomes after MI rather than merely a marker of other CV risk factors, the underlying mechanisms by which resting HR increased are unknown [16]. Although we could not determine the mechanism, our study has shown that mean and max HR were increased on Day 2 in the control group. Since the standard protocol on cardiac rehabilitation in Japan allows bathing and walking freely in the hospital 6 days after PCI [11], a day-to-day increase in the in-hospital activity may influence their HR on Day 2. After being divided into three time periods, the max HR was increased only during night-time. The fact that max HR could be attenuated by miglitol suggests that HR may be partly affected by a postprandial or post-supper phase-dependent cause. Koivikko et al. [17] reported that marked and prolonged hypoglycemia increases the HR and decreases the high-frequency oscillations of HR. In our control group, the increases in the SNS activity during day-time and the max HR during night-time were accompanied by the increases in both glucose fluctuation (∆glucose ranged from 1.71 to 2.20 mmo/L) and hypoglycemic episodes (6–29%) during bed-time on Day 2. While, the miglitol treatment significantly increased min glucose and reduced glucose fluctuations and hypoglycemic episodes during bed-time, accompanied by attenuation of both increase of mean and max HR and SNS activation. These results suggest that impaired hemodynamic and neurohormonal parameters may be, in part, associated with the increases in both glucose fluctuation and hypoglycemic episodes, possibly forming a vicious cycle in T2DM patients with recent ACS. In addition, these changes may contribute to the intrinsic disturbance of hemodynamic and neurohormonal systems, partly resulting in a short-term poor prognosis after onset of ACS [1‐6].

αGI mono- and combination therapy can delay glucose uptake in the intestine and lower the postprandial glucose spike with infrequent development of hypoglycemia in patients with T2DM or impaired glucose tolerance [10, 18, 19]. Kitano et al. [20] reported that short-term miglitol treatment in patients with postprandial hyperglycemia and recent ACS attenuated the postprandial glucose spike, accompanied by improvement of endothelial function. In addition, a previous large randomized controlled trial and a meta-analysis showed that αGI treatment could prevent the development of CV events [21, 22]. However, the precise mechanism underlying αGI-induced improvement of CV outcomes is yet to be determined. Moreover, little is known about miglitol-mediated effects on glucose fluctuations, HRV, and SNS activity in the acute phase of ACS. In the present study, miglitol exerted prompt beneficial effects on glucose fluctuations and reduced the number of hypoglycemic episodes. Furthermore, miglitol attenuated increases in the mean and max HR and the mean LF/FH throughout the day. Although the reason why glucose fluctuations parameters were not associated with time-domain HRV parameters remains to be explained, our results suggest that glucose fluctuations may be, at least partially, associated with the autonomic nerve activity. Given that oscillating glucose is associated with excess oxidative stress in vascular endothelial cells and consequent endothelial dysfunction [23], it is possible that suppression of oscillating glucose might contribute to CV protection after onset of ACS. Further studies would be required to evaluate possible association between glucose fluctuation and hemodynamics in the other classes of anti-hyperglycemic agents [24].

Our study has several limitations. First, although we detected extensive individual differences in HR and cardiac autonomic regulation during 24 h, it was a post hoc exploratory analysis of a small sample of a PROBE design study, but not a placebo-controlled study. Because unexpected bias might be introduced in the assessment of outcome, it may be difficult to reach strong clinical conclusion. In addition, we could not refer to the generalization of the current findings for other ethnicities because of no recruitment of the other ethnicities than Japanese, although previous studies demonstrated a common association between reduced HRV and impairment of glycemic metabolisms in Hispanic/Latino and Japanese adults [25, 26]. Although HRV parameters are influenced by physical and exercise intensity, those intensity was not compared between the groups. In addition, although guidelines for clinical use of HRV recommended that time-domain methods should be used when using long-term ECGs in clinical studies [27], but not frequency domain methods, time-domain parameters, such as SDNN and RMSSD, during 24-h were increased in both groups. Furthermore, the SDNN during night- and bed-time increased significantly only in the control group, although why such increases were less evident in the miglitol group remains to be clarified. In addition, several parameters on HRV and SNS activity increased during 24-h on Day 2 in the control group, while the time phases of their increases differed for HR, SDNN, and LH/FH, with the mean and max LF/HF increasing during day-time, and max HR increasing during night-time. Therefore, further trials equipped with more appropriate design and sample would be needed in the future. Second, the study was unable to determine if there was a cause-and-effect relationship. Although we obtained evidence that glucose fluctuations can be associated with changes in HR and HRV, we cannot determine whether glucose fluctuations, especially periods of hypoglycemia, can directly alter HR and HRV. We obtained no data on biomarkers of inflammation or oxidative stress during the study period. Third, because the aim of present study was to evaluate the effects of glucose fluctuations on HRV and SNS activity in T2DM patients with recent ACS, we did not recruit non-diabetic patients in this study. Accordingly, we could not elucidate enough the specific peculiarities of HRV in diabetes. Accumulated evidence suggest that disturbed HRV was associated with metabolic syndrome, insulin resistance and diabetes [28, 29], and it is also reported that glycemic variability was independently associated with the presence of cardiovascular autonomic neuropathy in patients with uncontrolled T2DM [30]. Therefore, further studies are also needed to elucidate the diabetes-specific peculiarities of HRV and the effective medical care. Fourth, the study protocol focused on only the short-term acute-phase after ACS. Hence, we cannot extrapolate the miglitol-mediated beneficial effect to the chronic-phase. However, our study provides novel observations during the acute-phase of ACS that may partly contribute to future improvement of CV outcomes. Fifth, hyperinsulinemia and insulin resistance may be related to glucose fluctuation and/or incidence of hypoglycemia especially in diabetic patients with obesity or visceral fat obesity, and miglitol can diminish it [31]. Although we did not measure beta cell function and insulin sensitivity in the current study, hyperinsulinemia/insulin resistance and/or leptin resistance may be related to glucose fluctuation and/or incidence of hypoglycemia. Sixth, the CGM system measured glucose levels in the interstitial fluid, not in plasma or in the capillary artery. Therefore, our data about daily glucose variations should be confirmed by serial measurements of plasma glucose levels.

In summary, this is the first study to evaluate the associations between glucose fluctuations, HRV, and SNS activity in T2DM patients with recent ACS. Subclinical hypoglycemia was more prevalent in ACS than expected. Even short-term treatment of miglitol could alter favorably HRV, SNS activity, and glucose fluctuation with avoiding hypoglycemic episodes. Disturbed hemodynamic and neurohormonal activities may be, in part, associated with increased glucose fluctuations and hypoglycemic episodes, suggesting a potential new strategy to improve outcomes in patients with recent ACS.