The relationship between insulin sensitivity and heart rate-corrected QT interval in patients with type 2 diabetes

This cross-sectional study included 2927 patients with type 2 diabetes who were followed-up at the outpatients of the Affiliated Haian Hospital and Second Affiliated Hospital of Nantong University from January 2011 to December 2015. The inclusion criteria were as follows: (1) diagnosis of type 2 diabetes according to the criteria of ADA in 2011 [13] and (2) current use of antidiabetic treatments for more than 3 months. The exclusion criteria were as follows: (1) type 1 diabetes, testing positive for glutamic acid decarboxylase antibody or insulin antibody; (2) type 2 diabetic patients, who presented with instability of glycemic control and high risks of hypoglycemia, and treated with basal-bolus insulin, could not tolerate the OGTT; (3) fibrillation or flutter, atrioventricular blocks, and bundle-branch blocks; (4) heart valve disease, myocardial infarction, and heart surgery; (5) use of any drugs known to affect the QT interval such as tricyclic antidepressants; (6) chronic hepatic disease and kidney disease or malignancy; (7) excessive drinking (alcohol consumption more than 40 g of ethanol daily for women or 60 g daily for men); and (8) acute complications of diabetes, such as hyperglycemic hyperosmolar state and diabetic ketoacidosis; and (9) other endocrine disorders may have effect on glycaemic metabolism, such as hypothyroidism and hyperthyroidism. And informed consents in writing were received from all participants. The study protocol was reviewed and approved by the Medical Ethics Committee of the Affiliated Haian Hospital and Second Affiliated Hospital of Nantong University.

Upon enrolment, all participants were interviewed by trained investigators to record their age, sex, medication use (antidiabetic treatments, hypertensive treatment, and statin medications), health behaviours (smoking and drinking), and medical history of coronary heart disease (CHD). The antidiabetic treatments included lifestyle intervention alone, insulin injection, insulin secretagogues, and insulin sensitizers. CHD was proven by coronary angiography. Body mass index (BMI) was calculated as weight divided by the height square for further analysis. Those with SBP ≥ 140 mmHg, with DBP ≥ 90 mmHg, or receiving antihypertensive agents were considered as hypertensive.

After an overnight fast, the 75-g OGTT was performed during the early morning. All antidiabetic treatments were withheld at least 24 h before the OGTT. Blood samples were collected at basal, and 30, 60, 120, and 180 min after glucose ingestion for the determinations of plasma glucose and insulin levels. Insulin sensitivity was assessed using the insulin sensitivity index (ISI) proposed by Matsuda and DeFronzo [5]. ISI_Matsuda = 10,000/square root of (basal Insulin × basal glucose) × (mean glucose × mean insulin during the OGTT).

Standard resting 12-lead ECGs (FX-7402, CardiMax, FuTian Beijing Ltd., China) was performed for all participants. The ECG from each participant was recorded on a standard paper with a waveforms-amplitude of 10 mm/mV and a travelling-rate of 25 mm/s. The QT and RR intervals from ten consecutive beats were simultaneously assessed on the ECG. And QT and RR intervals were measured in lead II by two independent experienced physicians who were blinded to the personal information of participants. The QT interval was defined as the duration from the beginning of the QRS complex to the end of the T wave. The beginning of the QT interval was defined as the first negative deflection of the QRS complex, and the end was defined as the point that slope of T wave merged with the baseline [14]. The QT interval was corrected with the RR interval using Bazett’s formula, where QTc (ms) = QT/square root of RR (seconds). QTc interval was the mean of QTc interval from ten consecutive beats. For each participant the QTc interval represented average measurements of the two independent physicians. A QTc interval more than 440 ms was considered abnormally prolonged [9].

Serum insulin level was determined using magnetic beads-based enzymatic spectrofluorometric immunoassay (AIA360, TOSOH). Plasma glucose level (mmol/L) was determined using method of the glucose oxidase (Model 7600 Series, Hitachi). HbA1c concentration was determined by the method of high-performance liquid chromatography (D-10 system, Bio-Rad). Lipid profiles including triglycerides(TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDLC) and high-density lipoprotein cholesterol (HDLC), and serum uric acid (UA) were simultaneously determined by an automatic analyser (Model 7600 Series, Hitachi).

All analyses were conducted using SPSS Statistics V19.0 software (IBM SPSS Inc., USA). Clinical variables were calculated for the total subjects and across the ISI_Matsuda quartiles. Continuous variables with normal distributions are presented as the means and standard deviation (SD), whereas skewed distributions were presented as median and interquartile range. Categorical variables were presented as a frequency and percentage. Log-transformations were applied to all variables with skewed distributions for further analyses. The differences in continuous variables between the ISI_Matsuda quartiles were compared by One-way analysis of variance (ANOVA), and the categorical variables between the four groups were compared by Chi squared test. The correlation between the log ISI_Matsuda and the QTc interval was calculated with Pearson’s correlation test. Multiple linear regression analysis was conducted to compare the influence of the ISI_Matsuda and other metabolic factors on the QTc interval. Multiple logistic regression analysis models were also applied to investigate the associations of ISI_Matsuda quartiles (Q1–Q3) with the prolonged QTc interval (≤440 vs. >440 ms) relative to Q4, and the odds ratio (OR) and 95% confidence interval (95% CI) were determined. A value of p < 0.05 was considered statistically significant.

The clinical characteristics of the total participants and four subgroups according to the ISI_Matsuda quartiles are shown in Table 1, and the distribution of the QTc interval and ISI_Matsuda are shown in Figs. 1, 2. The mean QTc interval among all participants was 419 ± 32 ms, and the prevalence of the prolonged QTc interval in our study was 23.4%. The QTc interval significantly increased from 408 ± 31 to 414 ± 31 ms, 423 ± 29 and 432 ± 33 ms from ISI_Matsuda Q4 to Q1, and the corresponding proportion of patients with prolonged QTc interval significantly increased from 12.1% to 17.9%, 25.6%, and 37.9% from ISI_Matsuda Q4 to Q1, respectively. As the ISI_Matsuda quartiles decreased, the ratio of females, BMI, SBP, ratio of hypertension and statin medication, frequency of drinking and CHD, and values of TG, TC, LDLC, serum UA and HbA1c significantly increased, whereas the HDLC level decreased. Age, DBP, diabetic duration and the frequency of smoking did not show differences among the ISI_Matsuda quartiles. Comparisons of hypoglycaemic treatments showed that the frequency of insulin treatment increased as the ISI_Matsuda quartiles decreased, whereas secretagogues and sensitizers use were comparable among the ISI_Matsuda quartiles. The tendency of lifestyle intervention alone was increased when the ISI_Matsuda quartiles increased, but the difference was not significant (p for trend = 0.067).

The correlation between the QTc interval and the ISI_Matsuda is presented in Fig. 3. The QTc interval was significantly and negatively correlated with the ISI_Matsuda (r = −0.296, p < 0.001). The proportion of patients with prolonged QTc interval (>440 ms) increased with the ISI_Matsuda quartiles decreased (p for trend <0.001) (Fig. 4).

The QTc interval was significantly correlated with the ISI_Matsuda in the univariate analysis, and a multiple linear stepwise regression analysis was further performed to assess the associations of the ISI_Matsuda and other clinical risk factors, as independent variables, with the QTc interval as the dependent variable for the participants. The independent factors included age, female, BMI, SDP, DBP, diabetic duration, antidiabetic treatments, statins medication, hypertension, drinking, smoking, CHD, TG, TC, HDLC, LDLC, serum UA, and HbA1c. After adjusting for the metabolic risk factors in the multiple linear regression analysis, ISI_Matsuda, female gender, age, hypertension, insulin treatments and serum UA (β = −0.23, 0.22, 0.13, 0.078, 0.062 and 0.059, respectively, p < 0.005, total partial R ² = 18.4%) were the major independent contributors to the increase in the QTc interval (Table 2), and the ISI_Matsuda was the main independent contributor (standardized coefficient β = −0.23, t = −12.63, p < 0.001, partial R ²  = 8.5%).

Table 3 also shows the ORs of the prolonged QTc interval according to the ISI_Matsuda quartiles. Compared with participants in Q4 of ISI_Matsuda, the ORs of prolonged QTc interval for Q1, Q2 and Q3 of ISI_Matsuda were 4.45 (95% CI 3.41–5.81), 2.50 (1.89–3.30) and 1.59 (1.19–2.13), respectively. After adjustment in the multiple logistic regression, the corresponding ORs of the prolonged QTc interval for Q1, Q2 and Q3 of ISI_Matsuda versus Q4 were 3.11 (2.23–4.34), 2.09 (1.51–2.88) and 1.53 (1.09–2.14), respectively.

The prevalence of the QTc interval prolongation is considerably high in patients with type 2 diabetes, and many diabetes-related risks may contribute to the increase in the QTc interval. An increased QTc interval may be related to cigarette smoking [15], obesity [16], non-alcoholic fatty liver disease [17], hypertension [18], UA [19], dyslipidemia [20], hyperinsulinemia [21], glycaemic status [22], coronary artery disease [10], carotid intima media thickness [23], diabetic neuropathy [24, 25] and diabetic retinopathy [14]. Meanwhile, reduced insulin sensitivity is the pathophysiologic basis of type 2 diabetes and may underlie above cited risks factors. These risks factors that are associated with decreased insulin sensitivity are the same as those favoring the prolongation of the corrected QT interval. In the present study, insulin sensitivity assessed by the ISI_Matsuda, insulin treatment, hypertension and serum UA, was significantly associated with an increase in the QTc interval apart from non-modifiable risk factors including age and female. The associations of hypertension, serum UA, age and female with an increased QTc interval were consistent with previous studies. Gastaldelli et al. [26] demonstrated that physiological hyperinsulinemia induced by the euglycaemic insulin clamp acutely prolonged ventricular repolarization as assessed by the QTc in healthy volunteers. Our study revealed that insulin treatment was an independent risk for an increase in the QTc interval. Insulin may prolong the QTc interval in both healthy subjects and diabetic patients. With regard to the relationship between insulin sensitivity and the QTc interval, Shin et al. [27] showed that insulin resistance was an important risk for the prolongation of the QTc interval in normoglycaemic female subjects, and Festa et al. [12] found a weak association of the increased QTc interval with blunted insulin sensitivity, as determined by an intravenous glucose tolerance test in established diabetic patients (r = −0.15). Our study revealed that reduced insulin sensitivity assessed by the ISI_Matsuda was a major independent contributor to the increase in the QTc interval and accounted for 8.5% of its variation. Strategies targeting to improve insulin sensitivity may provide therapeutic methods to ameliorate the prolongation of the QTc interval and its associated prognosis in type 2 diabetes patients.

Reduced insulin sensitivity is the pathophysiologic basis of type 2 diabetes and may underlie a host of metabolic and cardiovascular disorders including glycaemic abnormality, dyslipidemia, hypertension and abdominal adiposity, which comprise the basis of the metabolic syndrome [2]. Each component of metabolic syndrome, which is characterized by reduced insulin sensitivity, is a significant risk factor for CVD. Metabolic syndrome can promote both atherosclerosis and atherosclerotic plaque formation, and the mechanisms possibly involve interaction of the components of metabolic syndrome that promote these processes [28]. Several prospective studies have demonstrated an association between reduced insulin sensitivity and severity of cardiovascular diseases in type 2 diabetes patients [2]. The Verona Diabetes Complications Study by Bonora et al. [29] showed insulin resistance that derived from HOMA was a major predictor for CVD in population with type 2 diabetes. In type 2 diabetes patients, the presence of metabolic syndrome was associated with a nearly fivefold increase in risk of cardiovascular diseases [30]. Our study documented that reduced insulin sensitivity as assessed by the ISI_Matsuda was significantly associated with an increase in the QTc interval, which  represented the ventricular myocardial membrane electrical stabilization, in type 2 diabetes patients.

The QTc interval reflects the total time taken for ventricular myocardial depolarization and repolarization, and metabolic, morphological, functional and structural abnormalities of the myocardium may induce ventricular myocardial membrane electrical destabilization and a subsequent increase in the QT interval. In an animal study, Lin et al. [31] observed that obese, insulin-resistant, 16 to 17-week-old rats developed cardiac hypertrophy, exhibited defective inactivation of current, and presented altered electrophysiology characterized by a prolongation of QTc interval. This study suggests that defective calcium inactivation can cause prolongation of the QT interval in patients presented with insulin resistance. Type 2 diabetes is associated with a high prevalence of left ventricular hypertrophy [32], and left ventricular mass is a strong determinant of the QT interval in these patients [9]. Insulin resistance may be important in the development of left ventricular diastolic dysfunction and structure in patients with type 2 diabetes mellitus [33‐35]. Insulin resistance and associated hyperinsulinemia in type 2 diabetes can promote the development of a specific form of cardiomyopathy, which is also termed diabetic cardiomyopathy, manifested by left ventricular hypertrophy and diastolic dysfunction [36]. The myocardial triglyceride content is increased in type 2 diabetes patients with insulin resistance [37] and could contribute to concentric remodelling and contractile dysfunction of the left ventricle [38].