Insulin resistance and dysglycemia are associated with left ventricular remodeling after myocardial infarction in non-diabetic patients

This study complies with the Declaration of Helsinki. The study protocol was approved by the local hospital ethics committee, and written informed consent was obtained from all participants.

We consecutively enrolled 1125 subjects with the first ever acute STEMI with the onset of symptoms within 12 h preceding hospital admission and received primary PCI from Jan, 2014 to Dec, 2017 in the Department of Cardiology, Rui Jin Hospital, Shanghai Jiao Tong University School of Medicine (Fig. 1). A total of 483 patients comorbid with diabetes, chronic or acute infection, prior myocardial infarction, chronic heart failure or cardiomyopathy, liver disease, malignancy and diseases requiring steroid therapy, were excluded. Additionally, 63 patients who did not have biochemical indices including fasting glucose and insulin, as well as echocardiography parameters obtained at discharge were not enrolled. Another 94 patients without follow-up echocardiography at 12-month were also excluded. Thus, 485 patients comprised the final enrollment. HOMA-IR levels were reassessed in 168 subjects within the study population at follow-up. The changes in echocardiography parameters were calculated. The association between changes in echocardiograph parameters and basal HOMA-IR as well as glucometabolic disorders was analyzed. The diagnosis of diabetes was made according to the criteria of American Diabetes Association and prediabetes was defined by fasting blood glucose of 100 mg to < 126 mg/dL, 2-h plasma glucose of 140 to < 200 mg/dL, or HbA1C of 5.7 to < 6.5% [14]. Hypertension and dyslipidemia were diagnosed according to seventh report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure (JNC 7) and guideline of the National Cholesterol Education Program (ATP III), respectively [15, 16].

All the blood samples were drawn and echocardiography was performed at discharge except that the peak cardiac troponin I (cTnI) was recorded after admission. The blood samples were collected in a quiet, air-conditioned room after overnight fasting and after at least 20 min supine rest. Serum glucose, insulin, blood urea nitrogen, creatinine, uric acid, total cholesterol, low-density lipoprotein-cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides, apolipoprotein A-I and apolipoprotein B were assessed (HITACHI 912 Analyzer, Roche Diagnostics, Germany). The estimated glomerular filtration rate (eGFR) was computed using the Chronic Kidney Disease Epidemiology Collaboration equation [17]. Blood HbA1c concentration was measured using ion-exchange high performance liquid chromatography with Bio-rad Variant Hemoglobin Testing System (Bio-Rad Laboratories, USA). Serum N-terminal pro-B-type natriuretic peptide (NT-proBNP) was determined using a commercially available electrochemiluminescence immunoassay kit (Roche Diagnostics). Serum levels of high sensitive C-reactive protein (hsCRP) were determined by ELISA (Biocheck Laboratories, Toledo, OH, USA). HOMA-IR was calculated according to the formula: fasting insulin (µU/L) × fasting glucose (mmol/L)/22.5. The detailed information about medical history and lifestyles including smoking status was obtained using a standard questionnaire by trained physicians. Body mass index (BMI) was calculated using the formula of weight/height² (kilograms per square meter). Body surface area (BSA) was calculated using the formula of 0.0061 × height + 0.0128 × weight − 0.1529. Blood pressure was measured on the non-dominant arm in a seated position after a 10-min rest, using an electronic blood pressure monitor (OMRON Model HEM-752 FUZZY’ Omron Co., Dalian, China). Three measurements were taken at 1-min intervals, and the average was used for analysis.

Transthoracic echocardiography was performed, at least, at the time of enrollment and 12-month follow-up, using a commercially available system (Vivid-I, GE Healthcare, Milwaukee, WI) with a 1.9- to 3.8-mHz phased-array transducer. Echocardiography was performed by a single sonographer credentialed in cardiac ultrasound. Two-dimensional (2D), pulsed-Doppler imaging was performed from standard parasternal and apical transducer positions with 2D frame rates of 60 to 100 frames/s. All data were stored digitally, and offline data analysis was performed (EchoPac, version 7; GE Healthcare) by two cardiologists at the conclusion of the study, blinded to the study time point.

The LV ejection fraction (LVEF) was calculated using the modified Simpson’s biplane technique. The LV length was measured in the apical 4-chamber view. To facilitate application of clinical normality cut points (20), LV end-diastolic volume (LVEDV) and LV end-systolic volume (LVESV) were indexed by BSA calculated at each study time point. LV mass was estimated from M-mode measurements by the formula:where LVEDD is LV end-diastolic diameter, IVST is interventricular septal thickness, LVPWT is LV posterior wall thickness. LV mass was indexed by BSA. Relative wall thickness (RWT) was determined by the formula:where IVST is interventricular septal thickness, LVPWT is LV posterior wall thickness, LVEDD is LV end-diastolic diameter.

Continuous variables were presented as median (interquartile range) or mean (standard deviation), and categorical data were summarized as frequencies (percentages). For continuous variables, normal distribution was evaluated with Kolmogorov–Smirnov test. Differences among groups were analyzed by Student’s t-test or one-way analysis of variance (ANOVA) followed by post hoc Bonferroni test. Correlation between HOMA-IR and changes in echocardiography parameters was determined by Pearson’s correlation test. Furthermore, multivariate linear regression was implemented to interrogate the association between insulin resistance and LV remodeling parameters. In model 1, changes in LVEDD were employed as dependent variables, and covariates, including male gender, age, history of hypertension, the presence of chronic kidney disease (CKD), levels of lymphocytes, HDL cholesterol, TIMI flow grade after PCI, log-transformed cTnI and baseline LVEDD was adjusted. In model 2, we further adjusted the presence of impaired glucose tolerance (IGT) and BMI (dichotomized by median), as well as interaction terms of each maker with tertiles of HOMA-IR in the model. Regression coefficients of interaction term represent the difference in coefficient of IGT (or BMI) between subgroups stratified by tertiles of HOMA-IR. Coefficients of interaction terms corresponding to intermediate and high tertiles vs. low tertile of HOMA-IR were calculated. All statistical analyses were performed using the SPSS 23.0 for Windows (SPSS, Inc., Chicago, IL, USA). A 2-tailed < 0.05 was considered statistically significant.

A total of 485 non-diabetic patients with STEMI who underwent PCI were enrolled in this study (Table 1). The level of HOMA-IR (2.36 [1.56–3.92]) was higher than the normal range according to previous reports [10, 18] and the prevalence of impaired fasting glucose (IFG) and IGT was 40.0% and 40.2%, respectively. After dividing the study population into three groups according to HOMA-IR tertiles, we found subjects with high HOMA-IR levels tended to have higher BMI, diastolic blood pressure, liver enzymes, triglyceride and cTnI levels, but were younger and had lower NT-proBNP levels. There was no difference in renal function, smoking history, comorbidities such as hypertension, atrial fibrillation, and CKD between the three groups. On admission, dual antiplatelet agents, statins, ACEI/ARBs and β-blockers were prescribed unless contraindicated. A total of 8.7% of patients were prescribed spirolactone depending on clinical scenario. No significant difference in medication use was detected between different groups.

HOMA-IR was reassessed in 168 subjects within the study population at the 12-month follow-up. There was no statistic difference in baseline HOMA-IR levels between the selected subjects and the overall population (P = 0.122). Although HOMA-IR levels were lower at follow-up (2.12 [1.61–3.09]) compared to those in the baseline (2.61 [1.77–3.99]; P = 0.022), there was still a stepwise increase in the follow-up HOMA-IR levels across increasing tertiles of the baseline values (P < 0.001; Additional file 1: Figure S1).

Baseline and 12-month follow-up LV geometric and functional parameters were assessed. Changes in echocardiography parameters were compared in subjects stratified by HOMA-IR tertiles (Table 2). There was an upward trend in post-infarction LV dilation with increasing tertiles of HOMA-IR (LVEDD, LVESD, LVEDV index and LVESV index, all P < 0.001). Especially, subjects with IGT presented greater LV dilation than those without at intermediate and high tertiles of HOMA-IR (Fig. 2; intermediate tertile: 1.78 vs. 0.13, P = 0.001; high tertile: 2.52 vs. 0.77, P = 0.010). When stratified by BMI, however, no difference was detected at any tertile. On the other hand, the post-infarction LV wall thinning appeared to be more prominent (Δ IVST, P = 0.003; Δ LVPWT, P = 0.006; Δ RWT, P = 0.001) in subjects with high HOMA-IR, suggesting a trend toward eccentric remodeling of LV after STEMI in subjects with insulin resistance. No significant difference in changes in LV mass index (P = 0.130) and LVEF recovery (P = 0.089) during follow-up was detected between different tertiles.

In addition, when grouping the overall study population based on the presence of IFG or IGT, we found subjects with IFG or IGT present more severe LV dilation than those without these glucometabolic abnormalities (Additional file 1: Tables S1, S2). Dividing subjects according to prediabetes status showed a modest but non-significant increase in LVEDD (P = 0.051) and LVEDV index (P = 0.091) in subjects with prediabetes versus those with normoglycemia (Additional file 1: Table S3).

Correlation analyses showed that Log-transformed HOMA-IR was correlated positively with changes in LV diameter and volume (LVEDD: r = 0.172, P < 0.001; LVESD: r = 0.164, P < 0.001; LVEDV index: r = 0.154, P = 0.001; LVESV index: r = 0.167, P < 0.001; Table 3 and Fig. 3), and inversely with LV wall thickness (IVST: r = − 0.129, P = 0.004; LVPWT: r = − 0.144, P = 0.001; RWT r = − 0.193, P < 0.001). After adjusting for confounding risk factors, these correlations remained significant. In line with previous reports [19], an inverse correlation was also present between Log-transformed HOMA-IR and changes in LVEF.

Finally, we performed multivariate linear regression to interrogate the association between insulin resistance parameters and LV remodeling (Table 4). After adjusting for confounding clinical variables in Model 1, tertiles of HOMA-IR were independently associated with changes in LVEDD (P for trend < 0.001). Compared with the low tertile, intermediate and high tertiles of HOMA-IR corresponded to 1.196-mm and 1.926-mm increases in LVEDD, respectively. By including IGT, BMI (dichotomized by median) and the interaction terms of these variables with tertiles of HOMA-IR in Model 2, we found the presence of IGT (P = 0.007), but not BMI (P = 0.140), was independently associated with changes in LVEDD. Furthermore, IGT modified the association of HOMA-IR with changes in LVEDD (P = 0.001), suggesting that subjects with high HOMA-IR levels perceive greater risk of LV dilation when coinciding with IGT. In contrast, the interaction term between HOMA-IR and BMI was non-significant (P = 0.212), suggesting that HOMA-IR was associated with post-infarction LV dilation irrespective of BMI.

Insulin resistance serves as the pathophysiological basis of type 2 diabetes, as well as the primary metabolic disorders in a great number of patients long before progression into overt diabetes [20]. While diabetes is an established risk factor for the development of heart failure after myocardial infarction [21, 22], the role of insulin resistance in post-infarction LV remodeling in non-diabetic patients is less understood. Existing evidence showed that insulin resistance predicted subsequent CHF incidence independently of established risk factors including diabetes in a large community-based cohort [23]. A separate study of the same cohort revealed that insulin resistance was associated with LV concentric remodeling rather than LV hypertrophy [12]. However, the remodeling process triggered by acute myocardial infarction may be different from that occurred in chronic conditions.

In the present study, we demonstrated that LV dilation was more prominent in subjects presenting with insulin resistance and dysglycemia. There was an upward trend in LV dilation with increasing tertiles of baseline HOMA-IR levels in non-diabetic patients with STEMI. HOMA-IR was positively correlated to changes in LV dimensions. By constructing multivariate regression models, we found HOMA-IR was independently associated with LV dilation after STEMI. On the other hand, the post-infarction LV dilation was more severe in subjects with IFG or IGT. These data are consistent with previous reports that basal glucose levels and poor glycemic control are predictors of adverse LV remodeling and cardiovascular events [24‐27]. Noteworthy, glycemic variability after the onset of acute coronary events was shown to predict patient prognosis and subsequent LV dilation [28, 29]. Taken together, these findings provide evidence that patients with insulin resistance, dysglycemia and high glycemic variability perceive higher risk of developing adverse LV remodeling and poor cardiovascular outcomes.

We found the association between LV dimension and HOMA-IR level was modified by the presence of IGT in the regression model. Subjects in the intermediate and high tertiles of HOMA-IR have greater LV dilation when coinciding with IGT, implying that insulin resistance and dysglycemia may have additive effects on adverse LV remodeling after myocardial infarction. In contrast, no significant interaction term was present between HOMA-IR and BMI with changes in LV dimension. This may be due to the reason that BMI is a crude measure of general adiposity and its impact on LV remodeling is, at least partly, attributable to its association with insulin resistance (Pearson r = 0.344, P < 0.001). In other words, subjects with different BMI but identical HOMA-IR level may develop LV dilation to a similar extent (Fig. 2b). A substantial body of evidence showed that abdominal obesity would be more strongly related to risk of mortality [30], myocardial infarction [31] and type 2 diabetes [32]. Although we did not record waist circumstance or waist-to-hip ratio in this study, they are expected to provide more information than BMI in relation to LV remodeling.

The mean HOMA-IR level in this study is beyond the normal range as defined by a substantial number of epidemiological studies [10, 18]. There was also high prevalence of IFG (40.0%), IGT (40.2%) and prediabetes (72.6%) in the study population even though subjects who met the diagnosis of diabetes had already been excluded. These data concur with previous reports demonstrating high prevalence of glucometabolic abnormalities and undiagnosed diabetes in patients with acute myocardial infarction [33‐35]. Although the observed hyperglycemia and compromised insulin action might be due to activated production of catecholamines and cortisol in response to infarct extension and myocardial dysfunction, Choi et al. [33] showed that the proportion of abnormal glucose tolerance kept similar at admission and 3 months after discharge when the effects of acute stress and inflammation should have already been lessened. In the present study, HOMA-IR levels were reassessed at 1-year follow-up in 168 subjects within the study population. Although there was a decrease in HOMA-IR at follow-up compared to the baseline, subjects with high HOMA-IR levels in the acute phase were still more likely to exhibit insulin resistance after 1 year.

We showed that subjects with insulin resistance were at higher risk to develop eccentric LV remodeling than those without after STEMI, which was different from previous findings [12] that insulin resistance led to concentric LV remodeling in chronic conditions. Different mechanisms may predominate in distinct pathophysiological conditions that may underlie the paradox. In the chronic phase, insulin resistance leads to altered substrate metabolism, maladaptive immune responses, mitochondrial dysfunction and endoplasmic reticulum stress. These mechanisms promote reactive oxygen species (ROS) production and inflammation, thereby resulting in interstitial collagen deposition, crosslinking, and finally myocardial hypertrophy and relaxation deficiency [36‐38]. On the other hand, in the setting of myocardial infarction, insulin resistance is associated with poor myocardial reperfusion, impaired coronary microcirculation [39] and collateralization [40], and reduced collagen deposition in the scar [41]. These factors potentially lead to greater infarct size and post-infarction LV dilation, and finally a higher incidence of heart failure [42]. Nevertheless, the implicated mechanisms await precise characterization in future studies.

We appreciate limitations in our study. First, this study was a retrospective analysis based on prospectively collected data, and all the enrolled patients were from a single center. Second, data were analyzed according to HOMA-IR levels near the acute phase, which may over-estimate insulin resistance status of STEM patients. Third, we evaluated LV remodeling by calculating changes in echocardiography parameters. Performing cardiac magnetic resonance would provide more information. Fourth, anthropometric parameters such as waist circumstance and waist-to-hip ratio were not recorded in this study. They are better measures of central obesity than BMI and should provide more information. Further prospective studies are warranted to analyze the causal link between insulin resistance and LV remodeling, and the prognostic value of HOMA-IR for hard cardiovascular events in subjects with STEMI.