- Email: [email protected]
Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after ST-Elevation Myocardial Infarction
IJCA-25746; No of Pages 7 International Journal of Cardiology xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after ST-Elevation Myocardial Infarction Filipe A. Moura a, Riobaldo Cintra a, Luiz S.F. Carvalho a, Simone N. Santos a, Rodrigo Modolo a, Daniel B. Munhoz a, Jose Carlos Quinaglia e Silva b, Otavio R. Coelho a, Wilson Nadruz a, Andrei C. Sposito a,⁎,1 a b
Cardiology Division, Faculty of Medical Sciences, State University of Campinas Medical School (Unicamp), Campinas, SP, Brazil Hospital de Base do Distrito Federal, Brasilia, DF, Brazil
a r t i c l e
i n f o
Article history: Received 3 February 2017 Received in revised form 19 July 2017 Accepted 30 November 2017 Available online xxxx Keywords: Insulin sensitivity HOMA Myocardial infarction STEMI Mortality
a b s t r a c t Background: Although stress hyperglycemia after myocardial infarction (MI) is consistently associated with increased mortality, recent studies suggest that the addition of upstream markers of glucose metabolism may improve risk identification. Hence, our aim was to evaluate the association between insulin sensitivity changes during MI hospitalization and outcomes. Methods: A prospective cohort of 331 consecutive ST-Elevation MI (STEMI) patients without insulin provision therapy was used for the analyses. Blood samples were collected upon admission (D1) and after 5 days (D5) of the inciting event. We measured blood glucose and insulin to estimate insulin sensitivity using the updated Homeostasis Model Assessment (HOMA2S). Patients were assessed for intra-hospital death and major adverse cardiac events (MACE) during follow-up. Results: HOMA2S was 62% ± 52% on D1 and 86% ± 57% on D5 (p b 0.001). Total follow-up was a median of 2 (0.9– 2.8) years and found a U-shaped relation between the change in HOMA2S from D1 to D5 (ΔHOMA2S) and major adverse cardiac events (MACE) (p = 0.017). Fully adjusted cox-regression models showed that patients from T1 and T3 were about 2.5 times more prone to suffer from MACE than those in T2. Net Reclassification Index adding ΔHOMA2S as a categorical variable dichotomized as T2 and T1 or T3 to a model of GRACE risk score with glucose D1 yielded a better predictive model (0.184 [95% CI 0.124–0.264]; p = 0.032). Conclusion: A U-shaped curve describes the relation between insulin sensitivity change and MACE during acute phase STEMI and, thus indicating that acute dysglycemia must be appreciated in light of a time spectrum and insulin levels. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The incidence of stress hyperglycemia in patients admitted with myocardial infarction (MI) has doubled in the last 20 years, nowadays occurring in about half of patients [1]. The biggest concern with this trend lies in the fact that hyperglycemia during acute coronary syndromes (ACS) is associated with increased mortality at 30 days, 1 year, and even 20 years after the inciting event [1], [2]. Furthermore, this association exists in both diabetic and non-diabetic patients with MI [3]. From a pathophysiological standpoint, hyperglycemia can feasibly affect outcome by decreasing collateral circulation, endothelial function, and ischemic preconditioning [4]. It can also increase myocardial cell
⁎ Corresponding author at: Cardiology Division, Faculty of Medical Sciences, State University of Campinas (Unicamp), 13084-971 Campinas, SP, Brazil. E-mail address: [email protected] (A.C. Sposito). 1 On behalf of the Brasília Heart Study Group.
apoptosis, thrombogenicity, and systemic inflammatory activity [4]. The only existing randomized study with ACS patients to date in which the glucose level was significantly lowered in the intervention group suggests there may be a causal role between the reduction in blood glucose and reduced mortality [5]. On the other hand, the association between hyperglycemia and mortality persists even after controlling for comorbidities, severity of coronary disease, and residual ventricular dysfunction, which gives way to the possibility of an indirect association or an unexplored mechanism [6]. Although the actual contribution of hyperglycemia in MI remains unclear, there are particular features that stand out. First, pre-existing diabetes seems to mitigate the risk associated with hyperglycemia [2]. Second, the increase in blood glucose during the first days after MI has greater impact on prognosis than admission blood glucose [6]. Thirdly, the degree of insulin resistance estimated during MI is associated with early mortality regardless of admission glycemia [7]. Taking all this evidence under consideration, it is reasonable to infer that the temporal variation in ST-elevation MI (STEMI) -related insulin resistance may influence the adverse effects of stress hyperglycemia.
https://doi.org/10.1016/j.ijcard.2017.11.111 0167-5273/© 2017 Elsevier B.V. All rights reserved.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
2
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
In order to assess insulin sensitivity in a larger number of STEMI patients, we conducted a preliminary study comparing surrogate insulin sensitivity indexes with the hyperinsulinemic euglycemic clamp in STEMI patients [8]. In this preliminary study, we found that the second version of the Homeostasis Model Assessment (HOMA2S) best estimates insulin sensitivity in these patients. In this context, the aim of the present study was primarily to both verify the existence of a change in insulin sensitivity, as measured by HOMA2S, during the acute phase of MI and its implications on clinical outcome. As secondary objectives, we sought to explore potential mechanisms that may mediate the association between dysglycemia and unfavorable clinical outcome.
2. Methods 2.1. Study patients Study participants were from the Brasilia Heart Study [9] (ClinicalTrials.gov Identifier: NCT02062554). It is an ongoing observational prospective cohort of consecutive ST-elevation MI (STEMI) patients admitted to Hospital de Base do Distrito Federal, a reference hospital located in Brasilia, Brazil. The current study included 331 patients enrolled between May 2006 and December 2015 and the flow diagram is illustrated in Fig. 1. Inclusion criteria were as follows: (i) b24 h after the onset of MI symptoms, (ii) ST-segment elevation ≥ 1 mm (limb leads) or ≥2 mm (precordial leads) in two contiguous leads, (iii) myocardial necrosis, as evidenced by an increase to at least one value above the 99th percentile above the reference limit of CK-MB (25 U/L) and troponin I (0.04 ng/mL) followed by a decline of both, (iv) and absence of hindrances for maintaining clinical follow-up. Individuals with previous use of insulin, insulin providers such as sulphonylureas or incretin-mimetic agents, i.e. dipeptidyl peptidase-4 inhibitors or glucagon-like peptide-1 agonists, were excluded from the study. The rationale behind this is because these represent an extremely heterogeneous population and these
pharmacotherapies alter endogenous insulin levels and intrinsic IS, thus potentially directly interfering with the interpretation of our data. Diabetes status was defined as previously diagnosed diabetes, use of hypoglycemic agents, or HbA1c ≥ 6.5%. In-hospital patient treatment was always conducted according to in-house assistant physicians and study investigators played a purely observational role. The physicians involved in the treatment of the patients were blind to all analyses performed in the study. The local Ethics Committee approved the study and all participants were required to sign an informed consent.
3. Biochemical analysis Blood samples were obtained at D1, i.e. b 24-h from MI onset, and at day five following MI (D5). Collected blood samples were centrifuged for 10 min at 3500 rpm and plasma was aliquoted for storage at −80 °C. The following biochemical panel was obtained from plasma samples and measured using standard laboratory methods: glucose, total cholesterol, triglycerides, high-density lipoprotein cholesterol, and high-sensitivity C-reactive protein. HbA1c was measured using highperformance liquid chromatography. A NO chemiluminescence analyzer (model NOA, Sievers Instruments, Boulder, CO) was used to determine the plasma pool of nitrite and nitrate (NOx) after reduction with acidic vanadium (III) chloride. Plasma insulin and C-peptide concentrations were respectively assessed by electrochemiluminescence and imunoquimioluminescence. The Homeostasis Model Assessment version 2 (HOMA2) was used to estimate insulin sensitivity (HOMA2S) using fasting plasma insulin and glucose and calculator version 2.2 [10]. The validity of HOMA2 index during STEMI was previously verified in STEMI patients by euglycemic hyperinsulinemic clamp in our laboratory [8].
Fig. 1. Flow chart of study patients.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
3
3.1. Coronary angiography
3.5. Study outcomes
Coronary angiography was systematically performed on all patients enrolled to assess coronary thrombotic burden, degree of reperfusion, and severity of atherosclerotic disease. Thrombotic burden was estimated at the culprit coronary artery according to the Thrombolysis In Myocardial Infarction (TIMI) Thrombus Grade Scale as follows: (i) Grade 0 – no angiographic evidence of thrombus is present; (ii) Grade 1 – possible thrombus, with reduced contrast density, haziness, irregular lesion contour, or a smooth convex at the site of total occlusion suggestive but not diagnostic of thrombus; (iii) Grade 2 – definite thrombus, with greatest dimensions less than half of vessel diameter; (iv) Grade 3 – definite thrombus but with greatest dimensions less than half but greater than two vessel diameters; (v) Grade 4 – definite thrombus greater than two vessel diameters; (vi) Grade 5 – total occlusion [11]. The Gensini score, a well-validated and widely used method for calculation angiographic CAD burden, was estimated coronary disease severity [12]. Coronary flow in the culprit artery was measured using TIMI flow grade and Myocardial Blush Grade (MBG), and was assessed after angioplasty in patients who underwent primary percutaneous coronary intervention (PCI) or at the earliest possible angiogram performed after thrombolysis.
The primary end point of this study was the occurrence of major adverse cardiac events (MACE), defined as in-hospital death (any death before hospital discharge or within 30 days of the index event), new fatal or non-fatal MI and sudden cardiac death (non-traumatic unexpected event occurring within 6 h of symptom onset). After discharge from the hospital, all events were assessed during the previously described outpatient visits. Absentee patients (or relatives) were contacted by telephone for evaluation of vital status. All events were adjudicated by an independent adjudication committee that did not participate in data analysis, and only the first event was chosen when a patient experienced more than one event. Patients who died from non-cardiac causes were censored on the day of their deaths.
3.2. Brachial artery reactivity Brachial artery measurements were performed using a highresolution ultrasound (IE33 and 3–9 MHz linear transducer; Philips Medical Systems, Bothell, Washington, USA) after over-night fasting 30 days after admission (D30). Vasoactive medications were withdrawn 24 h prior to assessment. After a 10-min rest period in a quiet room with the temperature controlled around 22 °C, the brachial artery was located above the elbow, and a longitudinal image of 6 to 8 cm was obtained as the resting scan. A blood pressure cuff was placed on the forearm and inflated to 50 mmHg above the systolic blood pressure for 5 min. The cuff was deflated, and the flow-mediated dilation (FMD) scan was obtained for 2 min. The percentage change in diameter for FMD was calculated in relation to the respective baseline scans. The same experienced physician, who was blinded to patients' data, made all assessments. The intra-observer reproducibility was 95%.
3.3. Cardiac magnetic resonance imaging (CMRI) CMRI studies were carried out in a subgroup of 65 patients using a MRI scanner with a 1.5-T (Signa CV/i, GE Medical Systems, Waukesha, WI), equipped with a gradient of high performance (gradient strength 40mT/m; maximum slew rate 150mT/m/s) and a four elements phased array cardiac coil. Areas of MI were quantified on the gadolinium-based delayed enhancement myocardial images at D30 to quantify MI mass, left ventricle (LV) volumes, and ejection fraction. Areas of microvascular obstruction were defined as subendocardial hypo-enhanced regions surrounded by hyperenhancement and were included as part of the core infarct. On cine-CMRI left ventricle (LV) volumes and ejection fraction were measured by ReportCard software (GE Medical Systems, Waukesha, WI), applying Simpson's method.
3.4. Follow up After hospital discharge, patients were reevaluated during visits every three months at the study outpatient clinic. All patients received guideline-based therapy and lifestyle counseling for diet, smoking cessation, regular physical activity, and weight loss. The physicians involved in outpatient care were blind to all analyses performed in the study.
3.6. Statistical analysis Groups were separated according to tertiles of ΔHOMA2S, which represents the change in HOMA2S from D1 to D5. The change in values for dependent variables was calculated as the difference between D5 and D1. Data distribution was assessed for normality with the use of histograms and Kolgomorov-Smirnov test. For continuous variables, normally distributed data are presented as mean ± standard deviation and non-normally distributed data are presented in the form of median (interquartile range). Proportional differences between groups were evaluated using chi-square. Log transformation was applied to variables in order to use parametric tests. Mean differences between groups were evaluated with one-way analysis of variance and analysis of covariance (ANCOVA) where there was need for adjustments. Mean differences between variables that persisted with skewed distribution after logtransformation were compared using the Kruskal-Wallis test (glucose, HDL-C, TG, CK-MB peak, and infarction mass). ANCOVA was used for FMD, NOx, and CRP and adjusted for age, sex, BMI, DM, hypertension, dyslipidemia, and reperfusion therapy. ANCOVA was used to compare these variables between groups after analysis with histograms, normality plots, and residual scatter plots that tested for linearity, normality, and variance. Post-hoc testing (Bonferroni) was performed to determine which groups were different when there were significant differences from ANOVA and ANCOVA analyses. Before conducting survival analysis, we explored the association between events and ΔHOMA2S because previous data has shown a J-shaped association between glucose levels and mortality [13]. The found curvilinear association between ΔHOMA2S and outcomes was assessed using restricted cubic splines with 3 knots adjusted for age, sex and baseline log-transformed HOMA2S values. Given the distribution that was found, separated the cohort in tertiles. Long-term survival between tertiles was assessed using Kaplan-Meier method and the difference between survival curves was tested with the log-rank test. Cox regression models were used to evaluate the association between tertiles of ΔHOMA2S and MACE. The number of MACE for this analysis suggested that the model would support only 4 to 5 covariates. Therefore, a propensity score for ΔHOMA2S tertiles (T2 = 1 and T1 or T3 = 0) was generated using age, sex and variables that showed significant or marginal differences among the ΔHOMA2S tertiles: hypertension, diabetes mellitus, obesity, dyslipidemia, triglycerides and reperfusion therapy. Four Cox-regression models were constructed: model 1 included HOMA2S D1; model 2 included HOMA2S D1, age and sex; model 3 included HOMA2S D1 and the propensity score; and model 4 included: HOMA2S D1, glucose D1 and the propensity score. Net reclassification improvement (NRI) and integrated diagnostic improvement (IDI) were used to assess risk reclassification when adding ΔHOMA2S dichotomized as T2 versus T1 or T3 to the GRACE score and glucose D1. A two- sided p-value of b 0.05 was considered statistically significant. All analyses were conducted using SPSS software version 20.0 or STATA software version 10.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
4
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
4. Results
Table 1 Characteristics of study patients stratified by tertiles of ΔHOMA2S.
4.1. Baseline characteristics Patients were organized according to tertiles of ΔHOMA2S, which was defined as the difference between D5 and D1 HOMA2S levels. Cut-offs for tertiles of ΔHOMA2S were as follows: Tertile 1 (T1) ≤ 2.9% (n = 111), 2.9% b Tertile 2 (T2) ≤ 41.0% (n = 111), and Tertile 3 (T3) N 41.0% (n = 109). In a very simplified manner, the tertiles may be interpreted in the following way: T1 - little to no improvement in insulin sensitivity; T2 - improvement in insulin sensitivity; T3 marked improvement in insulin sensitivity. Clinical characteristics are shown on Table 1. Of note, patients from T1 and T2 had lower prevalence of hypertension. Also, patients from T3 demonstrated lower BMI than T2 (p = 0.005), but there was no difference between patients from T1 and T2 (p = 0.883) or between T1 and T3 (p = 0.109). Waist circumference behaved in a similar fashion; there were only differences between T2 and T3 (p = 0.001). There was also a lower proportion of patients with BMI ≥ 30 kg/m2 in T3. Patients from T2 had a lower prevalence of sedentary lifestyle than other groups. No other differences between demographics, medical history, medical evaluation, or hemodynamics were found. Finally, Patients from T1 and T2 received higher rates of tenectaplase, but considering both tenecteplase and PCI, reperfusion therapies were similar between groups. There were no other significant differences in treatment between groups. 4.2. Glucose metabolism and biochemical data We found that 82% of patients presented with declined insulin sensitivity (HOMA2S b 100%) on D1 while 68% of patients presented this on D5 after STEMI. Accordingly, there was improvement of HOMA2S from D1 to D5 of STEMI (D1 62 ± 52% vs. D5 86 ± 57%, p b 0.001). Patients at the higher tertile of ΔHOMA2S had higher levels of glucose D1 and lower levels of glucose D5. There was also higher prevalence of glucose ≥120 mg/dL in T3, but there was no difference in prevalence of glucose ≥140 mg/dL. Also, groups from T2 and T3 had higher values of insulin on D1 and lower levels of insulin D5. These findings were expected due to the manner in which groups were separated. Of note, T1 (p = 0.061) and T2 (p = 0.001) had lower TG levels when compared to T3 using post-hoc tests. 4.3. Endothelial function, infarction size, and angiographic assessment We found no differences in endothelial function, as assessed by NOx or FMD, and in inflammatory markers, according to ΔHOMA2S tertiles. There was also no significant difference in infarct size between groups as estimated by CKMB-peak or by CMRI. Angiographic data showed that patients from T2 had a higher proportion of patients with significant left main coronary lesion. All other characteristics were similar between groups (Supplementary Table 1). 4.4. Clinical follow-up and outcomes Clinical follow-up time lasted for a median of 2.0 (0.9 to 2.8) years. During follow-up, 49 MACE occurred (17 intra-hospital deaths, 9 fatal MI, 17 nonfatal MI, and 6 sudden cardiac deaths). Kaplan-Meier curve estimation (Fig. 2A) showed that patients in T2 have a higher rate of survival free of MACE (T1 18.9% vs. 7.2% vs. T3 18.3%, p = 0.025). Multiple cox-regression models showed that patients from T1 and T3 were about 2.5 times more prone to suffer from MACE (Supplementary Table 2). Finally, adding only T2 to the cox-regression model and adjusting for age, sex, HOMA2S D1, hypertension, diabetes, dyslipidemia, BMI, tenecteplase, and beta-blocker use yielded a HR of 0.449 (95% CI 0.204–0.986; p = 0.046), further suggesting better survival in this group. Additional adjustment for admission glucose (HR 0.439 [95% CI
ΔHOMA2S ≤2.9%
N2.9 and ≤41%
N41%
N
111
111
109
Demographics Age, years Male, % BMI, kg/m2 BMI ≥ 30 kg/m2, % Waist circumference, cm
60 ± 12 78 26.8 ± 4 23 96.8 ± 11
57 ± 11 73 27.4 ± 4 19 98.9 ± 11
60 ± 11 77 25.7 ± 4 10 93.4 ± 10
0.064 0.571 0.006 0.035 0.001
10.6
6.2
6.1
0.347
37.2 65.5 14.2 38.9 63.7 54.9 131(180)
38.1 49.6 12.4 29.2 47.8 53.1 90(158)
46.9 51.8 5.3 24.6 57.9 38.9 120(192)
0.258 0.033 0.07 0.057 0.05 0.057 0.286
133 ± 29 85 ± 17 76 ± 15 92
134 ± 32 83 ± 20 75 ± 18 91
138 ± 30 86 ± 21 76 ± 17 86
0.413 0.375 0.959 0.32
5.9(0.8) 113(33) 40.4 17.5 113(32) 11.1 ± 8 21.2 ± 18 86.3(87) 44.7(45) −37.2(45)
5.8(0.9) 119(37) 49.6 24.8 103(25) 24.5 ± 13 14.1 ± 6 34.0(29) 56.4(33) 61.2(57)
5.8(0.7) 127(31) 62.6 28.7 97(16) 24.8 ± 15 6.4 ± 3 31.7(38) 126.0(83) 251.6(236)
0.954 b0.001 0.003 0.132 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
69 ± 19 36(13) 116 ± 37 133(124) 0.8(1.5) 3.47(5.9) 1.90(4.9) 214(261) 14.6(12)
71 ± 19 35(13) 125 ± 43 138(103) 0.6(1.3) 3.60(5.3) 2.14(4.3) 224(239) 18.3(11)
69 ± 23 38(17) 123 ± 40 109(71) 0.7(1.3) 3.27(7.8) 2.35(6.1) 250(288) 13.5(11)
0.782 0.216 0.201 0.001 0.750 0.087 0.087 0.903 0.633
64.9 11.6 76.8 99.1 69.4 51.4 62.2
67.3 12.8 81.7 97.3 67.3 61.9 62.8
52.2 17.6 69.4 99.1 64.0 54.9 74.6
0.042 0.404 0.106 0.447 0.693 0.265 0.084
Medical history Prior MI, % Risk factors Smoking, % Hypertension, % Diabetes Mellitus, % Dyslipidemia, % Sedentarity, % Family history for CAD, % Arrival time, min Hemodynamics Systolic Blood Pressure, mmHg Diastolic Blood Pressure, mmHg Heart Rate, bpm Killip class I, % Biochemical analyses HbA1c, % ^ Glucose D1, mg/dL Glucose D1 ≥ 120 mg/dL, % Glucose D1 ≥ 140 mg/dL, % ^ Glucose D5, mg/dL Insulin D1, μU/mL Insulin D5, μU/mL HOMA2S D1, % HOMA2S D5, % (HOMA2S D5 – HOMA2S D2)/ HOMA2S D1 Glomerular filtration rate, units ^ HDL-C, mg/dL LDL-C, mg/dL ^ TG, mg/dL ⁎CRP D1, mg/L ⁎CRP D5, mg/L ⁎ΔCRP, mg/L ^ CK-MB peak, mg/dL ^ Infarction mass (CMRI), g Treatment Tenecteplase, % Primary PCI, % Reperfusion therapy, % Aspirin, % Simvastatin, % ACEi or ARB, % Betablocker, %
p Value
⁎ Analysis of covariance adjusted by age, sex, smoking, BMI, reperfusion therapy, and basal levels (where fitting). D1; 1st day of MI; D5: 5th day post-MI; BMI: body mass index; CAD: coronary artery disease; CRP: C reactive protein; CMRI: Cardiac Magnetic Resonance Imaging. ^ Kruskal-Wallis test.
0.202–0.953; p = 0.037]) or ΔGlucose (HR 0.437 [95% CI 0.202–0.946; p = 0.036]) yielded similar results. As shown in Fig. 3A, when treating ΔHOMA2S as a continuous variable, the relationship of ΔHOMA2S to MACE was U shaped. The same result was obtained using the Quantitative Insulin Check Index (QUICKI) model and long-term MACE (Supplementary Fig. 1). Furthermore, reclassification analysis adding ΔHOMA2S as a categorical variable dichotomized as T2 and T1 or T3 to a model of GRACE risk score with glucose D1 yielded better predictive models with increases in NRI (index =
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
%Survival free of events
A
5
100
ΔHOMA2S T1 ΔHOMA2S T2 ΔHOMA2S T3
80 60 40 20 0
0
400
800
1200
1600
Days post STEMI
%Survival free of events
B
100
Δ HOMA2S T1 Δ HOMA2S T2 Δ HOMA2S T3
80 60 40 20 0
0
400
800
1200
1600
Days post STEMI Fig. 2. Kaplan-Meier curves. A) Kaplan-Meier curves showing survival free of MACE of STEMI patients (not taking insulin or secretagogues) according to tertiles of ΔHOMA2S. B) KaplanMeier curves showing survival free of MACE of non-diabetic STEMI patients according to tertiles of ΔHOMA2S.
Fig. 3. Adjusted hazard ratio of major adverse cardiac events (MACE) in function of difference in HOMA2S between D5 and D1 (ΔHOMA2S) in the whole population. Data were adjusted for age, sex, and baseline HOMA2S. The continuous line represents the hazard ratio (HR) and the black dashed lines the 95% confidence interval (CI). ΔHOMA2S = 21.5% (median) was considered the reference (HR = 1).
0.184 [95% CI 0.124–0.264]; p = 0.032) and IDI (index = 0.056 [95% CI 0.0390–0.073]; p = 0.001). This was further observed after KaplanMeier assessment stratified by the presence of hyperglycemia (defined as glucose D1 ≥ 140 mg/dL). Once again, patients from T2 demonstrated increased survival free of MACE both in patients with and without hyperglycemia upon hospital admission (p = 0.017). The trend for increased MACE-free survival of T2 was found both in patients with high HbA1c (≤5.8%) and those with low HbA1c (N 5.8%) after Kaplan-Meier curve estimation stratified by HbA1c status (p = 0.018). We also conducted an analysis using the Forest Plot method in order to better understand which type of events (intra-hospital death or postdischarge events) were influencing differences in MACE between groups. Post-discharge MACEs seem to be more pronouncedly different between groups (Supplementary Fig. 2). Finally, as shown in Fig. 2B, we ran the same analysis strictly with non-diabetic patients (n = 258) and found a similar difference between groups (T1 19.8% vs. T2 6.8% vs. T3 22.6%, p = 0.011). After cox-regression analysis, patients from T1 (HR 2.913, 95% CI 1.139– 7.446; p = 0.026) and T3 (HR 3.753, 95% CI 1.498–9.404; p = 0.005) presented with higher long-term risk than patients from T2. This observation persisted even after adjusting for age and sex (T1 HR 2.521, p = 0.057; T3 HR 3.271, p = 0.012) (Supplementary Table 3).
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
6
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
5. Discussion Using a subgroup of acute-phase STEMI patients that did not previously use insulin provision or incretin-based therapies, we have shown that there is a pronounced change in insulin sensitivity between the first and the fifth day of STEMI. This very change in insulin sensitivity during the acute phase of STEMI is strongly tied to outcome, even after adjusting for more established markers such as admission hyperglycemia. In particular, the relation between the recovery in insulin sensitivity and adverse outcome is not linear but is in fact U-shaped. In light of this, patients in whom insulin sensitivity recovers either too vigorously or who don't recover at all both have at least approximately a 3-fold risk of intra-hospital death and a 2-fold increase in incidence of MACE during long-term follow-up. Over the past years, the lack of consensus in insulin-mediated glucose control trials during MI has resulted in hyperglycemia loosing traction as a causal agent of adverse clinical outcome after MI. However, the derangement in glucose metabolism that occurs during MI is complex and recent evidences suggest that efforts should not be merely focused on blood glucose level control. For example, we have previously shown that endogenous insulin secretory function is associated with intrahospital mortality [9]. Furthermore, a previous group has also shown that admission HOMA-IR, another popular surrogate for insulin sensitivity, may be directly associated to risk after MI [7,14]. Hence, a fresh perspective on the subject was warranted to help unveil more of what occurs in this setting. As commented above, to better assess the role of insulin sensitivity during STEMI, we first investigated whether surrogate indexes were feasible for assessing insulin sensitivity during STEMI. We found that HOMA2S and QUICKI were the methods that best corresponded to hyperinsulinemic euglycemic clamps and that HOMA-IR performed inferiorly to all others [8]. Looking at the cohort as a whole, we found that there was a high proportion of patients presenting with insulin resistance at hospital admission, which is comparable to previous studies [15]. There was also a considerable drop in the proportion of patients with insulin resistance at D5. Accordingly, there was approximately a 23% absolute rise in HOMA2S between D1 and D5, which suggests that there is a recovery in insulin sensitivity shortly after STEMI. We have identified two different profiles of patients that behave with higher risk after STEMI. Patients from the extreme ends of the ΔHOMA2S spectrum behave similarly regarding adverse outcome, and it seems that a moderate recovery of insulin sensitivity is more desirable. As shown above, this observation remained true even after adjustment for other relevant baseline differences between groups. More importantly, our results remained significant even after adjusting for Glucose D1, ΔGlucose, and HOMA2S D1 [14], which suggests that a proper metabolic status assessment in this setting should at the very least not be limited only to blood glucose nor to an isolated admission evaluation. Although we evaluated insulin sensitivity, this assumption can be backed up by a previous study that found that multiple estimates of glucose levels outperformed admission and fasting glucose levels in post-STEMI risk assessment [13]. Another important finding was that even after stratifying long-term survival analysis by hyperglycemia status and HbA1c status, there was a clear distinction in outcomes where patients from T2 had increased survival free of MACE. Accordingly, we were able to significantly reclassify patients even after the use of GRACE score and chronic and acute dysglycemia (as assessed by HbA1c and glucose D1, respectively). In all, these observations suggest that the acute metabolic milieu at the time of STEMI is both a complex and dynamic phenomenon and that it has a tight association with clinical outcome. To the best of our knowledge, there is no prior description of a U-shaped association between the change of insulin sensitivity and clinical outcome after MI. Although in a different physiologic context, prior studies have demonstrated a J-shaped association between admission glucose levels and outcome after MI [13]. Hence, there is need of further
studies to fully explain our findings. Potentially, the absence of significant improvement in insulin sensitivity can mean that these patients aren't benefitting from positive effects that insulin exerts on the cardiovascular system. For example, insulin increases NO production [16] and myocardial perfusion [17], it enhances cardiac contractility [18], and inhibits platelet aggregation [19]. In addition, during surgically induced MI in rats, the decline in myocardial-specific insulin sensitivity occurs early after MI even without any detectable changes in systemic insulin sensitivity; this change can contribute to post-MI adverse remodeling and heart failure [20]. It is also possible that the two ends of the spectrum of insulin sensitivity change during the acute phase of MI identify different clusters of pathophysiological mechanisms that are equally significant in influencing outcome. For example, it is possible that patients who demonstrated a robust improvement in insulin sensitivity may also be more susceptible to higher amplitude of glycemic excursions, which has been significantly tied to adverse outcome in MI [21]. Another point to consider is that during myocardial ischemia glucose becomes the heart's most important source of energy through activation of MAPK [22]. Therefore, a quick recovery of peripheral insulin sensitivity may decrease glucose availability to the myocardium by driving glucose to skeletal muscle and adipose tissue. Finally, we cannot rule out the possibility that the U-shaped curve reflects STEMI-related patterns of neurohumoral activation that may simultaneously influence cardiovascular risk and insulin sensitivity. Nevertheless, this study advances a step in understanding the nature of the adaptive change in insulin sensitivity during MI and opens a door to new strategies in mechanistic and clinical studies. A few limitations of our study deserve to be addressed. First of all, the current analysis only included patients that lived until D5 of STEMI. This in itself may have introduced a selection bias to our analysis. A second issue is that the study was not designed or powered to define mechanistic aspects that can explain this finding. As far as we can tell, there were no apparent differences in inflammatory activity, endothelial function, ejection fraction, or infarct size. Thus, it is important to highlight that this is a hypothesis-generating study that describes a novel unknown element in the relationship between insulin sensitivity and STEMI and its relation to mortality. The initial motivation to perform this study stemmed from inconsistencies in subsequent glucose lowering trials, which suggested that targeting hyperglycemia is rather too simplistic. Glucose metabolism during MI is clearly not limited to glucose levels and targeting insulin sensitivity may be a more reasonable option. Hence, the estimated effect size must be considered as a motivation for further confirmatory and mechanistic studies. In conclusion, this study describes the existence of a U-shaped curve for the association between the change in insulin sensitivity during STEMI and risk. The improvement in insulin sensitivity seems to be a desirable event, but at some point there is a tipping of the scale. This study also serves to further support the notion that acute dysglycemia during STEMI is complex and it should also be further considered under the light of a time spectrum. Author contributions F.A.M researched data, conducted data analyses, interpreted results, and prepared the manuscript. R.C. and L.S.F·C researched, aided in result interpretation, and critically revised the manuscript. R.M, D.B.M., and J.C.Q.S. researched data and reviewed the manuscript. W.N.J. aided in data analysis and critically revised the manuscript. A.C.S. conceptualized the study, aided in data analysis and interpretation and critically revised the manuscript. Conflict of interest The authors report no relationships that could be construed as a conflict of interest.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.ijcard.2017.11.111. References [1] J.W. Deckers, R.T. van Domburg, M. Akkerhuis, S.T. Nauta, Relation of admission glucose levels, short- and long-term (20-year) mortality after acute myocardial infarction, Am. J. Cardiol. 112 (2013) 1306–1310. [2] M. Kosiborod, S.S. Rathore, S.E. Inzucchi, et al., Admission glucose and mortality in elderly patients hospitalized with acute myocardial infarction: implications for patients with and without recognized diabetes, Circulation 111 (2005) 3078–3086. [3] L.P. Hoebers, P. Damman, B.E. Claessen, et al., Predictive value of plasma glucose level on admission for short and long term mortality in patients with ST-elevation myocardial infarction treated with primary percutaneous coronary intervention, Am. J. Cardiol. 109 (2012) 53–59. [4] L.H. Opie, Metabolic management of acute myocardial infarction comes to the fore and extends beyond control of hyperglycemia, Circulation 117 (2008) 2172–2177. [5] K. Malmberg, Prospective randomised study of intensive insulin treatment on long term survival after acute myocardial infarction in patients with diabetes mellitus. DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction) Study Group, BMJ 314 (1997) 1512–1515. [6] A. Goyal, K.W. Mahaffey, J. Garg, et al., Prognostic significance of the change in glucose level in the first 24 h after acute myocardial infarction: results from the CARDINAL study, Eur. Heart J. 27 (2006) 1289–1297. [7] C. Lazzeri, S. Valente, M. Chiostri, P. Attana, A. Mattesini, G.F. Gensini, Acute insulin resistance assessed by the homeostatic model assessment in acute coronary syndromes without previously known diabetes, Angiology 65 (2013) 519–524. [8] F.A. Moura, L.S. Carvalho, R.M. Cintra, et al., Validation of surrogate indexes of insulin sensitivity in acute phase of myocardial infarction based on euglycemichyperinsulinemic clamp, Am. J. Physiol. Endocrinol. Metab. 306 (2014) E399–403. [9] R. Cintra, F.A. Moura, L.S. Carvalho, et al., TCF7L2 polymorphism is associated with low nitric oxide release, endothelial dysfunction and enhanced inflammatory response after myocardial infarction, BBA Clin 5 (2016) 159–165. [10] J.C. Levy, D.R. Matthews, M.P. Hermans, Correct homeostasis model assessment (HOMA) evaluation uses the computer program, Diabetes Care 21 (1998) 2191–2192.
7
[11] J.P. Henriques, F. Zijlstra, A.W. van't Hof, et al., Angiographic assessment of reperfusion in acute myocardial infarction by myocardial blush grade, Circulation 107 (2003) 2115–2119. [12] G.G. Gensini, A more meaningful scoring system for determining the severity of coronary heart disease, Am. J. Cardiol. 51 (1983) 606. [13] M. Kosiborod, S.E. Inzucchi, H.M. Krumholz, et al., Glucometrics in patients hospitalized with acute myocardial infarction: defining the optimal outcomes-based measure of risk, Circulation 117 (2008) 1018–1027. [14] C. Lazzeri, A. Sori, M. Chiostri, G.F. Gensini, S. Valente, Prognostic role of insulin resistance as assessed by homeostatic model assessment index in the acute phase of myocardial infarction in nondiabetic patients submitted to percutaneous coronary intervention, Eur. J. Anaesthesiol. 26 (2009) 856–862. [15] D. Trifunovic, S. Stankovic, D. Sobic-Saranovic, et al., Acute insulin resistance in ST-segment elevation myocardial infarction in non-diabetic patients is associated with incomplete myocardial reperfusion and impaired coronary microcirculatory function, Cardiovasc. Diabetol. 13 (2014) 73. [16] G. Zeng, M.J. Quon, Insulin-stimulated production of nitric oxide is inhibited by wortmannin. Direct measurement in vascular endothelial cells, J. Clin. Invest. 98 (1996) 894–898. [17] J. Sundell, J. Knuuti, Insulin and myocardial blood flow, Cardiovasc. Res. 57 (2003) 312–319. [18] D. Li, G. Sweeney, Q. Wang, A. Klip, Participation of PI3K and atypical PKC in Na +-K +-pump stimulation by IGF-I in VSMC, Am. J. Phys. 276 (1999) H2109–2116. [19] N.N. Kahn, W.A. Bauman, V.B. Hatcher, A.K. Sinha, Inhibition of platelet aggregation and the stimulation of prostacyclin synthesis by insulin in humans, Am. J. Phys. 265 (1993) H2160–2167. [20] F. Fu, K. Zhao, J. Li, et al., Direct evidence that myocardial insulin resistance following myocardial ischemia contributes to post-ischemic heart failure, Sci. Rep. 5 (2015), 17927. [21] G. Su, S.H. Mi, H. Tao, et al., Impact of admission glycemic variability, glucose, and glycosylated hemoglobin on major adverse cardiac events after acute myocardial infarction, Diabetes Care 36 (2013) 1026–1032. [22] F. Iliadis, N. Kadoglou, T. Didangelos, Insulin and the heart, Diabetes Res. Clin. Pract. 93 (Suppl. 1) (2011) S86–91.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after ST-Elevation Myocardial Infarction Filipe A. Moura a, Riobaldo Cintra a, Luiz S.F. Carvalho a, Simone N. Santos a, Rodrigo Modolo a, Daniel B. Munhoz a, Jose Carlos Quinaglia e Silva b, Otavio R. Coelho a, Wilson Nadruz a, Andrei C. Sposito a,⁎,1 a b
Cardiology Division, Faculty of Medical Sciences, State University of Campinas Medical School (Unicamp), Campinas, SP, Brazil Hospital de Base do Distrito Federal, Brasilia, DF, Brazil
a r t i c l e
i n f o
Article history: Received 3 February 2017 Received in revised form 19 July 2017 Accepted 30 November 2017 Available online xxxx Keywords: Insulin sensitivity HOMA Myocardial infarction STEMI Mortality
a b s t r a c t Background: Although stress hyperglycemia after myocardial infarction (MI) is consistently associated with increased mortality, recent studies suggest that the addition of upstream markers of glucose metabolism may improve risk identification. Hence, our aim was to evaluate the association between insulin sensitivity changes during MI hospitalization and outcomes. Methods: A prospective cohort of 331 consecutive ST-Elevation MI (STEMI) patients without insulin provision therapy was used for the analyses. Blood samples were collected upon admission (D1) and after 5 days (D5) of the inciting event. We measured blood glucose and insulin to estimate insulin sensitivity using the updated Homeostasis Model Assessment (HOMA2S). Patients were assessed for intra-hospital death and major adverse cardiac events (MACE) during follow-up. Results: HOMA2S was 62% ± 52% on D1 and 86% ± 57% on D5 (p b 0.001). Total follow-up was a median of 2 (0.9– 2.8) years and found a U-shaped relation between the change in HOMA2S from D1 to D5 (ΔHOMA2S) and major adverse cardiac events (MACE) (p = 0.017). Fully adjusted cox-regression models showed that patients from T1 and T3 were about 2.5 times more prone to suffer from MACE than those in T2. Net Reclassification Index adding ΔHOMA2S as a categorical variable dichotomized as T2 and T1 or T3 to a model of GRACE risk score with glucose D1 yielded a better predictive model (0.184 [95% CI 0.124–0.264]; p = 0.032). Conclusion: A U-shaped curve describes the relation between insulin sensitivity change and MACE during acute phase STEMI and, thus indicating that acute dysglycemia must be appreciated in light of a time spectrum and insulin levels. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The incidence of stress hyperglycemia in patients admitted with myocardial infarction (MI) has doubled in the last 20 years, nowadays occurring in about half of patients [1]. The biggest concern with this trend lies in the fact that hyperglycemia during acute coronary syndromes (ACS) is associated with increased mortality at 30 days, 1 year, and even 20 years after the inciting event [1], [2]. Furthermore, this association exists in both diabetic and non-diabetic patients with MI [3]. From a pathophysiological standpoint, hyperglycemia can feasibly affect outcome by decreasing collateral circulation, endothelial function, and ischemic preconditioning [4]. It can also increase myocardial cell
⁎ Corresponding author at: Cardiology Division, Faculty of Medical Sciences, State University of Campinas (Unicamp), 13084-971 Campinas, SP, Brazil. E-mail address: [email protected] (A.C. Sposito). 1 On behalf of the Brasília Heart Study Group.
apoptosis, thrombogenicity, and systemic inflammatory activity [4]. The only existing randomized study with ACS patients to date in which the glucose level was significantly lowered in the intervention group suggests there may be a causal role between the reduction in blood glucose and reduced mortality [5]. On the other hand, the association between hyperglycemia and mortality persists even after controlling for comorbidities, severity of coronary disease, and residual ventricular dysfunction, which gives way to the possibility of an indirect association or an unexplored mechanism [6]. Although the actual contribution of hyperglycemia in MI remains unclear, there are particular features that stand out. First, pre-existing diabetes seems to mitigate the risk associated with hyperglycemia [2]. Second, the increase in blood glucose during the first days after MI has greater impact on prognosis than admission blood glucose [6]. Thirdly, the degree of insulin resistance estimated during MI is associated with early mortality regardless of admission glycemia [7]. Taking all this evidence under consideration, it is reasonable to infer that the temporal variation in ST-elevation MI (STEMI) -related insulin resistance may influence the adverse effects of stress hyperglycemia.
https://doi.org/10.1016/j.ijcard.2017.11.111 0167-5273/© 2017 Elsevier B.V. All rights reserved.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
2
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
In order to assess insulin sensitivity in a larger number of STEMI patients, we conducted a preliminary study comparing surrogate insulin sensitivity indexes with the hyperinsulinemic euglycemic clamp in STEMI patients [8]. In this preliminary study, we found that the second version of the Homeostasis Model Assessment (HOMA2S) best estimates insulin sensitivity in these patients. In this context, the aim of the present study was primarily to both verify the existence of a change in insulin sensitivity, as measured by HOMA2S, during the acute phase of MI and its implications on clinical outcome. As secondary objectives, we sought to explore potential mechanisms that may mediate the association between dysglycemia and unfavorable clinical outcome.
2. Methods 2.1. Study patients Study participants were from the Brasilia Heart Study [9] (ClinicalTrials.gov Identifier: NCT02062554). It is an ongoing observational prospective cohort of consecutive ST-elevation MI (STEMI) patients admitted to Hospital de Base do Distrito Federal, a reference hospital located in Brasilia, Brazil. The current study included 331 patients enrolled between May 2006 and December 2015 and the flow diagram is illustrated in Fig. 1. Inclusion criteria were as follows: (i) b24 h after the onset of MI symptoms, (ii) ST-segment elevation ≥ 1 mm (limb leads) or ≥2 mm (precordial leads) in two contiguous leads, (iii) myocardial necrosis, as evidenced by an increase to at least one value above the 99th percentile above the reference limit of CK-MB (25 U/L) and troponin I (0.04 ng/mL) followed by a decline of both, (iv) and absence of hindrances for maintaining clinical follow-up. Individuals with previous use of insulin, insulin providers such as sulphonylureas or incretin-mimetic agents, i.e. dipeptidyl peptidase-4 inhibitors or glucagon-like peptide-1 agonists, were excluded from the study. The rationale behind this is because these represent an extremely heterogeneous population and these
pharmacotherapies alter endogenous insulin levels and intrinsic IS, thus potentially directly interfering with the interpretation of our data. Diabetes status was defined as previously diagnosed diabetes, use of hypoglycemic agents, or HbA1c ≥ 6.5%. In-hospital patient treatment was always conducted according to in-house assistant physicians and study investigators played a purely observational role. The physicians involved in the treatment of the patients were blind to all analyses performed in the study. The local Ethics Committee approved the study and all participants were required to sign an informed consent.
3. Biochemical analysis Blood samples were obtained at D1, i.e. b 24-h from MI onset, and at day five following MI (D5). Collected blood samples were centrifuged for 10 min at 3500 rpm and plasma was aliquoted for storage at −80 °C. The following biochemical panel was obtained from plasma samples and measured using standard laboratory methods: glucose, total cholesterol, triglycerides, high-density lipoprotein cholesterol, and high-sensitivity C-reactive protein. HbA1c was measured using highperformance liquid chromatography. A NO chemiluminescence analyzer (model NOA, Sievers Instruments, Boulder, CO) was used to determine the plasma pool of nitrite and nitrate (NOx) after reduction with acidic vanadium (III) chloride. Plasma insulin and C-peptide concentrations were respectively assessed by electrochemiluminescence and imunoquimioluminescence. The Homeostasis Model Assessment version 2 (HOMA2) was used to estimate insulin sensitivity (HOMA2S) using fasting plasma insulin and glucose and calculator version 2.2 [10]. The validity of HOMA2 index during STEMI was previously verified in STEMI patients by euglycemic hyperinsulinemic clamp in our laboratory [8].
Fig. 1. Flow chart of study patients.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
3
3.1. Coronary angiography
3.5. Study outcomes
Coronary angiography was systematically performed on all patients enrolled to assess coronary thrombotic burden, degree of reperfusion, and severity of atherosclerotic disease. Thrombotic burden was estimated at the culprit coronary artery according to the Thrombolysis In Myocardial Infarction (TIMI) Thrombus Grade Scale as follows: (i) Grade 0 – no angiographic evidence of thrombus is present; (ii) Grade 1 – possible thrombus, with reduced contrast density, haziness, irregular lesion contour, or a smooth convex at the site of total occlusion suggestive but not diagnostic of thrombus; (iii) Grade 2 – definite thrombus, with greatest dimensions less than half of vessel diameter; (iv) Grade 3 – definite thrombus but with greatest dimensions less than half but greater than two vessel diameters; (v) Grade 4 – definite thrombus greater than two vessel diameters; (vi) Grade 5 – total occlusion [11]. The Gensini score, a well-validated and widely used method for calculation angiographic CAD burden, was estimated coronary disease severity [12]. Coronary flow in the culprit artery was measured using TIMI flow grade and Myocardial Blush Grade (MBG), and was assessed after angioplasty in patients who underwent primary percutaneous coronary intervention (PCI) or at the earliest possible angiogram performed after thrombolysis.
The primary end point of this study was the occurrence of major adverse cardiac events (MACE), defined as in-hospital death (any death before hospital discharge or within 30 days of the index event), new fatal or non-fatal MI and sudden cardiac death (non-traumatic unexpected event occurring within 6 h of symptom onset). After discharge from the hospital, all events were assessed during the previously described outpatient visits. Absentee patients (or relatives) were contacted by telephone for evaluation of vital status. All events were adjudicated by an independent adjudication committee that did not participate in data analysis, and only the first event was chosen when a patient experienced more than one event. Patients who died from non-cardiac causes were censored on the day of their deaths.
3.2. Brachial artery reactivity Brachial artery measurements were performed using a highresolution ultrasound (IE33 and 3–9 MHz linear transducer; Philips Medical Systems, Bothell, Washington, USA) after over-night fasting 30 days after admission (D30). Vasoactive medications were withdrawn 24 h prior to assessment. After a 10-min rest period in a quiet room with the temperature controlled around 22 °C, the brachial artery was located above the elbow, and a longitudinal image of 6 to 8 cm was obtained as the resting scan. A blood pressure cuff was placed on the forearm and inflated to 50 mmHg above the systolic blood pressure for 5 min. The cuff was deflated, and the flow-mediated dilation (FMD) scan was obtained for 2 min. The percentage change in diameter for FMD was calculated in relation to the respective baseline scans. The same experienced physician, who was blinded to patients' data, made all assessments. The intra-observer reproducibility was 95%.
3.3. Cardiac magnetic resonance imaging (CMRI) CMRI studies were carried out in a subgroup of 65 patients using a MRI scanner with a 1.5-T (Signa CV/i, GE Medical Systems, Waukesha, WI), equipped with a gradient of high performance (gradient strength 40mT/m; maximum slew rate 150mT/m/s) and a four elements phased array cardiac coil. Areas of MI were quantified on the gadolinium-based delayed enhancement myocardial images at D30 to quantify MI mass, left ventricle (LV) volumes, and ejection fraction. Areas of microvascular obstruction were defined as subendocardial hypo-enhanced regions surrounded by hyperenhancement and were included as part of the core infarct. On cine-CMRI left ventricle (LV) volumes and ejection fraction were measured by ReportCard software (GE Medical Systems, Waukesha, WI), applying Simpson's method.
3.4. Follow up After hospital discharge, patients were reevaluated during visits every three months at the study outpatient clinic. All patients received guideline-based therapy and lifestyle counseling for diet, smoking cessation, regular physical activity, and weight loss. The physicians involved in outpatient care were blind to all analyses performed in the study.
3.6. Statistical analysis Groups were separated according to tertiles of ΔHOMA2S, which represents the change in HOMA2S from D1 to D5. The change in values for dependent variables was calculated as the difference between D5 and D1. Data distribution was assessed for normality with the use of histograms and Kolgomorov-Smirnov test. For continuous variables, normally distributed data are presented as mean ± standard deviation and non-normally distributed data are presented in the form of median (interquartile range). Proportional differences between groups were evaluated using chi-square. Log transformation was applied to variables in order to use parametric tests. Mean differences between groups were evaluated with one-way analysis of variance and analysis of covariance (ANCOVA) where there was need for adjustments. Mean differences between variables that persisted with skewed distribution after logtransformation were compared using the Kruskal-Wallis test (glucose, HDL-C, TG, CK-MB peak, and infarction mass). ANCOVA was used for FMD, NOx, and CRP and adjusted for age, sex, BMI, DM, hypertension, dyslipidemia, and reperfusion therapy. ANCOVA was used to compare these variables between groups after analysis with histograms, normality plots, and residual scatter plots that tested for linearity, normality, and variance. Post-hoc testing (Bonferroni) was performed to determine which groups were different when there were significant differences from ANOVA and ANCOVA analyses. Before conducting survival analysis, we explored the association between events and ΔHOMA2S because previous data has shown a J-shaped association between glucose levels and mortality [13]. The found curvilinear association between ΔHOMA2S and outcomes was assessed using restricted cubic splines with 3 knots adjusted for age, sex and baseline log-transformed HOMA2S values. Given the distribution that was found, separated the cohort in tertiles. Long-term survival between tertiles was assessed using Kaplan-Meier method and the difference between survival curves was tested with the log-rank test. Cox regression models were used to evaluate the association between tertiles of ΔHOMA2S and MACE. The number of MACE for this analysis suggested that the model would support only 4 to 5 covariates. Therefore, a propensity score for ΔHOMA2S tertiles (T2 = 1 and T1 or T3 = 0) was generated using age, sex and variables that showed significant or marginal differences among the ΔHOMA2S tertiles: hypertension, diabetes mellitus, obesity, dyslipidemia, triglycerides and reperfusion therapy. Four Cox-regression models were constructed: model 1 included HOMA2S D1; model 2 included HOMA2S D1, age and sex; model 3 included HOMA2S D1 and the propensity score; and model 4 included: HOMA2S D1, glucose D1 and the propensity score. Net reclassification improvement (NRI) and integrated diagnostic improvement (IDI) were used to assess risk reclassification when adding ΔHOMA2S dichotomized as T2 versus T1 or T3 to the GRACE score and glucose D1. A two- sided p-value of b 0.05 was considered statistically significant. All analyses were conducted using SPSS software version 20.0 or STATA software version 10.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
4
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
4. Results
Table 1 Characteristics of study patients stratified by tertiles of ΔHOMA2S.
4.1. Baseline characteristics Patients were organized according to tertiles of ΔHOMA2S, which was defined as the difference between D5 and D1 HOMA2S levels. Cut-offs for tertiles of ΔHOMA2S were as follows: Tertile 1 (T1) ≤ 2.9% (n = 111), 2.9% b Tertile 2 (T2) ≤ 41.0% (n = 111), and Tertile 3 (T3) N 41.0% (n = 109). In a very simplified manner, the tertiles may be interpreted in the following way: T1 - little to no improvement in insulin sensitivity; T2 - improvement in insulin sensitivity; T3 marked improvement in insulin sensitivity. Clinical characteristics are shown on Table 1. Of note, patients from T1 and T2 had lower prevalence of hypertension. Also, patients from T3 demonstrated lower BMI than T2 (p = 0.005), but there was no difference between patients from T1 and T2 (p = 0.883) or between T1 and T3 (p = 0.109). Waist circumference behaved in a similar fashion; there were only differences between T2 and T3 (p = 0.001). There was also a lower proportion of patients with BMI ≥ 30 kg/m2 in T3. Patients from T2 had a lower prevalence of sedentary lifestyle than other groups. No other differences between demographics, medical history, medical evaluation, or hemodynamics were found. Finally, Patients from T1 and T2 received higher rates of tenectaplase, but considering both tenecteplase and PCI, reperfusion therapies were similar between groups. There were no other significant differences in treatment between groups. 4.2. Glucose metabolism and biochemical data We found that 82% of patients presented with declined insulin sensitivity (HOMA2S b 100%) on D1 while 68% of patients presented this on D5 after STEMI. Accordingly, there was improvement of HOMA2S from D1 to D5 of STEMI (D1 62 ± 52% vs. D5 86 ± 57%, p b 0.001). Patients at the higher tertile of ΔHOMA2S had higher levels of glucose D1 and lower levels of glucose D5. There was also higher prevalence of glucose ≥120 mg/dL in T3, but there was no difference in prevalence of glucose ≥140 mg/dL. Also, groups from T2 and T3 had higher values of insulin on D1 and lower levels of insulin D5. These findings were expected due to the manner in which groups were separated. Of note, T1 (p = 0.061) and T2 (p = 0.001) had lower TG levels when compared to T3 using post-hoc tests. 4.3. Endothelial function, infarction size, and angiographic assessment We found no differences in endothelial function, as assessed by NOx or FMD, and in inflammatory markers, according to ΔHOMA2S tertiles. There was also no significant difference in infarct size between groups as estimated by CKMB-peak or by CMRI. Angiographic data showed that patients from T2 had a higher proportion of patients with significant left main coronary lesion. All other characteristics were similar between groups (Supplementary Table 1). 4.4. Clinical follow-up and outcomes Clinical follow-up time lasted for a median of 2.0 (0.9 to 2.8) years. During follow-up, 49 MACE occurred (17 intra-hospital deaths, 9 fatal MI, 17 nonfatal MI, and 6 sudden cardiac deaths). Kaplan-Meier curve estimation (Fig. 2A) showed that patients in T2 have a higher rate of survival free of MACE (T1 18.9% vs. 7.2% vs. T3 18.3%, p = 0.025). Multiple cox-regression models showed that patients from T1 and T3 were about 2.5 times more prone to suffer from MACE (Supplementary Table 2). Finally, adding only T2 to the cox-regression model and adjusting for age, sex, HOMA2S D1, hypertension, diabetes, dyslipidemia, BMI, tenecteplase, and beta-blocker use yielded a HR of 0.449 (95% CI 0.204–0.986; p = 0.046), further suggesting better survival in this group. Additional adjustment for admission glucose (HR 0.439 [95% CI
ΔHOMA2S ≤2.9%
N2.9 and ≤41%
N41%
N
111
111
109
Demographics Age, years Male, % BMI, kg/m2 BMI ≥ 30 kg/m2, % Waist circumference, cm
60 ± 12 78 26.8 ± 4 23 96.8 ± 11
57 ± 11 73 27.4 ± 4 19 98.9 ± 11
60 ± 11 77 25.7 ± 4 10 93.4 ± 10
0.064 0.571 0.006 0.035 0.001
10.6
6.2
6.1
0.347
37.2 65.5 14.2 38.9 63.7 54.9 131(180)
38.1 49.6 12.4 29.2 47.8 53.1 90(158)
46.9 51.8 5.3 24.6 57.9 38.9 120(192)
0.258 0.033 0.07 0.057 0.05 0.057 0.286
133 ± 29 85 ± 17 76 ± 15 92
134 ± 32 83 ± 20 75 ± 18 91
138 ± 30 86 ± 21 76 ± 17 86
0.413 0.375 0.959 0.32
5.9(0.8) 113(33) 40.4 17.5 113(32) 11.1 ± 8 21.2 ± 18 86.3(87) 44.7(45) −37.2(45)
5.8(0.9) 119(37) 49.6 24.8 103(25) 24.5 ± 13 14.1 ± 6 34.0(29) 56.4(33) 61.2(57)
5.8(0.7) 127(31) 62.6 28.7 97(16) 24.8 ± 15 6.4 ± 3 31.7(38) 126.0(83) 251.6(236)
0.954 b0.001 0.003 0.132 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
69 ± 19 36(13) 116 ± 37 133(124) 0.8(1.5) 3.47(5.9) 1.90(4.9) 214(261) 14.6(12)
71 ± 19 35(13) 125 ± 43 138(103) 0.6(1.3) 3.60(5.3) 2.14(4.3) 224(239) 18.3(11)
69 ± 23 38(17) 123 ± 40 109(71) 0.7(1.3) 3.27(7.8) 2.35(6.1) 250(288) 13.5(11)
0.782 0.216 0.201 0.001 0.750 0.087 0.087 0.903 0.633
64.9 11.6 76.8 99.1 69.4 51.4 62.2
67.3 12.8 81.7 97.3 67.3 61.9 62.8
52.2 17.6 69.4 99.1 64.0 54.9 74.6
0.042 0.404 0.106 0.447 0.693 0.265 0.084
Medical history Prior MI, % Risk factors Smoking, % Hypertension, % Diabetes Mellitus, % Dyslipidemia, % Sedentarity, % Family history for CAD, % Arrival time, min Hemodynamics Systolic Blood Pressure, mmHg Diastolic Blood Pressure, mmHg Heart Rate, bpm Killip class I, % Biochemical analyses HbA1c, % ^ Glucose D1, mg/dL Glucose D1 ≥ 120 mg/dL, % Glucose D1 ≥ 140 mg/dL, % ^ Glucose D5, mg/dL Insulin D1, μU/mL Insulin D5, μU/mL HOMA2S D1, % HOMA2S D5, % (HOMA2S D5 – HOMA2S D2)/ HOMA2S D1 Glomerular filtration rate, units ^ HDL-C, mg/dL LDL-C, mg/dL ^ TG, mg/dL ⁎CRP D1, mg/L ⁎CRP D5, mg/L ⁎ΔCRP, mg/L ^ CK-MB peak, mg/dL ^ Infarction mass (CMRI), g Treatment Tenecteplase, % Primary PCI, % Reperfusion therapy, % Aspirin, % Simvastatin, % ACEi or ARB, % Betablocker, %
p Value
⁎ Analysis of covariance adjusted by age, sex, smoking, BMI, reperfusion therapy, and basal levels (where fitting). D1; 1st day of MI; D5: 5th day post-MI; BMI: body mass index; CAD: coronary artery disease; CRP: C reactive protein; CMRI: Cardiac Magnetic Resonance Imaging. ^ Kruskal-Wallis test.
0.202–0.953; p = 0.037]) or ΔGlucose (HR 0.437 [95% CI 0.202–0.946; p = 0.036]) yielded similar results. As shown in Fig. 3A, when treating ΔHOMA2S as a continuous variable, the relationship of ΔHOMA2S to MACE was U shaped. The same result was obtained using the Quantitative Insulin Check Index (QUICKI) model and long-term MACE (Supplementary Fig. 1). Furthermore, reclassification analysis adding ΔHOMA2S as a categorical variable dichotomized as T2 and T1 or T3 to a model of GRACE risk score with glucose D1 yielded better predictive models with increases in NRI (index =
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
%Survival free of events
A
5
100
ΔHOMA2S T1 ΔHOMA2S T2 ΔHOMA2S T3
80 60 40 20 0
0
400
800
1200
1600
Days post STEMI
%Survival free of events
B
100
Δ HOMA2S T1 Δ HOMA2S T2 Δ HOMA2S T3
80 60 40 20 0
0
400
800
1200
1600
Days post STEMI Fig. 2. Kaplan-Meier curves. A) Kaplan-Meier curves showing survival free of MACE of STEMI patients (not taking insulin or secretagogues) according to tertiles of ΔHOMA2S. B) KaplanMeier curves showing survival free of MACE of non-diabetic STEMI patients according to tertiles of ΔHOMA2S.
Fig. 3. Adjusted hazard ratio of major adverse cardiac events (MACE) in function of difference in HOMA2S between D5 and D1 (ΔHOMA2S) in the whole population. Data were adjusted for age, sex, and baseline HOMA2S. The continuous line represents the hazard ratio (HR) and the black dashed lines the 95% confidence interval (CI). ΔHOMA2S = 21.5% (median) was considered the reference (HR = 1).
0.184 [95% CI 0.124–0.264]; p = 0.032) and IDI (index = 0.056 [95% CI 0.0390–0.073]; p = 0.001). This was further observed after KaplanMeier assessment stratified by the presence of hyperglycemia (defined as glucose D1 ≥ 140 mg/dL). Once again, patients from T2 demonstrated increased survival free of MACE both in patients with and without hyperglycemia upon hospital admission (p = 0.017). The trend for increased MACE-free survival of T2 was found both in patients with high HbA1c (≤5.8%) and those with low HbA1c (N 5.8%) after Kaplan-Meier curve estimation stratified by HbA1c status (p = 0.018). We also conducted an analysis using the Forest Plot method in order to better understand which type of events (intra-hospital death or postdischarge events) were influencing differences in MACE between groups. Post-discharge MACEs seem to be more pronouncedly different between groups (Supplementary Fig. 2). Finally, as shown in Fig. 2B, we ran the same analysis strictly with non-diabetic patients (n = 258) and found a similar difference between groups (T1 19.8% vs. T2 6.8% vs. T3 22.6%, p = 0.011). After cox-regression analysis, patients from T1 (HR 2.913, 95% CI 1.139– 7.446; p = 0.026) and T3 (HR 3.753, 95% CI 1.498–9.404; p = 0.005) presented with higher long-term risk than patients from T2. This observation persisted even after adjusting for age and sex (T1 HR 2.521, p = 0.057; T3 HR 3.271, p = 0.012) (Supplementary Table 3).
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
6
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
5. Discussion Using a subgroup of acute-phase STEMI patients that did not previously use insulin provision or incretin-based therapies, we have shown that there is a pronounced change in insulin sensitivity between the first and the fifth day of STEMI. This very change in insulin sensitivity during the acute phase of STEMI is strongly tied to outcome, even after adjusting for more established markers such as admission hyperglycemia. In particular, the relation between the recovery in insulin sensitivity and adverse outcome is not linear but is in fact U-shaped. In light of this, patients in whom insulin sensitivity recovers either too vigorously or who don't recover at all both have at least approximately a 3-fold risk of intra-hospital death and a 2-fold increase in incidence of MACE during long-term follow-up. Over the past years, the lack of consensus in insulin-mediated glucose control trials during MI has resulted in hyperglycemia loosing traction as a causal agent of adverse clinical outcome after MI. However, the derangement in glucose metabolism that occurs during MI is complex and recent evidences suggest that efforts should not be merely focused on blood glucose level control. For example, we have previously shown that endogenous insulin secretory function is associated with intrahospital mortality [9]. Furthermore, a previous group has also shown that admission HOMA-IR, another popular surrogate for insulin sensitivity, may be directly associated to risk after MI [7,14]. Hence, a fresh perspective on the subject was warranted to help unveil more of what occurs in this setting. As commented above, to better assess the role of insulin sensitivity during STEMI, we first investigated whether surrogate indexes were feasible for assessing insulin sensitivity during STEMI. We found that HOMA2S and QUICKI were the methods that best corresponded to hyperinsulinemic euglycemic clamps and that HOMA-IR performed inferiorly to all others [8]. Looking at the cohort as a whole, we found that there was a high proportion of patients presenting with insulin resistance at hospital admission, which is comparable to previous studies [15]. There was also a considerable drop in the proportion of patients with insulin resistance at D5. Accordingly, there was approximately a 23% absolute rise in HOMA2S between D1 and D5, which suggests that there is a recovery in insulin sensitivity shortly after STEMI. We have identified two different profiles of patients that behave with higher risk after STEMI. Patients from the extreme ends of the ΔHOMA2S spectrum behave similarly regarding adverse outcome, and it seems that a moderate recovery of insulin sensitivity is more desirable. As shown above, this observation remained true even after adjustment for other relevant baseline differences between groups. More importantly, our results remained significant even after adjusting for Glucose D1, ΔGlucose, and HOMA2S D1 [14], which suggests that a proper metabolic status assessment in this setting should at the very least not be limited only to blood glucose nor to an isolated admission evaluation. Although we evaluated insulin sensitivity, this assumption can be backed up by a previous study that found that multiple estimates of glucose levels outperformed admission and fasting glucose levels in post-STEMI risk assessment [13]. Another important finding was that even after stratifying long-term survival analysis by hyperglycemia status and HbA1c status, there was a clear distinction in outcomes where patients from T2 had increased survival free of MACE. Accordingly, we were able to significantly reclassify patients even after the use of GRACE score and chronic and acute dysglycemia (as assessed by HbA1c and glucose D1, respectively). In all, these observations suggest that the acute metabolic milieu at the time of STEMI is both a complex and dynamic phenomenon and that it has a tight association with clinical outcome. To the best of our knowledge, there is no prior description of a U-shaped association between the change of insulin sensitivity and clinical outcome after MI. Although in a different physiologic context, prior studies have demonstrated a J-shaped association between admission glucose levels and outcome after MI [13]. Hence, there is need of further
studies to fully explain our findings. Potentially, the absence of significant improvement in insulin sensitivity can mean that these patients aren't benefitting from positive effects that insulin exerts on the cardiovascular system. For example, insulin increases NO production [16] and myocardial perfusion [17], it enhances cardiac contractility [18], and inhibits platelet aggregation [19]. In addition, during surgically induced MI in rats, the decline in myocardial-specific insulin sensitivity occurs early after MI even without any detectable changes in systemic insulin sensitivity; this change can contribute to post-MI adverse remodeling and heart failure [20]. It is also possible that the two ends of the spectrum of insulin sensitivity change during the acute phase of MI identify different clusters of pathophysiological mechanisms that are equally significant in influencing outcome. For example, it is possible that patients who demonstrated a robust improvement in insulin sensitivity may also be more susceptible to higher amplitude of glycemic excursions, which has been significantly tied to adverse outcome in MI [21]. Another point to consider is that during myocardial ischemia glucose becomes the heart's most important source of energy through activation of MAPK [22]. Therefore, a quick recovery of peripheral insulin sensitivity may decrease glucose availability to the myocardium by driving glucose to skeletal muscle and adipose tissue. Finally, we cannot rule out the possibility that the U-shaped curve reflects STEMI-related patterns of neurohumoral activation that may simultaneously influence cardiovascular risk and insulin sensitivity. Nevertheless, this study advances a step in understanding the nature of the adaptive change in insulin sensitivity during MI and opens a door to new strategies in mechanistic and clinical studies. A few limitations of our study deserve to be addressed. First of all, the current analysis only included patients that lived until D5 of STEMI. This in itself may have introduced a selection bias to our analysis. A second issue is that the study was not designed or powered to define mechanistic aspects that can explain this finding. As far as we can tell, there were no apparent differences in inflammatory activity, endothelial function, ejection fraction, or infarct size. Thus, it is important to highlight that this is a hypothesis-generating study that describes a novel unknown element in the relationship between insulin sensitivity and STEMI and its relation to mortality. The initial motivation to perform this study stemmed from inconsistencies in subsequent glucose lowering trials, which suggested that targeting hyperglycemia is rather too simplistic. Glucose metabolism during MI is clearly not limited to glucose levels and targeting insulin sensitivity may be a more reasonable option. Hence, the estimated effect size must be considered as a motivation for further confirmatory and mechanistic studies. In conclusion, this study describes the existence of a U-shaped curve for the association between the change in insulin sensitivity during STEMI and risk. The improvement in insulin sensitivity seems to be a desirable event, but at some point there is a tipping of the scale. This study also serves to further support the notion that acute dysglycemia during STEMI is complex and it should also be further considered under the light of a time spectrum. Author contributions F.A.M researched data, conducted data analyses, interpreted results, and prepared the manuscript. R.C. and L.S.F·C researched, aided in result interpretation, and critically revised the manuscript. R.M, D.B.M., and J.C.Q.S. researched data and reviewed the manuscript. W.N.J. aided in data analysis and critically revised the manuscript. A.C.S. conceptualized the study, aided in data analysis and interpretation and critically revised the manuscript. Conflict of interest The authors report no relationships that could be construed as a conflict of interest.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111
F.A. Moura et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.ijcard.2017.11.111. References [1] J.W. Deckers, R.T. van Domburg, M. Akkerhuis, S.T. Nauta, Relation of admission glucose levels, short- and long-term (20-year) mortality after acute myocardial infarction, Am. J. Cardiol. 112 (2013) 1306–1310. [2] M. Kosiborod, S.S. Rathore, S.E. Inzucchi, et al., Admission glucose and mortality in elderly patients hospitalized with acute myocardial infarction: implications for patients with and without recognized diabetes, Circulation 111 (2005) 3078–3086. [3] L.P. Hoebers, P. Damman, B.E. Claessen, et al., Predictive value of plasma glucose level on admission for short and long term mortality in patients with ST-elevation myocardial infarction treated with primary percutaneous coronary intervention, Am. J. Cardiol. 109 (2012) 53–59. [4] L.H. Opie, Metabolic management of acute myocardial infarction comes to the fore and extends beyond control of hyperglycemia, Circulation 117 (2008) 2172–2177. [5] K. Malmberg, Prospective randomised study of intensive insulin treatment on long term survival after acute myocardial infarction in patients with diabetes mellitus. DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction) Study Group, BMJ 314 (1997) 1512–1515. [6] A. Goyal, K.W. Mahaffey, J. Garg, et al., Prognostic significance of the change in glucose level in the first 24 h after acute myocardial infarction: results from the CARDINAL study, Eur. Heart J. 27 (2006) 1289–1297. [7] C. Lazzeri, S. Valente, M. Chiostri, P. Attana, A. Mattesini, G.F. Gensini, Acute insulin resistance assessed by the homeostatic model assessment in acute coronary syndromes without previously known diabetes, Angiology 65 (2013) 519–524. [8] F.A. Moura, L.S. Carvalho, R.M. Cintra, et al., Validation of surrogate indexes of insulin sensitivity in acute phase of myocardial infarction based on euglycemichyperinsulinemic clamp, Am. J. Physiol. Endocrinol. Metab. 306 (2014) E399–403. [9] R. Cintra, F.A. Moura, L.S. Carvalho, et al., TCF7L2 polymorphism is associated with low nitric oxide release, endothelial dysfunction and enhanced inflammatory response after myocardial infarction, BBA Clin 5 (2016) 159–165. [10] J.C. Levy, D.R. Matthews, M.P. Hermans, Correct homeostasis model assessment (HOMA) evaluation uses the computer program, Diabetes Care 21 (1998) 2191–2192.
7
[11] J.P. Henriques, F. Zijlstra, A.W. van't Hof, et al., Angiographic assessment of reperfusion in acute myocardial infarction by myocardial blush grade, Circulation 107 (2003) 2115–2119. [12] G.G. Gensini, A more meaningful scoring system for determining the severity of coronary heart disease, Am. J. Cardiol. 51 (1983) 606. [13] M. Kosiborod, S.E. Inzucchi, H.M. Krumholz, et al., Glucometrics in patients hospitalized with acute myocardial infarction: defining the optimal outcomes-based measure of risk, Circulation 117 (2008) 1018–1027. [14] C. Lazzeri, A. Sori, M. Chiostri, G.F. Gensini, S. Valente, Prognostic role of insulin resistance as assessed by homeostatic model assessment index in the acute phase of myocardial infarction in nondiabetic patients submitted to percutaneous coronary intervention, Eur. J. Anaesthesiol. 26 (2009) 856–862. [15] D. Trifunovic, S. Stankovic, D. Sobic-Saranovic, et al., Acute insulin resistance in ST-segment elevation myocardial infarction in non-diabetic patients is associated with incomplete myocardial reperfusion and impaired coronary microcirculatory function, Cardiovasc. Diabetol. 13 (2014) 73. [16] G. Zeng, M.J. Quon, Insulin-stimulated production of nitric oxide is inhibited by wortmannin. Direct measurement in vascular endothelial cells, J. Clin. Invest. 98 (1996) 894–898. [17] J. Sundell, J. Knuuti, Insulin and myocardial blood flow, Cardiovasc. Res. 57 (2003) 312–319. [18] D. Li, G. Sweeney, Q. Wang, A. Klip, Participation of PI3K and atypical PKC in Na +-K +-pump stimulation by IGF-I in VSMC, Am. J. Phys. 276 (1999) H2109–2116. [19] N.N. Kahn, W.A. Bauman, V.B. Hatcher, A.K. Sinha, Inhibition of platelet aggregation and the stimulation of prostacyclin synthesis by insulin in humans, Am. J. Phys. 265 (1993) H2160–2167. [20] F. Fu, K. Zhao, J. Li, et al., Direct evidence that myocardial insulin resistance following myocardial ischemia contributes to post-ischemic heart failure, Sci. Rep. 5 (2015), 17927. [21] G. Su, S.H. Mi, H. Tao, et al., Impact of admission glycemic variability, glucose, and glycosylated hemoglobin on major adverse cardiac events after acute myocardial infarction, Diabetes Care 36 (2013) 1026–1032. [22] F. Iliadis, N. Kadoglou, T. Didangelos, Insulin and the heart, Diabetes Res. Clin. Pract. 93 (Suppl. 1) (2011) S86–91.
Please cite this article as: F.A. Moura, et al., Adverse outcome has a U-shaped relation with acute phase change in insulin sensitivity after STElevation Myocardial Infarction, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.11.111