Original article
Glycosylated haemoglobin is associated with neurohormonal activation and poor outcome in chronic heart failure patients with mild left ventricular systolic dysfunction Luigi Emilio Pastormerloa, Chiara Mamminia, Alberto Giannonib, Alessandro Valleggia, Concetta Pronteraa, Alessandra Gabuttia, Roberta Polettia, Luigi Padelettib, Michele Emdina and Claudio Passinoa Aims We aimed to evaluate the impact of glycometabolic imbalance as assessed by glycosylated haemoglobin [HbA(1c)] on neurohormonal activation and outcome in chronic heart failure (CHF). Methods and results Nine hundred and twenty CHF patients (65 W 12 years, left ventricular ejection fraction 33 W 10%, 29% diabetic patients) underwent a thorough humoral and clinical characterization, including HbA(1c), and were then followed up for the endpoint of cardiac death. In the whole population, diagnosis of diabetes resulted in no difference in neurohormonal or echocardiographic data, or in outcome. Conversely, the diabetic patients with HbA(1c) above 7% showed, in comparison to both diabetic patients with HbA(1c) below 7% and non-diabetic individuals, higher plasma renin activity (1.81, 0.48–5.68 vs. 1.23, 0.43–2.8 and 1.29, 0.44–5 ng/ml/h, respectively; P < 0.01 for both), Nterminal pro-brain natriuretic peptide (NT-pro-BNP) (1602, 826–3498 vs. 1022, 500–3543 and 1134, 455–3545 ng/l, respectively; P < 0.01 for both) and worse symptoms with a higher rate of cardiac mortality vs. both diabetic patients with HbA1(c) below 7% and non-diabetic individuals (P < 0.05 for both). In the left ventricular ejection fraction 38–50% tertile (mild left ventricular dysfunction), elevated HbA(1c) was associated with higher NT-pro-BNP and PRA
Introduction The prevalence and incidence of diabetes mellitus are higher among chronic heart failure (CHF) patients than in the general population.1,2 The previous studies aimed to assess the prognostic weight of diabetes mellitus in CHF showed conflicting results. Whereas some studies found diabetes mellitus to be an independent prognostic marker in CHF patients,3,4 other studies failed to find any significant difference in survival between CHF patients with and without diabetes mellitus.5,6 The most likely explanation for this might be searched in the study characteristic heterogeneity and, particularly, in different severity of left ventricular dysfunction. A recent metaanalysis demonstrated that the risk of all-cause mortality and hospitalization in CHF patients with diabetes mellitus is significantly higher than in non-diabetic individuals,7 whereas the reasons for this excess mortality have not yet been fully understood. 1558-2027 Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved.
(P < 0.01), and, alongside NT-pro-BNP, resulted the only independent predictor of outcome beyond diagnosis of diabetes. HbA(1c) failed to show up differences in neuroendocrine activation or in outcome in moderate and severe left ventricular dysfunction tertiles. Conclusion Glycometabolic imbalance, as represented by HbA(1c), is associated with neurohormonal activation and poor prognosis in CHF patients, beyond diabetes. The impact of metabolic derangement on prognosis appears greater at the early stages of CHF, when it might exacerbate neurohormonal activation. J Cardiovasc Med 2015, 16:423–430 Keywords: diabetes, glycaemic control, heart failure, NT-pro-BNP, plasma renin activity a Division of Cardiovascular Medicine, Fondazione G. Monasterio CNR-Regione Toscana, Pisa and bDepartment of Heart and Vessels, University of Florence, Florence, Italy
Correspondence to Luigi Emilio Pastormerlo, MD, Division of Cardiovascular Medicine, Fondazione Toscana Gabriele Monasterio CNR-Regione Toscana, Via Giuseppe Moruzzi 1, 56124 Pisa, Italy Tel: +39 050 3152216; fax: +39 050 3152109; e-mail: [email protected] Received 6 February 2014 Revised 5 May 2014 Accepted 6 May 2014
Neurohormonal activation as a maladaptive response to cardiac damage plays a key role in the evolution of CHF disease.8 Patients’ outcome strictly depends on chronic activation of the neuroendocrine axes, namely of the sympathetic and renin–angiotensin–aldosterone (RAA) systems, counteracted by cardiac endocrine function.9,10 Diabetes mellitus has been demonstrated to influence the function of these systems, through several pathways.11,12 Chronic glycometabolic derangement, measured by glycosylated haemoglobin A1c [HbA(1c)] assay, could impact on disease evolution and patient outcome, possibly sustaining neurohormonal activation, influencing cardiovascular remodelling, and causing other organ damage. However, the scarce previous literature provided contradictory results.13,14 Therefore, in this study, we aimed to evaluate the impact of glycometabolic imbalance, beyond the diagnosis of DOI:10.2459/JCM.0000000000000159
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
424 Journal of Cardiovascular Medicine 2015, Vol 16 No 6
diabetes mellitus, on neurohormonal activation and outcome in CHF patients, and to assess whether this effect would be dependent on the severity of left ventricular systolic dysfunction.
Methods Ethics statement
The present study complies with the Declaration of Helsinki; the Local Ethics Committee has approved the research protocol and informed consent has been obtained from all the patients. Study population
Between June 2004 and November 2010, we prospectively enrolled 920 consecutive patients with CHF [left ventricular ejection fraction (LVEF) 50%], who were admitted to the cardiology subintensive unit or in the outpatient clinic of the Division of Cardiovascular Medicine at Fondazione G. Monasterio in Pisa for heart failure management. The diagnosis of heart failure was determined by history, symptoms, and physical and instrumental findings for the assessment of structural myocardial involvement. Acute coronary syndrome within 6 months before the enrolment was the only exclusion criterion. Cardiac morphology and systolic and diastolic function were assessed by two-dimensional Doppler echocardiography. Patients were classified as diabetic if they were treated with oral hypoglycaemic drugs or insulin, or had a documented history of elevated (>126 mg/dl) fasting blood glucose on at least two separate occasions in conjunction with ongoing dietary measures, or yet if they had documented serum glucose above 200 mg/dl 2 h after the oral administration of 75 g of glucose. Seven per cent of HbA(1c) was used as a cut-off, indicating adequate glycaemic control, as suggested by guidelines.15 At the time of the HbA(1c) assay, patients were not on tight control of blood glucose. Anti-diabetic therapy was adjusted, if needed, at the first characterization and during the follow-up according to good clinical practice, with no constraints. All patients were on optimal pharmacologic treatment for CHF at the maximum tolerated dose of beta-blockers, and angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers. Biohumoral evaluation
All patients underwent humoral evaluation at baseline – blood chemistry, natriuretic peptides, catecholamines, plasma renin activity (PRA), aldosterone – as previously described in detail.8 The blood samples were drawn at 8:00 a.m. after overnight fasting and a 30-min rest in supine position, through an indwelling venous cannula, according to a standardized experimental protocol.16 N-terminal pro-brain natriuretic peptide (NT-pro-BNP) was assessed with ECLIA (Elecsys 2010 analyzer, Roche Diagnostics, Basel, Switzerland); norepinephrine was assayed by HPLC (Chromosystems Diagnostics, GmbH, Munchen, Germany). PRA and aldosterone were
measured by radioimmunoassay (DiaSorin S.r.l., Saluggia, Italy). Thyroid-stimulating hormone and free thyroid hormones were measured by sandwich and competitive enzyme immunoassays, respectively (EurogeneticsTosoh, Turin, Italy). Finally, HbA(1c) was evaluated by HPLC (HLC-723 G8, automated glycohaemoglobin analyser, Tosoh, Turin, Italy). This is a standardized dosage, as certified by the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC). The correlation coefficient and coefficient of variation were, respectively, 0.9998 and 0.21%. Estimated glomerular filtration rate (eGFR) was estimated using Cockroft–Gault equation. Follow-up
Follow-up started at blood sampling and continued until study termination (i.e. June 2011). Independent interviewers obtained follow-up data directly from patients, relatives, Institute cardiologists or general practitioners. The primary endpoints were all-cause death and cardiac death (due to myocardial infarction, worsening heart failure or sudden death). Information about the time and cause of death was obtained from death certificates, post-mortem reports or family doctors. Patients who underwent heart transplantation or ventricular assistance device implantation were considered censored at the time of the event. No patient was lost to follow-up. Statistical analysis
Statistical analysis was carried out using SPSS 12.0 (SPSS inc., Chicago, Illinois, USA). Groups were compared for categorical data or frequency of events using the chisquare test and for continuous variables using the Student’s t test or analysis of variance, as appropriate. Logtransformed values of the original data were used in parametric and regression statistical analyses for continuous variables (such as NT-pro-BNP, norepinephrine and PRA), known to share a log-normal distribution. Receiver-operating curve analysis for cardiac mortality with the evaluation of area under the curve was also performed for NT-pro-BNP and HbA(1c). For subgroup analysis, the patients were divided into tertiles according to LVEF. Univariate and multivariate survival analyses were performed with the Cox proportional-hazards model; proportional-hazards assumptions were checked by the use of Shoenefeld’s residuals. The continuous variables [age, LVEF, free tri-iodothyronine, free tetraiodothyronine, thyroid-stimulating hormone, BMI, eGFR, haemoglobin, NT-pro-BNP, norepinephrine, PRA, aldosterone, and HbA(1c)] and the dichotomized variables [sex, hypertension, diabetes mellitus, dyslipidaemia, New York Heart Association (NYHA) functional class I–II/III–IV] were entered into the Cox proportional-hazards regression model to identify univariate predictors of mortality. All variables significantly associated with outcome at univariate analysis were then entered into multivariate
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Glycometabolic imbalance in heart failure Pastormerlo et al. 425
analysis. In the tertile survival analysis, a maximum of two covariates (with all possible combinations of univariate predictive variables) was used because of the limited number of events. Survival curves were analysed using the Kaplan–Meier estimate. Differences in survival curves were tested with the log-rank test. Values are presented as mean SD, or median and interquartile range (for variables with non-normal distribution). A P value less than 0.05 was considered significant.
Results Patients’ clinical and neurohormonal characteristics
The baseline characteristics of the entire study population are presented in Table 1, according to the presence/absence of diabetes mellitus and, as concern diabetic patients, HbA(1c) at least or less than 7%. Out of the 920 patients, 276 (29%) presented with a diagnosis of diabetes mellitus. The diabetic patients were older and had a higher BMI than the non-diabetic ones. Overall, the two groups did not show significant differences in CHF cause, symptom severity, renal and thyroid function, neurohormonal activation, and echocardiographic findings. Conversely, diabetic patients with HbA(1c) at least 7% (n ¼ 118) showed, in comparison to both diabetic patients with HbA(1c) less than 7% (n ¼ 158) and non-diabetic individuals, higher values of PRA (1.81, 0.48–5.68 vs.
1.23, 0.43–2.8 and 1.29, 0.44–5 ng/ml/h, respectively; P < 0.01 for both) and NT-pro-BNP (1602, 826–3498 vs. 1022, 500–3543 and 1134, 455–3545 ng/l, respectively; P < 0.01 for both) with worse clinical status [43% of patients with HbA(1c) 7 being in NYHA class III/IV vs. 36% of patients with HbA(1c)
Glycosylated haemoglobin is associated with neurohormonal activation and poor outcome in chronic heart failure patients with mild left ventricular systolic dysfunction Luigi Emilio Pastormerloa, Chiara Mamminia, Alberto Giannonib, Alessandro Valleggia, Concetta Pronteraa, Alessandra Gabuttia, Roberta Polettia, Luigi Padelettib, Michele Emdina and Claudio Passinoa Aims We aimed to evaluate the impact of glycometabolic imbalance as assessed by glycosylated haemoglobin [HbA(1c)] on neurohormonal activation and outcome in chronic heart failure (CHF). Methods and results Nine hundred and twenty CHF patients (65 W 12 years, left ventricular ejection fraction 33 W 10%, 29% diabetic patients) underwent a thorough humoral and clinical characterization, including HbA(1c), and were then followed up for the endpoint of cardiac death. In the whole population, diagnosis of diabetes resulted in no difference in neurohormonal or echocardiographic data, or in outcome. Conversely, the diabetic patients with HbA(1c) above 7% showed, in comparison to both diabetic patients with HbA(1c) below 7% and non-diabetic individuals, higher plasma renin activity (1.81, 0.48–5.68 vs. 1.23, 0.43–2.8 and 1.29, 0.44–5 ng/ml/h, respectively; P < 0.01 for both), Nterminal pro-brain natriuretic peptide (NT-pro-BNP) (1602, 826–3498 vs. 1022, 500–3543 and 1134, 455–3545 ng/l, respectively; P < 0.01 for both) and worse symptoms with a higher rate of cardiac mortality vs. both diabetic patients with HbA1(c) below 7% and non-diabetic individuals (P < 0.05 for both). In the left ventricular ejection fraction 38–50% tertile (mild left ventricular dysfunction), elevated HbA(1c) was associated with higher NT-pro-BNP and PRA
Introduction The prevalence and incidence of diabetes mellitus are higher among chronic heart failure (CHF) patients than in the general population.1,2 The previous studies aimed to assess the prognostic weight of diabetes mellitus in CHF showed conflicting results. Whereas some studies found diabetes mellitus to be an independent prognostic marker in CHF patients,3,4 other studies failed to find any significant difference in survival between CHF patients with and without diabetes mellitus.5,6 The most likely explanation for this might be searched in the study characteristic heterogeneity and, particularly, in different severity of left ventricular dysfunction. A recent metaanalysis demonstrated that the risk of all-cause mortality and hospitalization in CHF patients with diabetes mellitus is significantly higher than in non-diabetic individuals,7 whereas the reasons for this excess mortality have not yet been fully understood. 1558-2027 Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved.
(P < 0.01), and, alongside NT-pro-BNP, resulted the only independent predictor of outcome beyond diagnosis of diabetes. HbA(1c) failed to show up differences in neuroendocrine activation or in outcome in moderate and severe left ventricular dysfunction tertiles. Conclusion Glycometabolic imbalance, as represented by HbA(1c), is associated with neurohormonal activation and poor prognosis in CHF patients, beyond diabetes. The impact of metabolic derangement on prognosis appears greater at the early stages of CHF, when it might exacerbate neurohormonal activation. J Cardiovasc Med 2015, 16:423–430 Keywords: diabetes, glycaemic control, heart failure, NT-pro-BNP, plasma renin activity a Division of Cardiovascular Medicine, Fondazione G. Monasterio CNR-Regione Toscana, Pisa and bDepartment of Heart and Vessels, University of Florence, Florence, Italy
Correspondence to Luigi Emilio Pastormerlo, MD, Division of Cardiovascular Medicine, Fondazione Toscana Gabriele Monasterio CNR-Regione Toscana, Via Giuseppe Moruzzi 1, 56124 Pisa, Italy Tel: +39 050 3152216; fax: +39 050 3152109; e-mail: [email protected] Received 6 February 2014 Revised 5 May 2014 Accepted 6 May 2014
Neurohormonal activation as a maladaptive response to cardiac damage plays a key role in the evolution of CHF disease.8 Patients’ outcome strictly depends on chronic activation of the neuroendocrine axes, namely of the sympathetic and renin–angiotensin–aldosterone (RAA) systems, counteracted by cardiac endocrine function.9,10 Diabetes mellitus has been demonstrated to influence the function of these systems, through several pathways.11,12 Chronic glycometabolic derangement, measured by glycosylated haemoglobin A1c [HbA(1c)] assay, could impact on disease evolution and patient outcome, possibly sustaining neurohormonal activation, influencing cardiovascular remodelling, and causing other organ damage. However, the scarce previous literature provided contradictory results.13,14 Therefore, in this study, we aimed to evaluate the impact of glycometabolic imbalance, beyond the diagnosis of DOI:10.2459/JCM.0000000000000159
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
424 Journal of Cardiovascular Medicine 2015, Vol 16 No 6
diabetes mellitus, on neurohormonal activation and outcome in CHF patients, and to assess whether this effect would be dependent on the severity of left ventricular systolic dysfunction.
Methods Ethics statement
The present study complies with the Declaration of Helsinki; the Local Ethics Committee has approved the research protocol and informed consent has been obtained from all the patients. Study population
Between June 2004 and November 2010, we prospectively enrolled 920 consecutive patients with CHF [left ventricular ejection fraction (LVEF) 50%], who were admitted to the cardiology subintensive unit or in the outpatient clinic of the Division of Cardiovascular Medicine at Fondazione G. Monasterio in Pisa for heart failure management. The diagnosis of heart failure was determined by history, symptoms, and physical and instrumental findings for the assessment of structural myocardial involvement. Acute coronary syndrome within 6 months before the enrolment was the only exclusion criterion. Cardiac morphology and systolic and diastolic function were assessed by two-dimensional Doppler echocardiography. Patients were classified as diabetic if they were treated with oral hypoglycaemic drugs or insulin, or had a documented history of elevated (>126 mg/dl) fasting blood glucose on at least two separate occasions in conjunction with ongoing dietary measures, or yet if they had documented serum glucose above 200 mg/dl 2 h after the oral administration of 75 g of glucose. Seven per cent of HbA(1c) was used as a cut-off, indicating adequate glycaemic control, as suggested by guidelines.15 At the time of the HbA(1c) assay, patients were not on tight control of blood glucose. Anti-diabetic therapy was adjusted, if needed, at the first characterization and during the follow-up according to good clinical practice, with no constraints. All patients were on optimal pharmacologic treatment for CHF at the maximum tolerated dose of beta-blockers, and angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers. Biohumoral evaluation
All patients underwent humoral evaluation at baseline – blood chemistry, natriuretic peptides, catecholamines, plasma renin activity (PRA), aldosterone – as previously described in detail.8 The blood samples were drawn at 8:00 a.m. after overnight fasting and a 30-min rest in supine position, through an indwelling venous cannula, according to a standardized experimental protocol.16 N-terminal pro-brain natriuretic peptide (NT-pro-BNP) was assessed with ECLIA (Elecsys 2010 analyzer, Roche Diagnostics, Basel, Switzerland); norepinephrine was assayed by HPLC (Chromosystems Diagnostics, GmbH, Munchen, Germany). PRA and aldosterone were
measured by radioimmunoassay (DiaSorin S.r.l., Saluggia, Italy). Thyroid-stimulating hormone and free thyroid hormones were measured by sandwich and competitive enzyme immunoassays, respectively (EurogeneticsTosoh, Turin, Italy). Finally, HbA(1c) was evaluated by HPLC (HLC-723 G8, automated glycohaemoglobin analyser, Tosoh, Turin, Italy). This is a standardized dosage, as certified by the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC). The correlation coefficient and coefficient of variation were, respectively, 0.9998 and 0.21%. Estimated glomerular filtration rate (eGFR) was estimated using Cockroft–Gault equation. Follow-up
Follow-up started at blood sampling and continued until study termination (i.e. June 2011). Independent interviewers obtained follow-up data directly from patients, relatives, Institute cardiologists or general practitioners. The primary endpoints were all-cause death and cardiac death (due to myocardial infarction, worsening heart failure or sudden death). Information about the time and cause of death was obtained from death certificates, post-mortem reports or family doctors. Patients who underwent heart transplantation or ventricular assistance device implantation were considered censored at the time of the event. No patient was lost to follow-up. Statistical analysis
Statistical analysis was carried out using SPSS 12.0 (SPSS inc., Chicago, Illinois, USA). Groups were compared for categorical data or frequency of events using the chisquare test and for continuous variables using the Student’s t test or analysis of variance, as appropriate. Logtransformed values of the original data were used in parametric and regression statistical analyses for continuous variables (such as NT-pro-BNP, norepinephrine and PRA), known to share a log-normal distribution. Receiver-operating curve analysis for cardiac mortality with the evaluation of area under the curve was also performed for NT-pro-BNP and HbA(1c). For subgroup analysis, the patients were divided into tertiles according to LVEF. Univariate and multivariate survival analyses were performed with the Cox proportional-hazards model; proportional-hazards assumptions were checked by the use of Shoenefeld’s residuals. The continuous variables [age, LVEF, free tri-iodothyronine, free tetraiodothyronine, thyroid-stimulating hormone, BMI, eGFR, haemoglobin, NT-pro-BNP, norepinephrine, PRA, aldosterone, and HbA(1c)] and the dichotomized variables [sex, hypertension, diabetes mellitus, dyslipidaemia, New York Heart Association (NYHA) functional class I–II/III–IV] were entered into the Cox proportional-hazards regression model to identify univariate predictors of mortality. All variables significantly associated with outcome at univariate analysis were then entered into multivariate
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Glycometabolic imbalance in heart failure Pastormerlo et al. 425
analysis. In the tertile survival analysis, a maximum of two covariates (with all possible combinations of univariate predictive variables) was used because of the limited number of events. Survival curves were analysed using the Kaplan–Meier estimate. Differences in survival curves were tested with the log-rank test. Values are presented as mean SD, or median and interquartile range (for variables with non-normal distribution). A P value less than 0.05 was considered significant.
Results Patients’ clinical and neurohormonal characteristics
The baseline characteristics of the entire study population are presented in Table 1, according to the presence/absence of diabetes mellitus and, as concern diabetic patients, HbA(1c) at least or less than 7%. Out of the 920 patients, 276 (29%) presented with a diagnosis of diabetes mellitus. The diabetic patients were older and had a higher BMI than the non-diabetic ones. Overall, the two groups did not show significant differences in CHF cause, symptom severity, renal and thyroid function, neurohormonal activation, and echocardiographic findings. Conversely, diabetic patients with HbA(1c) at least 7% (n ¼ 118) showed, in comparison to both diabetic patients with HbA(1c) less than 7% (n ¼ 158) and non-diabetic individuals, higher values of PRA (1.81, 0.48–5.68 vs.
1.23, 0.43–2.8 and 1.29, 0.44–5 ng/ml/h, respectively; P < 0.01 for both) and NT-pro-BNP (1602, 826–3498 vs. 1022, 500–3543 and 1134, 455–3545 ng/l, respectively; P < 0.01 for both) with worse clinical status [43% of patients with HbA(1c) 7 being in NYHA class III/IV vs. 36% of patients with HbA(1c)
