Canadian Journal of Cardiology 30 (2014) 1529e1534
Clinical Research
Osteoprotegerin Is Associated With Subclinical Left Ventricular Systolic Dysfunction in Diabetic Hypertensive Patients: A Speckle Tracking Study Ezgi Kalaycıoglu, MD,a Tayyar Gökdeniz, MD,b Ahmet Çagrı Aykan, MD,a Engin Hatem, MD,c Mustafa Ozan Gürsoy, MD,d Asım Ören, MD,e Hüseyin Yaman, MD,e Ays¸e Gül Karadeniz, MD,f and S¸ükrü Çelik, MDa a
Department of Cardiology, Ahi Evren Chest and Cardiovascular Surgery Education and Research Hospital, Trabzon, Turkey b c
e f
Department of Cardiology, Kafkas University, Faculty of Medicine, Kars, Turkey
Department of Cardiology, Erzurum Regional Training and Research Hospital, Erzurum, Turkey d Department of Cardiology, Gaziemir Salih Nevvar I_ ¸s gören State Hospital, I_ zmir, Turkey
Department of Medical Biochemistry, Karadeniz Technical University Faculty of Medicine, Trabzon, Turkey
Department of Radiology, Ahi Evren Chest and Cardiovascular Surgery Education and Research Hospital, Trabzon, Turkey
See editorial by Putko et al., on pages 1492-1495 of this issue. ABSTRACT
RESUM E
Background: Recently, the role of osteoprotegerin (OPG) in the pathogenesis of heart failure through different mechanisms has received much attention. Subclinical changes in left ventricular (LV) function can be identified using quantification of myocardial strain, and global longitudinal strain (GLS) is a superior predictor of outcomes than ejection fraction. We hypothesized that increased OPG levels could predict subclinical LV systolic dysfunction in treated diabetic hypertensive patients with preserved LV ejection fraction. Methods: The study was composed of 86 diabetic hypertensive and 30 nondiabetic hypertensive patients. All patients underwent echocardiography and venous blood samples were taken for determination of OPG. The relation between OPG levels and LV GLS was investigated using 2-dimensional speckle tracking echocardiography. Results: Diabetic hypertensive patients had higher diastolic peak
cemment, le rôle de l’oste oprote ge rine (OPG) dans la Introduction : Re pathogenèse de l’insuffisance cardiaque par le biais de ses divers canismes a fait l’objet d’une grande attention. La quantification de me formation myocardique peut de terminer les modifications subla de formation cliniques de la fonction du ventricule gauche (VG), mais la de dit mieux les re sultats que la fraction longitudinale globale (DLG) pre jection. Nous avons pose l’hypothèse que l’augmentation des cond’e dire la dysfonction systolique subclinique centrations de l’OPG peut pre tiques traite s ayant une fraction du VG des patients hypertendus diabe jection VG pre serve e. d’e thodes : L’e tude e tait compose e de 86 patients hypertendus diaMe tiques et de 30 patients hypertendus non diabe tiques. Tous les be chocardiographie et un pre lèvement patients ont subi une e chantillons de sang par ponction veineuse pour de terminer l’OPG. d’e
Osteoprotegerin (OPG) is a soluble member of the tumour necrosis factor (TNF) receptor superfamily with pleiotropic effects on bone metabolism, endocrine function, and the immune system.1 OPG plays a regulatory role in inflammatory pathways, acting as a decoy for the receptor activator of nuclear factor kb (RANK) ligand (RANKL), competitively inhibiting interactions between RANK and RANKL.2
Recently, the role of the OPG/RANKL/RANK axis in the pathogenesis of heart failure (HF) through different mechanisms such as promotion of matrix degradation and inflammation1 has received much attention.3 OPG has been linked to diabetes mellitus (DM), silent myocardial ischemia, acute myocardial infarction,4 and microvascular complications.5 It has been shown that increased plasma OPG is a strong predictor of all-cause mortality independent of conventional risk for cardiovascular disease (CVD) in patients with type 2 DM (T2-DM).6 Previously, an association between OPG and left ventricular (LV) dysfunction in the general population was also described.1 However, there is no previous study regarding the relationship between OPG and LV dysfunction in diabetic hypertensive patients. T2-DM is one of the major risk factors
Received for publication June 22, 2014. Accepted August 19, 2014. Corresponding author: Dr Ezgi Kalaycıoglu, Ahi Evren Chest and Cardiovascular Surgery Education and Research Hospital, Soguksu M, Camlık Cad, 61040 Trabzon, Turkey. Tel.: þ90-5305606260. E-mail: [email protected] See page 1533 for disclosure information.
http://dx.doi.org/10.1016/j.cjca.2014.08.018 0828-282X/Ó 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.
1530
Canadian Journal of Cardiology Volume 30 2014
early/early diastolic tissue velocity and lower systolic tissue velocity, GLS, GLS rate systolic, and GLS rate early diastolic than nondiabetic hypertensive patients (P ¼ 0.009, P ¼ 0.049, P < 0.001, P ¼ 0.004, and P < 0.001, respectively). Diabetic hypertensive patients were divided into 2 groups according to median GLS value (> 18.5 and 18.5). The patients with GLS 18.5 had higher diastolic blood pressure (mm Hg; P ¼ 0.048), OPG (pmol/L; P < 0.001), and hemoglobin A1c (%; P ¼ 0.042) values than those with GLS > 18.5. In multivariate logistic regression analysis, OPG was found to be an independent predictor of impaired GLS (P ¼ 0.001). Receiver operating characteristic curve analysis revealed that OPG values of > 6.45 (pmol/L) identified the patients with GLS 18.5. Conclusions: Plasma OPG values could predict subclinical LV systolic dysfunction in diabetic hypertensive patients.
te La relation entre les concentrations d’OPG et la DLG du VG a e tudie e par e chocardiographie bidimensionnelle Speckle Tracking e (suivi de pixel). sultats : Les patients hypertendus diabe tiques avaient une ve locite Re coce maximale/une ve locite tissulaire diatissulaire diastolique pre coce plus e leve es et une ve locite tissulaire systolique, une stolique pre DLG, un taux de DLG durant la systole et un taux DLG durant la diastole coce plus faibles que les patients hypertendus non diabe tiques (P pre ¼ 0,009, P ¼ 0,049, P< 0,001, P ¼ 0,004, et P< 0,001, respectiques ont e te divise s en 2 tivement). Les patients hypertendus diabe diane de la DLG (> 18,5 et 18,5). Les groupes selon la valeur me patients ayant une DLG 18,5 avaient des valeurs de pression rielle diastolique (mm Hg; P ¼ 0,048), d’OPG (pmol/l; P< 0,001) arte moglobine A1c (%; P ¼ 0,042) plus e leve es que ceux ayant une et d’he gression logistique multivarie e, l’OPG DLG > 18,5. Dans l’analyse de re tait ave re e un pre dicteur inde pendant de la de te rioration de la DLG s’e ristique d’efficacite du (P ¼ 0,001). L’analyse de la courbe caracte cepteur re ve lait que des valeurs d’OPG > 6,45 (pmol/l) re terminaient les patients ayant une DLG 18,5. de dire la Conclusions : Les valeurs plasmatiques de l’OPG pourraient pre dysfonction systolique subclinique du VG des patients hypertendus tiques. diabe
for the development of CVD, particularly when it is associated with hypertension (HT)7 and markers of CVD are needed in DM to identify patients at risk of severe complications.4 Global LV systolic function, most commonly assessed according to echocardiographic ejection fraction (EF), is an important predictor of cardiovascular outcome. However, the measurement of EF presents a number of challenges related to image quality, assumption of LV geometry, and expertise.8 Also, EF is known not to be a sensitive marker for the detection of subclinical LV systolic dysfunction.9 Twodimensional speckle tracking echocardiography (2D-STE) is a novel technique used for the measurement of cardiac mechanics. It assesses myocardial deformation and the myocardial deformation rate by tracking speckles in the myocardium on greyscale (B-mode) images and can be used to evaluate global and regional myocardial strain and strain rate.10 Longitudinal tissue deformation is evaluated using frame-by-frame tracking of individual speckles throughout the cardiac cycle, and global longitudinal strain (GLS) is calculated.8 Previous studies have reported that subclinical changes in LV function can be identified using quantification of myocardial strain,9 and GLS is superior to EF as a predictor of outcomes.8 In addition, this technique has some advantages over conventional echocardiography including angle independency, free of tethering and translation effects, low signal-to-noise ratio, and low measurement variability.11,12 We hypothesized that increased OPG levels could predict subclinical LV systolic dysfunction in treated diabetic hypertensive patients with preserved LV EF. This study was the first to evaluate this issue in current literature.
diabetic hypertensive patients and 30 consecutive nondiabetic hypertensive patients. All patients were receiving antihypertensive and/or antidiabetic treatment for at least 1 year. Exclusion criteria were echocardiographic evidence of either regional or global wall motion abnormalities, LV EF < 50%, history of significant coronary artery disease (CAD) and coronary artery bypass graft surgery, angina pectoris, positive treadmill test or nuclear perfusion stress test, ischemic electrocardiographic findings, chronic renal failure, atrial fibrillation/flutter, second- or third-degree atrioventricular block, moderate to severe valvular heart disease, secondary HT, pregnancy, neoplasia, systemic inflammatory disease, and known immune system or connective tissue disease. HT was diagnosed when the mean of 3 separate blood pressure (BP) measurements were 140 mm Hg for systolic BP (SBP) and/or 90 mm Hg for diastolic BP (DBP).13 Diabetes was diagnosed when fasting glucose was 126 mg/dL or 200 mg/dL 2 hours after oral glucose overload (repeated twice) or after detection of symptoms of diabetes and random blood glucose 200 mg/dL.14 Informed consent was obtained from the patients and the study was approved by the local ethics board. Body mass index (BMI) was calculated as weight (kg)/height (m2). The body surface area was calculated according to the following formula in square metres: 0.007184 weight (kg)0.425 height (cm)0.725.
Methods Study population From September 2012 to June 2013, we conducted a single-centre cross-sectional study including 86 consecutive
Laboratory measurements Estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease study equation.15 Blood samples for determination of OPG were collected in EDTA tubes and centrifuged, and the isolated plasma samples were stored at 80 C until analysis. Plasma OPG was measured using a sandwich enzyme-linked immunosorbent assay using commercially available antibodies (Biovendor). Urine albumin was assessed quantitatively using an immunoturbidimetric test (Beckman Coulter).
Kalaycıo glu et al. Osteoprotegerin and Left Ventricular Dysfunction
Assessment of carotid intima media thickness Carotid arteries were investigated in the longitudinal and the transverse projections with B-mode sonography (Mindray M7 DC7) using a 7-MHz linear probe. The intima media thickness of both carotid arteries was always measured manually on the common carotid artery outside the plaque, if any was present. Each measurement was calculated using the average of 3 readings.16 Conventional echocardiography The echocardiographic studies were performed using a commercially available echocardiography machine (VIVID S-5 General Electric Medical System, Vingmed Ultrasound AS, Horten, Norway) equipped with a 3.6-MHz transducer and tissue Doppler imaging. Measurements were performed according to the American Society of Echocardiography guidelines,17 by a cardiologist. LV dimensions (end-diastolic and end-systolic) and wall thickness (septum and posterior wall) were obtained from the parasternal long axis with an Mmode. LV EF was measured using the modified Simpson method. LV mass was calculated according to Devereux formula and was indexed to the body surface area. The relative wall thickness was calculated as the ratio (2 posterior wall thickness/LV internal dimension at end-diastole). Mitral inflow velocities were evaluated using pulse-wave Doppler with the sample volume placed at the tip of the mitral leaflets from the apical 4-chamber view. Using the average of 3 beats, we measured diastolic peak early (E), peak late transmitral flow velocity (A), peak E to peak A velocities (E/A) and deceleration time of peak E velocity. Systolic tissue velocity (S’) and early diastolic tissue velocity (E’) were acquired from the septal and lateral sides of the annulus using the tissue Doppler imaging and the pulse wave Doppler mode and measured using averaged values. To evaluate LV filling pressures, the ratio of E/E’ was calculated. Two-dimensional speckle tracking analysis is shown in the Two-dimensional Speckle Tracking Analysis section of the Supplementary Material. Details on reproducibility are shown in the Reproducibility section of the Supplementary Material. Statistical analysis SPSS 17.0 statistical software (SPSS Inc, Chicago, IL) was used for statistical analysis. The Kolmogorov-Smirnov test was used to test the normality of distribution of continuous variables. Continuous variables are expressed as mean standard deviation (SD) or median (interquartile range), and categorical variables are expressed as percentage, as appropriate. Continuous variables were compared using Student t test or the Mann-Whitney U test. Comparisons for categorical variables were evaluated using the c2 test. To avoid collinearity in assessing the multivariate model, independent variables were tested for intercorrelation. Pearson or Spearman correlation analysis was used for assessing correlation between GLS and continuous variables depending on Gaussian distributions. Diabetic hypertensive patients were divided into 2 groups according to the median value of GLS (> 18.5 and 18.5). Multivariate logistic regression analysis was performed to find independent associates of impaired GLS ( 18.5).
1531
Independent variables such as DBP, SBP, S’, OPG, hemoglobin (Hb) A1c, age, female sex, E/A ratio, and eGFR were entered into the model. Interobserver agreement of echocardiographic parameters obtained from 2D-STE data was calculated using a Bland-Altman analysis and intraclass correlation coefficient was used to assess intraobserver agreement. Receiver operating characteristic curve analysis was performed to detect the cutoff value of the OPG in predicting GLS ( 18.5). For further analysis of the independent predictors of GLS, multivariate linear regression analysis was performed. Logarithmic transformation was performed for variables that were not normally distributed including S’ and HbA1c before they were entered into multivariate regression analysis. DBP, SBP, Log10-S’, OPG, Log10-HbA1c, age, female sex, glomerular filtration rate, and E/A ratio were entered into the model. A 2-tailed P < 0.05 was considered statistically significant. Results The study population consisted of 86 (55.8% female) diabetic hypertensive and 30 (73.3% female) nondiabetic hypertensive patients. The clinical, biochemical, and echocardiographic characteristics of the study population are presented in Supplemental Table S1. Diabetic hypertensive patients had lower eGFR (86.40 16.20 mL/min vs 92.00 9.40 mL/min; P ¼ 0.024) and higher microalbumin (1.1 [2.87] mg/dL vs 0.5 [0.64] mg/dL; P ¼ 0.001), carotid intima media thickness (CIMT; 0.8 [0.3] mm vs 0.5 [0.32] mm; P < 0.001), OPG (7.37 2.53 pmol/L vs 5.32 2.17 pmol/L; P < 0.001) values than nondiabetic hypertensive patients. Diabetic hypertensive patients had higher E/E’ and lower S’ (m/s) than nondiabetic hypertensive patients (7.06 [3.38] vs 6.2 [2.47]; P ¼ 0.009 and 0.09 0.08 vs 0.12 0.02; P ¼ 0.049). Furthermore, LV systolic functions were more impaired in diabetic hypertensive patients compared with nondiabetic hypertensive patients (GLS, 18.74 2.89% vs 20.87 1.96%; P < 0.001; GLS rate systolic, 1.10 0.24 L/s vs 1.24 0.19 L/s; P ¼ 0.004; GLS rate early diastolic, 1.12 0.35 L/s vs 1.46 0.32 L/s; P < 0.001, respectively). Diabetic hypertensive patients were divided into 2 groups according to median GLS values (> 18.5 and 18.5). The characteristics of diabetic hypertensive patients according to GLS values are shown in Supplemental Table S2. The patients with GLS 18.5 had higher DBP (91.89 10.20 mm Hg vs 87.31 10.94 mm Hg; P ¼ 0.048), OPG (8.40 2.20 pmol/L vs 6.30 2.45 pmol/L; P < 0.001) and HbA1c (7.90 [1.49]% vs 7.25 [1.54]%; P ¼ 0.042) values than those with GLS > 18.5. To find independent predictors of impaired GLS ( 18.5), multivariate logistic regression analysis was performed. Independent variables such as DBP (mm Hg), SBP (mm Hg), S’ (m/s), OPG (pmol/L), HbA1c (%), age (years), female sex, E/ A ratio, and eGFR (mL/min) were entered into the model. OPG was found to be an independent predictor of impaired GLS ( 18.5; P ¼ 0.001; Supplemental Table S3). The correlation analysis between OPG and GLS values is shown in Figure 1A. A multivariate linear regression model was built to find the independent determinants of GLS. DBP (mm Hg), SBP (mm Hg), Log10-S’, OPG (pmol/L), age (years), female sex, eGFR
1532
Canadian Journal of Cardiology Volume 30 2014
Figure 1. (A) The correlation between OPG and GLS values. (B) ROC curve analysis of OPG in predicting GLS ( 18.5). GLS, global longitudinal strain; OPG, osteoprotegerin; ROC, receiver operating characteristic.
(mL/min), and E/A ratio were entered in the model. OPG and Log10-S’ (systolic tissue velocity; m/s) were independent predictors of impaired GLS (b ¼ 0.534; P < 0.001 and b ¼ 4.591; P ¼ 0.003, respectively; Supplemental Table S4). Receiver operating characteristic curve analysis was performed in diabetic hypertensive patients to detect the cutoff value of OPG in predicting patients with GLS ( 18.5). The analysis showed that OPG values of > 6.45 (pmol/L) identified the patients with GLS ( 18.5) with a specificity of 69.1% and a sensitivity of 79.5% (area under the curve, 0.755; 95% confidence interval, 0.650-0.859; P < 0.001; Fig. 1B). Partial correlation analysis was performed in all subjects to determine the partial correlation between OPG and GLS adjusted for smoking, fasting plasma glucose, BMI, age, and female sex. We found that OPG was negatively correlated with GLS (P < 0.001; r ¼ 0.653). Univariate analysis showed that OPG (pmol/L) was correlated with HbA1c percentage (r ¼ 0.347; P < 0.001), CIMT (mm; r ¼ 0.243; P ¼ 0.009), eGFR (r ¼ 0.194; P ¼ 0.037) and E/E’ (r ¼ 225; P ¼ 0.015); but not correlated with SBP (mm Hg; P ¼ 0.067), DBP (mm Hg; P ¼ 0.157), pulse pressure (PP; r ¼ 0.113; P ¼ 0.226) and LV mass index (g/m2; P ¼ 0.754) in all study populations. OPG (pmol/L) was correlated with E/E’ (r ¼ 223; P ¼ 0.039), but no correlation was detected between OPG and HbA1c percentage (P ¼ 0.087), CIMT (mm; P ¼ 0.209), SBP (mm Hg; P ¼ 0.098), DBP (mm Hg; P ¼ 0.104), PP (r ¼ 0.074; P ¼ 0.501), and LV mass index (g/m2; P ¼ 0.113) in diabetic hypertensive patients. Discussion To our knowledge we showed for the first time in this study, that plasma OPG values could predict subclinical LV systolic dysfunction in diabetic hypertensive patients who were receiving medical treatment and had no previous history of CVD. A cutoff value > 6.45 pmol/L was able to identify the patients with subclinical LV systolic dysfunction. Diabetes is known to be associated with the development of HF even without the presence of coexisting CAD.9,18 In addition, the clinical and morphologic features of heart disease in diabetic hypertensive patients are more severe than in those with hypertension or diabetes alone.19 DM is associated with myocardial structural alterations, leading to progressive
deterioration of cardiac hemodynamics. Diabetic cardiomyopathy seems to be composed of 2 stages. In the first stage, physiological adaptation occurs because of metabolic changes. In the second stage, degenerative changes start and the myocardium might have limited capacity for repair. Therefore, early treatment might delay the progression of the severity of cardiomyopathy.19 The early stage of diabetic cardiomyopathy includes minor changes in myocardial structure (such as normal LV dimension, wall thickness, and mass) and cardiac dysfunction usually can only be detected using sensitive methods such as strain and strain rate in this stage.19,20 Systolic dysfunction might be initially apparent in the longitudinal direction, because subendocardial fibres, which are the ones more vulnerable to myocardial ischemia and fibrosis, are longitudinally oriented.21 It has been shown that GLS correlates well with EF measured using echocardiography22 and magnetic resonance imaging (MRI).23 However, they measure different aspects of myocardial function, with EF measuring radial and partly longitudinal function, and GLS measuring longitudinal function.8 Persistent inflammation appears to play a role in the development of HF.24 In particular, TNF-a has been implicated as a possible pathogenic factor in this process.24 OPG is a member of the TNF receptor superfamily.2 Recent studies have shown that other members of the TNF/TNF receptor superfamily also play a pathogenic role in chronic HF. Several authors suggested that among several TNF superfamily members that have been shown to be activated in HF, the OPG/RANK/RANKL axis might be of importance in the pathogenesis of this disorder3 by mediating autocrine, paracrine, and endocrine interactions between cells expressing RANK.3 The binding of RANKL to RANK promotes the activity of matrix metalloproteinase, leading to degradation of extracellular matrix and increased apoptosis.2 Previous studies suggested that experimental and clinical HF were associated with increased expression of the OPG/RANKL/RANK axis. Because RANKL and RANK are difficult to be measured in vivo2 and OPG circulates at much higher levels than RANKL it might be a more stable overall measure of RANKL/RANK activity and more suited for clinical use than quantification of circulating RANKL concentrations.1 Omland et al.1 evaluated the relationship between OPG levels and cardiac magnetic resonance indices of LV structure and functions in the general population. They suggested that
Kalaycıo glu et al. Osteoprotegerin and Left Ventricular Dysfunction
higher levels of OPG were independently associated with higher LV end-systolic volume and lower EF in both sexes. OPG levels were found to be increased in subjects with only mildly impaired LV systolic function, suggesting that the OPG/RANKL/RANK system activation is an early phenomenon in the process of ventricular dysfunction and HF development.1 To generalize our findings, we evaluated the potential relationship between OPG and GLS in all hypertensive subjects. After adjustment for potential confounders (smoking, fasting plasma glucose, BMI, age, and sex), OPG remained significantly associated with GLS. The association between OPG levels and endothelial dysfunction, arterial stiffness, and atherosclerosis was shown previously.4 Avignon et al.25 investigated OPG concentrations and signs of myocardial ischemia on myocardial perfusion scintigraphy in patients with DM and concluded that OPG measurements can help to better define the diabetic population with an increased risk of developing silent myocardial ischemia. Higher levels of OPG were documented in T2-DM patients with asymptomatic CAD, and also in experimental and clinical HF, even in the absence of ischemic cardiomyopathy.26 The pathophysiological connection between plasma OPG concentrations and CVD is not clearly known, but a correlation with arterial and myocardial disease has been suggested.4 We did not evaluate the inducible myocardial ischemia in our study population, so we can speculate that the relationship between the OPG/RANK/RANK axis and subclinical LV dysfunction might be attributed to endothelial dysfunction and/or the negative remodelling effect on the LV myocardium. Blázquez-Medela et al.27 found significantly higher levels of OPG in diabetic hypertensive patients than in hypertensive, nondiabetic patients and in diabetic, nonhypertensive patients. They also found that HbA1c was significantly correlated with OPG in all study populations (including healthy control subjects) and in the hypertensive group, but there was no correlation in the diabetic hypertensive patients group. Consistent with the findings of this study, our results showed that plasma OPG levels were higher in diabetic hypertensive patients than in nondiabetic hypertensive patients, and HbA1c was found to be significantly correlated with OPG in all study populations but there was no correlation in diabetic hypertensive patients. The underlying mechanisms behind the correlation between OPG and glycemic status remain unknown, but might be related to possible regulatory effects of OPG production from vascular cells and osteoblasts because OPG synthesis is regulated by insulin-like growth factor-I in osteoblasts, and by insulin and TNF-a in endothelial cells.28 Several previous studies reported that OPG was related to insulin resistance in diabetic and nondiabetic subjects.29 Altinova et al.29 found a significant positive correlation between serum OPG and a homeostasis model assessment of insulin resistance in diabetic patients. O’Sullivan et al.30 found that OPG was negatively correlated with fasting plasma glucose in healthy subjects, but there was no correlation between OPG and fasting plasma glucose and HbA1c in diabetic patients, suggesting that serum OPG might be determined by different factors in different populations. Taken together, the authors suggested that more studies are needed to understand the exact role of OPG in glucose metabolism.29 Several studies examined the association between OPG and BP with conflicting results.1 In the current study, we found
1533
that the diabetic hypertensive patients with impaired LV systolic function (GLS 18.5) had more elevated DBP values. In correlation analysis, we did not find any correlation between OPG and SBP-DBP, or PP values. In concordance with our study, Blázquez-Medela et al.27 suggested that OPG values were correlated with SBP and PP in all study populations (including healthy control subjects), but there was no correlation in diabetic hypertensive patients. Kim et al.31 suggested that OPG values were significantly correlated with SBP, but their study population included patients newly diagnosed with DM and they excluded patients with previously diagnosed HT who were receiving medical treatment. Coutinho et al.32 reported a relationship between PP and OPG in hypertensive patients. The association between OPG and HT was explained by the generalized matrix changes, including OPG accumulation in the arterial wall, and therefore increased arterial stiffness.28 However, it was previously shown that angiotensin II blockade could downregulate OPG secretion in vitro.27,33 In the current study, no association was detected between OPG and BP or PP; this was probably because of the use of angiotensin-converting enzyme inhibitors-angiotensin receptor blockers in most of the patients. In the study by Coutinho et al.32 a minority of the subjects were treated with angiotensin converting enzyme inhibitors-angiotensin receptor blockers. Therefore, we can speculate that a relationship between OPG and BP or PP might be detected in newly diagnosed hypertensive subjects who are not yet taking medication or treated with angiotensin converting enzyme inhibitors-angiotensin receptor blockers. Study limitations Several limitations are present in our study. The sample size of subjects was relatively small to generalize our findings. Although MRI is the gold standard for myocardial strain imaging, 2D-STE correlates well with MRI. The inclusion of the healthy control subjects would have increased the power of the study. Conclusion Our study shows the emerging role of OPG as an indicator of subclinical LV systolic dysfunction in diabetic hypertensive patients. It is evidently crucial to establish biochemical markers of increased risk for CVD events. These markers could be used in the clinical setting for the early diagnosis of subclinical LV systolic dysfunction, which would allow for strategies to be designed to reduce the cardiovascular event rate in those patients. Further studies are needed to establish whether increased OPG levels in diabetic hypertensive patients can in fact predict later development of LV systolic dysfunction. Disclosures The authors have no conflicts of interest to disclose. References 1. Omland T, Drazner MH, Ueland T, et al. Plasma osteoprotegerin levels in the general population: relation to indices of left ventricular structure and function. Hypertension 2007;49:1392-8.
1534
Canadian Journal of Cardiology Volume 30 2014
2. Gupta S, Drazner MH, de Lemos JA. Newer biomarkers in heart failure. Heart Fail Clin 2009;5:579-88.
19. Fang ZY, Prins JB, Marwick TH. Diabetic cardiomyopathy: evidence, mechanisms, and therapeutic implications. Endocr Rev 2004;25:543-67.
3. Ueland T, Yndestad A, Dahl CP, Gullestad L, Aukrust P. TNF revisited: osteoprotegerin and TNF-related molecules in heart failure. Curr Heart Fail Rep 2012;9:92-100.
20. Fang ZY, Yuda S, Anderson V, et al. Echocardiographic detection of early diabetic myocardial disease. J Am Coll Cardiol 2003;41:611-7.
4. Augoulea A, Vrachnis N, Lambrinoudaki I, et al. Osteoprotegerin as a marker of atherosclerosis in diabetic patients. Int J Endocrinol 2013;2013:182060. 5. Knudsen ST, Foss CH, Poulsen PL, et al. Increased plasma concentrations of osteoprotegerin in type 2 diabetic patients with microvascular complications. Eur J Endocrinol 2003;149:39-42. 6. Reinhard H, Lajer M, Gall MA, et al. Osteoprotegerin and mortality in type 2 diabetic patients. Diabetes Care 2010;33:2561-6. 7. U.K. Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPD 38. BMJ 1998;317:703-13. 8. Stanton T, Leano R, Marwick TH. Prediction of all-cause mortality from global longitudinal speckle strain: comparison with ejection fraction and wall motion scoring. Circ Cardiovasc Imaging 2009;2:356-64. 9. Nakai H, Takeuchi M, Nishikage T, Lang RM, Otsuji Y. Subclinical left ventricular dysfunction in asymptomatic diabetic patients assessed by two-dimensional speckle tracking echocardiography: correlation with diabetic duration. Eur J Echocardiogr 2009;10:926-32. 10. Kalaycıoglu E, Gökdeniz T, Aykan AC, et al. Evaluation of right ventricle functions and serotonin levels during headache attacks in migraine patients with aura. Int J Cardiovasc Imaging 2014;30:1255-63. 11. Crendal E, Walther G, Dutheil F, et al. Left ventricular myocardial dyssynchrony is already present in nondiabetic patients with metabolic syndrome. Can J Cardiol 2014;30:320-4. 12. Gökdeniz T, Erkol A, Kalaycıoglu E, et al. Relation of epicardial fat thickness to subclinical right ventricular dysfunction assessed by strain and strain rate ımaging in subjects with metabolic syndrome: a twodimensional speckle tracking echocardiography study [Epub ahead of print]. Echocardiography http://dx.doi.org/10.1111/echo.12635.
21. Henein MY, Gibson DG. Long axis function in disease. Heart 1999;81: 229-31. 22. Delgado V, Mollema SA, Ypenburg C, et al. Relation between global left ventricular longitudinal strain assessed with novel automated function imaging and biplane left ventricular ejection fraction in patients with coronary artery disease. J Am Soc Echocardiogr 2008;21:1244-50. 23. Cho GY, Chan J, Leano R, Strudwick M, Marwick TH. Comparison of two-dimensional speckle and tissue velocity based strain and validation with harmonic phase magnetic resonance imaging. Am J Cardiol 2006;97:1661-6. 24. Ueland T, Yndestad A, Øie E, et al. Dysregulated osteoprotegerin/RANK ligand/RANK axis in clinical and experimental heart failure. Circulation 2005;111:2461-8. 25. Avignon A, Sultan A, Piot C, et al. Osteoprotegerin: a novel independent marker for silent myocardial ischemia in asymptomatic diabetic patients. Diabetes Care 2007;30:2934-9. 26. Chen WJ, Rijzewijk LJ, van der Meer RW, et al. Association of plasma osteoprotegerin and adiponectin with arterial function, cardiac function and metabolism in asymptomatic type 2 diabetic men. Cardiovasc Diabetol 2011;10:67. 27. Blázquez-Medela AM, García-Ortiz L, Gómez-Marcos MA, et al. Osteoprotegerin is associated with cardiovascular risk in hypertension and/or diabetes. Eur J Clin Invest 2012;42:548-56. 28. Rasmussen LM, Tarnow L, Hansen TK, Parving HH, Flyvbjerg A. Plasma osteoprotegerin levels are associated with glycaemic status, systolic blood pressure, kidney function and cardiovascular morbidity in type 1 diabetic patients. Eur J Endocrinol 2006;154:75-81. 29. Altinova AE, Toruner F, Akturk M, et al. Relationship between serum osteoprotegerin, glycemic control, renal function and markers of atherosclerosis in type 2 diabetes. Scand J Clin Lab Invest 2011;71:340-3.
13. Mancia G, De Backer G, Dominiczak A, et al. 2007 ESH-ESC Practice Guidelines for the Management of Arterial Hypertension: ESH-ESC Task Force on the Management of Arterial Hypertension. J Hypertens 2007;25:1751-62.
30. O’Sullivan EP, Ashley DT, Davenport C, et al. Osteoprotegerin and biomarkers of vascular inflammation in type 2 diabetes. Diabetes Metab Res Rev 2010;26:496-502.
14. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 2003;26:5-20.
31. Kim SM, Lee J, Ryu OH, et al. Serum osteoprotegerin levels are associated with inflammation and pulse wave velocity. Clin Endocrinol (Oxf) 2005;63:594-8.
15. Imai E, Horio M, Nitta K, et al. Modification of the Modification of Diet in Renal Disease (MDRD) Study equation for Japan. Am J Kidney Dis 2007;50:927-37.
32. Coutinho T, Turner ST, Mosley TH, Kullo IJ. Biomarkers associated with pulsepressure in African-Americans and non-Hispanic whites. Am J Hypertens 2012;25:145-51.
16. Leoncini G, Ratto E, Viazzi F, et al. Increased ambulatory arterial stiffness index is associated with target organ damage in primary hypertension. Hypertension 2006;48:397-403.
33. Moran CS, McCann M, Karan M, et al. Association of osteoprotegerin with human abdominal aortic aneurysm progression. Circulation 2005;111:3119-25.
17. Gottdiener JS, Bednarz J, Devereux R, et al. American Society of Echocardiography recommendations for use of echocardiography in clinical trials. J Am Soc Echocardiogr 2004;17:1086-119. 18. Alpaydın MS, Aksakal E, Erol MK, et al. Assessment of regional left ventricular functions by strain and strain rate echocardiography in type II diabetes mellitus patients without microvascular complications. Turk Kardiyol Dern Ars 2011;39:378-84.
Supplementary Material To access the supplementary material accompanying this article, visit the online version of the Canadian Journal of Cardiology at www.onlinecjc.ca and at http://dx.doi.org /10.1016/j.cjca.2014.08.018.
Clinical Research
Osteoprotegerin Is Associated With Subclinical Left Ventricular Systolic Dysfunction in Diabetic Hypertensive Patients: A Speckle Tracking Study Ezgi Kalaycıoglu, MD,a Tayyar Gökdeniz, MD,b Ahmet Çagrı Aykan, MD,a Engin Hatem, MD,c Mustafa Ozan Gürsoy, MD,d Asım Ören, MD,e Hüseyin Yaman, MD,e Ays¸e Gül Karadeniz, MD,f and S¸ükrü Çelik, MDa a
Department of Cardiology, Ahi Evren Chest and Cardiovascular Surgery Education and Research Hospital, Trabzon, Turkey b c
e f
Department of Cardiology, Kafkas University, Faculty of Medicine, Kars, Turkey
Department of Cardiology, Erzurum Regional Training and Research Hospital, Erzurum, Turkey d Department of Cardiology, Gaziemir Salih Nevvar I_ ¸s gören State Hospital, I_ zmir, Turkey
Department of Medical Biochemistry, Karadeniz Technical University Faculty of Medicine, Trabzon, Turkey
Department of Radiology, Ahi Evren Chest and Cardiovascular Surgery Education and Research Hospital, Trabzon, Turkey
See editorial by Putko et al., on pages 1492-1495 of this issue. ABSTRACT
RESUM E
Background: Recently, the role of osteoprotegerin (OPG) in the pathogenesis of heart failure through different mechanisms has received much attention. Subclinical changes in left ventricular (LV) function can be identified using quantification of myocardial strain, and global longitudinal strain (GLS) is a superior predictor of outcomes than ejection fraction. We hypothesized that increased OPG levels could predict subclinical LV systolic dysfunction in treated diabetic hypertensive patients with preserved LV ejection fraction. Methods: The study was composed of 86 diabetic hypertensive and 30 nondiabetic hypertensive patients. All patients underwent echocardiography and venous blood samples were taken for determination of OPG. The relation between OPG levels and LV GLS was investigated using 2-dimensional speckle tracking echocardiography. Results: Diabetic hypertensive patients had higher diastolic peak
cemment, le rôle de l’oste oprote ge rine (OPG) dans la Introduction : Re pathogenèse de l’insuffisance cardiaque par le biais de ses divers canismes a fait l’objet d’une grande attention. La quantification de me formation myocardique peut de terminer les modifications subla de formation cliniques de la fonction du ventricule gauche (VG), mais la de dit mieux les re sultats que la fraction longitudinale globale (DLG) pre jection. Nous avons pose l’hypothèse que l’augmentation des cond’e dire la dysfonction systolique subclinique centrations de l’OPG peut pre tiques traite s ayant une fraction du VG des patients hypertendus diabe jection VG pre serve e. d’e thodes : L’e tude e tait compose e de 86 patients hypertendus diaMe tiques et de 30 patients hypertendus non diabe tiques. Tous les be chocardiographie et un pre lèvement patients ont subi une e chantillons de sang par ponction veineuse pour de terminer l’OPG. d’e
Osteoprotegerin (OPG) is a soluble member of the tumour necrosis factor (TNF) receptor superfamily with pleiotropic effects on bone metabolism, endocrine function, and the immune system.1 OPG plays a regulatory role in inflammatory pathways, acting as a decoy for the receptor activator of nuclear factor kb (RANK) ligand (RANKL), competitively inhibiting interactions between RANK and RANKL.2
Recently, the role of the OPG/RANKL/RANK axis in the pathogenesis of heart failure (HF) through different mechanisms such as promotion of matrix degradation and inflammation1 has received much attention.3 OPG has been linked to diabetes mellitus (DM), silent myocardial ischemia, acute myocardial infarction,4 and microvascular complications.5 It has been shown that increased plasma OPG is a strong predictor of all-cause mortality independent of conventional risk for cardiovascular disease (CVD) in patients with type 2 DM (T2-DM).6 Previously, an association between OPG and left ventricular (LV) dysfunction in the general population was also described.1 However, there is no previous study regarding the relationship between OPG and LV dysfunction in diabetic hypertensive patients. T2-DM is one of the major risk factors
Received for publication June 22, 2014. Accepted August 19, 2014. Corresponding author: Dr Ezgi Kalaycıoglu, Ahi Evren Chest and Cardiovascular Surgery Education and Research Hospital, Soguksu M, Camlık Cad, 61040 Trabzon, Turkey. Tel.: þ90-5305606260. E-mail: [email protected] See page 1533 for disclosure information.
http://dx.doi.org/10.1016/j.cjca.2014.08.018 0828-282X/Ó 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.
1530
Canadian Journal of Cardiology Volume 30 2014
early/early diastolic tissue velocity and lower systolic tissue velocity, GLS, GLS rate systolic, and GLS rate early diastolic than nondiabetic hypertensive patients (P ¼ 0.009, P ¼ 0.049, P < 0.001, P ¼ 0.004, and P < 0.001, respectively). Diabetic hypertensive patients were divided into 2 groups according to median GLS value (> 18.5 and 18.5). The patients with GLS 18.5 had higher diastolic blood pressure (mm Hg; P ¼ 0.048), OPG (pmol/L; P < 0.001), and hemoglobin A1c (%; P ¼ 0.042) values than those with GLS > 18.5. In multivariate logistic regression analysis, OPG was found to be an independent predictor of impaired GLS (P ¼ 0.001). Receiver operating characteristic curve analysis revealed that OPG values of > 6.45 (pmol/L) identified the patients with GLS 18.5. Conclusions: Plasma OPG values could predict subclinical LV systolic dysfunction in diabetic hypertensive patients.
te La relation entre les concentrations d’OPG et la DLG du VG a e tudie e par e chocardiographie bidimensionnelle Speckle Tracking e (suivi de pixel). sultats : Les patients hypertendus diabe tiques avaient une ve locite Re coce maximale/une ve locite tissulaire diatissulaire diastolique pre coce plus e leve es et une ve locite tissulaire systolique, une stolique pre DLG, un taux de DLG durant la systole et un taux DLG durant la diastole coce plus faibles que les patients hypertendus non diabe tiques (P pre ¼ 0,009, P ¼ 0,049, P< 0,001, P ¼ 0,004, et P< 0,001, respectiques ont e te divise s en 2 tivement). Les patients hypertendus diabe diane de la DLG (> 18,5 et 18,5). Les groupes selon la valeur me patients ayant une DLG 18,5 avaient des valeurs de pression rielle diastolique (mm Hg; P ¼ 0,048), d’OPG (pmol/l; P< 0,001) arte moglobine A1c (%; P ¼ 0,042) plus e leve es que ceux ayant une et d’he gression logistique multivarie e, l’OPG DLG > 18,5. Dans l’analyse de re tait ave re e un pre dicteur inde pendant de la de te rioration de la DLG s’e ristique d’efficacite du (P ¼ 0,001). L’analyse de la courbe caracte cepteur re ve lait que des valeurs d’OPG > 6,45 (pmol/l) re terminaient les patients ayant une DLG 18,5. de dire la Conclusions : Les valeurs plasmatiques de l’OPG pourraient pre dysfonction systolique subclinique du VG des patients hypertendus tiques. diabe
for the development of CVD, particularly when it is associated with hypertension (HT)7 and markers of CVD are needed in DM to identify patients at risk of severe complications.4 Global LV systolic function, most commonly assessed according to echocardiographic ejection fraction (EF), is an important predictor of cardiovascular outcome. However, the measurement of EF presents a number of challenges related to image quality, assumption of LV geometry, and expertise.8 Also, EF is known not to be a sensitive marker for the detection of subclinical LV systolic dysfunction.9 Twodimensional speckle tracking echocardiography (2D-STE) is a novel technique used for the measurement of cardiac mechanics. It assesses myocardial deformation and the myocardial deformation rate by tracking speckles in the myocardium on greyscale (B-mode) images and can be used to evaluate global and regional myocardial strain and strain rate.10 Longitudinal tissue deformation is evaluated using frame-by-frame tracking of individual speckles throughout the cardiac cycle, and global longitudinal strain (GLS) is calculated.8 Previous studies have reported that subclinical changes in LV function can be identified using quantification of myocardial strain,9 and GLS is superior to EF as a predictor of outcomes.8 In addition, this technique has some advantages over conventional echocardiography including angle independency, free of tethering and translation effects, low signal-to-noise ratio, and low measurement variability.11,12 We hypothesized that increased OPG levels could predict subclinical LV systolic dysfunction in treated diabetic hypertensive patients with preserved LV EF. This study was the first to evaluate this issue in current literature.
diabetic hypertensive patients and 30 consecutive nondiabetic hypertensive patients. All patients were receiving antihypertensive and/or antidiabetic treatment for at least 1 year. Exclusion criteria were echocardiographic evidence of either regional or global wall motion abnormalities, LV EF < 50%, history of significant coronary artery disease (CAD) and coronary artery bypass graft surgery, angina pectoris, positive treadmill test or nuclear perfusion stress test, ischemic electrocardiographic findings, chronic renal failure, atrial fibrillation/flutter, second- or third-degree atrioventricular block, moderate to severe valvular heart disease, secondary HT, pregnancy, neoplasia, systemic inflammatory disease, and known immune system or connective tissue disease. HT was diagnosed when the mean of 3 separate blood pressure (BP) measurements were 140 mm Hg for systolic BP (SBP) and/or 90 mm Hg for diastolic BP (DBP).13 Diabetes was diagnosed when fasting glucose was 126 mg/dL or 200 mg/dL 2 hours after oral glucose overload (repeated twice) or after detection of symptoms of diabetes and random blood glucose 200 mg/dL.14 Informed consent was obtained from the patients and the study was approved by the local ethics board. Body mass index (BMI) was calculated as weight (kg)/height (m2). The body surface area was calculated according to the following formula in square metres: 0.007184 weight (kg)0.425 height (cm)0.725.
Methods Study population From September 2012 to June 2013, we conducted a single-centre cross-sectional study including 86 consecutive
Laboratory measurements Estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease study equation.15 Blood samples for determination of OPG were collected in EDTA tubes and centrifuged, and the isolated plasma samples were stored at 80 C until analysis. Plasma OPG was measured using a sandwich enzyme-linked immunosorbent assay using commercially available antibodies (Biovendor). Urine albumin was assessed quantitatively using an immunoturbidimetric test (Beckman Coulter).
Kalaycıo glu et al. Osteoprotegerin and Left Ventricular Dysfunction
Assessment of carotid intima media thickness Carotid arteries were investigated in the longitudinal and the transverse projections with B-mode sonography (Mindray M7 DC7) using a 7-MHz linear probe. The intima media thickness of both carotid arteries was always measured manually on the common carotid artery outside the plaque, if any was present. Each measurement was calculated using the average of 3 readings.16 Conventional echocardiography The echocardiographic studies were performed using a commercially available echocardiography machine (VIVID S-5 General Electric Medical System, Vingmed Ultrasound AS, Horten, Norway) equipped with a 3.6-MHz transducer and tissue Doppler imaging. Measurements were performed according to the American Society of Echocardiography guidelines,17 by a cardiologist. LV dimensions (end-diastolic and end-systolic) and wall thickness (septum and posterior wall) were obtained from the parasternal long axis with an Mmode. LV EF was measured using the modified Simpson method. LV mass was calculated according to Devereux formula and was indexed to the body surface area. The relative wall thickness was calculated as the ratio (2 posterior wall thickness/LV internal dimension at end-diastole). Mitral inflow velocities were evaluated using pulse-wave Doppler with the sample volume placed at the tip of the mitral leaflets from the apical 4-chamber view. Using the average of 3 beats, we measured diastolic peak early (E), peak late transmitral flow velocity (A), peak E to peak A velocities (E/A) and deceleration time of peak E velocity. Systolic tissue velocity (S’) and early diastolic tissue velocity (E’) were acquired from the septal and lateral sides of the annulus using the tissue Doppler imaging and the pulse wave Doppler mode and measured using averaged values. To evaluate LV filling pressures, the ratio of E/E’ was calculated. Two-dimensional speckle tracking analysis is shown in the Two-dimensional Speckle Tracking Analysis section of the Supplementary Material. Details on reproducibility are shown in the Reproducibility section of the Supplementary Material. Statistical analysis SPSS 17.0 statistical software (SPSS Inc, Chicago, IL) was used for statistical analysis. The Kolmogorov-Smirnov test was used to test the normality of distribution of continuous variables. Continuous variables are expressed as mean standard deviation (SD) or median (interquartile range), and categorical variables are expressed as percentage, as appropriate. Continuous variables were compared using Student t test or the Mann-Whitney U test. Comparisons for categorical variables were evaluated using the c2 test. To avoid collinearity in assessing the multivariate model, independent variables were tested for intercorrelation. Pearson or Spearman correlation analysis was used for assessing correlation between GLS and continuous variables depending on Gaussian distributions. Diabetic hypertensive patients were divided into 2 groups according to the median value of GLS (> 18.5 and 18.5). Multivariate logistic regression analysis was performed to find independent associates of impaired GLS ( 18.5).
1531
Independent variables such as DBP, SBP, S’, OPG, hemoglobin (Hb) A1c, age, female sex, E/A ratio, and eGFR were entered into the model. Interobserver agreement of echocardiographic parameters obtained from 2D-STE data was calculated using a Bland-Altman analysis and intraclass correlation coefficient was used to assess intraobserver agreement. Receiver operating characteristic curve analysis was performed to detect the cutoff value of the OPG in predicting GLS ( 18.5). For further analysis of the independent predictors of GLS, multivariate linear regression analysis was performed. Logarithmic transformation was performed for variables that were not normally distributed including S’ and HbA1c before they were entered into multivariate regression analysis. DBP, SBP, Log10-S’, OPG, Log10-HbA1c, age, female sex, glomerular filtration rate, and E/A ratio were entered into the model. A 2-tailed P < 0.05 was considered statistically significant. Results The study population consisted of 86 (55.8% female) diabetic hypertensive and 30 (73.3% female) nondiabetic hypertensive patients. The clinical, biochemical, and echocardiographic characteristics of the study population are presented in Supplemental Table S1. Diabetic hypertensive patients had lower eGFR (86.40 16.20 mL/min vs 92.00 9.40 mL/min; P ¼ 0.024) and higher microalbumin (1.1 [2.87] mg/dL vs 0.5 [0.64] mg/dL; P ¼ 0.001), carotid intima media thickness (CIMT; 0.8 [0.3] mm vs 0.5 [0.32] mm; P < 0.001), OPG (7.37 2.53 pmol/L vs 5.32 2.17 pmol/L; P < 0.001) values than nondiabetic hypertensive patients. Diabetic hypertensive patients had higher E/E’ and lower S’ (m/s) than nondiabetic hypertensive patients (7.06 [3.38] vs 6.2 [2.47]; P ¼ 0.009 and 0.09 0.08 vs 0.12 0.02; P ¼ 0.049). Furthermore, LV systolic functions were more impaired in diabetic hypertensive patients compared with nondiabetic hypertensive patients (GLS, 18.74 2.89% vs 20.87 1.96%; P < 0.001; GLS rate systolic, 1.10 0.24 L/s vs 1.24 0.19 L/s; P ¼ 0.004; GLS rate early diastolic, 1.12 0.35 L/s vs 1.46 0.32 L/s; P < 0.001, respectively). Diabetic hypertensive patients were divided into 2 groups according to median GLS values (> 18.5 and 18.5). The characteristics of diabetic hypertensive patients according to GLS values are shown in Supplemental Table S2. The patients with GLS 18.5 had higher DBP (91.89 10.20 mm Hg vs 87.31 10.94 mm Hg; P ¼ 0.048), OPG (8.40 2.20 pmol/L vs 6.30 2.45 pmol/L; P < 0.001) and HbA1c (7.90 [1.49]% vs 7.25 [1.54]%; P ¼ 0.042) values than those with GLS > 18.5. To find independent predictors of impaired GLS ( 18.5), multivariate logistic regression analysis was performed. Independent variables such as DBP (mm Hg), SBP (mm Hg), S’ (m/s), OPG (pmol/L), HbA1c (%), age (years), female sex, E/ A ratio, and eGFR (mL/min) were entered into the model. OPG was found to be an independent predictor of impaired GLS ( 18.5; P ¼ 0.001; Supplemental Table S3). The correlation analysis between OPG and GLS values is shown in Figure 1A. A multivariate linear regression model was built to find the independent determinants of GLS. DBP (mm Hg), SBP (mm Hg), Log10-S’, OPG (pmol/L), age (years), female sex, eGFR
1532
Canadian Journal of Cardiology Volume 30 2014
Figure 1. (A) The correlation between OPG and GLS values. (B) ROC curve analysis of OPG in predicting GLS ( 18.5). GLS, global longitudinal strain; OPG, osteoprotegerin; ROC, receiver operating characteristic.
(mL/min), and E/A ratio were entered in the model. OPG and Log10-S’ (systolic tissue velocity; m/s) were independent predictors of impaired GLS (b ¼ 0.534; P < 0.001 and b ¼ 4.591; P ¼ 0.003, respectively; Supplemental Table S4). Receiver operating characteristic curve analysis was performed in diabetic hypertensive patients to detect the cutoff value of OPG in predicting patients with GLS ( 18.5). The analysis showed that OPG values of > 6.45 (pmol/L) identified the patients with GLS ( 18.5) with a specificity of 69.1% and a sensitivity of 79.5% (area under the curve, 0.755; 95% confidence interval, 0.650-0.859; P < 0.001; Fig. 1B). Partial correlation analysis was performed in all subjects to determine the partial correlation between OPG and GLS adjusted for smoking, fasting plasma glucose, BMI, age, and female sex. We found that OPG was negatively correlated with GLS (P < 0.001; r ¼ 0.653). Univariate analysis showed that OPG (pmol/L) was correlated with HbA1c percentage (r ¼ 0.347; P < 0.001), CIMT (mm; r ¼ 0.243; P ¼ 0.009), eGFR (r ¼ 0.194; P ¼ 0.037) and E/E’ (r ¼ 225; P ¼ 0.015); but not correlated with SBP (mm Hg; P ¼ 0.067), DBP (mm Hg; P ¼ 0.157), pulse pressure (PP; r ¼ 0.113; P ¼ 0.226) and LV mass index (g/m2; P ¼ 0.754) in all study populations. OPG (pmol/L) was correlated with E/E’ (r ¼ 223; P ¼ 0.039), but no correlation was detected between OPG and HbA1c percentage (P ¼ 0.087), CIMT (mm; P ¼ 0.209), SBP (mm Hg; P ¼ 0.098), DBP (mm Hg; P ¼ 0.104), PP (r ¼ 0.074; P ¼ 0.501), and LV mass index (g/m2; P ¼ 0.113) in diabetic hypertensive patients. Discussion To our knowledge we showed for the first time in this study, that plasma OPG values could predict subclinical LV systolic dysfunction in diabetic hypertensive patients who were receiving medical treatment and had no previous history of CVD. A cutoff value > 6.45 pmol/L was able to identify the patients with subclinical LV systolic dysfunction. Diabetes is known to be associated with the development of HF even without the presence of coexisting CAD.9,18 In addition, the clinical and morphologic features of heart disease in diabetic hypertensive patients are more severe than in those with hypertension or diabetes alone.19 DM is associated with myocardial structural alterations, leading to progressive
deterioration of cardiac hemodynamics. Diabetic cardiomyopathy seems to be composed of 2 stages. In the first stage, physiological adaptation occurs because of metabolic changes. In the second stage, degenerative changes start and the myocardium might have limited capacity for repair. Therefore, early treatment might delay the progression of the severity of cardiomyopathy.19 The early stage of diabetic cardiomyopathy includes minor changes in myocardial structure (such as normal LV dimension, wall thickness, and mass) and cardiac dysfunction usually can only be detected using sensitive methods such as strain and strain rate in this stage.19,20 Systolic dysfunction might be initially apparent in the longitudinal direction, because subendocardial fibres, which are the ones more vulnerable to myocardial ischemia and fibrosis, are longitudinally oriented.21 It has been shown that GLS correlates well with EF measured using echocardiography22 and magnetic resonance imaging (MRI).23 However, they measure different aspects of myocardial function, with EF measuring radial and partly longitudinal function, and GLS measuring longitudinal function.8 Persistent inflammation appears to play a role in the development of HF.24 In particular, TNF-a has been implicated as a possible pathogenic factor in this process.24 OPG is a member of the TNF receptor superfamily.2 Recent studies have shown that other members of the TNF/TNF receptor superfamily also play a pathogenic role in chronic HF. Several authors suggested that among several TNF superfamily members that have been shown to be activated in HF, the OPG/RANK/RANKL axis might be of importance in the pathogenesis of this disorder3 by mediating autocrine, paracrine, and endocrine interactions between cells expressing RANK.3 The binding of RANKL to RANK promotes the activity of matrix metalloproteinase, leading to degradation of extracellular matrix and increased apoptosis.2 Previous studies suggested that experimental and clinical HF were associated with increased expression of the OPG/RANKL/RANK axis. Because RANKL and RANK are difficult to be measured in vivo2 and OPG circulates at much higher levels than RANKL it might be a more stable overall measure of RANKL/RANK activity and more suited for clinical use than quantification of circulating RANKL concentrations.1 Omland et al.1 evaluated the relationship between OPG levels and cardiac magnetic resonance indices of LV structure and functions in the general population. They suggested that
Kalaycıo glu et al. Osteoprotegerin and Left Ventricular Dysfunction
higher levels of OPG were independently associated with higher LV end-systolic volume and lower EF in both sexes. OPG levels were found to be increased in subjects with only mildly impaired LV systolic function, suggesting that the OPG/RANKL/RANK system activation is an early phenomenon in the process of ventricular dysfunction and HF development.1 To generalize our findings, we evaluated the potential relationship between OPG and GLS in all hypertensive subjects. After adjustment for potential confounders (smoking, fasting plasma glucose, BMI, age, and sex), OPG remained significantly associated with GLS. The association between OPG levels and endothelial dysfunction, arterial stiffness, and atherosclerosis was shown previously.4 Avignon et al.25 investigated OPG concentrations and signs of myocardial ischemia on myocardial perfusion scintigraphy in patients with DM and concluded that OPG measurements can help to better define the diabetic population with an increased risk of developing silent myocardial ischemia. Higher levels of OPG were documented in T2-DM patients with asymptomatic CAD, and also in experimental and clinical HF, even in the absence of ischemic cardiomyopathy.26 The pathophysiological connection between plasma OPG concentrations and CVD is not clearly known, but a correlation with arterial and myocardial disease has been suggested.4 We did not evaluate the inducible myocardial ischemia in our study population, so we can speculate that the relationship between the OPG/RANK/RANK axis and subclinical LV dysfunction might be attributed to endothelial dysfunction and/or the negative remodelling effect on the LV myocardium. Blázquez-Medela et al.27 found significantly higher levels of OPG in diabetic hypertensive patients than in hypertensive, nondiabetic patients and in diabetic, nonhypertensive patients. They also found that HbA1c was significantly correlated with OPG in all study populations (including healthy control subjects) and in the hypertensive group, but there was no correlation in the diabetic hypertensive patients group. Consistent with the findings of this study, our results showed that plasma OPG levels were higher in diabetic hypertensive patients than in nondiabetic hypertensive patients, and HbA1c was found to be significantly correlated with OPG in all study populations but there was no correlation in diabetic hypertensive patients. The underlying mechanisms behind the correlation between OPG and glycemic status remain unknown, but might be related to possible regulatory effects of OPG production from vascular cells and osteoblasts because OPG synthesis is regulated by insulin-like growth factor-I in osteoblasts, and by insulin and TNF-a in endothelial cells.28 Several previous studies reported that OPG was related to insulin resistance in diabetic and nondiabetic subjects.29 Altinova et al.29 found a significant positive correlation between serum OPG and a homeostasis model assessment of insulin resistance in diabetic patients. O’Sullivan et al.30 found that OPG was negatively correlated with fasting plasma glucose in healthy subjects, but there was no correlation between OPG and fasting plasma glucose and HbA1c in diabetic patients, suggesting that serum OPG might be determined by different factors in different populations. Taken together, the authors suggested that more studies are needed to understand the exact role of OPG in glucose metabolism.29 Several studies examined the association between OPG and BP with conflicting results.1 In the current study, we found
1533
that the diabetic hypertensive patients with impaired LV systolic function (GLS 18.5) had more elevated DBP values. In correlation analysis, we did not find any correlation between OPG and SBP-DBP, or PP values. In concordance with our study, Blázquez-Medela et al.27 suggested that OPG values were correlated with SBP and PP in all study populations (including healthy control subjects), but there was no correlation in diabetic hypertensive patients. Kim et al.31 suggested that OPG values were significantly correlated with SBP, but their study population included patients newly diagnosed with DM and they excluded patients with previously diagnosed HT who were receiving medical treatment. Coutinho et al.32 reported a relationship between PP and OPG in hypertensive patients. The association between OPG and HT was explained by the generalized matrix changes, including OPG accumulation in the arterial wall, and therefore increased arterial stiffness.28 However, it was previously shown that angiotensin II blockade could downregulate OPG secretion in vitro.27,33 In the current study, no association was detected between OPG and BP or PP; this was probably because of the use of angiotensin-converting enzyme inhibitors-angiotensin receptor blockers in most of the patients. In the study by Coutinho et al.32 a minority of the subjects were treated with angiotensin converting enzyme inhibitors-angiotensin receptor blockers. Therefore, we can speculate that a relationship between OPG and BP or PP might be detected in newly diagnosed hypertensive subjects who are not yet taking medication or treated with angiotensin converting enzyme inhibitors-angiotensin receptor blockers. Study limitations Several limitations are present in our study. The sample size of subjects was relatively small to generalize our findings. Although MRI is the gold standard for myocardial strain imaging, 2D-STE correlates well with MRI. The inclusion of the healthy control subjects would have increased the power of the study. Conclusion Our study shows the emerging role of OPG as an indicator of subclinical LV systolic dysfunction in diabetic hypertensive patients. It is evidently crucial to establish biochemical markers of increased risk for CVD events. These markers could be used in the clinical setting for the early diagnosis of subclinical LV systolic dysfunction, which would allow for strategies to be designed to reduce the cardiovascular event rate in those patients. Further studies are needed to establish whether increased OPG levels in diabetic hypertensive patients can in fact predict later development of LV systolic dysfunction. Disclosures The authors have no conflicts of interest to disclose. References 1. Omland T, Drazner MH, Ueland T, et al. Plasma osteoprotegerin levels in the general population: relation to indices of left ventricular structure and function. Hypertension 2007;49:1392-8.
1534
Canadian Journal of Cardiology Volume 30 2014
2. Gupta S, Drazner MH, de Lemos JA. Newer biomarkers in heart failure. Heart Fail Clin 2009;5:579-88.
19. Fang ZY, Prins JB, Marwick TH. Diabetic cardiomyopathy: evidence, mechanisms, and therapeutic implications. Endocr Rev 2004;25:543-67.
3. Ueland T, Yndestad A, Dahl CP, Gullestad L, Aukrust P. TNF revisited: osteoprotegerin and TNF-related molecules in heart failure. Curr Heart Fail Rep 2012;9:92-100.
20. Fang ZY, Yuda S, Anderson V, et al. Echocardiographic detection of early diabetic myocardial disease. J Am Coll Cardiol 2003;41:611-7.
4. Augoulea A, Vrachnis N, Lambrinoudaki I, et al. Osteoprotegerin as a marker of atherosclerosis in diabetic patients. Int J Endocrinol 2013;2013:182060. 5. Knudsen ST, Foss CH, Poulsen PL, et al. Increased plasma concentrations of osteoprotegerin in type 2 diabetic patients with microvascular complications. Eur J Endocrinol 2003;149:39-42. 6. Reinhard H, Lajer M, Gall MA, et al. Osteoprotegerin and mortality in type 2 diabetic patients. Diabetes Care 2010;33:2561-6. 7. U.K. Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPD 38. BMJ 1998;317:703-13. 8. Stanton T, Leano R, Marwick TH. Prediction of all-cause mortality from global longitudinal speckle strain: comparison with ejection fraction and wall motion scoring. Circ Cardiovasc Imaging 2009;2:356-64. 9. Nakai H, Takeuchi M, Nishikage T, Lang RM, Otsuji Y. Subclinical left ventricular dysfunction in asymptomatic diabetic patients assessed by two-dimensional speckle tracking echocardiography: correlation with diabetic duration. Eur J Echocardiogr 2009;10:926-32. 10. Kalaycıoglu E, Gökdeniz T, Aykan AC, et al. Evaluation of right ventricle functions and serotonin levels during headache attacks in migraine patients with aura. Int J Cardiovasc Imaging 2014;30:1255-63. 11. Crendal E, Walther G, Dutheil F, et al. Left ventricular myocardial dyssynchrony is already present in nondiabetic patients with metabolic syndrome. Can J Cardiol 2014;30:320-4. 12. Gökdeniz T, Erkol A, Kalaycıoglu E, et al. Relation of epicardial fat thickness to subclinical right ventricular dysfunction assessed by strain and strain rate ımaging in subjects with metabolic syndrome: a twodimensional speckle tracking echocardiography study [Epub ahead of print]. Echocardiography http://dx.doi.org/10.1111/echo.12635.
21. Henein MY, Gibson DG. Long axis function in disease. Heart 1999;81: 229-31. 22. Delgado V, Mollema SA, Ypenburg C, et al. Relation between global left ventricular longitudinal strain assessed with novel automated function imaging and biplane left ventricular ejection fraction in patients with coronary artery disease. J Am Soc Echocardiogr 2008;21:1244-50. 23. Cho GY, Chan J, Leano R, Strudwick M, Marwick TH. Comparison of two-dimensional speckle and tissue velocity based strain and validation with harmonic phase magnetic resonance imaging. Am J Cardiol 2006;97:1661-6. 24. Ueland T, Yndestad A, Øie E, et al. Dysregulated osteoprotegerin/RANK ligand/RANK axis in clinical and experimental heart failure. Circulation 2005;111:2461-8. 25. Avignon A, Sultan A, Piot C, et al. Osteoprotegerin: a novel independent marker for silent myocardial ischemia in asymptomatic diabetic patients. Diabetes Care 2007;30:2934-9. 26. Chen WJ, Rijzewijk LJ, van der Meer RW, et al. Association of plasma osteoprotegerin and adiponectin with arterial function, cardiac function and metabolism in asymptomatic type 2 diabetic men. Cardiovasc Diabetol 2011;10:67. 27. Blázquez-Medela AM, García-Ortiz L, Gómez-Marcos MA, et al. Osteoprotegerin is associated with cardiovascular risk in hypertension and/or diabetes. Eur J Clin Invest 2012;42:548-56. 28. Rasmussen LM, Tarnow L, Hansen TK, Parving HH, Flyvbjerg A. Plasma osteoprotegerin levels are associated with glycaemic status, systolic blood pressure, kidney function and cardiovascular morbidity in type 1 diabetic patients. Eur J Endocrinol 2006;154:75-81. 29. Altinova AE, Toruner F, Akturk M, et al. Relationship between serum osteoprotegerin, glycemic control, renal function and markers of atherosclerosis in type 2 diabetes. Scand J Clin Lab Invest 2011;71:340-3.
13. Mancia G, De Backer G, Dominiczak A, et al. 2007 ESH-ESC Practice Guidelines for the Management of Arterial Hypertension: ESH-ESC Task Force on the Management of Arterial Hypertension. J Hypertens 2007;25:1751-62.
30. O’Sullivan EP, Ashley DT, Davenport C, et al. Osteoprotegerin and biomarkers of vascular inflammation in type 2 diabetes. Diabetes Metab Res Rev 2010;26:496-502.
14. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 2003;26:5-20.
31. Kim SM, Lee J, Ryu OH, et al. Serum osteoprotegerin levels are associated with inflammation and pulse wave velocity. Clin Endocrinol (Oxf) 2005;63:594-8.
15. Imai E, Horio M, Nitta K, et al. Modification of the Modification of Diet in Renal Disease (MDRD) Study equation for Japan. Am J Kidney Dis 2007;50:927-37.
32. Coutinho T, Turner ST, Mosley TH, Kullo IJ. Biomarkers associated with pulsepressure in African-Americans and non-Hispanic whites. Am J Hypertens 2012;25:145-51.
16. Leoncini G, Ratto E, Viazzi F, et al. Increased ambulatory arterial stiffness index is associated with target organ damage in primary hypertension. Hypertension 2006;48:397-403.
33. Moran CS, McCann M, Karan M, et al. Association of osteoprotegerin with human abdominal aortic aneurysm progression. Circulation 2005;111:3119-25.
17. Gottdiener JS, Bednarz J, Devereux R, et al. American Society of Echocardiography recommendations for use of echocardiography in clinical trials. J Am Soc Echocardiogr 2004;17:1086-119. 18. Alpaydın MS, Aksakal E, Erol MK, et al. Assessment of regional left ventricular functions by strain and strain rate echocardiography in type II diabetes mellitus patients without microvascular complications. Turk Kardiyol Dern Ars 2011;39:378-84.
Supplementary Material To access the supplementary material accompanying this article, visit the online version of the Canadian Journal of Cardiology at www.onlinecjc.ca and at http://dx.doi.org /10.1016/j.cjca.2014.08.018.
