With overweight and obesity reaching epidemic proportions,1 the risks of cardiovascular morbidity and mortality have been increasingly related to adverse metabolic effects.2,3 Metabolic syndrome (MS), a cluster of cardiovascular risk factors including abdominal obesity, elevated blood pressure, impaired glucose tolerance, insulin resistance, elevated triglycerides, and low high-density lipoprotein cholesterol concentrations,4 affects about 25% of the adult population and increases the risk of heart failure,5 increased epicardial fat,6 diabetes, myocardial infarction, stroke, and death.3,7

Metabolic syndrome has been associated with systolic and diastolic dysfunction identified by tissue Doppler imaging8,9; however, ejection fraction, the most widely used parameter for evaluation of systolic function, has low sensitivity for the assessment of early dysfunction in myocardial contractility.10 New echocardiographic techniques such as global longitudinal strain (GLS) by 2-dimensional speckle-tracking echocardiography (2D-STE) allow the assessment of subclinical left ventricular dysfunction by quantitative assessment of myocardial deformation,11 although the proper GLS cutoff to define subclinical left ventricular dysfunction has not yet been established.12

Given the major obesity epidemic and related metabolic syndrome and the paucity of studies examining the impact of MS components on myocardial dysfunction,13 our aim was to investigate the association between MS and its components with impaired left ventricular GLS, using different GLS cutoffs, in middle-aged adults free of prevalent heart disease participating in the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil).

From our analytic sample of 1055 participants (53% women; 52±9 years), 444 (42%) fulfilled the diagnostic criteria for MS. The clinical and demographic characteristics of all participants according to GLS quartiles are shown in Table 1. Worse GLS (less negative values) was associated with an increased prevalence of MS and its components, except for low high-density lipoprotein cholesterol (all other P values <.001). The baseline sociodemographic characteristics between the data of this random sample and the available data published in the cohort profile of ELSA-Brasil were similar and are shown in Table 1 of the supplementary material.

In the overall sample, the mean GLS was –18.6%±2.5%. Individuals with MS had worse left ventricular systolic function as measured by GLS (–18.0%±2.5%) compared with participants without MS (–19.0%±2.4%, P <.0001). On multiple linear regression analysis (Table 2), the presence of MS was associated with a 0.86 absolute reduction in GLS (GLS in percentage units, P <.0001) after adjustment for age, sex, race/color, educational level, and study center. This absolute GLS impairment with MS was attenuated to 0.60 (GLS percentage units) after adjustment for body mass index, but remained highly statistically significant (P=.0005).

Similarly, we found about twice the prevalence of subclinical systolic dysfunction among participants with MS (Table 3, model 2), corresponding to higher adjusted prevalence ratios (PR) for impaired GLS as estimated with robust Poison regression. Using increasingly more stringent cutoffs for GLS (1, 1.5 and 2 standard deviations [SD]), the associations were as follows: for GLS –16.1% (PR, 1.76; 95%CI, 1.30-2.39), for GLS –14.8% (PR, 2.35; 95%CI, 1.45-3.81), and for GLS –13.5% cut-point (PR, 2.07; 95%CI, 0.97-4.41). After inclusion of body mass index in the model, associations were attenuated and were no longer statistically significant (Table 3).

Figure 2 shows quantile regression differences in strain (in GLS percentage units) between participants with and without MS or with its components across the range of GLS distribution (5th to 95th quantiles), when adjusted for age, sex, race/color, educational level and study center, and in the individual components analyses additionally for the remaining MS criteria. We observed a significant difference in strain in participants with MS, which appeared to increase with worsening GLS (Figure 2). Among individual MS components, only elevated waist circumference was associated with poorer GLS. As in the other analyses, these associations were mostly attenuated when body mass index was included in the models (Table 2 of the supplementary material).

Figure 2.

Adjusted quantile regression differences in GLS (in GLS percentage units and 95% confidence intervals) related to the presence of metabolic syndrome and of its components across the 5th and 95th quantiles of GLS. Adjusted for sex, age (years), race/color, educational level and ELSA-Brasil center, and in analyses of individual components of metabolic syndrome additionally for the remaining metabolic syndrome components. GLS, global longitudinal strain; HDL, high-density lipoprotein.

(0.29MB).

The variables included in the initial investigation and removed from all final regression models, as they did not modify the association of MS with GLS, were heart rate, left ventricular mass, mitral E to e’ ratio, left ventricular diastolic diameter, and ejection fraction. Despite this, we also investigated the role of ejection fraction as a potential confounder in the association of MS with GLS. In linear regression, when we added 2D-STE left ventricular ejection fraction in the model, the association remained statistically significant (GLS difference 0.62%; P <.0001). The ejection fraction only minimally affected the size and significance of the association between MS and GLS in robust Poisson regression when the same GLS cutoffs were used in 3 separate models (1 SD: PR, 1.49; 95%CI, 1.12-1.97; 1.5 SD: PR, 1.95; 95%CI, 1.25-3.02; and 2 SD: PR, 1.66; 95%CI, 0.85-3.23).

In this multicenter cohort of middle-aged individuals, one of the largest reported MS samples of nonreferred participants, we investigated the association of MS and its components with GLS, an indicator of subclinical systolic dysfunction. Participants with MS (42%) had worse left ventricular systolic function assessed by different GLS parameters than those without MS. Participants with MS had about twice the prevalence of subclinical systolic dysfunction, using different GLS cutoff points. MS was independently associated with impaired GLS. After inclusion of body mass index in the models, the associations diminished somewhat, suggesting that, at least in part, they are explained by obesity. Furthermore, on analysis of each individual component of MS, elevated waist circumference was the only component that remained independently associated with impaired GLS.

In 2009, major health organizations proposed consensus criteria for the clinical diagnosis of MS.16 However, as we have shown, different components have distinct associations with GLS.13 Thus, it may make more sense to identify the association of impaired GLS with MS components than with the resultant, more complex definition.21–25

The associations of MS with myocardial dysfunction have been previously shown in smaller samples using conventional echocardiography, tissue Doppler imaging,8,9 and STE.13 However, few studies using 2D-STE have examined which components of the MS are independently associated with the impairment of systolic function. Previous studies demonstrated the unfavorable influence of different MS risk factors on left ventricular deformation such as hypertension,22 diabetes,23 obesity,24 and dyslipidemia.25

In addition, some investigations have demonstrated the existence of stronger associations between the components of MS and myocardial dysfunction with an increasing number of MS risk factors. Tadic et al.26 showed that an increasing number of MS criteria is associated with cardiac diastolic dysfunction. In the Multiethnic Study of Atherosclerosis,27 waist circumference and fasting glucose were significantly associated with left ventricular impairment of circumferential and longitudinal strain, respectively. That study also showed that GLS changed from –14.2% in participants with ≤ 1 MS component, to –13.4%, in participants with 2, and to –12.1% in those with ≥ 3 MS criteria (P <.01). Recently, Tadic et al.28 and Wang et al.29 found that among all MS components, blood pressure, waist circumference, and fasting plasma glucose were the components most closely associated with left ventricular deformation indices. Obesity and cardiometabolic risk factors have been implicated in myocardial dysfunction such as coronary dysfunction and diabetic cardiomyopathy.30,31 Nonetheless, the mechanisms contributing to these changes remain incompletely understood, and are believed to be the result of a complex interplay of hemodynamic and neurohumoral factors as well as of inflammation and oxidative stress, which contribute to cellular apoptosis, hypertrophy, and interstitial fibrosis.30-32

The 2D-STE has emerged as a more robust technique to detect subclinical left ventricular dysfunction by quantitative assessment of myocardial deformation.11 This technique has the advantages of angle independency, and less dependency on loading conditions and geometry than traditional parameters. Additionally, it permits postprocessing in regular 2D acquired images, thus making highly feasible the reproducible examination of a major component of regional myocardial performance. The 2D-STE is becoming widely available, due to technological development and its inclusion in most commercially available echocardiographic systems. Despite speckle-tracking software standardization concerns,11,33 the distribution of GLS values in our study was similar to that found in a recent meta-analysis,12 as was its measurement variability,34 which supports the quality of our data.

Left ventricular function assessed by GLS was significantly impaired in MS participants, and the prevalence of impaired left ventricular GLS was higher in MS participants at several different GLS cutoffs. Moreover, we identified elevated waist circumference as the main MS component associated with impaired left ventricular GLS. Among the MS criteria, central obesity best depicted the link between metabolic derangement and cardiac function.

This work was supported by the Brazilian Ministry of Health (Science and Technology Department), the Brazilian Ministry of Science, Technology and Innovation (Financiadora de Estudos e Projetos-grants 01 06 0010.00, 01 10 0643.00 RS, 01 06 0212.00 BA, 01 06 0300.00 ES, 01 06 0278.00 MG, 01 06 0115.00 SP, 01 06 0071.00 RJ), CNPq (the Brazilian National Council for Scientific and Technological Development), and Coordination for the Improvement of Higher Education Personnel-CAPES (grants 1407742 [W. Cañon-Montañez] and 1556322 [M. Bessel]).

None declared.