Relationship between overweight and obesity and cardiac dimensions and function in a paediatric population

Obesity in adults is associated with left ventricular hypertrophy, dilatation, and myocardial fibrosis, as well as heart failure and coronary heart disease. These associations have been studied to a lesser extent in the paediatric population. This study aims to investigate the relationship between obesity and cardiac structure and function in the paediatric population. In a southern Spanish village, we selected all inhabitants aged 6–17 years stratifying by age, gender, and educational centres. We performed a complete transthoracic echocardiogram evaluating all the cardiac morphological and functional parameters commonly measured in an echocardiographic study. There were 212 children and adolescents included. Of them, 48.1% were males. The mean age was 10.9 ± 3.0 years. A total of 106 (50%) were normal weight, 57 (26.9%) overweight, and 49 (23.1%) obese. Sex and age were similar in all three groups. Overweight and obesity were associated with larger left ventricular end-diastolic and end-systolic volumes (p < 0.0005), greater left ventricular mass (p < 0.0005), and smaller ejection fraction (p < 0.0005). They were also associated with larger atrial, aortic, and right ventricular size. Lateral and mean E/e′ ratios were higher (p = 0.007 and p = 0.01 respectively). Body mass index was independently associated with all cavity size variables as well as left ventricular ejection fraction. Conclusion: Childhood obesity is independently associated with larger heart chambers, greater left ventricle mass, and smaller left ventricle ejection fraction. What is Known: • Childhood obesity is related to the development of cardiovascular risk factors and is considered an epidemic of the twenty-first century; its prevalence is rising. What is New: • Childhood overweight and obesity lead to changes in cardiac structure and function which, although not considered clinically pathological, are significant and a result of obesity, and which behave as unfavourable incipient alterations at an early age. What is Known: • Childhood obesity is related to the development of cardiovascular risk factors and is considered an epidemic of the twenty-first century; its prevalence is rising. What is New: • Childhood overweight and obesity lead to changes in cardiac structure and function which, although not considered clinically pathological, are significant and a result of obesity, and which behave as unfavourable incipient alterations at an early age.


Introduction
Childhood obesity is one of the most significant public health issues of the twenty-first century. In Spain, the prevalence of overweight and obesity in the paediatric population is 34.1% and 10.3% respectively [1]. This prevalence is increasing at an alarming rate: childhood overweight and obesity are estimated to have increased globally by 47.1% over the past three decades [2].
Overweight and obesity are cardiovascular risk factors and are associated with the presence of further cardiovascular risk factors in the paediatric age group, such as hypertension, dyslipidaemia, hyperinsulinemia, and chronic inflammation [3,4]. Childhood obesity has also been linked to the development of cardiovascular risk factors in adulthood and thus to increased morbidity and mortality [5,6].
Obesity in adulthood has been identified as a causative agent of cardiovascular disease independently of other classical risk factors. Obesity-related clinical entities include the development of coronary heart disease, heart failure, atrial fibrillation, and sudden death [7].
In adult obesity, there is an increase in cardiac output because of the body's increased metabolic demands. This haemodynamic alteration secondary to obesity has been linked to left ventricular (LV) dilatation and hypertrophy, as well as to the development of obesity-related cardiomyopathy. This pathological entity involves an increase in ventricular mass and dilatation of the cardiac chambers, as well as myocyte hypertrophy and the development of myocardial fibrosis. The presence of diastolic dysfunction related to this cardiomyopathy accounts for 50% of cases of heart failure in this group [8][9][10][11].
In childhood, the relationship between overweight and obesity with cardiac morphological and functional alterations is not completely defined. There are few studies relating overweight and obesity to cardiac geometry and function, although LV enlargement and alterations in diastolic function have been reported [12][13][14].
The aim of this study is to investigate the relationship between overweight and obesity, and changes in cardiac structure and function in a paediatric population in our environment.

Methods
We carried out a cross-sectional study. All primary and secondary school children and adolescents from a village of around 3000 inhabitants in Andalusia were invited to participate. The sample was stratified by age, gender, and educational degrees of school. Healthy children and adolescents aged 6-17 years who accepted and signed consent by their parents or legal guardians were included. Those who were outside the age range, had previously known acquired or congenital cardiac pathology, did not give consent, or presented cardiac pathology in the echocardiography study were excluded.
The anthropometric data of the children and adolescents were collected using standardized measurements of weight and height. As for the definition of childhood obesity, there is no agreed criterion for establishing overweight and obesity based on body mass index (BMI). In this study, the definition proposed by the World Health Organisation was used: overweight was defined for sex-and age-specific BMI values greater than one standard deviation and less than two standard deviations, and obesity was defined for sex-and age-specific BMI values greater than two standard deviations [15]. Blood pressure was measured, and blood biochemistry was performed to determine fasting basal glycaemia and lipid profile.
Each subject underwent a standard transthoracic echocardiographic study with a Philips iE33 and an S5-1 probe (Philips Medical Systems, Amsterdam, The Netherlands). All images were recorded and saved for a posteriori digital analysis. Parasternal long-axis M-mode images were obtained to measure left ventricular (LV) diastolic and systolic diameter, interventricular septal thickness, LV posterior wall thickness, and aortic size. Parasternal long-axis view was used because we felt easier to obtain true perpendicular measures of the LV from this echocardiographic window. Calculation of LV mass was performed by the Devereux formula [16]: 0.8 × {1.04 [(LVDD + PW + IVS) 3 -(LVDD) 3 ]} + 0.6; with LVDD corresponding to LV diastolic diameter, PW to posterior wall thickness, and IVS to interventricular septal thickness.
Two-, three-, and four-chamber images were obtained through the apical window. The end-diastolic and endsystolic volumes were obtained using the biplane Simpson's method, and the left ventricular ejection fraction (LVEF) value was calculated from these volumes. The basal diameter of the right ventricle was measured in the apical four-chamber plane focused on the right ventricle. Left atrial volume was calculated by the Simpson biplane method. Right atrial area was obtained by apical fourchamber planimetry.
By pulsed Doppler, LV diastolic function was analysed through mitral filling flow, placing the sample volume at the level of the free edges of the mitral leaflets in the apical four-chamber plane. E-wave and A-wave were measured, and the E/A ratio was calculated. By tissue Doppler, lateral and medial e′ waves were measured, and average e′ wave and lateral, medial, and average E/e′ ratios were calculated according to previous measurements. With continuous Doppler through the tricuspid valve in apical four chambers, the pressure gradient from the right ventricle to the right atrium was estimated by means of the maximum tricuspid regurgitation velocity.
For the descriptive analysis, quantitative variables are represented by their mean and standard deviation and proportions by frequency and percentage. The Kolmogoroff-Smirnoff test was used to check which variables followed a normal distribution. For the comparison of subgroups, we used the ANOVA test for quantitative variables and the chisquare test for proportions, and we used Bonferroni-Dunn post hoc test. Univariate regression analysis was performed to assess the relationship of BMI with the different echocardiographic parameters. Multiple regression analysis was performed to assess the independent influence of BMI on echocardiographic parameters, after adjusting for other baseline covariates age, body mass index, systolic blood pressure, diastolic blood pressure, blood glucose, HDL, LDL, and triglyceridaemia. SPSS v 21.0 (SPSS Inc., Chicago, USA) was used for the analysis. A p < 0.05 value was considered significant.

Results
A total of 265 children and adolescents were initially selected, the final resulting sample was 212, after 53 subjects were excluded: 26 cases did not give consent, 4 cases did not meet age criteria, 8 cases presented cardiac pathologies in the echocardiographic study, 14 cases showed unacceptable quality or other technical problems in the echocardiographic study, and in one case, the anthropometric variables were missing.
Of the 212 children and adolescents, 106 (50.0%) were normal weight, 57 (26.9%) were overweight, and 49 (23.1%) were obese. The mean age was 10.9 ± 3.0 years, and 48.1% were boys. The main clinical characteristics of the different groups are shown in Table 1. Age and sex were similar in the three groups. Statistically significant differences were found in anthropometric parameters (weight, height, and body mass index), blood pressure, blood glucose, HDL, LDL, and triglyceridaemia, with higher values in the overweight and obese groups for all parameters except HDL. Table 2 shows the comparison of echocardiographic structural variables and LVEF in the three groups. For both the LV size and LV thickness variables, the overweight group had higher values than the normal weight group, and the obese group had higher values than the overweight group. Comparisons of diastolic diameter (p = 0.003), end-diastolic volume (p < 0.0005), end-systolic volume (p < 0.0005), interventricular septum size (p = 0.01), posterior wall (p < 0.0005), and LV mass (p < 0.0005) were significantly different. Left atrial volume (p = 0.004), right atrial area (p = 0.001), basal diameter and right ventricular outflow tract (p = 0.001 and p < 0.0005 respectively), and aortic size (p < 0.0005) behaved in the same way. LVEF was significantly lower in the obese group than in the overweight group, and significantly lower in the overweight group than in the normal weight group (p < 0.0005), with no significant post hoc comparison of the overweight and obese groups. Table 3 shows the diastolic function variables. E-wave and A-wave velocities were not significantly different among groups; furthermore, lateral and medial e′ wave velocities also were not significantly different. E/e′ lateral ratio (p = 0.007) and E/e′ average ratio (p = 0.01) showed significant differences, with no significant post hoc comparison of normal weight and overweight groups in both cases. The pressure gradient between the right ventricle and right atrium was higher in the obese group (p = 0.05).
The association of BMI with the different echocardiographic variables was assessed by univariate regression analysis (Table 4). All chamber size variables assessed in the study were significantly correlated with BMI. The association was strongest for LV ventricular volumes (LV end-diastolic volume r 2 = 0.48, p < 0.0005, and LV end-systolic volume r 2 = 0.51, p < 0.0005), whilst for all other variables the association was weaker (LV mass r 2 = 0.32, p < 0005; LV diastolic diameter r 2 = 0.29, p < 0005; LVEF r 2 = 0.12, p < 0005; left atria volume r 2 = 0.37, p < 0005; right atria area r 2 = 0.26, p < 0005; right ventricle basal diameter r 2 = 0.16, p < 0005; aortic diameter r 2 = 0.36, p < 0005). For diastolic function parameters, the association was particularly weak, and only the right cavity pressure gradient was significantly correlated with BMI (estimated pressure gradient r 2 = 0.04, p < 03).
Multivariate analysis was performed on those variables that were significant in the univariate analysis (Table 5). BMI was independently associated with all chamber size variables, as well as with LVEF, after adjusting for other baseline covariates. The best explained variables were LV mass and volumes (LV mass: r 2 = 0.52 p < 0.0005, LV enddiastolic volume: r 2 = 0.66 p < 0.0005, LV end-systolic volume: r 2 = 0.63 p < 0.0005), aortic size (r 2 = 0.56, p < 0.0005), and LVEF (r 2 = 0.56, p < 0.0005).

Discussion
The main finding of this study is that overweight and obesity in children and adolescents are independently and negatively associated with cardiac morphology and function parameters measured by echocardiography. In the study by Sánchez-Cruz et al. [17], the prevalence of childhood overweight was 26% and that of childhood obesity was 12.6%. Data from the recently published ENPE study showed a prevalence of childhood obesity of 10.3% [1]. The percentage of overweight in our sample was 26.9%, and the percentage of obesity was 23.1%, higher than that reported in previous population-based studies in Spain. These findings could be correlated with a heterogeneous distribution of obesity in the different areas of our country. The enKid study [18], designed to assess the dietary and nutritional habits of children and young people, reported that the Canary Islands and Andalusia had the highest rates of childhood obesity. These data are consistent with the high rates of overweight and obesity in the present study. Furthermore, the results of this study show that higher BMI in childhood is independently associated with increased ventricular wall thickness and LV mass, as well as with increased LV size in both diastolic and systolic volumes. Similarly, higher BMI is associated with increased size of both atria, right ventricle, and ascending thoracic aorta. Regarding LV systolic function, the overweight and obese groups had a lower ejection fraction compared to the normal weight group, and increased BMI was independently associated with decreased LVEF with acceptable goodness-of-fit. It is important to note that the results show a very modest correlation (only two variables with r 2 > 0.5 in univariate analysis, although six in multivariate analysis -a finding that is common in multiple linear regression modelling, as variability is better explained by several parameters); consequently, its clinical impact is minimal. Nevertheless, its epidemiological impact might well be relevant.
Previous studies have described a relationship between childhood obesity and increased left ventricular mass and altered diastolic parameters. The group of Saltijeral et al. [12] described an increase in LV size and LV wall thickness in the childhood obesity group compared to the control group. Mangner et al. [13] described LV and left atrial enlargement and worse diastolic function in obese children and adolescents. Dias et al. [14] agreed with previous studies in that adolescents with obesity showed an increase in left ventricular thickness and size. However, these previous studies were case-control studies with small samples, and only the study by Saltijeral et al. was conducted in Spain. With this respect, our study provides relevant information from a large, randomly selected sample of a rural Spanish child population. So, although not unexpected, our results confirm previous preliminary data with a higher grade of certainty and external validity. Furthermore, none of these studies reported a significant worsening of LVEF in the obese group, nor an independent relationship of BMI with worsening LVEF. The association of left ventricular hypertrophy with worse systolic function had been previously described in other entities, such as systemic arterial hypertension and hypertrophic cardiomyopathy in adults [19][20][21].
Based on our results, in childhood obesity, we observed an adaptation of the LV by dilation and hypertrophy, in a similar way that is reported in adult obesity, and it may be related to the increase in cardiac output needed to meet the increased metabolic demands. Although these structural and functional changes are far from clinical significance, it is likely that these changes, present an early age, may persist over time, and may represent incipient changes for the development of obesity-related cardiomyopathy in adulthood.
It has been described that therapeutic intervention on obesity in adulthood and the reduction of BMI is accompanied by significant structural improvements, even reaching complete normalisation of the cardiac structure [22]. Extrapolating this information to children, therapeutic intervention on childhood obesity might contribute in the primary prevention of cardiovascular events in this population.
As a main limitation of the study, we consider that the study has been designed as a cross-sectional study and it does not really allow to assess whether structural changes will be maintained over time into adulthood. However, the tracking phenomenon is well known, as an example, in the case of arterial hypertension [23]. Similarly, it does not allow to assess whether an intervention on obesity in childhood could reverse the structural and functional changes observed. Further sample follow-up or obesity intervention studies are needed to answer these questions. Furthermore, this is a small study in a very specific, selected population in southern Spain and the results may not be extrapolated to other regions or countries. Another point to consider is that the body surface area and LV mass were calculated with the Dubois and Dubois and Devereux formulas because they are widely used in previous literature, but we acknowledge that the Haycock and de Simone formulas would have been good alternatives. Finally, it is important to note that methodologically there are some measurements that were performed as recommended in adult guidelines and deviate from current paediatric guidelines recommendations, as LV diastolic and systolic diameters, which were obtained in long-axis M-mode images.

Conclusion
The results of this study show that obesity in childhood is independently associated with significant changes in the structure and function of several cardiac chambers.
Availability of data and material All necessary information and material can be requested from the author.
Code availability Not applicable.

Declarations
Ethics approval It has been included the Ethical Approval by Public Health System's approval in Andalusia (Ethical Approval number 2353).
Consent to participate and publication Verbal consent was obtained from all study participants as well as written consent from their parents or legal guardian for participation and use for scientific purposes.

Conflict of interest
The authors declare no competing interests.