Association of metabolically healthy obesity in young adulthood with myocardial structure and function

Obesity is major cause of cardiovascular diseases. Metabolically healthy obesity (MHO) may increase heart failure risk early in life, and may be reflected in impaired cardiac structure and function. Therefore, we aimed to examine the relationship between MHO in young adulthood and cardiac structure and function. A total of 3066 participants from the Coronary Artery Risk Development in Young Adults (CARDIA) study were included, who completed echocardiography in young adulthood and middle age. The participants were grouped by obesity status (body mass index ≥30 kg/m2) and poor metabolic health (≥2 criteria for metabolic syndrome) into four metabolic phenotypes as follows: metabolically healthy non-obesity (MHN), MHO, metabolically unhealthy non-obesity (MUN), metabolically unhealthy obesity (MUO). The associations of the metabolic phenotypes (MHN serving as the reference) with left ventricular (LV) structure and function were evaluated using multiple linear regression models. At baseline, mean age was 25 years, 56.4% were women, and 44.7% were black. After a follow-up 25 years, MUN in young adulthood was associated with worse LV diastolic function (E/é ratio, β [95% CI], 0.73 [0.18, 1.28]), worse systolic function (global longitudinal strain [GLS], 0.60 [0.08, 1.12]) in comparison with MHN. MHO and MUO were associated with LV hypertrophy (LV mass index, 7.49 g/m2 [4.63, 10.35]; 18.23 g/m2 [12.47, 23.99], respectively), worse diastolic function (E/é ratio, 0.67 [0.31, 1.02]; 1.47 [0.79, 2.14], respectively), and worse systolic function (GLS, 0.72 [0.38, 1.06]; 1.35 [0.64, 2.05], respectively) in comparison with MHN. These results were consistent in several sensitivity analyses. In this community-based cohort using data from the CARDIA study, obesity in young adulthood was significantly associated with LV hypertrophy, worse systolic and diastolic function regardless of metabolic status. Relationship of Baseline Metabolic Phenotypes with Young Adulthood and Midlife Cardiac Structure and Function. Adjusted for year 0 covariates: age, sex, race, educational level, smoking status, drinking status, and physical activity; metabolically healthy non-obesity was used as a reference category for comparison. † Criteria for metabolic syndrome are listed in Supplementary Table S6. MUN metabolically unhealthy non-obesity, MHO metabolically healthy obesity, LVMi left ventricular mass index, LVEF left ventricular ejection fraction, E/A early to late peak diastolic mitral flow velocity ratio, E/é mitral inflow velocity to early diastolic mitral annular velocity, CI confidence interval. Relationship of Baseline Metabolic Phenotypes with Young Adulthood and Midlife Cardiac Structure and Function. Adjusted for year 0 covariates: age, sex, race, educational level, smoking status, drinking status, and physical activity; metabolically healthy non-obesity was used as a reference category for comparison. † Criteria for metabolic syndrome are listed in Supplementary Table S6. MUN metabolically unhealthy non-obesity, MHO metabolically healthy obesity, LVMi left ventricular mass index, LVEF left ventricular ejection fraction, E/A early to late peak diastolic mitral flow velocity ratio, E/é mitral inflow velocity to early diastolic mitral annular velocity, CI confidence interval.


INTRODUCTION
Obesity affects at least 30% of adults in many countries worldwide and is the major cause of cardiovascular diseases (CVD) [1,2]. Of note, obesity is a heterogeneous condition, and some individuals with obesity don't develop metabolic and cardiovascular complications [3,4]. Such individuals are defined as having metabolically healthy obesity (MHO) and account for at least one-third of adults with obesity [5]. Recent studies indicate that individuals with MHO have an increased risk of CVD, primarily driven by heart failure (HF) [6,7].
Asymptomatic left ventricular (LV) hypertrophy or dysfunction is defined as "pre-HF", characterized by the high risk of developing HF [8]. Prior cross-sectional studies assessed the relationship between MHO and LV hypertrophy or dysfunction [9][10][11][12]. However, findings from these studies were inconsistent. In addition, the relationship is unknown in young adults who may have more cardiovascular benefits from healthy lifestyle intervention, which may hinder the optimal management of HF [13]. Therefore, a larger prospective study that assesses the association of MHO in young adulthood with LV structure and function is needed to reduce the global burden of HF.
To better understand the role of MHO in the development of HF, we focused on the association of MHO with cardiac structure and function in young adulthood and middle age by analyzing data from the Coronary Artery Risk Development in Young Adults (CARDIA) study.

SUBJECTS AND METHODS Study population
The CARDIA study was a community-based, prospective, cohort study that examined trends in cardiovascular risk factors during young adulthood. In the CARDIA study, a total of 5115 healthy participants were recruited from 4 centers in the United States (Birmingham, AL; Oakland, CA; Chicago, IL; and Minneapolis, MN). Compared to the established National Heart Lung and Blood Institute cohorts, CARDIA participants (born between 1955 and 1968) were significantly influenced by the obesity epidemic at younger ages. The participants were balanced according to race (white and black), sex, and education (high school or less and more than high school) at the baseline examination in [1985][1986]. Further details of this study's participant selection and design have been previously published [14]. The CARDIA study was approved by the Institutional Review Boards of all four centers, and all participants provided written informed consent.
A total of 5113 participants underwent baseline examination (year 0); 4242 underwent echocardiography assessment at year 5, and 3133 completed echocardiography measurements at year 25. The present study excluded 67 participants missing covariates at year 0 for a final sample size of 3066 participants.

Definition of obesity and metabolic health
The study population was divided into four metabolic phenotypes by obesity status and poor metabolic health at the baseline examination (year 0): MHO, metabolically unhealthy obesity (MUO), metabolically healthy non-obesity (MHN), and metabolically unhealthy non-obesity (MUN).

Echocardiographic assessment
Participants were evaluated with 2-dimensional guided M-mode and Tissue Doppler echocardiography at years 5 and 25. Each field center used a standardized protocol, and all sonographers received centralized training. Well-trained sonographers conducted echocardiographic examinations with the ACUSON cardiac ultrasound system (Siemens, Erlangen, Germany) at year 5 and with the ARTIDA ultrasonographic system (Toshiba Medical Systems, Japan) at year 25. This echocardiographic examination followed the guidelines of the American Society of Echocardiography and had good reproducibility [17]. Further details of the year 5 and year 25 echocardiographic examinations have been previously published [17,18]. LV mass was calculated from the Devereux formula, and the LV mass index (LVMi) was obtained using LV mass indexed to body surface area, as calculated by the Mosteller formula. LV ejection fraction (LVEF) was calculated from LV volumes, which were measured from apical views. The peak early (E) and late (A) diastolic velocity and early peak diastolic mitral annular velocity (é) were assessed by Tissue Doppler at the septal mitral annulus. The septal E/A and E/é ratios indicate diastolic function. Global longitudinal strain (GLS) represents LV shortening in the longitudinal planes and was derived from the average of segmental peak values for each phase (%). GLS was calculated from the speckle tracking echocardiography data in Advanced Cardiology Package 2D Wall Motion Tracking software version 3.0 (Tochigi, Toshiba Medical Systems).

Covariates
Data on covariates were obtained from standardized questionnaires, and clinical and laboratory data. Sex, race, and age at baseline were obtained through an interview. Height, weight, blood pressure, glucose levels, lipid levels, and education (number of years) were measured by standardized protocols [14]. Physical activity was quantified in exercise units using a modified version of the Minnesota Leisure Time Physical Activity questionnaire. Antidiabetic, antihypertensive and cholesterol-lowering medication use, drinking status (current drinker or current nondrinker) and smoking status (current smoker or current nonsmoker) were selfreported at baseline. The definition of hypertension was systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or the current use of antihypertensive medication. Diabetes was defined as fasting plasma glucose ≥126 mg/dL, glucose tolerance ≥200 mg/dL, glycated hemoglobin ≥6.5%, or use of insulin or oral hypoglycemic agents.

Statistical analyses
We used Stata (version 16) to perform statistical analyses, and a two-tailed P value <0.05 was considered statistically significant. Descriptive statistics are presented as the mean (SD) for continuous variables and number (%) for categorical variables. The primary goal of our analysis was to assess the relationship between metabolic phenotypes and echocardiographic outcomes. Therefore, we used multivariate linear regression models to assess the association of metabolic phenotypes (MHO, MUO, and MUN; MHN served as the reference) in young adulthood with the following echocardiographic outcomes (year-5 and 25): myocardial structure (LVMi), systolic function (LVEF and GLS) and diastolic function (the E/e ratio and the E/A ratio). The models were adjusted for the following covariates based on previous studies: age, sex, race, education, smoking status, drinking status, and physical activity at baseline (year 0).
To examine the possibility that glucose-lowering, blood pressurelowering, or cholesterol-lowering medication may have residual effects after statistical adjustment, we conducted a sensitivity analysis in participants without glucose-lowering, blood pressure-lowering, or cholesterol-lowering medication. Further, due to noticeable effect of hypertension and diabetes on our analysis, we carried out a sensitivity analysis in participants without hypertension or diabetes, respectively. To assess the robustness of our results, echocardiographic outcomes were also set to binary variables. Table 1 shows the baseline characteristics of the included participants grouped by obesity and metabolic health. Among the 3066 participants from the CARDIA study (Fig. 1 (Fig. 3). In addition, we analyzed the relationship between the metabolic phenotypes and GLS in year 25, using the MHN group as the reference. The 14], P < 0.001, respectively) (Fig. 3). The β-coefficients of the E/A ratio in Year 5 and the E/é ratio in Year 25 were roughly twice as large in the MUO group than in the MHO or MUN group.

Sensitivity analyses
In the sensitivity analyses, we found that the association of MHO with cardiac structure and function was not altered by excluding individuals with hypertension or diabetes, or those who used glucose-lowering, blood pressure-lowering, or cholesterollowering medication (Supplementary Tables S2, S3 and S4). When echocardiographic parameters were set as categorical variables, MHO was associated with LV hypertrophy in young adulthood and associated with LV hypertrophy, systolic dysfunction, and diastolic dysfunction in middle age (Supplementary Table S5). The prevalence of LV hypertrophy and dysfunction at year 25 across the metabolic phenotypes at baseline was shown in Supplementary Fig. S1.

DISCUSSION
In this community-based prospective study, we found that the MUN in young adulthood was related to significantly worse LV function but not LV hypertrophy compared to the MHN. Furthermore, we found that obesity in young adulthood was significantly associated with LV hypertrophy and worse LV function irrespective of metabolic status. These findings suggested that young adults should not only maintain metabolic health, but also, and more importantly, normal weight.
Obesity and metabolically healthy obesity in young adults Obesity, a burden of epidemic proportions, continues unabated in the United States, and it is estimated that the prevalence of obesity will reach 48.9% by 2030 [19]. Current guidelines consider people with obesity to be in stage A of HF [8]. A prospective study demonstrated that greater BMI in early adulthood is associated with impaired LV function and LV hypertrophy [20], which may be driven by obesity-related risk factors, such as hyperglycemia, hypertension, or dyslipidemia [21]. Young adults gain weight faster than other age groups but have not yet reached the stage of life that metabolic syndrome typically emerges [22], The current guideline defines obesity without metabolic risk factors as MHO [23]. The prevalence of MHO ranged from 6-60% in adults with obesity [24,25], and was higher in young adults than older adults [26][27][28]. However, data on the association of MHO in young adulthood with cardiac remodeling and dysfunction are still lacking. Therefore, our study, evaluating the effects of metabolic phenotypes of young adulthood on cardiac structure and function, provided further evidence that obesity is associated with unfavorable alteration in LV structure and function regardless of metabolic status.

Comparison with previous studies
In recent years, MHO has attracted increasing interest. Several studies have investigated associations of metabolic phenotypes in middle-aged and older populations with cardiac structure and function [9][10][11][12]. Two prior cross-sectional studies reported that MUO was associated with preclinical impaired LV systolic and diastolic function [9,12]. However, the relationship between MHO and impaired diastolic function, as reflected by the E/e′ ratio, was not significant. Furthermore, in a small study (n = 67) that investigated the effect of BMI and metabolic health on myocardial function, impaired myocardial function was associated with poor metabolic health rather than obesity [10]. In addition, a relatively large crosssectional study involving 789 Korean individuals showed that poor metabolic health was related to more significant adverse effects on LV structure and function than obesity [11]. Findings from these cross-sectional studies suggested that poor metabolic health in middle-aged and older individuals was more closely associated with LV hypertrophy and dysfunction, and implied that we should pay more attention to maintaining metabolic health rather than normal weight. However, a prospective longitudinal study emphasized the adverse effects of obesity in young adulthood on cardiac structure and function [20]. Of note, this study did not consider the effects of different obesity phenotypes. Our findings bridge a gap in the literature by providing evidence that MHO in young adulthood is associated with LV hypertrophy and incipient dysfunction in midlife and that the association is stronger in individuals with MUO, which emphasizes the need for young adults to maintain a healthy weight.

Implications
Findings from our study link obesity in young adulthood with LV hypertrophy and incipient dysfunction in midlife regardless of metabolic status, which may make young adults more aware of the adverse effects of obesity. That is particularly concerning with the marked growth of prevalence of overweight and obesity in adolescents and young adults. For example, the prevalence of overweight and obesity was over 40% among African-Americans [29]. Previous studies indicated poor metabolic health was more significantly related to LV remodeling and dysfunction in middleaged and older adults [9,10]. However, young adults have not yet reached the stage of life that metabolic syndrome typically emerges, and the prevalence of MHO in young adulthood was higher than that of older adults [26][27][28]. Furthermore, individuals with MHO may not receive early and more aggressive treatment for weight loss. Therefore, our study focused on MHO in young adulthood and proved that obesity is associated with LV hypertrophy and incipient dysfunction irrespective of metabolic status, consistent with current guideline [8]. Emphasizing obesity-related cardiac damage may motivate young adults to more aggressively control weight.
MHO is not stable and more than one-third of individuals with MHO develop MUO [30]. In our study, the relationship between obesity and LV hypertrophy and dysfunction was more evident in participants with poor metabolic health. Previous studies demonstrated that individuals who maintained MHO have a lower risk of cardiovascular disease and all-mortality than individuals with MHO converted to MUO [7,30]. Fortunately, MHO is not irreversible and could return to MHN after treatment. A large nationwide population-based study including 7,148,763 Korean individuals demonstrated that the transition from MHO to MHN protected against hospitalization for HF [31]. Moreover, moderately 3-10% weight loss can bring significant improvement of metabolic health [32,33]. Therefore, young adults with obesity should give adequate attention to lose weight irrespective of metabolic status, which may contribute to minimizing the burden of HF.

Strengths and limitations
The strengths of our study included using data from a standardized echocardiographic assessment in a large, community-based sample with a long-term follow-up of 25 years and the ability to account for confounding factors. In addition, to our knowledge, this may be the first study to comprehensively investigate the association of different metabolic phenotypes in young adulthood with cardiac structure and function.
However, our study also had several limitations. First, because the CARDIA trial was an observational study and our study did not analyze the change in echocardiographic indices at follow-up, our results could not determine causal inferences or reverse causation. We attempted to adjust for common confounding factors, but the effects of residual confounding factors may still have been present. Second, we could not assess the association of obesity or metabolic health with HF because there were limited HF cases in the CARDIA cohort. Third, there is no unified consensus on the definition of MHO [5]. To overcome this limitation, we used the most common definition, which defines MHO as obesity without metabolic syndrome [6,34,35]. Fourth, since our study was only performed in young adults who were white or black, the relationship between metabolic phenotypes and cardiac remodeling in other racial groups and at other ages warrants future investigation. Fifthly, subgroup analysis was not performed according to ethnicity groups because of the limited sample size. However, we adjusted race in a multiple linear regression model. Finally, our study lacked data, and any consideration in the interpretation of our results, of differences in the composition of body fat depots, which define not only the different phenotypes but also links between obesity and poor metabolic health. Future studies may be needed to consider and interpret the nature and origins of cardiometabolic risk from the perspective of differences in the composition of body fat depots.

CONCLUSION
Among the young adults, the MHO during young adulthood was associated with LV hypertrophy, and reduced LV systolic and diastolic function in middle age. Our study provided evidence based on young adults that MHO should not be regarded as a safe condition and reinforced the need for weight management in young adults irrespective of metabolic status.

DATA AVAILABILITY
The data used and/or analyzed in this study are available to all researchers upon reasonable request.