Differential Cardiac Remodeling in Overweight and Obesity: Insights from Noninvasive Imaging Open Access

The escalating prevalence of overweight and obesity (O&O) has heightened concerns regarding associated health risks in China [1, 2]. Despite government initiatives, the rate of O&O among Chinese adults has exceeded 50% and is projected to reach 65.3% by 2030 [3]. This condition significantly affects cardiac morphology and function, contributing to an increased incidence of cardiovascular diseases (CVDs). A study by Afshin et al. [4] indicated that 60% of deaths linked to O&O are attributable to CVDs. Another meta-analysis demonstrated that the relative risk per 5-unit increase in body mass index (BMI) was 1.41 (95% confidence interval [CI], 1.34–1.47) for heart failure incidence and 1.26 (95% CI, 0.85–1.87) for heart failure mortality when BMI exceeded the threshold of 23 to 24 kg/m² [5]. Consequently, O&O imposes a substantial burden on CVD incidence and disability-adjusted life years [6].

Cardiac abnormalities associated with O&O are primarily manifested as structural changes in the left ventricle (LV) and impaired myocardial function [7]. These alterations often result in chronic volume overload and elevated LV wall stress, which can lead to compensatory eccentric LV hypertrophy [8, 9]. Advances in cardiac imaging and detection techniques have significantly improved our capacity to identify early subclinical LV involvement in individuals with obesity [10‒12]. Despite these advancements, a critical gap remains in our understanding of how early myocardial abnormalities differ between those classified as overweight or obesity. Rapid economic development has contributed to an increasing number of individuals with varying degrees of O&O, but research on exploring the specific cardiac impacts at different stages of obesity remains limited [13]. It is essential to investigate at what point these myocardial changes become pronounced, as this knowledge could guide early intervention strategies. By clarifying the nuances of cardiac characteristics across different levels of obesity, we can better address the health risks associated with this escalating epidemic.

Cardiac magnetic resonance (CMR) is a crucial and well-established clinical tool for diagnosing and managing CVDs, serving as the standard reference for evaluating cardiac morphology and function [14]. By comprehensively analyzing the results of noninvasive cardiac tests, this study aimed to elucidate the impact of O&O on the cardiac characteristics and to explore the specific effects of obesity at various stages on heart health. This research is essential for developing targeted early intervention strategies for populations affected by O&O.

A retrospective study was conducted at The First Affiliated Hospital of Xi’an Jiaotong University, focusing on O&O adults who had resided in Shaanxi Province over 6 months between August 2021 and July 2023. Participants were categorized based on their BMI into the overweight group (25.0∼27.9 kg/m²) and the obesity group (≥28.0 kg/m²) [15]. We also recruited healthy weight individuals (BMI = 18.0∼24.9 kg/m²) as controls for O&O. The detailed inclusion and exclusion criteria for each group can be found in online supplementary material A (for all online suppl. material, see https://doi.org/10.1159/000546406).

The demographic information such as age, gender, height, etc., was obtained from individuals’ hospitalization records. Serum biomarkers, including brain natriuretic peptide, high-sensitivity cardiac troponin T, high-sensitivity C-reactive protein, etc., were derived from historical laboratory results. LV end-diastolic and end-systolic diameter, left atrial (LA) diameter, left ventricular septum (LVS), left ventricular posterior wall (LVPW), and left ventricular fractional shortening (LVFS) were collected from echocardiographic reports directly.

CMR imaging serves as a noninvasive and reproducible quantitative assessment tool, exhibiting high sensitivity, specificity, and accuracy [16, 17]. All participants underwent scanning with a 3.0 T Ingenia CX scanner (Philips Healthcare, Best, The Netherlands). Steady-state free-precession cine imaging was conducted utilizing a 32-element phased-array body coil with ECG gating in both cardiac vertical and horizontal short-axis and long-axis orientations. The scanning parameters employed were as follows: time of repetition of 3.5 ms, time of echo of 1.72 ms, field of view of 320 mm × 320 mm, layer thickness of 8 mm, 8 layers in the LV short axis, 3 layers in the remaining images, a flip angle of 45°, and 30 cardiac cycles for the number of excitations. CMR data were analyzed using QMass MR 7.5 software (Medis, Netherlands) in accordance with the manufacturer’s instructions. The QMass module in Medis Suite software was utilized to delineate the endocardium and epicardium of the myocardium at the LV short-axis level (excluding papillary muscles) to derive cardiac structural and functional indicators such as left ventricular ejection fraction (LVEF), left ventricular cardiac output (LVCO), left ventricular stroke volume (LVSV), LV end-diastolic/systolic volume index, LV mass and LV mass index, etc. For assessment of intra- and interobserver variability, two investigators performed CMR QMass measurement independently of the other, to provide a blinded assessment in randomly selected 50 individuals from all participants. Intra-observer and interobserver variability was assessed by calculating the intra- and inter-class correlation coefficients and their 95% limits. During the measurements, we evaluated all contours labeled by the software and made manual adjustments when necessary [18, 19] (online suppl. material B, Fig. S1).

Baseline characteristics are presented as the mean ± standard deviation, median (interquartile range), or number (%), as appropriate. Demographic and clinical information was compared among the subgroups using two-sample independent t tests, analysis of variance (ANOVA) with post hoc Bonferroni correction, and Kruskal-Wallis test. Correlations between BMI and cardiac structural and functional indicators were assessed using Pearson’s correlation analysis (if normally distributed) or Spearman’s method. Univariate linear regression model was used to estimate the association of BMI and cardiac structural and functional indicators, where BMI was analyzed as continuous variables and triple categorized. To adjust for potential confounding factors, we established two backward stepwise multivariable linear regression models, and the details are provided in the online supplementary material C. The regression coefficient and its 95% CI are presented. In addition, we established a Loess spline (locally weighted regression) to illustrate the nonuniform changes in cardiac structural and functional indicators with BMI. All the analyses presented above were conducted by using R software (version 4.2.3). p values were two-sided and were considered statistically significant if p < 0.05.

Between August 2021 and July 2023, 346 participants with O&O were enrolled in our study. After applying the eligibility criteria, 175 individuals (mean age: 42.19 ± 11.37 years; 69.74% were male) were ultimately included in the final analysis (Fig. 1). Specifically, 100 (57.14%) were in the overweight group (mean age: 42.78 ± 11.36 years; 64.00% were male) and 75 (42.86%) were in the obesity group (mean age: 39.53 ± 10.65 years; 77.33% were male). We also enrolled 100 individuals as controls in the healthy weight group (mean age: 37.59 ± 12.04 years; 60.00% were male).

Demographically, there were no significant differences in height and history of alcoholism among the healthy weight, overweight, and obesity groups (all p > 0.05, ANOVA). However, a stepwise increase was observed across the groups in weight, systolic and diastolic blood pressure (all p < 0.001, ANOVA). Additionally, there was an increase in the frequency of hypertension, hyperlipidemia, and diabetes from healthy weight to obesity, with more significant differences observed in the obesity group (both p < 0.05 compared to healthy weight) than in the overweight group (p < 0.05 for hypertension and hyperlipidemia compared to healthy weight). Serum biomarkers such as hemoglobin A1c, triglycerides, and high-density lipoprotein cholesterol exhibited a gradual increasing or decreasing trend across groups, with abnormalities in some indicators beginning in the overweight group. Specific information is summarized in Table 1.

Figure 2 summarizes the cardiac characteristics of individuals from echocardiographic reports. Both the O&O group showed significant structural remodeling compared with the healthy weight group (Fig. 2a–e, p < 0.05 for overweight group and obesity group vs. healthy weight). Notably, LVS and LVPW did not differ between overweight group and obesity group (p = 0.41, p = 0.56, respectively). As for functional indicators, LVFS displayed a stepwise decrease from healthy weight to obesity group, with obesity group showing significant differences versus healthy weight and overweight group (p < 0.001, p = 0.001, respectively, Fig. 2f).

The CMR post-processed measurements showed that the O&O groups exhibited worse LVEF and LVSV index, larger LV end-diastolic/systolic volumes, and LV mass compared to the healthy weight (both p < 0.05). However, there were no significant differences in LVCO, LVSV, and LV end-diastolic volume index among the subgroups (both p > 0.05). When LVCO was adjusted to the LV cardiac index and LV end-systolic volume was adjusted to the LV end-systolic volume index, the only nonsignificant difference was observed between overweight group and obesity group (p = 0.73, p = 0.058, respectively). The specific characteristics are shown in Figure 3a–k.

Table 2 illustrates the correlations between cardiac structural and functional remodeling with BMI. Overall, BMI was correlated with all the included indicators except for LVCO and LVSV (r = 0.01 and r = −0.05, respectively; both p > 0.05). Specifically, BMI demonstrated a significant positive correlation with cardiac structural remodeling including LV end-diastolic/systolic diameter, LA diameter, etc. (r = 0.4233, r = 0.4221, and r = 0.5044, respectively; both p < 0.001). In terms of cardiac functional remodeling, BMI displayed a significant negative correlation with the LVFS, LVEF, LV cardiac index, and LVSV index (r = −0.3055, r = −0.4479, r = −0.2118, and r = −0.2634, respectively; both p < 0.001), and positively correlated with LV end-diastolic/systolic volume index, etc. (r = 0.1314 and r = 0.2669, respectively; both p < 0.05).

We further conducted a univariate regression analysis on cardiac structural and functional indicators that had correlation coefficients greater than 0.3 with BMI (online supp. material C, Table S1). The results showed one-unit increase in BMI was associated with a 0.84-mm, 1.05-mm, 0.70-mm, and 0.09-mm increase in LV end-diastolic/systolic diameter, LA diameter, and LVPW respectively; a 0.81 and 1.52 decrease in LVFS and LVEF; a 6.20-mL, 6.34-mL, 5.14-g, and 1.77-g/m² increase in LV end-diastolic/systolic volume, LV mass, and LV mass index, respectively. After adjusting for individuals’ demographic and clinical characteristics in multivariable linear regression analysis, increasing BMI still affects these abnormalities. However, this effect was only pronounced in the obesity group, particularly in model 2, where adjustments were made for potential metabolic abnormalities (systolic blood pressure, diastolic blood pressure, hemoglobin A1c, fasting glucose, total cholesterol, and triglycerides), and there were no differences in LVFS, LV end-diastolic volume, LV mass, and LV mass index between overweight group and the healthy weight group (online suppl. material C, Table S2).

The Loess spline demonstrated the nonuniform changes in selected cardiac structural and functional remodeling as BMI increases. The change in the slope of the curves indicated that all indicators underwent significant changes when the BMI reached the O&O stage. In addition, with the exception of LV end-diastolic diameter, the trends of the remaining indicators appeared more pronounced in the overweight stage than in the obesity stage (Fig. 4a–j, red points).

Intra-observer and interobserver variability of cardiac function was shown in online supplementary material D, Table S3. All the measurements of parameters showed good reproducibility, with inter-class correlation coefficient ranging from 0.617 to 0.980 in intra-observer and from 0.763 to 0.990 in interobserver, respectively.

The association between the progression of obesity and altered cardiac abnormalities has been recognized since the mid-19th century, with early autopsy studies revealing excessive fat accumulation on the heart’s surface and septum in individuals with obesity [20, 21]. As cardiac analysis techniques have advanced, we now understand that obesity can lead to various cardiac issues, including increased cardiac output, reduced peripheral vascular resistance, elevated LV diastolic pressure, and severe pulmonary hypertension. These conditions can eventually progress to left ventricular hypertrophy and systolic or diastolic dysfunction [22, 23]. Furthermore, individuals with obesity are at a heightened risk for neurohormonal and metabolic disorders, such as hypertension and leptin resistance, compared to those of normal weight [20, 24‒26]. Our observations indicate that individuals with O&O exhibit subclinical changes in metabolism, serum biomarkers, and cardiac structure and function as BMI increases. Notably, the differences in cardiac structure and function indicators between the O&O stages highlight significant distinctions in cardiac characteristics across these phases.

Specifically, through ANOVA, we found that in terms of cardiac structural changes, although alterations in chamber size were most evident at the obesity stage (Fig. 2a–c), changes in wall thickness appeared to be more pronounced at the overweight stage (Fig. 2d–e). Compared to structural changes, cardiac function exhibited more complex and irregular patterns. For example, changes in LVCO, LVSV, and LV end-diastolic volume index showed no significant differences across subgroups (Fig. 3b, d, g). However, the LV cardiac index and LV end-systolic volume index displayed more noticeable differences at the overweight stage compared to both the healthy weight and obesity groups (Fig. 3c, i), while other indicators, including LVFS, LVEF, and the LVSV index, were more prominent at the obesity stage. This observation has previously been observed primarily in adolescents with obesity, making our findings crucial for reevaluating cardiac health in adults affected by O&O [27, 28]. A possible explanation for this phenomenon is that adults with O&O experience initial hemodynamic changes within the LV, leading first to increased afterload and LV dilation. As a result, significant structural changes may occur before the development of overt LV dysfunction [22]. Furthermore, as the risk of metabolic conditions like hypertension or type 2 diabetes increases among individuals with O&O, metabolic abnormalities may cause inflammatory myocardial edema, elevating the extracellular volume and potentially accelerating cardiac remodeling before functional changes become apparent [12, 29, 30]. This finding underscores the potential of identifying new biomarkers to detect early cardiac abnormalities, highlighting the need for further prospective studies to validate their clinical relevance.

Subsequently, our multivariable linear regression analysis continued to show an association between BMI growth and changes in cardiac structure and function. An in-depth analysis using Loess spline, as a tool for identifying trends and patterns in cardiac structural and functional changes accompanying the increase in BMI, revealed that the primary remodeling of cardiac structure and function throughout the progression of O&O was more pronounced at the overweight stage compared to the obesity group (Fig. 4b–j). This phenomenon could be due to the higher weight-adjusted basal and resting metabolic rates in overweight individuals, resulting in more frequent and intense cardiac contractions, which may accelerate the progression of LV hypertrophy or LV systolic/diastolic dysfunction [31, 32]. Additionally, the overweight populations tend to have a younger age of onset, which may increase the duration of overweight status in an individual’s average lifespan (online supplementary material E, Table S4). This prolonged overweight state also contributes to increased cardiac volume and pressure overload, elevated demand for blood and oxygen supply, as well as heightened inflammation and oxidative stress [33‒36].

Abnormal changes in cardiac structure and function caused by O&O pose a threat to cardiac health. In addition to promptly addressing the direct effects of O&O, it is essential not only to focus on O&O itself but also to consider the social factors closely related to it, as well as changes in individuals’ lifestyles, behaviors, and psychological aspects, such as socioeconomic status, culture, occupation, diet, exercise, and sleep [2, 37]. These changes can be reflected in our results, including variations in participants’ demographic information and specific serum biomarkers (such as hemoglobin A1c, total cholesterol, triglycerides, and high-density lipoprotein cholesterol) across different subgroups (Table 1). Specifically, the unique dietary patterns such as high-fat, high-calorie carbohydrates, and meat, combined with unhealthy behaviors such as lack of exercise, late nights, smoking, and drinking, alongside increased psychological stress, anxiety, and depression due to social development, further exacerbate this issue [38, 39]. Fortunately, the changes in cardiac structure and function induced by O&O can be reversed through effective early intervention strategies, which also improve the neurohormonal and metabolic environment [40]. Therefore, early intervention and preventive strategies are crucial for patients with O&O. This is particularly important for overweight individuals, as early identification and management of risk factors may be more beneficial in reducing the risk of severe cardiovascular complications in the future than solely targeting those who are already experiencing obesity.

Our study had three main limitations. First, the definition of O&O or healthy weight in our study is based on BMI, which has been used as an international standard for assessing fat content for decades [41]. However, in the Chinese population, fat accumulation tends to be concentrated around the waist and abdominal area; therefore, waist or abdominal circumference may be more accurate indicators compared to BMI. The use of BMI could potentially lead to the misdiagnosis of individuals with higher muscle mass (such as athletes) [42, 43]. Second, our participants were screened from hospitalized patients, unavoidable confounding factors may affect the results of comparisons between different subgroups, so we established a multivariable linear regression model to avoid the effects of confounding factors as much as possible. Third, the period we selected for the study coincided with the COVID-19 pandemic, and related studies have shown that COVID-19 can also affect cardiac health [44, 45]. Even with strict exclusion criteria, we are still unable to completely eliminate this potential influence. Additionally, it would have been valuable to include right ventricular dimensions and function, as well as native T1 times and extracellular volume measurements across different subgroups, to assess whether cardiac fibrosis is more pronounced in individuals with O&O [46, 47]. Future research could also be enhanced by incorporating functional tests, such as cardiopulmonary exercise tolerance, to further evaluate cardiac performance [48]. These areas warrant further investigation, and additional longitudinal studies are needed to validate the findings of our current study.

Our study confirms that the cardiac structural and functional remodeling in O&O individuals is significantly associated with BMI. Additionally, alterations in cardiac structure and function differed between O&O, with some of cardiac structural and functional remodeling appearing more pronounced in the overweight state. Thus, early detection of heart damage in overweight individuals may facilitate the adoption of preventive strategies and potentially yield greater benefits in reducing CVD risk, rather than solely targeting populations with obesity and higher BMI.

The authors thank all coinvestigators for their generous dedication and support.

The current study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of the Ethics Committee at The First Affiliated Hospital of Xi’an Jiaotong University (Approval No. XJTU1AF2025LSYY-252). As this research involved only retrospective analysis of existing, de-identified data and posed no risk to participants, the need for informed consent was waived by the Institutional Review Board of the Ethics Committee at The First Affiliated Hospital of Xi’an Jiaotong University.

The authors declare that they have no competing interests.

This work was supported by the Foundation of Key Research and Development Plan of Shaanxi Province of China (2023-YBSF-403), Key Research and Development Program of Shaanxi (2021LL-JB-06).

Yue Wu, Zhijie Jian, and Wenjun Liu conceived the idea and designed the study. Shanshan Li, Yue Yu, Zixuan Meng, Yaxuan Xue, Renjie Liu, and Hui Liu were responsible for the data collection and analysis. Lele Cheng and Jian Yang provided technical support for radiographic method. Wenjun Liu, Shanshan Li, and Yue Yu wrote the manuscript draft. Zhijie Jian, Yue Wu, and Jian Yang critically revised the manuscript. All authors reviewed and approved the final manuscript.

Wenjun Liu and Shanshan Li contributed equally to this work.

Due to privacy concerns, all data included in this article and its online supplementary material file for generation or analysis are not publicly available. For further inquiries, contact the corresponding author.