Introduction

Obesity has been increasing globally and it is known to have adverse impacts on public health. Obesity is associated with increased cardiovascular risk factors such as hypertension, diabetes, dyslipidemia, and inflammation [1, 2]. Also, it can be related to changes in cardiac structure and function as well as hemodynamic alteration [3–5]. Obesity can increase total blood volume and cardiac output, thus increasing cardiac workload. As cardiac loading and left ventricular (LV) filling pressure have increased, the LV chamber can be enlarged and hypertrophied as a compensatory mechanism. These structural changes can lead to LV diastolic dysfunction and then subsequently systolic dysfunction in some cases. In addition, the left atrium (LA) can enlarge, therefore increasing the risk of atrial arrhythmia as well as heart failure [6].

Bariatric surgery is being considered as a therapeutic option for morbidly obese patients. Several studies have demonstrated lower overall mortality and cardiovascular events in patients who have undergone bariatric surgery [7, 8]. Data are accumulating showing that this surgical intervention can improve obesity-related morphologic and functional changes [9–11].

In the current study, we assessed the changes of left ventricular (LV) structure and function after bariatric surgery in the obese patients. Additionally, we utilized speckle tracking echocardiography to investigate the subtle changes in LV systolic function after surgical weight loss.

Methods

Study Population

Thirty-seven patients (age = 36 ± 10 years; male:female = 11:26) who had undergone a Roux-en-Y gastric bypass for obesity were included. Patients who had an atrial fibrillation or other significant arrhythmia, LV ejection fraction less than 50 %, or significant valvular disease were excluded. Clinical and echocardiographic parameters were assessed before and after bariatric surgery for at least 1 year from the surgery. In addition, echocardiographic assessment was done in age 37 and sex matched healthy controls with a BMI comparable to those after bariatric surgery (BMI 27.4 ± 4.2 in controls vs. 27.9 ± 3.5 in postoperative patients). This study was approved by an institutional review committee.

Blood Test and Body Fat Measurement

Fasting blood tests for glucose, insulin, and cholesterol were performed. An estimate of insulin resistance was calculated using a computer model for the homeostasis model assessment of insulin resistance (HOMA2-IR) [12, 13].

The distribution of abdominal fat was assessed by a computed tomography (CT) scan (Lightspeed VCT; GE Healthcare, Waukesha, WI, USA). A 5 mm CT slice scan was acquired at the L4-L5 level with the subject in a supine position. The attenuation interval was set between −250 and −50 Hounsfield units to indicate adipose tissue. The area of visceral adipose tissue (VAT) was measured in the muscle wall surrounding the abdominal cavity.

Echocardiography

Transthoracic echocardiograms were performed with a frame rate of 40 ~ 60 Hz. Standard echocardiographic parameters, including LV dimension and ejection fraction (EF), peak early (E) and late (A) mitral inflow velocities, deceleration time of early mitral flow velocity (DT), and early mitral annular velocity (e’) at the septal annular site were analyzed with an offline analysis workstation. LV internal dimension (LVID) and septal (SWT) and posterior wall thickness (PWT) were measured at end diastole according to American Society of Echocardiography recommendations [14]. LV mass was calculated with the following formula and corrected for the body surface: LV mass (g) = 0.8{1.04 x [(LVID + PWT + SWT)3 - (LVID)3]} + 0.6 [14]. Relative wall thickness (RWT) was also calculated by (2 x PWT)/LVID [14]. In addition, 2D speckle tracking (Artida TMR, Toshiba) was utilized to measure LV myocardial longitudinal, circumferential, and radial deformation. For 2D speckle tracking measurements, the software automatically tracked the motion through the cardiac cycle after tracing the subendocardium of the LV manually. Tracking quality was verified for each segment with manual adjustment if needed. The LV global longitudinal strain rate was averaged from apical 4- and 2-chamber views, and circumferential and radial strains were obtained from the parasternal short axis views at the level of the papillary muscle. Five patients were excluded in strain analysis because of poor speckle tracking.

Statistical Analysis

Continuous variables were expressed as mean ± standard deviation and categorical variables were expressed as numbers and percentages. We checked skewness and kurtosis for normality of our data. We assessed the changes of the parameters after bariatric surgery using a paired t test for normally distributed datasets or a Wilcoxon signed-rank test for non-normally distributed datasets where applicable. A Pearson’s correlation coefficient was calculated to investigate the relationship between the change in anthropometric variables and the change in hemodynamic or echocardiographic parameters. All statistical analyses were performed using STATA 14.0 (STATA Corp., College Station, TX, USA). For all tests, a P value <0.05 was considered to be significant.

Results

Patients Characteristics

The mean age of the patients was 36 ± 10 years (11 men). Among them, 19 patients (51 %) had hypertension, 16 patients (43 %) had diabetes, and 22 patients (57 %) had dyslipidemia. Mean BMI was 39.7 ± 5.5 and mean LVEF was 64.3 ± 3.1 %.

Clinical and Laboratory Changes After Bariatric Surgery

Bariatric surgery resulted in a significant decrease of BMI by 11.8 ± 4.7 over 15.6 ± 5.5 months (p < 0.001). After bariatric surgery, both the total fat area and the visceral fat area were reduced (all p < 0.001). Systolic and diastolic blood pressure decreased, and fasting glucose HOMA2-IR, and cholesterol levels were significantly reduced after surgical intervention (Table 1). Bariatric surgery resulted in the resolution of hypertension in 11/19 patients (58 %), diabetes in 14/16 patients (88 %), and dyslipidemia in 20/22 patients (91 %).

Table 1 Changes of clinical and laboratory parameters after bariatric surgery
Full size table

While change in diastolic BP was related to both changes in BMI and VAT area (r = 0.45, p = 0.005 for BMI; r = 0.34, p = 0.04 for VAT), change in HOMA2-IR was more related to change in the VAT area but not to change in BMI (r = 0.39, p = 0.02 for VAT; r = 0.33, p = 0.06 for BMI). Changes in lipid parameters were more associated with a change in BMI (Table 2).

Table 2 Relation between changes in BMI or VAT and changes in blood pressure, HOMA2-IR or lipid parameters
Full size table

Echocardiographic Parameters After Bariatric Surgery

Table 3 showed the changes of echocardiographic measurements after bariatric surgery. LV end-diastolic and end-systolic dimensions significantly decreased after weigh loss by surgery (51.0 ± 3.3 to 49.1 ± 3.4 mm, p < 0.001 for the LV end-diastolic dimension; 32.5 ± 3.5 to 30.2 ± 3.7 mm, p < 0.001 for the LV end-systolic dimension). LV mass as well as RWT decreased postoperatively (192.6 ± 33.5 to 146.2 ± 29.1 g, p < 0.001 for LV mass; 0.40 ± 0.06 to 0.35 ± 0.05, p < 0.001 for RWT). In addition, there was a significant increase in the ratio of early-to-late diastolic mitral inflow (E/A) and early diastolic tissue Doppler velocity (e’) at the follow-up (1.42 ± 0.52 to 1.59 ± 0.56, p = 0.04 for E/A ratio; 9.7 ± 2.0 cm/s to 11.0 ± 2.4 cm/s, p < 0.001 for e’). E/e’ decreased significantly postoperatively (p = 0.04). While the LV ejection fraction did not significantly change, LV longitudinal strain significantly increased after surgery (64.3 ± 3.1 to 64.2 ± 3.1 %, p = 0.77 for LV ejection fraction; 14.1 ± 1.9 to 16.2 ± 1.4 %, p < 0.001 for longitudinal strain). LV circumferential and radial strains were comparable at follow-up. While most echocardiographic parameters were not significantly different between postoperative patients and healthy subjects with comparable BMI, left atrial size was smaller and LV longitudinal strain was higher in the controls (p < 0.001 for left atrial size; p = 0.004 for longitudinal strain, Table 3).

Table 3 Changes in cardiac geometry and function after bariatric surgery
Full size table

Factors Related to Changes in LV Geometry and Function

LV mass reduction was related to the changes both in BMI and in VAT (r = 0.53, p < 0.001 for BMI; r = 0.47, p = 0.005 for VAT). Multivariate analysis showed both changes in BMI and VAT as the independent predictors of a change in LV mass (p = 0.016 for BMI, p = 0.03 for VAT). However, changes in RWT and e’ were related to a change in BMI (Fig. 1). Further, the change in LV longitudinal strain was associated with the change in LV mass (r = 0.37, p = 0.04) but not with the change in VAT or HOMA2-IR. An improvement of LV longitudinal strain was more prominent in the patients with depressed LV longitudinal strain (p < 0.001).

Fig. 1
figure 1

Relation between changes in BMI or VAT and echocardiographic parameters

Full size image

Discussion

Obesity can increase preload and afterload, thus resulting in increased LV wall stress. As a result, LV dilatation and a compensatory LV hypertrophy can occur. Eventually, this process can increase myocardial stiffness, leading to LV diastolic and systolic dysfunction [2, 3, 15]. Conversely, obesity can be associated with insulin resistance and neurohormonal adaptation such as activation of the renin-angiotension-aldsterone and sympathetic nervous system, which can also affect cardiac structure and size [16–19]. Weight loss by bariatric surgery is expected to improve these obesity-related changes.

In our study, we demonstrated a substantial reverse remodeling of LV size and mass after bariatric surgery in obese patients. In addition, surgical weight loss resulted in improvement of LV mechanics, represented by increased longitudinal strain. Our findings are comparable with prior studies which showed favorable effects on LV size and mass after weight loss, suggesting that it might be the possible mechanism by which bariatric surgery improves survival in patients with morbid obesity [9, 11, 20, 21]. We also found improvement in diastolic function and decreased LV filling pressure estimated by E/e’ in bariatric patients. Regression of LV mass and thickness would partly play a role in this favorable change in diastolic function. In contrast to LV diastolic function, change of LV systolic function remains controversial. Many studies did not show significant change in LV contractility, mostly as assessed by LVEF [9, 11, 21]. However, most studies have included patients with preserved LVEF, and this might impose limitations on evaluating LV systolic function with only LVEF in this population. Speckle tracking echocardiography is an emerging technique for detecting subtle change in LV systolic function. Previous studies have shown the utility of strain measurements by speckle tracking echocardiography beyond LVEF [22–24]. In our study, we applied speckle tracking echocardiography to assess the changes in LV deformation on bariatric patients with preserved EF and showed an increase of LV longitudinal function, but not circumferential or radial strain and EF postoperatively. It is not surprising when considering the observations that LV longitudinal strain mostly represents the function of subendocardial longitudinal fibers, which could be more sensitive to change of wall stress according to obesity status [25]. The LV myocardial fibers are arranged in a double helical orientation, where right-handed helical geometry in the subendocardium gradually changes into left-handed helix in the subepicardium [26]. This counterdirectional helical arrangement of myofibers results in energetically great efficiency with multidirectional deformation of the LV in both long and short axes. While the subendocardial fibers are arranged mostly parallel to the long-axis of the LV, mid-layer and subendocardial fibers run relatively parallel to the short-axis of the LV. The subendocardium is generally more vulnerable to the pathologic injury than the subepicardium. Thus, the long axis function can be impaired first in the disease process [27, 28]. At this stage, circumferential and radial contraction remains preserved or may even be accentuated to compensate impaired longitudinal function [23, 24]. As the disease progress, both the subendocardium and the subepicardium are involved, and, subsequently, LV short-axis function as well as long-axis function gets compromised.

In our data, changes in BP and LV mass were related to change in both BMI and visceral adiposity; whereas, metabolic change such as insulin sensitivity are more related to change in visceral adiposity rather than body mass itself. Weight loss by bariatric surgery seemed to impact glucose metabolism and insulin resistance by decreases in fat mass and changes in the release of adipocytokines [29, 30], Many studies have previously reported the beneficial effects of weight loss by bariatric surgery on cardiovascular risk factors and marked rates of remission or improvements for hypertension, diabetes, and dyslipidemia [31, 32]. We also confirmed decreases in systolic and diastolic blood pressure as well as improvement of insulin sensitivity and lipid parameters after surgery. Physical changes of restrictive and malabsorptive properties can influence body mass. In addition to physical change after bariatric surgery, changes of adipoinsular and adipoinsular axes could contribute to this improvement of systemic metabolism. The direct effects of substantial weight loss accompanied by other beneficial hemodynamic changes, such as a decrease of plasma volume and blood pressure, can favorably impact LV mass and regression. Also, bariatric surgery might lead to additional beneficial effects on cardiac size and function by modulating metabolism by gut hormones [20, 21, 33]. Interestingly, our data showed that change in LV longitudinal strain was more related to the change in LV mass rather than changes in VAT or HOMA-IR. While more long-term data with larger population would be necessary to determine this relationship, LV mechanic seemed to be changed more sensitively by reduction of LV mass than metabolic change.

Our study also had several limitations. First, the study population was small. Second, we only included the patients who could be followed up at least for 1 year to investigate the change of LV structure and function. Thus, it might lead to selection bias. Third, we only studied patients who had bariatric surgery and could not elucidate if surgical weight loss would have different impacts compared to dietary weight loss in our data. Finally, we did not evaluate long term cardiovascular morbidity and mortality in our data. Further study would be necessary to elucidate if this favorable LV reverse remodeling would lead to improved clinical outcome.

In conclusion, significant weight loss by bariatric surgery led to a regression of LV hypertrophy and improvement of LV longitudinal deformation and diastolic function among obese patients. These results might be a plausible mechanism yielded from the beneficial effects of bariatric surgery for future cardiovascular outcomes in obese patients.