Impact of Diabetes and Increasing Body Mass Index Category on Left Ventricular Systolic and Diastolic Function

CARDIOMETABOLICS

Impact of Diabetes and Increasing Body Mass Index Category on Left Ventricular Systolic and Diastolic Function Arnold C. T. Ng, MBBS, PhD, Francesca Prevedello, MD, Giulia Dolci, MD, Cornelis J. Roos, MD, Roxana Djaberi, MD, Matteo Bertini, MD, PhD, See Hooi Ewe, MBBS, Christine Allman, MSc, Dominic Y. Leung, MBBS, PhD, Nina Ajmone Marsan, MD, PhD, Victoria Delgado, MD, PhD, and Jeroen J. Bax, MD, PhD, Woolloongabba, Brisbane, and Sydney, Australia; Leiden, The Netherlands; and Padova and Verona, Italy

Background: Diabetes and obesity are both worldwide growing epidemics, and both are independently associated with increased risk for heart failure and death. The aim of this study was to examine the additive detrimental effect of both diabetes and increasing body mass index (BMI) category on left ventricular (LV) myocardial systolic and diastolic function. Methods: The present retrospective multicenter study included 653 patients (337 with type 2 diabetes and 316 without diabetes) of increasing BMI category. All patients had normal LV ejection fractions. LV myocardial systolic (peak systolic global longitudinal strain and peak systolic global longitudinal strain rate) and diastolic (average mitral annular e0 velocity and early diastolic global longitudinal strain rate) function was quantified using echocardiography. Results: Increasing BMI category was associated with progressively more impaired LV myocardial function in patients with diabetes (P < .001). Patients with diabetes had significantly more impaired LV myocardial function for all BMI categories compared with those without diabetes (P < .001). On multivariate analysis, both diabetes and obesity were independently associated with an additive detrimental effect on LV myocardial systolic and diastolic function. However, obesity was associated with greater LV myocardial dysfunction than diabetes. Conclusion: Both diabetes and increasing BMI category had an additive detrimental effect on LV myocardial systolic and diastolic function. Furthermore, increasing BMI category was associated with greater LV myocardial dysfunction than diabetes. As they frequently coexist together, future studies on patients with diabetes should also focus on obesity. (J Am Soc Echocardiogr 2018;31:916-25.) Keywords: Cardiomyopathy, Diabetes mellitus, Obesity, Ventricles

Department of Cardiology, Princess Alexandra Hospital, The University of Queensland, Woolloongabba (A.C.T.N.), and Translational Research Institute, Brisbane (A.C.T.N.); Department of Cardiology, Liverpool Hospital, Sydney (C.A., D.Y.L.), Australia; Department of Cardiology, Leiden University Medical Center, Leiden, The Netherlands (A.C.T.N., F.P., G.D., C.J.R., R.D., M.B., S.H.E., N.A.M., V.D., J.J.B.); University of Padua, Department of Cardiac, Thoracic and Vascular Sciences, Padova (F.P.); Section of Cardiology, Department of Medicine, University of Verona, Verona (G.D.), Italy. Conflicts of Interest: The Department of Cardiology at Leiden University Medical Center receives unrestricted research grants from Biotronik, Medtronic, Boston Scientific, Edwards Lifesciences, and General Electric. Victoria Delgado received speaker fees from Abbott Vascular. The remaining authors reported no potential or actual conflicts of interest. Reprint requests: Jeroen J. Bax, MD, PhD, Leiden University Medical Center, Department of Cardiology, Albinusdreef 2, 2333 ZA Leiden, The Netherlands (E-mail: [email protected]). 0894-7317/$36.00 Copyright 2018 by the American Society of Echocardiography. https://doi.org/10.1016/j.echo.2018.02.012

916

There is currently a worldwide epidemic of obesity and type 2 diabetes. The latest projection by the World Health Organization estimated that globally in 2005, approximately 1.6 billion adults >15 years of age were overweight, and at least 400 million adults were obese. Because of the obesity epidemic, there is a concomitant increase in the prevalence of type 2 diabetes. In 2000, the World Health Organization estimated that more than 170 million people worldwide had diabetes, and the prevalence was projected to double in the next 20 years.1 Both obesity and diabetes are independently associated with increased risk for heart failure.2 Although the pathophysiologic mechanisms underlying obesity and diabetic cardiomyopathy are not identical, the combination of insulin resistance, hyperinsulinemia, and hyperglycemia leads to inflammation, neurohormonal activation of the renin-angiotensin-aldosterone system, and eventual myocardial structural and functional changes.3-8 Despite previous studies showing obesity to be an independent risk factor for subsequent development of diabetes and heart failure,2,9,10 few have examined the simultaneous impact of increasing body mass index (BMI) category and concomitant diabetes on changes in left ventricular (LV) myocardial function. We hypothesized that both increasing BMI

Journal of the American Society of Echocardiography Volume 31 Number 8

category and diabetes are independently associated with 2D = Two-dimensional progressive impairment of LV myocardial systolic and diastolic ANOVA = Analysis of functions and that the association variance is additive and not synergistic. BMI = Body mass index Thus, we conducted a multicenter retrospective study BSA = Body surface area (Leiden University Medical LV = Left ventricular Center, The Netherlands, and Liverpool Hospital, Australia) LVEF = Left ventricular ejection fraction whereby both patients with diabetes and those without LVESV = Left ventricular enddiabetes without coronary artery systolic volume disease were evaluated in order to (1) examine the impact of increasing BMI category on LV myocardial systolic (peak systolic global longitudinal strain and peak systolic global longitudinal strain rate) and diastolic function (average mitral annular e0 velocity and early diastolic global longitudinal strain rate) as quantified by echocardiography in patients with type 2 diabetes, (2) compare LV myocardial systolic and diastolic function with increasing BMI category between patients with diabetes and those without diabetes, and (3) determine the independent and additive detrimental effect of increasing BMI category and diabetes on LV myocardial systolic and diastolic function. Abbreviations

METHODS Patient Population The overall patient population consisted of 653 patients recruited from two institutions (104 from Liverpool Hospital and 549 from Leiden University Medical Center). All patients were identified over a 10-year period from the Australian and Dutch departmental combined echocardiographic and clinical databases. Of these, 337 had type 2 diabetes, which was diagnosed according to World Health Organization criteria.11 Although BMI does not take into account the wide variation in body fat distribution, it is the most useful population-level measure of obesity and is recommended by the World Health Organization to define overweight and obesity within a population and the risks associated with it.12 Furthermore, several multicenter and epidemiologic studies have demonstrated the independent prognostic value of BMI as a measure of general obesity for predicting all-cause mortality.13-15 As there were only two patients with diabetes with BMI < 20 kg/m2 measured at the time of echocardiography, all 337 patients with diabetes were divided into three categorical groups: 80 lean patients with diabetes (BMI < 25 kg/m2), 139 overweight patients with diabetes (BMI 25– 29.9 kg/m2), and 118 obese patients with diabetes (BMI $ 30 kg/m2). Patients with type 2 diabetes were compared against 316 patients without diabetes of similar age, gender, and BMI. All patients without diabetes were clinically referred for assessment of LV and/or valvular function and had structurally normal hearts on echocardiography. Similarly, as there were only five patients without diabetes with BMI < 20 kg/m2, all 316 patients without diabetes were divided into three categorical groups: 89 lean patients without diabetes (BMI < 25 kg/ m2), 134 overweight patients without diabetes (BMI 25–29.9 kg/ m2), and 93 obese patients without diabetes (BMI $ 30 kg/m2). The exclusion criteria for all patients with diabetes and without diabetes included age < 18 years, rhythm other than sinus rhythm, LV

Ng et al 917

ejection fraction (LVEF) < 50%, moderate or severe valvular stenosis or regurgitation, and congenital heart disease. To avoid coronary artery disease as a potential confounding factor for any changes observed in myocardial function, all patients with known significant underlying coronary artery disease, previous myocardial infarction, previous coronary artery bypass surgery or percutaneous coronary intervention, presence of segmental wall motion abnormalities on echocardiography, or positive results on stress testing were excluded. All patients underwent history, physical, biochemical, and transthoracic echocardiographic examinations. Baseline biochemical analyses included hemoglobin level, glomerular filtration rate calculated by the Modification of Diet in Renal Disease formula as recommended by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative Guidelines16 and glycated hemoglobin level. The definition of hypertension was different between patients with and those without diabetes. In patients with diabetes, the cutoff was >130/80 mm Hg on two separate occasions after >5 min of rest. In patients without diabetes, the cutoff was >140/90 mm Hg on two separate occasions after >5 min of rest. All clinical and biochemical variables were collected by an independent observer blinded to the echocardiographic results. The impact of increasing BMI categories on LV structure (LV volumes and mass) and function was initially assessed in the patients with type 2 diabetes. LV myocardial function within each BMI category (lean, overweight, and obese) in the diabetic population was then compared against patients without diabetes. Finally, to determine the independent and additive detrimental effect of increasing obesity and diabetes on LV myocardial function, multivariate analysis was performed with BMI categories and the presence or absence of diabetes entered as covariates, adjusted for baseline age, gender, systolic blood pressure, heart rate, LV mass, and LV volume. Echocardiographic analyses for all patients with and without diabetes, including two-dimensional (2D) speckle-tracking, were performed offline. Therefore, the present evaluation does not tabulate the results summarized in clinical reports. The institutional review boards approved the study. The institutional review board of the Leiden University Medical Center waived the need to obtain patient written informed consent for retrospective analysis of clinically acquired data anonymously handled. Echocardiography Transthoracic echocardiography was performed in all subjects at rest using commercially available ultrasound systems (Vivid 7 and E9, GE Vingmed Ultrasound, Horten, Norway). All images were digitally stored on hard disks for offline analysis (EchoPAC version 108.1.5; GE Vingmed Ultrasound). A complete 2D, color, pulsed, and continuous-wave Doppler echocardiographic examination was performed according to standard techniques.17,18 LVend-diastolic volume and LV end-systolic volume (LVESV) were calculated using the Simpson biplane method of disks. LVEF was calculated and expressed as a percentage. LV mass was calculated from the formula as recommended by the American Society of Echocardiography.19 Transmitral inflow velocities were recorded using conventional pulsed-wave Doppler echocardiography in the apical four-chamber view using a 2-mm sample volume. Transmitral early (E-wave) and late (A-wave) diastolic velocities as well as deceleration time were recorded at the mitral leaflet tips. Average mitral annular e0 velocity and average E/e0 ratio were obtained from the septal and lateral annulus as recommended by current guidelines.20 Maximal left atrial volume was calculated using the Simpson biplane method of discs in the fourand two-chamber views.

918 Ng et al

HIGHLIGHTS  Diabtetics have more impaired LV function in all categories of BMI than controls.  BMI and diabetes are independent predictors of impaired LV myocardial function.  Increasing BMI and diabetes have an additive detrimental effect on LV function. Myocardial Functional Assessment by 2D Speckle-Tracking Quantification of longitudinal myocardial function was performed using 2D speckle-tracking echocardiography in the three apical (two-, three-, and four-chamber) views. During image analysis, the LV endocardial border was manually traced at end-systole, and the region of interest width was adjusted to include the entire myocardium. The 2D speckle-tracking software then automatically tracks the motion of LV myocardial segments throughout the entire cardiac cycle. LV myocardial segments of good tracking quality were automatically accepted for further analyses, whereas poorly tracked segments were rejected, while simultaneously allowing the user to manually override the software’s decisions on the basis of visual assessments of tracking quality. From the three individual apical views, peak systolic global longitudinal strain, peak systolic global longitudinal strain rate, and early diastolic global longitudinal strain rate were calculated. Variability Analysis Intraobserver and interobserver measurement variabilities were performed in 20 randomly selected patients and expressed as mean absolute difference 6 1 SD. The respective intraobserver and interobserver measurement variabilities were 1.2 6 0.6% and 1.2 6 1.0% for peak systolic global longitudinal strain, 0.10 6 0.06 and 0.11 6 0.08 sec1 for peak systolic global longitudinal strain rate, and 0.09 6 0.05 and 0.16 6 0.09 sec1 for early diastolic global longitudinal strain rate. Statistical Analysis All continuous variables were tested for Gaussian distribution using the Kolmogorov-Smirnov test. Continuous variables are presented as mean 6 1 SD and categorical variables as frequencies and percentages. Unpaired Student’s t test was used to compare two independent groups of continuous variables, and the c2 test with Yates’s correction was used to compare categorical variables. To assess the univariate linear relationship between two variables, Pearson correlation was performed between two continuous variables, Spearman correlation was performed between one continuous variable and one ordinal variable (e.g., BMI categories), and point-biserial correlation was performed between one continuous and one dichotomous variable (e.g., gender and presence of diabetes). One-way analysis of variance (ANOVA) was initially used to examine the influence of increasing BMI subgroup categories on LV myocardial function (peak systolic global longitudinal strain, peak systolic global longitudinal strain rate, average mitral annular e0 velocities, average E/e0 ratio, and early diastolic global longitudinal strain rate) in patients with type 2 diabetes. Next, factorial

Journal of the American Society of Echocardiography August 2018

ANOVA was used to compare peak systolic global longitudinal strain, peak systolic global longitudinal strain rate, average mitral annular e0 velocities, average E/e0 ratio, and early diastolic global longitudinal strain rate changes with increasing BMI category between patients with and those without diabetes. Finally, multiple linear regression analyses were used to determine the independent and additive detrimental effect of increasing BMI category and diabetes on peak systolic global longitudinal strain, peak systolic global longitudinal strain rate, average mitral annular e0 velocities, and early diastolic global longitudinal strain rate, with correction for baseline age, gender, systolic blood pressure, heart rate, LV mass and LVESV. Standardized coefficients were presented to demonstrate the relative contribution of each variable to the multivariate linear regression model. To avoid multicollinearity, a tolerance of >0.5 was set. All post hoc multiple pairwise comparisons were performed with Bonferroni corrections. A two-tailed P value of <.05 was considered to indicate statistical significance. All statistical analyses were performed using SPSS version 17 for Windows (SPSS, Chicago, IL).

RESULTS Increasing BMI Category in Patients with Diabetes Table 1 outlines the clinical, biochemical, and echocardiographic characteristics of the entire cohort of patients with type 2 diabetes and the three diabetic groups categorized according to BMI category. The mean age was 57 6 12 years, and 63.2% were men. With increasing BMI category, there were progressive increases in systolic (P = .002, one-way ANOVA) and diastolic (P < .001, one-way ANOVA) blood pressures and higher glycated hemoglobin level (P = .046, one-way ANOVA). On echocardiography, increasing BMI category was associated with LV structural changes with progressive increases in LV enddiastolic volume, LVESV, and LV mass. Assessment of LV systolic function showed that LVEF did not differ significantly across the diabetic BMI subgroups. However, increasing BMI category was significantly associated with progressively more impaired peak systolic global longitudinal strain (P < .001, one-way ANOVA). Multiple pairwise comparisons with Bonferroni corrections showed that peak systolic global longitudinal strain became increasingly more impaired with each increase in BMI category (P < .001 for both). Similarly, increasing BMI category was also significantly associated with progressively more impaired peak systolic global longitudinal strain rate (P < .001, one-way ANOVA), and multiple pairwise comparisons with Bonferroni corrections showed that peak systolic global longitudinal strain rate was significantly more impaired in overweight versus lean patients with diabetes (P = .008) and in obese versus overweight patients with diabetes (P = .003). Assessment of LV diastolic function showed that increasing BMI category in patients with diabetes was significantly associated with progressive prolongation of transmitral deceleration time and a nonsignificant trend toward worsening of transmitral E/A ratio. Increasing BMI category was also significantly associated with progressively more impaired average mitral annular e0 velocity (P < .001, one-way ANOVA) and average E/e0 ratio (P = .004, oneway ANOVA). Similarly, increasing BMI was significantly associated with progressively more impaired early diastolic global longitudinal strain rate (P < .001, one-way ANOVA). On multiple pairwise comparisons with Bonferroni corrections, early diastolic global

Journal of the American Society of Echocardiography Volume 31 Number 8

Ng et al 919

Table 1 Clinical, biochemical, and echocardiographic characteristics of patient with diabetes

Variable

Total diabetic population (N = 337)

Lean diabetic (BMI < 25 kg/m2) (n = 80)

Overweight diabetic (BMI 25–29.9 kg/m2) (n = 139)

Obese diabetic BMI ($30 kg/m2) (n = 118)

P*

Clinical 57 6 12

Age (y) Male

56 6 12

57 6 12

63.2

67.5

74.1

56 6 11 47.5

.63 <.001

Height (cm)

173 6 10

173 6 10

172 6 11

173 6 10

.83

Weight (kg)

84 6 17

69 6 10

81 6 11

101 6 15

<.001

BMI (kg/m2)

28.9 6 5.6

22.8 6 1.5

27.2 6 1.4

35.0 6 4.4

<.001

Hypertension

56.4

42.5

56.1

66.1

.005

Hyperlipidemia

51.0

40.0

54.0

55.1

.076

Family history of ischemic heart disease

15.7

20.0

16.5

11.9

.29 .37

Current smoker

29.4

28.8

25.9

33.9

Systolic blood pressure (mm Hg)

139 6 20

134 6 20

139 6 19

144 6 20

.002

Diastolic blood pressure (mm Hg)

81 6 11

78 6 11

81 6 10

85 6 11

<.001

Hemoglobin (g/dL)

13.9 6 1.6

13.6 6 1.8

14.1 6 1.5

13.9 6 1.6

.076

Glomerular filtration rate (mL/min/1.73 m2)

87.7 6 26.9

85.2 6 22.8

87.1 6 27.7

90.2 6 28.6

.42

7.3 6 1.5

7.1 6 1.4

7.2 6 1.5

7.6 6 1.4

.046

Biochemical

HbA1c (%) Echocardiography Heart rate (beats/min)

74 6 13

72 6 13

74 6 12

75 6 14

.20

LV end-diastolic volume (mL)

93 6 24

88 6 23

93 6 25

98 6 24

.020

LVESV (mL)

38 6 12

36 6 12

37 6 12

40 6 12

.048

LVEF (%)

59 6 5

59 6 5

60 6 5

59 6 5

.53 <.001

LV mass (g)

183 6 49

161 6 42

182 6 47

200 6 48

Transmitral E/A ratio

0.97 6 0.32

1.04 6 0.40

0.96 6 0.28

0.94 6 0.30

.084

Deceleration time (msec)

198 6 54

190 6 58

192 6 45

209 6 58

.016

Average mitral annular e0 velocity (cm/sec)

6.7 6 2.0

7.5 6 2.2

6.5 6 1.9

6.3 6 1.8

<.001

0

Average E/e ratio

10.9 6 5.3

9.5 6 4.6

10.7 6 4.1

12.1 6 6.6

.004

Maximal left atrial volume (mL)

58.5 6 17.9

54.6 6 14.7

57.3 6 16.7

62.5 6 20.3

.006

Peak systolic global longitudinal strain (%)

17.6 6 2.3

18.9 6 2.2

17.7 6 1.9

16.6 6 2.3

<.001

Peak systolic global longitudinal strain rate (sec1)

0.93 6 0.16

1.01 6 0.17

0.94 6 0.14

0.88 6 0.16

<.001

0.99 6 0.27

1.13 6 0.28

1.00 6 0.25

0.89 6 0.25

<.001

Early diastolic global longitudinal strain rate (sec1) HbA1c, Glycated hemoglobin. Data are expressed as mean 6 SD or as percentages. *P value by one-way ANOVA.

longitudinal strain rate was significantly more impaired in overweight versus lean patients with diabetes (P = .001) and in obese versus overweight patients with diabetes (P = .004). Comparisons between Patients with and Those without Diabetes Table 2 compares the clinical and echocardiographic characteristics of patients with versus those without diabetes. There were no significant differences in age, gender, and BMI. Patients with diabetes were more likely to have histories of hypertension and hyperlipidemia. However, there were no significant differences in systolic and diastolic blood pressure at the time of echocardiographic examination. On echocardiography, there were no significant differences in LV end-diastolic volume, LVESV, and LV mass. Assessment of LV systolic function demonstrated that there was no significant difference in

LVEF between patients with and those without diabetes (59 6 5% vs 60 6 5%, P = .40). However, patients with diabetes had significantly more impaired peak systolic global longitudinal strain compared with those without diabetes (17.6 6 2.3% vs 18.9 6 2.4%, P < .001). Furthermore, patients with diabetes had more impaired peak systolic global longitudinal strain than those without diabetes across all BMI subgroups (P < .001, factorial ANOVA), and there was no significant interaction between the presence of diabetes and BMI categories (P = .31, factorial ANOVA; Figure 1). Thus, lean patients with diabetes had similar peak systolic global longitudinal strain as overweight patients without diabetes (18.9 6 2.2% vs 19.0 6 2.2%, P > .99 with Bonferroni correction), overweight patients with diabetes had similar peak systolic global longitudinal strain as obese patients without diabetes (17.7 6 1.9% vs 17.4 6 2.3%, P = .70 with Bonferroni correction),

920 Ng et al

Journal of the American Society of Echocardiography August 2018

Table 2 Comparison of clinical and echocardiographic characteristics between patients without and those with diabetes Variable

Nondiabetic (n = 316)

Diabetic (n = 337)

P

Clinical Age (y) Male

57 6 14

57 6 12

62.7

63.2

.83 .89

28.0 6 4.9

28.9 6 5.6

.074

Lean (BMI < 25 kg/m2)

28.2

23.8

.24

Overweight (BMI 25–29.9 kg/m2)

42.4

41.2

—

BMI (kg/m2)

Obese (BMI $ 30 kg/m2) Hypertension

29.4

35.0

—

27.5

56.4

<.001

Hyperlipidemia

14.2

51.0

<.001

Current smoker

11.4

15.7

.11

Systolic BP (mm Hg)

136 6 24

139 6 20

.08

Diastolic BP (mm Hg)

82 6 12

81 6 11

.29

Heart rate (beats/min)

71 6 13

74 6 13

.010

LV end-diastolic volume (mL)

95 6 25

93 6 24

.30

LVESV (mL)

39 6 12

38 6 12

.57

LVEF (%)

60 6 5

59 6 5

.40

LV mass (g)

183 6 51

183 6 49

.94

Transmitral E/A ratio

1.10 6 0.47

0.97 6 0.32 <.001

Deceleration time (msec)

212 6 60

198 6 54

.001

Average mitral annular e0 velocity (cm/s)ec

7.5 6 2.5

6.7 6 2.0

<.001

Average E/e0 ratio

10.2 6 4.1

10.9 6 5.3

.06

Maximal left atrial volume (mL)

59.4 6 21.3

58.5 6 17.9

.56

Echocardiography

Peak systolic global longitudinal strain (%)

18.9 6 2.4

Peak systolic global longitudinal strain rate (sec1)

0.99 6 0.16 0.93 6 0.16 <.001

Early diastolic global longitudinal strain rate (sec1)

1.15 6 0.32

17.6 6 2.3

<.001

0.99 6 0.27 <.001

BP, Blood pressure. Data are expressed as mean 6 SD or as percentages.

and obese patients with diabetes had the most impaired peak systolic global longitudinal strain (16.6 6 2.3%). Similarly, patients with diabetes had more impaired peak systolic global longitudinal strain rate than those without diabetes across all BMI subgroups (P < .001, factorial ANOVA), and there was no significant interaction between the presence of diabetes and BMI categories (P = .67, factorial ANOVA; Figure 2). Thus, lean patients with diabetes had similar peak systolic global longitudinal strain rate as overweight patients without diabetes (1.01 6 0.17 vs 0.98 6 0.16 sec1, P = .68 with Bonferroni correction), overweight patients with diabetes had similar peak systolic global longitudinal strain rate as obese patients without diabetes (0.94 6 0.14 vs. 0.92 6 0.15 sec1, P = .80 with Bonferroni correction), and obese

patients with diabetes had the most impaired peak systolic global longitudinal strain rate (0.88 6 0.16 sec1). On assessment of LV diastolic function, patients with diabetes had more impaired LV diastolic function as measured by average mitral annular e0 velocity (6.7 6 2.0 vs 7.5 6 2.5 cm/sec), P < .001, and across all BMI categories (P < .001, factorial ANOVA; Figure 3). There was no significant interaction between the presence of diabetes and BMI categories in average mitral annular e0 velocity (P = .63, factorial ANOVA). As shown in Figure 3, lean patients with diabetes had similar average mitral annular e0 velocity as overweight patients without diabetes (7.5 6 2.2 vs 7.3 6 2.8 cm/sec, P > .99 with Bonferroni correction), overweight patients with diabetes had similar peak systolic global longitudinal strain rate as obese patients without diabetes (6.5 6 1.9 vs 6.8 6 2.0 cm/sec, P = .56 with Bonferroni correction), and obese patients with diabetes had the most impaired peak systolic global longitudinal strain rate (6.3 6 1.8 cm/sec). In contrast, there was no significant difference in the assessment of LV filling pressure between patients with and those without diabetes as quantified by average E/e0 ratio in patients with diabetes (10.9 6 5.3 vs 10.2 6 4.1, P = .06). Finally, patients with diabetes also had more impaired early diastolic global longitudinal strain rate than those without diabetes across all BMI categories (P < .001, factorial ANOVA). Interestingly, there was a significant interaction between the presence of diabetes and increasing BMI categories (P = .005, factorial ANOVA). Evaluation of Figure 4 demonstrated that patients without diabetes had a greater initial decline in early diastolic global longitudinal strain rate from normal weight to overweight compared with patients with diabetes. Similarly, lean patients with diabetes had similar early diastolic global longitudinal strain rate as overweight patients without diabetes (1.13 6 0.28 vs 1.11 6 0.29 sec1, P > .99 with Bonferroni correction), overweight patients with diabetes had similar early diastolic global longitudinal strain rate as obese patients without diabetes (1.00 6 0.25 vs 0.99 6 0.26 sec1, P > .99 with Bonferroni correction), and obese patients with diabetes had the most impaired early diastolic global longitudinal strain rate (0.89 6 0.25 sec1). Diabetes, Obesity, and Myocardial Dysfunction To examine the independent and additive detrimental effect of both increasing obesity and diabetes on peak systolic global longitudinal strain, peak systolic global longitudinal strain rate, average mitral annular e0 velocity and early diastolic global longitudinal strain rate, multiple linear regression analyses were performed with the presence of diabetes and BMI categories entered as covariates, corrected for baseline age, gender, systolic blood pressure, heart rate, LV mass, and LVESV. Table 3 shows that both increasing BMI category and the presence of diabetes were independently associated with peak systolic global longitudinal strain (model R = 0.56, P < .001). Furthermore, increasing BMI (standardized b = 0.379, P < .001) was associated with greater LV myocardial dysfunction than diabetes (standardized b = 0.231, P < .001). There was no significant interaction between the presence of diabetes and increasing obesity, suggesting an additive (not synergistic) detrimental effect of diabetes and obesity on LV myocardial function. Similar results were also obtained when BMI was modeled as a continuous variable (standardized b = 0.355, P < .001), when systolic blood pressure was substituted by diastolic blood pressure or history of hypertension or when all patients with hypertension were excluded from the multivariate analysis.

Journal of the American Society of Echocardiography Volume 31 Number 8

Ng et al 921

Figure 1 Comparisons of peak systolic global longitudinal strain (GLS) in patients with versus those without diabetes across BMI categories. Increasing BMI was associated with progressive impairment of peak systolic global longitudinal strain. Furthermore, patients with diabetes had more impaired peak systolic global longitudinal strain across all BMI categories. Table 4 shows the multivariable regression analysis for peak systolic global longitudinal strain rate (model R = 0.53, P < .001). Similarly, increasing BMI category (standardized b = 0.285, P < .001) was associated with greater LV myocardial dysfunction by peak systolic global longitudinal strain rate compared with the presence of diabetes (standardized b = 0.180, P < .001). Similar results were also obtained when BMI was modeled as a continuous variable (standardized b = 0.247, P < .001). Table 5 shows the multivariate regression analysis for average mitral annular e0 velocity (model R = 0.68, P < .001). Increasing

BMI (standardized b = 0.157, P < .001) was associated with greater LV diastolic dysfunction by average mitral annular e0 velocity compared with the presence of diabetes (standardized b = 0.146, P < .001). Finally, Table 6 shows the multivariate regression analysis for early diastolic global longitudinal strain rate (model R = 0.64, P < .001). Increasing BMI category (standardized b = 0.337, P < .001) was associated with greater LV myocardial dysfunction by early diastolic global longitudinal strain rate compared with the presence of diabetes (standardized b = 0.230, P < .001). Finally, similar results were also

922 Ng et al

Figure 2 Comparisons of peak systolic global longitudinal strain rate in patients with versus those without diabetes across BMI categories. Increasing BMI was associated with progressive impairment of peak systolic global longitudinal strain rate. Patients with diabetes had more impaired peak systolic global longitudinal strain across all BMI categories.

Journal of the American Society of Echocardiography August 2018

Figure 4 Comparisons of early diastolic global longitudinal strain rate in patients with versus those without diabetes across BMI categories. Increasing BMI was associated with progressive impairment of early diastolic global longitudinal strain rate. Patients with diabetes had more impaired early diastolic global longitudinal strain across all BMI categories. However, patients without diabetes with normal BMI had a greater decline in early diastolic global longitudinal strain rate with increasing BMI.

Table 3 Independent determinants of LV peak systolic global longitudinal strain Peak systolic global longitudinal strain Variable

obtained when BMI was modeled as a continuous variable (standardized b = 0.297, P < .001).

DISCUSSION The present multicenter observational study demonstrated that increasing BMI category was associated with progressive and detrimental changes in LV structure and systolic and diastolic function in patients with diabetes. Compared with patients without diabetes of similar age, gender, and BMI, those with dia-

P

Standardized b

P

0.027

.49

0.027

.49

Male

0.063

.11

0.063

.087

Systolic blood pressure

0.124

.002

0.022

Heart rate

0.196

<.001

0.190

LV mass

0.200

<.001

0.035

LVESV

0.265

<.001

0.246

Age

Figure 3 Comparisons of average mitral annular e0 velocity in patients with versus those without diabetes across BMI categories. Increasing BMI was associated with progressive impairment of average mitral annular e0 velocity. Patients with diabetes had more impaired average mitral annular e0 velocity across all BMI categories.

Univariate r*

.55 <.001 .41 <.001

Presence of diabetes

0.274

<.001

0.231

<.001

BMI categories

0.434

<.001

0.379

<.001

*Pearson correlation performed for two continuous variables, Spearman correlation performed for one continuous variable and one ordinal variable (i.e., BMI categories), and point-biserial correlation performed for one continuous variable and one dichotomous variable (i.e. gender and presence of diabetes).

betes had more impaired LV function at all categories of BMI. Multivariate analysis showed that both increasing BMI category and diabetes were independent predictors of impaired LV myocardial systolic and diastolic function despite preserved LVEF. Furthermore, both diabetes and increasing BMI category had an additive detrimental effect on LV myocardial function, and increasing BMI was a stronger determinant of impaired LV myocardial function than diabetes. Pathogenesis of Diabetic and Obesity Cardiomyopathy Various mechanisms underlie the etiology of diabetic cardiomyopathy, including processes such as altered myocardial metabolism with subsequent steatosis and lipotoxicity, endothelial dysfunction

Journal of the American Society of Echocardiography Volume 31 Number 8

Ng et al 923

Table 4 Independent determinants of LV peak systolic global longitudinal strain rate

Table 6 Independent determinants of LV early diastolic global longitudinal strain rate

Peak systolic global longitudinal strain rate Univariable r*

P

Age

0.019

.63

Male

0.118

.003

0.177

0.088

.027

0.014

0.199

<.001

0.202

LV mass

0.145

<.001

0.069

LVESV

0.324

<.001

0.332

<.001

Presence of diabetes

0.179

<.001

0.180

BMI categories

0.352

<.001

0.285

Variable

Systolic blood pressure Heart rate

Standardized b

P

<.001

0.399

<.001

0.068

.08

0.071

.039

Systolic blood pressure

0.298

<.001

0.020

.56

Heart rate

0.128

.001

0.098

.002

LV mass

0.313

<.001

0.043

.28

LVESV

0.095

.015

0.119

.002

<.001

Presence of diabetes

0.257

<.001

0.230

<.001

<.001

BMI categories

0.392

<.001

0.337

<.001

.72 <.001 .12

Standardized b

P

0.591

<.001

0.520

<.001

0.043

.21

Systolic blood pressure

0.367

<.001

0.078

.027

Heart rate

0.165

<.001

0.105

.001

LV mass

0.240

<.001

0.103

.010

0.167

<.001

0.080

.034

LVESV

P

0.400

P

.09

Standardized b

Male

Univariable r*

0.069

P

Age

Average mitral annular e0 velocity

Age

Univariable r*

.13

Table 5 Independent determinants of average mitral annular e0 velocity

Male

Variable

<.001

0.060

*Pearson correlation performed for two continuous variables, Spearman correlation performed for one continuous variable and one ordinal variable (i.e., BMI categories), and point-biserial correlation performed for one continuous variable and one dichotomous variable (i.e., gender and presence of diabetes).

Variable

Early diastolic global longitudinal strain rate

Presence of diabetes

0.175

<.001

0.146

<.001

BMI categories

0.238

<.001

0.157

<.001

*Pearson correlation performed for two continuous variables, Spearman correlation performed for one continuous variable and one ordinal variable (i.e., BMI categories), and point-biserial correlation performed for one continuous variable and one dichotomous variable (i.e., gender and presence of diabetes).

with microvascular disease, autonomic neuropathy, altered myocardial structure with fibrosis, and atherosclerosis.21-26 Often, these processes act together and result in myocardial hypertrophy, increased interstitial fibrosis with increased LV stiffness, and manifesting as diastolic dysfunction in early diabetic heart disease.5 Over time, there is progressive loss of myocardial contractile function, resulting in global LV systolic dysfunction. Recent large epidemiologic studies had unequivocally demonstrated increased all-cause mortality in patients with higher levels of BMI.14,15 Furthermore, increasingly obese patients had progressively higher cardiovascular mortality.15 Compared with diabetic cardiomyopathy, obesity cardiomyopathy is a clinically less well-recognized phenomenon. Obesity cardiomyopathy is characterized by a variety of cardiac structural and hemodynamic changes.6 Previous studies have shown that obese patients have larger LV chamber sizes and increased hypertrophy compared with their lean counterparts.27-30 These cardiac structural changes are associated with changes in

*Pearson correlation performed for two continuous variables, Spearman correlation performed for one continuous variable and one ordinal variable (i.e., BMI categories), and point-biserial correlation performed for one continuous variable and one dichotomous variable (i.e., gender and presence of diabetes).

cardiovascular hemodynamics, including increased total blood volume and cardiac output and reduced systemic vascular resistance.6,31 Although it has been suggested that the association between obesity and incident heart failure may be secondary to these hemodynamic and cardiac structural changes,6 recent evidence suggests that the relationship may be mediated by obesity-related metabolic, inflammatory, and neurohormonal changes.2,7,8 The presence of obesity-related insulin resistance, systemic vascular and adipose tissue inflammation, increased free fatty acid delivery, and activation of the renin-angiotensin-aldosterone system all contribute to atherosclerosis, myocardial hypertrophy, increased interstitial fibrosis, and subsequent myocardial dysfunction.2,7,8,28 Although the underlying pathogenesis of diabetic and obesity cardiomyopathy may not be identical, the final common pathway appears to be an increased interstitial fibrosis followed by initial myocardial diastolic dysfunction and eventually systolic dysfunction. However, few studies to date have evaluated the combined detrimental effects of both diabetes and obesity on LV function. In the present study, both diabetes and increasing BMI category were associated with LV myocardial systolic and diastolic dysfunction. Figure 4 demonstrated that patients without diabetes had a greater initial decline in early diastolic global longitudinal strain rate from normal weight to overweight compared with those with diabetes. This suggests that early diastolic global longitudinal strain rate is most sensitive to the detrimental LV myocardial functional changes secondary to diabetes or increasing BMI and is in agreement with the pathophysiological process of diastolic dysfunction occurring before systolic dysfunction. Diabetes, Obesity, and Myocardial Dysfunction Numerous echocardiographic studies examining diabetes have demonstrated the presence of both LV diastolic and systolic dysfunction compared with normal control subjects.32-34 Similarly, studies on obese subjects without diabetes have also demonstrated LV diastolic and systolic dysfunction compared with lean control subjects.28,29,35 Although diabetes and obesity frequently coexist,36 few studies to date have evaluated the combined impact of both diabetes and obesity on LV function.37,38 Kuperstein et al.37 recruited a large

924 Ng et al

number of predominately obese subjects and demonstrated a synergistic effect of obesity and diabetes on LV mass. However, the study was confounded by the small number of patients with diabetes (6% of the total study population) and the fact that all of the obese patients with diabetes were women.37 Furthermore, the combined impact of diabetes and obesity on LV function in their study was unknown. Di Stante et al.38 demonstrated LV diastolic dysfunction in 40 obese patients with diabetes compared with 93 obese patients without diabetes. However, the impact of increasing obesity on LV dysfunction was not evaluated, and multivariate analysis identifying independent determinants of LV dysfunction was not performed. In contrast, the present study is the largest to date evaluating the combined impact of diabetes and increasing BMI category on LV myocardial function. Both diabetes and increasing BMI were independent determinants of LV myocardial dysfunction. Furthermore, both increasing BMI and obesity had an additive detrimental effect on LV myocardial function, and increasing BMI was associated with a greater impairment of LV myocardial function than diabetes. This was shown by the standardized b coefficients demonstrating that increasing BMI had a greater relative contribution to the multiple linear regression models than the presence of diabetes, despite similar P values. Importantly, the present study excluded all patients with known or suspected significant underlying coronary artery disease. Therefore, it was unlikely that the presence of undiagnosed significant coronary artery disease could have significantly influenced the results. Similarly, although increasing BMI and diabetes were associated with a higher incidence of hypertension, the independent associations between diabetes, obesity, and myocardial dysfunction were still significant despite adjusting for differences in systolic blood pressure on multivariate analysis. Similar results were also obtained when systolic blood pressure was substituted by diastolic blood pressure or history of hypertension or when all patients with hypertension were excluded from analysis. Study Limitations Because the present study was a cross-sectional analysis of the effects of diabetes and BMI on myocardial function, there were no longitudinal follow-up clinical outcome data. LV volumes and mass were not indexed to body surface area (BSA), because the most frequently used formulas for BSA calculation only approximate, and no true gold standard exists. In the present study population, the Mosteller and Du Bois formula to calculate BSA showed excellent absolute agreement (intraclass correlation coefficient = 0.99). However on the basis of a Bland-Altman plot, there was an increasing bias at extremes of BSA. Therefore, the unindexed values of LV volumes and mass were provided. Similarly, there was no information on the influences of weight loss and intensity of diabetic control on myocardial function. Finally, there was a potential enrollment or selection bias, as clinical patients were referred for clinical echocardiography, thereby possibly increasing the incidence of subclinical myocardial dysfunction despite normal LVEF. Clinical Implications Both obesity and diabetes are worldwide growing epidemics. Previous epidemiologic studies demonstrated that both diabetes and obesity were independent risk factors for the development of heart failure.9,13,39,40 These studies have shown that diabetes and obesity were each associated with an approximately twofold increased risk for heart failure.9,13,40 However, these studies

Journal of the American Society of Echocardiography August 2018

generally examined the prognostic impact of diabetes and obesity separately. The present study demonstrated that diabetes and obesity were independently associated with an additive detrimental effect on LV myocardial function, and increasing obesity was associated with greater LV myocardial dysfunction than diabetes. Thus, future clinical studies of patients with diabetes should also focus on the frequently associated obesity problem in this patient population.

CONCLUSION Both diabetes and obesity were independently associated with an additive detrimental effect on LV myocardial function. Furthermore, increasing BMI category was associated with greater LV myocardial dysfunction than diabetes. Therapies aiming at reducing cardiac morbidity and mortality in diabetic patients should also focus on weight reduction.

REFERENCES 1. Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes. Diabetes Care 2004;27:1047-53. 2. Horwich TB, Fonarow GC. Glucose, obesity, metabolic syndrome, and diabetes: relevance to incidence of heart failure. J Am Coll Cardiol 2010;55:283-93. 3. Aneja A, Tang WH, Bansilal S, Garcia MJ, Farkouh ME. Diabetic cardiomyopathy: insights into pathogenesis, diagnostic challenges, and therapeutic options. Am J Med 2008;121:748-57. 4. Braga MFB, Leiter LA. Role of renin-angiotensin system blockade in patients with diabetes mellitus. Am J Cardiol 2009;104:835-9. 5. Maya L, Villarreal FJ. Diagnostic approaches for diabetic cardiomyopathy and myocardial fibrosis. J Mol Cell Cardiol 2010;48:524-9. 6. Alpert MA. Obesity cardiomyopathy: pathophysiology and evolution of the clinical syndrome. Am J Med Sci 2001;321:225-36. 7. Gustafson B, Hammarstedt A, Andersson CX, Smith U. Inflamed adipose tissue: a culprit underlying the metabolic syndrome and atherosclerosis. Arterioscler Thromb Vasc Biol 2007;27:2276-83. 8. Andersson CX, Gustafson B, Hammarstedt A, Hedjazifar S, Smith U. Inflamed adipose tissue, insulin resistance and vascular injury. Diabetes Metab Res Rev 2008;24:595-603. 9. Kenchaiah S, Evans JC, Levy D, Wilson PW, Benjamin EJ, Larson MG, et al. Obesity and the risk of heart failure. N Engl J Med 2002;347: 305-13. 10. Hu G, Jousilahti P, Antikainen R, Katzmarzyk PT, Tuomilehto J. Joint effects of physical activity, body mass index, waist circumference, and waist-to-hip ratio on the risk of heart failure. Circulation 2010; 121:237-44. 11. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998;15:539-53. 12. World Health Organization. Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation. WHO Technical Report Series 894. Geneva, Switzerland: World Health Organization; 2000. 13. Pischon T, Boeing H, Hoffmann K, Bergmann M, Schulze MB, Overvad K, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008;359:2105-20. 14. Berrington dG, Hartge P, Cerhan JR, Flint AJ, Hannan L, MacInnis RJ, et al. Body-mass index and mortality among 1.46 million white adults. N Engl J Med 2010;363:2211-9.

Journal of the American Society of Echocardiography Volume 31 Number 8

15. Zheng W, McLerran DF, Rolland B, Zhang X, Inoue M, Matsuo K, et al. Association between body-mass index and risk of death in more than 1 million Asians. N Engl J Med 2011;364:719-29. 16. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis 2002;39:S1-266. 17. Nishimura R, Miller FJ, Callahan M, Benassi R, Seward J, Tajik A. Doppler echocardiography: theory, instrumentation technique and application. Mayo Clin Proc 1985;60:321-43. 18. Tajik A, Seward J, Hagler D, Mair D, Lie J. Two dimensional real-time ultrasonic imaging of the heart and great vessels: technique, image orientation, structure identification and validation. Mayo Clin Proc 1978;53:271-303. 19. Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2015;28:1-39. 20. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF III, Dokainish H, Edvardsen T, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2016;29:277-314. 21. McGavock JM, Victor RG, Unger RH, Szczepaniak LS. Adiposity of the heart, revisited. Ann Intern Med 2006;144:517-24. 22. Creager MA, Luscher TF, Cosentino F, Beckman JA. Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part I. Circulation 2003;108:1527-32. 23. Vinik AI, Freeman R, Erbas T. Diabetic autonomic neuropathy. Semin Neurol 2003;23:365-72. 24. Goraya TY, Leibson CL, Palumbo PJ, Weston SA, Killian JM, Pfeifer EA, et al. Coronary atherosclerosis in diabetes mellitus: a population-based autopsy study. J Am Coll Cardiol 2002;40:946-53. 25. Kawaguchi M, Techigawara M, Ishihata T, Asakura T, Saito F, Maehara K, et al. A comparison of ultrastructural changes on endomyocardial biopsy specimens obtained from patients with diabetes mellitus with and without hypertension. Heart Vessels 1997;12:267-74. 26. Marwick T. The diabetic myocardium. Curr Diab Rep 2006;6:36-41. 27. Alpert MA, Terry BE, Mulekar M, Cohen MV, Massey CV, Fan TM, et al. Cardiac morphology and left ventricular function in normotensive morbidly obese patients with and without congestive heart failure, and effect of weight loss. Am J Cardiol 1997;80:736-40.

Ng et al 925

28. Wong CY, O’Moore-Sullivan T, Leano R, Byrne N, Beller E, Marwick TH. Alterations of left ventricular myocardial characteristics associated with obesity. Circulation 2004;110:3081-7. 29. Di Bello V, Santini F, Di Cori A, Pucci A, Palagi C, Delle Donne MG, et al. Obesity cardiomyopathy: is it a reality? An ultrasonic tissue characterization study. J Am Soc Echocardiogr 2006;19:1063-71. 30. Turkbey EB, McClelland RL, Kronmal RA, Burke GL, Bild DE, Tracy RP, et al. The impact of obesity on the left ventricle: the multi-ethnic study of atherosclerosis (MESA). JACC Cardiovasc Imaging 2010;3:266-74. 31. de Divitiis O, Fazio S, Petitto M, Maddalena G, Contaldo F, Mancini M. Obesity and cardiac function. Circulation 1981;64:477-82. 32. Zabalgoitia M, Ismaeil MF, Anderson L, Maklady FA. Prevalence of diastolic dysfunction in normotensive, asymptomatic patients with wellcontrolled type 2 diabetes mellitus. Am J Cardiol 2001;87:320-3. 33. Boyer JK, Thanigaraj S, Schechtman KB, P€erez JE. Prevalence of ventricular diastolic dysfunction in asymptomatic, normotensive patients with diabetes mellitus. Am J Cardiol 2004;93:870-5. 34. Ng ACT, Delgado V, Bertini M, van der Meer RW, Rijzewijk LJ, Shanks M, et al. Findings from left ventricular strain and strain rate imaging in asymptomatic patients with type 2 diabetes mellitus. Am J Cardiol 2009;104: 1398-401. 35. Tumuklu MM, Etikan I, Kisacik B, Kayikcioglu M. Effect of obesity on left ventricular structure and myocardial systolic function: assessment by tissue Doppler imaging and strain/strain rate imaging. Echocardiography 2007;24:802-9. 36. Burke GL, Bertoni AG, Shea S, Tracy R, Watson KE, Blumenthal RS, et al. The impact of obesity on cardiovascular disease risk factors and subclinical vascular disease: the multi-ethnic study of atherosclerosis. Arch Intern Med 2008;168:928-35. 37. Kuperstein R, Hanly P, Niroumand M, Sasson Z. The importance of age and obesity on the relation between diabetes and left ventricular mass. J Am Coll Cardiol 2001;37:1957-62. 38. Di Stante B, Galandauer I, Aronow WS, McClung JA, Alas L, Salabay C, et al. Prevalence of left ventricular diastolic dysfunction in obese persons with and without diabetes mellitus. Am J Cardiol 2005;95:1527-8. 39. Kannel WB, McGee DL. Diabetes and cardiovascular disease. The Framingham study. JAMA 1979;241:2035-8. 40. Coughlin SS, Pearle DL, Baughman KL, Wasserman A, Tefft MC. Diabetes mellitus and risk of idiopathic dilated cardiomyopathy. The Washington, DC dilated cardiomyopathy study. Ann Epidemiol 1994;4:67-74.