© 2013, Wiley Periodicals, Inc. DOI: 10.1111/echo.12389

Echocardiography

Evaluation of Subclinical Left Ventricular Dysfunction in Diabetic Patients: Longitudinal Strain Velocities and Left Ventricular Dyssynchrony by Two-Dimensional Speckle Tracking Echocardiography Study Arezoo Zoroufian, M.D., Tannaz Razmi, M.D., Mohsen Taghavi-Shavazi, M.D., Masoumeh Lotfi-Tokaldany, M.D., M.P.H. and Arash Jalali, P.H.D. Catheterization Laboratory of Tehran Heart Center, Tehran University of Medical Sciences, Tehran, Iran

Background: We evaluated left ventricular (LV) subclinical systolic dysfunction in diabetes mellitus patients using two-dimensional speckle tracking echocardiography (STE) for early detection of changes in LV longitudinal strain (ST) or synchronized contraction. Methods: To determine ST and LV dyssynchrony, 37 normal coronary and normotensive diabetes mellitus patients with LV ejection fraction >50% were enrolled and compared to 39 nondiabetic normal coronary and LV function subjects. The cases underwent standard conventional transthoracic echocardiography and tissue Doppler imaging (TDI) and STE. End-systolic ST and time-to-peak systolic strain (Ts) were measured in 18 LV segments. Results: Conventional parameters were similar between diabetic and nondiabetic subjects. In diabetic patients, significant reduction in global and segmental ST adjusted for age and body mass index, independently correlated with early diastolic velocity at the septal mitral valve annulus by TDI (P = 0.001), ratio of transmitral early and late diastolic velocities (P < 0.001), relative wall thickness (P = 0.014), glycosylated hemoglobin (P < 0.001), and fasting blood sugar (P < 0.001). These correlations were not found in the nondiabetic patients. After adjustment, presence of diabetes mellitus remained an independent correlate of reduced LV global longitudinal ST (R = 0.688, P = 0.003). Delay of Ts between the anteroseptal and posterior walls and all the LV segments was markedly higher in the diabetic group regardless of diastolic dysfunction. Conclusion: In diabetic patients with normal coronary and ejection fraction, segmental and global end-systolic longitudinal ST decreased and differences between Ts among LV segments increased irrespective of diastolic dysfunction at early stage. These results suggest that there might be early detectable changes in systolic function in the natural course of diabetes mellitus by STE study. (Echocardiography 2014;31:456–463) Key words: diabetes mellitus, strain, tissue Doppler imaging Diabetes mellitus (DM) is considered a major contributing factor for heart failure even in the absence of coronary artery disease and obesity,1,2 natural history of type 1 DM showed that diastolic dysfunction appears early and precedes the onset of systolic dysfunction.3 With respect to high mortality and morbidity and health costs, it is too important to detect and follow up myocardial dysfunction in the diabetic patients before clinical evidence of impaired cardiac function appears. As the subendocardial myocardial fibers which are more vulnerable to early manifestation of ischemia and left ventricular (LV) systolic dysfunction stand longitudinally, many projects have been designed to detect minimal changes in Address for correspondence and reprint requests: Mohsen Taghavi-Shavazi, M.D., Tehran Heart Center, North Kargar Street, Tehran, Iran. Fax: 0098-21-88029257; E-mail: [email protected]

456

these fibers’ function. In recent years, several studies have identified tissue Doppler imaging (TDI) as a sensitive technique for the evaluation of early diastolic and longitudinal systolic myocardial dysfunction in patients with type 2 diabetes even in the presence of normal cardiac function in conventional echocardiography.4,5 Since introduction of TDI, several studies have demonstrated that subclinical systolic longitudinal dysfunction can be identified using TDI in patients with DM.6,7 But numerous technical challenges such as angle dependency have limited TDI acquisition and data reproducibility of this technique. Newly introduced speckle tracking method is a type of strain imaging and has advantage of calculating myocardial strain independent of angle of incidence. Therefore, speckle tracking echocardiography (STE) may be potentially a valuable tool for screening diabetic

Early Left Ventricular Dysfunction in Diabetic Patients

patients susceptible for developing myocardial dysfunction if found to be reliable. To our knowledge, few studies have assessed parameters of mentioned methods in diabetic patients, therefore we decided to evaluate LV myocardial systolic and diastolic functions in a group of uncomplicated and normotensive DM patients using conventional echocardiographic method and two-dimensional STE (2DSTE). Methods: Study Population: A total of 37 patients (30.8% males) with type 2 DM were recruited in this study between January 2012 and January 2013. The inclusion criteria for all diabetic patients were type 2 DM diagnosed according to the World Health Organization criteria,8 normal coronary arteries proved by coronary angiography and LV ejection fraction (LVEF) >50%, and no regional wall-motion abnormality. Coronary angiography had been performed in catheterization laboratory of Tehran Heart Center, Tehran, Iran, mainly due to the presence of frequent typical or atypical chest pain 1–6 weeks prior to echocardiographic evaluations. The exclusion criteria included history of hypertension with or without antihypertensive drug consumption, absence of stable sinus rhythm, conduction and rhythm disturbances, abnormal serum liver enzyme levels, endocrine and other than DM systemic disease, renal impairment with serum creatinine level ≥1.5 mg/dL, and valvular or congenital heart disease. A total of 39 sexmatched control subjects (45.9% males) with LVEF >50% and without diabetes were included. Control subjects had also normal coronary angiography (performed in the same center) and their exclusion criteria were as same as the patients group. Conventional and Doppler Echocardiography: Standard transthoracic echocardiography with Doppler studies was performed for the subjects at rest using a commercially available EKO 7 diagnostic ultrasound system (Samsung Medison, Seoul, Korea) with 2.2 MHZ multifrequency transducer. All images were digitally stored in 3 consecutive beats on hard disks for offline analysis. A complete 2D, color, pulsed, and continuous-wave Doppler echocardiogram (Samsung Medison) was performed according to echocardiography guidelines.9 The LV dimensions and wall thickness were measured from two-dimensionally targeted M-mode tracing from parasternal longaxis views. Relative wall thickness calculated as the ratio 29 posterior wall thickness in diastole/ LV internal diameter.9 LV volumes and LVEF were calculated by calculating the auto-EF data from stored apical two-, three-, and four-chamber

view. The mitral inflow velocity was 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 the deceleration time, were recorded at the mitral leaflet tips. By TDI, early LV diastolic velocity of the mitral annulus was obtained from the apical four-chamber view with the 1.5 mm sample volume placed in the septal mitral valve annulus at frame rates >100 frames/ sec (septal E’). Diastolic dysfunction was graded as follows: stage 1, abnormal filling pattern is the delayed relaxation that results in a reversed E/A ratio (E/ A < 1), being the earliest stage of heart disease; stage 2, abnormalities in both relaxation and compliance and is known as pseudonormalization because of an apparently normal E/A ratio (E/A > 1); stage 3, restrictive filling pattern found in patients with severely compromised LV compliance and elevated ventricular filling pressures, reflecting an advanced stage of disease.10 To assess LV myocardial performance index (MPI), the LV isovolumic contraction time (IVCT) and the LV isovolumic relaxation time (IVRT) were measured as the interval from the end of the aortic flow to the onset of mitral inflow with use of the pulsed Doppler recordings. Ejection time (ET) was evaluated as the time from beginning to ending of S-wave and MPI (Tei index) calculated by the following formula: (IVCT + IVRT)/ET.11. Two-Dimensional STE: Speckle tracking analyses were performed, using an EKO 7 diagnostic ultrasound system machine with acquiring standard 2D grayscale LV images from apical two-, three-, and four-chamber views with a frame rate at least 30 frame/sec. Speckle tracking analysis began by manually tracing the endocardial border at end-systolic frame in counterclockwise direction starting from the righthand mitral annulus in each 3 apical views by a point-and-click approach12; thereafter, the machine displayed the epicardial line 10 mm outside the created endocardial outline that manually adjusted and rechecked approximate to epicardium. The software then automatically divided the LV into 6 equidistant segments (apical, mid, and basal) and accepted segments of good tracking quality and rejected poorly tracked segments, allowing the observer to manually override its decisions according to visual assessments of the tracking quality. Peak systolic strain, peak systolic strain rate, and time-to-peak systolic strain were calculated for each of the 18 segments. The global longitudinal strain/strain rate was calculated from the average of 18 LV segments values. Longitudinal dyssynchrony were 457

Zoroufian, et al.

defined as maximum time-to-peak strain delay among opposing wall segments using a predefined cutoff ≥130 ms as significant dyssynchrony with standard deviation (SD) cutoff ≥76 ms.12,13 Statistical Analysis: The results are presented as mean  SD for the numeric variables and are summarized by absolute frequencies and percentages for the categorical variables. All the statistical analyses were carried out using SPSS software (version 20, IBM Corp., Armonk, NY, USA). Differences in the continuous variables between both groups were evaluated using the independent sample t-tests or Mann–Whitney U-test if needed. The categorical variables were compared using the chisquare test or the Fisher’s exact test whenever appropriate. The Pearson or Spearman correlation coefficients were used to evaluate the linear relationship between the independent and dependent variables. The multiple linear regression with inter-method was used to adjust the correlation between the parameters. A P < 0.05 constituted statistical significance. Results: The demographic and clinical characteristics of the study patients are summarized in Table I. The median time duration for the diagnosis of DM was 76 months (IQR: 36 and 136 months).The diabetic patients were older and had higher mean diastolic blood pressures (but not more than 85 mmHg) and higher rates of hypercholesterolemia than the controls. The conventional echocar-

diographic characteristics of the 2 groups are listed in Table II. The parameters accessed via conventional echocardiography were similar in the diabetic and nondiabetic subjects. However, in Doppler and TDI records, the diabetic subjects showed a remarkably lower peak A and E/A ratio and higher IVCT, IVRT, E/E’ ratio, and declaration time. The diabetic subjects had a greater frequency of grade 1 diastolic dysfunction than the controls (53.85% vs. 13.8%, respectively) (Table II). The results of speckle tracking longitudinal strain imaging for each individual LV segment are depicted in Tables III and IV. The diabetic patients showed significantly reduced strain and strain rate in most of the LV segments; in some of them, however, the reduction was not significant (Table III). The lowest value of strain belonged to the anterior wall in the diabetics and to the lateral wall in the nondiabetics. Overall mean strain in both groups was at its lowest in the mid-level of the LV. Among the diabetic patients, reduction in global peak systolic strain, adjusted for age and the body mass index, was independently correlated with decreased septal E’ (P = 0.001), E/A ratio (P < 0.001), relative wall thickness (P = 0.014), glycosylated hemoglobin (P < 0.001), and fasting blood sugar (P < 0.001). These correlations were not found in the nondiabetic patients. A cut-off value of 14.91% was calculated using the value of strain in the control subjects via the following formula: mean 2 SD. By this cut-off value, 53.8% (21/39) of the diabetics and none of the controls had abnormal strain.

TABLE I Demographic Characteristics of Study Patients

Mean age (year) Male gender, n (%) Body mass index Heart rate (beats/min) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Fasting blood sugar (mg/dL) Hemoglobin A1C (glycosylated hemoglobin) Duration of diabetes (year) Cigarette smoking, n (%) Hypertriglycemia, n (%) Hypercholesterolemia, n (%) Drugs, n (%) Aspirin b‐blockers Oral antihyperglycemic agent Insulin Antihyperlipidemic agent

458

Diabetic (n = 39)

Nondiabetic (n = 37)

55.79  8.48 12 (30.8) 29.57  4.71 74.59  11.68 116.03  7.96 74.62  6.43 163.62  63.94 7.93  1.62 7.69  5.62 1 (2.6) 3 (7.7%) 13 (33.3)

51.51  8.14 17 (45.9) 26.82  3.31 70.43  10.04 116.62  6.24 77.84  2.77 95.89  8.39 – – 2 (5.4) 1 (2.7) 2 (5.4)

21 (53.8) 6 (15.4) 33 (84.6) 7 (17.9) 13 (33.3)

6 (16.2) 0 0 0 2 (5.4)

P-Value 0.028 0.173 0.005 0.100 0.880 0.035