Diastolic dysfunction is considered the first marker of diabetic cardiomyopathy. However, preclinical systolic alteration was also recently described by strain, but its association with diastolic dysfunction has never been investigated.
One hundred fourteen patients with type 2 diabetes mellitus (DM) with controlled blood pressure and without overt heart disease were prospectively enrolled and compared with 88 age-matched controls. All subjects underwent comprehensive echocardiography, including diastolic evaluation according to current recommendations and speckle-tracking imaging. The prevalence of diastolic dysfunction, the determinants of diastolic parameters, and the association between preclinical systolic and diastolic dysfunctions were studied.
Diastolic parameters were altered in patients compared with controls, with lower E/A ratios, longer mitral deceleration and isovolumic relaxation times, and higher E/e′ ratio. Diastolic dysfunction occurred in 47% of patients with DM (33% and 14% with grade I and II diastolic dysfunction, respectively) and systolic alteration (longitudinal strain ≥ −18%) in 32% of patients. Whereas longitudinal systolic strain was independently associated with DM and gender, diastolic parameters were influenced by many factors, including age, rate-pressure product, history of hypertension, and body mass index. Systolic alteration occurred in 28% of patients with DM with normal diastolic function and in 35% with diastolic dysfunction.
Diastolic dysfunction diagnosed according to current recommendations is frequent in patients with DM but is also influenced by other factors. Systolic strain alteration may exist despite normal diastolic function, indicating that diastolic dysfunction should not be considered the first marker of a preclinical form of diabetic cardiomyopathy.
Patients with type 2 diabetes mellitus (DM) may develop cardiomyopathy independently of coronary artery disease and associated hypertension. This “diabetic cardiomyopathy” is partly responsible for the high incidence of heart failure and subsequent mortality. However, this concept of diabetic cardiomyopathy on the basis of experimental observations needs to be better defined in the clinical setting.
Several studies in patients with DM have identified diastolic dysfunction as the earliest functional alteration in the course of diabetic cardiomyopathy. The prevalence of diastolic dysfunction has been reported in 23% to 75% of patients with DM, with most studies using single parameters such as E/A or E/e′ ratio or combined indices derived from mitral inflow pattern and pulmonary venous flow. However, the existence of diastolic dysfunction has not yet been assessed using the new American Society of Echocardiography and European Association of Echocardiography recommendations. Recently, diastolic dysfunction has been established as an important prognostic parameter in a large cohort of patients with DM. However, diastolic function is also influenced by many other factors, such as age, hypertension, and left ventricular (LV) hypertrophy. Therefore, this raises the question of diastolic dysfunction as the hallmark of diabetic cardiomyopathy.
Recent studies have shown that preclinical systolic dysfunction assessed by systolic strain may also occur in patients with DM on normal conventional echocardiography. Moreover, our group recently demonstrated that in patients with DM without overt heart disease, systolic longitudinal strain alteration is determined only by DM and gender. However, the association between preclinical diastolic and systolic dysfunction has never been investigated.
In the present study, we sought to evaluate more precisely diastolic function in patients with DM by analyzing diastolic parameters as defined by the recent American Society of Echocardiography and European Association of Echocardiography recommendations. Our aims were (1) to assess the prevalence of diastolic dysfunction in patients with DM, (2) to investigate the independent factors of diastolic dysfunction in this population, and (3) to establish the potential association between diastolic dysfunction and systolic strain alterations.
A total of 114 consecutive patients with type 2 DM referred to our institution (Louis Pradel Hospital, Lyon, France) between February 2006 and March 2008 were prospectively recruited and described in a previous report regarding radial and longitudinal function. In the present study, we specifically focused on diastolic function and its relation with systolic function in the same patients and controls.
The recruitment of patients with DM was based on referral to the outpatient clinical department of diabetology of our institution. Patients were included if they met the following inclusion criteria: age between 35 and 60 years, no symptoms or history of heart disease, LV ejection fraction (LVEF) > 55%, absence of regional LV wall motion abnormalities assessed by echocardiography, and no myocardial ischemia, defined by normal results on exercise electrocardiography (54 patients) or dobutamine stress echocardiography (60 patients) in addition to myocardial perfusion (17 patients) within the month before inclusion. Exclusion criteria were the absence of sinus rhythm, coronary and valvular heart diseases, severe renal failure, echocardiographic images unsuitable for quantification, type 1 DM, severely uncontrolled DM (glycosylated hemoglobin < 12%), and uncontrolled blood pressure at rest (defined as systolic blood pressure > 180 mm Hg and/or diastolic blood pressure > 100 mm Hg). Patients had a mean DM duration of 11 ± 7 years and a mean glycosylated hemoglobin value of 7.7 ± 1.4%.
In addition, 88 age-matched healthy controls, previously enrolled in the Asklepios cohort, were included in this study. An overview of the recruitment, exclusion criteria, screening, and participant examination procedures of the Asklepios study has been provided elsewhere. Briefly, the Asklepios cohort is a cohort of 2,524 community-dwelling male and female volunteers, aged approximately 35 to 55 years at study initiation (October 2002), recruited from the twinned communities of Erpe-Mere and Nieuwerkerken, near the Belgian capital of Brussels. This cohort is representative of the general population free from overt cardiovascular disease. From this database, we selected a group of age-matched subjects with the DM group and without cardiovascular risk factors. Asklepios subjects were eligible if they me the following criteria: never smokers, systolic blood pressure < 140 mm Hg and diastolic blood pressure < 85 mm Hg on triplicate measurement, no drug treatments for hypertension, no diabetes, and normal fasting glycemia (glycemia < 110 mg/dL), triglycerides < 150 mg/dL, total cholesterol < 230 mg/dL, low-density lipoprotein cholesterol < 160 mg/dL, high-density lipoprotein cholesterol > 38.5 mg/dL, and serum creatinine < 1.25 mg/dL.
The study protocol was approved by the local ethics committees of Louis Pradel Hospital and the University of Ghent, and all subjects gave written informed consent.
Transthoracic echocardiography was performed in patients with DM and controls at rest using a commercially available ultrasound system (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway). All echographic acquisitions were digitally stored from at least three consecutive heartbeats for offline analysis (EchoPAC; GE Vingmed Ultrasound AS), which was performed at Louis Pradel Hospital by two investigators (L.E., C.B.).
Conventional measures (LV end-diastolic and end-systolic diameters, LV mass [LVM], LV volumes, LVEF, and stroke volume [SV]) were obtained as previously described. Briefly, LV diameters were measured using M-mode echocardiography according to the recommended criteria, and LV volumes and LVEF were calculated from the apical views using the modified Simpson’s biplane method. LVM was determined by Devereux’s formula and index to body surface area.
In addition, longitudinal and radial systolic strains of the left ventricle were assessed by speckle-tracking imaging as previously described. As previously reported, a longitudinal systolic strain value ≥ −18% (mean value in controls − 2 SDs) was indicative of preclinical systolic dysfunction, whereas a radial systolic strain value ≤ 45% was considered to indicate preclinical radial systolic dysfunction.
Diastolic Functional Analysis
Diastolic functional parameters in patients and controls were assessed according to the current recommendations of the American Society of Echocardiography and the European Association of Echocardiography.
Diastolic Functional Parameters
Using the apical four-chamber view, we measured the left atrial volume by the area-length method (indexed to body surface area), mitral inflow velocities by pulsed Doppler (early diastolic velocity [E], peak and duration of late diastolic velocity [A], and E-wave deceleration time [mDT]), mitral annular diastolic velocity (e′ and the E/e′ ratio) by pulsed-wave Doppler tissue imaging, and pulmonary venous flow (Ap duration, S/D ratio, and Ap velocity) and isovolumic relaxation time (IVRT) by pulsed Doppler. The E/e′ ratio divided by SV was used as a surrogate of LV diastolic elastance (stiffness). The time interval between the peak of the R wave and the onset of mitral E velocity and between the peak of the R wave and the onset of e′ was measured, and the difference between these time intervals ( T E−e′ ) was calculated to obtain the IVRT/ T E−e′ ratio.
Longitudinal diastolic strain rate (SR) analysis was performed using speckle-tracking imaging (EchoPAC) in the apical four-chamber view (frame rate, 75 images/sec). Peak SRs during early (SR E ) and late (SR A ) diastole from the six segments of the inferoseptal and anterolateral walls were measured, and the mean values were calculated. The E/SR E ratio was then calculated.
Diastolic Dysfunction Grading
Using the current guidelines, we assigned patients with DM to one of the four following groups: normal diastolic function or grade I, grade II, or grade III diastolic dysfunction. The first step of this evaluation consisted of the classification of patients according to septal and lateral e′ velocities and left atrial volume index. Patients with septal e′ ≥ 8 cm/sec and lateral e′ ≥ 10 cm/sec and left atrial volumes < 34 mL/m 2 constituted the group with normal diastolic function. For the remaining patients, diastolic dysfunction was graded according to the following criteria: E/A ratio, mDT, average E/e′ ratio, and Ar-A time interval ( Figure 1 ).
Fasting blood lipid profile, glycemia, and creatinine were measured using standard procedures. In addition, glycosylated hemoglobin and microalbuminuria were also measured in patients.
Statistical analysis was performed using SPSS version 17.0 (SPSS, Inc., Chicago, IL). Normally distributed data are expressed as mean ± SD. Data deviating from normality are expressed as median (interquartile range), and categorical variables are expressed as percentages. Baseline data in the DM group and control group were compared using unpaired t tests for continuous variables and χ 2 tests for categorical variables. Then, comparison between patients with DM with and without diastolic dysfunction was also tested using the same procedures.
Multivariate linear regressions were performed in patients with DM and controls to assess independent variables associated with each diastolic functional parameter as follows: E/A ratio, mDT, IVRT, e′, E/e′ ratio, E/e′ ratio/SV, and E/SR E . The following variables were tested: age, sex, rate-pressure product (systolic blood pressure × heart rate), DM, smoking, history of hypertension, dyslipidemia, body mass index (BMI), triglycerides, low-density lipoprotein cholesterol, creatinine, LVEF, LVM index, and wall stress. All variables significantly associated with each diastolic functional parameter on univariate analysis were entered in each corresponding general linear model. We used a backward elimination procedure with a cutoff of P < .05. In addition, using the same method in patients with DM, we analyzed the association between the same diastolic parameters and the following variables: DM duration, diabetic complications (retinopathy, neuropathy, microalbuminuria), DM imbalance (glycosylated hemoglobin), and medications (insulin, metformin, sulfonylurea, glitazones, angiotensin-converting enzyme inhibitors or angiotensin II antagonists, calcium channel blockers, β-blockers, diuretics, and statins).
Then, the prevalence of preclinical radial and longitudinal systolic dysfunction according to diastolic functional grade was calculated. P values < .05 were considered statistically significant.
Table 1 displays the characteristics of the study population. Patients with DM had normal but higher blood pressure than controls and presented higher heart rates and rate-pressure products. In addition, patients with DM had normal conventional parameters of systolic function including LVEF, and a normal LVM index, compared with controls. However, on the basis of systolic strains, the prevalence of radial and longitudinal systolic dysfunction among patients with DM as 38% (43 of 114 patients) and 32% (36 of 114 patients), respectively.
|Variable||Controls ( n = 88)||Patients with DM ( n = 114)||P|
|Age (y)||51.7 ± 2.6||52.0 ± 4.5||.59|
|BMI (kg/m 2 )||24 ± 3||29 ± 5||.0001|
|SBP (mm Hg)||120 ± 9||128 ± 14||.0001|
|DBP (mm Hg)||74 ± 6||77 ± 10||.02|
|HR (beats/min)||63 ± 8||75 ± 12||.001|
|RPP (mm Hg/min)||7,539 ± 1,121||9,620 ± 1,828||<.0001|
|Calcium blockers||0||21 (18%)||—|
|Antiplatelet agents||0||15 (13%)||—|
|LVEDD (mm)||48 ± 4||49 ± 5||.32|
|LVESD (mm)||29 ± 4||29 ± 4||.39|
|LVM index (g/m 2 )||80 ± 19||84 ± 18||.12|
|LVEDV index (mL/m 2 )||43 ± 1||45 ± 1||.77|
|LV systolic function|
|LVEF (%)||68 ± 7||67 ± 6||.12|
|Longitudinal systolic strain (%)||−22 ± 2||−19 ± 3||<.001|
|Radial systolic strain (%)||56 ± 12||50 ± 16||.003|
|LA volume index (mL/m 2 )||22 ± 6||24 ± 6||.16|
|LA area (cm 2 )||14 ± 3||17 ± 3||<.0001|
|E/A ratio||1.2 ± 0.2||1.1 ± 0.2||<.001|
|mDT (msec)||180 ± 27||225 ± 52||<.0001|
|IVRT (msec)||74 ± 10.5||83 ± 11.0||<.0001|
|e′ (cm/sec)||10.3 ± 2.3||8.3 ± 2.4||<.0001|
|E/e′ ratio||7.7 ± 1.7||10.9 ± 3.6||<.0001|
|E/e′ ratio/SV||0.102 ± 0.03||0.143 ± 0.056||<.0001|
|T E−e′ (msec)||15.1 ± 12.7||15.9 ± 17.0||.07|
|IVRT/ T E−e′||8.0 ± 7.0||10.2 ± 1.2||.11|
|SR E (s −1 )||−1.5 ± 0.2||−1.5 ± 0.3||.07|
|SR A (s −1 )||−0.9 ± 0.2||−1.2 ± 0.3||<.0001|
|E/SR E ratio||524 ± 97||602 ± 170||.002|
Prevalence of Diastolic Dysfunction
Diastolic parameters were altered in patients compared with controls, with lower E/A ratios, longer mDTs and IVRTs, lower e′, and higher E/e′ ratios ( Table 1 ). In addition, the surrogate measure of LV diastolic elastance (E/e′ ratio/SV) was higher in patients than in controls.
According to the current recommendations, 60 patients (53%) presented normal diastolic function and 54 (47%) diastolic dysfunction. Most patients with diastolic dysfunction were grade I (33%), 14% of patients were grade II, and none were grade III ( Figure 1 ). Compared with patients with normal diastolic function, patients with diastolic dysfunction were older, had higher blood pressures and rate-pressure products, and had higher indexed LV end-diastolic volumes ( Table 2 ).
|Characteristic||Normal diastolic function ( n = 60)||Diastolic dysfunction ( n = 54)||P|
|Age (y)||51.0 ± 4.6||53.2 ± 4.2||.009|
|SBP (mm Hg)||123 ± 13||134 ± 14||<.001|
|DBP (mm Hg)||75 ± 11||76 ± 11||.04|
|HR (beats/min)||75 ± 12||76 ± 11||.67|
|RPP (mm Hg/min)||9,193 ± 1,847||10,088 ± 1,700||.009|
|LVEDD (mm)||48.5 ± 4.6||49.3 ± 4.5||.38|
|LVESD (mm)||28.4 ± 4.3||30.0 ± 4.2||.06|
|LVM index (g/m 2 )||83.7 ± 1.9||86.1 ± 1.8||.49|
|LVEDV index (mL/m 2 )||42.8 ± 7.9||47.4 ± 1.1||.01|
|LV systolic function|
|LVEF (%)||67 ± 6||66 ± 5||.37|
|Longitudinal systolic strain (%)||−20.0 ± 2.7||−18.8 ± 2.9||.03|
|Radial systolic strain (%)||52 ± 14||47 ± 17||.07|
|LA volume index (mL/m 2 )||43 ± 12||48 ± 15||.09|
|LA area||17 ± 5||17 ± 3||.7|
|E/A ratio||1.2 ± 2.4||1.0 ± 2.2||<.001|
|mDT (msec)||219 ± 48||231 ± 55||.20|
|IVRT (msec)||81 ± 10||85 ± 12||.1|
|e′ mean (cm/sec)||10.6 ± 1.9||7.5 ± 1.4||<.001|
|E/e′ ratio||9.0 ± 2.7||12.9 ± 3.3||<.001|
|E/e′ ratio/SV||0.115 ± 0.039||0.171 ± 0.057||<.001|
|T E−e′ (msec)||17.6 ± 18.3||14.1 ± 15.5||.29|
|IVRT/ T E−e′||8.5 ± 7.8||12.1 ± 1.5||.14|
|SR E (s −1 )||1.6 ± 0.3||1.3 ± 0.3||<.001|
|SR A (s −1 )||1.2 ± 0.3||1.3 ± 0.3||.13|
|E/SR E ratio||55.5 ± 14.2||64.7 ± 18.3||.01|
Independent Factors Associated with Diastolic Functional Parameters
Multivariate analysis was performed to determine independent factors associated with diastolic functional parameters ( Table 3 and Supplementary Table 1 ). By univariate analysis, most diastolic functional parameters were associated with DM. However, by multivariate analysis, none of the diastolic parameters except IVRT was independently associated with only DM. E/e′ ratio, mDT, and E/e′ ratio/SV were determined by cofactors other than DM, including age, rate-pressure product, smoking, and history of hypertension. Finally, the following diastolic parameters were not significantly associated with DM: E/A ratio (associated only with age and rate-pressure product), e′ velocity (associated only with age and rate-pressure product), E/SR E ratio (associated only with history of hypertension and BMI), and E/e′ ratio/ SV (associated only with age, heart rate, and history of hypertension).
|Diastolic parameter||E/A ratio||E/e′ ratio||e′||mDT||IVRT||E/SR E ratio||E/e′ ratio/SV|
|Diabetes (yes vs no)||1.8||.001||39||<.0001||0.36||<.001||0.02||.03|
|Smokers (yes vs no)||−1.8||.001|
|Hypertension (yes vs no)||1.8||.001||8.5||.004||0.03||.003|