This study was approved by the Institutional Review Board of Ilsan-Paik Hospital. The study subjects consisted of 65 type 2 diabetic patients without hypertension or IHD (M : F = 45 : 20; mean age 51 [26-76] years; mean body mass index [BMI] 25.0 ± 2.5 kg/m²) who initially visited the outpatient diabetes clinic at Inje University Ilsan-Paik Hospital between January 2006 and April 2006. Inclusion criteria were: i) normal arterial blood pressure (< 130/85 mm Hg) without antihypertensive medication, ii) no symptoms or signs of heart disease, iii) no history of coronary heart disease (stable angina, unstable angina, myocardial infarction, or revascularization) or valvular heart disease, iv) sinus rhythm and no evidence of IHD on resting 12-lead ECG, v) no evidence of severe medical illness including liver cirrhosis, end-stage renal disease, or cancer. Patients were excluded from participation in the study if either of the following criteria applied: i) diabetes diagnosed before the age of 26, or ii) history of type 1 diabetes or diabetic ketoacidosis.

Height and weight were measured in the morning with the subjects wearing light clothing but no shoes. Blood pressure was measured with a mercury sphygmomanometer on the right arm with the subjects in a sitting position after a 5-minute rest. BMI was calculated as weight in kilograms divided by the square of the height in meters. Diabetes duration and diabetic complications were ascertained from review of medical records. Diabetic complications were recorded as retinopathy, neuropathy (absence of ankle jerks and reduced vibratory sensation, or impairment on nerve conduction studies or autonomic function tests), overt proteinuria (24 hour urine protein ≥ 300 mg/day), or microalbuminuria. Fasting plasma glucose (FPG), total cholesterol, high-density lipoprotein (HDL)-cholesterol, low-density (LDL)-cholesterol, lipoprotein a [Lp(a)], triglycerides, and uric acid were measured with an autoanalyzer (Beckman Coulter, Miami, FL, USA). Level of hemoglobin A₁C (HbA₁C) (high performance liquid chromatography; Tosoh, Tokyo, Japan), serum insulin (two-site chemiluminescent immunometric assay; Roche, Basel, Switzerland), and high-sensitivity C-reactive protein (hs-CRP, enzyme-linked immunosorbent assay; DRG Diagnostics, Marburg, Germany) were also measured.

All echocardiographic examinations were performed on a GE VingMed Vivid 5® Echocardiography System (GE Medical Systems, Milwaukee, WI, USA) with a 2.5 MHz transducer. All smokers were prevented from smoking for at least 30 minutes before the examination. The examination was performed while the patient was in a period of quiet respiration. Echocardiograms were stored digitally and analyzed by one examiner. All recordings were performed at a high sweep speed (100 mm/sec) with simultaneous electrocardiographic (ECG) recording and included complete M-mode, 2-dimensional, and Doppler echocardiographic examinations, with emphasis on evaluation of LV diastolic function [10-12]. For evaluation of diastolic myocardial function, mitral inflow velocities (E- and A-waves, cm/sec), deceleration time (DT), isovolumic relaxation time (IVRT), systolic (S) and diastolic (D) pulmonary venous forward flow, pulmonary venous atrial reversal velocity (Ar), and early (E') and late (A') diastolic velocity of mitral annulus were measured in compliance with the standard protocols [10-12]. Left ventricular systolic function was determined by estimation of left ventricular ejection fraction (LVEF). The normal range of LVEF was 65 ± 10%.

The following criteria were used for the diagnosis of LV diastolic function [15]: impaired relaxation pattern was defined as E/A ratio < 1.0 and DT > 200 ms; pseudonormal pattern as E/A ratio from 1.0 - 2.0 with at least two of the following: S/D ratio < 1, or Ar ≥ 35 cm/sec, or E' < A' and E/E' ratio > 10; and restrictive pattern as E/A ratio > 2.0 and DT < 150 ms. All diastolic dysfunctions in this study were impaired relaxation pattern.

Data are presented as mean ± SEM. Statistical analysis was performed using SPSS for Windows (SPSS Inc., Chicago, IL, USA). Differences in variables between patients with and without LV diastolic dysfunction were analyzed using Mann-Whitney test and Fisher's exact test. Analysis of covariance (ANCOVA) test was used for age and sex-adjusted characteristics according to the presence of diastolic dysfunction (Table 1). ANCOVA test was also used to assess the independent association of duration of diabetes with the presence of diastolic dysfunction, with age, sex, and BMI as covariates (Fig. 1). Logistic regression analysis was used to examine independent determinants for diastolic dysfunction, with age, sex, BMI, and duration of diabetes as covariates. All probability values were two-tailed, and statistical significance was defined as P < 0.05.

The median duration of diabetes for the study subjects was 5 years (range, 0 to 26). Mean HbA₁C was 8.0 ± 2.1%. Fifteen patients (23.1%) showed diastolic dysfunction on Doppler echocardiographic studies. The characteristics of the patients with and without prevalent diastolic dysfunction are compared in Table 1. Patients with LV diastolic dysfunction were older than patients without LV diastolic dysfunction (60.0 ± 2.5 vs. 50.5 ± 1.9 years; P < 0.01). There was no difference in the sex ratio between the two groups. After adjusting for age and sex, BMI was found to be higher (26.6 ± 0.7 vs. 24.6 ± 0.3 kg/m²; P < 0.01) and diabetes duration was found to be longer (9.65 ± 1.48 vs. 4.71 ± 0.78 years; P < 0.01) in patients with LV diastolic dysfunction than in those without LV diastolic dysfunction. There were no differences in smoking, blood pressure, lipid profiles, HbA₁C, fasting plasma glucose, fasting serum insulin, or diabetic microvascular complications between the LV diastolic dysfunction group and the normal LV diastolic function group. There was no difference between the two groups in medications for hyperglycemia or dyslipidemia, including thiazolidinediones, statins, and insulin (data not shown).

No subject had LV systolic dysfunction and no differences in LVEF were found between the LV diastolic dysfunction group and the normal LV diastolic function group (P = 0.18). The E/A ratio was lower in patients with LV diastolic dysfunction (0.70 ± 0.04 vs. 1.07 ± 0.29; P < 0.001). There were no significant differences in other parameters between the LV diastolic dysfunction group and the normal LV diastolic function group (Table 2).

Logistic regression analysis for diastolic dysfunction with age, sex, BMI, and diabetes duration as covariates showed that diabetes duration and BMI were independent determinants (Table 3). When participants were classified into four groups based on the duration of diabetes (≤ 1, > 1 and ≤ 5, > 5 and ≤ 10, > 10 years), the frequency of LV diastolic dysfunction increased with increasing duration of diabetes after adjusting for age, sex, and BMI (P for trend = 0.005; Fig. 1).