Impaired Left Atrial Conduit Function in Coronary Artery Disease Patients With Poorly Controlled Diabetes: Two-Dimensional Speckle-Tracking Echocardiographic Study

Authors


Abstract

Objectives

The myocardium can be affected by diabetes mellitus. The effects of blood glucose control on some organs such as the kidney and eye have been previously reported. The aim of our study was to evaluate left atrial function via 2-dimensional (2D) speckle-tracking echocardiography in a group of coronary artery disease (CAD) patients with well-controlled diabetes (hemoglobin A1c [HbA1c] < 7%) and to compare it with that in a group of CAD patients with poorly controlled diabetes.

Methods

This cross-sectional study included 110 CAD patients, comprising 44 euglycemic control patients, 33 patients with well-controlled diabetes (HbA1c < 7%), and 33 patients with poorly controlled diabetes. The study population thereafter underwent 2D speckle-tracking echocardiography for an evaluation of their left atrial function.

Results

Our findings showed that the absolute values of early diastolic strain and early diastolic strain rate were lower in the CAD patients with poorly controlled diabetes than in the euglycemic control patients with CAD. Moreover, early diastolic strain in the CAD patients with poorly controlled diabetes was lower than that in the CAD patients with well-controlled diabetes. Multivariable analysis revealed that poorly controlled diabetes was an independent determinant of early diastolic strain and strain rate.

Conclusions

The conduit function of the left atrium was impaired in the CAD patients with poorly controlled diabetes compared with that in the euglycemic control patients with CAD and the CAD patients with well-controlled diabetes.

Abbreviations
CAD

coronary artery disease

HbA1c

hemoglobin A1c

LA

left atrial

LAD

left anterior descending artery

LAV

left arterial volume

LV

left ventrical

PCBS

poorly controlled blood sugar

STE

speckle-tracking echocardiography

WCBS

well-controlled blood sugar

2D

2-dimensional

Estimations show that diabetes mellitus had affected 6.4% of the global adult population in 2010 and that the figure may rise to 7.7% by 2030.[1] Many body systems such as the cardiovascular system are affected by diabetes mellitus, with cardiovascular disease constituting the leading cause of mortality among diabetic patients.[2] For example, the risk of coronary heart disease mortality is 2 to 4 fold in diabetic compared with non-diabetic subjects.[3] Also, the risk of developing heart failure is increased in patients with diabetes mellitus.[4] Left atrial (LA) enlargement on echocardiography has been previously reported in diabetic patients.[5] Diabetes mellitus is a risk factor for the occurrence of atrial fibrillation,[6] and hyperglycemia is allied to atrial fibrillation through several mechanisms.[7] Hemoglobin A1c [HbA1c] is referred to as “glycated hemoglobin;” in the non-enzymatic pathway, glucose joins hemoglobin to form HbA1c. HbA1c value can display average blood glucose over the preceding 3 months and is used for an evaluation of glycemic control.[8] It has been demonstrated that there is a 14% increased risk of the occurrence of atrial fibrillation in diabetic patients for every 1% increase in HbA1c.[9] Furthermore, intensive glycemic control (HbA1c < 6% vs HbA1c = 7.0 to 7.9%) cannot reduce the incidence of atrial fibrillation in diabetic patients.[10] The impairment of LA function in the presence of atrial fibrillation rhythm has been revealed.[11-13]

Hyperglycemia is one of the important factors contributing to the induction of systolic and diastolic myocardial dysfunction through metabolic disturbance.[3] The effects of blood glucose control on some organs such as the kidney (microalbuminuria) and eye (visual acuity and cataract extraction) have been previously reported.[14] There is an increased risk of the incidence of heart failure in diabetic patients with follow-up HbA1c > 7% compared with those with follow-up HbA1c < 7% (21% vs 7%).[15]

There are only a few studies on myocardial function as assessed by speckle-tracking echocardiography (STE) in diabetic patients with controlled blood glucose.[16, 17] The assessment of myocardial deformation via 2-dimensional (2D) STE is an objective method for the evaluation of myocardial function. Also, it is an angle-independent technique for the evaluation of tissue deformation in contrast to color-coded tissue Doppler imaging.[18] This echocardiographic method has been commonly used for the assessment of ventricular myocardial function in diabetic patients.[19-22] The effects of diabetes on LA function as assessed by echocardiography have been explored in several studies,[23-27] but the effects of well-controlled blood sugar on LA function in the short term have not been evaluated. Previously, it was demonstrated that coronary artery disease (CAD) affected LA function.[28, 29] Nonetheless, there is a paucity of data on LA function in CAD patients with well- or poorly controlled diabetes. The aim of our study was to evaluate LA function via 2D STE in CAD patients with well-controlled blood sugar (WCBS) (HbA1c < 7%) and compare it with that in CAD patients with poorly controlled blood sugar (PCBS) (HbA1c > 7%).

Materials and Methods

Study Population

Between January and September 2015, 110 consecutive patients admitted to our hospital for coronary artery bypass grafting surgery were included. Recruitment was done first in PCBS diabetic patients between January and March 2015 and then in WCBS diabetic patients and euglycemic control patients between March and September 2015. History taking and physical examinations were done for these patients. Venous blood was drawn for biochemistry studies, including HbA1c and cell blood count after 12 hours of fasting. The inclusion criteria were left ventricular (LV) ejection fraction > 50% and normal sinus rhythm. The exclusion criteria were comprised of history of myocardial infarction; history of cardiac surgery; increased serum creatinine > 1.5 mg/dL; increased alanine and aspartate transferase > twofold the upper limit of the normal range; history of pacemaker implantation; congenital heart diseases; any degree of valvular stenosis; more-than-mild valvular regurgitation; estimated pulmonary artery pressure > 34 mm Hg measured by echocardiography; history of atrial fibrillation, flutter, and tachycardia; left bundle-branch block; LV diastolic dysfunction above grade I according to the recommendations of the American Society of Echocardiography[30]; history of cancer; type 1 diabetes mellitus; history of hyper- and hypothyroidism; HbA1c > 11%; cardiomyopathy; blood pressure > 160/90 mm Hg at the time of echocardiography; poor echocardiography window, and body mass index > 35 kg/m2. Two patients were excluded because of poor echocardiography window. Diabetes was defined as receiving oral antidiabetic agents or insulin or fasting blood sugar > 126 mg/dL or HbA1c > 6.5% in 2 separate examinations. Diabetic patients with admission time HbA1c > 7% were defined as those with PCBS, and diabetic patients with admission time HbA1c < 7% were defined as those with WCBS. Nondiabetic CAD patients with HbA1c < 5.7% were allocated to the euglycemic control group. These definitions were in accordance with the guidelines of the American Diabetes Association.[2] Accordingly, patients with prediabetes (HbA1c = 5.7 to 6.4%) were excluded from our study. In total, we recruited 110 CAD patients, comprising 44 euglycemic control patients, 33 diabetic patients with WCBS, and 33 diabetic patients with PCBS. The biochemistry laboratory staff was blinded to the echocardiography data. Insulin and oral antidiabetic agents were administered to the patients at 6 AM and 6 PM in divided doses, in accordance with our hospital's protocol. All the echocardiographic examinations were done between 8 AM and 4 PM. On the day of echocardiography, no patients received continuous insulin infusions. All the echocardiographic examinations were done up to 48 hours after venous sample taking.

Standard Echocardiography

All the transthoracic echocardiographic examinations were done in the left lateral decubitus position with a commercial echocardiography setting (2–4 MHz probe; GE Medical Systems S5, Wauwatosa, WI) before the cardiac surgeries. One experienced echocardiologist blinded to the biochemistry data performed all the echocardiographic examinations. LV linear internal diameters and wall thickness were measured in the parasternal long-axis view. LV volume in systole and diastole and ejection fraction were measured according to the modified biplane Simpson method (in 2- and 4-chamber apical views). LV mass was calculated according to the recommendations of the American Society of Echocardiography.[31] The peak velocity of mitral flow wave in early (E) and late diastolic phase (A), deceleration time of E wave (DT), and peak velocity of pulmonary vein flow wave in systole (S) and diastole (D) were measured according to the recommendations of the American Society of Echocardiography.[30] In pulsed-wave tissue Doppler imaging, the peak velocities of mitral annulus (lateral and septal) motion in systole (s′), early diastole (e′), and late diastole (a′) were measured according to the recommendations of the American Society of Echocardiography,[30] and the averages of these measurements in the septal and lateral mitral annulus were reported.

Left arterial volumes (LAVs) were measured using another commercial setting (Samsung Medison, EKO 7, Seoul, South Korea). For these measurements, the “Auto EF” option was used. So, a volume-time curve was created with the possibility of obtaining LAVs in a selected time. Next, maximal and minimal LAVs (Max.LAV and Min.LAV) and LAV at the start of P wave (P.LAV) in electrocardiography were measured. All the LAVs were measured in the 2- and 4-chamber apical views. The averages of these measurements in the mentioned views were computed. The volumetric parameters of LA function were evaluated through the calculation of the following indices: filling volume = Max.LAV − Min.LAV; expansion index = 100 × (Max.LAV − Min.LAV)/Min.LAV; diastolic emptying index = 100 × (Max.LAV − Min.LAV)/Max.LAV as the marker of the reservoir function of the LA; passive emptying percent of total emptying = 100 × (Max.LAV − P.LAV)/(Max.LAV − Min.LAV); passive emptying index = 100 × (Max.LAV − P.LAV)/Max.LAV as the marker of the conduit function of the LA; booster active emptying percent of total emptying = 100 × (P.LAV − Min.LAV)/(Max.LAV − Min.LAV); and active emptying index = 100 × (P.LAV − Min.LAV)/P.LAV as the marker of booster pump function.

Two-Dimensional Speckle-Tracking Echocardiography

Two-dimensional STE was done using a commercial echocardiography setting (Samsung Medison, EKO 7). The position of the electrocardiography leads was adjusted to obtain the highest P wave, and then the voltage of electrocardiography was increased to the maximum. LA images from the 2- and 4-chamber apical views at the end of expiration were obtained, and 3 cardiac cycles were stored. Maximal effort was applied to delete the LA appendage and the pulmonary vein orifice. The frame rate of these stored images was 50 to 80 frames per second. The endocardial border of the LA at the end of ventricular systole was traced. Then, the epicardial border of the LA was automatically traced by software. In the next step, the following of the endocardial and epicardial borders of the LA from the traced line was checked. These steps were repeated if the traced line could not be followed. Each LA wall was divided into 3 segments automatically by software. If there was uninterpretable signal in 1 segment after applying multiple efforts, this segment was deleted from the calculation. Consequently, the average measurements of the accepted segments in 3 cardiac cycles were reported. If there were 3 or more deleted segments in 1 patient, the patient was excluded from the study. The zero level was set according to the start of QRS wave, so there were 1 positive systolic peak in LA longitudinal strain curve (SS) and 2 diastolic phases: early diastole (ED) plateau between the end of T wave and initiation of P wave and peak after onset P wave in late diastole (LD), and 1 positive systolic peak (SRs) and 2 negative diastolic peaks in strain rate curve: early (SRe) and late (SRa), see Figure 1. The reservoir function of the LA was evaluated by the peak of systolic strain (SS) and peak systolic strain rate (SRs). The conduit function of the LA was evaluated as the difference between SS and ED in strain curve (EDS) and peak early diastolic strain rate (SRe). The booster pump function of the LA was evaluated as the difference between LD and ED in strain curve (LDS) and late diastolic strain rate (SRa). In sum, 1276 segments (96.7%) were accepted for analysis. Our hospital research committee approved the research proposal, and informed written consent was obtained from all the patients before echocardiography.

Figure 1.

A, Longitudinal strain curve of the left atrium presents peak systolic strain (SS), early diastolic strain (EDS), and late diastolic strain (LDS). B, Longitudinal strain rate curve of the left atrium presents peak systolic strain rate (SRs), early diastolic strain rate (SRe), and late diastolic strain rate (SRa).

Statistical Analysis

The categorical variables are expressed as absolute frequencies and percentages. The continuous variables are expressed as means ± standard deviations if normally distributed; otherwise, they are expressed as medians and interquartile ranges. The continuous variables in the normal control group, WCBS diabetic patients, and PCBS diabetic patients were compared using the one-way analysis of variance if normally distributed; otherwise, they were compared using the non-parametric Kruskal–Wallis H test. When the omnibus test was statistically significant, the Bonferroni adjusted method was employed for pairwise comparisons. The chi-square test or the Fisher exact test was used for the comparison of the categorical variables. The Pearson correlation coefficient and Spearman rank test were employed to evaluate the correlation coefficient between the variables. Multivariable linear regression models were used to determine the association between the controlling blood sugar and 2D STE-derived indices of LA function adjusted for sex, hypertension, duration of diabetes, hematocrit, and e′ (a marker of LV diastolic function) as the potential confounders. Because some variables were not normally distributed, we transformed them logarithmically for normalization. The statistical analyses were conducted using IBM SPSS statistics for Windows (version 21.0) (IBM Corp, Armonk, NY). P values ≤ .05 were considered statistically significant. Eighteen patients were analyzed for the assessment of interobserver and intraobserver variability after 2 months. The average measurements of the accepted segments in 3 cardiac cycles were used in assessment of inter- and intraobserver variability. Interobserver variability for SS, EDS, LDS, SRs, SRe, and SRa was 8.5%, 9.8%, 8.0%, 9.9%, 8.6%, and 8.1%, respectively. Also, intraobserver variability for SS, EDS, LDS, SRs, SRe, and SRa was 9.8%, 8.0%, 7.9%, 6.8%, 8.5%, and 10.0%, respectively.

Results

The study population comprised 33 women and 77 men at a mean age of 61.9 ± 6.9 and 61.4 ± 8.4 years, respectively. The demographic, clinical, and laboratory information of the patients in all 3 groups is depicted in Table 1. The diabetic patients with PCBS had lower hematocrit and hemoglobin levels than did the other groups and had a higher triglyceride level than did the other groups. Male gender ratio was different between the 3 groups. The duration of diabetes was the same in the 2 diabetic groups. The other characteristics were, however, similar between the 3 groups. Glibenclamide and metformin were the only oral antidiabetic agents consumed by the patients. Insulin NPH with and without regular insulin was the only formulation insulin administered to our patients.

Table 1. Demographic Characteristics, Clinical Profile and Laboratory Data in the Diabetic Patients (Poorly Controlled Blood Sugar and Well-controlled Blood Sugar) and the Euglycemic Patients
CharacteristicsEuglycemic Patients (n = 44)Well-controlled Diabetic Patients (n = 33)Poorly Controlled Diabetic patients (n=33)P
  1. a

    PCBS vs GCBS; P = .160.

  2. Data are presented as means ± SD for the normally distributed continuous variables, medians (interquartile boundaries) for the skew-distributed continuous variables, and frequencies (%) for the categorical variables. ACEI/ARB indicates angiotensin-converting enzyme inhibitor/angiotensin receptor blocker; DBP, diastolic blood pressure; FH, family history of coronary artery disease; FBS, fasting blood sugar; HR, heart rate; HDL, high-density lipoprotein; LAD, left anterior descending artery; LCX, left circumflex artery; LDL, low-density lipoprotein; NYHA, New York Heart Association; RCA: right coronary artery; and SBP, systolic blood pressure.

Male gender36 (82%)23 (70%)18 (55%).035
Age, y61.4 ± 7.262.2 ± 7.661.9 ± 7.4.874
SBP, mm Hg125.1 ± 16.2121.5 ± 15.2123.6 ± 14.4.609
DBP, mm Hg78.0 ± 10.074.9 ± 10.473.2 ± 9.8.102
HR, bpm67.3 ± 11.172.7 ± 14.472.8 ± 13.7.107
Body mass index, kg/m226.9 ± 2.827.3 ± 3.127.0 ± 3.8.828
Body surface area, m21.8 ± 0.21.8 ± 0.21.7 ± 0.2.337
NYHA class   .022
I5 (11%)13 (39%)8 (24%) 
II34 (78%)14 (43%)22 (67%) 
III5 (11%)6 (18%)3 (9%) 
Hypertension29 (66%)20 (61%)19 (58%).747
Cigarette smoking19 (43%)8 (24%)4 (12%).009
FH15 (34%)5 (15%)3 (9%).018
ACE or ARB use36 (82%)21 (64%)26 (79%).161
Calcium blocker use7 (16%)9 (27%)8 (24%).452
Beta-blocker use39 (89%)26 (79%)23 (70%).118
Statin use42 (95%)29 (88%)30 (91%).558
Diuretic use3 (7%)6 (18%)5 (15%).295
Nitrate use34 (78%)23 (70%)26 (79%).724
Aspirin use44 (100%)30 (91%)30 (91%).055
Insulin04 (12%)13 (39%)<.001
Oral antidiabetic agents018 (55%)18 (55%)<.001
Number of diseased vessels   .270
One-vessel disease4 (9%)1 (3%)0 (0%) 
Two-vessel disease9 (20%)8 (24%)12 (36%) 
Three-vessel disease31 (71%)24 (73%)21 (64%) 
LAD44 (100%)33 (100%)33 (100%)>.999
LCX38 (86%)28 (85%)27 (82%).860
RCA33 (75%)28 (85%)26 (79%).574
Right dominancy39 (89%)27 (82%)27 (82%).669
Duration of diabetes03.0 (0.2–10.0)5.2 (2.3–15.8)<.001a
Hemoglobin level, mg/dL14.5 ± 1.713.7 ± 1.512.5 ± 1.9<.001
Hematocrit level, %42.1 ± 4.040.8 ± 5.137.1 ± 4.9<.001
FBS level, mg/dL88.0 (78.5–91.8)103.0 (94.0–125.0)133.0 (108.5–169.0)<.001
Serum triglyceride level, mg/dL117.0 (84.5–150.5)144.0 (104.5–167.5)152.0 (109.5–258.5).030
Serum cholesterol level, mg/dL143.8 ± 27.2156.7 ± 46.2151.0 ± 38.8.340
Serum HDL level, mg/dL38.9 ± 8.637.2 ± 9.437.1 ± 10.9.641
Serum LDL level, mg/dL82.0 (62.5–110.0)84.0 (68.5–109.0)84.0 (70.5–103.5).749
Serum urea level, mg/dL37.3 ± 9.137.9 ± 12.039.8 ± 13.2.630
Serum creatinine level, mg/dL0.9 ± 0.20.9 ± 0.20.9 ± 0.3.755

Serum hemoglobin A1c, %

5.5 ± 0.26.3 ± 0.48.4 ± 1.1<.001

Apropos stenosis, the left anterior descending artery (LAD) had intraluminal stenosis >70% in all the patients. The left circumflex artery and the right coronary artery had intraluminal stenosis > 70% in 85% and 79% of the patients, respectively. Intraluminal stenosis between 50 and 70% was seen only in vessels with > 70% stenosis in another site of the same vessels. There were no statistically significant differences in the number of diseased vessels, location of significant stenosis, and dominance of the coronary artery.

Table 2 presents the standard echocardiographic data. The 3 groups were similar in these characteristics except for A wave and a′ wave (the patients with PCBS had a larger A wave and a′ wave than did the other 2 groups). An LA volumetric comparison of these 3 groups showed no differences between the groups. Moreover, the 2D STE-derived indices of LA function showed that EDS, as a marker of the index of LA conduit function, was lower in the PCBS patients than in the other 2 groups (euglycemic vs PCBS; P = .001 and PCBS vs GCBS; P = .014). In the diabetic patients, the correlations between insulin or oral antidiabetic usage and 2D STE-derived indices of LA function were not statistically significant. The absolute value of SRe, as an index of LA conduit function, was lower in the PCBS patients than in the euglycemic control patients (P= 0.026). The multivariate analysis showed that after adjustment for sex, hypertension, duration of diabetes, hematocrit, and e′, PCBS, female sex, and e′ were the independent determinants of log EDS and PCBS, female sex, and hematocrit were the independent determinants of the absolute value of SRe. The duration of diabetes and history of hypertension were not determinants of log EDS and absolute value SRe (Table 3).

Table 2. Echocardiographic Characteristics of the Diabetic Patients (Poorly Controlled Blood Sugar and Well-controlled Blood Sugar) and the Euglycemic Patients
CharacteristicsEuglycemic Patients (n = 44)Well-controlled Diabetic Patients (n = 33)Poorly Controlled Diabetic Patients (n = 33)P
  1. a

    Euglycemic vs PCBS; P = .001, PCBS vs GCBS; P = .014, euglycemic vs GCBS; P > .999.

  2. b

    Euglycemic vs PCBS; P = 0.026, PCBS vs GCBS; P = .282, euglycemic vs GCBS; P > .999.

  3. Data are presented as means ± standard deviations for the normally distributed continuous variables and medians (interquartile boundaries) for the skew-distributed continuous variables.

  4. DT indicates deceleration time; EDS, left atrial early diastolic longitudinal strain; IVST, interventricular septal thickness; LA, left atrial; LDS, left atrial end-diastolic longitudinal strain; LV, left ventricle; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; SPAP, systolic pulmonary artery pressure; SRa, left atrial late diastolic longitudinal strain rate; SRs, left atrial systolic longitudinal strain rate; SRe, left atrial early diastolic longitudinal strain rate; and SS, left atrial systolic longitudinal strain.

LVEDV index, mL/m265.1 ± 11.960.7 ± 14.560.3 ± 8.9.149
LVESV index, mL/m225.2 ± 5.723.9 ± 7.222.7 ± 3.8.169
LVEF,%60.1 ± 7.960.5 ± 7.861.7 ± 8.2.818
Posterior wall, mm9.2 ± 1.19.1 ± 1.19.1 ± 1.4.940
IVST, mm9.2 ± 1.39.3 ± 1.39.1 ± 1.4.809
LV mass index, g/m284.8 ± 18.282.2 ± 22.376.7 ± 24.6.268
E, cm/s61.3 ± 15.863.1 ± 13.464.65 ± 16.2.658
A, cm/s70.1 ± 16.371.0 ± 17.784.6 ± 20.2.001
DT, msec224.2 ± 66.3239.4 ± 54.3247.8 ± 52.7.213
E/A ratio0.8 (0.7–1.0)0.8 (0.7–1.1)0.8 (0.6–0.9).078
S, cm/s50.9 ± 13.354.5 ± 12.256.3 ± 13.6.186
D, cm/s34.6 ± 9.535.7 ± 8.235.7 ± 12.4.869
SPAP, mm Hg25.8 ± 4.725.6 ± 3.926.3 ± 4.7.580
s′, cm/s7.9 ± 1.67.6 ± 1.47.9 ± 1.4.653
e′, cm/s7.8 ± 1.97.0 ± 1.57.4 ± 1.5.149
a′, cm/s9.2 ± 1.89.1 ± 1.810.1 ± 1.3.041
E/e′ ratio8.3 ± 2.79.3 ± 2.39.0 ± 2.8.229
e′/a′ ratio0.9 ± 0.20.8 ± 0.20.7 ± 0.2.053
Maximum LA volume, mL52.4 ± 15.149.7 ± 14.649.9 ± 14.9.655
Minimum LA volume, mL14.5 (11.2–20.3)14.4 (9.9–19.1)14.6 (10.7–18.9).955
Pre A LA volume, mL31.7 ± 9.931.4 ± 10.733.1 ± 10.6.786
Filling volume, mL37.1 ± 11.534.3 ± 9.934.1 ± 9.0.364
Diastolic emptying index72.6 (63.9–77.9)69.8 (64.5–76.0)69.5 (65.0–75.6).503
Expansion index265.5 (176.7–352.1)231.1 (181.4–316.9)227.0 (186.0–309.3).503
Passive emptying percent,%54.4 (49.0–63.4)50.5 (44.2–59.9)47.8 (34.5–62.8).126
Passive emptying index38.3 (32.5–47.6)35.2 (28.0–45.3)30.1 (24.4–43.2).102
Booster active emptying percent total emptying %45.6 (36.6–51.0)49.5 (40.1–55.8)52.2 (37.2–65.5).127
Active emptying index52.6 ± 12.554.2 ± 12.052.1 ± 13.8.828
SS, %39.7 ± 11.641.5 ± 13.037.0 ± 9.9.293
EDS,%23.6 (19.3–26.5)21.0 (16.3–29.0)17.6 (13.8–20.7).001a
LDS,%18.4 (15.3–22.4)20.6 (15.8–26.9)20.7 (17.5–25.9).129
SRs, s−12.4 (2.2–2.8)2.6 (2.2–3.1)2.6 (2.1–2.8).514
SRe, s−1−2.0 ± 0.5−1.9 ± 0.6−1.7 ± 0.5.030b
SRa, s−1−2.6 (−3.1 – −2.2)−2.9 (−3.6 – −2.2)−2.9 (−3.4 – −2.2).553
Table 3. Multivariable Analysis of Left Atrial Conduit Function Indices
VariableLog EDSSRea
βPβP
  1. a

    Absolute value of SRe has been shown. HCT indicates hematocrit; PCBS, poorly controlled blood sugar; and WCBS, well-controlled blood sugar.

PCBS−2.85.005−2.11.029
WCBS−0.13.895−0.45.656
Sex (female)−3.62<.001−2.73.007
Hypertension0.81.4210.83.411
Duration of diabetes0.68.496−0.03.980
e′2.63.0101.97.051
HCT−1.18.242−2.62.010

Discussion

To our knowledge, this is the first study of its kind to compare LA function between CAD patients with PCBS and WCBS diabetes. In this study, we evaluated LA function in euglycemic patients and PCBS and WCBS diabetic patients via volumetric assessment and 2D STE-derived indices. Our findings showed that in PCBS, LA conduit function indices—including EDS and SRe—were impaired. Also, PCBS was an independent determinant of LA conduit function: EDS and SRe. Accordingly, our study demonstrated that the conduit function of LA was reduced in the diabetic patients with PCBS compared with that in the diabetic patients with WCBS.

Two studies compared 2 groups of subjects with and without CAD.[28, 29] Both studies found that SRe, as index of LA conduit function, was impaired in the CAD patients. In one of these studies, SS, SRs, and SRa were reduced in these patients. LAD stenosis was correlated to SRa in 1 study,[28] and in the other study, no correlations were found between the location of coronary artery stenosis and 2D STE-derived indices of LA.[29] In our study, all the patients had significant stenosis in the LAD. There were no statistically significant differences between the 3 groups in our study concerning the location of coronary artery stenosis and the number of coronary arteries with significant stenosis.

There are several studies on the comparison of LA function between nondiabetic and diabetic patients.[23-27] In most of the available studies, while there are no data on HbA1c in nondiabetic patients,[23-26] there is information regarding HbA1c in some diabetic patients.[25] It is, consequently, possible that prediabetic patients were included in the normal groups. The effects of prediabetes on LA function has been previously shown.[27] Another issue in some of these studies is that a significant proportion of the diabetic patients had PCBS.[23, 26, 27] In only 1 study, only WCBS patients were included in the diabetic group.[24] The fact that these studies excluded patients with suspected CAD means that the available literature contains no research on the comparison of LA function between CAD patients with and without diabetes. Also, in some mentioned studies, CAD was excluded only by history taking[27] or by exercise test[23, 24] or stress echocardiography.[25] Only in 1 study was CAD excluded by coronary computed tomography angiography.[26] It seems whereas patients with CAD and patients without CAD were not completely separated in those studies, all of our patients had confirmed CAD. The following is a comparison of our findings with those reported by the aforementioned studies as well as possible explanations for the dissimilarities between the results.

Muranaka et al[23] showed that the absolute values of SRs and SRe were less in their diabetic patients than in their normal group. The authors, however, did not report whether or not diabetes mellitus was an independent determinant of SRe. The following points should be considered in the interpretation of that early study. The investigators utilized color-coded tissue Doppler imaging, which is angled-dependent compared with 2D STE. In addition, they measured the deformation indices of LA only in the mid segments of the LA walls. Also, whereas they reported that mean HbA1c in their diabetic patients was 8.5%, they did not measure HbA1c in their normal control group. Furthermore, in their diabetic patients, HbA1c was not correlated with LA strain rates.

Mondillo et al[24] reported that the absolute values of SS, EDS, LDS, SRs, and SRe were more in their nondiabetic patients with a normal LA volume and diabetic patients. Also, there was no difference between these 2 groups concerning the volumetric parameters of LA function. Although in that study PCBS diabetic patients were excluded, only patients with a small LA size (Max.LA < 28 cc/m2) were included. This LA size is relatively small compared with the new recommended limit (Max.LA < 35cc/m2) for the normal LA size.[31]

Kadappu et al[25] found that the absolute values of SS, SRs, SRe, and SRa were less in their diabetic patients than in their nondiabetic group. The authors measured LA strains and strain rates only in the apical 4-chamber view, and they did not measure HbA1c in their diabetic patients. The participants in that study had different levels of LV diastolic dysfunction, whereas patients with grades II and III were excluded from our study. The authors reported a high level of LV diastolic function and stated that Max.LAV index in their diabetic patients was greater than that in their nondiabetic group. This finding was not repeated in our study.

Liu et al[26] reported that the absolute values of EDS and SRe in their hypertensive nondiabetic patients were higher than those in their hypertensive diabetic patients. The investigators found that LV early diastolic strain rate, Max.LAV index, E/A ratio, and diabetes mellitus were the independent determinants of EDS, while LV early diastolic strain rate, hypertension, body mass index, age, E/e′ ratio, and E/A ratio were the independent determinants of SRe. Nevertheless, diabetes mellitus was not an independent determinant of SRe. The authors also reported that passive emptying index was lower in the hypertensive diabetic patients than in the nonhypertensive diabetic patients. HbA1c in the diabetic patients was 6.7 ± 1.4%, denoting that there was a combination of WCBS and PCBS patients in the diabetic group.

Tadic et al[27] showed that all 2D STE-derived indices and emptying passive index in their diabetic patients were different from those in their euglycemic patients. In that study, although prediabetic patients were excluded from the euglycemic control group, it seems that PCBS diabetic patients were dominant in the diabetic groups (HbA1c = 8.1 ± 1.3%).

In all the studies mentioned previously, LA conduit function in diabetic patients was reduced compared with that in nondiabetic patients. In our study, we showed that LA conduit function in our CAD patients with PCBS was reduced compared with that in our CAD patients with WCBS and euglycemic control patients with CAD. Also, this function was preserved in the CAD patients with WCBS. According to our findings, it seems that LA conduit function is the most sensitive function of the LA that is impaired by PCBS. Also, LV diastolic function as evaluated by e′ was an independent determinant of LA conduit function.

It has been shown that one of the major determinants of LA function is LV function.[32] In the case of LA conduit function, LV relaxation is one of the determinant factors. Our results are aligned to this finding. It has been shown that the prevalence of diastolic dysfunction in type 2 diabetic patients is correlated to HBA1c; it may, therefore, explain persevered LA conduit function in WCBS patients.[33-35] It can be due to an increased formation of nonenzymatic advanced glycation end products and the production of free oxygen radicals in PCBS.[36]

In a recent study, it was revealed that in pediatric patients with type 1 diabetes, blood sugar at the time of echocardiography was able to affect LV systolic deformation.[37] According to that study, diabetic patients with real-time blood sugar >150 mg/dL had a lower absolute value of LV systolic deformation. In our study, it was not possible for us to obtain real-time blood sugar, but echocardiography was done <48 hours after venous blood-sample taking. It was the time as soon as possible for us to do echocardiography. Given the type of oral antidiabetic agents and insulin formulation and the administration time of these drugs to the patients, it seems that the administration time of these drugs could not have significantly affected our results. Another point deserving of note vis-à-vis the aforementioned study is that the implications of its results in adult patients with type 2 diabetes are not clear.

LV relaxation is one of the known determinants of LA conduit function.[32] LV relaxation in postmenopausal women is lower than that in men, which may be because of gender differences in Ca handling, nitric oxide system, brain natriuretic peptide regulation, myocardial matrix turnover, adrenergic receptor system, and rennin-angiotensin activity. Estrogen is involved in most of these mechanisms.[38] That most of the women in our study were in postmenopausal age (61.4 ± 8.4) may explain the negative effect of the female gender on LA conduit function. However, LV relaxation is not the sole determinant factor of LA conduit function.[32] Indeed, there is evidence that LA structural changes can affect LA conduit function.[39] Furthermore, it has been demonstrated that LA conduit function is related to fibrosis and apoptosis.[39] Diabetes is also associated with fibrosis and apoptosis.[40, 41]

The benefits of metformin for cardiomyocyte relaxation have been previously demonstrated.[42] Nevertheless, there are different opinions as regards the putative cardioprotective properties of glibenclamide.[43] Insulin can induce oxidative stress,[44] cardiomyocyte hypertrophy, and extracellular matrix expansion, thereby aggravating cardiomyocyte relaxation.[45] However, diabetic patients with WCBS usually have more hypoglycemic events. These events can induce LV remodeling, which may aggravate LA function inasmuch as LV function is one of the major determinants of LA function.[46]

HbA1c is an indicator of blood sugar control in diabetic patients over the preceding months (about 3 months). The microvascular- and macrovascular complications of diabetes are lessened by controlling blood sugar (HbA1c < 7.0%) if implemented in the early years after the diagnosis of diabetes.[2]

Longitudinal myocardial fibers are the dominant fibers in the subendocardial layer of the LA.[47] It is thus reasonable to assume that the mechanisms damaging LV subendocardial layer in diabetes mellitus can damage LA myocardial fibers. Also, LV function plays a significant role in LA function.[32] These proposed mechanisms are delay in myocardial relaxation due to the modified handling of calcium ions,[48] fibrosis,[41] microvasculopathy,[49] and oxidative stress.[50]

The major limitation of our study was its small sample size. Also, we used software designed for the evaluation of the LV. It was not possible for us to evaluate LA function by 3D STE and in other directions such as radial and circumferential. HbA1c shows the time span of glycemic control up to 3 months and we measured HbA1c at the time of hospital admission. Accordingly, our results showed the effects of blood sugar control in the short term on LA function. Our study participants were affected by CAD and were candidates for coronary artery bypass grafting surgery; the results of our study are, therefore, limited to these patients.

In conclusion, our findings revealed that LA conduit function was impaired in the CAD patients with PCBS compared with that in the CAD patients with WCBS and the euglycemic control patients with CAD.

Ancillary