Background
Coronary artery disease (CAD) is commonly cited as a mechanism underlying diastolic dysfunction. However, the association of CAD without ischemia and left ventricular (LV) diastolic dysfunction has not been convincingly demonstrated in asymptomatic patients. The objective of this study was to determine if such a relation exists using coronary artery calcium score (CACS) as a surrogate for coronary atherosclerosis burden.
Methods
Consecutive eligible patients with normal ejection fraction who underwent CACS assessment, echocardiography, and stress testing with negative results for obstructive CAD between August 2006 and September 2007 were included in this retrospective study. Clinical variables were collected from the medical record. Diastolic function classification was based on established echocardiographic guidelines recommended by the American Society of Echocardiography. Statistical analysis was used to identify predictors of CACS.
Results
A total of 349 subjects (302 men) aged 58 ± 6 years were studied. Risk factors included hyperlipidemia ( n = 202 [58%]), hypertension ( n = 127 [36%]), impaired fasting glucose ( n = 78 [22%]), and diabetes ( n = 21 [6%]). Left atrial volume index was weakly correlated with CACS ( r = 0.26, P < .001). There was no significant relationship between CACS and LV diastolic function grade in the entire group ( P = .14) or in a subgroup of younger patients ( n = 140) who matched the ages qualifying for premature CAD ( P = .17). After stepwise elimination multivariate analysis, five variables independently predicted CACS: age ( P < .001), hyperlipidemia ( P < .001), LA volume index ( P < .001), male gender ( P = .01), and LV posterior wall thickness ( P = .03).
Conclusions
In asymptomatic patients with normal LV ejection fraction and negative cardiac stress test results, CACS does not correlate with LV diastolic function as defined by established Doppler echocardiographic criteria. In the absence of ischemia, postinfarction LV remodeling, or previous coronary artery bypass surgery, CAD does not appear be a cause of LV diastolic dysfunction in asymptomatic patients.
Left ventricular (LV) diastolic dysfunction is a common condition associated with increased risk for heart failure and mortality. The earliest manifestation of LV diastolic dysfunction is characterized by delayed LV relaxation and is most commonly seen with hypertension or LV hypertrophy. Symptomatic diastolic heart failure occurs when LV stiffness increases and an elevation in mean left atrial (LA) pressure is superimposed on the impaired LV relaxation.
Population-based studies indicate that at least one third of all patients with symptomatic heart failure have normal LV ejection fractions. Because at least half of these patients have prevalent coronary artery disease (CAD) in the form of angina, previous myocardial infarction, or previous coronary artery bypass surgery, CAD is commonly cited as a mechanism underlying diastolic dysfunction. However, the association of LV diastolic dysfunction with stable or non-flow-limiting asymptomatic CAD, independent of other known risk factors, has not been established.
Computed tomographic coronary artery calcium score (CACS) is a surrogate for coronary atherosclerosis burden and independently predicts future cardiovascular (CV) risk. Many of the same factors that contribute to atherosclerosis may also produce LV diastolic dysfunction by either direct mechanisms (e.g., hypertension and age-related vascular stiffening) or secondarily via CAD progression and resulting changes in myocardial compliance. Both CACS and echocardiography provide useful but somewhat different prognostic information for future adverse CV events. Whether there is an independent association between CACS (plaque burden) and LV diastolic function has not been established. If a strong association between CACS and LV diastolic dysfunction exists, this would have important implications for the use of these tests in the risk stratification of individuals for the primary prevention of CAD, atrial fibrillation, congestive heart failure, and stroke. The objective of this study was therefore to determine if a meaningful relation between CACS and LV diastolic function and other echocardiographic measures of risk exists using a study group of asymptomatic adults with normal ejection fractions and a low probability of obstructive CAD on the basis of negative cardiac stress test results.
Methods
Study Design and Population
A retrospective cohort design was used. After obtaining approval from the institutional review board, all asymptomatic self-referred adults aged ≥18 years who had undergone preventive CV testing including CACS assessment and echocardiography from August 1, 2006, to September 1, 2007, were included in the study. A total of 349 consecutive patients were eligible.
Clinical Data
The medical records were reviewed for clinical data. Age, gender, height, and weight were recorded at baseline. Body surface area was calculated as previously reported. Systemic hypertension was defined as the presence of a positive history or antihypertensive treatment. Diabetes mellitus was defined on the basis of abnormal fasting blood sugar and positive history of oral or insulin treatment. Hyperlipidemia was defined as a total serum cholesterol >200 mg/dL or the presence of appropriate drug therapy. Family history of CAD was considered positive if there were men aged <55 years or women aged <65 years old with coronary heart disease. Smoking status was positive if the patient was currently smoking and past if the patient had ceased for >1 year. None of the patients had histories of CAD (defined as a history of acute coronary syndrome, percutaneous coronary intervention, or coronary artery bypass surgery) or cerebrovascular disease. Cardiac stress testing type and results within 1 year of CACS evaluation for each patient were included. Patients without stress testing or with stress testing with positive results for ischemia were excluded from the study. Patients not in normal sinus rhythm at the time of testing were excluded from the study. All included patients had normal ejection fractions and no LV regional wall motion abnormalities.
Echocardiography
Echocardiographic variables were measured using current recommendations of the American Society of Echocardiography and the European Society of Echocardiography. Measurements of LV end-diastolic dimension (LVEDD) and LV end-systolic dimension from the parasternal long-axis view were incorporated into the Quinones formula to calculate ejection fraction: LV ejection fraction = (%∆ D 2 ) + (1 − ∆ D 2 )/(%∆ L ), where %∆ D 2 = [LVEDD 2 − (LV end-systolic diameter) 2 ]/LVEDD 2 × 100%, and %∆ L is the apical correction factor. LV mass was determined using the Troy formula according to the recommendations of the American Society of Echocardiography: LV mass (g) = 1.05[(LVEDD + interventricular septum + posterior wall) 3 − LVEDD 3 ]. LV mass was divided by body surface area to obtain the LV mass index.
Maximal LA volume was measured in all patients using a modified biplane area-length method. Apical four-chamber and two-chamber views were obtained for the determination of LA area and length (from the middle of the plane of the mitral annulus to the posterior wall). LA volume was calculated on the basis of the algorithm ([0.85 × A 1 × A 2 ]/ L ), where A 1 is the four-chamber area, A 2 is the two-chamber area, and L is the average of the two lengths obtained from the apical view. LA volume was indexed according to body surface area.
Diastolic Function
Mitral inflow was assessed using pulsed-wave Doppler from the apical four-chamber view. The Doppler beam was aligned parallel to the direction of flow, and a 1-mm to 2-mm sample volume was placed between the tips of mitral leaflets during diastole. From the mitral inflow profile E-wave and A-wave peak velocities, mitral deceleration time (DT) and the E/A ratio were determined. Mitral annular motion lateral and septal velocity obtained using the tissue Doppler technique were measured, and e′ and the E/e′ ratio were calculated. Diastolic function grade was assessed using the methods and criteria recommended in the updated American Society of Echocardiography guidelines. This method integrates Doppler measurements of mitral inflow and Doppler tissue imaging of the mitral annulus with estimations of LV filling pressures (as previously validated by invasive catheterization data ) and LA volume. An average E/e′ ratio ≤8 identified patients with normal filling pressures, whereas a ratio ≥13 indicates an increase in LV filling pressures. When the E/e′ ratio ranged from 9 to 12, maximal LA volume index ≥34 mL/m 2 and estimated pulmonary artery systolic pressure ≥35 mm Hg were used to identify patients with high LV filling pressures. The variables Ar-A and Valsalva ΔE/A were not available in all patients and were not incorporated into the diastolic function assessment.
Diastolic function was divided into five categories: (1) normal diastolic function, (2) mild diastolic dysfunction (impaired relaxation without evidence of increased filling pressures), (3) moderate diastolic dysfunction (impaired relaxation or pseudonormal with moderate elevation of filling pressures), (4) severe diastolic dysfunction (advanced reduction in compliance), or (5) indeterminate diastolic function. Normal diastolic function included subjects with septal e′ ≥8, lateral e′ ≥10, and LA volumes <34 mL/m 2 . Those with one or more elevated values for these variables were considered abnormal, and additional measurements were used to determine the grade of diastolic dysfunction. Mild diastolic dysfunction (grade I) was classified as a mitral E/A ratio <0.8, DT >200 msec, and an E/e′ ratio <8 (septal and lateral). Moderate diastolic dysfunction (grade II) was classified as a mitral E/A ratio of 0.8 to 1.5, average E/e′ ratio of 9 to 12, and DT of 160 to 200 msec. Severe diastolic dysfunction (grade III) was characterized by restrictive filling with an E/A ratio ≥2, DT <160 msec, and average E/e′ ratio >13 or septal E/e′ ratio ≥15 and lateral E/e′ ratio >12. Subjects were required to meet two Doppler criteria for moderate or severe diastolic dysfunction to be so classified. Subjects meeting one criterion for moderate or severe diastolic dysfunction or those with borderline parameters were classified as indeterminate rather than normal. Diastolic function was categorized as indeterminate in the presence of missing data.
CACS
Multidetector computed tomography was performed to determine CACS on the basis of the x-ray attenuation coefficient, or computed tomographic number measured in Hounsfield units, and the area of calcium deposits (Agatston score ) in a non-contrast-enhanced scan (Brilliance 64; Philips Medical Systems, Cleveland, OH). The Agatston score is a pseudocontinuous variable derived from plaque densities and their areas in all coronaries.
Statistical Analysis
Descriptive statistics (expressed as mean ± SD for continuous variables and as frequencies and percentages for categorical variables) were used to summarize baseline patient characteristics. A natural logarithm was applied to CACS to normalize the distribution. A two-sample t test was used for two-level variables, one-way analysis of variance used for three-level to four-level variables, and simple regression was used for continuous variables to identify predictors of CACS. Spearman’s correlation test was used for skewed continuous variables. Pearson’s correlation test was used for sets of normal variables. On the basis of guidelines for the assessment of CV disease in asymptomatic individuals, patients with low (CACS ≤ 100) versus high (CACS > 100) coronary plaque burden were compared to determine if there were differences in diastolic function grade. A subgroup of men aged <55 years and women aged <65 years was also analyzed to evaluate the relationship between CACS and echocardiographic variables in a population that is considered to be at lower risk for CV events, with less prevalence of confounding factors such as hypertension and LV hypertrophy, which may contribute to diastolic dysfunction. A final multivariate linear regression analysis with stepwise elimination technique was used to find the most important predictors among those were important at the univariate stage. Any P value < .05 was considered statistically significant. A box plot was used to depict the relationship between diastolic function class and CACS. All statistical analyses were performed using SAS version 9.1.3 (SAS Institute Inc., Cary, NC).
Results
There were 349 subjects (301 men, 48 women), ranging in age from 49 to 78 years (mean age, 58 ± 6 years). Risk factors included hyperlipidemia ( n = 202 [58%]), hypertension ( n = 127 [36%]), impaired fasting glucose ( n = 78 [22%]), and diabetes ( n = 21 [6%]) ( Table 1 ). All patients had cardiac stress test results interpreted as negative for ischemia within 1 year of CACS (340 underwent treadmill echocardiography, 6 underwent treadmill electrocardiography, 2 underwent treadmill nuclear perfusion scans, and 1 underwent dobutamine stress echocardiography).
| Variable | Value |
|---|---|
| Age (y) | 58 ± 6 |
| Men | 301 (86%) |
| Hypertension | 127 (36%) |
| Hyperlipidemia | 202 (58%) |
| Diabetes mellitus | 21 (6%) |
| Impaired fasting glucose | 78 (22%) |
| Family history of premature CAD | 24 (7%) |
| Ejection fraction (%) | 64 ± 5 |
| LA volume index (cm 3 /m 2 ) | 30 ± 6 |
| Interventricular septal thickness (mm) | 10.4 ± 1.6 |
| Posterior wall thickness (mm) | 10.1 ± 1.4 |
| LV mass index (g/m 2 ) | 110 ± 29 |
| Mitral e velocity (cm/sec) | 0.70 ± 0.16 |
| Mitral a velocity (cm/sec) | 0.63 ± 0.16 |
| Septal e′ velocity (cm/sec) | 7.3 ± 1.9 |
| Lateral e′ velocity (cm/sec) | 9.4 ± 2.4 |
| DT (msec) | 218 ± 121 |
| Septal E/e′ ratio | 10.0 ± 3.4 |
| Lateral E/e′ ratio | 7.8 ± 2.6 |
| Average E/e′ ratio | 8.9 ± 2.8 |
| Right ventricular systolic pressure (mm Hg) | 27 ± 5 |
| CACS | 175 ± 588 |
| Computed tomographic coronary calcium percentile | 38 ± 33 |
| Log CACS | 2.74 ± 2.45 |
Diastolic Function and CACS
The average CACS was 175 ± 588. The median CACS was 14.2 (interquartile range, 0—135.8). One hundred eight patients had CACS of 0. Of the remaining patients, 103 had CACS >100. The patients were assessed as an entire study group and also as two individual groups of lower risk CACS (≤100; n = 246) and higher risk CACS (>100; n = 103).
The diastolic function grades of all patients are shown in Figure 1 . The majority of patients ( n = 219) had either mild ( n = 65) or moderate ( n = 154) diastolic dysfunction, while normal diastolic function was seen in 115 individuals. No asymptomatic individual had severe diastolic dysfunction, and only 15 had indeterminate diastolic function. For the entire group, there was no significant relationship between CACS and diastolic function grade ( Figure 2 ). When considering patients with lower risk CACS of ≤100, 91 (37%) had normal diastolic function, 39 (16%) had mild diastolic dysfunction, 105 (43%) had moderate diastolic dysfunction, none had severe diastolic dysfunction, and 11 (5%) had indeterminate diastolic function. Of patients with higher risk CACS of >100, 24 (23%) had normal diastolic function, 26 (25%) had mild diastolic dysfunction, 49 (48%) had moderate diastolic dysfunction, none had severe diastolic dysfunction, and 4 (4%) had indeterminate diastolic function. On comparison of the two groups, patients with CACS ≤100 were more likely to have normal diastolic function ( P = .01); otherwise, the prevalence of mild, moderate, severe, and indeterminate diastolic grade was not different in the two groups. Although there was a trend toward a positive correlation between CACS and increasing severity of diastolic dysfunction, this relationship did not meet statistical significance ( P = .14; Figure 2 ).
Univariate and Multivariate Analysis
In the entire study group, increasing LA volume index was weakly correlated with CACS ( r = 0.26, P < .001; Figure 3 ), but there was no significant relation between CACS and diastolic function class ( P = .14) in the overall study group ( Figure 2 ). After stepwise elimination multivariate analysis, five variables independently predicted CACS: age ( P < .001), hyperlipidemia ( P < .001), LA volume index ( P < .001), male gender ( P = .01), and LV posterior wall thickness ( P = .03) ( Table 2 ).
