Correlation Between Diastolic Dysfunction and Coronary Artery Disease on Coronary Computed Tomography Angiography.

AIMS: We investigated the relationship between coronary artery calcium score (CACS) and coronary artery disease (CAD) on coronary computed tomography angiography (CCTA), and measures of left ventricular diastolic function (DD).
METHODS: We included 429 consecutive patients (39% women; mean age 49 ± 12 years) without known CAD, who underwent CCTA and transthoracic echocardiography (TTE) within 1-month. Evaluation of CCTA was per vessel, and per segment basis for intraluminal diameter stenosis. We also used the 16-segment model to determine overall coronary plaque burden with segment involvement score (SIS). DD on TTE was assessed using mitral inflow E wave-to-A wave ratio (EAR) and tissue Doppler early mitral annual tissue velocity axial excursion.
RESULTS: A total of 293 (68.4%) patients had DD, 15.4% had more than stage 2 DD. The presence of DD was associated with increasing CACS (P < 0.001). Similarly, there was a statistically significant correlation between EAR and CCS (r = -0.147, P = 0.004) and SIS (r = 0.536, P < 0.001). The prevalence of more than stage 2 DD was associated with higher prevalence of obstructive CAD (26.2% vs. 11.7%, P < 0.0001). In multivariable analyses, the independent predictors of more than stage 1 DD are the age (P < 0.001), and diabetes (P = 0.010), while the CACS and SIS were not independently associated with DD.
CONCLUSION: Our analysis suggests that CACS, as well as CAD by CCTA, are not independently associated with measures of DD on echocardiography.

BACKGROUND

Atherosclerotic vascular disease starts to develop early in life and continues to progress for decades as a silent process. This condition, so-called subclinical atherosclerosis, is highly prevalent in the general adult population.[1] During this period, nonobstructive atherosclerotic plaques may not cause ischemia,[2] but promote functional changes in vascular tone, related to vasoactive mediators release and impaired production of nitric oxide by a dysfunctional endothelium.[34] Computed tomographic coronary artery calcium score (CACS) is a surrogate for coronary atherosclerosis burden and independently predicts future cardiovascular (CV) risk.[567]

Abnormal diastolic filling of the left ventricle (LV), as evaluated by transthoracic echocardiography (TTE), is associated with worse prognosis.[89] Diastolic function (DD) is the first cardiac function to be impaired in ischemic heart disease.[10] It is possible that the pathophysiological process of subclinical atherosclerosis, usually not sufficient to reduce systolic function, is able to alter the DD of the LV.

The relationship between subclinical stages of atherosclerosis and DD has not well been established in subjects without prior coronary artery disease (CAD). There is limited and contradicting evidence suggesting that atherosclerotic burden may or may not have an inverse relationship with measurements of diastolic dysfunction.[1112] Thus, the objective of this analysis is to investigate the relationship between CAD and CACS on coronary computed tomography angiography (CCTA) and measures of left ventricular diastolic dysfunction (DD) on echocardiography.

METHODS

Study design and population

This is retrospective crosssectional study, conduct in a tertiary care center between 2012-2013, which included consecutive patients who underwent a clinically indicated CCTA and TTE within 1-month apart. Patients with prior history of CAD (prior myocardial infarction, prior percutaneous stenting or coronary artery bypass grafting surgery), patients with systolic dysfunction with ejection fraction <50%, severe valvular heart disease, severe restrictive, constrictive or hypertrophic cardiomyopathy, and patients with congenital heart disease were excluded. All data were collected after obtaining King Abdallah International Medical research Center (KAIMRC) approval.

The past medical history, clinical characteristics, and computed tomography (CT) finding were obtained by trained research associates. Systemic hypertension was defined as the presence of a known to be hypertensive or use of antihypertensive treatment. Diabetes mellitus was defined as abnormal fasting blood sugar and/or 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 smoking within 1-year and the past if the patient stopped smoking for > 1-year.

Computed tomography and calcium score

All patients underwent CCTA by a standard protocol.[13] They received intravenous metoprolol to achieve a heart rate of 65 beats/min and 0.4 mg of sublingual nitroglycerin before CCTA. All CCTA scans were performed by 64-detector row CT scanners (GE Healthcare, Milwaukee, WI) or a dual source second generation scanner (FLSH, SIEMENS, Erlangen, Germany) with double-phase contrast protocol as follows: 60 mL of Iodixanol, followed by a 50 mL saline flush. CCTAs were scanned from the carina to the cardiac apex with the use of high pitch scanning or prospective gating for which the scan parameters were 64, 0.625 mm collimation, tube voltage of 120 mV, and 350–750 mA.

Multiphasic reconstruction, postprocessing, and image interpretation were performed by a cardiologist blinded to clinical history and echo findings. All images were examined on an advantage workstation (GE Healthcare, Milwaukee, USA). Agatston scores were calculated with semiautomatic software (Smart Score; GE Healthcare). Coronary segments showing coronary disease were visually scored as mild (1–49%, score 1), moderate (50–69%, score 2), or severe (>70%, score 3). Intraluminal stenosis using the 16-segment coronary model was recommended by the society of CV CT. The segment involvement score (SIS)[14] was calculated as the total number of coronary artery segments exhibiting plaque, irrespective of the degree of luminal stenosis within each segment (minimum = 0; maximum = 16).

Diastolic function and echocardiography

All transthoracic echocardiographic studies were performed by experienced cardiac sonographers using standardized protocols. All echocardiographic studies were interpreted by an expert echocardiographer blinded to coronary CTA data. In addition to standard M-mode, two-dimensional and color Doppler imaging, continuous-wave Doppler examination of tricuspid flow, pulsed-wave Doppler examination of mitral inflow, and Doppler tissue imaging of the septal and lateral mitral annulus in 4-chamber view were performed in each subject in accordance with American Society of Echocardiographic and European Association of Echocardiography guidelines.[15]

Mitral E and A waves were measured, with an E wave-to-A wave ratio (EAR) calculated to determine DD. DD by EARs were assessed to consider the presence of other supporting metrics of DD including septal and lateral axial excursion by tissue Doppler imaging, deceleration time, isovolumic relaxation time, pulmonary venous systolic-to-diastolic ratio, and left atrial enlargement.

In addition, chamber dimensions and LV ejection fraction will be measured by M-mode echocardiogram and left atrium (LA) volume was determined by Simpson's rule using the 4-chamber view indexed to body surface area. LV mass will be calculated according to Devereux's equation and indexed to body surface area. DD was categorized as normal, abnormal relaxation (stage I), pseudo normalization (stage II), and restrictive (stage III) as per the guidelines.[15]

Statistical analysis

Descriptive statistics was expressed as a mean and standard deviation for continuous variables and as frequencies and percentages for categorical variables and used to summarize baseline patient characteristics. Continuous variables were compared using Student's t-test. A two sample t-test was used for two-level variables, one-way analysis of variance used for three-level to four-level variables. Pearson's correlation test was used to determine the correlation between continuous variables. To determine if diastolic dysfunction was independently associated with CAD, a multivariable logistic regression was used adjusting for potential confounder. For all multivariate modeling, the threshold for variable entry into models was P < 0.05 and the threshold for variable removal was P > 0.20. A selection of variables for entry consideration was based on the clinical judgment, results of previous publications, and the expertise of the investigators. Care was given to avoid model over fitting by maintaining an events-to-covariate ratio of at least 10:1. All statistical analyses were carried out by SPSS version 21. (Chicago, IL, USA). A P < 0.05 was considered significant.

RESULTS

Baseline characteristics

A total of 429 patients without prior CAD were included. Their baseline characteristics and demographic data are shown in Table 1. The mean CACS was 28 ± 69 (ranging from 0–473) and 384 (89.5%) patients has a CACS < 100. On CCTA, the SIS ranged from 0 to 12 with mean of 1.9 ± 2.8, and 263 (61.3%) patients has SIS <2. Most patients had either normal or mild diastolic dysfunction.

Table 1

Baseline characteristics and demographic data

The mean mitral inflow E/A ratio was 1.2 ± 0.4 (minimum 0.4 and maximum 4.1), mean LA volume index was 15.7 ± 6.7, while E/e' ratio ranged from 0.6 to 27.1 with mean of 8.1 ± 3.1.

Univariate analysis

In univariate analysis, comparison of the absence and presence of CACS and DD with characteristics and risk factors of patients were shown in Tables 2a and b. There was a significant relationship between CACS and DD grade [Figure 1a]. Patients with a zero CACS on CCTA have a higher prevalence of normal DD (54.7% vs. 76.7%, P = 0.001). Conversely, patients had a CACS > 100 on CCTA have a higher prevalence of diastolic dysfunction (17% vs. 7%, P = 0.001) [Figure 1b].

Table 2a

Comparison of the study cohort based on the calcium score

Table 2b

Comparison of the study cohort based on the diastolic function

Figure 1

Bar graphs shows measurement of diastolic dysfunction and relation with multiple parameters, (a) Mean coronary calcium score and different types of diastolic dysfunction, (b) Diastolic dysfunction in categorizing groups of coronary calcium score, Relationship between coronary calcium score (CCS) and diastolic dysfunction (DD), (c) Diastolic dysfunction in categorizing groups of segment involvement score, a correlation between segment involvement score (SIS) and diastolic dysfunction (DD)

Similarly, patients with SIS more than two have a higher prevalence of diastolic dysfunction (43.3% vs. 20.8%, P < 0.001). Oppositely, patients with zero SIS have a higher prevalence of normal DD (34.1% vs. 59.2%, P < 0.001) [Figure 1c]. The increased CACS was associated with higher E/e' ratio, higher mitral E/A ratio, and larger left atrial volume index [Figure 2a-c].

Figure 2

Scatter plots shows relation of total coronary calcium score with different measurement of echocardiography. (a) Correlation between total calcium score with left atrial volume index, (b) Correlation between total calcium score with mitral inflow E/A ratio, (c) Correlation between total calcium score with E average e' ratio

Multivariate analysis

Using multivariate logistic regression adjusting for age, sex, diabetes, hypertension, and body mass index, there was no independent association between CCS (odds ratio [OR] =1.003, (95% confidence interval [CI] =0.998 – 1.009), P value = 0.225), obstructive CAD (OR = 1.26, (95% CI= 0.71–2.24), P = 0.423), and SIS (OR = 1.14, (95% CI = 0.91–1.42), P = 0.259) and diastolic dysfunction.

DISCUSSION

Our analysis showed that despite sharing common risk factors, atherosclerotic measures on CCTA, and measures of diastolic dysfunction are not independently correlated. Many of the same factors that contribute to atherosclerosis may also result in left ventricular DD by either direct mechanisms (e.g. hypertension and age-related vascular stiffening) or secondarily via CAD progression and resulting changes in myocardial ischemia.[16] Because at least half of these patients have prevalent CAD,[1718] in the form of angina, previous myocardial infarction, or previous coronary artery bypass surgery,[181920] CAD is commonly cited as a mechanism underlying diastolic dysfunction. However, the association of DD with stable or nonflow-limiting asymptomatic CAD, independent of other known risk factors, is not well-established.

Prior studies had contradicting data about the relationship of DD and CAD. Some of the prior data have shown that CAD has an inverse relationship with measurements of DD. Prior studies assessing the association between DD and CAD came to conflicting conclusions. Garcia et al. demonstrated that early stage of subclinical atherosclerotic disease is negatively associated with DD parameters in healthy individuals, regardless of age and clinical characteristics. However, they used carotid intima-media thickness as a surrogate for subclinical atherosclerosis instead of CACS.[3] In another study, Eleid et al. did not find consistent relation between coronary artery plaque burden as assessed by CACS and echocardiographic grade of LV DD in the population of asymptomatic adults with normal LV ejection fraction and negative cardiac stress test results.[16] Their study population had lower risk factor profile than our study: Only 6% are diabetic and 36% hypertensive. Recently, Lin et al. concluded that the extent and severity of obstructive as well as nonobstructive CAD by coronary CTA are associated with increased LV end diastolic pressure and measures of worsening DD.[21] However, their protocol and patient population are different from our study. The time interval between the CCTA and TTE was 1-year. In addition, patients with known CAD or LV dysfunction were not excluded from their analysis, which may have contributed to their positive findings.

In our study, we observed a direct relationship between coronary artery plaque burden and measures of DD on univariate analysis. However, many risk factors, such as age, hypertension, diabetes mellitus, and dyslipidemia, contribute to both CAD and DD. After adjusting for these confounders, we could not detect an independent association between atherosclerosis and DD.

Limitations

Our study is not without limitations. It is a retrospective cross sectional analysis. In the future, prospective longitudinal studies are needed to bring more insights about the relationship between atherosclerotic burden and diastolic dysfunction. In addition, both CACS and echocardiography provide different prognostic information for future adverse CV events. Our analysis lacks outcomes data, and it is not clear from this analysis if diastolic dysfunction adds incremental prognostic value over CAD on CCTA.

CONCLUSIONS

Our analysis suggests that CACS, as well as CAD by CCTA, is not independently associated with measures of DD on echocardiography among patients without prior CAD or LV dysfunction.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.