Diagnostic Accuracy of Computed Tomography Coronary Angiography in Patients Presenting with Heart Failure of Unknown Etiology in the Middle East.

OBJECTIVES: The purpose of this study is to assess the diagnostic performance of coronary artery calcium score (CACS), computed tomography coronary angiography (CTCA), and the prevalence of coronary artery disease (CAD) as etiology of heart failure (HF) in the Middle Eastern population.
BACKGROUND: CTCA has several advantages compared to invasive coronary angiography (ICA). However, studies on the diagnostic accuracy of CTCA and CACS in detecting the prevalence of CAD in patients with newly diagnosed HF are lacking in the Middle East.
METHODS: This study included 204 patients with symptoms of HF and ejection fraction (EF) of <50% by echocardiography who underwent CTCA for diagnosis of CAD within 3 months. The exclusion criteria were defined as patients with a history of CAD, percutaneous coronary intervention, or coronary artery bypass grafting. All patients with obstructive CAD based on CTCA were referred for ICA. In addition, 30 patients with normal CTCA also underwent ICA for verification.
RESULTS: The mean age was 48 ± 13 years, 69% (n = 141) were male and 31% (n = 73) were female, mean left ventricular EF was 31% ± 9%, and mean CACS was 58 ± 120. Based on the CTCA results, 169 patients had normal or nonobstructive CAD, whereas 35 patients had obstructive CAD. ICA was performed in all 35 patients with obstructive CAD; 30 of them were confirmed as having abnormal ICA, and only 5 had nonobstructive CAD. In addition, 30 patients with normal CTCA underwent ICA testing and were confirmed as having normal ICA. The CTCA had 100% sensitivity, 84% specificity, 86% positive predictive value, and 100% negative predictive value. Of the total population, 30 (15%) who were documented as having obstructive CAD were classified as CAD HF based on ICA. The remaining 174 (85%) patients were classified as having no CAD HF based on normal CTCA and/or ICA. The prevalence of CAD HF based on ICA was 15%. There was a strong correlation between CACS and both CTCA and ICA, with P = 0.001 and 0.0048, respectively.
CONCLUSION: In patients with newly diagnosed HF, CACS and CTCA had a 100% sensitivity and negative value as well as overall excellent diagnostic accuracy. CACS = 0 excluded CAD as the etiology of HF with correlation between CACS groups and both CTCA and ICA. The prevalence of CAD as etiology of HF in the study population was 15%. Copyright:

INTRODUCTION

The current guidelines recommend patients with newly diagnosed heart failure (HF) to be tested for coronary artery disease (CAD), even if the likelihood of CAD is low.[1] The prevalence of angiographically significant CAD in patients with HF with reduced left ventricular ejection fraction (LVEF) has been investigated by a few studies in the last several years. One study reported that CAD was the etiology of HF with reduced EF in one-third of patients without angina and no prior cardiac events.[2] Even without any cardiac risk factors, >10% of both men and women had significant CAD.[2]

Ischemic heart disease is the number one reason for systolic HF. Therefore, it is important that all patients with newly diagnosed HF should be assessed for underlying CAD. The association of coronary artery calcification (CAC) and HF has a sensitivity of 90% for detection of ischemic cardiomyopathy.[3] In addition, a prior study reported that coronary artery calcium score (CACS) =0 ruled out ischemic cardiomyopathy.[4] However, positive CACS does not estimate the extent of significant CAD. Conventional invasive coronary angiography (ICA) is often performed on patients with new-onset HF to exclude CAD.[5]

However, ICA is invasive, expensive, inconvenient for the patient and requires routine follow-up care.[6] Computed tomography (CT) coronary angiography (CTCA) has several advantages over ICA. CTCA, with the newest generation multidetector CT, is fast, patient-friendly, noninvasive, and associated with decreasing doses of ionizing radiation.[7] Several studies have shown the excellent accuracy and feasibility of CTCA for anatomic evolution of coronary circulation in patients with HF.[8] In addition, studies have shown that CTCA can be used to determine CAD HF with high accuracy.[9]

Data on the diagnostic utility of CACS and CTCA in detecting the prevalence of CAD in patients with newly diagnosed HF are lacking in the Middle East. Therefore, this study sought to asses CAD as etiology of new-onset HF, the diagnostic performance of CACS and CTCA, and the correlation between different clinical and diagnostic variables in predicting CAD in this group of patients.

METHODS

Study population

From January 2013 to June 2017, 754 patients were diagnosed with new-onset HF by clinical symptoms and EF <50% by echocardiography. Of these 754 patients, 204 met the study criteria. The inclusion criteria were defined as patients with symptoms of HF and EF of <50% by echocardiography who underwent CTCA for diagnosis of CAD within 3 months after echocardiography. The exclusion criteria were defined as patients with a history of myocardial infarction, CAD, percutaneous coronary intervention, coronary artery bypass grafting, iodine indolence, or poor renal function, or patients who refused to have a CTCA or who were hemodynamically unstable.

All patients in this study underwent CACS and CTCA. Any patients discovered to have obstructive CAD based on CTCA were referred for ICA. An additional 30 patients with normal CTCA also underwent ICA. There was no specific criteria for patients with normal CTCA to have additional diagnostic ICA; however, they may have been referred to ICA based on physician preference, symptoms, particularly angina, or abnormal CACS.

Coronary artery calcium score acquisition and analysis

All patients with heart rates >70 bpm received oral β-blocker therapy with 50 or 100 mg of metoprolol tartrate (AstraZeneca, Zoetermeer, Netherlands) after stress testing to achieve a heart rate <70 bpm for the CAC scan. A nonenhanced CT scan (High Definition CT XT; GE Healthcare) was performed during breath-hold at end expiration to calculate the total CACS. CACS was acquired with electrocardiogram (ECG) triggering at 75% of the R-R interval and with the following scanning parameters: 40 or 48 sections of 2.5 mm thickness, gantry rotation time of 330 ms, tube voltage of 120 kV, and a tube current of 125 mA. If patients were not able to hold their breath, a free-breathing CT was performed. Postprocessing was conducted at a dedicated workstation using SmartScore software (GE Healthcare, Milwaukee, Wisconsin, USA). The CACS was calculated using the standard Agatston criteria.[10]

Computed tomography coronary angiography acquisition

Patients without contraindications received metoprolol, targeting a heart rate of ≤ 65 bpm, and nitroglycerin 0.8 mg sublingually before image acquisition. A bolus tracking technique was used to calculate the time interval between intravenous contrast (Visipaque 320, GE Healthcare, Milwaukee, Wisconsin, USA) infusion and image acquisition.

Final images were acquired with a triphasic protocol (100% contrast, 40%/60% contrast/saline, and 40 cc saline). The contrast volume and infusion rate (5–6 cc/s) were individualized according to scan time and patient body habitus. Retrospective ECG-gated data sets were acquired with a GE high-definition CT (GE Healthcare; Milwaukee, Wisconsin, USA) with 64 mm × 0.625 mm slice collimation and a gantry rotation of 350 ms (mA = 300–800, kV = 120). Pitch (0.16–0.24) was individualized to the patient's heart rate.

The CTCA data sets were reconstructed with an increment of 0.4 mm using the cardiac phase with the least cardiac motion.

Computed tomography coronary angiography image analysis

Images were processed using the GE advantage volume share workstation (GE Healthcare, Milwaukee, Wisconsin, USA) and visually interpreted by two expert observers, blinded to all clinical data. A 17-segment model of the coronary arteries and 4-point grading score (normal or mild [< 50%], moderate [50%–69%], and severe [≥70%]) were used for the evaluation of coronary stenosis. Obstructive CAD was defined as coronary diameter stenosis ≥ 50%.

Invasive coronary angiography

Conventional ICAs were performed within 3 months post-CTCA. The coronary arteries were divided into segments according to the American Heart Association for CTCA analysis.[11] The angiograms were analyzed by two interventional cardiologists blinded to CTCA results. The stenosis was classified as significant if the lumen reduction was >50%.

Statistical analysis

The data were analyzed using the IBM Corp., which was released 2011. IBM SPSS Statistics for Windows, Version 20.0. (Armonk, NY: IBM Corp). Descriptive statistics were presented as means and standard deviations. Independence between categorical variables was tested using Pearson's Chi-squared test. Equality of two means was tested using the independent samples t-test. When there were more than two groups, equality of means was tested using ANOVA. Multivariate logistic regression model was used to assess the joint effect of groups of potential risk factors. The Type I error rate was set at the conventional 5%.

RESULTS

The total number of patients who met the entry criteria was 204. The mean age was 48 ± 13 years, 141 (69%) were male and 73 (31%) were female, mean LVEF was 31% ± 9%, and mean CACS was 58 ± 120. The baseline patient characteristics are presented in Table 1. Based on the CTCA results, 169 patients had normal or nonobstructive CAD, whereas 35 patients had obstructive CAD. ICA was performed on all patients with abnormal and/or obstructive CAD based on CTCA.

Table 1

Baseline patient characteristics of the total study population (n=204)

Factors	Total number (%)
Gender/male	141 (69)
Age (years)	48±13
BMI	32±8
Weight/kg	86±13
Creatinine	94±82
BNP	2384±5336
Total cholesterol	4.7±1
Echocardiography
LVEF	31±9
EKG
Abnormal EKG	48 (24)
LBBB	34 (17)
Q wave	10 (5)
CAD risk actors
Hypertension	80 (39)
Diabetes mellitus	60 (29)
Smoking	30 (15)
Family history of CAD	8 (4)
CHF symptoms
SOB	124 (61)
Fatigue	59 (12)
Peripheral edema	42 (29)
Palpitation	14 (7)
CACS	58±120

BMI: Body mass index, LVEF: Left ventricular ejection fraction, EKG: Electrocardiogram, LBBB: Left bundle branch block, CAD: Coronary artery disease, CHF: Congestive heart failure, CACS: Coronary artery calcium score, BNP: Shortness of Breath, SOB: Brain natriuretic peptide

It was discovered that 30 of 35 patients had confirmed abnormal ICA (true positive), and only five patients had normal and/or nonobstructive CAD (false positive). Another thirty patients with normal and/or nonobstructive CTCA underwent ICA, and all of them were found to have normal ICA. The CTCA had 100% sensitivity, 84% specificity, 86% positive predictive value, and 100% negative predictive value [Table 2].

Table 2

Diagnostic accuracy of computed tomography coronary angiography for distinguishing between coronary artery disease heart failure and noncoronary artery disease heart failure

	n	TP	TN	FP	FN	Sensitivity	Specificity	PPV	NPV
CTCA	61	30	26	5	0	100%	84%	86%	100%

TP: True positive, TN: True negative, FP: False positive, FN: False negative, PPV: Positive predictive value, NPV: Negative predictive value, CTCA: Computed tomography coronary angiography

Of all patients, 30 (15%) documented with obstructive CAD were classified as CAD HF (ischemic cardiomyopathy) based on ICA, and the remaining 174 (85%) patients had no CAD HF based on normal CTCA and/or normal ICA. The prevalence of CAD HF based on ICA was 15%.

The association between patients' characteristics, echocardiography, clinical symptoms, coronary artery calcium score, coronary artery disease heart failure, and noncoronary artery disease heart failure

Age, CACS, and the presence of diabetes mellitus were found to be correlated with CAD HF with P = 0.000, 0.000, and 0.007, respectively. However, there were no correlations between CAD HF and many other factors, such as LVEF, angina, and other clinical symptoms. In multivariable logistic regression, CACS was the only predictive factor for CAD HF [Table 3].

Table 3

Clinical data and coronary artery disease risk factors in both noncoronary artery disease and coronary artery disease heart failure

	Non-HF (n=174)	CAD HF (n=30)	Total (n=204)	P
Gender
Male	123	18	141	0.242
Female	51	12	63
Age	46±13	59±10		0.000
BMI	31±8	33±6		0.316
CAC score	23±61	259±170		0.000
BNP	2539±5723	1482±1704		0.318
Creatinine	95±88	87±26		0.645
LVEF	32±10	29±0.9		0.122
Total cholesterol	4.7±9	4.60.8		0.072
LVEDD	5.8±0.9	6.1±1.2		0134
Angina
Negative	153	25	178	0.951
Positive	20	5	25
Fatigue
Negative	121	24	145	0.243
Positive	53	6	59
Peripheral edema
Negative	141	21	162	0.167
Positive	33	9	32
Palpitation
Negative	162	28	190	0.963
Positive	12	2	14
Diabetes mellitus
Negative	129	15	144	0.007
Positive	45	15	60
Hypercholesterolemia
Negative	79	12	84	0.023
Positive	28	12	37
Hypertension
Negative	109	15	124	0.190
Positive	65	15	80
Smoking
Negative	144	28	172	0.328
Positive	28	2	30
Family history of CAD
Negative	167	28	195	0.406
Positive	6	2	8
SOB
Negative	66	14	80	0.365
Positive	108	16	124
Abnormal EKG
Negative	134	22	156	0.581
Positive	38	8	46
LBBB
Negative	148	22	170	0.112
Positive	26	8	34
Q wave
Negative	148	22	170	0.667
Positive	26	8	34

BMI: Body mass index, LVEF: Left ventricular ejection fraction, EKG: Electrocardiogram, LBBB: Left bundle branch block, CAD: Coronary artery disease, HF: Heart failure, CAC: Coronary artery calcification, LVEDD: Long-axis views at end-diastole, SOB: Shortness of breath, BNP; Brain natriuretic peptide

Of the total study population, 128 (63%) patients had CACS = 0, and none of these patients had abnormal CTCA, yielding 100% negative predictive value. CACS was between 1 and 300 in 64 (31%) patients and >300 in 12 (6%) patients. Although CACS = 0 excluded CAD HF, CACS >1 did not discriminate between ischemic and nonischemic HF. There was a strong correlation between CACS and both CTCA and ICA with P = 0.001 and 0.0048, respectively [Figure 1 and Table 4].

Figure 1

Relationship between computed tomography coronary angiography, coronary artery calcium score, and invasive coronary angiography

Table 4

Relationship between computed tomography angiography and invasive coronary angiography and coronary artery calcium score groups

	CACS=0 (n=128; 63%)	CACS=1-300 (n=64; 31%)	CACS >300 (n=12; 6%)	P
CTA	128	64	12
Normal	128	39	2	0.001
Abnormal	0	25	10
ICA	5	44	12
Normal	5	24	2	0.0048
Abnormal	0	20	10

CACS: Coronary artery calcium score, ICA: Invasive coronary angiography, CT: Computed tomography, CTA: CT angiography

DISCUSSION

This study had five major findings. In newly diagnosed HF, CTCA had 100% sensitivity and negative predictive value. CACS = 0 excluded CAD as etiology of HF. A positive correlation was found between CACS groups and CAD HF. Age, CACS, and diabetes were strong predictors of CAD HF. The prevalence of CADHF was only 15% in this studied population.

Diagnostic accuracy of CTCA in HF is a clinical necessity. These findings support a high accuracy of CTCA in this clinical setting, with both a sensitivity and negative predictive value of 100%. These data are in agreement with previously published studies. ten Kate et al. reported very similar diagnostic accuracy of CTCA in CAD HF of unknown origin, with reported sensitivity of 99% and negative predictive value of 99.8%.[9] Based on this study's findings and those of previous studies, it is well established that CTCA distinguishes between CAD and non-CAD HF, and it may be appropriate to conduct CTCA tests on patients with newly diagnosed HF of unknown origin.

This study confirmed that CACS = 0 excludes obstructive CAD with very high negative predictive value. The association of CACS = 0 and CAD has been previously studied by the researchers of the current study; it was found that CACS = 0 in stable patients with low-to-intermediate risk for CAD indicated a very low likelihood (<1%) of obstructive CAD.[12]

In addition, these results are in line with other published studies. Abunassar et al. reported that Agatston score = 0 confers a very low likelihood for obstructive CAD and appears to rule out high-risk CAD, and thus, it may be used to rule out ischemic cardiomyopathy.[4]

The analysis of CAD risk factors, HF symptoms, and echocardiographic findings has been studied to discriminate between CAD and non-CAD HF. In this study, age, CACS, and diabetes were the only factors that were predictive of CAD HF. In agreement with previous studies, symptoms of HF or LVEF, abnormal EKG or left bundle branch block, and the presence of Q wave did not discriminate between CAD and non-CAD HF.[13] Studies have shown a consistent high prevalence of DM and high CACS in patients with CAD HF, but inconsistent correlation between other factors such as smoking and hypertension with CAD HF.

The relationship between CACS >0 and both CTCA and ICA has been demonstrated in many previous studies. It is well known that there is a strong positive correlation between CACS severity and CTCA results. Studies have shown that CACS is more accurate than myocardial perfusion imaging and echocardiography in predicting severity of CAD. Budoff et al. reported that CACS >0 had a sensitivity of 97% and specificity of 68% for defining myocardial ischemia compared to a sensitivity of 56% and specificity of 50% for nuclear stress testing.[14] Furthermore, it has been shown that CACS is superior to the segmental wall motion abnormalities detected by echocardiography for distinction between ischemic and nonischemic cardiomyopathy.[15]

The prevalence of CAD HF in the studied population was 15%, which was much less than reported by some previous studies. Silva et al. reported that, of a total of 168 patients in an HF clinic, CAD was the etiology of HF with reduced EF in 1/3 of the patients without angina and no prior cardiac events.[2]

Comprehensive studies evaluating the presence of HF, its associated mortality, and causes are lacking or inconsistent in the Middle East. For example, the prevalence of HF in Oman was reported as low as 5.17/100,000 individuals.[16] Recently, in a prospective study of HF patients in 108 centers in 16 countries from 2012 to 2014, ischemic cardiomyopathy was the most common cause (50%) in Middle Eastern participants, followed by dilated idiopathic nonischemic cardiomyopathy (19%).[17]

This study also found that the majority of Middle Eastern participants had diabetes mellitus (56%) and hyperlipidemia (57%).[17] It appears that the low prevalence of CAD HF in the current study was most likely due to bias in the study population, as the patients were all from the same tertiary care center.

Study limitations

There were some limitations to this study. This was a retrospective study in a single tertiary care center; as such, selection bias must to be considered in interpretation of the results of this study. Furthermore, the study population was small and cannot be generalized to a larger population. However, the data were highly supportive for high diagnostic value of CTCA and CACS in patients with newly diagnosed HF. Another limitation was that the results obtained were not compared with other imaging modalities, such as stress nuclear myocardial perfusion, stress echocardiography, or cardiac magnetic resonance imaging.

The prevalence of CAD HF in this study is relatively low compared to other published studies; this is most likely due to patient selection bias, as all patients with known CAD were excluded from the study. Finally, another important limitation was the standard definition of CAD based on CTCA, ICA, and EF by echocardiography. Patients with CAD may have had concomitant etiology of cardiomyopathy. Conversely, myocardial infarction may occur due to coronary spasm or coronary plaque rupture in the absence of obstructive CAD.[18]

CONCLUSION

This study found that CACS and CTCA had a 100% sensitivity and negative value, as well as overall excellent diagnostic accuracy, in patients with newly diagnosed HF. CACS = 0 excluded CAD as the etiology of HF. Furthermore, there was a strong positive correlation between CACS group and both CTCA and ICA. Patient's age, CACS, and diabetes mellitus were the only factors that were able to discriminate between CAD and non-CAD HF. The prevalence of CAD as etiology of HF was only 15% for this study population. Thus, further prospective studies with larger populations from multiple centers are highly recommended to investigate the role of CTCA and CACS, as well as the prevalence of CAD in newly diagnosed HF in the Middle Eastern population.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.