Cardiac computed tomography for rule-out of ischaemic heart disease in patients with new-onset heart failure

Abstract

Aims

In patients with new-onset heart failure (HF), invasive coronary angiography (ICA) is recommended to rule out coronary artery disease (CAD). The objective is to investigate the utility of a coronary artery calcium score (CACS) and coronary computed tomography angiography (CCTA) for rule-out of obstructive CAD in patients with new-onset HF.

Methods and results

Patients with new-onset HF referred for cardiac computed tomography (CT) were included (2008–22). Patients were grouped according to CACS and CCTA findings. Stenosis on CCTA was defined as ≥1 vessel with ≥50% luminal diameter stenosis. Obstructive CAD was defined as ≥1 vessel with ≥50% luminal diameter stenosis at ICA performed within 120 days from cardiac CT. Revascularization procedures within 120 days from cardiac CT were identified. Overall, 3336 patients were eligible. Obstructive CAD was ruled out in 2332/2780 patients (83.8%) with complete cardiac CT. A total of 1032 (30.9%) patients had CACS = 0, and 377 (11.3%) patients had CACS ≥ 1000. A total of 18.0% of patients had stenosis on CCTA, ranging from 2.8% to 71.7% in patients with CACS = 0 and CACS ≥ 1000, respectively. Obstructive CAD at second-line ICA was diagnosed in 11.5% of patients, ranging from 1.2% to 47.2% in patients with CACS = 0 and CACS ≥ 1000, respectively. Revascularization was performed in 6.9% of patients, ranging from 0.6% to 26.5% in patients with CACS = 0 and CACS ≥ 1000, respectively.

Conclusion

In stable patients with new-onset HF, cardiac CT may be considered as the primary imaging modality to rule out ischaemic heart disease, and implementation of a CT-based strategy for ischaemia rule-out may substantially reduce the need for invasive examination.

Graphical Abstract

Open in new tabDownload slide

Introduction

European and North American guidelines on heart failure (HF) recommend that patients with new-onset HF undergo further investigation to clarify the underlying aetiology.^1,2 Coronary artery disease (CAD) is the most frequent cause of HF in the adult population,³ and invasive coronary angiography (ICA) is recommended in patients with a high risk of CAD.^1,2,4 Identifying the origin of HF is pivotal for patient-centred information and shared decision-making, as it may necessitate revascularization, implantation of cardiac defibrillator device, and/or necessitate preventive medical therapy.^1,2

Despite these recommendations, randomized controlled trials in patients with HF and stable CAD found no benefit of revascularization in terms of major adverse cardiovascular events or overall survival within the first years after the HF diagnosis.^5,6 However, revascularization with coronary-artery bypass grafting compared with optimal medical therapy alone lowered rates of major adverse cardiovascular events among patients with ischaemic HF during long-term follow-up.⁷ Further, a recent meta-analysis demonstrated a modest prognostic benefit of revascularization in patients with chronic HF when combined with optimal guideline-directed medical therapy.⁸ On average, patients with ischaemic HF are at greater risk of sudden death than patients with non-ischaemic HF, and therefore, although the relative benefits are similar, the absolute benefit of revascularization is greater in patients with ischaemic HF.

Treatment with aspirin and statin reduces the risk of myocardial infarction in patients with HF of ischaemic origin, whereas these drugs are not routinely recommended in patients with non-ischaemic HF.^1,9,10 Therefore, identification of patients with new-onset HF of ischaemic origin is necessary to initiate proper medical therapy.

Although ICA is generally considered a safe procedure, a routine invasive strategy increases the risk of procedure-related complications compared with a non-invasive strategy.¹¹

Cardiac computed tomography (CT), including the quantification of a coronary artery calcium score (CACS) and a coronary computed tomography angiography (CCTA), is increasingly used as first-line imaging modality in HF patients, owing to the diagnostic and prognostic value demonstrated in patients with symptoms suggestive of obstructive CAD.^12,13 However, a cardiac CT strategy only holds a class IIa, level C recommendation in European guidelines,¹ whereas examination by ICA holds a class I, level B recommendation. Studies have demonstrated the potential of CACS = 0 to rule out obstructive CAD in small HF study population.^14,15

Therefore, we aimed to describe and evaluate the utility of cardiac CT to rule out obstructive CAD in patients with new-onset HF using real world clinical data.

Methods

Study population

The study cohort comprised patients referred for elective cardiac CT with a first-time diagnosis of HF registered in the Western Denmark Heart Registry (WDHR) between 7 January 2008 and 27 October 2022.¹⁶ The WDHR covers all 13 hospitals with a cardiology unit in Western Denmark (3.3 million inhabitants).

Both HF with reduced and preserved ejection fraction were included. Patients referred for primary investigation by other imaging modalities (e.g. myocardial perfusion imaging tests) or ICA were not included. For patients undergoing more than one cardiac CT during the study period, the first cardiac CT was used as the index-scan. Patients with a history of myocardial infarction or revascularization therapy were not considered eligible for study inclusion. Exclusion criteria were missing information on cardiac CT and death or acute myocardial infarction occurring within 120 days after performed cardiac CT and before scheduled ICA.

The study was approved by the Danish Data Protection Agency. No informed consents or ethical approval are required in observational, register-based studies according to Danish regulations.

Comorbidities

Baseline characteristics including sex, age, body mass index, left ventricular ejection fraction, family history of premature CAD, smoking status, dyslipidaemia, hypertension, diabetes mellitus, and renal function were retrieved from the WDHR. Renal function was expressed as estimated glomerular filtration rate calculated according to the 2021 EPI-CDK Creatinine equation.¹⁷

CCTA

Cardiac CT data were obtained from the WDHR. Cardiac CT was performed according to standard clinical practice using CT scanners with a minimum of 64 detector rows, and the results were analysed and registered in the WDHR by local cardiologists. The WDHR did not provide information on scanning protocols, but all scans were prospectively acquired and beta-blockers and ivabradine was not routinely utilized to reduce heart rate in these patients with HF. Non-diagnostic scans were handled according to local clinical practice. In general, most patients would be referred for ICA if obstructive CAD could not be excluded.

CACS results were grouped into 0, 1–99, 100–399, 400–999, ≥1000, or missing. CCTA results were categorized as (i) no CAD (CACS = 0 and no stenosis on CCTA), (ii) non-obstructive CAD (diffuse vessel disease with <50% luminal stenoses), or (iii) CCTA stenosis (≥1 vessel with ≥50% luminal diameter stenosis). The categorization of CAD was based on CCTA results alone if CACS data were missing. Cardiac CT ruled out obstructive CAD when CACS = 0 and no stenosis on CCTA. CCTA stenosis ruled in obstructive CAD.

ICA

According to contemporary clinical practice, a subset of patient with cardiac CT received subsequent ICA as presented in Table 3. Only elective ICAs performed on the indication of HF within 120 days of the cardiac CT were considered. The 120-day window after cardiac CT was applied to ensure that ICAs were performed as a result of the cardiac CT findings. The 120-day window was based on results evaluating prognostic value of myocardial perfusion imaging after first-line cardiac CT in patients with suspected CAD from the WDHR.¹⁸ A sensitivity analysis was performed to access the revascularization rates from 120 to 365 days that includes revascularization of due to both chronic and acute coronary syndrome. ICA results were categorized as non-obstructive CAD (no vessels with ≥50% luminal diameter stenosis or diffuse CAD) or obstructive CAD (≥1 vessel with ≥50% luminal diameter stenosis). Fractional flow reserve measurements were performed at the discretion of the investigating physician.

Coronary revascularization procedures were similarly considered within a 120-day window from cardiac CT. Coronary revascularization was defined as either percutaneous coronary intervention or coronary-artery bypass grafting. Procedure-related data and clinical findings were analysed locally and registered in the WDHR by the treating physician.

Statistical analyses

Categorical variables are presented as prevalences with percentages and continuous variables as mean ± standard deviation (SD) or median (25th and 75th percentile). Radiation dose was calculated as effective dose and expressed as millisievert (mSv). For conversion of radiation dose to mSv, dose length product was multiplied by 0.014 and Gy·cm² by 0.16.^19,20 Statistical analyses were performed using Stata version 18.0 (StataCorp LLC).

Results

Study population

The study population comprised 3336 patients with new-onset HF (Figure 1). Baseline characteristics are outlined in Table 1. In total, 1143 of 3336 (34.3%) patients were female, mean age was 61 ± 4 years, and median left ventricular ejection fraction was 35% (25.0–40.0%).

CACS and CCTA

Complete cardiac CT was available for 2780 of 3336 (83.3%) patients (Table 2), CACS was available for 3182 of 3336 (95.4%) patients, and CCTA was available for 2934 of 3336 (87.9%) patients. The distribution of CACS is presented in Figure 2A. Overall, 1032 of 3336 (30.9%) patients had CACS = 0 and 377 of 3336 (11.3%) patients had CACS ≥ 1000.

Of the complete study population (n = 3336), cardiac CT ruled out obstructive CAD in 954 (28.6%) patients; 1378 (41.3%) patients had non-obstructive CAD; and CCTA stenosis ruled in obstructive CAD in 602 (18.0%) patients (Figure 2B). Thus, 2332 of 3336 patients (69.9%) had no significant CAD according to cardiac CT and thereby had low risk of CAD as potential cause of HF. Further, obstructive CAD was ruled out in 2332 of 2780 (83.3%) patients with complete cardiac CT.

Among patients with CACS = 0, 28 of 982 (2.8%) patients had a stenosis on CCTA. Among patients with CACS ≥ 1000, 142 of 198 (71.7%) patients had ≥1 stenosis on CCTA, with the majority having ≥50% luminal diameter stenosis in more than one coronary vessel (Table 2).

ICA and revascularization

The number of ICAs performed within 120 days of cardiac CT is presented in Table 3. Among patients with a cardiac CT-based rule-out of CAD, 84 of 2332 (3.6%) patients received additional ICA. Overall, prevalences of obstructive CAD at ICA and revascularization increased with increasing CACS. Among patients with CACS = 0, 12 of 1032 (1.2%) patients had obstructive CAD at ICA. Among patients with CACS ≥ 1000, 178 of 377 (47.2%) patients had obstructive CAD at ICA. Revascularization was performed in 6 of 1032 (0.6%) patients with CACS = 0 and in 100 of 377 (26.5%) patients with CACS ≥ 1000. Similar results are found among patients with complete cardiac CT (Figure 2C).

Supplementary analyses

The absolute number of CAD stenosis in the WDHR increased during the study period (see Supplementary data online, Figure S1). The relative number of CCTAs leading to revascularization remained more or less constant during the study period (see Supplementary data online, Figure S2). The radiation dose used at CCTA increased according to level of CACS, but the radiation dose was low in the group without performed CCTA (see Supplementary data online, Table S1). There was no difference in prevalence of patients diagnosed with obstructive CAD at ICA and performed revascularization when comparing results from the study population and in patients with or without complete cardiac CT (See Supplementary data online, Table S2 and Table S3). A sensitivity analyses illustrated that the number of patients undergoing revascularization did not change from 120 to 365 days (see Supplementary data online, Table S4). Overall, when extending from 120 to 365 days, only an additional 22/3336 (0.7%) patients underwent revascularization.

Characteristics of performed cardiac CT stratified by year are presented in Supplementary data online, Table S5. Herein, it is also presented that the number of performed cardiac CT increased during the study period.

Discussion

In this registry-based study, we found that obstructive CAD was ruled out in 2332 of 2780 (83.9%) of the cohort with complete cardiac CT. Further, patients with CACS = 0 had a very low prevalence of stenosis by CCTA; conversely, among patients with CACS ≥ 1000, a very high proportion had stenosis at CCTA and underwent revascularization.

According to these results, cardiac CT may be used as a first-line non-invasive imaging modality to detect non-obstructive CAD and to rule out obstructive CAD in stable patients with new-onset HF.

Previous studies on CACS

CACS is a surrogate measure of calcified plaques burden and does not quantify non-calcified plaque burden or degree of luminal diameter stenosis. Though CACS correlates with both the presence of coronary stenosis and prognosis, CACS is most often combined with CCTA to obtain a more thorough analysis of the atherosclerotic plaque burden as well as visualization of the coronary artery lumen.²¹

There are limited data on the diagnostic value of CACS in the context of new-onset HF, but studies have demonstrated that CACS = 0 may be used to rule out ischaemic aetiology of HF.^14,15,22 It has also been demonstrated that 46% of patients with new-onset HF had CACS = 0.¹⁵ Furthermore, a review study showed that CACS = 0 excluded ischaemic HF with a specificity and positive predictive value of 98.4% and 98.3%.¹⁴ In the present study, using a larger study population, 30.9% of the cohort had CACS = 0 and the prevalences of obstructive CAD among these patients were 2.8% and 1.2% at CCTA and ICA, respectively.

In non-HF patients with a low clinical likelihood of CAD, the combination of a cardiovascular risk factor assessment together with CACS = 0 has been shown to have a high negative predictive value and enables the deferral of further examination in a large proportion of patients.²³

Based on the present study results using a larger study population with complete cardiac CT, the combination of CACS and CCTA may be a feasible way to exclude ischaemic HF in patients with new-onset HF and a low pre-test likelihood of CAD.

Previous studies on CCTA

In the setting of new-onset HF, the diagnostic utility of CCTA is currently increasing but is sparsely elucidated. In contrast to ICA, CCTA is a non-invasive procedure without risk of vascular complications and while the initial cost of first-line CCTA is lower than ICA downstream costs may increase after the initial CCTA.²⁴ However, various factors may limit the use of CCTA in patients with new-onset HF, including high resting heart rate causing low image quality and risk of overestimating CAD severity, slow passage of contrast and thereby poor contrast filling of the coronary arteries due to low cardiac output, and arrhythmias including premature ventricular complexes and atrial fibrillation,²⁵ which lowers both sensitivity and specificity. Nonetheless, dedicated scanning protocols are often used to compensate for these factors and ongoing technical refinements, including the increased use of artificial intelligence, will likely improve the diagnostic accuracy of CCTA in these challenging settings.

Results from the present study found that cardiac CT could be used to rule out CAD in patients with new-onset HF, and this non-invasive imaging procedure could thereby avoid potentially unnecessary and harmful ICAs in this population.

Potential advancement in cardiac CT

Continued advancement of CCTA technology may increase the number of diagnostic scans and reduce the need for downstream testing. Increased gantry rotation time reduces motion artefact in patients with high heart rates that is a frequent problem in HF patients. Advancement of the photon counting detector technology reduces pixel size compared with conventional CT detectors, enabling higher spatial resolution and improved contrast-to-noise ratio. The potential reduced beam-hardening artefacts using photon counting CT will also increase the CT ability to rule-out obstructive CAD in patients with moderate/severe coronary calcification together with reduced radiation dose and contrast dose.²⁶ Increased focus on CT myocardial tissue characterization and CT-derived myocardial strain measurements can potentially also impact patient management in the future.

In contrast to anatomical testing, myocardial perfusion imaging test enables diagnosis of perfusion abnormalities and myocardial infarction scar tissue.²⁷ Limiting these modalities is a currently non-defined cut-off for myocardial blood flow in HF patients and lack of diagnosis of non-obstructive CAD. No studies have compared the effectiveness of myocardial perfusion imaging compared with CCTA as first-line diagnostic test in HF patients.

Recent European cardiomyopathy guidelines advocate use of cardiac magnetic resonance imaging (CMR) in diagnostic work-up of cardiomyopathies.²⁸ Stress CMR may have a potential role when assessing HF aetiology but needs to be evaluated with prospectively studies.

Clinical implications

Current guidelines recommend accelerated ICA in patients with new-onset HF and a high risk of ischaemic aetiology. An initial diagnostic approach with CACS measurement, which is feasible despite tachycardia in the acute/subacute phase, may rule out ischaemic origin in approximately one-third of patients by precluding coronary calcium (CACS = 0). Conversely, CACS measurement will document extensive coronary calcium in 10% of patients, in whom referral for ICA seems indicated. In the remaining patients, CCTA seems prudent as coronary stenosis is ruled out in most patients (Table 2). Furthermore, only coronary-artery bypass grafting has been found beneficial in reduction of major adverse cardiovascular events when compared with optimal medical therapy in patients with HF.⁷ This finding emphasizes the need of standardizing the use of a first-line non-invasive imaging modality to evaluate ischaemic aetiology of HF in patients with low risk of obstructive CAD. The challenge is to choose the best strategy to evaluate HF aetiology and aim to restrict ICA to patients with HF and high risk of obstructive CAD that mandate revascularization.

Study limitations

The following limitations should be considered when interpreting our results. First, the observational study design may introduce referral and selection bias. The study population included only patients referred for cardiac CT. Referral for elective investigation to outline ischaemic vs. non-ischaemic HF was done at the discretion of the treating clinician. Patients with very high CAD risk e.g. clinical instability of strong suspicion of previous myocardial infarction e.g. suspicious ECG or echocardiography findings were referred for invasive angiography directly.

As patients referred directly for myocardial perfusion imaging or ICA might have a different risk profile, selection bias cannot be excluded. This selection criteria impact the generalizability of the study to only involve patients who were referred direct to CCTA with a low-to-intermediate risk of CAD. However, the selection did not introduce informative bias as potential revascularization had not occurred at the time of study enrolment.

Secondly, ICA was only performed in patients with abnormal cardiac CT, which may introduce verification bias for the ICA stenosis prevalence. Despite the fact that CCTA has previously been shown to have high negative predictive values in other populations,²³ the prevalence of ICA stenosis is likely to be underestimated.

Conclusion

Cardiac CT may be considered as the primary imaging modality to rule out ischaemic heart disease in stable patients with new-onset HF. Randomized studies comparing cardiac CT with ICA for rule-out of ischaemic heart disease in these patients are warranted.

Clinical perspectives

Clinical competencies

ICA is recommended with higher weight than cardiac CT (class I level B vs. IIa level C) for assessment of ischaemic aetiology in patients with new-onset HF.^1,2 Cardiac CT may be considered as a first-line diagnostic tool to rule out obstructive CAD in stable patients with new-onset HF and a low-to-intermediate risk of CAD.

Translational outlook

Randomized controlled trials comparing cardiac CT and ICA in the context of new-onset HF are warranted.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

The study was funded by the University Clinic for Cardiovascular Research, Aarhus University, Aarhus, Denmark, and Goedstrup Hospital, Herning, Denmark.

Data availability

According to Danish legislation, data are not publicly available but can be applied for of the Regions Denmark.

References

Author notes