Discussion
We have developed a simple clinical score based on known risk factors of CAD that can predict the risk of CAD in patients with newly diagnosed HF. Patients with a low score may benefit from an MSCT-first approach, potentially saving money, radiation as well as complications.
There are several limitations to our findings. First, while the two databases are similar in many respects, unfortunately, the Western Danish Heart Registry include no classification of symptoms of angina or extracardiac arteriopathy, both factors included in our score. We have chosen to include these factors in our analysis since they are both strong predictors of CAD in the Web-PATS registry. Another limitation is the possible information bias as not all patients diagnosed with HF receive CAG or, in fact, any sort of evaluation of possible CAD. Whereas the patient population in this study is already subject to a clinical judgement that they are at risk of having CAD. Thus, our findings might not be generalisable to all patients with HF. However, as these patients were evaluated, they probably represent a higher risk of CAD than a complete population of HF patients. Thus, if calculated on all patients, we would expect a lower average CT-HF score as well as a lower average prevalence of CAD. All patients in this study received a CAG and therefore, we do not know the true re-evaluation rate with CAG if an MSCT-first approach was implemented.
MSCT is a Class IIb recommendation in the latest European Society of Cardiology guidelines for patients with a low-to-intermediate risk of CAD.7 However, not all patients with HF are suitable for an MSCT-first approach. Patients with a history of CAD were excluded from this analysis. Further, we chose to exclude patients with a history of chronic kidney disease. This was done mainly because of the risk of contrast-induced nephropathy associated with the increased use of contrast in MSCT and in eventual re-evaluation. Additionally, patients with chronic kidney disease have a higher risk of coronary calcifications and were, thus, more likely to have blooming artefacts, potentially leaving the procedure inconclusive. Clinically unstable patients should be spared the possible extra waiting time of an MSCT and instead be sent directly to CAG where acute revascularisation is available. Also, though recent studies have shown the ability of MSCT to exclude CAD in patients with atrial fibrillation,18 this is not routinely done in all centres due to technical issues. Some of the specific clinical characteristics of HF patients may in itself limit the accuracy and reproducibility of MSCT, such as inability to hold their breath as well as difficulties controlling the heart rate. HF patients also represent a population at a higher risk of contrast-induced nephropathy,19 something to keep in mind when sending patients to an MSCT that in general uses more contrast than a CAG. This underscores the importance of patient selection as unnecessary examinations with MSCT could lead to additional cases of contrast-induced nephropathy.
The most common complications of a CAG are vascular (about 4%–7% of procedures).5 Most of these are minor bleedings, but some are larger or more complicated (eg, pseudoaneurysm or arteriovenous fistula) and may even result in the patient needing a transfusion and prolonging the hospital stay. More serious complications are rare, such as stroke (0.07%), myocardial infarction (<0.05%) and death (<0.08%), though they may be more frequent in patients with congestive HF.5 MSCT is non-invasive, making it less expensive20 and associated with a lower mean radiation dose.15
Large systematic reviews have shown MSCT to be a highly sensitive test for the exclusion of CAD, with a sensitivity of >98%.21 22 However, MSCT still only provide anatomical information and physiological significance, and obstructiveness cannot be routinely assessed from the scan. For this reason, positive findings still have to be confirmed with a conventional invasive CAG. With the addition of fractional flow reserve, MSCT accuracy may improve further limiting the number of re-evaluations with CAG needed. Though still primarily used as a test of CAD, MSCT does provide additional data among patients with HF regarding morphology (eg, hypertrophic cardiomyopathy or dilated cardiomyopathy) as well as aetiology (eg, sarcoidosis).23 Functional imaging tests, such as Rubidium-82 positron emission tomography imaging, is a possible alternative to MSCT as a gatekeeper for CAG. However, MSCT is generally viewed as the most cost-effective gatekeeper.24
An advantage of MSCT is higher availability since the procedure is more easily implemented at centres than invasive procedures. Another potential advantage is the ability to visualise and diagnose significant extracardiac pathology in surrounding structures, mostly mediastinum and lungs. The prevalence of such findings varies in the literature from about 14% up to 28%.25 26
The newly developed score, the CT-HF score, is an easily calculated clinical score that stratifies HF patients according to their risk of CAD. What constitutes low to intermediate on the score remains to be determined. A cut-off of 9 points identifies about 56%–64% of patients with a risk of 13%–21% of CAD, while a lower cut-off of 7 points yields 36%–38% of patients with a risk of 9%–14%. Using the mean price of CAG and MSCT across the seven countries, a cut-off of 8 would save about 14% of the cost and 20% of the radiation. At 7 points, this drops to 14% of the cost and 18% of the radiation.
With the improved survival after AMI and improved treatment of patients with HF, the prevalence of HF is only expected to rise in the coming decades.1 2 CAD is among the most common causes of HF contributing in as many as 65% among male HF patients.1 With the publishing of the STICH trial 10-year follow-up, we now have data for patients with HF and CAD treated with CABG showing a robust long-term benefit.3 Further, HF patients with concomitant CAD have an indication for implantable cardioverter-defibrillator,27 which have shown a clear prognostic benefit unlike patients with non-ischaemic HF for whom the benefit seems to be limited to younger patients.28 This underscores the importance of identifying patients with HF and CAD for whom treatment is clinically relevant.
A possible alternate approach for low-risk patients could be the evaluation of coronary artery calcium (CAC) as the first step before MSCT. Though cut-offs for different levels of risk of CAD varies in the literature, a CAC of 0 is universally associated with a low risk of events29 30 even among patients with HF,31 with long-term follow-up studies finding a low risk of mortality as far as 12 years ahead of a measurement.32 Thus, patients with a low CT-HF score (<4) may avoid a full MSCT in the presence of a CAC of 0.