Should NICE Guidelines Be Universally Accepted for the Evaluation of Stable Coronary Disease?

A Debate

Harvey S. Hecht; Leslee Shaw; Y.S. Chandrashekhar; Jeroen J. Bax; Jagat Narula


Eur Heart J. 2019;40(18):1440-1453. 

In This Article

Pro: Nice Guidelines Represent a Nice Evolution

Computed Tomography Angiography Offers Improved Accuracy

Extensive review of the meta-analysis literature reveals that with ICA with >50% diameter stenosis as the reference, CCTA, stress MRI, and PET emerge as the most sensitive and specific modalities, compared with single-photon emission computed tomography (SPECT) and stress echo imaging (Table 1).[6–9] On the other hand, with invasive FFR ≤0.80 as the reference (Table 2),[9–13] CCTA, and stress MRI and PET prove to be the most sensitive, and the latter two to be the most specific. Coronary CT angiography is the least specific but CT derived FFR and CT perfusion (CTP) increase the specificity and positive likelihood ratio closer to the level of MRI and PET imaging without loss of sensitivity (Table 2). Single-photon emission computed tomography and stress echo appear to be the least sensitive (Table 2).[9–13] Thus, while the virtue of CTA is popularly considered to be ruling out the presence of significant obstructive disease by virtue of its extraordinary negative predictive value (NPV), the enhanced specificity provided by CT derived FFR and CTP make it equally effective in accurately identifying FFR positive stenoses (Table 2).[10–13] Moreover, in the presence of persistent symptoms in the absence of ischaemia by non-invasive functional testing, proceeding to CTA is appropriate in view of the lower sensitivity for >50% DS and invasive FFR<0.80 of non-invasive functional imaging compared with CTA.

In short-term randomized controlled trials of CCTA vs. usual care (UC) with functional testing, the former strategy has a lower rate of myocardial infarction, lower downstream testing, greater implementation of medical therapy, possible improvement in quality of life and angina, without an increase in the referral for invasive coronary angiograms.[14–21]

Computed Tomography Angiography Offers Improved Outcomes

More important than diagnostic accuracy is the effect of different testing strategies on cardiac outcomes and resource utilization (Table 3). Individual[14,16–18] and pooled[19]results for CTA vs. UC are shown in Table 3 for four short-term prospective randomized trials totalling 14 187 stable chest pain patients and the only long-term trial (SCOT-HEART 5-year follow-up).[20] In the PROMISE trial,[14] 10 000 symptomatic patients were randomized to CTA or functional testing (SPECT 67.5%, SE 22.4%, ETT 10.2%) and followed for a mean of 25 months. There were no outcome differences, very likely because of the low incidence of events and obstructive disease or ischaemia, but CTA had better prognostic value than functional testing (c-index 0.72 vs. 0.64, P = 0.04).[15] In the Scottish Computed Tomography of the HEART (SCOT-HEART) trial, 4146 subjects were randomized to CCTA plus ETT vs. ETT alone with a 21 month median follow-up,[16] followed by a 5 year median follow-up.[20] The CAPP trial randomized 489 patients to CCTA vs. ETT and followed for a median of 12 months.[17] In the pilot trial by Min et al.,[18] 190 patients were randomized to CTA vs. SPECT and followed for a median of 2 months. Pooled results of the shorter term trials revealed significantly fewer myocardial infarctions with CTA (3.8% vs. 5.6%, P = 0.038), as well as higher rates of revascularization (49% vs. 21%, P = 0.01) compared with UC. There were no differences in death, ICA, or chest pain readmission.[19]

Several trials were published after the cut-off date for the meta-analysis. Computed Tomography vs. Exercise Testing in Suspected Coronary Artery Disease (CRESCENT) randomized 350 patients in a 2:1 ratio to CAC + CTA if CAC 1–400 vs. functional testing (ETT 50%, SPECT 29%, SE 7%, ICA 11%) with a median 1.2 year follow-up.[21] It reported significant reductions for CTA in major adverse cardiac events (3% vs. 10%, P = 0.004) with an event free hazard ratio of 0.36, (96.7% vs. 89.8%, P = 0.01) compared with UC, as well as a significant cost reduction (P < 0.0001) without change in ICA frequency. The Danish National Registry compared 53 744 stable CAD patients evaluated by functional testing (80% treadmill testing, 20% MPI) and 32 961 patients evaluated by CTA, with a 3.6 year follow-up period.[22] There was a lower risk of myocardial infarction following CTA (hazard ratio 0.71; 95% confidence interval 0.61–0.82) and comparable all-cause mortality (hazard ratio 0.96; 95% confidence interval 0.88–1.05). Finally, the recently published SCOT-HEART 5 year follow-up reported hazard ratios of 0.59 (P = 0.004) for CTA compared with UC for the primary endpoint of death from CAD or non-fatal myocardial infarction and 0.60 for non-fatal myocardial infarction without differences in ICA or revascularization.[20]

Computed Tomography Angiography is Associated With Less Non-obstructive Invasive Coronary Angiography, Downstream Testing, and Angina and Increased Quality of Life and Medical Therapy

The percentage of patients who underwent ICA that revealed no significant obstructive disease (<50% DS) in the prospective trials discussed above, with the addition of CRESCENT,[21] is displayed in Table 4. In three studies, there was significantly more non-obstructive ICA following UC than CTA; there were no differences in CRESCENT. Downstream testing was significantly lower for CTA in the two trials in which it was reported. Significantly improved quality of life following CTA was reported by CAPP and decreased angina frequency following CTA was noted in CRESCENT. In the two trials that reported changes in medical therapy, CTA was more effective than UC. In particular, in the SCOT-HEART 5 year study, improved outcomes were attributed to the greater increase in medical treatment since, there were no differences in revascularization.[20]

Computed Tomography Angiography Should Improve Compliance With Appropriate use Criteria (AUC)

The multiple choice and less prescriptive array of possible first line non-invasive tests, such as ETT, vasodilator and dobutamine myocardial perfusion imaging, exercise and dobutamine SE, vasodilator, dobutamine and exercise MRI, vasodilator PET and CTA, would be reduced to the single choice of CTA except for those with previously known disease, contrast allergy, or renal functional limitations.

In the United States, the number of stress tests has increased from 1.6 million annually in 1993–1995, to 3.8 million per year in 2008–2010, and the accompanied imaging tests increasing from 59% to 87%, respectively.[23] However, the increased utilization has not been associated with an increase in their appropriate use application. A meta-analysis of 59 reports of 103 567 tests, published from 2000 to 2012, reported 53% appropriate use of SE in 2008 and 55% in 2011, and 72% appropriate use of MPI in 2005 to 77% in 2009. During almost the same time, appropriate use of CTA increased from 37% in 2006 to 55% in 2010.[24] A more recent meta-analysis of 34 studies with 41 578 from academic centres reported SE appropriate testing rates of 53.0% and 50.9% using 2008 and 2011 criteria, respectively, and appropriate MPI testing rates of 71.1% and 72.0% by the 2005 and 2009 AUC, respectively.[25] There was no temporal improvement in appropriateness for either modality.

Expected Financial Implications

The impact on cardiologists in the private sector is likely to be profound because ETT, SE, and MPI would be substantially reduced in accordance with an increase in appropriateness; there are very few CT scanners owned and operated by individual cardiology practices. Except for ETT, CTA provides the lowest reimbursement, but requires far higher acquisition and maintenance costs than functional testing modalities. On the other hand, imaging sections including radiologists in the hospital outpatient settings will benefit. The resultant decrease in office revenues may accelerate the already rapid United States trend of cardiologist employment by hospital systems and the decline of individual or group practices. The Centers for Medicare and Medicaid Services and private insurance companies would be the executors of the new policy by restricting reimbursement for non-designated first line imaging. Major reorganization of disease states cost estimates would only follow after analysis of the first few years of implementation. The appropriate use of testing and restriction in numbers might improve incentive payments and at least partially compensate for the loss of revenue. The financial impact is likely to be considerably less severe in Europe where the vast majority of care is rendered by national health systems.

Potential Incorporation of Fractional Flow Reserve by Computed Tomography With Computed Tomography Angiography

A NICE guideline for CT-FFR was also published recently.[3] Its approval as a second line modality in the United Kingdom would probably further decrease utilization of the existing functional imaging tests. Use of FFRCT elsewhere might also accelerate reimbursement by CMS and the US insurance industry. It is currently utilized more commonly in Europe than in the US.

Radiation Burden and Incidental Findings

The 2012 NICE guideline recommendation that the risk from radiation procedure be considered when choosing a diagnostic test[2] has since been deleted from their latest iteration. To quote the 2016 Update '… because CT coronary angiography is now the recommended first line diagnostic test and is considered to pose minimal risk of radiation exposure relative to the yield of important diagnostic information'.[1] Advances in imaging technology and surveillance measures incorporated in imaging platforms will further help optimize imaging parameters and reduce radiation exposure.

Incidental non-coronary chest findings will also significantly increase. While this has been raised as a possible harm in the context of CAC CTA screening in a presumably healthy asymptomatic population, it has not been an issue for CTA use in symptomatic patients.

In summary, the superior accuracy and outcomes associated with CTA compared with the most widely used and available functional tests of exercise electrocardiography, SE and SPECT, as well as the likelihood of improved compliance with AUC, justify the NICE placement of CTA as the test of choice for chest pain evaluation in patients without documented obstructive CAD. MRI and PET are comparable to CTA in accuracy (and superior in specificity compared with CTA alone without CT-FFR or CTP), but their expense and lack of widespread availability render them less attractive as the first line test.