This report evaluates the current state of evidence regarding effectiveness and harms of noninvasive technologies for the diagnosis of coronary artery disease (CAD) or dysfunction that results in symptoms attributable to myocardial ischemia in stable symptomatic patients who have no known history of CAD.
Systematic searches of the following databases were conducted through July 2015: Ovid MEDLINE®, Cochrane CENTRAL, Cochrane Database of Systematic Reviews, and Evidence-Based Medicine Reviews–Health Technology Assessment. Bibliographies of relevant articles were also reviewed.
Using predefined criteria, randomized controlled trials (RCTs) and observational studies comparing the effectiveness or safety of noninvasive cardiac testing--stress electrocardiography (ECG), stress echocardiography, single-photon emission computed tomography (SPECT), positron emission tomography, coronary computed tomography angiography (CCTA), and calcium scoring via computed tomography--with other noninvasive tests, usual care, or no testing were included. Analyses were stratified by pretest risk of CAD as reported by the authors. The quality of included studies was assessed, data extracted, and results summarized qualitatively and using meta-analysis where feasible. The strength of the evidence was assessed for primary outcomes to reflect the confidence in effect estimates: high strength of evidence (greatest confidence), moderate (moderate confidence), low (low confidence), and insufficient (no evidence or no confidence in the estimate).
From 17,146 citations identified, 46 studies were included. Definition of pretest risk across studies varied. There was no clear difference in myocardial infarction (MI) or in all-cause mortality between different testing strategies across settings or pretest risk groups that included patients with intermediate pretest risk, based on low- to moderate-strength evidence from nine trials. Across studies, the frequency was low for all-cause mortality (0%–1.5% in outpatient settings, 0%–1.1% in emergency department [ED] settings past the initial visit) and for MI (0%–0.8% in outpatients, 0%–3% in ED settings). Invasive coronary angiography (ICA) was more common following CCTA than following various functional tests, with a large trial of CCTA versus functional testing providing high-strength evidence. Revascularization referral was more common following CCTA versus functional testing in general (high strength of evidence) and versus exercise ECG (low strength of evidence) but was similar compared with SPECT and usual care (low strength of evidence). In ED settings, additional testing was more common following CCTA than following SPECT (high strength of evidence) but less common versus usual care (moderate strength of evidence). Hospitalization was less common following CCTA than following usual care at the initial ED visit (moderate evidence for intermediate pretest risk; low evidence for low to intermediate pretest risk), but similar for CCTA and functional testing in outpatient settings (moderate strength of evidence). Few studies compared functional tests, and findings were inconsistent for ICA and revascularization referral; however, additional noninvasive testing was less common with SPECT than with exercise ECG (low strength of evidence for all outcomes). The impact of testing on post-test probability of CAD and subsequent clinical decisions regarding treatment or further testing was not described in RCTs. Harms were rarely reported, and limited information regarding radiation exposure was provided.
A review of current studies found no clear differences between testing strategies across settings with regard to clinical or management outcomes on which to base recommendations for one strategy over another for any given pretest risk group that included patients with intermediate pretest risk. No conclusions regarding low-risk patients or high-risk patients without ACS are possible. Limited evidence from RCTs found no clear differences between CCTA and other strategies in clinical outcomes across risk groups, although anatomic testing may result in a higher frequency of referral for ICA and revascularization. The frequency of all-cause mortality and MI was low across studies in all settings. The absence of information on post-test risk stratification and subsequent decisionmaking precluded evaluation of the impact of testing on patient management or outcomes. Testing strategies vary in radiation exposure; there is inadequate comparative evidence to make judgments regarding exposure for the initial test or downstream testing. Assessment of harms was limited. Future research using more refined evidence-based definitions of pretest risk, coupled with information on post-test risk stratification, its impact on clinical management (treatment and referral for additional testing), and longer term followup to assess clinical outcomes, is needed to determine optimal testing strategies and roles of tests in different pretest risk groups.