It has now been almost 20 years since one of the most sentinel papers regarding cost-effectiveness in cardiac imaging was published. The Economics of Noninvasive Diagnosis (END) Multicenter Study Group examined the cost impact of an approach using an initial referral for stress myocardial perfusion tomography versus direct referral for coronary angiography without initial noninvasive testing.16 In this study, 11,372 patients were prospectively evaluated for costs of care both in their initial diagnostic testing and in composite costs over a 3-year period. The costs of care using radionuclide myocardial perfusion imaging (range, $2,387–$3,010) were lower than for patients undergoing direct angiography (range, $2,878–$4,579) (p < 0.0001) despite similar rates of death or myocardial infarction (p > 0.20). Part of this increased cost was due to higher rates of revascularization among patients in the invasive arm. Reductions in cost were also seen due to only one-third of patients assigned to the initial imaging strategy undergoing subsequent coronary angiography. Although some interpreted these results as a license to use nuclear stress testing as a "gatekeeper," it should instead be viewed as a precursor to trials such as COURAGE and BARI-2D, in that revascularization based solely on anatomic disease increases cost without a clinical benefit in terms of event-free survival (i.e., death or myocardial infarction-free survival).17–19
Figure 20-2 shows a graph from the END Multicenter Study Group comparing both the diagnostic and follow-up costs for different clinical risk subsets of patients for stress-first versus invasive-first approaches. Upon review of this finding, several factors have to be kept in mind: (1) The dollar amounts in this figure reflect health care costs from 1995 which would clearly underestimate current costs16; (2) Both guideline-directed medical therapy and revascularization technologies have improved, which also underestimate the costs to both arms if reproduced at present day; (3) Referral patterns for patients undergoing ischemic heart disease evaluation have changed with recent studies showing lower abnormality rates for single-photon emission computed tomography (SPECT) testing.20 These factors taken together would suggest a greater cost-effectiveness for a cardiac imaging first approach in stable ischemic heart disease.
Diagnostic and follow-up costs for direct catheterization and initial stress perfusion imaging in the END trial based on pretest clinical risk. (Reproduced with permission from Shaw LJ, Hachamovitch R, Berman DS, et al. The economic consequences of available diagnostic and prognostic strategies for the evaluation of stable angina patients: an observational assessment of the value of precatheterization ischemia. Economics of Noninvasive Diagnosis (END) Multicenter Study Group. J Am Coll Cardiol. 1999;33(3):661–669.)
In 1999, the Economics of Myocardial Perfusion Imaging in Europe (EMPIRE) study retrospectively examined 396 patients for the cost-effectiveness of several diagnostic strategies, as well as the quality of diagnosis in a patient population undergoing initial evaluation of suspected coronary artery disease.13 The diagnostic approaches evaluated in this trial were exercise stress testing, nuclear stress testing and coronary angiography with differentiation between institutions using nuclear stress testing routinely and for those who did not. The results from the EMPIRE study revealed that costs associated with both initial diagnostic approaches were reduced for institutions utilizing SPECT stress testing as well as a 32% lower 2-year diagnostic and management costs in these populations among patients without identified obstructive coronary artery disease. In patients where the algorithm was SPECT stress testing followed by coronary angiography versus angiography alone, 2-year costs were reduced 54% and 36% in patients without disease and those with disease, respectively, when using SPECT first (Fig. 20-3).13
Mean 2-year diagnostic and management costs for patients with and without coronary artery disease (CAD) based on testing approach utilized in the EMPIRE study. (Data from Underwood SR, Godman B, Salyani S, et al. Economics of myocardial perfusion imaging in Europe–the EMPIRE Study. Eur Heart J. 1999;20(2):157–166.)
More recent data presented in the Study of Myocardial Perfusion and Coronary Anatomy Imaging Roles in Coronary Artery Disease (SPARC) registry evaluated 1703 patients who were prospectively monitored after SPECT, positron emission tomography (PET), or 64-slice coronary computed tomography angiography (CCTA) for their 90-day cardiac catheterization rates.3,21 The results showed that referral to catheterization was relative to the severity of the noninvasive test abnormalities across all modalities, 2.8% catheterization rate in those with normal/nonobstructive results, 20.3% catheterization rate in those with mildly abnormal results, and 48.2% catheterization rate in those with moderate or severe abnormal results (p < 0.001). However, the referral rate to catheterization for patients in the CCTA arm (13.2%) was higher when compared to SPECT (4.3%) and PET (11.1%) (p < 0.001).3
For the purposes of this discussion on cost-effectiveness, subsequent analysis from the SPARC registry was performed to evaluate resource use, medical costs, and clinical outcomes for this cohort, allowing calculation of cost-effectiveness for the varying diagnostic test strategies.22 These results revealed that 2-year costs were lowest for those who underwent initial SPECT testing compared to CCTA and PET (mean $3,965 vs. $4,909 vs. $6,647, respectively), which persisted after multivariable adjustment for clinical characteristics. Furthermore, patients undergoing CCTA had costs that were 15% higher than those undergoing SPECT. However, after controlling for longer survival, the cost-effectiveness of one modality over another became uncertain.22 Notably, this registry was based on physician choice for test selection, and not randomization, and thus selection biases and additional confounders may have been present.
In the Cost-Effectiveness of Noninvasive Cardiac Testing (CECat) trial, 898 patients at high risk for obstructive coronary artery disease were randomized to SPECT, stress echocardiography, magnetic resonance perfusion imaging or index coronary angiography.23 The goal was to compare the index noninvasive test versus direct coronary angiography. Survival was generally similar across the patients randomized to stress echocardiography and radionuclide stress testing when compared to coronary angiography (hazard ratio [HR] 1.0, 95% confidence interval [CI] 0.4–2.9; and HR 1.6, 95% CI 0.6–4.0, respectively). However, magnetic resonance perfusion imaging was noted to have a nearly threefold increased risk of death (HR 2.6, 95% CI 1.1–6.2). In the quality-of-life analysis, no difference was found in any of the four arms, thus revealing a similar pattern across all of the compared patient subsets. Notably, the cost of the index coronary angiography compared to index noninvasive testing was lower, but this was partially attributed to the high rate of subsequent coronary angiography among patients undergoing index noninvasive testing (75–80%). Of note is a key statement made in their report, "Our data clearly demonstrate a limited future role for cost-effective noninvasive imaging if referring physicians are not willing to accept a negative result as ground truth."23 This underlines the utilization of noninvasive testing results in determining patients who would benefit from coronary angiography.
A different conclusion was reached from the results of the single-center clinical evaluation of magnetic resonance imaging in coronary heart disease (CeMARC). A total of 752 patients with suspected angina and at least one cardiovascular risk factor were recruited to undergo SPECT, CMR, or coronary angiography in order to evaluate the sensitivity, specificity, positive predictive value, and negative predictive value of SPECT versus CMR when using coronary angiography as the gold standard.24 A subsequent decision analytic model was applied to this data to identify the most cost-effective diagnostic approach in a similar patient population.25 Two diagnostic approaches using CMR (one with index ETT testing followed by CMR after a positive or inconclusive ETT and the other with index CMR) were concluded to be the most cost-effective with a threshold range cost of £20,000 (index ETT) to £30,000 (index CMR). In the recently published CE-MARC 2 trial, CMR and SPECT both reduced the probability of unnecessary angiography compared with NICE guideline-directed care. There was no statistical significance between the CMR and SPECT arms.26
Importantly, one of the least costly of initial noninvasive tests has been the exercise treadmill test (ETT). In the What Is the Optimal Method for Ischemia Evaluation in Women (WOMEN) trial, 824 women with intermediate pretest likelihood of coronary artery disease were randomized to an index diagnostic test of ETT versus exercise SPECT imaging.12 Of note, all enrolled patients had interpretable EKGs for ischemia and predetermined exercise tolerance >5 METs. Two-year major adverse cardiac event (MACE)-free survival was similar in the two groups with 98.0% in the ETT arm and 97.7% in the MPI arm (p = 0.59). The index testing costs for patients undergoing ETT averaged $154.28 compared to exercise MPI of $495.24 (p ≤ 0.001). This cost benefit was sustained when follow-up testing was also added with the ETT arm averaging $337.80 versus the exercise MPI at $643.24 (p ≤ 0.001) (Table 20-2). These results support the cost-effectiveness of index diagnostic testing using ETT in lower-risk women who have an ability to exercise and have an interpretable EKG. Although not a cost-effectiveness trial by design, Borque et al. evaluated a similar question of additive benefit of myocardial perfusion imaging among patients who were able to exercise beyond 10 METS without ischemic ECG changes.27 This study demonstrated that the prevalence of >10% ischemia in this group was very low (0.6%) with an annual cardiac mortality rate of 0.1% per year. Given these low event rates, it appears that there is little incremental value of SPECT MPI for patients who can perform a high-level exercise without ECG evidence for myocardial ischemia.
Table 20-2Diagnostic Costs for Women Undergoing Initial ETT Versus Exercise MPI Testing. Results of the WOMEN Trial ||Download (.pdf) Table 20-2 Diagnostic Costs for Women Undergoing Initial ETT Versus Exercise MPI Testing. Results of the WOMEN Trial
|Test Strategy ||Index Testing Mean (SD) ||Follow-Up Testing Mean (SD) ||Total Costs |
|ETT ||$1454.28 ($30.42) ||$179.97 ($413.64) ||$337.80 ($416.26) |
|Exercise MPI ||$495.24 ($8.54) ||$144.77 ($407.75) ||$643.24 ($411.51) |
| ||p < 0.001 ||p = 0.0008 ||p < 0.001 |
In 2015, the Prospective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) trial published results comparing noninvasive anatomic assessment versus functional testing for patients presenting with chest pain.28 A total of 10,003 symptomatic outpatients who required nonurgent, noninvasive cardiac evaluation were randomized to initial evaluation with CCTA versus stress testing (67% stress radionuclide imaging, 23% stress echocardiography, and 10% ETT). The primary endpoint of a composite of death, myocardial infarction, hospitalization for unstable angina, or major procedural complication was similar between the diagnostic arms with an event rate of 3.3% in the CCTA group and 3.0% in the functional-testing group (HR 1.04; 95% CI 0.83–1.29, p = 0.75). An economic analysis of 9649 of the enrolled patients was subsequently published.29 The costs were divided into three sections: First, was the cost of the index testing. The range of mean costs for different modalities ranged from $174 for ETT and up to $1132 for pharmacologic SPECT. The second analysis examined cumulative costs through 90 days. In comparing treatment approaches, the CCTA strategy had mean costs of $2494 compared to the functional strategy costs of $2240. A major component of this cost difference was attributed to coronary angiography and revascularization. In the CCTA group, 12.2% of the patients underwent invasive angiography with 6.2% undergoing revascularization compared to 8.1% of the functional testing group undergoing invasive angiography and 3.2% undergoing revascularization. Through 3 years of follow-up, using bootstrap analysis, the cost differences between anatomic versus functional diagnostic testing were minimal after the first year ($26 in year 2 and $91 in year 3 after removal of an outlier representing a noncardiovascular hospitalization). Figure 20-4 shows the cost differences between the two diagnostic approaches. Similarly, a retrospective analysis of Medicare spending from a sample of 282,830 patients from 2005 to 2008 showed that those undergoing evaluation with CCTA had a higher rate of subsequent invasive angiography compared to patients undergoing SPECT evaluation (22.9% vs. 12.1%, p < 0.001),30 which led to higher total costs for CCTA. Importantly, medication use was not included as part of the cost calculations in either of these studies. As revealed in a secondary analysis of SPARC, aspirin and statin utilization were greater following an index CCTA as compared to an index SPECT.3 Whether these changes in medication use could have prevented clinical worsening, additional downstream testing, or coronary hospitalizations are unknown, but are important variables to consider in cost-effectiveness analysis.
Mean cost differences between CTA and functional approaches in the PROMISE trial. Values above the baseline show higher cost for CTA compared to below the baseline show higher cost for the functional test. (Reproduced with permission from Mark DB, Federspiel JJ, Cowper PA, et al. Economic outcomes with anatomical versus functional diagnostic testing for coronary artery disease. Ann Intern Med. 2016;165(2):94–102.)