Cost-effectiveness for imaging stable ischemic disease

Abstract

Stable ischemic heart disease remains a major cause of morbidity and mortality. Although there are multiple imaging modalities to diagnose and/or assist in the clinical management, the most cost-effective approach remains unclear. We reviewed the relevant and recent evidence-based clinical studies and trials to suggest the most cost-effective approach to stable ischemic heart disease. The limitations of these studies are discussed. Incorporating the results of recent multicenter trials, we suggest that for appropriate patients with coronary artery disease with any degree of stenosis or presence of coronary calcium, optimal medical therapy may be most cost-effective. Invasive coronary angiography and/or coronary revascularization would be primarily for non-responders or >/=50% left main stenosis. Stress cardiac magnetic imaging would be performed for those patients with non-diagnostic coronary CT angiography from motion and non-responders from optimal medical therapy in non-diagnostic coronary CT angiography group from high coronary calcium. These paths seem to be safe and cost-effective but requires modeling for confirmation.

Introduction

Stable ischemic heart disease (SIHD) is one leading causes of death and public health burdens. The current diagnostic and treatment approaches to SIHD are variable. Prevention by lifestyle changes and/or medication is important and cost-effective for asymptomatic patients at all ages either by population-based or personalized approach.^1,2 Non-invasive cardiovascular imaging such as vascular ultrasound of carotid/femoral arteries and coronary artery calcium (CAC) score can help assess cardiovascular risk and evaluate the extent of the atherosclerotic disease. The CAC score may be considered for asymptomatic individuals to reclassify low to intermediate cardiovascular risk.^1,2 Other imaging studies are suitable for symptomatic patients.³ 

Clinical information and pretest probability (PTP) are still mandatory to approach SIHD as well as non-invasive functional imaging according to 2014 ACC/AHA/AATS/PCNA/SCAI/STS guidelines⁴ and the 2019 ESC guideline⁵ while 2016 updated NICE guideline no longer recommended PTP as a part of the diagnostic approach.⁶ The Diamond and Forester Score played an essential role in 2010 NICE guidelines. It was created when the prevalence of coronary artery disease (CAD) was high and PTP also has gradually changed due to changes in lifestyles and statin therapy, therefore PTP is no longer suitable for the current situation according to the updated 2016 NICE guidelines.

These guidelines recommend non-invasive imaging such as cardiac stress imaging (CSI) and coronary CT angiography (CCTA) for diagnostic purposes. The 2014 ACC/AHA/AATS/PCNA/SCAI/STS guideline included interpretable ECG as a part of the work-up since it affected the option of imaging tests.⁴ Recently published ESC guidelines recommend CSI or CCTA as diagnostic imaging tests in symptomatic patients with suspected CAD or CSI if CCTA shows CAD of uncertain functional significance or non-diagnostic result.⁵ 

The concepts of these guidelines are supported by the SCOT-HEART trial.⁷ This multicenter trial (4146 patients) compared a standard of care (SOC) arm to SOC plus CCTA (1:1) and showed that the addition of CCTA increased the diagnosis of CAD and decreased the rate of death from CAD or non-fatal myocardial infarction (MI). There were similar rates of invasive coronary angiography (ICA in the two arms at 5 years but more preventative treatment was prescribed following CCTA.   

Cost-effectiveness analysis of non-invasive cardiac imaging for SIHD   

Increased growth of non-invasive cardiac imaging and new treatments bring up challenges with choosing most cost-effective approach for SIHD based on available resources. The cost-effective analysis of imaging approaches to SIHD remains a complex issue. The actual charges of imaging studies and treatment are frequently not publicly reported and extremely vary in Europe and the USA. Great variation depends on several external factors such as practice, facility location (states), inpatient or outpatient setting, insurance companies, reimbursement and healthcare systems.^8,9

The general concepts and limitations of cost-effectiveness analyses may be found in recently published articles.^10,11 The Second Panel on Cost-Effectiveness in Health and Medicine recommended that all analyses include a discussion of relevant limitations and any efforts to compensate for shortcomings.¹² It was also recommended that the analyses be provided for a healthcare sector perspective that incorporates the costs, benefits, and harms that are incurred by a payer and from a societal perspective in which incorporates all costs and health effects. Cost-effectiveness from a societal perspective includes diagnosis and treatments of disease and any (short- and long-term) complications.  Few cost-effectiveness studies for imaging follow these recommendations.   Additionally, many cost-effectiveness studies have been criticized for extreme inhomogeneity of diagnostic studies due to poor methodology, diversity of the target population/microstimulation models and absent several key parameters.¹³ Unfortunately, these are issues for the cost-effectiveness studies reviewed below.

Some of the publications for cost-effectiveness of imaging SIHD included the only cost of imaging tests^9,14,15 while some investigators added PTP and willingness to pay (WTP)^16–19 and cost of treatment of complications and long-term treatment into their evaluation.^19–24 Short-term complications include contrast-related renal toxicity requiring hemodialysis, anaphylactic reaction requiring resuscitation, fatal and non-fatal MI requiring cardiac intervention. Long-term complications include cost of hospital admission and stay from complications, malignancy resulting from radiation exposure, and work-up for incidental findings. The diagnostic imaging approach to SIHD is commonly multimodality and hence the designs of most of these studies were sequential multimodality imaging. The diagnostic layering of additional imaging tests can also add to costs. The highlights and conclusions of the studies are summarized below.    

The Cost-Effectiveness of non-invasive Cardiac Testing (CECaT) study group conducted a prospective study to compare outcomes and cost-effectiveness of initial imaging in stable chest pain.¹⁵ At the time, this study was the first study to include cost-effectiveness of stress cardiac magnetic resonance imaging (sCMR) into their analysis. The study included 898 patients (27 had expired at 2 year follow-up). Subjects were randomized into four arms, invasive coronary angiography (ICA) (n = 222), single photon emmision emmision CT (SPECT) (n = 224), sCMR (n = 226) and stress echocardiography (SE) (n = 226) then received subsequent tests and initial management. 69% of patients in each group were high risk for CAD. 78% of SPECT, 80% of sCMR and 75% of ECHO groups underwent ICA. 20% to 25% of patients did not need further investigation. 21% of negative CSI patients in each modality underwent ICA and positive ICA was found 31% in SPECT, 52% in sCMR and 48% in SE. Total vascularization rate in all modalities was 34% (310/890). There was no significant difference in patient management among these groups. Although the overall mortality rate was low in all groups, sCMR had the highest mortality rate up to 8%, followed by SE at 5%, ICA and SPECT at 3%. ICA was the most costly exam, however the initial ICA strategy had lower mean overall cost than SE and was similar to sCMR up to 3 years. They found that CSI (SE, sCMR and stress SPECT) can be used to defer ICA without changes in clinical outcomes and extra costs in the outpatient setting.

The fractional flow reserve (FFR) PLATFORM trial (Prospective Longitudinal Trial of FFR_CT: Outcome and Resource Impacts) included 584 symptomatic patients and showed that FFR_CT is more cost-effective than ICA within 90 days and also is associated with improvement of quality of life compared to usual non-invasive imaging.¹⁹ The medical costs of FFR_CT guided group in the first 90 days was found to be lower than the usual care group ($7,343 vs $ 10,734, p < 0.0001).

A study from Euro-CMR registry showed a cost reduction compared to a hypothetical ICA + FFR strategy in patients with low to intermediate disease prevalence with minimal cost saving in the subgroup typical angina.²⁵ The cost was calculated from the imaging costs, revascularization and treatment of complications during 1 year follow-up. The study included 3647 patients from 18 countries. The sCMR + ICA strategy reduced the cost compared to ICA + invasive FFR up to 14% in Germany, 34% in UK, 27% in Swiss and 24% in the USA and showed significant cost reduction when compared to ICA strategy up to 59% in Germany, 52% in UK, 61% in Swiss and 71% in the USA.  

Ferreira et al created a model of 100 patients with 10–90% PTP using different seven imaging strategies: (1) Ergometric test (ET) followed by SPECT (positive or inconclusive ET), (2) ET followed by CCTA (positive or inconclusive ET), (3) SPECT (as first option), (4) echocardiography with dobutamine (as first option), (5) CCTA (as first option), (6) coronary artery calcium scan (CAC) followed by CCTA (when CAC > 0) and (7) ICA.²⁰ The investigators included the direct cost (cost of the tests) and indirect costs in the analysis. Indirect cost included cost associated with incidental findings during CAC and CCTA such as pulmonary nodule(s), cost associated with complications of ICA such as major adverse cardiovascular events (MACE) and its compensation for absence from work and vascular access complications and cost of false-negative tests such as repeated exams or further work-up for non-cardiac chest pain symptoms. Cost of treatment of acute or serious complications was not included due to its low incidence. Since this was a short-term analysis, potential harms from radiation were not calculated. It was found that PTP and WTP played important roles in the analysis. For those with PTP</=50% CAD, CCTA was found to be most cost effective whereas ICA seemed to be most cost-effective for those with PTP>/=60%. WTP also affected the imaging approach. At similar WTP of € 60,000 (total cost per 100 patients), CCTA was found to be most cost-effective for those with PTP 10–30% and ICA was most cost-effective for those with PTP 70–90%. 

Turchetti et al not only evaluated cost per correct diagnosis but also included cost per QAYL and reclassification of risk and cost saving per year in their analysis.¹⁶ The authors suggested CCTA followed by ICA was most cost-effective for low to intermediate PTP and CSI may be cost-effective for higher PTP.

A cohort study by Petrov et al evaluated the effectiveness of sCMR as a gatekeeper for ICA.²⁶ The investigators randomized 1158 patients who presented with suspected CAD into two arms (sCMR and ICA) with median follow-up period of 7.9 years. The survival rate was similar in both groups but patients who underwent sCMR had a lower rate of catheterization and significantly lower costs from hospital stay and follow-up compared to direct ICA arm (2904 € vs 3421 € p = 0.018 for hospital stay and 2045 € vs 3318 €, p = 0.037 for follow-up). They found that sCMR prior to ICA saved 12,466 € of hospital costs per life year and also low costs at follow-up.

The study by van Waardhuizen et al suggested that CCTA may be cost-effective for initial imaging for patients with intermediate PTP.¹⁴ SE or SPECT myocardial perfusion imaging (MPI) was preferred for functional imaging for patients with intermediate PTP and ICA for patients with high PTP. Limitations of this analysis included heterogeneity of the studies from different countries and health care/reimbursement systems and lack of WTP.  

Genders at al created a model of 60-year-old patients with low to intermediate PTP (>/=50% stenosis) from three different countries: the USA, UK and the Netherlands.²¹ The authors included cost of treatment of complications and long-term treatment in the analysis. Five strategic initial imaging approaches were included, no imaging, CCTA, CSI, CCTA and CSI and ICA. The decision to proceed with a second test depended on the result of the initial test.  Strategies were considered cost-effective when incremental cost-effectiveness ratio (ICER)s were less than $50,000/quality adjusted life year (QALY) (the USA), £ 25,000/QALY (UK) and €80,000 (the Netherlands). The US had the highest cost of imaging and treatment approach, for example, CABG ($ 38,217 vs   £ 7,318 (UK) vs € 11,887 (the Netherlands)), MI ($ 10,208 vs £ 5,195(UK) vs € 8,151 (the Netherlands)), outpatient percutaneous coronary artery intervention ($, 6529 vs   £ 3,676 (UK) vs € 4,168 (the Netherlands)) and annual medication ($ 24–587 vs £ 9–126  (UK) vs € 34–135 (the Netherlands)). Echocardiography was the least expensive while ICA was the most expensive for all three countries. The authors found that a multimodality approach was more cost-effective than single modality approach and CCTA as an initial test followed by CSI (if >/=50% stenosis of at least one coronary artery) followed by ICA (if ischemia is present on CSI) was most cost effective in the US and Netherlands while OMT was most cost-effective for UK males with moderate CAD on CCTA or mild ischemia on CSI. SE followed by ICA was most cost-effective for UK females with mild to moderate ischemia on SE.   

An analytic model of 60-year-old patients with prevalence of CAD between 20 and 70% by Bertoldi et al used 11 imaging strategic approaches (double and triple modality).¹⁷ The performance of the second and third tests depended on the results of the initial and second tests respectively. Based on the imaging results, the patients were categorized into different health states: low, medium and high risk, medium risk with false negative, high risk with false negative and low risk with false positive. PTP and WTP significantly influenced the optimal imaging approach. CCTA followed by ICA was most cost effective at low WTP (1X GDP/capita) and gradually decreased but still high in high WTP (5X GDP/capita) while sCMR followed by ICA showed gradually increased cost-effectiveness from low WTP to high WTP and was slightly more cost-effective than CCTA followed by ICA when WTP was more than 5X GDP/capita.  The individual health status influenced the long-term outcomes such as MACE and revascularization. The investigators concluded that CCTA was most cost-effective for low to intermediate PTP. sCMR was an attractive alternative. SE was found to be inferior to sCMR, however it performed better than SPECT MPI.  

Pletscher et al created an analytic model for 30% PTP of IHD in 60-year-old male patients using eight different imaging strategies (single, double and triple modalities).²² ICA was the only single modality approach and also served as the final test for double or triple modality approaches. QALY, cost of imaging test, revascularization, treatment of IHD, acute care for fatal and non-fatal events and follow-up after complication were included in the calculation. The most cost-effective strategies were exercise treadmill test (ETT)–sCMR (if ETT was equivocal)–ICA (if ETT was positive or CMR was positive or equivocal) and ETT-ICA (if ETT was positive). ETT–sCMR (if ETT was positive or equivocal)–ICA (if CMR was positive or equivocal), CMR-ICA (if CMR was positive or equivocal) and ETT-SPECT MPI (if ETT was equivocal)-ICA (if ETT was positive or SPECT is equivocal or positive) were slightly less cost-effective. The authors concluded sCMR seemed to be most cost-effective and should be more widely used as a diagnostic approach to SIHD. 

Bertoldi et al created an analytic model using PTP ranging from 20 to 70% and 11 imaging strategies (double and triple modalities).¹⁸ The cost of treated complications such as MI was included in the evaluation. The PTP greatly affected the imaging approach. Three most cost-effective strategic approaches were Exercise-ECG, CCTA-ICA, Exercise-ECG-ECHO-ICA and CMR-ICA for low PTP (20%) and intermediate PTP (50%). Two most cost-effective imaging approaches for high PTP (70%) were ICA-ECHO and ICA-SPECT. They found that CCTA was a cost-effective anatomic strategy. Echo-based functional imaging performed equally well. CMR had higher yield at intermediate to high PTP and acceptable yield at low PTP. 

Min et al generated an economic model based on 1000 55-year-old males at with 20% likelihood of obstructive CAD (>/=50% luminal stenosis of LM or >/=70% in any other major epicardial artery) with 12 diagnostic pathways.²³ The authors included the prevalence of CAD, WTP, QALYs, cost of imaging tests, treatment cost, cost of false-negative diagnosis, cost of treatment of clinical events such as revascularization, myocardial infarction and death. Different PTP and WTP altered the cost-effectiveness of the imaging approach. ETT-CCTA-ICA and CCTA-SE-ICA strategies were cost-effective with ICERs of $49,021 and $52,899 respectively and CCTA-ICA was most effective but more costly with an ICER of $ 233,138 for low PTP (20%). ICERs of similar strategies, ETT-CCTA-ICA and CCTA-SE-ICA increased ($63,294 and $ 73,734 respectively) in intermediate PTP (50%). The most cost-effective imaging approach was ETT-CCTA-ICA in low and intermediate PTP. For high PTP (80%), ETT-SE-ICA was most cost effective with ICER of $38,234 compared to ETT-CCTA-ICA (ICER of $78,404). In conclusion, initial ETT followed by imaging (if ETT is non-diagnostic) seemed to be more cost-effective than any strategies.

Van der Bruel et al evaluated the cost-effectiveness of various diagnostic pathways for SIHD.²⁴ They reviewed meta-analyses and randomized controlled trials (RCTs) and modeled 13 imaging strategies (single, double and triple modalities). Most were triple modality approaches. The only single modality was ICA. The authors included the costs of imaging tests, primary outcomes such as MACE and harms from the tests such as contrast reaction and secondary outcomes such as hospitalization and follow-up visits in the calculation in the analysis. For patients with low to low- intermediate prevalence of CAD 0f 10–30% with gradually increased diagnostic cost per patient from 0 to € 14,000, the most cost-effective imaging was CCTA-ICA followed by ICA. The cost-effectiveness of ICA was slightly increased when the prevalence of CAD increased but still was not as cost-effective as the CCTA-ICA approach. The patient outcomes analysis was analyzed from 14 studies with a high level of bias risk and it was found that the rates of revascularization of usual care, SPECT strategy and CCTA strategy were similar. MRI/SPECT had high positive predictive value but had higher cost. The ICA strategy has the highest positive predictive value but also highest mortality rate. CCTA-ICA and ICA had relatively higher, radiation exposure (4–21 mSvs) while the lowest radiation exposure was found for ECHO-ICA. CCTA followed by ICA if positive provided nearly as many correct diagnoses in their analysis as ICA alone but at reduced cost (Figure 1).

Figure 1.

Number of true positives and diagnostic costs per patient for various imaging strategies. The most expensive and effective test was ICA alone. CCTA followed by ICA when CCTA was positive was found to have an acceptable number of correct diagnoses with less cost per patient. Other strategies were found to have less favorable characteristics.  From Van der Bruel A, Neyt M, Smit Y, Perera R, Van Der Weerdt A, Beek A, et al. Diagnostiek coronaire hartziekten: ME-TA. https://www.dropbox.com/sh/mbo4cppfwis9re9/bnvURkxxco/Diagnostiek coronaire hartziekten.pdf; 2016. CCTA, coronary CTangiography; ICA, invasive coronary angiography

Lorenzoni et al evaluated the cost-effectiveness of various imaging approached to SIHD.⁹ The parameters included costs of the tests and ICER. A total 350 patients with 28% prevalence of obstructive CAD were included with the diagnostic end point as obstructive CAD (>50% stenosis of LM or at least one major coronary artery at ICA). The definition of effectiveness was the percentage of correct diagnosis and costs. The reference was no imaging. The overall prevalence of CAD was 28%. Cost-effective strategies were CCTA followed by SE, SPECT, PET or sCMR followed by CCTA. Cost saving comparison to “no imaging” of sCMR-CCTA, CCTA-PET, CCTA-SPECT and CCTA-SE were € 969, € 1,490, € 309 and € 3,776 respectively.  In conclusion, CCTA followed by CSI was found to be the most cost-effective approach for low prevalence CAD.    

A cost-effectiveness analysis was performed of the updated 2016 NICE guidelines based on four studies.²⁷ The definition of significant stenosis was >50% stenosis. Only imaging costs was included in the analysis while the costs of treatment (CAD, complication from CAD and procedure, hospital stay) and follow-up visits were not included. Two studies from NICE group and one study from external publication had major limitations and provided limited data.

Discussion

Our review of the cost-effectiveness literature shows a number of important limitations of these analyses. These include heterogeneity of the studies from different countries and health care/reimbursement systems and lack of WTP. An additional issue is the reference standard used for diagnostic accuracy of testing. Most have used ICA as the gold-standard. Some considered 50% stenosis to be significant and others 70%. It is widely accepted that FFR is better than luminal stenosis when evaluating for hemodynamic significance. With the exception of the study from Euro-CMR registry, none of the reviewed cost-effectiveness analyses used FFR as the reference standard. The diagnostic accuracy of MPI is diminished when luminal stenosis is used as the reference standard and therefore the conclusions for cost-effectiveness of various diagnostic tests may not apply for clinical use or be useful for policy decision-making. Given these limitations, it is worth considering how recent high-quality trials may inform future cost effectiveness analyses for imaging SIHD.

The MRI-INFORM trial demonstrated noninferiority of sCMR compared with invasive FFR in terms of MACE in patients with SIHD.²⁸ sCMR had a lower rate of revascularization than invasive FFR. This multicenter randomized controlled trial included 918 patients. The MACE rate and percentage of patients free from angina at 1 year in both groups were similar.

The ISCHEMIA trial randomized patients with documented moderate to severe myocardial ischemia to being initially treated with OMT or OMT + revascularization. The recently presented results reported that there was no significant difference of MACE between OMT and Invasive arms.²⁹ OMT included lifestyle counseling (smoking cessation, nutrition, exercise and mediation adherence) and pharmacological therapy [lipid control (statin therapy), blood pressure control, anginal symptom control and antiplatelet therapy]. A total 5179 out of 8518 patients with moderate to severe ischemia who underwent CCTA to exclude >/=50% LM stenosis or <50% stenosis of the rest of coronary arteries were eligible and included in the study. Several exclusion criteria included left ventricular ejection fraction <35%, severely symptomatic patients, recent MI and recent invasive treatment and advanced renal disease. The trial did not demonstrate a different reduction in MACE (primary outcome: cardiovascular death, MI, resuscitated cardiac arrest or hospitalization for unstable angina or heart failure) between 2 groups after 3.3 years follow-up (13.3% in invasive therapy vs 15.5% in OMT, p = 0.34) as well as secondary end point (cardiovascular death or MI, all-cause death, periprocedural and spontaneous MI).     

Given that all ISCHEMIA patients had documented moderate–severe ischemia, establishing the presence of a significant stenosis may not be as important as commonly thought for SIHD. A more cost-effective approach for diagnosis and treatment of SIHD may be CCTA followed by OMT for those with evidence of CAD (plaques or any degree of luminal stenosis of the coronary arteries) (Figure 2). Unlike the studies which apply CCTA to identify significant stenoses with limited diagnostic accuracy, we make use of the fact that the diagnostic accuracy of identifying atherosclerotic plaque is very high and that atherosclerosis is treated with OMT. Such an approach would need to consider the exclusion criteria of the ISCHEMIA trial. If CCTA was not possible to perform or diagnostic, sCMR we believe is preferred since it has been proven to non-inferior to invasive FFR, the accepted gold-standard for hemodynamically significant CAD. This approach requires modeling and validation and is unique in that the downstream treatment would be included in the cost-effectiveness analysis.

Figure 2.

Proposed cost effective imaging strategy for SIHD inferred from the ISCHEMIA trial. Patients with any degree of coronary artery stenosis or positive coronary calcium should be treated medically. ICA and/or cardiac Intervention is limited to non-responders and those with >/=50% LM stenosis. When CCTA is non-diagnostic secondary to motion, or because of high calcium score and the patient is not responding to medical treatment, sCMR should be performed to confirm the diagnosis. *, exclusion criteria include LVEF <35%, prior cardiac intervention within 1 year, prior MI within 1 month **, any degree of luminal narrowing or stenosis ***, no luminal stenosis +, positive inducible ischemia seen as positive perfusion defect 2.5 consecutive segments on stress compared to rest perfusion according to MRI-INFORM trial -, no inducible ischemia ++, ICA shows severe stenosis (>/=50% LM stenosis) or complete occlusion of major coronary arteries --, ICA shows any degree of stenosis or luminal narrowing of coronary arteries except severe LM stenosis (>/=50%). CCTA, coronary CTangiography; ICA, invasive coronary angiography; LM, left main; LVEF, leftventricular ejection fraction; SIHD, stable ischemic heart disease.