If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Clinical trials conducted before the introduction of modern medical management to prevent stroke demonstrated that carotid endarterectomy (CEA) and carotid artery stenting (CAS) prevent stroke following transient ischemic attack (TIA). We compared the cost-effectiveness of CEA, CAS, and modern medical management in two secular settings of medical management in individuals with incident TIA and type 2 diabetes.
Methods
Using simulation modeling, our base-case analyses were performed from the healthcare sector perspective over a 20-year time horizon with an annual 3% discount rate applied to both costs and quality-adjusted life years (QALYs). Outcomes depended on age, sex, biomarkers associated with cardiovascular risk, and treatment effects based on a validated model of type 2 diabetes. Our simulation population was drawn from the National Health and Nutrition Examination Survey (NHANES) 2014 cohort. Costs for modern medical management were based on average wholesale prices, and revascularization costs were derived from published literature. One-way and probabilistic sensitivity analyses were conducted.
Results
Compared to all other strategies, historical medical management plus CEA was either cost-saving or cost-effective at a threshold of $100,000 per QALY gained. Modern medical management was cost-effective compared to historical medical management without revascularization at a $100,000 acceptability threshold. However, both revascularization approaches (plus medical management) were cost-saving compared to modern medical management alone.
Conclusion
Among individuals requiring carotid revascularization, carotid endarterectomy is the cost-effective strategy to treat individuals with type 2 diabetes following a TIA. For individuals for whom revascularization is contraindicated, modern medical therapy is cost-effective.
Cardiovascular disease (CVD) is an important cause of morbidity, mortality, and health care expenditures in US adults with type 2 diabetes (T2DM). CVD in Medicare beneficiaries with T2DM accounts for 50% of total Medicare expenditures.
The risk of stroke in people with T2DM is twice that of individuals without diabetes, and stroke is the leading cause of serious long-term adult disability in the US.
Stroke prevention among individuals with T2DM may reduce the subsequent morbidity, mortality, and cost of CVD. Because the risk of stroke within 6 months of transient ischemic attack (TIA) has been estimated to be as high as 15%,
carotid revascularization following TIA may be a particularly effective and cost-effective strategy to prevent stroke and reduce disability.
Over the past 2 decades, studies have defined preventive strategies to improve long-term CVD outcomes. Carotid endarterectomy (CEA) and carotid artery stenting (CAS) are well established treatments to prevent stroke following TIA, even in individuals over 70 years of age.
Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. 2014.
Guidelines for the primary prevention of stroke: A statement for healthcare professionals from the American heart association/American stroke association.
When is carotid angioplasty and stenting the cost-effective alternative for revascularization of symptomatic carotid stenosis? A Canadian health system perspective. [Internet].
AJNR. Am. J. Neuroradiol.2014; 35 (Available from): 327-332
Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. 2014.
Guidelines for the primary prevention of stroke: A statement for healthcare professionals from the American heart association/American stroke association.
) were conducted in the 1980s and 1990s, before the widespread use of newer anti-platelet therapies and high-intensity statins (atorvastatin, rosuvastatin) and before the introduction of newer, more aggressive guidelines for blood pressure control. CAS, in turn, has been shown to be non-inferior to CEA.
However, medical treatment has improved in the decades since these trials, and neither CEA nor CAS have been compared to modern medical treatment alone. Thus, some researchers and policy makers question the utility of CEA and CAS relative to modern medical management alone.
Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. 2014.
Guidelines for the primary prevention of stroke: A statement for healthcare professionals from the American heart association/American stroke association.
The Carotid Revascularization Endarterectomy versus Stenting Trial 2 (CREST-2), a large and well-designed clinical trial, is testing the efficacy of CEA and CAS vs modern medical management.
Costs and cost-effectiveness of carotid stenting versus endarterectomy for patients at standard surgical risk: Results from the Carotid Revascularization Endarterectomy Versus Stenting Trial (CREST).
However, the CREST-2 trial is not assessing the long-term cost-effectiveness of these treatments, and by the time long-term outcomes of CREST-2 are available, standards of care may have changed again. Fortunately, fully developed and validated simulation models of CVD that are based on the best available observational and clinical trial data and account for the trajectory of complications and comorbidities of both diabetes and CVD can be used to inform practice guidelines.
Our aim was to simulate the cost-effectiveness of three competing treatment paradigms: CEA, CAS, and medical management alone in each of two settings: historical medical management and modern medical management. This led to a total of 6 treatment scenarios for cost-effectiveness analysis. Our population of interest was people with incident TIA and T2DM who were candidates for carotid revascularization, defined as either CEA or CAS.
METHODS
Our cost-effectiveness analysis was conducted using simulation modeling with 1,000 replications. Our base case scenario involved 20 years of follow-up, and we made modifications to the base model for each of 5 additional scenarios.
Simulation Specifications for Base Case Model
The Michigan Model for Diabetes (MMD), an externally-validated model of diabetes progression, was the basis for our simulation. It describes the trajectory of patients with T2DM as they age and develop risk factors, complications, and comorbidities of diabetes. These include hyperglycemia, hypertension, dyslipidemia, retinopathy, nephropathy, neuropathy, cerebrovascular disease, and coronary heart disease (CHD).
The model includes event states including TIA, ischemic stroke, and myocardial infarction. Individuals experiencing acute events then progress to annual states including history of TIA, history of ischemic stroke with minor disability, major disability, and death. Time-varying biomarkers (e.g., BMI, HbA1c, blood pressure, lipids) are updated annually as a function of time and treatment.
These time-varying covariates, in turn, modify the risk for progression to complications and CVD. The biomarker model also triggers a model of treatment and adherence, as outlined in our CHD model, which includes prescription of and adherence to treatments (e.g., antiplatelet agents, glucose-lowering medications, blood pressure-lowering medications, and lipid-lowering medications).
These treatments directly and indirectly influence individual risk factors and disease trajectories over time. Medical management is modeled in the MMD by changing risk factor levels (i.e., fasting glucose, blood pressure, and lipids) and through direct effects of pharmacologic therapies (such as antiplatelet agents) on outcomes.
Unlike simulation models that are constrained to a single longitudinal data set, the MMD integrates estimates from multiple data sources to create a single self-consistent system of disease progression equations. Finally, we used cutting-edge calibration techniques
and integrate the best epidemiological studies with the results of modern clinical trials.
Treatment Scenarios
Our treatment scenarios considered 3 competing treatment paradigms: medical management alone, medical management plus CEA, and medical management plus CAS in each of two settings: historical medical management (insulin, metformin, sulfonylureas, aspirin, atenolol, and moderate-intensity statin therapy) versus modern medical management (semaglutide, clopidogrel, losartan, and high-intensity statin therapy). Thus, we have a total of 6 scenarios and 15 possible pairwise comparisons.
Historical medical management used the default diabetes treatments described above, based on clinical practice guidelines. For modern medical management, we changed glycemic management from generic sulfonylureas to semaglutide
Cardiovascular morbidity and mortality in patients with diabetes in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): A randomised trial against atenolol.
Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): Multicentre randomised placebo-controlled trial.
The stroke risk reduction due to these therapies relative to the historical therapies were based on the results of modern clinical trials and standards of care for patients with type 2 diabetes
Cardiovascular morbidity and mortality in patients with diabetes in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): A randomised trial against atenolol.
Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): Multicentre randomised placebo-controlled trial.
(Table 1). These therapies may have either a direct effect on stroke risk, or an indirect effect through risk-factor management. Combinations of multiple treatments were considered to be additive. Likewise, differences in quality-of-life following CEA and CAS have been shown to dissipate within 12 months.
Cardiovascular morbidity and mortality in patients with diabetes in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): A randomised trial against atenolol.
Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): Multicentre randomised placebo-controlled trial.
as displayed in Table 1. This included both short and long term risk of stroke. A meta-analysis of CAS vs CEA demonstrated no difference in myocardial infarction between the two treatments, so we did not model an additional cardiac effect.
To acquire baseline data for correlated risk factors, we sampled 400 individuals with diabetes from the 2014 NHANES cohort as our simulation population. Because we were interested in risk among people with type 2 diabetes, and because NHANES did not have an indicator for TIA or eligibility for CAS or CEA, we sampled the NHANES cohort with diabetes to replicate the demographic characteristics of CREST participants. We weighted observations by the inverse distance from CREST's mean of age, sex, hypertension status, smoking status, systolic blood pressure, and diastolic blood pressure. To simulate a more racially diverse sample than enrolled in CREST, we did not sample based on race. These baseline data were used for all simulated scenarios to compare medical management among individuals with T2DM eligible for carotid revascularization procedures.
All subjects, on initiating the simulation, were assumed to have had an incident TIA at baseline which triggered treatment with medical management, CEA, or CAS. Other baseline demographics, risk factors, medical history, and medical management were determined from the NHANES data.
Simulation Costs
We used an Impact Inventory (Supplemental Table 1), as recommended by the Second Panel on Cost-Effectiveness in Health and Medicine,
and adopted a health care sector perspective. We applied an annual 3% discount rate to both costs and QALYs. The economic analyses were reported in compliance with the Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022).
Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 Explanation and Elaboration: A Report of the ISPOR CHEERS II Good Practices Task Force [Internet].
For the carotid revascularization procedures, costs were based on total in-hospital patient cost, including fixed costs (administrative, capital, and utilities) and variable costs (labor and supply) as noted in Table 2.
Table 2Intervention costs of historical, modern medical, and surgical management
To identify the best treatments, we used a 2-stage approach. First, we ordered the 6 strategies by average total cost and eliminated competing treatments that were both more expensive and less effective than another treatment. Second, we used incremental cost-effectiveness ratios (ICERs) to measure the relative cost-effectiveness of the treatments that were not eliminated in the first step. The ICER represents the additional cost required to achieve one additional unit of health benefit (QALY) when comparing treatment strategies.
Costs are reported in 2018 dollars, and health benefits were monetized using a US$100,000 per QALY threshold.
Sensitivity Analysis
We performed sensitivity analyses assessing the length of follow-up by simulating results for 10 and 30 years as well as our base case of 20 years. Sensitivity analyses were also performed with respect to disease progression and costs. We limited our sensitivity analyses to treatment strategies which were not dominated by some other strategy. One-way sensitivity analyses for the probability of stroke and for costs were performed by increasing and decreasing the simulation parameters by one standard error based on the literature
When is carotid angioplasty and stenting the cost-effective alternative for revascularization of symptomatic carotid stenosis? A Canadian health system perspective. [Internet].
AJNR. Am. J. Neuroradiol.2014; 35 (Available from): 327-332
Health utility scores for people with type 2 diabetes in U.S. managed care health plans: Results from translating research into action for diabetes (TRIAD).
as displayed in Table 3. Multiple estimates were available for the probability of incident stroke without carotid revascularization treatment and its cost . We used the range of reported costs divided by 4 to estimate the standard error
. In addition, we performed a probabilistic sensitivity analysis that varied the parameters related to the probability of stroke, QALYs, and costs simultaneously using the same variances. Each sensitivity analysis was replicated 100 times to estimate the cost-effectiveness of each treatment strategy.
Table 3Standard deviations used for probabilistic sensitivity analysis
Supplemental Table 2 compares the demographic characteristics of our simulated population to those of the CREST cohort. Individuals in our simulation population were similar to the CREST population for measures of age, sex, smoking status, hypertension status, systolic blood pressure, and diastolic blood pressure. In the CREST study, 93% of participants were white. In contrast, due to our sampling design from NHANES, only 70% of individuals in our simulation population were white.
Table 4 summarizes the ICERs for all pairwise comparisons of the 6 treatment-scenarios over 20 years. Compared to all other strategies, historical medical management plus CEA was either cost-saving or cost-effective at a threshold of $100,000 per QALY gained. Modern medical therapy alone had the highest total cost per person due to the increased incidence of stroke and its associated cost.
Table 4Incremental cost-effectiveness ratios for comparisons between 6 scenarios
Base case: 20 yearsLower total cost Higher total cost
MM alone
MM+CAS
Cost-saving
MM+CEA
Cost-saving
Cost-saving
HM alone
Not cost-effective
Not cost-effective
Not cost-effective
HM+CAS
Cost-saving
$198,123
$224,653
Cost-saving
HM+CEA
Cost-saving
Cost-saving
$239,093
$267,043
Cost-saving
Sensitivity analysis: 10 years
MM alone
MM+CAS
Cost-saving
MM+CEA
Cost-saving
Cost-saving
HM alone
Not cost-effective
Not cost-effective
Not cost-effective
HM+CAS
Cost-saving
$288,525
$358,787
Cost-saving
HM+CEA
Cost-saving
Cost-saving
$408,438
$489,104
Cost-saving
Sensitivity analysis: 30 years
MM alone
MM+CAS
$3,585
MM+CEA
Cost-saving
$970
HM alone
Not cost-effective
Not cost-effective
Not cost-effective
HM+CAS
Cost-saving
$161,111
$179,749
Cost-saving
HM+CEA
Cost-saving
Cost-saving
$188,567
$207,361
Cost-saving
Abbreviations: HM, historical medical management; CAS, carotid artery stenting; CEA, carotid endarterectomy; MM, modern medical management.
Notes: 1) Strategies are reported by the order of total cost in the base case analysis. Incremental cost-effectiveness ratios are reported as the row-defining strategy compared with the column-defining strategy. 2) The white cells indicate comparisons within a setting. The light grey cells indicate comparisons between the two settings for a given treatment modality, while the dark grey cells indicate comparisons between different settings and treatment modalities.
Considering the clinical scenario where carotid revascularization procedures are contra-indicated, the ICER for modern medical management alone versus historical medical management alone was $32,037 per QALY gained, which is cost-effective at an acceptability threshold of $100,000 per QALY gained. Considering clinical instances where individuals are already receiving modern medical management, adding CEA was cost-saving compared to either adding CAS or medical management alone. In particular, the QALYs associated with modern medical management plus CEA were nearly identical to modern medical management plus CAS. This near equality in QALYs was driven by the complication rate in the simulated outcomes. For 20-year outcomes, the simulated incidence of ischemic stroke was 0.52% and 0.31% respectively in modern medical management plus CEA and modern medical management plus CAS. However, the rates of myocardial infarction (0.75% and 0.74% per person-year, respectively) and coronary heart failure (1.8% and 1.8% per person-year, respectively) were substantially higher than the rates of ischemic stroke and drove the QALYs for these two scenarios. Thus, the combination of lower cost of CEA and very similar QALYs resulted in ICERs of extremely large magnitude (Table 5).
Table 5Simulated costs and QALYs per person for 6 treatment scenarios
In sensitivity analyses, we compared historical medical management plus CEA and modern medical management plus CEA, because these strategies were the only ones which were not strictly dominated by some other strategy. In the one-way sensitivity analyses, the ICERs ranged from $51,442 to $303,893 per QALY gained. Only in scenarios where the stroke incidence was at its lower limit, and where semaglutide had 4 generic competitors, did the ICER fall below a $100,000 per QALY gained threshold (ICERS of $51,442 and $82,932 respectively). In probabilistic sensitivity analysis, at thresholds of $50,000 and $100,000 per QALY gained, historical medical management plus CEA was considered cost-effective compared to modern medical management plus CEA in 6.3% and 32.0% of model iterations from a health care sector perspective, respectively. Figure 1 shows a plot of the outcomes for the sensitivity analysis.
Figure 1caption: Probabilistic Sensitivity Analysis between Historical Medical Management plus CEA and Modern Medical management plus CEA.
We simulated the long-term health and economic outcomes of carotid revascularization procedures following incident TIA among people with type 2 diabetes and compared two types of medical management: historical medical management with aspirin, sulphonylurea, atenolol, and moderate-intensity statin therapy versus modern medical management with clopidogrel, semaglutide, losartan, and high-intensity statin therapy. We found that, in scenarios with both historical and medical management, CEA is either cost-saving or cost-effective at a $100,000 acceptability threshold per QALY gained, compared to other strategies. For a patient who has already undergone CEA, changing from historical to modern medical management is not cost-effective (ICER > 100,000 per QALY). However, modern medical management is cost-effective compared to historical medical management without carotid revascularization procedures at a $100,000 acceptability threshold. Therefore, our results suggest that CEA is the most important element of management for carotid revascularization candidates with T2DM following a TIA.
The cost-effectiveness of CEA in the historical setting was driven by a reduction in down-stream complications combined with the low cost of treatment. Not only did the total cost of care decrease with CEA, but the QALYs increased compared to other treatments within a historical context. This is consistent with the decreased risk of subsequent stroke observed with CEA in CREST.
Over the course of a 20-year simulation, a small reduction in stroke rates yields lower complication costs and greater health related quality-of-life.
The literature has reported mixed results when estimating the cost-effectiveness of CAS versus CEA. The SAPPHIRE trial found that CAS was cost-effective relative to CEA (ICER= $67,891 per QALY for CAS vs CEA).
In studies that recruited subjects with lower CVD risk, CAS typically was not cost-effective. For example, one 4-year simulation study based on the CREST population, in which 40% of subjects had histories of cardiovascular disease, reported an ICER of >$200,000 per QALY for CAS vs CEA.
Costs and cost-effectiveness of carotid stenting versus endarterectomy for patients at standard surgical risk: Results from the Carotid Revascularization Endarterectomy Versus Stenting Trial (CREST).
Costs and cost-effectiveness of carotid stenting versus endarterectomy for patients at standard surgical risk: Results from the Carotid Revascularization Endarterectomy Versus Stenting Trial (CREST).
We look forward to comparing our results for modern medical management to the CREST-2 results.
When we compared historical medical management versus modern medical management, the difference in ICERs was driven primarily by medication costs. Over the course of 20 simulated years, the medication costs under modern medical management were almost $60,000 greater than for historical medical management. These additional costs translated into a total cost difference of less than $5,500 for each of the treatment modalities due to a reduction in complications. Therefore, for individuals who are not carotid revascularization candidates, modern medical management is cost-effective. Moreover, the primary cost of modern medical management was semaglutide. Our sensitivity analysis suggested that if there were 4 generic options for semaglutide on the market, modern medical management plus CEA would be cost-effective relative to historical medical management plus CEA.
Our choice of modern medical management was based on guidelines for diabetes management, which may have a direct or indirect effect on stroke risk. For example, although there is some evidence that clopidogrel is only slightly superior to aspirin for stroke prevention, evidence from the CHANCE and POINT trials suggests that dual antiplatelet therapy with aspirin and clopidogrel for 21-30 days after high-risk TIA reduces stroke risk.
Our model accounts only for treatment effects over each year, not the short-term treatment effects of dual antiplatelet therapy with aspirin and clopidogrel, suggesting that our estimates of the cost-effectiveness of CEA plus modern medical therapy were conservative.
Limitations of our study include the assumptions inherent in any simulation model. We have managed this limitation by using an externally validated model and by performing sensitivity analyses. Although our results assume that prices do not change and new treatments do not come onto the market during the simulation period, our simulations provide an estimate of the long-term benefits of these competing treatments without additional intervention. As prices change and modern medical therapy improves over time, further comparisons can be considered. In addition, our sensitivity analyses suggested that our results are robust to variability in the input parameters. We found that adding Proprotein convertase subtilisin/kexin type 9 serine protease (PCSK9) inhibitors to the modern medical management of dyslipidemia was not cost-effective at a $100,000 acceptability threshold compared to modern medical management without PCSK9 inhibitors due to their high cost (data not shown). Another limitation is that we simulated a racially diverse population using treatment effects that were derived from a study performed in a largely white population. Thus, our results represent the economic outcomes assuming that racial disparities are not present. When data become available to estimate the effect of racial disparities, we can update our model to reflect those disparities and to estimate the benefit of eliminating racial disparities in the treatment of TIA.
In conclusion, we found that the effect of CEA dominates the effect of modern medical therapy in people with diabetes and incident TIA. However, if revascularization is contra-indicated, modern medical management is cost-effective compared to historical medical management alone.
Potential Conflicts of Interests
DM: None
SK: None
WY: None
DL: None.
WH: Dr. Herman is a member of a Data Safety Monitoring Board for Merck Sharp & Dohme
Author role
DI: acquisition and analysis of data, drafting a significant portion of manuscript and tables
SK: drafting a significant portion of manuscript and tables
WY: critical review of manuscript
DL: significant clinical review
WH: conception and design of the study, significant clinical review
Acknowledgements
Funded by the National Institutes of Health NIA R21 AG-060277 and NIDDK P30 DK92926
Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. 2014.
Guidelines for the primary prevention of stroke: A statement for healthcare professionals from the American heart association/American stroke association.
When is carotid angioplasty and stenting the cost-effective alternative for revascularization of symptomatic carotid stenosis? A Canadian health system perspective. [Internet].
AJNR. Am. J. Neuroradiol.2014; 35 (Available from): 327-332
Costs and cost-effectiveness of carotid stenting versus endarterectomy for patients at standard surgical risk: Results from the Carotid Revascularization Endarterectomy Versus Stenting Trial (CREST).
Cardiovascular morbidity and mortality in patients with diabetes in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): A randomised trial against atenolol.
Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): Multicentre randomised placebo-controlled trial.
Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 Explanation and Elaboration: A Report of the ISPOR CHEERS II Good Practices Task Force [Internet].
Health utility scores for people with type 2 diabetes in U.S. managed care health plans: Results from translating research into action for diabetes (TRIAD).
00600-0&id=ga1.jpg){kind=link}
00600-0&id=gr1.jpg){kind=link}