Percutaneous Transluminal Angioplasty (PTA) of the Carotid Artery Concurrent with Stenting

Table 1 below provides a timeline of carotid stent device approvals to date.

Table 1 – FDA approved carotid stent devices

VI. General Methodological Principles

When making national coverage determinations (NCDs) under section 1862(a)(1)(A) of the Social Security Act, CMS generally evaluates relevant clinical evidence to determine whether or not the evidence is of sufficient quality to support a finding that an item or service falling within a benefit category is reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member.  The critical appraisal of the evidence enables us to determine to what degree we are confident that: 1) the specific assessment questions can be answered conclusively; and 2) the intervention will improve health outcomes for beneficiaries.  An improved health outcome is one of several considerations in determining whether an item or service is reasonable and necessary.

A detailed account of the methodological principles of study design that the Agency utilizes to assess the relevant literature on a therapeutic or diagnostic item or service for specific conditions can be found in Appendix A.

Public comments sometimes cite published clinical evidence and give CMS useful information. Public comments that give information on unpublished evidence such as the results of individual practitioners or patients are less rigorous and therefore less useful for making a coverage determination.  Public comments that contain personal health information that cannot be redacted will not be made available to the public.  CMS responds in detail to the public comments on a proposed national coverage determination when issuing the final national coverage determination.

VII. Evidence

A. Introduction

This section provides a summary of the evidence we considered during our review.  The evidence reviewed to date includes the published medical literature on pertinent clinical trials of carotid artery stenting compared to carotid endarterectomy.  Our assessment focuses on the key evidence questions below.

B. Discussion of Evidence

1. Evidence Questions

Q1.  Does carotid artery stenting with embolic protection improve health outcomes similarly to carotid endarterectomy in symptomatic patients with carotid artery stenosis?

Q2.  Does carotid artery stenting with embolic protection improve health outcomes similarly to carotid endarterectomy in asymptomatic patients with carotid artery stenosis?

Q3.  Do specific patient characteristics impact which procedure, CAS or CEA, results in better health outcomes for individual patients?

Q4.  Are NCD criteria for physicians, care teams, and facilities performing CAS necessary or helpful to ensure that beneficial health outcomes seen in trials are achieved in broad community practice?

2. External Technology Assessments

CMS did not request an external technology assessment (TA) on this issue.

3. Internal Technology Assessment

We searched the databases PubMed and Embase for English language articles in peer-reviewed journals published since the last NCD, from 2009-2023, using the search terms ‘carotid artery stenting’, ‘carotid endarterectomy’, ‘transcarotid artery revascularization’ and “healthcare disparities” in various combinations. To ensure that we captured all the relevant articles, the search was conducted independently by the contractor International Consulting Associates (ICA), and the CMS Coverage and Analysis Group (CAG).  We incorporated all the distinct, relevant references into a single reference database.  We included other, relevant articles and documents cited by those in the reference database, or cited by the requester and public commenters, as well as expert opinion or commentary from online sources.  We also searched for the most recent guidelines by medical societies, based in the U.S. and abroad, as well as relevant government reports (e.g., by the FDA or CDC).  The final result was the identification of over 70 peer-reviewed documents relevant to the NCD analysis.

4. Medicare Evidence Development & Coverage Advisory Committee (MEDCAC)

A MEDCAC meeting was not convened on this issue.

5. Evidence Table

VIII. Public Comment

Public comments sometimes cite the published clinical evidence and give CMS useful information. Public comments that give information on unpublished evidence such as the results of individual practitioners or patients are less rigorous and therefore less useful for making a coverage determination.

CMS uses the initial public comments to inform its proposed decision. CMS responds in detail to the public comments on a proposed decision when issuing the final decision memorandum. All comments that were submitted without personal health information may be viewed in their entirety by using the following link: https://www.cms.gov/medicare-coverage-database/view/ncacal-public-comments.aspx?ncaid=311.

Initial Comment Period: 1/12/2023-2/11/2023

During the 30-day public comment period following the release of the tracking sheet, CMS received 193 timely comments.  All comments were published on the CMS website and considered for this proposed decision.  The majority of comments, 146 comments, supported the expansion of coverage criteria in one or more areas.  Of these 146 comments, 75 comments specifically supported removal of the requirement that patients are at high risk for CEA, 21 comments specifically supported removal of the operator and facility standards and approval requirements, and 3 comments supported expansion of coverage only for TCAR and not transfemoral CAS.  Twenty-three comments did not support expansion.  Of these 23 comments, 2 comments did not support removal of the requirement that patients are at high risk for CEA, and 4 comments did not support the removal of the operator and facility standards and approval requirements.  Three comments supported expansion of coverage criteria in some areas and non-expansion in others.  Twenty-one comments did not state a clear position regarding coverage criteria. Of these 21 comments, 10 comments were generally positive toward transfemoral CAS and 3 comments were generally positive toward TCAR.  One comment supported independent facility accreditation for all carotid stenting procedures and operator guidelines provided by professional societies.

One hundred fifty-six comments were submitted by healthcare professionals, with the majority (at least 145 comments) provided by physicians.  Seven comments were provided by medical device companies.  Seventeen commenters did not specify their titles and/or organizations.  Thirteen comments were provided by national associations, professional societies, commissions, foundations, including the Intersocietal Accreditation Commission (IAC), the Multispecialty Carotid Alliance (MSCA), the Society for Cardiovascular Angiography and Interventions (SCAI), the Society of Interventional Radiology (SIR), the Society of NeuroInterventional Surgery (SNIS), the Society of Vascular and Interventional Neurology (SVIN), the Society for Vascular Medicine (SVM), the Society for Vascular Surgery (SVS), the Society for Vascular Surgery (SVS) Patient Safety Organization (PSO) Vascular Quality Initiative (VQI), the Vascular InterVentional Advances (VIVA) Foundation, a joint comment from the American College of Cardiology (ACC) and the American Heart Association (AHA), and a joint comment from the American Association of Neurological Surgeons (AANS), the Congress of Neurological Surgeons (CNS), and the AANS/CNS Joint Section on Cerebrovascular Neurosurgery.

Numerous commenters provided references for our deliberation of this NCA.  We very much appreciate this information.  All such references were assessed for inclusion in our evidence review.

Second Comment Period: 7/11/2023-8/10/2023

During the second 30-day comment period, after the posting of the proposed decision memorandum, CMS received 760 timely comments, including one comment with 88 cosigners from the Global Expert Collaboration.  Comments were closely split on support and opposition to the proposed decision memorandum, with 381 comments expressing at least some degree of support for expanded coverage of CAS and 358 expressing opposition to expanded coverage.  The remaining 21 comments did not express support or opposition and included three comments adding cosigners to the Global Expert Collaboration comment letter, other comments asking questions or for clarification of the proposed policy language and unintelligible comments (for example, one commenter appeared to paste a file location in the comment submission tool instead of the content of that file).

Of 760 timely comments, 672 comments were provided by health care professionals, with most of these comments (at least 644 comments) provided by physicians.  Ten comments were provided by academia.  Six comments were provided by the Global Expert Collaboration.  Two comments were provided by consumer representatives.  Six comments were provided by medical device companies.  Three comments were provided by trade associations.  One comment was provided by a think tank.  Four comments were provided by health care systems/organizations, including Cleveland Clinic, the OhioHealth Vascular Institute (OHVI), the University of Connecticut Healthcare System, and UnityPoint Health.  Fifteen comments were provided by national associations/professional societies/foundations, including the American Academy of Neurology (AAN), the Multispecialty Carotid Alliance (MSCA), the Neurocritical Care Society (NCS), the Neurohospitalist Society (NHS), the Society for Cardiovascular Angiography and Interventions (SCAI), the Society of Interventional Radiology (SIR), the Society of NeuroInterventional Surgery (SNIS), the Society of Vascular and Interventional Neurology (SVIN), the Society for Vascular Medicine (SVM), the Vascular and Endovascular Surgery Society (VESS), the Vascular InterVentional Advances (VIVA) Foundation, a joint comment from the Society for Vascular Surgery (SVS) and the SVS Patient Safety Organization (PSO) Vascular Quality Initiative (VQI), a joint comment from the American College of Cardiology (ACC) and the American Heart Association (AHA), and a joint comment from the American Association of Neurological Surgeons (AANS), the Congress of Neurological Surgeons (CNS), and the AANS/CNS Joint Cerebrovascular Section.  Forty-one comments did not specify a title and/or organization.

Numerous commenters submitted multiple comments.  For the purpose of summarizing and responding to the comments in a timely manner given the large volume received, each comment submitted is counted as a single comment and not combined with other comments submitted by the same commenter.  The largest number of comments submitted by a single commenter, was seven.  Some individual commenters were also cosigners on comment letters submitted by organizations or larger groups of individuals.  Additionally, multiple form letters were submitted reflecting different opinions on the proposals.  Many commenters appeared to modify some of the form letter content or the suggested language provided by larger organizations to both add and remove content.  As such, distinct language varied across comments, even when conveying similar sentiments.  Despite the large volume of comments, specific themes stood out across the comments which we have summarized and responded to below.

Population

Comment
CMS received over 250 comments specifically addressing the appropriate patient population for expanded coverage.  Of these, about 75 express specific support for coverage in patients at standard surgical risk, about 140 express specific support for coverage in patients with symptomatic carotid artery stenosis of 50% or more and about 145 express specific support for coverage in patients with asymptomatic carotid artery stenosis of 70% or more. Several commenters express general support for expanded coverage in asymptomatic and symptomatic patients and others recommend coverage in asymptomatic patients with 80% or more stenosis.

Commenters that oppose expanded coverage largely offer no populations in which they believe coverage is appropriate, albeit in doing so, they focus on transfemoral CAS (TF-CAS) while separately categorizing and generally supporting coverage for TCAR. However, several of these commenters note that there may be patients who benefit from TF-CAS that are symptomatic, at high risk and cannot undergo CEA or TCAR. We summarize and respond to the points made by commenters in opposition to the proposed decision memorandum throughout the sections below.

Response
For the reasons detailed in the analysis section of the proposed and this final decision memorandum, as well as in our responses to comments below, we are finalizing coverage for patients with symptomatic carotid artery stenosis ≥ 50% and asymptomatic carotid artery stenosis ≥ 70% without restrictions based on risk for CEA.  As discussed throughout this decision memorandum, a combination of contemporary RCTs, updated meta-analyses, and large registry data all support that CAS is a reasonable and sometimes preferable option compared to CEA for many patients, depending on their individual anatomy, pathophysiology, and preferences, all of which we believe should be considered by patients in shared decision-making with their practitioner(s).  This is the basis for expanded coverage and related criteria provided in the NCD.

Comment
Several commenters request that CMS specify in the NCD that coverage is for an FDA-approved indication with one recommending language stating that coverage is for carotid stents “used in accordance with the FDA approved or cleared labeling and indications for use.”  This commenter further requests that CMS include the FDA indications and labeling information in the NCD.  Another commenter requests that CMS not limit use to “on-label” as the treating physician is best suited to determine which device is best for each patient.

Response
We believe that the coverage parameters specified in the NCD language, which do not include reference to FDA labeling or indications, accurately reflect what the clinical evidence supports to be reasonable and necessary under §1862(a)(1)(A) of the Act.  Furthermore, we are allowing Medicare Administrative Contractors (MACs) to make reasonable and necessary determinations under section 1862(a)(1)(A) for any other beneficiary seeking coverage for PTA of the carotid artery concurrent with stenting in addition to coverage specified in the NCD.  We believe there may be scenarios where “off-label” use may be reasonable and necessary to ensure optimal treatment for beneficiaries.

Imaging

Comment
CMS received about 100 comments addressing the imaging related proposals in the proposed decision memorandum.  Key points of criticism include:

• Many believe there may be instances when digital subtraction (catheter) angiography is needed, including when computed tomography angiography (CTA) or magnetic resonance angiography (MRA) are contraindicated (e.g., for patients at risk for contrast nephropathy).  Thus also, use of CTA or MRA should not be an absolute requirement.
• Some support using both duplex ultrasound (US) and cross-sectional imaging (either CTA or MRA), and not just cross-sectional imaging alone, if there are no contraindications.
• Some believe that catheter angiography remains the gold standard, that other imaging modalities should thus not be required, and that it is useful to perform diagnostic angiography while the patient is on the procedure table (allowing the physician to both diagnose carotid stenosis and do the intervention at the same time). 
• Some also express concern about radiation exposure with CTA, which they say is unneeded anyway because some combination of duplex US and catheter angiography provides a better diagnosis.  As such, they argue that CTA would add time, expense and radiation exposure, while MRA is even less accurate; neither CTA or MRA add value.

Response
We agree that intra-arterial digital subtraction (catheter) angiography may still be necessary, not only if there is a significant discrepancy between non-invasive imaging results, but also if patients have contraindications to CTA or MRA (including risk factors for contrast-induced nephropathy).  Exceptions and flexibility for patients with contraindications is basic to all of imaging – but we agree this needs to be explicitly stated; therefore, we have added relevant language to our NCD criteria.  We note that conventional catheter angiography uses contrast and causes radiation exposure as well, but effective doses may be less, and are more in the control of the operator.

We also agree that a combination of duplex US and either CTA or MRA are needed for greater accuracy and to evaluate other relevant arterial anatomy in addition to the carotid stenosis.  This again applies only when there are no contraindications to CTA or MRA.  As such, we disagree that a combination of duplex US and catheter angiography provide a better default diagnostic work up; rather, we believe catheter angiography should be used only when CTA or MRA are contraindicated.  The limited radiation from CTA may be acceptable given that catheter angiography also involves radiation and has association with stroke.  We have updated the NCD accordingly.

This combination of non-invasive imaging is consistent with ACST-2 trial inclusion criteria (the latest and largest ever CAS v CEA trial), and with the most recent guidelines (Naylor 2022) that considered ACST-2 results.  For ACST-2, duplex US was the primary imaging modality, and to be eligible for the trial patients also needed “CT or MRI confirmation of suitability for CAS and for CEA (which would also have been used to exclude from trial entry any patient without sufficient stenosis to justify intervention)” (Halliday 2021).  The guidelines in turn recommend: “For a patient where carotid artery stenting is being considered, it is recommended that any duplex ultrasound study be followed by computed tomographic angiography or magnetic resonance angiography, which will provide additional information on the aortic arch, as well as the extra- and intracranial circulation” (Naylor 2022).  Again, we agree there needs to be an explicit exception for contraindications.

As for special circumstances where more information may be needed by digital subtraction (catheter) angiography, which could not be obtained by two non-invasive imaging modalities, we note that the MACs are structured to consider such case-by-case situations when the need arises.

However, we remain firm that digital subtraction (catheter) angiography should be the exception, not the rule, for diagnostic imaging to assess patients for possible carotid procedures.  Here we are distinguishing between, on the one hand, diagnostic imaging used for assessing stenosis, other plaques and their characteristics, and general arterial anatomy – all of which should inform shared decision-making (SDM) before the patient is on a procedure table – and on the other hand, use of minimal angiography during an interventional procedure for purposes of stent placement.  Rarely, there may be instances when the latter leads to recognition that the non-invasive imaging (required by this NCD) that got the patient onto the procedure table was clearly incorrect.  The expectation is that the procedure would then not be performed.

We noted in our NCD analysis that digital subtraction (catheter) angiography itself is associated with stroke; non-invasive imaging is not, and has seen significant technological advances, with more in the pipeline.  When a patient is receiving their diagnostic imaging test on a procedure table (with catheter angiography), there is increased incentive and pressure to just do the procedure.  There is also less ability to then decide that CEA (or OMT) may, after all, be more appropriate.  Indeed, this would undercut the whole point of SDM, at the outset, which should involve discussion of the patient’s anatomy and other factors, and ideally would involve the patient’s primary care physician as well as specialists.  A scenario where it would be helpful to do diagnostic imaging, combined with possible intervention, with the decision making happening while the patient is on the procedure table, is an acute setting; however, this NCD applies to chronic carotid artery stenosis only, not emergent care.

Finally, while our conclusions rely on our independent analysis based on our review of the literature, along with consideration of public comments, we include the below from the 2023 ESVS guidelines (Naylor 2022) for the public to see how another entity has independently evaluated this topic, in their own words:

“2.5. Imaging strategies in carotid artery disease
During ECST and NASCET, all participants underwent intraarterial angiography. This policy has now been abandoned because of angiogram related stroke. In the Asymptomatic Carotid Atherosclerosis Study (ACAS), 30 day death/stroke after CEA was 2.3%, but half of the peri-operative strokes were angiogram related.¹⁹⁵ Colour DUS [Duplex US] is the first line imaging modality due to low cost and accessibility and there are consensus criteria for diagnosing stenosis severity.^196-198  Alternatives include CTA or MRA which can simultaneously image the aortic arch, supra-aortic trunks, carotid bifurcation, distal ICA and intracranial circulation, which is important if CAS is being considered. Contrast enhanced MRA (CEMRA) has higher accuracy than non-contrast MRA (time of flight) but requires paramagnetic contrast agents (gadolinium). In a Health Technology Assessment meta-analysis of 41 nonrandomised studies, DUS, MRA and CTA were equivalent in detecting significant stenoses,¹⁹⁹ but it was advised that centres relying on DUS before CEA should perform a second DUS, preferably by a second operator.¹⁹⁹ A combination of two imaging modalities (DUS + CTA or DUS + MRA) improves accuracy and is routine practice in many centres²⁰⁰” (Naylor 2022).

Neurological Assessment

Comment
CMS received about 20 comments regarding the proposed requirement that an independent neurological assessment be performed before and after CAS.  Of these comments, one expresses support for the proposed requirement. Fourteen commenters disagree with this proposal.  Of these, one states that there is no need for an independent neurologist unless there is a major event. Ten commenters disagree with an independent assessment because a neurological assessment is well within the scope of practice for the physicians involved in the intervention. One commenter opposes because not all facilities have a nearby neurologist to fulfill the independent neurologic assessment requirement and another commenter asserts that asymptomatic patients typically have not seen a neurologist so adding a mandated visit before and after CAS will increase cost, delay care, and add burden to patients.

Four commenters acknowledge that independent neurological examinations before and after treatment are an optimal method of assessment but ask for clarification regarding how CMS defines who performs the test, for example, a board-certified neurologist, an associate partner not treating the patient, a physician of another specialty or a certified nurse practitioner stroke specialist?  Another commenter asks if CMS will stipulate that the independent neurological assessments may be conducted in the normal course of patient care without requiring a unique visit.

One commenter contends that “independent” is vague and problematic when used in policies and recommends including language to specify that “a neurologic assessment” be performed by an NIH stroke scale trained health professional, including a PA, NP or CNS which is current practice for CEA.

Response
Based on public comments, we recognize that our proposed language was too vague, and that there are certain real-world clinical scenarios where a truly “independent” neurological assessment would be logistically problematic.  We also note that there is no requirement for an independent neurological assessment before and after CEA, and that in the large German registry (discussed in our Analysis) this was not mandated, and in an early analysis of that registry only about half of the patients received a neurological assessment (Theiss 2004).

Nonetheless, such assessments before and after the procedure are recommended by the latest guideline (Naylor 2022) and have variably been a requirement in trials.  A key reason CMS is taking a step back from our previous, more intensive oversight of facilities and physicians is the belief that there is adequate infrastructure and oversight in place to identify proceduralists who are not meeting acceptable outcomes and take appropriate action.  What constitutes acceptable outcomes is evolving as technology, techniques, training, and patient selection all improve, and these outcomes are explicitly stated in guidelines.  We believe evaluation of neurological outcomes is important to the integrity of this system.

To strike this balance, we therefore are requiring that “Neurological assessment by a neurologist or NIH stroke scale (NIHSS) certified health professional before and after carotid artery stenting (CAS) must be performed” for the purpose of assessing procedural harms.  We are thus striking the word “independent” from our proposed criteria and clarifying who may perform the assessment. Neurological assessments may be conducted in the normal course of patient care without requiring a unique visit.

Oversight

Facility Standards and Approval Requirements
Comment
About 110 comments support the proposal to remove the facility standards and approval requirement from the NCD.  Nine commenters assert that establishing facility and operator requirements is more appropriately addressed by the hospital credentialing process and medical society guidelines, with some noting that this is consistent with other well-established procedures.  Three commenters similarly support removal of the facility standards and approval requirements stating that it should be handled instead by the credentialing processes of the hospitals and hospital systems where CAS is performed and two support deferring specific clinician, care team and facility requirements to societal guidelines and decision of local facilities.  Three commenters contend that removing these requirements from the NCD is consistent with the current state of published literature and standard clinical practice. One commenter supports the removal of minimum standards for facilities as it will allow facilities to rely on well-established facility credentialing requirements and quality assurance programs.  Another commenter agrees that physician, care team and facility requirements are not needed since standards adequately exist in societal guidelines and outcomes at local facilities can be audited or reviewed by appropriate entities, like with other well-established procedures.  One commenter supports elimination of these requirements in favor of enforcing hospitals to be accountable for quality oversight and monitoring in accordance with professional society guidelines and third-party accreditation, which is standard for most procedures.  Four commenters assert that it is incumbent on relevant vascular societies to continue to refine clinical practice guidelines/appropriate use criteria for all therapeutic options (medical, CAS, TCAR, and CEA) as additional data becomes available and for hospitals to adhere to this and insist on rigorous credentialing (i.e. requiring formal training) and peer review processes to ensure high standards of quality are met.

About 180 comments oppose removing facility standards and approval.  Of these comments, about 150 assert that continued credentialing is critical to ensure a high degree of patient safety.  Six commenters assert that facilities should be required to meet standards. One commenter states that procedures should only be allowed at certified stroke centers.  One commenter supports maintaining these requirements to ensure CAS is performed in controlled centers by trained staff, noting that facility approval requirements can help deter inexperienced physicians from “dabbling” in CAS raising patient safety concerns.

About 155 comments contend that if coverage is expanded, the final NCD must include a definition for “qualified physician” and demonstrated core competency standards relating to carotid PTA concurrent with stenting.  These commenters assert that CMS work with relevant stakeholders to develop core competency standards.  Seven other commenters support requirements for operators such as specialized training and case requirements.  Some commenters cite the steep learning curve associated with TF-CAS to support the need for credentialing and training requirements.  Other commenters note that outcomes depend on operator expertise so removing standards will remove the brake on centers performing these procedures, increasing stroke and death rates.  One commenter contends that only a qualified provider should be able to perform TF-CAS and another commenter asserts that rigorous credentialing requirements are needed for TF-CAS.  One commenter states that carotid procedures are best performed by providers with training and experience in all aspects of the management of cerebrovascular disease.  Another commenter specifically states that implanters should be trained and perform over 50 carotid angios with intracranial and extracranial interpretations.  One commenter requests that CMS explain what “qualified physician” means.  Nine commenters assert that only vascular surgeons can make the best decision for patients because they offer all three revascularization options.  One commenter contends that only interventional neurologists and vascular surgeons should be allowed to perform procedures.  One commenter recommends requiring a consultation with a surgeon if the procedure is to be performed by an interventionalist specialty.  Two commenters state that vascular surgeons should not be the only specialty allowed, with one noting that cardiologists have the wire skills and vascular surgery only recently started to furnish endovascular procedures as their surgical volume decreased.  Six commenters contend that the treating provider has primary responsibility for care management without requiring patients to see multiple physician specialists with one noting that such a requirement could delay care.  Two other commenters recommend against any requirement for other physicians from different specialties to sign off prior to performing CAS and one asserts that the treating surgeon will do what is best for the patient and not only what he can offer.

Response
We appreciate the thoughtful comments aimed at ensuring safety for Medicare beneficiaries. When we proposed to remove the facility standards and approval requirements, we did not envision that CAS procedures would be performed in facilities and by physicians that did not meet standards and requirements as directed by medical specialty societies and other guidelines as implemented by facilities locally.  Our intention was to move the responsibility and direction from CMS to facilities. We recognize that the proposal appeared to abandon standards.

In response to public comments highlighting significant concerns about procedures being performed in inappropriate facilities and/or by physicians without training or experience, we have decided to maintain standards for facilities to perform CAS procedures and have finalized standards largely consistent with those established in 2005.  Since facilities have over 18 years of experience establishing and maintaining CAS programs as specified in the 2005 NCD, we no longer believe that it is necessary for facilities to undergo an approval process by CMS to perform CAS procedures.  Given the maturity of the technology, we believe facilities are capable and should be fully responsible for ensuring that the CMS standards are met and adhered to as specified.

The modified facility standards retain the basic elements of the 2005 NCD, while removing the facility approval and data submission process and allowing flexibility so facilities may establish customized programs based on up-to-date specialty society guidelines that are tailored to their institutional, community and practitioner goals and needs, ensuring optimal outcomes and addressing any suboptimal results swiftly.  These standards do not specify exact training, experience, volume or other requirements that must be met by individual operators or facilities, but instead, similar to the 2005 requirements, direct facilities to have a dedicated CAS program that institutes such requirements for the facility and its operators.  As such, we are not establishing a definition of “qualified physician” but encourage specialty societies to identify appropriate physician standards, as well as other standards, to assist facilities in developing, establishing and maintaining high-quality programs.

Mandatory Data Collection/Registry Participation
Comment
More than 180 commenters support requiring data collection, registry use (many support mandated data submission for CAS procedures to the Society for Vascular Surgery Vascular Quality Initiative (SVS VQI)), and/or data monitoring.  About 150 commenters recommend that CMS require real time data collection and reporting if coverage is expanded.  Commenters contend that such mandated data collection is important for patient safety, tracking short and long-term outcomes, measuring the impact of reimbursement, confirming/disproving if the randomized trial results are generalizable to real world practice, to inform changes in policy or develop initiatives to address bad outcomes.  Some note that requiring participation in a registry that reports outcomes would deter unskilled and inexperienced operators from performing CAS.  About 110 other commenters support coverage without requiring participation in a clinical trial (as coverage under section B3 of NCD 20.7 for certain TCAR procedures, due to coverage restrictions since 2005 under section B4, requires that procedures are performed in studies including the SVS VQI TransCarotid Revascularization Surveillance Project which involves registry participation).  Several commenters support expanding coverage without requiring additional evidence collection and others contend that CMS should not require data registry participation, with one commenter noting that they add significant cost to hospitals.

Response
In the Analysis section of the proposed decision memorandum, and repeated in our response to comments on trial evidence below, we note that a standardized, nationwide registry – for all carotid artery procedures – would be helpful to monitor procedural safety, further evolve patient risk stratification, and to facilitate auditing and quality improvement, including comparison of local outcomes to national and other benchmarks.  However, for a technology that has been investigated and used widely in various forms, we do not believe mandated registry participation or data submission is necessary for coverage purposes.  Recognizing and supporting the value of registries is separate and distinct from mandating participation for coverage purposes under Medicare.

We acknowledge that some interested parties strongly disagree with use of CAS entirely, most specifically, TF-CAS, unless data is collected in a national registry.  We understand the desire for ongoing research and information gathering to optimize patient treatment options, and we believe, for a technology that has gone through nearly 30 years of clinical investigation with FDA marketing approvals beginning in 2004, that coverage is reasonable and necessary without further mandated evidence generation for optimally selected patients.  CMS believes that there is value in registry participation, albeit not for coverage purposes in this NCD.

Predicted Consequences of Coverage Expansion
Comment
Many commenters opposing the proposed decision to expand coverage and discontinue facility standards and approval requirements with no required data submission assert that such changes in policy will result in inexperienced interventionalists performing CAS (many commenters specify TF-CAS) on inappropriate patients in inadequate facilities resulting in poor outcomes, including significant increases in strokes and high costs to society.  Commenters warn that expanding coverage without any “guardrails” (largely in the form of facility and operator requirements and data submission and tracking) will lead to overuse of CAS in the hands of incapable and unethical operators, who some commenters suggest are motivated by greed, and ultimately, devastating outcomes in patients who commenters contend may not even require treatment.  These comments illustrate a future for CAS under our proposals with unrestricted, unmonitored use that leads to large, widespread increases in both stroke risk and stroke resulting in extreme financial strain to both pay for purportedly inappropriate procedures and treat high rates of poor outcomes.  Numerous commenters describe similar concerns and outcomes in many different ways, but ultimately communicate to CMS that, beyond simply disagreeing with an expansion of coverage, any expansion of coverage without strict oversight is unconscionable.

Some commenters cite the Hippocratic Oath to “First, do no harm” in their comments on both the proposed decision to expand coverage and the operators who they predict will abuse an expansion of coverage in the interest of personal and financial gains.  Some commenters cite news articles about improper overutilization of atherectomy procedures ultimately resulting in limb amputations in patients who were victimized by unscrupulous operators to bolster their claims of inevitable dire outcomes should coverage of CAS be expanded without strict oversight.

Response
We disagree that the proposed expansion of Medicare coverage for a technology that has been investigated for nearly 30 years and first received FDA approval nearly 20 years ago, in 2004, will result in this predicted “unfettered” use.  A multidisciplinary, team-based approach in the treatment of any clinical condition affords patients the best care possible and offers greater opportunity for physicians to ensure that they and their fellow physicians actually do no harm.  There is no basis for the comparison commenters make between CAS and the improper overutilization of atherectomy procedures which occurred following a change in Medicare payment rules (unrelated to an NCD).  Further, there is no indication that CAS would be next. We believe that, again, there is a role for physicians to play in working together to optimize outcomes for all patients.

Comment
Some commenters warn of the negative impact an expansion of coverage for Medicare patients could have beyond the US. Commenters assert that US health policy impacts other countries so inappropriate health policy decisions in the US will negatively influence other countries.

Response
We agree that US decisions may impact other countries.  While our focus is on benefits and harms to our Medicare population, in answering that question, our NCD analysis did present and discusses the latest European guidelines and trials. So that influence, when evidence-based, can and should cross the ocean both ways.  The concern about “inappropriate health policy decisions” is often tied in comments to the assertation that CMS should wait for the results of the CREST-2 trial.  CREST-2 is addressed in a separate response, but note here that the results of CREST-2 will be seen when the trial is completed and results published, which is likely years from now.  This, and similar trials, may or may not impact guidelines and CMS coverage alike, and we cannot infer results based on any interim analysis while CREST-2 is ongoing.

Comment
Several commenters assert that expanding coverage will negatively impact ongoing and future research.  Commenters reference the ongoing CREST-2 and ECST-2 trials, in these statements, which we discuss in the next section on Evidence Analysis.  Commenters contend that an expansion of coverage will end research on carotid artery disease treatment options and weaken evidence-based medicine in other specialties.  Commenters indicate that physicians do not participate in research unless payment is contingent on participation.  Additionally, commenters state that there will also be a negative impact on research globally.

Response
We discuss in more detail below our response regarding comments that urge CMS to wait until the CREST-2 trial is complete and/or the results are published to consider changes to this NCD.  We disagree with this recommendation for the reasons listed below.  Likewise, we disagree with predictions that an expansion in Medicare coverage will compromise completion of ongoing carotid trials, end research for carotid artery disease and compromise evidence-based medicine in other specialties and negatively impact research globally.  Research continues across all areas of medicine independent of payer coverage determinations and we disagree that this single expansion of coverage will negatively affect research on carotid artery disease or the vast landscape of clinical research.  Withholding coverage for services supported by contemporary RCTs solely for the sake of forcing continued research is not in the best interest of Medicare beneficiaries.

Evidence Analysis

Comment
Many comments made at least some reference to the evidence to justify support or opposition to the proposed decision.  Interestingly, many commenters cite the same evidence base in expressing opposing views.  In other words, both comments in support of coverage expansion and opposed to coverage expansion use the same evidence base to support their positions. CMS received about 200 comments indicating that evidence supports expanding coverage and about 130 comments indicating that evidence does not support expanding coverage.  Comments in support citing evidence are largely consistent with our assessment of the evidence base in the Analysis section.  Some commenters, most of whom oppose coverage expansion, include explanations detailing how and why the evidence does not support an expansion in coverage.  These arguments largely involve the following critiques: there is no level 1 evidence to support CAS; evidence demonstrates TF-CAS is not safe or equivalent to CEA/TCAR/medical therapy; studies are underpowered; highly trained operators and strict clinical study parameters cannot be replicated in the real world, thus good outcomes cannot be replicated either; patient populations in clinical studies are not representative of the Medicare population so the same results cannot be expected outside of trials; composite endpoints appear close, but stroke rates alone in the same studies are higher than CEA (and TCAR in registry data); minor strokes are not minor to patients and are not acceptable.

Response
Because comments cite the same evidence base to support opposing opinions on the proposed decision, below, we address the issues raised in comments, often to oppose coverage, related to the evidence base as well as our assessment of it. We summarize key points made by commenters and provide a comprehensive response to these points under each subheading. We encourage readers to read the public comments directly to fully review the statements and conclusions made.

Optimal Medical Therapy
Comment
Multiple commenters express skepticism on whether, given vast improvements in optimal medical therapy (OMT), any procedure is better than contemporary OMT alone. A related concern by commenters is that this proposed decision could prevent these ‘OMT trials’ from going to completion; or at best is premature – and more than 40 commenters ask CMS to wait until those trial results, like CREST-2, are in. Conversely, other commenters request that CMS not wait for the results of CREST-2 as they are years away and one reminds readers that CREST-2 is not powered to compare CEA and CAS.  One commenter notes that even if CREST-2 results show equivalency of surgical and medical therapy, patients would still have access to CEA and should to CAS as well.  Another commenter asserts that CREST-2 results are not relevant to the current decision because CREST-2 seeks to define which patients are appropriate for medical therapy as opposed to revascularization whereas the NCD expands coverage options for patients already deemed appropriate for revascularization.

Response
We believe skepticism on whether, given vast improvements in OMT, any procedure is better than contemporary OMT alone, has merit. It provides the “clinical equipoise” that is the ethical foundations for the two ongoing trials (CREST-2 and ECST-2, discussed in the Analysis section of this decision memorandum), in which OMT is a distinct arm compared to CEA and CAS, respectively. We did not mince words in our proposed decision about our support for these trials. While the commenters and cosigners believe interim analyses appear promising, they can appreciate that the well-established evidence-based approach we adhere to precludes drawing conclusions based on this alone.

We considered the impact that expanded coverage could have on ongoing studies before publishing our proposed decision and we recognized concerns that premature coverage of CAS could hinder further enrollment in CREST-2 (and similar) trials.  In previous years, we did not reopen this NCD in part because of the clear evidence gap for asymptomatic patients, and in part because of concern, among government agencies and CREST-2 trial investigators alike, that premature coverage of CAS could hinder further enrollment in this (and similar) trials. However, years have passed, the enrollment picture has changed, and this consensus has changed with it.  While enrollment is encouraging, trial publications are likely several years away. We considered this fact in 2022, along with the recently published ACST-2 trial (the largest CAS v CEA trial ever, which doubled the existing data on asymptomatic patients). Such RCTs “represent the highest form of evidence” (from another commenter, and consistent with our Appendix A hierarchy of evidence). Since new RCT evidence has been published since the last reconsideration of this policy in 2009, in which we requested more evidence, we agreed with the NCD requestors that it was appropriate to reconsider the NCD in response to their complete and formal request even with the expectation of new related evidence being released several years in the future.

As such, our focus is on the evidence that exists today – from completed trials – and driven by five new papers from large, multicenter RCTs published since the last CMS reconsideration in 2009. We stated in our analysis that: “We recognize that the medical therapy that served as the control group treatment in the early, foundational trials has also vastly improved. Thus, contemporary trials comparing CEA to CAS reviewed in this NCD analysis, with their wide use of modern antithrombotic, antihypertensive, and lipid-lowering drugs, and lifestyle risk-modification, as background medical therapy for both groups, can fairly compare the two procedures. However, these trials do not answer the question of whether any procedure produces better health outcomes than contemporary OMT alone for these patients. This highlights the importance of completing and publishing ongoing trials which compare each procedure to OMT alone (e.g., CREST-2 and ECST-2).”

We agree with commenters who point out that this expansion of coverage will offer more options for patients already deemed appropriate for revascularization.  We do not support CAS in patients who should not be revascularized and this is a critical aspect of both local CAS program standards and SDM interactions.

Meta-analyses (and the case for asymptomatic patients)
Comment
Some commenters state that CAS is inferior to CEA as demonstrated by meta-analyses.

Response
Assertions are made based largely on two meta-analyses (Batchelder 2019 and Muller 2020), selected from many. The more recent of these two meta-analyses had already been updated and we used the updated version in our analysis (the older version being obsolete).  This is specifically the Cochrane Review meta-analysis (some comments refer to it as “Bonati 2020” in line with comment text, but correctly cite as Muller 2020 in endnotes), which Bonati later updated using “all trials of CAS versus CEA found by the literature-searching strategy of the 2020 Cochrane review.”  This updated meta-analysis was reported along with the ACST-2 trial report in the same paper (Halliday 2021, with Bonati as coauthor). While the previous meta-analysis included just over 3,300 asymptomatic patients, the updated version incorporated the latest and largest trial, ACST-2, included over 6,200 patients– and reached a different conclusion.

This updated meta-analysis reported that: “Overall, the ratio (CAS vs CEA) of long-term stroke incidence rates is 1.11 (95% CI 0.91–1.32; p=0.21).  As previous studies have shown successful CEA to be substantially protective,^4,5 this RR of 1.11 (which includes the ACST-2 result) shows that the protective effects of CAS and CEA are similar for at least the first few years” (Halliday 2021, with Bonati as co-author).  The meta-analysis results for asymptomatic patients (ratio of long-term stroke incidence rates: 1.10, 95% CI 0.87–1.39 in 6,266 patients), and for symptomatic patients (1.13, 95% CI 0.89–1.43 in 4,714 patients) were similar to the overall results above, which applied to a total of 10,980 patients.

We state in our NCD analysis, “while meta-analyses of pooled trial data have generally supported the findings of the individual trials, the strongest evidence comes from the most recent trials themselves.”  In general, meta-analyses that lump together trials launched many years (even decades) apart must be treated with caution given differences in patient selection, techniques, training, and background medical therapy.  In this case, this updated and much larger meta-analysis, the ACST-2 trial itself, and a large contemporary registry, all point to a similar conclusion about the use of CAS for asymptomatic patients with carotid stenosis.  Our conclusions in this NCD draw on the totality of this evidence.

The ACST-2 trial for asymptomatic patients with carotid stenosis in turn reported: “Overall, 1% had disabling stroke or death procedurally (15 allocated to CAS and 18 to CEA) and 2% had non-disabling procedural stroke (48 allocated to CAS and 29 to CEA).  Kaplan-Meier estimates of 5-year non-procedural stroke were 2·5% in each group for fatal or disabling stroke, and 5·3% with CAS versus 4·5% with CEA for any stroke (rate ratio [RR] 1·16, 95% CI 0·86–1·57; p=0·33).” Thus, there was no significant difference between CAS and CEA for the primary procedural (30-day) composite outcome or the long-term (5-year) primary outcome.  But there was a statistically significant, but small difference in absolute terms, for procedural non-disabling or minor stroke, which was 2.7% for CAS vs 1.6% for CEA, p=.03 (Table 2, Halliday 2021), thus approximately a 1% absolute difference between groups.

Real World Clinical Practice (and the case for registries)
Comment
Some commenters contend that differences between CAS and CEA, with CAS leading to worse outcomes, will be clearer in real world clinical practice in millions of patients where procedural standards are lower than in trials.

Response
The ACST-2 authors state that registries give a better assessment of procedural harms in diverse, “real world” clinical practice (and we agree), while randomized controlled trials remain the best measure of comparative long-term benefits and risks: “The current hazards of these two procedures may be better estimated by evidence from large, representative, up-to-date registries than by evidence from randomised trials. What trials can achieve, however (and analyses of registries or other health-care databases cannot),⁷ is a reliably unbiased comparison between the long-term protective effects of CAS and CEA” (Halliday 2021). 

“Analyses of registries or routine health-care databases cannot reliably assess moderate differences in long-term stroke rates, as there may well be systematic differences between the types of patients who undergo CAS and CEA that cannot be sufficiently controlled by mathematical modelling or propensity matching” (Halliday 2021).

CMS has long recognized in its NCDs, across medical fields, that clinical trials and registries have different but complementary strengths and weaknesses. In this instance, the ACST-2 authors report that the results from these two different types of research studies are consistent with each other. “In the national German registry, asymptomatic patients undergoing CAS or CEA during 2014–19 had in both cases an in-hospital risk of disabling stroke or death of 0·7%, with median time to discharge of 4–5 days³ (appendix p 9). A risk of 0·7% within 4–5 days suggests a 30-day risk of disabling stroke or death of about 1% for each procedure, which is similar to that in ACST-2.  Both in the German registry data and in ACST-2, CAS was associated with a slightly greater risk than CEA of non-disabling stroke” (Halliday 2021).

The CREST Trial
Comment
Some commenters assert that secondary analysis of 10-year follow-up data from the CREST trial demonstrates that CAS is inferior to CEA for long-term stroke outcomes.

Response
We discuss the CREST trial further in the Analysis section below (which was also included in the proposed decision memorandum).  The main results of CREST were: “there was no significant difference in the rate of the primary composite end point [stroke, myocardial infarction, or death] between the stenting group (11.8%; 95% confidence interval [CI], 9.1 to 14.8) and the endarterectomy group (9.9%; 95% CI, 7.9 to 12.2) over 10 years of follow-up (hazard ratio, 1.10; 95% CI, 0.83 to 1.44).  With respect to the primary long-term end point, postprocedural ipsilateral stroke over the 10-year follow-up occurred in 6.9% (95% CI, 4.4 to 9.7) of the patients in the stenting group and in 5.6% (95% CI, 3.7 to 7.6) of those in the endarterectomy group; the rates did not differ significantly between the groups (hazard ratio, 0.99; 95% CI, 0.64 to 1.52)” (Brott 2016).

Commenters appear to ignore these primary trial findings, highlighting instead the result of one secondary analysis, which was barely statistically significant: “The risk of periprocedural stroke or death and subsequent ipsilateral stroke was 37% higher in the stenting group than in the endarterectomy group (hazard ratio, 1.37; 95% CI, 1.01 to 1.86; P = 0.04)” (Brott 2016).  This difference was driven by periprocedural events, mostly minor strokes.  As we discussed in our analysis, it has been widely acknowledged, in contemporary trials for both symptomatic and asymptomatic patients, and in meta-analyses (some of which include older trials), that there is a pattern of greater incidence in the procedural (30-day) period of minor strokes after CAS, and of non-fatal MIs and cranial nerve palsies after CEA. 

Such secondary analyses, like subgroup analyses, do not invalidate the main outcomes of a trial.  They do, however, provide additional information that patients and their physicians (primary care and specialists) should consider in formal SDM, along with a host of patient-specific risk factors.

Appropriate Trial Outcomes
Comment
Commenters appear to disagree with the short-term (procedural) and long-term primary outcomes that the ACST-2 investigators selected for their trial.

Response
All trials must, at the outset, select and justify one or very few primary outcomes – those which are the most clinically meaningful to patients – to be the measuring stick of whether trial results are “positive.”  We believe the primary outcomes selected for ACST-2 are reasonable and consistent with other trials we have seen (in this field and in others).  As stated in our Analysis: “The primary early outcome was a composite of disabling stroke or death from any cause at 30-days (i.e., procedure-related events). The strength of this composite is that these two harms are important, relevant, and roughly equivalent (indeed many older Americans prefer death to a major, disabling stroke; CMS 2019). The primary late outcome was “long-term (up to 5 or more years) prevention of stroke, particularly disabling or fatal stroke” (Halliday 2021).”

The authors justify separating analysis of early and late outcomes (rather than combining them): “Proportional-hazard methods are not used for analyses that combine procedural hazards with long-term stroke rates.  For early risk from an intervention may be followed by later benefit, so the hazard ratio comparing one treatment versus another could well go first in one direction then in another (invalidating methods that assume approximately constant hazard ratios)” (Halliday 2021).

We note again that some commenters prefer larger composite results (e.g., any stroke, death, or MI) but we caution that results for such a broad composite are often driven by just one component of that composite (such as non-disabling or minor stroke).  The fact that, historically in CAS v CEA trials, non-disabling stroke is more common in the procedural period with CAS, while non-fatal MIs and cranial nerve palsies are more common with CEA, was acknowledged both by the ACST-2 authors and in our analysis (where we consider the totality of evidence, not just one trial).  This discrepancy is, again, part of the reason we are requiring SDM.  Our intent is for patients and their primary care physicians, as well as specialists, to discuss the different risks these procedures appear to entail, tailored to the individual patient.

We do not “discount” non-disabling or minor strokes but acknowledge that not all patient outcomes are equal – and agree (as discussed above) that death or disabling stroke is an important, patient-centered composite health outcome and a legitimate one for ACST-2 investigators to have preselected for their trial.  We discuss non-disabling strokes further below.

Statistical “Powering” of Trials
Comment
Commenters contend that individual randomized trials (like SAPPHIRE, CREST-1 and ACST-2) were underpowered and hence their results are unreliable.  Additionally, some assert, related to powering, that few subpopulations have shown benefit from a carotid artery procedure (and here only CEA).

Response
Underpowering of trials occurs when there are not enough enrolled patients to confidently trust the trial's statistical results.  There are two key aspects of this with respect to ACST-2: first, what the statistical trial results are; and second, how the trial investigators interpreted their results in combination with two other key sources of evidence. 

As context for the first issue, the ACST-2 trial investigators initially sought to enroll 5,000 asymptomatic patients with severe carotid artery stenosis.  Like many trials, this one did not achieve its aspirational goal, and ended up with 3,625 patients (which still makes it the largest as well as the most recent of any CAS v CEA trial).  The change in sample size resulted in protocol amendments.  Note that trial statistics are not ‘wrong’ because the sample size (“N”) was not the one initially sought; rather, they mathematically reflect patient outcomes for the sample size actually achieved.  The ACST-2 trial’s statistical results were:  For the primary 30-day (procedural) outcome, death or disabling stroke, there were 15 (0.9%) for CAS, and 18 (1.0%) for CEA, p=0.77, indicating no significant difference between these groups, in a trial that was the largest CAS v CEA trial to date (N=3,625).  For the primary late outcome, Kaplan-Meier estimates of 5-year non-procedural stroke were 2.5% in each group for fatal or disabling stroke.

Some commenters assert that a larger trial would have detected significant differences between the treatment groups (on more than non-disabling procedural strokes).  We note that such speculation could cut both ways: perhaps the greater number of procedural death or disabling stroke (the most serious outcomes) for CEA (18 patients) compared to CAS (15 patients) seen in ACST-2 would widen further.  Even if, in a larger hypothetical trial, nonsignificant differences on primary outcomes turned significant and favored CEA, it is unclear how individual patients and their physicians in real world clinical care would weigh what might be small absolute differences between overall trial arms, with other legitimate considerations, such as individual patient anatomy.  In any case, this NCD bases coverage policy on trial evidence that exists today, not on speculation of results of hypothetical or ongoing trials. 

The second key issue is the ACST-2 authors’ decision to present and interpret the trial results in combination with two other sources of evidence.  As stated in the Authors’ Reply, “The discussion of [ACST-2] findings drew on two other sources of evidence; first, the procedural hazards seen in large population registries, and second, our meta-analysis of all the properly randomised trials” (Correspondence 2022).  The authors continue: “To compare procedural hazards, large population registries might well be more reliably informative than trials, partly because of sample size considerations and partly because trial participants are highly selected and may well have procedural hazards that differ substantially from those in the general population.

We therefore interpreted the ACST-2 findings on procedural hazards (which are underpowered) from the ACST-2 trial in light of the procedural hazards recorded in the legally mandated nationwide German registry of carotid procedures.  This registry involves vastly larger numbers than any randomised trial and reported similar immediate risks of disabling stroke with CEA and with CAS.  Such registries do not allocate patients randomly between CEA and CAS; therefore, some unknown systematic differences must remain between individuals who undergo these procedures.  For comparing the short-term procedural hazards of CEA with those of CAS, however, the effects of any such differences can be limited by excluding the few patients known to be at particularly high risk (American Society of Anesthesiologists grade 3 or worse), by adjusting for age and sex (although neither materially affected procedural stroke rates in the German registry), and by concentrating on disabling strokes (given that these are particularly important, and are likely to be reported).

By contrast, to help compare the long-term effects of successful CEA and successful CAS on stroke rates, the effects of systematic differences between the types of patient undergoing the two procedures could well moderately bias non-randomised comparisons in ways that cannot be reliably allowed for. Hence, it is necessary to rely on the magic of randomisation rather than the myth of what gets misleadingly described as real-world evidence.² We therefore interpreted the ACST-2 findings on long-term stroke rates (which are also underpowered) in light of our meta-analysis of the ACST-2 findings on long-term stroke incidence and the corresponding findings from all other randomised comparisons of CEA versus CAS, which did not confirm the apparent difference in ipsilateral stroke rates that had been suggested (non-significantly) by ACST-2.  Overall, this meta-analysis¹ showed no material difference between successful CEA and successful CAS in their effects on long-term stroke rates” (Correspondence 2022, Halliday 2021).

Another issue raised in public comments, and related to trial underpowering, is lack of sufficient data for certain subgroups.  We note that trials commonly do not produce definitive results for specific subgroups (as seen in forest plots in trial publications), as they would be impractically large if required to achieve narrow confidence intervals for outcomes for each subgroup that could be of interest (e.g., each medical condition or demographic characteristic).  In that light, numerous RCTs across multiple medical fields have been ‘underpowered’ compared to what we would ideally like to see.

Symptomatic Patients
Comment
Commenters assert that CAS is inferior to CEA for symptomatic patients – considering both individual trials and meta-analysis.

Response
We noted above that a recent meta-analysis (the update in 2021 performed by the senior author, Bonati, of the Cochrane review meta-analysis presented in Muller 2020), demonstrated no significant differences between CAS and CEA in terms of long-term stroke rates for symptomatic as well as asymptomatic patients (Halliday 2021, with Bonati as co-author).  Notwithstanding, we consider publications of contemporary RCTs to be the primary and strongest evidence for both patient categories.  Our 2009 NCD covered symptomatic patients with ≥70% carotid stenosis and high surgical risk, and this 2023 NCD updates that to ≥50% stenosis and includes patients with standard as well as high surgical risk.  As further background, the FDA approved a carotid stent for symptomatic patients with ≥50% stenosis as early as 2004, and then updated that approval to include patients with standard as well as high surgical risk (section V of this decision memorandum).

As stated in our proposed NCD analysis summary (and later discussed in more detail, specifically on subgroup analysis): “Based primarily on the CREST trial (half of which were symptomatic patients) and ICSS trial (symptomatic patients only), the evidence is stronger for patients with carotid artery stenosis ≥70%; however, we believe it is sufficient for patients with ≥50% stenosis as well.  The reason for the relatively weaker evidence for the ≥50% stenosis mark is that while inclusion criteria for both trials was carotid stenosis ≥50%, in fact most patients had stenosis ≥70% (86% for CREST and 89% for ICSS). This may reflect reluctance of physicians to refer patients with moderate stenosis at that time. However, there is evidence supporting CAS for carotid stenosis of 50-69% and, importantly, we believe it is less helpful if CEA were the only procedural option for these patients, regardless of individual patient anatomy and other characteristics.” 

To clarify further, CEA is currently an option for Medicare patients with 50-69% stenosis and there is no NCD specifying coverage criteria for CEA.  Commenters do not request that CMS provide restrictive coverage criteria for CEA, despite also arguing that OMT is better than any procedure (CMS does not have an opinion on the latter, as those trials are years away from publication).  Commenters opposing expanded coverage instead simply argue that there should be no coverage or expanded coverage for CAS.

Finally, we believe there is room for disagreement about which patient may be better suited for which procedure – or no procedure at all – when individual patient anatomy, pathophysiology, procedural concerns (e.g., greater risk of non-fatal MIs and cranial nerve palsies after CEA, versus greater risk of non-disabling stroke after CAS), patient preferences, and other factors are considered, and as risk assessment models are constantly improving.  That is why we proposed formal SDM, which we are now finalizing and discuss further below, although without a requirement to use a specific tool. 

Strokes
Comment
Commenters contend that strokes with a modified Rankin score of < 3 are not necessarily ‘nondisabling’ or minor with respect to a patient’s quality of life and state that strokes that do not lead to death or obvious physical disability should not be discounted to make CEA and CAS look similar.  Commenters assert that appropriate primary outcomes should include all stroke, not just major or disabling stroke.

Response
Minor or non-disabling strokes are defined in the trials according to standardized criteria and are not “discounted” by CMS or by trial investigators.  We agree minor or non-disabling strokes are bad outcomes, as are myocardial infarctions and cranial nerve palsies.  Death and major, disabling stroke however are considered by our Medicare beneficiaries to be worse outcomes than minor stroke, and roughly equivalent to each other, with many beneficiaries considering disabling stroke a worse outcome than death (CMS 2019).

Overall (and as discussed above and in our NCD analysis), there has been greater incidence during the procedural (30-day) period of minor or non-disabling strokes after CAS, and of non-fatal MIs and cranial nerve palsies after CEA. These differences render the question of ‘which procedure is superior’ more complex; and our goal is to have patients and their physicians (both primary care and those who do procedures, whether interventional or surgical) consider these, along with the individual patient’s anatomy, pathophysiology, and preferences, when choosing which procedure – if any – is best for them.

TCAR
Comment
Many commenters, both for and against an expansion of coverage, offer various comments about TCAR.  Some commenters identify concerns with the TCAR evidence base, notably that no RCT has ever been done for TCAR, while other commenters cite real world data from registries to assert that TCAR is superior to TF-CAS.  Some commenters recommend more research on TCAR before an expansion of coverage and others support expanded coverage in TCAR and not TF-CAS.

Response
We acknowledged that despite numerous publications and a large, US national registry for TCAR, that there have been no RCTs comparing TCAR directly to CEA or traditional, TF-CAS, or OMT.  That fact can be raised in SDM, and we anticipate patients will discuss such matters with their primary care physicians as well as specialists.  While we are not mandating use of a specific decision tool, we believe that such tools, developed by experts and based on contemporary evidence, could better inform both patients and their primary care physicians. As explained in section III.A. of this final decision memorandum, the scope of this NCD analysis includes TCAR.  CMS considers TCAR a type of CAS and this NCD applies to all types of CAS regardless of delivery system. As such, coverage requirements for TCAR fall under the final NCD.  We included evidence specific to TCAR in our review and discuss TCAR in the Analysis section below.

Cost
Comment
As noted above, many commenters opposed to the proposed decision warn against expanding coverage as they predict there will be significant increases in costs if coverage is expanded.  Some commenters present estimates for those increased costs due to both increases in the number of TF-CAS procedures that will be performed on Medicare patients as well as costs associated with treatment and care for predicted increases in stroke resulting from more TF-CAS procedures being performed.

Response
While we appreciate commenters sharing this information, we note that CMS does not consider cost or price of services or procedures when making NCDs.

Shared Decision-Making

Comment
CMS received over 360 comments on the SDM proposals.  Most of these commenters note that a validated SDM tool does not exist, but express general support for SDM. Of these comments, about 160 support mandating the use of an SDM tool designed in collaboration with applicable medical specialty societies and/or relevant stakeholders and several more commenters support requiring the use of an SDM tool.  About 120 comments support SDM without requiring use of a validated tool.  About 65 support requiring physicians to have an SDM interaction with patients, but without delaying coverage until a validated SDM tool has been developed.  Another 17 commenters support SDM but recommend that expanded coverage should not be delayed because a validated tool does not exist, with some noting that CMS has not enforced the use of SDM tools under other NCDs when one did not exist.  About 10 commenters oppose requiring the use of an SDM tool.  Three commenters oppose SDM and one commenter notes that SDM would not replace standard informed consent process required by tort law.

Response
CMS agrees with commenters that support using SDM prior to CAS and is finalizing this requirement as proposed, but without the requirement that a validated SDM tool be used during the SDM interaction. Additionally, the required SDM interaction in this final NCD is in addition to and does not replace standard informed consent.

Comment
Several commenters share concerns about SDM noting skepticism about whether patients can trust their doctors to communicate the higher risk for stroke and death from CAS given the confusion and potential conflicts of interest. Another commenter asserts that patients do not want an option that increases stroke risk or to take responsibility for a decision that could result in an unnecessary procedure which could absolve others from accountability and want the physician to choose the option that is best for them.  Two commenters caution that SDM could be a way of making patients take responsibility for treatment that is more likely to harm than help.  Two commenters note that patient consent to treatment is heavily dependent on what information is presented or omitted by their doctors making them highly vulnerable and CMS has no validated SDM tool to ensure patients give properly informed consent.  One commenter stresses that SDM relies on health literacy so if patients do not understand the complexities of the physician provided medical information, there is no SDM.  Some commenters contend that all carotid stenosis treatment options should require an SDM interaction before proceeding (i.e., TCAR, CEA and medical therapy). 

Response
CMS believes that to most patients, informed decision making using SDM that takes patient preferences into consideration is important and of value.  We expect that true SDM involves clear explanations to improve patients’ health literacy so they may make informed decisions with their practitioners, which includes discussion of risks and benefits associated with each treatment option.  It is each practitioner’s responsibility to progress through the SDM process clearly and carefully by engaging in a dialogue with each patient, as opposed to simply providing information without engagement and discussion.

CMS encourages interested parties to work together to develop a tool for use across all specialties to assist practitioners and patients in making the best treatment choice considering each individual patient’s clinical situation.  However, since no such tool currently exists, we have removed this requirement in the final NCD.  Under NCD 20.7, this requirement is specific to all forms of CAS including TCAR.  This reconsideration is limited to CAS, including TCAR, and cannot establish a national, identical SDM requirement for CEA or medical therapy.

Comment
About 30 commenters assert that documentation of an SDM interaction in the medical record is sufficient, many stressing the importance of this being a comprehensive discussion regarding patient specific risks, benefits, and alternatives including revascularization and medical therapy.  One commenter recommends CMS accept medical record documentation that an SDM visit with a physician or NPP occurred and another commenter asserts that an in person patient/physician discussion should suffice to meet an SDM requirement.  One commenter asks what CMS considers sufficient documentation of SDM to be and how will CMS ensure a formal SDM interaction with the beneficiary meets the NCD requirements.  Two commenters ask if CMS requires the use of a validated SDM tool, who will take notice, with one further asking how insurers will realistically manage compliance and payment at the local level.  Another commenter cautions that even if CMS mandates SDM, it may not be adopted by all insurance companies, including Medicare Advantage (MA) plans, which may introduce confusion if not in coordination with other insurance plans.

Response
The NCD requirement that a formal SDM interaction must take place prior to CAS is a binding requirement for Medicare coverage of CAS.  This means that failure to fulfill this requirement as directed in the NCD would result in subsequent CAS procedures being noncompliant with mandatory coverage requirements.  CMS’ Center for Program Integrity has various tools for addressing situations of noncompliance which may result in recoupment of Medicare payments.  When providers submit claims to Medicare for services or procedures it is expected that all Medicare requirements, including but not limited to any applicable NCDs, are met when the service or procedure is furnished.  As such, we are not requiring anything beyond normal  medical record or claims documentation that the required SDM interaction occurred.

NCDs apply to Medicare beneficiaries, including those who have MA plans. MA plans are required to comply with NCDs and update their policies accordingly when NCDs are finalized. NCDs do not apply to other insurance companies so they do not need to include a formal SDM interaction requirement prior to CAS in their policies.

Comment
About 10 commenters express support for the four core SDM elements identified in the proposed decision. One commenter contends that CMS should not specify what to include in the SDM interaction while another commenter states that CMS needs to better define what it means to provide SDM.

Response
We believe it is important for all SDM interactions to include at least the four elements specified in the proposed decision memorandum to optimize treatment for each individual patient. As such we are finalizing these elements for inclusion in the required SDM interaction prior to CAS.

Comment
Six commenters assert that all qualified treating physicians, regardless of specialty, are able to provide patients with current information and personalized advice on treatment options and risks/benefits even if they do not personally offer all options. One commenter requests that CMS clarify who “the physician” is that performs the SDM. Other commenters offer suggestions on who is best to perform SDM. For example, some commenters contend that only physicians who can offer all treatments options can truly offer SDM, while others recommend a team-based approach for adequate discussion of all options with one commenter noting the importance of a multi-specialty discussion in cases with complex anatomic or clinical disease. Others see the role as best filled by a surgeon and non-surgeon, and another commenter supports that vascular surgeons are best to perform SDM. One commenter recommends requiring evaluation and discussion (for both CAS and CEA) with a neurologist trained in vascular neurology but who will not perform the procedure to discuss all options with the patient with no financial incentive. Two commenters caution that requiring patients to be counseled on procedural risks is naïve as it is unlikely non-surgical specialists will present information objectively to patients with another commenter echoing concerns that interventionalists that only perform CAS will present patients with all options. One commenter agrees that surgeons trained in all treatment options will likely not be consulted to provide true informed consent to patients should coverage be expanded.

Response
As stated in the proposed decision Analysis section, “when appropriate, the SDM interaction could offer information and recommendations from experts across specialties (for example from surgeons and non-surgical interventionalists) who perform procedures in accordance with contemporary standards of care as well as independent physicians with expertise in medical management of carotid artery disease. The goal of this type of interaction is to ensure beneficiaries have all relevant information and expert viewpoints to help them make an informed choice.”

However, we did not propose a requirement for multiple physicians from different specialties to be involved in the SDM interaction. We recognize that such a requirement is challenging for many facilities and practices to meet and would likely impart added burden to some beneficiaries in the form of extra appointments, travel, delayed treatment, etc. Likewise, we did not propose to define specifically which practitioner must perform the SDM interaction and we will not add such specificity in this final decision; we thus eschew a strict requirement on which practitioner or specialty must perform the SDM interaction with the patient. We expect this interaction will likely occur with a physician already involved in treating the patient, but do not want to establish this limitation in the event practices develop comprehensive, multispecialty teams of physicians and other health professionals to bolster and enhance SDM interactions with patients to achieve optimal informed consent and optimize individualized treatment for every patient.

We disagree that requiring an SDM interaction is naïve and that non-surgical specialists will not present all information objectively because, as explained above, the SDM interaction, including the four specified elements, is a binding requirement for Medicare coverage.  Additionally, and as noted above, we are not specifying what specialty must perform the SDM interaction so there is no requirement that surgeons trained in all treatment options must be consulted.

Comment
Multiple comments include suggestions or recommendations for validating or developing an SDM tool.  Two commenters direct CMS to a comprehensive review article by Abbott that outlines what information should be conveyed to patients who are offered carotid artery revascularization which they state is a published, open access tool for assisting properly informed consent that is valid with respect to the evidence base and needs to be validated for clinician use.  No other comments identify potential decision-making aids, but some stress the need for and call for the development of a validated SDM tool.  Four commenters contend that SDM is easier and more uniformly accomplished when the treating physician provides all three options like vascular surgeons and note that SDM by interventionalists lacking surgical skills becomes more challenging where local medical facilities do not support easy access to multi-discipline committees or desire to participate in such arenas is lacking.  One commenter states that developing and agreeing on a tool will be a long process but worthy endeavor and another commenter notes that the development of an SDM tool is desirable but it is difficult for experts to agree on the information a tool should convey plus the tool would need to be flexible to accommodate improvements in treatments.  One commenter recommends that CMS encourage clinicians and societies to develop a tool and allow a future validated tool to meet the NCD SDM requirement.  One commenter asserts that professional societies can be tasked with development of an SDM tool and another encourages CMS to develop one.  One commenter contends that CMS should encourage development of a validated tool that includes the four elements in the NCD and state in the final decision that use of a validated SDM tool would provide adequate documentation of the interaction.  One commenter stresses that SDM tools must be vetted and free of bias before being implemented and nine other commenters contend that any tool is subject to unintended bias in its synthesis of the evidence which warrants review and validation by multiple specialty societies representing the physicians who intend to use the tool.  One comment requests clarification on what constitutes an “evidence-based tool” or “validated SDM tool” as it applies to multiple NCDs, future policies, etc.

Response
We are not finalizing a requirement for the SDM interaction to include the use of a validated SDM tool.  As commenters have reminded CMS, one does not exist.  Since we are not finalizing the proposal of requiring the use of a validated SDM tool, and because one does not exist, we cannot offer responses to comments requesting more detail or specificity on the tool as proposed.  We do, however, encourage interested parties to work together, across specialties, to explore development of a tool or decision aid to assist practitioners with accurately communicating critical and unbiased information about all treatment options for carotid artery disease.  We believe this is an endeavor best led by experts in treating carotid artery disease and should neither focus on nor exclude any single specialty.  Better health outcomes, which is the ultimate goal of any treatment plan, are most likely to be achieved when all specialties involved or that could be involved in treatment, come together instead of work to oppose involvement of certain specialties. This is particularly pertinent to the treatment of carotid artery disease as multiple specialties have been involved in treating patients for decades, despite assertions that only one specific specialty is capable.  CMS appreciates that commenters request more guidance and direction overall regarding SDM and tools as discussed in other NCDs and in any potential future NCDs. We cannot provide further guidance and direction at this time, but will take these requests into consideration in future NCD related activities.

Medicare Administrative Contractor (MAC) Discretion

Comment
CMS received about 100 comments in support of allowing the MACs to determine coverage for patients not described in the NCD. No comments opposed this proposal.

Response
CMS is finalizing that in addition to coverage specified in the NCD, MACs may make reasonable and necessary determinations under section 1862(a)(1)(A) for any other beneficiary seeking coverage for PTA of the carotid artery concurrent with stenting.

Comment
Two commenters request clarification that the NCD requirements do not apply to emergent cases for treating acute stroke.  These commenters specify situations including: emergent cases where PTA is used at physician judgement in treating acute stroke; for emergency cerebral embolectomy/thrombectomy for acute stroke when a carotid stent is needed to be able to pass embolectomy/thrombectomy devices into the intra cerebral circulation and to provide an adequate and smooth lumen in the ICA for the patient post intervention.

Response
CMS agrees that NCD 20.7 does not apply to the treatment of acute stroke so coverage is at the discretion of the MACs.

Disparities

Comment
Some commenters align increased access resulting from expanded coverage with the potential to reduce disparities while others disagree.  One commenter contends that data indicates structurally marginalized communities are more likely to present with symptomatic (as opposed to asymptomatic) carotid disease and robust data indicate that individuals in structurally vulnerable communities often live near or receive care in low volume, low quality centers with variable expertise.  This commenter explains that this presents a patient safety concern, particularly for Black and Brown patients who present with stroke and a symptomatic carotid lesion and asserts that all patients deserve the right treatment at the right time by the right team, but the proposals threaten such optimal care, particularly in structurally vulnerable communities.

Response
CMS appreciates the discussion around goals to decrease disparities and improve care and treatment options especially in populations with access to historically lower volume and quality centers.  We believe the expansion of coverage, coupled with maintaining facility standards and requiring an SDM interaction will lead to improved access to treatment options in higher quality programs with operators that meet and are held to locally implemented and monitored standards. 

Access/Options

Comment
About 70 comments address access and treatment options related to the proposed decision. Most of these commenters cite expanded access and availability of CAS or the importance of patients having options and the ability to choose the best treatment option.  One commenter notes that self reported data indicate TCAR availability is limited where about 25% of hospitals that treat carotid disease have access to TCAR; five states do not have any hospitals offering TCAR and 21 states have three or fewer hospitals that offer TCAR which are small numbers compared to the 1,318 hospitals in the country that have billed for carotid treatment.  One commenter contends that there is no unmet need for TF-CAS when CEA and TCAR are available and another commenter asserts that there are no issues with access.

Response
We agree with commenters that expanded coverage will improve access and options for patients and believe evidence supports affording patients the opportunity to consider all treatment options with their practitioners.

Conflict of Interest

Comment
Some commenters share concerns regarding conflicts of interest.  These involve conflicts of interest of NCD requestors and CMS staff.  Commenters note that no conflicts of interest are disclosed by the NCD requestors nor are conflicts of interest disclosed by CMS staff involved in this NCD analysis.  Commenters assert that coverage recommendations from individuals with conflicts of interest should not drive health policy and contend that conflicts of interest of CMS should be disclosed as well as how intellectual and financial conflicts at CMS have been avoided.  One commenter states that “at least four of the eleven members/experts of the Board offering ‘opinions’ that CMS is using in this decision process have a financial interest in a startup company bringing a new trans-femoral stent product to market resulting in a direct conflict in their participation in this Board.”  Another asks if “the people making this decision involved with a company that has developed a new transfemoral carotid stent platform.”

Response
The MSCA submitted a complete, formal request to reconsider NCD 20.7.  Because the request met the requirements specified in the Federal Register instructions for requesting an NCD (https://www.cms.gov/medicare/coverage/determinationprocess/downloads/fr08072013.pdf), CMS accepted the request and later opened this reconsideration.  As evident from the Federal Register instructions, there are no requirements for NCD requestors to disclose any conflict of interest information.  Furthermore, NCD requestors are not “members” or “experts” on a “Board” that guides or directs CMS policy making.  NCD requestors make the request, provide information to support their request and interact with CMS as any other interested party. CMS reviews all public comments, including those submitted by NCD requestors and all other timely comments submitted during two 30-day public comment periods.  CMS also meets with any interested parties that request meetings.  Following a comprehensive review of the evidence, CMS issued a proposed decision, and is now issuing a final decision.

IX. CMS Analysis

Introduction

National coverage determinations (NCDs) are determinations by the Secretary with respect to whether or not a particular item or service is covered nationally by Medicare (§1869(f)(1)(B) of the Act).  In order to be covered by Medicare, an item or service must fall within one or more benefit categories contained within Part A or Part B, and must not be otherwise excluded from coverage.  Moreover, with limited exceptions, the expenses incurred for items or services must be reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member (§1862(a)(1)(A) of the Act).

When making NCDs, we evaluate the evidence related to our analytic questions based on the quality, strength and totality of evidence presented in the reviewed literature.  As part of this evaluation, it is important to consider whether the evidence is relevant to the Medicare beneficiary population.  In determining the generalizability of the results of the body of evidence to the Medicare population, we consider, at minimum, the age, race and sex of the study participants.

Summary

We reconsidered certain aspects of section 20.7 of the Medicare National Coverage Determinations manual, Percutaneous Transluminal Angioplasty (PTA) of the Carotid Artery Concurrent with Stenting.  Specifically, we reconsidered NCD 20.7, B4, which covers stenting with embolic protection devices (EPDs) for symptomatic patients with carotid artery stenosis ≥70% who are considered “high surgical risk” for carotid endarterectomy (CEA).  Section B4 of the NCD also covers other, including asymptomatic, patients in the context of studies (CMS 2009).  We are finalizing our proposal at section B4 of the NCD to cover PTA of the carotid artery concurrent with stenting with EPDs for symptomatic patients with carotid artery stenosis ≥50% or asymptomatic patients with carotid artery stenosis ≥70%.  We are finalizing the proposal to require an SDM interaction prior to CAS, but, based on public comments, and because no such tool exists, we are not requiring the use of a validated SDM tool during this interaction.  We are not finalizing our proposal to eliminate facility standards, and instead, based on public comment, are maintaining standards for facilities to perform CAS procedures, but removing the requirement for CMS to approve facilities.  This change means that all facilities that perform CAS procedures must comply with all facility standards as specified in the final NCD, however facilities are no longer required to submit attestations to CMS to be approved and maintain approval.  As such, the list of approved CAS facilities will no longer be updated and will be removed from the CMS website. 

We are finalizing our proposal to add language at NCD 20.7, D (“Other”) to enable MACs to make reasonable and necessary determinations under section 1862(a)(1)(A) for other Medicare beneficiaries who do not meet the coverage criteria in section B but are seeking Medicare coverage for PTA of the carotid artery concurrent with stenting.

The 2009 NCD was based on earlier trials comparing CAS to CEA.  These in turn were based on the superiority of CEA over medical therapy in foundational trials launched three decades ago, in patients who were symptomatic (NASCET, ECST trials) or asymptomatic (ACAS trial).  Even for asymptomatic patients, CEA “halved the subsequent rate of disabling or fatal stroke” compared to medical therapy alone (Halliday 2010), although the absolute difference (6%) was small (Walker 1995, Halliday 2004).

Since the last CMS reconsideration in 2009, papers on five major RCTs comparing CAS to CEA have been published along with large, prospective registry-based studies, and meta-analyses of pooled trial data.  We find that this new evidence, especially now for asymptomatic patients, suffices to demonstrate that CAS and CEA are similarly effective with respect to the composite primary outcomes of recent trials (which have included various combinations of disabling stroke, death, any stroke, and myocardial infarction) in patients with either standard or high surgical risk, and who are symptomatic with carotid artery stenosis ≥50% or asymptomatic with stenosis ≥70%.

“Symptomatic” patients are those with recent transient ischemic attack, stroke, or other relevant neurologic signs or symptoms as defined in contemporary trials.  “Asymptomatic” patients are those without a recent neurological event but with carotid atherosclerosis on imaging.  We note here that screening for carotid artery stenosis is not recommended and is not a covered service under Medicare.  If carotid artery stenosis is incidentally seen on other imaging, the further imaging work up as specified in this NCD would apply to the patient.  Certain diagnostic imaging and other patient assessment requirements in our proposed decision were modified based on public comments, to provide greater accuracy, and flexibility in diverse clinical scenarios.  We believe these modifications will further support SDM, designed to assist patients in making better treatment choices, as discussed below.

Consistent with every recent societal guideline, all patients should have a reasonable life expectancy (>3 or >5 years depending on the guideline) in order to experience benefit from the procedure (i.e., to overcome procedure-related adverse events, which medical therapy does not entail).  We are finalizing our proposal to require a formal SDM interaction between the practitioner and the patient because of the importance of individual patient anatomy, pathophysiology, and preferences, and as CAS and CEA entail different procedural (30-day) risks (minor stokes generally being more common with CAS; myocardial infarction (MI) and cranial nerve palsies, with CEA).  We are not requiring the use of a validated SDM tool as one has not been developed at the time of writing this final decision.

While not a requirement under this NCD, individuals undergoing a carotid revascularization procedure should be provided with optimal medical therapy (OMT) for atherosclerotic disease, including lifestyle/behavioral modification counseling, tailored to the individual patient.  The aim is to reduce all-cause and cardiac mortality in addition to carotid-related stroke and mortality, and (perhaps) to protect against long-term cognitive decline (Paraskevas 2022, Lancet 2021, Balestrini 2013).  Such medical therapy is often reflected in the latest societal guidelines, if published recently (e.g., the 2023 European Society of Vascular Surgeons (ESVS) clinical guideline is now available (Naylor 2022); it is also the only guideline to date that incorporates the most recent CAS vs CEA trial for asymptomatic patients, ACST-2).

We recognize that the medical therapy that served as the control group treatment in the early, foundational trials has also vastly improved.  Thus, contemporary trials comparing CEA to CAS reviewed in this NCD analysis, with their wide use of modern antithrombotic, antihypertensive, and lipid-lowering drugs, and lifestyle risk-modification, as background medical therapy for both groups, can fairly compare the two procedures.  However, these trials do not answer the question of whether any procedure produces better health outcomes than contemporary OMT alone for these patients.  This highlights the importance of completing and publishing ongoing trials which compare each procedure to OMT alone (e.g., CREST-2 and ECST-2).

Since we proposed to remove facility standards and approval requirements, we did not propose specific physician, care team, or facility standards in this NCD.  Based on public comments, we have decided to maintain standards for facilities to perform CAS procedures and have finalized standards largely consistent with those established in 2005.  Since facilities have over 18 years of experience establishing CAS programs as specified in the 2005 NCD, we no longer believe that it is necessary for facilities to undergo an approval process by CMS to perform CAS procedures. Given the maturity of the technology, we believe facilities are capable and should be fully responsible for ensuring that the CMS standards are met and adhered to as specified.  The modified facility standards retain the basic elements of the 2005 NCD, while removing the facility approval and data submission process and allowing flexibility so facilities may establish customized programs tailored to their institutional, community and practitioner goals and needs while ensuring optimal outcomes and addressing any suboptimal results swiftly.

Standards for CAS programs are important, and adequately exist in societal guidelines, are backed by evidence from contemporary randomized trials, continue to evolve, and provider, patient, and outcomes data housed at local facilities can be audited or reviewed by appropriate entities.  That said, we note that a standardized, nationwide registry – for all carotid artery procedures – would be helpful to monitor procedural safety, further evolve patient risk stratification, and to facilitate auditing and quality improvement, including comparison of local outcomes to national and other benchmarks.  However, for a technology that has been investigated and used widely in various forms, we do not believe it is necessary for coverage purposes.  Recognizing and supporting the value of registries is separate and distinct from mandating participation for coverage purposes under Medicare.

We recognize some interested parties strongly disagree with utilization of CAS entirely, most specifically, TF-CAS, unless data is collected in a national registry.  We understand the desire for ongoing research and information gathering to improve patient treatment options, and we believe, for a technology that has gone through nearly 30 years of clinical investigation with FDA marketing approvals beginning in 2004, that coverage is reasonable and necessary without further mandated evidence generation for optimally selected patients.

It is the responsibility of facilities and practitioners, with professional medical society support, to ensure that the right patient receives the right procedure at the right time. The SDM requirements of this NCD coupled with the facility standards, are intended to guide this, but an NCD cannot be so prescriptive as to actually practice medicine. Regardless of personal opinions of who should do what to whom and when, the fact is that CAS, transfemoral, transradial or TCAR, is a revascularization option that many specialties have been involved with for decades. It is prudent for all specialties to work cohesively together to ensure each patient is able to receive the treatment that is right for them, given their individual context.

CMS can ensure that the tools, supported by strong clinical evidence, as TF-CAS is, are available, and even establish parameters within which their use is reasonable and necessary. But CMS cannot set forth policy so prescriptive and prohibitive based on a prediction of overuse by bad actors, that patients who could benefit from a procedure are not afforded the option. As CMS believes that there is value in registry participation, albeit not for coverage purposes in this NCD, it would be helpful wherever possible that large-scale, data-collection enterprises (registries) collaborate, include all patients, and build off existing infrastructure rather than create anew. Similar collection of outcomes data for patients on OMT alone would be informative. Results of ongoing trials, such as CREST-2 and ECST-2, may better inform registry content.

Our analysis and final decision do not:

• Distinguish, for purposes of coverage criteria, among different:
  • approaches for deploying FDA-approved stent systems (e.g., transfemoral, transradial, or transcarotid);
  • types of FDA-approved stents (e.g., open cell, closed cell, dual layer); and
  • types of FDA-approved devices and methods for embolic protection (e.g., proximal or distal protection, or use of flow reversal).
• Address the timing of carotid artery procedures after neurological events, or the appropriate sequence for multiple procedures (e.g., for patients in need of both carotid and coronary revascularization).
• Specify performance criteria for physicians, care teams, or facilities.
• Assess carotid procedures compared to OMT alone.  While observational studies support the case for OMT, these merely provide the clinical equipoise for the definitive, ongoing randomized trials.  Assessment awaits trial results.

Answers to the Four Reconsideration Questions

Our answers to the four key questions for this NCD reconsideration are based on: a review of the trial evidence and relevant reports (discussed in greater depth in the Evidence Review below) since the 2009 reconsideration; internal review across CMS coverage policies to promote consistency; and feedback during the public comment period.  All four questions are with respect to Medicare beneficiaries.

Q1.  Does carotid artery stenting with embolic protection improve health outcomes similarly to carotid endarterectomy in symptomatic patients with carotid artery stenosis?

Our answer is yes, and CAS does so for symptomatic patients with standard as well as high surgical risk (as these terms are described in recent trials).  Based primarily on the CREST trial (half of which were symptomatic patients) and ICSS trial (symptomatic patients only), the evidence is stronger for patients with carotid artery stenosis ≥70%; however, we believe it is sufficient for patients with ≥50% stenosis as well.  The reason for the relatively weaker evidence for the ≥50% stenosis mark is that while inclusion criteria for both trials was carotid stenosis ≥50%, in fact the vast majority of patients had stenosis ≥70% (86% for CREST and 89% for ICSS).  This may reflect reluctance of physicians to refer patients with moderate stenosis at that time.  However, there is evidence supporting CAS for carotid stenosis of 50-69%, and we believe it is less helpful if CEA was the only procedural option for these patients, regardless of patient anatomy and other characteristics (see Q3 below).

How carotid artery stenosis is measured matters.  In contemporary practice and trials alike, carotid stenosis is appropriately measured by noninvasive means, primarily duplex (Doppler vascular and B mode) ultrasound, with confirmation by CT or MR angiography (CTA or MRA).  This is because of improvements in non-invasive imaging modalities, and because interventional (catheter) angiography is itself associated with increased risk of stroke.  Both CTA and MRA also provide additional information about anatomy (e.g., aortic tortuosity), and characteristics and locations of atherosclerotic plaques, which are crucial for selecting the right procedure for the individual patient.  The importance of patient selection is a primary reason for SDM, discussed in Q3 below.

Q2.  Does carotid artery stenting with embolic protection improve health outcomes similarly to carotid endarterectomy in asymptomatic patients with carotid artery stenosis?

Our answer here is yes as well, and again for patients with standard as well as high surgical risk.  The publication of ACST-2, the largest and most recent of any CAS vs CEA trial, has reversed the historical trend in which there has always been better evidence for carotid procedures in symptomatic than in asymptomatic patients.  The ACST-2 trial supports coverage of asymptomatic patients with standard risk and carotid artery stenosis ≥70% (again, by duplex ultrasound along with CT or MR angiography).  Similar outcomes from a large, national registry support the conclusion that trial findings are generalizable to broad “real-world” clinical practice.

For both symptomatic and asymptomatic patients with carotid artery stenosis, while guidelines and meta-analyses of trial data are supportive, the most convincing evidence comes from the most recent individual trials themselves.  However, the evidence to date does not compare either procedure to OMT alone, which has also vastly improved over the last decade; that comparison awaits publication of trials like CREST-2 and ECST-2.  We look forward to reading those trial reports.

Q3.  Do specific patient characteristics impact which procedure, CAS or CEA, results in better health outcomes for individual patients?

Yes. There is a growing body of evidence on which patients might be better suited for the respective procedures. Contemporary risk stratification includes not just patient age and comorbidities (e.g., presence and severity of heart disease, lung disease, diabetes, hypertension, obesity, history of stroke or MI), which are typically assessed before any surgery, but also, for carotid procedures, relevant arterial anatomy, and characteristics, locations and extent of atherosclerotic plaques to include above and below the carotid bulb. 

For example, a recent review states, “there are a variety of clinical and imaging features associated with the risk of future stroke in asymptomatic carotid stenosis. They include: (1) stenosis progression; (2) plaque characteristics including morphology, size, and vulnerability; (3) accompanied stroke-related conditions; and (4) microembolization” (Baek 2023).  However, the evidence on patient risk stratification is still growing and has not matured to the extent that would allow us to codify stable, evidence-based criteria into NCD coverage criteria at this time.

Individual patient characteristics thus matter, but so do patient concerns and preferences, especially as there are three management strategies (CAS, CEA, OMT alone), and the two procedures entail different procedural (30-day) risks (minor stokes generally being more common with CAS; MI and cranial nerve palsies, with CEA).  As a clinical example, a patient with no aortic tortuosity, good femoral access, no extensive carotid calcifications, no hypertension or significant white matter disease on brain MRI, but (stable) coronary artery disease, and whose father died in his 50s from an MI, might reasonably select CAS over CEA.  A different patient with a different clinical picture and set of concerns may reasonably select CEA. Many patients will (and probably should) select OMT alone.

We thus are requiring a formal SDM interaction to ensure that the beneficiary receives up to date information and personalized advice.  Ideally, the interaction would involve the patient’s primary care physician, and expert advice might come directly from the competing experts: a surgeon, an interventionalist who is not a surgeon, both of whom perform their procedures in accordance with contemporary standards of care (more on this in Q4 below), and an independent physician with expertise in medical management of carotid artery disease, who does not perform procedures.  We note that such advice from various experts has been integrated along with short videos into interactive, electronic decision tools in other medical fields.  However, we are not requiring use of such a tool in this NCD (although we encourage multidisciplinary development of one).  Even without a specific tool, we realize that all elements of the ideal interaction(s) are not equally feasible for all patients and clinical settings, and so we do not specify which practitioner or specialties are to provide the SDM interaction. 

Q4.  Are NCD criteria for physicians, care teams, and facilities performing CAS necessary or helpful to ensure that beneficial health outcomes seen in trials are achieved in broad community practice?

We proposed to discontinue the requirement that facilities meet specific standards and be approved by CMS to perform CAS procedures.  However, based on public comments, we have decided to maintain standards for facilities to perform CAS procedures, and have finalized standards largely consistent with those established in 2005 as follows:

Facilities must establish and maintain institutional and physician standards to support a dedicated carotid stent program. These standards must at least include and ensure the following:

• Facilities have a clearly delineated program for granting carotid stent privileges and for monitoring patient outcomes for individual physicians and the program as a whole.
• The oversight committee for this program shall be empowered to identify the minimum case volume for a physician to maintain privileges, as well as the (risk-adjusted) threshold for complications that the institution will allow before suspending privileges or instituting measures for remediation.  Committees are encouraged to apply published standards from specialty societies and widely-used, published professional society guidelines to determine appropriate physician qualifications.
• Facilities have appropriately trained staff capable of fulfilling roles and responsibilities as delineated under the dedicated carotid stent program.
• Facilities have appropriate supporting personnel and equipment for imaging, emergency management, advanced physiologic monitoring, and other ancillary care.
• Facilities must ensure continuous quality improvement by assessing procedural outcomes and making necessary programmatic adjustments to assure patient safety.

Please note: CMS or third-party facility approval, accreditation or certification are not required by this NCD for coverage purposes.

Since facilities have over 18 years of experience establishing CAS programs as specified in the 2005 requirements, we no longer believe that it is necessary for facilities to undergo an approval process by CMS to perform CAS procedures. Given the maturity of the technology, we believe facilities are capable and should be fully responsible for ensuring that the CMS standards are met and adhered to as specified.  The modified facility standards retain the basic elements of the 2005 NCD, while removing the facility approval and data submission process and allowing flexibility so facilities may establish customized programs based on up-to-date specialty society guidelines that are tailored to their institutional, community and practitioner goals and needs while ensuring optimal outcomes and addressing any suboptimal results swiftly.  Facilities approved by CMS to perform CAS procedures prior to the effective date of this updated NCD, must comply with the above standards, but will no longer submit letters for recertification or be listed on the CMS website as approved CAS facilities going forward.

Analysis of data from large registries support the conclusion that benefits and harms of CAS compared to CEA seen in trials are generalizable to broad community practice.  This is not surprising, as 20 years have passed since the first investigational device exemption (IDE) study was conducted for CAS.  The standard of care for diagnosis and treatment of carotid artery disease, and training to enter this field, has furthermore matured since our 2009 NCD reconsideration.  Currently, provider and individual patient data are maintained at these local facilities, and could be used for continuous quality improvement.  Team approaches to care and acceptable patient outcomes are backed by contemporary trials and guidelines alike.  Thus, appropriate physician and facility experience and outcomes are adequately reflected in current standards of care, and institutional processes (public and private) exist to monitor practice and enforce standards. We expect facilities to use these resources in developing and updating their carotid stent programs.

The standards themselves are constantly improving.  For example, the current rate for procedural (30-day) disabling stroke or death in asymptomatic patients is now near 1%, and for any stroke or death, below 4% (for either procedure, in trials and registries alike).  The ESVS recommends that providers have a <3% rate for any stroke or death for asymptomatic patients (and <6% for symptomatic ones) – but have hinted that this may change (Naylor 2022).  The SVS guidelines have yet to be updated after ACST-2.  We would not want to codify in an NCD a numeric threshold that may be constantly moving so the final facility standards allow flexibility for programmatic adjustments as the carotid stenting landscape evolves.

Evidence Review

The five major, relevant new reports from trials comparing CAS to CEA since the last NCD reconsideration are for: ICSS, CREST, ACT-1, SPACE-2, ACST-2 (see Evidence Table 1 in this decision memorandum).  These five RCTs are all in patients with standard (also called average) surgical risk.  One trial has symptomatic patients (ICSS); one has both symptomatic and asymptomatic patients (CREST), and three have asymptomatic patients (ACT-1, SPACE-2, ACST-2).  As reflected in the 2009 NCD, the evidence for procedural treatment has previously been stronger for symptomatic than for asymptomatic carotid artery disease (CMS 2009).  The ICSS and CREST trials brought additional evidence for CAS treatment of symptomatic patients, including those at average surgical risk (not just those at high risk for adverse events with CEA).  And with the publication of ACST-2 in 2021, the evidence gap for asymptomatic patients with average surgical risk has largely been filled, with this one trial alone doubling the number of total asymptomatic trial patients (Halliday 2021). 

Additional key evidence includes studies from large registries, both in the U.S., and in Germany (where by law key data from all carotid procedures must be submitted to one national registry (Klein 2023)). While both early and late outcomes of RCTs remain the gold-standard for unbiased comparison of interventions in terms of safety, efficacy, and clinical utility, the analysis of large, observational registry data may better assess the generalizability of trial data to broad (“real-world”) community practice, which involves greater diversity of patients, clinicians, and treatment centers.

Finally, while meta-analyses of pooled trial data have generally supported the findings of the individual trials, the strongest evidence comes from the most recent trials themselves, using contemporary devices and standards of care, which includes physician credentialing and optimal medical therapy.  We discuss these key trials in greater depth below.

ICSS, CREST and ACT-1 trials (earlier trials, for symptomatic and asymptomatic patients)

Of the trial reports published after the 2009 NCD, these constitute the “early” ones, and have a mix of symptomatic and asymptomatic patients.  ICSS is symptomatic only, ACT-2 is asymptomatic only, and CREST is both.  These “early” trials have in common use of background OMT for both CAS and CEA that is not the contemporary OMT seen, for instance, in the ACST-2 trial (discussed later), or ongoing trials in which OMT is a separate treatment arm directly compared to each procedure.  Late outcomes of the ICSS trial were published in Lancet in 2015, and early and late outcomes of CREST and ACT-1 were published in the same New England Journal of Medicine issue in 2016 (Rosenfield 2016, Brott 2016). 

ICSS

ICSS randomized 1,710 symptomatic patients with standard surgical risk and “atherosclerotic carotid stenosis causing at least 50% reduction in carotid artery lumen diameter” to CAS or CEA, with the primary outcome of “fatal or disabling stroke in any territory to end of follow up.”  The final, 10-year report published in 2015 is our focus (Bonati 2015).

Standard surgical risk was described as those patients “deemed equally suited for both treatments” (Bonati 2015), and where the investigators were “uncertain which of the two treatments was the best option for the patient” (ICSS 2010).  This would become the usual wording in later trials for patients not at “high surgical risk” for CEA.  Note that ICSS does not actually use the terms “standard” or “average” regarding patient risk, but the description is consistent with this, and subsequent editorials and reviews use those terms in referring to ICSS (White 2022).  Thus, “patients unsuitable for stenting because of tortuous anatomy proximal or distal to the stenosis, visible thrombus, proximal common carotid artery stenosis, or internal carotid artery pseudo-occlusion were excluded, as were patients unsuitable for endarterectomy because of the distal site of the stenosis, a rigid neck, or risk factors for surgical complications” (with “risk factors” including patient comorbidities; ICSS 2010).  This was a precursor to contemporary risk stratification, itself still evolving. 

Stenosis “of at least 50% reduction in carotid artery lumen diameter” refers to the use of interventional (catheter) angiography to measure lumen diameter (a non-interventional equivalent was acceptable but not typical), using NASCET criteria (Ederle 2010, Bonati 2015).  ICSS ended up with roughly 89% of patients having stenosis ≥70%, with the remainder at 50-69% stenosis (Ederle 2010).

“Symptomatic” carotid stenosis was defined as any one or more of the following symptoms or events “attributable to the randomized artery” in the past 12 months: amaurosis fugax, transient ischemic attack (TIA), ischemic hemispheric stroke, or retinal infarction; patients with major prior strokes were excluded (Ederle 2010).  This is consistent with how trials (both early and contemporary) in this 2010-present era would define symptomatic versus asymptomatic patients.  Physicians in turn had to meet certain experience/volume criteria (Ederle 2010).

The ICSS trial found no significant difference between CAS and CEA at 10 years for the primary outcome of fatal or disabling stroke, or for either ipsilateral (same-sided) stroke, or severe restenosis of the treated carotid artery.  CAS had more non-disabling stroke at 30 days and 5 and 10 years, while CEA had more cranial nerve palsies (but not more MIs, as was seen in prior trials) (Bonati 2015).

CREST

CREST evaluated CAS (with embolic protection) vs CEA in 2,502 patients with standard surgical risk who were either symptomatic (and with ≥50% stenosis of the arterial diameter on angiography, by NASCET criteria) or asymptomatic (with stenosis of ≥60% on angiography).  Overall, 86% of patients had carotid stenosis ≥70%.  Associated criteria for stenosis by duplex ultrasound, CTA and MRA were included.  “Symptomatic” patients were defined as having had a transient ischemic attack, amaurosis fugax, or minor nondisabling stroke involving the ipsilateral carotid artery within 6 months before randomization.  “Asymptomatic” patients were those without a recent neurological event but with evidence of carotid atherosclerosis.  Patients with a major stroke at any previous time were excluded from the trial.  Surgeons and interventionalists had to meet experience and outcome criteria, and stenting used embolic protection devices whenever feasible (Brott 2010, 2016; Hopkins 2010).

In the 1-year and 4-year follow-up reports, the trial found no significant difference between CAS and CEA for the primary composite outcome of any stroke, MI, or death from any cause (all-cause mortality) during the periprocedural period (first 30 days post procedure), or for the long-term outcome of any subsequent ipsilateral stroke during 4 years of follow-up (Brott 2010).  At 30-days, CAS had relatively more strokes, while CEA had more MIs and cranial nerve palsies (Figure 1, Supplementary Appendix, Brott 2010).  The early report was evaluated in the 2009 NCD (CMS 2009). 

The long-term outcome (ipsilateral stroke after the periprocedural period among patients who had had no periprocedural event) was chosen with the rationale that the purpose of the procedure is to prevent future stroke in the cerebral artery territory fed by the treated carotid artery (Brott 2016).

Ten years was chosen for follow up because the mean age of patients at the start of the trial was 69 years, and “at that age the average life expectancy is 15 years for men and 17 years for women” (Brott 2016).  At 10 years, the authors found that for 2,502 asymptomatic (47.2%) and symptomatic (52.8%) patients, there was no significant difference for the early primary composite outcome between CAS (11.8%) and CEA (9.9%), or for the late primary outcome (any ipsilateral stroke) between CAS (6.9%, or 0.7% per year) and CEA (5.6%, or ~0.6% per year).  The combined risk of periprocedural stroke or death and subsequent ipsilateral stroke was 37% higher for CAS than CEA; this was driven by the higher CAS 30-day (periprocedural) stroke rate.  The trial was criticized (by some) for including MI as a component of the primary outcome.  There was no significant difference between CAS and CEA in rates of restenosis or revascularization (Brott 2016).

Prespecified subgroup analysis found no significant between-group differences between symptomatic and asymptomatic patients (Silver 2011).  However, the trial was not powered to detect outcome differences in the asymptomatic population subgroup; thus, comparison between CAS and CEA specifically for asymptomatic patients with severe carotid stenosis and standard surgical risk remained hypothesis-generating.  

ACT-1

ACT-1 randomized 1,453 patients who were asymptomatic (i.e., had not had a stroke, transient ischemic attack, or amaurosis fugax in the 180 days before enrollment), with severe stenosis (≥70% by carotid ultrasound or angiography), and standard surgical risk, to CAS with an embolic protection device or to CEA.  While the trial had a short “run in” phase, there was no physician credentialing criteria as done in CREST.  The trial had planned for 1,658 patients but was stopped early due to slow enrollment and “not related to patient safety, futility, or concerns regarding the study devices” (Rosenfield 2016); hence the trial was slightly underpowered (landing at 75% power).  The primary and some secondary outcomes were also different from other trials.  The trial used a non-inferiority design[1] with a margin of 3% for the primary outcome (a composite of any death, stroke [ipsilateral or contralateral], or MI within 30 days of the procedure), or any ipsilateral stroke within 1 year (Rosenfield 2016).

The trial found, for the primary composite outcome, that CAS (at 3.8%) was noninferior to CEA (at 3.4%).  There was no significant difference, in the context of a non-inferior design, between CAS and CEA for:

• the composite of any stroke or death within 30 days (2.9% for CAS, 1.7% for CEA);
• the composite of major stroke or death within 30 days (0.6% for each);
• freedom of ipsilateral stroke from 30 days to 5 years after the procedure (97.8% for CAS, 97.3% for CEA);
• overall survival (87.1% for CAS, 89.4% for CEA); and
• the cumulative 5-year rate of stroke-free survival (93.1% for CAS, 94.7% for CEA).  (Rosenfield 2016.)

Within 30 days of the procedure, CAS had more strokes, while CEA had more MIs (but these were not statistically significant).  CEA had a small but statistically higher rate of cranial nerve palsies (1.1% vs 0.1% for CAS) (Rosenfield 2016).

Evidence gaps after CREST, ICSS, and ACT-1

An editorial on both CREST and ACT-1 (as these were published in the same issue of the NEJM) concluded that the question about the long-term durability of stents was now answered (Spence and Naylor 2016).  The editorial then highlighted remaining evidence gaps.

The first major gap was “the issue of the generalizability of randomized-trial findings into routine clinical practice . . .” Citing a recent systematic review (Paraskevas 2016), they stated that “9 of 21 large administrative data-set registries (43%) reported rates of death and stroke in excess of the 3% risk threshold that is recommended by the American Heart Association in asymptomatic patients undergoing stenting, as compared with 1 of 21 registries (5%) after endarterectomy” (Spence and Naylor 2016).

The second gap, and related to the issue of generalizability, was a question of appropriate patient selection in broad practice, specifically in the United States.  In the U.S., “more than 90% of carotid-artery interventions are performed in asymptomatic patients, even though evidence suggests that up to 90% of them will undergo an ultimately unnecessary and potentially harmful procedure [Bogiatzi 2012, Spence 2005].  By contrast, the percentage of interventions that are performed for asymptomatic stenoses is approximately 60% in Germany and Italy, 15% in Canada and Australia, and 0% in Denmark [Vikatmaa 2012]” (Spence and Naylor 2016).

The third gap, and related to appropriate selection of patients, was the question of whether any procedure was necessary for patients with carotid artery stenosis, particularly asymptomatic ones, given vast improvements in OMT since the foundational trials demonstrated the clinical utility of CEA over 30 years ago.  The editorial stated that “data from both randomized trials and nonrandomized studies suggest that the annual rate of stroke among medically treated asymptomatic patients has declined over the past two decades, regardless of the severity of stenosis at baseline [Naylor 2015].  Evidence now suggests that the annual rate of ipsilateral stroke may be as low as 0.5 to 1% [Naylor 2015] — a rate that is very similar to that observed in ACT I and CREST after successful stenting or endarterectomy [Rosenfield 2016, Brott 2016]” (Spence and Naylor 2016).

The editorial thus concluded: “Outside clinical trials, endarterectomy and stenting should be reserved for patients with symptomatic severe stenosis or for asymptomatic patients who are shown to be at higher risk for stroke with medical therapy than with intervention [Spence 2005, 2010; Markus 2010; Naylor 2014]” (Spence and Naylor 2016).

In sum, after the longer-term results of ICSS (2015) and CREST (2016), and the ACT-1 trial (2016), there was stronger evidence for CAS (compared to CEA, but not for either procedure compared to OMT) in symptomatic patients with standard surgical risk and carotid artery stenosis. 

Although both ICSS and CREST included patients with carotid stenosis ≥50%, most patients (89% in ICSS, 86% in CREST) had stenosis ≥70%.  A secondary analysis of pooled data of symptomatic patients from these two trials (and including two prior trials) supported that rates for stroke recurrence for CAS and CEA were similar and low out to 10 years (Brott 2019), but given its design, this study is hypothesis-generating.  However, there is face value in the overall result for the primary outcome of a randomized controlled trial population.  While it is important to look at forest plots and other tools for subgroup analysis for indications of which patients in the trial benefitted more or less, caution must be taken in drawing conclusions from these.  That caution cuts both ways.  One cannot definitively reject or accept a certain subpopulation of a trial based on non-powered subgroup analysis.  Nor can one reject or accept a conclusion about a subpopulation based on analysis of data pooled with other trials; for each trial has greater validity internally within the randomized population sample, and less validity outside of it. Such analyses are useful, but mostly for forming hypotheses.

More significant evidence gaps remained for asymptomatic patients (with CREST not powered for subgroup analysis of the asymptomatic patient subgroup, and irregularities in the ACT-1 trial) and for generalizability of trial results to broader (“real-world”) community practice.  Viewing the randomized trials up to that time as a whole, there were limitations due to varying patient populations, interventionalists’ experience (e.g., CREST credentialed physicians while ACT-1 did not), primary outcomes, device types, and use of embolic protection devices.  Finally, comparison of these procedures to vastly improved, contemporary OMT was needed. 

SPACE-2 and ACST-2 (contemporary trials, asymptomatic patients)

In 2022, CMS accepted a complete, formal request to reconsider this NCD, due to the new body of evidence available since the 2009 analysis.  In addition to the SPACE-2 trial (Reiff 2022), ACST-2, the most recent and also the largest of any CAS vs CEA trial, and designed to address key concerns about asymptomatic patients, had recently been published (Halliday 2021).  We consider these two the “contemporary” trials as background OMT, patient selection, and physician criteria were all improved from prior trials. 

SPACE-2

SPACE-2 was designed as a three-armed RCT (BMT alone vs. CEA plus BMT vs. CAS plus BMT) in asymptomatic patients with moderate-to-severe carotid artery stenosis, ≥50% by NASCET trial criteria.  (Note that BMT or best medical therapy, is also called OMT or optimal medical therapy, throughout the literature.) The primary outcome was “the cumulative incidence of any stroke or death from any cause within 30 days or any ipsilateral ischemic stroke within 5 years” (Reiff 2022).

However, because of slow patient recruitment, the trial was changed to two parallel RCTs (BMT alone vs CEA and BMT, and BMT alone vs CAS and BMT).  The authors reported that “Trial recruitment ceased after recruiting 513 patients over a 5-year period (CEA vs. BMT (n=203); CAS vs. BMT (n=197), and BMT alone (n=113)).  The 30-day rate of any death or any stroke was 2.0% for patients undergoing CEA, and 2.5% for patients undergoing CAS.  No strokes or deaths occurred in the first 30 days after randomization in patients randomized to BMT” (Reiff 2022).  Originally, the trial planned to include 3,640 patients.

At the 5-year mark, there was no significant difference among the three arms for the primary outcome (the cumulative incidence of any stroke or death from any cause within 30 days or any ipsilateral ischemic stroke within 5 years), with 2.5% for CEA, 4.4% for CAS, and 3.1% for BMT (Reiff 2022).

The authors concluded that “CEA plus BMT or CAS plus BMT were not found to be superior to BMT alone regarding risk of any stroke or death within 30 days or ipsilateral stroke during the 5-year observation period.  Because of the small sample size, results should be interpreted with caution.”

Explaining the recruitment problem, an editorial stated: “SPACE-2 concluded that recruitment largely failed because (1) many non-trial clinicians believed that interventions were warranted in the majority of asymptomatic patients; (2) patients expected to undergo CEA/CAS following referral and did not want to be randomized to BMT; (3) CAS was already reimbursed across the SPACE-2 countries for the treatment of asymptomatic patients outside the trial; and (4) randomization to BMT (which was not reimbursed) meant a loss of income to randomizing hospitals, surgeons, and interventionists” (Naylor 2016).

ACST-2

We discuss the ACST-2 trial in greater depth as it is both the largest, and the most recent, of any CAS vs CEA trial.  It fills a major evidence gap by doubling the total number of asymptomatic trial patients, and its results complement those of a recent large (“real-world”) registry study, as well as an updated meta-analysis of contemporary RCTs using Cochrane Review methods, as reported in the ACST-2 trial paper itself (Halliday 2021).

ACST-2 randomized 3,625 asymptomatic patients with severe carotid artery stenosis (97% of patients had stenosis ≥70% by duplex ultrasound) and standard surgical risk to CAS with embolic protection or CEA with outcome assessments at 30 days, 5 years, and 10 years.

As with earlier trials, the terms “standard” or “average” risk were not used, but the description matched their meaning.  Duplex (Doppler vascular plus B mode) ultrasound became the dominant imaging modality, replacing interventional (catheter) angiography, because it is readily available, has improved in accuracy, and most importantly, is non-invasive; it was found that interventional angiography itself was associated with higher stroke risk (Naylor 2023, Halliday 2021).  Ultrasound stenosis was confirmed by CT or MR angiography (CTA or MRA); the latter provided additional anatomical information as well (e.g., aortic tortuosity, and plaque characteristics, length, and burden proximally and distally).  Physicians had to meet outcome criteria (a first in any trial): “For participation, the risks of any stroke or death had to be 6% or lower for symptomatic patients and 3% or lower for asymptomatic patients” (Halliday 2021).

The primary early outcome was a composite of disabling stroke or death from any cause at 30-days (i.e., procedure-related events).  The strength of this composite is that these two harms are important, relevant, and roughly equivalent (indeed many older Americans prefer death to a major, disabling stroke; CMS 2019).  The primary late outcome was “long-term (up to 5 or more years) prevention of stroke, particularly disabling or fatal stroke” (Halliday 2021).

Prespecified meta-analysis combining data on symptomatic and asymptotic patients across recent trials, including ACST-2, and comparison of ACST-2 trial results to those of a large registry, were performed with the aim of improving both subgroup analysis and generalizability to broad community practice.  Finally, the study sponsors “had no role in design, data collection, analysis, interpretation, or report writing” (Halliday 2021) – aimed at reassuring that this was an unbiased, investigator-led trial.

The ACST-2 trial found, for early (30-day, procedural) outcomes, that 1% overall had disabling stroke or death (with no significant difference between CAS and CEA), while 2% had non-disabling stroke (with CAS having a statistically significant higher rate of 2.7% vs 1.6% for CEA).  CEA in turn had a greater number of MIs (0.7% CEA vs 0.3% CAS), which was not statistically significant, and of cranial nerve palsies (5.4% vs 0), which was.  This pattern of adverse events was consistent with prior trials, including those with symptomatic patients (e.g., Ederle 2010, Economopoulos 2011, Rosenfield 2016, Cui 2018).  The net crossover from CAS to CEA was 53 patients (3% of the CAS group); CAS unsuitability or failure was typically due to unfavorable anatomy (e.g., extensive arterial calcifications, or tortuosity of vessels including the aorta) (Halliday 2021).

For late outcomes, the 5-year non-procedural rates for fatal or disabling stroke was 2.5% (or 0.5% per year) for each procedure, and for any stroke, 5.3% for CAS and 4.5% for CEA (with no significant difference).  Patients are being followed for 10-year outcomes.  Meta-analysis evaluating any non-procedural stroke in all recent CAS vs CEA trials demonstrated no significant difference between symptomatic and asymptomatic patients (at their respective degrees of stenosis specified in trial inclusion criteria) (Halliday 2021).

The ACST-2 results are best viewed against the background of three facts.  The first was the relative lack of asymptomatic patients in prior trials comparing CAS to CEA (also noted in a 2012 Medicare Evidence Development and Coverage Advisory Committee (MEDCAC) meeting).  ACST-2 has now doubled the number of total asymptomatic trial patients.

The second fact was the similar results of a German nationwide registry (IQTiG 2023), which found that “asymptomatic patients undergoing CAS [n=18,000] or CEA [n=86,000] during 2014–19 had in both cases an in-hospital risk of disabling stroke or death of 0.7%, with median time to discharge of 4–5 days³ (appendix p 9).  A risk of 0.7% within 4–5 days suggests a 30-day risk of disabling stroke or death of about 1% for each procedure” (Halliday 2021).  In addition, in-hospital non-disabling stroke was 1.1% for CAS and 0.7% for CEA (IQTiG 2023, Halliday 2021). 

The third fact was the emergence of studies suggesting that with contemporary OMT alone, patients with severe carotid stenosis may have a risk of about 1% per year of disabling stroke or death (Naylor 2015, Halliday 2021, Reiff 2022).

The authors concluded, “The main finding from the ACST-2 trial of CAS versus CEA is that the effects of the two procedures on disabling or fatal events are approximately equal in terms of procedural hazards (about 1% for each treatment, in line with findings from large, representative registries) and of 5-year disabling stroke rates (which were about 0.5% per year with either procedure, suggesting that they would have been about 1% per year with neither procedure)” (Halliday 2021).  

Our conclusion: with the ACST-2 trial, there is now sufficient evidence for CAS (compared to CEA, but not for either procedure compared to OMT) in asymptomatic patients with standard as well as high surgical risk, and carotid artery stenosis ≥70%.  As noted in greater detail in our response to public comments, we believe the updated and much larger meta-analysis, the ACST-2 trial itself, and the large contemporary registry, all point to a similar conclusion about use of CAS for asymptomatic patients with carotid stenosis. Our conclusions in this NCD draw on the totality of this evidence, not one source of evidence alone.

Recent systematic reviews and guidelines

Recent systematic reviews and meta-analyses (e.g., Hasan 2021, Howard 2021, Müller 2020, Galyfos 2019, Cui 2018) were written before ACST-2, and while helpful, are outdated with respect to treatment of asymptomatic patients with carotid artery disease.  The Oxford Vascular Study, a large prospective, population-based study in the UK, “aimed to establish whether there is any association between the degree of asymptomatic stenosis and ipsilateral stroke risk in patients on contemporary medical treatment” (Howard 2021).  The study included patients who were currently asymptomatic, but who were referred because of a “recent suspected transient ischaemic attack or stroke” – and as such reflected a slightly different population from those in the ACST-2 trial.

The Oxford study concluded that “the stroke risk reported in cohort studies was highly dependent on the degree of asymptomatic carotid stenosis, suggesting that the benefit of endarterectomy might be underestimated in patients with severe stenosis.  Conversely, the 5-year stroke risk was low for patients with moderate stenosis on contemporary medical treatment, calling into question any benefit from revascularisation” (Howard 2021).  This further supports the inclusion criteria in ACST-2 of patients with severe stenosis only. 

The difference in findings about the relationship between the degree of carotid stenosis and risk of ipsilateral stroke in the ACST-2 trial (no relationship found), and the Oxford cohort study (a positive relationship found) is unclear but could be due in part to differences in patient populations as the Oxford patients were referred originally because of prior symptoms (neurological events).  However, both the Oxford and ACST-2 authors agree that it is unclear which patients may still benefit from any carotid artery procedure, compared to contemporary optimal medical therapy alone – and hence on the importance of ongoing trials (such as CREST-2 and ECST-2) which include OMT alone as a separate trial arm.

Recent guidelines, in turn, specifically the 2021 U.S. Society for Vascular Surgery (SVS), the 2021 European Stroke Organization (ESO), and the 2020 German-Austrian guidelines, recommend CEA (plus OMT) for asymptomatic patients with standard (average) surgical risk, and carotid stenosis ≥70% (AbuRahma 2022a, 2022b; Bonati 2021; Eckstein 2020).  None of these consider ACST-2 trial results, published later.  These societies also recommend that providers should have outcomes below certain thresholds for procedural disabling stroke or death (<2% in the 2020 German-Austrian and 2021 ESO, and <3% in the 2021 SVS and 2023 ESVS guidelines), and that patients have a reasonable life expectancy (>5 years typically, but not by every guideline).  The German-Austrian and ESO guidelines dropped to <2% rate of stroke/death from the previous threshold of <3% based on the lower rates seen in recent CAS vs CEA trials, and in recent observational studies of patients on OMT alone (Eckstein 2020, Bonati 2021). 

U.S. Society for Vascular Surgery (SVS) Guidelines (AbuRahma 2022a, 2022b)

The SVS (U.S.) guideline has yet to be updated after the ASCT-2 trial publication.  The current guideline recommends the following:

• For neurologically symptomatic patients with stenosis less than 50% or asymptomatic patients with stenosis less than 60% diameter reduction, optimal medical therapy is indicated. There are no data to support transfemoral CAS, TCAR, or CEA in this patient group.
• Neurologically asymptomatic patients with a 70% or greater diameter stenosis should be considered for CEA, TCAR, or transfemoral CAS for reduction of long-term risk of stroke, provided the patient has a 3- to 5-year life expectancy and perioperative stroke/death rates can be 3% or less. Perhaps future models to help estimate life expectancy based on calculating various physiologic comorbidities such as cardiac, pulmonary, renal, malignancy, will be available in the future. The determination for which technique to use should be based on the presence or absence of high risk criteria for CEA, TCAR, or transfemoral CAS (Table I).  [Table I appears in-line below.]
• CEA is preferred over transfemoral CAS in symptomatic patients with 50% or greater stenosis who are candidates for both procedures. TCAR is preferred over transfemoral CAS but not CEA in symptomatic patients with 50% or greater stenosis.
• Transfemoral CAS is preferred over CEA in symptomatic patients with 50% or greater stenosis and tracheal stoma, situation where local tissues are scarred and fibrotic from prior ipsilateral surgery or external beam radiotherapy. CEA or TCAR may be preferable in situations where ipsilateral tissue planes remain relatively intact.
• TCAR is preferred over CEA and transfemoral CAS in symptomatic patients with 50% or greater stenosis and lesion above C2.
• TCAR is preferred over CEA and transfemoral CAS in high surgical risk (both anatomically and physiologically).
• CAS is preferred over CEA in symptomatic patients with 50% or greater stenosis and severe uncorrectable coronary artery disease, congestive heart failure, or chronic obstructive pulmonary disease. The committee recognized the difficulty in clearly defining this group of individuals, both in symptomatology and risk assessment, and acknowledged the potential increased role of aggressive medical management as primary therapy in this high-risk group.
• Neurologically asymptomatic patients deemed high risk for CEA, TCAR, and transfemoral CAS should be considered for primary medical management. Intervention can be considered in these patients only with evidence that perioperative morbidity and mortality is less than 3%. CAS should not be performed in these patients except as part of an ongoing clinical trial.
• There are insufficient data to recommend transfemoral CAS as primary therapy for neurologically asymptomatic patients with 70% to 99% diameter stenosis. Data from CREST, ACT, and the VQI suggest that in properly selected asymptomatic patients, CAS may be equivalent to CEA in the hands of experienced interventionalists. Operators and institutions performing CAS must exhibit expertise sufficient to meet the previously established AHA guidelines for treatment of patients with asymptomatic carotid stenosis. Specifically, the combined stroke and death rate must be less than 3% to ensure benefit for the patient.

Table I. Revascularization techniques with high-risk criteria

European Society for Vascular Surgery (ESVS) 2023 Clinical Practice Guidelines (Naylor 2022)

The most recent of any societal guideline, the 2023 ESVS clinical guidelines recommends, for symptomatic patients: “for patients reporting carotid territory symptoms within the preceding six months” and who have either a 50-69%, or a 70-99%, carotid stenosis, “carotid endarterectomy is recommended provided the 30-day risk of death/stroke rate is <6%” (Naylor 2022). 

For asymptomatic patients, the ESVS recommends that “for average surgical risk patients with an asymptomatic 60-99% stenosis, carotid endarterectomy should be considered,” while carotid stenting “may be an alternative to carotid endarterectomy,” in both cases “in the presence of one or more imaging or clinical characteristics that may be associated with an increased risk of late stroke, provided 30-day stroke/death rates are <3% and patient life expectancy exceeds five years” (Naylor 2022).  For asymptomatic patients “deemed by the multidisciplinary team to be ‘high risk for surgery’ and who have an asymptomatic 60-99% stenosis in the presence of one or more imaging/clinical characteristics that may be associated with an increased risk of late stroke on best medical therapy, carotid stenting may be considered provided anatomy is favourable, 30-day death/stroke rates are <3% and patient life expectancy exceeds five years” (Naylor 2022).

As for diagnostic work up and patient selection, ESVS recommends the following (Naylor 2022):

• For patients undergoing evaluation of the extent and severity of extracranial carotid stenoses, duplex ultrasound, computed tomographic angiography and/or magnetic resonance angiography are recommended.
• For a patient where carotid artery stenting is being considered, it is recommended that any duplex ultrasound study be followed by computed tomographic angiography or magnetic resonance angiography, which will provide additional information on the aortic arch, as well as the extra- and intracranial circulation.
• In units [centers] which base management decisions in patients with atherosclerotic carotid disease on duplex ultrasound measurement, it is recommended that reports should state which measurement method is used.
• For patients with atherosclerotic disease being considered for revascularisation, intra-arterial digital subtraction [catheter] angiography is not recommended, unless there are significant discrepancies on non-invasive imaging.
• Multidisciplinary team review is recommended to reach consensus decisions regarding the indications for, and treatment of, patients with carotid stenosis regarding carotid endarterectomy, carotid stenting or optimal medical therapy.
• Independent neurological assessment before and after carotid interventions is recommended to audit periprocedural risks.
• For patients with asymptomatic and symptomatic carotid disease, behavioural counselling to promote healthy diet, smoking cessation and physical activity is recommended.”  This is in addition to extensive ESVS recommendations on optimal medical therapy, which go beyond the scope of this NCD.

The ESVS states a preference for CEA over CAS (where the latter could include TCAR as well as transfemoral CAS) in symptomatic patients age >70yrs. 

Considering Societal Guidelines and Trial Evidence as a Whole

As noted above, our coverage criteria for CAS – the focus of this NCD – is based primarily on the evidence from large RCTs, and thus on the major patient inclusion and exclusion criteria for these trials.  However, there is significant overlap among recommendations by SVS and ESVS, and the trial evidence itself (as might be expected), with some discrepancies, and we note that ACST-2 trial results for asymptomatic patients were not available for the SVS committee to consider when writing its recommendations. 

No guideline (or any of the five new major RCT papers discussed above) recommends or provides evidentiary support for treatment with any procedure for any patient with carotid artery stenosis <50%.  All support OMT combined with lifestyle/behavioral modifications (and no procedure) in patients with mild stenosis as such, to prevent progression of carotid atherosclerotic disease.  Our final decision reflects these recommendations.

The ESVS recommends independent neurological assessment before and after carotid interventions to assess procedural harms and facilitate audits (Naylor 2022).  This was a requirement we proposed but are now modifying based on public comments.  We acknowledge that our proposed language was too vague, and that there are certain real-world clinical scenarios where a truly “independent” neurological assessment would be logistically problematic.  We also note that there is no requirement for an independent neurological assessment before and after CEA, and that in the large German registry (discussed earlier) this was not mandated, and in an early analysis only about half of the patients received this (Theiss 2004). 

Nonetheless, such assessments are recommended by the latest guideline (Naylor 2022) and have been an integral part of trials.  A key reason CMS is taking a step back from our previous, more intensive oversight of facilities and physicians is the belief that there is adequate infrastructure and oversight in place to identify proceduralists who are not meeting acceptable outcomes and take appropriate action.  What constitutes acceptable outcomes is evolving as technology, techniques, training, and patient selection all improve, and these outcomes are explicitly stated in guidelines.  We believe evaluation of neurological outcomes is important to the integrity of this system. 

To strike this balance, we therefore are requiring that “Neurological assessment by a neurologist or NIH stroke scale (NIHSS) certified health professional before and after carotid artery stenting (CAS) must be performed” for the purpose of assessing procedural harms.  We are thus striking the word “independent” that appeared in our proposed language, and clarifying who may perform the assessment.  Neurological assessments may be conducted in the normal course of patient care without requiring a unique visit.

As for thresholds for carotid stenosis, we derive our criteria principally from the randomized trials themselves (as reviewed earlier), as these provide the primary evidence base.   Largely based on those trials, we concluded that there is sufficient, new RCT evidence for CAS to be considered for certain symptomatic patients, across the spectrum of patient risk for adverse events with CEA, with carotid artery stenosis ≥50%.  There is also sufficient evidence now for CAS to be considered for certain asymptomatic patients, also across the risk spectrum, with carotid artery stenosis ≥70%.  Hence, we have removed risk categories for symptomatic and asymptomatic patients from the 2009 NCD, and updated criteria for the degree of carotid stenosis required for each group.  We do not establish coverage criteria based in comparison of TCAR to CEA or transfemoral CAS, as CAS vs CEA have multiple RCTs for symptomatic and asymptomatic patients, while TCAR has none.  As we discuss in the TCAR section below, TCAR remains from a regulatory standpoint a type of CAS that uses a different delivery approach.

For diagnostic imaging, the evidence supports primary reliance on non-invasive imaging modalities for diagnosis of carotid artery stenosis, but here again we are modifying our proposed requirements based on public comments (please see the Public Comment subsection on Imaging above).  We agree that arterial digital subtraction (catheter) angiography may still be necessary, not only if there is a significant discrepancy between non-invasive imaging results, but also if patients have contraindications to CTA or MRA (including risk factors for contrast-induced nephropathy).  Exceptions and flexibility for patients with contraindications is basic to all of imaging – but we agree this needs to be explicitly stated in our NCD criteria.  We note that conventional catheter angiography uses contrast and causes radiation exposure as well, but effective doses may be less, and are more in control of the operator.

We also agree that a combination of duplex US and either CTA or MRA are needed for greater accuracy and to evaluate other relevant arterial anatomy in addition to the carotid stenosis.  This again applies only when there are no contraindications to CTA or MRA.  This combination of non-invasive imaging is consistent with ACST-2 trial inclusion criteria (the latest and largest ever CAS v CEA trial), and with the most recent guidelines (Naylor 2022) that considered ACST-2 results.  For ACST-2, duplex US was the primary imaging modality, and to be eligible for the trial patients also needed “CT or MRI confirmation of suitability for CAS and for CEA (which would also have been used to exclude from trial entry any patient without sufficient stenosis to justify intervention)” (Halliday 2021).  The guidelines in turn recommend: “For a patient where carotid artery stenting is being considered, it is recommended that any duplex ultrasound study be followed by computed tomographic angiography or magnetic resonance angiography, which will provide additional information on the aortic arch, as well as the extra- and intracranial circulation” (Naylor 2022).  Again, we agree there needs to be an explicit exception for contraindications.

As for special circumstances where more information may be needed by digital subtraction (catheter) angiography, which could not be obtained by two non-invasive imaging modalities, we note that the Medicare Administrative Contractors (MACs) are structured to consider such case-by-case situations when the need arises.

However, we remain firm that digital subtraction (catheter) angiography should be the exception, not the rule, for diagnostic imaging to assess patients for possible carotid procedures.  Here we are distinguishing between, on the one hand, diagnostic imaging used for assessing stenosis, other plaques and their characteristics, and general arterial anatomy – all of which should inform SDM before the patient is on a procedure table – and on the other hand, use of minimal angiography during an interventional procedure for purposes of stent placement.  Rarely, there may be instances when the latter leads to recognition that the non-invasive imaging (required by this NCD) that got the patient onto the procedure table was clearly incorrect.  The expectation is that the procedure would then not be performed.

We note that digital subtraction (catheter) angiography itself is associated with stroke; non-invasive imaging is not, and has seen significant technological advances, with more in the pipeline.  When a patient is receiving a diagnostic imaging test on a procedure table (with catheter angiography), there is increased incentive and pressure to just do the procedure.  There is also less ability to then decide that CEA (or OMT) may, after all, be more appropriate.  Indeed, this would undercut the whole point of SDM, at the outset, which should involve discussion of the patient’s anatomy and other factors, and ideally would involve the patient’s primary care physician as well as specialists.  A scenario where it would be helpful to do diagnostic imaging, combined with possible intervention, with the decision making happening while the patient is on the procedure table, is an acute setting; however, this NCD applies to chronic carotid artery stenosis only, not emergent care.

Finally, while our conclusions rely on our independent analysis based on our review of the literature, along with consideration of public comments, we include the below from the 2023 ESVS guidelines (Naylor 2022) for the public to see how another entity has independently evaluated this topic, in their own words:

“2.5. Imaging strategies in carotid artery disease
During ECST and NASCET, all participants underwent intraarterial angiography. This policy has now been abandoned because of angiogram related stroke. In the Asymptomatic Carotid Atherosclerosis Study (ACAS), 30 day death/stroke after CEA was 2.3%, but half of the peri-operative strokes were angiogram related.¹⁹⁵ Colour DUS [Duplex US] is the first line imaging modality due to low cost and accessibility and there are consensus criteria for diagnosing stenosis severity.^196-198  Alternatives include CTA or MRA which can simultaneously image the aortic arch, supra-aortic trunks, carotid bifurcation, distal ICA and intracranial circulation, which is important if CAS is being considered. Contrast enhanced MRA (CEMRA) has higher accuracy than non-contrast MRA (time of flight) but requires paramagnetic contrast agents (gadolinium). In a Health Technology Assessment meta-analysis of 41 nonrandomised studies, DUS, MRA and CTA were equivalent in detecting significant stenoses,¹⁹⁹ but it was advised that centres relying on DUS before CEA should perform a second DUS, preferably by a second operator.¹⁹⁹ A combination of two imaging modalities (DUS + CTA or DUS + MRA) improves accuracy and is routine practice in many centres.²⁰⁰” (Naylor 2022).

All imaging reports of stenosis should specify the method or criteria used (e.g., “NASCET criteria” for a direct measure of luminal stenosis, and for duplex ultrasound, the method for correlating metrics for peak-systolic velocity, etc., with degree of carotid stenosis). 

We use the term “certain” symptomatic and asymptomatic patients above as there is a growing body of evidence about which patients might be better suited for the respective procedures (or OMT alone, with no procedure).  Contemporary risk stratification includes not just patient age and comorbidities (e.g., presence and severity of heart disease, lung disease, diabetes, hypertension, obesity, history of stroke or MI), which are typically assessed before any surgery, but also, for carotid procedures, relevant arterial anatomy, and characteristics, locations and extent of atherosclerotic plaques to include above and below the carotid bulb.

For example, a recent review states, “there are a variety of clinical and imaging features associated with the risk of future stroke in asymptomatic carotid stenosis. They include: (1) stenosis progression; (2) plaque characteristics including morphology, size, and vulnerability; (3) accompanied stroke-related conditions; and (4) microembolization” (Baek 2023).  However, the evidence on patient risk stratification is still growing and has not matured to the extent that would allow us to codify stable, evidence-based criteria into NCD coverage criteria at this time.

Individual patient characteristics thus matter, but so do patient concerns and preferences, especially as there are three management strategies (CAS, CEA, OMT alone), and the two procedures entail different procedural (30-day) risks (minor stokes generally being more common with CAS; MI and cranial nerve palsies, with CEA).  As a clinical example, a patient with no aortic tortuosity, good femoral access, no extensive carotid calcifications, no hypertension or significant white matter disease on brain MRI, but (stable) coronary artery disease, and whose father died in his 50s from an MI, might reasonably select CAS over CEA, other factors being equal.  A different patient with a different clinical picture and set of concerns may reasonably select CEA. Many patients will (and probably should) select OMT alone.

We agree with both of the SVS and ESVS (and all other) guidelines that asymptomatic patients should have some reasonable degree of life expectancy to overcome the procedural (30-day) risks of adverse events such as death, stroke, MI, and cranial nerve palsies demonstrated in multiple trials.  Most guidelines recommend >5 years, while the SVS (U.S.) recommends at least “3-5 years.”  We encourage continual updates to the criteria, as clinical outcomes improve (and adverse events decline) and acknowledge that there is currently no agreed-upon, evidence-based formula for determining life expectancy. 

In the final NCD we include modified facility standards that retain the basic elements of the 2005 NCD and remove the facility approval and data submission process.  This will allow flexibility so facilities may establish customized programs tailored to their institutional, community and practitioner goals and needs while ensuring optimal outcomes and addressing any suboptimal results swiftly.  We are not establishing more detailed standards in this NCD.  Large registries support the conclusion that benefits and harms of CAS compared to CEA seen in trials are generalizable to broad community practice.  This is not surprising, as 20 years have passed since the first IDE study was conducted for CAS.  The standard of care for diagnosis and treatment of carotid artery disease, and training to enter this field, has furthermore matured since our 2009 NCD reconsideration.  Provider and individual patient data are maintained at these local facilities, and team approaches to care and acceptable patient outcomes are backed by contemporary trials and guidelines alike.  Appropriate physician and facility experience and outcomes are adequately reflected in current standards of care, that data can be used for continuous quality improvement, and institutional processes (public and private) exist to monitor practice and enforce standards.  We expect facilities to use these resources in developing and updating their carotid stent programs.

Also, the standards themselves are constantly improving.  For example, the current rate for procedural (30-day) disabling stroke or death in asymptomatic patients is now near 1%, and for any stroke or death, below 4% (for either procedure, in trials and registries alike).  The ESVS recommends that providers have a <3% rate for any stroke or death for asymptomatic patients (and <6% for symptomatic ones) – but hinted that this could change (Naylor 2022).  The SVS guidelines have yet to be updated after ACST-2.  We agree that provider criteria are important (physicians should only be doing a procedure if they can do it well), but we would not want to codify in an NCD a numeric threshold that may be constantly moving.  For these reasons, while we believe thresholds for provider outcomes are valuable, it is in part because societal guidelines continually (and appropriately) update them that we do not feel a need to include them as specific NCD coverage criteria at this time.  The final facility standards allow flexibility for programmatic adjustments as the carotid stenting landscape evolves.

Transcarotid artery revascularization

Transcarotid artery revascularization (TCAR) is a hybrid procedure that combines elements of CAS and CEA.  A stent is placed in the carotid artery through a surgical incision, and cerebral protection is provided by proximal carotid artery clamping and reversal of cerebral arterial flow (these are combined in the ENROUTE Transcarotid Stent and Neuroprotection System).  TCAR was initially performed with a stent that had already been FDA-approved, a shorter version of the same stent used in the CAS arm of the CAS v CEA SAPPHIRE trial (Kashyap 2022).  As such, TCAR was regarded essentially as an approved stent using a different route and method of insertion compared to traditional CAS.  ROADSTER, a small observational study, was the first to use the ENROUTE transcarotid neuroprotection system (NPS), with a variety of carotid artery stents.  Promising 30-day results (Kwolek 2015), and 1-year follow-up of a subset of these patients (Malas 2019) led to the much larger ROADSTER 2 observational study, which used only the ENROUTE Transcarotid Stent and Neuroprotection System (Kashyap 2020, 2022).  This NCD remains consistent with the initial categorization of TCAR as a stent system.  We do not distinguish for purposes of coverage criteria, among different: approaches for deploying FDA-approved stent systems (eg, transfemoral, transradial, or transcarotid); types of FDA-approved stents (eg, open cell, closed cell, dual layer); or types of FDA-approved devices and methods for embolic protection (eg, proximal vs distal, use of blood flow reversal).

The Vascular Quality Initiative (VQI) registry reportedly captures 95% of all TCAR procedures performed in the U.S.  Scores of observational studies have analyzed data collected in this registry.  The registry is commendable and valuable due to its size (>15,500 patients as of 2022), diversity (21% of registry participants are Black, although only 8% are Hispanic), and degree of standardization (Zhang 2022).

Using the registry platform, the ROADSTER 2 prospective observational study included 632 patients at high risk for adverse events with CEA, with symptomatic carotid artery stenosis ≥50%, or asymptomatic stenosis ≥80%.  Early (30-day, periprocedural) outcomes demonstrated a “composite 30-day stroke/death rate of 0.8% and a stroke/death/myocardial infarction (MI) rate of 1.7%” using per-protocol, not intention-to-treat, statistical analysis (Kashyap 2020).  One-year follow up of a small subset of 155 of those ROADSTER 2 patients reported no subsequent ipsilateral stroke (post-procedural, so between 31and 365 days), although 4 patients (2.6%) died, all deemed to be due to non-neurological causes (Kashyap 2022). 

A recent retrospective cohort study using propensity-score methods matched 2,962 TCAR patients with standard surgical risk, “defined as those lacking Medicare-defined high medical or surgical risk characteristics,” with similar CEA patients in a 1:3 ratio (Liang 2023).  The study found no statistically significant difference in the risk of the primary composite outcome of 30-day stroke, death, or MI or 1-year ipsilateral stroke (TCAR 3.0%; CEA 2.6%).  TCAR was associated with a statistically significant higher risk of 1-year ipsilateral stroke (TCAR 1.6%; CEA 1.1%), but no difference in 1-year all-cause mortality (Liang 2023).  An accompanying editorial commented that “an important caveat is the completeness of data at 1 year was poor with only 33% of patients who underwent TCAR with available information and 50% of patients who underwent CEA” (Chaturvedi 2023).

The totality of evidence demonstrates that TCAR is promising as a suitable alternative for certain patients in need of carotid artery revascularization.  There is however the caveat that not a single pivotal RCT assessing TCAR has been performed, limiting its comparison to CEA or transfemoral CAS, which in contrast are supported by multiple contemporary randomized trials.  The authors of ROADSTER 2 concluded that “Larger, randomized controlled trials are needed to confirm the safety and efficacy of TCAR, as well as understand the comparative effectiveness with other carotid interventions” (Kashyap 2022).  We agree.  We also believe that pending CAS or CEA vs OMT trials (e.g., CREST-2 and ECST-2) may better inform patient risk stratification for all carotid artery procedures, TCAR included.

Ongoing trials comparing carotid procedures to OMT

“As with many procedures in the elderly, the key question is should it be done and not whether it can be done” (CMS 2009).

Based on marked improvement in OMT for patients with atherosclerotic disease over the last dozen years (as discussed earlier), two ongoing RCTs are designed to reassess the foundational trials, launched three decades ago, comparing revascularization (now with CEA or CAS) to contemporary OMT.

ECST-2

The 2nd European Carotid Surgery Trial (ECST-2) is randomizing 2,000 patients with asymptomatic or symptomatic carotid stenosis ≥50% and an estimated 5-year risk of stroke of <20% using the Carotid Artery Risk (CAR) score, to revascularization (CEA or CAS, whichever is deemed most appropriate for the patient) or to OMT.  The primary outcome is the combined 2-year rate of any clinically manifest stroke, new cerebral infarct on MRI, MI or periprocedural death (Cheng 2022).

The CAR score was adapted from Rothwell’s model, developed from ECST trial data and validated with NASCET trial data, which showed that several factors such as carotid plaque ulceration, age and associated comorbidities could stratify patients into low, intermediate or high risk of future stroke in patients treated with medical therapy alone and also the amount of risk reduction through CEA (Rothwell 2005, Cheng 2022).  This is especially important given marked improvements in contemporary OMT using cholesterol-lowering statin, antihypertensive, and antithrombotic (anticoagulation and antiplatelet) medications, as well as behavioral modification (e.g., diet, exercise, smoking cessation).  The trial was estimated to be completed in 2025.  However, during the public comment period, we learned that it is now unclear whether ECST-2 as originally designed will be completed at all.  This is regrettable, but not a course CMS can correct through this NCD.

CREST-2

Carotid revascularization for primary prevention of stroke (CREST-2) is two independent randomized controlled trials (CEA v OMT, and CAS v OMT) in 2,480 patients with asymptomatic carotid stenosis ≥70%.  The primary outcome is the composite of stroke and death within 44 days after randomization and ipsilateral stroke thereafter up to 4 years.  Intensive medical therapy is used in all arms, including medical and lifestyle modification and treatment for stroke risk factors such as hypertension, elevated cholesterol, diabetes, tobacco use, excess body weight, and sedentary lifestyle, with use of a patient coach.  At the time of this writing the trial was estimated to be completed in 2026, with trial publication following within a year.  There is a parallel CREST-2 registry designed in part to answer questions about generalizability.  (Turan 2020, 2012; Mott 2017; ClinicalTrials.gov accessed April and September, 2023.)

Shared Decision-Making

As noted above, appropriate patient selection is critical to ensure optimal outcomes from CAS or any other treatment for carotid artery stenosis.  Individual patient characteristics, along with patient preferences, are important to consider when assessing treatment options.  Because there are multiple treatment options for carotid artery stenosis and patient characteristics are complex and varied in this space, it is essential that patients are informed of all treatment options and corresponding risks and benefits.  This is critical to ensure that not only are treatment options assessed to achieve optimal clinical outcomes, but beneficiary preferences and priorities are also incorporated in the decision-making process.

As such, we are finalizing that prior to furnishing CAS, practitioners must engage in a formal SDM interaction with the beneficiary (to also include family and/or medical representatives as needed) to ensure the beneficiary is familiar with and aware of all treatment options.  As discussed in response to public comments, we are not requiring the use of a validated SDM tool.

The formal SDM interaction must include:

• Discussion of all treatment options including carotid endarterectomy (CEA), CAS (which includes transcarotid artery revascularization (TCAR)), and optimal medical therapy (OMT)).
• Explanation of risks and benefits for each option specific to the beneficiary’s clinical situation.
• Integration of clinical guidelines (e.g., patient comorbidities and concomitant treatments).
• Discussion and incorporation of beneficiary’s personal preferences and priorities in choosing a treatment plan.

While not requiring use of a validated SDM tool in the NCD, we would like to note that we believe CMS is not the appropriate entity to develop a tool or specify how to develop a tool to assist in decision-making for carotid artery disease treatments.  We believe this role is best left to clinical societies and other expert bodies, but we note the following criteria are informative:

“These criteria include: 1) transparency regarding the development and validation process (including availability of documentation of how development occurred, when updates were made, and any conflicts of interest within the development team); 2) direct relevance to the specific clinical decision patients are facing; 3) comprehensiveness of the discussion of available treatment options (including alternative pharmacomedical interventions and symptom-focused therapy where applicable); and 4) availability within the public domain” (Knopke 2019).

Additionally, when appropriate, the SDM interaction could offer information and recommendations from experts across specialties (for example from surgeons and non-surgical interventionalists) who perform procedures in accordance with contemporary standards of care as well as independent physicians with expertise in medical management of carotid artery disease.  The goal of this type of interaction is to ensure beneficiaries have all relevant information and expert viewpoints to help them make an informed choice.  We are not requiring distinct interactions with each of these types of practitioners for coverage purposes, but strongly encourage a multidisciplinary approach to the diagnosis and treatment of this patient population (consistent with ESVS recommendations). 

Health Disparities

We have stated that our NCD analysis considers the key randomized controlled trials (RCTs) as the strongest evidence base, but views these results through the prism of large registry data (which provide greater generalizability), recent guidelines from medical societies, and meta-analyses, systematic reviews, secondary analyses and the like.  On race and ethnicity, participation in trials of carotid artery disease treatments (and in all trials across all of medicine for that matter) has typically not reflected what the Medicare population ‘looks like’, and even if it did so, such ‘proportionate representation’ would generally not provide the subgroup sample sizes needed for definitive subgroup analysis. 

While we strongly encourage better strategies for recruitment and retention of underrepresented populations in clinical trials, we have also not wanted that shortfall to itself create more barriers to access.  Thus, in practice, if there is no other evidence suggesting that a particular subgroup defined by race, ethnicity, or gender would not achieve meaningful clinical benefit from a treatment that shows overall positive results in a well-designed trial, our approach is to provide access to those subgroups.  We may encourage or require more research, but will err on the side of access for interventions which for the overall trial population result in meaningful, improved net health outcomes.

We take that approach – ‘leaning toward access’ – in this NCD, while acknowledging that the peer-reviewed medical literature (beyond the trials themselves) has reported differences in patients undergoing carotid revascularization and outcomes across race, ethnicity, gender, and insurance status.  We discuss some key studies in that literature below.

A retrospective analysis of State Inpatient Databases (SIDs) from five U.S. states assessed disparities across race and insurance status in postoperative outcomes of carotid revascularization (CEA or CAS) (Panchap 2020).  The authors found that Black and Hispanic patients had higher odds of stroke and combined death/stroke than White patients.  This disparity was also observed for a combined group of patients reported as having other non-White races. Black and other non-White patients also had increased odds of in-hospital mortality.  These disparities were observed after stratification by procedure type (CEA vs. CAS) and by symptoms on presentation (symptomatic vs. asymptomatic).  The authors reported that Black, Hispanic, and other non-White patients were more likely to undergo CAS than CEA.  The authors found that insurance status was not significantly associated with outcomes in this study.

A recent retrospective analysis of data from the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) assessed racial disparities in perioperative outcomes of CEA and CAS (Cui 2022).  The authors compared outcomes between White patients and a combined group of patients who were American Indian or Alaskan Native, Asian, Black or African American, or Native Hawaiian or Pacific Islander.  The authors found non-White patients were more likely than White patients to experience 30-day postoperative stroke, unplanned return to the operating room, and restenosis following carotid revascularization.  In a stratified analysis of CAS procedures, non-White patients had increased odds of postoperative stroke, stroke/death, and stroke/transient ischemic attack.  This disparity was not observed in a stratified analysis of CEA, but the authors found that non-White patients had increased odds of unplanned return to the operating room and restenosis.  The authors reported that non-White patients had fewer risk factors for CEA but were more likely to undergo CAS.

An analysis of data from self-referred, self-paid vascular screening assessments showed that Black and Hispanic participants had a more adverse cardiovascular risk profile, including a higher prevalence of hypertension or diabetes, but a lower prevalence of high-grade carotid stenosis than White participants (Lal 2021).  The authors also found that Asian participants had a generally lower prevalence of high-grade stenosis and participants reported as Native American had a higher prevalence of high-grade stenosis.  The authors concluded that a lower prevalence of high-grade carotid stenosis may contribute to the lower rates of carotid revascularization in Black and Hispanic patients in the United States.

A recent systematic review and meta-analysis of RCTs and observational studies showed that Black and Hispanic patients who underwent surgery for carotid artery stenosis were more likely symptomatic and had an increased risk for all-cause mortality and stroke or transient ischemic attack compared to White patients (Lee 2023).

A retrospective analysis of data from the Society for Vascular Surgery Vascular Quality Initiative (VQI) assessed sex-based disparities in outcomes following carotid revascularization (CEA or CAS) (Dansey 2020).  The authors found that among symptomatic patients who underwent CEA, female patients had increased odds of 30-day perioperative mortality compared to male patients.  Among asymptomatic patients who underwent CAS, female patients had increased odds of in-hospital stroke and perioperative stroke/death.  The authors observed that female patients were less likely than male patients to receive perioperative statin and antiplatelet therapy.

A recent retrospective analysis of data from NSQIP assessed gender-based differences in postoperative complications of carotid intervention (CEA or CAS) (Goicoechea 2022).  The authors found that female gender was associated with a higher rate of 30-day postoperative cerebrovascular accident/stroke, bleeding complication, and urinary tract infection.  In asymptomatic patients, female gender was also associated with a higher rate of readmission.  In an analysis specific to CAS, the authors found that female gender was associated with higher odds of readmission and urinary tract infection. The majority of patients in this study (98.0%) underwent carotid endarterectomy.

A recent retrospective analysis of data from the National Inpatient Sample (NIS) assessed disparities across race and insurance status in the utilization of high-volume centers for carotid revascularization (Mabeza 2022).  The authors also assessed differences in clinical outcomes between high-volume and low-volume centers.  High-volume centers were hospitals in the highest quartile of annual carotid revascularization caseload (combined CEA and CAS); low-volume centers were in the lowest quartile of annual caseload.  The authors found that Black and Hispanic patients had lower odds of carotid revascularization at high-volume centers than White patients.  This disparity was also observed for a combined group of patients reported as Asian, Pacific Islander, Native American, and other races.  The authors reported that patients with Medicaid or no insurance were less likely to undergo carotid revascularization at high-volume centers than patients with private insurance.  The authors also found that high-volume centers were associated with lower odds of in-hospital mortality/stroke/MI than low-volume centers.  In an analysis specific to CAS, the authors found no significant differences in mortality/stroke/MI between high-CAS and low-CAS centers.  TCAR was not captured in this study.

Differences in health literacy, socioeconomic status, timely surgical access, physician referral patterns, provider factors, and institutional barriers have been suggested in the literature as potential contributors to the observed health disparities.  Further research is required to describe other circumstances and patient characteristics that may impede the delivery of carotid revascularization to those who can benefit from it.  CMS believes the expansion of coverage for CAS to our beneficiaries will serve to allow greater numbers of individuals to access this procedure for stroke prevention.

X.   Conclusion

The Centers for Medicare & Medicaid Services (CMS) finds that coverage of percutaneous transluminal angioplasty (PTA) of the carotid artery concurrent with stenting is reasonable and necessary with the placement of a Food and Drug Administration (FDA) approved carotid stent with an FDA-approved or cleared embolic protection device, for Medicare beneficiaries under the following conditions:

1. Patients with symptomatic carotid artery stenosis ≥50%; and
2. Patients with asymptomatic carotid artery stenosis ≥70%.

   For both A and B above:

   1. Neurological assessment by a neurologist or NIH stroke scale (NIHSS) certified health professional before and after carotid artery stenting (CAS) must be performed.
   2. First-line evaluation of carotid artery stenosis must use duplex ultrasound.
   3. Computed Tomography angiography or magnetic resonance angiography, if not contraindicated, must be used to confirm the degree of stenosis and provide additional information about the aortic arch, and extra- and intracranial circulation.
   4. Intra-arterial digital subtraction (catheter) angiography may be used only when there is significant discrepancy between non-invasive imaging results, or in lieu of computed tomography angiography or magnetic resonance angiography if these are contraindicated.

Prior to furnishing CAS, the practitioner must engage in a formal shared decision-making interaction with the beneficiary. The shared decision-making interaction must include:

• Discussion of all treatment options including carotid endarterectomy (CEA), CAS (which includes transcarotid artery revascularization (TCAR)), and optimal medical therapy (OMT)).
• Explanation of risks and benefits for each option specific to the beneficiary’s clinical situation.
• Integration of clinical guidelines (e.g., patient comorbidities and concomitant treatments).
• Discussion and incorporation of beneficiary’s personal preferences and priorities in choosing a treatment plan.

Facilities must establish and maintain institutional and physician standards to support a dedicated carotid stent program. These standards must at least include and ensure the following:

• Facilities have a clearly delineated program for granting carotid stent privileges and for monitoring patient outcomes for individual physicians and the program as a whole.
• The oversight committee for this program shall be empowered to identify the minimum case volume for a physician to maintain privileges, as well as the (risk-adjusted) threshold for complications that the institution will allow before suspending privileges or instituting measures for remediation.  Committees are encouraged to apply published standards from specialty societies and widely-used, published professional society guidelines to determine appropriate physician qualifications.
• Facilities have appropriately trained staff capable of fulfilling roles and responsibilities as delineated under the dedicated carotid stent program.
• Facilities have appropriate supporting personnel and equipment for imaging, emergency management, advanced physiologic monitoring, and other ancillary care.
• Facilities must ensure continuous quality improvement by assessing procedural outcomes and making necessary programmatic adjustments to assure patient safety.

Please note: CMS or third-party facility approval, accreditation or certification are not required by this NCD for coverage purposes.

In addition to the national coverage described above, Medicare Administrative Contractors (MACs) may make reasonable and necessary determinations under section 1862(a)(1)(A) for any other beneficiary seeking coverage for PTA of the carotid artery concurrent with stenting.

See Appendix B for the NCD manual language.

APPENDIX A

General Methodological Principles of Study Design
(Section VI of the Decision Memorandum)

When making national coverage determinations, CMS evaluates relevant clinical evidence to determine whether or not the evidence is of sufficient quality to support a finding that an item or service is reasonable and necessary. The overall objective for the critical appraisal of the evidence is to determine to what degree we are confident that: 1) the specific assessment questions can be answered conclusively; and 2) the intervention will improve health outcomes for patients.

We divide the assessment of clinical evidence into three stages: 1) the quality of the individual studies; 2) the generalizability of findings from individual studies to the Medicare population; and 3) overarching conclusions that can be drawn from the body of the evidence on the direction and magnitude of the intervention’s potential risks and benefits.

The methodological principles described below represent a broad discussion of the issues we consider when reviewing clinical evidence. However, it should be noted that each coverage determination has its unique methodological aspects.

Assessing Individual Studies

Methodologists have developed criteria to determine weaknesses and strengths of clinical research. Strength of evidence generally refers to: 1) the scientific validity underlying study findings regarding causal relationships between health care interventions and health outcomes; and 2) the reduction of bias. In general, some of the methodological attributes associated with stronger evidence include those listed below:

• Use of randomization (allocation of patients to either intervention or control group) in order to minimize bias.
• Use of contemporaneous control groups (rather than historical controls) in order to ensure comparability between the intervention and control groups.
• Prospective (rather than retrospective) studies to ensure a more thorough and systematical assessment of factors related to outcomes.
• Larger sample sizes in studies to demonstrate both statistically significant as well as clinically significant outcomes that can be extrapolated to the Medicare population. Sample size should be large enough to make chance an unlikely explanation for what was found.
• Masking (blinding) to ensure patients and investigators do not know to that group patients were assigned (intervention or control). This is important especially in subjective outcomes, such as pain or quality of life, where enthusiasm and psychological factors may lead to an improved perceived outcome by either the patient or assessor.

Regardless of whether the design of a study is a randomized controlled trial, a non-randomized controlled trial, a cohort study or a case-control study, the primary criterion for methodological strength or quality is to the extent that differences between intervention and control groups can be attributed to the intervention studied. This is known as internal validity. Various types of bias can undermine internal validity. These include:

• Different characteristics between patients participating and those theoretically eligible for study but not participating (selection bias).
• Co-interventions or provision of care apart from the intervention under evaluation (performance bias).
• Differential assessment of outcome (detection bias).
• Occurrence and reporting of patients who do not complete the study (attrition bias).

In principle, rankings of research design have been based on the ability of each study design category to minimize these biases. A randomized controlled trial minimizes systematic bias (in theory) by selecting a sample of participants from a particular population and allocating them randomly to the intervention and control groups. Thus, in general, randomized controlled studies have been typically assigned the greatest strength, followed by non-randomized clinical trials and controlled observational studies. The design, conduct and analysis of trials are important factors as well. For example, a well-designed and conducted observational study with a large sample size may provide stronger evidence than a poorly designed and conducted randomized controlled trial with a small sample size. The following is a representative list of study designs (some of that have alternative names) ranked from most to least methodologically rigorous in their potential ability to minimize systematic bias:

Randomized controlled trials
Non-randomized controlled trials
Prospective cohort studies
Retrospective case control studies
Cross-sectional studies
Surveillance studies (e.g., using registries or surveys)
Consecutive case series
Single case reports

When there are merely associations but not causal relationships between a study’s variables and outcomes, it is important not to draw causal inferences. Confounding refers to independent variables that systematically vary with the causal variable. This distorts measurement of the outcome of interest because its effect size is mixed with the effects of other extraneous factors. For observational, and in some cases randomized controlled trials, the method in that confounding factors are handled (either through stratification or appropriate statistical modeling) are of particular concern. For example, in order to interpret and generalize conclusions to our population of Medicare patients, it may be necessary for studies to match or stratify their intervention and control groups by patient age or co-morbidities.

Methodological strength is, therefore, a multidimensional concept that relates to the design, implementation and analysis of a clinical study. In addition, thorough documentation of the conduct of the research, particularly study selection criteria, rate of attrition and process for data collection, is essential for CMS to adequately assess and consider the evidence.

Generalizability of Clinical Evidence to the Medicare Population

The applicability of the results of a study to other populations, settings, treatment regimens and outcomes assessed is known as external validity. Even well-designed and well-conducted trials may not supply the evidence needed if the results of a study are not applicable to the Medicare population. Evidence that provides accurate information about a population or setting not well represented in the Medicare program would be considered but would suffer from limited generalizability.

The extent to that the results of a trial are applicable to other circumstances is often a matter of judgment that depends on specific study characteristics, primarily the patient population studied (age, sex, severity of disease and presence of co-morbidities) and the care setting (primary to tertiary level of care, as well as the experience and specialization of the care provider). Additional relevant variables are treatment regimens (dosage, timing and route of administration), co-interventions or concomitant therapies, and type of outcome and length of follow-up.

The level of care and the experience of the providers in the study are other crucial elements in assessing a study’s external validity. Trial participants in an academic medical center may receive more or different attention than is typically available in non-tertiary settings. For example, an investigator’s lengthy and detailed explanations of the potential benefits of the intervention and/or the use of new equipment provided to the academic center by the study sponsor may raise doubts about the applicability of study findings to community practice.

Given the evidence available in the research literature, some degree of generalization about an intervention’s potential benefits and harms is invariably required in making coverage determinations for the Medicare population. Conditions that assist us in making reasonable generalizations are biologic plausibility, similarities between the populations studied and Medicare patients (age, sex, ethnicity and clinical presentation) and similarities of the intervention studied to those that would be routinely available in community practice.

A study’s selected outcomes are an important consideration in generalizing available clinical evidence to Medicare coverage determinations. One of the goals of our determination process is to assess health outcomes. These outcomes include resultant risks and benefits such as increased or decreased morbidity and mortality. In order to make this determination, it is often necessary to evaluate whether the strength of the evidence is adequate to draw conclusions about the direction and magnitude of each individual outcome relevant to the intervention under study. In addition, it is important that an intervention’s benefits are clinically significant and durable, rather than marginal or short-lived. Generally, an intervention is not reasonable and necessary if its risks outweigh its benefits.

If key health outcomes have not been studied or the direction of clinical effect is inconclusive, we may also evaluate the strength and adequacy of indirect evidence linking intermediate or surrogate outcomes to our outcomes of interest.

Assessing the Relative Magnitude of Risks and Benefits

Generally, an intervention is not reasonable and necessary if its risks outweigh its benefits. Health outcomes are one of several considerations in determining whether an item or service is reasonable and necessary. CMS places greater emphasis on health outcomes actually experienced by patients, such as quality of life, functional status, duration of disability, morbidity and mortality, and less emphasis on outcomes that patients do not directly experience, such as intermediate outcomes, surrogate outcomes, and laboratory or radiographic responses. The direction, magnitude, and consistency of the risks and benefits across studies are also important considerations. Based on the analysis of the strength of the evidence, CMS assesses the relative magnitude of an intervention or technology’s benefits and risk of harm to Medicare beneficiaries.

20.7 - Percutaneous Transluminal Angioplasty (PTA) (Various Effective Dates Below)

A. General 

This procedure involves inserting a balloon catheter into a narrow or occluded blood vessel to recanalize and dilate the vessel by inflating the balloon.  The objective of percutaneous transluminal angioplasty (PTA) is to improve the blood flow through the diseased segment of a vessel so that vessel patency is increased and embolization is decreased.  With the development and use of balloon angioplasty for treatment of atherosclerotic and other vascular stenoses, PTA (with and without the placement of a stent) is a widely used technique for dilating lesions of peripheral, renal, and coronary arteries. 

Indications and Limitations of Coverage 

B. Nationally Covered Indications 

The PTA is covered when used under the following conditions: 

1. Treatment of Atherosclerotic Obstructive Lesions 

-In the lower extremities, i.e., the iliac, femoral, and popliteal arteries, or in the upper extremities, i.e., the innominate, subclavian, axillary, and brachial arteries. The upper extremities do not include head or neck vessels.  

-Of a single coronary artery for patients for whom the likely alternative treatment is coronary bypass surgery and who exhibit the following characteristics:  

• Angina refractory to optimal medical management; 
• Objective evidence of myocardial ischemia; and 
• Lesions amenable to angioplasty.  

-Of the renal arteries for patients in whom there is an inadequate response to a thorough medical management of symptoms and for whom surgery is the likely alternative. PTA for this group of patients is an alternative to surgery, not simply an addition to medical management.  

-Of arteriovenous dialysis fistulas and grafts when performed through either a venous or arterial approach. 

2. Concurrent with Carotid Stent Placement in Food and Drug Administration (FDA)-Approved Category B Investigational Device Exemption (IDE) Clinical Trials 

Effective July 1, 2001, Medicare covers PTA of the carotid artery concurrent with carotid stent placement when furnished in accordance with the FDA-approved protocols governing Category B IDE clinical trials. PTA of the carotid artery, when provided solely for the purpose of carotid artery dilation concurrent with carotid stent placement, is considered to be a reasonable and necessary service when provided in the context of such a clinical trial. 

3. Concurrent with Carotid Stent Placement in FDA-Approved Post-Approval Studies 

Effective October 12, 2004, Medicare covers PTA of the carotid artery concurrent with the placement of an FDA-approved carotid stent and an FDA-approved or –cleared embolic protection device (effective December 9, 2009) for an FDA-approved indication when furnished in accordance with FDA-approved protocols governing post-approval studies. The Centers for Medicare & Medicaid Services (CMS) determines that coverage of PTA of the carotid artery is reasonable and necessary in these circumstances. 

4. Concurrent with Carotid Stent Placement

5. Concurrent with Intracranial Stent Placement in FDA-Approved Category B IDE Clinical Trials 

Effective November 6, 2006, Medicare covers PTA and stenting of intracranial arteries for the treatment of cerebral artery stenosis ≥50% in patients with intracranial atherosclerotic disease when furnished in accordance with the FDA-approved protocols governing Category B IDE clinical trials.  CMS determines that coverage of intracranial PTA and stenting is reasonable and necessary under these circumstances.  

C. Nationally Non-Covered Indications 

All other indications for PTA with or without stenting to treat obstructive lesions of the vertebral and cerebral arteries remain non-covered.  

All other indications for PTA without stenting for which CMS has not specifically indicated coverage remain non-covered. 

D. Other 

Coverage of PTA with stenting not specifically addressed or discussed in this NCD is at the discretion of the MACs.  

(This NCD last reviewed xx/xx/xx.)

APPENDIX C—NCD 20.7 (2009)

20.7 - Percutaneous Transluminal Angioplasty (PTA) (Various Effective Dates Below) (Rev. 212, Issued: 01-19-19, Effective: 02-19-19, Implementation: 02-19-19)

The term Medicare beneficiary identifier (Mbi) is a general term describing a beneficiary’s Medicare identification number. For purposes of this manual, Medicare beneficiary identifier references both the Health Insurance Claim Number (HICN) and the Medicare Beneficiary Identifier (MBI) during the new Medicare card transition period and after for certain business areas that will continue to use the HICN as part of their processes. 

A. General 

This procedure involves inserting a balloon catheter into a narrow or occluded blood vessel to recanalize and dilate the vessel by inflating the balloon.  The objective of percutaneous transluminal angioplasty (PTA) is to improve the blood flow through the diseased segment of a vessel so that vessel patency is increased and embolization is decreased.  With the development and use of balloon angioplasty for treatment of atherosclerotic and other vascular stenoses, PTA (with and without the placement of a stent) is a widely used technique for dilating lesions of peripheral, renal, and coronary arteries. 

Indications and Limitations of Coverage 

B. Nationally Covered Indications 

The PTA is covered when used under the following conditions: 

1. Treatment of Atherosclerotic Obstructive Lesions 

-In the lower extremities, i.e., the iliac, femoral, and popliteal arteries, or in the upper extremities, i.e., the innominate, subclavian, axillary, and brachial arteries. The upper extremities do not include head or neck vessels.  

-Of a single coronary artery for patients for whom the likely alternative treatment is coronary bypass surgery and who exhibit the following characteristics:  

• Angina refractory to optimal medical management;
• Objective evidence of myocardial ischemia; and
• Lesions amenable to angioplasty.  

-Of the renal arteries for patients in whom there is an inadequate response to a thorough medical management of symptoms and for whom surgery is the likely alternative. PTA for this group of patients is an alternative to surgery, not simply an addition to medical management.  

-Of arteriovenous dialysis fistulas and grafts when performed through either a venous or arterial approach. 

2. Concurrent with Carotid Stent Placement in Food and Drug Administration (FDA)-Approved Category B Investigational Device Exemption (IDE) Clinical Trials 

Effective July 1, 2001, Medicare covers PTA of the carotid artery concurrent with carotid stent placement when furnished in accordance with the FDA-approved protocols governing Category B IDE clinical trials. PTA of the carotid artery, when provided solely for the purpose of carotid artery dilation concurrent with carotid stent placement, is considered to be a reasonable and necessary service when provided in the context of such a clinical trial. 

3. Concurrent with Carotid Stent Placement in FDA-Approved Post-Approval Studies 

Effective October 12, 2004, Medicare covers PTA of the carotid artery concurrent with the placement of an FDA-approved carotid stent and an FDA-approved or –cleared embolic protection device (effective December 9, 2009) for an FDA-approved indication when furnished in accordance with FDA-approved protocols governing post-approval studies. The Centers for Medicare & Medicaid Services (CMS) determines that coverage of PTA of the carotid artery is reasonable and necessary in these circumstances. 

4. Concurrent with Carotid Stent Placement

Effective March 17, 2005, Medicare covers PTA of the carotid artery concurrent with the placement of an FDA-approved carotid stent with embolic protection for the following:  

• Patients who are at high risk for CEA and who also have symptomatic carotid artery stenosis ≥70%.  Coverage is limited to procedures performed using FDAapproved carotid artery stenting (CAS) systems and FDA-approved or -cleared (effective December 9, 2009) embolic protection devices.  If deployment of the embolic protection device is not technically possible, and not performed, then the procedure is not covered by Medicare (effective December 9, 2009); 

• Patients who are at high risk for CEA and have symptomatic carotid artery stenosis between 50% and 70%, in accordance with the Category B IDE clinical trials regulation (42 CFR 405.201), as a routine cost under the clinical trials policy (Medicare National Coverage Determination (NCD) Manual 310.1), or in accordance with the NCD on CAS post-approval studies (Medicare NCD Manual 20.7); 

• Patients who are at high risk for CEA and have asymptomatic carotid artery stenosis ≥80%, in accordance with the Category B IDE clinical trials regulation (42 CFR 405.201), as a routine cost under the clinical trials policy (Medicare NCD Manual 310.1), or in accordance with the NCD on CAS post-approval studies (Medicare NCD Manual 20.7). 

Coverage is limited to procedures performed using an FDA-approved CAS, stents and FDA-approved or -cleared embolic protection devices. 

The use of an FDA-approved or cleared embolic protection device is required. If deployment of the embolic protection device is not technically possible, and not performed, then the procedure is not covered by Medicare. 

Patients at high risk for CEA are defined as having significant comorbidities and/or anatomic risk factors (i.e., recurrent stenosis and/or previous radical neck dissection), and would be poor candidates for CEA. Significant comorbid conditions include but are not limited to:  

• Congestive heart failure (CHF) class III/IV;
• Left ventricular ejection fraction (LVEF) <30%;
• Unstable angina;
• Contralateral carotid occlusion;
• Recent myocardial infarction (MI);
• Previous CEA with recurrent stenosis;
• Prior radiation treatment to the neck; and,
• Other conditions that were used to determine patients at high risk for CEA in the prior carotid artery stenting trials and studies, such as ARCHER, CABERNET, SAPPHIRE, BEACH, and MAVERIC II. 

Symptoms of carotid artery stenosis include carotid transient ischemic attack (distinct focal neurological dysfunction persisting less than 24 hours), focal cerebral ischemia producing a non-disabling stroke (modified Rankin scale <3 with symptoms for 24 hours or more), and transient monocular blindness (amaurosis fugax). Patients who have had a disabling stroke (modified Rankin scale ≥3) shall be excluded from coverage. 

The determination that a patient is at high risk for CEA and the patient’s symptoms of carotid artery stenosis shall be available in the patient medical records prior to performing any procedure. 

The degree of carotid artery stenosis shall be measured by duplex Doppler ultrasound or carotid artery angiography and recorded in the patient’s medical records. If the stenosis is measured by ultrasound prior to the procedure, then the degree of stenosis must be confirmed by angiography at the start of the procedure. If the stenosis is determined to be <70% by angiography, then CAS should not proceed. 

In addition, CMS has determined that CAS with embolic protection is reasonable and necessary only if performed in facilities that have been determined to be competent in performing the evaluation, procedure and follow-up necessary to ensure optimal patient outcomes. Standards to determine competency include specific physician training standards, facility support requirements and data collection to evaluate outcomes during a required reevaluation. 

The CMS has created a list of minimum standards modeled in part on professional society statements on competency. All facilities must at least meet CMS’s standards in order to receive coverage for CAS for high-risk patients. 

• Facilities must have necessary imaging equipment, device inventory, staffing, and infrastructure to support a dedicated carotid stent program. Specifically, high quality x-ray imaging equipment is a critical component of any carotid interventional suite, such as high-resolution digital imaging systems with the capability of subtraction, magnification, road mapping, and orthogonal angulation. 

• Advanced physiologic monitoring must be available in the interventional suite. This includes real time and archived physiologic, hemodynamic, and cardiac rhythm monitoring equipment, as well as support staff who are capable of interpreting the findings and responding appropriately.

• Emergency management equipment and systems must be readily available in the interventional suite such as resuscitation equipment, a defibrillator, vasoactive and antiarrhythmic drugs, endotracheal intubation capability, and anesthesia support.

• Each institution shall have a clearly delineated program for granting carotid stent privileges and for monitoring the quality of the individual interventionalists and the program as a whole. The oversight committee for this program shall be empowered to identify the minimum case volume for an operator to maintain privileges, as well as the (risk-adjusted) threshold for complications that the institution will allow before suspending privileges or instituting measures for remediation. Committees are encouraged to apply published standards from national specialty societies recognized by the American Board of Medical Specialties to determine appropriate physician qualifications. Examples of standards and clinical competence guidelines include those published in the December 2004 edition of the American Journal of Neuroradiology, and those published in the August 18, 2004, Journal of the American College of Cardiology.

• To continue to receive Medicare payment for CAS under this decision, the facility or a contractor to the facility must collect data on all CAS procedures done at that particular facility.  This data must be analyzed routinely to ensure patient safety. This data must be made available to CMS upon request. The interval for data analysis will be determined by the facility but shall not be less frequent than every 6 months. 

Since there currently is no recognized entity that evaluates CAS facilities, CMS has established a mechanism for evaluating facilities. Facilities must provide written documentation to CMS that the facility meets one of the following:  

1. The facility was an FDA-approved site that enrolled patients in prior CAS IDE trials, such as SAPPHIRE, and ARCHER;  

2. The facility is an FDA-approved site that is participating and enrolling patients in ongoing CAS IDE trials, such as CREST;  

3. The facility is an FDA-approved site for one or more FDA post approval studies; or,  

4. The facility has provided a written affidavit to CMS attesting that the facility has met the minimum facility standards. This should be sent to:  
   
   
   Director, Coverage and Analysis Group
   7500 Security Boulevard, Mailstop S3-02-01
   Baltimore, MD 21244 

The letter must include the following information: 

• Facility's name and complete address; 
• Facility's national provider identifier (formerly referred to as the Medicare provider number); 
• Point-of-contact for questions with telephone number; 
• Discussion of how each standard has been met by the hospital; 
• Mechanism of data collection of CAS procedures; and,
• Signature of a senior facility administrative official.  

A list of certified facilities will be made available and viewable at: http://www.cms.hhs.gov/coverage/carotid-stent-facilities.asp.  In addition, CMS will publish a list of approved facilities in the Federal Register. 

Facilities must recertify every two (2) years in order to maintain Medicare coverage of CAS procedures. Recertification will occur when the facility documents that and describes how it continues to meet the CMS standards. 

The process for recertification is as follows: 

1. At 23 months after initial certification: 
   • Submission of a letter to CMS stating how the facility continues to meet the minimum facility standards as listed above. 

2. At 27 months after initial certification: 
   • Submission of required data elements for all CAS procedures performed on patients during the previous two (2) years of certification.  

     Data elements:

     1. Patients’ Medicare beneficiary identifier if a Medicare beneficiary;
     2. Patients’ date of birth;
     3. Date of procedure;
     4. Does the patient meet high surgical risk criteria (defined below)?
        • Age ≥80;
        • Recent (<30 days) MI;
        • LVEF <30%;
        • Contralateral carotid occlusion;
        • New York Heart Association (NYHA) Class III or IV congestive heart failure;
        • Unstable angina: Canadian Cardiovascular Society (CCS) Class III/IV;
        • Renal failure: end-stage renal disease on dialysis;
        • Common Carotid Artery (CCA) lesion(s) below clavicle;
        • Severe chronic lung disease;
        • Previous neck radiation;
        • High cervical Internal Carotid Artery (ICA) lesion(s);
        • Restenosis of prior CEA;
        • Tracheostomy;
        • Contralateral laryngeal nerve palsy. 
     5. Is the patient symptomatic (defined below)?
        • Carotid Transient Ischemic Attack (TIA) persisting less than 24 hours;
        • Non-disabling stroke: Modified Rankin Scale
        • Transient monocular blindness: amaurosis fugax. 
     6. Modified Rankin Scale score if the patient experienced a stroke.
     7. Percent of stenosis of stented lesion(s) by angiography.
     8. Was embolic protection used?
     9. Were there any complications during hospitalization (defined below)?
        • All stroke: an ischemic neurologic deficit that persisted more than 24 hours;
        • MI;
        • All death. 

Recertification is effective for two (2) additional years during which facilities will be required to submit the requested data every April 1 and October 1. 

The CMS will consider the approval of national CAS registries that provide CMS with a comprehensive overview of the registry and its capabilities, and the manner in which the registry meets CMS data collection and evaluation requirements. Specific standards for CMS approval are listed below. Facilities enrolled in a CMS-approved national CAS registry will automatically meet the data collection standards required for initial and continued facility certification. Hospitals’ contracts with an approved registry may include authority for the registry to submit required data to CMS for the hospital. A list of approved registries will be available on the CMS Coverage Web site. 

National Registries 

As noted above, CMS will approve national registries developed by professional societies and other organizations and allow these entities to collect and submit data to CMS on behalf of participating facilities to meet facility certification and recertification requirements. To be eligible to perform these functions and become a CMS-approved registry, the national registry, at a minimum, must be able to:  

1. Enroll facilities in every U.S. state and territory;
2. Assure data confidentiality and compliance with HIPPA;
3. Collect the required CMS data elements as listed in the above section;
4. Assure data quality and data completeness;
5. Address deficiencies in the facility data collection, quality, and submission;
6. Validate the data submitted by facilities as needed;
7. Track long term outcomes such as stroke and death;
8. Conduct data analyses and produce facility specific data reports and summaries;
9. Submit data to CMS on behalf of the individual facilities; and
10. Provide quarterly reports to CMS on facilities that do not meet or no longer meet the CMS facility certification and recertification requirements pertaining to data collection and analysis. 

Registries wishing to receive this designation from CMS must submit evidence that they meet or exceed our standards.  Though the registry requirements pertain to CAS, CMS strongly encourages all national registries to establish a similar mechanism to collect comparable data on CEA.  Having both CAS and CEA data will help answer questions about carotid revascularization, in general, in the Medicare population. 

The CAS for patients who are not at high risk for CEA remains covered only in FDA-approved Category B IDE clinical trials under 42 CFR 405.201. 

The CMS has determined that PTA of the carotid artery concurrent with the placement of an FDA-approved carotid stent and an FDA-approved or –cleared embolic protection device is not reasonable and necessary for all other patients. 

5. Concurrent with Intracranial Stent Placement in FDA-Approved Category B IDE Clinical Trials 

Effective November 6, 2006, Medicare covers PTA and stenting of intracranial arteries for the treatment of cerebral artery stenosis ≥50% in patients with intracranial atherosclerotic disease when furnished in accordance with the FDA-approved protocols governing Category B IDE clinical trials.  CMS determines that coverage of intracranial PTA and stenting is reasonable and necessary under these circumstances.  

C. Nationally Non-Covered Indications 

All other indications for PTA with or without stenting to treat obstructive lesions of the vertebral and cerebral arteries remain non-covered.  

All other indications for PTA without stenting for which CMS has not specifically indicated coverage remain non-covered. 

D. Other 

Coverage of PTA with stenting not specifically addressed or discussed in this NCD is at local A/B MAC discretion.