August 8th, 2011

Will ISCHEMIA Tell Us More Than COURAGE? Part I: Aiming to Beat Bias with Blinding

, and

On August 1, 2011, New York University announced that the National Heart, Lung, and Blood Institute would fund the ISCHEMIA study of an invasive strategy versus optimal medical management in patients with stable coronary artery disease and moderate-to-severe ischemia. We welcome the members of the study Executive Committee, chaired by Dr. Judith Hochman, to answer questions posed by L. David Hillis and Richard Lange, the co-moderators of the Interventional Cardiology blog on CardioExchange. The other members of the committee are William Boden, Gregg Stone, Bob Harrington, Bruce Ferguson, David Williams, and David Maron.

Hillis and Lange: What information will the ISCHEMIA study provide above and beyond what we learned from COURAGE?

ISCHEMIA Executive Committee: COURAGE taught us that as an initial management strategy for patients with stable ischemic heart disease (SIHD) and known coronary anatomy, adding PCI to optimal medical therapy did not reduce clinical events. ISCHEMIA builds upon that information and addresses gaps left by both COURAGE and BARI 2D. If PCI (or CABG in BARI 2D) isn’t necessary as an initial strategy, is it necessary to catheterize patients in the first place? Did the use of cardiac catheterization before randomization result in selection bias such that many patients, including those at the highest risk and those most anatomically suitable for revascularization, did not participate in these clinical strategy trials? In COURAGE, ischemia of any severity — even mild — was an entry requirement. ISCHEMIA is restricting entry to patients with at least moderate ischemia, a more homogeneous, higher-risk SIHD population, who may be more likely to derive benefit from revascularization. In the COURAGE nuclear substudy, there was a trend toward an independent association of less residual ischemia at 6–18 months and lower death/MI rates during follow-up. However, additional data demonstrate that there was no reduction in events with PCI for those with moderate-to-severe ischemia at baseline, without regard to ascertainment of a follow-up stress nuclear scan — hence the need for the ISCHEMIA trial.

ISCHEMIA will include a much higher proportion of DES recipients and provide the opportunity to see if the early greater improvement in angina and need for fewer antianginal medications associated with improved quality of life in the COURAGE PCI recipients, especially in those with greater degrees of angina at baseline, is durable. We will also examine the effect of CABG. Based on lessons from COURAGE, we are prespecifying angina-related quality of life as the major secondary endpoint.

Hillis and Lange: Are you concerned that cardiologists will find it difficult not to recommend revascularization to their patients with ischemia? Will they be blinded to the extent of ischemia?

ISCHEMIA Executive Committee: Yes, we are concerned that some will be biased toward revascularization. We believe that cardiologists will enroll patients for the following reasons: 1) A survey that we conducted of 500 cardiologists from around the world found that the vast majority would be willing to participate, provided that patients at very high risk for death — e.g., significant left main stenosis — would be excluded. 2) Prior to randomization, patients will undergo blinded coronary CT angiography to exclude patients with significant left main disease. 3) Results from COURAGE and BARI 2D demonstrate that it is safe to treat initially with optimal medical therapy alone and to defer revascularization. 4) It may be easier for cardiologists to enroll patients if they — and the patients — don’t see the anatomy. It is possible that this will prevent the well-described “oculostenotic reflex.”

Cardiologists will not be blinded to the extent of ischemia because the test will be ordered for clinical indications, and eligible patients will only be identified after the test results are known.

Hillis and Lange: Who decides the best method of revascularization? Are you encouraging (or mandating) a “Heart Team” approach?

ISCHEMIA Executive Committee: In ISCHEMIA, we are testing the concept of optimal revascularization as well as that of optimal medical therapy. Gregg Stone chairs the optimal revascularization committee, David Williams chairs its PCI subcommittee, and Bruce Ferguson and Philippe Menasche chair the CABG subcommittee. These expert committees of interventionalists and surgeons have developed trial guidelines for site investigators regarding the optimal method of revascularization. We will require that each site establish a Heart Team composed of an interventional cardiologist and cardiac surgeon. In most cases, it is desirable for the local Heart Team to discuss each case after diagnostic angiography and subsequent discussion with the patient to reach a consensus about the best revascularization modality. We recognize, however, that this is not always practical, and in some cases of non-complex coronary artery disease, the performance of ad hoc PCI immediately after diagnostic angiography may be preferred. As a general guideline, PCI is recommended for SYNTAX scores of 0–22, CABG for SYNTAX scores ≥33, and PCI or CABG for SYNTAX scores of 23–32 (if in doubt, CABG is recommended). Heart Team review will be expected for all patients with a SYNTAX score of 23–32.

Find more questions and answers from the ISCHEMIA Executive Committee in Part II: Banking on Eight Years of Equipoise.

2 Responses to “Will ISCHEMIA Tell Us More Than COURAGE? Part I: Aiming to Beat Bias with Blinding”

  1. David Powell , MD, FACC says:

    Maybe we will get this info later…but I am concerned about criteria for” crossover” from the conservative to invasive strategy. This may be fairly high as the cardiologists will know that there is substantial ischemia. What guidelines will be established for this?

  2. Judith Hochman says:

    On behalf of the ISCHEMIA Study Leadership:

    This is a great question that the research team considered carefully. We use the term “crossover” to refer to patients randomized to the conservative strategy (CON) who undergo cath before a primary endpoint event and “inappropriate crossover” to denote the subset of patients who do so in violation of the CON treatment protocol. Cath for refractory symptoms despite maximal medical therapy is an appropriate reason for cath and is, by definition, part of the CON strategy. In contrast, other reasons, including physician preference that do not meet protocol-defined specific criteria for symptoms refractory to OMT are a protocol violation; protocol violations require notification of the IRB. All such crossovers – both appropriate and inappropriate – may potentially reduce the event rate in the CON arm (if revascularization reduces events) and may dilute the power of the trial to test its primary hypothesis; this was factored in to our power calculations. Cath performed after a primary endpoint event does not impact statistical power for the primary endpoint (because the event is already counted) and is not considered a crossover. We have developed a number of measures to limit crossovers:
    1. Prospective patients will be excluded if they have substantial disappointment with their current management of angina (predicting a high likelihood of crossover).
    2. Site investigators will be discouraged from performing routine surveillance stress tests in clinically stable patients; this is consistent with guidelines.
    3. We will develop guidelines for management of patients with worsening symptoms in the absence of an acute cardiac event ; they will be disseminated with frequent reminders. If angina worsens, medical therapy will be intensified and If symptoms are refractory to maximum medical therapy, patients will undergo cath. Site investigators must provide documentation, including current intensity of medical therapy, heart rate, blood pressure, and a repeat Seattle Angina Questionnaire to justify crossover. Sites must call the 24-hour helpline if cath is being considered but is not emergent, and they must complete a checklist.
    4. Rates of crossover at each site will be monitored and sites will be contacted if high rates occur.

    As is common in any strategy trial, the trial leadership in the US and other countries will spend a great deal of ongoing educational effort devoted to this.