HEAT-PPCI Questions? More Answers from Dr Rod Stables

Shelley Wood

July 10, 2014

LIVERPOOL, UK — Full results for the controversial HEAT-PPCI trial were published last week in the Lancet[1]. Senior investigator for the study, Dr Rod Stables (Liverpool Heart and Chest Hospital, UK), spoke at length with heartwire editor Shelley Wood for our news story, addressing some of the concerns raised about trial design and conduct, as well as some of the proposed explanations for why the results came out the way they did. This interview has been redacted and edited for clarity.

heartwire: Some insist that the HEAT-PPCI results depart radically from previous trials because they showed no bleeding advantage with bivalirudin [Angiomax, the Medicines Company]: What is your response?

Stables: Let me establish two assertions that will find broad approval and that are backed by good evidence. Assertion number one is that with the use of unfractionated heparin [UFH], there is a strong relationship between the administered dose and the threat of bleeding complications. When you administer bigger and bigger doses of heparin, there is more bleeding. The Berger/Blankenship editorial mentions a 19 000 patient-meta-analysis that shows this effect[2]. In this, the dose of heparin varies almost threefold, and that can explain some of the historic bleeding seen in studies using higher doses. The ISAR trial group did a trial, ISAR-REACT 3 , where they looked at bivalirudin vs heparin with a high dose of heparin [140 U/kg] and found this was associated with higher risk of bleeding. Subsequently they did ISAR-REACT 3A , which looked at bleeding rates with a reduced dose of heparin and showed that with the heparin dose of 100 U/kg, the bleeding complications become comparable to bivalirudin. In all trials that have used heparin at a dose similar to HEAT and compared this against bivalirudin, bivalirudin doesn't have a bleeding advantage.

Dr Rod Stables

Assertion number two is about the impact of GP IIb/IIIa inhibitor [GPI] use on bleeding, and here again we have near-universal agreement that more widespread use of GPIs is associated with greater bleeding. This is true when used in combination with heparin or with bivalirudin. If you look at trials where there is comparable GPI use, bleeding rates are essentially identical.

So putting those things together, bivalirudin has the potential to show a bleeding advantage if the heparin dose is high, if there is differential use of GP IIb/IIIa inhibitors, or if both factors are in play.

heartwire: One concern, however, is whether all bleeding events were appropriately identified. Is this valid?

Stables: In HEAT, the patients were managed in an open clinical environment where documentation was recorded by staff not otherwise involved in the trial. All of these records were trawled and reviewed by trial-related staff, but in addition, the perceptions of the trial-related staff were cross-checked against the findings of the hospital clinical coding department, which is responsible for determining reimbursement. In addition to that, we arranged quarterly audits. These were performed by independent medical practitioners on random cross-section samples of the data. These detached and independent individuals reported very high levels of accuracy and completeness. All of these quality-control measures were reviewed by the [data safety and monitoring board] DSMB. We had one or two instances where the clinical coding department did identify bleeding that had otherwise been missed by the trial staff, but the published data reflect the totality of the effort, and therefore I put it to you that reported bleeding is not just the perceptions of trial staff.

It is impossible for me to give you a 100% guarantee that somewhere an event might have occurred and not been reported, but the chances of that are relatively low. More important, the subtext of the question is: do I believe in any way that some kind of malevolent and systemic bias was applied to selectively amplify bleeding in one group or suppress it in another or have a modest combination of both? I totally refute that assertion.

heartwire: A similar question has been raised about how the trial defined and adjudicated reinfarctions and reinterventions. Please address.

Stables: The reason this question is raised by lots and lots of people is that, if you think of other trials with softer outcomes where [target lesion revascularization] TLR is used, it is possible that the decision to return to the lab or the decision to perform a PCI on an area of perhaps marginal imperfection is to some extent a value judgment. But in HEAT, fortunately, that was not the case. Almost all the outcome measures reported in HEAT are overt clinical events substantiated by important and unequivocal clinical findings. TLR of what might be called 'value judgment' could only be questioned in a single case.

heartwire: But physicians making the decision to send a patient back to the cath lab were aware of what treatment the patient had received. Could the open design have had an impact on their decision-making?

Stables: That's definitely a theoretical possibility, but merely visiting the cath lab would not have been an outcome-measured event. You'd have to visit the cath lab and have a finding that then persuaded you to perform repeat intervention in the original target lesion, and these events were the subject of blinded adjudication. In almost every case, the findings were of unequivocal occlusion with thrombosis of the stent itself.

Don't forget we are talking here about a clinical story that starts with pain; it is then supported by new ECG change, which drives a trip to the cath lab with a new angiographic finding, followed by an intervention. The strength of the adjudication depends on obtaining what you might call the 'full house.'

The strength of the adjudication depends on obtaining what you might call the 'full house.'

There is a need for an international agreement of the framing of this definition of reinfarction in the acute phase . . . but we are very confident that our approach to the definition of reinfarction, published in detail in the online annex, is robust, mirrors clinical common sense, and will find near-universal approval with practicing cardiologists.

heartwire: Several experts have suggested that the early stent-thrombosis risk with bivalirudin could be reduced by using a prolonged infusion time. Thoughts?

Stables: I do believe that it would be possible to use bivalirudin on a regular basis and achieve results as good as heparin, and there are a number of ways in which this could be achieved. One option would be to coadminister heparin to provide additional antithrombotic coverage after the completion of the procedure, and another method would be to extend the bivalirudin infusion.

People are putting forward ways to battle to increase costs to improve the performance of the costly drug to match the one that costs pence. It's madness.

For the latter approach, the exact ideal dosing regime has never been established in proper randomized trials, but the current suggestions are two to four hours at the PCI procedure dose rate. If you use this strategy to make bivalirudin as safe as heparin, the cost differential of the drugs increases rapidly, because you will be forced to open additional vials of bivalirudin to maintain that infusion. So instead of a cost differential of 400-fold, this could rapidly become 800-fold or 1200-fold, depending on the body weight of the patient and the duration of the infusion. Normally in medicine the reverse argument is applied, if you have drug that is incredibly cheap but marginally less effective, you battle in every possible way to find ways to improve the performance of the cheap drug. Here people are putting forward ways to battle to increase costs to improve the performance of the costly drug to match the one that costs pence. It's madness.

heartwire: What about the high rate of radial procedures in the trial. Was this a key driver of the results?

Stables: This merits attention, and I hope this will be the subject of one of our more detailed secondary analyses. What the published Lancet paper shows is that heparin advantage appears to present for both radial and femoral cases, although you can't make any statistically significant statements about that. In HEAT, the majority of cardiologists are default radial operators, and most completed greater than 90% of all procedures by the radial route. When these operators perform a femoral case, it is usually for specific clinical reasons, and those reasons — in many cases — are associated with greater risk — cardiogenic shock, preexisting bypass grafts, and/or need for hemodynamic support.

heartwire: Another theory has been that the advantage of bivalirudin over heparin in earlier studies is due in part to the fact that they were done in a time of less potent antiplatelet therapies. In HEAT-PPCI, ticagrelor [Brilinta, AstraZeneca] or prasugrel [Effient, Lilly/Daiichi Sankyo] were used in about 90% of patients. Is that important?

Stables: I don't think it's possible to draw a conclusion on that from the HEAT data. The newer P2Y12 agents when compared with clopidogrel are associated with increased bleeding themselves. In their editorial, Berger/Blankenship explicitly ask whether or not these agents play a role in avoiding what they describe as late-catch-up stent thrombosis with heparin. Obviously, none of us know the truth and can only guess as to the etiology, but it is feasible that those events could actually be related to and perhaps secondary to bleeding. One of the great difficulties we face in this arena is that ischemic events beget bleeding and bleeding begets ischemic events.

heartwire: How so?

Stables: When a patient experiences an ischemic event, like stent thrombosis and reinfarction, this may prompt a chain of events. This may include a repeat emergency angiographic procedure, the use of additional antithrombotics, including systemic GP IIb/IIIa inhibitor agents, and it's very easy to see how these events could be associated with the development of some form of bleeding. But the reverse is also true. If a patient who doing quite well from a coronary point of view experiences a bleed, and if the bleed is catastrophic, they may become hypotensive, and this may promote thrombotic events. If they receive autologous blood transfusion products, this is also well known to be prothombotic to the patient with recent stents. Or it may be just that the investigator backs off on the oral antiplatelet therapy in an effort to reduce future bleeding and in so doing increases the rate of stent thrombosis. So it's difficult to know what in English law is called the 'chain of causation,' and it may be that so-called late-stent-thrombosis events could actually be secondary to bleeding.

heartwire: Are you planning to explore this further?

Stables: Something we did in HEAT that I hope to provide more information on at a later date is that when events were adjudicated, all events in the same patient were grouped, and the adjudication team, in addition to making a ruling as to whether an event had or had not occurred (according to the strict trial definitions), were additionally asked to provide a temporal sequencing of events so that all of the patient's events are classified in a 1-2-3-4 order. Later I'm hoping to try to shed some light on the relationship or the chain of causation between ischemia and bleeding and vice versa.

heartwire: Many people are saying we'll need another, multicenter study, a 'HORIZONS 2,' to replicate these findings, because these are from a single, UK center. How do you respond?

Stables: HEAT mirrors expected findings that can be inferred from the existing data and is not some unique venture that is at odds with the existing wisdom. It fits very nicely, and to me, it completes the picture rather than opens the question. When people say it's a single-center trial, I say: identify for me the specific concerns that you have and we'll discuss them and I will attempt to allay your concerns or acknowledge them. . . . Many of the limitations traditionally associated single-center trials are not material or legitimate concerns in HEAT.

heartwire: Critics of the trial have questioned whether HEAT's design and, in particular, its delayed informed-consent process were ethical. How does this debate relate to the overall trial findings?

Stables: I believe that it may be possible for some people to entertain two debates at the same time. They may be prepared to consider the clinical implications and ethical questions separately. These individuals may wish to debate the ethics but still be prepared to accept the clinical findings.

These individuals may wish to debate the ethics but still be prepared to accept the clinical findings.

By the same token, I also accept that some individuals — if they have sufficient ethical concerns — will find themselves unable to read, review, contemplate, or act on the scientific clinical findings. I understand that position and accept the implications.

heartwire: What will be the impact of HEAT-PPCI on future STEMI studies?

Stables: Both editorials picked up on aspects of HEAT that were original, using words like 'unparalleled' and 'ground-breaking.' I am delighted. This was a totally novel concept of trial design, strategy, and conduct. Delivery demanded a major effort and presented a range of completely new and difficult challenges. I am grateful for the cohesion, teamwork, discipline, and sustained support of so many different professional groups in the amazing hospital where I work.

Normally, when you consent a patient for a trial, you are, as an investigator, thinking, is this a reliable subject? Will he/she maintain contact? Will he/she comply? Will I lose touch with them in a year? In HEAT, people were randomized without any such consideration or 'safeguard.' We had to find ways to work with tourists from other nations or vagrants of no fixed abode. We even had one person who was living under an alias, and it was like some kind of detective novel to try to get their follow-up. So there is a whole range of issues and challenges if you want to pursue this. We are actually working on a paper that deals with the practicalities of how you might go about 'pulling off' a trial of this type.

heartwire: Are you concerned that the 'ethical concerns' raised over HEAT, particularly early on, were in part an effort to deflect from the main clinical trial results?

Stables: I'm sure that that must be true for a small minority of individuals, but I think it's difficult to wish to take that position. I believe it is important, in humanity, to respect peoples' ethical or religious or other important views and never seek to denigrate or dismiss those views on the basis of supposition or theory. So I will try to never take that view, so I can at least participate in the ethical debate in an open and collaborative manner.

All of our news and features coverage of HEAT-PPCI can be found here .

Stables disclosed receiving grants from the Medicines Company and grants and personal fees from AstraZeneca during the conduct of the study. All other authors had no conflicts of interest.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: