The Proper SYNTAX For the GOD Panelists

DrRich | September 17th, 2010 - 8:27 am

Podcast:

In his last post, DrRich suggested that the Guideline Wars (i.e. the bloody battles over who gets to establish the patient-care guidelines that determine which patients will get which medical services, and which medical specialists will get to provide them) are about to enter the Obamacare phase, in which those who make the guidelines will no longer be medical professional organizations, but agents of the federal government. DrRich helpfully labeled the various guideline panels provided for by the Obamacare legislation as the “GOD panels” (for Government Operatives Deliberating), in order to avoid using the more inflammatory “death panels” terminology favored by certain less sophisticated commentators.

In addition, DrRich pointed out that his own tribe of medical specialists – the cardiologists – may perhaps be in a better position than most other physician tribes to manipulate the deliberations of these GOD panelists. The cardiologists would attempt such manipulations, DrRich suggested, by “pre-spinning” certain critical data from clinical trials, before that data is taken up by the government panels.

From their long experience in fighting the Guideline Wars, cardiologists understand that data from clinical science does not invariably lead to a fixed conclusion (as most proponents of evidence-based medicine seem to believe), but rather, can often be shaped into whatever sort of conclusion one might want to reach. Just as different primitive cultures discerned different constellations when they looked up into the same night sky, so will different groups of experts come to different conclusions when they look at the same clinical data.

Accordingly (DrRich submits), cardiologists have already embarked on the task of pre-spinning the data, such that when the GOD panelists look for the first time up into the vast and chaotic sky of clinical evidence, they will have in hand a map of the constellations as seen by the cardiologists.

To illustrate what he means, DrRich calls his readers’ attention to the SYNTAX trial, a clinical trial designed by cardiologists and their industry partners for the purpose of reaching a specific conclusion, but which (unfortunately for cardiologists) reached the opposite conclusion. If the data from the SYNTAX trial should ever fall into the hands of the GOD panelists (or any other guideline panels) in a pristine fashion, it could spell disaster. So the cardiologists have spent nearly two years attempting to make the data say what they want it to say, and today, after continuously massaging the data, issuing press release after press release, making presentation after presentation, and publishing academic paper after academic paper, it would be at least a little surprising if the God panelists, surveying this body of pre-spun data, would fail to produce clinical guidelines which provide the cardiologists at least some of what they’re after.

The SYNTAX trial randomized 1800 patients with complex coronary artery disease (i.e., CAD that produces either significant blockage in the left main coronary artery, or severe triple-vessel disease) to therapy with either bypass surgery or drug-eluting stents (DES), and assessed their long-term outcomes.

In general, patients with stable CAD (i.e., those who are not currently having a heart attack or unstable angina) do just as well with aggressive medical therapy as they do with invasive therapy. People like the ones enrolled in SYNTAX, however, are the exception to this rule. That is, patients with either of these two specific patterns of complex CAD have been shown to have improved survival if they receive bypass surgery. Indeed, these patients represent a virtual “last stand” for cardiac surgeons – they are nearly the only patients cardiologists (at least some cardiologists) still feel obligated to refer for bypass surgery. And, as one might expect, in their decades-long turf war with cardiac surgeons (a war from which they will not desist until they see the great majority of cardiac surgeons seeking jobs as beer vendors at sporting events), cardiologists have long chafed at this singular remaining obligation to refer.

Accordingly, SYNTAX, a study instigated by cardiologists and sponsored by Boston Scientific (a manufacturer of DES), was intended to show that with modern cardiac stents, stenting yields outcomes that are not significantly inferior to bypass surgery in these patients. Specifically, that is, SYNTAX was designed as a “non-inferiority trial.” This was certainly a modest goal – some might say too modest – but a positive result would enable cardiologists (the gatekeepers to all invasive CAD therapy) to simply keep these patients for themselves, just as they now do with all the other CAD patients. The SYNTAX trial asked the question, “Do I really have to refer these patients to the cardiac surgeon?” And the desired answer was, “No, it apparently is not statistically provable that you absolutely have to refer them, one supposes.” Judging from the study design, that answer would have been plenty rigorous enough for the study designers.

Alas, however, when the one-year follow-up data for the SYNTAX study was analyzed, the results turned out to be negative; stenting was significantly inferior to surgery. The endpoint of the study was a composite called MACCE (Major Adverse Cardiac and Cerebrovascular Events), and in this study included death, heart attack, stroke, and the need for more revascularization procedures. The bottom line is that in SYNTAX, the risk of MACCE was significantly higher for stenting than for bypass surgery. The study failed to meet even the modest non-inferiority goal the cardiologists had devised for it.

The one-year data for the SYNTAX trial was published in the spring of 2009 in the New England Journal of Medicine, just as the battle over healthcare reform was taking shape, and nearly six months after Sarah Palin had reported her mysterious vision of death panels. But whether one wanted to call them death panels, GOD panels, or panels of distinguished monkeys, it was clear by the time SYNTAX was published that the government, and no longer the physicians’ own specialist organizations, would be manufacturing all clinical guidelines in the near future. So disaster loomed.

But, the cardiology community quickly rallied, and launched into a concerted effort to spin the results of SYNTAX from a disaster into a victory, or at least, to something akin to victory. And the efforts of the cardiologists in this regard have been impressive over the last 18 months.

Within minutes of the publication of the original SYNTAX article, scores of press releases were launched, and scores of “experts” were dispatched to give interviews, implying that the SYNTAX study was a major, ground-breaking victory for stenting.

For instance, here’s the link to an article in the New York Times (subsequently reproduced in hundreds of newspapers around the country) entitled, “Heart Stents Found As Effective As Bypass For Many Patients.”

And here’s a triumphant press release from Boston Scientific, the study’s sponsor, in which the negative overall results of SYNTAX are buried deep within the 6th paragraph (following all kinds of positive-sounding fluff), and are difficult to locate even if you are specifically looking for them.

And here’s the more-than-triumphant press release from one of the leading clinical sites for the SYNTAX trial, which reads, in part:

“At NorthShore, we experienced stunning outcomes [emphasis DrRich's] in patients whose only option would have previously been bypass surgery,” said Ted Feldman, M.D, F.S.C.A.I., Director of Cardiac Catheterization Laboratory, NorthShore University HealthSystem, and a lead investigator of the trial. “The data in this study will provide cardiologists with additional information as they determine treatment therapy for patients with complex CAD.”

Most remarkable of all, we have the spectacle of the lead author of the SYNTAX paper, Dr Patrick W Serruys himself, telling Heartwire immediately after publication of the paper, that the paper’s concluding sentence (i.e., “CABG as compared with PCI is associated with a lower rate of MACCE at one year among patients with three-vessel or left main coronary artery disease (or both) and should therefore remain the standard of care for such patients.”), is just plain wrong. Serruys declared that this concluding sentence actually “is not the essence of the trial.” He only allowed that concluding sentence to appear in the paper, he said, “because the New England Journal of Medicine wanted something more conservative.” (Apparently, having the paper appear in a prestigious journal overrode the necessity of having the paper accurately reflect what the authors meant to say.)

In any case, Serruys insisted (despite the conclusion expressed in his ink-not-yet-dry paper) that many patients like the ones enrolled in the SYNTAX trial can safely be treated with stents, and indeed, he announced that he and his co-investigators were hard at work teasing apart the SYNTAX data in order to develop a so-called “SYNTAX score,” that would help cardiologists determine which patients they can treat themselves, without referring them for surgery. So indeed, despite the negative results, and despite the conclusion written in their own paper, the SYNTAX trial was immediately spun by key trial participants themselves into a win (while not a complete victory, still a win).

Before his readers come down too hard on the cardiologists for such behavior, DrRich feels obligated to point out a partially mitigating truth. Namely, cardiologists believe to the depths of their souls, notwithstanding the largely negative body of medical literature to the contrary, that stenting coronary artery blockages – in virtually any configuration and any clinical situation – saves lives. And if they haven’t been able to prove that yet, it’s just because of the vagaries of clinical research. One must not let spurious results from imperfect research block the Truth, lest one allow great harm to come to humanity.

The results of the SYNTAX trial must simply be wrong, cardiologists believe, and so they would be gravely harming patients if they did not take whatever steps were necessary to render the results of SYNTAX more favorable to stenting. To do otherwise would cause thousands of clinicians to make inappropriate decisions.

In this way, DrRich believes, cardiologists are no more guilty than are Progressives, another category of humans who believe with their hearts and souls in something that is simply not true (in the case of Progressives, that the great mass of humanity will willingly suppress their own individual interests in favor of the interests of the collective). Like the Progressives, cardiologists are often very nice, well-meaning, sensitive and compassionate individuals, and some of them would even be fun to go out with for a beer.

In other words, DrRich pleads, cardiologists are not being particularly evil in spinning the SYNTAX trial results; they are simply doing what comes naturally, and what they deeply believe to be the right thing.

This is why the SYNTAX investigators were convinced that, buried within the vast body of clinical data the SYNTAX trial has generated, there simply MUST be something useful to cardiologists. Accordingly, the SYNTAX investigators dived head-first into the proverbial room full of manure, enthusiastically digging for the pony which simply must be in there somewhere.

And indeed they quickly found their pony.

And here it is: While a straightforward analysis of the SYNTAX study shows that bypass surgery wins hands down over stenting, if one delves a bit deeper into the data, one finds that one of the components of the MACCE endpoint – the incidence of stroke – was statistically higher among the patients randomized to bypass surgery. One also finds that the incidence of needing revascularization during follow-up was higher in the stent patients. So, taking these two interesting observations together, the cardiologsts have concluded that patients receiving bypass surgery are trading a reduced need for subsequent revascularization for an increased risk of stroke – a bad trade indeed. Therefore, despite the overall results of the study, they have concluded that stents are better than bypass surgery for at least some patients. (And they promised to discover for us, during the data-mining exercise from which their “SYNTAX score” was subsequently invented, which patients those are.)

And this – the reduced incidence of stroke seen in the stent patients – is the basis for the celebratory statements which were issued by the SYNTAX investigators upon publication of their original paper.

DrRich agrees that, as a general proposition, he would probably rather have an extra invasive cardiac procedure than a stroke, and suspects that most people would say the same thing. But before we all buy what cardiologists are selling here, DrRich would like to make a few observations.

First, the results of the SYNTAX trial are the results. Stenting did not meet even its modest non-inferiority endpoint, and it failed to meet it by quite a lot. Once the pre-designated endpoint of a randomized clinical trial is determined, any remaining observations that can be gleaned from the large amounts of data invariably generated by such trials must be viewed as inconclusive, as merely hypothesis-generating. Such observations are not to be regarded as having sufficient statistical surety to vastly change medical practice, or to figure into evidence-based guidelines. So, another clinical study would need to be conducted to prove the hypothesis that strokes are less frequent with stenting than with bypass surgery in patients like these.

Similarly, the “SYNTAX score” – which indeed was generated and subsequently published as a “guide” for cardiologists treating patients with complex CAD, and which is therefore presumably being used today by cardiologists all over the world to select which of these complex patients they can just go ahead and stent rather than refer for bypass surgery – has no business being incorporated into clinical practice. An exercise like this – in which investigators comb retrospectively through the clinical data, selecting out patients who had good results with stenting, then devising a group of characteristics that appears to differentiate them from those who did not – cannot possibly yield a validated, widely-applicable clinical tool. If they want to claim that their SYNTAX score is clinically useful, they need to conduct another randomized clinical trial to test that hypothesis.

Next, and most remarkably, there’s the almost universally-ignored fact (reported by Dr Friedrich W Mohr, co-principle investigator of the trial), that among patients assigned to bypass surgery in the SYNTAX trial who experienced a stroke, nearly half of them had their strokes PRIOR TO SURGERY. What this means is that, in reality, the bypass surgery itself did not cause those strokes, a fact that ought to cause serious damage to the chief assertion of the stent-proponents. This fact alone turns their pony into a pig. The claim that stenting instead of surgery would have avoided these strokes is largely, if not entirely, spurious. Indeed, if anything, the fact that patients “randomized” to surgery apparently had a lot of strokes in the brief period of time between the act of randomization and the surgery itself ought to make one question whether the selection of therapy was really and truly random, or whether, somehow, patients who looked particularly sick got sent preferentially to the surgeons.

And finally, the clinical choice as it has been starkly painted by many proponents of stenting – that the real trade-off in choosing between stenting and bypass surgery in these patients is the choice between the higher risk of stroke with surgery versus the higher risk of revascularization with stenting – is incomplete and misleading. Presenting the choice in this way clearly favors stenting, and this presentation entirely explains the positive press releases and subsequent media coverage of the SYNTAX trial. But this is not a valid comparison of risks for several reasons:

1) As noted above, the actual risk of stroke posed by performing bypass surgery in the SYNTAX trial has been substantially overstated for public consumption (by implying that the surgery caused those strokes, when half occurred prior to surgery).

2) Surgeons in the SYNTAX trial most often did not employ newer techniques now in routine use, such as off-pump surgery and LIMA grafts, both of which can substantially reduce the risk of stroke and other embolic phenomena.

3) Re-occlusion of the involved arteries (which occurred about equally in both groups in this study, and which spokespersons for SYNTAX seem to brush off as not such a big deal), is an entirely different phenomenon in patients who have received DES than it is in patients who have had bypass surgery. After bypass surgery, re-occlusion tends to occur gradually, and the patient generally experiences recurrent symptoms of angina. But in DES, re-occlusion much more commonly occurs acutely, and catastrophically, leading rapidly to permanent cardiac damage and often, to sudden death.

4) Item # 3 might explain why the composite endpoint of “death, heart attack and stroke” was equal in both groups, even though stroke was significantly higher in the surgery group. That is, in order for the math to work out, the remaining dyad of “death and heart attack” necessarily must have been higher in the stent group. As far as DrRich can tell, this point has never been discussed in public.

5) In order to avoid the catastrophic re-occlusions seen with DES, cardiologists now insist that their DES patients take long-term, even life-long, Plavix, a powerful blood thinner. As the purveyor of a patient-oriented website on heart disease, DrRich cannot tell you how many distressed and stented-up patients have written to him with the following lament:

“My doctor put in one of those drug coated stents and has me on Plavix. He says if I stop the Plavix I could die, and won’t let me stop it for any reason. But I need my gallbladder out because I keep having gallbladder attacks, and the last one gave me blood poisoning. My surgeon says I need the surgery but he won’t do it unless I stop Plavix, and my cardiologist says no stopping the Plavix for any reason. What can I do? Can they just take these stents out so I can stop the Plavix?”

There is no easy answer to this question, at least not that DrRich can find. The DES patient commonly is left in the middle of a pissing match between surgeon and cardiologist. The fear of the cardiologist is that when one stops Plavix, there is a risk of sudden, catastrophic thrombosis of the coronary artery. But surgeons simply cannot operate safely on patients taking this drug. Few cardiologists seem to explain this to their patients before placing DES.

And more to the point at hand, none of the cardiologists spinning the SYNTAX trial are explaining to the public the implications of long-term Plavix. Even if their claims that stenting yields significantly fewer strokes turned out to be accurate, the choice here is clearly NOT a simple one between a higher risk of stroke on one hand, and a higher risk of needing “revascularization” (if they survive the re-occlusion, that is) on the other. There’s a lot more to think about than that, and cardiologists who imply otherwise are being either disingenuous, or delusional.

Just last week, SYNTAX investigators reported on the three-year outcomes in patients enrolled in the trial. The results, similar to the one-year outcomes, remain strongly in favor of bypass surgery at three years, and indeed, the incidence of stroke in stented patients has “caught up” with the incidence of stroke in the surgery patients.

This persistently bad news still does not really phase the cardiologists, who are now saying that the results of SYNTAX don’t really apply any more in any case, because drug-eluting stents have been improved since the trial was done.

It would appear that the cardiologists are going for some sort of official announcement to the effect that that the results of SYNTAX are, for practical purposes, indeterminate, and that what is needed is a new clinical trial, in which patients randomized to DES will receive the latest generation of stents. (Since there is a new generation of stents every year or so, this entire process can be repeated as needed until the cardiologists finally get the results they’re looking for, at which point they can declare final victory and stop.)

Summary:

All medical specialists should take a lesson from the cardiologists. In an era in which specialist organizations will no longer be writing the clinical guidelines for their own specialty, it is necessary to aggressively pre-spin any important clinical data upon which the GOD panels will be deliberating.

Considering the SYNTAX trial as a case study, one sees how it is possible to take the most straightforward results from a very straightforward clinical trial and, if not turn a negative outcome into a positive one, at least introduce enough complications, nuance, spin and uncertainty to cause any self-respecting GOD panelist to hesitate in making a definitive pronouncement on those results. Then, if you couple all the uncertainty you’ve created with a loud call for yet another clinical trial – one that will take into account new equipment, new techniques, new scoring systems &c., and that promises to clear up all the confusion you’ve dug up as a result of the last clinical trial – then you stand a decent chance of at least getting a postponement on any new guidelines harmful to your cause.

And this, you neurologists, gastroenterologists, pulmonologists and all you other, less savvy medical specialists, is how one can manage the GOD panels.

5 Responses to “The Proper SYNTAX For the GOD Panelists”

  1. Dr Peter Alison says:

    sadly this article tells it how it is. sadly it will be ignored along with the vast literature that has ever shown a negative outcome with percutaneous intervention. there has never and will never be a satisfactory justification of “non inferiority”. any scientist would publish the etrue difference and let the reader determine whether a measured difference is relevant as preselection of an “acceptable” range is too easy to cloud with bias – innocent or otherwise.

    the only other regret is too few clinicians have the confidence to speak out – partly a fear of the seemingly never ending list of trial pneumonics. little do they realise that the interventionalists are equally poorly informed of the science they love to quote. the only other feature you failed to mention is the “patient preference”. patients almost universally prefer the treatment recommended by their doctor rather than some preconceived notion as evident in the registry arm of the syntax trial. CABG pts did actually better than in the randomised arm and only 55 of pts declined the recommendation for surgery.

    no one wants a CABG but I doubt anyone wants to have a DES either.

    regards
    Peter Alison FRACS

  2. GingerR says:

    I must say that this and your JUPITER study write-up keep me coming back for downloads.

    As someone with an interest in statistics it is amazing to hear from you what isn’t mentioned in the outputs of these studies!

  3. Plas says:

    So, as someone who just had a non-medicated stent placed in the Circumflex artery (80 to 85% blockage – other arteries are 10-25% blocked) I’m in danger of a major coronary event? That;s exactly the opposite of what I was lead to believe – that this stent would prevent a major problem.

    The more I travel down this path (which started with an MD’s observation of an erratic heartbeat, the more confused I’m becoming, especially when I read information such as you present.

    I must add that my situation is a bit more complex, as there is the atrial fib, and some mitral valve prolapse, but still – I’m at a point I don’t know what to think of all of the vast info I keep running into.

    • DrRich says:

      Plas,

      The stent is meant to treat the existing occlusion of the specific blockage that is stented. It obviously cannot affect partial blockages that are not stented. The unfortunate fact is that any plaque can rupture and cause acute coronary syndrome, and this fact probably accounts for the fact that placing stents in stable coronary artery disease does not appear to statistically reduce the overall risk of subsequent cardiac emergencies.

      There may be extenuating circumstances in your case that change this equation. You should talk to your cardiologist about it, and make sure you are both on the same page.

      Rich

  4. Plas says:

    Thanks. I do keep up a running dialogue with my cardiologist – sometimes more than he might like :-). I’m full of ideas and opinions. The more I read about, the more we discuss.

    I appreciate all the additional information I can get, although sometimes the information conflicts.

You can leave a response, or trackback from your own site.

Leave a Reply