Dallas, TX - A blistering debate at the American Heart Association Scientific Sessions 2005 on the value of enoxaparin vs unfractionated heparin (UFH) for acute coronary syndromes (ACS) rocked an amused audience with accusations of data dredging on the one side and tunnel vision on the other. The debate pitted Dr Robert Califf (Duke Clinical Research Institute, Durham, NC) against Dr Steven Nissen (Cleveland Clinic, OH).

Drs Steven Nissen (background) and Robert Califf (foreground)

Arguing in favor of enoxaparin, Califf declared, "If I have an acute coronary syndrome, don't just give me enoxaparin, give me a smart doctor who knows how to use enoxaparin appropriately and I'll be in good shape."

But Nissen, feigning sympathy for the investigators on the pivotal SYNERGY and A to Z trials, Califf among them, countered by asking, "What's a clinical trialist to do? You design a massive and costly clinical trial to demonstrate that a more expensive therapy is better than standard therapy, but unfortunately the study backfires miserably and shows that the new treatment is no better than standard therapy and carries substantially greater risk. So what you do is spin the trial . . . and SYNERGY becomes SPIN-ERGY."

Totality of data

Califf relied heavily on a meta-analysis published in the Journal of the American Medical Association and reported by heartwire that showed a significant reduction in the 30-day composite of death or MI and no difference in major bleeding or transfusion rate with enoxaparin in six trial populations, including the "contemporary" SYNERGY and A to Z trials [1].

"In other debates," said Califf, "I've ended up scratching my head about whether it's just a matter of opinions and looking at pretty pictures. In this case, we've actually done the clinical trials, and rather than picking and choosing studies, the most unbiased decisions come from weighing all the randomized trials and examining observational data."

Nissen, however, dismissed Califf's argument, pointing out that the meta-analysis included four older trials that do not reflect the more aggressive and invasive treatment approaches that characterize modern-day treatment of ACS.

"It's not an appropriate meta-analysis because it combines heterogeneous trials with outdated trials," he argued. "Older studies primarily evaluated enoxaparin in medically managed patients." This "dilutes the hazards of enoxaparin," because in the older studies very few patients underwent catheterization.

Califf showed data from the meta-analysis indicating a signal of harm among patients switched from one anticoagulant regimen to another. "In the real world, as we all know, people come into one hospital or are seen by emergency medical services and tend to get a jumble of different anticoagulant therapies. . . . In this case, when we put together all the randomized trials, with a proper subgroup analysis, we see that in those who were started on one of the antithrombin agents and continued on that throughout the treatment phase, we see a very significant reduction in death and MI, absolutely no difference in transfusion, although actually a slight trend toward a benefit of enoxaparin, and then for TIMI major bleeding, you see a slight excess with enoxaparin."

Nissen dismissed Califf's arguments out of hand. "You taught me, Rob, that subset analyses are dangerous, that they're only hypothesis generating, and I think you need to apply the same standard to SYNERGY."

Nissen continued: "This crossover hypothesis is making lemons into lemonade. It's never before been theorized, and it is not particularly biologically plausible. It really is data mining to look for a benefit. At best it is hypothesis generating, and it must be tested in a prospective, randomized trial."

Califf, in turn, pointed to the STEEPLE trial, which suggested dosing of enoxaparin was critical, and the RESCUE trial, stopped due to slow enrollment, in which patients were randomized to either enoxaparin or UFH in the emergency department, with no switching between treatments permitted. "I would argue [RESCUE] is going to provide a test of whether the concepts I've been giving here are actually applicable."

No common ground

In conclusion, said Califf, "We have a lot of data, it's not overwhelming, and I would argue that looking at the bright side of things, we can say that we can reduce death and MI modestly with enoxaparin and if used appropriately—as opposed to Dr Nissen's sour look of the world—we can benefit our patients."

But Nissen was unconvinced, declaring SYNERGY simply did not "pass the scratch-and-sniff test."

He accused enoxaparin supporters of "deferring attention away from SYNERGY with the meta-analysis."

"You can run but you can't hide," Nissen insisted. "This is what actually happened in SYNERGY: There was no difference on the primary end point: [it had] 10 000 patients but a p value of 0.4, and TIMI major bleeding had a p value of 0.008 that favored UFH. I don't know how you can have a more negative trial than this."

Not only was the drug more costly, argued Nissen, who estimated it cost more than two million dollars more than UFH in SYNERGY, it also seemed particularly unsafe in patients with renal insufficiency, a point also acknowledged by Califf. But as Nissen noted, physicians giving enoxaparin may not be aware of a patient's creatinine clearance before it's too late.

Ultimately, Nissen concluded, "Enoxaparin significantly increases the risk of major potential life-threatening bleeds. It significantly increases the cost of ACS treatment. It cannot be given in renal failure, and it is not easily reversed. Accordingly, enoxaparin is contraindicated in ACS until further studies can confirm or refute this crossover hypothesis."

In a final searing repartee, Nissen asked, "So what would you like, Dr Califf? An enzyme bump, to bleed like a stuck pig, or be dead? Choose your poison."

Califf, however, held firm: "If I come rolling in, don't give me this UFH then try to switch, that's for sure, because I know if I get enoxaparin all the way through, I'll have a lower risk of death or MI, I'll have no increase in the risk of transfusion, and I might have a bit more nuisance bleeding, but that really is minor to me compared with reducing my heart-attack risk. Early treatment really matters, and you need to give the same treatment the whole way through."

Bruised egos and an eye to future results

Speaking with heartwire after the debate, Nissen reiterated, "I don't know how you can lose a trial with any more robustness." He added that he himself doesn't use enoxaparin, and it is not used in the coronary care unit of the Cleveland Clinic. "There are obviously differences of opinion, and Dr Califf has looked at the meta-analysis and sees some benefit, but I just don't see it that way."

For his part, Califf told heartwire, "Dr Nissen's notorious selective use of data backed up by theatrics were certainly in evidence in the debate." Califf said he does, in fact, use enoxaparin "as agreed on by a consensus of Duke cardiologists in our ACS algorithm, which has been published and discussed in multiple forums."

Califf also took issue with Nissen's cost estimates for enoxaparin, saying the drug's higher initial cost is ultimately offset by improved long-term outcomes. As he told heartwire, "I am confident that for patients without moderate or worse renal dysfunction, enoxaparin offers better protection against death and MI than UFH during medical therapy and equivalent protection in the cath lab, at a lower overall cost with fewer errors induced by complex infusions."

More generally, Califf told heartwire, "I hope we can move toward a community that constantly examines the totality of data with its many nuances in an approach to understanding how to offer our patients the best treatment, and away from the PT Barnum approach to understanding, so as not to trivialize years of hard work representing the efforts of many investigators."