Some New 'Intriguing' Data: HORIZONS AMI at Three Years

June 17, 2011

June 16, 2011 (New York, New York) — Final three-year results from the HORIZONS AMI trial have been published online June 13, 2011 in the Lancet, showing a continued mortality benefit of bivalirudin compared with heparin plus GP IIb/IIIa blockade and also suggesting "intriguing" late benefits on stent thrombosis and repeat MI.

The results from the stent part of the trial suggest that paclitaxel-eluting stents are just as safe long-term as bare-metal stents with regard to stent thrombosis, although rates with both types of stents were high (almost 5%). Lead investigator Dr Gregg Stone (Columbia University Medical Center, New York) described the main benefit of the paclitaxel stent–-the reduction in target vessel revascularization--as "modest," adding that its use should probably be targeted to those at high risk of restenosis. "The important message from these results is that is that the drug-eluting stent looks just as safe as a bare-metal stent, so now we can decide who actually needs them."


The HORIZONS AMI trial randomized 3602 STEMI patients within 12 hours of symptom onset who were undergoing primary PCI to treatment with heparin plus a GP IIb/IIIa inhibitor or to bivalirudin alone. The two primary end points of the study were major bleeding and net adverse clinical events at 30 days, both of which were significantly reduced in the bivalirudin group. In addition, there were significant reductions in rates of death from cardiac causes and death from all causes, which continued to be seen at one- and two-year follow-up.

The trial also randomized patients to a paclitaxel-eluting stent or a bare-metal stent, with results at one and two years showing a reduced rate of target lesion revascularization procedures in the paclitaxel-stent group, with nonsignificant differences in rates of stent thrombosis and other adverse ischemic events.

Bivalirudin Continues to Impress

At three years, the results look broadly similar to the one- and two-year results, in that there were continued reductions in all-cause and cardiac mortality with bivalirudin. However, in contrast to previous results, by three years, there appeared also to be a reduction in repeat MI, and even stent thrombosis was showing a trend toward benefit in the bivalirudin group.

HORIZONS AMI: Antithrombotic Results at Three Years

Outcome Bivalirudin (%) Heparin+GP IIb/IIIa blocker (%) HR p
All-cause mortality 5.9 7.7 0.75 0.03
Cardiac mortality 2.9 5.1 0.56 0.001
Repeat MI 6.2 8.2 0.76 0.04
Stent thrombosis 4.5 5.1 0.89 0.49
Major bleeding 6.9 10.5 0.64 0.0001

Stone commented to heartwire : "The three-year results show that the reduction in mortality with bivalirudin is robust. The curves are not just parallel; they are diverging, if anything. So I think we can now say for sure that this is a clinically important finding."

He claims that most people should now accept that this mortality reduction is real. "This was a large randomized trial that has shown a sizable mortality reduction in both the short and long term. We have to be a little cautious, as the trial was not powered to show a difference in mortality, but other trials have also shown a reduction in bleeding to be associated with reduced mortality, and the mortality reduction with bivalirudin in PCI has been replicated in the nonrandomized PREMIER registry in >125 000 patients."

MI and Stent-Thrombosis Data Need Confirmation

On the repeat-MI and stent-thrombosis results, Stone was more cautious. 'We showed for the first time in these three-year results that there was less late MI and late stent thrombosis in the bivalirudin group. These are new findings and need replicating to confirm that they are not spurious. I'm not ready to consider these as real findings yet."

He added: "In the first four to five hours we saw an increase in stent thrombosis with bivalirudin, but by 30 days there was catch-up, also seen at one year. And now at three years there is a suggestion of benefit in stent thrombosis in the bivalirudin group. I'm not sure what to make of that and would prefer not to speculate about why that may be occurring."

In an accompanying editorial [2], Dr Debabrata Mukherjee (Texas Tech University Health Sciences Center, El Paso, TX) says the increasing difference in mortality over time between bivalirudin and heparin plus a GP IIb/IIIa inhibitor is "striking" and "supports the use of bivalirudin as the preferred antithrombotic agent."

To heartwire , Mukherjee commented: "The mortality benefit at three years is slightly larger, but that is quite difficult to explain for a therapy just given acutely. So I am more comfortable with [saying] consistent [with results of previous years]." He also found the stent-thrombosis data difficult to explain. "It could be a statistical anomaly, or there could be some scientific reason, such as better endothelialization of the stent. It would be interesting to use imaging techniques to look more closely at this."

Kaul Points Out Caveats

But pouring cold water on all the enthusiasm for bivalirudin, Dr Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA) noted some caveats of the trial results, which he maintains should cause the mortality finding to be regarded as hypothesis-generating. "Because of the open-label design, which precludes definitive conclusions due to inherent bias, the study not being powered for mortality assessment, and the fact that the results were not adjusted for multiple comparisons, the possibility of a false-positive finding due to play of chance cannot be ruled out," he claims.

The FDA is unlikely to view it as sufficiently credible enough to allow a mortality claim.

Kaul also pointed out that bivalirudin is not approved for primary PCI either by the European Medicines Agency or by US FDA. "Regardless of the favorable three-year-outcome data in HORIZONS, the FDA is unlikely to view it as sufficiently credible to allow a mortality claim," he says. But Stone replied that "bivalirudin is labeled for all cases of PCI in the US--so a specific STEMI label is not required (and thus hasn't been requested)."

Kaul also questions the HORIZONS authors' explanation that the mortality reduction was caused by the lower bleeding rate with bivalirudin. He told heartwire : "The precise mechanism explaining the link to mortality cannot be elucidated from the current data. The authors suggest a possible association with decreased bleeding, but in ACUITY a significant difference in bleeding at 30 days did not translate into a mortality advantage at one year for bivalirudin. This raises the question of association vs a causal relationship." But Stone responds: "A new analysis that we will be presenting soon demonstrates that >70% of the reduction in mortality due to bivalirudin is attributable to decreased bleeding."

Paclitaxel Stent: Use in Patients at Highest Restenosis Risk

In the stent part of the trial, the three-year results continue to show a lower rate of ischemia-driven target lesion revascularization, with no significant differences in the rates of death, repeat MI, stroke, or stent thrombosis. Stent thrombosis was high in both groups.

HORIZONS AMI: Stent Results at Three Years

Outcome Paclitaxel stent (%) Bare-metal stent (%) HR p
Ischemia-driven target lesion revascularization 10.2 14.8 0.67 0.006
Stent thrombosis 4.8 4.3 1.10 0.63

Stone commented to heartwire : "The major findings from the stent part of the trial are that the paclitaxel stent looks safe at three years. There have been concerns about late stent thrombosis with drug-eluting stents in STEMI patients. But we did not see any difference in this regard between the paclitaxel and the bare-metal stent, so that is reassuring. However, rates with both stents were high, approaching 5%. This is very concerning, but it does not appear to be an issue specific to drug-eluting stents."

He added that most people would expect the drug-eluting stent to be effective in reducing target vessel revascularization, as these stents are known to reduce restenosis. "There may have been some concerns about catch-up in this end point, but that was not seen. However, I would say that the 4.6% absolute reduction in target vessel revascularization with the paclitaxel stent, even though highly significant, was modest, so I think we have to stratify patients as to who is going to get the most benefit from a drug-eluting stent. Patients who have numerous risk factors for restenosis will have greater benefit from a drug-eluting stent, whereas it is probably not worth using such stents in nondiabetic patients with simple lesions."

Mukherjee said the high rate of stent thrombosis in this trial was "worrisome." He told heartwire : "We have seen similar results in other studies with different stents in the STEMI population. This is telling us that we need to do something extra for these patients--possibly use stronger antiplatelet agents. The primary-PCI population will be a good target group for both prasugrel and ticagrelor; both of these have shown large reductions in stent thrombosis compared with clopidogrel in such patients."

Stone has been a consultant for Osprey, Reva, Merck, CoreValve, Boston Scientific, Abbott Vascular, Xtent, Edwards, BTI, Asten Biopharma, ATI, SB Medical, Evalve, AstraZeneca, Prescient, Eli Lilly, Bristol-Myers Squibb, Biosensors, Otsuka, the Medicines Company, Ortho-McNeil, and Gilead; has received research grants from TherOx, the Medicines Company, Abbott Vascular, Atrium, Boston Scientific, Volcano, and InfraReDx; has received honoraria from Edwards and Vascular Solutions; and has equity in CoreValve, Savacor, Biostar I and II funds, Caliber, FlowCardia, MedFocus I, II, and Accelerator funds, Ovalum, MediGuide, Guided Delivery Systems, Arstasis, Micardia, and AccessClosure. Disclosures for the coauthors are listed in the paper. Mukherjee has published several textbooks of cardiovascular medicine and receives royalties from the publishers.