DANAMI-2 results picked over in the New England Journal of Medicine

Shelley Wood

December 02, 2003

Tue, 02 Dec 2003 20:30:00

Boston, MA - Results of the Danish Multicenter Randomized Study on Thrombolytic Therapy versus Acute Coronary Angioplasty in Acute Myocardial Infarction (DANAMI-2) trial came under fire this past week in the form of several letters to the New England Journal of Medicine. The DANAMI investigators, for their part, defend their findings.

The DANAMI-2 results first appeared in the Journal on August 21, 2003, as reported by heartwire ; the trial compared a strategy of hospital transfer for PCI with on-site fibrinolysis in 1572 AMI patients in Denmark. First presented at the 2002 ACC meeting, the DANAMI-2 results showed primary angioplasty to reduce the primary end point of death, reinfarction, or disabling stroke, compared with fibrinolysis, a finding driven by a 75% reduction in clinical reinfarction in the PCI-treated patients.

Disputing DANAMI-2

In one letter published in the November 27, 2003 issue of the Journal, Dr Kevin S Channer (Royal Hallamshire Hospital, Sheffield, UK) states that the choice of the primary end point "biased the results against thrombolysis," since rates of reinfarction and stroke would be higher after fibrinolysis than they would after angioplasty. Channer also asserts that the trial excluded patients who would have had a higher mortality rate following intervention, such as diabetics.

Channer, as well as Drs Paul W Armstrong (University of Alberta, Edmonton) and Elliott M Antman (Brigham and Women's Hospital, Boston, MA) in a separate letter, point out that only 37% of patients screened were included in the DANAMI-2 trial, raising questions about whether the excluded patients might have fared better with fibrinolysis. Patients deemed too "high risk" during ambulance transfer were not included, for example.

To heartwire , Armstrong pointed out that DANAMI-2 was stopped early, another factor that might have influenced the balance of benefit between the two treatment groups. "Every time you stop a trial early because of efficacy, you overestimate benefit, and then you begin to ask about inclusion/exclusion criteria. And you begin to see that if the patients were too sick, they didn't get entered in DANAMI-2, and if they had shock or pulmonary edema, or it was thought unsafe to transfer them, they weren't entered, so then you say, well how does that apply to patients I would transfer in a community hospital? And what that then influences is the general applicability of these results."

Armstrong and Antman note in their letter that the antithrombotic regimen used in DANAMI-2 exceeded the recommended dose set out in the ACC/AHA guidelines. As well, although patients with contraindications to fibrinolysis were excluded from the trial, patients who had had a previous stroke were not. "Prior stroke is a risk factor for intracranial hemorrhage, which just happens to be a major part of the end point of the DANAMI-2 study," Armstrong elaborated. "All I'm saying is that it makes counting intracranial hemorrhage a very tricky business."

The inclusion of previous stroke patients is "especially germane," Armstrong and Antman argue, because more patients in the fibrinolysis arm had previous stroke (4.0% vs 2.7%; p=0.06). In DANAMI-2, stroke occurred in 2.0% of the fibrinolysis group.

In addition, if a patient randomized to the fibrinolysis arm of the trial then had recurrent ischemia or reinfarction, the study protocol specified that this patient should receive repeat fibrinolysis. Armstrong told heartwire , "If you look at the study design, you see that a patient who had recurrent ischemia or a problem was to receive more lysis and shouldn't get PCI. But if you get lysis, you have a greater chance of intracranial hemorrhage and bleeding and a less good result in terms of reinfarction. However, most contemporary studies and practices, such as the CAPTIM trial, employ a crossover to PCI for patients who need rescue reperfusion or have recurrent ischemia."

A third letter, from Dr Azfar G Zaman (Freeman Hospital, Newcastle, UK) suggests that the reduction in reinfarction seen in the PCI-treated patients could have stemmed from the fact that patients treated with fibrinolysis received aspirin only, whereas the PCI-treated patients received aspirin plus a thienopyridine for one month. Since the study follow-up period was only 30 days, the reduction in cardiovascular events in the PCI group could have been due to thienopyridine treatment, Zaman postulates.

The DANAMI-2 investigators respond

DANAMI-2 investigators Drs Henning R Andersen and Torsten T Nielsen (Skejby Hospital, Aarhus, Denmark) respond to the various concerns defending the DANAMI-2 results and trial design. They note that, in the Primary Angioplasty in Patients Transferred from General Community Hospitals to Specialized PTCA Units with or without Emergency Thrombolysis (PRAGUE) trial, both the fibrinolysis and PCI groups received ticlopidine, yet angioplasty was superior in this trial, too.

They counter that diabetic patients were not, in fact, excluded from DANAMI-2 per se, but that metformin-treated diabetic patients were excluded out of concerns over increased risk of contrast-induced renal failure in these patients. They also take issue with the suggestion that an enrollment rate of 37% of patients screened was problematic. "In our opinion, 37% of the total screened population of patients with MI and ST-segment elevation represents a high inclusion fraction for a trial that necessitated the transportation of patients over long distances."

Finally, they refute the notion that the combined primary end point "biased the results against thrombolysis," noting that even if a reduction in reinfarction and stroke had been the only effect of PCI, "such an effect is of benefit to patients."

Andersen and Nielsen concede that their high unfractionated heparin dose and different previous stroke rates between the two treatment groups "may have influenced" the results. "We do not, however, consider the incidence of stroke (2.0%) in the fibrinolysis group to be substantially different from that in previous trials," they conclude.

Replying to the reply

To heartwire , Armstrong responded to some of the arguments used by Andersen and Nielsen, asserting that their reply falls short. Many of Armstrong's concerns stemmed from a careful reading of the DANAMI-2 trial design, which was published separately but almost simultaneously in the August 2003 issue of the American Heart Journal.

"[Andersen and Nielsen] don't actually address the point we raise, which is that if you start skimming the sickest patients off, how can you then say that the trial tells you that you can transfer patients safely, which is what many people have concluded?" Armstrong argues. "You have to read the American Heart Journal paper to see what patients were excluded, and then you're into the issue of how applicable this really is."

As an example, Armstrong notes that bleeding complications are one of the secondary end points listed in the description of DANAMI-2's design, "but if you look at the trial results, you can't find any mention of them." This is important, says Armstrong, since a meta-analysis of 23 PCI vs thrombolysis trials, conducted by Keeley, Boura, and Grines and published earlier this year, showed that major bleeds are more common in PTCA-treated patients than in lysis-treated patients, Armstrong noted, in contrast to intracranial hemorrhage, in which the reverse is true.

"From my perspective, the reason to pay attention to this is that here we have a trial that changed thinking and practice in the minds of many investigators and that was presented in March 2002," Armstrong said. "We finally have a peer-reviewed publication in August 2003, nearly 17 months later, at the same time that the protocol was published in the American Heart Journal. . . . It is a little unusual that you would publish the protocol almost simultaneously, but at any rate it is helpful in understanding this trial, which has been the subject of so much discussion on theheart.org and other places. People think transportation is safe, that this is the right thing to do, PCI is superior: those are the bottom lines. And it's clear when you go through and look at this, that a careful review leads to a rather different interpretation."

Related links

1. [HeartWire > News; Nov 18, 2003]

2. [Interventional cardiology > GuideWire; Sep 24, 2003]

3. [HeartWire > News; Aug 20, 2003]

4. [HeartWire > News; Jan 02, 2003]

5. [HeartWire > News; Sep 13, 2002]

6. [HeartWire > News; Mar 20, 2002]


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.