DIAS-2: No Benefit of Desmoteplase in Acute Ischemic Stroke

Susan Jeffrey

June 03, 2007

June 3, 2007 (updated June 4, 2007) — Preliminary results of the Desmoteplase In Acute Ischemic Stroke 2 (DIAS-2) trial show no benefit of either of 2 doses of desmoteplase, a novel thrombolytic agent derived from vampire-bat saliva, over placebo.

Mortality was actually increased with the 125-µg/kg dose vs both placebo and the 90-µg/kg dose, although the excess deaths occurred late after treatment and were predominantly nonneurological, and there was no excess in intracerebral hemorrhage (ICH) between groups.

Werner Hacke, MD, from the University of Heidelberg, in Germany, cochair of the steering committee for the DIAS-2 trial, presented the disappointing findings at the 16th European Stroke Conference, in Glasgow, Scotland.

"It simply does not make real sense, what we are seeing here," Dr. Hacke said.


DIAS-2 was a prospective, double-blind, single-bolus study investigating the efficacy and safety of 2 doses of desmoteplase, 90 and 125 µg/kg, given as an intravenous bolus. Two phase 2 trials of desmoteplase, DIAS and DEDAS, had been previously reported and provided encouraging data on both the safety and potential efficacy out to 9 hours after stroke onset with this agent.

Patients were eligible for DIAS-2 if they could be treated within 3 and 9 hours after the onset of stroke symptoms, had a score on the National Institutes of Health Stroke Scale (NIHSS) of 4 to 20, and had a distinct ischemic penumbra of at least 20% established by magnetic resonance imaging (MRI) (perfusion-weighted imaging /diffusion-weighted imaging ) or perfusion computed tomography (CT).

The primary outcome was clinical improvement at day 90, defined as having achieved all of the following: an improvement in NIHSS score of 8 points or more (or an NIHSS score of < 1), Barthel Index score of 75 to 100, and a modified Rankin Scale score of 0 to 2.

On October 25, 2006, Paion AG, in Aachen, Germany, announced that the trial had been halted by the data monitoring committee (DMC) for apparent safety concerns, the precise nature of which the DMC did not disclose. However, on October 30, the trial was allowed to resume enrollment without modification of the protocol, and the data remained blinded. Now it is clear what had given the DMC pause.

Using intention-to-treat analysis, investigators found no significant difference between the groups in clinical response rates, with numbers that contrasted sharply with their previous findings with this agent.

"This is far from our expectations," Dr. Hacke said. "This pattern makes no sense, specifically when we look at what we had seen in the previous randomized trials — everything is turned upside down." When they considered clinical response rates without dropouts, he noted, "The picture changes slightly; now we have an equal response in all treatment arms, but not what we had expected."

DIAS-2: Clinical Response Rates at 90 Days

End Point
Placebo, n = 63 (%)
90 µg/kg, n = 57 (%)
125 µg/kg, n = 66 (%)
Clinical response rate
Clinical response rate without dropouts

Survival was actually lower with treatment in the highest-dose group vs both placebo and the 90-µg/kg dose.

DIAS-2: All-Cause Mortality at 90 Days
End Point
Placebo, n (%)
90 µg/kg, n (%)
125 µg/kg, n (%)
All-cause mortality at 90 days
4 (6.3)
3 (5.3)
14 (21.2)

Of the 14 deaths in the 125-µg/kg group, 10 were considered by the investigators to be unrelated to the treatment, 9 were nonneurological, and 9 were late — that is, occurring more than 10 days after administration of the study drug. There was no excess in systemic bleeding with treatment, however, and although there was ICH in the 2 treatment groups, the rate was not high enough to explain the difference in mortality.

DIAS-2: Intracerebral Hemorrhage at 72 Hours
End Point
Placebo, n (%)
90 µg/kg, n (%)
125 µg/kg, n (%)
ICH at 72 h
0 (0)
2 (3.5)
3 (4.5)

"The next steps will be in-depth per-protocol analysis, PK and antibody assays, and analysis of the placebo response," Dr. Hacke said. They do not yet have the analysis of data on the imaging inclusion or adjudication of the late death cases, he added.

No Good, No Harm?

During the discussion period, an audience member pointed out that while disappointing, these results underline the importance of the need for small trials with positive results to be replicated and suggested that the most likely reason for these results is the play of chance. The excess in deaths is likely to be unrelated to treatment because this was not seen in the prior DIAS trial, and a connection between the 2 is biologically implausible, he said. "This excess of deaths is real, but it's very unlikely to be due to the drug, so the whole result of the 2 studies is a neutral result — no good, no harm," he said.

Dr. Hacke agreed that the most likely explanation is the play of chance but pointed out that given the highly technical nature of the inclusion process in this trial, it would have been difficult to include significantly larger numbers.

What was unsettling here was that there is no clear reason for the negative findings, Dr. Hacke added. "In other trials, we see it and we recognize patterns — an imbalance in stroke severity at baseline here; and there, more bleeding — but it's all not there," he said.

Cochair on the trial was Anthony Furlan, MD, from the Cleveland Clinic Foundation in Ohio. He told Medscape that they are disappointed with the results but need to do further analyses to better understand the findings. "Many of the primary results are at variance with prior trials," he said. "The placebo group had an almost 50% response rate, which is twice as high as the prior 2 trials and makes it difficult to detect any treatment effect with such a small sample size."

They are also awaiting a detailed analysis of the imaging data, comparing outcomes in MR vs CT patients. "After further analyses are completed, we'll decide on next steps," Dr. Furlan said. However, he added, "we remain convinced that penumbral imaging is the most promising means to select patients for reperfusion therapy after 3 hours."

Accumulation of Bad Luck

In a subsequent interview with Medscape, Dr. Hacke said he now feels that the trial was the victim of the accumulation of the play of chance, good outcomes in the placebo group, and completely unrelated events in 1 of the treatment groups.

He has no concern about the increased deaths with the 125-µg/kg dose, he said. "I know every single case; what the condition was, what the cause of death was, and whether or not it was in any way related to the presenting stroke or to the treatment. Most of it is just by chance — a strain of bad luck — rare complications and even rare diseases that occurred more often in the 1 treatment arm."

The trial included small numbers, but the power calculation had been based on the previous phase 2 trials. "The results from the previous trials told us we would be somewhere on the order of a 20% to 25% response rate with placebo, and based on that, and a cautious calculation of what to expect in the treatment arm — about 25% more, or a 50% response — we came up with the power calculation," he said. "Well, we have the 50% in the treatment groups, but we also have 50% in the placebo group."

"So it can happen, but that it happens in 1 trial at 2 ends — adding people whose death has nothing to do with the stroke or the treatment, and adding great responders in spite of no active treatment in the placebo group — that is the puzzling and unprecedented problem."

What happens now for desmoteplase is up in the air, but the investigators are still ready to try again with this drug, Dr. Hacke said. "Everyone is supportive of giving it another try," he said. "If we do a joint analysis of the now 3 trials, then there would still be a hint of efficacy, but we don't know what the companies will decide.

"I tell you, the investigators would buy this drug and do it by themselves," he added.

The study was supported by Paion Deutchland GmbH and Forest Laboratories. Dr. Hacke disclosed he has no stock options and is chair of other ongoing stroke trials, including ECASS III.

16th European Stroke Conference. May 29-June 1, 2007.


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: