New evidence suggesting that the benefit of ticagrelor outweighs the risk when added to aspirin for 30 days after a mild-to-moderate ischemic stroke or high-risk transient ischemic attack (TIA) has been reported from a secondary analysis of the THALES trial.
The new analysis was presented at the recent European Stroke Organisation Conference (ESOC) and published online in Stroke on September 3.
The main THALES trial, published last year, showed that ticagrelor plus aspirin reduced the risk for stroke or death compared with aspirin alone in patients presenting acutely with stroke or TIA. But the combination also increased the risk for major hemorrhage, leading to some uncertainty on how the results should be translated into clinical practice.
“The trial’s main results have been difficult to interpret,” said lead investigator Claiborne Johnston, MD, PhD, Dell Medical School, University of Texas, Austin.
“The primary efficacy outcome (stroke or death) included hemorrhagic stroke and hemorrhagic cause of death but these events were also included in the primary safety outcome of severe bleeding so there has been double counting of these events on both sides,” he explained.
Another thing that has caused confusion is the emphasis on relative risk, he noted. “There has been a real reduction in ischemic stroke and a real increase in hemorrhagic stroke but if we just look at relative risks, we don’t know how to balance the two.
“So, in this analysis, we have disentangled the outcomes and look at the disability of the hemorrhage and of the ischemia and show it by absolute differences. This was not a predefined analysis, but it was requested by the regulatory authorities, so it was guided by their recommendations,” he added.
Johnston reported that the original trial results showed an approximate 1% reduction in the composite of stroke and death (primary efficacy outcome) and a 0.4% increase in GUSTO severe bleeding (the main safety outcome).
Those outcomes were “entangled,” he said. When they are disentangled, “We see an approximate 1.2% absolute reduction in major ischemic events and a 0.3% increase in severe bleeding. That is a 4 to 1 ratio for benefit to risk.”
Because a major ischemic event may not exactly equal a severe hemorrhage in terms of disability, the investigators also report an analysis of these events taking into account various disability levels associated with them, and they found that the ratio of benefit to risk stayed the same at about 4 to 1.
“If we defined disability as a modified Rankin Scale (mRS) score of 2 or greater, then ticagrelor was associated with a reduction of disabling ischemic events of 0.8% with an increase in disabling bleeds of 0.2%. And if we define disability as mRS of 3 or greater (so a moderate to severe disability) then it’s still that same ratio — 0.7 vs 0.18,” he said.
“While these results from this new analysis are not that different from the original results, they clarify that while the benefit comes with a price, the benefit exceeds that price, even when you consider disability in the equation,” Johnston commented.
The researchers also looked at various subgroups using these new criteria and found no significant interactions.
“There was no subgroup in which ticagrelor was definitely more superior than in the whole population and there [were] no subgroups where you would definitely not want to treat with ticagrelor,” Johnston said.
But he pointed out some trends in the subgroups. “There are always trends. If you look at enough subgroups you are going to find trends so that shouldn’t be surprising,” he commented.
One subgroup that showed a trend toward a lower net benefit of ticagrelor was in those aged over 75 years. “There was a greater risk of hemorrhage in patients older than 75 years without the same increase in benefits. The net benefit was a little less in that group, but the interaction was not significant,” Johnston reported
Conversely, patients with a prior ischemic stroke or TIA showed a trend toward a greater net benefit. “They had a disproportionally greater net clinical benefit, but not so much that one shouldn’t treat those without a prior ischemic stroke/TIA, so it doesn’t really dictate a change in treatment,” Johnston noted.
Full results of the new analysis show that among the 11,016 patients in the trial a major ischemic event (the composite of ischemic stroke or nonhemorrhagic death) occurred in 5.3% in the ticagrelor-aspirin group and in 6.5% of the aspirin-alone group (absolute risk reduction, 1.19%; 95% confidence interval [CI], 0.31% 2.07%).
A major hemorrhage (including intracranial hemorrhage (ICH) or hemorrhagic death) occurred in 0.4% of the ticagrelor-aspirin group and 0.1% of the aspirin-alone group (absolute risk increase, 0.29%; 95% CI, 0.10% – 0.48%). Net clinical impact favored ticagrelor-aspirin (absolute risk reduction 0.97%; 95% CI, 0.08% – 1.87%).
Translating the absolute risk differences into patient numbers, treating 1000 patients with ticagrelor-aspirin for 30 days instead of aspirin alone is estimated to result in a reduction of 12 major ischemic events (composite of ischemic stroke and nonhemorrhagic death) and an increase of three major hemorrhages (composite of ICH and fatal bleeding), the THALES investigators report.
The net benefits accrued across a spectrum of demographic and presenting characteristics, they add.
Asked what the clinical implications of these new results would be, Johnston replied: “I think we’re still undertreating patients with antiplatelet therapy after a mild or moderate ischemic stroke or high-risk TIA. I think we need more data to show just how much we are undertreating them, but I think we will discover that it is substantial, and we need to address that as a quality issue now.”
“We have plenty of data that shows that either clopidogrel plus aspirin or ticagrelor plus aspirin should be used in the type of patients that were enrolled in these trials. Hopefully we will start to do this more and we will gather more data on which of these two agents to choose. I don’t think an indirect comparison is the right way to make that determination at this stage.”
Johnston added that future trials should use nonoverlapping safety and efficacy outcomes. “Separating them out just makes sense. It is the way we ultimately consume the data so why not prepare for that when defining the endpoints.”
He also said that trial investigators need to get more accustomed to using absolute risk reductions rather than relative risk reductions when reporting the primary results of a trial. “Maybe we can report both but let’s make it clear that relative risk reductions don’t tell us much when we’re thinking about trade-offs.”
Commenting on this new analysis, Rustam Al-Shahi Salman, PhD, University of Edinburgh, United Kingdom, said: “This type of secondary analysis is really valuable as it is addressing a common problem for those interested in not just protecting patients against one specific event but looking at benefit in the round. So, I think these are really helpful analyses. These are very reassuring findings for patients and clinicians but also for regulators.”
He added that he was hopeful that there would be “a new era of proper examination of net clinical benefit in secondary prevention trials.”
The THALES trial was supported by AstraZeneca. Johnston has received institutional research support from AstraZeneca.
European Stroke Organisation Conference 2021. Presented September 3.
Stroke. Published online September 3, 2021. Full text.