Prof Dominick J Angiolillo (University of Florida College of Medicine-Jacksonville, Jacksonville, FL, US) discusses the findings of the TWILIGHT study. 

The TWILIGHT study compared the use of ticagrelor alone versus ticagrelor and aspirin together.

Questions :
1. Can you remind us of the TWILIGHT study and what it showed?
2. Why have you conducted this sub-analysis in high risk patients with diabetes?
3. Can you describe cohort size, how they were randomised and your findings?
4. How should these results be interpreted?
5. Does this data justify the use of aspirin?
6. What further research should be conducted in light of this data?

Filmed remotely from Jacksonville, Florida.
Interviewer: Victoria Perroud

Transcript Below:

Question 1 : Can you remind us of the TWILIGHT study and what is showed ? 

[Angiolillo] As a way of background, we know that a dual antiplatelet therapy with aspirin and a P2Y12 inhibitor is associated with a reduction in ischaemic events. However, we know that with the prolongation of dual antiplatelet therapy, there's also an increased risk of bleeding. And we also know that bleeding is associated with adverse prognosis. Therefore, there is a need to define antithrombotic strategies associated with reduced bleeding. And amongst these, a strategy of dropping aspirin after a brief period of dual antiplatelet therapy has emerged. So, this was the foundation for developing the TWILIGHT trial, which was a prospective, randomised, double-blind study in which patients undergoing high-risk PCI were enrolled, and more specifically, high-risk PCI was defined by the presence of one clinical, and one angiographic criteria. And these patients meeting study entry criteria were treated for three months with dual antiplatelet therapy with aspirin and ticagrelor, and patients who were compliant to treatment and who had not experienced an adverse event, either ischaemic or bleeding, were then randomised in a double-blind fashion to ticagrelor plus aspirin, or ticagrelor plus placebo for a period of 12-months after this three-month timeframe. With regards to the key outcomes, the primary endpoint of BARC two, three, and five bleeding was significantly reduced in the ticagrelor monotherapy arm. This was a finding that was shown in all, with all exploratory bleeding definitions, including the more severe bleedings of BARC three, or five. Obviously, with the discontinuation of aspirin, there was also a necessity to evaluate if this occurred at the expense of an increased risk of ischaemic events, and therefore the secondary endpoint of non-ischaemic events was also evaluated and also met. In other words, there were no differences in the composite of death, MI, or stroke. So, this was the overall findings of the TWILIGHT trial. And, very exciting results. And, based on these findings there was interest in evaluating outcomes in specific cohorts of interest, such as patients with diabetes. 

Questions 2 : Why have you conducted this sub-analysis in high risk patients with diabetes?

[Angiolillo] Well, there are a number of reasons. One, patients with diabetes represented one of the pre-clinical entry criteria for the trial. Two, we know that patients with diabetes represent an emerging cohort of patients that we encounter in clinical practise. As we know, the prevalence of diabetes continues to grow, and expected to grow even more so in the years to come. Therefore, a lot of interest in this patient population. And last but not least, diabetes as a cohort is known to be at high-risk for bleeding, but especially ischaemic events. And therefore, understanding if this strategy of dropping aspirin not only would reduce bleeding, but we really wanted to make sure that there was not an increase in thrombotic complications. 

Questions 3 : Can you describe cohort size, how they were randomised and your findings? 

[Angiolillo] So, in the trial, as already mentioned, there were, all patients needed to adhere to treatment, and ultimately, and not have an ischaemic or bleeding complication. And of the diabetic cohort that did reach this criteria, there were 2,620 patients who were ultimately randomised, and represented the cohort of interest for our analysis. And what we observed in the trial, or better off, in this analysis from the overall trial is that bleeding was significantly reduced. And this not only according to the primary safety endpoint of BARC two, three, or five, but also using every other bleeding definition, a finding that was consistent also with the overall trial results. But what was also interesting is that we found that there was a numerical reduction in ischaemic events. So, we're just not powered to show any differences, but clearly there was no increase. If anything, the numbers were numerically reduced. And what was interesting, the components of the composite ischaemic endpoint were, for the most part, numerically reduced. So, overall providing overall reassurance that this strategy was not associated with any harm. Last but not least, in light of the findings, we performed an exploratory analysis, where we took the most severe bleeds, the BARC three or five, and ischaemic events, death, MI, or stroke, and looked at what we call a net adverse clinical event. And in this analysis, we saw that there was a 3.3% absolute reduction in net adverse clinical events. In other words, all the things that we don't want to happen to our patients, with a number needed to treat of 30. So, overall, very reassuring data for the diabetic cohort within the TWILIGHT trial. 

Question 4 : How should these results be interpreted?

[Angiolillo] The results need to be interpreted in the context of a subgroup analysis. So, again, the study was not specifically powered to look at outcomes in the diabetic cohort. So, they need to be interpreted as exploratory, and hypothesis-generating. Nevertheless, we conduct the specific subgroup analysis to provide important insights to cohorts of interest from the trial, and that we encounter in our clinical practise. And the results need to be interpreted as consistent with the overall trial results, without any signals of harm with this strategy in the diabetic cohort. 

Question 5 : Does this data justify the use of aspirin?

[Angiolillo] This is a very good question, and it's important for those who are listening to not interpret that in all patients we should be discontinuing aspirin. Aspirin remains a very important drug, and a very important drug that needs to be used in patients undergoing PCI. And, in the context of a dual antiplatelet therapy strategy for the first several months after a PCI. So, it's important to keep that in mind. The question becomes whether the utility for using aspirin after this three month run-in phase, and again, for patients meeting specific entry criteria who had complied with the treatment regimen, and who had not experienced an adverse event during this run-in phase. So, having said that, after if these criteria are met, one could argue whether there's any added benefit of continuing aspirin. And the answer from the TWILIGHT study is no. And if anything, there's clearly a signal, more than a signal, that you can reduce bleeding significantly, including the big bleeds by eliminating aspirin. 

Question 6 : What further research should be conducted in light of this data?

[Angiolillo] Well, there's already a number of ongoing studies of stopping aspirin in different patient populations, and also using, on the background of different P2Y12 inhibitors. So, I do believe that with the gathering of all this information, we will have a better understanding, and a better definition of the safety and efficacy of this approach, which, thus far, has showed rather consistent findings. So, I think that the next step is hopefully guideline writing committees will be able to collectively analyse this information, put it together, and formulate specific recommendations for the practising clinician.