Ticagrelor Risky in Diabetic Patients With Stable CAD: THEMIS

But in the THEMIS-PCI substudy, ticagrelor cut ischemic-event risk without upping bleeds in stable CAD patients with prior PCI.

Ticagrelor Risky in Diabetic Patients With Stable CAD: THEMIS

PARIS, France—Use of ticagrelor (Brilinta; AstraZeneca) in patients with stable coronary artery disease and type 2 diabetes significantly lowers the risk of major adverse cardiovascular events when compared with aspirin, according to the results of the THEMIS study. But this comes at the cost of more bleeding.

Presented today during the first Hot Line session of the European Society of Cardiology Congress 2019, and published simultaneously in the New England Journal of Medicine, investigators reported that the primary endpoint—a composite that included cardiovascular mortality, MI, or stroke—occurred in 7.7% of patients treated with ticagrelor and aspirin compared with 8.5% of patients treated with placebo and aspirin (P = 0.04).

The risk of TIMI major bleeding, however, was more than twofold greater among the ticagrelor-treated patients, and the risk of intracranial hemorrhage was also significantly greater. Unless the patient has a high risk of ischemic events, and low propensity for bleeding, use of ticagrelor in stable CAD patients with diabetes carries an unfavorable risk-benefit profile, the researchers said.

Deepak Bhatt, MD (Brigham and Women’s Hospital, Boston, MA), who led the THEMIS trial along with P. Gabriel Steg, MD (Hôpital Bichat, Paris, France), said that when weighing both efficacy and safety, a composite endpoint of net irreversible harm, ticagrelor was no better than placebo. For stable CAD patients with diabetes and without a prior MI or stroke, therefore, “we think the bleeding risks, in general, outweigh the benefits and ticagrelor shouldn’t be used.”

But in a provocative finding, the THEMIS investigators say they have identified a high-risk patient population that does benefit from long-term ticagrelor therapy: stable CAD patients with diabetes who had previously undergone PCI.

In THEMIS-PCI, a substudy of the main trial presented during the late-breaking clinical trial session and published in the Lancet, the researchers looked at clinical outcomes among the more than 11,000 patients who had been revascularized with stents, and in this setting the use of ticagrelor significantly lowered the risk of cardiovascular death, MI, or stroke by 15% when compared with patients treated with placebo (7.3% with ticagrelor vs 8.6% with placebo; P = 0.01).

Use of ticagrelor was associated with an increased risk of bleeding, including a more than twofold higher risk of TIMI major bleeding (HR 2.03; 95% CI 1.48-2.76), but there was no difference in the risk of fatal bleeding (BARC 5) or intracranial hemorrhage between the ticagrelor- and placebo-treated patients with a prior PCI.

In terms of net clinical benefit, an endpoint focusing on the composite endpoint of all-cause mortality, MI, stroke, fatal bleeding, or intracranial hemorrhage, treatment favored ticagrelor in patients with a prior PCI (9.3% with ticagrelor vs 11.0% with placebo; P = 0.005), but not in those without prior stenting.

“In our parsing through the data, it seemed [prior PCI] was the subgroup where the action was, so to speak,” Bhatt told TCTMD. “That is, the efficacy looked better than in the overall trial and the safety looked better, hence the net clinical benefit looked particularly good and not so much in the no-PCI patients. We think if you have had a stent in the past, that means you have passed a bleeding ‘stress test.’ That is, you’ve been on dual antiplatelet before and presumably didn’t have a big bleed.”

A prior PCI, said Bhatt, essentially marks a stable CAD patient at lower risk of bleeding who would benefit from the addition of ticagrelor. To TCTMD, Steg added that patients treated with a drug-eluting stent as opposed to a bare-metal stent had a larger clinical benefit. Like Bhatt, Steg said a prior DES is marker of withstanding protracted dual antiplatelet therapy and such patients could be considered for ticagrelor.

Richard Becker, MD (University of Cincinnati Heart, Lung, and Vascular Institute, OH), a spokesperson for the American Heart Association who was not involved in the study, praised the investigators, noting that THEMIS was a pragmatic trial given that physicians frequently encounter patients with stable CAD and diabetes. “I think we moved the needle in the right direction,” Becker told TCTMD. “The question now is: who are the patients most likely to benefit from the addition of ticagrelor? I think they identified those patients—those with a prior PCI, anywhere from 6 months to more than 3 years—with diabetes and a low risk for bleeding.” 

Stable CAD and Diabetes but No Prior MI or Stroke

THEMIS is a large-scale, international clinical trial that included 19,271 patients with stable coronary artery disease and type 2 diabetes mellitus treated with aspirin and ticagrelor 60-mg twice daily (the dose was lowered from 90-mg twice daily following the PEGASUS-TIMI 54 trial). Importantly, the patients did not have a prior history of MI or stroke, although 58% of patients in both treatment arms had undergone prior PCI. Roughly 56% of patients in the trial had angina and 62% had multivessel CAD. Patient had been diagnosed with diabetes for a median of 10 years.

The reduction in the primary endpoint was driven by reductions in the risk of MI and stroke. For those treated with ticagrelor, the risk of MI was reduced 16% compared with placebo while the risk of ischemic stroke was reduced by 20%. There was no effect on cardiovascular and all-cause mortality.

THEMIS: Primary and Secondary Outcomes



(n = 9,619)


(n = 9,601)

HR (95% CI)

Cardiovascular Death, MI, or Stroke



0.90 (0.81-0.99)

Cardiovascular Death



1.02 (0.88-1.18)




0.84 (0.71-0.98)

Ischemic Stroke



0.80 (0.64-0.99)

Death From Any Cause



0.98 (0.87-1.10)

Based on the modified intention-to-treat population, 93 patients would need to be treated to cause one major bleeding event. There was no difference in the risk of fatal bleeding with ticagrelor, although 70 patients treated with ticagrelor had an intracranial hemorrhage compared with 46 events in the placebo arm. In an analysis that focused on the net irreversible harm, a composite looking at the risks of death from any cause, MI, stroke, fatal bleeding, or intracranial hemorrhage, the rates were similar in the ticagrelor and placebo arms.

THEMIS: Safety Outcomes



(n = 9,562)


(n = 9,531)

HR (95% CI)

TIMI Major Bleeding



2.32 (1.82-2.94)

TIMI Major or Minor Bleeding



2.49 (2.02-3.07)

PLATO Major Bleeding



2.41 (1.98-2.93)

BARC 3, 4, or 5 Bleeding



2.36 (1.96-2.84)

Intracranial Hemorrhage



1.71 (1.18-2.48)

For stable CAD patients with diabetes, the data suggest that “dual antiplatelet therapy probably isn’t the way to go,” said Bhatt. “That’s not dissimilar to what we saw in CHARISMA many years ago with aspirin and clopidogrel.”

Sanjit Jolly, MD (McMaster University, Hamilton, Canada), who was not involved in the study, said there are data from randomized clinical trials, such as PEGASUS-TIMI 54 and PLATO, showing that post-MI and ACS patients benefit from long-term dual antiplatelet therapy and noted that physicians and patients have choices given the numerous therapies at their disposal.

While the THEMIS trial showed a modest reduction in ischemic events in stable CAD patients with diabetes, the higher risk of bleeding, particularly a higher risk of intracranial hemorrhage, is concerning, he said. In the COMPASS trial of patients with stable cardiovascular disease, low-dose rivaroxaban plus aspirin lowered the risk of cardiovascular death, MI, or stroke. While the risk of bleeding was higher in patients treated with the non-vitamin K antagonist oral anticoagulant, there was no increased risk of fatal bleeding or intracranial hemorrhage.   

“When you’re in front of the patient, you’re not just dealing with one trial, but with the entire literature,” Jolly told TCTMD.  

While comparisons across trials remain difficult, and there is no head-to-head study of ticagrelor versus rivaroxaban, there is some overlap between the COMPASS and THEMIS patients, said Jolly. In COMPASS, he noted, all-cause mortality was significantly lower in patients treated with rivaroxaban and aspirin compared with aspirin alone. “Mortality, as they say, is the ultimate net benefit,” he said. “If more patients are alive at the end of the study, you’re doing something right.”

If more patients are alive at the end of the study, you’re doing something right. Sanjit Jolly

Becker also pointed out there are different options for dual antiplatelet therapy in patients with stable CAD. Given the potential for bleeding with ticagrelor, only patients at high risk for ischemic events and low risk for bleeding should be treated with the potent P2Y12 antagonist. Like Jolly, he said low-dose rivaroxaban is also a potential option in stable CAD patients. “Cardiology has prided itself on having large clinical trials and guidelines geared nicely toward patient care, but we acknowledge that we don’t have carefully laid-out algorithms for every single patient,” said Becker. “It’s a changing dynamic.”


In the THEMIS-PCI study, investigators focused on the 11,154 patients with stable CAD and diabetes mellitus who previously underwent PCI. The median time since the most recent PCI was 3.3 years. In addition to the 15% relative reduction in the risk of the primary endpoint, treatment with ticagrelor reduced the risk of MI (3.1% in the ticagrelor arm vs 3.9% in the placebo arm; P = 0.027) including STEMI and the risk of stroke (1.7% in the ticagrelor arm vs 2.3% in the placebo arm; P = 0.024). There was also an 18% relative reduction in the risk of all-cause mortality, MI, or stroke in the previous PCI patients, but no effect was observed in those without prior stenting (P = 0.02 for interaction).

Similarly, there was a significantly lower risk of all-cause mortality, MI, stroke, acute limb ischemia, or major amputations with ticagrelor—a “mega” composite of ischemic events—among prior-MI patients treated with ticagrelor (HR 0.82; 95% CI 0.72-0.92). There was no effect on cardiovascular or all-cause mortality.

The risk of bleeding was roughly twofold higher with ticagrelor compared with placebo in patients with prior PCI. However, there was no significant interaction between those with and without a history of PCI with respect to TIMI major, TIMI major or minor, and PLATO major bleeding. Rates of fatal bleeding and intracranial hemorrhage were equivalent in ticagrelor- and placebo-patients with history of PCI. 

“We had hypothesized that THEMIS patients with a prior history of PCI previously treated with dual antiplatelet therapy and tolerated would most likely be a group who would have a favorable balance of efficacy and safety,” said Steg. In the THEMIS-PCI study, 84 patients would need to be treated with ticagrelor to prevent one cardiovascular death, MI, or stroke.

The data, taken alongside PLATO and PEGASUS-TIMI 54, suggest that dual antiplatelet therapy with aspirin and ticagrelor is effective in patients with acute coronary syndrome and previous MI as well as in stable CAD patients with diabetes who had previously undergone PCI, say investigators. The THEMIS-PCI patients represent an “easy-to-identify” patient group for whom extended dual antiplatelet therapy could be an option.

“Among the population of patients with diabetes and coronary artery disease, we think that those who have had stenting in the past, and who haven’t had any bleeding problems, would be the ones where [ticagrelor] should be considered,” said Bhatt.

To TCTMD, Jolly said dual antiplatelet therapy with ticagrelor is effective in preventing stent thrombosis, but that low-dose rivaroxaban may be more effective in reducing the overall risk of vascular events, such as stroke and PAD. “If I have a patient with a stent in the left main [artery], where it would be catastrophic for them to have a stent thrombosis, I’d be more likely to prolong or continue using low-dose ticagrelor,” he said. “On the other hand, if I have a patient with peripheral vascular disease, or who is at high risk of stroke, I’d be more likely to use rivaroxaban. I think there’s still an opportunity to use clinical information to guide your judgment.”  

Michael O’Riordan is the Associate Managing Editor for TCTMD and a Senior Journalist. He completed his undergraduate degrees at Queen’s…

Read Full Bio
  • Bhatt DL, Steg PG, Mehta SR, et al. Ticagrelor in patients with diabetes and stable coronary artery disease with a history of previous percutaneous coronary intervention (THEMIS-PCI): a phase-3, placebo-controlled, randomized trial. Lancet. 2019;Epub ahead of print.

  • Steg PG, Bhatt DL, Simon T, et al. Ticagrelor in patients with stable coronary disease and diabetes. N Engl J Med. 2019;Epub ahead of print.

  • AstraZeneca sponsored THEMIS and THEMIS-PCI.
  • Bhatt reports receiving research funding or unfunded research support from Abbott, Amarin, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chiesi, Eisai, Ethicon, Forest Laboratories, Idorsia, Ironwood, Ischemix, Lilly, Medtronic, PhaseBio, Pfizer, Regeneron, Roche, Sanofi, Synaptic, The Medicines Company, FlowCo, Merck, Novo Nordisk, PLx Pharma, and Takeda; being a site co-investigator for Biotronik, Boston Scientific, St. Jude Medical, and Svelte; being a trustee for ACC; serving as an advisory board member, director, or chair for Cardax, Elsevier Practice Update Cardiology, Medscape Cardiology, Regado Biosciences; the Boston VA Research Institute, the Society of Cardiovascular Patient Care, TobeSoft; the American Heart Association Quality Oversight Committee; serving on a range of data safety monitoring committees; receiving honoraria for editorial or committee activities for a range of publications and organizations; and receiving royalties from Elsevier.
  • Steg reports personal fees and non-financial support from AstraZeneca, during the conduct of the study; grants and personal fees from Bayer/Janssen and Servier, Merck, Sanofi, and Amarin, and personal fees from Amgen, Bristol-Myers Squibb, Boehringer Ingelheim, Pfizer, Novartis, Regeneron, Lilly, AstraZeneca, Novo Nordisk, and Idorsia, outside the submitted work.
  • Jolly reports no relevant conflicts of interest.