Conference Coverage

Ticagrelor: Modest benefit, bigger bleed risk in diabetes plus stable CAD



THEMIS was an international multisite double-blind, placebo-controlled study randomizing 19,220 patients 1:1 to receive aspirin, plus placebo (N = 9,601) or ticagrelor (N = 9,619). Patients were followed for a median of 39.9 months; those with previous myocardial infarction or stroke were excluded. Patients had to be at least 50 years old and on anti-hyperglycemic medications for at least 6 months to participate. Patients in the overall study had a baseline age of 66 years, and 31% were female. Most patients were white (71%).

Stable coronary artery disease (CAD) was defined by having any of a previous history of PCI, coronary artery bypass grafting, or angiographically documented stenosis of at least 50% in at least one coronary artery.

During the study period, Dr. Bhatt explained, ticagrelor dosage was reduced from 90 to 60 mg daily as other studies yielded data about improved safety and tolerability without compromise in efficacy at the lower ticagrelor dose.

Permanent treatment discontinuation was common, but more common in patients taking ticagrelor, compared with placebo (34.5% vs. 25.4%). The most frequent reasons for ticagrelor discontinuation were dyspnea and bleeding. All patients who were randomized, save those at a study site that was closed before unblinding, were included in the modified intention-to-treat population for calculation of efficacy outcomes for both THEMIS and THEMIS-PCI.

Given the large number of patients who discontinued the study drug, an estimation was made of the number of events that would have occurred had patients remained in the trial, and outcomes were calculated using these estimations to account for missing data.

Safety outcomes were calculated by including all patients who received at least one dose of a study drug.

An exploratory composite outcome of “net irreversible harm” included all-cause death, myocardial infarction, and stroke, but also fatal bleeding and intracranial hemorrhage. In the full study population, this outcome was seen in 10.1% of the placebo group and 10.8% of the placebo group, for a nonsignificant hazard ratio of 0.93, said Dr. Bhatt.

An additional composite pre-specified exploratory outcome included acute limb ischemia or major amputation; here, the HR of 0.45 favored ticagrelor.

Dr. Bhatt made the point that these pragmatic, patient-centered outcomes are valuable tools when weighing the potential risks and benefits of therapy for a particular patient, and provide a discussion point for individualized, shared decision making.

Results of a pre-specified subgroup analysis of the 58% of THEMIS participants (n = 5,558) with prior PCI were presented by THEMIS’ co-principal investigator, Philippe Gabriel Steg, MD, of the University of Paris and the French National Institute of Health and Medical Research.

“In the history of PCI subgroup, 92% of patients had a history of receiving a stent, and 61% had received at least one drug-eluting stent,” said Dr. Steg.

Patients with PCI saw a slightly greater reduction in relative risk for ischemic events when they received ticagrelor, compared with placebo; the PCI group had a HR of 0.85 for ischemic events (P = .013), compared with a HR of 0.98 for those with no PCI history (P = .76). This meant that ticagrelor DAPT’s efficacy as measured by the primary endpoint of ischemic events lost significance when the non-PCI group was evaluated (P = .76, with P for interaction between the groups of .16).

Some secondary endpoints showed statistical significance for the interaction between PCI status and study drug status. These included the composite outcome of all-cause death, MI, or stroke (P for interaction, .021), and another “mega-composite ischemia” outcome that folded in major amputation of vascular etiology along with all-cause death, MI, and stroke (P = .023).

Looking at bleeding endpoints, there was no significant difference between the groups for TIMI major bleeding, the primary safety endpoint. Patients in the full study cohort as well as the PCI subgroup had significantly more TIMI major bleeding on ticagrelor.

Bleeding measured by Bleeding Academic Research Consortium (BARC) criteria was a secondary endpoint, and the P for interaction just reached statistical significance for the aggregate of all levels of BARC bleeding.

“But the two observations I would draw your attention to are the fact that in patients with a history of PCI, fatal bleeding occurred in the same number of patients in each group – 6 patients in each group,” added Dr. Steg. “And even more importantly, intracranial hemorrhage occurred in 33 patients in the ticagrelor group and 31 patients in the placebo group for patients with a history of PCI, whereas it was 37 and 15 for patients without a history of PCI.” This yielded a significant P value for the interaction of .036.

The exploratory net clinical benefit score favored the PCI group, for a P for interaction of .012. Dr. Steg also shared an analysis showing a net benefit for ticagrelor vs. placebo as a function of the time elapsed between PCI and trial randomization, showing patient benefit to 6 years post drug initiation for the PCI group.

“The subgroup analysis of THEMIS PCI was pre-specified, from a large, clinically meaningful population; it’s plausible and it can be easily explained from the action of dual antiplatelet therapy, and it shows a net benefit,” Dr. Steg said.

The discussant for the presentations was Colin Baigent, , and he wasn’t convinced by the THEMIS-PCI data. He pointed out that looking at the absolute numbers overall for THEMIS yields an absolute benefit of about 8 per 1,000 participants, and an absolute risk of about 12 per 1,000 participants.

“The natural instinct is to then go to the subgroups and try to find people who will see a net benefit,” he said. “Why pick out ‘history of PCI?’” among the 18 pre-specified subgroups, he asked, noting that there was not significant evidence of heterogeneity of hazard ratios among the subgroups.

Overall, “The main results of THEMIS are consistent” with previous investigations into the benefits of ticagrelor DAPT, showing modest efficacy at the expense of a two-fold rise in major bleeding events, said Dr. Baigent, professor of epidemiology at the University of Oxford (England).

The THEMIS study and the subpopulation analysis were funded by AstraZeneca, which markets ticagrelor. Dr. Bhatt reported financial relationships with AstraZeneca and multiple other pharmaceutical companies. In addition to reporting a financial relationship with AstraZeneca, Dr. Steg also reported relationships with multiple pharmaceutical companies. Dr. Baigent reported a financial relationship with Boehringer Engelheim.

Source: Steg PG et al. N Engl J Med. 2019 Sep 1: DOI: 10.1056/NEJMoa1908077; Bhatt DL et al.Lancet. 2019 Sep 1: DOI:

Next Article: