The THEMIS trial demonstrated that in high-risk patients with stable coronary artery disease and diabetes without previous myocardial infarction or stroke, ticagrelor, in addition to aspirin, reduced the incidence of ischemic events but increased major bleeding. Identification of patients who could derive the greatest net benefit from the addition of ticagrelor appears important. We used the CRUSADE bleeding risk score to risk stratify the THEMIS population.

The population was divided into tertiles: score ≤22, 23 to 33, and ≥34. In each tertile, primary efficacy (composite of cardiovascular death, myocardial infarction, or stroke) and safety (TIMI major bleeding) outcomes were analyzed. NACE (net adverse clinical events) was defined as the irreversible harm composite, in which all-cause death, myocardial infarction, stroke, amputations, fatal bleeds, and intracranial hemorrhage were counted.

Patients in the lower risk tertile experienced fewer ischemic events with ticagrelor than placebo, whereas there was no significant benefit from ticagrelor in the other tertiles (P_interaction = .008). Bleeding rates were consistently increased with ticagrelor across all tertiles (P_interaction = .79). Ticagrelor reduced NACE in the first tertile (HR = 0.74, 95% CI = 0.61-0.90) but not in the others (HR = 1.03, 95% CI = 0.86-1.23 and HR = 1.05, 95% CI = 0.91-1.22, respectively; P_interaction = .012).

In patients with stable coronary artery disease and diabetes without a history of myocardial infarction or stroke, only those at the lower end of the bleeding risk spectrum according to the CRUSADE score derived net benefit from ticagrelor.