This study is a post-hoc analysis of the ISAR-REACT 5 trial and includes all patients with available BMI data. The design and outcomes of the ISAR-REACT 5 trial have previously been reported.20,21 Patients were eligible for enrollment if they were hospitalized for an ACS (ST-segment elevation myocardial infarction [STEMI], non–ST-segment elevation myocardial infarction [NSTEMI], or unstable angina) and were scheduled to undergo invasive evaluation. Patients were randomly assigned to either ticagrelor or prasugrel in a 1:1 ratio. The study had an open-label design. Patients in the ticagrelor arm received a loading dose of 180mg of ticagrelor as soon as possible after randomization and continued at a maintenance dose of 90mg twice daily. Patients in the prasugrel arm received a loading dose of 60mg of prasugrel (immediately after randomization in patients with STEMI, and after coronary angiography was known [ie, after coronary angiography] but before PCI in patients with NSTEMI or unstable angina) and continued at a maintenance dose of 10mg daily. For patients aged ≥ 75 years or those with a body weight lower than 60kg, a reduced maintenance dose of 5mg of prasugrel was recommended.22 All patients received a loading dose of 150 to 300mg of intravenous or chewed aspirin and a maintenance dose of 75 to 100mg once daily. Written informed consent was obtained from all patients. The study protocol was approved by the local ethics committee at each participating center. The study conformed to the Declaration of Helsinki.

BMI was calculated as a ratio of weight in kilograms to height in meters squared (kg/m2) using the patient's weight and height measured during the hospital stay. BMI data were available in 3987 of 4018 patients of the ISAR-REACT 5 trial. BMI ranged from 14.0kg/m2 to 55.0kg/m2. Patients were categorized into groups according to BMI categories: patients with a BMI of <25kg/m2 (normal weight group), patients with a BMI of ≥ 25kg/m2 to <30kg/m2 (overweight group), and patients with a BMI of ≥ 30kg/m2 (obesity group). Of the 1084 patients in the normal weight group, 27 patients (2.5% of the group with BMI <25kg/m2) had a BMI <18.5kg/m2.

The primary (efficacy) endpoint was the 12-month incidence of all-cause death, myocardial infarction, or stroke. The secondary (safety) endpoint was the 12-month incidence of Bleeding Academic Research Consortium (BARC) type 3 to 5 bleeding.23 Myocardial infarction was defined according to the Third Universal Definition of Myocardial Infarction.24 Stent thrombosis and cardiovascular death were defined according to the Academic Research Consortium (ARC) criteria.25 Detailed definitions are provided in the supplementary data. All outcomes assessed in the primary trial20 were included in this analysis. All primary and secondary outcomes were centrally adjudicated in a blinded manner by members of the event adjudication committee.

Follow-up was scheduled at 1 month, 6 months, and 1 year. In case of potential endpoint-related adverse events, source data were solicited. All serious adverse events, including outcomes analyzed in this study, were monitored on-site. Patients were monitored either via hospital visits, outpatient visits or through telephone and structured follow-up letters.

This study represents a post-hoc analysis of the ISAR-REACT 5 trial. Continuous variables are presented as mean with standard deviation or median with 25th-75th percentiles and were compared using the Student t test or the Wilcoxon rank sum test as appropriate. Categorical data are presented as counts and proportions and were compared using the chi-squared test. The primary endpoint and all-cause death are presented as cumulative incidence(s) and were calculated using the Kaplan-Meier method.26,27 All other endpoints are presented as cumulative incidence(s) after accounting for the competing risk of death. Competing risk analysis is used when there is an event (in our study death) whose occurrence precludes the occurrence of the primary event of interest.28 The comparison between the groups was performed using the Cox proportional-hazards model with the participating center and stratification according to clinical presentation (ST-segment elevation or non–ST-segment elevation ACS) entered into the model as covariates along with the study drug. The association of study drug and BMI category with the primary and secondary endpoints was adjusted for potential confounders using the Cox proportional-hazards model. The following variables were entered into the model: study drug, age, sex, diabetes mellitus, smoking, hypertension, hypercholesterolemia, history of MI, history of PCI, history of coronary artery bypass grafting, cardiogenic shock, heart rate, serum creatinine on admission, diagnosis on admission, type of treatment (conservative or PCI), prasugrel 5mg and treatment arm-by-BMI interaction. To assess the treatment arm-by-BMI interaction for study endpoints, an interaction term was entered into the Cox proportional-hazards model. Risk estimates are presented as hazard ratios [HR] with 95% confidence intervals [95%CI]. The primary (efficacy) endpoint was analyzed according to the intention-to-treat principle (ie, including all patients as initially assigned, irrespective of the actual treatment received). The secondary (safety) endpoint was analyzed in a modified intention-to-treat (mITT) population (ie, including all patients with at least 1 application of the study drug with bleeding assessed for up to 7 days after study drug discontinuation). The statistical analysis was performed using the R 3.6.0 Statistical Package (The R foundation for Statistical Computing, Vienna, Austria). A 2-sided P <.05 was considered to indicate statistical significance.

The study flowchart is shown in figure 1. Data on BMI were available in 3987 patients: 1084 patients with normal weight, 1890 patients with overweight and 1013 patients with obesity. Baseline data are shown in table 1 of the supplementary data. Patients with obesity were younger, less likely to be smokers and more likely to have diabetes, hypertension, hypercholesterolemia, or prior PCI. Cardiogenic shock was less frequent and serum creatinine values were higher in patients with obesity. The rate of coronary angiography and treatment strategy differed little across the BMI categories. Angiographic and procedural data are shown in table 2 of the supplementary data and table 3 of the supplementary data. Drug therapy at discharge is shown in table 4 of the supplementary data. Twenty-two of the 27 patients with a BMI <18.5kg/m2 were discharged on ticagrelor and 5 patients were discharged on prasugrel. Of the latter, 4 patients were discharged on a reduced dose of prasugrel (5mg). In patients with normal weight, overweight, and obesity, the primary endpoint occurred in 102 (9.6%), 136 (7.3%), and 80 (8.0%) patients, respectively (hazard ratio [HR]=0.77; 95%CI, 0.59-0.99; P=.047 for overweight vs normal weight and HR, 0.84; 95%CI, 0.62-1.13; P=.240 for obesity vs normal weight; figure 1 of the supplementary data). The secondary endpoint (BARC type 3-5 bleeding) occurred in 79 (7.4%), 104 (5.6%) and 41 (4.1%) patients, respectively (HR, 0.72; 95%CI, 0.54-0.97; P=.031 for overweight vs normal weight and HR, 0.52; 95%CI, 0.36-0.77; P=.0008 for obesity vs normal weight; figure 2 of the supplementary data).

Figure 1.

Study flowchart. BMI, body mass index.

(0.22MB).

Baseline data in patients assigned to ticagrelor or prasugrel in each BMI category are shown in table 1. In the group with normal weight, the proportion of patients with hypertension and prior PCI was significantly higher among ticagrelor-assigned patients. In the overweight group, baseline characteristics were well-balanced between the 2 treatment arms. In the group with obesity, patients assigned to ticagrelor were younger than those assigned to prasugrel. The remaining characteristics did differ little between patients assigned to ticagrelor and prasugrel across all BMI categories.

Angiographic data are shown in table 5 of the supplementary data. There were no significant differences in terms of access site, number of narrowed coronary arteries, or left ventricular ejection fraction between patients according to study drugs in groups with normal weight, overweight, or obesity. Procedural characteristics are shown in table 6 of the supplementary data. Drug therapy at discharge is shown in table 7 of the supplementary data. In the normal weight and overweight groups, patients assigned to prasugrel were more often discharged on aspirin (P=.023) and clopidogrel (P=.049), respectively.

Clinical outcomes according to study drugs across BMI categories are shown in table 2. In the group with normal weight, the primary endpoint occurred in 63 patients (11.7%) in the ticagrelor group and 39 patients (7.5%) in the prasugrel group (HR, 1.62; 95%CI, 1.09-2.42; P=.018; figure 2A). In the overweight group, the primary endpoint occurred in 78 patients (8.3%) in the ticagrelor group and 58 patients (6.2%) in the prasugrel group (HR, 1.36; 95%CI, 0.97-1.91; P=.076; figure 2B). In the group with obesity, the primary endpoint occurred in 43 patients (8.6%) in the ticagrelor group and 37 patients (7.3%) in the prasugrel group (HR, 1.18; 95%CI, 0.76-1.84; P=.451; figure 2C). There was no treatment arm-by-BMI interaction regarding the primary endpoint (P for interaction=.578).There were no significant differences regarding the incidence of death between ticagrelor and prasugrel across the BMI categories. In the group with normal weight, the incidence of myocardial infarction, and definite or probable stent thrombosis was higher in patients assigned to ticagrelor than those assigned to prasugrel (5.9% vs 2.5%; HR, 2.43; 95%CI, 1.27-4.63; P=.007), and (2.0% vs 0.4%; HR, 5.22; 95%CI, 1.15-23.62; P=.032), respectively. In the overweight group, the incidence of myocardial infarction, and definite or probable stent thrombosis in the ticagrelor and prasugrel groups was 3.3% vs 2.9%; HR, 1.18; 95%CI, 0.71-1.99; P=.520, and 1.1% vs 1.4%; HR, 0.79; 95%CI, 0.34-1.79; P=.567, respectively. In the group with obesity, the incidence of myocardial infarction, and definite or probable stent thrombosis in the ticagrelor and prasugrel groups was 6.6% vs 4.0%; HR, 1.68; 95%CI, 0.96-2.93; P=.067, and 1.0% vs 1.0%; HR, 1.01; 95%CI, 0.29-3.48; P=.991, respectively.

Figure 2.

Cumulative incidence of the primary endpoint (1-year incidence of death, myocardial infarction, or stroke) in the ticagrelor and prasugrel groups in A) patients with normal weight, B) patients with overweight and C) patients with obesity. 95%CI, 95% confidence interval; HR, hazard ratio.

(0.18MB).

Bleeding events are shown in table 2. In the group with normal weight, BARC type 3 to 5 bleeding occurred in 30 patients assigned to ticagrelor and 29 patients assigned to prasugrel (cumulative incidence accounting for competing risk 6.5% and 6.6%, respectively; HR, 1.00; 95%CI, 0.60-1.67; P=.990). In the overweight group, BARC type 3 to 5 bleeding occurred in 45 patients assigned to ticagrelor and 39 patients assigned to prasugrel (cumulative incidence accounting for competing risk, 5.6% and 5.0%, respectively; HR, 1.13; 95%CI, 0.74-1.74; P=.566). In the group with obesity, BARC type 3 to 5 bleeding occurred in 19 patients in the ticagrelor group and 11 patients in the prasugrel group (cumulative incidence accounting for competing risk, 4.4% and 2.8%, respectively; HR, 1.59 95%CI, 0.76-3.35; P=.219). Time-to-event curves are shown in figure 3. There was no treatment arm-by-BMI interaction regarding the secondary endpoint (P for interaction=.596).

Figure 3.

Cumulative incidence of the secondary endpoint (1-year incidence of Bleeding Academic Research Consortium type 3-5 bleeding) accounting for the competing risk of death in the ticagrelor and prasugrel groups in A) patients with normal weight, B) patients with overweight and C) patients with obesity. BARC, Bleeding Academic Research Consortium; 95%CI, 95% confidence interval; HR, hazard ratio.

(0.2MB).

To address an eventual impact of underweight patients (those with BMI <18.5kg/m2; n=27 patients; 2.5% of the group with BMI <25kg/m2), a sensitivity analysis was performed by excluding these patients from the group with normal weight (BMI <25kg/m2). Among patients with a BMI between 18.5kg/m2 and 25kg/m2 (n=1057 patients), the primary and secondary endpoints occurred in 99 patients (9.5%) and 76 patients (7.3%), respectively. The primary endpoint occurred in 60/532 patients assigned to ticagrelor and 39/525 patients assigned to prasugrel (11.5% vs 7.6%; HR, 1.58; 95%CI, 1.06-2.37; P=.026). The secondary endpoint (BARC type 3 to 5 bleeding) occurred in 28/521 patients assigned to ticagrelor and 29/466 patients assigned to prasugrel (6.5% vs 6.6%; HR, 0.96; 95%CI, 0.57-1.61; P=.877). Thus, in the current analysis, the exclusion of patients with a BMI <18.5kg/m2 appears to have no impact on the frequency of adverse events or on the efficacy and safety of ticagrelor vs prasugrel among patients with a BMI <25kg/m2.

The association of study drug and BMI category with the primary and secondary outcomes was adjusted for potential confounders using the multivariable Cox proportional-hazards model (see methods for variables entered in the model). After adjustment for potential confounders, ticagrelor was independently associated with the primary (adjusted HR, 1.52, 95%CI, 1.13-2.04), but not with the secondary (adjusted HR, 1.01; 95%CI, 0.71-1.37) endpoint. BMI category was not associated with the primary endpoint (adjusted HR, 0.88; 95%CI, 0.67-1.14; P=.331 for overweight vs normal weight and HR=1.07 [0.78-1.46]; P=.696 for obesity vs normal weight); however, it appears to be associated with the secondary endpoint (HR, 0.87; 95%CI, 0.64-1.18; P=.365 for overweight vs normal weight and HR, 0.64; 95%CI, 0.43-0.96; P=.031 for obesity vs normal weight).

Our study has limitations. The present analysis is a post-hoc analysis of a randomized trial. As such, it carries the inherent limitations of post-hoc analyses. The categorization of patients in groups according to BMI reduced the study power to detect differences in the efficacy and safety of ticagrelor vs prasugrel in BMI categories. In the normal weight group, 2.5% of patients had a BMI <18.5kg/m2. The sensitivity analysis showed that the exclusion of patients with a BMI <18.5kg/m2 from the group with normal weight had almost no impact (0.1% difference) on the occurrence of primary or secondary endpoints or on the efficacy or safety of ticagrelor vs prasugrel. However, these data do not negate an impact of underweight on the occurrence of adverse events or on the efficacy and safety of ticagrelor vs prasugrel in patients with ACS; they simply show that the proportion of underweight patients was too small to have an overall impact on the occurrence of events in the current study. Furthermore, the observed greater reduction of ischemic events by prasugrel as compared with ticagrelor in patients with normal weight may be the result of chance amplified by multiple testing.32 Since randomization was not performed according to BMI, a possible impact of hidden confounders cannot be entirely ruled out even though we adjusted for an array of clinical and demographic variables in the multivariable analysis. Although the treatment effect of ticagrelor vs prasugrel was directionally consistent, the study was underpowered to assess heterogeneity in the treatment effect across BMI categories. Finally, although all adverse events were centrally adjudicated in a blinded manner, the study had an open-label design. For these reasons, the study findings should be considered as exploratory or hypothesis-generating.