Restrictive vs liberal red blood cell transfusion strategies in patients with acute myocardial infarction and anemia: Rationale and design of the REALITY trial

Abstract Background Anemia is common in patients with acute myocardial infarction (AMI), and is an independent predictor of mortality. The optimal transfusion strategy in these patients is unclear. Hypothesis We hypothesized that a “restrictive” transfusion strategy (triggered by hemoglobin ≤8 g/dL) is clinically noninferior to a “liberal” transfusion strategy (triggered by hemoglobin ≤10 g/dL), but is less costly. Methods REALITY is an international, randomized, multicenter, open‐label trial comparing a restrictive vs a liberal transfusion strategy in patients with AMI and anemia. The primary outcome is the incremental cost‐effectiveness ratio (ICER) at 30 days, using the primary composite clinical outcome of major adverse cardiovascular events (MACE; comprising all‐cause death, nonfatal stroke, nonfatal recurrent myocardial infarction, or emergency revascularization prompted by ischemia) as the effectiveness criterion. Secondary outcomes include the ICER at 1 year, and MACE (and its components) at 30 days and at 1 year. Results The trial aimed to enroll 630 patients. Based on estimated event rates of 11% in the restrictive group and 15% in the liberal group, this number will provide 80% power to demonstrate clinical noninferiority of the restrictive group, with a noninferiority margin corresponding to a relative risk equal to 1.25. The sample size will also provide 80% power to show the cost‐effectiveness of the restrictive strategy at a threshold of €50 000 per quality‐adjusted life year. Conclusions REALITY will provide important guidance on the management of patients with AMI and anemia.

Methods: REALITY is an international, randomized, multicenter, open-label trial comparing a restrictive vs a liberal transfusion strategy in patients with AMI and anemia.
The primary outcome is the incremental cost-effectiveness ratio (ICER) at 30 days, using the primary composite clinical outcome of major adverse cardiovascular events (MACE; comprising all-cause death, nonfatal stroke, nonfatal recurrent myocardial infarction, or emergency revascularization prompted by ischemia) as the effectiveness criterion. Secondary outcomes include the ICER at 1 year, and MACE (and its components) at 30 days and at 1 year.
Results: The trial aimed to enroll 630 patients. Based on estimated event rates of 11% in the restrictive group and 15% in the liberal group, this number will provide 80% power to demonstrate clinical noninferiority of the restrictive group, with a noninferiority margin corresponding to a relative risk equal to 1.25. The sample size will also provide 80% power to show the cost-effectiveness of the restrictive strategy at a threshold of €50 000 per quality-adjusted life year.
Conclusions: REALITY will provide important guidance on the management of patients with AMI and anemia.

K E Y W O R D S
acute myocardial infarction, anemia, cost effectiveness, transfusion

| INTRODUCTION
Anemia is common in patients with acute myocardial infarction (AMI), 1 and is an independent predictor of cardiac events and increased mortality. [1][2][3][4] The association with increased mortality is strong, with a relative increase in mortality exceeding 20% for each 1 g/dL decrement in hemoglobin below 14 g/dL. 2 The antiplatelet and anticoagulant drugs used for the treatment of patients with AMI increase the risk of bleeding, which in turn increases the risk of ischemia and death. Whether this risk can be overcome by transfusion is debated.
Clinical data on the effects of blood transfusion in patients with AMI are observational 5,6 or based on small, underpowered randomized trials. 7,8 Previous results have been inconsistent, which prompted a call for large randomized trials. 9,10 As a consequence, only low levels of evidence are available to guide international guidelines on blood transfusion in patients with AMI and anemia, which are somewhat inconsistent. [11][12][13] Randomized trials comparing transfusion strategies in patients undergoing noncardiac 14 or cardiac surgery, 15,16 and in patients with upper gastrointestinal bleeding, 17 have generally found benefit from a restrictive transfusion strategy. However, these trials excluded patients with AMI. In the specific context of acute myocardial ischemia, the effect of transfusion strategies may differ markedly from what is observed in other settings. Apart from the clinical benefit of either strategy, other issues should be considered: blood is a costly and scarce resource; transfusion can be associated with adverse events (infectious and noninfectious); and transfusion mobilizes substantial resources and healthcare professional time, such that even if a restrictive strategy were noninferior to a liberal transfusion strategy, there may be value in preferring the former.
Given the high prevalence of anemia, its strong association with death and adverse cardiac outcomes, the cost implications of transfusion, and the lack of high-quality data to guide clinical practice in patients with AMI and anemia, we conducted an international, randomized clinical trial comparing a "restrictive" (triggered by hemoglobin ≤8 g/dL) vs a "liberal" (triggered by hemoglobin ≤10 g/dL) red blood cell (RBC) transfusion strategy in patients with AMI and anemia (defined as 7 g/dL < hemoglobin ≤10 g/dL). We hypothesized that a restrictive transfusion strategy would be clinically noninferior to a liberal transfusion strategy but more cost effective. The steering committee is responsible for the medical, scientific, and operational conduct of the study, and vouches for the integrity of the data analysis and the reporting of the results. The study adheres fully to the ethical principles of the Declaration of Helsinki. The duration of participation for each patient was 1 year.

| Study population
Patients ≥18 years with AMI and anemia who met the inclusion criteria and none of the exclusion criteria (Table 1)  Inclusions in the main study could be extended to attain this goal. All patients provided written informed consent.

| Study outcomes
The primary outcome is the incremental cost-effectiveness ratio (ICER) at 30 days, using major adverse cardiovascular events (MACE) (composite of all-cause death, nonfatal stroke, nonfatal recurrent myocardial infarction, or emergency revascularization prompted by ischemia) as the effectiveness criterion. The primary and various secondary outcomes are listed in Table 2.
All clinical outcomes (deaths, including cause of death, AMIs, strokes, ischemia-driven emergency revascularizations, and heart failures) will be reviewed and adjudicated by a CEC. The CEC will be given all relevant clinical and biochemical information and will be blinded to the randomization assignment. In addition, AMIs will be categorized by the CEC according to the Third Universal Definition. 18 Each triggered event will be reviewed by two CEC members. In case of discrepancy between the two adjudicators, an ad hoc committee with at least three CEC members will reach a final decision based on consensus.
Safety outcomes of special interest, which will be monitored during the index hospitalization, include hemolysis, infections, multiorgan system dysfunction, acute lung injury, acute heart failure, acute renal failure, and severe allergic reactions.

| Health-economic analysis
A cost-effectiveness primary outcome (30 day ICER) was chosen due to its frequent use in cardiology and the availability of benchmarks for the cost-effectiveness results. 19,20 The BASKET trial has provided a benchmark for the acceptable cost per MACE averted with an ICER of 64 732 euros to prevent one major adverse cardiac event. 21 The secondary health outcome for the cost-utility analyses is quality-adjusted life years (QALYs) at 1 year. These are calculated for both groups using health-related quality of life (HRQoL) scores from the EuroQoL five dimensions (EQ-5D) questionnaire and converted to utility scores using country weights. 22,23 The cost-effectiveness analysis of the costs and outcomes for patients in the restrictive and liberal transfusion groups will be T A B L E 1 Inclusion and exclusion criteria in the REALITY trial All-cause death, nonfatal stroke, nonfatal recurrent myocardial infarction, or emergency revascularization prompted by ischemia. All components of the composite outcome will also be analyzed separately.
conducted from the perspectives of both the French and Spanish governments (providing a healthcare system perspective for both countries). This perspective allows comparisons with the results of other economic evaluations in cardiology. The economic evaluation will be based on the entire population of patients in the trial. Unit costs are undiscounted and are presented in Tables S1 and S2. Resources will be collected prospectively at the patient level. Prespecified subgroup analyses by country are necessary due to potential differences in unit costs and resource utilization. 24

| Statistical analyses
Analyses will be performed in both the ITT and PP cohorts. The ITT population is defined as all randomized patients who have signed an informed consent form. In case of consent withdrawal, only data collected before withdrawal will be used. All patients will be analyzed in their randomized arm regardless of the strategy used during the trial.
The PP population is defined as all patients randomized and treated without major protocol violations or deviations. Predefined major protocol violations/deviations are: (1) violation of the selection criteria; (2) receipt of the alternative transfusion strategy (apart from in the situations listed below); (3) missing data for the primary efficacy outcomes; (4) inclusion in another interventional study; and (5) major protocol deviation identified during a blinded data review before database lock. Crossover from the restrictive to the liberal strategy will be allowed, and thus not considered a major protocol deviation (so these patients will be included in the PP analysis), in the following documented cases: massive overt active bleeding, presumed important fall in hemoglobin level and no time to wait for hemoglobin testing, or shock occurring after randomization. All statistical analyses will be performed using SAS 9.4 software.

| Cost-effectiveness analysis
A joint comparison of costs and MACE will be performed by nonparametric bootstrapping with 1000 re-samples. A distribution will be attributed to each variable according to accepted practice: normal distribution for MACE and log normal or Beta for costs. The results of the bootstrap replications will be presented on a cost-effectiveness plane. In addition to the cost-effectiveness plane, acceptability curves for French and Spanish values will be plotted.

| Cost-utility analysis
The cost-utility analysis will be performed with French and Spanish utility values estimated from EQ-5D scores. All individual components of the primary outcome will be analyzed by Chi-square or Fisher's exact tests. As these analyses will be exploratory, no adjustment of the alpha value for multiplicity will be made.

| Primary clinical outcome analysis
Times to death and MACE during 1-year follow-up will be described by Kaplan-Meier curves and tested by log-rank test. The reference date is defined as the randomization date. All patients will be censored at the time of the last observation.
All safety endpoints will be analyzed by Chi-square or Fisher's exact tests.

| Subgroup analyses
Interactions between outcomes and prespecified subgroups will be analyzed according to the following variables: age, sex, body weight, diabetes mellitus, smoking status, hypertension, dyslipidemia, Killip class, type of AMI (ST-segment elevation myocardial infarction [STEMI] or non-STEMI), renal function at baseline, use of revascularization for acute coronary syndrome (ACS), hemoglobin at randomization, and presence or absence of overt bleeding before randomization.

| Ancillary biological substudy
In France, an ancillary study (Impact of Transfusion of RBC on  The primary outcome is a comparison, between the two groups, of the variation in the level of platelet reactivity using vasodilatorstimulated phosphoprotein (VASP) between T0 and T1. Up to 100 patients will be included, which will give 85% power with a 15% absolute difference in platelet reactivity. Secondary objectives will be to explore the effect of transfusion on platelet aggregation (using the VerifyNow test), thrombosis biomarkers (CD40 ligand, P-selectin, plasminogen activator inhibitor-1, von Willebrand factor, and glycoprotein 6), inflammation biomarkers (tumor necrosis factor alpha, interleukin-6 and interleukin-10, and high-sensitivity C-reactive protein), and interaction with leukocytes (neutrophil extracellular traps).

| Funding
The study is funded via a grant from the Programme de Recherche Médico-Economique (PRME) from the French Ministry of Health, and a grant from the Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness), Grant n PI15/01543.

| DISCUSSION
The primary hypothesis of the REALITY trial (NCT02648113) is that a restrictive transfusion strategy (triggered by hemoglobin ≤8 g/dL) will be clinically noninferior to a liberal transfusion strategy (triggered by hemoglobin ≤10 g/dL), but will be less costly.
Physiopathologic considerations suggest that the effect of RBC transfusion may differ between patients with vs without AMI. In theory, transfusion should increase oxygen delivery to the myocardium.
However, recent data suggest that oxygen delivery is not increased in patients receiving transfusions, that RBCs are rapidly depleted of nitric oxide during storage, and that, conversely, transfusion may increase platelet activation and aggregation. 27 These consequences appear potentially even more deleterious in patients with cardiovascular disease. 28 RBC transfusion can also be associated with various noncardiac complications such as transfusion-associated infections, acute lung injury, and circulatory overload. 29 All of these potential complications will be carefully monitored in the REALITY trial.
The issue of a restrictive vs a liberal transfusion strategy is not only relevant in cardiology, but is a common clinical problem encountered in many other settings. 30 A randomized trial conducted in patients with acute upper gastrointestinal bleeding in Spain showed that a restrictive vs liberal (when hemoglobin is <7 vs <9 g/dL) transfusion strategy was associated with improved 6-week survival (95% vs 91%; P = .02). 17 However, patients with ischemic heart disease were excluded from enrollment. 17 A meta-analysis of randomized trials comparing restrictive vs liberal strategies and including more than 3000 patients with cardiovascular disease 14 (not only patients with coronary artery disease) found that the risk of ACS was increased in patients managed with the restrictive strategy. However, the studies included in this meta-analysis included patients in various contexts (eg, hip surgery), which is quite different from the scenario of patients with recent AMI and anemia.
In the field of cardiac surgery, two randomized trials did not find improved outcomes with either restrictive or liberal strategies. 15,16 However, a meta-analysis of randomized trials in cardiac surgery showed a trend towards benefit of a liberal transfusion strategy, whereas in observational studies, transfusion (vs no transfusion) was associated with higher mortality. 31  Transfusion is a costly treatment (RBC units cost more than 200 euros in France, 140 euros in Spain) and blood is a scarce resource. In addition, beyond the questions of clinical safety, efficacy, and RBC unit costs, blood transfusion is associated with an increased use of medical resources and increased length of stay; hence, the costs of RBC transfusion are probably underestimated. 41 Finally, blood bank organization, shortage of blood, and costs of storage, testing, and delivery remain major issues in the western world, but even more so in low-and middle-income countries. 42 For those reasons, a testing strategy combining clinical noninferiority and costeffectiveness superiority is legitimate, as establishing clinical noninferiority of a restrictive strategy but reduced costs would be sufficient to establish dominance of this strategy and change clinical practice.
The choice of an open-label design was justified by the difficulty in blinding the intervention (blood transfusion in the context of AMI).
However, the primary clinical effectiveness outcome is a composite of objective clinical outcomes (death and nonfatal major cardiac events: stroke, AMI, and emergency coronary revascularization prompted by myocardial ischemia), which are adjudicated by a CEC blinded to randomization and hemoglobin levels, providing reassurance that these should not be affected by investigator bias.

| CONCLUSION
REALITY is the largest randomized trial to date comparing transfusion strategies in patients with AMI. The trial will provide important clinical and cost-effectiveness information to guide the management of patients with AMI and anemia.

ACKNOWLEDGMENTS
Editorial support was provided by Jenny Lloyd and Sophie Rushton-Smith, MedLink Healthcare Communications, London, and was funded by the Programme de Recherche Medico-Economique (PRME) from the Health Ministry in France.