ES-BempeDACS is an independent clinical trial designed and conducted by researchers without the participation of the pharmaceutical industry. The trial follows a pragmatic, controlled, prospective, randomized, and unmasked design with blinded evaluation of endpoints (PROBE design).

Patients for inclusion will be identified within the first 72hours after hospitalization for ST-segment or non–ST-segment elevation AMI. Patients meeting the inclusion criteria (and none of the exclusion criteria) will be invited to participate. Patients can be selected after hospital discharge as long as randomization occurs within the first 72hours after symptom onset.

Inclusion criteria

The inclusion criteria are as follows (figure 1):

• •

  Men or women who are not pregnant or breastfeeding and who are aged ≥ 18 years.

• •

  Recent type 1 ST-segment or non–ST-segment elevation AMI (< 72hours). The concept of type 1 AMI is based on the fourth universal definition of AMI6 and is limited to infarcts caused by atherosclerotic coronary artery disease.

• •

  LDL-C ≥ 70mg/dL despite treatment with high-potency lipid-lowering statins at high doses in a stable baseline regimen (≥ 4 weeks) or statin-intolerant patients.

• •

  LDL-C ≥ 100mg/dL in patients under treatment with low-potency statins or high-potency statins at submaximal doses.

• •

  LDL-C ≥ 115mg/dL in statin-naïve patients.

• •

  Signed informed consent.

Figure 1.

Central illustration. ES-BempeDACS (Efficacy and Safety of BempeDoic Acid in Acute Coronary Syndrome) study algorithm. AMI, acute myocardial infarction; BA, bempedoic acid; CV, cardiovascular; EZ, ezetimibe; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein-cholesterol; PCSK9, proprotein convertase subtilisin/kexin type 9; TG, triglycerides. gr1.

(0.41MB).

For the purposes of the study, high doses of high-potency statins are considered to be 80mg/d for atorvastatin and 40mg/d for rosuvastatin. The aim of this criterion is to define profiles for patients with suboptimal metabolic control despite high-potency statins at high doses, the most complicated setting for achieving good metabolic control with oral lipid-lowering agents.

Patients will be selected during hospitalization for AMI. During this period (within the first 72hours after AMI onset) and after signing informed consent, patients will be randomized 1:1 to one of the following 2 treatment groups:

• •

  Intervention group: treatment with a single daily dose of BA 180mg + EZ 10mg (combined in a single tablet), in addition to conventional treatment with a high-potency statin at high doses.

• •

  Control group: no treatment with BA (usual care with high-potency statin ± EZ 10mg).

Randomization will be centralized through a specific study platform available online or via a smartphone application. Randomization will follow the permuted block design, considering 3 variables (sex [male or female], baseline lipid-lowering therapy with/without high-dose high-potency statin therapy, and LDL-C ≥ 100mg/dL). The block size will vary to ensure continual size equivalence for each group and with additional stratification by center.

The responsible physician will have the discretion to change each patient's treatment group. Patients and their responsible physician will be asked to contact the central coordinating center to report a group change before any decision is made. All decisions will be made by the responsible physicians but only if these changes are sufficiently indicated.

Group changes will be recorded, along with the main reason for the change. Because the main study end point is based on the “intention-to-treat” principle, patients will be analyzed according to randomization, independently of any group changes. A posteriori analysis will consider group changes and will analyze patients who remain in the original group randomization during follow-up (vide infra the prespecified groups).

In addition to BA/EZ therapy (or control), patients will be treated in accordance with current recommendations and the criteria of the responsible physician. Information will be collected from each patient on demographic and baseline variables considered to be clinically relevant for the subsequent analysis of study endpoints (table 1).

The primary endpoint is the percentage of patients achieving an LDL-C < 55mg/dL at 8 weeks of treatment. The following secondary endpoints have been established: a) absolute change in LDL-C at 8 weeks of treatment; b) percentage change in non-high-density lipoprotein-cholesterol (non-HDL-C) from baseline at 8 weeks of treatment; c) percentage change in HDL-C and triglycerides from baseline at 8 weeks of treatment; and d) CV events at 12 months: acute coronary syndrome, nonelective coronary revascularization, ischemic stroke, all-cause death, and CV death.

The study medication comprises BA 180mg and EZ 10mg, administered orally in a single tablet. The sponsor (Spanish Society of Cardiology) will provide and distribute the medication to participating centers for patient administration. Patients will remain on the assigned treatment for the entire study period (8 weeks) unless their grouping changes. At the end of the 8 weeks, the study treatment will be completed. Treatment will be continued with the medication considered most suitable for the patients’ condition by the responsible physician. Neither the researchers nor the sponsor commit to maintaining the treatment after study completion.

Follow-up will comprise telephone calls and a review of patients’ medical records (table 2). Follow-up will be centralized at the coordinating center, which will assess health status (vital status and hospitalization/emergency department visits) and study group adherence (BA/EZ vs not BA/EZ). If patients experience any events, including hospitalization or emergency department visits, the local investigator will be asked to review their medical records. Once a local investigator identifies that a patient has been admitted for any event potentially relevant for the study, the researcher responsible for data collection (eg, hospitalization report, test results) will send pseudonymzed data to the coordinating center for blinding and event adjudication. Follow-ups will be conducted at 2 time points: at 8 weeks after randomization and at 12 months.

The study sample size is based on the percentage of patients who achieve LDL-C values < 55mg/dL. The percentage achieving the primary end point (LDL-C < 55mg/dL) is predicted to be 30% in the control group and 50% in the BA/EZ group (ie, 40.3% of the total study population). To reduce the possibility of an underpowered study due to a lower than expected percentage of patients achieving an LDL-C < 55mg/dL, a blinded analysis will be performed of the overall percentage of the population with an LDL-C < 55mg/dL at 6 weeks instead of at 8 weeks (planned study completion date). We predict that at least 140 patients will complete the 8 weeks of follow-up. If the overall percentage of the population with an LDL-C < 55mg/dL is less than 40%, the data safety monitoring board will consider recalculation of the sample size in accordance with the actual percentage, and the president will decide if it is necessary to increase the sample size (and, therefore, prolong the selection period). The number of patients required was calculated for evaluating the primary end point, with an estimated sample size of 206 patients, power of 80%, and a predicted dropout/loss percentage of 10%.

Two interim analyses of both the primary endpoint and secondary endpoints will be conducted after validation of 33% (n = 68) and 66% (n = 136) of the analytical controls performed at 8 weeks of treatment. The stopping rules for the analysis of efficacy will be based on the Haybittle-Peto method, with a total alpha value of 0.05. Safety analysis will be performed by applying descriptive statistical methods to all data and will be accompanied by calculations of confidence intervals. No formal statistical stopping rules for safety data will be predefined.

SPSS 25 and Stata MP64 version 14 statistical software will be used for analysis. Analyses of the efficacy and safety endpoints will be conducted in the “complete analysis group”, following the “intention-to-treat analysis” principle; the intervention group will be compared with the control group. To preserve the advantages of randomization and minimize bias, the intention-to-treat analysis will also be applied to the evaluation of the safety end point relative to the occurrence of events at 12 months by comparing the intervention and control groups.

The percentage of patients with an LDL-C < 55mg/dL at 8 weeks of follow-up will be calculated in each treatment group by dividing the number of patients with an LDL-C < 55mg/dL by the total number of patients in each group. Comparisons between the 2 percentages will be performed using the Pearson chi-square test for independent data. Missing LDL-C values (and those of other continuous variables) will be imputed using Rubin's multiple imputation method. The Stata command “mi impute mvn” will be used.7 This command uses an iterative method based on Markov chains to generate multiple imputed datasets, assuming an underlying multivariable normal model. The command is designed to adjust to monotonic and arbitrary missing value patterns. For asymmetric data, log transformation will be used before the imputation process. To determine the number of imputed datasets, the percentage of missing data relative to the total available data will be considered. The number of datasets to be imputed will be at least as large as the percentage of individuals with missing values, which will result in estimates with lower variability (ie, the Monte Carlo error will be low). It is currently standard to use between 20 and 100 imputed datasets. In this study, the number of datasets to be imputed will initially be 10. To ensure that the imputation model will be sufficiently general to encompass subsequent analyses, the imputation model will include the outcome variable of interest, the interactions, and the covariables correlated with the variables with missing data. Potentially independent variables in the multivariate models for imputation will be statistically associated (P < .05) with the dependent variable (eg, LDL-C) in the univariable analysis, as well as variables with a known association in the literature with the dependent variable of interest, independently of their statistical significance in this study. All variables comprising the imputation model will be included in the model analyzing results.

The percentage change in non-HDL-C, HDL-C, and triglycerides at week 8 will be analyzed using analysis of covariance (ANCOVA). The ANCOVA model will include the treatment and the randomization stratum as factors, and the baseline value of the variable of interest as a covariate. The assumptions of the ANCOVA model will be assessed. If these assumptions exhibit severe violations, a nonparametric alternative approach will be used.

To better explore the effect of treatment under ideal circumstances (explanatory approach), the primary and secondary endpoints will also be analyzed in “per-protocol group” datasets. The above-described methods will be applied.

The safety endpoint, assessed by the change in transaminase, bilirubin, CPK, and uric acid levels, will be examined by comparing the baseline means of these parameters in each treatment group with those at week 8. For this comparison, we will use an independent samples t test.

Regarding the event analysis at 12 months, on the last day of follow-up, the data of all patients who have not experienced any event will be censored. The data of patients who have dropped out will be censored from the day of dropout. Given that the incidence of events such as acute coronary syndrome, nonelective coronary revascularization, ischemic stroke, and CV death compete with all-cause death, competing risk analysis will be used to provide a more accurate estimate of the risk of these events.8 The competing risk analysis will be performed to evaluate all events (acute coronary syndrome, nonelective coronary revascularization, ischemic stroke, and CV death) except all-cause death. The cause of death (CV vs non-CV) will be established according to the definition used by Hicks et al.9

Statistical significance and subhazard ratio values with their respective 95% confidence intervals (95%CIs) will be calculated and estimated using 1000 bootstrap replications (resampling with replacement). The incidence of each specific event will be estimated and presented as survival curves expressed as the function of cumulative incidence, which provide more accurate estimates than the Kaplan-Meier method in the context of a relatively long follow-up such as that established in this study.10 For the analysis of all-cause death, we will use Cox regression and Kaplan-Meier curves.

Due to the randomization process and the study sample size, no major imbalances in covariables are expected between the treatment groups. However, standardized differences will be calculated for exploratory endpoints. Previously identified prognostic factors with standardized differences > 0.1 will be included in a Fine-Gray multivariable competing risk model (the Cox model will be used for all-cause death). In all of these tests, 2-sided P values will be used with an alpha significance ≤ .05.

All subgroup analyses will be conducted using interaction models based on the main analysis model, including as factors the subgroup in question and the subgroup-treatment interaction, which will provide the statistics for each subgroup and the P for interaction. The subgroups to be analyzed will be women vs men, baseline treatment with vs without high-potency lipid-lowering statin therapy, baseline treatment with vs without EZ, and baseline LDL-C ≥ 100mg/dL.

The participants and their families will receive detailed information on the risk and possible benefits of their participation in the study. Consent documents have been approved by the reference institutional review board for medicinal products (CEIm-G, the Spanish acronym for Comité de Ética en la Investigación con Medicamentos). All participants will sign the informed consent document before participating in the study. The participants’ rights and wellbeing will be protected by stressing that the quality of medical care received will not be negatively affected if they decide not to participate in this study.

This work comprises an investigator-initiated clinical trial with a pragmatic design. Although this design allows for greater generalization of results due to less strict inclusion and exclusion criteria, it is associated with lower quality control in data collection and follow-up. However, we hope to compensate for this with the elevated motivation of the research team. In addition, the calculation of the sample size is largely centered on the primary end point (percentage of patients with LDL-C < 55mg/dL at 8 weeks after AMI). The remaining analyses of clinical and safety events are merely hypothesis generating because the study lacks the power to obtain conclusive results.