PRESC1SE-MI (PRospective Evaluation of the European Society of Cardiology 0/1h-algorithm`s Safety and Efficacy for Triage of Patients with Suspected Myocardial Infarction) is an investigator-initiated, international, multicenter, stepped-wedge, cluster randomized controlled trial. This specific study design is best suited to evaluate the implementation and implications of a novel algorithm in a real-world setting and allows a sequential implementation of the novel algorithm in participating centers. The study involves 3 groups of study centers starting recruitment at 3 different time points. Each group starts enrolment during a validation phase, where current practice with the ESC 0/3h-algorithm is continued. Centers of each group are being randomized to early vs late implementation of the ESC 0/1h-algorithm (implementation phase; Figure 1 ). Continuously collected individual patient data will be used to assess the primary outcomes based on patient-average treatment policy estimands. The use of a cluster-randomized design permits the inclusion of all patients with suspected MI, which prevents selection bias and confusion amongst clinicians who would otherwise be required to simultaneously assess patients using different diagnostic criteria for MI. The study started enrolment in December 2020. Patient recruitment was completed in December 2024 and follow-up is ongoing. The results for the primary outcome are expected to be available in 2026.

Simplified study diagram indicating the time points of study initiation for the 3 groups of centers. The numbers 1-3 refer to the groups of centers. Green bars indicate 6 months of current practice, grey-shaded bars indicate1 month of blanking phases, yellow bars indicate 6 months of implementation phases.

The study protocol was approved by all local Research Ethics Committees. The trial is registered at ClinicalTrials.gov ( clinicaltrials.gov/ct2/show/NCT05649384 ).

The study is performed at 19 sites in 10 countries, with 16 sites in 7 European countries (Austria, Finland, Italy, Spain, Switzerland, Romania, United Kingdom), 1 site in Asia (South Korea), 1 site in Australia, and 1 site in the United States. As it is crucial for this study to have unbiased patient enrolment and accurate assessment of recurrent MI and death during follow-up, centers were recruited based on the following methodological preferences: highest preference for centers which operate with a “general consent” (GC). In this setting, all patients are asked to sign the GC at hospital presentation as part of routine practice. No study-specific measures at the time of patient`s enrolment in the ED are performed. Only routine clinical data is collected. This allows unbiased evaluation of patient management during validation and implementation phases. Second highest preference is for centers which do not have a GC, but which do have regional and national registries documenting death and hospitalization for MI. The procedure in these centers (e.g., in the United Kingdom) is the same as successfully applied in previous trials. ^, Third highest preference is for centers which neither have a GC nor a regional or national registry for death and rehospitalization for MI. In this setting, all patients are asked to sign a study-specific Informed Consent Form (ICF) at hospital admission. The feasibility of reliable and reproducible semiautomated identification of patients with nontraumatic acute chest discomfort and suspected MI presenting to the ED was assessed prior to the activation of the study sites. The finally participating centers are shown in Figure 2 and further details are provided in Supplementary Table 1. One study site had to be closed on October 11, 2021 (prior to randomization) due to unforeseeable inability to systematically collect 30-day follow up information. All consecutive patients presenting with nontraumatic acute chest discomfort and suspected MI to the ED are identified using site-specific triage coding. The patient flow is shown in Figure 3 .

Study sites.

Patient flowchart. ESC, European Society of Cardiology; MI, myocardial infarction.

Study centers must use a hs-cTn assay with recommended assay-specific 99th percentile upper reference limits and the ESC 0/3h-algorithm as current practice and intend to implement the assay-specific ESC 0/1h-algorithm using the same hs-cTn assay.

Consecutive adult (age ≥18 years) patients presenting with nontraumatic acute chest discomfort and suspected MI to the ED will be enrolled in the trial. The identification of patients will be done via the triage coding in the ED ensuring comparable patient populations throughout the study. Patients with terminal kidney failure on chronic dialysis and those with out-of-hospital cardiac arrest or cardiogenic shock at presentation will be excluded, as these patients had been excluded also from all previous studies and no evidence supports the use of any triage algorithm in this highest-risk population. Moreover, patients who are nonresidents of a country where the study is being conducted are excluded, as these patients are not listed in national death registries which are mandatory for assessing death during follow-up. For sites that have only access to a regional death registry rather than the national death registry, also patients from other regions are excluded.

The study population is the one that is recommended for the use of the triage algorithms in current clinical practice guidelines. ^,, We do not expect any changes in the target population and the associated risk profile over the course of the trial, as eligible patients are identified by triage coding in the ED independent of research staff. The implementation of the ESC 0/1h-algorithm is the only intervention with an impact on patient care. The risks associated with MI lead to large numbers of patients presenting to the ED with undifferentiated chest pain of suspected cardiac origin. As symptoms suggestive of MI are one of the most common causes of presentation to the ED, safely accelerating the care for patients with suspected MI is likely to have beneficial downstream effects also to all other patients in the ED, e.g., due to increased availability of ED beds.

There is no restriction on type, manufacturer, or location (central laboratory vs point-of-care) of the hs-cTn assay used at each site. Similarly, each site is free to use either the uniform or sex-specific URL for the diagnosis of MI. ^, As recommended by the guidelines, the use-optimized and validated assay-specific decision thresholds are used for rule-out and rule-in of MI within the ESC 0/3h- and the ESC 0/1h-algorithms.

The co-primary safety outcome is defined as a composite of all-cause mortality or new type 1 MI within 30 days of index presentation. Death or type 1 MI that occur within 12 hours of index presentation are considered index events and not as endpoint events. In contrast, death or type 1 MI that occur later than 12 hours following index presentation are considered an endpoint event. Likewise, death or type 1 MI that occur within 12 hours following index presentation in patients that were initially discharged from hospital will also be considered an endpoint event. We do not differentiate whether type 1 MI occurred at a de-novo coronary lesion or related to a previous coronary intervention. Both components of the primary safety outcome are adjudicated by independent senior physicians blinded to group assignment according to the Fourth Universal Definition of MI.

The co-primary efficacy outcome is the time from ED presentation to ED discharge or transfer to another department or hospital.

Secondary outcomes include (1) the proportion of patients managed as outpatients (hospitalization is the major driver of costs and therefore an increase in the proportion of outpatients after implementation of the ESC 0/1h-algorithm could significantly reduce healthcare costs); (2) readmission (defined as a hospital stay >24 hours) for suspected MI within 30 days after the index presentation; (3) all-cause mortality at 30 days, 1 year and 3 years; (4) new type 1 MI at 30 days, 1 year and 3 years; (5) composite of all-cause mortality or new type 1 MI at 1 year and 3 years; (6) patient reported outcome measures (PROM) including patient satisfaction with the evaluation in the ED; (7) treatment costs in the ED and cost-effectiveness; (8) practical feasibility of the implementation of the ESC 0/1h-algorithm as measured by the following proportions among all patients who need a second blood draw as part of the ESC 0/1h-algorithm: (a) the percentage of patients in whom the 1h blood draw is actually done; (b) the percentage of patients in whom the 1h blood draw was performed within 1h ± 20 minutes from the 0h blood draw; (c) the percentage of patients in whom the 1h blood draw was performed within 40 and 100 minutes from the 0h blood draw; (9) effectiveness of rule-in (defined as the time from presentation to coronary angiography in patients ultimately diagnosed with type 1 MI) and of rule-out (defined as time from presentation to discharge in patients ultimately diagnosed with noncardiac disease); (10) satisfaction of ED nurses and physicians regarding the management of patients with suspected MI during both the validation and implementation phase and at 1 year and at 3 years after implementation of the ESC 0/1h-algorithm; (11) continuation of the use of the ESC 0/1h-algorithm in routine care 1 year and 3 years after implementation.

Participating study sites are stratified based on the expected number of patients presenting with acute chest discomfort and suspected MI annually and then randomly allocated to early or late implementation of the ESC 0/1h-algorithm using a computer-based randomization tool. The allocation sequence is generated at the Clinical Trial Unit of the University Hospital Basel, Switzerland.

As part of the study, time stamps of blood samples were recorded. As the turn-around time of hs-cTn testing (time from blood draw to availability of the results to clinicians) in many sites doing the hs-cTn measurement in the central laboratory is around 1h and often longer at some time, the participating sites were encouraged to train their nurse staff to obtain the 1h sample in all patients with suspected MI in order to allow the nurse team to function completely independent regarding blood sampling from the physician team. This approach includes the need to obtain a second blood draw at 1h also in the small proportion of patients in whom, according to the ESC 0/1h-algorithm, a decision of triage towards rule-out or rule-in could have been made with the hs-cTn concentration at 0h alone. The treating physician then has the choice in these patients to either make her/his definite decision only based on the 0h concentration, or what for reassurance of the 1h concentration, if obtained. This approach had also been used in most early adopting sites and was therefore recommended for use also in this study of late-adopting sites.

The 6-month validation phase provides baseline information and performance characteristics for the use of the ESC 0/3h-algorithm. Key aspects for successful implementation including acceptability, appropriateness, satisfaction with the respective algorithm leadership, resources, and innovation climate among all nurses and physicians working in the ED are assessed. At the randomization stage, participating study sites are randomly allocated to early or late implementation of the ESC 0/1h-algorithm. Before the early implementation starts, the respective study sites run a blanking phase of 1 month for intensive and standardized teaching prior to the start of the implementation phase. The other group of study sites continues to use the ESC 0/3h-algorithm and run an identical blanking phase before late implementation of the ESC 0/1h-algorithm ( Figure 4 ). Implementation success is monitored by regular upload of time stamps of the blood draws in the ED or the respective time stamps of arrival of the blood samples in the laboratory as a surrogate and educational webinars with case presentations by the local principal investigator.

Detailed study diagram indicating the design of the study with validation, first blanking, early implementation, second blanking, late implementation, and follow-up phase. These phases apply to each group of centers irrespective of the initiation time point. Green bars indicate current practice, grey-shaded bars indicate blanking phases, yellow bars indicate implementation phases and grey bars indicate follow-up for 30 days, 1-year and 3-years. ESC, European Society of Cardiology.

Follow-up is performed through access to regional and national registries documenting death and hospitalization for MI by the local investigators. At centers not using a GC or patient-level consent, patients are not directly contacted. Instead, for the 30-day follow-up, all information to assess the safety endpoints will be collected from regional and national registries documenting death and hospitalization for MI. At centers using a GC, trained and experienced study nurses performs follow-up telephone calls at 30 days according to regular quality assurance programs. Patients are interviewed on their health status and their overall satisfaction with the ED management. If allowed by the local ethical committee, study site performs a 1-year follow-up by mail, possibly including informed consent if required. If allowed by the local ethical committee, extended follow-up at 3-years is performed through access to regional and national registries documenting death and hospitalization for MI.

The study aims to test noninferiority in the composite of all-cause mortality or new type 1 MI at 30 days (safety) and superiority in the reduction of length of stay in the ED (efficacy) when implementing the ESC 0/1h-algorithm (intervention) compared to the ESC 0/3h-algorithm (current practice) in international routine care settings. Demonstration of both safety and efficacy is necessary to establish unequivocal clinical benefits of implementing the ESC 0/1h-algorithm.

a) Safety endpoint: All-cause mortality or new type 1 MI (=MACE) within 30 days

• H ₀ : odds ratio comparing the chances for MACE using the ESC 0/1h-algorithm to those using the ESC 0/3h-algorihm is 1.3 or larger

• VS

• H ₁ : odds ratio comparing the chances for MACE using the ESC 0/1h-algorithm to those using the ESC 0/3h-algorithm is smaller than 1.3.

H ₀ : median ED time using the ESC 0/1h-algorithm is larger than or the same as in the setting using the ESC 0/3h-algorithm

VS

H ₁ : median ED time using the ESC 0/1h-algorithm is smaller than in the setting using the ESC 0/3h-algorithm.

We expect that the ESC 0/1h-algorithm entails a comparable MACE rate to the current practice and that the median ED time is at least 1 hour shorter.

We estimated the sample size for 20 study sites using simulations, as advised for stepped-wedge randomized controlled trials in Baio et al. More precisely, we generated 1111 synthetic data sets for a range of potential sample sizes based on the below listed assumptions on MACE rates and median ED times. Then, the intended primary analysis is applied to each of the data sets simulated for a particular combination of simulation parameters and the power associated with a sample size is estimated as the proportion of statistically significant results across those simulated analyses. The simulations are based on the following assumptions: (1) the probability of MACE within 30 days is expected to be 3% in both study phases. Previous studies have shown event rates at 30 days among patients in whom an acute myocardial infarction was ruled-out by the ESC 0/3h-algorithm of 1.7% and by the ESC 0/1h-algorithm of 0.1%; ^,, In contrast to these studies, the PRESC1SE-MI trial includes event rates of all patients irrespective of whether they are classified as low-, intermediate- or high-risk. The rational for this is the fact that implementation of a novel pathway does not only affect the low-risk population but also likely changes triage and potential downstream investigations and treatments of patients at intermediate- or high-risk. This will also lead to differences in outcomes at 30 days. (2) the time spent by a patient in the ED follows a log-normal distribution and the corresponding median ED time with the standard of care is 5.3 hours, while it is 1 hour lower with the ESC 0/1h-algorithm, which is the smallest difference in median ED times observed in previous studies. ^,, Given these assumptions, at least 52,156 patients are required to be able to show noninferiority regarding safety as well as superiority regarding efficacy with a power of at least 90%.