Primary outcome measure
The primary aim of the present study is to determine the relative safety of the 0/1-h algorithm compared with the current standard of care for patients presenting with suspected NSTE-ACS to the ED. Primary outcome measure is the incidence of major adverse cardiac events (MACE), defined as the composite of all-cause mortality, new myocardial infarction (type 1 or 2), unstable angina, or revascularization within 30 days.
Secondary outcome measures:
The following predefined secondary outcomes are considered:
The proportion of patients managed as outpatients in order to quantify the efficacy of the investigated approaches in both arms.
The time from ED presentation to discharge in outpatients. The ED dwell time is derived from a common electronic patient record system used across all participating sites and is calculated as the time of the medical examination between the start time (sign-in time: defined as first medical contact with a MD or the triage nurse) and the end time (sign-off time: defined as the time when the attending physician has finished all medical procedures in the ED, including charting, with the patient to ready to leave the examination room either to be discharged or transferred to a different unit/ward within the hospital).
All-cause mortality (cardiovascular death and non-cardiovascular death) and new myocardial infarction (type 1 or 2) within 30 days after the index presentation.
Incidence of unplanned, urgent cardiac revascularization (e.g., PCI or CABG) within 30 days after the index presentation (defined as any unplanned hospitalization with cardiac revascularization performed within the first 12 h after hospital admission in the context of ACS).
Incidence of ED representation or hospitalization for unstable angina within 30 days after the index presentation.
Costs for healthcare resource use within 30 days following the index ED presentation are calculated based on each country’s guidelines and cost tables for hospitals. Different costs are used for academic and general hospitals. For each patient, the costs are calculated based on the observed number and type of healthcare resources used.
Incidence of cardiac examinations (e.g., stress testing or coronary angiography) within 30 days after the index presentation. Adherence to the intervention condition (the 0/1-h algorithm) was evaluated.
Study design and population (Fig. 1)
This is a prospective, international randomized trial involving eleven hospitals (Japan; Juntendo University Nerima Hospital, Juntendo University Urayasu Hospital, Sakakibara Heart Institute, Shimada general hospital, Teikyo University School of Medicine, Yokohama Rosai Hospital, Singapore; National Heart Centre Singapore and Sengkang General Hospital, Thailand; Maharaj Nakorn Chiangmai Hospital, Taiwan; Taipei City Hospital Yangming Branch, Malaysia; University Teknologi MARA Teaching Hospital, Kuala Lumpur Hospital, and Institute Jantung Negera) from five Asian countries. These hospitals provide advanced interventional therapy to patients presenting with suspected ACS. Patients of ≥ 18 years of age who presented to the ED of participating hospitals with symptoms suggestive of NSTE-ACS (e.g., chest pain, epigastric pain, radiating pain to the neck or shoulder, shortness of breath) are eligible for enrollment in the DROP-Asian ACS study, if the onset of symptoms occurred within the last 24 hours. Patients were excluded from the present study if they presented with ST-elevation myocardial infarction, cardiogenic shock, on chronic hemodialysis, or chest pain following chest trauma, or if they were unwilling or unable to provide their informed consent. All participating patients or their legally authorized representatives will provide written informed consent. This study is conducted according to the principles of the Declaration of Helsinki. Ethical approval has been received from the local ethics committees. This study is registered in the Japanese Trial Registry (Reg. No. UMIN ID; UMIN000042461; https://upload.umin.ac.jp/cgi-openbin/ctr_e/ctr_view.cgi?recptno=R000048466).
Routine clinical assessment
All study patients will undergo initial clinical assessments along with completion of the case record form, which includes a physical examination, documentation of symptoms, clinical history, cardiovascular risk factors, 12-lead electrocardiography (ECG), bedside cardiac ultrasound sonography (visual ejection fraction and an evaluation of the wall motion), chest X-ray, blood pressure, heart rate and standard blood tests. Hs-cTnT (manufactured by Roche Diagnostics, Rotkreuz, Switzerland) is routinely determined and results provided to the treating physicians as part of the local standard of care in all participating centers in both treatment groups. This hs-cTnT-assay has a 99th percentile concentration of 14 ng/L with a corresponding coefficient of variation of 10% at 13 ng/L (12). Limit of blank and limit of detection have been determined to be 3 ng/L and 5 ng/L. The physicians will complete the modified History, Electrocardiogram, Age, Risk factors and Troponin (HEART) score and the ED Assessment of Chest pain Score (EDACS) (15–17), and the HEART score will be further modified using the hs-cTnT level (18).
Stepped wedge design
This study has a stepped-wedge design, which is a type of cluster randomized trial. This design gives all participating general practitioners the opportunity to use the intervention during the study. The group of hospitals to which the intervention applies will be called a "cluster". Basically, each cluster contains one hospital, except for one cluster containing two hospitals—Shimada General Hospital and Juntendo University Urayasu Hospital—which are included in one cluster due to highly comparable geographical conditions (Chiba, Japan) and the expected relatively low number of enrollments because they can enroll the patients from Monday to Friday (9 a.m. to 4 p.m.) in both centers. Therefore, this study consists of 11 clusters with 12 hospitals. At study start, all clusters will apply 'usual care' (control condition) to all patients, i.e., risk assessment and subsequent management includes hs-cTnT but not the application of the 0/1-h algorithm. Every 1.5 months, one randomly allocated cluster will sequentially start to follow the 0/1-h algorithm after a transition period of 1.5 months in all chest pain patients (intervention condition; 0/1 care) (Fig. 2)
Sample size, level of significance, and power
A survey of ED physicians found the tolerable rate of missed major adverse cardiac events to be less than 1%(19), however, an analysis by Kline et al showed that a 2% rate of missed diagnoses of ACS might be considered acceptable considering the risk of harm from further testing(20), and more than 2% of ACS patients are inappropriately discarded annually according to a Medicare data base(21). Regional differences in the management of NSTE-ACS persist in Asian countries even after extensive adjustment for baseline characteristics and treatment(12). Indeed, in previous studies, the MACE rates tend to be higher than in Central European or North American countries and vary from 1.0–4.0%(10, 22–24). We estimated that, in the standard care arm, patients managed as outpatients by usual care would experience an event rate of 3.0%, for the primary outcome of MACE including all-cause death, subsequent myocardial infarction, unstable angina or revascularization in 30 days. The non-inferiority margin of 1.5% is based on clinical judgment and the available literature; thus, accepting an upper limit of the 95% confidence interval. No assumption was made about the level of intra-cluster correlation coefficient (ICC)(25). Based on these numbers, a one-sided significance level of 5% and a power of 80%, the required total sample size for a stepped wedge design with 11 clusters was determined to be 4080.
Control condition: Usual care
Based on current Asian guidelines, patients in the control condition will undergo ‘standard care’ including hs-cTnT testing at time of ED presentation, which will be repeated 2–3 h after presentation (with further testing at the clinician’s discretion) according to respective local standard operating procedures. Patients will have their disposition determined by the treating ED clinician. Access to outpatient care will be as per local practice. Subsequent care will be physician-determined (Table). We are opting out of reviewing the patient data. Subjects have the right to withdraw from the study at any time for any reason.
Intervention condition: the 0/1-h algorithm
The 0/1-h algorithm uses the blood concentrations of hs-cTnT obtained at time of ED presentation and their absolute changes within one hour to classify patients into the rule-out, observe, or rule-in group (8). Patients with an hs-cTnT concentration < 12 ng/L at 0h and a 1hΔ < 3 ng/L will be stratified into the rule-out group. In addition, a patient can be directly stratified into the rule-out group if hs-cTnT concentrations is < 5 ng/L at 0h and chest pain onset was more than three hours prior to ED presentation without the need for further serial hs-cTnT-testing. Those patients with hs-cTnT ≥ 52 ng/L or a 1hΔ ≥ 5 ng/L will be stratified into the rule-in group. The remaining patients will be placed into the observe group. Patients triaged towards rule-out may be considered for discharge from the ED after the exclusion of other life-threatening disorders causing chest pain (e.g., pulmonary embolism, aortic dissection, pneumothorax) according to the patients’ conditions and the judgment of the treating physicians. Treating physicians are free to overrule any triage recommendation of the ESC 0/1h-algorithm whenever deemed necessary. Written informed consent was obtained from all patients in this phase. We will evaluate the adherence to the protocol.
Transition period: Implementation support
To support the implementation of the 0/1-h algorithm during the transition period, we will provide written educational material and presentations at each site as well as training for clinical and laboratory staff. Educational material on the algorithm and decision thresholds will be presented at each ED handover (twice daily) during the transition period to ensure wide coverage of staff on all shift patterns. This will be reinforced by specialist chest pain nurses who will receive detailed training prior to implementation and who support ED clinicians in assessing patients with suspected ACS. Key details from the educational presentation will be formatted into a one-page reference guide that will be posted within each department and online in the hospital guidelines portal. This information will also be presented to the wider hospital teams in medical grand round presentations prior to implementation and circulated to all general practitioners. Laboratory staff will also receive training to ensure that any queries directed to the laboratory will be dealt with consistently. Finally, the research team will include senior cardiologists, emergency physicians and cardiology nurses who are clinically active within each of the hospital clusters; education will therefore be reinforced at a local level by these clinical leaders throughout the transition period. This period will not be subject to an analysis.
Simple randomization will be performed. The randomization-sequence of the participating clusters will be generated using computer-generated pseudo-random numbers by a statistician at the Japanese Organisation for Research and Treatment of Cancer (JORTC), who will be not otherwise involved in the study.
The adjudicated final diagnosis.
Both components of the primary outcome (all-cause mortality, new AMI type 1 or 2, unstable angina, or revascularization in 30 days) are adjudicated by two independent cardiologists. Interrater reliability is assessed by documenting the number of patients with mismatch in the final diagnosis of AMI by the 2 adjudicating cardiologists, which requires involvement of a third cardiologist. AMI will be defined according to the Fourth Universal Definition of Myocardial Infarction, which requires evidence of myocardial necrosis in addition to ischemia (as shown through ECG changes, positive findings of stress ECG, myocardial perfusion, coronary CTA or CAG) (26). In brief, necrosis will be diagnosed based on an increase or decrease (20% relative increase) in the hs-cTnT concentration with at least 1 measurement above the 99th percentile of the normal reference range at a level of assay imprecision near 10%. Type I myocardial infarction (MI) is characterized by atherosclerotic plaque rupture, ulceration, erosion or dissection, with the resulting intraluminal thrombus in one or more coronary arteries leading to myocardial necrosis (any hs-cTnT concentration above 14 ng/L, with an increase and decrease in the hs-cTnT concentration where serial testing is available). Type II MI is defined as myocardial necrosis in which a condition other than coronary plaque instability contributed to an imbalance between the myocardial oxygen supply and demand (e.g., coronary artery spasm, coronary embolisms, hypertension or hypotension). All other patients will be classified as “no AMI” in this analysis. Unstable angina (UA) will be diagnosed in patients with typical ischemic symptoms at rest or minor exercise with no evidence of acute myocardial necrosis (normal or elevated hs-cTnT levels without any relevant change during serial sampling). Vasospastic angina pectoris was diagnosed in patients with no obvious coronary stenosis, in which an ECG revealed ST elevation, or there was angiographic evidence of a coronary artery spasm that was released after the intracoronary administration of 1 mg of nitroglycerin (18, 27, 28). Atypical chest pain was diagnosed based on the absence of findings from laboratory tests, ECG and chest radiography at the 30-day follow-up.
The principal investigator and data committees have regular web meeting the months to monitor the current status of registration and remind to follow the protocols.
The steering committee of emergency physicians, emergency and cardiology nurses, cardiologists, implementation experts and clinical pathologists with representation from each of the participating hospitals is providing oversight to the study. Data management is centralized at Novelle PACE Inc (Tokyo, Japan). REDCap (Research Electronic Data Capture) will be used as electronic case report form (eCRF) to collect and manage study data. All local and server databases will be secured with password-protected access systems, access logging and encryption. Members of the endpoint committee have access to the patient information via encrypted server access and encrypted hard drives. A Clinical event adjudication committee, independent of the study management team, is providing blinded evaluations (events deidentified for treatment arm, hospital and patient details) of all components of the primary endpoint, including index (within 12 h of initial presentation) and subsequent MI. All study-related information will, until collection, be stored securely at the study sites in areas with limited access. After collection and entry, laboratory specimens, reports and data collection forms will be identified by a coded participant ID number to maintain participant confidentiality. Forms and any other listings that link participant ID numbers to other identifying information will be stored separately from study records with the participant ID number.
This is an investigator-initiated study funded by a Grant-in-Aid for Scientific Research (No. 18K09954)with additional funding through a restricted grant from Roche Diagnostics. Roche Diagnostics were approached after the study was designed, ethical approval gained, and enrollment commenced, and their contribution was not dependent on any protocol modification or direct access to the study data.
Continuous variables will be presented as the mean (standard deviation) or median (interquartile range [IQR]), and categorical variables will be presented as numbers and percentages. The 30-day incidence of MACEs will be analyzed using generalized linear mixed models after adjusting for background information and seasonal effects. The clustering effect by hospital will be considered in this analysis as a random effect. The binomial distribution and identity link will be used to directly estimate the absolute differences in MACE incidence between 0/1-h algorithm care and usual care patients. The main model will include conditions (control or intervention) and steps (time periods) as categorical variables and hospitals as clusters. Differences in MACEs with a 1-sided 95% CI will be estimated in order to evaluate non-inferiority. The non-inferiority margin is preset at 1.5%. The sensitivity and negative predictive values (NPVs) for MACEs in the rule-out group and the specificity and positive predictive values (PPVs) for MACEs in the rule-in group will be calculated. Multiple imputations will be used to handle missing data. The statistical analyses will be performed using the SAS software program, version 9.4 (SAS Institute, Cary, NC, USA), and R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).
Patient and Public Involvement
A patient review panel was consulted throughout the trial program and provided input on the educational advice provided to clinicians after the introduction of the new pathway. Patients were not involved in the conception or design of the trial.
These analyses will be performed in six hospitals in Japan, Taiwan and Thailand, only, as the remaining centers participating in this study lack a system for conducting cost analyses.
The diagnosis procedure combination (DPC; flat-fee payment system) is the bundled payment system of medical fees for acute inpatient medical care in Japan (29).
The bundled payment for each hospitalization will be calculated according to the codes in the International Classification of Diseases 10th revision (ICD-10) and the coefficient for each facility. The DPC database contains six categories of diagnoses, each with a limited number of recordable diseases. One diagnosis each is coded for the “main diagnosis,” “admission-precipitating diagnosis,” “most resource‐consuming diagnosis,” and “second-most resource‐consuming diagnosis.” A maximum of 4 diagnoses (10 diagnoses from 2016) each can be coded for “comorbidities present at time of admission” and “conditions arising after admission.” All procedures performed during hospitalization are recorded according to the Japanese fee schedule for reimbursement. The DPC component corresponds to the so-called “hospital fee,” including the basic hospital charge, pharmaceuticals, injections, laboratory examinations, and other related expenses, and is paid on a per day payment scheme. The fee-for-service (FFS) component corresponds to the charges for surgical procedures, angiography and other related expenses. The revenue equals the sum of the DPC and FFS components. “Material costs” include the costs of pharmaceuticals and medical materials. The total hospitalization costs are calculated as the sum of bundled payment and FFS, without the food fee (30).
The data will be anonymized and not include any information that might be used to identify individuals or hospitals. Each patient will have an identifier specific to the hospitals, with all patients regarded as a single individual, regardless of having both inpatient and outpatient data. The DPC data used for payments include patient demographics and select clinical information, admission and discharge statuses, diagnoses, surgeries and procedures performed, medications and special reimbursements for specific conditions. In this study, the DPC data will be evaluated from admission to the 30-day follow-up period. The data will be excluded if patients receive care from a non-cardiology department.
The DPC is the payment system of medical fees for acute patient medical care in Taipei City Hospital. The payment for hospitalization will be calculated according to the medical cost requested from the National Health Insurance program. The database contains cardiologic-related diagnoses. Each patient is coded for the “main diagnosis,” “admission-precipitating diagnosis”, “every comorbidity at every admission” and “second-most resource‐consuming diagnosis.” All procedures performed during hospitalization are recorded according to the Taiwanese fee schedule for reimbursement. The total cost component corresponds to the so-called “hospital fee,” including the basic hospital charge, pharmaceuticals, injections, laboratory examinations, surgical procedures angiography and other related expenses, and is paid on a per-day payment scheme. The total hospitalization costs are calculated as the sum of all medical costs, without the food fee, which is recorded in the hospital computer system.
All participants in this study will be identified from the hospital electronic database. The total cost will be disaggregated into individual cost items, including drugs, laboratory fees and service costs, such as interventions, operations and room costs, incurred at Maharaj Nakorn Chiang Mai Hospital, Faculty of Medicine, Chiang Mai University.