Chest Pain Triage at ED

Background

Acute chest pain triage in the Emergency Department (ED) may prove challenging, since similar symptoms can reveal both mild or life-threatening disorders, and the clinical presentation itself is known not to be reliable for evaluating the risk of acute coronary syndrome (ACS) .

The first priority in patients with suspected ACS is to identify those with ST-elevation myocardial infarction (STEMI), since these patients require urgent reperfusion therapy. Both European and American cardiology guidelines recommend implementation of a 12-lead electrocardiogram (ECG) within 10 min of arrival in the ED.

In contrast, in the absence of electrocardiographic changes consistent with ischemia, the ACS rate is low (1-8% depending on the number of risk factors) , and ACS are associated with lower in-hospital mortality.

Emergency Department (ED) triage systems for acute chest pain are either based on clinical features or on a 12-lead ECG. A recent study showed the systems having a similar diagnostic performance but different characteristics since clinical-based triage had a higher sensitivity and ECG-based triage had a higher specificity. The lack of sensitivity of ECG based triage was explained by the fact that patients with a normal or non-ischemic ECG were assigned to a low-acuity triage score, defined as inadequate (false negative) for ACS patients. Conversely, the lack of specificity of clinical based triage was due to a large number of patients with benign pathology assigned to a high-acuity triage score to perform a 12-lead ECG within the 10 minutes after the first medical contact as required.

The investigators therefore hypothesized that combination of 12-lead ECG and cardiovascular risk factor could increase the sensitivity of ECG based triage for chest pain patients with a normal or non-ischemic ECG.

Methods

Selection of participants

All consecutive patients above 18 years of age, presenting at ED with acute non-traumatic chest pain as main complaint will be systematically included in the study.

Study design

The investigators plan a prospective, single-center, observational study in a teaching hospital in Brussels, Belgium.

Triage of chest pain patients will be performed using the "French Emergency Nurses Classification version 2" (FRENCH) triage system, which is currently used in our ED. FRENCH score is a 5-levels triage system classifying patients from category 1 corresponding to the "immediately life-threatening condition" to category 5 corresponding to the "less or non-urgent situations". Maximum authorized waiting period before first medical assessment according to severity category are respectively: immediate response, 20 minutes, 60 minutes, 120 minutes and undefined. Initial triage of chest pain patients is based on vital signs, than if normal patients with a normal or unmodified ECG are sorted into category 3, those with an abnormal ECG without evidence of ischemia are classified into category 2, while those with ECG showing signs of ischemia (ST elevation, ST depression, new onset of a left bundle branch block, or T-wave inversions) are classified into category 1. Patients with lateral chest pain of a probable musculoskeletal nature are sorted into category 4 without performing an ECG.

The FRENCH triage (ECG-based triage) will be compared to a modified FRENCH triage system (ECGScore- based triage) upgrading high-risk patients with a normal ECG from category 3 to category 2. The investigators defined high-risk patients as patients having a Systemic Coronary Risk Estimation (SCORE) above 10 % according to the 2016 European Guidelines on cardiovascular disease prevention in clinical practice. To facilitate calculation of the SCORE at triage, the investigators defined one major or four minor criteria required to upgrade the patients.

The major criteria are:

1. History of Cardio-vascular disease
2. Acute coronary syndrome
3. Stroke or IAT.
4. Arterial revascularization procedure of peripheral artery.
5. Diabetes
6. End-stage renal disease or dialysis

The minor criteria are:

1. Man
2. Above 60 years of age
3. Hypertension
4. Smoker
5. Dyslipidemia

The above criteria will be collected by the triage nurse using a computerized form designed for the study. On this basis, the computer will additionally sort the patients using the modified triage scale, with blinded result avoiding bias in the initial nurse triage. This modified triage will therefore not be taken into account and management of the patients will not be modified.

Medical data will also be collected by physicians using a dedicated computerized form, particularly cardiovascular risks factors and complications occurring during the ED stay (arrhythmia, heart failure, shock).

The final diagnosis will be determined at the end of a 30-day follow-up, by either reviewing patient's files or a phone call to patients or relatives. ECG interpretation and diagnosis of acute myocardial infarction will be based on the universal definition of myocardial infarction. Pulmonary embolism and aortic dissection will be substantiated by computed tomography (CT), chest infection by relevant positive biomarkers and a chest X-ray, gastroesophageal reflux by gastroscopy or positive therapeutic test with a proton pump inhibitor, and abdominal pathology by biology, imaging, or both. Musculoskeletal pain will be diagnosed if all of the above-mentioned investigations are normal, and movement and palpation enhanced superficial pain. Finally, patients with a normal physical examination, ECG, chest X-ray, and blood tests who did not report any event during the 30-day follow-up, and who therefore have no definite cause for their chest pain, will be sorted into a dedicated group called 'unclear diagnosis'.

The local hospital ethics committee approved the study protocol (CEHF 2017/18AOU/407), and patients will provide their written informed consent to participate in this study.

Statistical methods

To evaluate triage system's diagnostic performance, the investigators will compare the triage score of patients with ACS (final diagnosis of STEMI, non-ST-elevation myocardial infarction [NSTEMI] or unstable angina [UA]) to a control group made up of patients with chest pain from mild severity diseases (including digestive diseases, chronic obstructive pulmonary disease (COPD) exacerbation, musculoskeletal chest pain, and all unclear diagnoses without any medical event in the follow-up period). Patients with other severe diseases, abnormal heart rate, abnormal blood pressure, respiratory distress, or hypoxemia will be excluded from the control group having a triage score driven by these conditions rather than by the ECG.

Since for ACS most of the arrhythmic events occur within 12 h of symptom onset and ACS patients should be monitored, the investigators defined for ACS a triage Score of 1 or 2 (high-acuity triage score) as adequate (true positive), and a triage Score of 3, 4, or 5 (low-acuity triage score) as inadequate (false negative). Conversely, for the control group with mild severity diseases, the investigators defined a triage Score of 1 or 2 (high-acuity triage score) as inadequate (false positive), and a triage Score of 3, 4 or 5 (low-acuity triage score) as adequate (true negative). Receiver operating characteristic (ROC) curves will be generated to assess each triage system's accuracy in classifying a patient in ACS versus control group.

Capability of triage nurses to collect cardiovascular risk factors will be assessed comparing data collected at triage to these collected by the physician in the medical computerized file.

Sample Size

Based on a previous study, FRENCH score AUC under the ROC curve was 0.69. The investigators hypothesized a corelation of 0.9 between the two tests. To show a difference of 0.06 between the AUC of the tests with a power of 0.8 (alpha 0.05) we calculated a sampled size of 233 patients. Given the uncertainty of the correlation, we planned to enrol 500 patients.