False ST-segment Elevation MI Score: A Simple Diagnostic Score to Predict False ST-segment Elevation MI Activation

Acute ST-elevation MI (STEMI) is a critical medical emergency and requires prompt cardiology consultation to facilitate timely revascularisation, either via primary percutaneous coronary intervention (PCI) or thrombolytic therapy. Timely intervention is essential to minimise myocardial damage and preserve cardiac function.^1,2 The initial medical evaluation, often conducted by a primary physician, such as a junior doctor at a community hospital, an emergency physician or an internal medicine resident, plays a critical role in assessing diverse presentations of chest pain. This evaluation is essential to accurately distinguish STEMI from other urgent cardiac conditions and non-emergency cases.³ While some STEMI activations result in confirmed STEMI diagnoses and appropriate emergency management, others are false alerts, which have several adverse consequences. These include patient confusion and diminished trust in healthcare providers, increased workload, fatigue and burnout among cardiac teams, potentially compromising performance, and substantial additional costs to the healthcare system.^4,5 Common ECG findings that may lead to misdiagnosis include left ventricular hypertrophy, early repolarisation changes and left bundle branch block (LBBB).⁶

Previous studies conducted on white populations report that the prevalence of false STEMI activations ranges from 10% to 40%.^6–8 However, in Thailand there is a limited number of centres capable of PCI and fewer economic resources.⁹ STEMI consultations in Thailand are primarily conducted by transmitting the ECG through a mobile app, a free messaging platform that allows users to communicate via voice and video calls and text messages. This is followed by a phone call to the cardiologist to discuss the patient’s medical history and relevant details. To date, there have been no studies examining the prevalence of false STEMI activations or the factors associated with false-positive STEMI diagnoses in the Thai population.

Aim

The objective of this study was to determine the prevalence of false-positive STEMI diagnoses in Thai patients. Additionally, we aimed to evaluate the relationship between false-positive activations and clinical and ECG factors available at the time of diagnosis, with the goal of developing a diagnostic score to predict false STEMI activations in this population.

Methods

This cross-sectional observational study was conducted at Chiangrai Prachanukroh Hospital, a tertiary care centre in Thailand, over a 1-year period from March 2023 to February 2024. Patients were eligible for enrolment if their STEMI activations, triggered by mobile app alerts, included data on baseline characteristics, comorbidities, time to consultation, initiated STEMI activations, clinical presentations and ECG characteristics. Patients were excluded if they were in cardiogenic shock, post-cardiac arrest, acute respiratory failure, or required emergency management for ventricular tachycardia/VF. All initial ECGs that triggered STEMI activation were retrospectively reviewed by at least two board-certified cardiologists blinded to the final diagnosis and study outcomes. Each ECG was evaluated to see if it met the American College of Cardiology (ACC) and the American Heart Association (AHA) STEMI criteria. The final diagnosis of true versus false STEMI activation was determined based on a comprehensive review of medical records, coronary angiography findings and the database from the referral system.

This study was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) and was approved by the Research Ethics Committee of Chiangrai Prachanukroh Hospital (Approval No. CR 0033.102/EC.66-222).

Definitions

These are the definitions used in the study:^10–13

• Typical chest pain: typically characterised by chest, arm or jaw pain described as dull, heavy, tight or crushing, often occurring at rest.
• Atypical chest pain: epigastric or back pain or pain described as burning, stabbing or resembling indigestion.
• Non-cardiac chest pain: musculoskeletal pain that may be elicited by percussion or direct pressure, or pain following a history of trauma around the chest. This can also include known cases of biliary obstruction presenting with epigastric pain.
• Absence of typical chest pain: any characteristics that do not match the definition of typical chest pain.

Acute STEMI Diagnosed by ECG

According to the ACC/AHA guidelines, STEMI is diagnosed when there is new ST elevation at the J-point in two or more contiguous leads. The threshold for ST elevation is ≥1 mm in all leads other than V2–V3. The following sex- and age-specific cut-offs apply for leads V2–V3:

• Men aged <40 years: ≥2.5 mm
• Men aged ≥40 years: ≥2.0 mm
• Women (regardless of age): ≥1.5 mm

Additionally, horizontal or down-sloping ST-segment depression of ≥1 mm in leads V1–V3 may be indicative of posterior STEMI, particularly when accompanied by tall R waves and upright T waves in these leads.

Acute STEMI is diagnosed based on a combination of clinical symptoms, ECG findings as outlined in the ACC/AHA guidelines and elevated cardiac biomarker levels. The diagnosis is confirmed by coronary angiography demonstrating coronary artery stenosis or occlusion.

For patients classified as false STEMI, acute coronary occlusion was excluded based on absence of culprit lesions in a coronary angiogram or a combination of normal biomarkers and follow-up clinical evaluation when angiography was not performed.

Patients with elevated troponin and non-obstructive coronary artery disease were classified separately as myocardial injury or MI with non-obstructive coronary arteries, based on clinical context, imaging and angiography findings.

Reciprocal changes were defined as ST depression in leads electrically opposite the leads with ST elevation, as per standard criteria, for example, ST depression in inferior leads when anterior ST-segment elevation (STE) is present.

Left ventricular hypertrophy (LVH) diagnosed by ECG is defined by the presence of any of the following criteria: Cornell criteria, Sokolow–Lyon criteria, or an R-wave amplitude greater than 11 mm in lead aVL. LVH is considered present by the Cornell criteria if the sum of the R wave in lead aVL and the S wave in lead V3 exceeds 28 mm in men or 20 mm in women. According to the Sokolow–Lyon criteria, LVH is present if the sum of the S wave in V1 or V2 and the R wave in V5 or V6 is at least 35 mm.

Early repolarisation is defined as J-point elevation manifested as terminal QRS slurring (the transition from the QRS segment to the ST segment) or notching (a positive deflection on the terminal QRS complex), accompanied by concave upward ST-segment elevation and prominent T waves in at least two contiguous leads.

Right bundle branch block is defined by a QRS duration of at least 120 ms in a majority of beats in any of leads I, II, III, aVL, and aVF, with one of the following criteria: R′>R in lead V1 or V2; an upright QRS with an R-peak duration of at least 60 ms in lead V1 or V2; or S duration exceeding R duration in all beats in lead I or II.

LBBB is defined by a QRS duration greater than 120 ms, with lead V1 showing either a QS or a small R wave accompanied by a large S wave, and lead V6 displaying a notched R wave without a Q wave.

Normal variant is characterised by the ST segment consistently elevated by 1 mm or more in lead V2 (up to 2–3 mm) and less in V3.

Left ventricular aneurysm is defined by a pattern of persistent anterior ST elevation along with pathological Q waves.

Statistical Analysis

Categorical data were reported as frequencies and percentages, while continuous variables were presented as medians with interquartile ranges. Fisher’s exact test was used to compare categorical variables and the Mann-Whitney U-test was employed for continuous variables. To analyse factors associated with false-positive activations, univariate logistic regression was performed for variables linked to false STEMI activations. Multivariate logistic regression was then conducted to adjust for potential confounders. A p-value <0.05 was considered statistically significant. All statistical analyses were performed using Stata software.

For sample size calculation, a one-sample comparison of proportions was performed. In our study, the estimated false STEMI activation rate was 38%, with an alpha level of 0.05 and a power of 90%, leading to a planned sample size of 226 participants. Following multivariable analysis, any identified factors predicting false STEMI activations were used to generate a diagnostic score for predicting false STEMI activations.

Results

All suspected STEMI cases activated through the mobile app between March 2023 and February 2024 were reviewed. After excluding patients with cardiogenic shock, post-cardiac arrest or acute respiratory failure, and those requiring emergency management for ventricular tachycardia/VF, a total of 301 patients were included in the study (Supplementary Figure 1).

In our cohort, 115 patients (38.21%) had false STEMI activation, while 186 patients (61.79%) had true STEMI. The baseline characteristics, summarised in Table 1, include 200 men (66.45%) with a median age of 65 years (IQR 56–72). Comorbidities, such as diabetes, hypertension, dyslipidaemia, chronic kidney disease, previous coronary artery disease, stroke and smoking status, did not show significant differences between true and false STEMI activations. The timing of STEMI consultations, whether in the morning, afternoon or after midnight, did not differ significantly between true and false STEMI activations. Among the 20 first-year interns who activated suspected STEMI cases through the mobile app, 14 (70.0%) triggered false STEMI diagnoses, compared to only one (6.2%) of the 16 internists (attending physicians).

Presentation with typical chest pain was associated with a higher likelihood of true STEMI diagnosis compared to false STEMI diagnosis (87.10% versus 6.09%; p<0.001). Patients with reciprocal changes on ECG were more likely to have true STEMI diagnoses than false STEMI diagnoses (73.66% versus 2.61%; p<0.001). Supplementary Figure 2 illustrates the ECG patterns associated with false STEMI activation. The most common pattern observed was LVH (n=39; 33.91%), followed by early repolarisation (n=26; 22.61%) and a normal variant (n=15; 13.04%).

We conducted both univariable and multivariable analyses to identify predictors of false STEMI activation, adjusting for age, sex, comorbid conditions, time of activation, the physician initiating STEMI consultations, clinical presentation and ECG patterns, as outlined in Table 2. After multivariable analysis, we identified several strong predictors of false STEMI activation: absence of typical chest pain (OR 260.73; 95% CI [37.36–1,819.48]; p<0.001), absence of reciprocal changes on the ECG (OR 249.29; 95% CI [26.84–2,315.42]; p<0.001) and localised STE in leads V1–V3 (OR 101.43; 95% CI [13.46–764.16]; p<0.001).

We used the regression coefficients to develop a diagnostic score for predicting false STEMI activations, assigning 1 point for each of the following criteria: absence of typical chest pain, absence of reciprocal changes and localised STE in leads V1-V3. A total score of at least 2 points suggests a potentially false STEMI activation. The receiver operating characteristic (ROC) curves for the false STEMI risk score demonstrated an impressive area under the curve of 0.983, with a sensitivity of 96.5% and specificity of 95.7% (Figures 1 and 2 and Table 3).

Discussion

False STEMI activation places significant strain on the healthcare system by diverting valuable resources away from true STEMI cases, including medical staff, diagnostic tools and treatment facilities. This results in increased healthcare costs due to unnecessary procedures and consultations, while also contributing to physician fatigue and burnout. Additionally, it can erode patient trust in the healthcare system due to misdiagnoses and unnecessary interventions. Moreover, the diversion of resources and attention from actual STEMI patients may lead to delayed care, worse outcomes and increased morbidity and mortality, underscoring the need for improved diagnostic accuracy and activation protocols.

The prevalence of false STEMI activation in Thai patients from our single-centre cohort was 38.2%, which is relatively higher than that observed in other studies.^6,7,14 The results of this study indicate that first-year interns demonstrated a tendency towards higher rates of false STEMI activations, whereas internists (specialist doctors) were more likely to activate true STEMI diagnoses. This finding suggests that diagnosing acute STEMI may pose challenges for less experienced physicians. However, with increased clinical experience, their diagnostic accuracy is expected to improve, leading to more precise STEMI activations.^15,16

Previous studies showed that the use of a predictive score could improve the accuracy of classifying patients with true STEMI.^14,17 Most studies consider multiple factors, including age, chest pain characteristics, concave-morphology ST elevation and the absence of reciprocal changes. However, these factors can be relatively difficult to apply in real-world clinical practice due to their complexity and variability.

Common causes of false activations of acute STEMI include LVH, early repolarisation and LBBB.^18,19 The ECG changes in these conditions can lead to STE in leads V1–V3, which may mimic acute anterior STEMI. For less experienced physicians, distinguishing between these conditions and true STEMI can be challenging, particularly when no previous ECG is available for comparison.²⁰ However, it is generally recognised that STE in true anterior STEMI typically affects more leads than in these conditions and is often accompanied by reciprocal changes in the ECG. This distinction explains why localised STE within V1–V3, without reciprocal changes, is a key factor contributing to false STEMI activations in our predictive score.

Therefore, developing a simple diagnostic score to exclude false STEMI activation offers several benefits. It aids in distinguishing true STEMI cases from false alarms, enhancing diagnostic accuracy. By incorporating easily identifiable criteria including absence of typical chest pain, absence of reciprocal changes and localised STE within V1–V3, the false STEMI score is user-friendly and can be applied quickly in clinical settings. This score has the potential to support more accurate decision-making, particularly among junior doctors, by aiding in the early identification of false STEMI activations. While its use may help reduce unnecessary interventions and improve triage accuracy, its impact on healthcare costs and clinician workload requires further validation through prospective studies.

Limitations

This study has several limitations. First, the retrospective single-centre design and relatively small cohort size may limit the generalisability of our findings to broader populations. Second, the exclusion criteria applied during sample selection – particularly the omission of critically ill patients and those requiring immediate intervention – may have led to under-representation of true STEMI cases. Additionally, as the study focused on patients for whom STEMI activation was triggered, it did not capture potential false-negative cases where STEMI was missed or under-recognised. As such, our findings may be subject to selection bias. Furthermore, although the final diagnosis was made by experienced cardiologists, coronary angiography was not performed in all cases, which may have limited the consistency of definitive diagnostic confirmation. These limitations may affect the internal validity and should be considered when interpreting the diagnostic performance of the proposed score.

Conclusion

In this study, 38% of STEMI activations were found to be false positives, with LVH being a common source of misinterpretation. To address this diagnostic challenge, we developed the false STEMI score, a simple tool based on clinical and ECG parameters that showed high sensitivity and specificity for identifying activations that were likely to be false. Prospective, multicentre validation studies with broader patient inclusion criteria are essential to confirm the score’s clinical usefulness and its potential to optimise STEMI triage and reduce inappropriate activations.

Click here to view Supplementary Material.