Current issues of ACP Journal Club are published in Annals of Internal Medicine


Diagnosis

Review: Clinical features and electrocardiographic changes predict myocardial infarction in adults with chest pain

ACP J Club. 1999 Mar-April;130:44. doi:10.7326/ACPJC-1999-130-2-044


Source Citation

Panju AA, Hemmelgarn BR, Guyatt GH, Simel DL. Is this patient having a myocardial infarction? JAMA. 1998 Oct 14;280:1256-63.


Abstract

Question

What features of the clinical history and examination and electrocardiogram (ECG) can be used to determine whether a patient with chest pain has had a myocardial infarction (MI)?

Data sources

Studies were identified with MEDLINE (from 1980) by using the index terms medical history taking, physical examination, myocardial infarction, chest pain, reproducibility of results, and observer variation and the text words interobserver, intraobserver, accuracy, precision, reliability, sensitivity, specificity, myocardial infarction, and chest pain. Bibliographies of relevant studies were also scanned.

Study selection

For assessment of precision, studies that assessed interobserver or intraobserver variation or both were included. For assessment of accuracy, studies were included if patients had chest pain suggestive of ischemia; if the history, physical examination, or ECG was described in detail; if the diagnosis of MI was based on World Health Organization criteria (confirmation by ECG and laboratory test results); and if ≥ 200 patients were studied.

Data extraction

Data on study quality, observer variation, inclusion criteria, rates of MI, age and sex of patients, clinical features and changes in ECG, and likelihood ratios (LRs) were extracted.

Main results

11 clinical features and 10 changes in ECG predicted the presence of MI, and 4 clinical features helped to rule it out. The features with a +LR > 3.0 are shown in the Table. Other features are radiation of chest pain to the right shoulder (+LR 2.9, 95% CI 1.4 to 6.0) or left arm (+LR 2.3, CI 1.7 to 3.1), pain in the chest or left arm (+LR 2.7), pulmonary crackles on auscultation (+LR 2.1, CI 1.4 to 3.1), history of MI (+LR range 1.5 to 3.0), chest pain as the most important symptom (+LR 2.0), diaphoresis (+LR 2.0, CI 1.9 to 2.2), nausea and vomiting (+LR 1.9, CI 1.7 to 2.3), new T-wave inversion (+LR 2.4 to 2.8), and any conduction defect (+LR 2.7, CI 1.4 to 5.4). Factors that decreased the likelihood of MI were pleuritic chest pain (-LR 0.2, CI 0.2 to 0.3), sharp or stabbing chest pain (-LR 0.3, CI 0.2 to 0.5), positional chest pain (-LR 0.3, 0.2 to 0.4), and chest pain reproduced by palpation (-LR 0.2 to 0.4).

Conclusion

11 clinical features and 10 electrocardiographic changes predict MI; 4 clinical features help to rule out MI.

Source of funding: Not stated.

For correspondence: Dr. A.A. Panju, McMaster Building, Room 201, Henderson General Hospital, 711 Concession Street, Hamilton, Ontario L8V 1C3, Canada. FAX 905-389-9849.


Table. Positive likelihood ratios (LRs) for predicting myocardial infarction in patients with chest pain using clinical or electrocardiographic (ECG) features

Clinical or ECG features +LR* (95% CI) or range for heterogeneous studies
Radiation of pain to both arms 7.1 (3.6 to 14.2)
Third heart sound on auscultation 3.2 (1.6 to 6.5)
Hypotension (systolic blood pressure ≤ 80 mm Hg) 3.1 (1.8 to 5.2)
Any ST-segment elevation 11.2 (7.1 to 17.8)
New conduction defect 6.3 (2.5 to 15.7)
New ST-segment elevation ≥ 1 mm 5.7 to 53.9
New Q wave 5.3 to 24.8
Any Q wave 3.9 (2.7 to 5.7)
Any ST-segment depression 3.2 (2.5 to 4.1)
T-wave peaking, inversion ≥ 1 mm, or both 3.1 (data unavailable)
New ST-segment depression 3.0 to 5.2

*LRs defined in Glossary.


Commentary

This thoughtful review by Panju and colleagues identifies clinical features, including history, physical examination, and findings on ECG, that will help identify patients with MI at initial presentation. Despite the diversity of studies, the authors skillfully synthesized the available data into a format that readers will find useful. Given the heterogeneity of the studies, it is reassuring that the results generally agree with clinical teaching and with the findings of other studies used to formulate clinical prediction models (1, 2).

The diagnosis of MI in patients with typical clinical features and changes on ECG suggestive of MI is shown to be relatively straightforward. They often benefit from early thrombolysis and angioplasty.

The study also shows the difficulties of evaluating patients who present without characteristic findings on ECG. Because many patients with MI present without these findings, much time and expense is spent on "rule-out MI" protocols, either in the emergency department or on the ward. Computer-based decision models may be helpful in determining the likelihood of MI in these patients and thus may allow earlier discharge (1, 2). Despite much recent research, the utility of rapid troponin T and troponin I tests in the triage of these patients remains unclear (3). For now, physicians must rely heavily on their clinical judgment to determine which patients require hospitalization and which can safely be sent home.

Kenneth A. Ballew, MD MS
University of VirginiaCharlottesville, Virginia, USA


References

1. Tierney WM, Roth BJ, Psaty B, et al. Predictors of myocardial infarction in emergency room patients. Crit Care Med. 1985;13:526-31.

2. Goldman L, Cook EF, Brand DA, et al. A computer protocol to predict myocardial infarction in emergency department patients with chest pain. N Engl J Med. 1988;318:797-803.

3. Hamm CW, Goldmann BU, Heeschen C, et al. Emergency room triage of patients with acute chest pain by means of rapid testing for cardiac troponin T or troponin I. N Engl J Med. 1997;337:1648-53.