Chapter 21
Chest Pain

Christopher R. Carpenter¹ and Ali S. Raja²

¹ Department of Emergency Medicine, Washington University School of Medicine, St. Louis, MO, USA

² Department of Emergency Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA

Highlights

Atraumatic chest pain has a broad differential diagnosis ranging from acute coronary syndrome (ACS) to pericarditis, pericardial effusion, pneumothorax, and pulmonary embolism.
Multiple emergency medicine clinical practice guidelines exist for the diagnostic approach to acute chest pain and do not recommend routine stress testing or admission for low‐risk patients.
Pericardial tamponade more commonly presents with dyspnea than chest pain.
Pulsus paradoxus is the most accurate physical exam finding for pericardial tamponade.
Point‐of‐care ultrasound can be diagnostic for pericardial effusion, but confirming tamponade may require quantifying the duration of right atrial collapse.
The HEART Score and HEART Pathway are becoming the most commonly used risk stratification instruments for ACS prediction of major adverse cardiac event risk.
High‐sensitivity troponin predominates in Europe and is increasingly prevalent in North America with equivocal impact on emergency department (ED) operational flow.
Computed tomography (CT) coronary angiography is superior to stress echocardiography or myocardial scintigraphy for the assessment of obstructive coronary artery disease in emergency department settings.
Safely disseminating guideline recommendations, risk assessment scores, and cardiac imaging decisions across dissimilar emergency departments will require patient‐shared decision‐making and implementation science approaches.

Background

Chest pain accounted for 4.7% of emergency department (ED) visits in 2017, and is the second most common presenting complaint after abdominal pain at an estimated cost of $10 billion annually in the United States.^1,2 ED physicians prioritize the identification of potentially life‐threatening etiologies of chest pain, including acute coronary syndrome (ACS), aortic dissection, pneumothorax, and pericardial effusion, in addition to lesser immediate threats like pericarditis, pneumonia, and gastroesophageal reflux disease. This chapter will focus on the diagnostic approaches for pericardial effusion, pericarditis, and ACS.

The first step in the ED evaluation and management of patients who present with chest pain or other symptoms concerning for ACS is a 12‐lead electrocardiogram (ECG). While findings on the initial ECG may be diagnostic or suggestive of ACS, the initial ECG can also be normal or nondiagnostic, underscoring the need for additional monitoring and testing of patients with suspected ACS. Evaluation of the initial ECG in patients with suspected ACS should focus on the presence of ST‐segment elevations, new left bundle branch blocks, or new dynamic ST changes, all indicative of acute myocardial infarction (AMI). In patients with known left bundle branch blocks or paced rhythm, the Sgarbossa criteria (ST‐segment elevation of 1 mm or more concordant with the QRS complex; ST‐segment depression of 1 mm or more in lead V1, V2, or V3; and ST‐segment elevation of 5 mm or more discordant with the QRS complex) can still be used to diagnose ACS.³ A modification of Sgarbossa criteria that replaces the ST‐segment elevation of 5 mm or more with the ratio of ST‐segment elevation divided by S‐segment depth (ST/S) less than −0.25 may improve overall accuracy for ST‐segment elevation myocardial infarction (STEMI).⁴ More recently, additional modification of the Sgarbossa criteria called the BARCELONA algorithm further improves sensitivity without significantly reducing specificity. The BARCELONA algorithm is “positive” for ACS with ST deviation ≥ 1 mm concordant with QRS polarity in any lead, or ST deviation ≥1 mm discordant with the QRS in leads in which the largest deflection of the R or S wave ≤ 6 mm.⁵ However, while the standard ECG is the single best test to identify patients with AMI upon their presentation to the ED, it has relatively low sensitivity.⁶ In patients with AMI, ST segments may be elevated in only 50% of initial ECGs.⁷ In addition, most left bundle branch block is not ACS and ST‐elevation can be observed with myopericarditis, early repolarization, hyperkalemia, Takotsubo cardiomyopathy, and left ventricular hypertrophy or ventricular aneurysm, among other causes.^8,9 Due to these limitations, distinguishing AMI and unstable angina from other noncardiac chest pain in patients at risk for ACS typically involves serial ECGs and/or serial serum biomarkers of myocardial injury, also with diagnostic imaging [provocative stress tests or computed tomography (CT) imaging] or cardiac catheterization.^10,11

Clinical questions relevant to the assessment of a patient with acute undifferentiated chest pain involve the test characteristics of history and physical examination findings, cardiac biomarkers, point‐of‐care ultrasound, and cardiac CT imaging.

Clinical question

Can emergency medicine physicians accurately and reliably distinguish pericardial effusion and pericarditis from acute coronary syndrome?

The pericardial sac consists of the outermost parietal pericardium and the inner visceral pericardium and usually contains less than 30 mL of fluid. The pericardium is compliant so large amounts of fluid can accumulate slowly without compromising cardiac function, but rapid accumulation of pericardial fluid over minutes to hours may exceed pericardial compliance. The incidence of pericardial effusion or tamponade among ED patients with acute chest pain has not been described, but one small single‐center prospective study of dyspnea patients reported an incidence of 14%.¹² In one case series of Spanish patients with moderate or large pericardial effusion, the most common causative diagnoses were pericarditis (20%), malignancy (13%), myocardial infarction (8%), end‐stage renal disease (6%), congestive heart failure (5%), collagen vascular disease (5%), and tuberculosis or bacterial infection (4%).¹³ Hypotension is uncommon in subacute pericardial effusion, whereas pericardial tamponade is typically associated with cardiogenic shock. Assessing for the presence or absence of pericardial tamponade is not a clinical diagnosis because both sensitivity and specificity are suboptimal for findings such as tachycardia, jugular venous distension, diminished heart sounds, and pulsus paradoxus.¹⁴ Even in tamponade, 27% of patients may present with paradoxical hypertension.¹⁵

One diagnostic accuracy systematic review in 2007 synthesized the diagnostic accuracy of history, physical exam, chest X‐ray, and ECG for the diagnosis of tamponade.¹⁴ For most findings, only sensitivity was reported. The exception was pulsus paradoxus which was evaluated in a single study of 65 pericardial tamponade and 101 control patients with positive likelihood ratio 5.9 (95% CI 2.4–14) and negative likelihood ratio 0.03 (95% CI 0–0.21) for pulsus paradoxus >12 mm Hg.¹⁶ Seven different studies found lower sensitivity for pulsus paradoxus, ranging between 56% and 86%.¹⁴ Pulsus paradoxus is not a paradox.¹⁷ Instead, this finding is an exaggeration of the normal inspiratory decrease in blood pressure. Pulsus paradoxus can be evaluated by auscultation of Korotkoff sounds during slow release of a blood pressure cuff. Pulsus paradoxus can be masked by the presence of hypotension, aortic regurgitation, atrial septal defects, or right ventricular hypertrophy.⁶

The most common presenting symptom in pericardial tamponade is dyspnea, which is found in 87% of cases. The next most common symptom is fever (25%) then chest pain (20%).¹⁸ The sensitivity of other physical examination findings for the diagnosis of tamponade are summarized in Table 21.1. The most sensitive ECG finding for tamponade is low voltage with a pooled sensitivity of 42% (95% CI 32–53).¹⁴ The sensitivity of electrical alternans ranges from 16% to 21%.^19,20 The sensitivity of cardiomegaly on chest X‐ray ranges from 68% to 100% for the diagnosis of tamponade.¹⁴

Point‐of‐care bedside ultrasound is the diagnostic test of choice to identify pericardial effusion and features of tamponade. The subcostal window is the recommended window because that positions the most dependent portion of pericardium adjacent to the probe. A common error is confusing epicardial fatty tissue with an effusion, which can be avoided by noting the heterogeneous sonographic texture of fat compared with fluid, as well as the failure of fat to track around the heart.²¹ Right atrial collapse is a sensitive marker of tamponade but is not specific.¹⁷ One sonographic approach to distinguish clinically inconsequential right atrial collapse associated with pericardial effusion from right atrial collapse due to tamponade is to quantify the duration of collapse during the cardiac cycle. Right atrial collapse exceeding one‐third of the cardiac cycle increases accuracy from 87% (using dichotomous collapse or no collapse) to 97%.²² Quantifying the duration of right atrial collapse can be accomplished using M‐mode.²³

Table 21.1 Sensitivity of physical examination for the diagnosis of tamponade

Source: Data from [14].

Sign	Number of studies	Sensitivity %
Tachycardia	4	65–87
Hypotension	4	14–35
Hypertension (>140 mm)	1	33
Tachypnea	1	80
Diminished heart sounds	3	24–34
Elevated jugular venous pressure	4	53–88
Peripheral edema	2	21–28
Pericardial rub	2	19–29
Hepatomegaly	2	28–55

Pericarditis, or inflammation of the layers surrounding the heart, usually affects young and middle‐aged individuals, is idiopathic in 90% of cases, and is diagnosed in approximately 5% of North American ED chest pain patients.^24,25 Pericarditis may be attributable to infection (usually viral), autoimmune, neoplastic, or medications, but is most commonly associated with metabolic (uremia, myxedema) or traumatic causes when an etiology can be identified.²⁵ Classically, the presenting complaint is chest pain described as “sharp” and pleuritic with exacerbation while lying flat and improvement leaning forward. Other features that may distinguish pericarditis from ACS include more prolonged duration of pain in pericarditis and radiation of pain to the trapezius ridge.²⁶ ECG classically demonstrates diffuse ST‐segment (J‐point) elevation with PR segment depression in leads other than V1 and aVR. Another finding is a down‐sloping ECG baseline and is called Spodick’s sign. In one retrospective review, Spodick’s sign was noted in 29% of acute pericarditis patients but also in 5% of STEMI patients. The ECG findings that most strongly distinguished pericarditis from STEMI were more profound ST‐depression in lead III than in lead II as well as the absence of PR‐depression (both favoring STEMI).²⁷ Unfortunately, inter‐rater agreement for Spodick’s sign (kappa 0.4–0.5) and PR‐depression (kappa 0.5) is suboptimal.²⁷

Clinical question

Can the HEART score accurately and reliably distinguish low‐ and high‐risk chest pain patients?

To appreciate the potential value of the HEART score and other chest pain decision‐aids in ED settings, understanding the limitations of isolated findings from the history and physical exam for the accurate diagnosis of ACS is essential. A large meta‐analysis by Panju et al. reviewed examination findings of patients with suspected AMI and was subsequently updated in 2008.⁶ Fourteen studies met their inclusion criteria, most of which used the World Health Organization definition of AMI as the criterion standard, which includes evolving changes on serial ECGs, a rise in serial biomarkers, and either chest pain with an abnormal ECG or other symptoms with evolving changes on serial ECGs. The analysis of the pooled data in the meta‐analysis (Table 21.2) demonstrates a number of historical and physical exam characteristics that increase and decrease the likelihood of AMI. For the sake of clarity, the authors chose to only report LRs that were either greater than 2.0 or less than 0.5.

Table 21.2 History and physical examination characteristics associated with either a greater or lesser risk of acute myocardial infarction

Source: Data from [6].

Clinical feature	Positive likelihood ratio (LR+) (CI)	Negative likelihood ratio (LR−) (CI)
Chest or center arm pain^*	2.7	—
Chest pain radiation to:
Right shoulder	2.2 (1.4–3.4)	—
Right arm	7.3 (3.9–14)	—
Left arm	2.2 (1.6–3.1)	—
Both left and right arms	9.7 (4.6–20)	—
Chest pain most important symptom^*	2.0	—
History of myocardial infarction^**	1.5–3.0	—
Nausea or vomiting	1.9 (1.7–2.3)	—
Diaphoresis	2.0 (1.9–2.2)	—
Third heart sound on auscultation	3.2 (1.6–6.5)	—
Hypotension^***	3.1 (1.8–5.2)	—
Pulmonary crackles on auscultation	2.1 (1.4–3.1)	—
Pleuritic chest pain	—	0.2 (0.2–0.3)
Sharp or stabbing chest pain	—	0.3 (0.2–0.5)
Positional chest pain	—	0.3 (0.2–0.4)
Chest pain reproducible by palpation^*	^*—	0.2–0.4

^* Data not available to calculate confidence intervals.

^** Reported as a range due to heterogeneity of the pooled studies.

^*** Defined as systolic blood pressure ≤80 mm Hg.

Numerous decision aids have emerged over the last 40 years to risk stratify ACS among ED chest pain patients. Since the Second Edition of this textbook was published, the History/ECG/Age/Risk factors/Troponin (HEART) score (Figure 21.1) and HEART Pathway (Figure 21.2) have emerged as the most frequently researched and emergency medicine guideline‐endorsed chest pain decision aid.^10,11 The American College of Emergency Physicians (ACEP) 2018 Clinical Policy on suspected non‐ST‐elevation ACSs gave a Level B recommendation to use HEART score ≤ 3 as a predictor of very low 30‐day major adverse cardiac events. In contrast, that same Clinical Policy provided a lower Level C recommendation for the use of the thrombolysis in myocardial infarction (TIMI) score for the same purpose.¹⁰ The ACEP Clinical Policy notes that the derivation and validation of the HEART Score in ED settings as opposed to the inpatient settings used in the older TIMI score are a strength of the HEART score. The Society for Academic Emergency Medicine (SAEM) Guidelines for Reasonable and Appropriate Care in the Emergency Department (GRACE) published recommendations for the management of recurrent low‐risk chest pain and defined “low‐risk” primarily by HEART score <4.¹¹