The skilled emergency clinician will not dismiss the complaint of chest pain because of the impression that it is rarely serious. While it is true that most children complaining of chest pain do not have a life-threatening condition, there are notable exceptions. The mandate of the emergency physician is to identify the child with a life-threatening condition in a sea of benignity. The purpose of this article, therefore, is to present an overview of the many causes of chest pain in children which should be considered, even if most can be excluded through a careful history and physical exam. What remains undiagnosed cannot be treated. While this review will focus and place disproportionate emphasis on serious causes of pediatric chest pain, the low likelihood of these diseases should be kept in context. Reassurance is important after the evaluation, and close attention should be paid to what are frequently subtle clues in the history, on the physical exam, and in the few diagnostic tests likely to be ordered in the emergency department. As with many conditions presenting to the emergency department, even if the likelihood of serious disease may be small, the implication of misdiagnosis is extremely high (e.g., aortic dissection, subarachnoid hemorrhage, meningococcemia).
|ACS:||acute chest syndrome|
|CHF:||congestive heart failure|
|LVH:||left ventricular hypertrophy|
|MVP:||mitral valve prolapse|
|NSAIDs:||nonsteroidal anti-inflammatory agents|
Sudden death is the dreaded and very public consequence of minimizing chest pain in children.1 As many as 50% of young victims of sudden death have symptoms suggestive of the underlying pathology prior to cardiac arrest.2 The rate of sudden death in individuals from 1-30 years of age ranges from 1.3 to 8.5 per 100,000 patient-years.2 From a quarter to two-thirds of these deaths are attributable to cardiac causes, including myocarditis, hypertrophic cardiomyopathy, coronary artery disease, congenital coronary artery anomalies, aortic lesions, and abnormalities of the conduction system.2-4 In athletes, hypertrophic cardiomyopathy and congenital coronary artery anomalies clearly predominate and are responsible for up to 49% of deaths.5 The highest risk of cardiac sudden death is in the teenage years and a family history of sudden death is not uncommon.2-3 Children and adolescents who have chest pain or syncope that is exertional and cannot be clearly attributed to a noncardiac cause should be taken very seriously, particularly if there is a family history of sudden death.
Shortly after midnight, just as you are making headway with the bolus of patients that the bus dropped off, you pick up one of your last charts for the evening. With relief, you note that the chief complaint is chest pain; your experience tells you that chest pain is almost never serious. But when you go in to see him, you note that despite a completely normal examination, he appears to be uncomfortable. He tells you that he was fine until he awoke shortly after going to bed with "severe" substernal chest discomfort that is slightly pleuritic in nature. He believably denies drug use and has no cardiopulmonary risk factors. You mark ECG and CXR on the order sheet and move on, reasonably sure he will have a pneumothorax on his chest film. Soon afterwards, the tech hands you an ECG that the computer interprets as "acute myocardial infarction." Indeed, the ECG shows focal ST segment elevation in the anteroapical leads with associated T wave inversions.211
Unlike the vast and monumental amount of work that has been done evaluating chest pain in the general adult population, there are few prospective studies of pediatric chest pain in the emergency (or primary care setting) and these have a relatively small numbers of patients.6-13,212 Many, if not most, of the final diagnoses represent clinical impressions rather than confirmed diagnoses. The result is that a majority of these children are believed to be suffering from musculoskeletal, functional, or idiopathic chest pain.7,10-12,14-16
This is understandable given that the above causes of chest pain are undoubtedly common and impossible to prove definitively. Results of such studies should be approached with caution, however, as none take a systematic approach in arriving at a diagnosis. Published series rely on the discharge diagnosis, which may be incorrect. Furthermore, few studies involve any follow-up past the primary visit. In one of the few studies that contained follow-up information, 9 out of 10 patients who sought further care received a different diagnosis from the followup clinician.17
These studies serve to emphasize the benign nature of this complaint. In contrast is the serious repercussion of missing subtle but potentially serious pathology. Sudden death in young athletes is catastrophic and highly visible, and media reports provide a sense that these events are common and preventable.1 The relatively prolific literature on adult and pediatric sudden death serves to reinforce such "exposure bias." Liberthson reminds us, however, to keep the risk of sudden death in perspective.2 There are only 10-25 sports-related sudden deaths from cardiac causes in the United States each year, and only 5 in 100,000 young athletes have a predisposing condition.
Finally, an inherent limitation to the literature on pediatric chest pain is that which is common to many pediatric conditions: extrapolation from the adult literature is a common theme. Many causes of chest pain that are common in adults are rare in children. Collecting large series or systematically assessing treatment regimens in children is challenging. Therefore, applying diagnostic and therapeutic interventions is initially more of an anecdotal than systematic endeavor. Some investigators have overcome this by creating national registries, such as the Canadian registry of venous thromboembolism, but these are the exception rather than the rule.18
Although comprising less than 1% of all pediatric ED visits, chest pain is a common complaint in children and is the second most common reason for referral to a pediatric cardiologist.7,10,13,19 Children are less likely to seek urgent medical attention for chest pain than adults, which is likely due to the belief that the condition is usually not serious. Prospective series and reviews tend to emphasis the benign nature of this complaint, with diagnoses such as costochondritis, idiopathic chest pain, and other chest wall syndromes being the most common etiologies listed.7,10-12,14-15,20 However, chest pain may originate from many different organ systems, including cardiac, pulmonary, gastrointestinal, musculoskeletal, and the central nervous system (psychosomatic). 7,10-12,15,21
Only a minority (4-15%) of patients will be found to have a cardiac etiology, but fear of underlying serious cardiac disease remains high.7,10-14,212 In one study, 70% of patients believed that adolescents could have heart attacks and 44% were concerned that this was their problem.15 In another study of adolescents with chest pain, 68% were concerned that their chest pain was cardiac and 44% had altered their behavior because of it.8 The serious attention paid to their complaint by physicians may reinforce this impression, even after a normal evaluation.8,15,22 These concerns are only amplified by sensationalistic media coverage of rare instances of sudden death in otherwise healthy adolescents.1
The anatomy of the chest wall: there are twelve ribs (seven of which are considered "true ribs");23 their cartilages attach to the sternum via chondrosternal synovial joints.24 The medial cartilaginous ends of ribs eight, nine, and ten articulate with the cartilage of the superior rib. Ribs eleven and twelve are "floating," without medial articulations. The first rib and xyphoid articulate with the manubrium as synchondroses, while the sternoclavicular joint is also contained in a synovial membrane.24 Multiple intercostal muscles underlie such larger muscular bodies as the pectoralis and serratus anterior.25
The chest wall is supplied in a linear dermatomal dermatomal fashion by the intercostal nerves. However, the nerve supply of many internal structures embryologically derive from the cervical region.24 The diaphragm and pericardium are innervated by the phrenic nerve which exits the spinal cord around C4, as do the nerves supplying the musculature around the shoulder regions, hence the variety of referred pain syndromes and difficulty isolating the source of pain.14 Cardiac pain, for example, is produced by excitation of a number of different pathways, including cells of the spinothalamic tract along the lower cervical and upper thoracic portions (but not at C7 or C8); this excitation of the upper cervical pathways is why patients with true cardiac pain (such as angina pectoris) often experience pain in the jaw and neck in addition to the chest.
As noted, one the foremost concerns of patients presenting with chest pain is whether they (or their children) are having a heart attack. While this is extremely uncommon, there are several conditions that can cause anginal chest pain (or myocardial infarction [MI]) in children, see Table 1. Infarction is extremely rare in adolescents but may occasionally occur.26-29 Children who present with typical symptoms, including exertional substernal chest pain, diaphoresis, radiation into the left arm, nausea and shortness of breath, should be aggressively investigated.
Obstruction of the left ventricular outflow tract, as with subaortic stenosis caused by hypertrophic cardiomyopathy (or aortic valve stenosis), can result in exertional chest pain. Previously referred to as idiopathic hypertrophic subaortic stenosis (IHSS), the term hypertrophic cardiomyopathy best describes a
spectrum of disease that is not always associated with dynamic left ventricular outflow tract obstruction.30-31 While left ventricular hypertrophy (LVH) may be symmetric (concentric), it is usually asymmetric, affecting the anterior septum or apex.31-32 Children younger than age 13 may show no clinical or diagnostic evidence of HCM, but the rapid growth associated with puberty may result in significant left ventricle (LV) remodeling with or without the onset of clinical symptoms.32
When symptomatic, the most common symptoms of HCM are exertional dyspnea from pulmonary venous congestion, anginal chest pain, nearsyncope, and palpitations.33 The etiology of the chest pain may be multifactorial. Exertion typically results in a fall in systemic vascular resistance, and the dynamic obstruction causes a fall in cardiac output with decreased coronary perfusion and ensuing ischemic pain. Alternatively, the coronary arteries are frequently abnormal in HCM. The result mimics small vessel disease, with a mismatch between perfusion and demand, resulting in myocardial ischemia.
Associated syncope should be considered an ominous finding and may represent critical ischemia or a potentially lethal arrhythmia. Intrinsically disorganized cellular architecture and fibrosis from ongoing, intermittent ischemia produces an electrically unstable substrate that results in life-threatening ventricular arrhythmias.31,34 Sudden death may be the catastrophic first indication of disease, and HCM is likely to be found in victims of sudden death involved in exertional activities.33,35
A harsh crescendo-decrescendo systolic ejection murmur from the outflow tract gradient may be present. The murmur is increased by maneuvers that decrease left ventricular filling or decrease afterload (e.g., standing or release of Valsalva, which increases dynamic outflow tract obstruction).32 The murmur of hypertrophic cardiomyopathy was distinguished from all other murmurs by an increase in intensity with the Valsalva maneuver (65% sensitivity, 96% specificity) and during squatting-to-standing action (95% sensitivity, 84% specificity) and by a decrease in intensity during standing-to-squatting action (95% sensitivity, 85% specificity), passive leg elevation (85% sensitivity, 91% specificity), and handgrip (85% sensitivity, 75% specificity).36
Likewise, maneuvers that increase left ventricular filling or decrease afterload will cause a decrease in the intensity of the murmur, e.g., squatting or passive leg raise. However, the exam is unreliable for the routine identification of asymptomatic HCM.31
While the electrocardiogram (ECG) may be normal, the majority of children have nonspecific abnormalities.32 Particular attention should be paid to left ventricular forces. Lateral ST-T wave flattening or inversion may also be present.2,32 In symptomatic patients, echocardiogram (ECHO) should be diagnostic, but this study may be normal prior to puberty.2,31
Identification of patients with HCM at high risk for sudden death is critical, as it may be sudden and unpredictable. Important risk factors include: history of cardiac arrest or ventricular tachycardia, family history of sudden death, and syncope/near-syncope (especially exertional).31 Ventricular tachycardia on Holter monitoring, hypotension with exercise testing, and extreme LVH noted on ECHO are significant findings.31
Interestingly, while there is no clear linkage between sudden death and outflow obstruction, myocardial bridging (when present) conveys a high risk of cardiac arrest.31,37 Myocardial bridging describes the tunneling of a segment of a major (normally epicardial) coronary artery within the myocardium.38-40,210 While systolic compression has little effect on diastolic coronary blood flow, patients may demonstrate prolonged compression well into diastole.38-40 Yetman et al found that myocardial bridging was common in children with HCM, and they demonstrated a much greater incidence of chest pain, ventricular tachycardia, and cardiac arrest.37
While treatment (prevention of sudden death) has traditionally consisted of beta-blockers and a variety of other agents to control dysrhythmias, little data exists to support this approach.31 Currently, implantable cardioverter-defibrillators appear to be the most effective means of preventing sudden death.31 In addition, a national consensus panel has recommended limiting participation in competitive sports and strenuous physical exertion.41
Another cause of exertional angina and sudden death in adolescents is congenital coronary artery abnormalities. Symptoms may be present in as many as 40% of patients before sudden death. Common congenital lesions include: anomalous origin of the left main coronary artery (ALMCA), anomalous origin of the right main coronary artery (ARCA), coronary ostial stenosis, and single coronary ostium.34,42-43 The incidence of anomalous origin of a main coronary artery is 0.17-1.2% on autopsy and catheterization studies.43 Anomalous origin of the left coronary artery may arise from the right sinus of Valsalva or pulmonary artery and the anomalous right coronary artery from the left sinus of Valsalva.26,43-44 Either may course between the aorta and pulmonary artery before branching.2,42-43,45
While most children remain asymptomatic, as many as 18-30% will have a prior complaint of cardiovascular symptoms (such as exertional chest pain, dyspnea, or syncope).2,31,42,46 In infants, anginal symptoms may manifest as crying spells easily confused with colic, while coincident pulmonary congestion may masquerade as bronchiolitis.11,44 Ventricular ectopy and other arrhythmias may be associated with this condition, and electrocardiography may show evidence of ischemia.42-43 Sudden death is likely the result of ischemia and infarction or fatal arrhythmia, as suggested by the results of stress testing and autopsy studies.42-43
Possible mechanisms of ischemia include compression of the anomalous artery between the aorta and pulmonary arteries, kinking or angulation of the vessel, or an acute angle of origin with a slit-like opening and diminished blood flow.2,36,42,47-48 Sudden death appears to be most frequently found when the vessel travels between the two great arteries, particularly when the vessel is dominant and perfusing a large area of myocardium.2-3,45,49 Patients with anomalous left coronary arteries seem to be at greatest risk of sudden death.36,45 Treatment with coronary artery bypass or coronary artery reimplantation appears to be effective in relieving the ischemia, although the long term patency of the grafts is unknown.36,43
Mitral valve prolapse and its association with chest pain has been well documented.50 Physiologically, the theory is that papillary muscle or left ventricular endocardial ischemia causes the pain.50 Diagnosis of mitral valve prolapse is suggested by the presence of a mid-to-late systolic click. A short systolic ejection murmur may be produced from mitral insufficiency.51 The intensity is increased by reducing left ventricular volume (standing, forced expiration, or Valsalva).
MVP as an etiology of chest pain has never been proven, however, and remains controversial.12,50,52-53 Similar to many other dubious relationships "uncovered" over the years, the relative frequency of both mitral valve prolapse and chest pain in the population make a clear association difficult to prove.12,15,54 In light of this controversy, one small study looked at 17 children with MVP and precordial chest pain. Fourteen had at least one abnormality on a comprehensive gastrointestinal evaluation, including esophageal manometry, Bernstein test, pH probe, and endoscopy. All but 1 improved with appropriate treatment directed towards the gastrointestinal abnormality.16
Patients with Kawasaki disease initially present with myocarditis and pericardial inflammation in addition to the textbook constellation of symptoms. Coronary aneurysms form early, leading to coronary thrombosis and myocardial infarction as both a short and long term concern.55-57 Untreated patients with Kawasaki disease have a 13-40% incidence of coronary artery aneurysms, depending on patient age and time since illness.56-58 Coronary arteries remodel with myointimal proliferation, developing segmental stenosis or dilatation at the site of an aneurysm and become a focus for accelerated atherosclerosis or thrombosis.56-57 The presence of giant coronary aneurysms (greater than 8 mm) is usually visible on echocardiography; they form at proximal branch points of both coronary arteries.57 Over time, these aneurysms (which are associated with the risk of myocardial infarction) persist.55,58 Since the 1980s, gamma globulin has been shown to be effective in significantly reducing aneurysm formation.56,58-59
Kato and Kawasaki examined a series of 195 patients with myocardial infarction complicating a diagnosis of Kawasaki disease.28 Children tended to have their myocardial infarction within the first year of the illness. Mortality was highest in children with obstruction of the left main coronary artery or the right coronary and left anterior descending arteries. While only a third of children described chest pain, this symptom was more common in survivors andolder children; the rest either had abdominal pain and vomiting or were asymptomatic.28
Burns et al studied the worldwide literature and identified 74 adolescent and adult patients with a history of Kawasaki disease presenting with cardiac sequelae.60 More than half with acute infarction presented with chest pain, while 10% presented with cardiogenic pulmonary edema or arrhythmias. Sixteen percent experienced sudden death. Almost 90% had a suggestive ECG and 24% had cardiomegaly on their chest x-ray (CXR). Interestingly, 82% were engaged in vigorous exercise at the time of their event. Aneurysms were described in more than 90% and included all three coronary arteries. A remarkable 42% had disease of their left main coronary artery, an extremely high risk location.28,60 Of those who died, all 18 had aneurysms.
Cocaine as a cause of chest pain and myocardial infarction of the young became widely recognized (and studied) during the cocaine epidemic of the 1980s. It typically affects otherwise young, healthy individuals without existing coronary artery disease. Intense coronary vasoconstriction can lead to intense vasospasm, coronary thrombosis, and cardiac ischemia, causing myocardial infarction and inducing life-threatening arrhythmias.61-63 Even in the absence of other cardiac risk factors, recurrent cocaine use can induce accelerated atherosclerosis and unstable plaques that are prone to rupture and acute thrombosis.61-63 It should be unnecessary to point out that adolescents may not provide a reliable history of drug use.
Accelerated atherosclerosis should only be seen in patients with a strong family history of early onset coronary artery disease. It is interesting to note that, in an autopsy study of 760 victims of accidents aged 15-34, advanced atheromas were seen in 2% of males aged 15-19 and 20% in ages 30-34 (women were 0 and 8%).64 These lesions represent the presence of a lipid laden plaque of necrotic debris vulnerable to rupture and sudden thrombosis. While atherosclerosis is uncommon in adolescents in the absence of familial hypercholesterolemia, even a 2% incidence of significant lesions in 15- to 19-year-old males represents a large group when spread across the entire population.64 Case reports document the lethal existence of accelerated coronary artery disease in adolescents with long standing diabetes.27 These reports serve as a reminder that adolescents with typical adult risk factors should be approached seriously.
Coronary vasospasm in the absence of known risk factors (e.g., drugs) has been reported.65-66 Substernal chest pain/pressure and diaphoresis with ischemic findings on ECG, relieved by nitroglycerine, was found in these cases. Cardiac catheterization demonstrated normal (clean) coronary arteries. Coronary vasospasm was demonstrated angiographically in one case with an acetylcholine challenge test.6
Chest pain from pericarditis classically localizes to the central chest, may be sharp or dull in quality, is reduced by leaning forward, and is increased in the supine position. It is often, but not always, pleuritic (respirophasic).32,67 When severe, it too may mimic acute MI, with severe substernal discomfort and the same patterns of radiation.67 Like myocarditis, there may be a viral prodrome with fever and myalgias.
In its more fulminant state (with cardiac tamponade) acute pericarditis may present with distant heart sounds, evidence of elevated systemic venous pressures, hypotension, and pulsus paradoxus.67,213 More commonly, a scratchy friction rub (with both a diastolic and systolic component) may be heard at the left lateral sternal base.68 A friction rub is virtually pathognomonic for pericarditis when present, but it may come and go and be missed.69 Having the patient lean forward or breathe in deeply may accentuate the rub by bringing the heart close to the chest wall.32 This may disappear as an effusion develops.
Typical ECG findings in acute pericarditis usually evolve through four stages, see Table 2.67 Patients may appear to skip or pass through stages rapidly.67,70 P-R depression, however, is found in the majority of patients at some point in the illness.67,70 These P-R shifts likely represent sub-epicardial atrial injury in the same way ST-segment changes represent sub-epicardial ventricular injury. Early ECG findings in pericarditis may be difficult to distinguish from AMI; clues are that ST-segment elevations tend to be concave up, resembling benign early repolarization, and there are no associated reciprocal T wave changes.67 If a large pericardial effusion is present, low voltage or electrical alternans may be noted, see Figure 2.
Cardiac enzymes may be elevated, depending on the degree of associated myocarditis. This may further confound the picture if ECG findings are asymmetric and suggest possible infarction. CXR may demonstrate an enlarged cardiac silhouette if the effusion is large but is otherwise normal. An ECHO is the gold standard to assess for the presence of pericardial effusion and associated myocardial dysfunction.32 Hemodynamically unstable pericardial tamponade should be a clinical diagnosis, and lifesaving pericardiocentesis should not depend on obtaining a diagnostic ECHO!
Management of idiopathic pericarditis (in the absence of tamponade) consists of NSAIDs (ibuprofen, aspirin) for analgesia. If there is a doubt that the presentation does not represent uncomplicated pericarditis (with no evidence of pericardial effusion) the patient should probably be hospitalized until the diagnosis can be confirmed and other serious diagnoses excluded.67 (While the prevalence of purulent pericarditis is rare in the modern era, it should not be missed as the management is obviously very different and beyond the scope of this article.)71 Therapy should continue for 1-2 weeks and taper over the same time period.67 Colchicine is sometimes added in refractory cases, while corticosteroids should be avoided. The majority of patients will improve within 3-4 weeks.32
Probably more common than is realized, myocarditis may be responsible for up to 8-44% of pediatric sudden deaths.2,13,72 Myocarditis generally follows a viral illness (not always recognized) and is most commonly caused by enterovirus, influenza, and other viruses.43,73 Enteroviruses, Coxsackie B virus, specifically, have been the most common viral etiology implicated, but the predominance of this microbial agent has recently been questioned.72-74 The role of autoantibodies (and molecular mimicry) is thought to be pivotal in many, if not most, patients.73,75-76 In endemic areas, Lyme disease is also a consideration.210 Other causes of myocarditis include HIV, diphtheria, drugs (e.g., chemotherapy, cocaine), connective tissue diseases, smallpox vaccination, and acute Kawasaki disease.72,75-76 Worldwide, Chagas disease, caused by Trypanosoma cruzi, is the most common cause of myocarditis and is endemic in rural Central and South America.76
Symptoms may be subtle or absent, but (when present) the most common symptoms of the patient with myocarditis are ventricular dilatation with congestive heart failure (CHF) and arrhythmias/palpitations. 73,76-78 In a young, healthy patient with no history of heart disease who presents with symptoms of CHF, with or without chest pain, myocarditis should be strongly considered.79 Typical CHF symptoms in these patients will include dyspnea on exertion, progressing to dyspnea at rest and orthopnea. Some patients may complain of RUQ pain from an enlarged liver. Others may complain of myalgias and a flu-like illness, particularly with an aggressive, fulminant variation.72 Infants will often present with respiratory distress that is easily confused with common respiratory ailments.68 Cardiopulmonary auscultation may reveal an S3 gallop and rales.32 A friction rub may be heard if an associated pericarditis is present. Wheezing may be cardiac in origin rather than asthma; consider the remainder of the history and physical to assess the etiology of the wheeze.
ECG abnormalities in myocarditis tend to be diffuse and nonspecific.80-81 The most common finding is sinus tachycardia.80 The ECG typically demonstrates ST-segment depression or elevation and T wave abnormalities, particularly in the inferior leads (II, III, aVF).72,81-82 Other findings include low voltage, prolonged QT interval, and atrioventricular block.32 A host of atrial and ventricular dysrhythmias are possible and may be associated with sudden death.81
Echocardiography typically reveals signs of cardiomyopathy and either a dilated or normal sized myocardium. Ventricular function may be so profoundly depressed as to be almost immeasurable.81 Significant valvular insufficiency is often noted as well. Typical laboratory findings include an elevated sedimentation rate, leukocytosis, eosinophilia, and cardiac enzyme abnormalities.76 Definitive diagnosis is made by endocardial biopsy.76
Patients may also present with findings more typical for acute MI, such as substernal chest pain, suggestive ECG changes, and elevated cardiac enzymes.82-85 In general, ECG changes are more global than focal, and ECG changes do not have the same progression as in acute MI. Myocarditis/pericarditis should be suspected in the young patient with the clinical picture of MI who otherwise has no risk factors (including cocaine use).72 Coronary angiography, when performed, should be normal.
While aortic dissection may be extremely uncommon in pediatric patients, it can strike those with connective tissue disorders such as Marfan's syndrome and Ehlers-Danlos syndrome.2,86-88 These disorders may cause aortic-root dilatation leading to aortic dissection. Other etiologies in the young include cystic medial necrosis, pregnancy, and aortitis from rheumatologic diseases.87,89 Using the DeBakey classification system, aortic dissection is generally classified as: type I, in which the primary tear is in the ascending aorta and extends into the arch; type II, in which the lesion is confined to the ascending aorta; and type III, in which the dissection begins in the descending thoracic aorta.90-91 The Stanford classification divides these lesions into only two categories: type I, involving the ascending aorta, and type II, originating in the descending aorta.
Interestingly, patients with Marfan's syndrome may be more likely to die of a lethal arrhythmia than aortic dissection.92 A structural fibrillin defect may result in an electrically unstable myocardium, allowing the development of ventricular tachycardia or fibrillation. Part of the explanation for this is that anatomic abnormalities of the aorta are followed closely and treated aggressively while there has been less awareness of the risk of sudden death in Marfan's due to ventricular arrhythmias.
While the clinical presentation of aortic dissection is generally striking, a high index of suspicion needs to be maintained in pediatric patients due to the extreme rarity of the disease.86 Even in adults, the initial clinical impression is often not aortic dissection.91 Dissections originating in the proximal aorticusually present with sudden, severe, sharp anterior chest pain while radiation to the back suggests more distal extension.91,93-94 Pain may migrate with extension of the dissection, and radiation into the jaw and neck is not uncommon.91 Associated physical findings may include hypotension, congestive heart failure, pulse deficits in the arms or legs, neurologic abnormalities (from vascular obstruction), or an aortic regurgitant murmur, all depending on the vessels involved. The physical exam may be entirely normal.91,93-94
Patients with preexisting aneurysmal dilatation may have a widened aortic diameter on CXR, but an aortic dissection does not always involve an aneurysm so mediastinal widening is not as common.91,93 Occasionally, the electrocardiogram will show evidence of acute myocardial infarction if the coronary arteries are involved.91,93 Diagnosis is typically made by computed tomography (CT) of the chest with intravenous contrast, although the use of transesophageal echocardiography or magnetic resonance imaging may have a role in select circumstances.88,93
Younger children may describe palpitations as chest pain, even when actual pain is not present.11,13,53 Regardless, any patient with chest pain and dizziness/syncope requires a cardiac evaluation. The ECG should be examined for any degree of heart block, QT prolongation, or evidence of accessory pathways.
While the incidence of pulmonary embolism is uncommon in pediatric patients, identification and anticoagulation of these children is vital. Most (but not all) children and adolescents with PE have identifiable risk factors, including the presence of a central venous catheter, malignancy, trauma, immobilization, and known hypercoagulable state.49,95 Additional risk factors in adolescent females include pregnancy/abortion and oral contraception use.96 The presence of a central venous catheter is clearly the most important factor associated with pediatric PE, with a mortality of close to 4%.97 In addition, an inherited thrombophilia (particularly factor V Leiden) is commonly identified in children and adolescents with pulmonary thromboembolism.98-99
Children with pulmonary embolism have clinical presentations similar to adults. Signs and symptoms include sudden onset pleuritic chest pain and shortness of breath; they may also be tachycardic and hypoxic. However, children have greater cardiopulmonary reserve and their presentation may be much more subtle.96 Pulmonary embolism should be considered in any child with chest pain or shortness of breath with risk factors or adolescents with sudden onset chest pain, even with a normal exam.95
Chest radiography, electrocardiograms, and arterial blood gases have little role in the definitive diagnosis of PE but are valuable to exclude alternative diagnoses.95,100-102 Diagnosis has traditionally been confirmed in pediatric patients by a positive (high probability) ventilation perfusion scan. In older adolescents and adults, the use of computerized tomography is becoming the standard approach.103-105 The d-dimer assay is a very sensitive test with a high negative predictive value in low risk populations and can be judiciously applied (as a screening test) to low risk patients to reliably exclude thromboembolism, obviating the need for further diagnostic study.106
Many patients with thromboembolism do not manifest life-threatening physiologic abnormalities.96 One of the main reasons to identify PE is to prevent recurrent disease, which may be more severe.18 When recognized, age appropriate heparinization is warranted; small series using thrombolytics have been reported in critically ill children.107-109
Primary spontaneous pneumothorax is quite rare in young children and becomes manifest in adolescents and young adults.110-112 Primary pneumothorax generally occurs in slender young males who are a bit taller than average, usually from rupture of a congenital bleb.113 According to Laplace's law, the tension placed on the wall of a curved surface increases proportionally to the diameter of the surface. These blebs are much larger than their alveolar counterparts and are ripe for rupture.
Classically, patients present with chest pain, nonproductive cough, and, at times, dyspnea.111,113 Onset is sudden and unilateral. Initially pleuritic, it may evolve into a dull ache or even disappear completely.111-112 Decreased breath sounds on the affected side and hyperresonance to percussion may be noted on exam.113 However, in a series of 1200 cases of pneumothorax, 30% had minor, nonspecific symptoms and the absence of any clinical signs; patients may wait days before presentation.111,113
Diagnosis is made on plain chest x-ray, but remember that a small pneumothorax is easily missed by the casual observer. If strongly suspected and not seen, an expiratory film or chest CT may reveal the pneumothorax.111,113 A recent report using CT as the criterion standard reported a 98% sensitivity using bedside ultrasonography to diagnose traumatic pneumothorax compared with 76% sensitivity with conventional radiography.114
A pneumothorax occupying less than 15-20% of the pleural space in the absence of shortness of breath or worsening symptoms can (and should) be managed without a chest tube.110-113 In patients managed without tube thoracotomy, oxygen significantly increases the rate of resorption of the intrapleural air.111 If tube thoracotomy is necessary, a small caliber chest tube or pigtail catheter may be used rather than the standard (large) chest tube.111,214
Tension pneumothorax is a very uncommon presentation of primary spontaneous pneumothorax; in one adult series, it constituted just 1% of 723 patients studied.113 However, when present, it represents the most serious of time-dependent medical emergencies. Where air enters the pleural space, a ball-valve effect prevents air from escaping the pleural cavity. The progressive accumulation of air in the chest cavity compresses venous return, critically impairing cardiac output.
Patients with tension pneumothorax are extremely ill appearing. Vital signs are invariably deranged with tachycardia, tachypnea/respiratory distress/failure, hypoxia, and hypotension. Physical exam findings may include tracheal deviation, displaced or distant heart sounds, jugular venous distention, and chest wall asymmetry. Diagnosis is confirmed by plain chest x-ray, but definitive treatment (needle thoracotomy in the second intercostal space) should never await the results of the chest radiograph if the diagnosis is suspected and the patient is in extremis. If confirmed, place a chest tube. If a pneumothorax is not confirmed, remove the thoracotomy needle and observe the patient closely for development of an iatrogenic pneumothorax.
Spontaneous pneumomediastinum tends to occur in adolescents and young adults.115 Instead of a subpleural bleb rupturing into the pleural space, air ruptures into the interstitium and dissects back through the hilum. Air can then spread through the soft tissues of the mediastinum, into the neck and subcutaneous tissues, and, at times, into the pericardium.
A common precipitant in children is an acute asthma attack in which increased alveolar pressure results in alveolar rupture.116-117 Another study identified illicit inhalational drug use as a common precipitant. 115 It was postulated that pneumomediastinum results from intense Valsalva activity.115
Patients typically complain of either dyspnea or chest pain.115 They may also complain of neck discomfort (neck pain, sore throat, dysphagia).115-117 Either subcutaneous emphysema or Hamman's crunch is found in the majority of patients.115-116 Hamman's crunch is crepitation on auscultation that varies with the heartbeat and is virtually pathognomonic. These findings may be subtle and missed on the initial exam.
Patients are generally admitted for fear that the pneumomediastinum will extend into the pleural or pericardial space or expand, creating upper airway or vascular compromise.116 This appears to be unlikely, and one group even suggests that routine hospitalization is unnecessary.115 Even so, pneumomediastinum in children admitted with asthma is a marker of prolonged hospital stay.117 This increase seems to be because concerned physicians are managing these children more conservatively (i.e., prolonged confinement) rather than any actual difference in the degree of illness.
Asthma is a well-recognized cause of chest pain in children and adults. Frequently, children will complain of chest pain in addition to, or instead of, shortness of breath. As with arrhythmias, young children may have no language to express dyspnea other than, "my chest hurts." Older children may complain of chest "pain" instead of chest "tightness," both of which frequently improve with therapy. As noted above, clinicians should be on the lookout for complications of asthma (such as pneumothorax or pneumomediastinum) in asthmatic children complaining of chest pain.
Exercise-induced asthma is a frequently overlooked cause of exertional dyspnea and chest discomfort, partly because it is not considered and because it is difficult to diagnosis in the ED.52,118 The primary stimulus for bronchospasm is the release of inflammatory mediators during exercise.119-120 Increased ventilation may result in water loss and cellular dehydration, increased osmolarity, and destabilization of eosinophils, mastocytes, and other cells.120
In one study of 88 otherwise healthy children without a prior history of asthma or heart disease referred to a pediatric cardiology clinic for chest pain, 73% demonstrated a fall in FEV1 on treadmill exercise testing.118 Most experienced their referral symptoms of chest pain or tightness and shortness of breath even in the absence of wheezing. Pain was noted to be midsternal in 80%, and chest tightness was reported in 22% (characteristics commonly associated with cardiac disease). All but 1 patient reported subjective improvement with bronchodilators and almost three-quarters demonstrated objective improvement in pulmonary function tests.
Nudel et al performed stress tests on 180 consecutive patients aged 5-22 years with exertional chest pain or dyspnea on exertion.121 Of 147 children with chest pain, 14 (9.5%) demonstrated EIA. Of 33 children with dyspnea on exertion, 7 (21%) had EIA with demonstrable decreases in peak expiratory flow rates. No cardiovascular abnormalities were otherwise identified in this cohort.
These studies both used exercise stress testing to confirm the diagnosis of EIA. Care should be taken in making a definitive diagnosis in the emergency department as the accuracy of clinically diagnosed EIA may not be high.122 Mast cell stabilizers, such as cromolyn or nedocromil, as well as albuterol, all effectively prevent the symptoms of EIA when inhaled prior to exercise.118-120
Pleurisy is the nonspecific term commonly used to describe inflammation of the pleural membranes.123 A host of conditions can cause pleural irritation, few of which are primary pleural processes. Some common or serious etiologies include: pulmonary infarction, adjacent pneumonia, mediastinitis, trauma (e.g., pneumothorax), malignancy (usually metastatic), connective tissue disorders, and acute hemorrhagic pancreatitis.123
Infection is the most common cause of pleuritic chest pain (with associated effusion).123 Viral and mycoplasma pneumonias frequently have unrecognized effusions which may be transiently uncomfortable.83,123 Treatment is directed towards adequate analgesia until the infection resolves. Large pleural effusions, empyema, and consolidated pneumonias require antibiotic therapy and measures beyond the scope of this article. In addition, take care not to miss serious systemic illnesses noted perviously when evaluating children with what appears to be a benign viral pleurisy.
While gastroesophageal disorders are responsible for a significant proportion of adult chest pain, little is written about their contribution to pediatric chest pain. The cardiac and esophageal plexi both arise from the vagus and sympathetic trunks, and localization of chest pain can be difficult and confused with cardiac.
Using esophageal manometry, endoscopy, and acid perfusion tests, Berezin et al found that 21/27 (78%) studied children with "idiopathic" chest pain had a likely gastrointestinal etiology.124 Sixteen had evidence of esophagitis, 4 had gastritis, and 1 demonstrated diffuse esophageal spasm. None of these patients had classical symptoms of gastrointestinal reflux or heartburn and all responded to appropriate therapy. In a follow-up study of children with atypical chest pain, the pain response to intraesophageal acid perfusion (Bernstein test) in 45 children with esophagitis demonstrated on endoscopy was compared to 15 children without esophagitis.125 Acid perfusion reproduced the chest pain in 18/45 of those with esophagitis and 0/15 of those without esophagitis. Their conclusions were that esophagitis is a significant cause of atypical chest pain in children, despite that lack of a suggestive history.
The presence of epigastric tenderness on exam was associated with a high likelihood of having a gastrointestinal etiology. In a study of 132 children referred to pediatric cardiology for evaluation of chest pain, one-third (44/132) were found to have epigastric tenderness.126 Endoscopy was positive in 93% (41), defined as the presence of some combination of "gastritis," "duodenitis," and "esophagitis." According to the authors, 38 of the 41 children (97%) had resolution of their chest pain symptoms with appropriate treatment.
Unusual violent vomiting can result in esophageal rupture, mediastinitis, and sepsis. Boerhaave's syndrome is named after Herman Boerhaave, a Dutch surgeon practicing in the early 18th century, after he described the death of Dutch Grand Admiral Baron J van Wassenaer who died soon after developing chest and abdominal pain after vomiting.127There is a single case report of an 8-year-old child who developed esophageal rupture and mediastinitis after vomiting from ruptured appendicitis.69
Even as we search vigorously for underlying cardiopulmonary disease, the most common discharge diagnosis will be musculoskeletal chest pain. Causes of musculoskeletal chest pain are multiple, and costochondritis is not the correct diagnosis for every patient who grimaces when their chest is poked by an overzealous clinician, see Table 3. Simple muscle strain may be responsible for the majority of musculoskeletal chest pain, and since few diagnostic tests exist to conclusively prove a musculoskeletal etiology, the diagnosis remains a clinical one after more serious causes of chest pain have been excluded. The following discussion focuses on a few less commonly recognized conditions.
First described by Davies Colley in 1922, slipping rib syndrome is characterized by a dull ache lasting hours or days that is punctuated by intermittent, sharp, stabbing pain and precipitated by certain postures and movements.128-129 The 8th, 9th, and 10th ribs (false ribs) are not attached to the sternum; they are connected anteromedially by fibrous tissue. Pain results from inadequate stabilization of the rib tips, allowing them to move upward and posteriorly, impinging upon the superior intercostal nerve.129-131 In some cases, this disruption may be posttraumatic.131
Patients present with mild tenderness of the lower costal cartilages (often perceived as upper abdominal pain) which is elicited by the "hooking" maneuver: grabbing under the costal margin and pulling anteriorly.132 Treatment consists of reassurance and NSAIDs.129,131,133 Persistent symptoms may be managed with local anesthetic injections or surgical resection in recalcitrant cases.110,140 As with many unusual conditions, the key to successful management is making the correct diagnosis.
Costochondritis: this poorly understood condition is defined as tenderness, not over a rib, but over the costochondral or chondrosternal joints.134
Xiphodynia is a form of costochondritis. The xiphoid is innervated by the 4th through 8th thoracic intercostal nerves and may mimic cardiac, gastric, or biliary pain.135 While xiphodynia may be confused with other disorders, reproducible tenderness of the xiphoid process to moderate pressure is suggestive.135
Frequently confused with costochondritis, Tietze's syndrome is a rare inflammatory condition of the costochondral junctions of the chest wall that is predominant in the teens and early twenties.136-137 Objective swelling of a costochondral junction is a mandatory feature of this condition and appears as a firm, tender, round swelling with warmth and erythema.137Generally, there is only a single articulation (usually the second costosternal junction), but multiple lesions are possible. The clinical course is benign and usually resolves over days to weeks.
The term "precordial catch" (or Texidor's twinge, named after T.A. Texidor) serves as a descriptive term for a brief, intense left precordial chest pain that is common in the young.138-140 The best description was that of Richard Asher:
"Have any of you ever had a very brief, sharp, needle-like pain, near the apex of the heart, acutely localized to one point seemingly inside the chest wall, but feeling as if something was adherent to it? Breathing sharpens it, so there is often a disinclination to take a deep breath while it lasts. It comes on out of the blue, it passes off in a few minutes, and although acute it is not at all distressing."139
It is nonexertional and may be exacerbated with inspiration before it resolves less than a minute later.138-139,141 Others note that it is extremely localized, with a sudden onset and resolution, lasting less than a couple of minutes.138 The etiology is unknown but the prognosis for a long and fruitful life is excellent.
Breast pain is not uncommon in the adolescent years and, while identifying the breast as the source of chest pain should not be difficult, it is worthy of mention. Physiological breast pain is not uncommon in adolescent females and is often cyclical (although young adolescents may lack the somatic insight to make this connection). Increased sensitivity of the breast to estrogen and progesterone is the likely etiology; "unrecognized" pregnancy should be strongly considered.
Breast pain that is associated with lumps may be fibrocystic in nature. Uncommon in adolescence, fibrocystic breast disease increases with age and presents as breast pain and tenderness that is greatest just prior to menses.142-143 Treatment consists of reassurance, nonsteroidal anti-inflammatory agents, and, possibly, evening primrose oil.142-143 When lesions persist longer than three months or continue to enlarge, the patient should see a breast specialist.142 Fibroadenomas are discrete, rubbery lesions that enlarge, slowly and tend not to be painful.143
Blunt trauma is a well-known cause of chest pain and rib fractures. But unrecognized trauma or strain may produce musculoskeletal chest pain. Stress fractures of the ribs, including the first rib and sternum, are a form of indirect trauma and are secondary to the extreme muscular forces produced by competitive young athletes.128 Excessive muscular activity, such as heavy lifting or vigorous physical activity, can cause injury to the intercostal muscles. Pain between the ribs may be persistent and exacerbated by movement, coughing, or breathing.128 The inciting event may be unknown or forgotten. Severe cough has also been reported to produce rib fractures in addition to intercostal muscle strain.118 (Don't forget pertussis!)
A substantial minority of children presenting with chest pain will be discharged without a clear diagnosis, particularly adolescents.7,11-12, 15 Functional or psychogenic and "idiopathic" chest pain will fall into this category. Adolescents frequently present with vague chest discomfort in the context of prominent emotional stressors.6
Psychogenic or functional chest pain need not be a diagnosis of exclusion, but the diagnosis should be made with caution. Psychogenic chest pain tends to be recurrent, often occurring for months before the patient presents to the ED. Other somatic complaints may be present (e.g., headache, abdominal pain). Often, a family history of chest pain is present.6,144 True psychogenic chest pain should be associated with clear, identifiable emotional stressors.6,144 These can be identified by performing a complete review of systems and skillfully slipping in appropriate psychosocial questions so as not to appear to minimize the complaint as "all in your head."
This is not a "diagnosis of dismissal." Psychosomatic symptoms cause very real discomfort and are common in older children and adolescents.6,144 Adolescents respond to emotional stressors with a broad array of cognitive, behavioral, and physiologic responses. The expression of a variety of physical symptoms, of which chest pain is one, is modified by social supports, the presence of real organic disease within the family, and behavioral patterns modeled by parents and guardians. Such presentations can be frustrating for physicians who function within a rigid diagnosis-treatment paradigm. Adolescents rarely malinger and remembered that even when a symptom or exam finding is clearly functional, they are manifesting signs of genuine distress and are in need of help.144
Hyperventilation is well recognized in patients complaining of shortness of breath, dizziness, and peripheral paresthesias (with or without tetany). Chest pain is also a symptom produced by hyperventilation, although the exact etiology of the pain is unclear.145 In adult studies, chest pain associated with panic disorder may be difficult to distinguish from anginal pain.146However, attributing a patient's chest pain to panic disorder or hyperventilation should be done with care as more serious disorders (such as pulmonary embolism or cardiac arrhythmias) may present with similar symptoms.
Studies have shown that a thorough history and physical are deemed the most important factors when evaluating a child in the emergency department with chest pain.21,147 Taking the history of a child with chest pain is no different from taking the history of any complaint, but some historical aspects should be more closely explored. What medications the child may be taking (e.g., tetracyclines can cause unrecognized erosive esophagitis and oral contraceptives are associated with PE) and whether or not the teenager has taken illicit drugs (e.g., cocaine, Coridicin®) are important considerations.19 The family should be queried about relatives who were victims of sudden death, cardiomyopathies, or exercise intolerance and about any history of early cardiac disease or inflammatory/rheumatic disease.19 One feature of the history that should be considered is whether the chest pain has had an impact on the child's activity, especially if the presence of the pain has forced lifestyle modifications.19,21
Cardiac examination is perhaps one of the most challenging aspects of the physical exam for the pediatric or emergency medicine physician. A step-wise approach is the simplest way to conduct a cardiac examination. After attending to the ABCs, the initial focus should be on a careful inspection of the general state of the child (i.e., degree of distress), the presence of pallor or cyanosis, and the respiratory rate and effort.
Palpation of the chest wall is important, but chest wall tenderness should be interpreted cautiously and in context with the complaint. Reproducible chest tenderness is one of the most common findings in children presenting with chest pain; one should not automatically assume that the presence of chest tenderness precludes the presence of a more serious etiology.148
Heart rate and regularity should be the first aspect of the auscultatory exam. Gallops, clicks, or fourth heart sound should prompt a more detailed evaluation. The presence of a widely split S1 or S2 and the timing of the split can be of importance, as this can be an indication of a congenital heart malformation (e.g., atrial septal defect). Murmurs are common in children and can be challenging to sort out; most are innocent in nature.149 However, in the context of chest pain, they may represent critical clues (e.g., hypertrophic cardiomyopathy). Any diastolic murmur should be investigated.9In regard to the emergency department setting, one study found that initial clinical evaluation distinguished between innocent and pathologic murmurs among experienced ED physicians, although this involved an adult population.150 A new onset murmur in a child who has chest pain, fever, or difficulty breathing warrants careful evaluation. Chest radiography and echocardiography in children with new onset murmurs continue to be valuable in the evaluation of such patients.151
The approach to the laboratory and radiographic evaluation of a child presenting to the emergency department with chest pain can sometimes be complicated. The majority of diagnostic studies on children with chest pain generally include an ECG and/or CXR.10,21,148,151 There are a variety of other diagnostic tools available to the clinician; choosing a study that will provide the clinician with useful information can be challenging.
Prospective studies suggest that the yield of electrocardiography is low in the absence of a suggestive history and physical.7,12 Most patients do not have underlying cardiac disease, and the ECG may actually be normal in patients with structural heart disease or intermittent arrhythmias. Nevertheless, it is relatively inexpensive, noninvasive, easy to perform, and one of the few diagnostic screens readily available in the emergency department and should be used liberally. One should not underestimate the psychological value of obtaining an ECG in children with chest pain. It demonstrates to the family that the complaint is being taken seriously and provides comfort that "everything is ok," particularly given the aforementioned concerns that families have about underlying cardiac disease. Routine ECG still has a role in the evaluation of nonspecific chest pain in children.151
Interpretation of the pediatric ECG is beyond the scope of this article, but a few important principles should be mentioned. Remember that ECG findings are age-specific. For example, infants have a right ventricular-dominated ECG and T waves are usually inverted. Certain ECG findings should raise concerns. While inverted T waves are normally found in the anterior leads V1 through V3, flat or inverted T waves in the lateral leads (V5, V6, I, and aVL) should provoke close scrutiny and may point to the presence of underlying cardiomyopathy or coronary artery abnormality.2,9 Abnormal rhythms are unusual on a routine ECG in the asymptomatic individual but evidence of atrioventricular node dysfunction, short runs of ventricular tachycardia, a prolonged QT interval, or a delta wave may be noted. Care must be taken in using computer programs to diagnose either MI or pericarditis in adolescents as computers are prone to overdiagnose.23,152-153
As with electrocardiography, indiscriminate use of chest radiography is of low yield.12,21One study suggested that while abnormalities such as atelectasis or hyperinflation may be noted with intense scrutiny, unexpected findings are unlikely.21 In another study, a large number of CXRs were performed in children presenting to the emergency department with chest pain (336 patients); the majority of positive results were due to infectious pulmonary pathology, and if this group was excluded, the rate of positive findings on CXR was only 2%, although there were no clinical predictors of which patients would have positive findings.10CXR may still have a role in the evaluation of chest pain, however. Cardiomegaly may point towards a previously unrecognized cardiac problem, particularly in children with new heart murmurs.151 CXR may demonstrate the presence of an unexpected pneumothorax. Finally, like ECG, it may be of psychological value in reassuring the patient and the family that there is no obvious pathology.
The most widely used biomarkers for cardiac injury are creatine kinase MB fraction (CK-MB) and cardiac troponins. Cardiac CK and troponins are released into the blood stream with cardiac injury over a predictable time period; elevated serum levels are 100% sensitive within 10-12 hours of injury.154 The major problem with relying on CK-MB levels for the accurate diagnosis of cardiac injury is one of specificity.155-157While skeletal muscle largely consists of CK-MM, small amounts of CK-MB are also released with skeletal muscle injury. Rhabdomyolysis, trauma, heavy exertion, myopathies, and renal failure can all lead to falsely elevated CK levels.103,158a,158b Furthermore, in an attempt to assure an acceptable level of specificity, the CK-MB must exceed a predetermined threshold at the expense of sensitivity and may therefore miss minor cardiac injury.155,157
While CK-MB is still widely used in the diagnosis of acute cardiac injury, cardiac troponins (cTnI and cTnT) have come into widespread and preferential use due to their superior sensitivity and specificity.158 Troponins are so sensitive for cardiac injury that extremely low levels become indicative of ischemic injury (as opposed to traditional myocardial infarction).158 However, false "positives" do exist so minor elevations should be interpreted carefully in traditionally low risk populations (i.e., children). Noncoronary (i.e., acute coronary ischemia) causes of elevated troponin levels include non-ischemic dilated cardiomyopathy, renal insufficiency, muscular disorders, CNS disease, HIV, sepsis, preeclampsia, and pulmonary disease.159
An understanding of the kinetics of cardiac enzyme release in acute cardiac injury is critical. While cardiac biomarkers are very sensitive, their sensitivity is a function of the timing of their acquisition. These biomarkers do not approach 100% sensitivity until 8-12 hours after the injury so they cannot be used to definitively exclude cardiac injury prior to this time period.160-161 In other words, a negative troponin assay drawn less than eight hours from the onset of chest pain cannot be used to exclude the possibility of acute coronary ischemia.
The use of d-dimers in pediatric patients is not as well studied as in the adult population, and definitive data regarding their use in pediatrics are lacking. D-dimers are the products of fibrinolysis and are clinically used most frequently to detect the presence of pulmonary embolism. A negative (normal) d-dimer has been found to be very effective in reliably excluding pulmonary embolism in appropriately selected populations (i.e., low pretest probability).162-163
Several types of d-dimer assays are available: latex agglutination assay, enzyme-linked immunosorbent assay (ELISA), turbidimetric assay, whole blood assay, and immunofiltration assay. While latex agglutination tests are not very sensitive or specific, the ELISA and turbidimetric assays have been shown to be more reliable.106,164 In one study of adult patients being evaluated for pulmonary embolism, ELISA testing was found to have a sensitivity of 97% and a specificity of 44%) for the detection of pulmonary embolism, while turbidimetric assays were reported to be 95% sensitive.106 Physicians must be aware of what assay their laboratories are using if they are relying on this information. It remains to be seen if d-dimers in the pediatric population will ultimately be proven to be useful.
Echocardiography should be obtained urgently in patients who have clinical signs suggesting cardiac tamponade (such as hypotension, tachycardia and pulses paradoxus) as this condition can quickly lead to hemodynamic compromise and death.155,215 If an effusion that is leading to tamponade is present, echocardiography is the study of choice to guide both diagnosis and therapeutic intervention. More commonly, echocardiography is useful to identify the presence of HCM, congenital coronary artery abnormalities, or wall motion abnormalities present in acute ischemic conditions.
Chest CT is a useful imaging modality to detect and further evaluate thoracic pathologies, including pneumothorax, empyema, pericardial disease, and aortic dissection. CT has been demonstrated to be extremely useful in the detection of pulmonary embolism in the adult population.105,164-165 With regard to the detection of pulmonary embolism, CT has not been well studied in the pediatric population; the diagnosis has been made primarily by V/Q scan.
V/Q scans have long been used in the detection of pulmonary embolism in adults, but interpretation can be fraught with difficultly and physicians should be very comfortable with their understanding of this diagnostic study. For instance, a high probability result in a low probability patient is not diagnostic of PE and a low probability scan in a child at high risk for pulmonary embolism is not reassuring. An intermediate probability scan is essentially non-diagnostic, see Table 4. A complete discussion is beyond the scope of this article and further reading is recommended.95,101
Chest pain is a symptom, not a diagnosis, and treatment obviously depends on the underlying cause. It is difficult to treat what you cannot put a name to the cause so a correct diagnosis is imperative, even if it is as nonspecific as "cardiac chest pain." Emergency medicine physicians are first and foremost diagnosticians. Once a correct diagnosis is arrived at, texts can be referred to and consultants conferred with. Having said this, some disorders are time dependent and may leave little time for research; most of these fall under the rubric of cardiac and will be discussed below, see Table 5.
Active cardiac ischemia and myocardial infarction demand a rapid and assertive response. In adults, treatment is directed towards rapidly restoring coronary blood flow (i.e., intravenous thrombolytics or percutaneous angioplasty). Experience with thrombolytics in pediatrics does exist and is useful in appropriate circumstances (e.g., aneurysmal thrombosis in Kawasaki disease and pulmonary embolism).102,107,109,166-167
However, acute ischemia and focal infarction (or ECG changes) in children is often due to nonthrombotic events, such as coronary vasospasm, myocarditis/ pericarditis, or a congenital arterial lesion. A higher level of proof is necessary prior to starting thrombolytics in pediatric and adolescent patients, such as cardiac catheterization and/or echocardiography, in conjunction with diagnostic ECG and laboratory findings.
Active chest pain in the setting of cocaine ingestion deserves mention. While aspirin and nitroglycerin are strongly recommended, treatment of tachycardia and increased blood pressure with beta-blockers (such as metoprolol or labetolol) must be avoided.168-169 Administration of these agents allows unopposed alpha adrenergic activity and potentiates the activity of cocaine on the coronary vasculature.170 Benzodiazepines attenuate the centrally mediated sympathomimetic effects of cocaine.171 Likewise, beta adrenergic blockade should be avoided in severe hypertension in the setting of cocaine intoxication; nitroprusside or phentolamine are more prudent choices.216 If the ECG suggests transmural myocardial infarction, immediate cardiac catheterization should be considered.168-169
Acute management of fulminant myocarditis is supportive and consists of therapies for CHF and dysrhythmias and inotropic support, if necessary.32 Critically ill children may require vasopressors, ventricular pacing, or even extracorporeal membrane oxygenation.72,81,172 Electrical instability may lead to lethal arrhythmias before the patient is transferred to the intensive care unit, and the myocardium may need to be stabilized with amiodarone.32 Intubation may result in refractory hypotension as sedation and paralysis suddenly removes maximal endogenous sympathetic support. If at all possible, intubation should be deferred until hemodynamic parameters are optimized.133
Aortic dissection is often a catastrophic event for which even perfect management does not guarantee a good outcome. Types I and II (involving the ascending aorta) require immediate surgical intervention, while a type III aortic dissection (of the descending arch) is often managed medically with aggressive control of blood pressure and heart rate. Generally, beta-blockade (e.g., esmolol) is begun to prevent reflex tachycardia prior to the initiation of vasodilators (e.g., nitroprusside). Aggressive control of blood pressure and heart rate are critical in the emergency department before transfer to an intensive care unit.
While it is often said that no patient should wait for a CXR before definitively treating tension pneumothorax, this over simplifies the situation. Frequently, signs and symptoms are nonspecific, particularly in their respective contexts of trauma or status asthmaticus. Having said that, any patient with a real possibility of tension pneumothorax who is critically ill and on the verge of cardiovascular collapse should be treated.173 Definitive treatment of a tension pneumothorax is straight-forward: a large bore needle (14- or 16-gauge) should be placed in the second or third intercostal space anteriorly or the fourth or fifth space laterally. Improvement in the heart rate and blood pressure as well as improved ventilation signifies success - if the diagnosis is correct. Published reports (as well as this authors' experience) should remind the reader that needle thoracostomy can be unsuccessful and rapid placement of a standard chest tube may be necessary.174
With the advent and success of pediatric cardiac surgical techniques, the survival of children with complex congenital heart disease is increasing.175 Consequently, the potential for their presentation to the pediatric emergency department can be expected to increase as well. There are many surgical procedures and a complete review of them is beyond the scope of this article.
One phenomenon that can be associated with any cardiac procedure is postpericardiotomy syndrome. This is essentially a delayed pericardial or pleural reaction whose etiology is unclear.175-176 The clinical presentation of this syndrome is a constellation of symptoms, including fever, chest pain, and friction rub; it is usually manifested at around 1-2 weeks post-surgery and has an incidence of approximately 7-30%.175-176 There is an increased incidence among patients who are younger, those with a history of prednisone use, those with aortic valve replacement, and those who were given haloflurane for anesthesia.177 History and physical exam will often help identify these patients, but echocardiography should be performed as these patients are at risk for pericardial effusion, which can lead to pericardial tamponade.176 Pericardial tamponade can indeed be a late complication of any cardiac procedure and can be associated with patients taking anicoagulants.179-180
Trauma is often a source of chest pain in pediatric patients. In instances of incidental or minor trauma, an exhaustive workup is not required. In cases of major trauma (i.e., motor vehicle accidents and severe blows to the chest), a more cautious approach is warranted.217 Although thoracic trauma does not account for a large portion of trauma in children (approximately 5%), mortality rates in pediatric victims of significant blunt chest trauma have been reported to be as high as 25%.181
Blunt chest trauma can be associated with cardiac injury. Usually, these patients can be identified by hemodynamic instability.178,182-184In a major thoracic trauma, the sternum can fracture, posing risk of injury to the right ventricle and the aorta; additionally, myocardial contusion can occur, although this is rarely a cause of death.184-185 Children who suffer rupture of a cardiac chamber usually die at the scene.158
Children are much more likely to sustain pulmonary injury than a cardiovascular insult.186Pulmonary contusion, pneumothorax, tracheobronchial disruption, and hemothorax, as well as aortic injury and rib fractures have all been well described.186-188 Pulmonary contusion is by far the most common sequelae of pediatric blunt chest trauma.186In children, there may be no external indications of underlying lung injury due to their more elastic, compliant chest walls, but thoracic injuries may be detected radiographically.186-189
Clinical decision rules for clinically identifying a thoracic injury have been proposed for children sustaining blunt thoracic trauma.186 The presence of low systolic blood pressure, tachypnea, femur fracture, and abnormal auscultatory examination have been shown to be sensitive predictors of intrathoracic injury.186 As noted, physical findings may be lacking in children and identification of pulmonary contusion, life-threatening hemothorax, and other injuries can be delayed; one study found a delay in diagnosis of up to 48 hours in some patients.187,190 In cases of pneumothorax and hemothorax, expectant management may be an option, although tube thoracostomy is often required.181,191-194
Initially, radiographic diagnosis of traumatic pulmonary injury can be difficult, but studies have shown that most children will have an abnormality on plain CXR.187-188 While helical CT scan has been shown to be highly sensitive for thoracic injury, CXR as an initial study choice is cost effective and remains clinically valuable; CT should be used in selective patient populations.195 One study specifically looking at whether or not CT was a useful adjunct to CXR found that when radiographic abnormalities were present or when there were clinical signs of thoracic injury, CT was useful in identifying additional occult injury 66% of the time; in the group whose CXRs were normal and who had no clinical signs of thoracic injury, CT identified occult injury 39% of the time.196 Furthermore, a significant change in management of the injury occurred 20% of the time in the first group and 5% of the time in the latter group; this suggests that the judicious use of CT appears to be most useful in patients who have either clinical or radiographic signs of thoracic injury.196 Although CT has been shown to be more sensitive for detecting abnormalities not seen on CXR, several other studies have found that the detection of such abnormalities rarely leads to a change in the management of the patient.197-199
Chest pain is not uncommon in patients with cystic fibrosis.218While chest pain may be chronic in some children, acute pneumothorax is a well-described complication of cystic fibrosis and has been found to be associated with a poor outcome (long term), including concomitant infection with Pseudomonas aeruginosa. Failure to resolve with conventional treatment and recurrence of the pneumothorax has been reported.200
Children with chest pain who have sickle cell disease should be evaluated for the presence of acute chest syndrome (ACS). ACS is generally defined as the presence of new pulmonary infiltrates in the context of fever or respiratory symptoms in patients with sickle cell disease.201 Symptoms have been shown to vary among different age groups; fever and cough are the most common findings in young children.202 Young children rarely have pain while older patients more often present afebrile and with severe pain.189 ACS in children is often associated with infections; mycoplasma has been found to be quite prevalent in pediatric ACS.189,203 In older populations, ACS may have infection as the inciting agent, but pulmonary fat embolism is often present.204 Children have a lower death rate and a shorter hospital stay compared to adults with ACS.189
Laboratory evaluation should include a complete blood count, blood cultures, and CXR. A study looking at the utility of various serum biomarkers to more accurately identify acute chest syndrome identified only secretory phospholipase A to be of any value (not available in the author's ED).205 While acute chest syndrome implies serious pulmonary pathology, myocardial involvement has been reported.206 Treatment is generally supportive and includes fluids, analgesia, bronchodilators, and broad spectrum intravenous antibiotics.201,204,207 On occasion, blood transfusion may be necessary.201,207An abnormal alveolar arterial oxygen gradient has been shown to be a clinical predictor of the need for blood transfusion, but transfusion is often indicated in hypoxic patients.201,206 Additionally, intravenous dexamethasone has been evaluated in children with mild to moderate ACS and was found to be of some benefit, but further study was suggested.207
Ultimately, disposition depends upon diagnosis — or lack of diagnosis. Unstable or ill patients with identified cardiopulmonary disease should obviously be admitted to the hospital under the care or consultation of the appropriate specialists. Children and adolescents who are less ill, with a clear diagnosis, (such as pneumonia or musculoskeletal chest pain) can have definitive treatment begun in the emergency department and referred back to their primary physician. Children with a presumed diagnosis should be referred back to their primary physician for reevaluation to either confirm the ED diagnosis or readdress the problem, see Table 6.
Many children, however, do not have a clear, identifiable reason for their chest pain. It is important to remember that many diagnoses are presumed, not clearly provable, and may ultimately prove to be incorrect.17 Humility is the essence of an honest and skilled physician. If there are any features that suggest the possibility of an underlying cardiac disorder, these children should be referred to their primary physician and probably a specialist as well (after a brief consultation from the department). These children should be restricted from strenuous physical activity until follow-up evaluation. While sudden death is a highly unlikely event, this possibility is obviously of critical medical and legal importance to the patient and physician.
Prospective series of pediatric chest pain clearly emphasize its benign nature.53,147,170,179,180,181,211 A clinician practicing in a busy practice could easily go through an entire career without seeing some of the conditions mentioned in this article. But these conditions will pop up, sometime and somewhere, and it is our fundamental professional responsibility to be aware of and on guard for the uncommon. This requires active, engaged learning and astute clinical acumen. Most pediatric chest pain is not of consequence. But do not be complacent. Few people miss the obvious; "listen" for subtle clues. Recognizing the ability of the adaptive unconscious, Malcolm Gladwell reminds us in his book Blink that our unconscious "often delivers a better answer than more deliberate and exhaustive ways of thinking."208
1. "A pediatric patient with chest pain has nothing wrong with them."
2. "The patient is here for a diagnosis and needs to leave with a definitive diagnosis."
3. "There is little value in ordering diagnostic studies for pediatric chest pain because the literature proves its low yield."
4. "A normal ECG and CXR eliminate the possibility of cardiopulmonary disease."
5. "I assumed that the underlying cause of chest pain in a hysterical adolescent was hysteria."
6. "It's not cardiac; therefore, it's not serious."
7. "Since it hurts when I touch, it must be costochondritis (I mean musculoskeletal)."
8. "Everybody with undifferentiated chest pain should have cardiology follow-up on the chart."
9. You have read this article too closely and think that everyone suffers from a fatal flaw in their anatomy.
The adolescent presented in the introduction had pericarditis. His presentation was not atypical for pericarditis, but could easily be confused with coronary ischemia, particularly given the focal ST changes. Cardiac catheterization demonstrated normal coronary arteries and, with time, the ST elevations became diffuse, confirming the diagnosis of acute pericarditis.
Evidence-based medicine requires a critical appraisal of the literature based upon study methodology and number of subjects. Not all references are equally robust. The findings of a large, prospective, randomized, and blinded trial should carry more weight than a case report.
To help the reader judge the strength of each reference, pertinent information about the study, such as the type of study and the number of patients in the study, will be included in bold type following the reference, where available. In addition, the most informative references cited in this paper, as determined by the authors, will be noted by an asterisk (*) next to the number of the reference.
Rick Place; Robert Vezzetti
August 1, 2007