Chapter 20
Evaluation of Chest Pain
The patient who presents with chest pain in the
outpatient setting poses a diagnostic challenge. The spectrum of
diagnostic possibilities ranges from life-threatening cardiac,
pulmonary, and aortic etiologies to esophageal and musculoskeletal
causes. Harmless conditions may mimic more serious disease. Atypical
chest pain can be especially problematic. The primary physician must be
skilled in quickly and accurately differentiating the patient who
requires immediate hospitalization from the person who can be safely
evaluated in the outpatient setting. Initial decision making depends
predominantly on a careful assessment of the history supplemented, when
possible, by checking for a few key physical and electrocardiographic
findings. Further testing must be selected judiciously to avoid
generating false-positive results.
PATHOPHYSIOLOGY AND CLINICAL PRESENTATION (1,2,3,4,5,6,7,8,9,10,11,12)
Chest pain may arise from chest wall, intrathoracic, abdominal, or even psychophysiologic sources.
P.126
Chest Wall
Pain originating in the chest wall is usually due to
musculoskeletal pathology, although occasionally nerve injury is
responsible. Typically, the pain can be pinpointed by the patient,
being somatic in nature. It is aggravated by deep inspiration, cough,
direct palpation, and movement. Common sites of involvement are the
costochondral and chondrosternal junctions. Duration ranges from a few
seconds to several days and quality from sharp to dull or aching.
Sometimes the patient complains of tightness. Vigorous and unaccustomed
exertion can lead to muscular and ligamentous strain, which may account
for some cases. Others are due to costochondritis (Tietze's syndrome),
which is an inflammatory condition that causes localized swelling,
erythema, warmth, and tenderness at the costochondral or chondrosternal
junction. Rib fracture may produce a similar picture, although location
is different, and there is a history of antecedent trauma or metastatic
cancer. Of interest, there is an increased frequency of musculoskeletal
pain in patients with angina, which causes a potentially confusing
clinical presentation.
Nerve injury due to a recrudescence of herpes zoster
infection can be very painful, with a dermatomal distribution being
characteristic. The pain may precede the typical rash (picturesquely
described as “dew drops on a rose petal”; see Chapter 193)
by 3 to 5 days. Neurologic complaints range from hypoesthesia to
dysesthesia and hyperesthesia. In the elderly, the pain may persist for
months, long after the rash resolves.
Nerve injury from cervical root compression (see Chapter 148) due to cervical spine disease or a thoracic outlet syndrome
can produce pain in the chest and upper arm, superficially resembling
angina. In the outlet syndrome, a cervical rib may compress part of the
brachial plexus, resulting in motor and sensory deficits in an ulnar
distribution at the same time that there is discomfort in the chest and
upper arm (see Chapter 167).
Lungs and Pleura
Inflammation or distention of the pleura produces true
“pleuritic pain,” which is worsened by deep inspiration and cough but
relatively unaffected by movement or palpation. A host of causes can
trigger the inflammatory process, including pneumonia, pulmonary embolization with infarction, neoplasm, uremia, and connective tissue disease.
The more florid the inflammation, the greater is the pain. An
infectious origin is more likely to cause pain than is a low-grade
serositis associated with connective tissue disease.
Pneumococcal Pneumonia and Pulmonary Tuberculosis
Pneumococcal pneumonia and pulmonary tuberculosis are
the archtypical pneumonias associated with pleural involvement. The
onset of symptoms in pneumococcal pneumonia (fever, chills, cough,
sputum production, pleuritic chest pain) may be acute and mimic
pulmonary embolization (see Chapter 52).
Pulmonary Embolization
Pulmonary embolization can cause pleuritic pain,
especially when embolization leads to parenchymal infarction and
pleural reaction. Pleural rub, effusion, low-grade fever, and
hemoptysis also herald pulmonary infarction with pleural involvement.
However, in some instances, the pain may be less clearly pleuritic and
is often absent; it is estimated that fewer than 10% of all embolic
episodes are accompanied by chest pain. The classic cardiopulmonary
manifestations of embolization — dyspnea, tachypnea, and tachycardia — are nearly universal but may be short lived. Hypoxemia
and oxygen desaturation may or may not be present, depending on the
degree of mismatch between ventilation and perfusion. Severe
embolization can cause acute pulmonary hypertension
manifested by systemic hypotension, jugular venous distention, an
accentuated pulmonic component of the second heart sound, acute
tricuspid regurgitation, chest x-ray abnormalities, and
electrocardiographic manifestations of acute right heart strain (see
later discussion).
Spontaneous Pneumothorax
Spontaneous pneumothorax stretches the pleura and
results in acute onset of pleuritic pain and dyspnea. The condition
occurs in young persons and those with emphysema, in which there can be
rupture of a bleb. If the pneumothorax is large, deviation of the
trachea may be observed.
Pleurodynia
Pleurodynia is a self-limited source of pleuritic pain,
most common in children and young adults and associated with a
respiratory viral infection, such as that due to coxsackie virus B. A
typical viral syndrome precedes the acute onset of chest pain. Chest
pain in the setting of a viral upper respiratory infection may also
occur from cough-initiated injury to the chest wall or from
bronchospasm. Young healthy persons sometimes note a sudden sharp
pleuritic episode relieved by taking a deep breath, referred to as the precordial-catch syndrome. Its mechanism is unclear, but a transient folding of the pleura on itself is hypothesized.
Heart and Pericardium
Angina Pectoris
Angina pectoris due to occlusive coronary artery disease is the most important cardiac source of chest pain. Coronary perfusion may be similarly compromised by critical aortic stenosis leading to angina (see Chapter 33).
The classic hallmarks of angina are its sudden onset with exertion,
emotional stress, or eating (usually a very large meal) and its relief
within minutes by rest or nitroglycerin. Patients usually describe
their chest pain as a squeezing, heaviness, or pressure, although it
may be burning or sharp. The quality of the pain is not diagnostic, and
many patients state the sensation is more a “discomfort” than a true
pain. Radiation to the jaw, neck, shoulder, arm, back, or upper abdomen
is common and may present in the absence of chest symptoms. At times,
the arm is reported to feel numb or tingling. Autonomic epiphenomena
such as diaphoresis and nausea may accompany the episode, as may
dyspnea if there is transient pump failure or marked anxiety. Episodes
last 2 to 20 minutes. Prompt response to nitroglycerin is
characteristic; relief is usually obtained within 5 minutes.
Gender and racial differences in presentation have been explored. The clinical presentation of myocardial ischemia in women,
particularly women younger than the age of 60 years, can differ from
that in men. Chest pain is more likely to be absent or atypical (see
later discussion) and may be overshadowed by exertional fatigue,
shortness of breath, diaphoresis, arm tingling, jaw discomfort, nausea,
or other epiphenomena of ischemia that are easy to dismiss as
“noncardiac.” Diabetes mellitus is a major risk factor for early onset
of ischemic heart disease in women. With regard to the effect of race
on presentation, acute chest pain presentations appear to be similar
among whites and African Americans.
P.127
Unstable Angina
Unstable angina is one of the acute coronary syndromes, along with non–Q-wave myocardial infarction and Q-wave myocardial infarction.
All are important causes of coronary chest pain and result from acute
plaque rupture, which triggers platelet activation, thrombin clot
formation, and active vasoconstriction. The clinical presentations of
unstable angina include onset of new chest pain within the last 2
months severe enough to inhibit activity; established angina now
increasing in frequency, severity, and duration (crescendo angina) and occurring with progressively less provocation; and development of rest pain or nocturnal angina
in a person with a previously stable anginal pattern. Immediate
mortality risk is high (up to 4%) but declines after 1 to 2 weeks.
Clinical features associated with greatest risk include rest pain in
excess of 20 minutes, signs of pump failure (hypotension, rales, S3), new or worsening mitral regurgitation, and 1 mm or more of ST-segment change with pain.
Women with unstable angina are less likely than men to
present with acute ST-segment elevation indicative of vessel-occluding
infarction. Compared with men presenting with unstable angina, they are
older and more likely to have diabetes, hypertension, and prior heart
failure, and typically present hours later into the episode.
Myocardial Infarction
Myocardial infarction is typically heralded by chest
pain exceeding that of unstable angina, but the presentation is often
more subtle or even silent, particularly in diabetics, the elderly, and
women. Bad prognostic signs include heart failure, hypotension, mitral
regurgitation, ST-segment elevation, and a new left-bundle-branch
block. Onset of postinfarction angina is also associated with high risk.
Variant Angina
Variant angina, as originally described by Prinzmetal,
refers to anginal pain occurring exclusively at rest in conjunction
with transient ST-segment elevation on electrocardiogram (ECG).
Classically, this syndrome was associated with coronary artery spasm at
the site of high-grade proximal fixed stenosis. However, other forms of
coronary disease may produce a similar clinical picture, and coronary
vasospasm may present in ways other than Prinzmetal's description. Cocaine abuse
can trigger ischemia by precipitating coronary vasoconstriction,
increasing myocardial oxygen demand, and enhancing platelet
aggregation. It may present as angina in a young person with no other
coronary heart disease (CHD) risk factors.
Atypical Angina (Atypical Chest Pain)
Atypical angina is a term
used to denote angina-like chest pain that differs in location,
quality, or other characteristics from more typical angina yet is still
suggestive by virtue of similar precipitants, timing, or other
features. Some define the term more precisely, indicating presence of
any two of angina's three cardinal features (substernal location,
exercise precipitation, prompt relief by rest or nitroglycerin). As
many as 50% of such patients who come to angiography prove to have
coronary disease. Among the remainder, there appear to be increased
incidences of panic disorder, major depression, esophageal disease, and
coronary microcirculatory dysfunction. The mechanisms for many of these
causes are not well understood, but recent interest has focused on the
coronary microcirculation.
Coronary Microvascular Dysfunction (Microvascular Angina, Coronary Syndrome X)
Coronary microvascular dysfunction (microvascular
angina, coronary syndrome X) has been noted among patients with typical
angina who exhibit an ischemic response to exercise stress testing yet
have entirely normal coronary angiograms. Proposed mechanisms include
abnormal microvascular responses to autonomic and biochemical stimuli.
Patients experience typical anginal chest pain and reductions in
subendocardial perfusion in response to adenosine infusion. Prognosis
is good and similar to normal persons without coronary disease. A few
develop mild left ventricular dysfunction or conduction abnormalities.
Microvascular dysfunction provides a possible explanation for the chest
pain of patients with hypertrophic cardiomyopathy.
Mitral Valve Prolapse
Mitral valve prolapse is notorious for its association
with atypical chest pain. The commonly held view of a link between the
two has been challenged by recent studies controlling more stringently
for selection bias. Some argue the apparent association is due to an
increased frequency of underlying psychopathology, such as panic
disorder (see later discussion), which may trigger chest pain. Symptoms
of autonomic dysfunction (e.g., palpitations, sweating, dizziness) may
sometimes accompany the chest pain and simulate an ischemic attack.
Pericarditis
Pericarditis may present with pleuritic pain, resulting
from spread of the inflammatory process from the relatively insensitive
pericardium to the adjacent pain-sensitive parietal pleura. The pain is
sharp, aggravated by respiratory activity, and sometimes precipitated
by swallowing if the posterior aspect of the heart is involved. When
the diaphragmatic surface of the pericardium is involved, pain will be
referred to the tip of the shoulder. Change in position may alter the
pain. Patients often note lessening of pain on sitting up and leaning
forward. Pericarditis can also produce a second type of pain that
mimics angina. Its most diagnostic physical finding is a two- or
three-component friction rub.
A vexing pericardial problem is the development of chest
pain after coronary bypass surgery. The return of typical angina raises
the specter of graft occlusion, but pleuritic pain suggests the postpericardiotomy syndrome.
Aorta
Aortic dissection is a
must-not-miss cause of chest pain. Almost invariably (70% to 90% of
cases), it begins with sudden onset of severe chest or interscapular
pain, maximal from the start, and tearing or ripping in quality. If it
begins in the chest, it may radiate to the interscapular region, neck,
jaw, lower back, or even down into the legs. Associated symptoms
include neurologic deficits from cutoff of blood supply to the brain,
spinal cord, or limb. Loss or diminution of a major peripheral pulse is
a key physical finding, as are new onset of aortic insufficiency and
pericardial tamponade due to dissection into the aortic root.
Esophagus
Esophageal pain can be the great mimicker of anginal
chest pain, producing chest discomfort that can resemble angina in
quality, location, radiation, and even precipitants (e.g., exposure to
cold, exertion). Unlike angina, esophageal chest pain is more likely to
persist as a dull sensation for several hours after an acute attack and
may occur with swallowing. The pain sometimes
radiates to the interscapular region. The chest pain may occur spontaneously or in the context of meals or acid reflux (manifested by retrosternal burning that may be brought on by a large meal, lying down, or bending over and is relieved by antacids). Some patients report dysphagia as an accompanying symptom. In some instances, studies of esophageal function reveal motor dysfunction (e.g., nonpropulsive contractions or “spasm”) and acid reflux from the stomach. Nitrates and calcium-channel blockers may provide relief in such cases, as they do for angina. Some patients with atypical chest pain and normal coronary angiograms manifest both esophageal spasm and microcirculatory dysfunction, raising the intriguing possibility of a generalized disorder of smooth muscle reactivity.
P.128
radiates to the interscapular region. The chest pain may occur spontaneously or in the context of meals or acid reflux (manifested by retrosternal burning that may be brought on by a large meal, lying down, or bending over and is relieved by antacids). Some patients report dysphagia as an accompanying symptom. In some instances, studies of esophageal function reveal motor dysfunction (e.g., nonpropulsive contractions or “spasm”) and acid reflux from the stomach. Nitrates and calcium-channel blockers may provide relief in such cases, as they do for angina. Some patients with atypical chest pain and normal coronary angiograms manifest both esophageal spasm and microcirculatory dysfunction, raising the intriguing possibility of a generalized disorder of smooth muscle reactivity.
About half of patients with noncardiac (i.e.,
angiogram-negative) angina-like chest pain report no concurrent
dysphagia or heartburn and manifest no signs of reflux or motor
dysfunction on detailed esophageal testing. In the past, such chest
pain was labeled as “noncardiac chest pain of unknown etiology”; however, controlled studies using impedance planimetry reveal esophageal hypersensitivity, hyperreactivity, and stiffness in a large proportion of previously undiagnosed patients. These findings suggest a sensory or “nociceptive”
etiology to much noncardiac chest pain. It appears that, in such
patients, normal degrees of esophageal distention result in exaggerated
perceptions of pain and in hyperreactivity.
Other Gastrointestinal Tract Sources
An attack of acute cholecystitis
may resemble angina by producing substernal discomfort that responds to
nitrates, which reduce cystic duct spasm. On rare occasions, pancreatitis or peptic ulcer disease
produces substernal chest pain. Even a patient with gaseous distention
of the bowel in the area of the splenic flexure may complain of
precordial discomfort.
Psychiatric Causes
Dramatic chest pain presentations are common among
patients with underlying psychopathology. In addition to presentations
that may be clinically indistinguishable from angina, patients with anxiety or depression
often describe feelings of chest heaviness or tightness that can last
for hours to days, unrelated to exertion and unrelieved by rest. In
patients with anxiety disorders, this sensation may be accompanied by a
feeling of inability to take in a deep breath. When there is associated
hyperventilation, the resulting hypocapnia leaves the patient
lightheaded and the extremities tingling.
Cardiac neurosis may lead to
reports of chest pain mimicking angina. At other times, the patient
misinterprets a noncardiac chest sensation. Patients with a personality disorder and somatization
may describe almost any form of chest pain, including some suggestive
of angina. A lifelong pattern of multiple refractory bodily complaints
is characteristic (see Chapter 230). Malingering
represents a conscious effort to feign illness for secondary gain. The
hallmark is inconsistency of the story. Although other forms of
psychogenic chest pain may bring secondary benefits to the patient,
there is no premeditated attempt to deceive.
Depression and panic disorder
can be sources of atypical chest pain. Patients with such conditions
tend to be younger, more often female, more apt to have a higher number
of accompanying autonomic symptoms, more bothered by phobias, and more
likely to describe an atypical form of chest pain than those with chest
pain and a positive coronary angiogram.
DIFFERENTIAL DIAGNOSIS
The differential diagnosis of chest pain can be organized along anatomic lines, as outlined in Table 20.1.
Must-not-miss diagnoses include CHD, critical aortic stenosis, aortic
dissection, pneumothorax, cholecystitis, pericarditis, and pleuritis
from pneumonia, embolization, or cancer. Underdiagnosis and delay in
diagnosis of CHD is a problem in women younger than the age of 60 years
because their clinical presentations may be atypical or ignored. A high
index of suspicion is warranted in such persons, especially when there
is preexisting diabetes, a major risk factor for early development of
CHD in women. Underdiagnosis of CHD is also a problem for African
Americans presenting with chest pain.
Table 20.1. Differential Diagnosis of Chest Pain | |
|---|---|
|
WORKUP (13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39)
The first priority is to determine the need for emergent hospitalization. This requires estimating the likelihood of myocardial ischemia, aortic dissection, and pulmonary
embolization, conditions that place the patient at high risk for a potentially life-threatening complication. Timeliness of the assessment is of the utmost importance because outcome often depends on prompt intervention. In many instances, the decision to immediately hospitalize will need to be made by history alone, often on the basis of a telephone call by the patient. Delaying hospitalization is a major cause of poor outcome. Delays tend to be particularly prevalent among women and African Americans. Postmenopausal women presenting with ischemic chest pain are one-third less likely to be admitted to the hospital than men with the same degree of cardiac risk.
P.129
embolization, conditions that place the patient at high risk for a potentially life-threatening complication. Timeliness of the assessment is of the utmost importance because outcome often depends on prompt intervention. In many instances, the decision to immediately hospitalize will need to be made by history alone, often on the basis of a telephone call by the patient. Delaying hospitalization is a major cause of poor outcome. Delays tend to be particularly prevalent among women and African Americans. Postmenopausal women presenting with ischemic chest pain are one-third less likely to be admitted to the hospital than men with the same degree of cardiac risk.
An efficient triage strategy for initial decision making
is one that determines and stratifies chest pain risk on the basis of
presenting history supplemented by physical examination, ECG, and on
occasion chest x-ray (if available and deemed necessary). Predictions
of risk based on this approach have proven to be extremely powerful and
can be used to determine who will benefit most from prompt
hospitalization and aggressive testing. Extensively testing
low-probability patients for ischemia, dissection, and embolization is
not only wasteful, it also leads to a high false-alarm rate with its
attendant adverse consequences. Although there may be psychological
pressure to proceed to elaborate diagnostic studies in all patients
with chest pain, only those with at least a modest pretest probability
(e.g., at least 20%) of serious pathology benefit from such testing
(see Chapter 2). The provision of meaningful
reassurance does not usually require exhaustive testing, but it is
aided by a careful clinical assessment (see the section Patient Education and Indications for Referral).
History
Estimating Probability of Coronary Disease by History
A careful chest pain description is critical. The
prevalence of angiographically confirmed CHD approaches 90% in persons
with a classic story for angina (see prior discussion). Prevalence
declines to less than 15% in those coming to catheterization who have
nonanginal chest pain. In the Framingham Study, patients presenting
with new onset of definite angina had a relative risk of a coronary
event over 2 years of 3.7 for men and 5.9 for women. Relative risk for
those with possible angina fell to 3.0 for men and 2.9 for women, and
fell to 1.3 for men and 0.8 for women with nonanginal chest pain.
Among the features of the chest pain description with the greatest discriminant value are its timing in relation to precipitating and alleviating factors.
Quality, location, radiation, and intensity of pain are notoriously
nonspecific. Precordial pain radiating down the left arm can occur with
almost any cause of chest pain. A common pitfall in taking the history
is to provide classic descriptions of chest pain to the patient who
cannot give a quick crisp account of his or her chest pain. Under the
duress of the physician's interrogation, the patient may agree to one
of these neat descriptions, leading to a false-positive diagnosis. The
initial vagueness may have been more useful.
Past medical history is reviewed for major cardiac risk factors
(e.g., hypertension, diabetes, smoking, hypercholesterolemia, and
obesity). Inquiry into cocaine use is essential, especially in young
persons presenting with ischemia-like chest pain. Family history is checked for premature coronary disease. Note is also taken of patient age and gender.
Postmenopausal women with coronary artery disease have an especially
poor prognosis. Awareness of the potential biasing effects of gender
and race on clinical thinking is critical to avoiding them. Race and
gender have been found to be independent determinants of how patients
are assessed. Because there are no differences in acute chest pain
presentations among whites and African Americans, the chest pain
presentation can be evaluated without need to adjust for race. However,
differences in CHD presentation between sexes need to be kept in mind.
Women are often underdiagnosed, in part because the story may be
atypical or vague (e.g., exertional fatigue, arm tingling, nausea,
shortness of breath). A high index of suspicion for CHD is required in
women younger than the age of 60 years presenting with chest pain,
especially if they have preexisting diabetes (which negates the
beneficial effects of estrogen).
Checking for an Acute Coronary Syndrome
Any person with angina-like pain and cardiac risk factors should be asked if the pain is of greater than 20 minutes' duration, crescendo in pattern, now occurring at rest or at night and with exertion, of new onset (especially if severe enough to limit activity), or associated with dyspnea.
An answer of “yes” to any of these questions on telephone triage should
prompt immediate hospitalization by ambulance because risk of an acute
coronary syndrome is sufficiently high, and time is of the utmost
importance. If evaluated in the office, the patient needs only a brief
physical examination and ECG (only if immediately available; see later
discussion) to complete the initial assessment.
Considering Noncardiac Etiologies in Persons with Angina-Like Pain
Some elements of the history suggestive of coronary disease are also important for the other causes they suggest. Pain brought on by exertion and relieved by rest
is certainly indicative of angina, but psychogenic disease and even
esophageal spasm may behave in similar fashion, necessitating at least
consideration of these alternative diagnoses. A check for anxiety,
depression, panic episodes, headache, nervousness, weakness, fatigue,
and lifelong history of multiple bodily complaints may help identify a
psychogenic origin. Heartburn, dysphagia, symptoms associated with
meals, and an absence of CHD risk factors raise the possibility of
esophageal disease. Recurrent episodes that last hours to days provide
further evidence of a noncardiac origin. Prompt response to nitroglycerin
is another characteristic feature of CHD, but esophageal spasm,
coronary microvascular disease, cystic duct spasm, and even some
psychogenic etiologies may also respond to nitrates. Chest pain brought on by eating
may be due to angina, but in the absence of other risk factors for CHD,
one needs to consider gastroesophageal or pancreaticobiliary pathology.
As noted earlier, response to nitroglycerin is not necessarily helpful
in differentiation.
Checking for Aortic Dissection
Attention to onset, radiation, and associated symptoms
is critical for early identification of aortic dissection. One checks
for sudden onset; maximum intensity from the start (often described as a catastrophic presentation of tearing or searing pain); radiation into the interscapular region, jaw, neck, or down into the lower back or legs; and any accompanying new neurologic deficit or syncopal episode.
Such a presentation should strongly suggest acute dissection of the
thoracic aorta and warrant consideration of immediate hospitalization.
Past medical history is reviewed for atherosclerotic risk
factors, existing vascular disease, blunt trauma to the chest, and connective tissue disease (e.g., Marfan's syndrome). Although many less serious conditions can cause chest pain that radiates into the back (esophageal pathology being the most common), the seriousness of aortic dissection mandates a careful review of the history and risk factors. Seventy percent to 90% of cases exhibit the characteristically dramatic presentation, but 10% to 30% are more subtle in their manifestations, necessitating a high index of suspicion for the condition.
P.130
factors, existing vascular disease, blunt trauma to the chest, and connective tissue disease (e.g., Marfan's syndrome). Although many less serious conditions can cause chest pain that radiates into the back (esophageal pathology being the most common), the seriousness of aortic dissection mandates a careful review of the history and risk factors. Seventy percent to 90% of cases exhibit the characteristically dramatic presentation, but 10% to 30% are more subtle in their manifestations, necessitating a high index of suspicion for the condition.
Evaluating Pleuritic Pain
Pain worsened by deep inspiration or cough
is a hallmark of pleural irritation, but such pain is also suggestive
of pericarditis and chest wall pathology. Even aortic dissection may
cause pain worsened by movement due to respiration. Focal chest wall tenderness
worsened by movement quickly narrows the differential to a chest wall
origin. In the absence of focal chest wall pain, one needs to search
promptly for evidence of intrathoracic pathology. Inquiry is needed
into fever, cough, sputum production, tuberculosis exposure,
hemoptysis, smoking, HIV exposure or high-risk behavior, unilateral leg
edema, calf tenderness, shortness of breath, past history of
embolization, recent orthopedic surgery, and oral contraceptive use.
Pneumothorax should come to mind when pleuritic pain is sudden in onset and accompanied by dyspnea
in a young patient with a previous history of pneumothorax or when the
patient has long-standing bullous emphysema. Precordial-catch syndrome
is suggested by brief, self-limited episodes in an otherwise healthy
young person. Pleuritic pain worsened by turning but relieved by sitting up and leaning forward is indicative of pericarditis, which can be further assessed by physical examination. Attention to the context
of the patient's pleuritic pain often suggests the diagnosis. Onset in
a person with known metastatic cancer may be due to a pathologic
fracture, pleural metastasis, or pulmonary embolization. The same pain
in an otherwise healthy young person with new onset of a dry cough,
low-grade fever, and myalgias is consistent with viral-induced
pleurodynia and muscle soreness from coughing.
Estimating the pretest probability of pulmonary embolization
is a critical task in the evaluation of the chest pain patient who also
presents with acute shortness of breath. Studies using multivariate
analysis have identified key elements of the history that make
important independent contributions to the probability of pulmonary
embolization (Table 20.2). These include sudden
onset or sudden worsening of dyspnea, pleuritic or noncardiac
(nonretrosternal) chest pain, hemoptysis, asymmetric leg edema or pain,
and presence of risk factors for thromboembolism (history of prior
thromboembolic disease; strongly positive family history of
thromboembolism; recent immobilization; recent surgery, particularly
orthopedic surgery; concurrent malignancy; lower extremity paralysis).
High fever, preexisting cardiopulmonary disease, and evidence for an
alternative etiology reduce pretest probability. Determination of the
pretest probability helps in decisions regarding subsequent workup for
thromboembolism (see later discussion).
Table 20.2. Estimating Clinical Pretest Probability of Pulmonary Embolization | ||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||
Physical Examination
In cases of acute chest pain where the history is very
suggestive of unstable angina, pulmonary embolization, or aortic
dissection, the decision to immediately hospitalize can be made on the
basis of history alone. There is no reason to delay admission. In the
setting of less acute chest pain or a more ambiguous story, the
physical examination can provide important evidence pertinent to the
differential diagnosis and the assessment of risk.
General appearance and vital signs
can be telling. Tachypnea and tachycardia in a person with acute
pleuritic pain are suggestive of pulmonary embolization, whereas an
anxious, sighing, hyperventilating individual who complains of constant
chest tightness is more likely to be suffering from an anxiety
disorder. Blood pressure is noted for elevation (an important risk
factor for cardiovascular disease), for hypotension (a bad prognostic
sign in acute coronary syndromes and pulmonary embolization), and for
asymmetry in the arms (a sign of thoracic aortic dissection).
The skin is checked for cyanosis, herpetic rash, pallor, jaundice, and xanthomata. Examination of the fundi may provide evidence of atherosclerotic, diabetic, or hypertensive disease. In the neck, the carotid pulse
is palpated for diminution or loss when considering thoracic aortic
dissection, and delay in upstroke when assessing for hemodynamically
significant aortic stenosis. In addition, jugular venous pressure
is determined; jugular venous distention may be noted in the setting of
pump failure due to acute ischemia, acute pulmonary hypertension due to
severe pulmonary embolization, and cardiac tamponade associated with
pericarditis.
The chest wall is examined carefully in the person reporting “pleuritic” pain, beginning with inspection for signs of trauma and the rash of herpes zoster,
and palpation for swelling and focal tenderness. If pain is elicited,
it is important to be sure the pain on palpation is identical to the
patient's presenting complaint.
One should next listen to the lungs for a pleural friction rub during inspiration and expiration and note any signs of consolidation or effusion. Hyperresonance, absent breath sounds, and tracheal deviation
suggest a significant pneumothorax that requires immediate attention,
especially if the patient is also tachycardic, hypotensive, and
cyanotic. Checking for rales (crackles) in
the patient with angina assesses the possibility of ischemic left
ventricular dysfunction (another sign of pump failure and poor
prognosis).
On examination of the heart,
the left ventricular impulse is observed and noted for signs of
hypertrophy (indicative of significant aortic stenosis, long-standing
hypertension, or a hypertrophic cardiomyopathy). Signs of ischemic
myocardial dysfunction, such as loss of physiologic splitting of the second heart sound, development of an S4, and presence of an S3,
are sought; they may be transient, occurring only during chest pain,
but their presence suggests considerable myocardium at risk. Listening
for the systolic ejection murmur of aortic stenosis should not be forgotten in the patient with angina nor should the systolic regurgitant murmur
of mitral regurgitation due to papillary muscle ischemia. Although the
relation between mitral valve prolapse and chest pain is questionable,
checking for its hallmark midsystolic click and late-systolic murmur
are indicated in persons with atypical chest pain. If the chest pain is
pleuritic, then carefully listening for the two- to three-component precordial friction rub
of pericarditis is indicated, as is checking for an accentuated
pulmonic component of the second heart sound and a new tricuspid
regurgitant murmur indicative of acute pulmonary hypertension from
severe pulmonary embolization.
The abdomen is checked for
epigastric and right upper quadrant tenderness (especially when
evidence suggests gastric or hepatobiliary disease) and for masses
(particularly an
abdominal aneurysm in the context of suspected dissection). The legs require careful examination for unilateral edema and other signs of phlebitis (see Chapter 22), a potential source of pulmonary embolization in the person presenting with pleuritic pain. All peripheral pulses are checked, noting any absences and evidence of acute ischemia, which might occur with an aortic dissection. The spine is palpated for areas of tenderness along the cervical and thoracic segments, and the neurologic examination needs to include a check for new focal deficits, another possible clue to dissection.
P.131
abdominal aneurysm in the context of suspected dissection). The legs require careful examination for unilateral edema and other signs of phlebitis (see Chapter 22), a potential source of pulmonary embolization in the person presenting with pleuritic pain. All peripheral pulses are checked, noting any absences and evidence of acute ischemia, which might occur with an aortic dissection. The spine is palpated for areas of tenderness along the cervical and thoracic segments, and the neurologic examination needs to include a check for new focal deficits, another possible clue to dissection.
Laboratory Studies
Outpatient Testing versus Immediate Transport to the Emergency Room
Although laboratory studies can be helpful, one should
not delay the decision to urgently transport the patient to the nearest
emergency room (ER) when the clinical picture suggests unstable angina,
acute pulmonary embolization, aortic dissection, or large pneumothorax.
In fact, any delay brought about by taking time to obtain
“nice-to-have” but not essential diagnostic studies in the office could
be life threatening. Only those studies that are essential to immediate
decision making should be considered. Even an ECG may be superfluous in
the face of a good story for unstable angina because a normal study
would not change the decision to get the patient quickly to the nearest
ER. Obviously, an ECG is essential to ER decision making and should be
obtained within minutes of the patient's arrival, but initial triage
decisions made outside the hospital will be based largely on the
patient's story.
Test Selection
An accurate estimation of pretest probability is
essential to the workup of chest pain, especially as it pertains to
effective test selection and interpretation (see Chapter 2).
The number of “must-not-miss” conditions is large, and there is
considerable pressure on the clinician to test extensively and
exhaustively. The risks inherent in ignoring pretest probability are
poor test selection and misinterpretation of test results (see Chapter 2)
leading to diagnostic errors and potentially harmful management
decisions. Accurate estimation of pretest probability is afforded by
focusing on the key history and physical examination findings noted
earlier.
Another potential source of error in test selection is
diagnostic bias associated with the patient's sociodemographic status
(e.g., race, gender, social group). Physician responses to
the chest pain presentation are sometimes affected by such factors and need to be kept in mind. For example, African American women are 60% less likely to be referred for cardiac catheterization than are white men with the same pretest risk of coronary disease. Women and non-whites are less likely to be hospitalized when presenting with acute ischemia.
P.132
the chest pain presentation are sometimes affected by such factors and need to be kept in mind. For example, African American women are 60% less likely to be referred for cardiac catheterization than are white men with the same pretest risk of coronary disease. Women and non-whites are less likely to be hospitalized when presenting with acute ischemia.
Testing for Suspected Coronary Artery Disease
The approach to laboratory workup for coronary disease
is determined by the condition's pretest probability and an
appreciation for the sensitivity, specificity, and cost of available
tests (see Table 20.3).
Table 20.3. Sensitivity, Specificity, and Costs of Stress Tests for the Detection of Coronary Disease | ||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||
High Pretest Probability: Unstable Angina Presentation
As noted earlier, the very high probability patient (multiple CHD risk factors, history very suggestive of unstable angina) requires immediate hospitalization without delay. Additional outpatient testing is unwarranted.
Electrocardiogram
Once the patient arrives in the emergency room, a resting ECG
is critical to initial decision making, used in most critical pathways
for optimal emergency room triage of chest pain patients. Any ST- or
T-wave changes indicative of ischemia are an indication for prompt
consideration of therapy, be it thrombolytic or revascularization (see
later discussion). Proceeding directly to coronary angiography is more cost-effective than stress testing in such very-high-probability patients with ECG changes.
A normal ECG is not by itself sufficient evidence for
discharge of the high-probability patient, but a shortened period of
observation (e.g., 6 hours) followed by stress testing
is being applied in some centers to safely minimize length of stay for
patients with no ECG changes and normal serum levels of markers of
myocardial injury (e.g., creatine kinase and troponins).
Exclusionary criteria for early stress testing include ongoing chest
pain, signs of heart failure, and elevations in creatine phosphokinase
or troponin levels.
Creatine Kinase MB Isozyme and Cardiac Troponins T and I
Creatine kinase MB isozyme (CK-MB) and cardiac troponins
T and I are macromolecular markers of myocardial injury that facilitate
the emergent evaluation of chest pain, especially in patients with high
or intermediate pretest probability. In cases of infarction, the CK-MB
level usually begins to rise within 4 hours of myocardial injury,
almost always turns positive by 12 to 24 hours, and then begins to
decline unless there is reinfarction. Although sensitivity is high,
specificity is not as good because other causes of myocardial injury
will also cause CK-MB elevation. Cardiac troponins T and I provide
enhanced sensitivity and specificity for detection of myocardial injury
and are more predictive of future coronary events. Levels rise at the
same rate as CK-MB but persist for days, making the test less useful
for detection of reinfarction. Rapid assays for troponins T and I show
respective sensitivities at 6 hours of 94% and 100%, and specificities
of 89% and 83% in persons with normal electrocardiograms. Although
troponin T is excreted renally, renal insufficiency does not impair its
predictive value. It is common practice to obtain CK-MB and troponin
levels simultaneously and serially, although some authorities suggest
reserving troponin determinations for situations where clinical
suspicion persists despite normal ECG and CK-MB levels.
Experimental Markers
The search continues for additional markers that can
independently contribute to the coronary risk assessment, particularly
in the ER setting, where the goal is not to discharge a patient at
risk. The C-reactive protein, erythrocyte sedimentation rate, and myeloperoxidase
(which derives from activated leukocytes, believed pathophysiologically
important in vulnerable plaques) have shown some promise in providing
additional predictive value in persons presenting with ischemic chest
pain, a nondiagnostic ECG, and normal levels of CK-MB and troponins.
Further confirmation of these findings is required before they can be
recommended for routine use in the workup of chest pain.
High Pretest Probability: Stable Angina Presentation
The patient with CHD risk factors and a clinical presentation that is classic for stable angina
does not need any testing to establish the diagnosis of coronary
disease (the pretest probability is already in excess of 90%). Instead,
the role of testing is to estimate prognosis and stratify risk, which inform selection of initial treatment (i.e., revascularization vs. medical therapy; see Chapter 30). A well-recognized prognostic determinant is amount of myocardium at risk, which can be assessed by electrocardiographic, radionuclide, or echocardiographic stress testing (see later discussion and Chapters 30 and 36).
A less well-recognized risk factor is prior silent myocardial
infarction, which may be detected by resting ECG (new Q waves or new
ST- or T-wave changes since last ECG) or cardiac ultrasound
(segmental wall akinesis). Postinfarction angina is a sign of
potentially severe coronary disease, especially in persons with prior
silent infarction (16-fold increase in mortality risk).
Intermediate Pretest Probability: Story Suggestive of Unstable Angina
The initial approach to this patient is the same as for
the high-probability patient (i.e., immediate emergency room
assessment; see prior discussion).
Intermediate Pretest Probability: Story Suggestive of Stable Angina
Patients in this category typically have a single
cardiac risk factor and give a history of atypical chest pain that is
noncrescendo in pattern. In such persons, a resting ECG may be helpful, but nonemergent stress testing is needed if the resting ECG is nondiagnostic. Several modalities are available for stress testing.
Stress Testing
Electrocardiographic stress testing (see also Chapter 36)
provides excellent sensitivity and specificity for detection of
high-risk coronary disease (i.e., left-main, left-main-equivalent, or
three-vessel disease) at low cost, but false positives occur in younger
women, sensitivity is lower in the setting of less serious forms of
coronary disease, and the test cannot be interpreted adequately unless
the resting ECG is normal. Stress echocardiography
can be performed by bicycle exercise or dobutamine stimulation; cost is
somewhat higher than with ECG, but sensitivity and specificity are
better (see Chapter 36). Radionuclide imaging with thallium or technetium sestamibi uses planar scanning and, more recently, single-photon emission computed tomography
(SPECT), which enhances sensitivity and specificity by providing
three-dimensional views. It can be performed by treadmill exercising or
adenosine injection, which causes a steal phenomenon and enhances
differences in uptake. Increased sensitivity is achieved with
radionuclide imaging but at increased cost. Positron emission tomography
(PET) is the most sensitive and specific test for coronary
insufficiency, but it is also the most expensive and is still limited
in availability. Overall, sensitivity for detection of coronary disease
by stress test ranges from 0.68 for ECG to 0.91 for PET, but all
stress tests are very sensitive for detection of left-main, left-main-equivalent, and three-vessel coronary disease, with sensitivities ranging from 0.86 for ECG stress testing to 0.98 for SPECT. Electrocardiographic stress testing is by far the lowest in cost but is also slightly less sensitive and specific than other modalities. Cost-effectiveness analyses identify ECG and echocardiographic stress testing as the most cost-effective for diagnosis of coronary disease in chest pain patients with an intermediate pretest probability of coronary disease. SPECT may also be cost-effective in settings where its cost is lower than average. The increased sensitivity and specificity associated with PET scanning are insufficient to overcome its very high cost (see Chapter 36).
P.133
stress tests are very sensitive for detection of left-main, left-main-equivalent, and three-vessel coronary disease, with sensitivities ranging from 0.86 for ECG stress testing to 0.98 for SPECT. Electrocardiographic stress testing is by far the lowest in cost but is also slightly less sensitive and specific than other modalities. Cost-effectiveness analyses identify ECG and echocardiographic stress testing as the most cost-effective for diagnosis of coronary disease in chest pain patients with an intermediate pretest probability of coronary disease. SPECT may also be cost-effective in settings where its cost is lower than average. The increased sensitivity and specificity associated with PET scanning are insufficient to overcome its very high cost (see Chapter 36).
Ambulatory Electrocardiographic Monitoring
Ambulatory electrocardiographic monitoring is not recommended because specificity is low and the false-positive rate is high.
Computed Tomography of the Coronary Arteries
Computed tomography (CT) of the coronary arteries has
been proposed as an adjunctive means of assessing CHD risk, but its
role remains to be precisely determined. The test is based on the
ability of CT to detect and quantify coronary artery calcification,
producing a coronary artery calcium score
(CACS) that correlates with degree of atherosclerosis. In studies of
sensitivity and specificity in symptomatic persons undergoing coronary
angiography, CT performs about the same as stress testing in predicting
coronary disease. However, unlike stress testing, it does not provide
anatomic or physiologic information. When used in asymptomatic persons
in conjunction with the Framingham risk score
(FRS)—which utilizes age, gender, blood pressure, total and
high-density-lipoprotein cholesterols, glucose, and smoking history to
assess CHD risk—the CACS only enhances CHD risk prediction in those
judged at intermediate risk by the FRS (i.e., FRS between 10 and 20),
providing a 3% to 9% increase in predicted 10-year CHD risk.
Unfortunately, the CT cannot rule out presence of
coronary disease; its false-negative rate is about 4% in asymptomatic
persons, which is due to presence of “soft” atherosclerotic plaque not
detectable by CT. Such soft plaque is less stable than calcified plaque
and more likely to rupture; it is the very lesion most likely to
present as unstable angina. Thus, a “negative” CT scan could be
misleading.
Low Pretest Probability
The patient with clearly nonanginal chest pain, no CHD
risk factors, and a normal cardiac examination has such a low
probability of CHD that testing is likely to generate only negative or
false-positive results and excessive medical bills. Even a CHD test
with high sensitivity and specificity will perform poorly and produce
an excessive proportion of false-positive results if applied to a
person with a very low pretest probability of CHD (see Chapters 2 and 36).
Occasionally, a resting ECG
is obtained to reassure the very anxious low-risk patient. Performing
this low-cost test has been found to speed resumption of normal
activity in the overly concerned patient without greatly increasing
cost. This approach should not be used unless there is strong patient
need for some “objective” reassurance and the patient is warned
beforehand that a positive result is most likely to be a false
positive. As lay knowledge of testing modalities for coronary disease
increases, requests among low-risk patients for stress testing are
likely to escalate but should be rebuffed because such testing will
only increase cost without improving diagnostic accuracy.
Testing for Suspected Esophageal Disease
The obvious case of esophageal disease (retrosternal
burning, difficulty swallowing) requires no testing unless symptoms are
refractory, in which case a search for malignancy is indicated (see Chapter 60).
More problematic diagnostically is the patient with angina-like pain, a
normal cardiac evaluation (including angiography), and no esophageal
symptoms. As many as 20% of patients with pain suspicious enough to
warrant coronary angiography have been shown to have esophageal
disorders. A convincing diagnosis of esophageal disease might save the
patient a cardiac catheterization.
Many provocative tests and esophageal function studies are available, including manometry, 24-hour pH monitoring, edrophonium provocation, and acid perfusion (see Chapter 60
for details). Although initial reports promoted their usefulness in the
evaluation of chest pain, more carefully controlled study has failed to
detect differences in esophageal function, either between periods of
pain and no pain or between asymptomatic, healthy control subjects and
patients. In about 25% of patients, edrophonium chloride or acid
provocation will trigger pain in patients and not in control subjects,
but there is no difference in motor response. Either motor dysfunction
is not the etiology in most esophageal cases or the tests of motor
function are not very good. In either case, there seems to be little
rationale for their routine application. Recent reports suggest that
measuring the response to esophageal distention
better differentiates patients from control subjects. Distention
testing is still a research tool and not available for clinical use,
but the future trend in diagnosis of esophageal chest pain is likely to
be toward identification of altered nociception and a heightened
sensory response.
Testing for Suspected Pulmonary Embolization
Just as for coronary disease, workup for suspected
pulmonary embolization benefits from assessment of pretest probability,
which recent research has shown can be readily and reliably determined
clinically. When combined with selective noninvasive testing, this
probabilistic approach can enhance diagnostic accuracy and greatly
reduce the need for radionuclide imaging, angiography, and unnecessary
anticoagulation.
Determining Pretest Probability
Rapid, accurate classification of patients into
categories of low, intermediate, and high pretest probability can be
achieved by attention to key clinical findings and a few readily
available tests such as chest x-ray, ECG, and pulse oximentry (see the
Wells prediction rule, Table 20.2).
Prospective validation testing of such rules finds 3-month rates of
pulmonary embolism of approximately 1.5% to 3.5% among patients
classified clinically as low risk, compared with 16% to 28% for those
categorized as intermediate risk, and 38% to 78% for those labeled high
risk.
Of the screening laboratory determinations that are
incorporated into most clinical prediction rules (i.e., ECG, chest
x-ray, pulse oximetry), no single abnormal test result is diagnostic
nor is a normal finding sufficient to rule out embolization, but the
results of these simple tests can contribute independently to the
probability estimate.
Chest X-Ray
Chest x-ray findings are usually nonspecific, but fewer
than 15% of patients with proven pulmonary embolization have normal
chest x-rays. The most common findings are unilateral pleural effusion
and atelectasis. More specific but much less common radiologic
manifestations include abrupt
amputation of a hilar artery and a wedge-shaped pleural-based infiltrate (“Hampton's hump”); both are seen in blockade of a large vessel.
P.134
amputation of a hilar artery and a wedge-shaped pleural-based infiltrate (“Hampton's hump”); both are seen in blockade of a large vessel.
ECG
ECG changes are also common but usually nonspecific;
minor abnormalities in ST and T waves are noted in upward of 70% of
cases with no history of prior cardiovascular disease. More suggestive
are findings associated with acute right heart strain, including the
classic S1Q3T3 pattern, new
right-bundle-branch block, and inverted T waves in leads V1 to V4,
believed due to posterior ischemia from compression of the right
coronary artery in the setting of right ventricular overload.
Arterial Blood Gases and Pulse Oximetry
Arterial blood gases and pulse oximetry
readings can be helpful, especially when hypoxemia is noted, but a
normal oxygen level does not rule out embolization. Hypoxemia that
requires substantial oxygen supplementation (>40% fraction of
inspired oxygen) is suggestive of significant embolization.
Low Pretest Probability (Fig. 20-1)
When pretest probability is low yet clinical suspicion remains, a D-dimer
determination is the test of choice. A negative result (<500 ng/dL)
reduces the posttest probability to less than 0.5% (negative predictive
value >99.5%) and obviates the need for further testing. Because
D-dimer lacks specificity, confirmatory testing (ventilation/perfusion
scan, venous ultrasound; see later discussion) is needed when a D-dimer
result is “positive” (i.e., >500 ng/dL).
 |
Figure 20-1.
Diagnostic algorithm for initial evaluation of patients with suspected
pulmonary embolism. Plus and minus signs indicate positive and negative
test results, respectively. DVT, deep vein thrombosis; PE, pulmonary
embolism; VQ, ventilation/perfusion lung scan. (From Wells PS, Anderson
DR, Rodger M, et al. Emergency diagnosis of pulmonary embolism. Ann
Intern Med 2001;135:100, with permission.) |
D-Dimer
D-Dimer is a degradation product of cross-linked fibrin.
Its formation occurs in the context of ongoing thrombosis and
fibrinolysis, making its measurement a potentially sensitive indicator
of active venous thrombosis and thromboembolization. Sensitivity in the
setting of pulmonary embolism ranges from 85% to 100%, depending on the
assay used. Specificity is not high (40% to 68%) and declines with
advancing age because D-dimer acts as an acute-phase reactant,
increasing in response to any event that induces fibrinolysis, such as
inflammation, cancer, or trauma.
There are two widely available assays. The enzyme-linked
immunosorbent assay is the more sensitive, but less specific, and must
be performed in the laboratory, taking about 3 to 4 hours. A rapid
bedside whole-blood assay (SimpliRed) is slightly less sensitive, but
more specific, cutting down on the number of false positives and more
useful in settings where the principal task is to rule out
embolization. Of concern are reports of unsatisfactory performance of
the whole-blood assay in cancer patients, where a negative result has
not been found to reliably exclude deep vein thrombosis.
These test characteristics make the D-dimer
determination most useful for rapidly ruling out pulmonary embolization
in patients with a low pretest probability.
If D-dimer testing is not available, ventilation/perfusion (V/Q) lung scan
can be obtained instead. A normal or nearly normal scan in a patient
with a low pretest clinical probability of embolization effectively
rules out the diagnosis (negative predictive value, 96%). If the V/Q
scan is not negative, additional testing is warranted (see intermediate pretest probability).
Intermediate Pretest Probability (Fig. 20-1)
Testing should start with a V/Q scan.
A normal V/Q scan rules out the diagnosis in persons with an
intermediate pretest probability; a positive or high probability scan
confirms the diagnosis. When an indeterminate or nonhigh-probability
V/Q scan is encountered in a patient with intermediate pretest
probabiity, venous ultrasound and D-dimer
testing can help refine the posttest probability assessment. If the
ultrasound is negative and D-dimer is negative, then embolism is ruled
out and no further testing is indicated. If D-dimer is not available or
is positive, then serial ultrasound testing over the next week to 10
days is a reasonable way to ensure absence of deep vein thrombosis and
low risk of embolization.
Ventilation/Perfusion Lung Scanning
This test is widely used in patients with suspected
embolization. As demonstrated in the landmark Prospective Investigation
of Pulmonary Embolism Diagnosis (PIOPED), a normal or nearly normal
scan in a patient with a low pretest clinical probability of
embolization effectively rules out the diagnosis (negative predictive
value, 96%). Likewise, a “high-probability” scan rules in the diagnosis
if the patient has a high pretest probability (positive predictive
value, 96%). Ventilation/perfusion scanning by itself can reduce the
need for angiography in the 40% of cases where the pretest
probabilities are strong and the scan results are concordant with these
probabilities.
Unfortunately, the V/Q scans of a large proportion of
patients produce either equivocal results or findings that are
discordant with pretest probabilities. High false-positive rates are
especially prevalent in those with preexisting lung disease and in the
elderly. The addition of ventilation scanning has not solved the
problem of false positives.
In the landmark PIOPED study, 40% of patients with a
“low-probability” scan and a high pretest probability proved to have
pulmonary embolization by angiography. Forty-four percent of patients
with a high-probability scan and a low pretest probability had no
embolization by angiography. The largest single group of patients were
those with scans of “intermediate” probability, in whom embolization
could be neither ruled in nor ruled out. These high rates of false
positives and false negatives among selected groups of patients and the
large number of indeterminate scans reflect shortcomings in the
sensitivity and the specificity of the test. Assuming any abnormality
on the scan to be a positive test, sensitivity would still be only
about 90% and specificity just 10%. V/Q scanning remains widely used as
a screening test for pulmonary embolism, but additional testing is
needed in many patients. The test appears to be safe in pregnant women, particularly after the first trimester.
Compression Venous Ultrasound of the Lower Extremities
Compression venous ultrasound of the lower extremities
provides some of the specificity lacking in V/Q scanning and D-dimer
testing. The test is very specific and sensitive for detection of
symptomatic deep vein thrombosis in the proximal veins (sensitivity
95%, specificity 96%). Application of serial ultrasound has been found
especially useful in patients with a moderate pretest probability and
indeterminate V/Q scan findings. In this setting, a normal serial
ultrasound rules out the diagnosis or at least obviates the need for
immediate anticoagulation or angiography; a positive study rules it in,
again avoiding the need for angiography.
Because almost all pulmonary thromboemboli originate
from the deep proximal veins of the legs, venous ultrasound has been
proposed as a means of improving noninvasive diagnosis of embolization.
However, the sensitivity of a single study is limited because up to 40%
of patients with embolism have no detectable clot remaining in the
proximal venous system after initial embolization. To overcome this
limitation, serial studies are done over 1 to 2 weeks. In most
instances, a clot needs to reform in the leg before reembolization.
Thus, sensitivity of the test for risk
of embolization can be greatly enhanced by performing serial ultrasound studies. Even if the diagnosis of embolism cannot be confirmed initially, the risk of reembolization can be estimated by serial study. Using serial ultrasound in conjunction with V/Q scanning and pretest probability for embolism, Canadian investigators made an accurate diagnosis in more than 96% of cases for suspected embolism; the false-negative rate was only 0.6%, with angiography needed in only 3.7% of cases.
P.135
of embolization can be greatly enhanced by performing serial ultrasound studies. Even if the diagnosis of embolism cannot be confirmed initially, the risk of reembolization can be estimated by serial study. Using serial ultrasound in conjunction with V/Q scanning and pretest probability for embolism, Canadian investigators made an accurate diagnosis in more than 96% of cases for suspected embolism; the false-negative rate was only 0.6%, with angiography needed in only 3.7% of cases.
High Pretest Probability (Fig. 20-1)
Such patients should be admitted and promptly heparinized pending the results of testing. V/Q scanning is the initial test of choice. A “high-probability” V/Q scan confirms the diagnosis, demonstrating a positive
predictive value of 96% in the setting of a high pretest probability; no additional testing is required. If the V/Q scanning result is indeterminate, then the next step is to proceed to venous ultrasound. If ultrasound is positive, the diagnosis is confirmed; if negative or equivocal, then angiography should be considered, aided by a D-dimer determination. A positive D-dimer raises the pretest probability sufficiently to warrant proceeding directly to angiography (helical CT scanning has been proposed as noninvasive alternative to angiography in selected patients; see later discussion). A negative D-dimer lowers the pretest probability sufficiently to allow withholding of anticoagulation and discharge of the patient. Nonetheless, close monitoring of the D-dimer–negative patient is required, including a repeat venous ultrasound in 1 week.
P.136
predictive value of 96% in the setting of a high pretest probability; no additional testing is required. If the V/Q scanning result is indeterminate, then the next step is to proceed to venous ultrasound. If ultrasound is positive, the diagnosis is confirmed; if negative or equivocal, then angiography should be considered, aided by a D-dimer determination. A positive D-dimer raises the pretest probability sufficiently to warrant proceeding directly to angiography (helical CT scanning has been proposed as noninvasive alternative to angiography in selected patients; see later discussion). A negative D-dimer lowers the pretest probability sufficiently to allow withholding of anticoagulation and discharge of the patient. Nonetheless, close monitoring of the D-dimer–negative patient is required, including a repeat venous ultrasound in 1 week.
Pulmonary Angiography
Pulmonary angiography remains the gold standard for
diagnosis of pulmonary embolism and should be used when definitive
diagnosis is essential, urgent, and not available from noninvasive
testing. Angiography without prior scanning is urged by some experts,
especially if the risk of anticoagulant therapy is high for the patient
and a definitive diagnosis is needed before commencing treatment.
Nonetheless, rational use of noninvasive approaches to testing can
greatly reduce the need for invasive study in most instances.
Helical (Spiral) Computer Tomography
Helical (spiral) CT is designed to detect emboli by
means of a two-dimensional volumetric image of the lung. Although rapid
to perform (about 30 seconds), the test requires administration of
intravenous contrast and is expensive. Some have proposed spiral CT as
an alternative to angiography and a primary testing modality in
conjunction with venous ultrasound. Reports of sensitivity range from
64% to 100%, depending on the critieria used for diagnosis; specificity
ranges from 89% to 100%. A negative test does not rule out
embolization, particularly in subsegmental vessels where detection by
CT is more problematic. The test provides sufficent imaging to detect
other important intrathoracic pathology that might account for the
patient's symptoms. Although the helical CT test shows promise and is
already used in many emergency departments, the test lacks sufficient
evidence of comparative efficacy from prospective outcome studies to
warrant its routine use at this time. Nonetheless, prospective study
results are promising when spiral CT is used as the primary testing
modality in conjunction with venous ultrasound. The next genertion of
spiral CT scanners uses multidetector methodology rather than a single
detector, providing shorter imaging time, thinner imaging sections,
more coverage of the thorax, and, it is hoped, better detection of
emboli in subsegmental vessels.
Testing for Suspected Pulmonary Infection and Other Nonembolic Causes of Pleuritic Chest Pain
The chest x-ray is the
initial test of choice. Pneumococcal pneumonia and tuberculosis often
present with acute pleuritic chest pain and may be mistaken clinically
for pulmonary embolism. Consequently, any patient with pleuritic pain,
sputum production, and an infiltrate of chest film should also have
both Gram's and acid-fast stains
made. A pleural effusion may also be detected on chest film. Any
nonloculated pleural effusion of unknown etiology should be tapped,
Gram-stained, cultured, examined microscopically, and sent for cell
count, glucose, lactic dehydrogenase, and protein determinations (see Chapter 43). Suspicion of pneumothorax
is also an indication for a chest film, but if radiography is not
immediately available and the patient is in respiratory distress,
decompression should not be delayed.
When pericarditis is under
consideration, an ECG is essential. However, the ECG changes of early
repolarization, a harmless finding seen in young men, may closely
resemble those of acute pericarditis. The presence of concave
ST-segment elevations in both limb and precordial leads and the
presence of PR-segment depressions in the precordial leads, if they
occur in the limb leads, distinguish pericarditis from early
repolarization. Cardiac ultrasonography
may reveal a pericardial effusion. An anti-nuclear antibody, blood urea
nitrogen, and tuberculin skin test are indicated when the cause of
pericarditis is not readily evident.
Suspected Aortic Dissection
A chest x-ray, which may
demonstrate a widened mediastinum (especially in traumatic aortic
rupture), is sometimes performed for screening purposes, but if
dissection is truly suspected on clinical grounds, then emergency
admission for aortic angiography is indicated. Delaying admission to
obtain a chest film is unwise. Patients with traumatic aortic
dissection may also be well served by transesophageal ultrasonography, which is preferred to transthoracic ultrasonography, which is less sensitive for detecting disease in the aortic arch. CT with contrast
is useful when the aortogram is not available or is negative, yet
clinical suspicion remains high. Many now proceed directly to CT
because it is less invasive and can be done on shorter notice than
angiography. Magnetic resonance imaging can also detect dissection and provides another noninvasive alternative to contrast angiography.
Other Conditions
Only a few musculoskeletal disorders require chest radiography: suspected rib fractures and cervical or thoracic spine disease. If a gastrointestinal
cause is suspected, a contrast study may be in order. The ECG may show
T-wave depression in cholecystitis and pancreatitis, and may mistakenly
be interpreted as evidence of coronary disease.
The anxious patient with psychogenic
pain may find a chest radiograph and/or ECG reassuring. In most
instances, however, a thorough history and careful physical examination
combined with a detailed explanation should suffice. Repeating tests
“just to be sure” may begin to undermine the patient's confidence in
the physician's explanation and even heighten anxiety, especially if
there are repeat studies.
It is important to realize that as many as 10% to 15% of
cases remain undiagnosed, even after careful and thorough evaluation.
Nevertheless, in such instances it is still possible to rule out the
presence of an acutely serious etiology. Most patients with chest pain
that initially eludes diagnosis can be followed expectantly for the
time being.
SYMPTOMATIC RELIEF
Relief of pain must be based on an etiologic diagnosis.
To simply suppress the pain with analgesics or sedatives before a
diagnosis is made may hide important clues and endanger the patient.
However, musculoskeletal forms of chest pain may benefit from
analgesia, especially if the patient is splinting and not ventilating
adequately. When the diagnosis of costochondritis is certain, local
injection with lidocaine into the point of maximal tenderness can
provide dramatic relief. An antacid regimen or histamine2-blocker therapy in conjunction
with other antireflux and acid-reducing measures are helpful in patients with esophagitis. Nitrates and calcium-channel blockers are sometimes of benefit to patients with esophageal spasm (see Chapter 61). Patients with depression or panic disorder require specific therapy directed at the underlying psychopathology; failure to treat etiologically may result in prolonged refractory disability from the chest pain (see Chapters 226 and 227).
P.137
with other antireflux and acid-reducing measures are helpful in patients with esophagitis. Nitrates and calcium-channel blockers are sometimes of benefit to patients with esophageal spasm (see Chapter 61). Patients with depression or panic disorder require specific therapy directed at the underlying psychopathology; failure to treat etiologically may result in prolonged refractory disability from the chest pain (see Chapters 226 and 227).
PATIENT EDUCATION AND INDICATIONS FOR REFERRAL
Teaching Recognition of Acute Coronary Syndromes
The availability of effective early interventions for
acute coronary syndromes and the risk of delay necessitate that
patients be taught to recognize the symptoms and to call for help
immediately. Those with multiple cardiac risk factors or known heart
disease deserve top priority for such an educational intervention.
Knowledge of key symptoms, especially duration in excess of 20 minutes,
new rest pain, and epiphenomena of ischemia (e.g., diaphoresis, nausea,
jaw or neck pain, lightheadedness) is weak, particularly among
socioeconomically disadvantaged groups, minorities, the elderly, and
younger patients. Patients should be instructed to call for an
ambulance and not delay. Delay has been documented in up to 50% of
cases and often compromises outcome. The importance of minimizing delay
in presentation makes patient education and telephone triage important
tools. Practices should be organized to handle telephone calls related
to chest pain without delay.
Providing Explanation and Meaningful Reassurance
A detailed review of clinical findings and their meaning
is as essential to the effective evaluation of chest pain as is a
correct diagnosis. When a harmless etiology is identified, it is not
enough to dismiss the chest pain as “nothing to worry about” because
the symptom is associated with too many fears for such words to
suffice. Meaningful reassurance requires eliciting a patient's concerns
about his or her chest pain and addressing these concerns by
specifically reviewing the pertinent clinical findings. Failure to do
so may lead to unnecessary activity restrictions (e.g., fear of
engaging in sexual relations, favorite sports, or important work
activities) or trigger requests for otherwise unnecessary testing.
Patients making repeated visits, asking for referrals, or requesting
elaborate testing usually harbor unexplored concerns. The effort taken
to provide meaningful explanation is usually well worth the time in
terms of patient appreciation, cost containment, and risk management.
INDICATIONS FOR ADMISSION AND REFERRAL
As noted earlier, immediate referral to the nearest
emergency room is essential to achieving the best possible outcome for
the patient with any clinical or laboratory evidence for an acute
coronary syndrome, pulmonary embolization, aortic dissection, or large
pneumothorax. Key features include a story of unstable angina, sudden
onset of severe tearing or searing chest pain with radiation into the
back, new pleuritic chest pain in a patient with risk factors for
pulmonary embolization or pneumothorax, severe dyspnea, signs of
respiratory or hemodynamic compromise, and ischemic ECG changes with
pain. Physician errors in triage are most prevalent among patients who
are female, non-white, present with a chief complaint of dyspnea, or
have a normal or nondiagnostic ECG.
Persons who present with acute ST-segment elevation
suggestive of acute myocardial infarction require consideration of
immediate referral to a center capable of performing angioplasty
because outcomes are often better than those achieved by thrombolysis,
especially when more than 3 hours has elapsed since onset of symptoms
or time from referral to balloon can be less than 2 hours. Referral to
the nearest emergency room for thrombolysis is reasonable when time
from onset of symptoms to administration of tissue plasminogen
activator will be less than 3 hours. For the chest pain patient who
presents to the office with evidence of postinfarction angina or ECG
findings of prior silent infarction, plans should be made for prompt
hospital admission and cardiac consultation—such patients are at high
risk. Acute chest pain in the context of cocaine use is also an
indication for immediate hospitalization, even in a young person with
no other cardiac risk factors. Referral for drug counseling should not
be overlooked; it, too, is critical to a successful outcome (see Chapter 235).
Patients at risk for a life-threatening chest pain etiology should be
taught to recognize key symptoms and instructed to call 911 first if
symptoms are acutely severe, even before calling the primary care
physician.
Outpatient workup is reasonable for the patient with
low-to-intermediate probability of pulmonary embolization, provided
there are no signs of respiratory or hemodynamic compromise and the
appropriate initial testing (e.g., D-dimer determination, V/Q scan,
venous ultrasound) can be completed within 2 to 3 hours. For the
patient with pleuritic chest pain due to pneumonia, knowledge of the
clinical features predictive of a complication is necessary to
determine candidacy for admission (see Chapter 52).
The utility of referral for esophageal function testing appears
limited, but endoscopy or barium swallow should be considered if there
is refractory reflux or dysphagia. The patient with panic disorder or
depression severe enough to cause disabling chest pain symptoms will
benefit from psychiatric referral (see Chapters 226 and 227).
RECOMMENDATIONS
- Perform a triage history focusing on evidence for acute coronary syndrome, pulmonary embolization, aortic dissection, tension pneumothorax, and acute cardiopulmonary compromise.
- Arrange immediate transport to the nearest ER by ambulance for the high-risk chest pain patient. Address common mistakes that cause delay and add to risk, such as reluctance of the patient to call an ambulance or having the patient be driven by a friend, a family member, or self.
- If the initial history is not strongly indicative of one of these etiologies, then proceed to evaluate the patient in the office with a more detailed history and careful physical examination, supplemented, by ECG, chest x-ray, and pulse oximetry, if available.
- Select the appropriate strategy for subsequent testing based on the pretest probability for the conditions in question. Be aware of common biases based on patient age, gender, and race.
- In formulating the initial differential diagnosis from these data, include a pretest estimate of the “must-not-miss”P.138
etiologies (i.e., myocardial ischemia, pulmonary embolization, aortic dissection, tension pneumothorax) and base immediate decision making and the optimal locus and approach to further workup on these probability estimates.
For Suspected Coronary Artery Disease
- Low pretest probability: Explain why the chest pain is unlikely to be of cardiac origin, and consider obtaining a resting ECG if it will help reassure the patient.
- Intermediate pretest probability: If the presentation is suggestive of unstable angina, proceed to immediate hospitalization and emergency room triage protocols. If the presentation is one of “stable” (i.e., noncrescendo) chest pain, proceed on a nonemergent basis to stress testing. Obtain either exercise ECG (if the resting ECG is normal) or exercise echocardiography; both are reasonably cost effective. If other noninvasive modalities (e.g., radionuclide stress testing) are available locally at lower cost or with enhanced sensitivity and specificity, then their use may also be cost effective and should be considered.
- High pretest probability: If the presentation is very suggestive of unstable angina, proceed to immediate hospitalization and emergency room triage protocols. For those with ECG changes on arrival in the emergency room, immediate angiography without prior noninvasive testing appears to be the more cost-effective diagnostic approach. For those whose presentation is indicative of stable angina, proceed to nonemergent stress testing for determination of amount of myocardium at risk (see Chapter 36 for test selection).
- To minimize delay in hospitalization, teach the high-risk patient and family the symptoms of acute coronary syndromes, and instruct them to call 911 immediately if such symptoms occur.
- Reduce requests for unnecessary testing in low-risk patients by eliciting a patient's concerns about his or her chest pain, performing a detailed history and physical examination, obtaining an ECG, and addressing patient concerns by a careful review of clinical findings.
- Omit routine use of esophageal studies in assessment of chest pain in patients who are ruled out for coronary disease.
For Suspected Pulmonary Embolization
- Low pretest probability: Obtain a D-dimer determination (or V/Q scan if D-dimer testing is not available). A negative test rules out embolism. A nondiagnostic V/Q scan or a D-dimer level in excess of 500 ng/dL necessitates additional testing, with serial venous ultrasound a reasonable next test.
- Intermediate pretest probability: Begin with V/Q scanning. A normal V/Q scan rules out the diagnosis; a high-probabilty study rules it in. A nondiagnostic V/Q result should be followed by D-dimer and serial venous ultrasound testing. If initial ultrasound is negative and D-dimer is positive, then serially repeat ultrasound over 1 week. If ultrasound is negative and D-dimer is negative, then embolism is excluded.
- High pretest probability: Admit to the hospital and begin anticoagulation, pending the results of testing. Start with V/Q scanning. If positive, continue anticoagulation; if totally negative, cease testing and anticoagulation. If indeterminate, follow with serial venous ultrasound and D-dimer testing. If ultrasound is positive, the diagnosis is confirmed; if negative or equivocal, then consider angiography (or helical CT) if the D-dimer is positive. If D-dimer and venous ultrasound are negative, cease anticoagulation and follow with serial venous ultrasound testing over the next week.
A. H. G.
ANNOTATED BIBLIOGRAPHY
1. Daton
W, Hall ML, Russo J. Chest pain: relationship of psychiatric illness to
coronary arteriographic results. Am J Med 1988;84:1. (There
is a greatly increased incidence of major depression and panic disorder
in patients with angiographically negative chest pain.)
2. DeSanctis RW, Dorochazi RB, Austen WG, et al. Aortic dissection. N Engl J Med 1987;317:1060. (A classic review, with emphasis on clinical features; 83 references.)
3. Epstein SE, Gerber LH, Borer JS. Chest wall syndrome. JAMA 1979;241:2793. (Best description of the condition and the physical maneuvers that might elicit the pain associated with it.)
4. Hochman
JS, Tamis JE, Thompson TD, et al. Sex, clinical presentation, and
outcome in patients with acute coronary syndromes. N Engl J Med
1999;341:226. (Data from the GUSTO
IIb [Global Utilization of Streptokinase and t-PA for Occluded Coronary
Arteries] study showing gender-based differences in presentation and
outcomes.)
5. Kayser HL. Tietze's syndrome—a literature review. Am J Med 1956;21:982. (Classic review; points out the often epidemic nature of the illness.)
6. Kloner RA, Rexkalla SH. Cocaine and the heart. N Engl J Med 2003;348:487. (A terse summary of acute and chronic effects; a major cause of ischemic chest pain in young persons.)
7. Panting
JR, Gatehouse PD, Yang GZ, et al. Abnormal subendocardial perfusion in
cardiac syndrome X detected by cardiovascular magnetic resonance
imaging. N Engl J Med 2002;346:1948. (Evidence for microvascular dysfunction).
8. Kirshenbaum HD, Ockene IS, Alpert JS, et al. The spectrum of coronary artery spasm. JAMA 1981;246:354. (Early paper emphasizing the variability of clinical presentation and the difficulties of diagnosis.)
9. Rao
SSC, Gregersen H, Hayek B, et al. Unexplained chest pain: the
hypersensitive, hyperreactive, and poorly compliant esophagus. Ann
Intern Med 1996;124:950. (Finds reductions in threshold for pain and compliance, and enhanced reactivity to distention compared with controls.)
10. Sahn SA, Heffner JE. Spontaneous pneumothorax. N Engl J Med 2000;342:868. (Practical review; 85 references.)
11. Spodick DH. Acute percarditis. JAMA 2003;289:1150. (Clinical update; 20 references.)
12. Wexler L. Studies of acute coronary syndromes in women—lessons for everyone. N Engl J Med 1999;341:275. (An
editorial summing up what is known about the differences between men
and women and the implications of these differences for treatment of
all patients.)
13. Bholasingh
R, Cornel JH, Kamp O, et al. The prognostic value of markers of
inflammation in patients with troponin-negative chest pain before
discharge from the emergency room. Am J Med 2003;115:521. (Prospective cohort study; both tests provided independent predictive value for future coronary events.)
P.139
14. Brennan
ML, Penn MS, Van Lente F, et al. Prognostic value of myeloperoxidase in
patients with chest pain. N Engl J Med 2003;349:1595. (Promising cohort study showing a strong independent contribution to the 6-month prediction of coronary event risk.)
15. Diamond
GA, Forrester JS. Analysis of probability as an aid in the clinical
diagnosis of coronary artery disease. N Engl J Med 1979;300:1350. (A classic paper on the importance of pretest probability in diagnosis, particularly for coronary disease.)
16. Feldman RL. Ambulatory electrocardiographic monitoring: the test for ischemia? Ann Intern Med 1988;109:608. (Editorial; urges caution in use for detection of ischemia, pointing out many problems in interpreting ST-segment depression.)
17. Garber
AM, Solomon NA. Cost-effectiveness of alternative test strategies for
the diagnosis of coronary artery disease. Ann Intern Med 1999;130:719. (Meta-analysis
and cost-effectiveness study; echocardiography and single-photon
emission computed tomography are the most cost-effective tests when the
pretest probability is intermediate; immediate angiography is best for
those with a high pretest probability.)
18. Kuntz
KM, Fleischmann KE, Hunink MGM, et al. Cost-effectiveness of diagnostic
strategies for patients with chest pain. Ann Intern Med 1999;130:709. (Excellent
decision-analysis study; electrocardiogram and echo stress testing
proved the most cost-effective in persons with an intermediate pretest
probability of coronary disease.)
19. Lee TH, Goldman L. Evaluation of the patient with acute chest pain. N Engl J Med 2000;342:1187. (Practical review of initial clinical evaluation and immediate management, with emphasis on the diagnosis of ischemic disease.)
20. Nallamothu
BK, Saint S, Bielak LF, et al. Electron-beam computed tomography in the
diagnosis of coronary artery disease: a meta-analysis. Arch Intern Med
2001;161:833. (Finds test characteristics to be similar to those for stress testing.)
21. Pope
JH, Aufderheide TP, Ruthhazer R, et al. Missed diagnoses of acute
cardiac ischemia in the emergency department. N Engl J Med
2000;342:1163. (Prospective cohort study identifying, by multivariate analysis, the causes of missed diagnosis.)
22. Pryor
DB, Shaw L, McCants CB, et al. Value of the history and physical in
identifying patients at increased risk of coronary disease. Ann Intern
Med 1993;118:81. (Clinical assessment effectively predicted coronary heart disease risk and need for further study.)
23. Schulman
KA, Berline JA, Harless W, et al. The effect of race and sex on
physicians' recommendations for cardiac catheterization. N Engl J Med
1999;340:618. (A videotape study of
simulated patients; race and sex were powerful independent predictors
of risk assessment and management recommendations.)
24. Frobert
O, Funch-Jensen P, Bagger JP. Diagnostic value of esophageal studies in
patients with angina-like chest pain and normal coronary angiograms.
Ann Intern Med 1996;124:959. (A
controlled study; no difference was found between patients and controls
in 24-hour esophageal monitoring studies and/or provocation testing;
the authors question the routine use of such testing in patients with
noncardiac chest pain.)
25. Chan WS, Ray JG, Murray S, et al. Suspected pulmonary embolism in pregnancy. Arch Intern Med 2002;162:1170. (Observational cohort study; prevalence of embolization was low; ventilation/perfusion [V/Q] scanning appeared safe to the fetus.)
26. Eisner MD. Before diagnostic testing for pulmonary embolism: estimating the prior probability of disease. Am J Med 2003;114:232. (An editorial emphasizing the ability and importance of making a pretest probability estimate.)
27. Frost
SD, Brotman DJ, Michota FA. Rational use of D-dimer measurement to
exclude acute venous thromboembolic disease. Mayo Clin Proc
2003;78:1385. (Terse review of test characteristics and evidence for its efficacy; 45 references.)
28. Kearon
C, Ginsberg JS, Hirsh J. The role of venous ultrasonography in the
diagnosis of suspected deep venous thrombosis and pulmonary embolism.
Ann Intern Med 1998;129:1044. (Reviews the evidence for the efficacy of the test in the diagnosis of pulmonary embolization.)
29. Krulp
MJHA, Leclercq MGL, van der Heul C, et al. Diagnostic strategies for
excluding pulmonary embolism in clinical outcome studies: a systematic
review. Ann Intern Med 2003;138:138. (Examines data on approaches that use pretest probability plus simple tests; finds substantial accuracy and efficacy.)
30. Perrier
A, Howarth N, Didier D, et al. Performance of helical tomography in
unselected outpatients with suspected pulmonary embolism. Ann Intern
Med 2001;135:88. (Prospective observational study; sensitivity, 70%; specificity, 91%; false-negative rate, 5%; false-positive rate, 7%.)
31. The
PIOPED Investigators. Value of the ventilation/perfusion scan in acute
pulmonary embolism: results of the Prospective Investigation of
Pulmonary Embolism Diagnosis (PIOPED). JAMA 1990;263:2753. (Best study of sensitivity and specificity of V/Q scanning.)
32. Rathbun
SW, Raskob GE, Whitsett TL. Sensitivity and specificity of helical
computed tomography in the diagnosis of pulmonary embolism: a
systematic review. Ann Intern Med 2000;132:227. (Critical analysis of the data, revealing inadequate information to determine the appropriate use of this new technology.)
33. Righini
M, Goehring C, Bounameaux K, et al. Effects of age on the performance
of common diagnostic tests for pulmonary embolism. Am J Med
2000;109:357. (A cross-sectional study; test yields and results are affected by age.)
34. Robin ED. Overdiagnosis and overtreatment of pulmonary embolism. Ann Intern Med 1977;87:775. (A classic discussion of the shortcomings of lung scans and arterial blood gases.)
35. Van
Strijen MJL, de Monye W, Schiereck J, et al. Single-detector helical
computed tomography as the primary diagnostic test in suspected
pulmonary embolism: a multicenter clinical management study of 510
patients. Ann Intern Med 2003;138:307. (Prospective obervational study of inpatients and outpatients.)
36. Wells
PS, Anderson DR, Rodger M, et al. Excluding pumonary embolism at the
bedside without diagnostic imaging: management of patients with suspect
pulmonary embolism presenting to the emergency department by using a
simple clinical model and D-dimer. Ann Intern Med 2001;135:98. (Prospective
cohort study; demonstrates the safety and the efficacy of combining
pretest probability and D-dimer testing; the negative predictive value
of the test was 99.5% in low-risk patients.)
37. Wells
PS, Ginsberg JS, Anderson DR, et al. Use of a clinical model for safe
management of patients with suspected pulmonary embolism. Ann Intern
Med 1998;129:997. (Multicenter
prospective cohort study demonstrating the efficacy and the safety of
combining clinical assessment of pretest probability, V/Q scanning, and
venous ultrasound to evaluate persons for pulmonary embolization.)
38. Andersen
HR, Nielson TT, Rasmussen K, et al. A comparison of coronary
angioplasty with fibrinolytic therapy in acute myocardial infarction. N
Engl J Med 2003;249:733. (A
randomized trial in the community setting; angioplasty produced
superior outcomes, despite the need to transfer patients to an
invasive-therapy center.)
39. Dracup
K, Alonzo AA, Atkins JM, et al. The physician's role in minimizing
prehospital delay in patients at high risk for acute myocardial
infarction: recommendations from the National Heart Attack Alert
Program. Ann Intern Med 1997;126:645. (Strategies for minimizing delay in hospitalization.)