Chest pain is one of the most frequent presenting
complaints of emergency department (ED) patients and often misdiagnosed.
Estimates of the incidence of missed diagnoses of acute myocardial
infarction (MI) in the ED range from 1 - 5%. A recent study found that
one-quarter of these missed MI patients were discharged with an ED
diagnosis of ischemic heart disease! The morbidity and short-term
mortality rates in patients sent home with MIs are significantly higher
than MI patients who are correctly admitted. On the other side of the
coin, low thresholds for admission of "rule-out" MI (ROMI) patients to
coronary care units (CCU) result in fewer than 30% "rule-in" MI rates.
Over-triage not only drives up the cost of medical
care at a time when managing health care costs seems to be the
preeminent goal but also limits the availability of intensive care beds
for other ED patients. Physicians must also become better
diagnosticians because of the advent of social trends, such as cocaine-
associated myocardial ischemia and more elderly patients with silent MIs
and because of the advent of time-dependent therapies such as
thrombolytic agents. Due to the difficulty of diagnosing acute MIs,
thrombolytic therapy is administered to fewer than 20% of all MI
patients. While clinical research has yet to identify a diagnostic
method superior to bedside judgment in determining which patients to
admit for possible MI, chest pain diagnostic testing is rapidly evolving
to enhance the physician's ability to better utilize resources and to
more rapidly detect patients in need of acute intervention.
The bedside determination of whether to admit a
patient to ROMI has traditionally been based upon the history,
supplemented by the electrocardiogram (ECG). The Multicenter Chest Pain
Study Group has collected data on over 10,000 ED visits of patients with
acute chest pain. This group has determined that it is possible to use
historical data to stratify patients into groups with higher or lower MI
probability, but there are no absolute factors in the history or
physical examination to ROMI.
For example a "burning" type of pain is
experienced by as many as 15% of MI patients. And chest pain
accompanied by tenderness of the chest wall is also found in as many as
15% of MI patients. Statistical predictive models have been developed
using the Chest Pain Study data (history and physical exam and ECG
results) including a model derived by logistic regression analysis and
another using recursive partitioning.
Compared with practicing physicians, all of these
computerized models show a trend toward higher specificity (i.e., fewer
ROMI admits). The sensitivity of these computerized models at detecting
MI is no better than that of physician judgment (i.e., no fewer missed
MIs) and may be worse. The most sensitive (97%) and specific (96%)
mathematical model reported to date for the prediction of MI is the
artificial neural network developed by Baxt.
However all of the published reports using
mathematical models have included patients with ECGs diagnostic for
acute MI which inflates their calculated accuracy. The sensitivity of
Baxt's neural network is only 80% in the subgroup of MI patients with
nondiagnostic ECGS, the results would have been 11 missed MI cases.
More relevant to current bedside practice, risk management analyses
reveal that the medical record documentation of the history in
particular is often inadequate. Authorities recommend computerized
dictation systems which prompt the physician to be complete.
Use of the ECG?
The ECG is the most widely used diagnostic study in
the ED evaluation of chest pain. Recently published studies question
the accuracy of ECG interpretation by emergency physicians, although the
actual percentage of clinically significant misreads is probably low,
especially when the emergency physician has specialty training or board
certification in emergency medicine. However, one study found that one-
quarter of all missed MIs are attributable to misinterpretation of the
initial 12-lead ECG.
Unfortunately the initial ECG performed in the ED
reveals ST-segment elevation or other ST-T wave changes in only about
half of all acute MIs. The sensitivity of the ECG for acute cardiac
ischemia can be enhanced by the use of serial ECGs, right-sided chest
leads (15 leads) and even back leads (22 leads). Continuous 12-lead ECG
trend monitoring with computerized programs can detect beat-to-beat
changes in QRS complex morphology and ST-segment shifts suggestive of
ischemia or infarction. But, until these refinements in
electrocardiography become standard practice, the best use of the 12-
lead ECG remains two-fold: to detect acute MI patients meeting ECG
criteria for thrombolytic therapy and to risk stratify ROMI patients to
the CCU or intermediate unit. Chest pain patients with abnormal ECGs
are at high risk for complications requiring critical care
Cardiac enzymes, specifically the creatine kinase
isoenzyme (CKMB), are the in-hospital gold standard for making the
diagnosis of MI. But their acceptance and applicability to the early
evaluation of the ED patient remains controversial, in part because of
previous experience with the nonspecific (skeletal as well as cardiac
muscle) and insensitive (not elevated in 15% of all MIs) creatine kinase
(CK). There are many factors to consider when interpreting CKMB.
Previously, CKMB, like CK was measured by electrophoresis, a cumbersome
and relatively insensitive assay technique.
Newer immunochemical assays have been developed
which can detect early small increases in CKMB. The multicenter
Emergency Cardiac Research Group has published several studies
suggesting that serial CKMB measurements over the first 3 - 4 hours
after presentation to the ED can significanlty improve the accuracy of
diagnosing MI patients with nondiagnostic ECGs. The rapid availability
of serial CKMB results appeared to affect clinical decision making in
about one-third of ED patients with chest pain and nondiagnostic ECGs.
Serical CKMB results were more accurate than serial 12-lead ECGs in the
ED diagnosis of MI.
The numbers of hours of chest pain is key to
interpreting cardiac enzyme results. MI patients with chest pain for 3,
6, 9 and 12 hours had elevated CKMB results in nearly half, >50%, >75%
and in 100% of cases, respectively. Myoglobin is released earlier than
CKMB from damaged myocardial cells (within 4 - 6 hours), but it
disappears with a few more hours (about 6 - 12 hours), about the time of
the expected CKMB elevation. Myoglobin is also nonspecific because it
also arises from skeletal muscle.
New rapid assays for myoglobin utilizing latex
agglutination, radioimmunoassay and immunoturbidimetric measurements are
promising, showing up to 100% sensitivity for detecting MI with serial
evaluations over a 4-hour period. Another report showed myoglobin to be
significantly more sensitive than either CKMB or CK in detecting MI
within the first 6 hours after onset of chest pain. Other serum markers
of myocardial damage like CK-MM subtypes, light chain myosin fragments
and the troponin proteins are currently under investigation. Troponin
appears to have similar early diagnostic characteristics as CKMB but it
remains elevated for at least a week. Troponin is more specific than
either CKMB or myoglobin in patients with concomitant skeletal muscle
and cardiac muscle damage and it can detect about one-third of patients
with unstable angina who are not detected by CKMB.
Cardiac Emergency Centers:
A trend in the evaluation of chest pain patients is
the introduction of cardiac emergency centers or chest pain units
usually within the confines of existing EDs. The goal is to administer
thrombolytic agents and other equally efficacious but less risky
medications (eg, aspirin, IV nitroglycerin, beta-blockers and magnesium)
as soon as possible to patients with ECGs diagnostic for acute MI as
well as to prepare MI patients for invasive procedures (eg, angiograms
and angioplasty) in the hospital's cath lab.
In other chest pain patients, the goal is to
combine as many diagnostic tools described above as well as those
usually associated with non-ED cardiac workups, such as echocardiograms
and treadmill stress tests, in order to "rule-out" cardiac ischemia as
soon as possible. The introduction of ED cardiac centers takes this
concept further by avoiding unnecessary CCU admits.
Static cardiac evaluation is being intensively
studied as a diagnostic tool in the ED. Echocardiography can reveal
regional wall motion abnormalities within the first hour after onset of
coronary artery occlusion. While the presence of wall motion
abnormalities cannot differentiate an old infarct from an evolving MI,
the presence of such changes may be diagnostic in certain patients.
Echo can also be used to rule out other critical pathology, such as
aortic dissection or cardiac tamponade as a cause of chest pain.
Thallium and technetium scintigraphy can show areas of decreased
myocardial perfusion with 4 - 6 hours of infarction but the dilemma of
distinguishing new from old infarct remains.
The results of recent studies of nuclear medicine
studies in ED patients conflict as to their diagnostic usefulness in
determining which patients with nondiagnostic ECG are likely to have an
MI. Dynamic cardiac evaluation may be appropriate for many stable chest
pain patients with otherwise negative initial workups. The preliminary
studies of treadmill exercise stress testing in ED patients with
atypical chest pain suggest that false positive results may be more
likely in this population of patients at low risk for significant
coronary disease. Further study on higher risk patients is necessary to
delineate the proper role of stress testing in the ED.
With demands for more rapid and accurate
diagnosis and for better resource utilization the future evaluation of
chest pain patients will most likely involve some type of outpatient
holding area in the ED.
* This article is presented and copyrighted by The 'Lectric Law Library
and Dr. Steven E. Lerner & Associates (www.drlerner.com)