~~ ~ ~

330 ACADEMIC EMERGENCY MEDICINE JULY/AUGUST 1994 VOL 1/NO 4

SCIENTIFIC ADVANCES
Prospective Multicenter Evaluation
of Cocaine-associated Chest Pain
Judd E . Hollander, MD, Robert S. Hoffman, MD, Paul
Gennis, MD, Phillip Fairweather, MD, Michael J. DiSano,
MD,David A . Schumb, MD, James A . Feldman, MD,
Susan S. Fish, PharmD, MPH, Sophia Dyer, BS, Paul Wax,
MD, Chris Whelan, MD, Evan Schwarzwald, DO
For the Cocaine Associated Chest Pain (COCHPA)
Study Group [See Appendix A ]

ABSTRACT
Objective: To describe a large cohort of patients who had chest pain
following cocaine use, and to determine the incidence of and clinical
characteristics predictive for myocardial infarction in this group of
patients.
Methods: A prospective observational cohort study of consecutive pa-
tients with cocaine-associated chest pain was conducted in six municipal
hospital emergency departments (EDs). Demographic variables, drug
abuse patterns, medical histories, chest pain characteristics, ECG results,
and laboratory data were recorded. Myocardial infarction was the prima-
ry endpoint.
Results: Fourteen of 246 patients (5.7%; 95% confidence interval [CI],
2.7-8.7%) had myocardial infarction, as diagnosed by elevated CK-MB
isoenzyme levels. There were two deaths (0.8%). The patients had a
median age of 33 years. The majority were male (71.5%), non-white
(83.3%), cigarette smokers (83.3%) who used cocaine regularly. Chest
pain began a median of 60 minutes after cocaine use and persisted for a
median of 120 minutes. Chest pain was most frequently described as
substernal (71.3%) and pressure-like (46.7%). Shortness of breath
(59.3%) and diaphoresis (38.6%) were common. There was no clinical
difference between patients who had myocardial infarctions and those
who did not. Twelve patients had arrhythmias and four had congestive
heart failure. All cases requiring intervention were evident upon presenta-
tion. An ECG revealing ischemia or infarction had a sensitivity of 35.7%
for predicting a myocardial infarction. The specificity, positive predictive
value, and negative predictive valueof the ECGs were 89.9%,17.9%, and
95.8%, respectively.
Conclusions: Myocardial infarction in patients who have cocaine-associ-
ated chest pain is not uncommon. No clinical parameter available to the
physician can adequately identify patients at very low risk for myocardial
infarction. Therefore, all patients with cocaine-associated chest pain
should be evaluated for myocardial infarction.
Acad. Emerg. Med. 1994; 11330-339.
Cocaine-associated Chest Pain, Hollander et al. 331

1 In the past decade the incidence of cocaine abuse in our
population has increased dramatically. An estimated 24
million Americans have used cocaine at least once,’ and
five million use it regularly.’ Concurrent with the increase
in recreational cocaine abuse, emergency physicians have
witnessed an almost 20-fold increase in the number of
cocaine-related complaints.3 By 1989, cocaine had become
the drug of abuse most commonly implicated in the presen-
tation of patients to emergency departments ( E D s ) , ~com-
prising over 40% of all such case^.^
Chest pain is the most frequent cocaine-associated
complaint, occurring in as many as 40% of such patients
presenting to the ED.6 Cocaine-associated myocardial in-
farction was first described in 1982.’ Since that time over
100 additional cases have been r e p ~ r t e d . As
~ , ~a result,
many patients who have cocaine-associated chest pain are
admitted to the hospital to “rule out” myocardial infarc-
tion.
Four previous clinical series have yielded disparate
results concerning the incidence of myocardial infarction
in patients with cocaine-induced chest pain. Zimmerman
et a1.I0 found that only three of 48 such patients (6%)
admitted to the CCU had sustained myocardial infarction.
Gitter et al.” reported no myocardial infarction in their
cohort of 101 patients with cocaine-associated chest pain
admitted to a monitored setting. Neither of these studies
included patients released from the ED. Since the percent-
age of patients admitted to the hospital with cocaine-
associated chest pain has been reported to be as low as
5% ,6 these studies suggest that myocardial infarction fol-
lowing cocaine use is a rare event.
In contrast, Amin et a1.12 found that 22 of 70 patients
(31%) admitted with chest pain and cocaine use had sus-
tained myocardial infarction. Tokarski et al.13 found that
eight of 42 patients (19%) with cocaine-associated chest
pain and normal or nonspecific ECGs had elevated cardiac
isoenzyme levels.
This prospective multicenter study was designed to
determine both the frequency and the clinical characteris-
tics of acute myocardial infarction in patients presenting to
the ED with chest pain following cocaine use.
I RESULTS
Patients
Two hundred forty-six patients were enrolled over 46
months at six hospital EDs. One patient was excluded due
to failure to participate in follow-up. Patient demographics
are summarized in Table 1.
Diagnosis
One hundred forty patients (56.9%) were admitted to
the hospital or observed in the ED observation unit (for up
to 24 hours). One hundred six patients (43.1%) were
released from the ED. Of the 246 patients enrolled in the
study, only two are known to have died. Both arrived in

TABLE 1 Demographic and Historical Characteristics of Patients with and Patients without Myocardial Infarction (MI)

Patients without
Patients with MI MI Significance of
(n = 14) (n = 232) Difference?
Age (median, IQ25.,s*) 36.5 (33-44) years 33 (26-38) years p = 0.025
Male 71.4% 71.6% NS
Race NS
African-American 85.7% 62.5%
Hispanic 14.3% 19.4%
Caucasian 0 17.7%
Other 0 0.4%
Cardiac risk factors
Hypertension 28.6% 18.1% NS
Hypercholesterolemia 7.1% 4.7% NS
Family history of coronary artery disease 21.4% 15.1% NS
Diabetes mellitus 7.1% 4.3% NS
Tobacco use 92.9% 82.8% NS
Past medical problems
Chest pain 21.4% 52.6% p = 0.047
Myocardial infarction 0 4.3% NS
Congestive heart failure 7.1% 3.4% NS
Arrhythmias 7.1% 3.4% NS
Cardiac medication 7.1% 8.6% NS
*IQ = interquartile range.
t N S = not significant (p > 0.05).
 332 ACADEMIC EMERGENCY MEDICINE JULY/AUGUST 1994 VOL 1/NO 4
I METHODS
Study Design
This was a prospective multicenter observational
study of consecutive eligible patients presenting to the
ED with cocaine-associated chest pain.
Participants
All patients presenting to the EDs of six inner-city
teaching hospitals who had anterior, precordial, or left-
sided chest pain unexplained by local trauma or radio-
graphic abnormalities and who had used cocaine in the
past week were eligible for inclusion. Patients with
isolated right-sided chest pain were not included be-
cause anecdotal experience suggested that the likeli-
hood of infarction in this population was extremely low.
Patients were identified prospectively and enrolled by
study investigators prior to hospital admission or re-
lease from the ED. Participating centers and dates of
commencement can be found in Appendix A. All cen-
ters terminated enrollment on May 29, 1992. The study
was approved by the institutional review boards of all
participating hospitals. Patients were assured confiden-
tiality of information, and all patients provided written
informed consent.
Data Collection
Historical and demographic data were recorded at
the time of the ED evaluation on a standardized data-
collection form. The data included: patient age, gender,
and race; cardiac risk factors (hypertension, hyper-
cholesterolemia, tobacco use, diabetes mellitus, and
family history of premature atherosclerotic heart dis-
ease); pattern of cocaine abuse (length of time used,
frequency of use, route of administration, estimated
quantity used in the last 24 hours, and time of the most
recent use); history of other substance abuse; and HIV
status. The duration, location, and quality of pain (pres-
sureltightness, aching/dull, burningtindigestion, sharp/
cardiac arrest and died within eight hours of arrival.
Neither of these patients had pathologic evidence of myo-
cardial infarction. Of the remaining 244 patients, follow-up
data were obtained for 223 (91.4%). All were alive at the
time of follow-up.
Fourteen patients (5.7%; 95% CI, 2.7-8.7%) met the
criteria for myocardial infarction. All 14 patients had
elevated CK-MB isoenzymes. No patient was diagnosed as
having a myocardial infarction based on electrocar-
diographic changes in the absence of an elevated CK-MB
stabbing, or other) and the presence of associated cardi-
ac symptoms (nausea, emesis, dyspnea, diaphoresis,
syncope, light-headedness, and palpitation) were re-
corded. Current medications and past medical history,
with an emphasis on chest-pain syndromes and heart
disease, were sought. The recorded physical examina-
tion included vital signs and the presence or absence of
jugular venous distention, rales, abnormal heart sounds,
and pedal edema. Also noted were the patient’s mental
status and pupil size and the presence of any abnormal
neurologic finding.
ECG
All patients received ECGs in the ED. Each ECG
was classified into one of six categories using a closed-
question format: 1) normal (no electrocardiographic
evidence of ischemia); 2) nonspecific (accepted devia-
tion from the norm with the lowest likelihood of isch-
emia, such as an abnormal T-wave axis in lead 111, sinus
tachycardia, or the early repolarization variant [elevated
takeoff of the ST segment at the J point of the QRS
complex greater than 0.1 m V relative to the isoelectric
line, downward concavity of the ST segment, and sym-
metrically limbed T-waves, often of large amplitude] );
3) abnormal but not diagnostic of ischemia (i.e., left-
ventricular hypertrophy or bundle-branch blocks);
4) consistent with ischemia but known to be old (A
comparative ECG [from a time when the patient was
not experiencing chest pain] was available and un-
changed.); 5 ) consistent with ischemia and not known to
be old; or 6) suggestive of acute myocardial infarction
(ST elevation greater than 0.1 mV measured 80 msec
from the J point, with or without T-wave inversions in
the distribution of one or more epicardial coronary
arteries).
Each ECG was read independently by two investiga-
tors and interpreted prior to knowledge of whether the
patient had sustained a myocardial infarction. When the
investigators’ initial interpretations were discordant,
the disagreement was discussed until a consensus diag-
nosis was agreed on.
isoenzyme level. Other diagnoses included ischemic heart
disease (35 patients), atypical chest pain (101 patients), and
non-cardiac chest pain (96 patients).
Cocaine Use
All but one of the patients were regular users of co-
caine, with a median length of cocaine use of 5.0 years
(IQ25-75, 2-7 years) and a median frequency of ten epi-
sodes of cocaine use per month (IQ25-75, 4-30 times/
Cocaine-associated Chest Pain, Hollander et al.
Chest Radiography
The need for chest radiography was left to the
judgment of the emergency physician caring for the
patient. If the patient had a radiographic abnormality
that was believed to account for the current episode of
chest pain (i.e., pneumonia or pneumothorax), he or she
was excluded from the study.
Laboratory Values and Cardiac Isoenzymes
All admitted patients had blood samples obtained for
evaluation of total creatine kinase every eight to 12
hours for a total of 24 hours. Total CK and CK-MB
samples were measured at each institution using the
institutional threshold for analysis based on normal
volunteers at that institution. When the CK level was
greater than the upper limits of normal, an MB fraction
was obtained. When the patient was released from the
ED, a CK sample was requested and the patient was
asked to return within 24 hours for a second sample.
When other laboratory values had been obtained at
the discretion of the treating physician, the hematocrit,
creatinine, glucose, prothrombin time, partial throm-
boplastin time, and cholesterol were recorded. When
the physician caring for the patient had obtained a urine
sample for cocaine metabolites, the results also were
recorded. We did not require urine toxicology for all
patients because it is unlikely that patients would claim
cocaine use in its absence.
FOllOW-Up
The decision to release the patient from the ED or
admit the patient was based on the attending physician’s
standard clinical evaluation, including review of the
patient’s prior medical record. The hospital courses of
admitted patients were tracked and recorded. Results of
tests assessing cardiac function (i.e., ECG, nuclear
study, stress test, or cardiac catheterization) were re-
corded, when obtained. Patients released from the ED
who did not return within 24 hours were assessed within
the next two weeks by telephone. They were questioned
month). One hundred seventy-nine patients (72.8%) smoked
cocaine, 67 (27.2%) insufflated cocaine, and 26 (10.6%)
injected cocaine intravenously. Twenty-three patients
(9.3%) used cocaine through at least two routes; three of
these patients (1.2%) used cocaine through three routes.
Patients presented a median of five hours after their most
recent cocaine use (IQ25-75, 2-22 hours). Myocardial in-
farction occurred after inhalation, insufflation, and intra-
venous administration of cocaine. No group demonstrated
a significantly higher rate of infarction. Ninety-one of 92
333
with regard to repeat hospital visits, recurrent chest
pain, shortness of breath, and continued cocaine use.
Patients who were ruled out for myocardial infarction in
the hospital or ED observation unit were not reassessed
by telephone.
Definitions
Myocardial infarction was considered to occur if the
patient had prolonged chest pain and either elevation of
CK-MB isoenzymes above the upper limits of normal or
standard electrocardiographic evolution of myocardial
infarction within 24 hours of ED presentation.
At the time of discharge from the hospital, a final
diagnosis of myocardial infarction, ischemic, “atypi-
cal,” or “non-cardiac” chest pain was assigned to each
case. The diagnosis of myocardial infarction was made
independently by the investigators according to the
above-mentioned criteria. Other diagnoses were made
based on the clinical impression of the attending physi-
cian managing the case. By definition, the diagnosis of
atypical chest pain was reserved for those patients
whose chest pain was potentially ischemic.
Statistical Analysis
Continuous nonparametric data are presented as
medians with interquartile ranges (IQ25-75). Categorical
data are presented as the frequencies of patients with the
assessed variables. The 95% CIS are also provided.
Characteristics of patients with myocardial infarc-
tion were compared with those of patients without
myocardial infarctions using the chi-square test or Fish-
er’s exact test for categorical data. Nonparametric com-
parisons of medians were performed using the Kruskal-
Wallis H test. Due to the small number of patients
having myocardial infarctions, we were unable to con-
struct a multivariate logistic regression model that
would be likely to be validated on an independent data
set.14 A p value of less than 0.05 was considered
significant for all tests.
patients (98.9%) tested were found to have cocaine metab-
olites present in their urines. The route of cocaine use,
length of use, frequency of use, or interval from last
cocaine use to presentation did not differ between patients
with and those without myocardial infarction.
History
Prior myocardial infarction, congestive heart failure,
arrhythmias, or use of cardiac medications was uncom-
334 ACADEMIC EMERGENCY MEDICINE JULY/AUGUST 1994 VOL l / N O 4

I TABLE 2 Chest-pain Characteristics among 246 Cocaine Abusers Chest Radiography

with and without Myocardial Infarction (MI)
. . . . ......,,,,.. ..........,. ,........ .. . .. ... . .... ... . . .... . .. .............. .. .
Chest radiography was performed for 191 patients; 170
Patients Patients (89%) had normal results. The remaining 21 patients had
with MI without MI Significance of abnormalities that could not sufficiently explain their chest
(n = 14) (n = 232) Difference+
pain.
Location* NS
Substernal 85.7% 70.4% Electrocardiography
Apical 0 18.4%
Other 14.3% 11.2% ECGs of 242 patients were available for analysis (see
Quality? NS Table 4). Only 68 patients (28.1%) had normal ECGs. Of
Pressure/tightness 28.6% 47.8% the 104 ECGs with nonspecific changes, 76 had ST-seg-
Sharplstabbing 42.9% 39.2%
ment elevations greater than 1 mm due to early repolariza-
Achingidull 21.4% 8.2%
Burning/indigestion 0 5.6% tion. Patients with myocardial infarction were more likely
Other 0 2.6% than those without myocardial infarction to have ECGs
Pleuritic component consistent with ischemia or infarction (p = 0.014), al-
Yes 28.6% 35.8% NS though no myocardial infarction was diagnosed based on
Associated symptoms electrocardiographic changes in the absence of elevated
Shortness of breath 28.6% 61.2% p = 0.033 cardiac isoenzymes. The sensitivity of an ECG revealing
Diaphoresis 50.0% 37.9% NS
ischemia or infarction for predicting an acute myocardial
Palpitations 35.7% 33.2% NS
Nausea 14.3% 28.4% NS infarction was only 35.7% (95% CI, 10.6-60.8%). Speci-
Vomiting 14.3% 10.8% NS ficity and positive and negative predictive values were
Syncope 21.4% 9.1% NS 89.9% (95% CI, 86-93.8%), 17.9% (95% CI, 3.7-32.1%),
*Location was not described for nine patients without MI. and 95.8% (95% CI, 93.1-98.5%), respectively.
t o n e patient with MI was unable to describe his chest pain. Several
patients without MI gave more than one description of their chest pain. Laboratory Evaluation
$NS = not significant (p > 0.05).
The hematocrit, creatinine, prothrombin time, partial
thromboplastin time, serum glucose, and serum choles-
mon. Most patients were cigarette smokers and had been terol were not helpful in identifying patients at high or low
for a median of 12 pack years (IQ25-,5, 5-20 pack years). risk of infarction. Cardiac isoenzymes were obtained for
Other cardiac risk factors were relatively uncommon, each 215 patients (87%). Two or more isoenzymes were obtained
occurring in less than 20% of patients. History of chest for 151 (70%) of these patients. Fourteen patients met
pain was less common for patients who had sustained criteria for myocardial infarction. When comparing pa-
myocardial infarction (see Table 1). tients with myocardial infarctions with those without myo-

Presentation I TABLE 3 Vital Signs at the Time of Presentation of Patients with

and without Myocardial Infarction (MI)
Chest-pain characteristics are summarized in Table 2.
Patients with and those without myocardial infarction were
Patients Patients with- Significance of
similar. Chest pain began a median of 60 minutes after with MI out MI Difference*
cocaine use (1Q25-75,3-645 minutes) and persisted for a
Systolic blood pressure (torr) NS
median of 120 minutes (IQ25-75, 30-367 minutes). There
> 140 5 (35.7%) 79 (34.1%)
was no difference between patients with and those without >91-139 8 (57.1%) 150 (64.5%)
myocardial infarction. Overall, 81% of patients presented <90 1(7.1%) 3 ( 0.4%)
to the hospital within 24 hours of the most recent use of Diastolic blood pressure (torr) NS
cocaine. All patients with myocardial infarctions had had >90 5 (35.7%) 48 (20.7%)
the onset of their chest pain, and presented to the hospital, <90 9 (64.3%) 184 (79.3%)
within 24 hours of cocaine use. The incidence of myocar- Pulse (beatdmin) NS
> 100 2 (14.3%) 79 (34.1%)
dial infarction in patients presenting within 24 hours of 12 (85.7%) 151 (65.0%)
60- 100
cocaine use was 7.8% (95% CI, 3.9-11.8%); in patients <60 0 2 ( 0.9%)
presenting within 12 hours of cocaine use the incidence of Respiratory rate (/min) NS
myocardial infarction was 7.4% (95% CI, 2.3-12.6%). >30 0 6 ( 2.6%)
Vital signs were obtained for all patients. Tachycardia 10-30 14 (100%) 224 (96.6%)
and hypertension were common in patients with and those < 10 0 2 ( 0.9%)
without myocardial infarction (see Table 3). No difference Temperature > 3 8 T 0 11 ( 4.7%) NS
in vital signs was noticed between the two groups. *NS = not significant (p > 0.05).
Cocaine-associated Chest Pain, Hollander et al.
cardial infarction, the patients with myocardial infarction
were more likely to have initial CK elevations (86% vs
4670,p = 0.009); higher initial CK levels (515 vs 177
mg/dL; Kruskal-Wallis, p = 0.0002); and initial CK lev-
els that were less likely to be the peak CK levels (85% vs
36%, p = 0.006). Ten of the 14 patients who had myocar-
dial infarction had initially elevated CK-MB level. All
patients who had myocardial infarction had elevated CK-
MB isoenzymes within 24 hours of arrival. For four
patients, only the initial CK-MB levels were elevated.
Disposition
One hundred forty patients (57%) were admitted to the
hospital or observed in an observation unit. One hundred
six patients (43%) were released from the ED. Patients
subsequently determined to have had myocardial infarc-
tion were more likely to have been admitted (86% vs 5 5 % ,
p = 0.027), but not to a monitored bed (58% vs 56%,
p = NS). Two patients with myocardial infarction were
released from the ED. One had CK isoenzymes investi-
gated only as part of the study protocol. This patient’s total
CK level was 2,150 mg/dL, with an MB fraction of 150
mg/dL (7%). He returned one week later with recurrent
chest pain and was admitted. A sub-maximal stress test
was negative at that time. The other patient had an initial
CK level of 258 mg/dL. He walked out after his second
sample, which subsequently revealed a CK level of 567
mg/dL and an MB fraction of 69 mg/dL (12%), was drawn.
This patient was unable to be contacted by telephone,
through relatives, or by a visit to his home.
Evaluation after Presentation
All 14 patients who had myocardial infarction had
elevated CK-MB isoenzymes. Five patients developed
Q-waves on their ECGs. Nine patients had non-Q-wave
myocardial infarctions. Seven infarctions involved the in-
ferior wall, including all the Q-wave infarctions. Three
non-Q-wave myocardial infarctions occurred in the ante-
rior distribution, and the locations of the remaining four
non-Q-wave infarctions were not obvious on ECGs.
No patient developed arrhythmias requiring interven-
tion, congestive heart failure, or infarction extension dur-
ing the hospitalization. Twelve patients developed arrhyth-
mias. Supraventricular arrhythmias occurred in four patients;
three patients had atrial fibrillation on arrival, and one
patient had an asymptomatic 12-beat run of supraventricu-
lar tachycardia 20 hours after arrival. Ventricular ar-
rhythmias were evident in four patients: two arrived in
cardiac arrest; one had stable ventricular tachycardia
treated with lidocaine prior to arrival in the ED; and one
had nonsustained ventricular tachycardia. Bradydysrhyth-
mias were evident in fourpatients: in two patients they were
present on arrival in the ED; the other two patients had
asymptomatic episodes evident on the telemetry monitor.
335
a TABLE 4 Classification of the Initial ECGs of Patients with and
without Myocardial Infarction (MI)
Patients Patlens with
with MI* out MI
fn = 141 (fl = 228)
~~~
Normal 3 (21.4%) 65 (29.0%)
Nonspecific 2 (14.3%) 102 (4S.070)
Abnormal, not diagnostic 4 (28.6%) 38 f16.7%)
Ischemic 1 ( 7.1%) 20 ( 8.8%)
Suggestive of acute infarction 4 (28.6%) 3 ( 1.3%)
*Patients with MI were more likely than those without MI to have
ECGs diagnostic of ischemia or infarction (p = 0.014).
All four patients with congestive heart failure presented
with clinically evident congestive heart failure. No patient
received thrombolytic therapy.
Ten patients without and six patients with myocardial
infarction had treadmill tests that were negative for exer-
cise-induced ischemia. No patient had a positive exercise
treadmill test. No cardiac catheterization was performed.
Twenty patients without infarction had echocardiography
performed. Eight of these patients had left ventricular
hypertrophy and three patients had depressed left ventricu-
lar function. Of the four patients with myocardial infarc-
tion to have echocardiography performed, three had focal
wall motion abnormalities and the other had right ventricu-
lar dilatation.
I DISCUSSION
...... ... . ..... .. .....
. , ....
.. ,, . ...., ., .....,,..., ., , .,. ., . , .... . .
Our study confirms the typical profile of a patient with
cocaine-associated chest pain. We found these patients to
be young (median age, 33 years), male (71.5%), cigarette
smokers (83.3%) with histories of repetitive cocaine abuse
(99.6%). Mean ages in previous series had ranged from 28
to 34 years.8J0-13 Males accounted for 57% to 88% of
patients.8J0-l3 Other than tobacco use, few traditional
cardiac risk factors have been found in this patient popula-
tion.8-loJ2 Although myocardial infarction has been re-
ported to occur in patients using cocaine for the first
time15.16 and in patients receiving cocaine as topical anes-
thesia,17.18 the majority of myocardial infarctions occur in
chronic users of cocaine.8
Coronary artery vasoconstriction, in-situ thrombus
formation, platelet aggregation, and accelerated ath-
erosclerosis are the principal mechanisms responsible for
cocaine-induced myocardial ischemia.8.9 Although clini-
cal evidence of coronary vasospasm secondary to cocaine
use has been demonstrated only infrequently during cardi-
ac catheterization,lgJ0 cocaine has been shown experi-
mentally to produce coronary vasoconstriction that can be
reversed by phentolamine, an alpha-adrenergic blocker,21
and exacerbated by propanolol, a beta-adrenergic block-
er.22 Cocaine-induced vasoconstriction occurs in diseased
and nondiseased coronary-artery segments, however, the
 336 ACADEMIC EMERGENCY MEDICINE JULY/AUGUST 1994 VOL 1/NO 4
magnitude of the effect is more pronounced in the diseased
segments.23
Patients in this and other series8 developed chest pain
within several hours of cocaine use. Recurrent coronary
vasoconstriction associated with the appearance of in-
creased quantities of benzoylecgonine and ethyl-methyl
ecgonine 90 minutes after the insufflation of cocaine24may
be responsible for this phenomenon.
Cigarette smoking has been shown to induce coronary
artery vasoconstriction through an alpha-adrenergic mech-
anism similar to that of cocaine.25~~6 Most patients with
cocaine-induced myocardial infarction in this and other
series8.9.12have been cigarette smokers. It has been postu-
lated that cocaine and nicotine may have a synergistic
effect on coronary vasoconstriction. However, in our co-
hort, patients who smoked cigarettes within the 12 hours
prior to presentation were not more likely to have had
myocardial infarction.
In-situ thrombus formation and platelet activation may
occur secondary to cocaine use. Cocaine has been shown to
activate platelets directly27 and indirectly through a alpha-
adrenergic-mediated increase in platelet aggregability.28
Adenosine diphosphate-induced platelet aggregation is
enhanced in the presence of cocaine.29 Anticardiolipin
antibody has also been detected in cocaine abusers.30
Layers of mural thrombus in two coronary arteries were
observed in a patient following a cocaine-induced myocar-
dial i n f a r ~ t i o nIn
. ~one
~ review, all patients who underwent
cardiac catheterization within four days of cocaine-in-
duced myocardial infarction were found to have near-
occlusion of the infarction-related vessel, and 35% of all
patients with cocaine-induced myocardial infarction in the
absence of atherosclerotic coronary artery disease were
reported to have thrombus present.8
Repeated injections of cocaine in rabbits have been
shown to induce arterial inJury,32 and cocaine has been
shown to accelerate atherogenesis in cholesterol-fed rab-
bits.33 Marked thickening of small-vessel walls has been
observed on endomyocardial biopsies of patients with
cocaine-induced chest ~ a i n . 3Premature~ atherosclerosis
has been well documented in several autopsy series of
cocaine u ~ e r s . ~ Thirty-one
5-~~ percent of reported patients
with cocaine-induced myocardial infarction who had coro-
nary angiography were found to have atherosclerotic coro-
nary artery disease, despite an average age of 32 years.8
The ability of cocaine to increase myocardial oxygen
demand and decrease coronary flow reserve through coro-
nary artery vasoconstriction, enhanced platelet aggrega-
tion, in-situ thrombus formation, premature atherosclero-
sis, left ventricular hypertrophy,38*39 hypertension, and
tachycardia makes it an ideal precipitant of myocardial
ischemia.
The incidence of myocardial infarction in our series
was 5.7%. The four prior clinical studies of patients with
cocaine-associated chest pain found incidences of myocar-
dial infarction ranging from 0 to 31%.l0-l3 The wide
variation probably resulted from selection bias secondary
to variable inclusion criteria, as well as from the small
numbers of patients in prior series. We report an incidence
of 5.7% (95% CI, 2.7-8.7%) based on a large unselected
cohort of patients with cocaine-associated chest pain. This
incidence is comparable to the incidence of 4.2% reported
for traditional patients presenting with chest pain between
the ages of 25 and 39 y e a s 4
However, our study may have underestimated the true
incidence of myocardial infarction. We were unable to
obtain complete sets of cardiac isoenzymes for all patients.
If we were to limit our analysis to patients with two or more
cardiac isoenzymes, the incidence of infarction would
increase to 7.3% (95% CI, 3.2-11.4%). If we were to limit
our analysis to only admitted patients (as other studieslo-12
have done), 8.6%of patients (95% CI, 3.9-13.2%) would
have myocardial infarctions. Although telephone follow-up
was performed for patients released from the ED, such
follow-up probably would not detect all instances of missed
myocardial infarction, especially given the apparent low
frequency of complications.
On the other hand, we cannot exclude the possibility
that some eligible patients were not enrolled in this study.
Clinicians may not have notified the investigators of pa-
tients considered to have very low risk for myocardial
infarction. If the clinicians were able to accurately identify
low-risk patients, the frequency of infarction reported in
this series would represent an overestimate.
We chose a history of cocaine use within the past week
as our cutoff for enrollment. Laboratory investigations
have demonstrated cocaine-induced coronary vasocon-
striction for only 90 minutes after cocaine use.24 Myocar-
dial infarction has been attributed to cocaine use several
days earlier,**41and cocaine-induced ischemia has been
documented by Holter monitor for several weeks following
cocaine use.42
In the present series, we could not identify any demo-
graphic or historical factor that could reliably predict or
exclude myocardial infarction (see Table 1). Location of
chest pain, duration of chest pain, quality of chest pain,
and symptoms associated with chest pain were not predic-
tive of infarction (see Table 2). This has been well studied
for the traditional patient presenting to the ED with chest
pain ,40.4345 and although low-risk groups have been de-
fined for patients to be admitted to intermediate care
units,46 a group of patients who can be safely released
based upon explicit criteria has not been clearly identified.
The ECG, though insensitive, has historically been the
best predictor of myocardial infarction.40 In our series, the
patients with infarction were more likely than those with-
out infarction to have initial ECGs consistent with isch-
emia or infarction (p = 0.014). However, the sensitivity
was only 35.7% and the positive predictive value was only
17.9% for predicting infarction. Patients with proven myo-
 ~ ~~~~
Cocaine-associated Chest Pain, Hollander et af.
cardial infarctions were as likely to present with normal or
nonspecific ECGs as with ischemic ECGs (see Table 4).
Though the ECG may still be the most sensitive test for
identifying patients with cocaine-induced myocardial in-
farction, it appears to be even less sensitive in this popula-
tion than in other patients presenting with chest ~ a i n . 4 ~
The addition of serial CK-MB testing to the ED evalua-
tion of patients with chest pain and non-diagnostic ECGs
has been shown to increase the sensitivity of acute myocar-
dial infarction detection to 80%.47In our series, ten of the
14 patients (71%) who had infarction had elevated initial
CK-MB levels. The routine use of an ED CK-MB deter-
mination may facilitate the early identification of some
patients who have cocaine-induced myocardial infarctions.
If accelerated atherosclerotic heart disease occurs sec-
ondary to cocaine, then it is reasonable to hypothesize that
patients with ischemic chest pain who are ruled out for
myocardial infarction may be at higher risk than the aver-
age population for the future development of ischemic
heart disease, even in the absence of continued cocaine use.
A history of prior chest pain (51%) or infarction (4%) was
common. Recurrent chest pain after cocaine-induced myo-
cardial infarction has been reported in patients who contin-
ue to use co~aine,Z.48-~~ as well as in those who abstain.49
The long-term prognosis and the appropriate diagnostic
evaluation of patients with cocaine-induced ischemia are
unclear. I11 a previous review,8 we were unable to find any
report of a patient with a cocaine-induced myocardial
infarction who had a positive stress test. Only one of eight
cocaine addicts in a detoxification center who had transient
ST-segment elevations on Holter monitoring had a positive
$tress test.42 In our series, no patient to undergo testing,
including six patients with myocardial infarction, had an
exercise treadmill test demonstrating myocardial isch-
emia. No patient in our series received cardiac catheteriza-
tion. However, recurrent myocardial infarction has even
been reported to occur in patients with normal coronary
angiograms who use cocaine.8 Long-term longitudinal
studies are needed to define the appropriate diagnostic
evaluation for this cohort of patients.
I FUTURE QUESTIONS
Future investigations should address the following issues:
1 . Do patients with cocaine-associated myocardial in-
farction have congestive heart failure, arrhythmias, or
infarction extension with more or less frequency than do
other patients with myocardial infarction? We observed no
life-threatening complications other than those present on
arrival to the ED; however, the size of our cohort with
infarction was small (n = 14), yielding a 95% CI of 0-21%
for the development of late complications.
2. Are patients who experience chest pain following
cocaine use at high risk for myocardial infarction in the
337
future? Patients without histories of prior chest pain were
more likely to have infarction. Is new-onset chest pain in
cocaine users analogous to new-onset chest pain (unstable
angina) in patients with atherosclerotic heart disease?
3 . Is short-term admission to an ED observation unit
appropriate for this patient cohort? If the risk of complica-
tions is low, a short-duration evaluation protocol might be
cost-effective.
4. What is the optimal treatment for patients with
cocaine-associated chest pain? Are sedative agents (i.e.,
benzodiazepines) of value for treating the symptoms of
cocaine-associated chest pain?
I CONCLUSIONS
Patients presenting to the ED with cocaine-associated
chest pain have approximately a 5.7% incidence of myocar-
dial infarction. No clinical parameter available to the
physician can adequately identify patients at very low risk
for myocardial infarction. As the consequences of releas-
ing a patient with myocardial infarction (unrelated to
cocaine) are ~ i g n i f i c a n t ,it~ ~is reasonable to rule out
myocardial infarction in all patients complaining of chest
pain within 24 hours of cocaine use. Given the apparent
low risk of complications, it may be appropriate to observe
most patients with cocaine-associated chest pain in an
intermediate care or short-term ED observation unit.
I REFERENCES
1 . National Institute of Drug Abuse. National household survey on drug
abuse. Population estimates, 1991. DHHS number (ADM) 92-1887,
1992.
2. Cregler LL, Mark H. Relation of acute myocardial infarction to
cocaine abuse. Am J Cardiol. 1985; 56:794.
3. Goldfrank LR, Hoffman RS. The cardiovascular effects of cocaine.
Ann Emerg Med. 1991; 20:165-75.
4. MacDonald DI. Cocaine heads ED drug visits. JAMA. 1987;
258:2029.
5 . Colliver JD, Kopstein AN. Trends in cocaine abuse reflected in
emergency room episodes reported to DAWN. Public Health Rep.
1991; 106:59-67.
6. Brody SL, Slovis CM, Wrenn KD. Cocaine-related medical prob-
lems: consecutive series of 233 patients. Am J Med. 1990; 88:325-
31.
7. Coleman DL, Ross TF, Naughton JL. Myocardial ischemia and
infarction related to recreational cocaine use. West J. Med. 1982;
136:444-6.
8. Hollander JE, Hoffman RS. Cocaine induced myocardial infarction:
an analysis and review of the literature. J Emerg Med. 1992; 10:169-
77.
9. Minor RL, Scott BD, Brown DD, Winniford MD. Cocaine-induced
myocardial infarction in patiens with normal coronary arteries. Ann
Intern Med. 1991; 115:797-806.
10. Zimmerrnan JL, Dellinger RF', Majid PA. Cocaine associated chest
pain. Ann Emerg Med. 1991; 20:611-5.
11. Gitter MJ, Goldsmith SR, Dunbar DN, Sharkey SW. Cocaine and
chest pain: clinical features and outcome of patients hospitalized to
rule out myocardial infarction. Ann Intern Med. 1991; 115:277-82.
 ~~
338 ACADEMIC EMERGENCY MEDICINE JULYlAUGUST 1994 VOL liNO 4
12 Amin M, Gableman G, Karpel J, Buttrick P. Acute myocardial
infarction and chest pain syndromes after cocaine use. Am J Cardiol.
1990; 66:1434-7.
13 Tokarski GN, Paganussi P, Urbanski R, et al. An evaluation of
cocaine induced chest pain. Ann Emerg Med. 1990; 19:1088-92.
14, Harrell FE Jr, Lee KL, Matcher DB, Reichert TA. Regression models
for prognostic prediction: advantages, problems, and suggested solu-
tions. Cancer Treat Rep. 1985; 69:1071-7.
15. Rollinger IM. Belzberg AS, MacDonald IL. Cocaine induced myo-
cardial infarction. Can Med Assoc J. 1986; 135:45-6.
16. Wehbie CS, Vidaillet HJ, Navetta FI. Peter RH. Acute myocardial
ihfarction associated with initial cocaine use. South Med J. 1987;
80:933-4.
17. Chiu YC, Brecht K, DasGupta DS, Mhoon E. Myocardial infatction
with topical cocaine anesthesia for nasal surgery. Arch Otolaryngol
Head Neck Surg. 1986; 112:988-90.
18. Ross GS, Bell J. Myocardial infarction associated with inappropriate
use of topical cocaine as treatment for epistaxis. Am J Emerg Med.
1992; 10:219-22.
19. Zimmerman FH, Gustafson GM, Kemp HG. Recurrent myocardial
infarction associated with cocaine abuse in a young man with normal
coronary arteries: evidence for coronary artery spasm resulting in
thrombosis. J Am Coll Cardiol. 1987; 9:964-8.
20. Ascher EK, Stauffer JC, Gaasch WH. Coronary artery spasm,
cardiac arrest, transient electrocardiographic Q waves and stunned
myocardium in cocaine associated acute myocardial infarction. Am J
Cardiol. 1988; 61:939-41.
21. Lange RA, Cigarroa RG, Yancy CW. et al. Cocaine induced coronary
artery vasoconstriction. N Engl J Med. 1989; 321:1557-62.
22. Lange RA, Cigarroa RG. Flores ED, et al. Potentiation of cocaine
induced coronary vasoconstriction by Beta-adrenergic blockade.
Ann Intern Med. 1990; 112:897-903.
23 Flores ED, Lange RA, Cigarroa, RC, Hillis LD. Effect of cocaine on
coronary artery dimensions in artherosclerotic coronary artery dis-
ease: enhanced vasoconstriction at sites of significant stenosis. J Am
Coll Cardiol. 1990; 16:74-9.
24, Brogan WC, Lange RA, Glamann B, Hillis LD. Recurrent coronary
vasoconstriction caused by intranasal cocaine: possible role for
metabolites. Ann Intern Med. 1992; 116:556-61.
25. Winniford MD, Wheelan KR, Kremers MS, et al. Smoking-induced
coronary vasoconstriction in patiens with atherosclerotic coronary
artery disease: evidence for adrenergically mediated alterations in
coronary artery tone. Circulation. 1986; 73:662-7.
26. Minor RL, Oskarrson H, Quillen JE. Rossen JD, Winniford MD.
Intracoronary phentolamine reverses cigarette smoking-induced epi-
cardial coronary vasoconstriction in humans. J Am Coll Cardiol.
1992; 19:285A.
27. Togna G , Tempesta E, Togna AR. Dolci N, Cebo B, Caprino I.
Platelet responsiveness and biosynthesis of thromboxane and pros-
tacyclin in response to in-vitro cocaine treatment. Haemostasis.
1985; 15:100-7.
28. Schnetzer GW. Platelets and thrombogenesis: current concepts. Am
Heart J. 1972; 8 3 3 2 - 6 4 .
29. Rezkalla S. Mazza JJ. Kloner RA, Tillema V, Chang SH. The effect
of cocaine on human platelets. Am J Cardiol. 1993; 72:243-6.
30. Fritsma GA. Leikin JB, Maturen AJ, Froelich CJ, Hryhorczuk DO.
Detection of anticardiolipin antibody in patients with cocaine abuse.
J Emerg Med. 1991; 9:37-43.
31. Stenberg RG, Winniford MD, Hillis LD, Dowling GP. Buja M.
Simultaneous acute thrombosis of two major coronary arteries fol-
lowing intravenous cocaine use. Arch Pathol Lab Med. 1989; 113:
521-524.
~
32. Bement CL, Cohen L, Nielson SW, Langer RO. Arterial injury in
rabbits following alternate day injections of cocaine. FASEB 1. 1989;
3:A297.
33. Langner RO, Bement CL, Cohen L, Nielson SW. Stimulation of
atherogenesis by cocaine in cholesterol fed rabbits. FASEB J. 1989;
3:A297.
34. Majid PA, Patel B, Kim H, Zimmerman JL, Dellinger RR An
angiographic and histologic study of cocaine induced chest pain. Am
J Cardial. 1990; 65:812-4.
35. Mittleman RE, Wetli CV. Cocaine and sudden “natural” death. J
Forensic Sci. 1987; 32:ll-9.
36. Dressler FA, Malekzadeh S, Roberts WC. Quantitative analysis of
amounts of coronary arterial narrowing in cocaine addicts. Am J
Cardiol. 1990; 65:303-8.
37. Tardiff K. Gross E, Wh J, Stajic M. Millman R. Analysis of cocaine
positive fatalities. J Forensic Sci. 1989; 3453-63.
38. Eichhorn EJ, Peacock E, Grayburn PA, et al. Chronic cocaine abuse
in association with accelerated atherosclerosis in human coronary
arteries. J Am Coil Chrdiol. 1992; 19:lOSA.
39. Brickner ME, Willard JE, Eichhorn EJ, Black J, Grayburn PA. Left
ventricular hypertrophy associated with chronic cocaine use. Circu-
lation. 1991; 84:1130-/5.
40. Lee TH, Cook EF, Wqisberg M , Sargent RK, Wilson C, Goldman L.
Acute chest pain in the emergency room: identification and examina-
tion of low risk patients. Arch Intern Med. 1985; 145:65-9.
41 * Del Aguila C , Rosman H. Myocardial infarction during cocaine
withdrawal. Ann Intern Med. 1990; 112:712.
42 Nademanee K, Gorelick DA, Josephson MA, et al. Myocardial
ischemia during cocaine withdrawal. Ann Intern Med. 1989;
111:876-80.
43 Goldman L. Weinberg M, Weisberg M, et al. A computer derived
I
protocol to aid in the diagnosis of emergency room patients with
acute chest pain. N Engl J Med. 1982; 307:588-96.
44 Pozen MW, D’Agostino RB, Selker HP, Sytkowski PA, Hood WB. A
I
predictive instrument to improve coronary care unit admission prac-
tices in ischemic heart disease: a prospective multicenter clinical
trial. N Engl J Med. 1984; 310:1273-8.
45. Goldman L, Cook EF, Brand DA, et al. A computer protocol to
predict myocardial infarction in emergency department patients with
chest pain. N Engl J Med. 1988; 318:797-803.
46. Lee T, Juarez G , Cook EF, et al. Ruling out acute myocardial
infarction: a prospective multicenter validation of a 12 hour strategy
for patients at low risk. N Engl J Med. 1991; 324A239-46.
47. Gibler WB. Young GP, Hedges JR, et al. Acute myocardial infarction
in chest pain patients with nondiagnostic ECG’s: serial CPK-MB
sampling in the emergency department. Ann Emerg Med. 1992;
21~504-12.
48. Kossowsky WA, Lyon AF, Chou SY. Acute non-Q-wave cocaine
related myocardial infarction. Chest. 1989; 96:617-21.
49. Smith HWB, Liberman, HA, Brody SL, et al. Acute myocardial
infarction temporally related to cocaine use: clinical, angiographic,
and pathophysiologic observations. Ann Intern Med. 1987; 107:13-8.
50. Kosowsky WA, Lyon AE Cocaine and acute myocardial infarction: a
probable connection. Chest. 1984; 86:729-31.
51. Zimmerman FH, Lipton B. Cocaine and MI. Chest. 1987; 92:767.
52. Lam D, Goldschlager N. Myocardial injury associated with poly-
substance abuse. Am Heart J. 1988; 115:675-80.
53. Lee TH, Rouan GW, Weisberg MC, et al. Clinical characteristics and
natural history of patients with acute myocardial infarction sent home
from the emergency room. Am J Cardiol. 1987; 60:219-24.

Cocaine-associated Chest Pain, Hollander et al.
A
APPENDIX
The following persons and institutions participated in
this study (lhe start date and number of patients enrolled
are denoted in parentheses):
Bronx Municipal Hospital Center, Bronx, New York
(April 30, 1991; n = 99): Judd E. Hollander, Paul Gennis,
Phillip Fairweather, Michael Lozano, E. John Gallagher;
New York City Poison Control Center, New York, New
York (July 1,1991; n = 13): Robert S. Hoffman; Highland
General Hospital, Oakland, California (January 1, 1992;
n = 57): Michael J. DiSano, David A. Schumb, Andrew
Levitt; Boston City Hospital, Boston, Massachusetts (Jan-
uary 29,1992; n = 45): James A . Feldman, Susan S. Fish,
Sophia Dyer, Nancy J. Battinelli; University of Rochester
Medical Center, Rochester, New York (October 25, 1991;
n = 16): Paul Wax; Queens Hospital Center, Queens, New
York (December 1, 1991; n = 16): Chris Whelan, Evan
Schwarzwald; and State University of New York Health
Sciences Center, Stony Brook, New York (statistician):
Henry Thode (statistical consultant).
Further Thoughts from the Reviewers
I This prospective multicenter observational cohort
study of chest pain patients with self-reported histories of
cocaine use is the most comprehensive analysis of this
challenging clinical problem to date. As mentioned by the
authors, the clinical presentation of those patients subse-
quently found to have myocardial infarction varied little
from that of the greater patient population ruled out for
myocardial infarction. In part, this may have been due to
the inclusion of significant numbers of patients with isch-
emic chest pain in the non-infarction group.
~ ~~ ~
339
The authors have identified the limitations of their
study well. While attempting to enroll a consecutive sam-
ple of patients, it is likely that their population represents a
sample of those patients believed by the evaluating clini-
cians to be most likely to have significant chest discomfort
in association with cocaine use. This selection bias applies
equally to other studies of chest pain patients requiring
informed consent, detailed data collection, observation,
and clinical follow-up. Further, although the authors report
only one refusal to participate, it seems unlikely that most
patients would admit to an illegal activity (i.e., cocaine
use) and agree to such a study unless thay were quite
symptomatic. In addition, a number of patients did not
return for follow-up studies and evaluation, requiring the
authors to use the telephone or the mail for follow-up
information. One patient with elevated CK-MB iso-
enzymes could not be located at follow-up. Nonetheless,
despite the enrollment and follow-up limitations, the au-
thors have collected a valuable data set for the evaluation of
the cocaine-using chest pain population.
The data suggest that myocardial infarction, while
infrequent, is certainly not unheard of in this patient
population. Further, cardiac complications are infrequent
and generally present on initial presentation. Cardiac en-
zymes appear useful in this population for identification of
infarction, although we were not provided with data on the
sensitivity of the initial CK-MB for infarction as a function
of interval duration from onset of discomfort until level
determination. Since patients may continue to use cocaine
after the onset of chest discomfort, the duration of isch-
emia may be considerably longer than the time since last
use of cocaine. This may explain the fairly good sensitivity
of the initial CK-MB for infarction in this study.
As noted by the authors, much remains to be deter-
mined regarding this patient population. The indications
for diagnostic studies and specific therapies in those pa-
tients with continued discomfort, electrocardiographic
changes, arrhythmias, elevated CK-MB isoenzymes, or
other significant findings remain to be determined.