Full text of DOI:10.1111/j.1553-2712.2001.tb01296.x

ACADEMIC EMERGENCY MEDICINE • March 2001, Volume 8, Number 3
211
BASIC INVESTIGATIONS
Cocaine, Ethanol, and Cocaethylene Cardiotoxity in an
Animal Model of Cocaine and Ethanol Abuse
L
ANCE D. WILSON, MD, JENNIFER
JEROMIN, MD,
L
EE ARVEY, MD, ALLISON ORBANDT, MD
G
D
Abstract. Objectives: Simultaneous abuse of co- The most dramatic hemodynamic changes occurred
caine and ethanol affects 12 million Americans an- after each cocaine bolus in the CϩE and C only
nually. In combination, these substances are substan- groups; however, persistent hemodynamic changes
tially more toxic than either drug alone. Their occurred in the CϩE group. Peak CE levels were as-
combined cardiac toxicity may be due to independent sociated with a 45% (SD Ϯ 22%, 95% CI = 22% to
effects of each drug; however, they may also be due 69%) decrease in cardiac output (p < 0.05), a 56% (SD
to cocaethylene (CE), a cocaine metabolite formed Ϯ 23%, 95% CI = 32% to 80%) decrease in dP/dtmax (p
only in the presence of ethanol. The purpose of this < .006), and a 23% (SD Ϯ 15%, 95% CI = 7% to 49%)
study was to delineate the role of CE in the combined decrease in SVO (p < 0.025). Ventricular arrhyth-
2
cardiotoxicity of cocaine and ethanol in a model sim- mias were primarily observed in the CϩE group, in
ulating their abuse. Methods: Twenty-three dogs which four of eight dogs experienced ventricular
were randomized to receive either 1) three intrave- tachycardia. Conclusions: Cocaine and ethanol in
nous (IV) boluses of cocaine 7.5 mg/kg with ethanol combination were more toxic than either substance
(
1 g/kg) as an IV infusion (CϩE, n = 8), 2) three co- alone. Co-administration resulted in prolonged car-
caine boluses only (C, n = 6), 3) ethanol infusion only diac toxicity and was dysrhythmogenic. Peak serum
E, n = 5), or 4) placebo boluses and infusion (n = 4). cocaethylene concentrations were associated with
(
Hemodynamic measurements, electrocardiograms, prolonged myocardial depression. Key words: co-
and serum drug concentrations were obtained at caine; ethanol; cocaethylene; cardiac arrhythmia;
baseline, and then at fixed time intervals after each myocardial depression. ACADEMIC EMERGENCY
drug was administered. Results: Two of eight dogs in MEDICINE 2001; 8:211–222
the CϩE group experienced cardiovascular collapse.
OCAINE and ethanol are the two most com- of sudden death when cocaine is used in combi-
1,2
C
monly abused drugs in the United States.
nation with ethanol than when it is abused alone
7
Twelve million Americans have used both cocaine in patients with coronary artery disease. Their
3
,4
and ethanol on a regular basis, and when cocaine combined use results in substantially more emer-
is abused, ethanol is also ingested 60–90% of the gency department (ED) visits requiring treatment
1
8
time. Cocaine and ethanol in combination have than any other drug combination. In addition, co-
been suggested to be more toxic than either drug caine and ethanol are the third most common fatal
5
,6
alone. There is a 21.5% increase in the incidence
substance abuse combination, behind cocaine and
heroin and heroin and alcohol, according to 1994
9
Drug Abuse Warning Network (DAWN) data. In
1
990, DAWN reported 1,005 deaths due to com-
From the Department of Emergency Medicine, MetroHealth
Medical Center, Cleveland, OH (LDW); Department of Emer-
8
bined cocaine and ethanol abuse, but in 1994,
gency Medicine, PHS–Mt. Sinai Medical Center, Cleveland, 1,471 cases were reported, and cocaine and alco-
9
OH (LDW, JJ); and Department of Emergency Medicine, Car-
hol-related deaths appear to be increasing. There
are numerous medical reports of sudden cardiac
olinas Medical Center, Charlotte, NC (LG, AD).
Received April 28, 2000; revision received October 23, 2000;
accepted October 31, 2000. Presented at the SAEM annual
meeting, Denver, CO, May 1996.
death related to combined cocaine and ethanol
10
abuse.
Supported by a Haas Foundation Grant from the Mount Sinai
Medical Center.
Address for correspondence and reprints: Lance D. Wilson,
MD, Department of Emergency Medicine, MetroHealth Medi-
cal Center, 2500 MetroHealth Drive, Cleveland, OH 44109.
Fax: 216-778-5349; e-mail: ldwilson@pol.net
The substantial toxicity of cocaine and ethanol
may be due to a synergistic interaction between
cocaine and ethanol. Cocaine produces both sym-
pathomimetic and local anesthetic cardiotoxic ef-
fects, and ethanol is also a myocardial depres-
11

212
COCAETHYLENE
Wilson et al. • COCAINE, ETHANOL, AND COCAETHYLENE CARDIOTOXICITY
1
2
sant. Other proposed causes of the combined travenous injection of alpha-chloralose, 100 mg/kg,
toxicity include alterations of cocaine metabolism dissolved in polyethylene glycol. Morphine was
1
3
in the presence of ethanol or additional interac- given approximately four hours prior to any data
tions with other drugs of abuse. collection. Alpha-chloralose anesthesia was chosen
The majority of individuals use these drugs four because of the drug’s ability to sustain autonomic
21
or more hours prior to the onset of symptoms and reflex control of the cardiovascular system. Sup-
1
4,15
presentation for emergency treatment.
This plemental doses of alpha-chloralose were given
prolonged time period from drug use to onset of throughout each experiment to maintain surgical
symptoms is surprising because the half-life of co- anesthesia.
1
6
caine is only one hour. To the best of our knowl-
After endotracheal intubation, each dog was
edge, the mechanisms of the delayed toxicity with ventilated on room air with a respirator (Harvard
combined cocaine and ethanol use are not known; Apparatus, Holliston, MA). At the initiation of each
however, the delayed toxicity observed clinically experiment, and every one hour thereafter, 1-mL
has raised suspicion that metabolites of these samples of arterial blood were collected for arterial
drugs may be involved.
blood gas analyses, which included pH and bicar-
Cocaethylene (CE) is a pharmacologically active bonate concentrations (Instrumentation Laborato-
cocaine metabolite that is formed in the liver by ries, Lexington, MA).
transesterification of cocaine in the presence of
The anterior neck and femoral areas were
1
0
ethanol. Cocaethylene is structurally similar to shaved and prepared with povidone–iodine solu-
1
6
cocaine and has similar pharmacologic effects on tion. The right femoral artery and vein as well as
central and peripheral neurotransmitters and on the right external jugular vein were isolated by
1
0,16–18
behavior.
Cocaethylene is equal to or greater surgical dissection. The left femoral vein was can-
than cocaine in producing addiction but more le- nulated with polyethylene tubing and the tubing
1
6–19
thal than cocaine
and has been found in the was connected to 5% dextrose and lactated
sera of patients with cocaine and ethanol toxic- Ringer’s solution. Sodium bicarbonate solution,
1
0,16,20
ity.
Most individuals who use cocaine and eth- 8.4%, was given intravenously to maintain normal
anol repeatedly binge on cocaine and drink etha- acid–base balance. The dextrose–Ringer–bicar-
nol, thereby perhaps accumulating high serum bonate solution was given to each animal at a rate
1
0,14
concentrations of CE.
Typically, abusers repeat- of 75 mL/hr to replace insensible and sensible fluid
edly administer cocaine because the euphoric ef- losses and to maintain the arterial pH between
fects of cocaine tend to be brief. They concomi- 7.35 and 7.45. The left femoral artery of each dog
tantly drink ethanol to prolong the euphoric effects was cannulated with a catheter that was advanced
of cocaine as well as ameliorate the dysphoria as- into the aorta to measure central arterial pressure.
1
0,14
sociated with cocaine withdrawal.
This catheter was connected to a fluid-filled P23
It was our aim to create an animal model that pressure transducer (Gould, Cleveland, OH). A
mimicked this pattern of drug abuse as well as cre- second catheter, with a microtip pressure trans-
ate clinically meaningful serum concentrations of ducer (Millar, Houston, TX), was advanced from
cocaine, ethanol, and CE. Such a model may be the left femoral artery into the left ventricle to
used to better delineate the role of CE in the com- measure the left ventricular systolic and diastolic
bined cardiotoxicity of cocaine and ethanol. It was pressures.
our hypothesis that the presence of CE would be
The entire left ventricular pressure waveform
associated with delayed myocardial toxicity due to of each of the left ventricular beats recorded by the
cocaine and ethanol.
Millar catheter was sampled at 500 Hz by an an-
alog-to-digital converter (Labview, National In-
struments, Austin, TX) and the digital information
was stored on the hard disk of a PC486 computer
M
ETHODS
Study Design. This was a prospective laboratory (Gateway, San Diego, CA). In each experiment, the
investigation. The experiments conformed to the rate of left ventricular pressure rise was deter-
guidelines established by the National Institutes mined as an index of left ventricular contractile
of Health for animal care and use and were ap- function (dP/dtmax). The rate of left ventricular
proved by the animal care and use committee of pressure decline was also determined as an index
our medical center.
of ventricular relaxation (dP/dtmin). Both dP/dtmax
and dP/dtmin were determined from the average of
Animal Subjects and Preparation. Experiments ten consecutive left ventricular pressure wave-
were performed on 23 mongrel dogs, with a mean forms that were recorded at each time interval.
body weight of 17.9 Ϯ 3.3 kg. Each dog was given
A fiberoptic, balloon-tipped catheter (Oximetric,
an intramuscular injection of morphine sulfate, 1.5 Abbott Laboratories, Mountain View, CA) was in-
mg/kg, for sedation and anesthetized with an in- serted into the right jugular vein of each dog and

ACADEMIC EMERGENCY MEDICINE • March 2001, Volume 8, Number 3
213
advanced into the main pulmonary artery. This dogs received ethanol, 1.5 g/kg in 5% dextrose in
catheter was connected to a P23 transducer (Gould) water solution (D W) as an intravenous bolus over
for measurements of pulmonary arterial pressure 20 minutes followed by an intravenous ethanol
and also pulmonary artery wedge pressure. drip at 20 mg/kg/hr in order to achieve a steady-
The fiberoptic catheter was also connected to an state ethanol level between 150 and 200 mg/dL. In
/CO computer (Oximetric) for continuous mea- the cocaine only group (C, N = 6), the dogs received
surements of mixed venous blood oxygen satura- W over 20 minutes and a maintenance fluid in-
tion and also measurements of cardiac output. fusion. Forty minutes after the D W bolus, the
5
SO
2
D
5
5
Cardiac output was determined by thermodilution dogs received 7.5 mg/kg of cocaine hydrochloride in
technique. An average of at least three measure- 0.9% normal saline as a 10-mL intravenous bolus.
ments of cardiac output that agreed within 10% The 7.5-mg/kg bolus was repeated two times at 30-
was used for the measurement of the cardiac out- minute intervals. In pilot experiments, we deter-
put.
mined that the dogs could easily tolerate three in-
An electrocardiogram (ECG) for heart rate mon- travenous boluses of 7.5 mg/kg of cocaine at
itoring and waveform analysis was obtained by at- 30-minute intervals with return of all hemody-
taching electrodes to the anterior chest wall of namic parameters to baseline (Wilson LD, unpub-
each dog in order to produce lead I of the standard lished data, 1993). In the cocaine ϩ ethanol group
ECG.
(CϩE, N = 8), the dogs received ethanol, 1.5 g/kg
Aortic pressure, left ventricular pressure, pul- in D W over 20 minutes, and were placed on an
5
monary artery and wedge pressures, and a surface ethanol infusion at 20 mg/kg/hr. Forty minutes af-
ECG were continuously monitored. The recording ter the ethanol bolus, the dogs received 7.5 mg/kg
system for the pressures was linear to a frequency of cocaine hydrochloride in 0.9% normal saline as
of 100 Hz. All hemodynamic and electrocardio- a 10-mL intravenous bolus. The 7.5-mg/kg bolus
graphic data were analyzed from an average of ten was repeated two times at 30-minute intervals. In
consecutive waveforms.
the control group (N = 4), the dogs received the
W bolus, maintenance infusion, and three pla-
D
5
Study Protocol. After completion of the instru- cebo boluses of 0.9% normal saline at 30-minute
mentation, each dog was allowed 30 to 60 minutes intervals. The doses of cocaine and ethanol were
for the monitored variables to achieve a steady chosen to produce serum concentrations in the
state. Each dog was then randomized to one of four study animals that would be comparable to the se-
groups. In the ethanol only group (E, N = 5), the rum cocaine, ethanol, and CE concentrations ob-
Figure 1. The plasma concentrations of ethanol in the cocaine ϩ ethanol and ethanol only groups as well as the
plasma concentrations of cocaethylene in the cocaine ϩ ethanol group are plotted as a function of time. Mean
ethanol concentrations are plotted in milligram percent and mean cocaethylene concentrations are plotted in nan-
ograms/milliliter. There were no statistically significant differences between ethanol levels in the two groups re-
ceiving ethanol.

214
COCAETHYLENE
Wilson et al. • COCAINE, ETHANOL, AND COCAETHYLENE CARDIOTOXICITY
Figure 2. The plasma concentrations of cocaine for the cocaine ϩ ethanol and cocaine only groups are plotted as
a function of time. There were no statistically significant differences in the cocaine levels in these two groups.
served in individuals who have used cocaine and until chemical analyses were performed. The
ethanol ‘‘recreationally’’ and who have experienced plasma cocaine, CE, and ethanol concentrations
toxic or fatal reactions to the drugs.
were determined by high-performance liquid chro-
22
When each dog achieved a steady state, base- matography (HPLC). Serum ethanol concentra-
line hemodynamic and electrocardiographic mea- tions were determined enzymatically with ethanol
surements were obtained. Data were then collected dehydrogenase and diphosphopyridine as coen-
23
from each dog at 30 and 60 minutes after baseline zyme.
measurements (10 and 40 minutes after complet-
ing ethanol or placebo bolus), 1, 2, 4, 6, 8, 15, and Data Analysis. Mean values and 95% confidence
3
0 minutes after each of the three cocaine or pla- intervals (95% CIs) are reported. Statistical signif-
cebo injections, then every 30 minutes for two icance was determined by repeated-measures anal-
hours, and then hourly for an additional two hours. yses of variance. Within any one treatment group,
Cardiac output determinations were performed in when group–time interactions were significant,
each dog at baseline, at 30 and 60 minutes after differences between specific means were deter-
baseline measurements (10 and 40 minutes after mined by Newman-Keuls post hoc test. In all our
completing ethanol or placebo bolus), 4–6, 8, 15, analyses, a value of p Յ 0.05 was considered sig-
and 30 minutes after each of the three cocaine or nificant.
placebo injections, then every 30 minutes for two
hours, and then hourly for an additional two hours.
In each dog, 4–5 mL blood samples for cocaine,
RESULTS
ethanol, and CE determinations were obtained at Two animals in the CϩE group experienced unsta-
baseline, at 30 and 60 minutes after baseline mea- ble dysrhythmias and died. They were excluded
surements (10 and 40 minutes after completing from analysis. One of these dogs experienced
ethanol or placebo bolus), 2, 8, 15, and 30 minutes marked aberrancy and sustained wide-complex
after each of the three cocaine or placebo injec- tachycardia immediately after the first cocaine bo-
tions, then every 30 minutes for two hours, and lus, which briefly improved. This dog then experi-
then hourly for an additional two hours.
enced torsades de pointes at 15 minutes after the
Each blood sample was immediately placed in first cocaine bolus and died. The other animal,
a chilled glass tube that contained potassium ox- shortly after the third cocaine bolus, also experi-
alate and sodium fluoride. The blood was then cen- enced marked aberrancy for 3 minutes and then
trifuged for 10 minutes. The plasma from each became asystolic and died. No other explanation
specimen was removed and placed in a second tube for their deaths was apparent other than a reac-
containing potassium oxalate and sodium fluoride. tion to the study drugs. No other study animals
The samples were then frozen at Ϫ80Њ centigrade died prematurely.

ACADEMIC EMERGENCY MEDICINE • March 2001, Volume 8, Number 3
215
Serum Cocaine, Ethanol, and CE Concentra- Ϯ 12%, 95% CI = 68% to 93%) from baseline (p <
tions. Serum ethanol levels were similar in the 0.05). By post hoc testing, the depression in dP/
two groups receiving ethanol, and serum cocaine dtmax observed after the cocaine boluses between
concentrations were also similar in the two groups the CϩE and the C only groups were similar. In
receiving cocaine (Figs. 1 and 2). Cocaethylene the ethanol only group, the dP/dtmax was depressed
formed early in the CϩE group, with the first de- relative to the control group with prolonged etha-
tectable serum concentration of CE observed at 2 nol administration. There were no significant
minutes after the first cocaine bolus in two of the changes over time in the control group.
six dogs receiving cocaine and ethanol. Cocaethyl-
Cocaine and alcohol also depressed ventricular
ene was detected by 90 minutes in all dogs receiv- relaxation as measured by the changes in dP/dtmin
ing both substances. No CE was detected in the (not shown). The depression in dP/dtmin paralleled
animals receiving only cocaine or only ethanol (Fig. those changes observed in dP/dtmax. The most sig-
1).
nificant depression in dP/dtmin was observed 1–4
minutes after the cocaine boluses in the CϩE and
Hemodynamic Effects. Hemodynamic changes C only groups (p < 0.05). The maximal decreases
were most pronounced during the first 30 minutes in dP/dtmin in the CϩE and C only group were 65%
after each intravenous cocaine bolus in both (SD Ϯ 21%, 95% CI = 43% to 88%) and 60% (SD Ϯ
groups receiving cocaine. These initial hemody- 19%, 95% CI = 40% to 80%) from baseline, respec-
namic effects were similar in the CϩE and the co- tively. By post hoc testing the changes observed in
caine alone groups. However, primarily in the CϩE these two groups were not significantly different.
ethanol group, prolonged hemodynamic changes In the E only group there was a significant de-
were observed.
crease in dP/dtmin at the end of the observation pe-
In all three study groups, some depression in riod. There were no significant changes in dP/dtmin
ventricular contractile function, as measured by in the control group.
reductions in left ventricular dP/dtmax, were ob-
The decreases in dP/dtmax and dP/dtmin were ac-
served. The most prolonged depression in dP/dtmax companied by substantial reductions in the cardiac
observed in the CϩE group (Fig. 3). In this group, output and ventricular stroke volume. The signif-
dP/dtmax returned to baseline at 15 minutes after icant decreases in cardiac output were primarily
the first cocaine bolus, but after the second cocaine observed in the CϩE group (Fig. 4). After the first
bolus, it remained depressed from baseline for the cocaine bolus, cardiac output decreased in the CϩE
rest of the observation period. The dP/dtmax in the group and remained depressed below baseline.
CϩE group was depressed by as much as 80% (SD Cardiac output was maximally decreased from
Figure 3. Changes in ventricular contractility, as measured by dP/dtmax in dogs. In this and all subsequent figures,
measurements were made in all dogs in each group at each time interval. In this and all subsequent figures, a
data point marked with * represents a statistically significant difference (p < 0.05) from the control group, while
†
represents a statistically significant difference from baseline and ‡ represents a statistically significant difference
from the other groups by analysis of variance.

216
COCAETHYLENE
Wilson et al. • COCAINE, ETHANOL, AND COCAETHYLENE CARDIOTOXICITY
Figure 4. Changes in cardiac output. The baseline values at 0 minutes for each group have been normalized to
00%. The decrease in cardiac output relative to control in all groups at time 390 can be partially explained by an
1
increase in cardiac output at this time in the control group. A data point marked with * represents a statistically
significant difference (p < 0.05) from the control group, while † represents a statistically significant difference from
baseline by analysis of variance.
baseline by 45% (SD Ϯ 22%, 95% CI = 22% to 69%) 5). These changes were significantly different from
at 210 minutes, at which time serum CE concen- the cocaine group by post hoc testing. There was
trations were also maximal.
also a significant decrease in MAP immediately af-
The depression in cardiac output was similar to ter the third cocaine bolus by 32% (SD Ϯ 15%, 95%
the depression observed in ventricular stroke vol- CI = 17% to 47%) from baseline in the cocaine only
ume; however, there was a decrease in stroke vol- group (p < 0.001). There were no significant
ume observed in all study groups (not shown). In changes in the ethanol or control groups.
the CϩE group, ventricular stroke volume signifi-
Cocaine and ethanol in combination caused the
cantly decreased after the first cocaine bolus, and greatest decreases in the mixed venous oxygen sat-
remained depressed throughout the entire obser- uration (Fig. 6). In the CϩE group, there were sig-
vation period, maximally by 60% (SD Ϯ 16%, 95% nificant decreases in SVO when compared with
2
CI = 44% to 77%) from baseline (p < 0.028), with the control group after the second and third co-
the exception of the 270-minute measurement. caine boluses. These changes returned to baseline,
Stroke volume was also decreased in the cocaine then the SVO again dropped below control and re-
2
only group, beginning at 105-minute measurement mained significantly decreased over the remaining
and basically persisting through 240 minutes max- observation period, maximally decreased from
imally decreased by 24% (SD Ϯ 38%, 95% CI = 16% baseline by 23% (SD Ϯ 15%, 95% CI = 7% to 49%)
to 65%) from baseline in this group (p < 0.012). In at the same time serum CE concentrations were
the ethanol only group, the stroke volume was de- highest (p < 0.025).
pressed at the 105-, 120-, 150-, and 180-minute
measurements relative to the control group. These Effects on the ECG. After the first bolus in the
relative decreases in stroke volume occurred dur- CϩE and the C only groups, there was a significant
ing increases in stroke volume observed in the con- increase in heart rate in the first 2 minutes after
trol group. The stroke volume then declined sig- the first cocaine bolus by 38 beats/min (SD Ϯ 40,
nificantly at 330 and 390 minutes.
95% CI = Ϫ4 to 80 beats/min) and 48 beats/min
The major observed changes in mean arterial (SD Ϯ 23, 95% CI = 24 to 71 beats/min) from base-
pressure (MAP) were significant decreases in MAP line, respectively (p < 0.04) (not shown). There was
in the CϩE group as compared with control, im- also a significant increase in heart rate at 330 and
mediately after the third cocaine bolus and at the 390 minutes in the ethanol only group by 60 (SD
final measurement, maximally by 33% (SD Ϯ 28%, Ϯ 41, 95% CI = Ϫ4 to 98 beats/min) from baseline.
95% CI = 4% to 62%) from baseline (p < 0.02) (Fig. Although the combination of cocaine and ethanol

ACADEMIC EMERGENCY MEDICINE • March 2001, Volume 8, Number 3
217
significantly depressed dP/dtmax, dP/dtmin, and sist (Figs. 7 and 8). There were no significant pro-
stroke volume, it did not result in prolonged in- longations of QRS or QTc interval duration in the
creases in heart rate.
ethanol or control groups. By post hoc testing,
Cocaine ϩ ethanol and cocaine alone caused there were no significant differences produced by
prolongations of the PR, QRS, and QTc intervals of cocaine and ethanol vs those produced by cocaine
the surface ECG. These changes occurred imme- alone. Statistically significant PR interval prolon-
diately after the cocaine boluses and did not per- gation occurred only in the CϩE group at 1 and 2
Figure 5. The changes in the mean arterial pressure (MAP). A data point marked with * represents a statistically
significant difference (p < 0.05) from the control group, while † represents a statistically significant difference from
baseline analysis of variance.
Figure 6. Changes in mixed venous oxygen saturation. A data point marked with * represents a statistically
significant difference (p < 0.05) from the control group, while † represents a statistically significant difference from
baseline and ‡ represents a statistically significant difference from the other groups by analysis of variance.

218
COCAETHYLENE
Wilson et al. • COCAINE, ETHANOL, AND COCAETHYLENE CARDIOTOXICITY
Figure 7. Changes in QRS interval duration. The nonsignificant increase in QRS duration at 61 minutes in the
ethanol only group reflects a short run of ventricular tachycardia at that time in one dog. A data point marked
with * represents a statistically significant difference (p < 0.05) from the control group, while † represents a
statistically significant difference from baseline by analysis of variance.
minutes after the second and third cocaine bolus tients who have complications or who died due to
1
0,16,20
maximally by 30% (SD Ϯ 22%, 95% CI = 5% to cocaine and ethanol.
In such patients, serum
cocaine levels range from 100 to 24,000 ng/mL, av-
52%) from baseline (p = 0.014).
24,25
In six dogs in the CϩE group and five dogs in eraging 5,300–6000 ng/mL.
Our data are con-
the cocaine only group, transient ECG ST-segment sistent with observational studies of ED patients
changes occurred after the cocaine boluses. How- who have used cocaine and ethanol in which CE
ever, the ST-segment changes occurred only when levels have been reported in the range of 3–622
14,26
the QRS intervals of the ECG were significantly ng/mL
and are typically lower than those of co-
prolonged. No dog developed persistent ST-seg-
ment changes or ECG evidence of acute myocardial
infarction.
caine. Much higher serum CE concentrations have
been reported, exceeding 2,300 ng/mL. Autopsy
and ED studies have reported levels of CE exceed-
10,16,20
Dysrhythmias were observed primarily in the
dogs receiving cocaine and ethanol. In the CϩE
group, both animals that died experienced marked
aberrancy followed by ventricular arrhythmias and
asystole. Two of the six animals that completed the
protocol experienced sustained ventricular tachy-
cardia, lasting 5 minutes in one dog and persisting
intermittently over a 15-minute period in another.
In the cocaine only group, two of six experienced
frequent premature ventricular contractions and
another animal experienced Mobitz type II atrio-
ventricular block. One animal in the ethanol only
group experienced a short run of ventricular tachy-
cardia. No arrhythmias were observed in the con-
trol group.
ing those of cocaine.
The wide range of re-
ported CE serum concentrations may be due to in-
dividual variability in drug metabolism and
27
intervals between drug use and presentation.
The concentration of CE at the time of symptoms
or death may be much greater than the actual con-
16
centrations reported. Because of a long half-life,
CE may gradually reach high serum concentra-
tions in individuals who repeatedly ‘‘binge’’ on co-
16,27
caine and ethanol.
One proposed mechanism for increased cardiac
toxicity of CϩE is that cocaine metabolism is de-
creased in the presence of ethanol, creating higher
3,4,14
serum concentrations of cocaine.
This may be
due to increased bioavailability of cocaine in the
presence of ethanol, and the route and order of ad-
13,14,28,29
DISCUSSION
ministration may play a role.
We did not ob-
serve a statistically significant increase in serum
We observed cocaine, ethanol, and CE serum con- cocaine levels in the dogs receiving concomitant
centrations consistent with those reported in pa- ethanol and do not believe that increased serum

ACADEMIC EMERGENCY MEDICINE • March 2001, Volume 8, Number 3
219
cocaine levels explain the hemodynamic or electro- shown to interfere with cardiac potassium chan-
cardiographic changes we observed. nels, affecting ventricular repolarization and pro-
Cocaine, ethanol and CE all interfere with the longing the QTc interval. In addition, CE has been
normal electrical conduction in the heart. Cocaine shown to block calcium channels in cardiac myo-
causes conduction abnormalities through its sym- cytes, which may partially explain the prolonga-
pathomimetic as well as its local anesthetic effects. tion of the myocardial refractory period and resul-
32
Cocaine is a potent blocker of fast sodium channels tant prolongation of the QTc interval.
in cardiac myocytes, which has been postulated to The present data suggest that CϩE and CE do
be a possible mechanism of cocaine-related dys- not significantly increase heart rate and may de-
3
0,31
rhythmias and sudden death.
Cocaethylene af- press cardiac compensatory chronotropic re-
fects the cardiac conduction system and myocytes sponses. Cocaine and CE have been shown to de-
3
2
by blocking sodium channels and is a more potent crease depolarization at the sinoatrial node,
30
blocker of sodium channels than cocaine. The possibly by depressing the slowly activated current
33
substantial acute prolongations of QRS intervals in the phase 4 action potential. Cocaine increases
in the present experiments were comparable in heart rate at lower serum concentrations due to its
dogs receiving cocaine and CϩE, suggesting that sympathomimetic effects and inhibition of cardiac
34
after the acute bolus, cocaine-induced sodium muscarinic cholinergic receptors. We observed a
channel blockade was the predominant factor. brief increase in heart rate due to cocaine alone,
However, QRS and QTc intervals did progressively consistent with other repeated bolus models in
35
increase after each bolus, as CE accumulated in which tachyphylaxis has occurred. In the CϩE
the serum. This effect may be secondary to a fur- group, the prolonged suppression of stroke volume
ther interaction of CE with the sodium channel or and decrease in contractility should have produced
increasing cocaine effects as cocaine concentra- a compensatory increase in heart rate. We do not
tions were also rising. No abnormalities of conduc- believe that chloralose anesthesia blunted the ef-
tion were observed after the acute boluses, when fect of cocaine, ethanol, or CE on the sinoatrial
21
serum cocaine levels were decreasing and CE con- node. Although anesthesia may blunt the adre-
centrations were greatest. It is possible that the nergic response to either cocaine or CE, we have
relatively low serum concentrations of CE ob- documented that significant tachycardia occurs in
1
1
served were not sufficient to cause observable dis- chloralose-anesthetized dogs given cocaine.
turbances of cardiac conduction.
3
0
Dysrhythmias were observed primarily in the
Both PR and QTc prolongations may be related cocaine and ethanol group. Sodium channel block-
to sodium channel blockade by cocaine, enhanced ade can result in reentry dysrhythmias and ven-
by the presence of ethanol or CE. Cocaine has been tricular tachycardia and could be a factor in dys-
Figure 8. Changes in QTc interval duration. A data point marked with * represents a statistically significant
difference (p < 0.05) from the control group, while † represents a statistically significant difference from baseline
by analysis of variance.

220
COCAETHYLENE
Wilson et al. • COCAINE, ETHANOL, AND COCAETHYLENE CARDIOTOXICITY
rhythmias secondary to cocaine or CE. QTc effect of each substance’s individual myocardial
5
prolongation occurred primarily in the CϩE group. toxicity, while others have suggested a synergistic
7,11
The prolonged QRS and QTc intervals may be due effect.
We observed comparable acute toxicity
to cocaine’s ability to block the delayed potassium due to CϩE and cocaine. However, we observed
rectifier current, prolonging the action potential prolonged depression of cardiac function only in
and inducing early afterdepolarizations as well as the CϩE group, greater than could be attributed
3
6,37
potassium-mediated repolarizations.
This could to the effects of cocaine or ethanol alone. In this
be a factor in CϩE-related ventricular dysrhyth- group, prolonged decreases in myocardial function
mias. Cocaine may induce dysrhythmias by its correlated with the highest CE levels, suggesting
sympathomimetic effects. However, we observed that the toxicity was in part due to CE. However,
only premature ventricular contractions in the co- ethanol alone most certainly was a factor in the
caine group, and no lethal dysrhythmias. We be- later changes observed in the CϩE group. We ob-
lieve that the dysrhythmias observed in the CϩE served decreases in cardiac function at the end of
group are likely related to an interaction between the observation period due to ethanol alone. The
these two drugs rather than cocaine alone.
toxicity observed due to CϩE is most likely mul-
Cocaine, CE, and ethanol all impair myocardial tifactorial, as the effects of CE and cocaine itself in
contractility. Cocaethylene has more negative ion- the prolonged presence of ethanol may have all
3
2
otropic effects on myocytes than cocaine. Both co- contributed to the cardiac toxicity observed.
caine and CE, by blocking sodium channels, cause
significant abnormalities in ventricular depolari-
zation and, by interfering with the normal pat-
terns of ventricular depolarization and repolari-
zation, affect contractility and relaxation. Both
cocaine and CE, because of their sodium channeϩl-
blocking effects, reduce available sodium for Na /
L
IMITATIONS AND
FUTURE UESTIONS
Q
It is possible that the dog may not form CE as
1
3
readily as does the human. The present study
was designed to achieve higher serum concentra-
tions of CE than were measured. If dogs do not
form CE as readily as humans do, the represen-
tative effects of CE might be underestimated in
this study. Certainly performing the study in an-
other species that may form CE at higher concen-
trations might improve the model; however, CE
formation in other large intact animal models has
not been studied.
These experiments were performed under an-
esthesia, which certainly may have confounded the
relative effects of cocaine and ethanol. Previous
authors have suggested that studies of cocaine-
related cardiotoxicity are significantly affected by
sodium pentobarbital anesthesia and that anesthe-
tized canine models tend to overemphasize the
myocardial depressant effects of cocaine and atten-
ϩϩ
Ca exchange. This would result in decreased cal-
cium influx into myocytes during membrane
3
0,32,38
13
depolarization.
Both may also directly inhibit
voltage-sensitive calcium channels in the sarco-
3
2,38
lemma.
By all of these mechanisms, cocaine
and CE could make less calcium available for load-
ing of the sarcoplasmic reticulum and the contrac-
tile apparatus. The negative ionotropic effects of
ethanol are well described, and are predominantly
due to decreased myofilament responsiveness to
ϩϩ
Ca and decreased calcium transients into myo-
1
2,39
cardial cells.
We observed significant decreases
in ventricular contractile function and stroke vol-
ume primarily after prolonged ethanol administra-
tion. The negative ionotropic effects of ethanol cer-
tainly contribute to the prolonged cardiac
depression seen in the CϩE group. Complemen-
40
uate the sympathomimetic effects. However, the
tary deleterious effects on myocardial function due current study was performed using alpha-chlora-
to cocaine, ethanol, and CE may be responsible for lose anesthesia specifically in order to minimize
the myocardial toxicity observed.
these effects as it sustains autonomic reflex control
The hemodynamic depression that occurred in of the cardiovascular system and has significantly
the present experiments could be due to cocaine-, less effect on ventricular function than pentobar-
21,41
ethanol-, or CE-induced myocardial ischemia. We bital.
Though significant changes in myocardial
believe that these ST changes observed were prob- function were not observed in the control group, we
ably due to aberrant impulse conduction but can- cannot rule out that anesthesia may have con-
not exclude the possibility that myocardial ische- founded the relative effects of cocaine, ethanol, and
mia due to cocaine, ethanol, and CE was a factor CE in these experiments.
in the myocardial depression observed.
In these experiments, attempts were made to
maintain normal acid–base status, primarily be-
cause alpha-chloralose anesthesia can cause met-
C
LINICAL ELEVANCE
R
21
abolic acidosis with prolonged administration.
One animal study suggested that the acute cardiac Cocaine has been implicated to cause metabolic ac-
42
toxicity due to cocaine and ethanol is an additive idosis and there are case reports of patients who

ACADEMIC EMERGENCY MEDICINE • March 2001, Volume 8, Number 3
221
present with severe cocaine toxicity who are sig- combination. Though most clinicians are very
nificantly acidemic. Sodium bicarbonate given to aware of the acute cardiac toxicity due to cocaine
these patients resulted in improvement of the co- abuse, delayed toxicity due to cocaine and ethanol
4
2,43
caine-related conduction abnormalities.
Co- in combination should also be considered in pa-
caine’s effects on cardiac sodium channels are af- tients presenting after using these drugs.
fected by pH, and increasing pH may improve
44
cocaine-related conduction abnormalities.
At-
The authors thank Elana Leanova, Dr. Robert Henning, Dr.
tempts to normalize pH may have diminished Matt Levy, Dr. Dave Igel, Don Wallick, Frank Walters, and
cocaine- or cocaine-and-ethanol-related ECG ab- Mike Ritzenthaller for their support of this project.
normalities and could affect the external general-
izability of this model.
References
Though significant decreases in myocardial
function were observed at the same time as peak 1. Grant BF, Harford TC. Concurrent and simultaneous use of
alcohol and cocaine: results of a national survey. Drug Alcohol
serum concentrations of CE, the relative impor-
tance of CE as an independent cause of this tox-
Depend. 1990; 25:97–104.
2. Weiss RD, Mirin SM, Griffin MI, Michael JL. Psychopa-
icity remains unclear. Other studies have reported thology in cocaine abusers: changing trends. J Nerv Ment Dis.
1
988; 176:719–25.
cardiac toxicity related to CE, but typically at
4
5
3. Farre M, De La Torre R, Llorent M, et al. Alcohol and co-
greater serum concentrations of CE. Serum co-
caine interaction in humans. J Pharmacol Exp Ther. 1993; 266:
caine levels remained significantly greater than 1364–73.
4
. Perez-Reyes M, Jeffcoat RA. Ethanol/cocaine interaction:
the serum cocaine concentrations later in the
study; however, the delayed decreases in myocar-
cocaine and cocaethylene plasma concentrations and their re-
lationship to subjective and cardiovascular effects. Life Sci.
dial function were not observed in the cocaine only 1992; 51:553–63.
5
. Uszenski RT, Gillis RA, Schaer GL, Analouei AR, Kuhn FE.
group. Other cocaine or ethanol metabolites were
not measured and potentially could have played a
Additive myocardial depressant effects of cocaine and ethanol.
Am Heart J. 1992; 124:1276–83.
role in the cardiotoxicity observed. The role of eth- 6. Foltin RW, Fischman MW. Ethanol and cocaine interactions
in humans: cardiovascular consequences. Pharm Biochem Be-
anol itself was likely significant in the delayed tox-
icity observed in the cocaine and ethanol group as
well.
hav. 1989; 31:877–83.
7
. Ruttenber J. Cocaine, alcohol mix in body to form even
longer lasting, more lethal drug. JAMA. 1992; 267:1043–4.
. Drug Abuse Warning Network (DAWN). Annual medical ex-
8
aminer data, 1990; National Institute on Drug Abuse, U.S. De-
partment of Health and Human Services Publication No.
C
ONCLUSIONS
(
ADM) 91–1840; 1991.
. Drug Abuse Warning Network (DAWN). Annual medical ex-
9
We observed that cocaine plus ethanol in combi-
nation was significantly more toxic than cocaine
aminer data, 1994; National Institute on Drug Abuse, U.S. De-
partment of Health and Human Services Publication, Series 1,
alone. After each cocaine bolus, both the cocaine No. 14-B; 1994.
1
0. Hearn WL, Flynn DD, Hime GW, et al. Cocaethylene: a
plus ethanol and cocaine groups had similar acute
toxicities; however, delayed toxicity was observed
unique cocaine metabolite displays high affinity for the dopa-
mine transporter. J Neurochem. 1991; 56:698–701.
in the cocaine plus ethanol group. Cocaethylene 11. Henning RJ, Wilson LD, Glauser JM. Cocaine plus alcohol
is more cardiotoxic than cocaine or alcohol alone. Crit Care
formed rapidly in this animal model in much lower
Med. 1994; 22:1896–906.
concentrations than its parent compound cocaine,
12. Danziger RS, Sakai M, Capogrossi MC, Spurgeon HA,
but well within the range that has been reported Hansford RG, Lakatta EG. Ethanol acutely and reversibly sup-
presses excitation–contraction coupling in cardiac myocytes.
clinically. The later toxicity in the cocaine plus eth-
Circ Res. 1991; 68:1660–8.
anol group occurred when CE serum levels were
13. Parker RB, Williams CL, Laizure SC, Mandrell TD, La-
highest. Cocaethylene may have been a factor in branche GS, Lima JJ. Effects of ethanol and cocaethylene on
cocaine pharmacokinetics in conscious dogs. Drug Metab Dis-
the combined toxicity of cocaine and ethanol; how-
pos. 1996; 24:850–3.
ever, myocardial depression was observed in the
14. McCance-Katz EF, Price LH, McDougle CJ, Kosten TR,
ethanol only group after prolonged administration Black JE, Jatlow P. Concurrent cocaine–ethanol ingestion in
humans: pharmacology, physiology, behavior, and the role of
and was also a factor in the prolonged toxicity in
cocaethylene. Psychopharmacology. 1993; 111:39–46.
the cocaine plus ethanol group. Prolonged cocaine-
15. Hollander JE, Hoffman RS, Gennis P, et al. Prospective
and-ethanol-induced myocardial depression is multicenter evaluation of cocaine-associated chest pain. Acad
Emerg Med. 1994; 1:330–7.
probably multifactorial. Reasons may include CE
toxicity in addition to primary toxicity of its parent
1
6. Jatlow P, Elsworth JD, Bradberry CW, et al. Cocaethylene:
a neuropharmacologically active metabolite associated with
compounds, synergistic or additive adverse effects concurrent cocaine–ethanol ingestion. Life Sci. 1991; 48:1787–
9
4.
of parent compounds, and possibly increased co-
caine serum concentrations in the setting of etha-
1
7. Katz JL, Terry P, Witkin JM. Comparative behavioral
pharmacology and toxicology of cocaine and its ethanol-derived
nol use. Lethal and malignant arrhythmias were metabolite, cocaine ethyl-ester (cocaethylene). Life Sci. 1992;
5
0:1351–61.
observed only in the cocaine and ethanol group and
certainly may contribute to the significant toxicity
1
8. Shelnutt S, Hudzik TJ, Lattin DL, McMillan DE. Effects
of cocaine and ethylcocaine on schedule-controlled responding
observed clinically due to cocaine and ethanol in in rats. Pharmacol Biochem Behav. 1992; 43:509–11.

222
COCAETHYLENE
Wilson et al. • COCAINE, ETHANOL, AND COCAETHYLENE CARDIOTOXICITY
2
ϩ
1
9. Hearn WL, Rose S, Wagner J, Ciarleglio A, Mash DC. Co- caethylene on intracellular Ca and myocardial contraction in
caethylene is more potent than cocaine in mediating lethality. cardiac myocytes. Br J Pharmacol. 1993; 109:293–8.
Pharmacol Biochem Behav. 1991; 39:531–3. 33. Wang J, Carpentier RG. Electrophysiologic in-vitro effects
0. Bailey DN. Plasma cocaethylene concentrations in pa- of cocaine and its metabolites. Int J Cardiol. 1994; 46:235–42.
tients treated in the emergency room or trauma unit. Am J 34. Sharkey J, Ritz MC, Schenden JA, Hanson RC, Kuhar MJ.
2
Clin Pathol. 1993; 99:123–7.
Cocaine inhibits muscarinic cholinergic receptors in heart and
2
1. Hulzgrefe HH, Everitt JM, Wright EM. Alpha-chloralose brain. J Pharm Exp Ther. 1988; 246:1048–52.
as a canine anesthetic. Lab Anim Sci. 1987; 37:587–94.
2
35. Hale SL, Alker KJ, Rezkalla SH, Eisenhauer AC, Kloner
2. Roy IM, Jefferies TM, Threadgill MD, Dewar GH. Analysis RA. Nifedipine protects the heart from the acute deleterious
of cocaine, benzoylecognine, ecognine methyl ester, ethylco- effects of cocaine if administered before but not after cocaine.
caine and norcocaine in human urine using HPLC with post- Circulation. 1991; 83:1437–43.
column ion-pair extraction and fluorescence detection. J Pharm 36. Przywara DA, Dambach GE. Direct actions of cocaine on
Biomed Anal. 1992; 10:943–8.
cardiac cellular electrical activity. Circ Res. 1989; 65:185–92.
2
3. Jung G, Ferard G. Enzyme-coupled measurement of eth- 37. Kimura S, Bassett AL, Xi H, Myerburg RJ. Early repolar-
anol in whole blood and plasma with a centrifugal analyzer. izations and triggered activity induced by cocaine: a possible
Clin Chem. 1978; 24:873–6.
mechanism of cocaine arrhythmogenesis. Circulation. 1992; 85:
2
4. Wetli CV, Wright RK. Death caused by recreational cocaine 2227–35.
use. JAMA. 1979; 241:2519–22.
2
38. Stewart G, Rubin E, Thomas AP. Inhibition by cocaine of
5. Virmani R, Robinowitz M, Smialek JE, Smyth DF. Cardi- excitation–contraction coupling in isolated cardiomyocytes. Am
ovascular effects of cocaine: an autopsy study of 40 patients. J Physiol. 1991; 260:H50–H57.
Am Heart J. 1988; 115:1068–76. 39. Katz AM. Effects of ethanol on ion transporting muscle
6. Bailey DN. Comprehensive review of cocaethylene and co- membranes. Fed Proc. 1982; 41:2456–8.
2
caine concentrations in patients. Am J Clin Pathol. 1996; 106: 40 Fraker TD, Temesy-Armos PN, Brewster PS, Wilkerson
7
01–4.
7. Jatlow P. Cocaethylene: pharmacologic activity and clinical dogs. Circulation. 1990; 81:1012–6.
41. Van Citters RL, Franklin DL, Rushmer RF. Left ventric-
8. Perez-Reyes M. The order of administration: its effects on ular dynamics in dogs during anesthesia with alpha-chloralose
RD. Mechanism of cocaine-induced myocardial depression in
2
significance. Ther Drug Monit. 1993; 15:533–6.
2
the interaction between cocaine and ethanol. Life Sci. 1994; 7: and sodium pentobarbital. Am J Cardiol. 1964; Mar:349–54.
5
2
41–50.
42. Wang RY. pH-dependent cocaine-induced cardiotoxicity.
9. Farre M, De La Torre R, Gonzalez ML et al. Cocaine and Am J Emerg Med. 1999; 17:364–9.
alcohol interactions in humans: neuroendocrine effects and co- 43. Jonsson S, O’Meara M, Young JB. Acute cocaine poisoning:
caethylene metabolism. J Pharm Exp Ther. 1997; 283:164–76. importance of treating seizure and acidosis. Am J Med. 1983;
3
0. Xu Y, Crumb WJ, Clarkson CW. Cocaethylene, a metabo- 75:1061–4.
lite of cocaine and ethanol, is a potent blocker of cardiac so- 44. Crumb WJ, Clarkson CW. The pH dependence of cocaine
dium channels. J Pharm Exp Ther. 1994; 271:319–25. interaction with cardiac sodium channels. J Pharm Exp Ther.
1. Crumb WJ, Clarkson CW. Characterization of the sodium 1995; 274:1228–37.
3
channel blocking properties of the major metabolites of cocaine 45. Wilson LD, Henning RJ, Suttheimer C, Lavins E, Balraj
in single cardiac myocytes. J Pharm Exp Ther. 1992; 261:910– E, Earl S. Cocaethylene causes dose-dependent reductions in
7
.
cardiac function in anesthetized dogs. J Cardiovasc Pharmacol.
3
2. Qiu Z, Morgan JP. Differential effects of cocaine and co- 1995; 26:965–73.