Effect of Cocaine on Left Ventricular Function
Relation to Increased Wall Stress and Persistence After Treatment
Background To determine whether alterations in left ventricular (LV) function after a cocaine infusion are due to reduced myocardial contractility or changes in loading conditions, we examined LV function in 30 morphine-sedated, closed-chest dogs. We also wanted to determine the time course of the effects of cocaine on LV function after the infusion was stopped.
Methods and Results Two-dimensional echocardiography and hemodynamics provided LV fractional shortening and end-systolic wall stress data. Radionuclide ventriculography was also performed. Four groups of dogs received saline or cocaine infusions of 10, 30, or 100 μg · kg−1 · min−1. Cocaine was infused for 90 minutes with ECG and arterial pressure monitoring. Animals were monitored for an additional 120 minutes after the infusion ended. Arterial pressure rose over the course of the experiment in all four groups, but saline and cocaine 10 μg · kg−1 · min−1 did not significantly change ejection fraction. Cocaine 30 and 100 μg · kg−1 · min−1 acutely increased arterial pressure and heart rate but decreased ejection fraction from 0.64±0.06 to 0.45±0.08 and from 0.65±0.10 to 0.46±0.11, respectively. Additionally, cocaine 100 μg · kg−1 · min−1 decreased fractional shortening from 36±9% to 23±12%. However, cocaine 30 and 100 μg · kg−1 · min−1 also increased wall stress from 42±15 to 65±11 g/cm2 and from 37±15 to 90±33 g/cm2, respectively. These results were analyzed by use of the relation between wall stress and fractional shortening as an index of contractility. Fractional shortening after cocaine infusion was displaced downward as a result of increased wall stress rather than changes in contractility. In addition, alteration of afterload with phenylephrine (6 μg/kg) and sodium nitroprusside (10 μg/kg) before and during infusion of cocaine 100 μg · kg−1 · min−1 showed similar regression lines for wall stress to fractional shortening.
Conclusions Ejection-phase indexes of LV function were reduced by cocaine in this model of conscious, sedated dogs, but effects were attributable to increased wall stress rather than to reduced myocardial contractility. These effects persisted for at least 2 hours after the infusion was stopped.
Cocaine abuse is an important public health problem, with an estimated 5 million regular users in the United States and increasingly recognized mortality resulting from overdose.1 2 3 4 Numerous reports in the literature suggest that cocaine abuse may precipitate myocardial infarction,1 5 6 ventricular arrhythmias,1 7 and left ventricular (LV) dysfunction.8 9 Cocaine abuse has been recognized as an independent risk factor for heart disease.5 However, rigorous investigation of cocaine cardiotoxicity in humans by controlled clinical trials has proved difficult. Differences in route of administration, drug purity, and other factors may also contribute to the cardiotoxic effects of cocaine.
Therefore, there has been considerable interest in studying the effects of cocaine on hearts in experimental animals. Hemodynamic studies showed increased heart rate and/or arterial pressure that was related to dose and experimental model.3 These hemodynamic effects could contribute to the adverse effects of cocaine abuse observed clinically. Several studies suggested that cocaine depresses myocardial function.10 11 12 13 However, the effect of cocaine on myocardial contractility is unclear. Also, direct effects of cocaine on regional myocardial blood flow are controversial, with some groups suggesting reduced coronary flow and provocation of spasm14 and others showing vasodilation or no change.15 16
In the present study, groups of closed-chest, sedated dogs received infusion of saline or cocaine in three doses. LV function was examined noninvasively with two-dimensional (2D) echocardiography and radionuclide angiography. Regional blood flow was measured with radiolabeled microspheres. We used an infusion model to simulate the steady-state “high” achieved by cocaine abusers through repetitive intranasal insufflation of cocaine, which presumably results in a sustained elevation of serum cocaine concentrations. This study was previously published in abstract form.17
The animal studies reported in this manuscript conform to the “Position of the American Heart Association on Research Animal Use” adopted November 11, 1984.
Mongrel dogs of both sexes, weighing 15 to 22 kg, were used. Animals were sedated with morphine 1 mg/kg SC to minimize discomfort and to limit vocalization. Animals were briefly anesthetized with thiamylal. Supplemental local anesthesia with lidocaine 4 mg/kg SC was administered as needed. An arterial catheter was placed in the femoral artery for arterial pressure measurement and reference blood flow withdrawal in some animals (n=20). Some animals also had a carotid artery dissected free and catheterized for LV pressure and injection of microspheres. Animals were allowed to recover for 30 to 45 minutes after catheter implantation.
The arterial pressure and lead II of the ECG were recorded continuously with Gould amplifiers (Gould Inc) and a Coulbourn chart recorder (Coulbourn Instruments).
Red blood cells were labeled in vivo with 15 to 25 mCi of 99mTc.18 Baseline equilibrium radionuclide angiography was performed by use of a Picker Dyna Camera 4 (Picker Corporation) with an Acusync 4R gating system (Advanced Medical Research) attached to a Paragon image analysis system (Medasys). A modified left anterior oblique view was used with the dog resting comfortably in a sling. Camera head position was adjusted to obtain optimal separation of the right and left ventricles for baseline radionuclide angiography. Images were acquired in frame mode at 64 frames per RR interval by summing the radioactivity in the ventricle to obtain 75 000 counts per frame. A general-purpose collimator was used.
Dogs were then positioned on a platform for 2D echocardiography with commercially available equipment (Hewlett-Packard 77020A Ultrasound Imager with a Sony videocassette recorder) and a 5-MHz transducer. With the dog lying on its right side, the area with the most prominent precordial pulse was located manually through an open area in the platform. Parasternal short-axis images just below the level of the mitral valve were obtained. This area was recorded for at least 10 seconds.
After baseline measurements were taken, an infusion of saline or graded doses of cocaine (Sigma Chemical Co) was administered for 90 minutes. Cocaine was prepared in physiological saline. Dogs were randomly assigned to infusion of saline or cocaine at 10, 30, or 100 μg · kg−1 · min−1. Hemodynamic parameters were recorded at intervals.
The dog was transferred to the sling for repeated radionuclide angiography 2 hours after the infusion was completed. The dog was then euthanatized with a concentrated sodium pentobarbital-KCl solution (J.A. Webster Co), and the heart was removed and weighed. Tissue sampling for regional flow studies was performed in 20 animals.
Assessment of LV Function
LV function was assessed by measurement of ejection fraction determined by radionuclide angiography and by fractional shortening from the 2D echocardiogram. For radionuclide angiography studies, a gated, 64-frame image was collected from the onset of the R wave of the ECG. This image was processed with background subtraction followed by spatial and temporal filtering. The region of interest was examined for each frame and adjusted as necessary. The LV ejection fractions from processed and raw images agreed closely.
Two-dimensional echocardiography was performed at baseline, during cocaine infusion, and at the end of infusion, providing results to complement findings from radionuclide angiography. Fractional shortening of the diameter from 2D echocardiographic images at the level of the papillary muscles was determined according to published techniques19 20 and was used as a second index of LV function. Ventricular cross-sectional diameter was also used as an index of LV size. Wall thickening was measured as a third index of LV function. Finally, LV wall stress was calculated according to the method of Grossman and coworkers21 22 :
where PES is end-systolic pressure in millimeters of mercury, DES and hES are end-systolic dimension and wall thickness in centimeters, and 1.35 is the factor to convert pressure from millimeters of mercury to grams per centimeter squared.
Regional Myocardial Blood Flow
Regional myocardial blood flow was determined by use of the reference blood sample technique.23 Carbonized plastic microspheres (10±3-μm diameter, New England Nuclear) labeled with 141Ce, 51Cr, 103Ru, 95Nb, or46Sc were obtained as 0.5 mCi of nuclide in 30 mL of isotonic saline to which 1 drop of Tween 80 was added to minimize aggregation. The mixture was agitated in a sonicator bath (Bransonic Ultrasonic Cleaner) and then in a vortex mixer (Fisher Scientific) before injection. Approximately 2×106 to 4×106 spheres were infused into the left ventricle in a total volume of 1.0 to 1.5 mL followed by an 8-mL flush with physiological saline. A few seconds before each microsphere injection, a timed collection of reference flow from the femoral artery was started and maintained at a constant rate (9 mL/min) for 3 minutes. Microsphere injections were made at baseline, 30 minutes into infusion, and 90 minutes after infusion.
After completion of the experiment, the heart was excised, washed with saline, and fixed in buffered 10% formalin for 24 hours. After the heart was rinsed in saline, epicardial fat, large vessels, the atrial cap, valves, and chordae tendineae were removed. Two regions of right ventricle were sectioned into endocardial and epicardial halves. In addition, two regions from the septum, anterior free wall, and posterior free wall were sectioned into subendocardium, midmyocardium, and subepicardium. All 22 tissue pieces were weighed and placed in polystyrene test tubes, and activity of the isotopes was determined at five energy windows in an autogamma spectrometer (Packard Autogamma 5650) attached to a preprogrammed computer. The true activity of each isotope in the tissue sample was separated by a preprogrammed computer and myocardial blood flow (Qm) calculated from the following equation: Qm=Qr/(Cm/Cr), where Qr is the withdrawal rate of the reference blood sample, Cm is the activity (counts per minute per gram) of the myocardial tissue sample, and Cr is the activity (counts per minute) of the reference blood sample. The two samples from each area were pooled for calculation of tissue blood flow in each heart.
The mean and SD are reported in text and tables. When two values from one group were compared, such as ejection fraction before and after cocaine infusion, paired t statistics were used. When more than one group was compared, such as wall stress after cocaine infusion in each dose group, one-way ANOVA was used. Statistical significance was analyzed with the Student-Newman-Keuls test.24 When multiple events within a group were examined, such as hemodynamic events at various times, two-way ANOVA followed by the Student-Newman-Keuls test was used. P<.05 was considered significant with the two-tailed test. For examination of the relation between fractional shortening and wall stress, the linear regression line for all baseline pooled with infusion values was calculated, along with the 95% CIs. This was compared with the regression line and 95% CIs for values measured at the two high drug doses. Comparisons between lines to assess differences in slope and y intercept were made with sas.
Table 1⇓ shows heart rate, mean arterial pressure, and the rate-pressure product in saline and cocaine experiments. No changes in heart rate were observed in the saline or cocaine 10 or 30 μg · kg−1 · min−1 group. In the cocaine 100 μg · kg−1 · min−1 group, a statistically significant increase in heart rate was observed at 90 minutes of infusion. No differences in heart rate were observed between groups at any time point. Also, heart rate tended to rise in the saline group to an extent similar to that seen in the cocaine 100 μg · kg−1 · min−1 group, but the change did not reach statistical significance because of within-group variability. In the saline and cocaine 10 μg · kg−1 · min−1 groups, arterial pressure rose over the 4-hour experimental period. In the cocaine 30 and 100 μg · kg−1 · min−1 groups, however, arterial pressure was increased acutely by cocaine. The increase was moderate, approximately 20 mm Hg. LV end-diastolic pressure was measured in saline and cocaine experiments in conjunction with microsphere determinations. No change in end-diastolic pressure was seen in the control group. In the cocaine 100 μg · kg−1 · min−1 group, end-diastolic pressure was significantly increased 90 minutes after infusion from baseline (4±2 to 10±4 mm Hg).
Ejection fraction was similar at baseline in the four groups (Fig 1⇓). At 2 hours after infusion of saline or cocaine 10 μg · kg−1 · min−1, ejection fraction was not significantly altered (from 0.61±0.10 to 0.59±0.11 and from 0.63±0.10 to 0.55±0.14, respectively). However, in the cocaine 30 and 100 μg · kg−1 · min−1 groups, ejection fraction was significantly reduced (from 0.63±0.6 to 0.44±0.08 and from 0.64±0.10 to 0.45±11).
Table 2⇓ shows LV dimension, wall thickness, and fractional shortening results. No differences were observed in end-diastolic dimension or in end-diastolic wall thickness in this study. In the cocaine 100 μg · kg−1 · min−1 group, end-systolic wall thickness was decreased 60 and 90 minutes after cocaine infusion. Also, end-systolic dimension was significantly elevated 90 minutes after cocaine infusion. No differences were observed in the saline and cocaine 10 μg · kg−1 · min−1 groups in fractional shortening.
Fractional shortening was decreased in the cocaine 30 μg · kg−1 · min−1 group 30 minutes after infusion (Table 2⇑). Fractional shortening was decreased in the cocaine 100 μg · kg−1 · min−1 group at 90 minutes of infusion and at later time points. LV end-systolic wall stress (Table 2⇑) was unchanged by saline or cocaine 10 μg · kg−1 · min−1. In the cocaine 30 and 100 μg · kg−1 · min−1 groups, wall stress was increased 90 minutes after infusion and from 90 minutes of infusion to 90 minutes after infusion, respectively.
The relation between fractional shortening and wall stress can be used as an index of the state of myocardial contractility. Fig 2⇓ shows this relation, with 95% tolerance intervals, for all baseline measurement points. A significant inverse relation with r=.7 was obtained. In Figs 3 (top) and 4 (top), the relation between wall stress and fractional shortening is shown for all experiments at 90 minutes of infusion and 90 minutes after infusion, respectively. Almost all points lie within the baseline 95% tolerance intervals. In addition, although fractional shortening is lower in the cocaine 100 μg · kg−1 · min−1 group, and to a lesser extent in the 30 μg · kg−1 · min−1 group, this change is related to the increased wall stress. In Figs 3 (bottom) and 4 (bottom), the relation between wall stress and fractional shortening is shown with CIs for baseline mean fractional shortening and mean values for the four treatment groups at 90 minutes of infusion and 90 minutes after infusion, respectively. The mean values±SEM overlap the baseline 95% CI.
Phenylephrine (6 μg/kg IV) and sodium nitroprusside (10 μg/kg IV) were used to alter loading conditions at baseline and 60 minutes into cocaine 100 μg · kg−1 · min−1 infusion. In this way, three to five baseline fractional shortening and wall stress values were obtained before and during infusion in each experiment. Fig 5⇓ shows these points and regression lines. The two sets of points overlap substantially. The regression lines were not different.
Regional Myocardial Blood Flow
Regional myocardial blood flow was unchanged by cocaine 100 μg · kg−1 · min−1. Table 3⇓ shows flow values for right ventricle and posterior free wall. There were no differences between the two groups at baseline. Regional flow was unchanged from baseline at 30 minutes of saline infusion and 60 minutes after infusion. Flow tended to increase 30 minutes into infusion of cocaine and return slightly below control values at 90 minutes after infusion. Because of variability, however, no significant differences were observed.
The present study examined the effects of cocaine infusion on myocardial function in closed-chest, sedated dogs. Cocaine infusion had mild hemodynamic effects, increasing mean arterial pressure and the rate-pressure product. Cocaine 30 and 100 μg · kg−1 · min−1 caused significant reduction of noninvasive parameters of LV function, namely fractional shortening and ejection fraction. The effects of the 90-minute infusion persisted for at least 90 minutes to 2 hours after infusion, with peak effects observed 90 minutes after the infusion ended. At this time, the LV end-diastolic pressure was elevated, LV end-systolic wall thickness reduced, and end-systolic dimension increased. Fractional shortening and ejection fraction were decreased. Plots of fractional shortening versus end-systolic wall stress indicate that the effects of cocaine on the ejection-phase indexes in our study were related to increased end-systolic wall stress and were not caused by altered myocardial contractile function.
Cardiovascular effects of cocaine are complex, involving a balance between direct actions of cocaine and indirect effects through stimulation of the central nervous system. Direct effects on the cardiovascular system result from the local anesthetic action of cocaine, which may act as a myocardial depressant, and the inhibition of norepinephrine reuptake, which may produce peripheral vasoconstriction.3 4 Increased mean arterial pressure and rate-pressure product generally have been found in conscious, sedated dogs but not in anesthetized dogs,12 13 25 presumably because of central nervous system stimulation. Prior studies have shown variable findings with other effects, such as changes in heart rate or LV end-diastolic pressure. It is likely that hemodynamic effects of cocaine are dose-related and may differ between bolus and infusion protocols. Also, the experimental model studied and the type of anesthesia will influence results. Wilkerson11 and Fraker et al14 showed that cocaine-induced hypertension is blunted in anesthetized dogs.
The direct effects of cocaine have been implicated as a cause of myocardial depression seen with its use. Cocaine has been reported to reduce LV positive dP/dt12 26 and increase LV end-diastolic pressure.12 In diazepam-sedated dogs receiving cocaine 0.5 mg · kg−1 · min−1 infusion, Bedotto et al13 showed a reduction of ejection fraction from 0.61 to 0.49, supporting the present results. Mean arterial pressure was increased, but heart rate and dP/dt were unchanged. Fraker et al14 showed a reduction in regional ejection fraction by echocardiogram in conscious dogs receiving 4-mg/kg bolus of cocaine, which was associated with increased arterial pressure, heart rate, and rate-pressure product. The reduction in regional ejection fraction occurred within 2 minutes of injection and tapered off at 10 minutes.
Other investigators challenged the concept that direct myocardial depression induced by cocaine is the principal mechanism for impaired LV function. Stambler and coworkers27 showed that cocaine 1 mg/kg IV caused a transient decrease in LV dP/dt followed by stimulation within 5 minutes, which they attributed to an early local anesthetic effect followed by adrenergic stimulation. They noted early and late significant increases in LV end-systolic wall stress after cocaine infusion. Liu et al25 showed that cocaine 3 mg/kg in pentobarbital-anesthetized dogs reduced ejection fraction, decreased LV dP/dt, and increased peripheral vascular resistance with significant effects 2 minutes after injection that persisted for 30 minutes. Analysis of pressure-volume loops showed brief reduction of contractility at 2 minutes, with recovery of contractile function within 10 minutes. The persistence of decreased myocardial function between 10 and 30 minutes was related to increased afterload. This was supported by studies in dogs with ganglionic blockade, whereas cocaine produced no peripheral vasoconstriction but produced a similar brief reduction in contractility. In the animals with ganglionic blockade, the effects of cocaine on ejection fraction and dP/dt were smaller and of shorter duration.25 In the present study, we did not use a bolus of cocaine, and we have no data related to myocardial effects of cocaine early in the infusion. However, the persistent reduction in ejection-phase indexes of LV function was due to increased wall stress without change in the relation between wall stress and fractional shortening, which agrees with the findings of Liu et al.25 Our results indicate that the primary effects of cocaine are due to altered loading conditions, although our data do not exclude the possibility of a transient decrease in myocardial contractility very early in the infusion.
We examined the effects of cocaine on LV end-systolic wall stress. In the cocaine 30 and 100 μg · kg−1 · min−1 groups, increased wall stress was observed. Because the reduction of fractional shortening was observed with increased wall stress, the possibility that contractility was unaltered was investigated by examination of the relation between wall stress and fractional shortening. The results show that contractility was not reduced but that reduction of fractional shortening was caused by increased wall stress. This was demonstrated by graphs of fractional shortening versus wall stress for 90 minutes of infusion and 90 minutes after infusion for the four groups within baseline tolerance intervals. Mean values for the four groups were within baseline CIs. Additionally, alteration of loading conditions at baseline and during infusion of cocaine 100 μg · kg−1 · min−1 gave similar regression lines, indicating that contractility was not altered despite significant decreases in fractional shortening and ejection fraction.
Our regional myocardial flow data support the lack of important myocardial depression, at least as a consequence of local ischemia. We found no tendency for reduced nutritive flow, which suggests that reduction of fractional shortening and ejection fraction was not caused by myocardial transmural ischemia. The lack of change in the ratio of endocardium to epicardium shows that subendocardial ischemia also was not a likely contributor. Effects of cocaine on regional myocardial blood flow were studied previously with various results. Using a bolus dose or infusion of 10 mg/kg, Hale et al12 showed a reduction in regional myocardial blood flow. More recently, Shannon et al28 reported a small increase in coronary flow after 1 mg/kg of cocaine infusion that was associated with increased coronary vascular resistance. Previous studies from our group showed no change in regional myocardial blood flow in dogs receiving a bolus dose of 0.25 to 2 mg/kg.16
Potential study limitations include the use of an infusion model in contrast to the bolus doses used by others. Infusions have the advantage of relatively steady-state effects, which were useful in examinations of hemodynamics during the infusion and allowed manipulation of loading conditions with phenylephrine and sodium nitroprusside. In addition, our model attempts to simulate the effects produced by repetitive intranasal use of cocaine by a drug abuser during a prolonged cocaine binge. Rich and Singer29 examined cocaine-related symptoms in patients presenting to an urban emergency department and more frequently found cardiopulmonary symptoms in patients who used intranasal cocaine versus other routes of administration. Epidemiological studies showed that cocaine hydrochloride, used by nonparenteral routes, is responsible for most cocaine-associated mortality.1 2 The persistence of depressed LV function after cocaine infusion found in our study may have significant implications for individuals involved in cocaine binges. This may be akin to the situation in which the effects of alcohol on alterations in mental capacity persist long after the last drink is consumed.
In conclusion, in morphine-sedated, closed-chest dogs, cocaine infusion had dose-dependent effects on cardiovascular function. Cocaine 30 and 100 μg · kg−1 · min−1 decreased ejection fraction and fractional shortening in proportion to increases in end-systolic wall stress, indicating that the observed depression in LV performance was not due primarily to an alteration in contractile state. The hemodynamic effects persisted for 2 hours after the cocaine infusion was stopped.
We wish to thank Charles Clark, MD, and Jayesh Patel, MD, for contributions made to this project and the staff of the East Carolina University Department of Comparative Medicine for their assistance. We also wish to thank Dr Robert A. Kloner for reviewing the data, and we greatly appreciate the valuable suggestions provided by Dr Joshua Wynne in revising the manuscript.
- Received September 12, 1994.
- Revision received December 20, 1994.
- Accepted December 27, 1994.
- Copyright © 1995 by American Heart Association
Kloner RA, Hale S, Alker K, Rezkalla S. The effects of acute and chronic cocaine use on the heart. Circulation. 1992;85:407-419.
Billman GE. Mechanisms responsible for the cardiotoxic effects of cocaine. FASEB J. 1990;4:2469-2475.
Cregler LL, Mark H. Relation of acute myocardial infarction to cocaine abuse. Am J Cardiol. 1985;56:794.
Benchimol A, Bartall H, Desser KB. Accelerated ventricular rhythm and cocaine abuse. Ann Intern Med. 1978;88:519-520.
Wilkerson RD. Cardiovascular effects of cocaine in conscious dogs: importance of fully functional autonomic and central nervous systems. J Pharmacol Exp Ther. 1988;246:466-471.
Chapman SF, Lust RM, Morrison RF, Etemad-Moghadam R, Chitwood WR, Austin EH. Possible mechanism of cocaine-induced sudden cardiac death. FASEB J. 1989;3:A407. Abstract.
Fraker TD, Temesy-Armos PN, Brewster PS, Wilkerson RD. Mechanism of cocaine-induced myocardial depression in dogs. Circulation. 1990;81:1012-1016.
Mehta PM, Grainger TA, Lust RM, Clark C, Movahed A, Jolly SR. Is cocaine-induced myocardial dysfunction a direct effect or due to increased wall stress? J Am Coll Cardiol. 1993;21:285A. Abstract.
Pavel DG, Zimmer AM, Patterson VN. In vivo labeling of red blood cells with 99mTc: a new approach to blood pool visualization. J Nucl Med. 1977;18:305-310.
Grossman W, Jones D, McLaurin LP. Wall stress and patterns of hypertrophy in the human left ventricle. J Clin Invest. 1975;56:56-64.
Domenech R, Hoffman JIE, Nobel M, Saunder K, Henson J, Subijanto S. Total and regional myocardial blood flow measured by radioactive microspheres in conscious and anesthetized dogs. Circ Res. 1969;25:581-596.
Glantz SA. Primer of Biostatistics. New York, NY: McGraw-Hill Publishing Co; 1987.
Stambler BS, Komamura K, Ihara T, Shannon RP. Acute intravenous cocaine causes transient depression followed by enhanced left ventricular function in conscious dogs. Circulation. 1993;87:1687-1697.
Shannon RP, Stambler BS, Komamura K, Ihara T, Vatner SF. Cholinergic modulation of the coronary vasoconstriction induced by cocaine in conscious dogs. Circulation. 1993;87:939-949.