- Email: [email protected]
Patients with detectable cocaethylene are more likely to require intensive care unit admission after trauma
American Journal of Emergency Medicine (2010) 28, 1051–1055
www.elsevier.com/locate/ajem
Original Contribution
Patients with detectable cocaethylene are more likely to require intensive care unit admission after trauma Sage E. Wiener MD a , Darrell Sutijono MD a , Cynthia H. Moon MD a , Ramanand A. Subramanian PhD a , Jim Calaycay MS b , Julie I. Rushbrook PhD b , Shahriar Zehtabchi MD a,⁎ a
Department of Emergency Medicine, State University of New York, Downstate Medical Center, Box: 1228, Brooklyn, NY 11203, USA b Department of Biochemistry, State University of New York, Downstate Medical Center, Brooklyn, NY 11203, USA Received 4 June 2009; revised 27 June 2009; accepted 27 June 2009
Abstract Cocaethylene (CE) is a toxic metabolite that is formed after simultaneous consumption of cocaine and ethanol. This potent stimulant is more toxic than cocaine and has a longer half-life. The deleterious hemodynamic and cardiovascular effects of CE have been proven in animal models. The aim of this study is to assess the impact of CE on clinical outcomes after trauma. We prospectively enrolled adult (≥13 years) trauma patients requiring admission. Predictor variables were age, sex, mechanism of injury, Injury Severity Score, base deficit, and toxicology groups (ethanol alone, cocaine alone, CE, and none). The outcomes examined were mortality, intensive care unit (ICU) admission, and length of hospital stay (LOS). We used nonparametric tests to compare continuous variables and χ2 test to compare categorical data. We constructed a logistic regression to identify variables that could predict mortality and ICU admission. We enrolled 417 patients (74% male; 70% blunt injury; median age, 40 [range, 13-95]; overall mortality, 2.2%). Urine toxicology and serum ethanol level screens classified patients into the following groups: 13.4% ethanol only, 4.1% cocaine only, 8.9% CE, and 46% none. Mortality and LOS were not statistically different among the groups. In logistic regression analysis, none of the variables were statistically significant in predicting mortality. However, the presence of CE significantly increased the likelihood of ICU admission (odds ratio, 5.9; 95% confidence interval, 1.6-22). The presence of detectable CE in the urine does not increase the mortality or LOS in trauma patients requiring admission but does increase the likelihood of ICU admission. © 2010 Published by Elsevier Inc.
1. Introduction The relationship between substance abuse and traumatic injuries has long been recognized. Many studies have demonstrated an association between drug use and an increased incidence of injuries with poor outcome [1-5]. Although patterns of abuse and injury have shifted some⁎ Corresponding author. Tel.: +1 718 789 1928; fax: +1 718 245 4799. E-mail address: [email protected] (S. Zehtabchi). 0735-6757/$ – see front matter © 2010 Published by Elsevier Inc. doi:10.1016/j.ajem.2009.06.014
what, alcohol and cocaine continue to be widely available and are among the most commonly abused substances in the United States according to the US Department of Health and Human Services [6]. Pharmacophysiologic studies in animals and to a lesser degree in humans have elucidated the potentially toxic influences of these 2 substances on hemodynamic and cardiac functions. When alcohol and cocaine are metabolized to cocaethylene (CE), these consequences are more severe and prolonged [7-16]. Psychopharmacologic literature [17,18] has also suggested
1052 harmful behavioral effects from concomitant use of alcohol and cocaine, possibly related to the formation of CE. Surprisingly, the causative relationship between substance abuse and trauma outcome has been difficult to establish. Furthermore, the mechanism by which these substances might exert effects on the outcome of human injuries is still unclear. Still, many institutions continue to screen routinely for substance abuse in trauma patients despite a lack of evidence that these tests are useful in predicting outcome or changing management. Very few studies have been undertaken to show the influence of substance abuse on care or outcome of patients in the in-hospital setting. Blaho et al [19] attempted to demonstrate the clinical relevance of cocaine metabolites but was not able to show a significant correlation between cocaine or metabolite concentration and the severity of disease or outcome. This study did not discriminate among the nature of illness and did not follow patients throughout the inpatient course. In our investigation, we examined trauma outcomes, namely, the in-patient course and mortality, and explored the role of CE formation in particular. Our objective was to study the relationship between CE and trauma outcomes as measured by mortality, need for intensive care unit (ICU) admission, and length of hospital stay in emergency department (ED) patients who test positive for this substance.
2. Materials and methods 2.1. Study design This was a prospective study using a convenience sample of trauma patients who presented to the ED. Our goal was to investigate the impact of CE on the clinical outcomes of trauma patients. Institutional review board approval was obtained, and waiver of written consent granted.
2.2. Study setting and population The study was conducted from January 2005 to August 2008 at a level I trauma center with approximately 140 000 annual ED visits and serving residents of a largely lowincome, minority community (about 80% African American/ Caribbean, 11% Hispanic, 5% white, 1% Asian, and 3% other races). We enrolled adult trauma patients (13 years and older) presenting to the ED with penetrating or blunt trauma of sufficient severity to require hospital or ICU admission. We excluded patients who died, who were discharged, or who were transferred to the operating room before completion of enrollment and data collection.
2.3. Study protocol Patients were enrolled by convenience sampling, mainly because research staff were not available at all times.
S.E. Wiener et al. However, when academic associates were available (generally weekdays from 10 AM to 10 PM), they enrolled a consecutive sample of trauma patients. Academic associates are medical students or undergraduate students who assist our department in data collection during their research elective. They were trained before the start of this study and were certified to assist after passing the “Protection of Human subjects” course offered by our institutional review board. Data abstractors were not involved in patient care. Treating physicians were not blinded to vital signs or results of urine toxicology, base deficit (BD), and lactate testing. Patient evaluation and treatment were not specified in the study protocol and were at the discretion of the attending emergency physician and trauma surgeon. On designated shifts, the academic associates recorded demographic data, vital signs, mechanism of injury, BD, and lactate. Once the results of imaging studies (radiographs and computed tomography scans) and all other diagnostic procedures were available, the data abstractors documented them. Upon completion of the trauma evaluation, Injury Severity Score (ISS) was calculated. The ISS was graded according to the 1990 revision of the Abbreviated Injury Scale [20,21]. From patients who tested positive for cocaine on routine urine toxicology screening (a routine part of the trauma protocol in our institution), another urine sample was obtained for testing for CE. If, for any reason, obtaining the second urine sample for CE analysis was not feasible in the ED, we allowed sample collection in the inpatient ward or in the ICU up to 72 hours after presentation. This decision was made based on the half-life of CE and its detectability in the urine. Because of the faster clearance of ethanol from serum and variability in time lapse in sending serum samples for ethanol measurement, the presence of detectable ethanol in the serum was not a requirement for enrollment in the study. This was to ensure capture of patients who consumed both substances but had sufficient passage of time before presentation to allow ethanol levels to drop below detectable levels. The outcome measures of interest included the following: in-hospital mortality, ICU admission, and length of hospital stay. The predictor variables included age, sex, mechanism of injury, ISS, BD, lactate, and toxicology group (ethanol alone, cocaine alone, cocathylene, cocaine and ethanol but no cocathylene, and none).
2.4. Measurements The urine toxicology for detection of benzoylecgonine (the primary cocaine metabolite) was performed using the kinetic interaction of microparticles in solution technique (Roche Modular ISE1800 and D2400; Roche Diagnostics, Indianapolis, IN). In the isolation and detection of CE in urine samples, urine samples varying in volume from 3 to 60 mL were centrifuged at 3500 rpm for 5 minutes at 5°C. Using a
Cocaethylene and Trauma peristaltic pump, the supernatant was infused at a flow rate of 0.75 mL/min onto a Sep-Pak C18 cartridge (Waters Corp, Milford, MA), preconditioned with 15 mL each of 90:9:1 CH3CN/H2O/TFA and 98:1:1 water/CH3CN/TFA. Selective elution of the CE was accomplished with 3.0 mL 81:18:1 water/CH3CN/TFA followed by concentration to dryness using a Savant vacuum evaporator. The samples were then redissolved in 100 μL of 90:9:1 CH3CN/H2O/HCOOH and analyzed using a mass spectrometer (a Finnigan LCQ Deca XP Ion Trap mass spectrometer equipped with a microspray ion source from ThermoFinnigan, San Jose, CA). Samples were introduced into the mass spectrometer using a syringe pump at a flow rate of 3 mL/min. Data were acquired in a 3stage consecutive reaction monitoring scan type (MS-MSMS), which provides a very high specificity for the analysis. Briefly, the mass of CE (m/z 318.2), designated as parent ion A, was selected and isolated by ejecting all other ions. Its excitation and collision with damping He gas resulted in unique fragment or daughter ions, the most abundant of which was m/z 196.1, designated as parent ion B. Parent ion B was selected, isolated, and subjected to further fragmentation, resulting again in daughter ions, the most abundant of which was m/z 150.1. The final product ion mass spectra showed only the ions with m/z 196.1 and 150.1 because the mass analyzer was set to monitor these ions only. To establish the limits of detection of the technique, 3 concentrations of CE were prepared by dilution of standard CE (1 mg/mL) with 0.5% formic acid to 0.39, 0.0982, and 0.00615 pmol/mL (12.5, 0.16, and 0.0004 ng/mL of CE, respectively). Formic acid (0.5% in water) provided the baseline. Positive signals at m/z 196.1 and 150.1 were obtained with all dilutions. A CE concentration of 0.0004 ng/ mL was arbitrarily defined as the lower limit of detection and urine samples that produced daughter ions of m/z 196.1 and 150.1 with ion intensities greater than obtained with this concentration were deemed positive.
2.5. Statistical methods Data are presented as median with interquartile range (IQR, 25%-75%). Ratio data are presented as percentages. The χ2 test was used to compare the categorical data. We used nonparametric tests (Man-Whitney U and KruskalWallis) to compare continuous variables when appropriate. Logistic regression was conducted on the following variables: whether admitted to the ICU (yes/no) (the dependent variable) and predictors as follows, presence of CE, injury mechanism (blunt/penetrating), BD, age, and ISS. Rather than automatically introducing continuous predictors (thus assuming linearity of relationship), the linearity assumption was first tested by dividing BD, age, and ISS into deciles and comparing odds of outcome across decile groups. For age, it was apparent that there was little association for patients younger than 60; the odds of admission increased in a more linear fashion above this age. It was therefore decided to model age using a piecewise
1053 regression structure, with one slope for patients younger than 60 years and a second slope for those 60 years or older. Similarly for ISS, separate slopes were modeled for ISS less than 6 and ISS 6 or higher. A single empirical cutoff point (−1.25) was established for BD, which was used as a categorical predictor. Lactate was removed from the model because of missing values for 17% of the patients. The presence of interactions among predictors was tested; the Hosmer-Lemeshow goodness-of-fit test was applied. The predictive utility of the logistic regression model was assessed using the c-statistic, which corresponds to the area under a receiver operating characteristic curve (a value of 0.5 indicates no utility; a value of 1.0 indicates perfect prediction). In a second analysis, the CE predictor was replaced by a variable defining 4 toxicology groups: CE present (CE+), ethanol only (ETOH+), cocaine alone (COC+), and none (no ethanol or cocaine). Five patients with positive results for both alcohol and cocaine but negative results for CE were excluded from this analysis. The Kruskal-Wallis test was used to determine whether the distribution of length of hospital stay differed for patients in the 4 chemical groups. Statistical significance was defined as P b .05. All statistical tests were 2-tailed. Calculations were done using SPSS for Windows, Rel. 14.0 (SPSS Inc, Chicago, Ill).
3. Results During the study period, 458 patients were enrolled. A total of 41 patients were excluded because of inconclusive results for cocathylene, absence of sample for urine toxicology screen, or death of subject in the ED. Three additional patients were excluded because of non–traumarelated admissions. The analysis was performed on the remaining 417 patients. Table 1 Baseline characteristics of CE-positive and CEnegative groups Variable
CE-positive (n = 37)
CE-negative (n = 380)
Pa
Age Age ≥60 Sex (male) Mechanism (blunt) BD (mmol/L) BD b1.25 Lactate (mmol/L) ISS ISS ≥6
47 (33 to 54) b 14% 86% 70%
39 (27 to 52) 13% 73% 70%
.1 .6 .06 1.0
−1.1 (−3.2 to 1.75) 49% 2.2 (1.3 to 3.0) 4.0 (2.0 to 9.0) 33%
0.2 (−2.5 to 2.2) 35% 2.1 (1.5 to 3.3) 4.0 (1.0 to 10.0) 42%
.2 .08 .5 .9 .2
a Man-Whitney U test for continuous variables and χ2 test for categorical variables. b Median and 25% to 75% interquartiles.
1054
S.E. Wiener et al.
Table 2 Comparison of the outcome measures in CE-positive and CE-negative groups (univariate analysis) Outcome
CE-positive CE-negative P (n = 37) (n = 380)
Mortality 1 (3%) Length of hospital stay (d) 5 (3-13) b ICU admission 16 (43%) a b c
11 (2.9%) 5 (2-7) b 90 (24%)
1.00 a .17 c .01 a
χ2 test. Median and IQR (25%-75%). Man-Whitney U test.
The sample consisted of 307 males (74%) and 110 females (26%), with ages ranging from 13 to 96 years (median, 40 years; IQR, 27-53 years). Seventy percent of injuries (n = 291) were inflicted by a blunt mechanism. Isolated head injury was reported in 84 patients (20%). Other patients had either multiple trauma involving head injury (113, 27%) or no head injury (174, 42%). A total of 37 patients (8.9%) tested positive for CE. Other groups included ethanol only (n = 56, 13.4%), cocaine only (n = 17, 4.1%), cocaine and ethanol (negative for cocathylene; n = 5, 1.2%), and none (n = 190, 46%). Table 1 presents the comparison of patients with and without CE. The in-hospital mortality in the study cohort was 3% (n = 12), with only 1 deceased patient belonging to the CEpositive group. Of 37 CE-positive patients, 16 (43%) were admitted to the ICU; of 380 CE-negative patients, 90 (24%) were admitted to ICU. We did not observe any significant difference between the groups comparing length of hospital stay or mortality (Table 2). However, ICU admission was significantly higher in patients with cocaetheylene (43% vs 24%, P = .01). In univariate analysis, ICU admission odds ratio (OR) for CE-positive patients relative to CE-negative patients was 2.6 (95% confidence interval [CI], 1.2-5.5). In multivariate analysis, the OR increased to 4.5 (95% CI, 1.613). The c-statistic for this model was 0.88. In a second analysis, we incorporated the various toxicology groups (CE-positive, cocaine alone, ethanol alone, and none) in the logistic regression analysis. This analysis was performed using age, mechanism, ISS, BD, and toxicology groups. The OR for ICU admission in CE-positive patients compared to other groups was 5.9 (95% CI, 1.6-22). Median (range) length of hospital stay for patients in the 4 chemical groups were the following: CE-positive, 5 days (IQR, 0-52); cocaine only, 4 days (IQR, 0-35); ethanol only, 7 days (IQR, 0-49); ethanol- and cocaine-positive (no cocathylene), 5 days (IQR, 0-128). No significant difference was found among the 4 distributions (P = .237).
4. Discussion The popularity and accessibility of cocaine in recent decades has created interest in its influence in human
injuries, especially when it is combined with the more common use of ethanol. Cocaine and ethanol in combination increases the potency, bioavailability, half-life, and volume of distribution of cocaine and its active metabolites [9,17,18,22]. The combination also results in the formation of the active metabolite CE [8], which in animal studies is suggested to exert harmful hemodynamic and behavioral effects. In dogs, CE causes profound cardiodepressive effects, decreasing contractility, stroke volume, and mean arterial pressure while increasing the incidence of dysrhythmias and electrocardiographic abnormalities [14-16]. Cocaethylene also causes neurologic dysfunction in rats by rendering higher dopamine concentration in brain tissue [10,11] and a decrease in task-specific behavior [13]. Although these studies suggest the existence of consequential CE or combined cocaine and ethanol toxicity, they cannot be unconditionally extrapolated to human cases because the animals studied were often given much higher doses of CE than would be present in humans after concomitant use of ethanol and cocaine. Furthermore, the chronicity and binge behavior of such abuse in humans must be taken into account, as well as possible species-specific differences in susceptibility. In some human studies, the combination of ethanol and cocaine use results in greater chronotropic effects than from either substance alone [7,23,24]. Only one study has looked at outcomes of inpatients with detectable CE, and no significant correlation between concentrations of cocaine or its metabolites and severity of disease or outcome was found [19]. However, no trauma patients were among the population studied, although CE has been identified more often in patients associated with traumatic events [25,26]. This is the first study to examine the correlation of CE and trauma outcomes. Although no difference in mortality or length of hospital stay was found, there was a marked increase in the odds of ICU admission after trauma when there was detectable CE in the urine. These odds were even greater when the presence of cocaine alone and ethanol alone were controlled for by multivariate analysis. Whether this was due to the cardiovascular and hemodynamic effects of CE as observed in animal models and human pharmacophysiology studies, to behavioral effects, or to some other mechanism yet unknown cannot be determined from the study. The presenting BD and lactate concentrations were similar between CE-positive and CE-negative groups, which would suggest a mechanism other than direct effects on the heart and vasculature. Also of interest were the 5 patients who tested positive for both cocaine and ethanol but negative for CE. There are several potential explanations for this. First, these patients may reveal the limitations of our assay. This is unlikely because in preliminary development of our assay, we were able to detect positive controls in concentrations as low as 0.0004 ng/mL, many orders of magnitude lower than the threshold for a positive benzoylecgonine screen. Another
Cocaethylene and Trauma possibility is that these patients may have used cocaine several days before the injury that led to admission but drank ethanol close to the time of admission. This would account for a positive benzoylecgonine screen, but the cocaine parent compound would not be present simultaneously with ethanol to form the CE metabolite. Finally, it is possible that individuals differ in their ability to produce this metabolite, and our population may be a genetically different group than those previously studied.
5. Limitation There were several limitations of the study. Most significant was the convenience sampling in enrolling patients. It is possible that there were differences in patterns of substance use among patients injured during the day compared to patients injured at night, which could have affected results. In addition, because treatment and disposition were left to the discretion of the treating physicians, it is possible that there were different patterns of practice between day and night physicians. If there were a greater tendency to admit to the ICU at night, when more patients using cocaine and alcohol tend to present, this could bias the results. It is unlikely, however, that knowledge of the patients' substance abuse biased the results. Although clinicians had access to the serum ethanol and urine benzoylecgonine screening results, they did not have access to the CE results. Because the difference in ICU admission rate was even greater when controlled for cocaine and ethanol, these test results could not have been responsible.
6. Conclusions Patients with detectable CE in their urine have greater odds of requiring ICU admission after traumatic injury than those without detectable CE. The mechanism by which this occurs is unclear. There was no significant difference in mortality or length of hospital stay.
References [1] Clark RF. Toxicology screening of the trauma patient: a changing profile. Ann Emerg Med 1991;20:151-3. [2] Lindenbaum GA, Carroll SF, Daskal I, Kapusnick R. Patterns of alcohol and drug abuse in an urban trauma center: the increasing role of cocaine abuse. J Trauma 1989;29(12):1654-8. [3] Marzuk PM, Tardiff K, Leon AC, et al. Fatal injuries after cocaine use as a leading cause of death among young adults in New York City. N Engl J Med 1995;332:1753-7. [4] Rivara FP, Mueller BA, Fligner CL, Luna G, Raisys VA, Copass M, et al. Drug use in trauma victims. J Trauma 1989;29(4):462-71.
1055 [5] Sloan EP, Zalenski RJ, Smith RF, et al. Toxicology screening in urban trauma patients: drug prevalence and its relationship to trauma severity and management. J Trauma 1989;29:1647-53. [6] U.S. Dept of Health and Human Services, Office of Applied Studies, 2007 National Survey on Results from the 2007 National Survey on Drug Use and Health: National Findings (NSDUH Series H-34, DHHS Publication No. SMA 08-4343), http://oas.samhsa.gov/ nsduh/2k7nsduh/2k7Results.cfm#TOC, (accessed on 5/1/09). [7] Foltin RW, Fischman MW, Levin FR. Cardiovascular effects of cocaine in humans: laboratory studies. Drug and Alcohol Dependence 1995;37:193-210. [8] Harris DS, Everhart ET, Mendelson J, Jones RT. The pharmacology of cocaethylene in humans following cocaine and ethanol administration. Drug Alcohol Depend 2003:169-82. [9] Hart CL, Jatlow P, Sevarino KA, McCance-Katz EF. Comparison of intravenous cocaethylene and cocaine in humans. Psychopharmacology 2000;149:153-62. [10] Pan WJ, Hedaya MA. Cocaine and alcohol interactions in the rat: effect on cocaine pharmacokinetics and pharmacodynamics. J Pharm Sci 1999;88:459-67. [11] Pan WJ, Hedaya MA. Cocaine and alcohol interactions in the rat: contribution of cocaine metabolites to the pharmacological effects. J Pharm Sci 1999;88:468-76. [12] Perez-Reyes M, Jeffcoat AR. Ethanol/cocaine interaction: cocaine and cocaethylene plasma concentrations and their relationship to subjective and cardiovascular effects. Life Sci 1992;51:553-63. [13] Sobel BF, Riley AL. The interaction of cocaethylene and cocaine and of cocaethylene and alcohol on schedule-controlled responding in rats. Psychopharmacology 1999;145:153-61. [14] Wilson LD, French S. Cocaethylene's effects on coronary artery blood flow and cardiac function in a canine model. J Toxicol Clin Toxicol 2002;40:535-46. [15] Wilson LD, Henning RJ, Suttheimer C, et al. Cocaethylene causes dose-dependent reductions in cardiac function in anesthetized dogs. J Cardiovasc Pharm 1995;26:965-73. [16] Wilson LD, Jeromin J, Garvey L, Dorbandt A. Cocaine, ethanol, and cocaethylene cardiotoxicity in an animal model of cocaine and ethanol abuse. Acad Emerg Med 2001;8:211-22. [17] McCance-Katz EF, Kosten TR, Jatlow P. Concurrent use of cocaine and alcohol is more potent and potentially more toxic than use of either alone—a multiple-dose study. Biol Psychiatry 1998;44:250-9. [18] McCance-Katz EF, Price LH, Kosten TR, Jatlow PI. Cocaethylene: pharmacology, physiology and behavioral effects in humans. J Pharmacol Exp Ther 1995;274:215-23. [19] Blaho K, Logan B, Winbery S, Park L, Schwilke E. Blood cocaine and metabolite concentrations, clinical findings, and outcome of patients presenting to an ED. Am J Emerg Med 2000;18:593-8. [20] Baker SP, O'Neill B, Haddon Jr W, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14:187-96. [21] Champion HR, Copes WS, Sacco WJ, Lawnick MM, Keast SL, Bain Jr LW, et al. The Major Trauma Outcome Study: establishing national norms for trauma care. J Trauma 1990;30:1356-65. [22] Andrews P. Cocaethylene toxicity. J Addict Dis 1997;16:75-84. [23] Farré M, De La Torre R, Gonzalez M, et al. Cocaine and alcohol interactions in humans: neuroendocrine effects and cocaethylene metabolism. J Pharmacol Exp Ther 1997;283:164-76. [24] Farré M, De La Torre R, Llorente M, et al. Alcohol and cocaine interactions in humans. J Pharmacol Exp Ther 1993;266:1364-73. [25] Signs SA, Dickey-White HI, Vanek VW, et al. The formation of cocaethylene and clinical presentation of ED patients testing positive for the use of cocaine and ethanol. Am J Emerg Med 1996;14:665-70. [26] Vanek VW, Dickey-White HI, Signs SA, et al. Concurrent use of cocaine and alcohol by patients treated in the emergency department. Ann Emeg Med 1996;28:508-14.
www.elsevier.com/locate/ajem
Original Contribution
Patients with detectable cocaethylene are more likely to require intensive care unit admission after trauma Sage E. Wiener MD a , Darrell Sutijono MD a , Cynthia H. Moon MD a , Ramanand A. Subramanian PhD a , Jim Calaycay MS b , Julie I. Rushbrook PhD b , Shahriar Zehtabchi MD a,⁎ a
Department of Emergency Medicine, State University of New York, Downstate Medical Center, Box: 1228, Brooklyn, NY 11203, USA b Department of Biochemistry, State University of New York, Downstate Medical Center, Brooklyn, NY 11203, USA Received 4 June 2009; revised 27 June 2009; accepted 27 June 2009
Abstract Cocaethylene (CE) is a toxic metabolite that is formed after simultaneous consumption of cocaine and ethanol. This potent stimulant is more toxic than cocaine and has a longer half-life. The deleterious hemodynamic and cardiovascular effects of CE have been proven in animal models. The aim of this study is to assess the impact of CE on clinical outcomes after trauma. We prospectively enrolled adult (≥13 years) trauma patients requiring admission. Predictor variables were age, sex, mechanism of injury, Injury Severity Score, base deficit, and toxicology groups (ethanol alone, cocaine alone, CE, and none). The outcomes examined were mortality, intensive care unit (ICU) admission, and length of hospital stay (LOS). We used nonparametric tests to compare continuous variables and χ2 test to compare categorical data. We constructed a logistic regression to identify variables that could predict mortality and ICU admission. We enrolled 417 patients (74% male; 70% blunt injury; median age, 40 [range, 13-95]; overall mortality, 2.2%). Urine toxicology and serum ethanol level screens classified patients into the following groups: 13.4% ethanol only, 4.1% cocaine only, 8.9% CE, and 46% none. Mortality and LOS were not statistically different among the groups. In logistic regression analysis, none of the variables were statistically significant in predicting mortality. However, the presence of CE significantly increased the likelihood of ICU admission (odds ratio, 5.9; 95% confidence interval, 1.6-22). The presence of detectable CE in the urine does not increase the mortality or LOS in trauma patients requiring admission but does increase the likelihood of ICU admission. © 2010 Published by Elsevier Inc.
1. Introduction The relationship between substance abuse and traumatic injuries has long been recognized. Many studies have demonstrated an association between drug use and an increased incidence of injuries with poor outcome [1-5]. Although patterns of abuse and injury have shifted some⁎ Corresponding author. Tel.: +1 718 789 1928; fax: +1 718 245 4799. E-mail address: [email protected] (S. Zehtabchi). 0735-6757/$ – see front matter © 2010 Published by Elsevier Inc. doi:10.1016/j.ajem.2009.06.014
what, alcohol and cocaine continue to be widely available and are among the most commonly abused substances in the United States according to the US Department of Health and Human Services [6]. Pharmacophysiologic studies in animals and to a lesser degree in humans have elucidated the potentially toxic influences of these 2 substances on hemodynamic and cardiac functions. When alcohol and cocaine are metabolized to cocaethylene (CE), these consequences are more severe and prolonged [7-16]. Psychopharmacologic literature [17,18] has also suggested
1052 harmful behavioral effects from concomitant use of alcohol and cocaine, possibly related to the formation of CE. Surprisingly, the causative relationship between substance abuse and trauma outcome has been difficult to establish. Furthermore, the mechanism by which these substances might exert effects on the outcome of human injuries is still unclear. Still, many institutions continue to screen routinely for substance abuse in trauma patients despite a lack of evidence that these tests are useful in predicting outcome or changing management. Very few studies have been undertaken to show the influence of substance abuse on care or outcome of patients in the in-hospital setting. Blaho et al [19] attempted to demonstrate the clinical relevance of cocaine metabolites but was not able to show a significant correlation between cocaine or metabolite concentration and the severity of disease or outcome. This study did not discriminate among the nature of illness and did not follow patients throughout the inpatient course. In our investigation, we examined trauma outcomes, namely, the in-patient course and mortality, and explored the role of CE formation in particular. Our objective was to study the relationship between CE and trauma outcomes as measured by mortality, need for intensive care unit (ICU) admission, and length of hospital stay in emergency department (ED) patients who test positive for this substance.
2. Materials and methods 2.1. Study design This was a prospective study using a convenience sample of trauma patients who presented to the ED. Our goal was to investigate the impact of CE on the clinical outcomes of trauma patients. Institutional review board approval was obtained, and waiver of written consent granted.
2.2. Study setting and population The study was conducted from January 2005 to August 2008 at a level I trauma center with approximately 140 000 annual ED visits and serving residents of a largely lowincome, minority community (about 80% African American/ Caribbean, 11% Hispanic, 5% white, 1% Asian, and 3% other races). We enrolled adult trauma patients (13 years and older) presenting to the ED with penetrating or blunt trauma of sufficient severity to require hospital or ICU admission. We excluded patients who died, who were discharged, or who were transferred to the operating room before completion of enrollment and data collection.
2.3. Study protocol Patients were enrolled by convenience sampling, mainly because research staff were not available at all times.
S.E. Wiener et al. However, when academic associates were available (generally weekdays from 10 AM to 10 PM), they enrolled a consecutive sample of trauma patients. Academic associates are medical students or undergraduate students who assist our department in data collection during their research elective. They were trained before the start of this study and were certified to assist after passing the “Protection of Human subjects” course offered by our institutional review board. Data abstractors were not involved in patient care. Treating physicians were not blinded to vital signs or results of urine toxicology, base deficit (BD), and lactate testing. Patient evaluation and treatment were not specified in the study protocol and were at the discretion of the attending emergency physician and trauma surgeon. On designated shifts, the academic associates recorded demographic data, vital signs, mechanism of injury, BD, and lactate. Once the results of imaging studies (radiographs and computed tomography scans) and all other diagnostic procedures were available, the data abstractors documented them. Upon completion of the trauma evaluation, Injury Severity Score (ISS) was calculated. The ISS was graded according to the 1990 revision of the Abbreviated Injury Scale [20,21]. From patients who tested positive for cocaine on routine urine toxicology screening (a routine part of the trauma protocol in our institution), another urine sample was obtained for testing for CE. If, for any reason, obtaining the second urine sample for CE analysis was not feasible in the ED, we allowed sample collection in the inpatient ward or in the ICU up to 72 hours after presentation. This decision was made based on the half-life of CE and its detectability in the urine. Because of the faster clearance of ethanol from serum and variability in time lapse in sending serum samples for ethanol measurement, the presence of detectable ethanol in the serum was not a requirement for enrollment in the study. This was to ensure capture of patients who consumed both substances but had sufficient passage of time before presentation to allow ethanol levels to drop below detectable levels. The outcome measures of interest included the following: in-hospital mortality, ICU admission, and length of hospital stay. The predictor variables included age, sex, mechanism of injury, ISS, BD, lactate, and toxicology group (ethanol alone, cocaine alone, cocathylene, cocaine and ethanol but no cocathylene, and none).
2.4. Measurements The urine toxicology for detection of benzoylecgonine (the primary cocaine metabolite) was performed using the kinetic interaction of microparticles in solution technique (Roche Modular ISE1800 and D2400; Roche Diagnostics, Indianapolis, IN). In the isolation and detection of CE in urine samples, urine samples varying in volume from 3 to 60 mL were centrifuged at 3500 rpm for 5 minutes at 5°C. Using a
Cocaethylene and Trauma peristaltic pump, the supernatant was infused at a flow rate of 0.75 mL/min onto a Sep-Pak C18 cartridge (Waters Corp, Milford, MA), preconditioned with 15 mL each of 90:9:1 CH3CN/H2O/TFA and 98:1:1 water/CH3CN/TFA. Selective elution of the CE was accomplished with 3.0 mL 81:18:1 water/CH3CN/TFA followed by concentration to dryness using a Savant vacuum evaporator. The samples were then redissolved in 100 μL of 90:9:1 CH3CN/H2O/HCOOH and analyzed using a mass spectrometer (a Finnigan LCQ Deca XP Ion Trap mass spectrometer equipped with a microspray ion source from ThermoFinnigan, San Jose, CA). Samples were introduced into the mass spectrometer using a syringe pump at a flow rate of 3 mL/min. Data were acquired in a 3stage consecutive reaction monitoring scan type (MS-MSMS), which provides a very high specificity for the analysis. Briefly, the mass of CE (m/z 318.2), designated as parent ion A, was selected and isolated by ejecting all other ions. Its excitation and collision with damping He gas resulted in unique fragment or daughter ions, the most abundant of which was m/z 196.1, designated as parent ion B. Parent ion B was selected, isolated, and subjected to further fragmentation, resulting again in daughter ions, the most abundant of which was m/z 150.1. The final product ion mass spectra showed only the ions with m/z 196.1 and 150.1 because the mass analyzer was set to monitor these ions only. To establish the limits of detection of the technique, 3 concentrations of CE were prepared by dilution of standard CE (1 mg/mL) with 0.5% formic acid to 0.39, 0.0982, and 0.00615 pmol/mL (12.5, 0.16, and 0.0004 ng/mL of CE, respectively). Formic acid (0.5% in water) provided the baseline. Positive signals at m/z 196.1 and 150.1 were obtained with all dilutions. A CE concentration of 0.0004 ng/ mL was arbitrarily defined as the lower limit of detection and urine samples that produced daughter ions of m/z 196.1 and 150.1 with ion intensities greater than obtained with this concentration were deemed positive.
2.5. Statistical methods Data are presented as median with interquartile range (IQR, 25%-75%). Ratio data are presented as percentages. The χ2 test was used to compare the categorical data. We used nonparametric tests (Man-Whitney U and KruskalWallis) to compare continuous variables when appropriate. Logistic regression was conducted on the following variables: whether admitted to the ICU (yes/no) (the dependent variable) and predictors as follows, presence of CE, injury mechanism (blunt/penetrating), BD, age, and ISS. Rather than automatically introducing continuous predictors (thus assuming linearity of relationship), the linearity assumption was first tested by dividing BD, age, and ISS into deciles and comparing odds of outcome across decile groups. For age, it was apparent that there was little association for patients younger than 60; the odds of admission increased in a more linear fashion above this age. It was therefore decided to model age using a piecewise
1053 regression structure, with one slope for patients younger than 60 years and a second slope for those 60 years or older. Similarly for ISS, separate slopes were modeled for ISS less than 6 and ISS 6 or higher. A single empirical cutoff point (−1.25) was established for BD, which was used as a categorical predictor. Lactate was removed from the model because of missing values for 17% of the patients. The presence of interactions among predictors was tested; the Hosmer-Lemeshow goodness-of-fit test was applied. The predictive utility of the logistic regression model was assessed using the c-statistic, which corresponds to the area under a receiver operating characteristic curve (a value of 0.5 indicates no utility; a value of 1.0 indicates perfect prediction). In a second analysis, the CE predictor was replaced by a variable defining 4 toxicology groups: CE present (CE+), ethanol only (ETOH+), cocaine alone (COC+), and none (no ethanol or cocaine). Five patients with positive results for both alcohol and cocaine but negative results for CE were excluded from this analysis. The Kruskal-Wallis test was used to determine whether the distribution of length of hospital stay differed for patients in the 4 chemical groups. Statistical significance was defined as P b .05. All statistical tests were 2-tailed. Calculations were done using SPSS for Windows, Rel. 14.0 (SPSS Inc, Chicago, Ill).
3. Results During the study period, 458 patients were enrolled. A total of 41 patients were excluded because of inconclusive results for cocathylene, absence of sample for urine toxicology screen, or death of subject in the ED. Three additional patients were excluded because of non–traumarelated admissions. The analysis was performed on the remaining 417 patients. Table 1 Baseline characteristics of CE-positive and CEnegative groups Variable
CE-positive (n = 37)
CE-negative (n = 380)
Pa
Age Age ≥60 Sex (male) Mechanism (blunt) BD (mmol/L) BD b1.25 Lactate (mmol/L) ISS ISS ≥6
47 (33 to 54) b 14% 86% 70%
39 (27 to 52) 13% 73% 70%
.1 .6 .06 1.0
−1.1 (−3.2 to 1.75) 49% 2.2 (1.3 to 3.0) 4.0 (2.0 to 9.0) 33%
0.2 (−2.5 to 2.2) 35% 2.1 (1.5 to 3.3) 4.0 (1.0 to 10.0) 42%
.2 .08 .5 .9 .2
a Man-Whitney U test for continuous variables and χ2 test for categorical variables. b Median and 25% to 75% interquartiles.
1054
S.E. Wiener et al.
Table 2 Comparison of the outcome measures in CE-positive and CE-negative groups (univariate analysis) Outcome
CE-positive CE-negative P (n = 37) (n = 380)
Mortality 1 (3%) Length of hospital stay (d) 5 (3-13) b ICU admission 16 (43%) a b c
11 (2.9%) 5 (2-7) b 90 (24%)
1.00 a .17 c .01 a
χ2 test. Median and IQR (25%-75%). Man-Whitney U test.
The sample consisted of 307 males (74%) and 110 females (26%), with ages ranging from 13 to 96 years (median, 40 years; IQR, 27-53 years). Seventy percent of injuries (n = 291) were inflicted by a blunt mechanism. Isolated head injury was reported in 84 patients (20%). Other patients had either multiple trauma involving head injury (113, 27%) or no head injury (174, 42%). A total of 37 patients (8.9%) tested positive for CE. Other groups included ethanol only (n = 56, 13.4%), cocaine only (n = 17, 4.1%), cocaine and ethanol (negative for cocathylene; n = 5, 1.2%), and none (n = 190, 46%). Table 1 presents the comparison of patients with and without CE. The in-hospital mortality in the study cohort was 3% (n = 12), with only 1 deceased patient belonging to the CEpositive group. Of 37 CE-positive patients, 16 (43%) were admitted to the ICU; of 380 CE-negative patients, 90 (24%) were admitted to ICU. We did not observe any significant difference between the groups comparing length of hospital stay or mortality (Table 2). However, ICU admission was significantly higher in patients with cocaetheylene (43% vs 24%, P = .01). In univariate analysis, ICU admission odds ratio (OR) for CE-positive patients relative to CE-negative patients was 2.6 (95% confidence interval [CI], 1.2-5.5). In multivariate analysis, the OR increased to 4.5 (95% CI, 1.613). The c-statistic for this model was 0.88. In a second analysis, we incorporated the various toxicology groups (CE-positive, cocaine alone, ethanol alone, and none) in the logistic regression analysis. This analysis was performed using age, mechanism, ISS, BD, and toxicology groups. The OR for ICU admission in CE-positive patients compared to other groups was 5.9 (95% CI, 1.6-22). Median (range) length of hospital stay for patients in the 4 chemical groups were the following: CE-positive, 5 days (IQR, 0-52); cocaine only, 4 days (IQR, 0-35); ethanol only, 7 days (IQR, 0-49); ethanol- and cocaine-positive (no cocathylene), 5 days (IQR, 0-128). No significant difference was found among the 4 distributions (P = .237).
4. Discussion The popularity and accessibility of cocaine in recent decades has created interest in its influence in human
injuries, especially when it is combined with the more common use of ethanol. Cocaine and ethanol in combination increases the potency, bioavailability, half-life, and volume of distribution of cocaine and its active metabolites [9,17,18,22]. The combination also results in the formation of the active metabolite CE [8], which in animal studies is suggested to exert harmful hemodynamic and behavioral effects. In dogs, CE causes profound cardiodepressive effects, decreasing contractility, stroke volume, and mean arterial pressure while increasing the incidence of dysrhythmias and electrocardiographic abnormalities [14-16]. Cocaethylene also causes neurologic dysfunction in rats by rendering higher dopamine concentration in brain tissue [10,11] and a decrease in task-specific behavior [13]. Although these studies suggest the existence of consequential CE or combined cocaine and ethanol toxicity, they cannot be unconditionally extrapolated to human cases because the animals studied were often given much higher doses of CE than would be present in humans after concomitant use of ethanol and cocaine. Furthermore, the chronicity and binge behavior of such abuse in humans must be taken into account, as well as possible species-specific differences in susceptibility. In some human studies, the combination of ethanol and cocaine use results in greater chronotropic effects than from either substance alone [7,23,24]. Only one study has looked at outcomes of inpatients with detectable CE, and no significant correlation between concentrations of cocaine or its metabolites and severity of disease or outcome was found [19]. However, no trauma patients were among the population studied, although CE has been identified more often in patients associated with traumatic events [25,26]. This is the first study to examine the correlation of CE and trauma outcomes. Although no difference in mortality or length of hospital stay was found, there was a marked increase in the odds of ICU admission after trauma when there was detectable CE in the urine. These odds were even greater when the presence of cocaine alone and ethanol alone were controlled for by multivariate analysis. Whether this was due to the cardiovascular and hemodynamic effects of CE as observed in animal models and human pharmacophysiology studies, to behavioral effects, or to some other mechanism yet unknown cannot be determined from the study. The presenting BD and lactate concentrations were similar between CE-positive and CE-negative groups, which would suggest a mechanism other than direct effects on the heart and vasculature. Also of interest were the 5 patients who tested positive for both cocaine and ethanol but negative for CE. There are several potential explanations for this. First, these patients may reveal the limitations of our assay. This is unlikely because in preliminary development of our assay, we were able to detect positive controls in concentrations as low as 0.0004 ng/mL, many orders of magnitude lower than the threshold for a positive benzoylecgonine screen. Another
Cocaethylene and Trauma possibility is that these patients may have used cocaine several days before the injury that led to admission but drank ethanol close to the time of admission. This would account for a positive benzoylecgonine screen, but the cocaine parent compound would not be present simultaneously with ethanol to form the CE metabolite. Finally, it is possible that individuals differ in their ability to produce this metabolite, and our population may be a genetically different group than those previously studied.
5. Limitation There were several limitations of the study. Most significant was the convenience sampling in enrolling patients. It is possible that there were differences in patterns of substance use among patients injured during the day compared to patients injured at night, which could have affected results. In addition, because treatment and disposition were left to the discretion of the treating physicians, it is possible that there were different patterns of practice between day and night physicians. If there were a greater tendency to admit to the ICU at night, when more patients using cocaine and alcohol tend to present, this could bias the results. It is unlikely, however, that knowledge of the patients' substance abuse biased the results. Although clinicians had access to the serum ethanol and urine benzoylecgonine screening results, they did not have access to the CE results. Because the difference in ICU admission rate was even greater when controlled for cocaine and ethanol, these test results could not have been responsible.
6. Conclusions Patients with detectable CE in their urine have greater odds of requiring ICU admission after traumatic injury than those without detectable CE. The mechanism by which this occurs is unclear. There was no significant difference in mortality or length of hospital stay.
References [1] Clark RF. Toxicology screening of the trauma patient: a changing profile. Ann Emerg Med 1991;20:151-3. [2] Lindenbaum GA, Carroll SF, Daskal I, Kapusnick R. Patterns of alcohol and drug abuse in an urban trauma center: the increasing role of cocaine abuse. J Trauma 1989;29(12):1654-8. [3] Marzuk PM, Tardiff K, Leon AC, et al. Fatal injuries after cocaine use as a leading cause of death among young adults in New York City. N Engl J Med 1995;332:1753-7. [4] Rivara FP, Mueller BA, Fligner CL, Luna G, Raisys VA, Copass M, et al. Drug use in trauma victims. J Trauma 1989;29(4):462-71.
1055 [5] Sloan EP, Zalenski RJ, Smith RF, et al. Toxicology screening in urban trauma patients: drug prevalence and its relationship to trauma severity and management. J Trauma 1989;29:1647-53. [6] U.S. Dept of Health and Human Services, Office of Applied Studies, 2007 National Survey on Results from the 2007 National Survey on Drug Use and Health: National Findings (NSDUH Series H-34, DHHS Publication No. SMA 08-4343), http://oas.samhsa.gov/ nsduh/2k7nsduh/2k7Results.cfm#TOC, (accessed on 5/1/09). [7] Foltin RW, Fischman MW, Levin FR. Cardiovascular effects of cocaine in humans: laboratory studies. Drug and Alcohol Dependence 1995;37:193-210. [8] Harris DS, Everhart ET, Mendelson J, Jones RT. The pharmacology of cocaethylene in humans following cocaine and ethanol administration. Drug Alcohol Depend 2003:169-82. [9] Hart CL, Jatlow P, Sevarino KA, McCance-Katz EF. Comparison of intravenous cocaethylene and cocaine in humans. Psychopharmacology 2000;149:153-62. [10] Pan WJ, Hedaya MA. Cocaine and alcohol interactions in the rat: effect on cocaine pharmacokinetics and pharmacodynamics. J Pharm Sci 1999;88:459-67. [11] Pan WJ, Hedaya MA. Cocaine and alcohol interactions in the rat: contribution of cocaine metabolites to the pharmacological effects. J Pharm Sci 1999;88:468-76. [12] Perez-Reyes M, Jeffcoat AR. Ethanol/cocaine interaction: cocaine and cocaethylene plasma concentrations and their relationship to subjective and cardiovascular effects. Life Sci 1992;51:553-63. [13] Sobel BF, Riley AL. The interaction of cocaethylene and cocaine and of cocaethylene and alcohol on schedule-controlled responding in rats. Psychopharmacology 1999;145:153-61. [14] Wilson LD, French S. Cocaethylene's effects on coronary artery blood flow and cardiac function in a canine model. J Toxicol Clin Toxicol 2002;40:535-46. [15] Wilson LD, Henning RJ, Suttheimer C, et al. Cocaethylene causes dose-dependent reductions in cardiac function in anesthetized dogs. J Cardiovasc Pharm 1995;26:965-73. [16] Wilson LD, Jeromin J, Garvey L, Dorbandt A. Cocaine, ethanol, and cocaethylene cardiotoxicity in an animal model of cocaine and ethanol abuse. Acad Emerg Med 2001;8:211-22. [17] McCance-Katz EF, Kosten TR, Jatlow P. Concurrent use of cocaine and alcohol is more potent and potentially more toxic than use of either alone—a multiple-dose study. Biol Psychiatry 1998;44:250-9. [18] McCance-Katz EF, Price LH, Kosten TR, Jatlow PI. Cocaethylene: pharmacology, physiology and behavioral effects in humans. J Pharmacol Exp Ther 1995;274:215-23. [19] Blaho K, Logan B, Winbery S, Park L, Schwilke E. Blood cocaine and metabolite concentrations, clinical findings, and outcome of patients presenting to an ED. Am J Emerg Med 2000;18:593-8. [20] Baker SP, O'Neill B, Haddon Jr W, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14:187-96. [21] Champion HR, Copes WS, Sacco WJ, Lawnick MM, Keast SL, Bain Jr LW, et al. The Major Trauma Outcome Study: establishing national norms for trauma care. J Trauma 1990;30:1356-65. [22] Andrews P. Cocaethylene toxicity. J Addict Dis 1997;16:75-84. [23] Farré M, De La Torre R, Gonzalez M, et al. Cocaine and alcohol interactions in humans: neuroendocrine effects and cocaethylene metabolism. J Pharmacol Exp Ther 1997;283:164-76. [24] Farré M, De La Torre R, Llorente M, et al. Alcohol and cocaine interactions in humans. J Pharmacol Exp Ther 1993;266:1364-73. [25] Signs SA, Dickey-White HI, Vanek VW, et al. The formation of cocaethylene and clinical presentation of ED patients testing positive for the use of cocaine and ethanol. Am J Emerg Med 1996;14:665-70. [26] Vanek VW, Dickey-White HI, Signs SA, et al. Concurrent use of cocaine and alcohol by patients treated in the emergency department. Ann Emeg Med 1996;28:508-14.