Cocaethylene cardiotoxicity in emergency department patients with acute drug overdose

Abstract

Objectives:

Cocaine use results in over 500,000 emergency department (ED) visits annually across the United States and ethanol co-ingestion is reported in 34% of these. Commingling cocaine with ethanol results in the metabolite cocaethylene (CE), which is metabolically active for longer than cocaine alone. Current literature on the cardiotoxicity of CE compared to cocaine alone is limited and lacks consensus. This study aims to fill this gap in the literature and examine cardiovascular events in cocaine use as confirmed by urine toxicology versus CE exposure.

Methods:

This was a secondary data analysis of a prospective cohort study of adult patients with acute drug overdose at two urban tertiary care hospital EDs over 4 years. Patients with positive urinary cocaine metabolites were analyzed, and outcomes were compared between patients with overdose and confirmed presence of cocaine on urine toxicology (cocaine group) and patients with cocaine and ethanol use (CE group). The primary outcome was cardiac arrest. Secondary outcomes included myocardial injury and hyperlactatemia. Data were analyzed using multivariable regression models.

Results:

We enrolled a total of 199 patients (150 cocaine, 49 CE). Rates of cardiac arrest were significantly higher in the CE group compared to cocaine (6.1% vs. 0.67%, p = 0.048). Cocaine was significantly associated with myocardial injury compared to CE exposure (mean initial troponin 0.01 ng/ml vs. 0.16 ng/ml, p = 0.021), while hyperlactatemia was associated with CE exposure (mean initial lactate 4.1 mmol/L vs. 2.9 mmol/L, p = 0.038).

Conclusions:

When compared to cocaine exposure alone, CE exposure in ED patients with acute drug overdose was significantly associated with higher occurrence of cardiac arrest, higher mean lactate concentrations, and lower occurrence of myocardial injury.

INTRODUCTION

Cocaine overdose is attributed to over 500,000 cases annually presenting to emergency departments (EDs) across the United States and co-ingestion with ethanol is reported in 34% of these cases.¹ The combination of cocaine and ethanol results in the production of cocaethylene (CE).^2,3 Cocaine and ethanol co-ingestion is also known to increase the amount of unchanged cocaine excreted in urine which is attributed to a decrease in cocaine hepatic clearance in the presence of ethanol.² Thus, the overall number of cocaine metabolites is decreased. Ethanol leads to nasal capillary vasodilation and is theorized to lead to more rapid cocaine absorption.² Some studies find an increase in serum cocaine concentrations by 30% if alcohol is consumed before cocaine insufflation.⁴ Studies have shown that CE is metabolically active for longer than either drug alone; cocaine has a half-life of 45–90 min, while for CE, it is 144 min.^4–7

Current literature regarding the relative cardiovascular effects of CE compared to cocaine alone are lacking in consensus. Some studies have identified higher rates of cardiotoxicity (cardiac arrest, myocardial injury, arrhythmia or corrected QT interval [QTc] prolongation) with larger quantities of cocaine (defined as 0.5 mg/kg) in comparison with CE³ while others note that CE is more cardiotoxic than cocaine alone and increases the likelihood of hospitalization.^6,8 There is a general lack of human studies that observe the effects of cocaine versus CE alone with most literature regarding adverse cardiovascular events such as cardiac arrest, QTc prolongation, arrhythmia, or myocardial injury limited to animal studies.^9,10 Human studies have exclusively been performed with intravenous formulations of CE and were controlled in the laboratory setting only.² There is a pressing need for studies on the cardiovascular impacts of CE in the context of real-world clinical ingestions.

This study aims to characterize the impact of CE toxicity compared to cocaine toxicity alone (as defined by overdose with confirmed presence of cocaine on urine toxicology) on cardiac arrest rates in adult ED patients with acute drug overdose. We hypothesized that there would be an increased rate of cardiac arrest in the population with CE toxicity compared to the population with cocaine toxicity alone.

METHODS

Study design and setting

This was a secondary data analysis from a prospective cohort study¹¹ at two academic, urban tertiary care EDs from March 2015 through March 2020. The ED at one site has annual patient volumes of approximately 145,000 visits; the second site has approximately 106,000 ED visits annually. Institutional review board approval was obtained with waiver of consent prior to data collection at the study institutions. Study findings are presented in this article in compliance with the STROBE guidelines.

Selection of participants

Patients in the cohort are consecutive ED patients over the age of 18 with acute overdose or poisoning; of these, a secondary data analysis was performed on those with a urine drug screen (UDS) that was positive for cocaine. UDSs were standard immunoassays run by the hospital clinical laboratory. Cohort enrollment and data collection have been previously described.¹¹ Briefly, ED patients with suspected acute drug overdose were screened prospectively by trained research assistants based on presenting chief complaint. Patients were enrolled consecutively 24 h a day over the study period. For enrollment within the original cohort, ED patients with suspected acute drug overdose were screened prospectively by trained research assistants based on presenting chief complaint and were included if history/clinical presentation confirmed patients were presenting within 24 h of exposure to an illicit or prescription drug with resulting clinical symptoms due to this exposure. Exclusion criteria from the cohort were initial ECG with QTc ≥ 500, age under 18, an alternative diagnosis (e.g., stroke, sepsis), nondrug exposure (e.g., caustic, plant), dermal exposure (due to the self-limiting nature of these exposures), chronic toxicity (i.e., symptom onset not within prior 24 h), prisoner status, or missing data. Urine and serum drug screen/level data for enrolled patients were available only if it was obtained as part of routine clinical care. Patients with QTc ≥ 500 were excluded from the parent study cohort as a primary outcome of the original study was QTc prolongation, requiring inclusion of patients with a baseline QTc < 500.

Patients with confirmed presence of cocaine on UDS and an elevated serum ethyl alcohol concentration (level greater than 0) were included in the CE cohort of this secondary analysis as prior data has demonstrated a large percentage of CE metabolites in volunteers given cocaine and alcohol together.¹² Those who did not have an elevated serum ethyl alcohol concentration were included in the cocaine alone cohort.

Data collection

Data were abstracted from the medical chart by several trained research assistants. Data collection from the medical chart occurred in accordance with accepted guidelines for valid medical chart abstraction, including training of abstractors and 95% agreement of a random sampling of 10 test charts prior to mass data abstraction.¹³ Abstractors were blind to study hypothesis. Observational data collected included demographics, exposure information, toxicant identification, initial mental status, prior history of coronary artery disease as determined by the patient’s medical history documented in the electronic health record, toxicology screen results, antidote administration, antidote dose and infusion duration, initial heart rate, initial systolic blood pressure, initial diastolic blood pressure, QTc, initial troponin I concentrations, peak troponin I concentrations, initial serum lactic acid level, vasopressor requirement, cardiopulmonary resuscitation (CPR) requirement, total arrhythmias (including ventricular fibrillation, torsades, and ventricular tachycardia), bicarbonate concentrations, and total number of patients who expired during hospitalization. For QTc, the computer-generated corrected QT interval (Bazett’s corrected QTc, QT/√RR) was used as it has been previously validated for this purpose.^11,14

Outcomes

The primary study outcome was the difference in occurrence of in-hospital cardiac arrest requiring CPR between the CE cohort and the cocaine-only cohort. Secondary outcomes were the difference in mean initial serum troponin concentration and mean lactate concentration between the cohorts. We opted to use initial rather than peak serum troponin concentration in our analyses based on prior data which has shown initial troponin to be superior to peak troponin in prediction of drug overdose mortality.¹⁵

Data analysis

Descriptive statistics examining patient demographics and clinical characteristics were calculated. Two-sample t-tests (for normally distributed variables) and chi-square tests (for categorical variables) were employed to compare demographic and clinical characteristics between groups. Chi-square and Fisher’s exact tests for categorical variables, t-tests for normally distributed continuous variables and Mann–Whitney U-tests for nonnormally distributed continuous variables were used to analyze differences in outcomes between the two cohorts.

We subsequently fit multivariable logistic regression models to adjust for demographic and clinical confounding variables. Troponin was dichotomized to a binary outcome with a cutoff point of 0.10 ng/ml. This cutoff value for troponin was a standard cutoff based on 99th percentile cutoff of the laboratory test and validated by prior literature identifying troponin 0.10 ng/ml and greater as an independent prognostic indicator of mortality.¹⁵ Based on this prior literature, we felt that dichotomizing troponin in this manner was the most valid means to draw clinically meaningful conclusions. Lactic acid was dichotomized to a binary outcome with a cutoff point of 5.0 mmol/L based on prior data.¹⁶

Relevant demographic and clinical covariates including age, race, sex, and history of coronary artery disease were selected a priori for inclusion in the models. Due to our relatively small sample size and limitations in serum testing, we decided not to adjust for coingestions as a confounding factor. While the study was performed at two hospital sites, these two sites are within the same geographic location and health system, reflecting a similar patient demographic and with substantial overlap in ED care providers between the two sites. Accordingly, we did not feel that hospital site needed to be incorporated as a confounding variable in our analysis. Analyses were conducted using SAS University Edition v.9.4 (SAS Institute) and SPSS v.24 (IBM).

Sample size and power

Based on a fixed sample size of 200 patients (which was the number of patients available in the database for secondary analysis), assuming 3:1 ratio of cocaine-to-CE patients, we had 80% power to demonstrate fivefold higher odds of the primary outcome in the CE group based on a post hoc power analysis.

RESULTS

Patient enrollment and baseline characteristics

Out of a total of 3138 screened and following application of enrollment criteria, there were a total of 199 patients analyzed. A total of 150 patients were in the cocaine only cohort and 49 patients were in the CE cohort. Study enrollment and application of inclusion/exclusion criteria is summarized in Figure 1. There was no significant differences in patient demographics, baseline coronary artery disease, or number of co-ingestions (based on patient self-report and UDS) between groups. In the cocaine group, most common classes of co-ingestions were opioids (n = 81), benzodiazepines (n = 45), cannabinoids (n = 30), antipsychotics (n = 13), and antidepressants (n = 10). In the CE group, most common classes of co-ingestions were opioids (n = 17), benzodiazepines (n = 12), cannabinoids (n = 7), acetaminophen (n = 7), and antihistamines (n = 4). Patient demographics and clinical characteristics are summarized in Table 1.

FIGURE 1.

Summary of study enrollment and patient inclusion/exclusion. Cocaine, UDS positive for cocaine; cocaine-only cohort, UDS positive for cocaine with no co-ingestion of ethanol; Cocaethylene cohort: UDS positive for cocaine with co-ingestion of ethanol. UDS, urine drug screen

TABLE 1.

Baseline clinical characteristics demographic and clinical characteristics of ED patients

Patient characteristic	Cocaine (n = 150)	CE (n = 49)
Age (years)	41.2 (±13.7)	42.5 (±13.6)
Female	57 (38)	15 (30.6)
Race/ethnicity
White	32 (21.3)	10 (20.4)
Black	35 (23.3)	6 (12.2)
Asian	4 (2.7)	3 (6.1)
Other/unknown	49 (32.7)	16 (32.7)
Hispanic	31 (20.7)	14 (28.6)
Co-ingestions	132 (88)	44 (90)
Coronary artery disease	7 (4.7)	0 (0)
Ethanol concentration	0	152.8 (±90.2)
Presenting vital signs
Pulse (beats/min)	95 (±24)	101 (±24)
MAP	91 (±19)	94 (±20)
Temperature (°F)	97.9 (±1.0)	97.6 (±1.1)
Oxygen saturation	96 (±7)	95 (±10)

Note: Data are reported as mean (±SD) or n (%).

Abbreviation: CE, cocaethylene.

Primary outcome

The proportion of patients who went into cardiac arrest requiring CPR was significantly higher in the CE group compared to cocaine only patients (6.12% [n = 3] vs. 0.67% [n = 1], p = 0.048) on unadjusted analysis. Results are summarized in Table 2. We subsequently performed a multivariable logistic regression to determine the association of CE with cardiac arrest following adjustment for age, race, sex, and history of coronary artery disease. CE remained significantly associated with cardiac arrest on adjusted analysis (adjusted odds ratio [aOR] 12.61, 95% confidence interval [CI] 1.10–144.18). Results are summarized in Table 3.

TABLE 2.

Unadjusted association between cocaine and CE use and study outcomes

Specific outcomes	Cocaine (n = 150)	CE (n = 49)
Cardiac arrest*	1 (0.67)	3 (6.12)
Lactate (mmol/L) initial serum concentration*	2.90 (±3.31)	4.10 (±3.16)
Lactate (dichotomized cutoff 5.0 mmol/L)	16 (10.67)	11 (22.45)
Troponin, initial*	0.01 (±0.03)	0.16 (±0.58)
Myocardial injury (troponin dichotomized cutoff 0.10 ng/ml)	17 (11.33)	1 (2.04)

Note: Data are reported as n (%) or mean (±SD). Myocardial injury is defined as initial serum troponin ≥ 0.10 ng/ml.

Abbreviation: CE, cocaethylene.

^*p < 0.05.

TABLE 3.

Adjusted association between cocaine and CE use and study outcomes

Specific outcomes	CE	Cocaine
Cardiac arresta	12.61 (1.10—144.18)	REF
Lactate elevationb	2.65(1.08–6.49)	REF
Myocardial injuryc	REF	8.96 (1.07–75.24)

Note: Data are reported as aOR (95% CI). Myocardial injury is defined as initial serum troponin ≥ 0.10 ng/ml.

Abbreviations: aOR, adjusted odds ratio; CE, cocaethylene; REF, reference.

^aModel covariates: age, race, sex, history of coronary artery disease.

^bModel covariates: age, race, sex.

^cModel covariates: age, race, sex.

Of the four patients with cardiac arrest, three were tachycardic upon ED arrival with heart rate ranging from 115–142 beats/min and normotensive with mean arterial pressure (MAP) upon arrival ranging from 79 to 88. The fourth patient was bradycardic with a heart rate of 40 beats/min and hypotensive with MAP of 49. Oxygen saturation upon ED arrival for the patients with cardiac arrest ranged from 92% to 100%. Three of the four patients with cardiac arrest experienced myocardial injury with peak troponin ranging from 0.26 ng/ml to 1.58 ng/ml. Three of the four patients with cardiac arrest ultimately expired. Two patients presented with prehospital cardiac arrest.

Secondary outcomes

There was significantly increased myocardial injury seen in the cocaine-only group compared to the CE cohort on unadjusted analysis (mean initial troponin 0.16 ng/ml vs. 0.01 ng/ml, p = 0.021). We subsequently performed a multivariable logistic regression to examine the association between cocaine and myocardial injury following adjustment for age, race, and sex with troponin dichotomized as a binary outcome with troponin of ≥0.1 considered positive for myocardial injury and troponin of <0.1 considered negative. Cocaine use remained significantly associated with myocardial injury on adjusted analysis (aOR 8.96, 95% CI 1.07–75.24) and on a sensitivity analysis incorporating hospital site as an additional covariate.

The CE cohort conversely had significantly higher serum lactate concentrations compared to the cocaine only cohort (mean initial lactate 4.1 mmol/L vs. 2.9 mmol/L, p = 0.038) on unadjusted analysis. Following adjustment for age, race, and sex using a multivariable linear regression model, CE remained significantly associated with higher serum lactate concentrations (aOR 2.65, 95% CI 1.08–6.49). CE also remained significantly associated with higher serum lactate on a sensitivity analysis incorporating hospital site as an additional covariate. Results are summarized in Tables 2 and 3.

Mean (±SD) peak serum troponin was 0.36 (±1.16) ng/ml in the cocaine group and 0.11 (±0.33) ng/ml in the CE group. Initial triage MAP was 91 mm Hg in the cocaine group (SD ± 18.9 mm Hg) and 94 (±19.5) mm Hg in the CE group. There was no significant difference in peak serum troponin or initial triage MAP between groups. Four patients in the cocaine group (2.68%) and two patients in the CE group (4.08%) experienced hypotension requiring vasopressors (p = NS). Results are summarized in Table 4.

TABLE 4.

Detailed clinical characteristics of ED patients

Patient characteristic	Cocaine	CE
Peak serum troponin (ng/ml)	0.36 (±1.16)	0.11 (±0.33)
MAP (mm Hg)	91 (±18.9)	94 (±19.5)
Hypotension requiring vasopressors	4 (2.68)	2 (4.08)

Note: Data are reported as mean (±SD) or n (%). Myocardial injury is defined as initial serum troponin ≥ 0.10 ng/ml.

Abbreviations: CE, cocaethylene; MAP, mean arterial pressure.

DISCUSSION

The main finding of this study was that ED patients with acute cocaine and ethanol ingestion, as confirmed by urine toxicology and serum ethanol level, had significantly increased occurrence of cardiac arrest requiring CPR (although overall rates of cardiac arrest in both groups were low, with only three patients in the CE group and one patient in the cocaine group experiencing cardiac arrest) and higher mean lactate concentrations than patients with cocaine use alone. In contrast, patients with cocaine use alone had significantly higher occurrence of myocardial injury than patients with cocaine and ethanol use.

Our findings are consistent with a prior cohort study noting 76% of cocaine-related sudden cardiac deaths occurred in patients concurrently using ethanol.¹⁷ Potential mechanisms mediating this increased occurrence of cardiac arrest with CE include the higher arrhythmogenicity of CE relative to cocaine as well as the longer elimination half-life and larger volume of distribution (with a larger volume of distribution denoting greater amount of tissue distribution of CE).¹⁸ Our adjusted analysis of CE with cardiac arrest should be interpreted cautiously, however, as only four patients experienced the outcome of interest (cardiac arrest), leading to a broad CI, limiting interpretability of this result.

We similarly found higher mean lactate concentrations in the CE cohort. This difference in lactate concentrations between cocaine and CE is likely explained by the type B lactic acidosis secondary to ethanol metabolism and thiamine deficiency often observed in acute alcohol intoxication due to underlying baseline thiamine deficiency.¹⁹ In the setting of alcohol intoxication, oxidation of ethanol leads to a low ratio of NAD⁺ to NADH, shifting pyruvate production to lactate production.²⁰ Increased lactate concentrations in the CE cohort may also potentially be due to increased agitation and need for use of physical restraints in the CE group; physical restraint use has previously been shown to be associated with increase lactate concentration.²¹

We found significantly higher occurrence of myocardial injury in the cocaine group compared to those with CE toxicity. Our findings differ from a prior study that found significantly higher troponin concentrations in patients using cocaine and ethanol, although this study was limited to a very small cohort of female patients only.²² These findings suggest that cocaine alone may lead to more cardiac myocyte necrosis than CE. Biologically, the interaction between cocaine and ethanol inhibits cocaine metabolism into benzoylecgonine and ecgonine methyl ester.²³ Benzoylecgnonine and ecgnonine methyl ester were previously documented in the literature to be inactive metabolites of cocaine.²⁴ However, prior literature has demonstrated a significant correlation between troponin and benzoylecgnonine concentrations suggesting that the metabolite may be more metabolically active than previously thought.²⁵ One possible explanation for the higher troponin concentrations in the cocaine only cohort is higher benzoylecgnonine concentrations may lead to either benzoylecgnonine-induced vasospasm or direct myocardial cell necrosis. Given that we observed higher rates of cardiac arrest in the CE cohort, the clinical relevance of the troponin elevation in the cocaine cohort remains unclear. Further studies are needed to more fully elucidate the association between CE and serum troponin concentrations and to determine the clinical relevance of this association. Myocardial injury (thought to be due to vasospasm as discussed above) was noted much more frequently in our study population than cardiac arrest; this follows previous literature which has demonstrated that cardiac arrest rates due to myocardial injury and coronary vasospasm are low, ranging from 3.7% to 12%. Accordingly, the relatively lower rates of cardiac arrest compared to myocardial injury appears consistent with prior literature.^26–28

Our study findings suggest potential increased severity of overdose and mortality in patients with CE toxicity compared to cocaine toxicity alone. Providers should be vigilant to assess for co-ingestion of ethanol in patients presenting with cocaine toxicity through targeted clinical history-taking and may consider longer periods of monitoring and observation of these patients although further studies are needed to determine specific duration of observation. While we noted increased troponin concentrations in patients using cocaine compared to patients using cocaine and ethanol, both populations should be considered at-risk for adverse cardiovascular events.

Larger, multicenter trials are needed to validate findings from this study and to improve generalizability. Larger future studies will additionally allow for explanatory analysis to determine mechanisms for increased rates of cardiac arrest and higher lactic acid concentrations due to CE toxicity. Subsequent studies should include serum analysis for presence and concentration of CE itself and other active metabolites to obtain more robust results.

LIMITATIONS

Our study is subject to all the known limitations of cohort studies as well as a relatively small sample size. Our study was also conducted at only two medical centers, limiting generalizability. We also only had access to routine drug assays that were obtained as part of clinical patient care and thus did not confirm the presence of CE itself with a specific assay or other metabolites that may have influenced our findings. Given that drug assay data were available only if collected as part of routine clinical care, not all patients in the original study cohort had UDSs or serum drug levels. Identification of patients with cocaine and CE exposure for this secondary analysis required urine and serum drug assays; accordingly, there may be an element of selection bias in our cohort as only patients who had urine and serum drug assays collected were included in our cohort. We also did not quantify cocaine and CE exposure as the data set only included drug levels that were collected as part of routine clinical care. Urine drug screening, which our study used as a means to identify presence of cocaine, has been shown to be a reliable means to identify presence of cocaine, with 100% positive predictive value for cocaine in prior literature.²⁹

We did not exclude patients with other co-ingestions; however, this did not appear to have biased our findings as there was no significant difference in the proportion of patients with co-ingestions in the CE and cocaine groups (Table 1). While we found significantly higher lactic acid concentrations in patients with CE exposure, this may be reflective of ethanol metabolism rather than organ dysfunction, which may limit clinical/prognostic relevance of this finding. We did not collect data on cardiac risk factors, history of cocaine/alcohol use, history of other substance use, etc., and thus were unable to account for these potential confounding factors in our analysis. Occurrence of the primary study outcome of cardiac arrest was rare with only four total patients in the study cohort undergoing cardiac arrest. This limits our study conclusions and raises the need for future studies with a larger sample size. We also did not obtain data on time to occurrence of cardiac arrest. Patients included in this secondary analysis all had an initial QTc < 500 because the parent study excluded patients with a QTc ≥ 500, which may have biased our study findings.

CONCLUSIONS

When compared to cocaine exposure alone, cocaine and ethanol exposure in ED patients was significantly associated with higher occurrence of cardiac arrest, higher mean lactate concentrations, and lower occurrence of myocardial injury. Further studies are needed to clarify the exact mechanism by which cocaethylene may increase risk of cardiac arrest and hyperlactatemia and to validate these study findings.