A commercially available lipid emulsion was able to delay progression of cocaine cardiac toxicity in vivo. Further, it improved acute cocaine‐ and cocaethylene‐induced cardiac toxicity in rat isolated heart while SBE‐β‐CD was effective only at the highest cocaine concentration. Further, both cocaine and cocaethylene inhibited lipid‐dependent mitochondrial respiration. Collectively, this suggests that scavenging‐independent effects of lipid emulsion may contribute to reversal of acute cocaine and cocaethylene cardiotoxicity, and the beneficial effects may involve mitochondrial lipid processing.

Treatment with lipid emulsion delayed onset of hypotension (140 seconds vs. 279 seconds; p = 0.008) and asystole (369 seconds vs. 607 seconds; p = 0.02) in whole animals. Cocaine and cocaethylene induced dose‐dependent decreases in RPP, +dP/dtmax, and –dP/dtmax abs (p < 0.0001) in Langendorff hearts; line pressure was increased by cocaine and cocaethylene infusion, but not altered by treatment. Lipid emulsion attenuated cocaine‐ and cocaethylene‐induced cardiac depression. SBE‐β‐CD alone evoked a mild cardiodepressant effect (p < 0.0001) but attenuated further cocaine‐ and cocaethylene‐induced decrements in cardiac contractility at high concentrations of drug (100 μmol/L; p < 0.001). Finally, both cocaine and cocaethylene, but not ecgonine and benzoylecgonine, inhibited lipid‐dependent mitochondrial respiration by blocking carnitine exchange (p < 0.05).

For whole animal experiments, Sprague‐Dawley rats were anesthetized, instrumented, and pretreated with lipid emulsion followed by a continuous infusion of cocaine to assess time of onset of cocaine toxicity. For ex vivo experiments, rat hearts were placed onto a nonrecirculating Langendorff system perfused with Krebs‐Henseleit solution. Heart rate, left ventricle maximum developed pressure (LVdevP), left ventricle diastolic pressure, maximum rate of contraction (+dP/dtmax), maximum rate of relaxation (–dP/dtmax), rate‐pressure product (RPP = heart rate × LVdevP), and line pressure were monitored continuously during the experiment. A dose response to cocaine (10, 30, 50, and 100 μmol/L) and cocaethylene (10, 30, and 50 μmol/L) was generated in the absence or presence of either 0.25% lipid emulsion or SBE‐β‐CD. Substrate‐specific rates of oxygen consumption were measured in interfibrillar cardiac mitochondria in the presence of cocaine, cocaethylene, ecgonine, and benzoylecgonine.

Cocaine intoxication leads to over 500,000 emergency department visits annually in the United States and ethanol cointoxication occurs in 34% of those cases. Cardiotoxicity is an ominous complication of cocaine and cocaethylene overdose for which no specific antidote exists. Because infusion of lipid emulsion (Intralipid) can treat lipophilic local anesthetic toxicity and cocaine is an amphipathic local anesthetic, the authors tested whether lipid emulsion could attenuate cocaine cardiotoxicity in vivo. The effects of lipid emulsion were compared with the metabolically inert sulfobutylether‐β‐cyclodextrin (SBE‐β‐CD; Captisol) in an isolated heart model of cocaine and cocaethylene toxicity to determine if capture alone could exert similar benefit as lipid emulsion, which exhibits multimodal effects. The authors then tested if cocaine and cocaethylene, like bupivacaine, inhibit lipid‐based metabolism in isolated cardiac mitochondria.

Cocaine is a commonly abused psychostimulant that acts on the central nervous system. Recreational use of cocaine contributes to more than half‐million emergency department (ED) visits annually in the United States, the most for any illicit drug of abuse.1 Cardiotoxicity is an ominous complication of cocaine overdose.2, 3 Sodium bicarbonate is used to overcome sodium channel blockade and reduce arrhythmias,4 along with benzodiazepines and α‐adrenergic antagonists to reduce the sympathomimetic effects of cocaine.5 However, no specific antidote exists to address cocaine toxicity. Further, in 34% of ED visits, cocaine toxicity presents with concurrent alcohol intoxication.1 Cointoxication slows clearance of cocaine,6, 7 produces the cardiotoxic metabolite cocaethylene,8, 9 and increases the likelihood of hospitalization.10 The complexity of molecular targets and cardiovascular effects of cocaine2, 11 makes treatment of overdose difficult; no single, site‐specific antagonist can address all of the associated clinical complications. Case reports have identified the potential efficacy of intravenous (IV) lipid emulsion (ILE) as a treatment for cocaine overdose in humans12, 13 and in animal models5, 14 based on ILE's ability to reverse local anesthetic toxicity15 via a number of mechanisms, including partitioning of toxic drugs16-20 and improved cardiac output.20-22 We wanted to assess how these effects might combine to combat cocaine‐induced cardiac depression. To this end, we tested the efficacy of ILE in preventing progression of cocaine toxicity to cardiac collapse in vivo. We further investigated mechanistic contributions from two perspectives. First, we tested the potential contribution of nonpartitioning effects by comparing ILE to the metabolically inert sulfobutyl‐β‐cyclodextrin (SBE‐β‐CD; Captisol, Ligand Tech, La Jolla, CA) in an isolated rat heart model. β‐Cyclodextrins are a class of macromolecular structures designed to bind lipophilic moieties and have been proposed as detoxification agents for drug overdose23 and for cocaine in particular.24, 25 They have the ability to scavenge drugs, but no metabolic activity. Second, because cocaine and other local anesthetics impair lipid‐based respiration,26, 27 we specifically tested the hypothesis that cocaine and its metabolites impair mitochondrial carnitine exchange to contribute to the cardiac depression that ILE may directly combat.

Methods Study Design This was a laboratory study of ILE versus SBE‐β‐CD for cocaine toxicity in a murine model. All experimental protocols were approved by the Chicago Veterans Affairs Health Care System Animal Care and Use Subcommittee. Rats were housed as pairs in the Veterinary Medical Unit at the Jesse Brown Veterans Affairs Medical Center (JBVAMC). Protocols were approved by the Institutional Animal Care and Utilization Committee of the JBVAMC. Animal Handling and Preparation We carried out lipid emulsion infusion experiments as previously described.28 Following induction in a Bell jar, we intubated animals (n = 12 total) and then mechanically ventilated and anesthetized them with 2% isoflurane in oxygen. We measured blood pressure via a femoral artery cannula and electrocardiogram by needle electrodes. Animals received infusions through separate femoral vein catheters, after 20‐minute equilibration of vital signs and arterial blood gas readings. We first randomized each animal to receive an infusion of either saline (n = 6) or 20% lipid emulsion (Intralipid, Baxter Pharmaceuticals, Deerfield, IL; n = 6) at 3 mL/kg/min for 5 minutes. Subsequently we delivered a continuous infusion of 10 mg/kg/min cocaine, until asystole was achieved. We assessed for cocaine toxicity based on both a drop in systolic blood pressure to <40 mm Hg and asystole of 5 seconds or more. We chose a blood pressure of <40 mm Hg to assess for profound hypotension that was indicative of poor perfusion and risk of ischemic sequelae.29, 30 Study Protocol Isolated Heart Experiments We performed isolated heart experiments as previously described.21 In brief, Sprague‐Dawley rats (n = 20) weighing between 350 and 430 g were heparinized and anesthetized with pentobarbital, and their hearts were removed and placed onto a Langendorff system perfused at a constant rate of 16 mL/min with Krebs‐Henseleit solution warmed to 37°C (pH adjusted to 7.40), and containing 100 mmol/L sodium chloride (NaCl), 4.74 mmol/L potassium chloride (KCl), 1.18 mmol/L monopotassium phosphate (KH 2 PO 4 ), 1.18 mmol/L magnesium sulfate (MgSO 4 ), 1.00 mmol/L calcium chloride (CaCl 2 ), 25.00 mmol/L sodium bicarbonate (NaHCO 3 ), 11.50 mmol/L glucose, 4.92 mmol/L pyruvate, and 5.39 mmol/L fumarate. Krebs solution was also equilibrated with a mixture of oxygen (95%) and carbon dioxide (5%) by passing through a membrane oxygenator. We measured pressure with a latex balloon in the left ventricle and recorded as in the intact animal experiments using LabChart 5.2 (ADInstruments, Colorado Springs, CO). We recorded heart rate, left ventricle maximum developed pressure (LVdevP), left ventricle diastolic pressure, maximum rate of contraction (dp/dt max), maximum rate of relaxation (–dP/dtmax), rate‐pressure product (RPP = heart rate × LVdevP), and line pressure continuously during the experiment. RPP is a measure of cardiac work or cardiac output that takes into consideration both the maximum pressure generated and the rate of contractions. It is important because it provides a good quantification of energy demand of the heart and its ability to perfuse tissues. We assessed the dose response of perfused hearts to cocaine (10, 30, 50, and 100 μmol/L; Sigma‐Aldrich, St. Louis, MO; n = 5) and cocaethylene (10, 30, and 50 μmol/L; Sigma‐Aldrich, St. Louis, MO; n = 5), in the absence or presence of either 0.25% lipid emulsion (n = 5) or 1.2 μmol/min SBE‐β‐CD (n = 5). Cocaethylene was made as a 5 mmol/L stock solution in 50% ethanol, with final concentrations of 0.1, 0.3, and 0.5% ethanol for 10, 30, and 50 μmol/L cocaethylene, respectively. A cocaethylene‐free ethanol control was run for the 0.1, 0.3, and 0.5% ethanol conditions. Mitochondrial Preps We isolated interfibrillar cardiac mitochondria as previously described27, 31 using differential centrifugation. Following calibration with a Clark oxygen electrode, the addition of 0.1 mmol/L adenosine diphosphate (ADP) depleted the endogenous substrates. Subsequently we added cocaine, cocaethylene, benzoylecgonine, or ecgonine and assessed mitochondrial respiration supported by nonlipid substrate (pyruvate) and by carnitine‐acylcarnitine translocase (CACT)‐dependent activity. We measured concentration‐dependent effects of cocaine (0 to 1 mmol/L), cocaethylene, and cocaine metabolites (benzoylecgonine, ecgonine at 1 mmol/L for both), for state III (ADP‐stimulated) respiration supported by pyruvate (10 mmol/L). Further, we assessed mitochondrial CACT activity by adding malonate, pyruvate, and carnitine to respiring mitochondria and determining the rate of carnitine‐stimulated pyruvate oxidation. Malonate inhibits the tricarboxylic acid cycle, so acetyl‐coenzyme A (acetyl‐CoA) accumulates in mitochondria and respiration is suppressed by feedback inhibition of pyruvate dehydrogenase, which catalyzes the reaction converting pyruvate to acetyl‐CoA, generating reduced nicotinamide adenine dinucleotide (NADH) in the process. The addition of carnitine to this system leads to the conversion of acetyl‐CoA to the acetylcarnitine, which is shuttled out of mitochondria by CACT, thereby relieving the inhibition of pyruvate dehydrogenase by acetyl‐CoA and allowing generation of NADH, which supports respiration by providing reducing equivalents to the hydrogen electron transport system. Under these conditions, activation of respiration by added carnitine is directly correlated to CACT activity. Kinetic analysis comprised measuring oxidative rates (nanogram atom of oxygen per minute per milligram of protein: ngatomolO × min−1 × mol−1 protein) at different respiratory substrate and inhibitor concentrations. We assembled Dixon plots for cocaine and cocaethylene by graphing reciprocal respiratory rate (1/V) against cocaine and cocaethylene concentrations (millimolar). Subsequently, we calculated inhibitory constants (Ki) based on intersection of fit curves for multiple acetylcarnitine and cocaine/cocaethylene concentrations as previously described.32, 33 Sample Size Calculations Sample size calculations were conducted based on previous experience and data for the same systems using bupivacaine as the local anesthetic including in vivo,28 ex vivo,21, 34 and mitochondrial experiments.27 For the in vivo experiments we assumed an effect size of 2, with α = 0.05 and power = 0.8 resulting in a total sample size of 12 (six in each group). Power calculations for isolated heart experiments were computed based on posttest criterion, assuming four posttests (10, 30, 50, and 100 μmol/L cocaine), splitting the alpha equally among them (α = 0.0125 per test), and checking for an effect at the highest concentration, assuming an effect size of three (30% difference in contractility with a maximal standard deviation of 10%). This yields a total sample size of 10 (five per group). For mitochondrial experiments we assumed complete inhibition (>60%) of carnitine‐stimulated pyruvate oxidation by cocaine and alpha split between the four posttest comparisons producing a sample size of three per group. Data Analysis We analyzed the toxicity data using survival analysis and quantified significance using the log‐rank (Mantel‐Cox) test. We conducted analysis of physiological data in Prism 6.0 (GraphPad, San Diego, CA) by two‐way analysis of variance (ANOVA), taking into account both drug dose (0 to 100 μmol/L) and treatment group (null vs. ILE, or null vs. SBE‐β‐CD) with parameters normalized to baseline. Cocaine and cocaethylene dose effects are reported as well as treatment effects, significance values from post hoc Sidak tests controlling for multiple comparisons for cocaine data and post hoc Tukey tests for multiple comparisons for cocaethylene and ethanol data. For mitochondrial data we used a Mann‐Whitney U‐test to compare baseline respiration levels with respiration levels after administration of accompanying drug (cocaine, cocaethylene, ecgonine, benzoylecgonine.)

Results Lipid Emulsion Pretreatment for Cocaine Toxicity in Intact Animals We previously found that pretreating rats with ILE increased the dose of IV bupivacaine (a lipophilic, cardiotoxic, local anesthetic) required to produce asystole.28 We studied whether this salutary effect could similarly attenuate the toxicity of cocaine. Blood pressure in controls (n = 6) dropped below 40 mm Hg after a median of 140 seconds (95% confidence interval [CI] = 103 to 240 sec) of cocaine infusion (p = 0.008), while the blood pressure of those pretreated with lipid emulsion (n = 6) reached 40 mm Hg after median 279 seconds (95% CI = 184 to 411 sec; Figure 1A). Saline‐treated controls developed asystole at 369 seconds (95% CI = 281 to 474) compared to 607 seconds (95% CI = 397 to 705; Figure 1B) among those pretreated with ILE (p = 0.0195). Thus, animals pretreated with lipid required roughly double the cocaine dose to reduce blood pressure to 40 mm Hg systolic and 50% more cocaine than controls to develop asystole. Figure 1 Open in figure viewerPowerPoint Lipid pretreatment increases time to <40 mm Hg and asystole in whole animal in vivo preparation. (A) Hypotension free progression, with a systolic blood pressure > 40 mm Hg following an infusion of 10 mg/kg/mL cocaine with a pretreatment of either 3 mL/kg/min of saline over 5 minutes (control) or 3 mL/kg/min of 20% lipid emulsion (Intralipid) over 5 minutes (lipid). Significantly prolonged median survival of 279 seconds in lipid‐treated versus 140 seconds in saline‐treated (p = 0.008, log‐rank test). (B) Asystole‐free progression. Significantly prolonged median survival to 607 seconds in lipid‐treated versus 369 seconds in saline‐treated (p = 0.0195, log‐rank test). Lipid Emulsion, Cocaine, and Cocaethylene in the Isolated Heart (Figure 2, Table 1) We studied the effects of cocaine in isolated rat hearts and further explored whether it exerted salutary effects on cocaine‐ or cocaethylene‐induced cardiac depression in a Langendorff, isovolumic, nonrecirculating, constant rate perfusion system. Cocaine induced dose‐dependent decreases in RPP (Figure 2A, p < 0.0001), dP/dtmax (Figure 2B, p < 0.0001), and –dP/dtmax absolute value (–dP/dtmax abs ; Figure 2C, p < 0.0001) and an increase in line pressure at a constant rate of coronary perfusion (Figure 2D, p = 0.004). Hearts cotreated with ILE exhibited less depression of cardiovascular parameters (RPP p = 0.01, dP/dtmax p = 0.0033, –dP/dtmax abs p = 0.0046) at cocaine concentrations of >10 μmol/L, but with no effect on line pressure. Cocaethylene also induced dose‐dependent decreases in RPP (Figure 2E, p < 0.0001) and dP/dtmax (Figure 2F, p < 0.0001), which were alleviated by cotreatment with ILE (p = 0.001, 0.0058 respectively). No differences were observed between ethanol‐treated controls and cocaethylene plus ILE groups. No interaction effects were noted in any condition. Figure 2 Open in figure viewerPowerPoint Lipid attenuates cocaine‐ and cocaethylene‐induced cardiovascular depression in the isolated heart. (A) Rate‐pressure product (RPP) following treatment with increasing concentrations of cocaine and concurrent 0.25% infusion of lipid emulsion (cocaine + 0.25% ILE) normalized to pre‐ILE baseline. Treatment with ILE improved cardiovascular parameters relative to cocaine alone (**p < 0.01, ***p < 0.001). (B) Maximal contraction force (dP/dt max, ***p < 0.001). (C) Maximal relaxation force (–dP/dt max abs , ***p < 0.001, ****p < 0.0001). (D) Line pressure. (E) RPP following infusions of different concentrations of cocaethylene (CE); ethanol control at 0.1, 0.3, and 0.5% final infusion (EtOH); or cocaethylene and concurrent 0.25% infusion of lipid emulsion (CE + 0.25% ILE) normalized to post‐ILE baseline. ILE improved cardiovascular parameters relative to CE alone (**p < 0.01 for CE vs. CE+0.25% ILE). EtOH was elevated relative to CE and indistinguishable from CE+0.25% ILE. (F) dP/dtmax during cocaethylene and lipid infusion (*p < 0.05, ***p < 0.001, ****p < 0.0001 for CE vs. CE+0.25% ILE). EtOH was elevated relative to CE and indistinguishable from CE+0.25%ILE. All values plotted as mean ± SEM. ILE = intravenous lipid emulsion. Table 1. Hemodynamic Parameters Following Cocaine and Cocaethylene With Treatment From ILE Cocaine Cocaethylene RPP dP/dt –dp/dt Line‐P RPP dP/dt 10 μmol/L 79 ± 5 85 ± 4 85 ± 5 107 ± 3 10 μmol/L 84 ± 5 90 ± 5 + ILE 90 ± 2 92 ± 3 93 ± 3 106 ± 5 0.1% EtOH 96 ± 2 a 99 ± 4 30 μmol/L 58 ± 2 67 ± 3 66 ± 3 130 ± 6 + ILE 101 ± 3 b 103 ± 4 a + ILE 78 ± 5 b 83 ± 5 c 85 ± 6 c 139 ± 17 30 μmol/L 69 ± 4 77 ± 3 50 μmol/L 51 ± 4 59 ± 3 59 ± 3 162 ± 12 0.3% EtOH 82 ± 2 a 94 ± 5 b + ILE 73 ± 4 c 80 ± 6 c 82 ± 7 d 170 ± 26 + ILE 94 ± 4 b 100 ± 5 c 100 μmol/L 35 ± 5 48 ± 4 47 ± 4 186 ± 20 50 μmol/L 59 ± 5 68 ± 4 + ILE 56 ± 3 c 64 ± 8 c 67 ± 8 c 198 ± 31 0.5% EtOH 79 ± 2 c 85 ± 1 c + ILE 84 ± 3 b 95 ± 5 d Cyclodextrin, Cocaine, and Cocaethylene in Isolated Heart (Figure 3, Table 2) SBE‐β‐CD alone evoked a mild cardiodepressant effect independent of cocaine treatment (p = 0.0001 compared to baseline) but attenuated further cocaine‐induced decreases in RPP (Figure 3A), dP/dtmax (Figure 3B), and –dP/dtmax abs (Figure 3C) at high cocaine concentrations. No significant effect was seen on line pressure (Figure 3D). At the highest cocaine concentration tested (100 μmol/L) SBE‐β‐CD alleviated the most pronounced cardiac depression for RPP, dP/dtmax, and –dP/dtmax abs . This differential effect of SBE‐β‐CD at low and high concentrations produced an interaction effect in the two‐way ANOVA for RPP (p < 0.0001), dP/dtmax (p = 0.0001), and –dP/dtmax abs (p = 0.0015) and prevented any overall treatment effect. Infusing SBE‐β‐CD also attenuated the cardiac depression associated with cocaethylene toxicity for RPP (Figure 3E) and dP/dtmax (Figure 3F). No differences were observed between ethanol‐treated controls and cocaethylene plus SBE‐β‐CD groups. Figure 3 Open in figure viewerPowerPoint Sulfobutylether‐β‐cyclodetrin (SBE‐β‐CD) induces mild cardiac depression but blocks further cardiac depression at high cocaine and cocaethylene concentrations. (A) Rate‐pressure product (RPP) following treatment with increasing concentrations of cocaine and concurrent SBE‐β‐CD (cocaine + SBE‐β‐CD) normalized to pre–SBE‐β‐CD baseline. SBE‐β‐CD diminished cardiovascular parameters at low cocaine concentrations but blocked additional cardiodepression at high cocaine concentrations (***p < 0.001). (B) Maximal contraction (dP/dtmax, **p < 0.01). (C) Maximal relaxation (–dP/dt max abs , *p < 0.05, **p < 0.01). (D) Line pressure. (E) RPP following infusions of different concentrations of cocaethylene (CE); ethanol control at 0.1, 0.3, and 0.5% final infusion (EtOH); and concurrent SBE‐β‐CD normalized to post–SBE‐β‐CD baseline (***p < 0.001 for CE vs. CE+SBE‐β‐CD). EtOH was elevated relative to CE and indistinguishable from CE+SBE‐β‐CD. (F) dP/dt max following cocaethylene and SBE‐β‐CD (**p < 0.01 for CE vs. CE+SBE‐β‐CD). EtOH was elevated relative to CE and indistinguishable from CE+SBE‐β‐CD. All values plotted as mean ± SEM. Table 2. Hemodynamic Parameters Following Cocaine and Cocaethylene With Treatment From Sulfobutylether‐β‐cyclodextrin Cocaine Cocaethylene RPP dP/dt –dP/dt Line‐P RPP dP/dt 10 mmol/L 80 ± 2 85 ± 1 78 ± 5 104 ± 1 10 mmol/L 82 ± 3 88 ± 2 + SBE‐β‐CD 67 ± 2 a 75 ± 2 b 70 ± 4 b 91 ± 2 0.1% EtOH 97 ± 1 b 100 ± 3 b 30 mmol/L 61 ± 4 71 ± 2 65 ± 7 115 ± 4 + SBE‐β‐CD 93 ± 4 94 ± 2 + SBE‐β‐CD 58 ± 2 71 ± 3 64 ± 4 96 ± 2 30 mmol/L 70 ± 3 73 ± 3 50 mmol/L 53 ± 3 68 ± 2 61 ± 5 116 ± 5 0.3% EtOH 82 ± 3 c 95 ± 5 a + SBE‐β‐CD 55 ± 2 70 ± 4 62 ± 4 102 ± 2 + SBE‐β‐CD 89 ± 5 a 91 ± 1 b 100 mmol/L 35 ± 3 51 ± 3 48 ± 2 120 ± 10 50 mmol/L 64 ± 4 73 ± 3 + SBE‐β‐CD 47 ± 2 a 62 ± 3 b 56 ± 3 c 109 ± 3 0.5% EtOH 78 ± 2 b 86 ± 1 b + SBE‐β‐CD 83 ± 4 a 88 ± 2 b Effect of cocaine and metabolites on mitrochondrial carnitine exchange Based on the alleviation of cocaine‐ and cocaethylene‐induced cardiac depression, we further tested if cocaine, cocaethylene, and the less cardiotoxic cocaine metabolites (benzoylecgonine and ecgonine) exhibited different effects on carbohydrate‐ versus lipid‐based mitochondrial respiration. Cocaine and cocaethylene (up to 1 mmol/L) did not inhibit respiration supported by the nonlipid substrate pyruvate (data not shown). However, when carnitine is added to pyruvate and malonate, both cocaine (1 mmol/L) and cocaethylene (1 mmol/L) inhibited respiration down to 20 (min:max, 12:21) ngatomO × min−1 × mg−1 (p < 0.05) and 9 (min:max, 6–9) ngatomO × min−1 × mg−1 (p < 0.05), respectively, from a baseline of 37 (min:max, 33:51) ngatomO × min−1 × mg−1; neither benzoylecgonine (1 mmol/L) nor ecgonine (1 mmol/L) inhibited CACT‐dependent respiration (Figure 4A). Using different concentrations of carnitine, we used the Dixon method32 to calculate the inhibitory constants (Ki) of both cocaine (Figure 4B) and cocaethylene (Figure 4C) for carnitine‐stimulated pyruvate oxidation—a measure of mitochondrial CACT activity and carnitine exchange. Both cocaine and cocaethylene had similar Ki (Figure 4D), with means of 253 (SEM ± 50 μmol/L) for cocaine (n = 4) and 237 (SEM ± 72 μmol/L) for cocaethylene (n = 4). Figure 4 Open in figure viewerPowerPoint Both cocaine and cocaethylene preferentially inhibit carnitine exchange in isolated mitochondria. (A) Carnitine‐acylcarnitine translocase (CACT)‐dependent respiration in rat, interfibrillar cardiac mitochondria at baseline, and in the presence of cocaine, cocaethylene, and the cocaine metabolites ecgonine and benzoylecgonine (all 1 mmol/L concentration). Both cocaine and cocaethylene significantly depress respiration (*p < 0.05). All data points plotted with median line. (B) Representative Dixon plot of cocaine‐based inhibition of carnitine‐dependent respiration. Cocaine produced a noncompetitive inhibition of carnitine‐dependent respiration. (C) Representative Dixon plot of cocaethylene based inhibition of carnitine dependent respiration. Cocaethylene produced a noncompetitive inhibition of carnitine dependent respiration. (D) Inhibitory constants (Ki) of CACT for both cocaine and cocaethylene. All data points plotted with median line.

Discussion Our results demonstrate that lipid emulsion is effective at relieving cardiac depression caused by either cocaine or cocaethylene. We further found that rescue of cardiac contractility by ILE is more effective than that provided by SBE‐β‐CD, a metabolically inert scavenging agent. This indicates the importance of cardiotonic21, 22 or direct effects17, 18 of the lipid on the heart during recovery. Like bupivacaine, both cocaine and cocaethylene potently inhibited mitochondrial carnitine exchange, while the less cardiotoxic metabolites35 ecgonine and benzoylecgonine did not. These parallels, shared with bupivacaine toxicity and the commonality of reversal by ILE, suggest that ILE‐based improvements in lipid metabolism during cocaine intoxication might underlie the cardiovascular recovery. Our results comport with prior clinical reports of the efficacy of ILE in combating cocaine cardiac toxicity12, 13 and with previous animal models14 that demonstrate that ILE can assist in recovery from toxicity. We have extended the results to cocaethylene, which is known to have significant detrimental effects on cardiovascular function,36, 37 and is found in a significant proportion of patients with cocaine‐related visits to the ED.1 Further, we identify CACT inhibition as a common mechanism of toxicity. Cocaine is a pharmacodynamically complex drug that binds and interacts with many different targets, including the serotonin‐norepinephrine‐dopamine transporter, in addition to the sodium, potassium, and calcium ion channels, leading to alterations in cardiac contractility, excitability, conduction, and repolarization.2, 11 The sodium channel–blocking capacity of cocaine underlies its historical use as the first local anesthetic. However, this beneficial effect of cocaine is offset by its potential as a severe cardiotoxin, first identified more than a century ago. Given that overdose caused by the toxic local anesthetic bupivacaine responds to ILE,15 it is reasonable to hypothesize that cocaine toxicity would respond similarly. Bupivacaine, like cocaine, exerts cardiotoxic effects at a variety of targets.38 The shared effect of bupivacaine, cocaine, and cocaethylene of inhibiting mitochondrial fatty acid metabolism points to a more general mechanism of local anesthetic induced cardiac toxicity. In particular, ATP supplementation to isolated myocardial strips can reverse bupivacaine‐induced contractile depression;39 this suggests that interference with mitochondrial energy generation is a key element of toxicity. Our results, coupled with prior work,27 could provide a common mechanism for the beneficial metabolic effect of lipid emulsion and a way to test what cardiac drug toxicities would derive benefit from lipid emulsion therapy. Scavenging of drug away from sites of toxicity15 is a putative mechanism underlying the efficacy of both ILE and SBE‐β‐CD. The drug delivery community is increasingly interested in developing improved drug uptake methods,40, 41 as sequestration by liposomes,42-44 microemulsions,45 macromolecular carrier systems,23 and other agents46, 47 all hold significant promise for the treatment of toxicity. However, in isolation, scavenging may not be enough. In whole animal models of bupivacaine toxicity, partitioning alone (the “lipid sink”) cannot account for the rapidity of recovery that ILE produces,48 and recent reports demonstrate that liposomes do not rescue cardiovascular parameters as effectively as lipid emulsions during acute toxicity.49 Recovery in these models requires an additional cardiotonic effect22 as observed in the absence of toxicity.21 In the case of overdose by drugs like cocaine and cocaethylene, toxicity is caused by altering a number of different targets,2, 50 and an effective treatment would need to address toxicity at many of these targets. A synergistic therapy including both indirect (e.g., drug‐scavenging) and direct effects (e.g., improvement in cellular energetics) would most likely be needed to achieve optimal recovery. While both the SBE‐β‐CD and the ILE can partition drug, only ILE contributes a cardiotonic effect.20 Interestingly, β‐cyclodextrins including SBE‐β‐CD and hydroxypropyl‐β‐cyclodextrin can improve recovery from toxicity from organophosphates,51 calcium channel blockers,52 and bacterial toxins,53 indicating the potential of scavenging agents. However, in more intelligently designed agents, molecular capture systems will likely complement agents that directly improve metabolism or physiologic performance in the same way that lipid emulsion provides benefit beyond just a partitioning effect.

Limitations In the whole animal experiments, our study is limited by the choice of pretreatment and potential clearance of lipid. We used a large dose (15 mL/kg total) based on prior experience in animal models14, 54, 55 and allometric scaling of human‐to‐rat doses.56 Assuming a half‐life of lipid (about 15 minutes), we expect a severely lipemic plasma, but the lack of continuous infusion could mean that lipid is cleared. Further, we restricted our model to pretreatment, while in clinical settings posttreatment is the required method of therapy. However, based on previous successful translation of bupivacaine pretreatment28 to clinical posttreatment,57 and the proposed mechanisms of action,20 we believe that evidence of efficacy in delaying cocaine toxicity indicates lipid will also be effective as a posttreatment for that indication. Finally, we did not blind the in vivo experiments, so experimenter effects may bias our results. Our isolated heart experiments are limited in terms of the choice of “pure capture” agent. SBE‐β‐CD interfered slightly with recovery at lower cocaine doses, possibly due to binding of extracellular calcium or uptake of cholesterol from cardiac membranes.58, 59 Moreover, SBE‐β‐CD might not be the most effective scavenging agent for cocaine or cocaethylene.60 Sugammadex, a γ‐cyclodextrin, was developed as a capture agent for rocuronim61 and is highly effective for that purpose. Presumably, a similar cocaine‐specific capture molecule could reverse cocaine‐induced cardiac toxicity with similar efficacy. Finally, we calculated our sample sizes to look for effects of lipid at the highest concentration of cocaine and cocaethylene. We might therefore miss beneficial effects or trends at the lowest cocaine and cocaethylene doses, but this is of less translational relevance as findings for highly toxic doses.

Conclusions We found that a commercially available lipid emulsion attenuates cocaine toxicity in the whole animal and improved acute cocaine‐ and cocaethylene‐induced cardiac toxicity in isolated rat heart. In contrast, a metabolically inert sulfobutyl‐β‐cyclodextrin (Captisol) was effective in reversing contractile depression in isolated heart only at the highest cocaine concentration. Cocaine and cocaethylene both impaired carnitine exchange, like bupivacaine, suggesting a shared metabolic target for causing cardiotoxicity. Collectively, these data suggest that effects other than drug scavenging contribute to lipid emulsion–induced reversal of cocaine cardiotoxicity. This work provides evidence in support of others,5, 12-14 that lipid emulsion could provide an effective antidote to cocaine overdose. Future studies should focus on the development of more targeted therapeutics and optimal treatment of this potentially fatal overdose.