Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Concurrent ingestion of cocaine and ethanol is an increasingly common presentation (Grant and Hartford, 1990). In a sample of cocaine-dependent patients, more than half of the subjects met criteria for current alcohol dependence, and in more than 50% of the occasions in which cocaine was consumed, both drugs had been used simultaneously (Higgins et al., 1994). In forensic studies, cocaine and ethanol are frequently identified in biological samples from fatally injured drivers (Budd et al., 1989; Marzuk et al., 1990).

Alcohol and cocaine interactions have been extensively investigated in humans. The drug combination produced greater increases in HR and blood pressure than those observed after the use of cocaine alone (Foltin and Fischman, 1989; Farré et al., 1993), reduced the feelings of drunkenness (Higgins et al., 1993), increased cocaine-induced euphoria (Perez-Reyes and Jeffcoat, 1992; Farré et al., 1993; McCance-Katz et al., 1993) and blunted some of the deleterious effects of alcohol on psychomotor performance tasks (Farré et al., 1993; Higgins et al., 1993).

The combined use of cocaine and alcohol increased cocaine and norcocaine plasma levels, reduced benzoylecgonine concentrations and induced the synthesis of cocaethylene. Cocaethylene is an active metabolite of cocaine formed in the presence of ethanol (Farré et al., 1993). The metabolic transformation of cocaine to benzoylecgonine is mediated in part by an hepatic carboxyltransferase, which also transforms cocaine into cocaethylene in the presence of alcohol (Brzezinski et al., 1994). Cocaethylene has been identified in forensic samples (Rafla and Epstein 1979; Jatlow et al., 1991) as well as in urine and blood samples from recreational users of alcohol and cocaine and from healthy volunteers taking part in experimental studies (De la Torre et al., 1991; Perez-Reyes and Jeffcoat, 1992; Farré et al., 1993; McCance-Katz et al., 1993). In humans, the i.v. administration of synthetic cocaethylene induced lower increases in HR and lower ratings of "high" than cocaine, reaching only 67% and 43% of the magnitude of the respective effects for cocaine (Perez-Reyes et al., 1994), whereas administration (0.95 mg/kg) by the intranasal route produced euphoria and cardiovascular effects similar to those caused by cocaine (0.92 mg/kg) (McCance et al., 1995).

With regard to neuroendocrine actions, the acute administration of alcohol to healthy subjects produced an increase in ACTH (Schuckit et al., 1988) and prolactin plasma levels (Schuckit et al., 1987a). In some studies, alcohol increased plasma cortisol levels, though others found no changes (Camí et al., 1988; Inder et al., 1995). The acute administration of cocaine induced an increase in ACTH and corticosterone levels in rats (Rivier and Vale, 1987; Borowsky and Kuhn, 1991; Levy et al., 1991; Torres and Rivier, 1992) and monkeys (Sarnyai et al., 1996). In humans, cocaine administration increased ACTH levels in cocaine-dependent subjects (Mendelson et al., 1992a; Teoh et al., 1994), and an increase in cortisol levels has been described in i.v. cocaine abusers (Baumann et al., 1995). Although a decrease in plasma prolactin concentrations has been documented in rhesus monkeys after the acute administration of cocaine (Mello et al., 1990a, 1990b, 1993), this inhibitory effect was not found in cocaine-dependent humans or i.v. cocaine abusers as compared with placebo (Mendelson et al., 1992a; Baumann et al., 1995). Hyperprolactinemia, however, has been described in cocaine-dependent subjects and explained in relation to a dopaminergic impairment (Mendelson et al., 1988; Mendelson et al., 1989).

This study was conducted to assess complementary aspects of a previous study on alcohol and cocaine interactions in healthy volunteers (Farré et al., 1993), in particular the effects of the combined use of cocaine and alcohol on neuroendocrine regulation, and to obtain further information on the metabolic pathway of cocaethylene. Subjective, cardiovascular and pharmacokinetic effects were also evaluated.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Subjects. Subjects were recruited "by word of mouth." Eligibility criteria required the recreational use of cocaine by the intranasal route on at least six occasions during the 3 months before participation in the study, a minimum daily alcohol consumption of 30 g and previous experiences in acute alcohol intoxication.

Study design. Subjects participated as outpatients in seven 8-hr experimental sessions in which the same doses and preparations of cocaine and alcohol were used. In each session, they drank a beverage containing alcohol or its placebo and snorted a powder containing cocaine or its placebo. Subjects arrived at the laboratory at 10:45 A.M. after an overnight fast and had an indwelling i.v. catheter inserted into a vein in the forearm of the nondominant arm. Thereafter, they remained seated in a quiet room throughout the session. Drug administration started around 12:00 A.M.. Three hours after drug administration, subjects had a light meal. Brief visits to the bathroom and tobacco smoking were permitted 3 hr after drug administration.

Drugs. Acute alcohol intoxication was induced by the ingestion of vodka (Stolichnaya red label) and tonic water containing a total alcohol dose of 0.8 g/kg. Several drops of aromatic bitters and lemon juice were added to mask the placebo drink, which contained tonic water only (Camí et al., 1988; Farré et al., 1993). The total volume of liquid was 450 ml. Subjects were given 30 min to consume the beverage, drinking 150 ml every 10 min.

Cardiovascular effects. Noninvasive HR and blood pressure were recorded at 30, 0 (immediately before beverage administration), 15, 28 (before powder snorting), 35, 40, 45, 50, 55, 60, 67, 75, 82, 90, 97, 105, 112, 120, 150, 180, 210, 270, 390 and 1470 min using a Dinamap 8100-T vital signs monitor (Critikon, Tampa, FL). For safety reasons, ECG, oral temperature and pulse oxymetry were continuously monitored during the first 3 hr of the session using a Dinamap Plus vital signs monitor (Critikon).

Subjective effects. Subjective effects were measured using the ARCI questionnaire and a set of 13 different visual analog scales. ARCI is a true-false questionnaire with empirically derived scales that are sensitive to the effects of a variety of classes of drugs of abuse (Haertzen, 1974). We used a short form of the inventory that consists of five scales with a total of 49 items (Martin et al., 1971). A Spanish validated version (Lamas et al., 1994) was administered. The five scales were PCAG (a measure of sedation), MBG (a measure of euphoria), LSD (a measure of dysphoria and somatic symptoms), BG (a stimulant scale consisting mainly of items related to intellectual efficiency and energy) and A (an empirically derived scale sensitive to the effects of d-amphetamine). ARCI was administered at 0 (immediately before beverage administration), 45, 90 and 390 min after drug administration.

Blood sampling. An indwelling i.v. catheter was inserted in a peripheral vein, and 0.9% sodium chloride solution was infused at a rate of 20 ml/hr. Blood samples (2 ml) were obtained for cortisol and prolactin analysis at 0, 40, 50, 60, 75, 90, 120 and 180 min after beverage administration. Samples were collected in chilled tubes and centrifuged immediately for 5 min at 3000 rpm at 4°C. The serum was removed and frozen at 20°C until analysis.

Hormone analysis. Plasma cortisol concentrations were determined by FPIA (Abbott Laboratories, Chicago, IL) according to the manufacturer's instructions. A good correlation between FPIA results and reference methods has previously been shown (Kobayashi et al., 1979). The intra-assay coefficients of variation (CV) were 2.92% (± 0.10), 7.78% (± 1.08) and 2.60% (± 0.96), for low (4.00 µg/dl), medium (15.00 µg/dl) and high (40.00 µg/dl) controls. Interassay CV were 17.43% (± 0.70), 10.31% (± 1.52) and 4.74% (± 1.81), respectively. The assay sensitivity is reported to be 0.45 µg/dl.

Drug and metabolite analyses. Blood alcohol levels were measured by gas-liquid chromatography (HP 5890 Hewlett-Packard, Palo Alto, CA) using a head-space injection technique (HP 19395A) and flame ionization detection. Separation was carried out using a cross-linked capillary column (RSL-160 Alltech, Deerfield, IL) 5 m long × 0.53 mm I.D. (film thickness 0.33 µm) (Farré et al., 1993). The method showed a good linearity (area ratio of the internal standard vs. concentration) from 50.6 to 1264.0 µg/ml (r = 0.9977, intercept = 0.053, slope = 0.0015). Intra-assay and inter-assay coefficients of variation for low concentrations were 7.8% and 10.3%, and for high concentrations were 0.9% and 3.3%, respectively.

Data analysis. Values from cardiovascular variables (including the rate-pressure product; Gobel et al., 1978), subjective variables and hormone concentrations were transformed to differences from base-line measures. The peak effect (maximum absolute change from base-line values) and the 3-hr AUC of effects vs. time calculated by the trapezoidal rule were determined for each variable except ARCI scores, which were evaluated only as peak effects. These transformations were analyzed by a one-factor repeated-measures ANOVA with drug condition as factor. When ANOVA results showed significant differences between treatment conditions, post-hoc multiple comparisons were performed using the Tukey test. Differences associated with P values lower than .05 were considered significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

HR and blood pressure. Cardiovascular effects after the administration of placebo, alcohol, cocaine and the drug combination are shown in table 1 and figure 1. Either alcohol or cocaine and the drug combination produced an increase in HR compared with placebo when the peak effects were considered, whereas for the AUC of pulse rate vs. time, only the drug combination increased the HR compared with placebo. The drug combination produced an increase in peak HR and in AUC of HR vs. time in comparison with either alcohol and cocaine. The peak difference between the conditions of alcohol and placebo was 17 bpm, between cocaine and placebo was 23 bpm and between placebo and the drug combination was 41 bpm. The difference between alcohol and the drug combination was 24 bpm, and that between cocaine and the drug combination was 18 bpm.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Time course (mean ± S.E.M., n = 8) of HR, systolic blood pressure and diastolic blood pressure after administration of placebo (), alcohol (), cocaine () and the combination of alcohol and cocaine (). The arrow signals cocaine administration.

Subjective effects. Alcohol increased peak PCAG score of the ARCI questionnaire as compared with placebo and with both conditions that included cocaine. The drug combination decreased peak PCAG score as compared with placebo and increased MBG score as compared with either placebo or alcohol. No differences between drug conditions were found in the LSD scores. Cocaine increased BG score as compared with alcohol. The drug combination increased BG and A scores as compared with all three other conditions, and cocaine increased A score as compared with either placebo or alcohol.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Time course (mean ± S.E.M., n = 8) of high and liking visual analog scales scores. Other details of the figure are similar to those for figure 1.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Peak effects (mean, n = 8) on ARCI scales after administration of placebo (, P), alcohol (, A), cocaine (, C) and the combination of alcohol and cocaine (, A/C). Filled symbols indicate a significant difference from placebo (P < .05). The letters a, b, and c indicate comparisons among the three active drug conditions; within the same panel, any two means designated with the same letter are not significantly different from each other at P < .05 (Tukey's post-hoc tests).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Peak effects (mean, n = 8) on visual analog scales. Other details of the figure are similar to those for figure 3.

Cortisol. Plasma cortisol concentrations were significantly higher (peak and AUC) after the administration of both conditions that included cocaine as compared with either placebo or alcohol. The combination condition produced a significantly higher increase in cortisol concentrations than did cocaine alone. Cortisol concentrations peaked 60 min after cocaine administration in the cocaine condition (90 min after alcohol administration), whereas in the combination condition, peak effects occurred 15 min earlier, 45 min after cocaine administration (75 min after alcohol ingestion). At 210 min after cocaine administration, cortisol concentrations were similar in the four treatment conditions. The mean peak difference in the plasma cortisol concentration between cocaine and placebo conditions was 10 µg/dl, and between cocaine and alcohol conditions was 9 µg/dl. The differences between the drug combination and placebo and between the drug combination and alcohol were 15 µg/dl and 14 µg/dl respectively. The difference between the drug combination and cocaine condition was 5 µg/dl. The effects of all treatment conditions on plasma cortisol levels are shown in table 1 and figure 5.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Time course (mean ± S.E.M., n = 8) of cortisol and prolactin plasma concentrations. Other details of the figure are similar to those for figure 1.

Prolactin. Prolactin plasma concentrations were increased after the administration of both conditions that included alcohol as compared with placebo (AUC) and cocaine (AUC and peak) conditions. No differences were found between alcohol and combination conditions. Placebo induced a slight decrease in plasma prolactin concentrations. After cocaine administration, plasma prolactin levels also decreased and remained below those achieved with placebo throughout the study period, although statistically significant differences were not reached. Prolactinemia peaked in both conditions that included alcohol 40 min after the beginning of alcohol drinking (10 min after beverage ending). The nadir concentrations were observed 90 min after cocaine administration and 120 min after placebo administration. Two hours after alcohol administration, prolactin concentrations were similar among all treatment conditions. The mean peak difference in prolactin plasma concentration between alcohol and placebo conditions was 6 ng/ml, and between alcohol and cocaine conditions was 8 ng/ml. The differences between the drug combination and placebo and between the drug combination and cocaine were 5 ng/ml and 8 ng/ml respectively. The peak difference between alcohol and drug combination conditions was negligible. The effects of all treatment conditions on plasma prolactine concentrations are shown in table 1 and figure 5.

Pharmacokinetic data. When the two conditions that included alcohol were compared, no significant differences were found in the pharmacokinetic parameters, except for C_max in the alcohol-alone condition (table 2).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Time course (mean ± S.E.M., n = 8) of cocaine and cocaethylene plasma concentrations. Symbols are used as follows: , cocaine concentrations after cocaine administration; , cocaine concentrations after cocaine and alcohol administration; , cocaethylene concentrations after cocaine and alcohol administration.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Time course (mean ± S.E.M., n = 8) of benzoylecgonine and ecgoninemethylester plasma concentrations. Symbols are used as follows: , benzoylecgonine concentrations after cocaine administration; , benzoylecgonine concentrations after cocaine and alcohol administration; , ecgoninemethylester concentrations after cocaine administration; , ecgoninemethylester concentrations after cocaine and alcohol administration.

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Time course (mean ± S.E.M., n = 8) of norcocaine and norcocaethylene plasma concentrations. Symbols are used as follows: , norcocaine concentrations after cocaine administration; , norcocaine concentrations after cocaine and alcohol administration; , norcocaethylene concentrations after cocaine and alcohol administration.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This study shows the increased pharmacological effects of cocaine and alcohol coadministration compared with the effects of both drugs given separately, a finding that confirms previous observations reported by our group (Farré et al., 1993) and by other authors (Perez-Reyes and Jeffcoat, 1992; Higgins et al., 1993; McCance-Katz et al., 1993). However, as far as we know, this is the first study in which the effects of both drugs on plasma concentrations of cortisol and prolactin have been examined in humans.

The alcohol dose (0.8 g/kg) used in this study produced a characteristic increase in the scores of drunkenness and sedation (ARCI-PCAG scale) as well as increases in HR and decreases in diastolic blood pressure as compared with placebo. These results are in agreement with the study of Higgins et al. (1993), who used a dose of alcohol of 1 g/kg. The intranasal administration of 100 mg of cocaine also reproduced the previously reported effects on subjective and cardiovascular variables (Resnick et al., 1976; Farré et al., 1993): increases in HR, diastolic blood pressure and subjective effects related to euphoria, well-being and central activation ("high," "good effects," "feeling good" and ARCI-A scale). Cocaine produced an increase in the scores on the MBG scale as compared with placebo but these differences were not statistically significant when intranasal cocaine was given. Only the i.v. administration of cocaine induced significant increases in scores on the MBG scale (Foltin and Fischman, 1991).

The combination of alcohol and cocaine produced a significant increase in the HR when compared with placebo, alcohol alone and cocaine alone. The increase in diastolic blood pressure was statistically significant when compared with placebo and with alcohol. These findings have been also observed in healthy volunteers by others (Foltin and Fischman, 1989; Perez-Reyes and Jeffcoat, 1992; Higgins et al., 1993; McCance-Katz et al., 1993). The elevations in HR and blood pressure with the drug combination result in increased myocardial oxygen consumption (measured by the rate-pressure product), which may be responsible for the observed increased risk of cardiovascular complications seen in those who use both cocaine and alcohol. However, in the study of Pirwitz et al. (1995), an increase in myocardial oxygen consumption was accompanied by coronary vasodilation when cocaine and i.v. ethanol were given concomitantly, whereas coronary vasoconstriction was observed after the use of intranasal cocaine alone.

The simultaneous use of alcohol and cocaine produced more marked subjective effects than alcohol or cocaine alone. Sedation ratings were lower than after alcohol administration (PCAG or "drowsy" scores) and were even lower than after placebo (PCAG scores). Drunkenness scores were reduced when cocaine was administered after alcohol, but these differences were not statistically significant. In our previous study, a similar result was obtained with 100 mg of cocaine and 1 g/kg of alcohol (Farré et al., 1993). In the study of Higgins et al. (1993), in which 96 mg of cocaine was given together with 1 g/kg of alcohol, alcohol-induced feelings of intoxication were significantly reduced by the stimulant drug. This difference may be explained by a lack of statistical power for this variable (1   = 0.69 for this comparison).

The drug combination was followed by higher ratings on the MBG scale, an objective measure of drug-induced euphoria (Martin et al., 1971), as compared with placebo and alcohol, but the difference between the two conditions that included cocaine was not statistically significant. The drug combination also raised the ratings of the BG and A scales in comparison with the other drug conditions. These effects have previously been described with the use of cocaine or other stimulant agents (Martin et al., 1971). Compared with cocaine alone, the drug combination also produced a generalized increase in self-reported subjective effects when AUC differences were considered, but not when peak differences were analyzed. This could be due to the longer duration of effects when alcohol and cocaine are given concurrently. Higher scores in all the visual analog scales except "worse performance," "hungry" and "drowsy" were found after the drug combination than after the administration of cocaine alone. These results are consistent with other reports (Higgins et al., 1993; McCance-Katz et al., 1993; Farré et al., 1993), but an enhancement in the scale "high" after cocaine and alcohol administration was found in one study only (Perez-Reyes and Jeffcoat, 1992). Taking into account the subjective profile we have cited, it seems that the drug combination produces more pleasurable effects, and this may explain why the drug combination is more likely to be abused than alcohol or cocaine alone.

With regard to the neuroendocrine effects of the testing conditions, one novel aspect of the study, the administration of ethanol (0.8 g/kg) was not followed by an increase in plasma cortisol concentrations. The effects of different doses of alcohol on the hypothalamic-pituitary-adrenal axis have been investigated extensively in experimental models. In rats, alcohol induced the release of CRF, ACTH and corticosterone. Alcohol perfusion on incubated hypothalami caused the release of CRF (Redei et al., 1988). Alcohol seems to have direct action on hypothalamic centers, but other regions must also be implicated in the response, because the alcohol-induced secretion of ACTH was partially attenuated, but not completely abolished, by lesions of the paraventricular nucleus or after inmunoneutralization of endogenous CRF (Rivest and Rivier, 1994). In humans, however, the effects of alcohol on the release of ACTH are conflicting: increases in ACTH levels have been found after the administration of alcohol at doses of 0.695 g/kg (Lukas and Mendelson, 1988) and 1.1 g/kg (Schuckit et al., 1988), and no increase at all has been found with doses between 0.75 and 1.3 g/kg (Ida et al., 1992; Waltman et al., 1993; Inder et al., 1995). On the other hand, elevations of cortisol concentrations in serum that seemed directly related to plasma levels of alcohol were described after the administration of 1.1 g/kg (Schuckit et al., 1987b; Lukas and Mendelson, 1988). Although a threshold of blood alcohol concentration of 100 mg/dl has been suggested to cause elevations of cortisol (Jenkins and Connolly, 1968), other studies have not been able to show an increase even after ethanol blood levels higher than this threshold (Ida et al., 1992; Waltman et al., 1993). Interestingly, the administration of alcohol attenuated naloxone-induced hypercortisolemia (Camí et al., 1988) and blunted the ovine corticotropin-releasing hormone-stimulated ACTH release in healthy volunteers (Waltman et al., 1993). It has been suggested that cortisol increases might appear only in subjects with gastrointestinal symptoms related to alcohol intolerance, such as nausea and vomiting (Inder et al., 1995). None of the subjects in our study had digestive complaints.

In healthy volunteers, the intranasal use of cocaine was followed by a marked increase in cortisol plasma levels. As far as we know, the results of cocaine in this population setting have not been previously reported. This finding is consistent with previous observations in animal models in which cocaine induced a release of CRF, ACTH, -endorphin and corticosterone (Rivier and Vale, 1987; Borowsky and Kuhn, 1991; Levy et al., 1991; Torres and Rivier, 1992; Sarnyai et al., 1996). The increase in ACTH secretion seems to be mediated by CRF release. It has been shown that cocaine-stimulated ACTH and corticosterone release are blunted by peripheral and i.c.v. administration of both CRF antiserum and a CRF receptor antagonist (Rivier and Vale, 1987; Sarnyai et al., 1992a; Sarnyai et al., 1992b). Moreover, bilateral lesions of paraventricular nucleus decreased the ability of cocaine to induce the release of ACTH (Rivier and Lee, 1994). The elevations of ACTH are probably dose-dependent (Borowsky and Kuhn, 1991; Torres and Rivier, 1992). It may be that this mechanism is mediated through effects on different sites, including dopaminergic (Borowsky and Kuhn, 1991), serotonergic (Borowsky and Kuhn, 1991; Levy et al., 1991) and N-methyl-D-aspartate (Torres et al., 1994) receptors, as well as by the local anesthetic properties of cocaine (Calogero et al., 1989). It is well known that cocaine binds specifically to uptake sites for dopamine, norepinephrine and serotonin. Acute administration of cocaine increases dopamine concentrations in synaptic clefts, mimicking the effects of dopamine and dopaminergic agonists (Gawin, 1991).

In cocaine-dependent subjects, the i.v. administration of 30 mg of cocaine produced an increase in ACTH plasma levels that was parallel to drug plasma concentrations and cocaine-induced cardiovascular and subjective effects (Mendelson et al., 1992a; 1992b). The administration of the partial opioid agonist buprenorphine suppressed the acute cocaine-induced stimulation of both ACTH and euphoria (Mendelson et al., 1992b). In i.v. cocaine abusers, the i.v. administration of 40 mg of cocaine produced an increase in cortisol plasma levels (Baumann et al., 1995). It has been postulated that cocaine increases ACTH pulse amplitude via stimulation of CRF release from the hypothalamus (Teoh et al., 1994). Some authors suggest that CRF could explain some of the behavioral actions of cocaine. In rats, cocaine-induced locomotor hyperactivity might be mediated, at least in part, through activation of the endogenous CRF system in the brain (Sarnyai et al., 1992a). Mendelson and associates (1988) found that during cocaine withdrawal, cortisol levels were within normal limits. In pulse frequency analysis of cortisol, no differences were found between cocaine users and age-matched controls (Mendelson et al., 1989). Accordingly, the increases in cortisol plasma levels found in our study may be attributable to the effects of cocaine on hypothalamic neurons releasing CRF, which, in turn, would stimulate the secretion of ACTH and cortisol.

Higher increases in cortisol plasma concentrations were found after the concurrent use of alcohol and cocaine than with cocaine alone. It is likely that this effect is related to the higher concentrations of cocaine and cocaethylene achieved with the drug combination. In experimental studies, cocaine-induced corticosterone increases are dose-dependent and are possibly related to cocaine plasma levels (Borowsky and Kuhn, 1991; Torres and Rivier, 1992). In humans, a concentration-effect relationship has been described for cardiovascular and subjective effects after the administration of i.v. cocaine, although the effects declined more rapidly than cocaine plasma concentrations. This result suggested the development of an acute tolerance phenomenon (Chow et al., 1985; Ambre et al., 1988; Foltin and Fischman, 1991). Although in our study, cortisol concentrations remained unchanged after the administration of alcohol, the presence of alcohol cannot be ruled out as a facilitating factor for the release of cortisol induced by cocaine.

Plasma concentrations of prolactin increased after the administration of alcohol, a pattern of response already observed by others after doses of alcohol between 0.75 and 1.3 g/kg (Schuckit et al., 1983, 1987a; Ida et al., 1992). The precise mechanism of this effect is poorly understood. An inhibition of dopamine secretion by the hypothalamus has been suggested (Ida et al., 1992). After alcohol exposure, numerous changes in the rate of turnover and in brain tissue levels of different neurotransmitters and neuromodulators (dopamine, ACh, norepinephrine, TRH, arginine vasopressin) have been described (Emanuele et al., 1992). Changes in dopamine, TRH or vasopressin levels may account for important variations in the plasma concentrations of prolactin. Increased basal prolactin levels and exaggerated responses to TRH have been described in alcoholics (Noth and Walter, 1984; Theoh et al., 1992).

It has been shown that dopamine reuptake in dopaminergic neurons is inhibited by cocaine. This results in a prolonged dopamine concentration at the neuronal synapses and in stimulation of dopamine receptors. After chronic administration, cocaine may cause a depletion of neuronal dopamine. A synaptic decrease of dopamine has been related to withdrawal symptoms associated with cocaine dependence (Gawin, 1991). In fact, drugs with dopamine agonist properties have successfully been used in the management of cocaine dependence (Gawin, 1991). Dopamine is the regulating factor of prolactin secretion by means of D₂ receptors. Substances that produce neuronal dopamine increases (e.g., dopamine agonists) inhibit prolactin release; by contrast, neuronal dopamine decreases are associated with hyperprolactinemia. Vasoactive intestinal peptides and TRH, both secreted by pituitary cells, are other important stimulants of prolactin secretion.

The acute administration of cocaine induces a decrease in prolactin concentrations both in experimental animals and in humans. In rhesus monkeys, either female in follicular phase or male, i.v. cocaine decreased prolactin levels (Mello et al., 1990a, 1990b; Mello et al., 1993). A rebound increase in prolactin concentration appeared 90 to 110 min after the administration of cocaine (Mello et al., 1990b). Chronic self-administration of cocaine in rhesus monkeys produced basal prolactin levels in the low normal range. The pattern of dopamine-induced supression of prolactin was similar to that described after acute administration in non-cocaine-dependent animals. However, postdopamine hyperprolactinemia showed an increase (Mello et al., 1994). In ovariectomized female rhesus monkeys, the administration of i.v. cocaine failed to decrease prolactin levels, a result that reflects the important role of gonadal steroids in anterior pituitary hormone secretion (Mello et al., 1995).

In this study, plasma prolactin concentrations were slightly lower as compared with placebo, although statistically significant differences were not found. Recently, Heesch et al. (1996), reported a significant decrease in prolactin concentrations in 12 healthy subjects after the intranasal administration of cocaine (2 mg/kg). The differences with placebo were observed at 60 to 120 min after cocaine administration. Several possibilities may be suggested to explain the discrepancy between their results and ours, such as dose of cocaine (100 mg vs. at least 140 mg [the volunteers' weight is not provided in their paper]), the number of subjects (8 vs. 12), which could decrease the variability and increase the statistical power, and the number of statistical comparisons (multiple in our investigation [four conditions, six comparisons using the conservative Tukey test], and a single comparison in their experiment). In cocaine-dependent subjects, the acute administration of cocaine (30 mg i.v.) was followed by a decrease in prolactin for 120 min as compared with base line (Mendelson et al., 1992a), although differences as compared with placebo were not statistically significant. Similar results were reported by Baumann et al. (1995) when 40 mg of cocaine was administered to i.v. cocaine abusers. Decreased, increased, and normal plasma prolactin concentrations have been reported after discontinuation of cocaine use (Gawin and Kleber, 1985; Mendelson et al., 1988; Swartz et al., 1990). Hyperprolactinemia seems to be a prognostic factor for relapse after cocaine withdrawal (Teoh et al., 1990; Weiss et al., 1994).

The effects of the alcohol and cocaine combination on prolactin regulation in humans have not been previously described. We found that alcohol-induced hyperprolactinemia was not altered by the drug combination. Although cocaine produced a slight decrease in plasma concentrations of prolactin when administered alone, it could not antagonize the effects of alcohol on prolactin secretion when alcohol and cocaine were given simultaneously.

The results of alcohol pharmacokinetics are in agreement with a previous study of our group (Farré et al., 1993). C_max was slightly lower in the combination condition, and there were no differences in T_max or AUC values. Other authors reported similar findings when both drug were administered simultaneously (McCance-Katz et al., 1993). The order of drug administration seems to be an important factor. The difference in peak blood levels of alcohol disappeared when the alcohol beverage was drunk 30 min after cocaine snorting (Perez-Reyes, 1994). Blood alcohol levels were marginally lower during the first hour in the kinetics of the drug combination condition. This differences could be related to changes in the alcohol absorption rate. The amount of alcohol diverted to cocaethylene synthesis does not seem to account for the differences in plasma levels of alcohol.

Cocaine and norcocaine blood levels were higher in the combination condition, as shown by the AUC values. Benzoylecgonine concentrations were lower after drug combination, with differences in C_max and AUC values. No differences were found in ecgoninemethylester pharmacokinetic parameters. Cocaethylene and norcocaethylene appeared only in the combination condition. The concentrations of cocaethylene increased slowly, with a maximal peak 2 hr after cocaine administration. As mentioned previously, cocaethylene is an active metabolite that produces euphoria and cardiovascular effects slightly less than or similar to those of cocaine (Perez-Reyes et al., 1994; McCance et al., 1995). Taking into account the ratio between cocaethylene and cocaine concentrations over the study period, we find that cocaethylene accounted for 8%, 12%, 18%, 24%, 36% and 39% of cocaine concentration at 45, 60, 90, 120, 180 and 240 minutes after drug administration. Because of the high sensitivity of the analytical methodology, it was possible to detect norcocaethylene in plasma after the consumption of cocaine doses compatible with the recreational use of alcohol and intranasal cocaine. This is the first description of norcocaethylene pharmacokinetics in humans. Other pharmacokinetic findings were consistent with previous reports (Perez-Reyes and Jeffcoat, 1992; Farré et al., 1993; McCance-Katz et al., 1993).

In terms of relative bioavailability, although the AUC for cocaine was greater in the combination condition, the AUC for benzoylecgonine (a major metabolite of cocaine) was higher in the cocaine condition. The sum of the AUCs of cocaine and its metabolites (benzoylecgonine, ecgoninemethylester, cocaethylene, norcocaine and norcocaethylene), as an index of absorption of the drug, was not different when cocaine alone (617,563 ng · min/ml [2183.8 nmol · min/ml]) and cocaine-alcohol conditions were compared (572,864 ng · min/ml [2019.4 nmol · min/ml]). These results of relative bioavailability agree with a previous study (Farré et al., 1993). Cocaine is metabolized to benzoylecgonine by spontaneous hydrolysis in plasma and especially by the action of an hepatic nonspecific cocaine carboxylesterase (Inaba et al., 1978; Inaba, 1989; Dean et al., 1991). In the presence of ethanol, this enzyme is responsible for the transesterification of cocaine, which forms the active metabolite cocaethylene (Brzezinski et al., 1994). Because the same enzyme regulates both metabolic pathways, an inhibition of metabolism (possibly a competitive inhibition) may explain our findings of increased cocaine concentrations, decreased benzoylecgonine levels and the appearance of cocaethylene when cocaine and alcohol were given simultaneously. This assumption may be supported by the observation of a decreased clearance of cocaine in the drug combination condition found in a previous study (Farré et al., 1993). As we have said, the order of drug administration could be a crucial factor in this interaction. Perez-Reyes and colleagues (1994) did not find differences in cocaine concentrations when alcohol was administered 30 min after cocaine snorting.

The higher plasma concentrations of cocaine in the drug combination condition could favor an increased availability of the substrate for the N-demethylation of cocaine, resulting in higher levels of norcocaine (Farré et al., 1993).

The time course of blood levels of norcocaethylene in experimental conditions in humans has not been described before. Cocaethylene is N-demethyled to norcocaethylene. The rate of metabolic biotransformation by oxidative N-demethylation seems greater for cocaethylene to norcocaethylene than for cocaine to norcocaine. Judging by the relation of AUCs between the substrates (cocaethylene or cocaine) and metabolites (norcocaethylene or norcocaine) as a possible index of oxidative metabolic rate, the N-demethylation of cocaethylene to norcocaethylene is 3-fold greater than that of cocaine to norcocaine.

When the overall effects of both conditions that include cocaine are considered, it seems that there is a relationship between cocaine levels and cardiovascular and subjective effects as well as plasma cortisol levels. Some of the effects reached their maximum around the peak of cocaine concentration, with similarities in effect-time and concentration-time curves. Statistically significant increases in HR, blood pressure, subjective euphoria-related effects and plasma cortisol found in the combination condition as compared with cocaine alone may be explained by pharmacodynamic and/or pharmacokinetic alcohol-cocaine interactions (e.g., a decrease in cocaine transformation to benzoylecgonine that may be due in part to a competitive inhibition of metabolism to form cocaethylene). The interaction resulted in a significant increase in plasma cocaine concentrations (20% higher in the combination condition), and the presence of cocaethylene, an active metabolite, that represented 20% of the cocaine concentrations. In our opinion, initially increased plasma cocaine levels, followed by the additive effect of cocaethylene, might explain the enhancement of cocaine effects in the combination condition.

In summary, the administration of alcohol and cocaine increased the cardiovascular, subjective and cortisol response to cocaine and decreased some of the subjective effects of alcohol intoxication. The administration of cocaine did not alter the effects of alcohol on prolactin secretion.