The Impact of Alcohol and Energy Drink Consumption on Intoxication and Risk-Taking Behavior


Reprint requests: Amy Peacock, PhD Candidate, School of Psychology, University of Tasmania, Private Bag 30, Hobart, Tas. 7001, Australia; Tel.: +61 3 6 226 7458; Fax: 61 3 6226 2883; E-mail:



It has been argued that consuming alcohol mixed with energy drinks (AmED) causes a subjective underestimation of intoxication and an increased level of risk-taking behavior. To date, however, there is mixed support for AmED-induced reductions in perceived intoxication, and no objective assessment of risk-taking following AmED consumption. Consequently, the present study aimed to determine the effect of alcohol and energy drink (ED) consumption on subjective measures of intoxication and objective measures of risk-taking.


Using a placebo-controlled, single-blind, cross-over design, participants (= 28) attended 4 sessions in which they were administered, in counterbalanced order: 0.5 g/kg alcohol, 3.57 ml/kg ED, AmED, and a placebo beverage. Participants completed the Biphasic Alcohol Effects Scale and a Subjective Effects Scale at baseline and 30 and 125 minutes postbeverage administration; risk-taking was measured using the Balloon Analogue Risk Task (BART).


Participants reported greater subjective intoxication, impairment, and sedation after active relative to placebo alcohol consumption, with no interactive AmED effects. However, a significant moderate magnitude increase in stimulation ratings was observed in the AmED relative to alcohol, ED, and placebo conditions. There was no independent effect of alcohol, or interactive effect with ED, on the BART. A significant, yet small magnitude, increase in risk-taking was evident in active relative to placebo ED conditions.


The interactive effect of AmED appears restricted to perceived stimulation, with alcohol-induced increases in subjective intoxication occurring regardless of presence or absence of ED. Engagement in risk-taking behavior was only increased by ED consumption; however, this effect was only of small magnitude; at these doses, alcohol consumption, with or without EDs, did not affect risk-taking. Further research assessing the dose-dependent effects of AmED on objectively measured risk-taking behavior could clarify whether the ED effect increases with higher doses and whether an interactive effect is observed with higher alcohol doses.

There is an increasing concern as to the impact of consuming alcohol mixed with energy drinks (AmED) on perceived and actual intoxication. Few studies have examined the effect of AmED relative to alcohol on subjective intoxication outcomes in a laboratory-based setting. Initial research revealed reduced ratings of motor coordination and dry mouth after AmED relative to alcohol consumption (Ferreira et al., 2006). However, ratings of other subjective outcomes (e.g., “tiredness,” “dizziness”) typically evident at the recorded peak breath alcohol concentrations (BrACs; 0.097 to 0.099 g/dl) did not differ significantly for AmED and alcohol conditions. Later research has also produced mixed findings, with several studies showing reduced ratings on select indices argued to index intoxication (i.e., “stimulation” and “mental fatigue”) post-AmED consumption (Marczinski et al., 2011, 2012), while others have revealed similar intoxication ratings across subjective measures for AmED and alcohol conditions (Alford et al., 2012).

Despite these disparate findings, several researchers (Arria and O'Brien, 2011; Weldy, 2010) have argued that AmED-induced underestimation of intoxication results in an increased likelihood of risk-taking behavior. The majority of research regarding AmED consumption and risk-taking has focused on comparison of alcohol and AmED consumers. AmED consumers report greater typical alcohol intake, maximum alcohol intake, number of days intoxicated, and number of heavy episodic drinking days relative to non-AmED consumers (Brache and Stockwell, 2011; O'Brien et al., 2008; Woolsey et al., 2010), and the odds of AmED use by hazardous drinkers is 4 times higher relative to nonhazardous drinkers (Berger et al., 2011). A field study in an American college bar district showed that bar patrons who had consumed AmED had a 3-fold increased risk of leaving an establishment highly intoxicated (BrAC ≥ 0.08 g/210 l) and a 4-fold increased risk of intending to drive while intoxicated compared with other bar patrons. Similarly, O'Brien and colleagues (2008) found that AmED users were generally more likely to report: (i) being taken advantage of sexually, (ii) taking advantage of someone sexually, (iii) driving while intoxicated, (iv) riding with a driver under the influence of alcohol, and (v) being hurt, injured, or required medical treatment.

Overall, these studies suggest greater risk-taking by AmED consumers. However, risk-taking behavior cannot be attributed to AmED's pharmacological effects as, with the exception of Thombs and colleagues' (2010) study, the outcomes reflect risk-taking across all alcohol-drinking sessions. In the case of Thombs and colleagues' (2010) study, the number of consumers who undertook the risk behavior is not known. Furthermore, few studies have controlled for systematic individual differences (e.g., risk-taking propensity, sensation-seeking) between consumer types, which could account for differences in risk-taking behavior.

Within-subject comparisons of risk-taking in AmED versus alcohol-drinking sessions circumvent these issues by controlling for individual differences between consumer types. However, the 2 studies published to date have shown mixed results. A study of American university athletes revealed that AmED users scored significantly higher on the Brief Comprehensive Effects of Alcohol questionnaire when reporting risk-taking expectations for AmED compared with alcohol-drinking sessions (Woolsey et al., 2010). However, retrospective report of actual risk-taking behavior in AmED and alcohol sessions was not undertaken. In contrast, a recent Australian community survey showed that AmED users reported significantly lower odds of engaging in 26 risk behaviors when consuming AmED relative to alcohol in the preceding 6 months (Peacock et al., 2012). To date, there has been no objective measurement of risk-taking following AmED ingestion. While laboratory-based risk-taking assessment may reduce ecological validity, the controlled environment permits direct measurement of AmED's pharmacological effects.

Given the divergent findings regarding AmED-induced intoxication misperception and lack of objective assessment of risk-taking outcomes following AmED consumption, the aims of the present study were to assess the effect of a moderate alcohol and energy drink (ED) dose on: (i) subjective measures of intoxication, and (ii) objective measures of risk-taking.

Materials and Methods


Twenty-eight healthy right-handed adults (14 males) aged between 18 and 25 years (= 19.5, SD = 1.8) participated in one 90-minute familiarization session and four 180-minute experimental sessions. The sample comprised self-reported regular caffeine (consumption of 5 to 28 caffeinated products per week) and ED (minimum consumption of 1 ED in the preceding month; maximum consumption of 1 ED per day in the preceding month) consumers. Exclusion criteria pertained to consumption of <2 standard alcoholic drinks in the preceding fortnight or an Alcohol Use Disorders Identification Test (AUDIT; Babor et al., 2001) score of 16 or higher. All participants recorded a body mass index between 18 and 30 and reported English as a first language, normal sleep patterns, normal or corrected-to-normal vision, and no history of substance abuse, neurological condition, or other serious physical condition. Exclusion was based on: (i) psychiatric diagnosis or Kessler Psychological Distress Scale (K10; Kessler et al., 2002) score of 30 or higher, and (ii) significant intellectual disability or Wechsler Test of Adult Reading (WTAR; Wechsler, 2001) quotient lower than 70. Additional self-report exclusion criteria concerned current regular tobacco or prescription medication (excluding the contraceptive pill) use, or illicit drug use in the preceding fortnight. No female participants self-reported being pregnant or currently lactating.

Recruitment occurred via university noticeboard advertisements; volunteers provided informed consent before participation. Ethics approval was granted by the Social Science Human Research Ethics Committee (Tasmania) Network. Participants received an honorarium of 30 AUD and task reimbursement (maximum 20 AUD) per experimental session.

Apparatus and Materials

Alcohol, Caffeine, and ED Intake Measures

The Timeline Follow-Back (Sobell and Sobell, 1992) questionnaire assessed alcohol consumption patterns. Participants provided retrospective daily standard alcoholic drink intake estimates for the preceding 30 days. Outcomes included: (i) total days consumed alcohol, (ii) total days alcohol consumption exceeded National Health and Medical Research Council's (NHMRC, 2009) lifetime low-risk guideline (i.e., 3 or more standard drinks per session), (iii) total days alcohol consumption exceeded NHMRC (2009) session low-risk guidelines (i.e., 5 or more standard drinks per session), (iii) average standard drinks per drinking day, and (iv) maximum standard drinks per drinking day.

A Caffeine and Energy Drink Use Questionnaire assessed average daily caffeine intake (mg); caffeine content of foods and beverages was based on the Australia New Zealand Food Standards (Food Standards Australia New Zealand, 2010) nutrient database or product packaging. ED consumption patterns in the preceding 30 days were determined by self-report of: (i) ED use frequency, (ii) typical ED intake per drinking day, and (iii) maximum ED intake per drinking day. ED estimates were expressed in standard sizes, where 1 standard drink was equivalent to 250 ml ED containing 80 mg caffeine.

Risk-Taking Measures

The Balloon Analogue Risk Task (BART; Lejuez et al., 2002) is an objective measure of sequential risk-taking operated via Inquisit Version software (Millisecond Software, Seattle, WA). Significant moderate positive correlations have been observed between the BART and self-reported real-world risk behaviors, including alcohol and substance use, cigarette use, gambling, aggressive and antisocial behavior, and sexual and driving risk-taking (= 0.25 to 48; Aklin et al., 2005 Lejuez et al., 2002, 2003a,b). In the present study, participants clicked on a pump to inflate a simulated balloon 1° and accrue 5 cents in a temporary bank. If the balloon was inflated beyond its predetermined break point (all accrued money was lost; if pumping was discontinued prior to the break point, the accrued money was added to a permanent bank. Participants completed 30 balloons, each with a different explosion probability based on a variable ratio schedule (average break point 64 pumps; for details of the algorithm, see Lejuez et al., 2002). Thus, each pump increased the accrued amount of money to be lost while decreasing the relative gain of additional pumps. Random selection of a trial number (1 to 30) at task cessation determined task reimbursement. The primary dependent risk-taking measure was the adjusted average number of balloon pumps (i.e., average number of pumps excluding those trials in which the participant was forced to stop pumping due to balloon explosion). Number of explosions and total earnings were also recorded.

The Risk-Taking Questionnaire-18 items (RT-18; de Haan et al., 2011) required participants to indicate the accuracy of 9 statements assessing risk behavior and 9 statements measuring risk assessment using a forced choice dichotomous response format (yes/no). Item score summation resulted in Risk Behaviour and Risk Assessment subscale scores (score range 0 to 9), with higher subscale scores indicative of greater behavioral risk-taking and less consideration of the consequences of risk-taking, respectively.

Subjective Intoxication Measures

The Biphasic Alcohol Effects Scale (BAES; Martin et al., 1993) assessed the subjective biphasic stimulation and sedation effects of alcohol. Participants rated the degree to which they were currently experiencing 7 stimulant (e.g., “energized”) and 7 sedation (e.g., “sluggish”) adjectives on an 11-point Likert scale (0 “not at all” to 10 “extremely”), with higher subscale scores (score range 0 to 70) indicating greater intensity of stimulation and sedation.

A Subjective Effects Scale (SES) assessed participants' perception of the treatment conditions' effects. Participants rated their level of intoxication, impairment, mental fatigue, and ability to drive on four 100-mm visual analog scales with left (0 mm) and right (100 mm) anchors designated “not at all” and “very much” (Beirness, 1987; Marczinski et al., 2011).

The Beverage Rating Scale (BRS; Fillmore and Vogel-Sprott, 2000) assessed participants' perceived alcohol and ED intake. Participants indicated the number of bottles of beer containing 4.8% alcohol (scale range 0 to 10) and number of standard 250 ml EDs containing 80 mg caffeine (scale range 0 to 3) consumed during beverage administration.

Treatment Conditions

Participants were randomly assigned a counterbalanced treatment administration order. While ED administration was double-blind, alcohol administration was single-blind. Doses were determined by body weight. The active alcohol conditions comprised 0.50 g/kg vodka (37.5% alcohol/volume Smirnoff Red Label®, No. 21; Smirnoff Co., Norwalk, CT), reduced to 85% for female participants (Baraona et al., 2001), with an intended peak BrAC of 0.050%, the Australian legal limit for drink driving. The active ED dose was 3.57 ml/kg Red Bull® (Red Bull GmbH, Fuschl am See, Austria), equivalent to 1 standard 250 ml ED per 70 kg person. The placebo alcohol condition was 5 ml vodka floated on each portion, with a light alcohol mist sprayed on the inner container (Marczinski and Fillmore, 2006). The placebo ED dose was 3.57 ml/kg Red Bull® minus caffeine, taurine, glucuronolactone, inositol, and B vitamin complex; active and placebo ED beverages were matched for sugar content (i.e., 27 g/250 ml). Vodka and Red Bull® were administered as a recent Australian survey study demonstrated that these are the most commonly used AmED constituents (Peacock et al., 2012).


Familiarization Session

Following initial eligibility confirmation, participants attended the familiarization session where they provided informed consent, completed additional screening assessments and sample characteristic measures, and practiced the BART.

Experimental Sessions

Experimental sessions were conducted between 09:30 am and 07:00 pm and separated by a minimum of 2 and maximum of 10 days. With the exception of consuming a standard breakfast bar 90 minutes prior to session commencement, participants fasted for 4 hours and abstained from caffeine for 8 hours, alcohol and prescription medication for 24 hours, and illicit drugs for the duration of participation. Participants signed a declaration confirming compliance, and a 0 BrAC was verified using an Alcolizer HH-2 (Alcolizer Pty Ltd., Brisbane, QLD, Australia) prior to session commencement.

After completing baseline BAES and SES measures, participants were administered the beverage in 2 portions served in opaque lidded cups, consuming each portion at an even pace within a 5-minute period. The BAES and SES were readministered 30 minutes after commencing beverage consumption, and the BART was commenced at 40 minutes; BrAC was recorded at these time points and at the conclusion of the BART (55 minutes). As 3 additional behavioral tasks were administered, subjective intoxication and BrAC were reassessed 125 minutes after beverage administration.

At each session's completion, participants received a detoxification meal and remained at leisure until recording 2 consecutive BrAC measurements of 0.03% or less over 15 minutes; debriefing occurred on conclusion of participation.

Data Analysis

Data were analyzed blind in IBM SPSS Statistics 19 (IBM Inc., Armonk, NY). Due to technical malfunction during electronic survey administration, 2 participants had missing data for the BAES and SES (= 26), and 1 participant had missing data for the BRS (= 27). Sample characteristics, objective risk-taking outcomes (BART adjusted average number of pumps, number of explosions, total earnings), and BRS ratings were analyzed using 2 (Alcohol: Active, Placebo) × 2 (ED: Active, Placebo) repeated measures analyses of variance (ANOVAs). Identical analyses were conducted for BAES and SES outcomes, with the dependent variables in these cases being change from baseline scores calculated for each time point (30- and 125-minute postbeverage administration). An additional variable, Sex, was included in all analyses. Alpha levels were maintained at < 0.050, with Bonferroni adjustments for follow-up paired and independent sample t-tests where necessary. Effect size was calculated using Hedges' g (Tamamoto et al., 2010). Pearson's product moment coefficients were calculated to determine the relationship between trait and objective risk-taking measures.


Demographic Characteristics and Self-Reported Alcohol, Caffeine, and ED Use

The demographic characteristics and self-reported alcohol, caffeine, and ED intake outcomes are displayed in Table 1. While the mean AUDIT score matched the cutoff score indicative of hazardous and harmful alcohol use (Babor et al., 2001), participants generally displayed above-average intelligence, low psychological distress, and a normal body mass index (World Health Organization, 2006). Participants consumed alcohol on a twice-weekly basis in the preceding month, typically ingesting 5 standard alcoholic drinks (9 at maximum), and exceeding the NHMRC (2009) lifetime and session low-risk guidelines on a weekly and fortnightly basis, respectively. The sample generally comprised moderate caffeine consumers. One-quarter (29%) ingested EDs on a monthly or less basis and one-third (32%) on a fortnightly to weekly basis; more frequent use was reported by two-fifths (39%) of the sample. Typical ED intake fell within the Australia New Zealand Food Authority (2009) recommendations; however, these daily intake guidelines were exceeded in maximum ED-drinking sessions.

Table 1. Demographic Characteristics and Self-Reported Alcohol Use, Caffeine, and ED Use (Standard Deviation in Parentheses; = 28)
Sample characteristicMeanRange
  1. Alcohol Use Disorders Identification Test (AUDIT) score range is 0 to 40, with a score of 16 or more indicative of hazardous or harmful alcohol use; Kessler Psychological Distress Scale (K10) score range is 10 to 50, with scores of 30 or higher indicative of a moderate to severe mental illness; Wechsler Test of Adult Reading (WTAR) standardized score is 100, with higher scores indicative of higher levels of general premorbid intellectual functioning; body mass index indicates a greater body mass, with scores between 17 and 29.9 indicating mild-thinness to pre-obese body mass; the Timeline Follow-Back reflects participants' alcohol consumption in the preceding month; National Health and Medical Research Council (NHMRC) lifetime low-risk guideline is a maximum of 2 standard alcoholic drinks on any day; NHMRC session low-risk guideline is a maximum of 4 standard alcoholic drinks on any day; the Caffeine Energy Drink Use Questionnaire definition of a standard energy drink (ED) was 250 ml ED containing approximately 80 mg caffeine.

AUDIT8.1 (3.0)3.0 to 14.0
K1015.8 (3.3)12.0 to 26.0
WTAR106.4 (10.3)87 to 126
Body mass index23.6 (3.0)18.3 to 30.0
Timeline Follow-Back (past month)
Days any alcohol7.5 (5.2)2.0 to 23.0
Days exceed NHMRC lifetime low-risk guideline4.4 (2.6)0.0 to 10.0
Days exceed NHMRC session low-risk guideline2.7 (2.3)0.0 to 9.0
Average standard alcoholic drinks per drinking day5.2 (3.2)1.3 to 14.9
Maximum standard alcoholic drinks per drinking day9.6 (5.1)1.9 to 22.0
Caffeine Energy Drink Use Questionnaire
Average daily caffeine intake (mg)236.1 (130.8)70.4 to 556.7
Average standard EDs1.3 (0.6)1.0 to 3.0
Maximum standard ED2.4 (1.3)1.0 to 6.0

Breath Alcohol Concentration

As no detectable BrACs were recorded in placebo alcohol conditions, analyses comprised a 2 (Condition: Alcohol vs. AmED) × 4 (Time: 30, 40, 55, 125 minutes) repeated measures ANOVA. The main effect of Time was significant, F(3, 81) = 62.674, < 0.001, with BrAC descending throughout the session (Table 2). No significant Condition main effect or Condition × Time interaction was observed (ps > 0.631). There was no significant difference in BrAC by Sex (ps > 0.258).

Table 2. Breath Alcohol Concentration, BART Adjusted Average Number of Pumps, Total Earnings, Number of Explosions, and Beverage Rating Scale Outcomes According to Treatment Condition (Breath Alcohol Concentration and BART n = 28; Beverage Rating Scale n = 27)
  1. BART, Balloon Analogue Risk Task; ED, energy drink; AmED, alcohol mixed with energy drink.

  2. No detectable breath alcohol concentrations were recorded in placebo and ED conditions. The Beverage Rating Scale range for standard alcoholic drinks and standard EDs was 0 to 10 and 0 to 3, respectively; a standard alcoholic drink contained 10 mg alcohol, and a standard ED was 250 ml ED containing approximately 80 mg caffeine.

Breath alcohol concentration
30 minutes (subjective intoxication measures)    0.0680.0190.0670.018
40 minutes (BART commencement)    0.0620.0140.0640.014
55 minutes (BART conclusion)    0.0580.0110.0600.007
 125 minutes (subjective intoxication measures)    0.0390.0090.0400.007
Balloon Analogue Risk Task
 Total earnings35.37.738.79.338.
 Number of explosions10.
Beverage Rating Scale
 Number of standard alcoholic drinks0.
 Number of standard energy drinks1.

Risk-Taking Outcomes

Balloon Analogue Risk Task

The adjusted average number of pumps for each treatment condition is displayed in Fig. 1. While there was no significant main effect of Alcohol (= 0.921, = 0.01), there was a significant main effect of ED, F(1, 27) = 4.335, = 0.047, = 0.28, revealing a small magnitude increase in the adjusted average number of pumps in active (= 44.5, SD = 40.3) relative to placebo (= 40.3, SD = 12.8) ED conditions. There were no interactive effects of alcohol and ED, as evidenced by a nonsignificant Alcohol × ED interaction (= 0.387).

Figure 1.

Mean Balloon Analogue Risk Task (BART) adjusted average number of pumps for each treatment condition (n = 28). Errors bars depict the standard deviation.

Similarly, the main effect of Alcohol and the Alcohol × ED interaction were not significant for total earnings and number of explosions (ps > 0.117; Table 2), nor was there any significant main effect of ED for these variables (ps > 0.364, gs > 0.17). There was no significant difference in the adjusted average number of pumps, total earnings, or number of explosions according to Sex (ps > 0.163).

Risk-Taking Questionnaire-18 Items

Table 3 displays participants' mean RT-18 subscale scores and the correlation between the RT-18 and BART adjusted average number of pumps. The correlations indicate negligible to weak associations between the RT-18 subscale scores and BART adjusted average number of pumps, with no consistent pattern across treatment conditions.

Table 3. Mean RT-18 Risk Behavior and Risk Assessment Subscale Scores, and Correlations with Adjusted Average Number of Pumps According to Treatment Condition (Standard Deviation in Parentheses; n = 28)
RT-18 SubscaleM (SD)Adjusted average number of pumps
  1. ED, energy drink; AmED, alcohol mixed with energy drink; RT-18, Risk-Taking Questionnaire-18 items.

  2. RT-18 subscale score range is 0 to 9, with higher scores indicative of greater risk behavior and risk assessment, respectively, on a continuum basis.

Risk behavior4.4 (2.2)0.0880.1390.053−0.197
Risk assessment3.3 (2.7)−0.2010.214−0.0680.211

Subjective Intoxication Measures

Biphasic Alcohol Effects Scale

Table 4 shows the mean BAES stimulation and sedation subscale change scores according to treatment condition. There was a significant main effect of Alcohol on stimulation ratings 30-minute postbeverage consumption, F(1, 25) = 6.303, = 0.019, = 0.47, with a moderate magnitude increase in stimulation ratings in active (= 7.4, SD = 12.5) relative to placebo (= 2.1, SD = 9.5) alcohol conditions; no significant main effect of Alcohol was evident at 125 minutes (= 0.392, = 0.17). However, a significant interaction of Alcohol and ED at 30 minutes was also observed, F(1, 25) = 8.447, = 0.008; follow-up comparisons revealed that stimulation ratings were significantly higher in the AmED condition relative to the alcohol condition (= 0.007, = 0.51), as well as the ED (< 0.001, = 0.84) and placebo (= 0.008, = 0.68) conditions. There was no significant Alcohol × ED interaction at 125 minutes (= 0.850). The main effect of ED was not significant for stimulation ratings at 30 (= 0.075, = 0.30) or 125 (= 0.105, = 0.26) minutes.

Table 4. Treatment Condition Baseline Ratings and Change from Baseline Ratings at 30- and 125-Minute Postbeverage Administration for BAES Stimulation and Sedation Subscales and SES Item (Standard Deviation in Parentheses; n = 26)
 Baseline30 Minutes125 Minutes
  1. ED, energy drink; AmED, alcohol mixed with energy drink; BAES, Biphasic Alcohol Effects Scale; SES, Subjective Effects Scale.

  2. BAES subscale scores range from 0 to 70, with higher scores indicating greater stimulation/sedation. SES item scores range from 0 to 100 with higher scores indicating greater intensity of intoxication, impairment, and mental fatigue/reduced ability to drive.

BAES Stimulation11.8 (11.6)9.5 (11.2)10.4 (10.0)10.0 (8.8)2.5 (12.3)1.8 (8.8)3.9 (15.1)10.9 (12.6)−1.0 (15.1)1.4 (10.7)−3.3 (14.1)0.0 (12.9)
BAES Sedation19.9 (12.6)23.4 (14.9)20.9 (12.7)21.0 (10.8)−2.3 (9.1)−0.6 (0.2)1.8 (11.8)4.3 (10.5)2.5 (12.0)−0.7 (8.6)6.5 (12.5)5.2 (10.9)
SES Intoxication1.3 (6.3)4.9 (15.5)4.1 (13.8)2.8 (9.9)1.4 (6.9)0.9 (19.9)45.3 (31.5)48.4 (30.0)−0.4 (6.9)−3.9 (16.0)18.5 (29.5)23.2 (29.9)
SES Impairment4.2 (12.4)4.4 (14.4)5.5 (14.4)6.4 (16.7)−0.4 (10.0)1.7 (18.3)33.9 (32.5)37.5 (32.4)2.9 (11.4)0.7 (18.3)20.0 (27.2)18.9 (26.1)
SES Mental fatigue13.2 (20.9)9.9 (14.0)12.5 (19.6)12.3 (20.0)0.9 (15.1)0.3 (16.4)6.8 (20.3)6.6 (12.5)12.8 (23.9)8.6 (22.1)18.0 (25.5)16.6 (22.6)
SES Ability to drive86.9 (24.7)70.9 (41.1)79.7 (34.4)75.2 (36.8)−13.2 (19.4)−6.2 (35.4)−55.8 (37.3)−50.3 (36.4)−14.0 (22.2)−5.0 (37.0)−43.6 (32.9)−43.6 (33.1)

While there was a trend toward a significant main effect of Alcohol on sedation ratings 30-minute postbeverage ingestion (= 0.057, = 0.52), with higher ratings in the active relative to placebo alcohol conditions, a significant main effect of Alcohol was observed at 125 minutes, F(1, 25) = 4.877, = 0.037, = 0.57, with moderate magnitude increase in sedation ratings in active (= 5.9, SD = 8.7) relative to placebo (= 0.9, SD = 8.7) alcohol conditions. The Alcohol × ED interaction was not significant at either time points (ps > 0.587), nor was the ED main effect (ps > 0.137, gs < 0.28). There was no significant differences according to Sex for BAES stimulation (ps > 0.092) or sedation ratings (ps > 0.093).

Subjective Effects Scale

The SES ratings according to treatment condition are displayed in Table 4. Intoxication ratings. There was a significant main effect of Alcohol on intoxication ratings at 30 minutes, F(1, 25) = 85.950, < 0.001, = 2.20, and 125 minutes, F(1, 25) = 20.932, < 0.001, = 1.10, following beverage administration, with high magnitude increases in intoxication ratings recorded in active (= 46.8, SD = 27.9 and = 20.9, SD = 27.7) relative to placebo (= 1.1, SD = 11.1 and = −2.2, SD = 8.5) alcohol conditions. However, the Alcohol × ED interaction was not significant at either time point (ps > 0.156), nor was the main effect of ED (ps > 0.689, gs < 0.07). While analyses according to Sex revealed a significant Alcohol × Sex interaction at the latter time point, F(1, 24) = 4.280, = 0.049, no follow-up tests were significant (ps > 0.148); there were no other significant differences according to Sex (ps > 0.169).

Impairment Ratings

Similarly, there was a significant main effect of Alcohol on impairment ratings at 30, F(1, 25) = 47.688, < 0.001, = 1.72, and 125 minutes, F(1, 25) = 18.315, < 0.001, = 0.97, following beverage administration, with large magnitude increases in impairment ratings in active (= 35.7, SD = 27.3 and = 19.4, SD = 22.9) relative to placebo (= 0.6, SD = 8.8 and = 1.8, SD = 11.7) alcohol conditions. The Alcohol × ED interaction was not significant at either time point (ps > 0.865), nor was the main effect of ED significant (ps > 0.434, gs < 0.16). While analyses according to Sex revealed a trend toward significant Alcohol × Sex interaction at 125 minutes, F(1, 24) = 3.679, = 0.067, no follow-up tests were significant (ps > 0.240); nor were any other interactions involving Sex significant (ps > 0.265).

Mental Fatigue Ratings

There was no significant main effect of Alcohol on mental fatigue ratings at 30 (= 0.062, = 0.53) or 125 (= 0.125, = 0.39) minutes after beverage consumption. The interaction between Alcohol and ED was not significant at either time point (ps > 0.761), nor was the main effect of ED significant (ps > 0.544, gs < 0.16). There was no significant difference in mental fatigue ratings according to Sex (ps > 0.273).

Ability to Drive Ratings

There was a significant main effect of Alcohol on ratings of ability to drive at 30, F(1, 25) = 50.525, < 0.001, = 1.54, and 125 minutes, F(1, 25) = 31.637, < 0.001, = 1.40, with strong magnitude decreases in ability to drive in active (= −53.0, SD = 32.1 and = −43.6 and SD = 26.0) relative to placebo (= −9.7, SD = 23.1 and = −9.5, SD = 22.5) alcohol conditions. However, the Alcohol × ED interaction (ps > 0.439) and the main effect of ED (ps > 0.221, gs < 0.23) were not significant across testing points. There was no significant difference in ratings of ability to drive according to Sex (ps > 0.196).

Beverage Rating Scale

Mean alcohol and ED beverage ratings are displayed in Table 2. There was a significant main effect of Alcohol for perceived alcohol intake, F(1, 26) = 152.164, < 0.001, = 3.13, with greater alcohol intake reported in active (= 2.9, SD = 1.0) relative to placebo (= 0.50, SD = 0.49) alcohol conditions. However, the Alcohol × ED interaction was not significant (= 0.532), nor was the main effect of ED (= 0.054, = 0.33). Analyses according to Sex revealed a trend toward a significant Alcohol × Sex interaction, F(1, 25) = 3.740, = 0.065; follow-up tests revealed that perceived active alcohol intake in alcohol conditions tended to be greater for males (= 3.2, SD = 1.0) compared with females (= 2.5, SD = 0.8; = 0.065, = 0.77); however, there were no sex differences in relation to perceived alcohol intake in placebo alcohol conditions (= 0.706). There were no other significant effects involving Sex.

There was no significant main effect of Alcohol (= 0.946, = 0.02), Alcohol × ED interaction (= 0.958), or main effect of ED (= 0.142, g = 0.25) on perceived ED intake. There was no significant difference in perceived ED intake according to Sex (ps > 0.457).


The results of the present study revealed an interactive effect of alcohol and EDs on perceived stimulation, with greater stimulation ratings in the AmED relative to the alcohol condition at 30-minute postbeverage administration. However, no interactive effects of alcohol and ED were observed for perceived sedation, impairment, mental fatigue, ability to drive and, most importantly, intoxication; with the exception of mental fatigue ratings, treatment effects were restricted to the independent effects of alcohol. Despite the escalation in perceived alcohol-induced impairment, a moderate alcohol dose (mean BrAC 0.062%) did not alter risk-taking behavior, nor did the interaction of alcohol and ED (mean BrAC 0.064%). While there was a significant increase in risk-taking evident in active relative to placebo ED conditions, the magnitude of difference was small.

The current results align with previous research involving moderate alcohol doses (peak mean blood alcohol concentration 0.071 to 0.089%) regarding the absence of an interactive alcohol and ED effect on perceived sedation, intoxication, impairment, and ability to drive (Marczinski et al., 2011, 2012). In regard to perceived stimulation, the authors of these studies reported an interactive effect of alcohol and EDs, with greater stimulation ratings observed in AmED conditions. In the former study by Marczinski and colleagues (2011), the reported interactive AmED effect was based on examination of descriptive data rather than direct statistical comparison of ratings in AmED and alcohol conditions. In the latter study, ratings were only provided postbeverage administration, and thus, baseline differences between treatment conditions in subjective state were not controlled. While conclusions regarding AmED-induced enhancement of stimulation have been challenged on the basis of these limitations (Peacock and Bruno, in press), the present study aligned with the interpretation of Marczinski and colleagues (2011, 2012) in that a significant moderate magnitude increase in perceived stimulation was evident 30 minutes after consumption of AmED relative to alcohol. These results also support those of Attwood and colleagues (2012), who observed greater stimulation ratings following co-ingestion of caffeine (2.0 mg/kg) with alcohol (0.6 g/kg) relative to independent alcohol ingestion, with no significant difference in ratings of intoxication. As such, Attwood and colleagues (2012) concluded that caffeine may change the nature, as opposed to the degree of intoxication; this same conclusion could be tentatively applied to the present subjective outcomes. As the stimulant effects of alcohol are a major predictor of subsequent alcohol intake (Corbin et al., 2008), ED enhancement of alcohol-induced stimulation could heighten the reinforcing effects of alcohol and increase alcohol intake. However, previous survey research has yielded mixed support; while Australian and Canadian studies have shown significantly greater alcohol intake in AmED sessions relative to alcohol sessions (Peacock et al., , in press; Price et al., 2010), an American study revealed the converse (Woolsey et al., 2010).

Marczinski and colleagues (in press) found that ED (1.82 ml/kg) co-ingested with alcohol (0.91 ml/kg) increased subjective ratings of “desire more alcohol” across more time points postdrink than alcohol only, suggesting that EDs may increase alcohol priming. However, between-condition analyses revealed no significant difference in ratings between alcohol and AmED conditions (Peacock and Bruno, in press). As such, conclusions regarding enhanced reinforcement by AmED remain tentative until laboratory studies are undertaken examining: (i) the dose-dependent effect of “real-life” doses on subjective intoxication indices, and (ii) the effect of AmED-increased stimulation enhancement on subsequent alcohol intake (Peacock and Bruno, in press), particularly in light of the moderate alcohol dose administered in the present study. While previous research has generally involved administration of a set dose (approximately 1 standard 250 ml ED per 70 kg person), research by Peacock and colleagues (in press) revealed that AmED consumers were typically ingesting 7.1 standard alcoholic drinks and 2.4 standard EDs in AmED-drinking session. As such, studies seeking to provide policy advice should extend further into these high-dosage domains for ecological validity, and because the propensity for risky behaviors is inflated at such high alcohol consumption levels (NHMRC, 2009), as is the potential for more complex interactive effects of ED.

In light of the equivalent perception of intoxication and impairment in AmED and alcohol sessions, it is not surprising that objective measurement of risk-taking revealed no interactive effect. These results contradict self-reported risk-taking behavior; Woolsey and colleagues (2010) observed an increased expectation of risk-taking for AmED relative to alcohol-drinking sessions, whereas Peacock and colleagues (2012) found lower odds of risk-taking in AmED relative to alcohol-drinking sessions. However, in the present study alcohol consumption did not alter risk-taking, regardless of the presence or absence of ED. Experimental research assessing the impact of alcohol on risk-taking has yielded equivocal results, with some studies revealing increased risk-taking (Lane et al., 2004; Liguori et al., 1999), while others have shown no significant effect (Breslin et al., 1999; George et al., 2005). The present results could be attributed to the moderate dose (0.5 g/kg) administered. For example, Lane and colleagues (2004) reported that alcohol (0.2, 0.4, and 0.8 g/kg) dose-dependently increased selection of the risky response option in a gambling task, with the highest dose increasing the probability of consecutive losing risky responses following a win on a risky response. Thus, the administered alcohol dose may not have been sufficient to result in alcohol-induced impairment.

However, the sensitivity of the BART could also explain the current results. Unlike other behavioral measures, the BART conceptualizes risk-taking as occurring on a continuum, with risk-taking becoming disadvantageous only at a certain point, which varies according to the circumstances (Lejuez et al., 2003a). Thus, the BART is advised for administration in nonclinical populations, as it captures risky behavior that is not necessarily disadvantageous (Skeel et al., 2008). However, despite significant correlations with self-reported real-world substance use behaviors (Aklin et al., 2005; Lejuez et al., 2003a,b), the BART has not consistently detected the acute effects of drugs on risk-taking. For example, while Acheson and de Wit (2008) reported decreased risk-taking for males and increased risk-taking for females on the BART following 20 mg d-amphetamine, but not diazepam, Menkes (2011) reported no significant effect of 20 mg diazepam. Similarly, Reynolds and colleagues (2006b) found that acute alcohol doses of 0.4 and 0.8 g/kg did not impact performance on the BART. As such, replication of the present study with an alternate measure of behavioral risk-taking may clarify the effect of acute AmED consumption on risk-taking.

The presence of an ED effect suggests that AmED can increase risk-taking via the ED component. However, the magnitude of effect was small, calling into question the practical implications for ED and AmED consumers. It is not known whether the effects of EDs on risk-taking would increase in magnitude with an increasing ED dose, or, indeed, with an increasing alcohol dose. Thus, further clarification is required assessing the dose-dependent effects of AmED on objectively measured risk-taking behavior, to determine: (i) whether the ED effect increases in magnitude with higher doses, and (ii) whether an interactive effect becomes apparent with higher alcohol doses.

The present results also showed negligible correlations between the BART adjusted number of pumps and the RT-18 subscale scores. These results align with several studies revealing no significant correlation between BART outcomes and trait sensation seeking and impulsivity measures (Aklin et al., 2005; Hunt et al., 2005; Lejuez et al., 2003a; Skeel et al., 2008). The majority of the literature points toward weak associations between psychometric and behavioral impulsivity measures among nonclinical samples, suggesting that the behavioral tendencies identified in self-report and laboratory measures may differ (Lane et al., 2003; Reynolds et al., 2006a, 2008).

While the present study was single-blind for alcohol administration, several procedures were enforced to minimize experimenter bias, including: (i) implementing systematic structures for participant interaction (e.g., standardized instructions), (ii) double-blinding of ED administration via coding, (ii) use of objective measurement procedures, and (iii) blinding treatment conditions throughout data processing and analysis. As the perceived alcohol intake in active alcohol conditions was approximately 2.5 standard drinks higher than placebo alcohol conditions, we cannot discount alcohol expectancy effects. However, the perceived alcohol intake and reported ratings of intoxication did not differ significantly for the alcohol and AmED conditions.

Another potential source of bias was the use of the BART adjusted average number of pumps. Although this measure is the primary outcome, trials on which the balloon exploded were excluded, thus discounting the participant's behavior on that trial and lowering the adjusted average (Euser et al., 2011; Pleskac et al., 2008). While approximately one-third of the trials were excluded across the treatment conditions, there was no significant difference across the treatment conditions in the number of explosions. More reliable estimates could be achieved by use of an automatic response mode, in which participants predetermine the number of pumps, allowing the balloon to automatically inflate until the pumps are completed or the break point is reached.

In conclusion, the present study's results suggest that the interactive effect of a moderate alcohol and ED dose were restricted to perceived stimulation, with no significant impact on perceived intoxication and impairment relative to alcohol alone. While no interactive AmED effect was evident for objectively measured risk-taking behavior, there was no effect of alcohol consumption in general on risk-taking outcomes. Engagement in risk-taking behavior was only increased by ED consumption; however, the magnitude of the effect suggests negligible implications for ED and AmED consumers. Conclusions regarding the link between AmED and risk-taking remain tentative until further research is undertaken: (i) under higher dosage domains, and (ii) with alternative validated behavioral measures of risk-taking.


This study was funded by the Alcohol, Tobacco and Other Drugs Council Tas Inc. Placebo samples were supplied by Red Bull GmbH. Research design, data collection, analysis interpretation, and manuscript preparation were the responsibility of the authors.