Caffeinated cocktails: energy drink consumption, high-risk drinking, and alcohol-related consequences among college students.

Abstract

The consumption of alcohol mixed with energy drinks (AmED) is popular on college campuses in the United States. Limited research suggests that energy drink consumption lessens subjective intoxication in persons who also have consumed alcohol. This study examines the relationship between energy drink use, high-risk drinking behavior, and alcohol-related consequences.
In Fall 2006, a Web-based survey was conducted in a stratified random sample of 4,271 college students from 10 universities in North Carolina.
A total of 697 students (24% of past 30-day drinkers) reported consuming AmED in the past 30 days. Students who were male, white, intramural athletes, fraternity or sorority members or pledges, and younger were significantly more likely to consume AmED. In multivariable analyses, consumption of AmED was associated with increased heavy episodic drinking (6.4 days vs. 3.4 days on average; p < 0.001) and twice as many episodes of weekly drunkenness (1.4 days/week vs. 0.73 days/week; p < 0.001). Students who reported consuming AmED had significantly higher prevalence of alcohol-related consequences, including being taken advantage of sexually, taking advantage of another sexually, riding with an intoxicated driver, being physically hurt or injured, and requiring medical treatment (p < 0.05). The effect of consuming AmED on driving while intoxicated depended on a student's reported typical alcohol consumption (interaction p = 0.027).
Almost one-quarter of college student current drinkers reported mixing alcohol with energy drinks. These students are at increased risk for alcohol-related consequences, even after adjusting for the amount of alcohol consumed. Further research is necessary to understand this association and to develop targeted interventions to reduce risk.

Full text

Page 1

CLINICAL PRACTICE
Caffeinated Cocktails: Energy Drink
Consumption, High-risk Drinking, and
Alcohol-related Consequences among
College Students
Mary Claire O’Brien, MD, Thomas P. McCoy, MS, Scott D. Rhodes, PhD, Ashley Wagoner, BS,
Mark Wolfson, PhD
Abstract
Objectives: The consumption of alcohol mixed with energy drinks (AmED) is popular on college cam-
puses in the United States. Limited research suggests that energy drink consumption lessens subjective
intoxication in persons who also have consumed alcohol. This study examines the relationship between
energy drink use, high-risk drinking behavior, and alcohol-related consequences.
Methods: In Fall 2006, a Web-based survey was conducted in a stratified random sample of 4,271 col-
lege students from 10 universities in North Carolina.
Results: A total of 697 students (24% of past 30-day drinkers) reported consuming AmED in the past
30 days. Students who were male, white, intramural athletes, fraternity or sorority members or pledges,
and younger were significantly more likely to consume AmED. In multivariable analyses, consumption
of AmED was associated with increased heavy episodic drinking (6.4 days vs. 3.4 days on average;
p < 0.001) and twice as many episodes of weekly drunkenness (1.4 days⁄week vs. 0.73 days⁄week;
p < 0.001). Students who reported consuming AmED had significantly higher prevalence of alcohol-
related consequences, including being taken advantage of sexually, taking advantage of another sexually,
riding with an intoxicated driver, being physically hurt or injured, and requiring medical treatment
(p < 0.05). The effect of consuming AmED on driving while intoxicated depended on a student’s reported
typical alcohol consumption (interaction p = 0.027).
Conclusions: Almost one-quarter of college student current drinkers reported mixing alcohol with
energy drinks. These students are at increased risk for alcohol-related consequences, even after adjust-
ing for the amount of alcohol consumed. Further research is necessary to understand this association
and to develop targeted interventions to reduce risk.
ACADEMIC EMERGENCY MEDICINE 2008; 15:1–8 ª 2008 by the Society for Academic Emergency
Medicine
Keywords: energy drinks, alcohol, injury
E
plant-based stimulants (e.g., guarana, yerba mate), sim-
ple sugars (e.g., glucose, fructose), glucuronolactone (a
naturally occurring glucose metabolite), amino acids
(e.g., taurine, carnitine, creatine), herbs (e.g., ginkgo
biloba, ginseng), and vitamins. The effects of these ingre-
dients are incompletely understood. A 6-ounce serving
of brewed coffee contains 100 mg of caffeine; the caf-
feine content of energy drinks varies considerably, with
some energy drinks containing more than 300 mg or
more per serving.1In 2006, Americans spent more than
$3.2 billion dollars on energy drinks.2Declining soda
nergy drinks are beverages that claim to provide
a burst of energy by using a combination of caf-
feine (the principal active ingredient), other
ª 2008 by the Society for Academic Emergency Medicine
doi: 10.1111/j.1553-2712.2008.00085.x
ISSN 1069-6563
PII ISSN 1069-6563583
1
From the Departments of Emergency Medicine (MCO), Social
Sciences and Health Policy (MCO, SDR, AW, MW), Biostatisti-
cal Sciences (TPM), and Pediatrics (MW), Wake Forest Univer-
sity School of Medicine, Winston-Salem, NC.
Received November 27, 2007; revision received January 11,
2008; accepted January 13, 2008.
Presented at the Annual Meeting of the American Public Health
Association, Washington, DC, November 4, 2007.
Supported by Grants R01 AA14007-2 and 2R01 AA014007-06A1
from the National Institute on Alcohol Abuse and Alcoholism
(NIAAA) and North Carolina DHHS⁄OJJDP EUDL Award 2004-
AH-FX-0014.
Addressforcorrespondence
O’Brien, MD; e-mail: mobrien@wfubmc.edu.
andreprints:Mary Claire

Page 2

sales have encouraged beverage companies to focus on
this lucrative market2and increasingly to market energy
drinks to specific segments of the population,2including
women, herb and vitamin enthusiasts, the affluent, and
youth.
Thirty-four percent of 18- to 24-year-olds are regular
energy drink users.2Many energy drinks are marketed
to young consumers through the careful selection of
‘‘taboo’’ drink names such as ‘‘Daredevil,’’ ‘‘Whoo-
pAss,’’ and ‘‘Cocaine,’’ and catchy slogans that empha-
size endurance and sexual prowess (e.g., ‘‘You can
sleep when you’re 30’’3; ‘‘Get it up and keep it up’’4).
Energy drinks are selectively and aggressively marketed
to college students. Red Bull, for example, enlists ‘‘stu-
dent brand managers,’’ whose duties include distribut-
ing free samples and gathering information for Red
Bull about individual college cultures to aid in targeted
marketing.5The company also employs ‘‘mobile energy
teams’’ to provide free energy drinks to the public and
to ‘‘educate’’ consumers about the benefits of energy
drinks. Young people are explicitly targeted at extreme
sports events (e.g., race car driving, waterfall kayaking,
parasailing), where energy drink sponsorships fuel a
belief in the invincibility and stamina of the average
energy drink consumer.
Consumingalcoholmixed
(AmED) is common on college campuses.6The alcohol
industry has been criticized for actively promoting the
consumption of AmED, by creating brand confusion
with nonalcoholic versions, pricing alcoholic energy
drinks below nonalcoholic versions, marketing alcoholic
drinks that use energy drinks as an ingredient (e.g.,
Jager Bombs, Red Bull and vodka), and using the stu-
dent-based marketing strategies already exploited by
energy drink companies.7
In two laboratory-based investigations, the consump-
tion of a caffeinated beverage has been shown to lessen
subjective intoxication in persons who also have con-
sumed alcohol.8,9A small double-blinded study detected
no significant differences in the physiologic and bio-
chemical parameters of volunteers who drank alcohol
compared to alcohol plus energy drink.10Caffeine has
been shown to promote voluntary ethanol consumption
in male rats.11
Alcohol is the major risk factor for injury. Emergency
physicians have an opportunity and a responsibility to
address unhealthy alcohol use with their patients. This
study examines the relationships between energy drink
use, high-risk drinking behavior, and alcohol-related
consequences within a sample of college students. We
hypothesizedthat college
AmED would have an increased risk of alcohol-related
consequences.
with energydrinks
studentswhoconsumed
METHODS
Study Design and Population
In Fall 2006, we invited a stratified random sample of
undergraduate college students attending 10 universi-
ties (8 public and 2 private) in North Carolina to com-
plete an online Internet-based survey of alcohol use
and other risk behaviors. The survey was part of the
Study to Prevent Alcohol-Related Consequences among
college students (SPARC), a National Institute on Alco-
hol Abuse and Alcoholism (NIAAA)-funded randomized
community trial to reduce high-risk drinking among
college students. Campus sizes ranged from approxi-
mately 5,375 to 44,841 students (graduate and under-
graduatestudents combined),
having a graduate program.
Campus administrators were approached by the
study team during the project protocol development
and asked whether they wanted their students to
participate. These administrators worked with the study
team to ensure that the project adhered to university
requirementsaswell as
recruitment and participation. Human subjects review
and study oversight were provided by the Wake Forest
University School of Medicine (WFUSM) institutional
review board (IRB). The review boards of other univer-
sities participating in this study either approved this
study or officially deferred to the WFUSM IRB. The
registrar at each campus provided the e-mail addresses
for each enrolled student.
Survey procedures have been described in detail
elsewhere.12–14We first placed posters in common
areas on each campus (e.g., residence halls and dorms,
student unions, and cafeterias) encouraging students
to check their e-mail accounts for an invitation to par-
ticipate in the study. We then randomly selected stu-
dents from each campus by academic classification
(‘‘freshman,’’ ‘‘sophomore,’’
and sent postcards asking them to check their e-mail
accounts for an invitation to complete the online Inter-
net-based assessment. These same randomly selected
students were sent messages by e-mail describing the
study and encouraging them to complete the online In-
ternet-based assessment. The e-mail messages contained
a link to a secure uniform resource locator (URL) where
the student completed the assessment. We sent nonre-
spondents up to four reminder e-mails. All students who
completed the survey were sent e-mails awarding them
$10.00 in PayPal?money. From the list of completions,
one student from each school was randomly selected to
receive $100.00 in PayPal?money.
The sample size varied between campuses based on
the enrollment. However, a target response quota of
33% was assumed so that the selected samples would
produce sufficient completed surveys to provide the
statistical power necessary for the SPARC study.14The
Web-site was shut down shortly after the target num-
ber from each school was achieved (i.e., 365 per
school). A total of 4,271 students completed the survey.
A description of the sample is provided in Table 1.
with alluniversities
expectations forstudent
‘‘junior,’’ and ‘‘senior’’)
Survey Content and Administration
The Web-based survey was developed after reviewing
the Harvard College Alcohol Survey,15
questionnaire,16the Youth Survey questionnaire used
in the National Evaluation of the Enforcing Underage
Drinking Laws program,17DeJong’s College Drinking
Survey,18and the Youth Risk Behavior Survey.19The
questionnaire had 307 items (with multiple skip pat-
terns based on the absence of reported behaviors) and
took between 17 and 24 minutes to complete.14The
survey measured demographic
the CORE
variables,attitudes
2
O’Brien et al.•CAFFEINATED COCKTAILS, RISKY DRINKING, AND INJURY

Page 3

about alcohol consumption, drinking behaviors, other
substance use behaviors, consequences experienced
from one’s own drinking, and consequences experi-
enced as a result of other students’ drinking. Students
were asked whether in the past 30 days they had expe-
rienced any of the following consequences as a result
of their drinking or the drinking of others: being taken
advantage of sexually, taking advantage of another
sexually, driving under the influence of alcohol, riding
with a driver who was under the influence of alcohol,
being hurt or injured, or requiring medical treatment.
Studentswere askedwhy
energy drinks; they were permitted to select multiple
reasons. Items to assess high-risk drinking included:
1) typical number of drinks in a drinking episode;
2) number of days of binge (or heavy episodic) drink-
ing in the past 30 days (range = 0–30 days); 3) number
of days drunk in a typical week (range 0–7 days); and
4) greatest number of drinks in a single episode in the
past 30 days. Binge drinking was defined in the widely
accepted manner as drinking four or more drinks in a
row for females and five or more drinks in a row for
males.20Students were also asked ‘‘In a typical week,
how many days do you get drunk?’’ where drunk was
defined in the standard way as ‘‘dizzy, unsteady, or
sick toyour stomach.’’21
defined as either heavy episodic drinking or drinking
to drunkenness.
they mixalcohol with
High-riskdrinkingwas
Data Analysis
The goals of the statistical analysis were to: 1) estimate
the prevalence of mixing alcohol with energy drinks
among past 30-day drinkers, 2) examine the association
of mixing alcohol with energy drinks and high-risk
drinking, and 3) examine the association of mixing alco-
hol with energy drinks and alcohol-related conse-
quences, after adjusting for drinking behaviors. We
sought to examine the additional risk of adding energy
drinks to alcohol; survey skip patterns restricted the
questions about energy drink use to those students
who reported that they had drunk alcohol at least once
in the past 30- days (N = 2,886; 68% of all students).
Drinking outcomes were analyzed using multivariable
linear mixed-effects modeling. These models adjusted
for student gender, age, race, fraternity or sorority sta-
tus (member, pledge, or neither), athlete status (varsity
athlete, intramural, or neither), and within-campus clus-
tering. Because students were nested within 10 univer-
sities,a mixed-modelapproach
statistical analysis. Specifically, a random intercept
model was fit to account for the within-university cor-
relation of student responses so that students were
considered nested within campus. Adjusted means and
standard errors were calculated for consumption of
AmED, where the observed marginals of the covariates
wereused in estimating.
regression models were fit for the six alcohol-related
wasusedinthe
Mixed-effectslogistic
Table 1
Characteristics of Students by Reporting of Mixing Alcohol and Energy Drinks (AmED) (N = 4,237)
Characteristics, n (%)
or mean ± SD Overall
Nondrinkers,
n = 1,351 (32%)
Non-AmED,
n = 2,189 (52%)
AmED,
n = 697 (16%)
Gender***
Male
Female
No response
Academic classification
Freshman
Sophomore
Junior
Senior
Other
No response
Age**
Race*
Non-Hispanic white
African American
Hispanic
Asian⁄Pacific Islander
Other
No response
Fraternity⁄sorority status**
Pledge
Member
Neither
Athletic status***
Intramural
Varsity
Nonathlete
Campus residence
On-campus
Off-campus
No response
1,638 (39)
2,572 (61)
27 (<1)
450 (33)
895 (66)
6 (<1)
857 (39)
1,317 (60)
15 (<1)
331 (47)
360 (52)
6 (<1)
1,110 (26)
1,078 (25)
1,045 (25)
840 (20)
129 (3)
35 (<1)
20.4 ± 2.8
481 (36)
356 (26)
334 (25)
158 (12)
16 (1)
6 (<1)
20.0 ± 2.9
468 (21)
546 (25)
546 (25)
515 (24)
91 (4)
23 (1)
20.7 ± 2.8
161 (23)
176 (25)
165 (24)
167 (24)
22 (3)
6 (<1)
20.3 ± 2.2
3,291 (78)
353 (8)
161 (4)
228 (5)
191 (5)
13 (<1)
939 (70)
184 (14)
58 (4)
94 (7)
72 (5)
4 (<1)
1,766 (81)
140 (6)
77 (4)
109 (5)
90 (4)
7 (<1)
586 (84)
29 (4)
26 (4)
25 (4)
29 (4)
2 (<1)
143 (3)
368 (9)
3,726 (88)
20 (1)
46 (3)
85 (4)
221 (10)
1,883 (86)
38 (5)
101 (14)
558 (80) 1,285 (95)
936 (22)
220 (5)
3,081 (73)
242 (18)
80 (6)
1,029 (76)
471 (22)
107 (5)
1,611 (74)
223 (32)
33 (5)
441 (63)
2,401 (57)
1,813 (43)
23 (<1)
911 (67)
435 (32)
5 (<1)
1,154 (53)
1,023 (47)
12 (<1)
336 (48)
355 (51)
6 (<1)
*p < 0.05, **p < 0.01, and ***p < 0.001 from mixed-effects logistic regression comparing AmED vs. non-AmED only.
ACAD EMERG MED•www.aemj.org
3

Page 4

consequences. The logistic models were used to assess
if there was a significant increase in the probability of
reporting the given consequence in the past 30 days,
adjusting for student age, gender, race, fraternity or
sorority status, athlete status, typical number of drinks
in a single episode (all as covariates), and within-cam-
pus clustering (school as random effect). The primary
independent variable was the indicator variable for
consuming energy drinks with alcohol on any day in
the past 30 days (AmED; 1 = yes, 0 = no). In the models
for alcohol-related consequences, the two-way interac-
tion of AmED and typical alcohol consumption (No. of
drinks) was first tested and retained in the model if sig-
nificant at the 5% level. Predicted probabilities were
calculated using observed marginals of the covariates.
Analysis of the outcome ‘‘driving while under the influ-
ence’’ resulted in a significant interaction effect between
energy drink consumption with alcohol and typical
number of drinks. To illustrate the nature of this inter-
action effect, predicted probabilities were calculated for
all combinations of these two variables (Figure 1). Uni-
variate analysis for checking normality was assessed
using box plots, Q-Q plots, and Shapiro-Wilk tests.
Regression assumptions were assessed with residual
analyses and influence diagnostics. Multicollinearity
diagnostics, such as variance inflation factors (VIFs),
were checked and modeling adequacy supported (all
VIFs < 1.3, below the diagnostic cutoff of 10).22,23
All analyses were performed using Stata v9.2 (Stata
Corp, College Station, TX) and the Generalized Linear
Latent and Mixed Models (GLLAMM) package (http://
www.gllamm.org/).24The criterion for statistical signifi-
cance was a two-sided p-value of <0.05.
RESULTS
A total of 4,271 students completed the survey. A total
of 4,237 students answered questions about past 30-day
drinking and consuming AmED (99.2%); 61% of these
students were female. The average student age was
20.4 years (range = 17–30 years). Of the 4,237 students,
2,886 (68%) reported drinking alcohol at least once in
the past 30 days. Eighty-seven percent of past 30-day
student drinkers were between the ages of 18 and
22 years. Among drinkers, 24% (697 students) reported
consuming AmED on at least 1 day in the past 30 days.
All subsequent analyses are restricted to past 30-day
drinkers.
In the bivariate analyses, students who were male
(p < 0.001),white(p = 0.040),
(p < 0.001), fraternity or sorority members or pledges
(p < 0.01), and younger (p < 0.01) were more likely to
consume AmED (see Table 1). The average age of first
drink was younger among students who reported con-
suming AmED (15.1 years vs. 16.0 years; p < 0.001).
Students who reported consuming AmED had more
drinking days on average during their last year of high
school than those who did not report consuming
AmED (1.7 days vs. 1.2 days; p < 0.001).
Fifty-five percent of students who consumed AmED
said they did so to hide the flavor of the alcohol (48%
of male mixers and 61% of female mixers.) Fifteen per-
cent of students reported mixing alcohol with energy
intramuralathletes
drinks to drink more and not feel as drunk; 5%
reported they mix to drink more alcohol and not
look as drunk. Seven percent of students reported
mixing energy drinks and alcohol so as not to get a
hangover. Forty-one percent of students provided other
reasons (e.g., ‘‘It was being served at a party’’; ‘‘It was
the only mixer available’’; ‘‘That’s how you make Jager
Bombs.’’)
In multivariable analyses, the consumption of AmED
was strongly associated with all six high-risk drinking
behaviors (see Table 2). Compared to current drinkers
who did not report consuming AmED, students who
reported consuming AmED drank significantly more
during a typical drinking session (5.8 drinks vs. 4.5
drinks⁄typical session; p < 0.001). They reported almost
twice as many heavy episodic drinking days in the past
30 days, as defined by the gender-specific measure of
four or more alcoholic beverages in a row for females
and five or more alcoholic drinks in a row for males
(6.4 days vs. 3.4 days; p < 0.001). They reported twice as
many episodes of weekly drunkenness (1.4 days⁄week
vs. 0.73 days⁄week; p < 0.001). Among students who
reported mixing alcohol and energy drinks, the greatest
number of drinks in a single episode of drinking was
36% higher, compared to drinking students who did
not report mixing alcohol with energy drinks (8.3 drinks
vs. 6.1 drinks; p < 0.001).
Students who reported
increased prevalence of all six alcohol-related conse-
quences due to their own drinking (see Table 3). They
were more likely to be taken advantage of sexually
(adjusted oddsratio [AOR] = 1.77;
advantageof someoneelse
p < 0.002), ride with a driver who was under the influ-
ence of alcohol (AOR = 2.20; p < 0.001), be hurt or
injured (AOR 2.25; p < 0.001), or require medical treat-
ment (AOR = 2.17; p = 0.007). For driving in a car under
the influence of alcohol, the effect of mixing alcohol
and energy drinks depended upon the typical number
ofdrinks thestudent drank
p < 0.001; typicalnumber
p < 0.001;interactionAOR = 0.93,
were significant main effects of consuming AmED
(p < 0.001) and typical number of drinks (p < 0.001), but
the risk for driving a car under the influence of alcohol
was greater at lower number of typical drinks for those
students who consumed AmED, compared to those stu-
dents who drank alcohol alone (see Figure 1).
consumingAmEDhad
p < 0.002),
(AOR = 2.18;
take
sexually
(AmED
drinks
p = 0.027).
AOR = 2.96,
AOR = 1.23, of
There
DISCUSSION
Studies of college students have demonstrated that the
riskof alcohol-related
increased at five or more drinks in a row for men and
four or more drinks in a row for women.25Serious
injury, sexual assault, drunk driving, and death are all
associated with heavy episodic drinking.26However,
even after adjusting for alcohol consumption, students
who consumed AmED had dramatically higher rates of
serious alcohol-related consequences. This is the first
study to suggest that consuming AmED constitutes
‘‘high-risk drinking’’ for college students, independent
of the quantity of alcohol consumed.
consequencesis greatly
4
O’Brien et al.•CAFFEINATED COCKTAILS, RISKY DRINKING, AND INJURY

Page 5

Table 2
Association of Energy Drink⁄Alcohol Mixers and Drinking Behaviors (n = 2,886)*
Drinking Behavior
Non-AmED,
n = 2,189 (76%)
AmED,
n = 697 (24%)b? 95% CI
z
Statistic p-Value?
Typical No. of drinks in single episode
No. of days with 5⁄4 heavy episodic
drinking in past 30 days
No. of days drunk in a typical week
Most No. of drinks in single episode
in past 30 days
4.5 ± 0.15
3.4 ± 0.17
5.8 ± 0.17
6.4 ± 0.23
1.4 (1.1, 1.6)
2.9 (2.5, 3.3)
11.69
14.21
<0.001
<0.001
0.73 ± 0.04
6.1 ± 0.15
1.4 ± 0.05
8.3 ± 0.19
0.70 (0.61, 0.79)
2.2 (1.9, 2.5)
15.44
14.28
<0.001
<0.001
Numbers given in first two columns are adjusted means ± SE from multivariable linear mixed-effects regression model adjusting
for student gender, age, race, fraternity or sorority status, athlete status, and within-campus clustering.
AmED = alcohol mixed with energy drinks; SE = standard error.
*Students who reported drinking in past 30 days only.
?b is the regression coefficient of the indicator variable comparing AmED to non-AmED drinkers.
?p-value is from comparing AmED vs. non-AmED for each drinking behavior outcome.
Table 3
Association of Consuming Alcohol Mixed with Energy Drinks (AmED) and Alcohol-related Consequences (n = 2,886)*
Consequences
Non-AmED,
n = 2,189 (76%)
(95% CI)
AmED,
n = 697 (24%)
(95% CI)AOR (95% CI)z Statistic p-Value?
Was taken advantage of sexually
Took advantage of another sexually
Rode with a driver who was under
the influence of alcohol?
Was hurt or injured
Required medical treatment
3.7% (2.9, 4.8)
1.7% (1.2, 2.4)
22.5% (18.6, 26.9)
6.4% (4.7, 8.7)
3.7% (2.5, 5.4)
38.9% (32.7, 45.6)
1.77 (1.23, 2.55)
2.18 (1.34, 3.55)
2.20 (1.81, 2.68)
3.05
3.13
7.83
0.002
0.002
<0.001
5.9% (4.8, 7.2)
1.2% (0.8, 1.8)
12.3% (9.9, 15.3)
2.6% (1.7, 4.1)
2.25 (1.70, 2.96)
2.17 (1.24, 3.80)
5.74
2.70
<0.001
0.007
AOR = adjusted odds ratio; CI = confidence interval.
*Students who reported drinking in past 30 days only.
?p-value is from comparing AmED users to Non-AmED for each drinking behavior outcome in a multivariable mixed-effects
logistic regression model adjusting for student gender, age, race, fraternity or sorority status, athlete status, typical number of
drinks in an episode, and within-campus clustering.
?Drove a car while under the influence of alcohol not reported in this table due to significant interaction effect of AmED use and
typical number of drinks.
Figure 1. Association of consuming alcohol mixed with energy drinks (AmED) and driving a car under the influence of alcohol.
ACAD EMERG MED•www.aemj.org
5