Longitudinal evidence from the National Longitudinal Study of Adolescent Health and the Adolescent Risk Study suggests that the co-occurrence of alcohol and tobacco use escalates during adolescence and reaches an asymptote by age 25 (Jackson et al. 2002). Moreover, these authors found that alcohol use strongly predicted tobacco use, more so than tobacco use predicted alcohol use. Additionally, both initiation and persistence of smoking were associated with prior drinking. Simon et al. (1995) demonstrated that alcohol use predicted post 1-year tobacco use in a large sample of adolescents. In college students, Sher et al. (1996) found reciprocal relationships between the development of alcohol and tobacco use disorders. Overall, these findings suggest that the concurrent use of alcohol and tobacco may facilitate the development of a tobacco use disorder.

Prevalence estimates collected over the prior decade find that approximately 30% of young adults are current smokers (Harrison et al. 2008; Johnston et al. 2001; Rigotti et al. 2000) and that young adults have the highest rates of smoking compared to older adults. Data from the 2007 National Health Interview survey shows that rates of current smoking among adults aged 18–24 were 20.7% higher than all adults. With regard to nondaily smoking, population studies have documented that 19–24% of current smokers are nondaily smokers (Hassmiller et al. 2003; Hennrikus et al. 1996), with younger age being significantly associated with this smoking pattern (Gilpin et al. 1997; Hassmiller et al. 2003; Hennrikus et al. 1996).

Research examining young adults has documented that nondaily smoking is most likely to occur in the context of alcohol use (Dierker et al. 2006; Nichter et al. 2006). We have found that young adult nondependent smokers report that 74% of all smoking episodes occurred while under the influence of alcohol (McKee et al. 2004). Interviews conducted with these smokers find that they are cognizant of the reasons why they co-use alcohol and tobacco and report that they experience increased pleasure from cigarettes and increased desire for cigarettes when drinking alcohol (McKee et al. 2004). When examining their episodic patterns of co-use, cigarette use was most likely to occur following heavy alcohol consumption (Harrison et al. 2009).

It is possible that alcohol may improve the subjective experience of cigarette smoking in heavy-drinking young adults who are still at an experimental stage of smoking, but this has yet to be investigated. A large proportion of smokers consolidate their smoking patterns during early adulthood, making this an important population to study. For example, Wechsler et al. (1998) reported that of college smokers, “11% had their first cigarette and 28% began to smoke regularly at or after the age of 19” (p. 1673). When examining adolescent to adult smoking trajectories, Chassin et al. (2000) characterized 39% (1,108/2,870) of smokers as late-stable smokers. These late-stable smokers had infrequent smoking during high school but increased to daily smoking over the course of early adulthood. It is possible that increased rates of drinking during early adulthood (Dawson et al. 2004) may be associated with this pattern of smoking behavior.

For the current study, we were interested in examining how alcohol might improve the subjective effects associated with cigarette smoking in heavy-drinking young adults (age 21–25) who are still at an experimental stage of smoking. The primary aim of this research was to investigate the effect of a high dose of alcohol, compared to a taste-masked placebo or a mixer beverage, on reactivity to tobacco and subsequent cigarette self-administration behavior. We hypothesized that alcohol (0.08 g/dL targeted blood alcohol level), compared to a placebo or a no-alcohol mixer beverage, would increase positive subjective effects (e.g., satisfaction) and decrease negative subject effects (e.g., nausea) associated with smoking a single cigarette. In young adult nondaily smokers, King et al. (2009) found that alcohol compared with placebo increased some positive effects of smoking. However, the King et al. (2009) study was designed to mask expectancy effects. It is possible that the expectation of alcohol consumption may be sufficient to alter subjective reactivity associated with cigarette smoking. Non-daily smoking has also been referred to as social smoking (Nichter et al. 2006; Shiffman et al. 2009) and may be associated with alcohol through nonpharmacological mechanisms. Thus, we included an expectancy manipulation. For the taste-masked placebo condition, subjects were told that the beverage contained alcohol, whereas the instructions for the alcohol and mixer conditions were consistent with the contents of the beverage. Secondary predictions included whether alcohol, compared to a placebo or mixer beverage, would increase ratings of tobacco and alcohol craving, the intensity of smoking behavior (i.e., greater puff volume, greater puff duration, shorter inter-puff interval, and greater depth of inhalation), and the number of cigarettes smoked. This initial study has implications for understanding how alcohol may facilitate the development of a tobacco use disorder and provide the basis for further investigations examining potential mechanisms for this effect.

Our findings are supported by prior studies examining the effect of alcohol and placebo beverages on craving responses (King and Epstein 2005; Epstein et al. 2007; Sayette et al. 2005). The Sayette et al. (2005) study included an expectancy manipulation for the taste-masked placebo beverage and found that at 10-min post-consumption, both alcohol and placebo beverages increased tobacco craving in light daily smokers. The authors suggest that alcohol or the expectation of alcohol functioned as a cue for smoking prior to subjects being able to discern pharmacological differences between alcohol and placebo beverages. In the latter post-beverage timepoints, Sayette et al. (2005) found that tobacco craving was greater following the alcohol beverage relative to the placebo beverage. In our relatively inexperienced smokers, the expectation of alcohol consumption also functioned as a cue for smoking during the ascending limb by increasing tobacco craving for positive reinforcement during the alcohol and placebo conditions. Following smoking, tobacco craving decreased during both the alcohol and placebo conditions but remained greater following alcohol consumption; supporting the Sayette et al. (2005) finding that the pharmacological effects of alcohol may augment continued desire to smoke. Other studies, which have not manipulated instructions concerning beverage content, have found that alcohol but not taste-masked placebo beverages increased tobacco craving (King and Epstein 2005; Epstein et al. 2007).

Following alcohol consumption, heavy-drinking experimental smokers smoked more cigarettes, which was due to a greater likelihood of choosing to smoke additional cigarettes. Subjects were most likely to choose to smoke additional cigarettes (58%) following alcohol consumption, relative to the placebo (42%) and mixer conditions (32%). While alcohol increased the amount smoked, there were no differences in how each cigarette was smoked (i.e., number of puffs, puff duration, puff volume, peak puff volume, and inter-puff interval). These findings are in contrast to those demonstrated by King et al. (2009). In their sample of young adult heavy-drinking nondependent smokers, men but not women increased puff count, volume, and duration following alcohol consumption. However, King et al. (2009) calculated topography measures based on the total amount smoked during the ad lib period, rather than separately for each cigarette. When using cumulative topography measures, amount smoked is merged with how each cigarette was smoked; thus, it is unknown whether alcohol altered per cigarette indices of smoking topography. Additionally, subjects in the King et al. (2009) study were older and smoked more frequently in greater quantities than the subjects in the current study, which also may account for the differences in the smoking topography findings.

One clear difference that was observed between the alcohol and placebo conditions concerned the negative subjective effects (i.e., nausea) associated with smoking. A possible explanation for the expectancy-related variation could lie in the nature of the effects. Significant positive self-reported effects (e.g., calm, satisfaction) could be considered to be more cognitive, while the negative effects (nausea) may be more visceral and interoceptive. When manipulating pharmacological and expectancy effects associated with alcohol, self-reported rather than physiological effects generally demonstrate stronger expectancy effects (see McKay and Schare 1999 for review). However, it should be noted that nausea ratings were generally low (<26 out of 100), suggesting that this finding may be of limited clinical significance.

It is known that individuals who initially experience cigarette smoking as a pleasurable rather than a negative experience are most likely to develop a pattern of daily smoking (Ríos-Bedoya et al. 2009). If alcohol reduces negative subjective experiences and the expectation of alcohol improves positive subjective experiences associated with tobacco smoking, it is possible that these factors may facilitate the transition to regular daily smoking. Many smokers stabilize their smoking patterns in early adulthood (Chassin et al. 2000), and the current findings support the notion that alcohol may play a role in this trajectory by increasing the reinforcing value of smoking. It is thought that alcohol potentiates the reinforcing effects of tobacco and vice versa (Rose et al. 2004; Shiffman and Balabanis 1995). Microdialysis studies with nicotine-naive rats have found that alcohol and nicotine additively increase extracellular dopamine release in the nucleus accumbens, providing a possible explanation for this relationship (Tizabi et al. 2002).

This study has several limitations. The present work included a single dose of alcohol (0.08 g/dL targeted BAL), less than what the participants reported that they generally drank in a drinking episode. Future research could incorporate higher and lower doses of alcohol to explore dose-dependent effects on reactivity to smoking, as well as explore the full blood alcohol curve as the current study focused on ascending limb effects only. Interview-based findings have indicated that inexperienced smokers tend to drink to binge levels prior to smoking during a drinking episode (consistent with a 0.08 g/dL blood alcohol level) and then to continue to smoke intermittently during alcohol consumption (Harrison et al. 2009). Future work could also manipulate the method of nicotine delivery to determine whether alcohol and the expectation of alcohol improve reactivity to smoking tobacco more generally or to nicotine specifically. Studies have demonstrated that nicotine increases alcohol consumption in non-nicotine-dependent male smokers (Acheson et al. 2006; Barrett et al. 2006) and decreases consumption in non-nicotine-dependent female smokers (Acheson et al. 2006). While we did not find any effects of gender on our primary outcomes, this study was not powered to examine gender differences. Also, our sample was primarily Caucasian, and results may be limited in their generalizability to other racial or ethnic groups. Finally, as outlined in the methods section and similar to other investigations (Acheson et al. 2006; Fillmore et al. 2009; Perkins et al. 2004), we altered the response format of some of our measures, which may change the psychometric properties of these scales.

In our sample, we attempted to recruit individuals who may develop into Chassin’s et al. (2000) “late-stable smokers”, which had infrequent smoking during high school but increased to daily smoking over the course of early adulthood. Due to alcohol consumption, our minimum age was 21. We hypothesize that our findings would also generalize to younger samples, but this would have to be examined with other methodological approaches (e.g., event sampling). While we did demonstrate that positive and negative subjective reactions to smoking were altered either by alcohol or the expectation of alcohol consumption, use of a CReSS smoking topography system may have reduced smoking enjoyment (Blank et al. 2009). Also, subjective reactivity to smoking a cigarette was measured immediately after smoking the single cigarette. While the instructions requested that subjects respond to how the cigarette made them feel, it is possible that subjects misattributed the effects of alcohol to this tobacco-related questionnaire. Ideally, this same item set should have been administered with regard to alcohol effects just prior to smoking. There were also issues with the alcohol blinding. While we did demonstrate significant expectancy effects, only 71% believed they had consumed alcohol in the placebo condition. Asking subjects to discern the amount of alcohol in a drink would have been a more sensitive estimation of subjects’ abilities to discriminate the high-alcohol from the placebo beverage, given that the placebo contained a small amount of alcohol. Also, examining lower doses of alcohol (e.g., 0.04 g/kg) which can be more effectively blinded would further help to disentangle expectancy and pharmacological effects.