We find that varenicline attenuates alcohol consumption in smokers. Varenicline also reduces reported cumulative consumption of alcohol and ongoing consumption, as measured using the objective marker ETG. Importantly, in view of the well-established efficacy of varenicline to reduce smoking, the effect on drinking uncovered in this study supports the hypothesis that varenicline will have clinical benefit for heavy-drinking smokers. Furthermore, it is consistent with animal studies indicating that alcohol and nicotine act through a common reward pathway that involves the NAChR (Bito-Onon et al. 2011
; Chatterjee et al. 2011
; Hendrickson et al. 2010
; Steensland et al. 2007
). Although our data indicate that the varenicline-induced reduction of alcohol consumption is independent of the effects of varenicline on smoking, additional studies must be conducted in nonsmoking drinkers to definitively determine whether varenicline would be effective in individuals who do not smoke.
Subjects were told that this was a smoking cessation study and, as such, were asked to identify a quit date. However, subjects were not excluded from the study if they failed to adhere to the quit date or if they chose not to attend the offered group therapy for smoking cessation, mitigating the pressure to actually quit smoking. Only three subjects participated in the offered group therapy sessions and only one subject adhered to their self-ascribed quit date, complicating the possibility of analysis based on smoking cessation. If more subjects had actually quit smoking, it might have been interesting to compare drinking between quitters and nonquitters. In the absence of this comparison, it is not possible to determine how much of the effect of varenicline on drinking was secondary to its effect on smoking.
In light of previously reported findings indicating the effectiveness of open-label varenicline in reducing heavy alcohol consumption among a treatment-seeking population (Fucito et al. 2011
), it was important here to assess the effects of varenicline on drinking in subjects who were not treatment-seeking for alcohol abuse and who were not self-described “alcoholics.” The use of a nontreatment-seeking population can control for sample bias and regression to the mean, as subjects are not explicitly attempting to attenuate their drinking and, therefore, are not necessarily expecting to drink less during the course of drug therapy. Our use of nontreatment-seeking drinkers also allowed us to isolate the effects of drug treatment without influence from other psychosocial strategies, such as Alcoholics Anonymous, group therapy, family support, etc. that treatment seekers may simultaneously pursue in an effort to decrease their intake. Additionally, from a public health perspective, it is important to identify treatments that can be used to effectively reduce drinking independent of abstinence. However, given that the present subject population was seeking treatment for smoking cessation, it remains possible that decreased alcohol craving and consumption was due in part to an attempt to quit smoking and to avoid places, such as bars, where smoking and drinking commonly occur together.
As in many such study populations, subject retention was problematic (Howard 1992
) and our attrition rate was magnified by the long study duration (16 weeks). Dropouts occurred with equal frequency in both groups, and there were no significant demographic differences between those subjects who failed to complete the 16-week study and completers. Subjects’ own reasons for leaving the study were rarely related to concerns about the efficacy, or lack thereof, of the study drug. Thus, we have no reason to believe that attrition biased our results.
Varenicline carries a black box warning for suicidal ideation and suicidal behavior and has also been reported to exacerbate depression, hostility, and aggression (Moore et al. 2010
; but see also Garza et al. 2011
). Though anecdotal, it is of interest that all four of the AEs that occurred in subjects on active study drug were associated with concurrent psychostimulant use. It is possible that some subjects’ use of cocaine and methamphetamine during the study exacerbated the potential for varenicline to induce these negative and potentially dangerous side effects. Although a previous study suggested that varenicline side effects are not exacerbated in cocaine users (Poling et al. 2010
), our results suggest that care should be taken when administering varenicline to alcohol abusers who are concurrently using psychostimulants. However, further studies are necessary to definitively assess the contribution of psychostimulants to varenicline AEs.
California has one of the smallest smoking populations in the USA (http://www.dhs.ca.gov/tobacco
) and, while identification of smokers seeking cessation treatment was not difficult in our population, it was uncommon to encounter an individual smoking a half a pack or more per week. Additionally, our cotinine results suggest that, although subjects were smoking significantly less at the end of the 12 weeks of varenicline treatment, they were, by and large, still smoking. Though our data indicates lower efficacy than previous reports on varenicline treatment of smoking (Gonzales et al. 2006
; Jorenby et al. 2006
), these data may also reflect our moderate inclusion criteria: subjects were asked to plan a quit date but were not dropped from the study if they did not meet this quit date nor were they dropped if they relapsed to cigarette smoking during treatment.