Mechanisms of topiramate effects: refining medications development for alcoholism

The current issue of Addiction Biology includes two manuscripts examining the mechanisms of topiramate (TOP) effects on alcohol use. The study by Kranzler et al. (2014) describes analyses of nighttime drinking behavior among treatment-seeking patients during a 12-week randomized and controlled trial of topiramate (200 mg/ day) for alcohol dependence. Following up on previous work from his group, Kranzler demonstrated that a genetic polymorphism (rs2832407) in the gene encoding the GluK1 kainate subunit (GRIK1) moderates response to topiramate such that C-allele homozygotes show lower levels of nighttime drinking across the 12-week trial. Interestingly, this study found that the pharmacogenetic effect (TOP × GRIK1 SNP) on nighttime drinking was accounted for by mean ratings of self-efficacy. The study by Miranda et al. (2014), in turn, focuses on nontreatment-seeking heavy drinkers assessed in the laboratory and in the natural environment using ecological momentary assessment (EMA) during a 5-week trial of topiramate (200 mg/day). While probing for medication effects on subjective responses to alcohol and alcohol craving, both in the laboratory and in the natural environment, an interesting pattern of results emerged. Topiramate was found to attenuate alcohol craving during a drinking episode, which, in turn, accounted for medication effects on alcohol use during the trial. These studies share a common focus on understanding the mechanisms by which topiramate may be an effective medication for alcoholism with the overarching goal of informing the use of topiramate through a personalized medicine approach. These two studies are methodologically sophisticated as they employ a host of techniques to assess both mechanisms and drinking behavior in a valid and reliable fashion, while also capturing the complex nature of these constructs. Kranzler and colleagues use interactive voice response (IVR) technology to obtain daily diary data in the context of a clinical trial, whereas Miranda and colleagues combine EMA assessments with human laboratory paradigms such as cue-exposure and alcohol challenge. One of the most notable strengths of both studies is their ability to test mechanisms in relation to carefully ascertained alcohol use over the course of each trial. As noted by Miranda and discussed in detail elsewhere (Ray, Hutchison & Tartter 2010), the association between a particular mechanism and alcohol use is often implied in human laboratory studies when in fact it should be explicitly examined. These two studies do a commendable job of capturing the dynamic nature of constructs such as craving and self-efficacy, while interrogating the association between these putative mechanisms of action and alcohol use. Using mediational models, Kranzler et al. and Miranda et al. take this line of inquiry one step further by testing whether medication effects (TOP versus Placebo) on drinking outcomes are in fact explained, or mediated, by self-efficacy and alcohol craving, respectively. The level of methodological rigor in these studies sets the standard for future studies of mechanisms of action of pharmacotherapies for alcoholism. In addition to learning from what these studies have in common, considering their differences may be equally important. To that end, the most notable distinction between these studies is that while Kranzler et al. enrolled treatment-seeking alcohol-dependent individuals, Miranda et al. recruited non-treatment-seeking heavy drinkers, most of whom were not alcoholdependent. Let us consider the issues of treatmentseeking status and alcohol use trajectory in studies of pharmacotherapy for alcoholism. Treatment-seeking is a key construct in that when individuals enroll in a treatment study, it is presumed that their primary motivation is to reduce their drinking or to abstain from alcohol. Conversely, non-treatment seekers are drawn to participate for monetary reinforcement. The smoking cessation literature has recognized that treatment-seeking status affects responses to pharmacotherapy in the clinic and in experimental laboratory paradigms alike (Perkins, Stitzer & Lerman 2006). For example, an early meta-analysis of nicotine replacement trials found that efficacy was greater among individuals who self-referred to a trial than those who were invited from hospital clinics (Tang, Law & Wald 1994). In medications development for alcoholism, less is known about the impact of treatmentseeking status, and clearly, some outcomes may be more vulnerable to the effects of treatment-seeking status than others. Further, within treatment seekers, there is also variability in treatment goals, which, in turn, predicts clinical outcome (Bujarski et al. 2012). The key point being that treatment-seeking status and, more broadly, motivation to change alcohol use represent a continuous, and rather volatile, construct requiring closer consideration in medications development for alcoholism. Next, let us examine the distinction between heavy drinkers and alcohol-dependent samples. The degree to which these samples can inform one another is poorly understood. Recently, it has been argued that even within COMMENTARY bs_bs_bannerAddiction Biology