Steven J. Nieto

Background: The U.S. Food and Drug Administration (FDA) recommends a minimum duration of 6 months for randomized controlled trials (RCTs) evaluating pharmacotherapies for alcohol use disorder (AUD). The relationship between trial length and treatment effects is not well understood. The current study conducted meta‐regression analyses of RCTs for AUD to examine the association between trial length and observed medication effect sizes in clinical trials. Methods: This secondary data analysis utilized data from a systematic literature review that identified 139 pharmacotherapy RCTs for AUD conducted between 1985 and 2023. Meta‐regressions were conducted using the metafor package in R to examine whether shorter (< 12 weeks) or longer (> 12 weeks) trials differed in observed medication effect sizes for abstinence‐ and drinking‐based endpoints, relative to 12‐week trials, which represented the median trial duration. Publication bias was addressed using a weight‐function model. Results: For abstinence‐based endpoints, neither shorter (< 12 weeks; β = 0.13, SE = 0.21, p = 0.54) nor longer (> 12 weeks; β = 0.11, SE = 0.11, p = 0.30) trials differed significantly from the 12‐week reference group in observed medication effect sizes. Similarly, for drinking‐based endpoints, neither shorter (< 12 weeks; β = −0.07, SE = 0.09, p = 0.39) nor longer (> 12 weeks; β = 0.06, SE = 0.05, p = 0.26) trials differed significantly from 12‐week trials in observed medication effect sizes. All results remained nonsignificant after adjusting for publication bias. In sensitivity analyses restricted to trials with statistically significant effect sizes, trials > 12 weeks showed smaller estimated treatment effects compared to trials ≤ 12 weeks. Conclusions: Observed medication effect sizes did not differ significantly across trial lengths for abstinence and drinking‐based endpoints, except for analyses restricted to trials with significant effect sizes, where there was an advantage for trials that were 12 weeks or shorter. These findings inform efforts to optimize AUD RCT design.

Despite increasing attention to the neurobiology of alcohol use disorder (AUD), development of effective pharmacotherapies remains limited. Many medications with promising preclinical profiles fail to translate into clinical efficacy. This scoping review characterizes preclinical studies evaluating medications for AUD that have also been tested in human laboratory paradigms or clinical trials. We focus on two widely used paradigms: the two‐bottle choice (2 BC ) task, which measures alcohol consumption and preference, and operant reinstatement models, which capture relapse‐like behavior. A systematic search and data extraction identified preclinical studies using 2 BC and reinstatement paradigms to evaluate medications that progressed to human testing. Study characteristics (species, strain, sex, outcomes, and sample sizes) were recorded and effect sizes (corrected Cohen's d ) were summarized across outcome domains. Ninety‐two unique studies met criteria: 74 employing the 2 BC paradigm and 18 using operant reinstatement models, including 1731 animals. Rats were the most common species and sex bias was pronounced: 77 2 BC studies using only males, five only females, and all 18 reinstatement studies used male rats exclusively. For 2 BC consumption, the largest effects were observed for medications with limited studies, such as levetiracetam ( d = −4.58), while more extensively tested medications showed moderate effects, including prazosin ( d = −0.74) and naltrexone ( d = −0.63). For 2 BC preference, levetiracetam again showed a large effect in a single study ( d = −4.29). In reinstatement models, rimonabant showed the largest effect ( d = −3.11), while naltrexone ( d = −0.70), baclofen ( d = −1.05), and varenicline ( d = −0.72) had moderate effects. This review highlights substantial heterogeneity in study design and underutilization of reinstatement and female animals. Medications with moderate, replicable preclinical effects, especially those tested across both 2 BC and reinstatement models, may hold greater translational promise. These findings offer key opportunities for enhancing the translational alignment of preclinical AUD pharmacotherapy research.

Background and aims: The alcohol cue‐reactivity paradigm is widely used to test the initial efficacy of potential medications for alcohol use disorder (AUD); however, there is limited quantitative evidence demonstrating that medication effects on cue‐induced craving are associated with efficacy in clinical trials. This meta‐analysis examined whether medication effects on cue‐induced alcohol craving are associated with medication effects in randomized clinical trials (RCTs). Methods: Follow‐up meta‐analysis of RCTs. Participants included people who engaged in heavy drinking or have AUD. Medications were compared with a placebo control. We computed medication effect sizes (Cohen's d ) for cue‐induced craving (k _{# of studies} = 36 studies; 15 medications) and for the following individual RCT endpoints (k _{# of studies} = 139 studies; 19 medications): percent days abstinent, percent heavy drinking days, the percentage of participants who returned to any drinking, the percentage of participants who returned to heavy drinking, drinks per day and drinks per drinking day. We applied Williamson‐York regression models to test the relationship between medication effects on cue‐induced craving and the six individual RCT endpoints. One‐sided P values were used to test directional hypotheses and two‐sided P values were also reported to allow for exploratory interpretation. Results: Medication effect sizes on cue‐induced craving were positively associated with medication effect sizes on percent participants who returned to heavy drinking [ = 1.06, standard error (SE) = 0.63, one‐sided P = 0.04, two‐sided P = 0.09; k _{# of effect size} = 74; 7 medications] in RCTs but no other RCT endpoints tested. Specifically, medications that reduced cue‐induced craving in the human laboratory also decreased the percentage of participants who returned to heavy drinking in RCTs. Conclusions: Medications that produce greater reductions in cue‐induced alcohol craving appear to be associated with lower rates of return to heavy drinking, but not consistently with other drinking outcomes. This pattern indicates that the cue‐reactivity paradigm has limited and outcome‐specific translational validity rather than functioning as a universal indicator of clinical efficacy.

Medication development for alcohol use disorder (AUD) represents a challenging, costly, and time‐consuming process. A typical development trajectory involves testing compounds in animal models, human laboratory studies, and randomized clinical trials (RCTs). This paper reviews a series of data‐driven studies seeking to evaluate best practices in medication development for AUD at each of the three levels of analysis. First, we evaluate the role of behavioral pharmacology paradigms in early efficacy testing and highlight the importance of preclinical and human laboratory evidence in making critical “go/no‐go” decisions throughout the development process. Second, we discuss a recent translational meta‐analytic approach that integrates preclinical, human laboratory, and RCT data to inform AUD medication development. This narrative review synthesizes findings from systematic reviews, meta‐regression analyses, and integrated data across stages of medication testing. Findings from this work suggest that factors such as sample characteristics and study design influence the detection of medication effects on AUD outcomes, emphasizing the need for standardized methodologies. Furthermore, findings suggest that early efficacy signals in animal models and human laboratory studies can predict specific clinical trial outcomes, offering valuable insights for data‐informed decision making in medication development. Together, these studies bridge the gap between the preclinical and clinical stages, facilitating more efficient medication development for AUD.

The alcohol cue‐exposure paradigm has a long and rich history in alcohol use disorder (AUD) research, contributing to the identification of risk factors, the recognition of craving as a core symptom, the understanding of AUD neurobiology, and the development of both pharmacological and behavioral treatments. The goal of this review was to evaluate the utility of the alcohol cue‐exposure paradigm as a screening tool for AUD medication development and to provide recommendations for refinement of the paradigm. First, we review evidence supporting the predictive validity and clinical applicability of the alcohol cue‐exposure paradigm. Second, we examine current practices in implementing the paradigm in AUD pharmacotherapy studies. Third, we highlight several areas for refinement and offer recommendations for the implementation of this paradigm. Finally, we outline key conclusions and actionable future directions. In summary, while the alcohol cue‐exposure paradigm has an established foundation in the study of AUD pharmacotherapies, its future success hinges on a concerted effort to refine and standardize protocols and validate outcome measures in large‐scale studies. Addressing these priorities could accelerate the development and regulatory approval of novel treatments for AUD.

Aims: Individuals with alcohol use disorder (AUD) often exhibit heightened delay discounting, a behavioral marker associated with poor treatment outcomes. Medications such as naltrexone and varenicline influence reward-related decision-making, but their effects on delay discounting remain unclear. This study examined whether these medications influence delay discounting rates. Methods: We conducted a double-blind, randomized, placebo-controlled trial in 34 treatment-seeking adults with AUD. Participants were assigned to naltrexone (50 mg/day), varenicline (2 mg/day), or placebo and completed a two-week medication titration followed by a six-day quit attempt. Delay discounting was assessed at baseline and post-treatment using the Monetary Choice Questionnaire (MCQ). General linear models tested medication effects on post-treatment discounting, controlling for baseline discounting, education, and income. Results: A significant interaction between medication and baseline delay discounting emerged (P = .03; η2 = .67). Among participants with lower baseline discounting, naltrexone reduced delay discounting compared to placebo and varenicline. However, no significant effects were observed in participants with higher baseline discounting. Varenicline did not significantly alter delay discounting compared to placebo. Conclusions: These findings suggest that naltrexone may reduce delay discounting in individuals with AUD, but primarily among those with lower discounting rates. The results highlight the importance of baseline traits in understanding medication effects on decision-making. Given the small sample size, future research should replicate these findings in larger trials and explore whether delay discounting could serve as a biomarker for personalized AUD treatment.

Background: The alcohol cue-exposure paradigm is widely used in alcohol use disorder (AUD) research. Individuals with AUD exhibit considerable variability in their alcohol cue-reactivity, highlighting the need to identify characteristics that contribute to this heterogeneity. This study applied machine learning models to identify clinical and sociodemographic predictors of subjective alcohol cue-reactivity (ALCUrge). Methods: Individuals with AUD (N = 139; 83 M/56F) completed an alcohol cue-exposure paradigm and a battery of clinical and sociodemographic measures. ALCUrge (primary outcome variable) was assessed using the Alcohol Urge Questionnaire following alcohol cue-exposure. We implemented three machine learning models (Lasso regression, Ridge regression, Random Forest) to identify clinical and sociodemographic predictors of ALCUrge and compared model performance (i.e. predictive accuracy). Results: Lasso regression had the strongest predictive accuracy, with a Root Mean Square Error (RMSE) of 9.48, followed by Random Forest (RMSE = 9.95), and Ridge regression (RMSE = 10.40). All models outperformed chance-level prediction (null baseline model RMSE = 14.80). Top predictors of ALCUrge across multiple models were alcohol urge prior to cue-exposure, compulsive alcohol-related behaviors/thoughts, tonic alcohol craving, cigarette smoking status, and biological sex. Higher pre-cue exposure alcohol urge, more compulsive alcohol-related tendencies, greater tonic craving, and occasional cigarette use was associated with greater predicted ALCUrge, while being female was associated with lower predicted ALCUrge. Conclusion: This study advances our understanding of the phenotypic overlap in the compulsive aspects of tonic craving and phasic cue-induced alcohol urge, and offers insight into additional factors, such as biological sex and cigarette smoking, that may contribute to variability in alcohol cue-reactivity.