Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Feb 1.
Published in final edited form as: Alcohol Clin Exp Res (Hoboken). 2024 Jan 11;48(2):420–429. doi: 10.1111/acer.15257

Reductions in WHO risk drinking levels correlate with alcohol craving among individuals with alcohol use disorder

Felicia R Tuchman 1,*, Kevin A Hallgren 2, Dylan K Richards 3, Arnie Aldridge 4, Raymond K Anton 5, Henri-Jean Aubin 6, Henry R Kranzler 7, Karl Mann 8, Stephanie S O’Malley 9, Katie Witkiewitz 1,3
PMCID: PMC10922776  NIHMSID: NIHMS1954835  PMID: 38149364

Abstract

Background:

Abstinence has historically been considered the preferred goal of alcohol use disorder (AUD) treatment. However, most individuals with AUD do not want to abstain and many are able to reduce their drinking successfully. Craving is often a target of pharmacological and behavioral interventions for AUD, and reductions in craving may signal recovery. Whether reductions in drinking during AUD treatment are associated with reductions in craving has not been well examined.

Methods:

We conducted secondary analyses of data from three AUD clinical trial s (N’s=1327, 346, and 200). Drinking reductions from baseline to end of treatment were measured via changes in World Health Organization (WHO) risk drinking levels; alcohol craving was measured using validated self-report measures. Regression analyses tested whether drinking reductions were associated with end-of-treatment craving reductions; moderation analyses tested whether associations between drinking reduction and end-of-treatment craving differed across AUD severity.

Results:

Reductions of at least 1 or at least 2 WHO risk drinking levels were associated with lower craving (all p’s<0.05). Results were substantively similar after removing abstainers at end-of-treatment. Associations between drinking reductions and craving were generally not moderated by AUD severity.

Conclusions:

WHO risk drinking level reductions are associated with significantly lower craving, as compared to those who did not achieve meaningful reductions in drinking. Results demonstrate the utility of WHO risk drinking levels as AUD clinical trial endpoints and provide evidence that drinking reductions mitigate craving.

Keywords: Alcohol use disorder, craving, alcohol, controlled drinking, non-abstinent recovery

1. Introduction

Alcohol use disorder (AUD) has a lifetime prevalence of 29.1% and often goes untreated (Grant et al., 2015). The 5th edition of the Diagnostic and Statistical Manual for Mental Disorders (DSM-5) includes alcohol craving as one of the criteria for a diagnosis of AUD (American Psychiatric Association, 2013). Alcohol craving is a subjectively distressing experience to many individuals and can lead to persistent or increased alcohol consumption (Serre et al., 2015, 2018; McHugh et al., 2016; Wemm et al., 2022). Thus, the reduction of alcohol craving is an important target for treatment of AUD, and offering treatments that target craving reductions may motivate individuals to participate in alcohol treatment.

Many behavioral and pharmacological treatments for AUD have been developed explicitly to reduce craving (Witkiewitz, Litten and Leggio, 2019; Bach et al., 2023). Craving may also be reduced by treatments that were not explicitly developed to have that effect (Erwin and Slaton, 2014; Schacht et al., 2014; Roberts, Harrison and McKee, 2017). Historically, the goal of most alcohol treatments has been for patients to become abstinent, and the Food and Drug Administration currently recommends abstinence or absence of heavy drinking days (defined as 4+/5+ drinks for females/males) as endpoints for alcohol clinical trials; however, drinking reductions short of abstinence are achievable and sustainable (Tucker & Witkiewitz, 2022). Based on this evidence, a recent definition of AUD recovery proposed by the National Institute on Alcohol Abuse and Alcoholism incorporates reductions in drinking short of total abstinence as reflecting recovery (Hagman et al., 2022). To quantify meaningful drinking reductions, the World Health Organization (WHO) risk drinking level metric has emerged as an alternative endpoint to abstinence or no heavy drinking days for clinical trials (Hasin et al., 2017).

Several recent empirical studies have shown that reductions in drinking, as assessed by WHO risk drinking levels, are associated with sustained improvements in quality of life, social functioning, physical health, mental health, reductions in other alcohol-related consequences, and health care costs–constructs increasingly considered central to the definition of recovery (Aldridge et al., 2021; Hasin et al., 2017; Knox et al., 2018, 2019, 2020; Tucker & Witkiewitz, 2022; World Health Organization, 2000; Witkiewitz et al., 2017, 2018). Importantly, in all of these studies, the effects of drinking reductions on improvements in functioning were not driven by abstainers, underscoring the association of non-abstinent drinking reductions with improved social, behavioral, and physical functioning.

Studies that have examined how non-abstinent drinking reductions are related to craving have shown lower levels of drinking to be associated with lower levels of alcohol craving and even greater reductions in alcohol craving for individuals who maintained abstinence (Hallgren, Delker, et al., 2018b; Kohen et al., 2023; Kuerbis et al., 2020; Roberts et al., 1999; Santos et al., 2022; Trela et al., 2018; Waddell et al., 2022; Witkiewitz et al., 2019). Notably, however, no prior studies have evaluated whether drinking reductions, quantified by reductions in WHO risk drinking levels, are associated with reductions in craving.

Few studies have evaluated the potential moderating role of AUD severity in the relationship between drinking reductions and craving, despite that higher craving levels are associated with greater AUD severity (MacKillop et al., 2010; Moak et al., 1998; Murphy et al., 2014; Yoon et al., 2006). In one study, AUD severity significantly moderated the association between heavy drinking contexts and craving such that the relationship between heavy drinking contexts and craving was stronger for individuals with higher AUD severity than for individuals with lower AUD severity (Kuerbis et al., 2020). A second study found evidence for incubation of craving, whereby increased exposure to alcohol cues observed in the environment was associated with increased craving over time among those with AUD, but not those who did not have AUD (Treloar Padovano and Miranda, 2021). The distinction between cue-induced craving and average levels of craving is important to note, given that non-human animal studies and some studies in humans have generally found cue-induced craving to be associated with dependence severity (Courtney et al., 2014). Yet other studies examining associations between WHO risk level reductions and biopsychosocial functioning variables have found no significant moderating effect of AUD severity (Witkiewitz, Falk, et al., 2019; Witkiewitz et al., 2020). The sparse human literature highlights the need for research that explores AUD severity as a potential moderator of the association between drinking reductions (here measured as WHO risk drinking level reductions) and craving.

In the present study, we investigated (1) the relationship between reductions in drinking and alcohol craving among adults receiving treatment for AUD, (2) the extent to which the relationship is driven by those who maintained abstinence, and (3) whether AUD severity moderates this relationship. We hypothesized that individuals who reported reductions in WHO risk drinking levels during treatment would report a reduction in alcohol craving and that this result would not be driven solely by individuals who abstained from alcohol over the course of treatment. Given the scant, contradictory literature on the role of AUD severity in relation to drinking reductions and craving, we did not advance hypotheses about AUD severity as a moderator. To test our research questions, we conducted secondary data analyses on datasets from three randomized clinical trials that evaluated pharmacological and behavioral interventions for AUD and systematically measured craving constructs.

2. Materials and Methods

2.1. Participants and Procedures

We conducted secondary data analyses on datasets from three randomized clinical trials that evaluated pharmacological and behavioral interventions for AUD: the COMBINE, Horizant, and Varenicline studies (Anton et al., 2006; Litten et al., 2013; Falk et al., 2019). Complete demographic data for participants in all studies can be found in Supplementary Table 1. Of note, participants in these clinical trials were largely male, non-Hispanic white, and employed; as such, results may not generalize to more racially and ethnically diverse samples of individuals who are female and with lower socioeconomic status.

The COMBINE study was a 16-week, multisite randomized double-blind clinical trial that evaluated combinations of medications (acamprosate, naltrexone, acamprosate plus naltrexone, or placebo) and behavioral interventions (medication management or combined behavioral intervention) in U.S. adults with DSM-IV alcohol dependence from 2001 to 2004 (first randomization to final follow-up) (Anton et al., 2006). Participants were randomized into 1 of 8 treatment conditions using a 2 × 2 × 2 design, and a ninth treatment condition received combined behavioral intervention with no study drug. Participants completed assessments at baseline, clinic visits, end of treatment, and 3 post-treatment follow-ups. For the present study, we used only data collected at the baseline (week 0) and end of treatment (week 16) timepoints. COMBINE participants (n = 1383) were largely male (68.8%) and non-Hispanic white (76.7%) with a mean age of 44.4 years (SD = 10.2).

The Horizant study was a 26-week, multisite randomized double-blind clinical trial that evaluated the safety and efficacy of gabapentin enacarbil extended-release (GE-XR) in treating U.S. adults with DSM-5 alcohol use disorder (Falk et al., 2019). Participants were recruited between 2015 and 2017 (Falk et al., 2019). Participants were randomized to receive either GE-XR or a matched placebo. Participants completed assessments at baseline, during treatment, and at follow-up interviews. For the present study, we used only data collected at the baseline (week 0) and end of treatment (week 26) timepoints. Horizant participants (n = 338) were largely male (66.0%) and non-Hispanic white (67.2%) with a mean age of 50.1 years (SD = 15.4).

The Varenicline study was a 13-week, multisite randomized double-blind clinical trial that evaluated the safety and efficacy of varenicline in treating U.S. adults with DSM-IV-TR alcohol dependence(Litten et al., 2013). Participants were recruited between 2011 and 2012 (Litten et al., 2013). Participants were randomized to receive either varenicline or placebo. Participants completed assessments at baseline, during treatment, and at a follow-up interview. For the present study, we used only data collected at the baseline (week 0) and end of treatment (week 14) time points. Varenicline participants (n = 200) were largely male (70.0%) and non-Hispanic white (65.5%) with a mean age of 45.0 years (SD = 16.5).

2.2. Measures

In all 3 studies, alcohol consumption was assessed at baseline and end of treatment using calendar based Timeline Follow-Back interviews (Sobell and Sobell, 1995; Miller, 1996). Average daily alcohol consumption over the prior 28 days was used to categorize participants into abstinence or one of four WHO risk drinking levels: very high, high, moderate, or low risk (World Health Organization, 2000). Very high risk was defined as daily consumption of >100 g of ethanol for men or >60 g for women, high risk as >60 to <100 gs for men or >40 to <60 g for women, moderate risk as >40 to <60 g for men or >20 to <40 g for women, low risk as 1 to <40 g for men or 1 to <20 g for women, and abstinence was defined as 0g for both men and women (World Health Organization, 2000). AUD severity was scored as a continuous measure of AUD criterion counts, as assessed by the Structured Clinical Interview (SCID) for DSM-IV in COMBINE and the MINI International Neuropsychiatric Interview of DSM-5 AUD in the Horizant and Varenicline studies (First et al., 1997; Sheehan et al., 1998). Notably, the MINI but not the SCID includes craving in its measurement of AUD severity.

In COMBINE, craving was assessed via the Obsessive Compulsive Drinking Scale, a 14-item self-report questionnaire that assess craving in the past week (OCDS; Anton, 2000). Items 7 and 8 of the OCDS assess drinking quantity and frequency, but not craving, and were excluded from the total OCDS score prior to analyses. Each item of the OCDS contains 5 response options scored on a scale of 0 to 4, and response options vary by item (Anton, 2000). In the present sample, internal consistency of OCDS items was high (ω = 0.96).

In Horizant, craving was assessed using the revised, short form version of the Alcohol Craving Questionnaire, a 12-item self-report measure that assesses momentary craving (“right now”) (ACQ-SF-R; Singleton, 1997). Each item of the ACQ-SF-R contains 7 Likert-type response options (i.e., strongly disagree to strongly agree; Singleton, 1997). In the present sample, internal consistency of ACQ-SF-R items was high (ω = 0.91).

In Varenicline, craving was assessed using the Penn Alcohol Craving Scale, a 5-item self-report questionnaire that assess craving in the past week (PACS; Flannery et al., 1999). Each item of the PACS contains 6 response options that vary by item (Flannery, Volpicelli and Pettinati, 1999). In the present sample, internal consistency of PACS items was high (ω = 0.93). Prior work has shown the PACS and OCDS to be related (Flannery et al., 1999; 2003).

2.3. Statistical Analysis

Data were analyzed using multiple regression in R version 2022.12.0+353 (R Core Team, 2023). Craving measures, which differed between studies and were the dependent variable in all analyses, were standardized as z-scores (with a mean of 0 and standard deviation of 1) to facilitate comparison of regression coefficients. AUD severity was mean centered prior to analyses. Change in WHO risk drinking level was calculated as a difference score between baseline and end-of-treatment WHO risk drinking levels. The Horizant and Varenicline samples comprised only individuals of moderate WHO risk level or above at baseline. To be consistent with the population in these trials, in our analyses, we included only individuals in COMBINE who were moderate WHO risk level or above at baseline (which was 95.0% of the COMBINE sample). In sensitivity analyses, we restricted the sample by excluding participants who achieved total abstinence from alcohol in the 28 days prior to the end-of-treatment time point to test whether non-abstinent drinking reductions were associated with reductions in craving. We used linear regression models to estimate the association between the change in WHO risk drinking level (the models’ independent variables) and end-of-treatment craving (the outcome variable). The models controlled for age, sex, treatment arm, baseline craving, AUD severity, baseline WHO drinking risk level, and the interaction between change in WHO risk drinking level and AUD severity.

By controlling for baseline craving, the results of regression analyses effectively model the relationship between 1- and 2-level WHO risk reductions and residualized changes in craving from baseline to end of treatment (controlling for demographics and baseline WHO drinking risk level). We estimated separate models for each study sample (COMBINE, Horizant, and Varenicline) and for each end-of-treatment WHO risk drinking level outcome of interest (at least a 1-level reduction and at least a 2-level reduction), and sensitivity analyses repeated these with abstinent participants removed for a total of 12 models. Significant interaction effects were examined using simple slopes analyses, conducted using the interActive Shiny app (McCabe et al., 2018).

COMBINE had 45.95% missingness in the craving variable at baseline and 57.1% missingness at end of treatment as well as 0% missingness in the WHO risk drinking level variable at baseline and 3.71% missingness at end of treatment. In Horizant, these values were 0%, 22.0%, 0%, and 18.8%, respectively, and in Varenicline, these values were 0%, 5.5%, 0%, and 8.0%, respectively. All missing data were handled using multiple imputation via the “mice” package in R (van Buuren & Groothuis-Oudshoorn, 2011) to minimize bias and retain participants with missing data in statistical analyses. Baseline and end-of-treatment craving, change in WHO risk drinking level, age, sex, treatment arm, baseline WHO risk drinking level, and AUD severity were used to impute missingness. This yielded results that were consistent with results using worst-case imputation and no imputation.

3. Results

Model results can be found in Supplementary Tables 2a–4b. In COMBINE, participants endorsed an average of 5.5 (SD = 1.3; range = 3–7) AUD criteria at baseline. In Horizant, participants endorsed an average of 7.5 (SD = 2.1; range = 6–11) AUD criteria at baseline. In Varenicline, participants endorsed an average of 5.6 (SD = 1.3; range = 6–11) AUD criteria at baseline.

In COMBINE, participants reported a mean OCDS craving score of 19.8 (SD = 7.4) at baseline and 14.1 (SD = 8.1) at end of treatment. In Horizant, participants reported a mean ACS craving score of 43.2 (SD = 14.5) at baseline and 28.8 (SD = 12.9) at end of treatment. In Varenicline, participants reported a mean PACS craving score of 17.2 (SD = 6.5) at baseline and 10.0 (SD = 6.6) at end of treatment.

3.1. COMBINE

Regression analysis indicated individuals with at least a 1-level WHO risk level reduction reported significantly lower craving than individuals whose WHO risk level did not change or increased (B = −1.51, CI: [−1.76, −1.25], p < 0.001; Figure 1a). Similarly, individuals with at least a 2-level WHO risk level reduction reported significantly lower craving than those whose WHO risk level decreased by 1 level, did not change, or increased (B = −1.46, CI: [−1.60, −1.31], p < 0.001; Figure 1b).

Figure 1a.

Figure 1a

Open in a new tab

COMBINE at Least 1-level WHO Risk Drinking Level Reduction Predicting End of Treatment Craving

Figure 1b.

Figure 1b

Open in a new tab

COMBINE at Least 2-level WHO Risk Drinking Level Reduction Predicting End of Treatment Craving

AUD severity significantly moderated the relationship between drinking reduction and craving for individuals with at least a 1-level WHO risk level reduction (B = −0.21, CI: [−0.41, −0.01], p = 0.042; Figure 1c). However, simple slopes analysis revealed no association between AUD severity and craving among individuals who achieved at least a 1-level WHO risk level reduction (B = 0.04, CI: [−0.02, 0.10], p = 0.21) and no association between AUD severity and craving among those who did not achieve at least a 1-level WHO risk level reduction (B = 0.03, CI: [−0.23, 0.30], p = 0.80). Null effects in the simple slopes analysis may be attributed to reduced power.

Figure 1c.

Figure 1c

Open in a new tab

COMBINE: Interaction between AUD Severity and End of Treatment Craving by at least 1-level WHO Risk level Reduction

AUD severity significantly moderated the relationship between drinking reduction and craving for individuals with at least a 2-level WHO risk level reduction (B = −0.24, CI: [−0.36, −0.11], p < 0.001; Figure 1d). Simple slopes analysis revealed no association between AUD severity and craving among individuals who achieved at least a 2-level WHO risk level reduction (B = 0.05, CI: [−0.01, 0.11], p = 0.10) and a positive association between AUD severity and craving among those who did not achieve at least a 2-level WHO risk level reduction (B = 0.17, CI: [0.03, 0.31], p = 0.019).

Figure 1d.

Figure 1d

Open in a new tab

COMBINE: Interaction between AUD Severity and End of Treatment Craving by at least 2-level WHO Risk level Reduction

3.2. Horizant

Regression analysis indicated individuals with at least a 1-level WHO risk level reduction reported significantly lower craving than individuals whose WHO risk level did not change or increased (B = −0.53, CI: [−0.79, −0.26], p < 0.001; Figure 2a). Individuals with at least a 2-level WHO risk level reduction also reported significantly lower craving than individuals whose WHO risk level reduced by 1 level, did not change, or increased (B = −0.52, CI: [−0.73, −0.32], p < 0.001; Figure 2b).

Figure 2a.

Figure 2a

Open in a new tab

Horizant at Least 1-level WHO Risk Drinking Level Reduction Predicting End of Treatment Craving

Figure 2b.

Figure 2b

Open in a new tab

Horizant at Least 2-level WHO Risk Drinking Level Reduction Predicting End of Treatment Craving

Among individuals with at least a 1-level WHO risk level reduction, AUD severity did not significantly moderate the relationship between drinking reduction and craving (B = −0.13, CI: [−0.38, 0.12], p = 0.30). The same was true for individuals with at least a 2-level WHO risk level reduction (B = −0.11, CI: [−0.31, 0.10], p = 0.30).

3.3. Varenicline

Regression analysis indicated individuals with at least a 1-level WHO risk level reduction reported significantly lower craving than individuals whose WHO risk level did not change or increased (B = −0.84, CI: [−1.08, −0.61], p < 0.001; Figure 3a). Individuals with at least a 2-level WHO risk level reduction also reported significantly lower craving than individuals whose WHO risk level decreased by 1 level, did not change, or increased (B = −0.87, CI: [−1.10, −0.64], p < 0.001; Figure 3b).

Figure 3a.

Figure 3a

Open in a new tab

Varenicline at Least 1-level WHO Risk Drinking Level Reduction Predicting End of Treatment Craving

Figure 3b.

Figure 3b

Open in a new tab

Varenicline at Least 2-level WHO Risk Drinking Level Reduction Predicting End of Treatment Craving

Among individuals with at least a 1-level WHO risk level reduction, AUD severity did not significantly moderate the relationship between drinking reduction and craving (B = 0.15, CI: [−0.10, 0.40], p = 0.24). The same was true for individuals with at least a 2-level WHO risk level reduction (B = 0.05, CI: [−0.18, 0.28], p = 0.68).

3.4. Sensitivity Analyses

We conducted sensitivity analyses to examine whether non-abstinent reductions in drinking were associated with reductions in craving. These analyses excluded participants who abstained from alcohol in the 28 days prior to the end-of-treatment time point. In all 3 trials, at least a 1-level or at least a 2-level reduction in drinking, short of total abstinence, were associated with significantly lower craving (COMBINE 1-level: B = −0.77, CI: [−1.05, −0.49], p < 0.001; Figure 1a; COMBINE 2-level: B = −0.81, CI: [−1.02, −0.60], p < 0.001; Figure 1b; Horizant 1-level: B = −0.36, CI: [−0.66, −0.05], p = 0.022; Figure 2a; Horizant 2-level: B = −0.46, CI: [−0.79, −0.13], p = 0.007; Figure 2b; Varenicline 1-level: B = −0.61, CI: [−0.92, −0.30], p < 0.001; Figure 3a; Varenicline 2-level: B = −0.81, CI: [−1.14, −0.47], p < 0.001; Figure 3b). In no cases was AUD severity a significant moderator of the relationship between drinking reductions and craving.

4. Discussion

Using data from three randomized clinical trials of different medications for treating AUD, we found that 1- and 2-level reductions in WHO risk drinking levels were consistently associated with greater reductions in alcohol craving at the end of treatment. Results were substantively similar with abstainers excluded from the analyses, indicating that these findings are neither specific to nor driven by those who maintained abstinence.

With the exception of COMBINE, AUD severity did not moderate the relationship between drinking reductions and craving. The moderation effect in COMBINE was such that there was a stronger association between AUD severity and craving among those who did not achieve WHO risk drinking level reductions. Additionally, although mean craving scores were lower with abstainers included, individuals who reduced drinking reported significantly lower craving than those who did not reduce their drinking. These results extend prior findings that reductions in WHO risk drinking levels following treatment for AUD are associated with benefits to client functioning and further validate WHO risk drinking levels as a treatment endpoint (Hasin et al., 2017; Knox et al., 2018; Witkiewitz et al., 2017, 2018).

Historically, abstinence was considered the only acceptable goal in many AUD treatment contexts (Davis et al., 2017; Davis & Lauritsen, 2016; Davis & Rosenberg, 2013). Even today, among treatment providers, non-abstinent drinking goals are more stigmatized than a goal of abstinence, particularly for individuals with severe AUD and those from minoritized racial and ethnic backgrounds (Davis et al., 2017; Rosenberg et al., 2020; Scott & Wahl, 2011; Xin et al., 2022). Most people with AUD do not want to be abstinent from alcohol, and the finding that drinking reductions are associated with lower craving––even in the absence of complete abstinence from alcohol––could encourage treatment providers to offer programs inclusive of non-abstinent goals and individuals with AUD to seek treatment (Davis and Rosenberg, 2013; Probst et al., 2015; Davis and Lauritsen, 2016; Davis, Rosenberg and Rosansky, 2017). That AUD severity generally did not moderate the association between drinking reductions and craving contradicts assumptions held by both trainees and professionals in the addiction field and can add to treatment providers’ and clients’ confidence in pursuing non-abstinent drinking goals (Davis and Lauritsen, 2016; Rosenberg, Grant and Davis, 2020; Xin, España and Davis, 2022).

4.1. Limitations and Future Directions

This study has several limitations. First, smaller sample sizes in the Horizant and Varenicline studies limited power for those analyses. This could have prevented us from detecting a significant moderating effect of AUD severity in Horizant and Varenicline similar to those found among individuals in COMBINE. Second, because COMBINE excluded individuals with comorbid substance use disorders, unstable medical conditions, and comorbid psychiatric disorders for which the individual was taking medication the generalizability of findings from that study is limited. However, it is promising that the main effect findings were replicated in the Horizant and Varenicline studies. Third, each of these samples was largely composed of non-Hispanic white participants, and individuals’ experiences of help-seeking and recovery from AUD vary widely by racial and ethnic background, which also limits the generalizability of the study findings (Zemore et al., 2009; Lee et al., 2018). Fourth, although measures were time-lagged and controlled for some potential confounds, the current study followed a non-experimental observational design. Thus, it is not possible to determine the directions of causality between drinking reductions and craving. Fifth, high levels of missingness in the COMBINE dataset may be non-random and may have biased the results of those analyses for reasons that we cannot identify. Finally, since the present study was a secondary data analysis, we were constrained to the measures that we had data available for. In each sample, craving was measured differently and each study only included aggregated measures of alcohol craving rather than measures of cue-induced craving. Specifically, in Horizant, the ACQ asked participants to report on their craving “right now,” while in COMBINE and Varenicline, the OCDS and PACS asked participants to report on their craving in the past week. That there was consilience across models using different craving measures supports the validity of our findings; however, if the relationship between craving and drinking reductions is time-dependent, these measures might be obscuring important nuance. Further, that a significant moderating effect of AUD severity was found only for COMBINE, which measured a narrower scope of AUD criteria than the other two trials, suggests that the constructs measured in the SCID may better represent moderators of the relationship between drinking and craving than those measured in the MINI.

The study also had noteworthy strengths. The use of data from three clinical trials allowed us to test for replication across independent treatment samples. Because the three trials used different instruments to quantify craving, we were able to test these relationships with different craving measures. The finding that AUD severity did not generally moderate the effects suggests that across the continuum of severity of AUD individuals who reduce their drinking may experience lower craving and that targeting drinking reductions may be beneficial for reducing craving even among those with the highest AUD severity. It may also be the case that average ratings of craving are less sensitive to the effects of AUD severity than measures of cue-induced craving or incubation of craving over time (Courtney et al., 2014; Treloar Padovano and Miranda, 2021).

Thus, future research using large samples is needed to test the moderating effect of AUD severity in the relationship between drinking reductions and craving. Additionally, subsequent studies should address how drinking reductions might be differentially related to craving in different sociodemographic groups and how drinking reductions may impact cue-induced craving or incubation of craving over time. Beyond overcoming limitations of the present study, future work ought to explore how, why, and for whom inclusive messaging both by treatment providers and society at-large about non-abstinent paths to recovery is helpful to AUD treatment- and non-treatment-seeking individuals. Moreover, given our finding that WHO risk drinking level reductions are a meaningful endpoint in clinical trials for AUD treatments as well as the specificity and international utility of this measurement (although standard drink sizes vary by country, WHO risk drinking levels are measured in grams, allowing for easier cross-cultural comparisons), we hope that clinical trials will more often implement WHO risk drinking level reductions as endpoints going forward (Hasin et al., 2017). Further, the utility of measuring WHO risk drinking levels as part of routine care for AUD treatment ought to be studied, as this clinically meaningful metric may help providers monitor changes in alcohol use that are associated with reduced craving and improved functioning. Finally, the present analyses evaluated the relationship between drinking reductions and end of treatment craving, but it is possible that the relationship between drinking reductions and craving is dynamic. Thus, future work should use a mixed-effects methodology to test the extent to which the relationship between drinking reductions and craving is time dependent and changes throughout treatment (Hallgren et al., 2015; Hallgren, Delker and Simpson, 2018b; Hallgren, Wilson and Witkiewitz, 2018).

4.2. Conclusions

Among treatment-seeking populations with AUD, drinking reductions are associated with less alcohol craving compared to no drinking reductions. This result is generally not driven by those who abstain completely and is generally not moderated by AUD severity. Further, total abstinence is associated with greater reductions in alcohol craving than drinking reductions short of total abstinence. Taken together, these findings support the WHO risk drinking levels as an endpoint that is associated with how patients feel and function and non-abstinent drinking reductions as associated with lower craving.

Supplementary Material

Supinfo

Funding:

This work was supported by NIAAA (R01 AA022328). DKR is supported by an early career development award from NIAAA (K01 AA030789). HRK is supported by the Mental Illness Research, Education and Clinical Center at the Crescenz VAMC in Philadelphia. These funding sources had no role in study design, analysis and interpretation of data, writing of the report, or the decision to submit it for publication.

Conflicts of Interest:

KW, SSO, RKA, HJA, AA, KM, and HRK are members of Alcohol Clinical Trials Initiative (ACTIVE) Workgroup, which is recently supported by Alkermes, Dicerna, Beam Diagnostics, Otsuka, and Kinnov Pharmaceuticals. In the past 36 months, its activities were supported by Alkermes, Dicerna, Ethypharm, Lundbeck, Mitsubishi, and Otsuka. HRK is a member of advisory boards for Dicerna Pharmaceuticals, Sophrosyne Pharmaceuticals, Enthion Pharmaceuticals, and Clearmind Medicine; a consultant to Sobrera Pharmaceuticals; the recipient of research funding and medication supplies for an investigator-initiated study from Alkermes; and a holder of U.S. patent 10,900,082 titled: “Genotypeguided dosing of opioid agonists,” issued on the 26th of January in 2021. HJA is a member of advisory boards, DSMB, or steering committees for Kinnov Pharmaceuticals, Bioprojet, and Ethypharm, and has received sponsorship for scientific meeting attendance, speaker honoraria, or consultancy fees from Ethypharm, Kinnov Pharmaceuticals, and Lundbeck. RKA is the Chair of ACTIVE and consultant for Imbrium/Purdue Pharma, Beam Diagnostics, Dicerna Pharmaceuticals, Denovo Biopharma, Kinnov Pharmaceuticals, Pear Therapeutics, Sobrera Pharma, and Sophrosyne Pharmaceuticals. KM is a member of an advisory board for Kinnov Pharmaceuticals.

References

  1. American Psychiatric Association. (2013) ‘Diagnostic and statistical manual of mental disorders’, 5th ed. 10.1176/appi.books.9780890425596 [DOI] [Google Scholar]
  2. Aldridge AP et al. (2021) ‘The Relationship Between Reductions in WHO Risk Drinking Levels During Treatment and Subsequent Healthcare Costs for the ACTIVE Workgroup’, Journal of Addiction Medicine [Preprint]. Available at: 10.1097/ADM.0000000000000925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anton RF (2000) ‘Obsessive-compulsive aspects of craving: development of the Obsessive Compulsive Drinking Scale’, Addiction (Abingdon, England), 95 Suppl 2, pp. S211–217. Available at: 10.1080/09652140050111771. [DOI] [PubMed] [Google Scholar]
  4. Anton RF et al. (2006) ‘Combined pharmacotherapies and behavioral interventions for alcohol dependence: the COMBINE study: a randomized controlled trial’, JAMA, 295(17), pp. 2003–2017. Available at: 10.1001/jama.295.17.2003. [DOI] [PubMed] [Google Scholar]
  5. Bach P et al. (2023) ‘Deep brain stimulation of the nucleus accumbens in treatment-resistant alcohol use disorder: a double-blind randomized controlled multi-center trial’, Translational Psychiatry, 13(1), pp. 1–11. Available at: 10.1038/s41398-023-02337-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Courtney KE et al. (2014) ‘The association between cue-reactivity in the precuneus and level of dependence on nicotine and alcohol’, Drug and Alcohol Dependence, 141, pp. 21–26. Available at: 10.1016/j.drugalcdep.2014.04.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Davis AK and Lauritsen KJ (2016) ‘Acceptability of non-abstinence goals among students enrolled in addiction studies programs across the United States’, Substance Abuse, 37(1), pp. 204–208. Available at: 10.1080/08897077.2015.1015702. [DOI] [PubMed] [Google Scholar]
  8. Davis AK and Rosenberg H (2013) ‘Acceptance of non-abstinence goals by addiction professionals in the United States’, Psychology of Addictive Behaviors, 27, pp. 1102–1109. Available at: 10.1037/a0030563. [DOI] [PubMed] [Google Scholar]
  9. Davis AK, Rosenberg H and Rosansky JA (2017) ‘American counselors’ acceptance of non-abstinence outcome goals for clients diagnosed with co-occurring substance use and other psychiatric disorders’, Journal of Substance Abuse Treatment, 82, pp. 29–33. Available at: 10.1016/j.jsat.2017.08.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Erwin BL and Slaton RM (2014) ‘Varenicline in the treatment of alcohol use disorders’, The Annals of Pharmacotherapy, 48(11), pp. 1445–1455. Available at: 10.1177/1060028014545806. [DOI] [PubMed] [Google Scholar]
  11. Falk DE et al. (2019) ‘Gabapentin Enacarbil Extended-Release for Alcohol Use Disorder: A Randomized, Double-Blind, Placebo-Controlled, Multisite Trial Assessing Efficacy and Safety’, Alcoholism, Clinical and Experimental Research, 43(1), pp. 158–169. Available at: 10.1111/acer.13917. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Flannery BA, Volpicelli JR and Pettinati HM (1999) ‘Psychometric properties of the Penn Alcohol Craving Scale’, Alcoholism, Clinical and Experimental Research, 23(8), pp. 1289–1295. [PubMed] [Google Scholar]
  13. Grant BF et al. (2015) ‘Epidemiology of DSM-5 Alcohol Use Disorder’, JAMA psychiatry, 72(8), pp. 757–766. Available at: 10.1001/jamapsychiatry.2015.0584. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hagman BT et al. (2022) ‘Defining Recovery From Alcohol Use Disorder: Development of an NIAAA Research Definition’, American Journal of Psychiatry, 179(11), pp. 807–813. Available at: 10.1176/appi.ajp.21090963. [DOI] [PubMed] [Google Scholar]
  15. Hallgren KA et al. (2015) ‘Trajectories of Drinking Urges During Individual- and Couple-based Cognitive-Behavioral Treatment for Alcohol Use Disorders.’, Alcoholism treatment quarterly, 33(2), pp. 161–184. Available at: 10.1080/07347324.2015.1018778. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hallgren KA, Delker BC and Simpson TL (2018b) ‘Effects of Initiating Abstinence from Alcohol on Daily Craving and Negative Affect: Results from a Pharmacotherapy Clinical Trial’, Alcohol: Clinical and Experimental Research, 42(3), pp. 634–645. Available at: 10.1111/acer.13591. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hallgren KA, Wilson AD and Witkiewitz K (2018) ‘Advancing analytic approaches to address key questions in mechanisms of behavior change research’, Journal of Studies on Alcohol and Drugs, 79(2), pp. 182–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hasin Deborah S et al. (2017) ‘Change in non-abstinent WHO drinking risk levels and alcohol dependence: a 3 year follow-up study in the US general population’, The Lancet Psychiatry, 4(6), pp. 469–476. Available at: 10.1016/S2215-0366(17)30130-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Knox J et al. (2018) ‘Reduction in nonabstinent WHO drinking risk levels and change in risk for liver disease and positive AUDIT-C scores: Prospective 3-year follow-up results in the U.S. general population.’, Alcoholism: Clinical and Experimental Research, 42(11), pp. 2256–2265. Available at: 10.1111/acer.13884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Knox J et al. (2019) ‘Reduction in non-abstinent WHO drinking risk levels and depression/anxiety disorders: 3-year follow-up results in the US general population’, Drug and Alcohol Dependence, 197, pp. 228–235. Available at: 10.1016/j.drugalcdep.2019.01.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Knox J et al. (2020) ‘Reduction in World Health Organization Risk Drinking Levels and Cardiovascular Disease’, Alcoholism, Clinical and Experimental Research, 44(8), pp. 1625–1635. Available at: 10.1111/acer.14386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kohen CB et al. (2023) ‘Alcohol craving in the natural environment: Moderating roles of cue exposure, drinking, and alcohol sensitivity’, Experimental and Clinical Psychopharmacology, 31, pp. 57–71. Available at: 10.1037/pha0000540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kuerbis AN et al. (2020) ‘Context and Craving among Individuals with Alcohol Use Disorder Attempting to Moderate Their Drinking’, Experimental and clinical psychopharmacology, p. 10.1037/pha0000349. Available at: 10.1037/pha0000349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lee MR et al. (2018) ‘Desistance and Severity of Alcohol Use Disorder: A Lifespan-Developmental Investigation’, Clinical Psychological Science, 6(1), pp. 90–105. Available at: 10.1177/2167702617736852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Litten RZ et al. (2013) ‘A double-blind, placebo-controlled trial assessing the efficacy of varenicline tartrate for alcohol dependence’, Journal of Addiction Medicine, 7(4), pp. 277–286. Available at: 10.1097/ADM.0b013e31829623f4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. MacKillop J et al. (2010) ‘Alcohol Demand, Delayed Reward Discounting, and Craving in relation to Drinking and Alcohol Use Disorders’, Journal of abnormal psychology, 119(1), pp. 106–114. Available at: 10.1037/a0017513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. McHugh RK et al. (2016) ‘Association between a brief alcohol craving measure and drinking in the following week’, Addiction, 111(6), pp. 1004–1010. Available at: 10.1111/add.13311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Miller WR (1996) ‘Form 90: A Structured Assessment Interview for Drinking and Related Behaviors: Test Manual: (563242012–001)’. American Psychological Association. Available at: 10.1037/e563242012-001. [DOI] [Google Scholar]
  29. Murphy CM et al. (2014) ‘Craving as an Alcohol Use Disorder Symptom in DSM-5: An Empirical Examination in a Treatment-seeking Sample’, Experimental and clinical psychopharmacology, 22(1), pp. 43–49. Available at: 10.1037/a0034535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Probst C et al. (2015) ‘Alcohol use disorder severity and reported reasons not to seek treatment: a cross-sectional study in European primary care practices’, Substance Abuse Treatment, Prevention, and Policy, 10, p. 32. Available at: 10.1186/s13011-015-0028-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Roberts W, Harrison ELR and McKee SA (2017) ‘Effects of varenicline on alcohol cue reactivity in heavy drinkers’, Psychopharmacology, 234(18), pp. 2737–2745. Available at: 10.1007/s00213-017-4667-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rosenberg H, Grant J and Davis AK (2020) ‘Acceptance of Non-Abstinence as an Outcome Goal for Individuals Diagnosed With Substance Use Disorders: A Narrative Review of Published Research’, Journal of Studies on Alcohol and Drugs, 81(4), pp. 405–415. Available at: 10.15288/jsad.2020.81.405. [DOI] [PubMed] [Google Scholar]
  33. Santos G-M et al. (2022) ‘Targeted Oral Naltrexone for Mild to Moderate Alcohol Use Disorder Among Sexual and Gender Minority Men: A Randomized Trial’, American Journal of Psychiatry, 179(12), pp. 915–926. Available at: 10.1176/appi.ajp.20220335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Schacht JP et al. (2014) ‘Varenicline effects on drinking, craving and neural reward processing among non-treatment-seeking alcohol-dependent individuals’, Psychopharmacology, 231(18), pp. 3799–3807. Available at: 10.1007/s00213-014-3518-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Serre F et al. (2015) ‘Ecological momentary assessment in the investigation of craving and substance use in daily life: A systematic review’, Drug and Alcohol Dependence, 148, pp. 1–20. Available at: 10.1016/j.drugalcdep.2014.12.024. [DOI] [PubMed] [Google Scholar]
  36. Serre F et al. (2018) ‘Predictors of craving and substance use among patients with alcohol, tobacco, cannabis or opiate addictions: Commonalities and specificities across substances’, Addictive Behaviors, 83, pp. 123–129. Available at: 10.1016/j.addbeh.2018.01.041. [DOI] [PubMed] [Google Scholar]
  37. Sobell MB and Sobell LC (1995) ‘Controlled drinking after 25 years: how important was the great debate?’, Addiction, 90(9), pp. 1149–1153. Available at: 10.1080/09652149541392. [DOI] [PubMed] [Google Scholar]
  38. Trela CJ et al. (2018) ‘Moderation of alcohol craving reactivity to drinking-related contexts by individual differences in alcohol sensitivity: An ecological investigation.’, Experimental and Clinical Psychopharmacology, 26(4), pp. 354–365. Available at: 10.1037/pha0000206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Treloar Padovano H and Miranda R (2021) ‘Incubation of alcohol craving as it naturally occurs in a developmentally diverse sample of dependent and nondependent drinkers’, Addiction Biology, 26(3), p. e12934. Available at: 10.1111/adb.12934. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Waddell JT, Bartholow BD and Piasecki TM (2022) ‘Changes in affect and alcohol craving during naturally occurring drinking episodes: The role of day-level drinking motives’, Experimental and Clinical Psychopharmacology, p. No Pagination Specified-No Pagination Specified. Available at: 10.1037/pha0000600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wemm SE et al. (2022) ‘Daily stress predicts later drinking initiation via craving in heavier social drinkers: A prospective in-field daily diary study’, Journal of Psychopathology and Clinical Science, 131(7), pp. 780–792. Available at: 10.1037/abn0000771. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Witkiewitz K et al. (2017) ‘Clinical validation of reduced alcohol consumption after treatment for alcohol dependence using the World Health Organization risk drinking levels’, Alcoholism: Clinical and Experimental Research, 41(1), pp. 179–186. Available at: 10.1111/ACER.13272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Witkiewitz K et al. (2018) ‘Drinking risk level reductions associated with improvements in physical health and quality of life among individuals with alcohol use disorder’, Alcoholism: Clinical and Experimental Research, 42(12), pp. 2453–2465. Available at: 10.1111/acer.13897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Witkiewitz K, Roos CR, et al. (2019) ‘Advancing Precision Medicine for Alcohol Use Disorder: Replication and Extension of Reward Drinking as a Predictor of Naltrexone Response’, Alcoholism, Clinical and Experimental Research, 43(11), pp. 2395–2405. Available at: 10.1111/acer.14183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Witkiewitz K, Falk DE, et al. (2019) ‘Maintenance of World Health Organization Risk Drinking Level Reductions and Posttreatment Functioning Following a Large Alcohol Use Disorder Clinical Trial’, Alcohol: Clinical and Experimental Research, 43(5), pp. 979–987. Available at: 10.1111/acer.14018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Witkiewitz K et al. (2020) ‘World Health Organization Risk Drinking Level Reductions Are Associated with Improved Functioning and Are Sustained Among Patients with Mild, Moderate, and Severe Alcohol Dependence in Clinical Trials in the United States and United Kingdom’, Addiction (Abingdon, England), 115(9), pp. 1668–1680. Available at: 10.1111/add.15011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Witkiewitz K, Litten RZ and Leggio L (2019) ‘Advances in the science and treatment of alcohol use disorder.’, Science advances, 5(9), p. eaax4043. Available at: 10.1126/sciadv.aax4043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Xin Y, España M and Davis AK (2022) ‘The acceptability of non-abstinent treatment goals among clinical social workers in the United States’, Journal of Social Work Practice, 0(0), pp. 1–16. Available at: 10.1080/02650533.2022.2034768. [DOI] [Google Scholar]
  49. Yoon G et al. (2006) ‘Alcohol craving in outpatients with alcohol dependence: rate and clinical correlates’, Journal of Studies on Alcohol, 67(5), pp. 770–777. Available at: 10.15288/jsa.2006.67.770. [DOI] [PubMed] [Google Scholar]
  50. Zemore SE et al. (2009) ‘Gender, acculturation, and other barriers to alcohol treatment utilization among Latinos in three National Alcohol Surveys’, Journal of Substance Abuse Treatment, 36(4), pp. 446–456. Available at: 10.1016/j.jsat.2008.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo
NIHMS1954835-supplement-Supinfo.docx (46.4KB, docx)

RESOURCES