Randomization was stratified by sex and participant report of experiences with withdrawal-related dysphoria. This a-priori stratification variable was intended to capture the “dark side of addiction” , whereby individuals reporting withdrawal-related dysphoria were estimated to experience greater dysfunction of the immune system. MediciNova, Inc. supplied ibudilast and placebo for the trial but did not provide any financial support for the study. The UCLA Research pharmacy prepared and dispensed all study medication in blister packs. Research staff, participants, and providers remained blind to medication condition during the trial. Participants were titrated on ibudilast as follows: 20 mg BID during days 1–2 and 50 mg BID during days 3–14. Target medication dose was selected based on safety considerations as well as preclinical and clinical data . Medication compliance was monitored through pill counts and self-report via DDA. Side effects were closely monitored and reviewed by study physicians. During the in-person screening visit, participants completed a set of assessments for individual differences and eligibility screening. Assessments included collection of demographic information , substance use characteristics and history, and psychological functioning and diagnoses. Surveys used to characterize the sample included the Beck Depression Inventory to assess levels of depression symptomatology, the Snaith-Hamilton Pleasure Scale to measure anhedonia , the Alcohol Use Disorders Identification Test to capture alcohol problem severity, Penn Alcohol Craving Scale to measure tonic craving levels, and the Reasons for Heavy Drinking Questionnaire to capture one’s motivations for heavy drinking. In addition, the RHDQ determined the presence of withdrawal-related dysphoria for randomization stratification as follows: raw scores ranging from 0 – 10 on the RHDQ question #6: “I drink because when I stop, I feel bad ”, were dichotomized into yes /no, based on a cut-off of 6+ points. Interviews used to determine eligibility criteria and determine baseline quantity and frequency of alcohol use were administered by clinical graduate students or trained research staff and included the Timeline FollowBack measuring alcohol, cigarette,hydroponic drain table and cannabis use over the previous 30 days , the Clinical Institute Withdrawal Assessment for Alcohol Scale – Revised assessing clinically significant alcohol withdrawal, and the Structured Clinical Interview for DSM-5 to determine current AUD diagnosis and severity and to screen for exclusionary psychiatric diagnoses.

Each morning throughout the two-week trial, participants were asked to retrospectively report on their previous day experiences by completing an electronic DDA survey . Study staff provided instructions on DDA completion and participants practiced filling out the survey at their randomization visit. Daily text messages or emails containing links to DDAs were sent to participants at 8am each morning during their 14-day medication period. Additional telephone or text reminders were sent by study staff as needed. At the start of each daily survey, participants were asked, “Did you drink any alcohol yesterday?” If participants endorsed alcohol use the previous day, they reported on drink type and quantity, and then completed two sets of items: 1) ratings of mood, craving, and urge before drinking, and 2) ratings of mood, craving, urge, stimulation, and sedation while drinking. For example, participants were asked: “Before you drank, how strong was your urge to drink alcohol yesterday?” and “While drinking, how strong was your urge to drink alcohol yesterday?” The current analyses focus primarily on drinking days, given our interest in medication-related changes in subjective response to alcohol. Mood states were assessed via the short form of the Profile of Mood States survey . POMS-SF is a standard, validated psychological rating scale that measures dimensions of transient mood states by asking subjects to indicate how well each item describes their mood on a 5-point Likert scale . To keep the survey brief and thus reduce the burden on participants, only select items from POMS-SF were chosen for DDAs . Reports of stimulation and sedation were assessed via the validated Brief Biphasic Alcohol Effects Scale . The B-BAES is a six-item measure on the acute stimulant and sedative effects of alcohol on an 11-point scale . Urge to drink was captured via the item, “How strong was your urge to drink alcohol yesterday” , in line with previously published reports . Phasic cravingwas assessed using the first and last items from the validated Alcohol Urge Questionnaire on a 7-point Likert scale . Participants reported the quantity of standard alcoholic drinks consumed according to established guidelines and provided details about non-standard drinks . Drink entries were reviewed and verified by study staff.

DDA Item Scoring—All descriptive and statistical analyses were completed in SAS Version 9.4 on the sample of participants who completed at least one DDA. Select items from the POMS-SF tension and depression subscales , were summed to form a negative mood state score and select items from the vigor subscale were summed to form a positive mood state score for each timepoint, consistent with previous reports . The two AUQ items were summed to form a craving score . Stimulation and sedation subscales from B-BAES were calculated using standard methods . Multilevel Models—Models were fit in SAS using the MIXED procedure and a multilevel framework, unstructured covariance matrix, residual maximum likelihood estimation, and random intercepts with observations nested within subjects to account for clustering and to preserve suitable Type-1 error rates . Kenward-Rogers degrees of freedom were chosen to reduce bias and obtain more accurate p-value estimates. Main and simple effects were probed by recentering dichotomous variables and using the simple slopes approach. Daily alcohol use quantity, mood states, craving, and urge data from non-drinking days were treated as missing. Comparable three-level models were fit for variables having both before and during drinking observations , such that these observations were nested within day and days were nested within subjects. All models were tested with the following level-2 covariates: sex, AUD severity , and baseline drinks per drinking day . In addition, daily number of drinks consumed during the trial was included as a predictor with random effect in all subjective response models to account and control for potential day-level drink quantity effects on subjective response. To examine both between- and within-subject effects and interactions, covariates were centered at the grand mean and focal within-subject variables were centered within cluster . Specifically, to assess the effect of medication on the acute stimulant and sedative effects of alcohol, one model for each B-BAES subscale was estimated in which stimulation or sedation served as the outcome and medication condition served as the focal predictor.

To assess the effect of medication on alcohol-induced changes in mood and craving, three-level models were run for each positive mood, negative mood, craving, and urge scores, as predicted by medication condition, time and a medication × time interaction.Two sets of exploratory analyses were conducted. First, to explore how medication effects might impact drinking outcomes, we tested whether ibudilast moderated the effect of stimulation/ sedation on same-day drinking during the trial, given support for these variables as strong predictors of alcohol use . As such, a within-subject cross-level interaction of medication × stimulation or sedation was added with random slopes, and same-day number of drinks served as the outcome. In a similar fashion,rolling benches hydroponics we also tested whether ibudilast moderated the effect on stimulation/ sedation on next-day drinking using cross-lagged logistic models; this analysis served to test whether subjective response predicted future drinking behaviors. Second, given the trial’s a priori interest in a withdrawal-related dysphoria characteristic, we tested whether dysphoria would moderate ibudilast’s effects on alcohol-induced changes in mood and craving. A three-way interaction was added to models estimating the outcomes- positive mood, negative mood, urge, and craving . Stimulation and sedation variables were limited to a single time point and were thus excluded from analyses testing before to during drinking changes. The final sample of randomized participants who completed at least one DDA, consisted of 50 non-treatment seeking individuals with current AUD . Overall, 66% of the sample reported their sex as male, 68% reported an annual household income < $60,000, and the average age was 32.7 years . Regarding race, participants most frequently identified as White , followed by 14% Black or African American, and 12% mixed race. In addition, 24% of the sample identified as Hispanic/ Latinx. Participants had an average of 5.6 DPDD in the month prior to their baseline visit. Medication adherence was high, as both medication groups exceeded 97% adherence rates.In this secondary analysis, we tested bio-behavioral mechanisms of ibudilast, a neuroimmune modulator, through naturalistic daily reporting of subjective response to alcohol collected during a two-week RCT enrolling 50 non-treatment seeking participants with AUD. Electronic DDAs were administered each morning to participants to capture their previous day drinking behaviors and subjective alcohol response measures. First, we were interested in understanding whether ibudilast altered average levels of stimulation and sedation during drinking episodes. Results showed that ibudilast treatment did not significantly change average levels of stimulation nor sedation during the trial compared with placebo.

These findings are consistent with an initial safety trial in which ibudilast did not significantly affect any subjective response variables during an experimentally controlled alcohol infusion in the laboratory . Relatedly, a trial combining laboratory and EMA methods showed that topiramate reduced drinking-related craving but not the stimulant or sedative effects of alcohol . However, animal literature shows that apremilast, another PDE inhibitor, did alter a wide range of ethanol-induced effects in mice, such as reducing acute functional tolerance and increasing the sedative, intoxicating effects, and aversive properties of ethanol . Perhaps unlike certain pharmacotherapies for AUD such as naltrexone, neuroimmune modulators, like ibudilast may not reduce drinking by robustly suppressing alcohol’s stimulant properties or amplifying its sedative effects. Rather, ibudilast may more directly alter other central mechanisms like alcohol craving or may exert a wider range of effects on multiple mechanisms that cumulatively impact drinking outcomes. Second, we tested a related exploratory aim examining the moderating effect of ibudilast on alcohol-related stimulation and sedation and same-day number of drinks consumed. Participants on ibudilast reported a significant, positive relationship between their stimulation and sedation ratings and same-day drinking levels, neither of which was observed in the placebo condition. This suggests that participants randomized to ibudilast consumed more alcohol on days when they retrospectively reported feeling more stimulated during a drinking episode than on days when they felt less stimulated . Yet for those on placebo, we did not detect a significant relationship between one’s feelings of stimulation or sedation and alcohol use. These findings are consistent with EMA data showing that naltrexone potentiated participant’s subjective “high” across rising levels of estimated BrAC . Similarly, topiramate was shown to strengthen2021. These results are also in line with a secondary analysis of our lab’s initial efficacy trial, whereby ibudilast potentiated the association between mood states and one’s craving for alcohol following a stress exposure paradigm compared with placebo . Mechanistically, PDE4 inhibitors attenuate alcohol-induced neuroimmune activation and dysregulation of GABAergic signaling . These important processes are connected to behavioral responses to ethanol . Thus, micro-longitudinal reports collected during the current trial helped to elucidate dynamic, day-to-day associations between within-person subjective effects and drinking, such that ibudilast seemed to moderate these relationships for a given individual, rather than by altering average subjective response levels across participants. For our second primary aim, we assessed whether ibudilast, compared with placebo, attenuated daily alcohol-induced changes in positive mood, negative mood, urge, and craving . Among the full sample, we found that ibudilast significantly dampened within-person alcohol-induced increases in craving seen under the placebo condition, but not other subjective response indicators. This suggests that one of the mechanisms by which ibudilast exerts its effects on drinking outcomes, such as reductions in heavy drinking, may be by diminishing one’s desire to continue drinking during an episode. Considering its immunomodulatory actions, ibudilast may reduce the acute and chronic proinflammatory effects of alcohol, either indirectly through suppression of peripheral inflammation or directly by altering cAMP signaling pathways and suppressing cytokine expression and in the brain . In return, acute alcohol-induced increases in craving are blunted. Supporting these findings is research on methamphetamine use disorder . An RCT for inpatients with MUD showed that ibudilast significantly blunted the rewarding effects of methamphetamine during an infusion in the laboratory and similarly diminished drug-induced increases in proinflammatory levels during infusion . Continuing, previous results from our group implicate ibudilast in the reduction of tonic craving and neural alcohol-cue reactivity, as evidenced by attenuation of cue-elicited activation in the ventral striatum compared with placebo . It is thus plausible that reductions in alcohol craving and reward, across these contexts, represent a primary mechanism of action of ibudilast for AUD.