CONSORT indicates Consolidated Standards of Reporting Trials; PCP, primary care physician.
Percentage of individuals with either (A) no heavy drinking days or (B) no drinking days (complete abstinence) over the course of the study by medication and alcohol withdrawal level (AWS low or high). At baseline (data not shown), all study participants had drinking days and heavy drinking days, so in this depiction they would start at 0%. AWS indicates alcohol withdrawal symptoms.
eFigure. A) Modified Alcohol Withdrawal Symptom Checklist (Pittman et al. 2007) given to participants prior to randomization and B) number of subjects in each group reporting that symptom.
Data Sharing Statement
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Anton RF, Latham P, Voronin K, et al. Efficacy of Gabapentin for the Treatment of Alcohol Use Disorder in Patients With Alcohol Withdrawal Symptoms: A Randomized Clinical Trial. JAMA Intern Med. Published online March 09, 2020. doi:10.1001/jamainternmed.2020.0249
Is gabapentin efficacious in the treatment of alcohol use disorder in adults with a history of alcohol withdrawal symptoms?
In this randomized clinical trial, gabapentin compared with placebo significantly increased the number of people with total abstinence and reduced drinking. This effect was most significantly observed in those with greater pretreatment alcohol withdrawal symptoms—41% of participants with high alcohol withdrawal symptoms had total abstinence on gabapentin compared with 1% of participants in the placebo arm.
This study showed that gabapentin is efficacious in promoting abstinence and reducing drinking in individuals with alcohol use disorder and especially so in those with more alcohol withdrawal symptoms.
Although an estimated 30 million people meet criteria for alcohol use disorder (AUD), few receive appropriate pharmacotherapy. A more personalized, symptom-specific, approach might improve efficacy and acceptance.
To examine whether gabapentin would be useful in the treatment of AUD, especially in those with the most alcohol withdrawal symptoms.
Design, Setting, and Participants
This double-blind randomized clinical trial conducted between November 2014 and June 2018 evaluated gabapentin vs placebo in community-recruited participants screened and treated in an academic outpatient setting over a 16-week treatment period. A total of 145 treatment-seeking individuals who met Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) criteria for AUD and were not receiving other AUD intervention were screened, and 96 who also met recent alcohol withdrawal criteria were randomized to treatment after 3 abstinent days. Daily drinking was recorded, and percentage of disialo carbohydrate-deficient transferrin in the blood, a heavy drinking marker, was collected at baseline and monthly during treatment.
Gabapentin up to 1200 mg/d, orally, vs placebo along with 9 medical management visits (20 minutes each).
Main Outcomes and Measures
The percentage of individuals with no heavy drinking days and those with total abstinence were compared between treatment groups and further evaluated based on prestudy alcohol withdrawal symptoms.
Of 96 randomized individuals, 90 were evaluable (44 in the gabapentin arm and 46 in the placebo arm), with a mean (SD) age of 49.6 (10.1) years; 69 were men (77%) and 85 were white (94%). The evaluable participants had 83% baseline heavy drinking days (4 or more drinks/day for women, 5 or more for men) and met 4.5 alcohol withdrawal criteria from the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition). More gabapentin-treated individuals had no heavy drinking days (12 of 44 participants [27%]) compared with placebo (4 of 46 participants [9%]), a difference of 18.6% (95% CI, 3.1-34.1; P = .02; number needed to treat [NNT], 5.4), and more total abstinence (8 of 44 [18%]) compared with placebo (2 of 46 [4%]), a difference of 13.8% (95% CI, 1.0-26.7; P = .04; NNT, 6.2). The prestudy high–alcohol withdrawal group had positive gabapentin effects on no heavy drinking days (P < .02; NNT, 3.1) and total abstinence (P = .003; NNT, 2.7) compared with placebo, while within the low–alcohol withdrawal group, there were no significant differences. These findings were similar for other drinking variables, where gabapentin was more efficacious than placebo in the high–alcohol withdrawal group only. Gabapentin caused more dizziness, but this did not affect efficacy.
Conclusions and Relevance
These data, combined with others, suggest gabapentin might be most efficacious in people with AUD and a history of alcohol withdrawal symptoms. Future studies should evaluate sleep changes and mood during early recovery as mediators of gabapentin efficacy.
ClinicalTrials.gov Identifier: NCT02349477
Up to 30 million people in the United States meet criteria for alcohol use disorder (AUD), and the number is increasing over time,1 accounting for considerable morbidity and mortality.2 However, only 20% of those who might benefit from treatment receive it, and of those individuals, only 20% (or less than 1 million individuals) receive medication to help them with maintaining abstinence or reducing drinking. One reason is that available medications are not universally efficacious.3 However, there is reason to believe that some subgroups of patients might respond better to available treatments.3
One understudied phenotype or subtype of AUD are people who experience alcohol withdrawal syndrome (AWS).4 Although the quantity and frequency of drinking might predict who is at risk for AWS, there is considerable variation (some likely genetic) as to who will experience AWS on the cessation of drinking. More than half of inpatients with AUD5 and 35% of community individuals with AUD reported alcohol withdrawal symptoms, leading to a higher rate of alcohol problems at follow-up.6 Similarly, those who have undergone previous medicated detoxifications relapse quicker after cessation of alcohol use.7 While up to 30% of individuals with AUD presenting for clinical trials (who are similar to those seen in addiction outpatient clinics and in primary care) have acute AWS, it has been recognized that lower-grade AWS might be present, and last longer, in considerably more individuals. Many individuals also stop or reduce drinking prior to initiating treatment and therefore may experience mild to severe AWS without being formally diagnosed. In addition, a constellation of problems, such as irritability, anxiety, dysphoria, difficulties with concentration, and insomnia, might persist for a time after initial abstinence. Some have labeled these lingering symptoms as protracted withdrawal8 or protracted abstinence.9 In general, this constellation of symptoms and any desire to drink emanating from them are not likely to be influenced by antireinforcement or anticraving medications, such as naltrexone, which are likely to target more reward-based craving.10-14 Consistent with this notion, AWS is thought to be mediated primarily by γ-aminobutyric acid (GABA) and glutamate brain signaling15-17 in contrast to reward-based craving, in which opioid and dopamine brain signaling play a larger role.18,19 Therefore, medications that target brain GABA and glutamate brain signaling systems might be particularly useful in the treatment of AWS and, by extension, in those who have previously shown a biological propensity to experience AWS.16,20,21
Gabapentin has unique pharmacologic characteristics, binding to voltage-sensitive calcium channels at the α2δ-1 site,22 affecting their function23 as well as influencing receptor trafficking.24 Through these mechanisms, gabapentin is thought to secondarily influence GABA and glutamate tone/activity,25-27 which can be clinically measured.28
Gabapentin is efficacious for the treatment of acute alcohol withdrawal symptoms29,30 and also provides short-term relapse prevention after medicated alcohol detoxification,31 perhaps by an effect on sleep normalization.32,33 Post hoc analysis has shown effectiveness of treatment with gabapentin, in combination with flumazenil34 or naltrexone,35 in patients with current or historic alcohol withdrawal symptoms, consistent with basic science findings.36 Recent randomized clinical trials37,38 of gabapentin had mixed results but did not take alcohol withdrawal symptoms into account. Gabapentin is nevertheless worthy of further study given its earlier suggested effectiveness in those with alcohol withdrawal symptoms, its minimal cognitive effects,39,40 its lack of significant adverse interaction with alcohol,30,41,42 and its kidney excretion, which makes it potentially safer for individuals with liver disease.
The current study was a double-blind, placebo-controlled randomized clinical trial of treatment with gabapentin in outpatient individuals with AUD who reported current or historical alcohol withdrawal symptoms. A secondary aim was to evaluate the relationship of recently experienced alcohol withdrawal symptoms to gabapentin response, with the hypothesis that gabapentin would be more efficacious in those with higher levels of self-reported alcohol withdrawal symptoms.
The study was a 16-week randomized clinical trial (ClinicalTrials.gov identifier: NCT02349477) of gabapentin vs placebo. The trial protocol was approved by the Medical University of South Carolina Institutional Review Board for Human Research and is included in Supplement 1. Participants had to be aged 18 to 70 years; meet Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) (DSM-5) criteria for AUD, including alcohol withdrawal, as determined by Structured Clinical Interview for DSM-543; have a minimum of 5 drinks per day in the 90 days prior to assessment; and be abstinent at least 3 days prior to randomization as measured by breath analysis and urinary ethyl glucuronide testing. Participants could be using cannabis or other drugs but not meet criteria for drug use disorder, except nicotine, and could have no other psychoactive drug detected in the urine. Participants could not be taking psychotropic medications other than antidepressants (with the dose stable for at least 1 month) or meet current criteria for any major depressive disorder, bipolar disorder, psychotic disorder, or eating disorder. A history of posttraumatic stress disorder with stable symptoms was allowed, given its comorbidity with AUD and the utility of gabapentin in some anxiety disorders.44 Participants had to be medically stable (including liver enzymes alanine aminotransferase and aspartate aminotransferase less than 3 times the upper limit of normal). Women could not be pregnant or breastfeeding, must be using a reliable form of contraception, or must be postmenopausal. A history of alcohol withdrawal seizure or a Clinical Institute Withdrawal Assessment for Alcohol–Revised (CIWA-Ar) scale45 score of 10 or more during assessment was exclusionary. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.
Participants were recruited from November 2014 through June 2018 primarily by community advertisement and were not engaged in other alcohol treatment. Participants provided written informed consent approved by the Medical University of South Carolina Institutional Review Board for Human Research before formal assessment. Individuals paid $50 for attending the end-of-study assessment to provide drinking and biological data.
After at least 3 days of abstinence and assessment, participants were randomized (using prespecified computer random assignment by the investigational pharmacy) to receive gabapentin (day 1: 300 mg at bedtime; day 2: 300 mg in the morning and at bedtime; days 3 and 4: 300 mg in the morning, at noon, and at bedtime; and days 5 through 112: 300 mg in the morning and at noon and 600 mg at bedtime for a total of 1200 mg) or identical placebo capsules (given at the same quantity and times) in blister packs for 16 weeks. Study medications were identically overencapsulated with 25 mg of riboflavin (added to measure adherence) and distributed in labeled blister packs. Medical management, which consisted of a relatively brief (15-20 minutes) educational/supportive and adherence-enhancing clinical interaction,46,47 was conducted on weeks 1, 2, 3, 4, 6, 8, 10, 12, and 16 when adherence (pill count and urine riboflavin collection) and medication adverse effects, using the Systematic Assessment for Treatment of Emergent Events (SAFTEE) instrument,48 were assessed.
Prior to randomization, multiple assessments were administered, including the Structured Clinical Interview for DSM-5, the Alcohol Withdrawal Symptom Checklist (AWSC),49 the Alcohol Dependence Scale (ADS),50 the Obsessive Compulsive Drinking Scale (OCDS),51 Form 90 (alcohol consumption, drug use, daily calendar method),52 the CIWA-Ar scale,45 and baseline physical complaints. Laboratory tests included a health screen, liver function tests, pregnancy test (in women), and alcohol use markers γ-glutamyltransferase (GGT), urine ethyl glucuronide, and percentage of disialo carbohydrate-deficient transferrin (%dCDT).53,54 White blood cell DNA samples were obtained for potential future analysis.
During each medical management session, the calendar-based timeline follow-back method55 was used to assess daily drinking since the last visit. In addition, OCDS and SAFTEE assessments were performed at each visit. To confirm verbal report of abstinence or heavy drinking, %dCDT was collected on weeks 3, 6, 10, and 16.
The primary a priori defined drinking outcome (efficacy) measure was the percentage of participants with no heavy drinking days (defined as 5 or more drinks per day for men and 4 or more drinks per day for women; standard drinks [defined as containing 14 g of ethanol, or translated into typical drinks: 1.5 oz of spirits, 5 oz of wine, or 12 oz of beer]) throughout the study.38,56,57 The main secondary outcome measure was the percentage of participants with no drinking days (ie, total abstinence) throughout the study. Verbal report was confirmed by %dCDT testing and drinking status corrected—that is, those who reported no drinking or no heavy drinking but had a %dCDT greater than 1.7% at any time during the study were considered to have heavy drinking days and not be abstinent. This led to 3 individuals being reclassified prior to unblinding of medication group allocation.
Additional analyses included the relationship of the level of reported recent alcohol withdrawal symptoms on the AWSC49 and medication response. Other drinking variables routinely reported in AUD randomized clinical trials that might be useful to clinicians were also evaluated. These included the percentage of heavy drinking days, percentage of days abstinent, number of drinks per day, and number of drinks per drinking day.
Urine riboflavin was assayed in the Medical University of South Carolina Clinical Neurobiology Laboratory (directed by R.A.) using standard/calibration curves constructed from known amounts of riboflavin against which unknown amounts of urine riboflavin were calculated by fluorescence detection (448 nm excitation/510 nm emission). Values greater than 1300 ng/mL indicated adherence. The %dCDT was measured with a reference high-performance liquid chromatography assay.54 Using this international standardized assay, a value greater than 1.7% is close to 100% specific for sustained heavy drinking in the weeks prior to testing and can be used to corroborate or independently evaluate drinking in clinical trials.58 Urine ethyl glucuronide (Microgenics Diagnostics) and other blood chemistries, including GGT, were measured with an autoanalyzer.
Data from individuals with an alcohol withdrawal history in 2 prior AUD trials that included gabapentin treatment34,35 indicated the percentage of individuals with no heavy drinking days (success rates) to be between 38% and 41% in the placebo-treated groups and between 71% and 79% in the gabapentin-treated group, estimating the power to detect a gabapentin effect to be between 0.83 and 0.98 at α = .05 with a sample size of 45 individuals per medication group.
For no heavy drinking (primary outcome) and no drinking/total abstinence (secondary outcome), the number of evaluable individuals who met that criteria by verbally reported drinking only or by %dCDT verification over the whole 16-week trial was analyzed using a 2-sample z test to produce confidence intervals and a χ2P value; P values less than .05 were considered significant. In addition, the number needed to treat (NNT) or the number needed to harm (NNH)59 is reported. For analytic purposes, the 9 placebo-treated and 5 gabapentin-treated individuals that had missing drinking data were considered to be drinking or heavy drinking. A sensitivity analysis evaluating the same effects in only those who completed the study and were adherent with medication was also performed. All outcome data were collected by study staff and analyzed without knowledge of medication group assignment.
For evaluation of the level of alcohol withdrawal symptoms predicting gabapentin response, the prestudy AWSC (see eFigure in Supplement 2) was divided into low or high alcohol withdrawal based on median split. A χ2 test evaluating the medication within the different alcohol withdrawal groups on no heavy drinking and total abstinence was conducted for the total study period.
Evaluation of other drinking data was done using a linear mixed model with an unstructured variance/covariance matrix in which drinking parameters (percentage of heavy drinking days, percentage of drinking days, number of drinks per day, and number of drinks per drinking day) were evaluated over the 4 months of the study. These models measured the main effects of time (month), medication group, and alcohol withdrawal symptom group (low or high) and their potential interactions. In these models, missing data were assumed to be missing at random. Statistical analyses were performed using SPSS version 24 (IBM).
Participant recruitment and disposition are summarized in the CONSORT diagram (Figure 1). Essentially, 145 individuals meeting DSM-5 AUD criteria provided written informed consent. Of these, 96 individuals were randomized to placebo (n = 50) or gabapentin (n = 46). Four participants receiving placebo and 2 receiving gabapentin had no valid follow-up data or an early protocol violation (disclosure of an excluded medical disease that was withheld at screening). Eighteen of 46 participants (39%) receiving placebo and 13 of 44 participants (30%) receiving gabapentin did not complete treatment (reasons given in Figure 1). Participants in both groups attended an average of 7 of 9 medical management sessions. Table 1 summarizes the study entry demographic information and drinking variables. Of the 90 participants included in the final analysis, the mean (SD) age was 49.6 (10.1) years; 69 (77%) were men; 85 (94%) were white; 63 (70%) were employed; and 39 (43%) used nicotine products. They drank a mean of 86% of pretreatment days, with 83% being heavy drinking days, consuming 13 drinks per drinking day, and 25 participants (28%) had past treatments and 12 (13%) had undergone past medical detoxifications. They met a mean of 4.5 of 8 possible DSM-5 criteria for alcohol withdrawal. Sixty-four participants (71%) had elevated %dCDT and 68 (76%) had elevated GGT, both markers of heavy drinking. There were no important significant differences in any prestudy variables between the placebo and gabapentin groups.
Data for the primary outcome variable, the percentage of individuals with no heavy drinking days, and the secondary outcome variable, the percentage of individuals with no drinking days, are summarized in Table 2. For no heavy drinking, using verbal report only, gabapentin (12 of 44 participants [27%]) was superior to placebo (6 of 46 participants [13%]) at the trend level (14.2% difference; 95% CI, −2.1 to 30.6; P = .09; NNT, 7.0), but when verbal report was confirmed by the highly specific %dCDT heavy drinking marker,60,61 the difference was statistically significant (18.6% difference; 95% CI, 3.1-34.1; P = .02; NNT, 5.4). For the secondary variable (no drinking days/total abstinence), when using verbal report only, gabapentin treatment (9 of 44 participants [21%]) was superior to placebo (2 of 46 participants [4%]) (16.1% difference; 95% CI, 2.8-29.4; P = .02; NNT, 6.2) and in the same direction after %dCDT confirmation (13.8% difference; 95% CI, 1.0-26.7; P = .04; NNT, 7.2).
Performing a sensitivity analysis on the 21 placebo-treated and 26 gabapentin-treated participants who completed the study (had all 16-week drinking data) and who adhered to pill taking (at least 75% riboflavin-positive urine samples), there were 12 gabapentin-treated individuals compared with 1 placebo-treated individual with no heavy drinking days (χ21 = 5.08; P = .02; NNT, 2.4) and 8 gabapentin-treated individuals compared with 1 placebo-treated individual who remained abstinent throughout the study (χ21 = 9.95; P = .002; NNT, 3.9). These results were similar or stronger when drinking status was adjusted based on within-treatment %dCDT levels.
Because the main hypothesis of this study was that gabapentin would be more efficacious in individuals with AUD with more alcohol withdrawal, as suggested in past work,29,30 and because ethical and feasibility issues limited the ability to enroll people in acute withdrawal (CIWA-Ar scores greater than 10), we undertook an additional analysis of the level of self-reported alcohol withdrawal symptoms when participants reduced or stopped drinking in the 2 weeks before randomization using the modified AWSC (eFigure in Supplement 2). Figure 2 shows the percentage of individuals with heavy drinking days and any drinking days in the medication groups predicated by prestudy alcohol withdrawal scores (median split greater or less than 8.5) over the study period. Those with high alcohol withdrawal scores had less relapse to heavy drinking (χ21 = 5.75; P = .02; NNT, 3.1) and more total abstinence (χ21 = 8.69; P = .003; NNT, 2.7) when treated with gabapentin (10 of 22 [46%] and 9 of 22 [41%], respectively) compared with placebo (3 of 23 [13%] and 1 of 23 [4%]), while those with low alcohol withdrawal scores had similar relapse to heavy drinking (χ21 = 0.18; P = .67; NNH, 25.3) and similar total abstinence (χ21 = 0.98; P = .32; NNH, 23.0) when treated with gabapentin (2 of 22 [9%] and 0 of 22 [0%], respectively) compared with placebo (3 of 23 [13%] and 1 of 23 [4%]). The same pattern was found when the data were corrected for %dCDT.
To evaluate gabapentin efficacy across a range of other drinking variables, we conducted sensitivity analyses taking AWSC score groups (as above) into account in a linear mixed model (Table 3). Although there was a main effect of time, it did not significantly interact with medication group or AWSC scores; thus, time interaction terms were dropped from the model. In this analysis, there was no main effect of gabapentin over placebo on any drinking variable, but there were significant interactions with AWSC scores such that in the high AWSC group, gabapentin compared with placebo reduced the percentage of heavy drinking days (F1,86.802 = 4.53; P = .04) and promoted more abstinence days (F1,86.419 = 5.68; P = .02) but did not decrease drinks per day (F1,85.252 = 3.40; P = .07) or drinks per drinking day (F1,71.659 = 0.14; P = .71).
Adverse effects as recorded with the SAFTEE interview at each study visit showed significantly more gabapentin-treated (n = 25) vs placebo-treated (n = 15) individuals reported mild to moderate dizziness (χ21 = 5.34; P = .02), but in a separate analysis, the presence or absence of dizziness did not significantly account for gabapentin’s effectiveness. While there was no other significant between-group differences in unique individuals reporting other symptoms, there were more within-individual reports of nervousness (χ21 = 10.62; P = .001) and headache (χ21 = 9.70; P = .002) in the gabapentin-treated group (none severe), and more reports of insomnia (χ21 = 10.81; P = .001) in the placebo-treated group (1 severe).
These results add to the growing literature on the use of gabapentin for the treatment of AUD. While some previous studies37 have shown efficacy, others have not,38 and a recent meta-analysis62 suggested weak support for its efficacy in treatment of AUD. Although multiple reasons might contribute to the variability of results, an important issue might be the individual’s alcohol withdrawal status/history.34,35 In the current study, we chose to explicitly study the hypothesis that a history of alcohol withdrawal symptoms would influence gabapentin efficacy by evaluating gabapentin only in those who met current or historical criteria for DSM-5 alcohol withdrawal. We further used a previously validated alcohol withdrawal self-rating instrument49 to gauge the relationship between the level of alcohol withdrawal symptoms in the weeks prior to randomization and gabapentin response.
In essence, we found gabapentin to be efficacious in preventing relapse to heavy drinking (NNT, 5.4) and, perhaps more importantly, in promoting abstinence (NNT, 7.2). These are conservative outcomes for AUD randomized clinical trials,56 but both are accepted by the US Food and Drug Administration as indicators of medication efficacy in AUD randomized clinical trials.57 However, when taking the amount of alcohol withdrawal symptoms into account, significant gabapentin effects were seen only in those with the higher levels of self-reported alcohol withdrawal symptoms. There was less relapse to heavy drinking (NNT, 3.1) and more total abstinence (NNT, 2.7) in those with more intense alcohol withdrawal symptoms treated with gabapentin, and we also found an interaction between the level of those symptoms on other continuous drinking outcomes, including percentage of heavy drinking days, percentage of days abstinent, and to some extent the number of drinks per day—all favoring gabapentin treatment over placebo. In fact, in our previous studies, we found more abstinence when gabapentin was combined with flumazenil34 compared with placebo and less relapse to heavy drinking when gabapentin was combined with naltrexone compared with naltrexone alone.35 In both studies, individuals with a greater history of alcohol withdrawal showed the most efficacy on those drinking outcomes. It is also of interest that Mason et al37 reported that gabapentin decreased relapse to heavy drinking and increased total abstinence in a less severe population, based on alcohol consumption levels, and without significant current alcohol withdrawal symptoms. However, in a meta-analysis62 in which the alcohol withdrawal status of the populations was not considered, the percentage of heavy drinking days was the only variable that showed superior gabapentin efficacy. In a larger clinical context, the NNT for gabapentin efficacy was equal to or better than reported for naltrexone in the COMBINE Study (NNT, 7),47 and with alcohol withdrawal status taken into account, gabapentin benefited even more people (NNT, 3).
Many studies have supported the efficacy of gabapentin in alcohol withdrawal treatment.44,63 Given gabapentin’s direct and indirect pharmacologic effects on brain GABA and glutamate systems15,17 and the role of these systems in alcohol withdrawal and relapse drinking,16,64 its efficacy makes sense. Of additional biological interest, the gene regulating the voltage-sensitive calcium channel α2δ-1 protein is upregulated by chronic alcohol exposure and withdrawal.65 Because this is the putative site of gabapentin action,23 this site might be a fruitful target for potential future AUD pharmacologic treatment development as well.
This study had several limitations. Although the noncompletion rate was similar to that for other AUD gabapentin treatment trials37,38 and a National Institutes of Health National Institute on Alcohol Abuse and Alcoholism–sponsored large multisite study,47 it should be noted that 13 of 44 (30%) of individuals on gabapentin and 18 of 46 (39%) on placebo did not complete the trial. Perhaps, adding other supportive counseling or Alcoholics Anonymous attendance could increase retention in treatment. Also, self-reported alcohol withdrawal symptoms prior to study entry might not fully capture the extent of withdrawal severity. In addition, those with complex psychiatric and medical conditions, including history of alcohol withdrawal seizures, were excluded. Furthermore, given its kidney excretion, gabapentin should be studied in patients with AUD with more severe liver disease, a condition in which medications are greatly needed.66
The weight of the evidence now suggests that gabapentin might be most efficacious after the initiation of abstinence to sustain it and that it might work best in those with a history of more severe alcohol withdrawal symptoms. To further confirm this, future studies should specifically evaluate symptoms related to protracted alcohol withdrawal during gabapentin treatment. Armed with this knowledge, clinicians may have another alternative when choosing a medication to treat AUD and thereby encourage more patient participation in treatment with enhanced expectation of success.
Accepted for Publication: January 23, 2020.
Corresponding Author: Raymond F. Anton, MD, Addiction Sciences Division, Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, 67 President St, MSC 861, Charleston, SC 29425 (email@example.com).
Published Online: March 9, 2020. doi:10.1001/jamainternmed.2020.0249
Author Contributions: Drs Anton and Hoffman had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Anton.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Anton, Book, Bristol.
Critical revision of the manuscript for important intellectual content: Anton, Latham, Voronin, Book, Hoffman, Prisciandaro.
Statistical analysis: Anton, Hoffman, Prisciandaro.
Obtained funding: Anton.
Administrative, technical, or material support: Anton, Latham, Voronin, Book, Prisciandaro, Bristol.
Study supervision: Anton, Book, Prisciandaro.
Conflict of Interest Disclosures: Dr Anton reported receiving grants from the National Institute on Alcohol Abuse and Alcoholism during the conduct of the study; receiving grants and consulting fees from Laboratorio Farmaceutico CT; receiving consulting fees from Alkermes, Allergan, Indivior, Insys, Life Epigenetics, XenoPort (Arbor Pharmaceuticals), and Alcohol Clinical Trials Initiative (ACTIVE) outside the submitted work; and serving as a member and chair of ACTIVE, which was partially supported by Alkermes, Amygdala Neurosciences, Arbor Pharmaceuticals, Ethypharm, Indivior, Lundbeck, Mitsubishi, and Otsuka. Dr Book reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Prisciandaro reported receiving grants and consulting fees from Farmaceutico Italiano and consulting fees from Laboratorio Farmaceutico CT outside the submitted work. No other disclosures were reported.
Funding/Support: Funding for this work was provided by the National Institute on Alcohol Abuse and Alcoholism (grant R01AA022364).
Role of the Funder/Sponsor: The National Institute on Alcohol Abuse and Alcoholism had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Data Sharing Statement: See Supplement 3.
Additional Contributions: Mark Ghent, BA, Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, assisted in data collection, and Patrick Randall, PhD, Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, assisted in initial power analyses and analytic planning and were compensated for their work.
Create a personal account or sign in to: