Improvement of Physical Health and Quality of Life of Alcohol-Dependent Individuals With Topiramate Treatment:  US Multisite Randomized Controlled Trial FREE

Alcoholism is a common disease.¹ Notwithstanding the primary aberration in drinking behavior, its extensive secondary physical and psychosocial consequences² are what make the burden of the disease so devastating. This realization is enshrined in practice in that no commonly used diagnostic scheme for alcoholism or alcohol dependence includes a quantification of drinking behavior—only its physical and psychosocial sequelae. Epidemiologically, these sequelae of alcoholism are manifested in the statistic that alcoholism as a risk factor ranks fifth on the global burden of disease.³

Topiramate, a sulfamate-substituted fructopyranose derivative, has been shown to decrease the primary aberration of excessive drinking by reducing the reinforcing effects of alcohol,⁴ presumably through 2 pharmacologic mechanisms: the contemporaneous facilitation of γ-aminobutyric acid function through a nonbenzodiazepine site on the γ-aminobutyric acid–A receptor⁵ and the antagonism of glutamate activity at α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid and kainate receptors.⁴

Indeed, the results of 2 recent clinical trials have demonstrated topiramate's efficacy at improving drinking outcomes. First, in a single-site, double-blind randomized controlled trial, topiramate was shown to be more efficacious than placebo at decreasing heavy drinking and promoting abstinence from alcohol.⁶ Second, in a subsequent 17-site, double-blind randomized controlled trial, topiramate again was shown to improve drinking outcomes and promote abstinence.⁷ In both the single-site and multisite trials, topiramate's estimated therapeutic effect size at reducing heavy drinking was in the medium range (0.61 and 0.52, respectively), and in the multisite trial, the number of participants who would need to be treated with topiramate to have 1 additional success compared with placebo in reducing heavy drinking days was 3.4.⁸ Thus, topiramate appears to be a promising medication for the treatment of alcohol dependence.

Nevertheless, to establish topiramate as being generally effective for the treatment of alcohol dependence would require a demonstration that it can reduce or attenuate the important physical and psychosocial consequences that are common among those with a pattern of severe, chronic, and pervasive alcohol consumption.

Physically, perhaps the best known sequelae of alcohol dependence seen in general medical practice include abnormalities of liver function (eg, elevated liver enzyme levels, fatty liver disease, and eventually liver cirrhosis) and cardiovascular disease due to hypertension and elevated cholesterol levels.⁹

Psychosocially, the sequelae of alcohol dependence include persistent obsessional thoughts and compulsions about using alcohol that can maintain or provoke relapse to heavy drinking¹⁰ and increased social disenfranchisement due to reduced capability to maintain school, work, and interpersonal relationships. All of these sequelae of alcoholism contribute to decreased quality of life.¹¹

Herein, we seek to determine whether topiramate is generally effective as a treatment for alcohol dependence that not only improves the “symptom” of drinking but also alleviates its important physical and psychosocial consequences.

We enrolled 371 men and women who had been diagnosed as having alcohol dependence according to the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) (DSM-IV).¹² Subjects were recruited across 17 sites in the United States between January 27, 2004, and August 4, 2006, by newspaper, radio, and television advertisements.

We included male and female subjects between the ages of 18 and 65 years who drank 35 or more (men) and 28 or more (women) standard drinks per week, as measured by the timeline follow-back method,¹³ during the 28-day period before the screening visit to assess study eligibility and during the 7-day period between the screening visit and randomization. One standard drink was defined as 14.8 mL of absolute alcohol, equivalent to 295.7 mL of beer, 118.3 mL of wine, or 29.6 mL of 100-proof liquor. Also, subjects had to have negative results of urine toxicologic screening for opioids, cocaine, amphetamines, antidepressants, propoxyphenes, and barbiturates at the time of randomization and before the beginning of the double-blind period in week 0. Participants with positive results of the urine drug screening for tetrahydrocannabinol or benzodiazepines in the week before randomization (week −1) could be enrolled if they had negative results of urine drug screening on retesting 7 days later and met all other enrollment criteria. Additional inclusion criteria were as described previously.⁷

We excluded subjects who had a current Axis I psychiatric diagnosis according to the DSM-IV other than alcohol, nicotine, or caffeine dependence; had a history in the past 6 months of substance abuse or dependence, not including dependence on alcohol, nicotine, or caffeine; had clinically significant alcohol withdrawal symptoms (revised Clinical Institute Withdrawal Assessment for Alcohol scale¹⁴ score > 10); had made more than 4 unsuccessful formal inpatient treatment attempts to curb alcohol dependence; had clinically significant depression, which was defined as a score of greater than 24 on the Montgomery-Asberg Depression Rating Scale¹⁵ or based on the impression of a study physician; were receiving treatment for alcohol dependence other than Alcoholics Anonymous; had been compelled to receive treatment for alcohol dependence to avoid imprisonment, parole, probation, or loss of employment; or were from the same household as another study participant. Additional exclusion criteria were as described previously.⁷

Ethical approval for this study was provided by the institutional review boards of all 17 participating sites.

At screening, the beginning of week −1 (approximately 7 days before randomization), after obtaining written informed consent, we assessed participants on medical eligibility to meet the study inclusion criteria. Subjects returned for a separate visit at the end of week −1 to complete the remainder of the screening process, which included assessment of physical health as measured by physical examination, electrocardiogram, vital signs (ie, blood pressure, pulse, and temperature), body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared), hematologic and biochemical screens (including plasma cholesterol), urine tests (including urine drug screen), and a urine pregnancy test for women of childbearing potential; and personal and psychosocial harm from alcohol as measured by the Alcohol Use Disorders Identification Test.¹⁶ Subjects were not allowed to provide written informed consent unless they had a breath alcohol concentration less than 0.02 g%.

During the double-blind period from weeks 0 to 14, subjects were assessed weekly on measures of vital signs (ie, blood pressure, pulse, and temperature) and BMI. At weeks 0, 4, 8, 12, and 14, we assessed overall clinical improvement as measured by Clinical Global Impression scales for improvement and severity (CGI-I and CGI-S, respectively)¹⁷; obsessional thoughts and compulsions about using alcohol on the Obsessive Compulsive Drinking Scale (OCDS)¹⁸; and liver enzyme levels (alanine aminotransferase, aspartate aminotransferase, and γ-glutamyltransferase) to provide an objective biochemical measure of alcohol consumption.¹⁹ Plasma γ-glutamyltransferase level is well validated and accepted widely as a biomarker of alcohol consumption. Our use of plasma γ-glutamyltransferase as a biochemical measure of transient alcohol consumption in this study also was designed to mirror what Johnson et al⁶ did in the previous single-site pharmacotherapy trial of topiramate for treating alcohol dependence. Hematologic and biochemical screens, including urine pregnancy tests, were repeated at weeks 0, 4, 8, 10, 12, and 14. At weeks 0, 6, 10, and 14 we assessed the harmful consequences of drinking (Drinker Inventory of Consequences scale, Recent Consequences form [DrInC-2R]),²⁰ general mood (Profile of Mood States),²¹ sleep quality (Medical Outcomes Study Sleep scale),²² and quality of life (Quality of Life Enjoyment and Satisfaction Questionnaire [Q-LES-Q]).²³ Additional health and safety measures included assessments of adverse events, concomitant medications, and breath alcohol concentration on a weekly basis and a second electrocardiogram and physical examination at the end of the trial. Subjects were not allowed to complete assessments during the double-blind period unless they had a breath alcohol concentration less than 0.02 g%.

Study procedures were as detailed previously.⁷ Briefly, subjects who continued to meet drinking eligibility criteria after the 7-day screening period were assigned at random in a 1:1 ratio to receive topiramate or placebo. Study medication was dispensed in double-blind fashion for the efficacy determination period from the beginning of week 0 to the beginning of week 14. The medication dose was titrated from the beginning of week 0 to the beginning of week 6 and then maintained until the beginning of week 14 (Table 1). All subjects had to achieve a minimum topiramate dosage of 50 mg/d or the placebo equivalent to remain in the trial. From weeks 14 to 16, study medication was tapered as a safety precaution. Topiramate and matching placebo tablets were provided by Ortho-McNeil Janssen Scientific Affairs LLC (Raritan, New Jersey).

All subjects received Brief Behavioral Compliance Enhancement Treatment (BBCET) as their psychosocial treatment. The BBCET program, a standardized, brief (delivered in about 15 minutes) psychosocial adherence enhancement procedure, emphasized that medication adherence is crucial to changing the drinking behavior of alcohol-dependent individuals.²⁴ Brief interventions,²⁵ such as BBCET, have been shown to be beneficial in treating alcohol dependence. The BBCET program was modeled on the clinical management condition in the US National Institute of Mental Health collaborative depression trial, which was used as an adjunct to the medication condition for that study.²⁶ Also, it was used successfully as the psychosocial treatment platform in the previous efficacy trial of topiramate for the treatment of alcohol dependence.⁶ Trained clinicians delivered manual-guided BBCET during each week of the double-blind period as described previously.⁷

Outcome measures of physical health included liver function tests (ie, plasma aspartate aminotransferase, alanine aminotransferase, and γ-glutamyltransferase), hematologic and biochemical measures (including plasma cholesterol and bicarbonate and urine pH levels), vital signs (ie, blood pressure, pulse, and temperature), and BMI.

Outcome measures of psychosocial function included obsessional thoughts and compulsions about using alcohol (OCDS), overall clinical improvement (CGI-I and CGI-S scales), harmful consequences of drinking (DrInC-2R), quality of life (Q-LES-Q), general mood (Profile of Mood States), and sleep quality (Medical Outcomes Study Sleep scale).

Additional health and safety measures included measurements of adverse events, concomitant medications, withdrawal symptoms, urine pregnancy tests, and physical examinations performed periodically during the trial. An independent data monitoring committee met periodically to ensure the safety of the study subjects.

We managed the data according to the International Conference on Harmonisation guidelines of good clinical practice. Individual plots were checked for unusual values and completeness. Efficacy values were validated as correct against case records. For all statistical tests, differences between the treatment groups were accepted as significant if they achieved the 5% level with 2-tailed tests. All efficacy analyses were conducted on the intent-to-treat population, which included all randomized subjects who returned for at least 1 double-blind visit and received at least 1 dose of the study medication. All safety analyses were conducted on the evaluable-for-safety population, which included all randomized subjects who took at least 1 dose of the study medication. Data were analyzed with SAS statistical software, version 9.1 (SAS Institute Inc, Cary, North Carolina).²⁷

The duration of the double-blind phase (ie, weeks 0 to 14) was calculated as the date of the first double-blind dose plus 1 day to the date when the last double-blind dose of medication was taken but before the 2-week period during which the study medication was tapered.

The null hypothesis of the study was that there would be no difference between topiramate and placebo on target physical and psychosocial outcome measures associated with alcohol dependence during the double-blind phase. The primary analytic approach was to conduct inferential analysis on all randomized participants returning for at least 1 double-blind visit and receiving at least 1 medication dose. Unless stated otherwise, we analyzed the difference between the treatment groups by means of a repeated-measures mixed model (SAS PROCMIXED; SAS Institute Inc) with treatment, center, week, baseline measure of the outcome variable, and treatment × week interaction as covariates. An unstructured covariance matrix was used to model the correlations between repeated measurements within subjects. Least-squares mean estimators and the 95% confidence intervals of the means were derived for each treatment group. Ninety-five percent confidence intervals also were calculated for the mean difference between treatment groups.

We used a fixed-sequence, multiple-testing procedure to control for type 1 error when determining the earliest time point at which the difference between treatment groups on an outcome measure became significant statistically (based on the repeated-measures model) and was sustained for the subsequent time points. The basis of the procedure was to test the between-groups difference on the outcome measure at week 14 at the .05 significance level (2-tailed). If there was a significant difference at week 14, this procedure would be repeated for preceding weeks until a time point was reached at which there was no difference. A finding of no significant difference at week 14 would have stopped the comparison for the preceding weeks.

While all the outcome measures of physical health were analyzed as outlined above, plasma γ-glutamyltransferase was examined as its incremental log ratio over time as done in the previous study.⁶

Obsessional thoughts and compulsions about using alcohol, as assessed with the OCDS, was measured on the total score and on 4 subscale factors derived empirically by principal components analysis.²⁸ These factors included Drinking Obsessions (obsessional thoughts related to drinking), Automaticity of Drinking (the extent to which drinking was controlled or uncontrolled), Interference Due to Drinking (the extent to which drinking interfered with work and social functioning and to which being prevented from drinking was distressing), and Alcohol Consumption (the quantity and frequency of drinking alcohol).

The differences between the topiramate and placebo groups on the ratings for overall clinical improvement (CGI-I and CGI-S) were analyzed with a Cochran-Mantel-Haenszel test stratified by sex and center using modified ridit scores. The CGI-I was scored on a 7-point scale ranging from 1 (“very much improved”) to 7 (“very much worse”) as a change from the baseline value at randomization and succeeding treatment weeks. The CGI-S is a 7-point scale ranging from 1 (reportedly abstinent and not seeking alcohol or “not addicted”) to 7 (reportedly drinking more and constantly seeking alcohol or “extremely severely addicted”) that was assessed during each double-blind treatment week. On the CGI-S, an individual without a clinically significant addiction was defined as scoring either 1 (reportedly abstinent and not seeking alcohol) or 2 (reportedly drinking less and occasionally seeking alcohol); individuals with higher scores were considered to have a clinically significant severity of alcohol dependence.

We assessed treatment differences between topiramate and placebo on DrInC-2R by means of the repeated-measures mixed model described earlier. The DrInC-2R was assessed on a 4-point scale ranging from “never” to “daily or almost daily” (0 to 3 points, respectively). It was subdivided further into 6 subscales: Physical Consequences (8 items; maximum score, 24), Intrapersonal Consequences (8 items; maximum score, 24), Social Responsibility Consequences (7 items; maximum score, 21), Interpersonal Consequences (10 items; maximum score, 30), Impulse Control Consequences (12 items; maximum score, 36), and Control Items (5 items; maximum score, 15), as well as a Total Consequences scale (45 items; maximum score, 135), which was the sum of all subscales other than the Control Items subscale. A lower score denoted a decreased adverse consequence of drinking as compared with a higher score. The Control Items subscale was the sum of 5 reverse-scored validity items. Also, over the double-blind treatment period, we examined the correlation between decreases in DrInC-2R score and percentage of heavy drinking days measured by the timeline follow-back method.

The quality-of-life measure (Q-LES-Q) was categorized as low and high scores. This categorization was based on the different percentiles of the maximum possible scores, ranging from the 50th to the 90th. For each study week, if a subject had a score higher than the given percentile of the maximum score, he or she was designated to have a high score; otherwise, it was determined that the score was low. Pearlstein et al²⁹ suggested that a score at the 70th percentile or higher represents a “normal” quality of life. Because we observed that increasingly stringent cutoff points demonstrated the contrasts between the treatment groups more effectively, we chose the 90th percentile of the maximum score as the final discrimination criterion to perform hypothesis testing. “High” scores were considered to be indicative of improvement. Thus, to analyze the Q-LES-Q, we calculated the relative likelihood (odds ratios) of topiramate compared with placebo to achieve the 90th percentile of improvement in Q-LES-Q subscales at the end of the study using the generalized estimating equation method, which accounted for the correlation of observations within individuals with autoregressive structure as the correlation matrix, as implemented by the SAS PROC GENMOD software (SAS Institute Inc)—the same procedure used in the previous single-site trial.³⁰ The Q-LES-Q was composed of 93 items, each containing responses on a 5-point scale that ranged from 1 (“not at all or never”) to 5 (“all the time”). Ninety-one of these items were grouped into 8 summary scales. Five of the 8 summary scales were scored for all subjects: physical health (13 items; maximum score, 65), subjective feelings (14 items; maximum score, 70), leisure-time activities (6 items; maximum score, 30), social relationships (11 items; maximum score, 55), and general activities (14 items; maximum score, 70). The 3 other summary scales were scored only for subjects for whom the activities were applicable: work (13 items; maximum score, 65), household duties (10 items; maximum score, 50), and school/course work (10 items; maximum score, 50). Two of the 93 items were scored separately as individual items: satisfaction with medication, and overall life satisfaction and contentment.²³

The measures of general mood (Profile of Mood States) and sleep quality (Medical Outcomes Study Sleep scale) were analyzed by the general repeated-measures approach described earlier.

To assess the consistency of our statistical inferences to different methods for handling missing data due to dropouts, we also performed a sensitivity analysis that involved reanalyzing all the outcome data after imputing the missing values for dropouts by means of the multiple imputation method. Multiple imputation^31- 32 replaces each missing value with a set of plausible values that represent the uncertainty about the right value to impute. It draws a random sample of the missing values from its distribution. The multiple imputed data sets are then analyzed separately by standard procedures for complete data, and the results are combined to produce the final inference. This process results in valid statistical inferences that properly reflect the uncertainty due to missing values, eg, confidence intervals with the correct probability coverage. Multiple imputation inference involves 3 distinct phases: (1) the missing data are filled in m times to generate m complete data sets, (2) the m complete data sets are analyzed by standard statistical methods, and (3) the results from the m complete data sets are combined to produce inferential results. For our application, we generated 10 complete data sets (m = 5 is usually sufficient). To create a monotone missing data pattern, intermittent missing values were imputed by means of the last-observation-carried-forward method. We then used the regression method for monotone missing data³² (SAS PROC MI, MONOTONE REG statement³³) to impute the missing data due to dropouts. We then performed the statistical analyses on each complete data set and combined the 10 analyses to arrive at the final results.

Cohen effect size³⁴ estimates have been provided where appropriate for the outcomes.

We assigned 183 subjects and 188 subjects (total N = 371) at random into the topiramate and placebo groups, respectively (Figure 1). The subjects in the topiramate and placebo groups had similar baseline characteristics (Table 2).

By the primary analytic approach, topiramate was more efficacious than placebo in decreasing all liver function test values, including the log plasma γ-glutamyltransferase ratio, which is the objective marker of alcohol consumption (Table 3). Although plasma bicarbonate levels were significantly lower for the topiramate group than for the placebo group, this did not require any medical intervention. Urine pH level did not differ significantly between the topiramate and placebo groups. There was a significant reduction in plasma cholesterol level for the topiramate group compared with the placebo group. Results of all other hematologic and biochemical tests did not differ between the topiramate and placebo groups (data not shown). Topiramate was more efficacious than placebo at reducing BMI as well as both systolic and diastolic blood pressure, with no effect on pulse (Figure 2) or temperature. Topiramate compared with placebo reduced significantly the total OCDS and subscale scores (Figure 3) and was associated with greater overall clinical improvement (both CGI-I and CGI-S) and fewer harmful consequences of drinking (measured on the DrInC-2R). There was an almost contemporaneous reduction in DrInC-2R scores and the percentage of heavy drinking days measured by the timeline follow-back method, with correlations of 0.46 and 0.45 for the topiramate and placebo groups (both P < .001), respectively (Figure 4). There was no significant difference between the topiramate and placebo groups on mood (measured on the Profile of Mood States) or on alcohol withdrawal scores (assessed on the revised Clinical Institute Withdrawal Assessment for Alcohol scale), which were exceedingly low.

The results of the sensitivity analyses from the multiple imputation method were similar to those obtained by the primary analytic approach, albeit somewhat smaller in magnitude (Table 4).

By the primary analytic approach, topiramate compared with placebo had greater quality of life (Q-LES-Q) improvement at the 90th percentile in general activities, leisure-time activities, and household duties, with odds ratios of 1.86 (P = .04), 1.94 (P = .03), and 1.78 (P = .02), respectively, along with a trend to reduce sleep disturbance, which reached significance at week 14 (mean difference, 6.37; 95% confidence interval, 2.10-10.64; P = .004). In comparison, with the secondary multiple imputation model, the odds ratios for topiramate vs placebo in general activities, leisure-time activities, and household duties on the Q-LES-Q were 1.45 (P = .23), 2.27 (P = .03), and 2.05 (P = .04), respectively, and topiramate compared with placebo had a greater tendency to reduce sleep disturbance (mean difference, 5.31; 95% confidence interval, 1.33-9.28; P = .01).

Rates of concomitant medication use for the topiramate and placebo groups were 88.0% and 95.7%, respectively. The following percentages of subjects within each group received these topiramate doses or equivalent placebo doses, respectively: 0 to 25 mg (3.8% and 1.6%), 25 to 50 mg (8.2% and 0.5%), 50 to 100 mg (12% and 3.2%), 100 to 150 mg (10.4% and 6.9%), 150 to 200 mg (13.7% and 10.1%), and 200 to 300 mg (50.8% and 77.7%). The rates for adherence to taking topiramate or placebo were similar (mean [SD], 91.5% ± 14.96% vs 90.1% ± 13.12%, respectively). There was no numerical difference between the topiramate and placebo groups in mean breath alcohol concentration, and the average reading was 0.002 g%.

Retention rates at the end of the study among those randomized were 61.2% (112 of 183) and 76.6% (144 of 188) (P < .001) for the topiramate and placebo groups, respectively.

Attrition rates due to adverse events were 18.6% (34 of 183) and 4.3% (8 of 188) (P < .001) for the topiramate and placebo groups, respectively. Table 5 shows that the adverse events reported to occur in 10% or more of subjects were paresthesia, headache, taste perversion, fatigue, anorexia, nausea, insomnia, difficulty with concentration and attention, nervousness, difficulty with memory, somnolence, diarrhea, sinusitis, dyspepsia, injury, dizziness, influenza-like symptoms, pruritus, and myalgia; all except headache, nausea, sinusitis, dyspepsia, injury, influenza-like symptoms, and myalgia were more frequent for the topiramate group than for the placebo group. Four subjects in each treatment group experienced a serious adverse event. In the topiramate group, 1 subject had myopia and another had cholelithiasis. Also in the topiramate group, 1 subject had convulsions and loss of consciousness; however, these could not be attributed to the study medication. In contrast, in the placebo group, 1 subject died after a cardiac arrest associated with gastrointestinal bleeding and seizures. The precipitating incident could not be determined. Also, 1 individual each had a tibial plateau fracture, abnormally elevated levels of serum liver enzymes, and diverticulitis.

We have reported that topiramate was more efficacious than placebo at improving all self-reported drinking outcomes in a heterogeneous and geographically diverse population of alcohol-dependent individuals receiving weekly BBCET during a 14-week period. Topiramate's therapeutic effect on drinking outcomes appears to be the largest reported for a medication in a multisite alcoholism trial.⁷

Topiramate administration was associated with improvements in physical health in the 3 domains of total cholesterol level, hepatic function, and hemodynamic cardiovascular status.

Long-term heavy drinking has been associated with elevated lipid levels and the development of fatty liver disease.³⁵ Thus, topiramate's ability to decrease cholesterol levels significantly—a feature that is perhaps related to its complex effects on weight and metabolism as well as its direct effect to reduce heavy drinking—adds considerably to its general medical utility in treating alcohol-dependent individuals.

Topiramate administration was associated with significant decreases in liver enzyme levels, probably due to the reduction in heavy drinking. Taken together with the lowered lipid levels and the consequent reduced risk of fatty liver degeneration, the propensity toward progressive liver disease and eventual cirrhosis could, perhaps, be diminished if alcohol-dependent individuals were treated with topiramate.

The average blood pressure readings in our subjects did not meet the diagnostic criterion for hypertension (ie, blood pressure of 140 mm Hg systolic and 90 mm Hg diastolic).³⁶ Nevertheless, they usually met the criterion for prehypertension (ie, blood pressure of 120 mm Hg systolic and 80 mm Hg diastolic), a level at which lifestyle modifications are required to prevent cardiovascular disease.³⁶ Therefore, it is of interest that, at the end of the study, the topiramate-treated group achieved blood pressure levels at about the threshold for prehypertension (mean, 123.0 mm Hg systolic and 75.5 mm Hg diastolic), whereas those who received placebo exceeded this cutoff (mean, 130.8 mm Hg systolic and 81.8 mm Hg diastolic). Presumably, these improvements in blood pressure were an added benefit of the decrease in BMI, reduction in drinking, or both among topiramate recipients. Nevertheless, these results suggest a possible added clinical benefit of an improvement in hemodynamic cardiovascular status if topiramate were to be used to treat alcohol-dependent individuals.

Excessive alcohol consumption is generally a risk factor for obesity³⁷ even though some individuals with severe alcoholism can become malnourished.³⁸ Topiramate-associated decreases in BMI, presumed to be partially due to a reduction in craving for food,³⁹ might, therefore, be of health benefit to alcohol abusers or dependent individuals who are overweight or obese or who have strong cravings for food.

Topiramate treatment was associated with an improvement in psychosocial factors that contributed to its therapeutic effect. Topiramate reduced all subscales on the OCDS²⁸ that measure obsessions and compulsions about using alcohol. While we have proposed that topiramate-induced decreases in obsessional thoughts and compulsions about drinking reduce alcohol consumption,⁴ the converse is also plausible. Similarly, the topiramate-associated reduction in sleep disturbance that we observed could have been the consequence of rather than the trigger for improved drinking outcomes. Nevertheless, because sleep disturbance is a feature of—and can be a trigger for—relapse among alcohol-dependent individuals,⁴⁰ formal research is needed on topiramate's effects on the sleep architecture of alcoholic individuals.

Topiramate's effect in reducing the psychosocial consequences of drinking increased as treatment progressed.⁷ As the frequency of heavy drinking decreased, there was a concomitant reduction in its psychosocial consequences. Because of the collinearity of these variables, it was not possible for us to attribute any cause-and-effect prediction to the relationship between psychosocial improvements and reduced heavy drinking. It was, however, tempting to speculate that the reductions in heavy drinking might explain a substantial part of this effect. This is because the principal target of the brief behavioral intervention was to enhance adherence to medication rather than to target any antecedents, triggers, or perpetuating psychosocial factors associated with the heavy drinking. Nevertheless, the possibility that the brief behavioral intervention led to some ongoing psychosocial improvement cannot be discounted entirely because such brief treatments have been shown previously to be effective treatment for alcoholism in a variety of settings.^41- 42 Also, we did not have an appropriate “control” group of participants who received the brief intervention alone (ie, without adjunctive placebo medication). Therefore, the independent effects of our brief intervention cannot be teased out from those of the medication alone.

Topiramate had a significant effect in improving some aspects of quality of life at the end of the trial, including general and leisure-time activities and household duties. We found this to be an encouraging effect given the relatively short duration of the trial.

Recently, the authors of an influential review of the literature suggested that harm reduction approaches might be as beneficial as abstinence-oriented strategies, especially if treatments are directed to accommodate the needs and preferences of the individual or target populations.⁴³ Indeed, Johnson et al³⁰ argued that a harm reduction approach offers the potential to sustain “partial” remission from drinking, perhaps even over the long term if therapeutically effective medication is taken continuously. Not only does this allow for other psychosocial influences to bring about complete remission, but it provides a framework for treating alcohol dependence in much the same way as any chronic relapsing disorder, such as diabetes or hypertension. Further elucidating the therapeutic benefit of long-term topiramate treatment on drinking behavior, and physical and psychosocial health, is, therefore, an important research goal.

We considered the possibility that the differential attrition rate, greater for the topiramate group than for the placebo group, might have an effect on outcome. Nevertheless, the consistency of the results between the primary analytic approach (mixed model) and the sensitivity analysis using the multiple imputation method presented herein, as well as information obtained by using other modeling techniques reported elsewhere,⁷ suggests that dropouts in this study were unlikely to be informative or affect the outcomes related to efficacy but were more the direct result of adverse events.

The robustness of topiramate's therapeutic efficacy was evidenced by the similarity of the results obtained for the primary and sensitivity analyses, and the effect sizes were typically in the medium range or better on most outcomes for which there was an improvement in physical or psychosocial health.

Our study had 4 limitations. First, while the pattern of adverse events was similar to that found in our previous study,⁶ the more rapid titration was associated with decreased study adherence. In that study, topiramate was titrated over an additional fortnight, and retention rates were similar between the topiramate and placebo groups. We observed that the clinical sites least familiar with topiramate experienced more difficulties with retention, but completion rates among some experienced groups approached 90% (data not shown). Therefore, an approach to more generalized use, especially among clinicians with less experience with topiramate, would be to propose a slower titration schedule, along with focused education about how best to manage emergent adverse events. Of course, future studies elucidating the optimum topiramate dose that balances efficacy, tolerability, and treatment adherence would be important, particularly with respect to the transient cognitive effects—typically word-naming difficulties—that require careful management and appropriate reassurance. Second, it is plausible that certain subtypes of alcoholic subjects might benefit the most from topiramate treatment. Although we did not observe a differential treatment response by age of problem drinking onset, subtype classifications that included genetics or other biomarkers might have provided additional information. Third, our study was not designed to examine topiramate's posttreatment effects; thus, it remains unknown how long these physical and psychosocial improvements will be sustained. Nevertheless, even when a brief adherence enhancement intervention is coupled with effective pharmacotherapy, there are important gains in physical and psychosocial health during treatment. Fourth, our specialists in this clinical trial treated alcohol-dependent individuals who might be generally healthier and more motivated to change than those in the community. Because topiramate pharmacotherapy can be paired with a brief intervention deliverable by nonspecialist health practitioners, a next step, therefore, would be to examine its efficacy in community practice settings.

In summary, our findings demonstrate that topiramate appears to be a generally effective treatment for alcohol dependence because it improves not only the “symptom” of drinking but also its physical and psychosocial sequelae.

Correspondence: Bankole A. Johnson, DSc, MD, PhD, MPhil, FRCPsych, Department of Psychiatry and Neurobehavioral Sciences, University of Virginia, PO Box 800623, Charlottesville, VA 22908-0623 (bankolejohnson@virginia.edu).

Accepted for Publication: December 30, 2007.

Author Contributions: Dr Johnson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Addolorato, Anton, Ciraulo, Kranzler, Mann, O’Malley, and Swift contributed equally as coauthors. Study concept and design: Johnson, Rosenthal, Capece, and Beyers. Acquisition of data: Johnson, Capece, Wiegand, Beyers, McKay, Ait-Daoud, Anton, Ciraulo, Kranzler, O’Malley, and Swift. Analysis and interpretation of data: Johnson, Rosenthal, Capece, Wiegand, Mao, Ait-Daoud, Addolorato, Anton, Ciraulo, Kranzler, Mann, O’Malley, and Swift. Drafting of the manuscript: Johnson and Rosenthal. Critical revision of the manuscript for important intellectual content: Johnson, Rosenthal, Capece, Wiegand, Mao, Beyers, McKay, Ait-Daoud, Addolorato, Anton, Ciraulo, Kranzler, Mann, O’Malley, and Swift. Statistical analysis: Johnson,Wiegand, and Mao. Obtained funding: Johnson, Rosenthal, Capece, Wiegand, and McKay. Administrative, technical, or material support: Johnson, Rosenthal, Capece, Wiegand, Beyers,McKay, Ait-Daoud, Addolorato, Anton, Ciraulo,Kranzler, Mann, O’Malley, and Swift. Study supervision: Johnson, Rosenthal, Capece, Wiegand, McKay, and Ait-Daoud.

Financial Disclosures: Dr Johnson has been a consultant for Ortho-McNeil Janssen Scientific Affairs LLC and is a consultant for Organon and TransOral Pharmaceuticals Inc. Drs Rosenthal, Wiegand, Mao, and McKay and Mss Capece and Beyers are employees of Ortho-McNeil Janssen Scientific Affairs LLC. Dr Ait-Daoud has been a consultant for Ortho-McNeil Janssen Scientific Affairs LLC. Dr Addolorato has been a consultant for Ortho-McNeil Janssen Scientific Affairs LLC and is a consultant for Organon. Dr Anton has received consulting fees and honoraria from Forest Laboratories and Alkermes Inc; consulting fees and grants from Bristol-Myers Squibb, Hythiam, and Pfizer; consulting fees, honoraria, and grants from Contral Pharma/Biotie Pharmaceuticals and Ortho-McNeil Janssen Scientific Affairs LLC; and consulting fees from AstraZeneca, Axis Shield, Cephalon, DrugAbuse Sciences, Sanofi-Aventis, Eli Lilly, and Solvay Pharmaceuticals. Dr Ciraulo has received consulting fees and clinical trial contracts from Janssen, Bristol-Myers Squibb, and Ortho-McNeil Janssen Scientific Affairs LLC; consulting fees from Cephalon; and clinical trial contracts from Alkermes Inc, AstraZeneca, Catalyst Pharmaceutical Partners, DrugAbuse Sciences, Lipha, and UCB Pharma. Dr Kranzler has served as a consultant for and received research support from Alkermes Inc, Ortho-McNeil Janssen Scientific Affairs LLC, Bristol-Myers Squibb, Forest Pharmaceuticals, and DrugAbuse Sciences, and served as a consultant for Elbion, Sanofi-Aventis, and Solvay Pharmaceuticals. Dr Mann has received consulting fees and clinical trial contracts from Alkermes Inc, Janssen, Sanofi-Aventis, Pfizer, Lipha Pharmaceuticals, Ortho-McNeil Janssen Scientific Affairs LLC, and Forest Pharmaceuticals. Dr O’Malley has served as a consultant for and received medication supplies and/or clinical trial contracts from Alkermes Inc, GlaxoSmithKline, Ortho-McNeil Janssen Scientific Affairs LLC, and Pfizer; served as a consultant for Eli Lilly and Johnson & Johnson; received medication supplies or clinical trial contracts from Lipha Pharmaceuticals, Bristol-Myers Squibb, Sanofi-Aventis, and Mallinckrodt; and received travel reimbursement from Alkermes Inc. Dr Swift has served as a consultant for and received grant funding from Ortho-McNeil Janssen Scientific Affairs LLC and Pfizer; served as a consultant and on speakers bureaus for Cephalon and Forest Laboratories; served as a consultant for Alkermes Inc, Organon, and TransOral Pharmaceuticals Inc; and received grant funding from US World Meds and Bristol-Myers Squibb.

Funding/Support: Ortho-McNeil Janssen Scientific Affairs LLC provided the medication and funding for this study.

Role of the Sponsor: The sponsor was involved in all stages from study design through interpretation of the results including critical review of the manuscript. Data were managed and analyzed by Ortho-McNeil Janssen Scientific Affairs LLC and PharmaNet Inc, a contract research organization, and were interpreted by authors of the study with input from Ortho-McNeil Janssen Scientific Affairs LLC clinical and statistical staff. The entirety of the first draft was prepared by the lead author (Dr Johnson). The draft was reviewed by all of the authors, who discussed it as a group at a scheduled meeting of the Topiramate for Alcoholism Advisory Board (see box on page 1198) in New York, New York, on February 2, 2007. Amendments pertained to style and presentation, and no changes were made to the results or their interpretation from the initial draft or the results as presented at that meeting.

Additional Contributions: We thank the staff at the participating sites for their skilled technical assistance, as well as Robert H. Cormier Jr, BA, and Chamindi Seneviratne, MD, for their assistance with preparing the report. Both Mr Cormier and Dr Seneviratne are employed by the University of Virginia.