Naltrexone is an opioid receptor antagonist that reduces

Transcription

1 Review Article The Status of Naltrexone in the Treatment of Alcohol Dependence Specific Effects on Heavy Drinking Helen M. Pettinati, PhD,* Charles P. O Brien, MD, PhD,*y Amanda R. Rabinowitz, BA,* Shoshana M. Wortman, BA,* David W. Oslin, MD,*y Kyle M. Kampman, MD,* and Charles A. Dackis, MD* *Center for the Study of Addictions, Department of Psychiatry, University of Pennsylvania School of Medicine and yphiladelphia Veterans Affairs Medical Center, Philadelphia, PA. Received January 5, 2006; accepted after revision September 1, This work was supported by grants from the National Institute on Alcohol Abuse and Alcoholism (R01-AA09544 to Dr Pettinati; R01-AA to Dr Oslin), the National Institute on Drug Abuse (P60 DA05186 to Dr O Brien; P50 DA12756 to Dr Pettinati), the National Institute on Mental Health (K08-MH A1 to Dr Oslin), and the VISN 4 Mental Health Research Education and Clinical Center (MIRECC) at the Philadelphia Veterans Affairs Medical Center. Address correspondence and reprint requests to Helen Pettinati, PhD, Department of Psychiatry, Treatment Research Center, University of Pennsylvania, 3900 Chestnut St, Philadelphia, PA Pettinati_H@mail.trc.upenn.edu. Copyright n 2006 by Lippincott Williams & Wilkins ISSN: /06/ DOI: /01.jcp Background: In almost 2 decades of naltrexone research for treating alcoholism, there have been 29 published randomized placebo-controlled trials of opioid antagonists, primarily naltrexone, for the treatment of. The present review builds on prior systematic reviews while maximizing the number of included studies to date, for the purpose of resolving inconsistencies in naltrexone s reported efficacy across trials. Clinical trial results in this article are evaluated by the type of outcome measure used to determine naltrexone s treatment advantage, that is, measures related to reducing heavy drinking versus those related to increasing abstinence. Methods: We conducted a Medline search to identify double-blind studies from 1990 to the present (2006) that evaluated the use of an opiate antagonist for the treatment of. There were 29 studies identified, representing 5997 alcohol-dependent, which met our study inclusion criteria for this review. Studies were evaluated in this review on 4 prespecified drinking outcomes 2 related to any drinking and 2 related to heavy or excessive drinking. Results: In the treatment of, we found that 19 (70%) of 27 clinical trials that measured reductions in heavy or excessive drinking demonstrated an advantage for prescribing naltrexone over placebo, whereas only 9 (36%) of 25 clinical trials that measured abstinence or any drinking found an advantage for medication over placebo. Conclusion: The majority of double-blind clinical trials in the literature favored prescribing naltrexone for to reduce heavy drinking. This finding is consistent with our understanding of naltrexone s mechanism of action of decreasing excessive drinking by reducing the reward associated with drinking alcohol. Thus, we conclude that outcome measures related to heavy or excessive drinking are most relevant to defining naltrexone s therapeutic effects. Factors influencing naltrexone response (treatment adherence and distinct patient subgroups) are also discussed. (J Clin Psychopharmacol 2006;26: ) Naltrexone is an opioid receptor antagonist that reduces heavy drinking by diminishing the rewarding neurobiological effect of alcohol. The discovery that naltrexone can aid in treating alcoholism was the direct result of a systematic progression of studies from the laboratory to the clinic in an almost legendary path. 1 This line of inquiry ultimately led to the Food and Drug Administration (FDA) in 1994 approving naltrexone for the treatment of alcoholism in the United States. However, despite its FDA approval for treating alcoholism, naltrexone has not been widely prescribed in the United States for this purpose. 2 Recently, naltrexone has emerged at the forefront of alcoholism treatment research. The FDA approval in April 2006 of Vivitrol (Alkermes, Inc, Cambridge, Mass and Cephalon, Inc, Frazer, Pa) (an injectable long-acting form of naltrexone) for treating, followed by the release of the results a month later from the national government-sponsored (National Institute on Alcohol Abuse and Alcoholism) COMBINE Study 3 showing that oral naltrexone given with a structured approach to medical management produced some of the best outcomes during 4 months of treatment, have reintroduced this relatively dated medication to the treatment community. In light of these latest developments, a thorough reevaluation of all of the existing literature on controlled naltrexone trials is warranted. In almost 2 decades of naltrexone research for treating alcoholism, there have been 29 published, randomized, placebo-controlled trials of opioid antagonists, primarily naltrexone. Many of these studies, but not all, support the advantage of prescribing opiate antagonists as part of the treatment of. There also have been a number of reviews that have favored naltrexone as a treatment of alcoholism, 4 8 including 2 published meta-analyses that quantified the advantage that naltrexone has, compared with placebo (plus counseling), in reducing heavy drinking, albeit modest. 9,10 Nonetheless, at least 1 large randomized, controlled naltrexone study has reported no advantage of 610 Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006

2 Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 Naltrexone Effects on Heavy Drinking naltrexone over placebo (eg, N = ), and naltrexone s inconsistent results across some studies may have discouraged the medication s use in the treatment community. The most frequent reason for treatment failure with naltrexone has been medication nonadherence Some major naltrexone trials have been published that have demonstrated that the benefits of naltrexone are only seen in subjects who are reliable in taking their study medication every day. Although there may be no demonstrable advantage for naltrexone in the study sample as a whole, a clear advantage for naltrexone emerges when outcome data are based only on who received adequate exposure to the medication In addition, recent studies suggest that specific patient subgroups may respond better to naltrexone than others. Subgroups of who appear to respond better to naltrexone include those with a specific genetic polymorphism in the m-opioid receptor gene 16 and those with enhanced opioidergic activity in response to alcohol ingestion. 17 A family history of alcoholism predicts a good naltrexone response, as does intense craving for alcohol. 20 As important as this topic is, however, constructing a profile for the optimal naltrexone responder will require further study. One approach to a better understanding of naltrexone s efficacy is to consider outcome measures that are sensitive to the medication s proposed mechanism of action. As the field has gained a better understanding of how naltrexone behaves in the at-risk alcoholic brain, it has become clear that its most robust effect is reduction of excessive or heavy drinking, which may in some cases lead to an increase in the number of abstinent days. Heavy drinking has been defined as drinking 5 or more alcoholic drinks for men and 4 or more for women in a day. Reductions in heavy drinking with naltrexone likely results from naltrexone s action on reducing the pleasure or the high that many alcoholic report when drinking alcohol. 19,21 Therefore, it could be predicted that naltrexone would more consistently provide a treatment advantage when treatment response is evaluated by measures related to reductions in heavy drinking rather than measures related to increased abstinence. However, many of the published naltrexone studies have either reported only 1 outcome measure or have evaluated several outcome measures without respect to naltrexone s purported action on reducing heavy drinking. The recently published meta-analyses 9,10 have come closest to evaluating naltrexone s specific effects on reducing drinking by comparing effect sizes across various outcome measures. However, the stringent requirements for conducting meta-analyses inevitably result in excluding a number of clinically relevant studies. The present review builds on the available meta-analytic reviews by including, in one place, the largest number of published clinical trials on naltrexone (United States and internationally). Using a complementary perspective, this review purports to resolve some of the troubling inconsistencies in naltrexone s efficacy across the double-blind, placebo-controlled clinical trials by systematically evaluating results by the type of outcome measure used to determine naltrexone s treatment advantage (measures related to reducing heavy drinking versus those related to increasing abstinence). In addition, for those studies in which the majority of the study population was nonadherent, results for adherent subjects will be considered as the main results in this review whenever these results were reported. Resolving some of the apparent discrepancies in naltrexone s efficacy results across the growing number of clinical studies will expand our knowledge base of this treatment, better define a naltrexone responder/nonresponder, and provide more comfort for clinicians deciding whether to prescribe naltrexone for an alcohol disorder. HISTORICAL AND CONTEMPORARY PERSPECTIVES ON NALTREXONE By way of background, before the proposed evaluation of the existing naltrexone literature, this article will provide abbreviated historical and current perspectives on the use of opioid antagonists, primarily naltrexone, in the treatment of. Naltrexone is the most commonly studied opioid antagonist for treating alcoholic. More information on the history and development of naltrexone is available. 1,4,5,7 Path of Discovery: Naltrexone Treatment for Alcoholism Naltrexone, an antagonist at the m-, d-, and k-opioid receptors, was initially investigated more than 30 years ago for its utility in treating heroin and other opiate addictions and subsequently received FDA approval in 1984 for the treatment of opiate dependence. In the 1970s, preclinical data emerged demonstrating that opioid antagonists, such as naltrexone, blocked the ability of ethanol to increase dopamine release in the dopamine reward pathways leading from the ventral tegmental area to the nucleus accumbens. 22,23 Other animal work showed that naloxone reduced alcohol preference in alcohol-dependent rats 24 and reduced drinking in rats bred for high levels of alcohol preference. 25 A seminal study by Altshuler and colleagues, 22 first presented in 1979 at the Committee on Problems of Drug Dependence, demonstrated that naltrexone blocked the selfadministration of alcohol in rhesus monkeys. Persuaded by results emerging in the animal literature, O Brien in 1983 initiated open dose ranging studies of naltrexone in alcoholic. Subsequently, he and Volpicelli initiated a placebo controlled clinical trial in chronic alcoholic at the Philadelphia Veterans Affairs Medical Center. 26,27 The results of this 12-week study were consistent with preclinical findings. That is, naltrexone significantly reduced rates of relapse to heavy drinking, compared with placebo, with minimal side effects. These results were replicated in an independent study by O Malley et al. 28 These 2 important studies provided the basis for the FDA approval of oral naltrexone in 1994 for the treatment of alcohol dependence in the United States. Naltrexone has also been approved for treatment of in Australia, Canada, and a number of European countries. Following naltrexone s FDA approval for alcohol use disorders, additional research has been conducted to clarify n 2006 Lippincott Williams & Wilkins 611

3 Pettinati and Associates Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 naltrexone s mechanism of action, 19,21,29 the effects of medication nonadherence on naltrexone outcomes, 12,15 and profiling naltrexone responders and nonresponders. 16,20 How Naltrexone Compares with Other FDA-Approved Pharmacotherapies for Alcoholism Naltrexone is the second (oral form) and fourth (injection lasting 30 days) of now 4 medications that have been approved in the United States specifically for treating. Disulfiram was the first medication approved by the FDA for treating alcoholism in 1951, 30,31 and the third medication, acamprosate, was approved by the FDA in 2004 for maintaining abstinence in the treatment of Although acamprosate and disulfiram are beyond the scope of this review, to provide a comparison to naltrexone, they are discussed briefly below. The mechanism of action for disulfiram is to create a physiologically aversive state when combined with alcohol that includes flushing, nausea, and headache. These symptoms are sufficiently strong to induce some to immediately stop drinking alcohol. In this way, disulfiram does not directly address the pathophysiological changes associated with. Evidence from clinical trials and private practice indicate that many skip medication doses so that they may consume alcohol without incident when they are experiencing an urge to drink alcohol. 30,38 Disulfiram is used today successfully in treatment when can be supervised regularly taking their medication. 39 Still, today fewer than 250,000 disulfiram prescriptions are written in a year s time. 2 A comprehensive review of the disulfiram literature can be found elsewhere. 40 Acamprosate is a putative glutamate modulator that promotes abstinence by alleviating aversive symptoms related to protracted alcohol withdrawal. Like naltrexone, acamprosate effects neural pathways related to ethanol s effects on the brain. Meta-analyses suggest that naltrexone and acamprosate both have significant but modest effects on drinking outcomes. However, in the recently completed COMBINE study, acamprosate failed to demonstrate superiority over placebo with respect to increasing the number of days of abstinence. 3 However, these results are in contrast to a number of European clinical trials that have consistently supported the efficacy of acamprosate. 41 Naltrexone Today: Pharmacology, Safety, and Medication Adherence Both theoretical and empirical research have supported the ideas that is a chronic disease and that biological vulnerabilities contribute to the pathogenesis of In addition, it has generally been thought that the identification of biological abnormalities in chronic diseases will allow more specificity in the pharmacological interventions used to remedy those conditions. One widely discussed area of biological vulnerabilities associated with centers around neurotransmitter abnormalities. More than 2 decades of research has shown that dopamine is a principal neurotransmitter identified with the rewarding effects of substances of abuse However, other neurotransmitters, such as endogenous opioids, also have been implicated in the selfadministration of alcohol and other substances, directly or indirectly through dopamine and other neurotransmitters involved in reward. 48 Pharmacology The most commonly studied opioid antagonist is naltrexone. When taken orally, naltrexone is quickly absorbed and undergoes first-pass metabolism in the cytosol system in the liver. 49 Naltrexone is then converted to several metabolites. The major metabolite of naltrexone is 6-bnaltrexol, which is also an opioid antagonist and by itself reduces alcohol drinking in a rat model. 50 Two other minor metabolites do not appear to significantly contribute to naltrexone s pharmacological activity or toxicity (2-hydroxy- 3-methoxy-6-b-naltrexol and 2-hydroxy-3-methyl-naltrexone). The mean serum elimination half-life after chronic administration of a 50-mg daily dose is 9.7 hours for naltrexone and is 11.4 hours for 6-b-naltrexol. 51 In contrast, PET studies using C11 carfentanil show significant blocking of brain m receptors for more than 72 hours after a single 50-mg dose. 52 The suspected mechanism of action of naltrexone is a blunting of the reinforcing properties of alcohol, which are mediated in part through the endogenous opioid system. Naltrexone has high affinity for m receptors and good affinity for k- and d-opioid receptors, modulating the endogenous opioid system in the brain. Animal studies suggest that blockade of multiple opioid receptors weakens the pleasurable effects or high alcoholic report when drinking alcohol. 29 Several laboratory studies have demonstrated that naltrexone decreases the reinforcing or pleasurable effects of alcohol in social drinkers 53,54 and alcoholic who sample alcohol. 19,21,29,55 Safety Profile A decided advantage of naltrexone in the treatment of is its safety profile. Historically, chronic drinkers often were denied medications (except for detoxification) due to safety concerns over the potential interaction of most medications with alcohol. However, naltrexone does not potentiate alcohol effects on motor skills or cognition or reduce seizure threshold; neither has naltrexone overdose produced any fatalities. In addition, naltrexone appears to have no significant abuse potential. That is, it has not been associated with pleasurable effects, nor physiological dependence, nor with tolerance to its opioid-blocking effects. 56,57 Finally, in the decade before FDA approval for treating, naltrexone had been given to more than 2000 detoxified intravenous-opioid abusers in numerous clinical trials. The frequency of side effects from naltrexone is relatively low (<15%). 58 Nausea and vomiting are the most common side effects reported with naltrexone. Headache, low energy, anxiety, depression, rashes, and decreased alertness are less common. 59 These side effects typically resolve spontaneously after a few doses, or by reducing the daily dosage. A few cases of dysphoria have 612 n 2006 Lippincott Williams & Wilkins

4 Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 Naltrexone Effects on Heavy Drinking been reported by with opioid addiction, 60,61 and there are reports of increased dysphoria in normal volunteers given naltrexone. 62 These reports may be indicative of a mild opioid withdrawal-like reaction in some that occurs when naltrexone abruptly blocks opiate receptors that are in an activated condition, possibly due to stress or substance intake. Nonetheless, most studies have not found evidence of dysphoria or other mood changes with the recommended dose of naltrexone when treating dependent on opiates 63,64 or alcohol. 26,28 In naltrexone s package insert, there is a black-box warning of possible hepatocellular injury (when taken in approximately 7 times the recommended daily dose). It also states that naltrexone should not be given in cases of acute hepatitis or liver failure. This warning of hepatotoxic effects was derived from early studies, which prescribed dosages of up to 350 mg/d for obesity and dementia. 68 There are no reports of hepatotoxicity at the recommended daily dosage of 50 mg. In fact, liver enzyme levels in alcoholic typically are lowered with naltrexone treatment (relative to placebo), probably because of reduced alcohol intake while in treatment. Consequently, there appears to be a much greater risk of hepatotoxicity from chronic excessive alcohol drinking than from treatment with naltrexone at the recommended dosage. 69 DuPont Pharma conducted a large safety study of naltrexone (approximately 570 alcohol-dependent ) and concluded that the medication was safe under a variety of circumstances. 58 Naltrexone also has been studied for use in several medical disorders other than alcohol and drug dependence, including obesity, bulimia, 70 dementia, 68 autism, 71 self-injurious behavior seen in autism, 72,73 and mental retardation. 74 Adherence Profile Medication nonadherence is a universal phenomenon in chronic medical illnesses. Frequently nonadherence is related to medication side effects; however, this is only one of many possible reasons that might skip days of medication or stop taking their medication entirely. For example, there is typically a bias against taking medication for too long, or the cost of medication can impact medication adherence in clinical settings where the patient pays out of pocket (see more on this subject in Refs. 15,75 78). In addition, historically, there has been a fundamental philosophical bias against taking medication for drinking problems. That is, medication nonadherence rates can be higher in alcoholic than in other patient populations, given the history of community attitudes toward alcohol treatment. In addition, unlike medications for most psychiatric conditions, opioid antagonists, such as naltrexone, neither alleviate distressful psychiatric symptoms nor inherently provide a sense of well-being or level of comfort. Rather, naltrexone s purported mechanism of action is to dampen the rewarding or euphoric feelings when drinking or thinking about drinking alcohol. Scientifically, the impact of blocking the high from alcohol can be enormous, given our knowledge about the brain, pleasurable reinforcement, 15,75 77 Now memory, and repetitive behaviors. However, naltrexone s mechanism of action essentially translates clinically to a medication that may not provide immediate relief and seems to prevent the pleasurable feelings from alcohol. Adding naltrexone to counseling, however, can be powerful in highly motivated and medication-adherent. 12,15 Nonetheless, getting to regularly take naltrexone over any length of time can be a problem in some settings and not taking naltrexone as prescribed is a major barrier to successful treatment response. That is, studies with high rates of medication nonadherence may inadvertently conclude that the medication is ineffective. The intent-totreat analyses, which can be comprised primarily of nonadherent in certain studies, may reveal no differences in drinking reductions between naltrexone- and placebo-treated subjects, unless a reanalysis is done to compare medication to placebo in who had good exposure to the medication/placebo (ie, were medication adherent, eg, Volpicelli et al 12 ). Future studies of the impact of medication nonadherence and methods to enhance adherence are needed. Suggestions for enhancing adherence with taking naltrexone as prescribed can be found elsewhere. that extended release naltrexone is available, monthly injections are likely to result in better adherence. This type of delivery system conveniently bypasses the burden of daily decisions to take medications, particularly a medicine that will reduce the enjoyment derived from drinking. The several studies that have published results using one of several types of a long-acting injectable naltrexone have been included in this review. OPIOID ANTAGONISTS FOR ALCOHOLIC PATIENTS: CLINICAL TRIALS In the past decade, there have been a number of well-controlled randomized clinical trials that have further tested the efficacy of opioid antagonists in treating alcohol dependence, with some of the studies failing to report any advantage for the medication over counseling alone. Inconsistent outcomes following FDA approval can be a major source of confusion and may unnecessarily discourage the use of an effective medication. On the other hand, the last decade has also provided us with an increased knowledge of how we think opiate antagonists work in treating alcoholism, allowing us to better target specific drinking behaviors (eg, reducing heavy drinking, which may or may not result in total abstinence from drinking). This knowledge has expanded the number and types of drinking measures being reported in more recent studies. Therefore, it is important to reevaluate the clinical trial evidence regarding opiate antagonists, primarily naltrexone, in the context of recently published studies, to fully understand the current status of naltrexone pharmacotherapy in the treatment of. Methods and Results We conducted a Medline search to identify studies between 1990 and the present (2006) evaluating the use of an opiate antagonist for the treatment of n 2006 Lippincott Williams & Wilkins 613

5 Pettinati and Associates Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 in human double-blind, placebo-controlled trials. There are 2 known opioid antagonists that have been tested for treating alcoholism: naltrexone and nalmefene. Using the keywords (naltrexone OR nalmefene) AND alcohol we identified a total of 489 search results, that we then limited to 95 by narrowing the search to human subjects and randomized controlled trials. The titles and abstracts for those 95 results were reviewed and limited to 27 appropriate trials. The reference lists of the 27 identified articles were also reviewed for relevant studies, and 2 additional trials were added based on that review for a total of 29 clinical trials. To be included in this review, trials had to be published, randomized, double-blind, placebo-controlled, and include subject populations who were identified as having an alcohol disorder. We also arbitrarily decided that this review would not include any published results where the minimum sample size was fewer than 20 subjects. This was based on the supposition that a study with less than 10 subjects per treatment group is more similar to a case study than to a randomized controlled trial. Also, given that our aim for this review was to evaluate studies based on the type of outcome measures collected, we could only include studies that reported at least 1 of 4 possible outcome measures of drinking (defined below). When multiple publications reported results on the same subject group, only the primary publication was considered. A low rate of medication adherence (<50%) was not exclusionary, but when drinking outcomes for treatment-adherent were reported or easily derived from the publication, our review evaluated those outcomes. Following these few guidelines, we identified a total of 29 published studies, representing 5997 alcohol-dependent, which fit these criteria. To obtain detailed drinking outcome data, all studies were searched for 4 prespecified well-known drinking outcome measures 2 related to any drinking and 2 related to excessive or heavy drinking. The majority of alcohol treatment studies have included some measure of abstinence or any drinking because these measures are often clinically relevant goals for treatment-seeking alcoholic. However, the effects of opiate antagonist pharmacotherapy on drinking are purported to be specific to reducing the reinforcement derived from consuming alcohol, which, in turn, predicts consuming less alcohol when drinking, but would not necessarily predict total abstinence. Little attempt has been made in clinical trials to tailor drinking outcome measures to purported mechanisms of specific pharmacotherapy, although many of the existing naltrexone trials included some measure of excessive or heavy drinking (eg, relapse-defined mainly as a return to heavy drinking, ie, increased volume). Very recent studies have included a measure that combines the frequency and intensity of drinking alcohol-days of heavy drinking. These latter 2 measures (ie, relapse and heavy drinking) are more specific to purported mechanisms of action of opiate antagonist pharmacotherapy. To this end, it was important in this review to evaluate the medications effects on reducing excessive or heavy drinking, as well as their potential effects on any drinking. The 2 measures of any drinking we preselected to report for all studies are (1) the percentage or number of days drinking (16/29, 55%, studies reported this measure) and (2) the percentage of subjects who were abstinent or, if not provided, the percentage or number of days abstinent was reported (15/29, 52%, studies reported this measure). The 2 measures of excessive or heavy drinking are (1) the percentage of subjects who relapsed or the time (in days) to relapse (23/29, 79%, studies reported this measure) and (2) the percentage or number of days of heavy drinking (9/29, 31%, of the studies reported this measure). Relapse and heavy drinking were typically defined as: after starting treatment, drinking 5 or more drinks for men or 4 or more drinks for women in 1 day. A few of the studies expanded upon the minimal definition to include more than 1 heavy drinking day in a specified period. Two of the included studies reported primary outcome measures other than those described above. One study reported the defined outcome measures as probabilities (eg, the probability of heavy drinking). 82 Another study reported the percent reduction in heavy drinking. 79 Wherever possible, we included the studies reported main outcome measure(s). Any major differences in definitions used in individual studies are documented (see footnotes to Table 2). However, other measures of drinking that may have been reported in the studies that could not be deemed a variation of the definitions we are using (eg, number of total drinks during treatment, liver enzyme levels) were not used as proxies for the predefined measures. Although biomarkers, such as liver enzymes, were not used as primary outcome measures in this review, it should be noted that they have been put forth as possibly superior indicators of change in drinking behavior than the self-report measures evaluated in this review. 83 Liver enzyme levels, such as g-glutamyl transpeptidase, aspartate aminotransferase, and alanine aminotransferase, offer a clear advantage over self-report measures, in that they can be an objective indicator of safety and efficacy in alcohol treatment studies. On the other hand, using biomarkers in this context does introduce certain challenges and limitations that are not encountered with self-reports. For example, (1) to detect a difference, the patient would need to have abnormal or elevated laboratory tests at treatment entry, and this is not the case for all alcohol-dependent entering treatment; (2) the patient s awareness that his or her drinking status may be verified by laboratory tests may influence attendance for such tests, as well as drinking behaviors; (3) biomarkers cannot be meaningfully evaluated too frequently as some period needs to pass in order for drinking behavior to affect them; (4) it is as yet unclear exactly how biomarkers, such as liver enzyme tests, can be quantified so that they are comparable across studies. 83 Liver enzyme levels are not typically viewed by the FDA as primary measures of efficacy for alcoholism clinical trials. It is for this reason, in conjunction with those limitations discussed above, that we chose not to include biomarkers in the outcome measures we systematically evaluated for the purposes of this review. However, biomarker results may 614 n 2006 Lippincott Williams & Wilkins

6 Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 Naltrexone Effects on Heavy Drinking illuminate otherwise confusing results in any given study included in this review. A notable example is the clinical trial results reported by Gastpar et al 84 (see summary tables). This study s primary drinking outcomes (self-report) after treatment with naltrexone indicated no advantage for naltrexone compared with placebo, but liver enzymes levels are significantly decreased in the naltrexone versus the placebo groups, suggesting that the naltrexone group likely had a substantial decrease in heavy or excessive drinking, compared with those treated with placebo. 84 There were 2 conclusion statements planned from our analysis of the 4 drinking outcome measures, regarding whether naltrexone/nalmefene had an advantage, or not, over placebo in treating for (1) any drinking or (2) excessive drinking. One conclusion represents 2 measures of any drinking, and the second conclusion represents 2 measures of excessive drinking. If either of the outcome measures within a drinking category statistically favored naltrexone/nalmefene, then the conclusion per drinking category was stated: naltrexone or nalmefene > placebo, respectively. If neither of the 2 outcome measures within a drinking category statistically favored naltrexone/nalmefene, then the conclusion per drinking category was stated: naltrexone = placebo. It is worth noting that there were no instances in any of the studies we reviewed where placebo-treated had statistically more favorable drinking outcomes than those treated with naltrexone (ie, naltrexone < placebo). Brief profiles of each of the 29 studies are presented in Table 1. 3,11 14,20,26 28,79 82, The drinking outcomes and conclusions about the efficacy of the medications summarized in Table 1 are from detailed drinking outcome data for each of the 29 studies, which are reported in Table 2. Inspecting the last 2 columns in Table 1 reveals that the conclusions about the efficacy of opiate antagonists depend heavily on what type of drinking is measured ( any vs. excessive ). That is, naltrexone s efficacy was far less clear when the outcome focused on abstinence: only 9 studies (2517 subjects) of 25 studies that measured any drinking reported an advantage of medication over placebo in reducing drinking (36% of the studies that measured abstinence). However, with respect to excessive or heavy drinking, 19 (3950 subjects) of 27 studies that measured excessive drinking reported an advantage of medication over placebo in reducing excessive or heavy drinking (70% of the studies that measure heavy drinking). Limitations There are some limitations of our methods that bear noting. We chose not to conduct a meta-analysis. At this time, well-conducted meta-analyses of naltrexone treatment of are available in the literature, and each of these supports the use of naltrexone clinically to reduce alcohol consumption in with alcohol problems, 9,10,101 although the effect sizes reported by these meta-analyses are modest. Our technique allowed us to simultaneously consider more trials than previous metaanalyses have been able to include. The most inclusive metaanalysis extant in the literature pools data from 9 trials (less than a third of the studies included in this review). 9 By including more trials, we have been able to present a more thorough summary of the literature and evaluate more types of outcome measures. The strength of this technique is that it has allowed us to be more comprehensive, but at the expense of employing more rigorous statistical methods, which might inspire stronger conclusions. There are some key differences between studies included in this review. An important factor on which these studies vary is dosing and administration of the medication. Three of the studies included in this review used a depot, as opposed to oral, administration of naltrexone As discussed earlier, long-acting injectable naltrexone may offer advantages over the orally taken pill form of the medication. There are also dosing inconsistencies among studies. Interestingly, a true dosing study was never conducted as part of the development of naltrexone. Variability in k- and d-receptor antagonism and intraindividual metabolism have led some investigators to consider dose to be important in understanding variability in outcomes. Although this is a reasonable hypothesis, only a few published studies prescribed more than 50 mg/d. Some investigators have posited that the discrepancy between studies with positive and negative results could be explained in part by the goal of the concomitant psychosocial intervention that typically improved with adjunctive medication. 84 Evidence supports that naltrexone is effective in settings that use a coping approach (ie, an approach focused on discontinuing drinking), more so than settings that exclusively focus on the patient achieving complete abstinence. 84 Most of the studies in this review had relatively similar study populations. Typically, subjects were treatment seeking, community-referred who met diagnostic (DSM) 102 criteria for, and no other substance dependence. However, there was 1 of the 29 studies with alcohol-dependent patient populations that is distinctively different from those included in the majority of the clinical trials, namely, with cocaine dependence. 87 This study found no advantage of 50 mg/d naltrexone for 12 weeks (over placebo) in reducing drinking for with both alcohol and cocaine dependence. This result is not surprising, given that alcohol-dependent with co-occurring cocaine dependence are more complicated at treatment entry, have higher attrition rates, and a poorer prognosis than with only alcohol dependence. 103,104 In addition, most of the naltrexone studies reported in this review excluded who were taking or needed to take a psychotropic medication for a concurrent psychiatric illness. However, some of the naltrexone studies did include with a concurrent psychiatric if they either did not need to take a medication or were stabilized for some time on a psychotropic medication. Although there is one controlled study that reported that naltrexone did not improve outcome more than placebo when added to an antidepressant (sertraline) in older adults with concurrent n 2006 Lippincott Williams & Wilkins 615

7 Pettinati and Associates Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 TABLE 1. Summary of 29 Randomized, Double-blind, Placebo-controlled Trials of Naltrexone or Nalmefene ( ) Effects on Drinking In-trial Naltrexone Conclusions Any Drinking Excess Drinking Double-blind Studies Subjects Target Population Psychosocial Study Design/Dosage Medication Adherence Treatment Completion Days Drinking (n or %) Abstinent Relapse Days Heavy Drinking Any Drinking Excess Drink Volpicelli et al, 26, O Malley et al, Mason et al, Oslin et al, N = 70 alcoholdependent veterans (21 65 yrs) N = 97 alcoholdependent (18 65 yrs) N = 21 alcoholdependent (18 65 yrs) N = 44 older alcoholdependent veterans (50 70 yrs) DSM-III-R No drug use in past 30 d DSM-III-R DSM-III-R DSM-III-R No drug use in 6 wk 1 4 wk: inpatient treatment 5 outpatient group 2/wk coping skills or supportive therapy 1/wk referred to Alcoholic Anonymous and other psychosocial treatments group therapy 1/wk; case management 2/mo 50 mg/d NAL or PLAC 50 mg/d NAL or PLAC NR NAL, 69% * NS * NAL > PLAC NAL > PLAC PLAC, 60% NAL, 92% 70% >10 treatment PLAC, 78% weeks: NAL/coping, 65.6% PLAC/coping, 60.0% NAL/support, 78.4% PLAC/support, 59.3% * * * NAL > PLAC for coping and support NAL > PLAC for coping and support 100% 38%: NS * NALM = PLAC NALM > PLAC mg/d NALM, 43% NALM PLAC, 29% mg/d NALM 3. PLAC NAL, 97.9% 61% NS NS NAL = PLAC NAL = PLAC 100 mg/d NAL PLAC, 94.5% or PLAC on Mondays and Wednesdays 150 mg/d NAL or PLAC on Friday Volpicelli et al, Hersh et al, Kranzler et al, Anton et al, N = 97 alcoholdependent (21 65 yrs) N = 64 comorbid alcohol-cocaine dependent (18 45 yrs) N = 20 alcoholdependent (18 60 yrs) N = 131 alcoholdependent (21 65 yrs) DSM-III-R No drug use in past 30 d DSM-III-R both cocaine and alcohol abuse/dependence No intravenous cocaine use DSM-IV DSM-III-R (first episode) No drug abuse/dependence 1 4 wk: relapse 50 mg/d NAL prevention 2/wk or PLAC 5 relapse prevention 1/wk 8 wk: 1 2/wk of individual relapse prevention treatment 8 wk: coping skills treatment 1/wk CBT 1/wk 8 wk: 50 mg/d NAL, 58%; 73%: * * NAL > PLAC NAL > PLAC PLAC, 47% NAL, 73% PLAC, 73% NAL, 97.9% 61%: NS NAL = PLAC NR PLAC, 94.5% NAL, 64.5% PLAC, 57.6% 8 wk: Not applicable NR NS * NAL = PLAC NAL > PLAC 2 wk: NAL, 50 mg/d 2 wk: washout 4 wk: injection, (206 mg NAL or PLAC) 50 mg/d NAL or PLAC NAL, 69% 83%: * NS * NAL > PLAC NAL > PLAC PLAC, 67% NAL, 87% PLAC, 78% Mason et al, N = 105 alcoholdependent (18 65 yrs) DSM-III-R CBT 1/wk 86.1% 89.9% NALM (2 groups mg/d NALM added), 64.3% mg/d NALM PLAC, 65.7% 3. PLAC NS * NALM = PLAC NALM > PLAC 616 n 2006 Lippincott Williams & Wilkins

8 Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 Naltrexone Effects on Heavy Drinking Chick et al, 13 United Kingdom, 2000 Kranzler et al, Heinala et al, 91 Finland, 2001 Krystal et al, 11 United States (15 sites), 2001 Lee et al, 92 Singapore, 2001 Monterosso et al, Monti et al, Morris et al, 93 Australia, 2001 Ahmadi and Ahmadi, 94 Iran, 2002 Gastpar et al, 84 Germany (7 sites), 2002 Guardia et al, 95 Spain (7 sites), 2002 N = 175 alcoholdependent (18 65 yrs) N = 183 alcoholdependent (18 60 yrs) N = 121 alcoholdependent (21 65 yrs) N = 627 alcoholdependent veterans (18 yrs +) N = 53 alcoholdependent male (21 65 yrs) N = 183 alcoholdependent N = 128 alcoholdependent (18 yrs +) N = 111 alcoholdependent male (18 65 yrs) N = 116 alcoholdependent male (23 56 yrs) N = 171 alcoholdependent in or out N = 202 alcoholdependent (18 60 yrs) DSM-III-R alcohol abuse or dependence DSM-III-R usual psychosocial treatment for center-not specified coping skills treatment 1/wk 50 mg/d NAL or PLAC Attended all visits and took >80% medications: NAL, 39% PLAC, 41% NAL, 66% 70.5%: NAL, 59.0% mg/d NAL NEFAZ, 76.8% 2. Up to 200 mg/d PLAC, 79.8% BID NEFAZ 3. PLAC NAL, 41% NS NS NS NAL = PLAC NAL = PLAC PLAC, 42% NEFAZ, 72.9% PLAC, 79.4% NS NS NS NS NAL = PLAC NAL = PLAC DSM-IV coping skills or abstinence support treatment 1/wk 20 wk: targeted medications DSM-IV 60 wk: 12-step facilitation therapy 1 4 mo: 1/wk 5 9 mo: 2/wk mo: 1/mo 32 wk: NR 83.5% completed 12 wk NAL 50 mg/d 69.4% completed or PLAC 32 wk 52 wk: 3 group comparison, 50 mg/d NAL or PLAC: 1. NAL for 12 mo 2. NAL for 3 mo then PLAC for 6mo 3. PLACfor12mo * NR NAL coping > PLAC coping 73% NS NS NAL = PLAC NAL = PLAC NAL (2 groups added), 72% days PLAC, 70% days 52 wk: NAL, 12, 44% days NAL, 3, 43% days PLAC, 42% days DSM-IV 4 wk: inpatient treatment 5 8 wk: psychiatrist visits 2/wk plus encouraged to attend support groups 50 mg/d NAL or PLAC (4 wk inpatient/8 wk outpatient) NR NAL, 40% NS NAL = PLAC NR PLAC, 22% DSM-IV MM 1/wk 100 mg/d NAL or PLAC % 82.1% * NR NAL > PLAC DSM-IV DSM-III-R Comorbid disorders included 3 mo stability on psychotropics Alcohol dependence ( criteria unspecified) No drug abuse 2 wk +: Partial hospital 50 mg/d program daily for 2 wk Aftercare observed for 12 wk group 1/wk individual counseling 1/wk 50 mg/d 50 mg/d 59% 91% NS * NR NAL > PLAC NAL, 98% 64%: NS * * NAL > PLAC NAL > PLAC PLAC, 97% NAL, 69% PLAC, 57% NR 61%: * NR NAL > PLAC NAL, 79.3% PLAC, 43.1% DSM-III-R No drug abuse DSM-IV at least 1 h/wk of nonspecific psychosocial treatment supportive group therapy 1/wk 50 mg/d 50 mg/d >80% NAL, 67% NS NS NAL = PLAC NAL = PLAC PLAC, 62% NAL, 78.4% NAL, 60.4% NS * NAL = PLAC NAL > PLAC PLAC, 78.8% PLAC, 58.4% n 2006 Lippincott Williams & Wilkins 617

9 Pettinati and Associates Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 TABLE 1. Continued Effects on Drinking In-trial Naltrexone Conclusions Any Drinking Excess Drinking Double-blind Studies Subjects Target Population Psychosocial Study Design/Dosage Medication Adherence Treatment Completion Days Drinking (n or %) Abstinent Relapse Days Heavy Drinking Any Drinking Excess Drink Latt et al, 96 Australia, 2002 Balldin et al, 97 Sweden (10 sites), 2003 Kiefer et al, 98 Germany, 2003 Kranzler et al, Anton et al, N = 107 alcoholdependent (18 70 yrs) N = 118 alcoholdependent (18 65 yrs) N = 160 alcoholdependent (18 65 yrs) N = 153 alcoholdependent (18 60 yrs) N = 270 alcoholdependent (21 yrs +) DSM-IV DSM-IV counseling and supportive behavioral treatment offered (not required), frequency not specified 24 wk: CBT or supportive 1 12 weeks: 1 DSM-IV every 2 wk weeks: 1 every 4 wk group CBT 1/wk 50 mg/d 24 wk: 50 mg/d 4 treatment groups mg/d NAL + PLAC mg/d ACAM + PLAC 3. PLAC only mg/d NAL and 1998 mg/d ACAM DSM-IV DSM-IV 8 wk: coping skills treatment 2/mo MET 4/12 wk NR 63%: NS * NAL = PLAC NAL > PLAC NAL, 59% PLAC, 67% 78%: 77% NS * * NAL = PLAC NAL > PLAC NAL/CBT, 84%; PLAC/CBT, 77% NAL/supportive, 77% PLAC/supportive, 75% 81.10% 46.9% * * NAL > PLAC NAL > PLAC 8 wk: 86%: 86%: * * NAL > PLAC mg/d NAL Daily NAL, 84.0% Daily NAL, 87.9% (daily) 2. PLAC/d Daily PLAC, 84.6 Daily PLAC, 87.2% Targeted NAL, 86.7% Targeted NAL, 88.1% mg NAL targeted 4. PLAC targeted (targeted = only in high-risk situations pill availability <1 pill/wk) 4 treatment groups 1. 5 mg/d NALM No drug abuse or dependence mg/d NALM mg/d NALM 4. PLAC/d Targeted PLAC, 88.6% >90% no difference between groups Targeted PLAC, 80.6% NAL > PLAC (daily) 74%: NS NS NALM = PLAC NALM = PLAC 5 mg NALM, 78% 20 mg NALM, 77% 40 mg NALM, 69% PLAC, 72% 618 n 2006 Lippincott Williams & Wilkins

10 Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 Naltrexone Effects on Heavy Drinking Killeen et al, Kranzler et al, Garbutt et al, Anton et al, N = 145 alcoholdependent N = 315 alcoholdependent (18 65 yrs) N = 627 alcoholdependent (18 yrs +) N = 1383 alcoholdependent DSM-IV No opiate dependence DSM-IV of DSM-IV of treatment as usual, included 12-step facilitation and relapse prevention (intensity and frequency varied) MET 5/12 wk 24 wk: supportive counseling 1/2 wk 3 treatment groups mg/d NAL 2. PLAC/d 3. No medication (treatment as usual) 150 mg/4 wk injection (at initial session subjects in NAL group receive mg injections) 24 wk: 3 treatment groups 190 mg or DSM-IV of 16 wk: 16 wk: 9 treatment groups 1. MM (4 grps) 2. CBI (1 grp) 3. MM + CBI (4 grps) 380 mg/4 wk NAL injection PLAC injection/ 4wk NAL dose = 100 mg/d ACAM dose = 1998 mg/d 9 treatment groups MM only or with CBI plus: 1 2. NAL+PLAC 3 4. ACAM + PLAC 5 6. NAL + ACAM or 7 8. PLAC + PLAC or 9. CBI plus no medications 58.62%: 72%: NS NS NAL = PLAC NAL = PLAC NAL, 53% NAL, 76% PLAC, 67% PLAC, 67% No medication, 75% Not Applicable 77.78%: * NS NAL > PLAC NAL =PLAC NAL, 80.4% PLAC, 75.2% Not applicable 64%: NS * NAL = PLAC NAL > PLAC 85.8% no significant difference between groups 380 mg NAL, 62.5% 190 mg NAL, 65.2% PLAC, 64.1% 94% research completion: no significant difference between groups * * * NAL > PLAC NAL > PLAC *There was a statistical advantage for the medication over placebo. ACAM indicates acamprosate; CBT, cognitive behavioral therapy; CBI, Combined Behavioral Intervention; DSM-III-R, Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition; MM, medical management; NAL, naltrexone; NAL or NALM > PLAC, medication has an advantage over placebo in at least 1 of 2 measures within drinking category ( any drinking or excess drinking ); NAL or NALM = PLAC, medication has no advantage in both measures within a drinking category; NALM, nalmefene; NEFAZ, nefazodone; NR, not reported; NS, there was no statistical advantage for the medication over placebo; PLAC, placebo;, no report on this variable in the published article. n 2006 Lippincott Williams & Wilkins 619

11 Pettinati and Associates Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 TABLE 2. Summary of Drinking Outcomes From 29 Randomized, Double-blind, Placebo-controlled Trials of Naltrexone or Nalmefene ( ) Any Drinking In-trial Excessive Drinking In-trial Study No. or % Days Drinking P % Subjects Abstinent or No. Days Abstinent (if days drinking was not reported) P % Relapsed or Time to Relapse (d) P No. or % Days of Heavy Drinking P Volpicelli et al, 26,27 NAL, 1.6% <0.025 NAL, 54% NS NAL, 23% <0.01 NR PLAC, 8.3% PLAC, 43% PLAC, 54% O Malley et al, 28 NAL/coping, 4.2% NAL/coping, 28% NAL/coping, NR PLAC/coping, 10.4% PLAC/coping, 21% PLAC/coping, 25 d NAL/supportive, 4.3% NAL/supportive, 61% NAL/supportive, NR PLAC/supportive, 9.4% PLAC/supportive, 19% PLAC/supportive, 35 d NAL vs. PLAC (main effect) <0.01 NAL/sup vs. PLAC/coping 0.01 NAL vs. PLAC (main effect) <0.05 NR Mason et al, 85 NR NALM (40 mg), 4.7 d NALM (40 mg), 29% NALM (10 mg), 2.4 d NALM (10 mg), 100% PLAC, 4 d PLAC, 67% NALM (40 mg) vs. PLAC NS NALM (40 mg) vs. PLAC <0.05 NR Oslin et al, 86 NAL, 1.9% NR NAL, 14.3% NR PLAC, 6.5% PLAC, 34.8% Volpicelli et al, 12 NAL, 2.83% 0.01 NR NAL, 14.3% 0.02 NR PLAC, 11.02% PLAC, 52.2% For treatment adherent For treatment adherent Hersh et al, 87 NAL, 17.9% 0.74 NR NR NR PLAC, 19.8% Kranzler et al, 80 NAL, 21% 0.42 NR NR NAL, 3.7% 0.03 PLAC, 28.4% PLAC, 5.3% Anton et al, 88 NAL, 10% 0.03 NAL, 47% 0.11 NAL, 38% 0.01 NR PLAC, 18% PLAC, 33% PLAC, 60% NAL, 60 d 0.02 PLAC, 48 d Mason et al, 89 NR NALM, 80 d 0.43 NALM (2 groups 0.02 NR PLAC, 83 d added), 37.1% PLAC, 58.8% Chick et al, 13 NAL, 12 d NS NAL, 18% NS NAL, 50% NS NR PLAC, 20 d PLAC, 19% PLAC, 65% For treatment adherent For treatment adherent Kranzler et al, 90 NAL, 28.9% NAL, 29.5% NAL, 7.2 wk NAL, 21.4% PLAC, 21.3% PLAC, 34.9% PLAC, 8 wk PLAC, 16.3% NEFAZ, 23.1% NEFAZ, 32.2% NEFAZ, 7.2 wk NEFAZ, 23% NAL vs. PLAC 0.43 NAL vs. PLAC NS NAL vs. PLAC 0.84 NAL vs. PLAC 0.43 Heinala et al, 91 NR NR NAL/coping, 73.5% <0.008 NR PLAC/coping, 97% NAL/supportive, 93% NS PLAC/supportive, 88% Krystal et al, 11 3-mo outcomes NAL (2 groups), 11.3% NS NR NAL (2 groups), 37.8% NS NR PLAC, 14.0% PLAC, 44.4% Lee et al, 92 NR NAL, 67% NS NR NR PLAC, 47% Monterosso et al, 20 NR NR NR NAL, 5% days 0.04 PLAC, 8.9% days Monti et al, 14 NR NR NAL, 28% NS NAL, 0.5% <0.03 PLAC, 32% PLAC, 1.7% For treatment adherent 620 n 2006 Lippincott Williams & Wilkins

12 Study Any Drinking In-trial Excessive Drinking In-trial No. or % Days Drinking P % Subjects Abstinent or No. Days Abstinent (if days drinking was not reported) P % Relapsed or Time to Relapse (d) P No. or % Days of Heavy Drinking P Morris et al, 93 NAL, 25% NS NAL, 21% 0.04 NAL, 50% 0.04 NR PLAC, 36% PLAC, 12% PLAC, 79% For treatment adherent For treatment adherent For treatment adherent Ahmadi and NR NR NAL, 79.3% <0.001 NR Ahmadi, 94 PLAC, 43.1% Gastpar et al, 84 NR NAL, 54% NS NAL, 38%* NS NR PLAC, 51% PLAC, 38% Guardia et al, 95 NR NAL, 65.3% days 0.56 NAL, 7.9% 0.05 NR PLAC, 63% days PLAC, 18.8% Latt et al, 96 NAL, 2.20 d/wk NS NR NAL, 33.9% 0.04 NR PLAC, 2.26 d/wk PLAC, 52.9% Balldin et al, 97 NAL/CBT, 32% NR NAL/CBT, 57 d <0.007 NAL/CBT, 21% PLAC/CBT, 47% PLAC/CBT, 22 d PLAC/CBT, 38% NAL/supportive, 44% NAL/supportive, 20 d NS NAL/supportive, 34% PLAC/supportive, 50% PLAC/supportive, 18 d PLAC/supportive, 39% NAL vs. PLAC (main effect) <0.109 NAL vs. PLAC (main effect) Kiefer et al, 98 NR NAL, 35% NAL, 48% NR PLAC, 7% PLAC, 75% NAL + ACAM, 58% NAL + ACAM, 35% ACAM, 27% ACAM, 60% NAL vs. PLAC 0.03 NAL vs. PLAC 0.02 Kranzler et al, 82 Daily NAL, 0.72 Targeted NAL, 0.62 PLAC groups, NR NR NR Daily NAL, 0.16 Daily PLAC, Targeted NAL, 0.22 Targeted PLAC, 0.21 Anton et al, 99 NR All groups, 57.1% days NS NALM (all doses), 9 d NS NR PLAC, 4 d Killeen et al, 100 NAL, 8% NS NR NR NAL, 4 NS PLAC, 10% PLAC, 5 Kranzler et al, 81 NR NAL, 18% NAL, 77% NS NR PLAC, 10% PLAC, 84% Garbutt et al, 79 NR NAL 380 mg, 7% NS NR NAL 380 mg 0.03 NAL 190 mg, 6% PLAC, 5% vs. PLAC, 25% reduction in heavy drinking days NAL 190 mg vs. PLAC, 17% reduction in heavy drinking days <0.045 NS NS Anton et al, 3 Main effect NR Main Effect only P value reported NAL, 21.2% NAL, 68.2% ACAM, 21.6% ACAM, 69.6% PLAC, 21.2% PLAC, 71.4% NAL vs. PLAC (main effect) 0.25 NAL vs. PLAC (main effect) 0.02 NAL Psychosocial interaction NAL Psychosocial interaction NAL (nocbi), 19.4% NAL (no CBI) 66.2% PLAC (no CBI), 24.9% PLAC (no CBI) 73.1% NAL plus CBI, 22.9% NAL plus CBI 70.2% PLAC plus CBI, 20.8% PLAC plus CBI, 69.7% NAL CBI (interaction effect) NAL CBI (interaction effect) *Significantly lower g-glutamyl transpeptidase compared with the PLAC group. y There is a statistical advantage for the medication over placebo. ACAM indicates acamprosate; NAL, naltrexone; NALM, nalmefene; NR, not reported; NS, no statistical advantage for the medication over placebo; PLAC, placebo; CBI, combined behavioral intervention. y y Journal of Clinical Psychopharmacology Volume 26, Number 6, December 2006 Naltrexone Effects on Heavy Drinking n 2006 Lippincott Williams & Wilkins 621