Medication combined with brief, manual-based intervention therapy

Researchers in Finland have found disulfiram to be superior to both naltrexone and acamprosate in the first randomized comparison of disulfiram, naltrexone, and acamprosate with brief, manual-based intervention therapy (Laaksonen, Koski-Jännes, Salaspuro, Ahtinen, & Alho, 2007). This open-label, naturalistic trial, which took place at six different alcohol treatment and healthcare units, involved 243 Caucasian subjects who met the International Statistical Classification of Diseases and Related Health criteria for alcohol dependence; the subjects were voluntarily seeking treatment. Patients visited a physician at scheduled intervals during two phases that lasted 52 weeks; about 67 weeks after completing the study, patients were contacted for followup information—a total of 2.5 years after starting the study. This study experienced a low rate of dropout (25.1 percent after 12 weeks and 51.8 percent by the end of the 52-week study period).

During the Phase 1 continuous medication period (weeks 1–12), patients designated a contact person to be responsible for supervising and controlling their daily study medication. Patients were randomly assigned to receive (1) 50 mg of naltrexone once a day, (2) 666 mg of acamprosate 3 times a day or 1,333 mg/day for those weighing less than 60 kg, or (3) 100–200 mg of disulfiram once a day or 2 tablets (400 mg) twice a week. At regular visits throughout the 52 weeks, patients brought in a diary of their alcohol consumption and medication intake and also used a manual with homework based on cognitive-behavioral principles. During the Phase 2 targeted medication period (weeks 13–52), patients were asked to take a daily medication dose in any “craving situation” when they perceived their propensity to drink was high. During this targeted medication phase, the study center no longer provided free medication.

The main conclusion of this study was that all three medications—disulfiram, naltrexone, and acamprosate, combined with brief manual-based intervention extended over time—significantly reduce heavy drinking, reduce craving for alcohol, and increase the quality of life. However, disulfiram was superior to the other medications, especially during the continuous medication period. The study data show that, during Phase 1, disulfiram was significantly better than naltrexone and acamprosate as follows:

• In time to first heavy drinking day—46.6±27.5 days for disulfiram versus 22.0±22.0 for naltrexone and 17.6±22.0 for acamprosate:
• In time to first drink—30.4±27.8 days for disulfiram versus 16.2±20.2 for naltrexone and 11.4±17.0 for acamprosate
• In average weekly alcohol intake (g/ethanol per week)—52.0±90.7 for disulfiram versus 183.7±174.1 for naltrexone and 203.2±180.2 for acamprosate.

During Phase 2, the targeted medication period, there were no significant differences among the three medication groups on time to first heavy drinking day or in days to first drinking. Average alcohol consumption in all groups remained significantly below the baseline. However, those in the disulfiram group had significantly more abstinence days than those in the other two groups.

Medication for patients with co-occurring alcohol dependence and depression

Petrakis and colleagues (2007) conducted a secondary analysis of 139 male veterans with alcohol dependence and current major depression, assessing the effectiveness of naltrexone and disulfiram in this population. As in their original large-scale study comparing naltrexone and disulfiram among 254 male veterans who had alcohol dependence and comorbid mental disorders (Petrakis et al., 2005), they found no advantage of one medication over the other. In comparison with outcomes found in the original study, a person’s having a diagnosis of co-occurring depression had no significant effect on retention in treatment or on drinking outcomes, including maximum consecutive days of abstinence, percentage of heavy drinking days, or abstinence throughout the entire study period. Petrakis and colleagues (2007) concluded that both disulfiram and naltrexone are safe pharmacotherapeutic agents for treating alcohol dependence in dually diagnosed individuals with depression.

As in the original study, an unexpected finding was that patients with depression who received disulfiram reported lower craving over time than subjects with depression who received naltrexone (Petrakis et al., 2007). (For additional information on this study, see New Findings on Combined Medication Therapy below.)

Effects of patient compliance on outcomes

Research suggests that the effectiveness of naltrexone in clinical trials—and probably also in clinical treatment—can be greatly influenced by the subjects’ adherence to the medication. A new study reanalyzed data, expanding the variable drinking outcomes reviewed, from an alcohol treatment trial involving 160 participants (Anton et al., 2005). This reanalysis looked specifically at how much patient compliance with naltrexone influenced the outcomes and compared two methods for measuring compliance (Baros, Latham, Moak, Voronin, & Anton, 2007). The researchers conclude that, because patient adherence to naltrexone has such a large influence on treatment outcomes, practitioners need to give utmost attention to methods for enhancing their patients’ compliance. The article also summarizes evidence from the literature on strategies to use for improving compliance (Baros et al., 2007).

This study evaluated outcomes for 137 randomized patients with alcohol dependence who completed 12 weeks of naltrexone or placebo, combined with either cognitive-behavioral therapy (CBT) or motivational enhancement therapy (MET). Compliance was monitored and compared using urine riboflavin measurements during study weeks 2, 6, and 12, as well as a medication event monitoring system (MEMS) that provided a detailed computerized record of when patients opened their medication bottles. Findings included the following:

• Accounting for adherence and compliance with naltrexone changed the outcomes (not significant in the original study) to demonstrate a significant drug therapy interaction for percentage of days abstinent, number of heavy drinking days, or total standard drinks.
• MEMS and urine riboflavin measures of compliance provided similar estimates of treatment effectiveness, although combining these two measures yields the most conservative, stringent index of medication compliance.
• The size of the treatment effect approximately doubled in the most compliant individuals.
• Patients treated with naltrexone and CBT showed more days of abstinence, less relapse to heavy drinking days, and fewer total drinks than the other groups (those receiving naltrexone plus MET or placebo plus psychotherapy).
• Older age predicted pill-taking compliance.

Effects of naltrexone on specific subgroups or populations

The treatment effects of naltrexone have recently been examined in several alcohol-dependent subgroups, including (1) those with a family history of alcohol problems and/or antisocial traits, (2) individuals with at least one copy of the G allele of the OPRM1 gene, and (3) women, including those with a comorbid eating disorder.

1. Family history of alcohol problems and/or antisocial traits Prior research has suggested that naltrexone may be significantly more effective in moderating heavy drinking among patients with certain characteristics, including a family history of drug problems, early onset of alcohol problems, high degree of antisocial traits, and comorbid drug use. Rohsenow, Miranda, McGeary, & Monti (2007) tested the contribution of these factors to naltrexone effectiveness by reanalyzing data from 128 patients with alcohol dependence enrolled in a 12-week, double-blind clinical trial of naltrexone. Participants in the original study had been recruited from a substance abuse day treatment program (Monti et al., 2001). Findings of the reanalysis included the following:
   • Having a high percentage (at least 20 percent) of first- and second-degree family members with problem drinking significantly affected naltrexone’s effects, resulting in lower drinking rates among those patients
   • Having more antisocial traits resulted in less heavy drinking on naltrexone than on placebo when the patient took at least 70 percent of the medication, whereas more socialized patients had no benefit from naltrexone regardless of compliance (degree of socialization was measured using the California Personality Inventory Socialization scale).
   • Age of onset of alcohol problems and comorbid cocaine or marijuana use had no interaction effect with the medication.
2. This study suggests that a meaningful (and inexpensive) way to match patients to naltrexone is by identifying those who have 20 percent or more relatives with alcohol problems and/or have high antisocial scores. Patients with alcohol dependence who use marijuana or cocaine can also be benefited by naltrexone.
3. Effects of naltrexone on alcohol sensitivity and genetic modulators of medication response. To better understand naltrexone’s mechanisms of action, particularly its effects on alcohol sensitivity and craving in persons with the A118G single nucleotide polymorphism (SNP) of the OPRM1 gene, Ray and Hutchison (2007) studied naltrexone’s effects in a within-subject, double-blind, placebo-controlled laboratory trial. Naltrexone was found to blunt alcohol’s effects on subjective feelings of stimulation, positive mood, craving, enjoyment, and vigor. This study suggests that, during treatment with naltrexone, carriers of the A118G SNP (the G allele) experience a more pronounced reduction in alcohol reward than others, which may explain the lower relapse rate with naltrexone treatment in these individuals. Naltrexone’s dampening of the rewarding subjective effects of alcohol may thus reduce the likelihood that a slip will trigger a full-fledged relapse into heavy drinking for carriers of the G allele.
   This laboratory study tested the effects of naltrexone in 40 subjects who drank heavily (15 of whom had at least one copy of the G variant and 25 who were homozygous for the A allele) after a 5-percent ethanol solution was infused intravenously. Findings included the following:
   • Naltrexone was differentially effective based on the individual’s genotype; it significantly reduced the self-reported, alcohol-induced high in participants with at least one copy of the G allele but had no effect on participants who were homozygous for the A allele.
   • After taking naltrexone, subjects with the G allele demonstrated greater blunting of the alcohol-induced high when the breath alcohol concentration (BrAC) reached 0.06 mg/L, with greatest effects at highest BrAC. This suggests that the effects of naltrexone may be alcohol-dose dependent (Ray & Hutchison, 2007).
4. Effects of naltrexone on women who are alcohol dependent. Relatively little is known about the efficacy of naltrexone for treating women who are alcohol dependent. The Women’s Naltrexone Study investigated the safety and efficacy of naltrexone combined with cognitive-behavioral coping skills therapy (CBCST) in 103 women who are alcohol dependent, 29 of whom had comorbid eating pathology (O’Malley et al., 2007b). Participants were randomized to receive weekly group CBCST plus either naltrexone (50 mg) or placebo for 12 weeks.
   This study, which essentially measured the added effect of naltrexone over CBCST, found no significant differences between naltrexone and placebo on the primary study outcomes: time to first drinking day, time to first day of heavy drinking, or percentage of participants who continued to meet criteria for alcohol dependence. It should be noted that all groups in this study showed large improvements in drinking behavior: overall, the percentage of days abstinent during treatment more than doubled from baseline, and the number of drinks during drinking day decreased from 7.12 drinks at baseline to 1.83 drinks during treatment. However, positive outcomes specifically with naltrexone included the following:
   • Naltrexone significantly delayed the time to second and third drinking days for women who did not maintain abstinence from alcohol.
   • Symptoms of eating pathology decreased during treatment among all groups (e.g., the frequency of binge eating decreased by almost 70 percent). This suggests that treatment for alcohol dependence may be associated with improvements in eating pathology (O’Malley et al., 2007b).
5. The outcomes of this study have been reexamined in a reanalysis of the data, described in the following study.

Combining naltrexone and disulfiram for patients with co-occurring alcohol dependence and depression

Petrakis and colleagues (2007) concluded that the combined use of naltrexone and disulfiram offered no advantage over either medication alone for subjects who are alcohol dependent with co-occurring major depression. This secondary analysis of data from a large randomized controlled trial looked at 139 male veterans who were alcohol dependent with a Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1994) diagnosis of current major depression. These subjects represented 54.7 percent of the 254 participants in the original study of veterans with alcohol dependence and comorbid mental disorders (Petrakis et al., 2005).

This secondary analysis showed that subjects with current co-occurring depression achieved positive outcomes comparable to all subjects in the trial—a trial in which almost 70 percent of subjects achieved complete abstinence during the 12-week study period. In addition, on the Hamilton Depression Rating Scale, these subjects with depression also showed a significant decrease in depression from baseline to posttreatment. Side effects of the combined medications, as well as naltrexone and disulfiram alone, were tolerated and consistent with those seen in patients who do not have a dual diagnosis. Because there was no advantage to the combined medication or to one pharmacotherapeutic agent over another, the choice of medication to treat AUDs in patients with depression can depend on such factors as patient preference (Petrakis et al., 2007).

Combining medication with psychosocial treatment

In a review article designed for physicians, Weiss and Kueppenbender (2006) describe the significant advances made in the development, standardization, and rigorous testing of the psychotherapeutic approaches used to treat alcohol dependence. Medical management interventions are available to physicians, as are strategies for improving medication adherence. The authors discuss the evidence from the literature since 1984, particularly clinical trials, on the interactions and efficacy of disulfiram, oral naltrexone, and acamprosate with particular psychosocial treatments. Weiss and Kueppenbender (2006) recommend that physicians use these medical management techniques when prescribing pharmacologic agents to patients with alcohol dependence. They also recommend that physicians become knowledgeable about the various psychotherapies as background for referring their patients who are alcohol dependent to concurrent psychosocial treatment. Some of the conclusions made by Weiss and Kueppenbender (2006), based on their review of the literature, include the following:

• Adding medical management therapy and pills to a specialty psychosocial therapy improves outcomes for patients who are alcohol dependent.
• Psychosocial interventions, ranging from brief medical management to more intensive manual-based psychotherapies, have all been shown to produce positive outcomes in certain studies, depending on the specific medication and the study context.
• No evidence suggests that one single form of psychosocial treatment is a criterion standard for patients with alcohol dependence who receive pharmacotherapy.
• For disulfiram, a successful and promising adjunctive approach is behavioral marital therapy augmented with a disulfiram contract by the couple.
• For naltrexone, the evidence suggests (although not conclusively) that CBT may be particularly effective as adjunct therapy.
• For acamprosate, few studies have been done in combination with structured, controlled psychosocial interventions. The limited evidence suggests that acamprosate may be used equally effectively with a variety of psychosocial treatments and that little psychosocial treatment may be needed beyond medical management.

Building on the existing literature about psychosocial approaches combined with pharmacotherapy, two recent articles describe additional possible psychotherapy approaches to augment the medication. The proposed therapies, described below, include (1) the trial of a second-generation CBT combined with oral naltrexone and (2) the use of contingency management (CM) with medications for treating substance abuse.

Broad-spectrum treatment (BST). A 3-month, randomized controlled trial explored whether a broad-spectrum CBT would be more effective than MET for patients who are alcohol dependent treated with naltrexone (Davidson, Gulliver, Longabaugh, Wirtz, & Swift, 2007). This initial trial suggests that, at least when combined with naltrexone, a second-generation CBT may have a meaningful clinical advantage over brief interventions such as MET (Davidson et al., 2007).

This research group developed a unique CBT manual-based protocol for alcohol dependence that combined components of the three psychotherapies demonstrated to be effective in NIAAA’s Project MATCH: CBT, 12-Step facilitation, and MET. This new BST approach incorporates such content material as cognitive restructuring, drink refusal, and assertiveness training with a patient-specific selection of session modules. The treatment matching uses a decision tree, with modules tailored for the individual through a psychometric assessment of each patient’s need.

In this study, 149 patients with alcohol dependence were randomly assigned to receive either BST and naltrexone or MET and naltrexone. Patients who received BST had a significantly higher percentage of days abstinent than patients receiving MET. Treatment was tailored in response to an assessment of the patients’ psychosocial resources, and the differential advantage for BST was most marked for those patients with social networks that supported drinking.

Contingency management. Clinical trials have demonstrated the effectiveness of CM procedures—an approach in which patients receive concrete rewards or reinforcers for discrete targeted behaviors. However, few trials have assessed the value of CM procedures when combined with pharmacotherapy for alcohol dependence. Carroll and Rounsaville (2007) review the existing evidence and suggest that CM would be an ideal platform for addressing the weaknesses of many pharmacotherapies used to treat drug abuse. CM can directly reinforce medication adherence, which may substantially improve compliance in treatment where unpleasant side effects must be overcome or where compliance is not strongly reinforced by rapid benefits from the treatment itself. The authors describe a variety of CM strategies used to improve compliance with disulfiram among patients with alcohol dependence.

A recent pilot double-blind trial of memantine—a selective noncompetitive N-methyl-D­aspartate (NMDA) receptor antagonist—among 34 individuals with alcohol dependence did not support the use of memantine for treating patients who are actively drinking. However, this study did support the use of voucher incentives to facilitate retention. With voucher incentives for clinic attendance, 80 percent of subjects completed the 16-week trial (Evans, Levin, Brooks, & Garawi, 2007).

Topiramate

Recent research shows that topiramate, a drug with complex actions that include activity at the GABA and glutamate receptors, is a promising treatment for alcohol dependence. Although only two major studies have been conducted, the consistency and size of topiramate’s clinical efficacy suggest the need for further research, particularly on the most efficacious ceiling dose, the impact of longer periods of treatment, and the subtypes of alcoholism most benefited by treatment with topiramate (Johnson et al., 2007).

A 17-site, double-blind, placebo-controlled trial with 371 men and women who were alcohol dependent found that up to 300 mg per day of topiramate reduced the percentage of heavy drinking days from baseline to week 14 and produced significant and meaningful improvement in a wide variety of self-reported drinking outcomes (Johnson et al., 2007). Topiramate compared with placebo treatment was associated with a significantly higher rate of achieving 28 or more days of continuous nonheavy drinking and 28 or more days of continuous abstinence. Furthermore, using two different analytic approaches, the topiramate group reached 28 or more days of continuous abstinence significantly faster than the placebo group (Johnson et al., 2007). Topiramate also decreased plasma GGT in the heterogeneous and graphically diverse population. These positive findings replicated the results of a smaller, randomized controlled trial (Johnson et al., 2003b). Topiramate’s therapeutic effect was evident no later than week 4. At the end of the 14-week trial, differences between topiramate and placebo were still increasing, suggesting that even more improvement may occur with longer administration (Willenbring, 2007).

Two additional factors make topiramate seem particularly promising for treating alcohol dependence in primary care settings. First, topiramate proved effective with patients who were actively drinking rather than abstinent at the time medication was started. Patients in the multisite study were drinking heavily at the time of enrollment and study randomization (men were drinking 35 or more and women 28 or more standard drinks per week). These patients were not required to stop drinking before entering the study, although they had to express a desire to stop or reduce their consumption of alcohol with the possible long-term goal of abstinence. Second, the study provided only minimal behavioral support that focused on enhancing medication compliance and encouraging abstinence—a brief intervention that could be provided by nonspecialist health practitioners (Johnson et al., 2007).

At least 10 percent of participants reported adverse events; the events reported most by those on topiramate, as compared with placebo, included paresthesia, taste perversion, anorexia, difficulty with concentration and attention, and pruritus (Johnson et al., 2007). In the multisite study, where topiramate was titrated over a 6-week period, the attrition that was due to adverse events was 18.6 percent for the topiramate group and 4.3 percent for controls. In the earlier study (Johnson et al., 2003a), where adverse events were similar but titration occurred over a longer, 8-week period, retention rates for the topiramate and control groups were similar. However, in the multisite study, the researchers found that completion rates approached 90 percent among practitioners experienced in administering topiramate, whereas less experienced practitioners had more difficulty with retention. To enhance adherence, the authors advise clinicians to use a slow titration schedule over 8 weeks and to provide focused education for patients on how to manage emergent adverse events (Johnson et al., 2007).

The efficacy of topiramate was also supported in the following two small studies:

• Topiramate as add-on therapy for patients with alcohol dependence who do not respond to standard treatment. In an observational open-label, multisite study in Spain, 64 patients who were alcohol dependent with poor outcomes in standard treatment were provided with a mean dose of almost 200 mg per day of topiramate and monitored over a 12-month period (Fernández Miranda et al., 2007). The addition of topiramate resulted in a significant decrease in all outcomes measured—number of drinking days per month and standard drinking units consumed per day, craving, priming, dependence intensity scales, and serum transaminase levels.
• Ability of topiramate to increase periods of continuous “safe” drinking (defined by NIAAA as 1 or fewer standard drinks per day for women and 2 or fewer standard drinks per day for men). Some patients with alcohol dependence do not achieve abstinence during treatment. Researchers carried out a secondary analysis of data from a double-blind, randomized, controlled 12-week trial (Johnson et al., 2003b) to determine whether topiramate recipients were able to achieve longer continuous periods of “safe” drinking than those on placebo (Ma, Ait-Daoud, & Johnson, 2006). The analysis found that soon after topiramate was administered, recipients began to achieve increasing lengths of “safe” drinking relative to placebo. Furthermore, these early treatment gains appear to be predictive of continuing improvement as the length of time in treatment increases (Ma et al., 2006).

Other medications under development

In addition to topiramate, the 2007 literature search identified positive reports on controlled clinical trials of two potential medications for treating alcohol dependence: (1) nalmefene, an opioid antagonist, and (2) quetiapine, an atypical antipsychotic that targets both dopamine and serotonin receptors. Findings were as follows:

• Targeted nalmefene. A multisite, randomized double-blind study in Finland found that 10 to 40 mg doses of nalmefene were safe and reduced heavy drinking among 242 subjects with self-identified drinking problems (Karhuvaara et al., 2007). Subjects received minimal psychosocial intervention and took nalmefene only when drinking seemed imminent. After 28 weeks, 57 subjects on nalmefene continued into a 24-week extension period with randomization to continued nalmefene or placebo. Decrease in drinking was significantly greater for subjects on nalmefene than on placebo, which was corroborated by significant decreases in alanine aminotransferase and GGT. During this randomized withdrawal period, subjects remaining on nalmefene maintained the drinking level achieved in the initial 28 weeks, whereas those switched to placebo seemed to return to more frequent heavy drinking.
• Quetiapine. According to a double-blind, placebo-controlled 12-week trial among 61 subjects who were alcohol dependent, quetiapine (400 mg per day) may be more effective in treating people with the more severely affected Type B alcoholism compared with those with Type A alcoholism. This small study found a significant interaction between quetiapine and alcoholic subtype (Kampman et al., 2007). As predicted, Type B subjects treated with quetiapine had significantly fewer days of drinking and fewer days of heavy drinking than Type B subjects on placebo. Compared with those on placebo, people with Type B alcoholism who were treated with quetiapine had alcohol craving significantly reduced. Among the patients with Type A alcoholism, quetiapine offered no advantage over placebo in improving drinking outcomes. Nine patients treated with quetiapine (31 percent) maintained complete abstinence compared with two patients on placebo (6 percent).

Attitudes of professional alcohol counselors

Community-based addiction treatment centers rarely use pharmacotherapies for treating their patients who are alcohol dependent (Thomas, Wallack, Lee, McCarty, & Swift, 2003). Before initiating pharmacotherapy education at six community-based addiction treatment centers, Thomas and Miller (2007) collected baseline data on the knowledge and attitudes of 84 counselors and administrators attending a staff education project. Respondents came only from centers that had no on-staff medical provider. The data showed the following:

• These counselors and administrators, with just one exception, had very little or no knowledge about naltrexone.
• Most believed that adjunctive pharmacotherapy is ineffectual in treating alcohol dependence.

The authors concluded that lack of knowledge and confidence about pharmacotherapy by counselors is a barrier to more widespread referral and use of pharmacotherapies in alcohol treatment centers. Focused education will be needed for both counselors and administrators.

The study suggests that educational efforts do not need to overcome negative opinions about adjunctive pharmacotherapies. Instead, the intent should be to convey accurate and empirically supported information about the value of current medications. The respondents’ personal recovery status from addiction did not appear related to their valuation of pharmacotherapies. The most senior addiction professionals—those with more than 10 years of experience in the addiction field—were generally more positive in their valuation of adjunctive pharmacotherapy (Thomas & Miller, 2007).

Attitudes of patients toward pharmacotherapy

A second study looked at whether medically hospitalized patients with alcohol dependence are interested in pharmacotherapy and primary care to treat their alcoholism (Stewart & Connors, 2007). This survey covered 50 inpatients identified as alcohol dependent; all were receiving internal medicine services in a university-affiliated public hospital. Most survey participants were socioeconomically disadvantaged males admitted with disorders that heavy alcohol use would typically cause or exacerbate. In the month before being admitted to the hospital, these patients on average had been drinking on 86 percent of days and had averaged 8.4 drinks per drinking day. Their responses suggest that many such patients will be interested in receiving medication for alcoholism:

• 84 percent agreed they needed to stop drinking.
• 50 percent agreed that medication helps prevent drinking.
• 66 percent agreed they would like to receive an effective medication to help prevent drinking.

Interest in receiving effective pharmacotherapy was positively associated with addiction severity, adverse consequences, recognition of the problem, and drinking frequency. The reaction to primary care was mixed; only 32 percent of these patients were interested in primary care treatment for their alcoholism. The authors conclude that primary care followup alone may not adequately address patients’ perceived needs; many patients may also require prompt referral to specialty care after hospitalization (Stewart & Connors, 2007).

Research on disulfiram

Research studies and clinical experience over 55 years offer valuable information about the efficacy and safety of disulfiram. Almost 40 years elapsed from the time disulfiram became available before the first multisite, randomized clinical trial covering 605 participants was published (Fuller et al., 1986). Many large double-blind studies of disulfiram show no therapeutic effect compared with placebo (reviewed by Myrick, 2002). There is still no unequivocal evidence from randomized, controlled trials to show that disulfiram improves abstinence rates over the long term (Mann, 2004).

Brewer, Meyers, and Johnsen (2000) reviewed all published clinical studies in which there had been attempts to directly supervise the administration of disulfiram at least weekly. Adequate supervision included appropriate training of supervisors and review of their ability to supervise. These researchers found 13 controlled and 5 uncontrolled studies with supervised disulfiram administration and reported positive findings in all but 1 study. In general, the better the supervision, the better the outcomes. Under supervised situations, disulfiram reduced drinking, prolonged remissions, improved treatment retention, and facilitated compliance with psychosocial interventions.

Anton (2001), in a review of the literature, concluded that the evidence for disulfiram is mixed. According to Anton (2001), the most reliable study suggests that disulfiram is not better than placebo. In the reviewed studies, Anton (2001) reported that the factors that seem to improve treatment effectiveness with disulfiram include patient motivation, patient monitoring, being an older man, and concomitant treatment with acamprosate.

The most recent comprehensive review of the literature, done by Suh, Pettinati, Kampman, & O’Brien (2006) and covering the literature from 1937 to 2005, concluded that supervised disulfiram can be an effective treatment for alcohol dependence. The reviewers recommended that more research be done on disulfiram combined with other—and especially newer—psychotherapies.

Disulfiram use in primary care

Adverse events. Disulfiram is well tolerated in most patients, with the most common adverse effects being tiredness, headache, and sleepiness (Chick, 1999). Toxicities such as psychotic reactions, confusional states, and neuropathy are rare and appear to be dose related (Bevilacqua, Diaz, Diaz, Silva, & Fruns, 2002; Chick, 1999).

Disulfiram hepatitis is a very rare, sometimes fatal complication that particularly affects women (Brewer & Hardt, 1999). A Danish study of adverse reactions to disulfiram over a 22-year period estimated the rate of fatal disulfiram-induced hepatitis to be 1 per 25,000 patients treated per year, with the peak of hepatotoxicity occurring 60 days after the beginning of treatment. Because hepatotoxicity can usually be reversed if disulfiram is stopped before liver disease is clinically evident, Wright, Valfier, and Lake (1988) recommend liver function testing before treatment, at 2-week intervals for 2 months, and at 3- to 6-month intervals thereafter. Chick (1999) recommends informing the patient and the patient’s family and physician of the risk and immediately stopping the drug if adverse effects, such as fever preceding jaundice, are noted. Fuller and Gordis (2004) recommend supervised administration of disulfiram, along with careful monitoring for hepatotoxicity.

Conditions excluding treatment with disulfiram. Patients with cardiovascular or cerebrovascular disease are excluded from treatment because hypotension can occur during a disulfiram–alcohol interaction (Fuller & Gordis, 2004). Disulfiram has been reported to cause fetal abnormalities, so pregnant women should not use it. Disulfiram is also contraindicated in patients who have an idiopathic seizure disorder or cannot understand the risks associated with use of the drug. Disulfiram may influence adversely the pharmacokinetics and, therefore, the effects of medications metabolized by the cytochrome p450 system, such as warfarin, phenytoin, amitriptyline, and benzodiazepines (including chlordiazepoxide and diazepam but not lorazepam and oxazepam). Disulfiram also interferes with the pharmacokinetics of the tricyclic antidepressants. Fuller and Gordis (2004) report that the literature indicates disulfiram is unsafe to use concomitantly with monoamine oxidase inhibitors.

Patients appropriate for disulfiram. Data suggest that disulfiram is most effective in older, motivated individuals and in those who are supervised during daily ingestion. Predictors of efficacy with disulfiram include patients highly motivated for abstinence, people who are married or have a good support system, people with behavioral contracts to take the medication, and people legally compelled to take disulfiram (Myrick & Anton, 2004; O’Farrell, Allen, & Litten, 1995). Disulfiram can also support abstinence when people who are alcohol dependent attend events that involve alcohol, such as family celebrations.

In general, disulfiram seems to have limited acceptance in the treatment of alcohol dependence (Anton & Swift, 2003). Several recent, small pilot studies suggest that disulfiram might be safe and useful for the following types of patients:

• Patients who are positive for the hepatitis C virus (HCV). A recent review of the literature recommends monitored disulfiram treatment for patients positive for HCV (Kulig & Beresford, 2005).
• Patients who are court ordered.Martin, Clapp, Alfers, and Beresford (2004) found that compliance with treatment was 61 percent after 18 months for those with court-ordered, supervised disulfiram treatment. This compared with 18-percent compliance among those in a voluntary, supervised disulfiram group.
• Adolescent patients. In a small, preliminary study, Niederhofer and Staffen (2003) compared 13 adolescents ages 16–19 on disulfiram with 13 controls. After 90 days, the mean abstinence duration was significantly greater for the disulfiram group than for the placebo-treated controls (68.5 days [SD 37.5] vs. 29.7 days [SD 19.0]). Disulfiram was well tolerated in adolescents, except for occasional diarrhea.
• Patients with severe mental illness. A preliminary study of 33 patients with severe mental illness and alcohol dependence found that supervised disulfiram treatment was associated with decreases in the number of days hospitalized. Controlled research is needed to evaluate the effects of disulfiram in this population (Mueser, Noordsy, Fox, & Wolfe, 2003).
• Patients codependent on alcohol and cocaine. Disulfiram’s main effects in initiating abstinence in cocaine and alcohol use were still maintained a year after patients with codependence received short-term (12-week) treatment with disulfiram combined with psychotherapy (Carroll et al., 2000). Ninety-six patients who were codependent randomly received CBT either with or without disulfiram, 12­Step facilitation with or without disulfiram, or clinical management with disulfiram. Carroll and associates (2000) concluded that this randomized controlled trial supports the efficacy of disulfiram with this challenging codependent population. The findings suggest long-term benefits can result from comparatively brief treatments that facilitate the initiation of abstinence.

Research needs

The effects of disulfiram on craving have not been widely studied, but disulfiram is unlikely to be very powerful in reducing craving, especially if a patient has not achieved sustained abstinence (Anton & Swift, 2003). For better information about the use of disulfiram today, randomized clinical trials need to determine whether supervised ingestion of disulfiram:

• Would be useful to ensure sobriety for such high-risk groups as criminal offenders and those who have failed previous attempts at treatment
• Is better if supervision is performed by a clinic staff member or by a relative
• Would improve treatment outcomes when combined with newer pharmacotherapies (Fuller & Gordis, 2004).

Research on naltrexone

In the last decade, the efficacy of naltrexone for alcohol dependence has been extensively studied, particularly in the United States. At least 19 published controlled studies of about 3,200 patients have compared the effects of oral naltrexone with placebo; nearly all showed efficacy in the treatment of alcohol dependence (Garbutt et al., 2005). The majority of clinical trials support the hypothesis that naltrexone can reduce the urge to drink, increase the number of days abstinent, and minimize the risk of relapse to heavy drinking in some patients (O’Malley & Froehlich, 2003). However, two recent studies, including a large, multisite VA study, have reported no or minimal effectiveness in reducing drinking behavior as compared with placebo (reviewed by Krystal et al., 2001). One reason for this ineffectiveness may be the high rate of noncompliance among patients in the VA study. Lack of compliance with oral naltrexone is a problem that varied greatly across studies, with 40 to 90 percent of subjects completing treatment in the studies.

The lack of consistent findings on the effects of oral naltrexone may be the result, at least in part, of variations in how compliant patients are with the medication. A number of studies indicate that poor compliance with therapy can limit the effectiveness of oral naltrexone (Johnson & Ait-Daoud, 2000; Kranzler, Wesson, & Billot, 2004). For example, in a 3-month followup study, Volpicelli and colleagues (1997) found that only patients who took their oral daily dose on at least 90 percent of study days improved their drinking outcomes. No differences were found between the placebo group and those who took naltrexone on fewer than 90 percent of study days on any drinking measure; 50 percent of these subjects relapsed, changing from abstinence to clinically significant drinking during the study. In a large, 1-year collaborative study in the United Kingdom, only patients who took at least 80 percent of their naltrexone tablets experienced better drinking outcomes than those on placebo (Chick et al., 2000). The outcomes in 17 clinical trials of naltrexone at a dose of 50 mg per day are shown in Exhibit 1 (Mann, 2004).

Exhibit 1

Published Placebo-Controlled Clinical Trials of Naltrexone 50 mg/day in Alcohol Dependence.

The extensive research on oral naltrexone has produced numerous review articles and several meta-analyses of the literature. Three meta-analyses concluded that the effect of naltrexone is significantly greater, on average, than that of placebo (Kranzler & Van Kirk, 2001; Srisurapanont & Jarusuraisin, 2005; Streeton & Whelan, 2001). The meta-analysis by Streeton and Whelan (2001) found that, after 12 weeks of naltrexone treatment, patients experience significantly fewer episodes of relapse and significantly more remain abstinent compared with subjects on placebo. The meta-analysis by Srisurapanont and Jarusuraisin (2005), which covered 3,048 subjects in 27 randomized controlled trials from 34 published and unpublished papers, concluded the following:

• Short-term treatment. Naltrexone should be accepted as a short-term treatment for alcoholism. In comparison with placebo, short-term treatment significantly reduces the chance of alcohol relapse for 36 percent of patients, is likely to reduce the chance of returning to drinking for 13 percent, and can lower the risk of withdrawing from treatment for 28 percent of patients.
• Medium-term treatment. Medium-term naltrexone treatment gives no benefit over placebo in terms of relapse prevention, but it does increase the time to first drink and diminishes craving. In this regard, naltrexone plus intensive psychosocial treatment is superior to naltrexone plus a simple psychosocial treatment. Naltrexone is also superior to acamprosate in reducing relapses, number of drinks, and craving.
• Concomitant treatment strategies. To improve treatment adherence and to ensure that real-world treatment is as effective as research findings, some form of psychosocial intervention and management of adverse effects needs to accompany naltrexone therapy.
• Unanswered questions. The existing research is limited because many trials are of short duration and have small sample sizes. Important areas of concern include the lack of data on different psychosocial benefits and on how long patients who respond to naltrexone should continue their treatment. The evidence for longer term (more than 8 months) efficacy of naltrexone remains to be demonstrated (Mason, 2003).

Recently, Pettinati and colleagues (2006) reviewed all published, randomized placebo-controlled trials of naltrexone to resolve inconsistencies in naltrexone’s reported efficacy across trials. Drinking outcomes measured in these studies related to four outcomes: two pertaining to “any drinking” and two pertaining to “heavy or excessive drinking.” This review found an advantage for naltrexone over placebo in 70 percent of clinical trials that measured reductions in “heavy or excessive drinking” but in only 36 percent of trials that measured abstinence or “any drinking.” Pettinati and colleagues (2006) concluded that naltrexone’s therapeutic effects are most related to outcomes pertaining to heavy or excessive drinking.

Naltrexone use in primary care

Of particular interest to physicians in primary healthcare settings is one recent study that looked at whether general internists and primary care physicians, using naltrexone, could treat patients who are alcohol dependent as effectively as addiction specialists can (O’Malley et al., 2003). Results indicated that primary care counseling with naltrexone pharmacotherapy can be effective in select patients. Study subjects were recruited from newspaper ads and required to be abstinent from alcohol for 5 to 30 days before initiating treatment. In this nested sequence of randomized, placebo-controlled trials, patients received the following:

• Phase I—197 patients received 10 weeks of naltrexone and either (1) brief counseling from a primary care physician (an initial 45-minute visit followed by 15-to 20-minute sessions in weeks 1 through 4, 6, 8, and 10) or (2) CBT from an addiction specialist (an initial 1.25-hour session followed by weekly 50-minute sessions for 10 weeks).
• Phase II—Responders from both groups received 24 weeks of continuing maintenance with naltrexone.

In Phase I, the results were comparable in the two groups, with 84.1 percent of primary care patients and 86.5 percent of CBT patients avoiding persistent heavy drinking. Persistent heavy drinking was defined as more than 2 days of heavy drinking (5 or more drinks per day for men and 4 or more drinks per day for women) during the last 28 days of Phase I. In Phase II, the response to naltrexone maintenance was better maintained among those who received primary care than in those with counseling appointments (O’Malley et al., 2003). Monterosso and colleagues (2001) also found a significant advantage of naltrexone use over placebo in patients who received 12 weeks of concurrent primary counseling.

The results are available from NIAAA’s COMBINE study, a multisite, randomized, controlled trial that evaluated medical management with naltrexone, acamprosate, or both, with or without additional specialist treatment (combined behavioral intervention). Participants received interventions over a 4-month period and were evaluated for up to 1 year after treatment. Findings from this study suggest that naltrexone with medical management can be delivered successfully in healthcare settings, which would greatly expand the number of people receiving treatment (Anton et al., 2006). In fact, the COMBINE data suggest that naltrexone can be effective in the context of medical management without specialized behavioral treatment.

In the COMBINE study, participants taking naltrexone received 25 mg on days 1 through 4, 50 mg on days 5 through 7, and 100 mg on days 8 through 112. Doses were chosen based on preliminary evidence that doses higher than those commonly prescribed could be more efficacious and provide better coverage for missed doses; two pilot studies confirmed the tolerability of these doses (Anton et al., 2006). Ongoing or recurrent dose reductions could be made for individual participants and were made in 12.1 percent of patients for naltrexone, compared with 11.9 percent for acamprosate, 20.9 percent for acamprosate plus naltrexone, and 7.8 percent for placebo. On average, 88 mg of naltrexone was taken daily, and the mean medication adherence rate for naltrexone was 85.4 percent—similar to the adherence rates for those receiving acamprosate or combined behavioral interventions. The COMBINE study confirmed the efficacy of naltrexone in reducing drinking among volunteers who were newly abstinent from alcohol. Key findings included the following:

• Participants receiving naltrexone plus medical management had a higher percentage of days abstinent (80.6 percent) than those receiving placebos and medical management only (75.1 percent).
• Naltrexone reduced the risk of a first heavy drinking day over time; the reduction in risk was 0.28, consistent with meta-analyses of other naltrexone trials that used 50 mg per day and included specialist care (Anton et al., 2006).

Other findings from the COMBINE study are detailed in the sections on acamprosate and on combined medication therapy.

Adverse events. Some researchers have attributed the low degree of compliance with naltrexone to poor tolerability and hepatic toxicity (Volpicelli et al., 1997). However, a recent meta-analysis of naltrexone studies concluded that only 10 percent of patients fail to complete treatment because of one or more adverse drug effects and that hepatic toxicity is very unlikely at the current dose of 50 mg of oral naltrexone daily (Bouza, Magro, Muñoz, & Amate, 2004; Yen, Ko, Tang, Lu, & Hong, 2006). Yen and colleagues (2006) concluded that naltrexone is not hepatotoxic at the recommended daily dose and may be beneficial for patients with elevated liver enzymes. In the COMBINE study, with its higher naltrexone dosage, only 1 of 70 serious adverse events could have been related to the medication. Of the 601 participants, 12 (primarily those in the naltrexone groups) had treatment-emergent levels of liver enzymes (aspartate aminotransferase or alanine aminotransferase) greater than five times the upper limit of normal; these cases resolved once the medication was discontinued except for two cases (one participant did not return for retesting and the other continued heavy drinking) (Anton et al., 2006).

Conditions excluding treatment with naltrexone. Patients are ineligible for naltrexone if they have poor liver function or a history of liver disease, have recent prescribed or nonprescribed opioid use, and, for women, are pregnant or not using adequate birth control (Rohsenow, 2004). Absolute contraindications to naltrexone use include acute hepatitis, liver failure, active opioid withdrawal, and the current use of methadone or opioid-containing medications prescribed to manage pain and treat serious medical conditions, such as heart disease, severe arthritis, sickle cell anemia, and recurrent congestive heart failure (CSAT, 1998). A relative contraindication applies to patients who have an anticipated need for opioids to treat an identified medical problem, because use of naltrexone can impede the effectiveness of prescription and over-the-counter analgesics, cough medicines, and pain medications that contain opioids (CSAT, 1998).

Patients appropriate for naltrexone. Current research suggests that the patients most likely to benefit from naltrexone are those who have close relatives with alcohol problems, have particularly strong urges to drink, or have limited cognitive abilities (Rohsenow, 2004). Naltrexone is also well tolerated in older patients (Oslin et al., 1997a). People with lower blood concentrations of the drug may benefit from a larger dose, and those with good results on naltrexone may benefit from longer maintenance (Rohsenow, 2004).

Research needs

More research is needed on which subgroups of patients are most likely to respond well to naltrexone, as well as to other pharmacotherapies. In a recent controlled trial in Germany, Kiefer, Helwig, Tarnaske, Otte, and Wiedemann (2005a) looked at the response to naltrexone and acamprosate by patients who had (1) low versus high baseline somatic distress, depression, and anxiety, (2) low versus high baseline craving, and (3) typological differentiation according to the subtypes proposed by Cloninger and Lesch (Lesch & Walter, 1996). A comparison of the course of abstinence rates indicated that naltrexone was effective particularly in patients with high baseline depression, whereas acamprosate was mainly efficacious in patients with low baseline somatic distress. Baseline craving showed no predictive value. Naltrexone revealed best treatment effects in Lesch’s types III and IV typology, whereas acamprosate was mainly effective in type I (Kiefer et al., 2005a).

Some researchers hope that it may become possible to choose therapy based on identification of genetic subtypes of the specific molecular targets for drugs. For example, because a family history of alcohol problems is a predictor of naltrexone response, it is hypothesized that a gene variation of the OPRM1 may increase an individual’s susceptibility to substance dependence, as well as increase the response to naltrexone. To test this, a recent randomized study of patients with alcohol dependence examined the association between their treatment outcomes and two specific polymorphisms of the gene encoding the OPRM1. In subjects of European descent, individuals with one or two copies of the Asp40 allele who were treated with naltrexone had significantly lower rates of relapse and a longer time to return to heavy drinking than those homozygous for the Asn40 allele (Oslin et al., 2003). If these results are replicated, then gene testing may be a feasible and cost-effective way to identify individuals who are most likely to respond to naltrexone treatment.

Research by McGeary and colleagues (2006) found that, among non-treatment-seeking heavy drinkers, all of naltrexone’s moderating effects on craving and on a cue-elicited urge to drink could be accounted for by Asn40Asp polymorphisms in the OPRM1 gene. However, in a study of veterans being treated for alcohol dependence, Gelernter and colleagues (2007) found no significant interactions between the OPRM1 Asn40Asp polymorphisms and the response to naltrexone treatment. Oslin, Berrettini, and O’Brien (2006) reviewed the current research agenda and the biological correlates of the receptor genes that have been demonstrated to predict clinical response to naltrexone in individuals who are dependent on alcohol.

Research on injectable naltrexone

Several formulations of injectable naltrexone have been tested in pilot studies and clinical trials since 1998, and all studies have shown the injectable long-acting formulation to be safe, well tolerated, and effective in reducing heavy drinking days and other measures of problem drinking. Although most studies involved small numbers of subjects, developers of the two investigational formulations have published multicenter, randomized, placebo-controlled clinical trials (Garbutt et al., 2005; Kranzler et al., 2004). Studies done to date include the following:

• Preliminary studies. In a 12-week study of an injectable naltrexone formulation (Depotrex) combined with 8 weekly coping skills sessions, the 15 patients on 206 mg of daily naltrexone had fewer drinking days than 5 patients on placebo, supporting continued research on the sustained-release drug (Kranzler, Modesto-Lowe, & Nuwayser, 1998). An 8-week study of the same injectable formulation among 12 subjects who were heroin dependent showed that both low- (192 mg) and high-dose (384 mg) injections were safe and effective and produced long-lasting antagonism to the effects of heroin (Comer et al., 2002). Few adverse effects were reported except for mild discomfort at the injection site, and blood plasma levels remained above 1 ng/ml for 3 to 4 weeks after the injection.
• DrugAbuse Sciences Naltrexone Depot Study Group. This group conducted the first multicenter study of injectable naltrexone (Naltrel) for alcohol dependence, randomly assigning 315 patients either to an intramuscular injection of naltrexone monthly for 3 months or to placebo; all subjects received five sessions of manual- guided MET (Kranzler et al., 2004). The medication was well tolerated, with approximately 74 percent of subjects receiving all injections. For those taking injectable naltrexone, there was a significant advantage over placebo in time to first drinking day, fewer drinking days during treatment, and a significantly greater abstinence rate than for the placebo group (18 vs. 10 percent). Earlier studies, including a 6-week, open-label trial of one 300 mg injection among 16 subjects, combined with weekly individual counseling sessions, found no serious adverse events and a significant reduction in the number of drinks per day, heavy drinking days, and the proportion of drinking days compared with baseline (Galloway, Koch, Cello, & Smith, 2005; Modesto-Lowe, 2002).
• Vivitrex^® Study Group. This group conducted a 6-month, double-blind, placebo-controlled trial of long-acting injectable naltrexone, using two different doses, at 24 U.S. public hospitals, private and VA clinics, and tertiary care medical centers. Adults who were actively drinking were randomized to naltrexone treatment or placebo, and 624 received at least one injection. All subjects received 12 sessions of a low-intensity psychosocial intervention. Compared with placebo, the high (380 mg) dose resulted in a 25-percent decrease in the event rate of heavy drinking days, whereas the low (190 mg) dose resulted in a 17-percent decrease. The long-acting naltrexone was well tolerated, and there was no evidence of hepatotoxicity (Garbutt et al., 2005). An earlier 16-week, multisite pilot study had shown the formulation to be both safe and well tolerated (Johnson et al., 2004).

Injectable naltrexone use in primary care

Now that it has been approved for marketing by FDA, injectable naltrexone is available as a treatment option that can be used by primary care practitioners and addiction specialists. In the multisite trial, the efficacy of the 380 mg dose was evident in the first month after the initial injection and was maintained over the 24-week treatment period (Garbutt et al., 2005). Unlike patients in the oral naltrexone trials, the majority of patients were actively drinking when they started injectable naltrexone treatment. However, the FDA Center for Drug Evaluation and Research (CDER) analysis of the study data concluded that injectable naltrexone was effective only in those who were abstinent at baseline. CDER’s analysis emphasized the proportion of patients who did not relapse to heavy drinking (FDA CDER, personal communication, 2008).

Adverse events. In patients using oral naltrexone, high urinary levels of 6-β-naltrexol have been associated with adverse events, such as headache, anxiety, nausea, and spontaneous erection (King, Volpicelli, Gunduz, O’Brien, & Kreek, 1997). In a multicenter study, at least 15 percent of individuals withdrew from oral naltrexone treatment because of adverse events, particularly nausea (Croop, Faulkner, & Labriola, 1997). In contrast to oral naltrexone, plasma levels of extended-release naltrexone remain relatively constant among patients taking the injectable formulation, which may be one reason for its milder adverse effects. The lack of first-pass metabolism with injectable formulations, with reduced levels of 6-β-naltrexol, may also contribute to its improved adverse-event profile (Johnson, 2006). The peak plasma concentration of injectable preparations exceeds that of oral naltrexone during the days immediately following the injection. The higher tolerability of injectable naltrexone may be because such peaks occur daily with oral therapy but only early in treatment with the injectable formulations (Johnson, 2006). The following side effects were observed in the clinical trials of injectable naltrexone:

• Vivitrol. High doses (400 mg) of Vivitrol seemed to be safe and well tolerated in the 16-week clinical trial, with the four most common side effects among 40 subjects being nausea, headaches, nonspecific abdominal pain, and pain at the injection site. Two subjects dropped out from adverse effects—one from induration at the injection site and one from an allergic reaction resulting in angioedema (Johnson et al., 2004). In the longer 24-week clinical trial, subjects who received the high (380 mg) dose of Vivitrol were significantly more likely to report nausea, fatigue, decreased appetite, dizziness, and pain at the injection site than those in the low (190 mg) dose or placebo groups (Garbutt et al., 2005). Among subjects in the high-dose Vivitrol group, 14.1 percent discontinued treatment compared with 6.7 percent of those in both the low-dose and the placebo groups. Two of the high-dose subjects dropped out because of serious adverse events—allergic-type eosinophilic pneumonia and interstitial pneumonia—which resolved following medical treatment. The most frequent reasons for dropping out of the study were nausea, injection site reactions, and headaches (Garbutt et al., 2005).
• Naltrel. In general, the first 12-week clinical trial showed Naltrel to be safe and well tolerated at an initial dose of 300 mg (one 150 mg injection in each buttock), with subsequent doses being only 150 mg (Kranzler et al., 2004). Side effects, including upper abdominal pain, chest pain, and injection site reactions, were significantly more common in the group taking Naltrel than in those taking placebo. Reasons for discontinuing treatment were similar in the Naltrel and placebo groups, although 13 subjects taking Naltrel (8.2 percent) dropped out of treatment, compared with 6 subjects (3.8 percent) in the placebo group. No serious adverse events were reported in a subsequent 6-week, open-label trial. Sixteen subjects, followed for 6 weeks after a single 300 mg dose of Naltrel intramuscularly, reported 198 side effects. The 17 side effects rated as severe included nausea, flatulence, gastrointestinal pain, fatigue, lethargy, somnolence (two reports), depression, irritability, headache (four reports from three participants), back pain, injection site mass, injection site pain, and an elevated GGT level (Galloway et al., 2005).

In light of one diagnosed and one suspected case of eosinophilic pneumonia in the Vivitrol trials, the manufacturer recommends that physicians consider a diagnosis of eosinophilic pneumonia in any patient receiving injectable naltrexone who develops progressive dyspnea and hypoxemia, as well as the possibility of eosinophilic pneumonia in patients who do not respond to antibiotics.

Conditions excluding treatment with injectable naltrexone. More research is needed to determine whether injectable naltrexone is associated with unexpected adverse or allergic reactions because three subjects in the Vivitrol trials had angioedema or pneumonia, an allergic-type reaction rate of 1:218. The Vivitrol manufacturer states that patients should not be actively drinking at the time Vivitrol is initially administered and that the medication is contraindicated in patients who have previously exhibited hypersensitivity to naltrexone, polylactide-co-glycolide, carboxymethylcellulose, or any other components of the diluent.

Injectable naltrexone is a potent opioid antagonist. Contraindications to the oral form of naltrexone also pertain to the injectable formulations, including the following:

• In patients with acute hepatitis or liver failure and patients with active liver disease, injectable naltrexone must be carefully considered, given naltrexone’s hepatotoxic effects. Use of injectable naltrexone should be discontinued in the event of symptoms or signs of acute hepatitis.
• In patients who are receiving opioid analgesics, patients with current physiologic opioid dependence, and patients in acute opioid withdrawal.
• In individuals who have failed the naloxone challenge test or have a positive urine screen for opioids.

There appears to be at most a fivefold margin between a safe dose of naltrexone and a dose that can cause hepatic injury. Injectable naltrexone at recommended doses does not appear to be a hepatotoxin.

Patients appropriate for injectable naltrexone. In reviewing the clinical evidence on injectable naltrexone, Johnson (2006) concludes that injectable naltrexone could benefit individuals who have failed at outpatient treatment using adjunctive medication, as well as other target populations. Such patients with alcohol dependence include the following:

• Patients who show low compliance with medication resulting from nonspecific factors, such as memory impairment
• Patients who experience marked or prolonged side effects from taking oral naltrexone
• Individuals who experience relatively low therapeutic effects from oral naltrexone, suggesting that a trial with an injectable preparation be done to rule out fluctuating blood levels of naltrexone as a possible cause
• Individuals who expect to be in situations where oral naltrexone might be unavailable or difficult to obtain if lost, such as overseas travelers or military personnel on short assignments
• Individuals detoxified in hospitals and awaiting referral to outpatient treatment so that medication can be available to these patients during the hiatus between detoxification and treatment
• Individuals with co-occurring alcohol and psychiatric disorders, for whom the injectable form would reduce the need for additional pills
• Offenders in forensic facilities or drug courts who could be offered the option of imprisonment or supervised treatment with injectable preparations; clear guidelines and protocols must guide the ethical use of injectable naltrexone in forensic settings.

As reported in the section on oral naltrexone, research demonstrates that people with a family history of alcoholism seem to respond best to this medication. The trials on injectable naltrexone have not explored the connection between such characteristics in men and women and the efficacy of the injectable formulation.

Estimates of efficacy—with a small to medium effect size—seem to be comparable in men taking Vivitrol or Naltrel (Johnson, 2006). However, the potential benefit of injectable naltrexone for women is unclear. The multisite clinical trials showed that injectable naltrexone can reduce heavy drinking in men, but no significant effects were shown in women. Injectable naltrexone seems to effectively reduce relapse and promote abstinence among individuals who are alcohol dependent, but the two published trials did not explore the effect of gender on treatment outcomes.

Research needs

The Vivitrol trial, which represents one of the largest samples ever treated with a medication for alcohol dependence, shows that this formulation could improve intervention strategies for alcohol dependence because it can provide a predictable pharmacological foundation for treatment. In addition, it has the clinical benefit of providing a firm basis for combination with other treatments, including psychotherapy, other medications, or both. Additional research will resolve the following issues raised by the multisite trial:

• Better understanding of the effects of injectable naltrexone on women. Treatment effects were highly significant among men taking 380 mg injectable naltrexone but not significant in women. Because only a small number of women were included in the Vivitrol study, they may not be representative of women with alcohol dependence in general. Also, women’s typical heightened response to psychosocial interventions may obscure the medication effects (Garbutt et al., 2005). A recent study found that drinking outcomes with oral naltrexone seemed to be superior for women compared with men (Kiefer, Jahn, & Wiedemann, 2005b). In light of this finding, it has been suggested that the injection delivery method may inhibit its effectiveness for women. The injections may have more frequently been delivered subcutaneously rather than intramuscularly in women, thereby slowing absorption (Johnson, 2006). Recent research on the efficacy of medications injected intramuscularly in the buttocks showed that the higher percentage of body fat in women frequently causes injections into fat rather than into muscle, which can be prevented through use of longer needles. More research is needed.
• More knowledge about treatment duration and special populations. Additional research is needed to determine the optimal duration of treatment with long-acting naltrexone, as well as indicators that treatment can be discontinued. The usefulness of injectable naltrexone for special populations, such as people with a major mental disorder or those in the criminal justice system, is yet to be examined.

Research on acamprosate

Over the past 15 years, the safety and efficacy of acamprosate for alcohol dependence have been well established in multiple double-blind, placebo-based trials (Mason, 2005). Overall, in numerous European trials, acamprosate has been consistently associated with greater beneficial effects than placebo on the following measures of alcohol abstinence: greater rates of complete abstinence, longer times to first drink, and/or an increased duration of cumulative abstinence (Mason, 2005). Surprisingly, two recent well-designed, multisite studies of U.S. patients have not shown the level of efficacy for acamprosate that is consistently demonstrated among European patients (Anton et al., 2006; Mason et al., 2006). In addition, a recent study in Australia found that use of acamprosate did not further improve the significant change that outpatients reported in their subjective health status and psychological well-being as a result of receiving CBT alone for their alcohol dependence (Feeney, Connor, Young, Tucker, & McPherson, 2006b).

Acamprosate has been used in conjunction with psychosocial or behavioral counseling to promote abstinence in 26 countries, producing an extensive body of data. By 2000, acamprosate had been studied in 17 randomized, placebo-controlled clinical trials performed in 11 European countries and South Korea and covering approximately 5,000 male and female outpatients. Reviews and several meta-analyses have been done on these trials, which all support the therapeutic effect of acamprosate (Mann, 2004; Mason, 2001; Soyka & Chick, 2003). Outcomes of the clinical trials of acamprosate at a dose of 1,998 mg per day are listed in Exhibit 2 (Mann, 2004).

Exhibit 2

Published Placebo-Controlled Clinical Trials of Acamprosate 1,998 mg/day in Alcohol Dependence.

The European acamprosate studies varied in duration from 3 months to more than a year. In 13 of 15 studies, subjects treated with acamprosate had a higher rate of treatment completion, longer time to first drink, and higher abstinence rates compared with subjects treated with placebo (Mason, 2001). In the combined studies, the abstinence rate at the end of treatment in the acamprosate groups was 35 percent versus 21 percent in the placebo groups.

One review of the clinical data concluded that there was some evidence in three studies (those of Chick, Pelc, and Paille and their colleagues) that acamprosate could reduce craving (Mann, 2004). However, other reviewers consider it inaccurate to refer to acamprosate as an anticraving agent; they believe the evidence supports its efficacy only as a medication to prevent relapse, possibly by blocking prolonged withdrawal symptoms (Mason, 2001). Mann (2004) also concluded that, for three studies that failed to show beneficial effects of acamprosate over placebo, the possible reasons were that patient numbers were too small, a 2-month treatment period was used rather than the longer treatment periods used in the other studies, and acamprosate was not started until 25 days after patients had been weaned from alcohol, by which time many subjects were no longer abstinent.

In a recent meta-analysis of 16 studies, the relative benefit of remaining continually abstinent for 6 months after detoxification was quantified as 1.47 for subjects treated with acamprosate compared with subjects receiving placebo (Mann, Lehert, & Morgan, 2004). This meta-analysis also suggested that the relative benefit attributable to acamprosate may increase over time.

Acamprosate use in primary care

Current data suggest that acamprosate may be equally useful in primary care and in specialized substance abuse treatment settings (Mann, 2004). A number of Phase IV studies of acamprosate have been made under naturalistic practice conditions that basically confirm the abstinence rates found in the placebo-controlled trials (Pelc et al., 2002; Soyka, Preuss, & Schuetz, 2002). A pragmatic trial in France compared results when 149 general practitioners, who were accustomed to managing patients in their practice, added acamprosate to standard treatment. A very high percentage of patients successfully completed the 1-year followup period (348 of 422 patients or 82.5 percent). The duration of abstinence compared well with the clinical trials: 0.67 for standard care and 0.81 for acamprosate. In clinical practice, this means that patients taking acamprosate could be expected to remain abstinent for 23 percent longer on average than patients on standard care and to experience about 2 months more abstinence during a 1-year treatment period (Kiritze-Topor et al., 2004). However, the principal finding was that adjunctive therapy with acamprosate was associated with a significantly better outcome in patients’ quality of life, based on social, medical, and economic measures.

The first U.S. study to evaluate the clinical efficacy of acamprosate compared the safety and effects of the standard 2 g dose, an exploratory 3 g dose, or placebo in a double-blind, 6‐month trial conducted among 601 volunteers in 21 outpatient clinics across the United States (Mason et al., 2006). All patients received the drug or placebo plus eight concomitant sessions of brief, manual-guided counseling (http://www.alcoholfree.info). The main outcome measure was the percentage of alcohol-free days over the 6-month period. Surprisingly, the percentage of abstinent days did not differ significantly across groups in the a priori analysis (54.3 percent for placebo, 56.1 percent for 2 g acamprosate, and 60.7 percent for 3 g). However, the researchers used standardized assessments to characterize the subjects at baseline according to such potential covariates as baseline goal of total abstinence, alcoholism severity, stage of readiness to change, treatment exposure, and such psychological precedents as psychiatric hospitalizations or suicide attempts. Analysis of these covariates showed that acamprosate was associated with a significantly higher percentage of abstinent days than placebo in the subgroup of patients who had a baseline goal of abstinence (58.1 percent for placebo, 70.0 percent for 2 g acamprosate, and 72.5 percent for 3 g). Researchers concluded that acamprosate has an appreciable treatment effect among patients who have abstinence as a treatment goal.

In the COMBINE study, all groups showed substantial reduction in drinking. This multisite, U.S. study used a 3 g daily dose of acamprosate. However, the study found no evidence of efficacy for acamprosate and no evidence of incremental efficacy for combinations of naltrexone, acamprosate, and combined behavioral intervention (Anton et al., 2006). The lack of acamprosate efficacy was unexpected, given the positive results of many previous trials (Anton et al., 2006). The substantial improvement shown by all COMBINE groups, possibly in part as a result of the attention within the study itself (the “Hawthorne effect”), may have lessened the study’s power to show an impact from the acamprosate.

Adverse events. The international and U.S. clinical trials demonstrated a favorable safety and tolerability profile (Garbutt, West, Carey, Lohr, & Crews, 1999; Mason, 2001). Side effects are generally mild, with the most frequent side effect being short-term diarrhea that is dose related and transient (Boothby & Doering, 2005). Very few patients drop out of treatment because of adverse effects. There is no risk of alcohol interactions with acamprosate, and there is no abuse potential (Mason, 2001). Participants in the first U.S. multisite trial experienced no deaths or serious drug-related adverse events (Mason et al., 2006). The COMBINE trial, using acamprosate at a 3 g dosage, found no problems with either adverse events or medication adherence (Anton et al., 2006).

Conditions excluding treatment with acamprosate. Acamprosate is contraindicated in patients with severe renal impairment and requires a dose reduction for patients with moderate renal impairment (Forest Laboratories, Inc., 2005). However, this medication may be particularly useful in patients with hepatic impairment and/or liver disease (Scott, Figgitt, Keam, & Waugh, 2005).

Patients appropriate for acamprosate.Mason and colleagues (2006) suggest that their U.S. multisite trial was “perhaps the most definitive evidence to date that acamprosate is not an effective treatment for alcohol dependence in non-motivated and non-abstinent populations.”

Acamprosate is a proven effective intervention for treatment of alcohol dependence. However, acamprosate prevents lapses or relapses only in a minority of patients. Two important questions, therefore, are (1) whether acamprosate is more effective when combined with particular types of psychosocial treatment and (2) whether specific subgroups of patients respond particularly well to acamprosate.

Three Phase IV studies failed to find any significant differences in outcome among various psychosocial treatment groups, which included individual therapy, group therapy, brief therapy, and CBT. A pooled analysis of seven trials, covering 1,485 patients, was unable to identify a positive predictor of efficacy with acamprosate treatment, suggesting that acamprosate can be considered a potentially effective pharmacotherapy for all patients (Verheul, Lehert, Geerlings, Koeter, & van den Brink, 2005). The variables looked at, none of which predicted efficacy with acamprosate, included family history of alcoholism, late age of onset, female gender, high physiological dependence, serious anxiety symptomatology, and severe craving at baseline.

Soyka and Chick (2003) recommend that patients be given an initial prescription trial of acamprosate and, if they manage to abstain, they should continue receiving the drug for 1 year. Both the European and U.S. studies suggest that treatment needs to be initiated as soon as possible after the period of alcohol withdrawal, once the patient has achieved abstinence (Soyka & Chick, 2003). Acamprosate should not be stopped if the patient lapses because this medication appears to have a small effect in reducing drinking following a relapse (Chick, Lehert, & Landron, 2003).

Research needs

More research is needed to understand the different outcomes of the international versus the U.S. trials on efficacy of acamprosate. Understanding why the research results are discrepant can have important clinical implications. A number of possible reasons have been postulated for the failure of U.S. studies to show the efficacy for acamprosate shown in nearly all international studies:

• Differences in U.S. and European drinking patterns
• Length of clinical trials (European trials are generally longer than U.S. trials)
• More standardized, manual-based psychosocial treatments in U.S. trials, which may result in more consistently improved patient outcomes that reduce the perceived effect of the added medication (Mason et al., 2006)
• Differences in length of pretreatment abstinence preceding the medication (COMBINE required only 4 days of abstinence, achieved primarily on an outpatient basis, whereas most positive studies of acamprosate have a longer pretreatment abstinence period established during inpatient treatment) (Anton et al., 2006).