Effectiveness of Naltrexone in a Community Treatment Program

0145-6008/04/2811-1710$03.00/0 ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH Vol. 28, No. 11 November 2004 Effectiveness of Naltrexone in a Community Treatment Program Therese K. Killeen, Kathleen T. Brady, Paul B. Gold, Kit N. Simpson, Richard A. Faldowski, Clare Tyson, and Raymond F. Anton Objective: In several large, well-designed, randomized, double-blind studies, the opiate antagonist naltrexone demonstrated efficacy in the treatment of alcohol dependence. Specifically, when combined with certain psychosocial therapies, naltrexone reduces the number of drinking days, heavy drinking, and time to relapse to alcohol use in alcohol-dependent individuals. Whether this efficacy can be generalized to individuals who have alcohol use disorders and present for treatment at front-line community treatment programs has not been well established. Methods: A total of 145 patients who presented for treatment at a rural community substance abuse treatment center were randomized to receive naltrexone 50 mg daily plus usual program treatment (n 54), placebo plus usual treatment (n 43), or usual treatment alone (n 48) for 12 week. A total of 133 participants had at least one follow-up visit. Primary outcome measures included percent days drinking, average drinks per drinking day, average drinks per day, heavy drinking days (four or more for women and six or more for men), and time to first heavy drinking day. Secondary measures included changes in serum biological markers (alkaline phosphatase, alanine transaminase, aspartate transaminase, and -glutamyltransferase), craving, and psychosocial functioning. Results: In the intention-to-treat analysis, there were no between-group differences for any of the primary drinking outcomes at 12 weeks. In post hoc exploratory analyses, the entire sample of participants was divided into two new groups: (1) people who drank during the 2 weeks before the start of medication (entry drinkers) and (2) people who did not drink during this interval (entry abstainers). Entry abstainers were at an advantage at study entry in that they were significantly more likely to have an inpatient hospitalization immediately before entry into outpatient treatment. Mixed-model analysis of variance revealed a main effect for entry group at the 12-week treatment endpoint on the primary outcome measures of percent days drinking, average drinks per drinking day, average drinks per day, heavy drinking days, and time to first heavy drinking day. Participants in any of the randomized groups who were entry abstainers had significantly better improvement on all of the primary outcome measures. The abstainer groups that were randomized to placebo and usual treatment had significantly better outcomes than the entry drinkers in those perspective groups. However, for the naltrexone-treated group, entry drinkers and entry abstainers had similar improvement in drinking-related outcomes. Conclusions: These data suggest that naltrexone may offer particular benefit to patients who continue to drink during the early stages of the trial as compared with those who have achieved abstinence before treatment entry. Key Words: Naltrexone, Community Treatment, Alcoholism, Treatment Outcome. NALTREXONE, AN OPIATE antagonist, was approved by the Food and Drug Administration for the treatment of alcoholism in 1994. Specifically, in several large, well-controlled, double-blind studies, naltrexone was shown to decrease percent of days drinking (PDD) and drinks per drinking day (DDD), as well as delay relapse to From the Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, Charleston South Carolina. Received for publication March 24, 2004; accepted September 23, 2004. This research was supported by a NIAAA funded grant (R01 AA11747) awarded to KTB. Reprint requests: Therese Killeen, PhD, APRN, Department of Psychiatry and Behavioral Science, Medical University of South Carolina, 67 President Street, Charleston, SC 29425; Fax: 843-792-3982; e-mail: killeent@musc.edu. Copyright 2004 by the Research Society on Alcoholism. DOI: 10.1097/01.ALC.0000145688.30448.2C heavy drinking (Anton et al., 1999; O Malley et al., 1992; Volpicelli et al., 1992). Furthermore, naltrexone may have a role in preventing progression to heavy drinking in individuals who sample alcohol (Anton et al., 1999; Guardia et al., 2002; Heinala et al., 2001; Volpicelli et al., 1992). Other studies have demonstrated that the effects of naltrexone are most robust in participants who adhere closely to the medication regimen (Chick et al., 2000; Oslin et al., 2002; Volpicelli et al., 1997). The psychosocial therapy used in conjunction with naltrexone has an impact on outcome. Cognitive behavioral therapy was used effectively in conjunction with naltrexone in studies done by Anton et al. (1999), Balldin et al. (2003), Heinala et al. (2001), Kiefer et al. (2003), and O Malley et al. (2003). O Malley et al. (1992) found that the combina- 1710 Alcohol Clin Exp Res, Vol 28, No 11, 2004: pp 1710 1717

NALTREXONE IN A COMMUNITY TREATMENT PROGRAM 1711 tion of supportive therapy and naltrexone was superior to naltrexone and coping skills in producing abstinence over a 12-week treatment period. However, in participants who sampled alcohol, naltrexone and coping skills were better in preventing relapse to heavy drinking. In many of the positive studies, manualized therapies were used by masters- or doctoral-level therapists who received intensive therapy training in an academic medical center (Anton et al., 1999; Balldin et al., 2003; Heinala et al., 2001; O Malley et al., 1992, 2003). Studies that used less intensive or less specified treatment had mixed results with regard to drinking outcomes (Chick et al., 2000; Gastpar et al., 2002; Guardia et al., 2002). In a study by Krystal et al. (2001), 627 veterans were randomized to naltrexone 50 mg daily for 12 months, naltrexone 50 mg daily for 3 months followed by 9 months of placebo daily, or placebo daily for 12 months. All three groups received 12-step facilitation therapy throughout the study. There were no differences in drinking outcomes between groups. In one commentary on the study, it was suggested that the lack of efficacy of naltrexone in this study might be related to the abstinence-based approach of the adjunctive 12-step therapy used (Sinclair et al., 2002). It was hypothesized that naltrexone may be effective only in programs that focus on drink reduction and preventing slips from becoming relapses. The notion of the moderating impact of adjunctive psychosocial treatment in determining the efficacy of naltrexone treatment is important because most treatment for alcoholism is delivered in community treatment programs that use a wide variety of psychotherapeutic treatment approaches, including strong emphasis on abstinence and 12-step therapy. More recently, a naltrexone depot was developed to address the problem of nonadherence to oral medications. In a large, double-blind, placebo-controlled study, 624 male and female participants were randomized to receive monthly injections of vivitrex 380 mg (naltrexone in injectable suspension), vivitrex 190 mg, or placebo injections for a total of 6 months. More than 60% of participants received all six injections. Heavy drinking was significantly diminished in the vivitrex 380 mg group versus the placebo group. This effect was significant in men but not in women. Men in the vivitrex 190 mg group also had a significant reduction in heavy drinking compared with the placebo group (Garbutt et al., 2004). A major limitation of previous naltrexone studies is the generalizability of findings to the typical alcohol-abusing patient commonly seen in community treatment programs. Interpretation of these studies is limited by the stringent inclusion and exclusion criteria used for enrollment. This limitation is compounded by the suggestion that naltrexone may be effective only in combination with specific psychosocial therapies. Effectiveness studies have the advantage of exploring medication usage in a more naturalistic setting. In the real-world treatment setting, patients with alcohol use disorders present into treatment with many problems, including psychiatric and medical comorbidity. Most studies required at least 5 days of abstinence and psychosocial stability, as well as exclusion of other substances of abuse, psychotropic medications, and legal involvement (Anton et al., 1999; Chick et al., 2000; Guardia et al., 2002; Heinala et al., 2001; Krystal et al., 2001; O Malley et al., 1992; Volpicelli et al., 1992). In addition, some studies discontinued medication in patients who relapsed (Guardia et al., 2002; Kiefer et al., 2003). Exploring subgroups of patients who respond to naltrexone may help to explain the conflicting results seen in previous studies. In this study, the safety and effectiveness of naltrexone treatment combined with usual treatment was compared to the usual treatment with placebo and usual treatment alone in a front-line community treatment program. To enhance the generalizability of the findings, we applied minimal exclusion criteria. MATERIALS AND METHODS The study was submitted and approved by a convened full Institutional Review Board. Individuals who entered treatment at an outpatient community treatment program for an alcohol use disorder were recruited for this study. Participants were either approached by the research assistant at initial intake or referred by their counselor within the first 2 weeks of treatment. The research assistant explained the study to patients who entered treatment for alcohol use disorders. When an individual was interested in participation, informed consent was obtained. A brief inclusion/exclusion questionnaire was used to determine initial study eligibility. Because this was an effectiveness study in a naturalistic clinic setting, only exclusion criteria that were essential for patient safety and study validity were applied. Inclusion criterion was a current alcohol use disorder with reported drinking in the 30 days before study entry. Exclusion criteria were (1) current addiction to opiates; (2) women who were pregnant, breastfeeding, or of childbearing age and not practicing effective birth control; (3) serious medical conditions or liver enzymes more than three times the normal range; (4) cognitive dysfunction to an extent that impaired understanding of informed consent or assessments; and (5) having had 10 days of outpatient treatment in the past 3 months. Participants were evaluated by a psychiatrist or a psychiatric clinical nurse specialist for physical examination and psychiatric diagnoses. Substance use disorders and co-occurring psychiatric diagnoses were made using DSM-IV criteria (American Psychiatric Association, 1994). Participants were randomized into one of three conditions: naltrexone 50 mg daily for 12 weeks plus treatment as usual, placebo daily for 12 weeks plus treatment as usual, and treatment as usual alone. The average time that elapsed between enrollment and start date was 6.61 6.6 days. For participants in the usual treatment alone group, study start dates were determined by yoking them to the number of days between consent date and medication start for the previous participant in the medication treatment group. Urn randomization was used to balance certain variables across groups (Stout et al., 1994; Wei and Lachin, 1988). Urn variables were gender, having one or more other Axis I psychiatric diagnoses, having one or more other substance use disorders, alcohol dependence severity as determined by a score of 22 or higher on the Alcohol Dependence Scale (ADS), court-ordered treatment, treatment intensity ( 6 hr/week versus 6 hr/ week), and being in a controlled environment for 3 days in the past 30 days before study entry. This study was conducted between February 1998 and June 2003 in a public treatment program in a rural community ~30 miles from Charleston, South Carolina. The program serves ~1000 patients per year in a variety of services, including intensive outpatient, court related, preven-

1712 KILLEEN ET AL. tion, aftercare, and counseling programs. Study participants were enrolled in several different treatment programs offered at the clinic. Treatment intensity ranged from 1 to 2 hr per week of group and/or individual therapy to intensive group programs for 3 to 4 hr, 4 days per week for ~6 weeks. Therapy was eclectic and included 12-step and relapse prevention approaches. Patients were strongly encouraged to attend Alcoholics Anonymous self-help groups in the community. As per clinic policy, breathalyzer and random urine drug screens were done routinely, and patients who continued to drink or use other substances while in treatment were often referred by their counselors to inpatient treatment. Assessments The Time Line Follow-Back (TLFB) self-report drinking was the primary outcome measure (Sobell and Sobell, 1995). Using a calendar, individuals recall estimates of their daily drinking over a specified time period. Relevant calendar events are used as probes to enhance recall. This was collected at baseline for the 90 days before study entry and at 12, 24, and 48 weeks of follow-up. The following measures were collected at baseline and 12, 24, and 48 weeks. Psychosocial functioning was assessed with the Addiction Severity Index (ASI), a semistructured interview that collects data in seven different psychosocial areas that are adversely affected by alcohol and drugs (McLellan et al., 1992). The Obsessive Compulsive Drinking Scale (OCDS), a 14-item self-report questionnaire, was used to assess alcohol craving. Specifically, subscale scores address the obsessive thoughts associated with craving for alcohol and compulsive drinking behavior (Anton et al., 1996). Alcohol severity was assessed using the ADS, a 25-item self-report questionnaire (Skinner and Allen, 1982). A symptom checklist to evaluate medication adverse effects and liver enzymes, including -glutamyl transferase (GGT), alanine aminotransferase (ALT), and aspartate aminotransferase (AST), were collected during the treatment study at weeks 4, 8, and 12. Carbohydrate-deficient transferrin, a biological drinking marker, was collected at baseline and week 12. Participants in the two medication groups were seen at baseline and 4, 8, and 12 weeks by the physician or the clinical nurse specialist for medication management. At each study visit, 40 tablets of the medication were dispensed. Adherence to medication was measured using the APREX Medication Electronic Monitoring System (MEMS), an electronic system designed to monitor pill bottle openings. Service utilization, TLFB, symptom checklist, liver enzymes, vital signs, breathalyzer reading, pill count, and MEMS cap readings for the two medication groups were obtained at the monthly visits. Participants were not given any feedback about adherence information obtained from the MEMS cap reading but were encouraged to take their medications. Participants in the usual treatment alone were interviewed for service utilization and self-report drinking (TLFB) at 4 and 8 weeks either by telephone or in person if they were attending their treatment program. At 12 weeks, all three groups were seen in person for a comprehensive follow-up assessment that included TLFB, ASI, OCDS, ADS, and biological markers. Follow-up visits were also conducted at 24 and 48 weeks after study entry for all study participants. These data will be reported in a later publication. Analysis Baseline group differences were examined with the use of analysis of variance for continuous data and 2 tests for dichotomous data for the intention-to-treat analysis. Primary and secondary outcome variables were analyzed using analysis of covariance with baseline measures as covariates. Primary dependent measures included ADD, PDD, DDD, and heavy drinking days (HDD) as measured by the TLFB. Heavy drinking days consisted of four or more drinks per day for women and six or more drinks per day for men. Secondary outcome measures included biological markers as measured by GGT, ALT, and AST; craving as measured by the OCDS; and psychosocial functioning as measured by the ASI. Kaplan- Meier survival analysis was performed on time to first drink and time to first HDD. In conducting the initial data analysis, it was noted that a cluster of participants were drinking at the time of entry into the study, whereas the other participants reported no alcohol consumption after signing informed consent. Because this phenomenon of entry drinking differences might have an impact on outcome, in a post hoc analysis, participants were divided into two groups called entry drinkers and entry abstainers. Entry drinkers were participants who reported any drinking in the 2 weeks before medication initiation. Entry abstainers were participants who were abstinent during this same interval. Treatment outcome data analysis was performed on the three randomized groups categorized into entry drinkers and entry abstainers. These six groups were naltrexone entry drinkers, naltrexone entry abstainers, placebo entry drinkers, placebo entry abstainers, usual treatment entry drinkers, and usual treatment entry abstainers. RESULTS Of the 191 patients screened, 145 were enrolled and randomized into one of the three groups. A total of 133 participants had at least one follow-up visit. Baseline demographic characteristics of the randomized groups are displayed in Table 1. No between-group baseline differences were detected. In general, participants mostly were male (63%), were white (76%), were an average of 37.3 ( 8.75) years of age, and had an average of 11.7 ( 2.2) years of education. There were no demographic differences between participants who were randomized and participants who were screening dropouts. The gender and ethnicity of this sample are representative of the county clinic population (62% male, 72% white). Sixty-six percent of the state clinic population is male and 62% is white (Department of Alcohol and Other Drug Abuse Services, 2003). Forty-nine percent of the state population is male and 70% is white. Blacks constitute the majority of the minority population in the state. The number of blacks who reside in rural counties across the state varies from as little as 7 to as much as 70% (US Census Bureau; quickfacts.census.gov/qfd/ststes/45000.html). Sixty-eight (51%) participants had a comorbid psychiatric disorder, and 47 (35%) had a substance use disorder in addition to alcohol abuse/dependence. The majority (88%) of participants were enrolled in an intensive outpatient (4 days/week) program. The only difference between groups was employment status. There was a trend (p 0.06) for more participants in the naltrexone group to be unemployed than in the other two groups. Medication adherence rates are displayed in Table 2. Adherence to medication was assessed by the percentage of MEMS cap openings, which represents the percentage of days that participants took their medication. There were no significant adherence differences between the naltrexone and placebo groups. The mean percentage of days that participants on naltrexone were adherent was 0.51 ( 0.35) versus 0.61 ( 0.31) for placebo. Sixty-seven percent of participants who stopped medications dropped out of clinic treatment. There were no differences in reasons for termination of medication between the groups. Follow-up for the entire group was 79% at 4 weeks, 74% at 8 weeks, and 72% at 12 weeks. Participants in the

NALTREXONE IN A COMMUNITY TREATMENT PROGRAM 1713 Table 1. Baseline Characteristics by Treatment Group Total (n 133) Naltrexone (n 51) Placebo (n 36) Usual treatment (n 46) Statistic p value Age 37.3 8.75 37.9 7.8 35.5 7.8 38.0 10.3 F 1.41 p 0.25 Gender Male 84 (63%) 30 (58%) 20 (56%) 34 (74%) 2 3.59 Female 49 (37%) 21 (41%) 16 (44%) 12 (26%) p 0.17 Ethnicity White 101 (76%) 39 (76%) 29 (81%) 33 (72%) 2 2.12 Black 29 (22%) 11 (22%) 7 (19%) 11 (23%) p 0.16 Other 3 (2%) 1 (2%) 2 (4%) Marital status Married 37 (28%) 20 (39%) 13 (29%) 11 (23%) 2 7.21 Separated/divorced 55 (41%) 21 (41%) 14 (31%) 22 (47%) p 0.51 Single/widow 41 (31%) 9 (18%) 16 (36%) 13 (28%) Education 11.7 2.2 11.66 1.9 12.0 2.9 11.9 2.0 F 0.218 p 0.8 Employed 73 (55%) 22 (43%) 20 (56%) 31 (67%) 2 5.75 p 0.06 Other Axis I 68 (51%) 27 (53%) 18 (50%) 23 (50%) 2 0.11 p 0.95 Other substance use disorder 47 (35%) 18 (35%) 14 (39%) 15 (33%) 2 0.38 p 0.84 ADS score 22 33 (25%) 12 (24%) 6 (17%) 15 (33%) 2 2.82 p 0.24 Court ordered 77 (58%) 29 (57%) 24 (67%) 24 (52%) 2 1.77 p 0.41 Controlled environment 3 days in past month 102 (77%) 11 (22%) 10 (22%) 10 (28%) 2 0.55 p 0.76 Previous treatments 70 (53%) 24 (47%) 24 (67%) 22 (48%) 2 5.53 p 0.24 Treatment intensity 6 hr per week 117 (88%) 45 (88%) 32 (89%) 40 (87%) 2 0.08 p 0.96 Table 2. Medication Adherence Between Groups Naltrexone (n 51) Placebo (n 36) Statistic p value Stopped medication 24 (47%) 12 (33%) 2 1.639 p 0.27 Stopped because of side effects 9 (38%) 4 (25%) 2 0.563 p 0.71 Mean % days taking 51% 0.35 61% 0.31 F 2.639 medication p 0.17 placebo group had a slightly lower follow-up rate (67%) than the treatment as usual (75%) and the naltrexone (76%) groups at 12 weeks. Those with 12-week follow-up were not significantly different from dropouts with regard to demographics or baseline drinking variables. Dropouts were distributed equally across the three randomized groups. However, there was a trend for dropouts to be actively drinking at the time of study entry (56% for dropouts vs. 38% for those who completed 12-week follow-up; 2 3.98; df 1; p 0.06). Table 3 displays adverse events experienced by participants in both medication groups during the 12-week treatment period. In general, naltrexone was well tolerated despite that many participants were actively drinking and/or using other substances of abuse, experiencing other comorbid disorders, and/or taking other psychotropic medications. Significantly more participants in the naltrexone group experienced daytime sleepiness, Symptom Table 3. Adverse Events Naltrexone (n) Placebo (n) Statistic Trouble falling asleep 5 3 NS Trouble staying asleep 4 4 NS Daytime sleepiness 10 2 0.039 Nervousness 3 1 NS Irritability 9 2 NS (0.064) Depressed 2 0 NS Inability to relax 2 2 NS Abdominal pain 4 2 NS Nausea and/or vomiting 10 3 NS (0.098) Decreased appetite 4 1 NS Fatigue 5 0 0.041 Headache 8 7 NS Dizziness 5 0 0.041 Itching 1 2 NS NS p 0.05. fatigue, and dizziness as compared with the placebo group (p 0.05). There was also a trend for more participants in the naltrexone group to experience irritability and nausea as compared with the placebo group. There were no significant differences in liver enzyme values (GGT, AST, and ALT) at 12 weeks between the two medication groups. All three groups had lower mean GGT (67.9 74.8 at baseline vs. 48.4 70.1 at 12 weeks), ALT (41.4 35.7 at baseline vs. 32.2 18 at 12 weeks), and AST (33.4 20.4 at baseline vs. 30.8 15.2 at 12 weeks) at 12 weeks compared with baseline (p 0.001).

1714 KILLEEN ET AL. Fig. 1. Drinking outcomes for randomized groups. TAU, treatment as usual; PLB, placebo; NTX, naltrexone. *p 0.001 within treatment differences from baseline to 12 weeks. Primary Outcome TLFB data for baseline and 12-week treatment period for all three groups are displayed in Fig. 1. For all three groups, there was a significant decrease in alcohol consumption during the 12-week period as compared with alcohol consumption in the 90 days before informed consent (baseline). There were no significant between-group differences in ADD (F 2.02; df 2; p 0.14), DDD (F 1.59; df 2; p 0.21), PDD (F 1.15; df 2; p 0.32), and HDD (F 1.48; df 2; p 0.23) during the 12-week study period. There were no between-group differences in time to first heavy drinking day ( 2 2.36; p 0.3) for those who had 12-week data available or up until the time of dropout. Forty-eight percent of participants in the usual treatment group, 41% of participants in the naltrexone group, and 33% of participants in the placebo group had at least one episode of heavy drinking during the 12-week treatment period ( 2 1.17; p 0.28). Craving, as measured by the OCDS, was significantly reduced at 12 weeks for all groups, but there were no between-group differences. Similarly, there was significant improvement for all groups in psychosocial function as measured by the ASI but no between-group differences. One of the difficulties in detecting between-group differences lies in the dramatic decrease in drinking in all groups during the 12-week study period. A large subgroup of individuals stopped drinking after study entry during the assessment period before medications were initiated, and we were interested in further exploring and characterizing this group. For a post hoc exploratory analyses, the entire sample of participants was divided into two new groups: (1) people who drank during the 2-week period after signing consent before the start of medication (entry drinkers) and (2) people who did not drink during this interval (entry abstainers). Thus, each treatment group was subdivided into two groups on the basis of entry drinking, resulting in six groups. Baseline characteristics of participants in these groups are displayed in Table 4. There were no betweengroup differences in age, gender, ethnicity, marital status, employment, education, comorbid psychiatric disorder, previous treatment, and treatment intensity. There was, however, a significantly larger number of individuals who had been in a controlled environment for 3 days in the month before study entry in the entry abstainer group, indicating that a higher number of these individuals had been in inpatient alcohol treatment. Entry drinkers were also more likely to be enrolled in less intense treatment as usual. An analysis of the TLFB data then was performed taking into account entry drinking status as well as randomized groups. There were no significant differences between randomized groups in the proportion of participants who were entry drinkers and entry abstainers ( 2 1.54; df 2; p 0.5). Analysis of covariance revealed a main effect for entry group and for interaction between entry group and treatment groups as randomized for the 12-week treatment endpoint on PDD (F 5.66; df 5; p 0.001), ADD (F 3.69; df 5; p 0.004), HDD (F 3.76; df 5; p 0.003), and DDD (F 2.62; df 5; p 0.027). As might be expected, the entry drinker group had higher PDD, DDD, HDD, and DDD during the 12-week treatment period as compared with entry abstainers in all three treat-

NALTREXONE IN A COMMUNITY TREATMENT PROGRAM 1715 Table 4. Baseline Characteristics by Entry Abstainer and Entry Drinker Groups Entry abstainers (n 81) Entry drinkers (n 52) Statistic p value Age 38.1 8.25 35.9 9.38 t 1.44 0.15 Gender Male 48 (59%) 36 (69%) Female 33 (41%) 16 (31%) 2 1.353 0.27 Ethnicity White 65 (80%) 36 (69%) Black 15 (19%) 14 (27%) Other 1 (1%) 2 (4%) 2 2.490 0.29 Marital status Married 23 (28%) 14 (27%) Separated/divorced 36 (44%) 19 (37%) Single 19 (23%) 19 (37%) 2 3.88 0.42 Education 11.6 2.4 12.2 2.0 F 2.30 0.13 Employed 45 (55%) 34 (56%) 2 0.01 1.0 Other axis I disorder 45 (56%) 23 (44%) 2 1.62 0.22 Other substance use 30 (37%) 17 (33%) 2 0.262 0.71 ADS score 22 22 (27%) 11 (21%) 2 0.612 0.5 Court ordered 43 (53%) 34 (65%) 2 1.970 0.21 Controlled environment 3 days in past month 25 (31%) 6 (12%) 2 6.617 0.01* Previous treatment 43 (53%) 27 (52%) 2 0.110 0.94 Treatment intensity 6 hr per week 6 (7%) 10 (19%) 2 4.183 0.06 Data are n SD or %. Fig. 2. Drinking outcomes at 12 weeks adjusted for baseline in randomized treatment groups and entry status. ED, entry drinkers; EA, entry abstainers. ment groups. These differences were significant for the treatment as usual and placebo groups. For the naltrexone group, however, there were not significant differences between the entry drinkers and entry abstainers in drinking outcomes during the 12-week study. Entry drinkers in the naltrexone group had significantly lower ADD (p 0.01), PDD (p 0.01), and HDD (p 0.006) than entry drinkers in the treatment as usual group. However, drinking outcomes in the naltrexone entry drinkers were not significantly different from the placebo entry drinkers. As can be seen in Fig. 2, for the naltrexone-treated group, the improvement in drinking-related outcomes in the entry drinker group was similar to that seen in the abstainer group. In the placebo and usual treatment groups, entry abstainers had significantly better drinking outcomes than the entry drinkers. Survival analysis curves for time to first HDD for the randomized treatment groups divided into entry abstainer and entry drinkers are presented in Fig. 3. For the placebo and naltrexone groups, there were no differences between time to first HDD for entry drinkers and entry abstainers (p 0.12, p 0.2; log-rank 2, respectively). For the

1716 KILLEEN ET AL. Fig. 3. Meyer-Kaplan survival curves for days to first heavy drinking in randomized groups and entry status. usual treatment group, there was a significantly longer time to first HDD in the entry abstainers compared with the entry drinkers ( p 0.0002; log-rank 2 ). There were no differences between the drinking entry groups on any of the craving, psychosocial functioning, or biological measures. DISCUSSION This study was designed to explore the use of naltrexone in a group of alcohol-dependent individuals who presented for treatment at front-line community treatment programs using minimal exclusion/inclusion criteria and not controlling the adjunctive psychosocial treatment. As such, this study is an effectiveness trial designed to build on the demonstrated efficacy of naltrexone in well-controlled clinical trials. Typically, individuals who present for treatment at community treatment programs have numerous psychosocial, medical, and psychiatric comorbidities that often make them ineligible for efficacy studies. The philosophy, intensity, and content of therapy offered in the treatment program can affect outcomes. It has been shown that the efficacy of naltrexone varies depending on the adjunctive psychosocial therapy being delivered. Previous studies have shown lower relapse rates when cognitive behavioral therapy or relapse prevention is used in conjunction with naltrexone (Anton et al., 1999; Heinala et al., 2001; Krystal et al., 2001). One study found higher rates of continuous abstinence with naltrexone plus supportive therapy (O Malley et al., 1996). O Malley et al. (2003) found comparable improved drinking outcomes using primary care management plus naltrexone as with cognitive behavioral therapy plus naltrexone. In a large study that used 12-step facilitation therapy aimed at abstinence with naltrexone or placebo, there were no differences in drinking outcomes between groups at 13 and 52 weeks (Krystal et al., 2001). In the present study, 85% of the participants were referred to intensive treatment services (3 4 hr, 4 days/week). Treatment was eclectic and abstinence based. Patients were strongly encouraged to attend community 12-step meetings. As in other studies dealing with heterogeneous and real life treatment settings (Chick et al., 2000; Krystal et al., 2001), we found no differences on drinking outcome measures among the naltrexone, placebo, and usual treatment groups. As with these studies, we had a substantial attrition rate, marginal medication adherence, and a treatment program that was abstinence based. As in other effectiveness trials, patients with psychosocial instability, polysubstance use, and psychiatric comorbidity are less likely to adhere to medications. The use of specific adherence enhancement procedures or the use of depot (injectable) naltrexone might circumvent those barriers. In the treatment setting for this study, participants who came to treatment and reported drinking/drug use or tested alcohol/drug positive on biological indices were often referred by their counselors to inpatient treatment. This clinic policy is likely to have contributed to the high study attrition. Of interest in the present study, participants who were taking naltrexone and reported drinking just before the initiation of study medications did as well on drinking outcomes as participants who were abstinent immediately before initiation of study medications. In the treatment as usual and placebo groups, the entry drinkers had substantially poorer outcomes as compared with entry abstainers. This suggests that naltrexone may have preferential efficacy on the subgroup of alcohol-dependent individuals who are actively drinking while in treatment. This finding is consistent with the findings by Heinala et al. (2001), who found that naltrexone plus coping skills therapy was beneficial in patients who were actively drinking. They proposed that naltrexone might react by extinguishing reinforced drinking behavior, making it essential that an individual continue with some drinking after naltrexone initiation for the medication to exert its therapeutic effect. Anton et al. (1999) found that participants in the naltrexone group had prolonged time to relapse after sampling alcohol than participants in the placebo group. This outcome has clinical significance in that it prevents a slip from becoming a

NALTREXONE IN A COMMUNITY TREATMENT PROGRAM 1717 full-blown relapse. Taken together, these studies offer some explanation for the conflicting results reported in the literature concerning the effectiveness of naltrexone in alcohol use disorders. Importantly, these studies also suggest that there may be a subgroup of alcoholics who are more likely to respond to naltrexone. Subtyping of alcoholdependent individuals such that specifically tailored treatment can be delivered is an area that has received much attention and could improve the cost-effectiveness of treatment enormously. One of the limitations of this study is related to overall study and subgroup sample size. The study was powered to detect a small to moderate effect size (0.4 on Cohen s range) for the psychosocial variables based on previous studies at our institution. However, the heterogeneity in our population resulted in large SDs in most of our measures. Furthermore, the large impacts of both usual treatment alone and usual treatment with placebo therapy eroded the number of subjects in the active drug group who could benefit from the effect of naltrexone. For example, at an level of 0.05 with a two-sided test, we had only 51% power to detect a decrease of 1.4 on ADD at 12 weeks as a result of naltrexone once the effect of usual treatment was accounted for. The only reason that we found a statistically significant difference for the subgroup of clients who drank at study entry and were randomized to naltrexone was that the incremental effect size as a result of naltrexone for this group was very large (1.1 on Cohen s range). The present study has important implications for the use of naltrexone in community treatment programs. Naltrexone that is given to patients who are having difficulty achieving abstinence at the onset of outpatient treatment may prevent such patients from being prematurely referred to expensive inpatient services. REFERENCES American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association, Washington, DC. Anton RF, Moak DH, Latham PK (1996) The obsessive-compulsive aspects of craving: A new method of assessing outcome in alcoholism treatment studies. Arch Gen Psychiatry 53:225 231. Anton RF, Moak DH, Waid LR, Latham PK, Malcolm RJ, Dias JK (1999) Naltrexone and cognitive behavioral therapy for the treatment of outpatient alcoholics: Results of a placebo-controlled trial. Am J Psychiatry 156:1758 1764. Balldin J, Berglund M, Borg S, Mansson M, Bendtsen P, Franck J, Gustafsson L, Hallidin J, Nilsson L-H, Stolt G, Willander A (2003) A 6-month controlled naltrexone study: Combined effect with cognitive behavioral therapy in outpatient treatment of alcohol dependence. Alcohol Clin Exp Res 27:1142 1149. Chick J, Anton R, Checinski K, Croop R, Drummond DC, Farmer R, Labriola D, Marshall J, Moncrieff J, Morgan MY, Peters T, Ritson B (2000) A multicentre, randomized, double-blind, placebo-controlled trial of naltrexone in the treatment of alcohol dependence or abuse. Alcohol 35:587 593. Department of Alcohol and Other Drugs Abuse Services (2003) Total client intakes for 2003: country race gender. Substance Abuse Agency Management System. Report, Columbia, SC. Garbutt, JC, Kranzler H, Gastfriend D, O Malley S, Pettinati H, Silverman B, Loewy J Ehrich E (2004) A phase III study of vivitrex (naltrexone in injectable suspension) in alcohol-dependent adults: Efficacy and tolerability (abstract). Alcohol Clin Exp Res 28:148A. Gastpar M, Bonnet U, Boning J, Mann K, Schmidt LG, Soyka M, Weterling T, Keilstein V, Labriola D, Croop R (2002) Lack of efficacy of naltrexone in the prevention of alcohol relapse: Results from a German multicenter study. J Clin Psychopharmacol 22:592 598. Guardia J, Caso C, Arias F, Gual A, Sanahuja J, Ramirez M, Mengual I, Gonzalvo B, Seguro L, Trujols J, Casas M (2002) A double-blind, placebo-controlled study of naltrexone in the treatment of alcoholdependence disorder: results from a multicenter clinical trial. Alcohol Clin Exp Res 26:1381 1387. Heinala P, Alho H, Kiianmaa K, Lonnqvist J, Kuoppasalmi K, Sinclair JD (2001) Targeted use of naltrexone without prior detoxification in the treatment of alcohol dependence: A factorial double-blind, placebocontrolled trial. J Clin Psychopharmacol 21:287 292. Kiefer F, Jahn H, Tarnaske T, Helwig H, Briken P, Holzbach R, Kampf P, Stracke R, Baehr M, Naber D, Widemann K (2003) Comparing and combining naltrexone and acamprosate in relapse prevention of alcoholism: A double-blind, placebo-controlled study. Arch Gen Psychiatry 60:92 99. Krystal JH, Cramer JA, Krol WF, Kirk GF, Rosenheck RA (2001) Naltrexone in the treatment of alcohol dependence. N Engl J Med 345: 1734 1739. McLellan AT, Kushner H, Metzger D, Peters R, Grissom G, Pettinatti H, Argeriou M (1992) The Fifth Edition of the Addiction Severity Index. J Subst Abuse Treat 9:199 213. O Malley SS, Jaffe AJ, Chang G, Rode S, Schottenfeld RS, Meyer RE, Rounsaville B (1996) Six month follow-up of naltrexone and psychotherapy for alcohol dependence. Arch Gen Psychiatry 53:217 224. O Malley SS, Jaffe AJ, Chang G, Schottenfeld RS, Meyer RE, Rounsaville B (1992) Naltrexone and coping skills therapy for alcohol dependence. A controlled study. Arch Gen Psychiatry 49:881 887. O Malley SS, Rounsaville BJ, Farren C, Namkoong K, Wu R, Robinson J, O Conner PG (2003) Initial and maintenance naltrexone treatment for alcohol dependence using primary care vs specialty care. Arch Intern Med 163:1695 1704. Oslin DW, Pettinati H, Volpicelli JR (2002) Alcoholism treatment adherence: Older age predicts better adherence and drinking outcomes. Am J Geriatr Psychiatry 10:740 747. Sinclair J, Alho H, Shinderman M (2002) Naltrexone in the treatment of alcohol dependence. N Engl J Med 346(17):1130. Skinner HA, Allen BA (1982) Alcohol dependence syndrome: Measurement and validation. J Abnorm Psychol 91:199 209. Sobell LC, Sobell MB (1995) Alcohol Timeline Follow Back Users Manual. Addiction Research Foundation, Toronto, Ontario, Canada. Stout RL, Wirtz PW, Carbonari JP, Del Boca FK (1994) Ensuring balanced distribution of prognostic factors in treatment outcome research. J Stud Alcohol Suppl 12:70 75. Volpicelli JR, Alterman AI, Hayashida M, O Brien CP (1992) Naltrexone in the treatment of alcohol dependence. Arch Gen Psychiatry 49:876 880. Volpicelli JR, Rhines KC, Rhines JS, Volpicelli LA, Alterman AI, O Brien CP (1997) Naltrexone and alcohol dependence. Role of subject compliance. Arch Gen Psychiatry 54:737 742. Wei L, Lachin J (1988) Properties of the urn randomization in clinical trials. Control Clin Trials 9:345 364.