Extended-Release Injectable Naltrexone (XR-NTX) with Intensive Psychosocial Therapy for Amphetamine Dependent Persons Seeking Treatment: A Placebo-Controlled Trial

Abstract

Objective

Explore the efficacy of XR-NTX for preventing relapse to amphetamine use.

Method

Clinical trial of 100 treatment-seeking patients with amphetamine dependence that were randomized to 6 monthly 380 mg doses of XR-NTX or matching placebo before entering intensive outpatient after varying lengths of inpatient treatment in Reykjavik, Iceland. Weekly urine drug tests, retention, and standardized instruments assessed efficacy.

Results

Of 169 approached, 100 were randomized. Though amphetamine dependence was the main reason for seeking treatment, three quarters or more of participants had one or more other substance dependencies. Of 51 randomized to XR-NTX, 20 received 4 or more injections; of 49 assigned to placebo, 26 received 4 or more injections. Of the planned 2400 weekly urine drug tests, 1247 were collected (52%); 4% of these were positive for amphetamine, 8% for benzodiazepine, 7% for marijuana, 1% for cocaine, and 1% for opioid. XR-NTX had no effect on amphetamine positive tests, retention, or other outcomes. Those providing half or more of their tests attended more weeks of treatment than those providing less than half of their tests (m = 10.76 vs. 3.31; t (92) = 5.91, p < .0001), and 92 participants provided at least one test.

Conclusions

Adding XR-NTX to the usual combination of inpatient and intensive outpatient treatment did not reduce amphetamine use. The low prevalence of substance use among collected urine samples, and the association between collected samples and weeks in treatment, was consistent with other studies showing that staying in treatment is associated with better outcomes.

Amphetamine and methamphetamine dependence are growing problems with over 34 million individuals estimated to be using them non-medically worldwide. This figure is greater than the number using heroin and cocaine combined, thus placing amphetamines second only to cannabis as the most commonly used illicit drug class (1–2). Treatment for amphetamine use disorders is typically based on psychosocial interventions that include individual and/or group therapy guided by the 12-Steps, motivational interviewing, cognitive behavioral therapy, relapse prevention, and/or contingency management. Studies have shown that these treatments improve outcomes but gains are typically modest and short-lived (3).

Several pharmacotherapies have been studied in hopes of finding one that improves outcomes. Among them have been opioid antagonists based on the idea that blockade of opioid receptors may reduce the rewarding effects of endorphin-stimulated dopamine release. For example, a preclinical study showed that naloxone pre-treatment reduced amphetamine-induced increases in extracellular dopamine in rats (Schad et al, 1995); another showed that naltrexone attenuated reinstatement of amphetamine self-administration in rats (Haggkvist et al); and a third showed that naltrexone decreased amphetamine and ethanol self-administration in primates (Jiminez-Gomez et al, 2011). In clinical studies, Jayaram-Lindstrom et al found that naltrexone reduced amphetamine-reinforcing effects in healthy volunteers (5); was well tolerated by amphetamine dependent patients seeking treatment (2005); and blocked amphetamine induced craving in detoxified amphetamine dependent persons waiting to begin treatment for Attention Deficit Hyperactivity Disorder (2008). Ray et al (2015) and Marks et al (2016) were also encouraging as they showed that naltrexone reduced amphetamine subjective effects in non-treatment seeking individuals meeting DSM-IV criteria for amphetamine abuse or dependence.

Jayaram-Lindstrom et al further explored these findings in a 12-week open trial of 20 amphetamine dependent patients treated with 50mg/day oral naltrexone and weekly cognitive behavior therapy. Most patients tolerated naltrexone, complete abstinence was obtained by two patients, and there was a significant decrease in the level and frequency of amphetamine use across the entire sample with a marked decrease in amphetamine craving (6). A third study found that naltrexone attenuated amphetamine subjective effects in amphetamine dependent patients (7), and a fourth study was a randomized trial of 20 amphetamine dependent patients who received naltrexone/amphetamine or placebo/amphetamine with a one-week washout period between conditions and showed reductions in amphetamine subjective effects and craving in the naltrexone condition as compared to placebo (8).

The fifth and largest study by Jayaram et al (9) randomized 80 amphetamine dependent, treatment-seeking patients to a 12-week course of 50 mg/day oral naltrexone or matching placebo. Patients had to meet DSM-IV criteria for amphetamine dependence, to have used amphetamines on at least 12 days in the past 12 weeks, and to have been able to maintain abstinence for two weeks prior to study entry. As in most of the other studies, patients were excluded if they had a DSM diagnosis of another substance use disorder except nicotine and if they had serious medical or psychiatric problems. Study participants provided urine samples twice weekly – once when they met with a psychologist for relapse prevention therapy and then when they met with a study nurse for vital signs and a check on progress and adverse events.

The primary outcome was amphetamine-negative urine tests with missing counted positive and results showed that naltrexone patients had more amphetamine-negative tests than placebo patients (65% vs 48%; p<0.05). Patients who provided at least 16 of the 24 scheduled urines were analyzed separately, and naltrexone patients had significantly more negative tests than placebo patients (79.7% vs. 64.1%; p<0.05). Secondary outcomes showed longer time to relapse, reductions in craving and self-reported use, and more continuous abstinence in naltrexone compared to placebo patients. There was no difference in retention between groups but there was a positive correlation between adherence to naltrexone and amphetamine-negative tests (9). These findings were consistent with preclinical studies summarized by Ray et al (10) and later explored in a double-blind, randomized crossover, placebo-controlled study in which 50 mg/day oral naltrexone blunted cue-induced methamphetamine craving and attenuated several hedonic effects of methamphetamine in 30 non-treatment seeking methamphetamine dependent individuals (10). Though some studies did not find a naltrexone effect (Grant et al, 2010; Stoops et al, 2014), both were underpowered (31 and 7 subjects, respectively), thus most of the available data indicate that naltrexone could be a meaningful addition to psychosocial treatment for amphetamine dependence.

Here, we present findings from a study of amphetamine dependent patients in Reykjavik, Iceland where, unlike those summarized above, patients were not selected for having only amphetamine dependence and/or being able to maintain abstinence during a pre-enrollment outpatient phase. Iceland was chosen as the study site due to the likely ethnic similarity of patients to those in the Jayaram-Lindstrom et al randomized trial (9) and the availability of a structured treatment programs that is operated by the the national center for addiction medicine at the Society for Alcohol and other Addictions (SAA) and is free of charge to all Icelandic citizens. The medication was extended release injectable naltrexone (XR-NTX), a product with a more reliable and quantifiable delivery system than the 50 mg tablet used in earlier studies.

Similar to global trends, Iceland has seen a growing amphetamine problem since the 1990’s with about 40% (N=800) of annual admissions having a diagnosis of amphetamine dependence at the time the study began (Society for Alcohol & Addiction report, 2010). Based on the findings from earlier studies, we hypothesized that a 24-week trial of monthly XR-NTX would prevent relapse, improve retention, and reduce craving when added to usual treatment. In Iceland, treatment begins with a 7–10 day course of inpatient detoxification and assessment at Vogur Hospital, a specialty addiction treatment facility, followed by referral to 30 days of residential or intensive outpatient treatment with transition to twice weekly group and/or individual therapy over 3–12 months, encouragement to participate in 12-Step programs, and provided housing and social support services as needed. The study protocol was approved by regional authorities, the Icelandic National Bioethics Committee and Data Protection Authority and the University of Pennsylvania IRB, and conducted in accord with Good Clinical Practice Guidelines and the Declaration of Helsinki.

METHOD

Participants

Persons aged 18 or above seeking treatment for amphetamine dependence at the SAA were approached by research and clinical staff about study participation. Study procedures were explained to those that expressed interest and written informed consent was obtained. Inclusion criteria were: primary diagnosis of amphetamine dependence according to DSM-IV-TR; 10 or more days of amphetamine use in the past month; successful completion of baseline measures during a 7–10 day inpatient detoxification and assessment at Vogur; and confirmed abstinence from all substances (except nicotine) unless used to treat alcohol withdrawal. Exclusion criteria were: physiological dependence on opioids; an opioid positive urine or positive naloxone challenge prior to receipt of study medication; AST or ALT >5 times the upper limit of normal; creatinine clearance <50; history of a major psychiatric disorder (schizophrenia, bipolar I); current severe depression; suicidal or homicidal ideation; dementia or a medical or social condition likely to interfere with the ability to consent or participate; needing opioids for chronic pain; likely to have surgery in the next 6 months; known hypersensitivity to naltrexone or other components of XR-NTX; and pregnancy, lactation, or not willing to use acceptable methods of birth control if female.

Treatment

Usual Treatment at SAA begins with inpatient assessment and detoxification, as mentioned above, and is based on the principle of equal access in the context of a national consensus that substance use disorders are health problems needing long-term treatment, and that treatment helps. It is delivered by a team of physicians, nurses and addiction counselors and free of charge except for outpatient services that require a small co-payment. Medication is used when indicated but the backbone of therapy is a combination of motivational, cognitive behavioral and relapse prevention counseling, introduction to the 12-steps and encouragement to participate in meetings, educational lectures, and group and individual therapy that are delivered by medically supervised, certified counselors that have completed a 2-year course in addiction treatment that includes an emphasis on substance use disorders as health problems needing long-term treatment (11).

Treatment begins with 7–10 days of detoxification and assessment at Vogur Hospital and an introduction to the psychosocial treatment program. Patients are encouraged to transition from Vogur to a 4-week residential or intensive outpatient program (IOP) and about two thirds of the patients follow these recommendations; the rest leave after completing detoxification. Both the residential and IOP settings deliver at least 60 hours of treatment of similar content over 4 weeks; the main difference is the setting. Assignment is based on patient preference and professional assessments that typically encourage the most intensive treatment the patient will accept and try to match patient needs to treatment setting in a collaborative process. Upon completing the first and most intensive treatment phase, patients are stepped down to 1–2 times a week of individual and/or group therapy for 3–12 months. Those who relapse can be readmitted to Vogur for detoxification and re-engagement in the program.

Medication

XR-NTX is a combination of naltrexone-containing microspheres suspended in a diluent that is delivered in a 380 mg dose by monthly injection into the muscles of the buttock. Plasma concentrations of naltrexone and 6-beta naltrexol (its main metabolite) are detectable for at least 30 days, thus reducing adherence problems that commonly occur with the 50-mg/day tablets. Naltrexone has caused hepatocellular injury, but this problem was almost always associated with oral doses of 1400 to 2100 mg per week and was not seen in a large study of heavily drinking alcoholics (12). Alkermes provided XR-NTX and matching placebo free of charge; more detailed product information is available at Vivitrol.com.

Study Design

Enrollment began in July 2010 and ended in February 2013. Participants were randomized using an urn schedule stratified according to male/female and injecting/non-injecting and given their first dose of study medication before leaving Vogur hospital and going on to IOP, or before leaving residential on the way to outpatient treatment. Medication was stopped after two consecutive missed injections but patients were encouraged to continue with the usual psychosocial treatment and study assessments. No reimbursement was given for study participation but travel expense vouchers for follow-up appointments were available. The primary outcome was proportion of amphetamine negative urine tests during weeks 1–24 of outpatient treatment. Secondary outcomes reported here are receipt of study medication, retention, other drug use, craving, relapse, and adverse events.

Assessments

The baseline assessment consisted of a medical history and physical; the DSM-IV Checklist (13) to identify amphetamine and other substance use disorders; a urine drug screen (Cozart Multi Test; amphetamine); a methylphenidate test that was done by the department of pharmacology and toxicology at the University of Iceland; an alcohol breath test; urine pregnancy test; CBC, ALT, AST, and creatinine; hepatitis C and HIV; Time-Line Follow-Back (14) covering the past 30 days at baseline and then weekly; Fagerstrom for nicotine dependence (15); Addiction Severity Index (16); HIV Risk Assessment Battery (17, 18); Visual Analog Scale for amphetamine craving (19); Alcohol Use Disorders Identification Test (20); Beck Depression Inventory (21); EuroQol (22); and Treatment Services Review (23).

Study nurses did weekly checks of injection sites and took vital signs, asked about adverse events, and did weekly urine drug screens and alcohol breath tests. Research assistants administered the TLFB, VAS and TSR and repeated the ASI and Fagerstrom at weeks 12 and 24, and the RAB at week 24. The liver panel was repeated at weeks 3, 11, and 23; hepatitis C and HIV testing were repeated at week 24; and a urine test for methylphenidate was done at baseline and at weeks 12 and 24.

Power Calculation

Participants were asked to provide one urine sample each week for drug testing, thus the primary outcome had possible 24 time points. Using two-sided tests and the methods of Diggle (24), group sizes of 50 provided 80% power to detect a difference of 20 to 25% at a level of 0.05 assuming a 15% dropout rate. For within-subject correlations of about 0.6, there was power for a difference of 25%; for within-subject correlations of about 0.4 there was power for differences of about 20%.

Data Analysis

A Cox proportional hazards regression model tested for differences in time to dropout between groups. Intent to treat analysis was performed on urine test results that were actually collected and with missing considered positive. A negative binomial regression model compared the number of amphetamine-positive tests in the two groups and included an offset term to account for the log transformed total number of tests actually provided. To account for dropout, inverse probability weighting based on a logistic model predicting completion was used. The lack of variability in urinalysis results precluded us from examining the data in a longitudinal model (i.e. generalized estimating equations (GEE) model, mixed effects model). A GEE model examined the relationship between attendance at weekly group and individual counseling (a binary variable reflecting whether or not they attended any group or individual session) and urinalysis results. This model included terms for therapy attendance (i.e., the binary variable described above), time (i.e., study week), and the attendance by time interaction with treatment condition as a covariate, specified as an exchangeable covariance structure. All analyses were carried out using SAS 9.3.

RESULTS

Randomization and Patient Characteristics

Of 169 patients approached for participation, 69 (41%) were not eligible, mostly for not completing the baseline assessment or other inclusion criteria.

Average age was 31.6 years, 75% were male, and the mean days of amphetamine use in the month prior to admission was 20. Other baseline characteristics are seen in Table 1.

Table 1.

Baseline Characteristics

Characteristics	Naltrexone (n=51)	Placebo (n=49)	Total (N=100)
Age, mean (SD), years	31.5 (7.91)	31.6 (9.4)	31.6 (8.6)
Gender male/female	36 /15	39/10	75/25
Ethnicity Caucasian	51	49	100
Education
Not completed general education	7	6	13
General education (10 years)	28	31	59
Some college	13	10	23
College degree or more	3	2	5
Housing
Independent	25	19	44
With parents	19	25	44
Halfway house or no housing	7	5	12
Employment past month
None	36	33	69
Partial	6	5	11
Full time	9	11	20
Amphetamine use
Days use in past month
10	7	8	15
11–20	27	18	45
21 or more	17	23	40
Mean # days past month amphetamine use	18.65 (1.08)	21.20 (1.21)	19.93
Route of administration
Oral	2	3	5
Nasal	37	38	75
Smoking	1	1	2
Injecting	11	7	18
Amphetamine Craving (0–100)	40.90 (4.35)	46.9 (4.38)	43.9
Clinical variables:
# Prior admissions SAA	5.8 (0.87)	4.9 (0.63)	5.4
# Days alcohol used past month	10.9 (8.75)	9.7 (8.86)
# Days cannabis used past month	16,8 (1,83)	18,2(1,84)	17.5
#Days benzodiazepines used past month	6,8 (1,38)	9,3 (1,56)	8.0
#Days cocaine used past month	5,6 (1,17)	5,2 (1,20)	5.4
Treatment status at baseline
Detox to IOP	23 (45%)	16 (32%)	39
Detox to residential	28 (55%)	33 (67%)	61
HIV positive			0
HCV positive			9

No between group differences were significant. Though amphetamine dependence was the main drug problem, 75% met DSM-IV-TR criteria for alcohol dependence, 69% for cannabis dependence, 30% for sedative (mainly benzodiazepine) dependence, 26% cocaine dependence, and 15% for methylphenidate dependence. As with other baseline features, these disorders were equally distributed across the two groups.

Receipt of Study Medication

Survival analyses examined time to dropout from study medication. The log-rank test of survival distributions showed no difference in time to dropout between the groups (chi square [1] = 0.57, p=. 45). Figure 2 depicts the number of participants receiving study medication over the course of the study

Figure 2.

Number of Patients Receiving Study Medication

Urine Test Results

We obtained 1247 (52%) of the planned 2400 urines; these included 53% from the XR-NTX group and 47% from the placebo group; 92 participants provided at least one urine test.

Of the missing 1153 samples, 867 (75%) were from dropout and 286 (25%) due to intermittent missed appointments prior to drop-out. Of the 1247 tests obtained, 1194 were amphetamine negative and 53 positive. Of the 92 participants that provided at least one urine specimen, 69 (75%) provided no amphetamine positive urine samples, 11 (12%) provided one positive sample, 6 (7%) provided two positive samples, 3 (3%) provided three positive samples, and one provided 4, 8, and 9 (1% each) positive samples, respectively.. Across both groups, 4% of the tests were positive for amphetamine, 8% for benzodiazepine, 7% for marijuana, 1% for cocaine, and 1% for opioid; only one of 104 methylphenidate tests was positive. A negative binomial regression model compared the two groups on the number of non-imputed amphetamine positive urines actually provided and no -significant naltrexone effect relative to placebo was seen (β = .003; IRR = 1.00, 95% CI=0.38–2.68, p = 0.99).

Therapy Attendance and Urine Test Results

Results revealed no main effect of treatment assignment (p = .22), week (p = .18), group (p = .91), or therapy attendance by week (p = .42) on amphetamine positive tests with no evidence that patients who participated in the residential program had fewer amphetamine positive tests than patients that went directly from Vogur to IOP. Patients that provided at least half of their urine tests, regardless of XR-NTX or placebo assignment, had significantly more weeks in which they attended at least one therapy session than those who provided fewer than half of their tests (m = 10.76 vs. 3.31; t (92) = 5.91, p < .0001).

Craving and Relapse

Craving on a self-reported scale of zero (none) to 100 (maximum) dropped from an average of 44 at baseline to 14 at 24 weeks for the approximately 50% of patients on which 24-week data were available with no significant difference between groups. Amphetamine relapse was defined as 3 or more consecutive days of self-reported use. Twenty nine of the patients on which we had these follow-up data reported amphetamine relapse at some point during the 24 weeks; 14 from the XR-NTX group and 15 from the placebo group.

Adverse Events (AE’s)

A total of 382 adverse events (AE’s) were reported over the course of the study. The most frequent were injection site reactions − 67 in the XR-NTX group and 77 in the placebo group; headaches −12 in the XR-NTX group and 10 in the placebo group; and nausea − 13 in the XR-NTX group and 6 in the placebo group. Overall, 80 individuals reported at least one AE − 41 XR-NTX and 39 placebo; NS difference between groups (p = .92). Approximately 55% (n = 28) of XR-NTX patients reported injection site reactions compared to 51% (n = 25) of placebo patients (p = .70). Almost 18% (n = 9) of XR-NTX patients reported headaches compared with 12% (n = 6) of placebo patients (p = .45), and 20% (n = 10) of XR-NTX patients reported nausea compared to 10% (n = 5) of placebo patients (p = .19). A few patients in each study arm withdrew consent while in the study. No consistent or objective reason were identified; some did not want to be bothered and others appeared to have lost interest in participating

DISCUSSION

Adding XR-NTX to usual treatment did not improve outcomes, as seen by the similarity in urine test results and treatment retention between XR-NTX and placebo patients. Though the study was underpowered, the similarity in outcomes would have made it difficult to find a difference even with a much larger sample. Research staff made frequent phone calls and sent letters asking patients to return for follow-up assessments but with limited success since only about half of the assessments were completed. Payments for completing assessments are known to facilitate completion of assessments (25) but were not included in this study because there is no tradition of of this practice in Iceland due concerns about creating an unrealistic impression of patient interest in the interventions being tested.

Though these results did not confirm the findings from other studies, there were significant patient and treatment differences that make comparisons difficult. Regarding patients, the Jayaram-Lindstrom et al. studies and the Ray et al. study excluded persons that were dependent on non-amphetamine substances except nicotine, and the largest of the Jayaram-Lindstrom et al studies (9) excluded patients that could not maintain abstinence for two weeks prior to enrollment. The result was that 214 of 294 (73%) persons that were approached for the study were excluded − 97 for inability to maintain abstinence and 77 for dependence on other drugs. In this study, 69 of 169 (41%) patients were excluded, 27 for not completing detoxification, 28 for other reasons, and none for another substance use disorder except opioid dependence. The result was a high concentration of more severely and multiply dependent patients, as seen in the baseline data where amphetamine use averaged 20 (66%) days in the past 30 as compared to 26–28% days in the past 12 weeks in Jayaram-Lindstrom et al (9), with 3/4ths or more of the participants in this study dependent on one or more non-nicotine substances. The potential of such patient differences to obscure naltrexone effects was mentioned by Brensilver et al (4) in their discussion of the 2008 Jayaram-Lindstrom study where they opined that “it is possible that inclusion of…more severely amphetamine-dependent individuals may erode these (positive) outcomes”.

Regarding treatment differences, the combination of a 7–10 day inpatient detoxification and assessment followed by 30-days of residential or IOP, and then 1–2 group and/or individual sessions over the next 3–12 months with encouragement to participate in 12-Step meetings was much more intensive than the 1–2 cognitive behavioral and/or supportive sessions over 12 weeks in the studies summarized above and likely contributed to the low level of amphetamine and other drug use in the urine tests that were obtained. Further support for the idea that the intensive psychosocial treatment was a meaningful contributor to positive outcomes is reflected in the finding that patients who provided at least half of their urine tests had significantly more weeks in which they had at least one therapy session as compared to those who provided less than half of their urine tests. Brensilver et al (4) mentioned this possibility in their earlier review (4). Though methylphenidate has become the most common stimulant used in Iceland (26) only 15% of study patients were dependent on it at the time of enrollment and only one of the 104 methylphenidate urine tests that we obtained was positive for it, thus suggesting that methylphenidate use was minimal and did not obscure a naltrexone effect.

Tihonen et al found a significant reduction in amphetamine and heroin use in multiply dependent patients that were treated with a naltrexone implant in Finland (#?). Though suggesting a naltrexone effect, an equally, or more likely explanation, is that the reduction was not specific to an amphetamine but from less heroin use opioid blockade since amphetamines are often used with opioids as a “speedball” to magnify their effects.

The most significant weaknesses was the low followup rate. Staff made extensive efforts to contact patients but did not have the option of using the contingency of patient payments, as mentioned above. Another weakness was that methamphetamine may be less responsive to naltrexone than amphetamine and we did not test for methamphetamine, thus have no way of knowing if methamphetamine use occurred and obscured an effect on amphetamine. Similarly, we tested for methylphenidate only 100 times because it was very expensive and could have missed a switch from amphetamine to methyphenidate use among some patients.

Strengths include XR-NTX as a way to document medication adherence; doing the study in a Nordic population that was likely similar to the Jayaram-Lindstrom et al patients; using evidence based psychosocial treatments and a chronic disease model; and the Icelandic social environment where substance use disorders are considered treatable health problems, treatment is considered helpful, and is available without restriction at little out of pocket cost. These patient, treatment and contextual factors likely made it difficult to detect a naltrexone effect unless it is robust, which does not appear to be the case based on the significant but relatively modest findings of the studies summarized above.

Figure 1.

Screening, randomization and follow-up of participants

Figure 3.

Number of urine specimens at each assessment point

Figure 4.

Observed and imputed amphetamine negative urinalysis results by treatment arm