Special Issue

Longitudinal Effects of LAAM and Methadone Maintenance on Heroin Addict Behavior M. Douglas Anglin, PhD Bradley T. Conner, PhD Jeffrey J. Annon, MA Douglas Longshore, PhD Abstract Levo-alpha-acetylmethadol maintenance (LAAM) was compared to methadone maintenance (MM) on the behavioral performance of 315 heroin addicts before, during, and after 12 months of fully subsidized treatment. Assessments of drug use, criminal behavior, HIV risk behaviors, and employment and residential status were obtained at treatment intake and at 6, 12, and 18 months after admission. Treatment retention and in-treatment suppression of heroin use were significantly better for the LAAM group than for the MM group. Improvements were also noted during treatment in criminal behavior, criminal justice involvement, and employment status, and there were reductions in injection HIV risk and number of sexual partners. Most significant effects were primarily related to active participation in maintenance treatment. Under subsidized treatment, retention rates were two to four times that of similar clients in local community programs during the same period. LAAM was a useful and a potentially important addition to treatment options for opiate addiction, conferring greater retention and opiate suppression benefits. Its removal from application provides a historical lesson concerning the introduction of new medications into addiction health services.

Address correspondence to M. Douglas Anglin, PhD, Integrated Substance Abuse Programs, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, UCLA, 1640 S. Sepulveda Blvd., Ste. 200, Los Angeles, CA 90025, USA. Phone: +1-310-6179922; Fax: +1-310-4737885; E-mail: [email protected]. Bradley T. Conner, PhD, Integrated Substance Abuse Programs, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA. Phone: +1-310-2675379; Fax: +1-310-4737885; E-mail: [email protected] Jeffrey J. Annon, MA, Integrated Substance Abuse Programs, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA. Phone: +1-310-2675333; Fax: +1-310-4737885; E-mail: [email protected] Douglas Longshore, PhD, Integrated Substance Abuse Programs, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA. Douglas Longshore died December 2005. Journal of Behavioral Health Services & Research, 2008 Healthcare.

)c 2008 National Council for Community Behavioral

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

267

From its introduction into clinical practice in the mid-1960s, narcotic replacement therapy (NRT) was primarily effected through the administration of the opiate agonist methadone, which became one of the most frequently evaluated modalities for the treatment of heroin addiction.1,2 In 1993, levo-alpha-acetylmethadol (LAAM), a longer acting agonist medication, also was approved for NRT following a large phase III safety trial.3,4 To date, most studies of the differential effectiveness of LAAM in comparison to methadone have focused on treatment retention and the suppression of opiate use under active medication. Along this line, the authors have published two prior articles examining differential in-treatment performance under the two medications. In the first,5 the authors briefly reprised the history of LAAM approval, clinical use, and subsequent discontinuation from manufacture, provided a literature review of LAAM treatment outcome studies, and reported 6-month outcomes of a randomized clinical comparison of LAAM and methadone maintenance (MM). There were equivalent retention rates for the LAAM and MM groups (about 77%) at 6 months. During treatment, LAAM was a significantly better suppressor of heroin use than methadone, as indicated by weekly urinalysis results that were stably positive for heroin metabolites at about 40% for the LAAM group versus 60% for the MM group (pG0.05). The second article6 further elaborated on the literature and reported additional follow-up outcomes. At 12 months, retention was modestly but significantly better for the LAAM group (57.4%) versus 46.2% for the MM group (pG0.05), primarily due to fewer terminations as a result of arrest and incarceration. The rates of heroin use during treatment continued at the 6-month levels reported above over the full 12-month period (40% LAAM, 60% MM), with the exception of increases for both groups during the last month of subsidized treatment. This increase may have been reflective of increased client anxiety about the stability of treatment as full subsidization was ending. Overall, evidence was found for superior effectiveness of LAAM in reducing opiate use during treatment, both as measured by urinalysis of a specimen collected weekly and at two follow-up points by research staff. Further analysis emphasized that the differential effects of LAAM and MM were contingent on remaining in treatment; among participants no longer in treatment at 12 months, relapse to opiate use, as determined by the research urinalysis, occurred at high rates among the LAAM (88.9%) and MM groups (75.0%; difference not significant). The results of these two studies, along with prior research7 in which LAAM outperformed MM, particularly for methadone “non-holders,” indicate LAAM to be at least as effective, and likely more effective, than MM at reducing opiate use in community-based clinical settings. Expanding beyond retention and opiate suppression, this article examines the differential effects of the two medications and of active NRT participation on other behaviors important in assessing treatment effectiveness. As noted and criticized by some,8 drug treatment outcomes, especially those regarding NRT, have primarily focused on social objectives, with little focus on patient quality of life. This omission is regrettable, but the sociopolitical forces that led to the development of a nationwide community-based drug treatment system in the early 1970s, especially in the provision of maintenance medication, promoted the public funding of treatment as a policy reducing the social ills associated with drug dependence, particularly crime.9 Once established, these criteria have continued to be the primary outcome measures of evaluation studies.10 Within this emphasis, considerable published literature has established that maintenance treatment reduces criminal behavior,11–13 promotes employment and improved family functioning,14 and reduces HIV and other disease risk related to injection drug use15–17 and unprotected sexual activities.18,19 These same criteria have been infrequently applied to the assessment of the effects of LAAM and are the focus, in part, of the present article. Although accessible throughout the study period, LAAM is no longer available as a maintenance medication since the Federal Drug Administration required a “black box” warning of potentially fatal heart arrhythmias to the package insert in 2001.4 Subsequently, in 2003, the sole manufacturer of LAAM withdrew the medication from the world market.20 It is important, however, to maintain full documentation of both medications’ differential effectiveness on such disparate outcome

268

The Journal of Behavioral Health Services & Research

36:2

April 2009

measures. The historical summary provided in two earlier articles by the authors and updated herein yield important lessons regarding the introduction of new medications into addiction health services. Treatment status, subsequent return to maintenance treatment, and opiate and other illicit drug use at a follow-up interview occurring 18 months subsequent to treatment intake (approximately 6 months after subsidized treatment ended) are reported. Over the study period, changes in criminal behavior; HIV risk behaviors including injection, needle sharing and hygiene practices, and sexual risk behaviors; improvements in employment rates; and residential stability were examined. The primary comparison was between the randomly assigned maintenance medication groups, but whether patients were actively participating in treatment at each follow-up point was also assessed since medication effects on opiate use and other measures may be most apparent among those who remain in treatment. Accordingly, interaction of current treatment status (either LAAM or MM treatment versus no treatment) and original medication group (LAAM or MM) were examined to determine the degree to which effects were contingent on concurrent maintenance treatment. It should be noted that at the end of 12 months of fully subsidized maintenance, support for services was terminated as planned, and patients had to secure other funding or pay for services themselves. As is typically the case when secure sources of funding end, treatment continuity may have suffered over the subsequent follow-up period of 6 months and relapse may have been higher, particularly among those who could not sustain treatment through fee-for-service.21 This context should be kept in mind when interpreting the results.

Method Study design NRT-seeking heroin addicts were randomly assigned at a 2:1 ratio to either LAAM or MM for a 12-month period of subsidized care including standard clinic services. Among the 315 patients with complete intake data, 209 were assigned to treatment with LAAM and 106 to MM. Both groups began with a 2-week induction period to methadone. Patients were considered to have stabilized on methadone when dose levels were both high enough to suppress craving for heroin, withdrawal symptoms, or drug-seeking behavior and low enough to avoid sedation from the medication. After 2 weeks, LAAM was substituted for methadone as per assignment. Dose could be adjusted at any time following stabilization according to clinical judgment. Dosing did in fact vary within patients over time and among patients: 77% of LAAM patients and 85% of MM patients had dose adjustments after stabilization. Average maximum dose for the LAAM group was 77.5 mg, and variability was wide (SD=17.4). Average maximum dose for the MM group was 67.4 mg, and variability was again relatively wide (SD=17.3). Participants Patient recruitment occurred between February 1997 and January 1999 as described by Longshore et al.5 and further elaborated by Anglin et al.6 Over the study period, UCLA’s Institutional Review Board maintained oversight to ensure the ethical treatment of participants. Average age was 46 years (SD=7.4; range 22–66 years), and 60% of the sample had completed high school. Only the racial/ethnic breakdown was significantly different. The LAAM group comprised 92 (44%) African American, 79 (38%) Hispanic (mostly Mexican Americans), 34 (16%) non-Hispanic white, and four (2%) individuals of “other” race/ethnicity. The comparative percentages for the MM group were 37 (35%), 39 (37%), 24 (22%), and 6 (5%; pG0.05). Only 25% of patients were employed at intake. Illegal income in the 6 months prior to the study was reported by 32% of the participants. Criminal justice involvement was common: 95% had been arrested at least once and 77% had been incarcerated at least once.

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

269

Participants had extensive drug use histories. Almost all (95%) had used heroin by injection, 74% had used cocaine combined with heroin by injection, 75% had used crack cocaine, and a majority had used non-prescription opiates. Almost all (97%) had been in opiate drug treatment previously, nearly all of which was MM. At intake, 60% reported using heroin daily for the past 30 days; the mean period of addiction to heroin was 21 years. When comparing the study sample to all those in Los Angeles County entering MM for the same period on a second or later admission, no significant differences were found on the above noted variables (analyses conducted by the authors using data from the Los Angeles County Participant Reporting System).22 Overall, the sample consisted of typical long-term heroin addicts for whom prior interventions had been unsuccessful in maintaining either continuing treatment participation or abstinence from opiate use. Approximately 6 months after subsidized treatment ended, participants were contacted to complete an 18-month post-admission follow-up interview. Some 24 participants (17 who received LAAM and seven MM) could not be located, yielding an interviewed sample of 285 (90.5%): 188 from LAAM and 97 from MM. This sample did not differ significantly from the originally randomized sample on any descriptive measures. Additionally, those not located did not significantly differ from those interviewed on any demographic variables. Table 1 provides information on the follow-up status of participants in the two study conditions (i.e., LAAM and MM).

Table 1 Results summary of 18-month status Follow-up status

LAAM

MM

Intent to treat, N 209 106 % Not located 8.1 6.6 % Deceased 1.9 1.9 % In jail or prison 7.7 8.5 Interviewed, N 188 (89.9%) 97 (91.5%) % in LAAM 6.4 0 % in MM 30.3 42.3 History of maintenance treatment % discharged by 6 months without return 15.1 15.1 % discharged by 6 months with return 9.6 12.2 % discharged by 12 months without return 20.4 18.0 % discharged by 12 months with return 6.5 8.7 % discharged after 12 months without returna 18.3 24.4 % continuously in treatment at 18 months 30.1 21.5 Total % in treatment at 18 monthsb 36.7 42.3 Research Urine, N 135 (71.8%) 76 (78.4%) Percent missing/refused research Urine sample Not in maintenance treatment 42.2 38.6 χ2 =15.4, pG0.001c In maintenance treatment 16.3 9.4 χ2 =11.7, pG0.001d a

Fully subsidized treatment ended at 12 months Total includes some from above groups showing a return to treatment by 18 months c Comparing those in the LAAM group who were in treatment versus those who were not in treatment d Comparing those in the MM group who were in treatment versus those who were not in treatment b

270

The Journal of Behavioral Health Services & Research

36:2

April 2009

Data collection Intake interviews were conducted on all 315 participants. Follow-up interviews were conducted for all participants who could be located at each time point, including those incarcerated. Intake interview All patients reported their personal background and drug use and drug treatment histories; criminal behavior and criminal justice involvement; social and family relationships; HIV risk behavior and HIV-related perceptions; education and employment history; motivation for treatment; and mental health status. Additionally, all participants voluntarily provided a urine specimen within the first week after randomization. This specimen was used to detect recent illicit drug use. Follow-up interviews At approximately 6, 12, and 18 months after intake, staff re-interviewed participants for 90 to 120 min to update information on drug use, other status measures, treatment services received, and utilization of health and mental health services outside the clinic. When possible, a urine specimen was voluntarily obtained at each interview to assess recent drug use. For a complete description of intake and 6-month interview data, see Longshore et al.5 Anglin et al.6 provide a complete description of the 12-month follow-up data. Here, the focus is primarily on data from the 18-month interview and the overall pattern of selected behaviors over the study period. The 18-month interview was conducted, on average, 19 days (SD=22) prior to the 18-month anniversary of the intake date.

Measures Measures that were repeated across the four interview time points (intake, 6, 12, and 18 months) were used in the analyses reported here. Some measures indicated status at the interview point, while others asked about events and behaviors in the prior 6 months. Treatment status Participation in LAAM, MM, or other treatments was assessed. If a participant had been discharged from study treatment, subsequent participation in maintenance or other treatment was ascertained by self-report at each interview and by the presence of LAAM or methadone in the research urine. In a few cases, however, the specimen did not contain metabolites for either methadone or LAAM, and these participants were considered not to be in maintenance treatment at the time of the follow-up interview. Treatment records from other maintenance programs were not able to be obtained. Drug use Detection of recent opiate, codeine, and cocaine use was accomplished using two methods: (1) analysis of the 18-month follow-up interview urine specimen and (2) analysis of selfreported use during the 30 days prior to the interview. Urine samples were sent to a National Institute on Drug Abuse (NIDA)-certified testing laboratory, PharmChem, Inc. Participants also completed a self-report questionnaire on which they indicated the number of days they used each type of drug during the 4 weeks prior to each follow-up interview. Criminal behavior Participants were asked to indicate if they were incarcerated or had been incarcerated, the number of times they had committed a crime, the type of crime, and if they had been arrested for committing the crime. At intake, the information was requested for the participant’s lifetime, whereas for later interviews, participants were asked to report incidents since the last interview. Employment history and income At each interview, participants were asked if they were currently employed. They were also asked to indicate participation in and the amount of money earned from

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

271

both legal and illegal sources in the last 6 months, at intake, and since the last interview at the two follow-up interviews. HIV risk behaviors Participants completed a self-report questionnaire assessing their participation in HIV risk behaviors, including items that asked about the number of times they had injected or shared needles or works, their use of needle hygiene methods, and the number of sexual partners, and the number of times they engaged in certain types of high-risk sexual encounters. Questions were asked for the last 6 months, at intake, and since the last interview at the two follow-up interviews. Residential stability Participants were asked where they lived (e.g., in a home they owned or rented, in an apartment they owned or rented, with family, with friends, in temporary/unstable housing, or if they were homeless). Questions were asked for the last 6 months, at intake, and since the last interview at the two follow-up interviews.

Statistical analyses Primary analyses were performed on an intent-to-treat basis, including all participants with intake and follow-up data required for the analyses, regardless of whether they entered or subsequently left the assigned treatment, how much of that treatment they received, how much of any other treatment they received, or whether they were still in treatment at the time of the followup interview. Analyses were conducted using standard procedures available in SPSS for Windows (release 14.0),23 except where noted. As previously reported,5,6 there were significant differences in the racial/ethnic composition between the MM and LAAM groups; as a result, race/ethnicity was controlled in all analyses. If there was an effect based on the different racial/ethnic compositions of the groups, that effect is reported. Some items (particularly criminal history items) asked about cumulative lifetime behavior at the intake interview, but only asked about frequencies of events since the last interview at the follow-up interviews. In these cases, intake data were used as a covariate to control for baseline differences. The detection of drug used followed the common NRT policy that urine samples were considered positive for a drug when drug metabolites were present in the sample or if the participant was present at the clinic or interview but refused to submit a sample. This type of classification is typical within MM programs and is the most conservative approach for analytic purposes. However, further analyses (not reported) reclassifying refusals as missing did not affect results. If the sample was not collected for some other reason, then data were considered missing. Only heroin, codeine, and cocaine use were observed at levels to warrant reporting. Multilayer chi-square analyses were used to determine if drug use as measured by urinalysis was contingent on medication group, concurrent maintenance treatment participation, and race/ ethnicity. Analyses of variance (ANOVAs) and multilayer chi-square analyses were used to determine if there were significant differences in self-reported criminal behavior, employment status, income sources, and risk of exposure to HIV through needle use/cleaning and sexual behaviors between medication group, concurrent maintenance treatment participation, and race/ ethnicity.

Results First, status results for the participants at 18 months after intake (Table 1, with further results in Tables 2 and 3) are presented. Next, behavior patterns are reported for the full study period, first

272

The Journal of Behavioral Health Services & Research

36:2

April 2009

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

273

88.7 – – 21.2 (11.7) – – 35.8 – –

90.9 – – 22.9 (10.8) – –

44.7 – –

MM

17.3 25.7 15.3

42.5 77.1 34.0c 4.1 (8.1) 10.5 (12.3) 2.4 (5.4)f

LAAM

20.9 37.5 17.3i

61.5b 75.0 58.7d 5.5 (9.0) 10.6 (13.2) 4.5 (7.3)

MM

6 months

24.8 26.3 24.2

58.6 70.2 52.5e 6.8 (10.5) 10.9 (13.4) 3.5 (6.4)g

b

27.6 43.5 20.8j

65.8 69.6 64.2 7.8 (10.7) 8.5 (12.0) 7.4 (10.1)h

MM

12 monthsa LAAM

End of fully subsidized maintenance treatment Significantly different between the medication groups, χ2 =8.79, pG0.01 c Significantly different within LAAM group between maintenance treatment status, χ2 =21.4, pG0.001 d Significantly different within those in treatment between medication groups, χ2 =12.3, pG0.001 e Significantly different within LAAM between maintenance treatment status, χ2 =4.7, pG0.05 f Significant main effect for maintenance treatment status, F(1, 270)=29.8, pG0.001 g Significant interaction between medication group and maintenance treatment status, F (1, 238)=4.6, pG0.05 h Significant main effect for maintenance treatment, F(1, 238)=8.5, pG0.01 i Significantly different within MM between maintenance treatment status, χ2 =3.3, p=0.07 j Significantly different within MM between maintenance treatment status, χ2 =4.1, pG0.05

a

Heroin use % Research urine-positive % Not in maintenance % In maintenance Days used past month Not in maintenance In maintenance Codeine use % Research urine-positive % Not in maintenance % In maintenance

LAAM

Intake

Table 2 Cross-sectional drug use results

23.0 22.7 23.2

63.7 68.2 59.4 9.3 (11.8) 11.3 (13.1) 7.5 (10.2)

LAAM

21.1 14.3 26.8

68.4 68.6 68.3 9.1 (11.4) 9.6 (12.8) 8.6 (10.2)

MM

18 months

274

The Journal of Behavioral Health Services & Research

36:2

April 2009

Criminal justice involvementa % Arrested Not in Maintenance In Maintenance % Convicted Not in Maintenance In Maintenance % Incarcerated Not in Maintenance In Maintenance % Reporting any illegal income Not in maintenance In maintenance % Reporting any dealing income Not in maintenance In maintenance % Reporting any property crime income Not in maintenance In maintenance HIV risk % Reporting any needle hygiene risk behaviort Not in maintenance In maintenance Sexual partnersv Not in maintenance In maintenance Employment % Employed Not in maintenance In maintenance

Table 3

95.7 – – 87.7 – – 82.1 – – 41.5k – – 20.8o – – 25.5 – – 29.5 – – 17.9 – – 19.8 – –

25.8 – – 15.3 – – 22.5 – –

MM

98.1 – – 84.2 – – 78.5 – – 31.7 – – 12.4 – – 18.2 – –

LAAM

Intake

32.1 28.9 34.7

8.7 10.3 8.3 10.4 20.5 7.6w

13.2 21.1 11.1 6.0 23.7 1.4d 12.1 21.1 9.7h 19.9 34.2 16.0l 8.2 17.9 5.6p 10.9 15.4 9.7

LAAM

MM

6.1 10.6 3.0u 12.7 16.7 10.1 31.1 24.6 42.9z

22.6x 27.8 21.3y

13.5 21.5 8.2b 5.1 10.8 3.1f 9.8 15.4 6.1i 17.2 20.0 15.3 8.5 10.6 7.1 8.5 10.6 7.1

LAAM

26.6 26.9 26.4aa

12.7 19.2 9.4 10.1 11.5 9.4

19.0 26.9 15.1 5.1 3.8 5.7 16.5 23.1 13.2 22.8 34.6 17.0n 11.4 19.2 7.5 11.4 19.2 7.5

MM

12-Montha

12.9 22.2 10.7 10.8 11.1 10.7

20.4 27.8 18.7 11.8 16.7 10.7e 16.1 27.8 13.3 29.0 27.8 29.3m 9.7 16.7 8.0 23.7 27.8 22.7s

6-Months

Cross-sectional behavioral results

30.3 35.4 27.5

8.9 12.1 5.8 8.9 12.1 5.8

11.2 16.9 5.8c 7.5 10.8 4.3g 10.4 13.8 7.2j 22.4 24.6 20.3 6.7 9.1 4.3 14.8 19.7 10.1

LAAM

32.0 34.3 30.0

9.2 5.7 12.2 13.2 11.4 14.6

13.3 17.1 10.0 5.3 5.7 5.0 17.3 25.7 10.0 18.7 14.3 22.5 11.8 8.6 14.6q 7.9 5.7r 9.8

MM

18-Months

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

275

8.6 – – 54.5 – –

5.7 – – 58.5 – –

9.1 7.9 11.1 59.7 55.3 62.5

4.3bb 5.6 4.0 65.0 66.7 64.0

b

Intake information is lifetime; follow-up interviews are since last interview Significantly different within LAAM between maintenance treatment status, χ2 =6.0, pG0.05 c Significantly different within LAAM between maintenance treatment status, χ2 =4.2, pG0.05 d Significantly different within LAAM between maintenance treatment status, χ2 =26.3, pG0.001 e Significantly different within those in treatment between medication groups, χ2 =9.7, pG0.01 f Significantly different within LAAM between maintenance treatment status, χ2 =4.0, pG0.05 g Significantly different within LAAM between maintenance treatment status, χ2 =4.2, pG0.05 h Significantly different within LAAM between maintenance treatment status, χ2 =3.6, pG0.05 i Significantly different within LAAM between maintenance treatment status, χ2 =3.8, p=0.06 j Significantly different within LAAM between maintenance treatment status, χ2 =3.2, p=0.07 k Significantly different between medication groups, χ2 =3.0, p=0.09 l Significantly different within LAAM between maintenance treatment status, χ2 =6.3, pG0.05 m Significantly different for those in treatment between medication group, χ2 =5.4, pG0.05 n Significantly different within MM between maintenance treatment status, χ2 =3.1, p=0.08 o Significantly different between medication groups, χ2 =3.8, p=0.05 p Significant main effect within LAAM between maintenance treatment status, χ2 =6.3, pG0.05 q Significantly different within maintenance treatment status between medication groups, χ2 =3.6, p=0.06 r Significantly different for those not in treatment between medication groups, χ2 =3.5, p=0.06 s Significantly different for those in treatment between medication groups, χ2 =6.8, pG0.01 t “How many times in the last 6 months have you used a needle after someone else used it?” u Significantly different within LAAM between maintenance treatment status, χ2 =4.0, pG0.05 v Percentage reporting two or more partners in the last 6 months w Significantly different within LAAM between those between maintenance treatment status, χ2 =5.53, p=0.06 x Significantly different for those in treatment between medication groups, χ2 =4.1, pG0.05 y Significantly different between maintenance treatment status, χ2 =4.2, pG0.05 z Significantly different within LAAM between maintenance treatment status, χ2 =5.7, pG0.05 aa Significantly different for those in treatment between medication groups, χ2 =4.0, pG0.05 bb Significantly different between medication groups, χ2 =21.9, pG0.001 cc Significantly different within MM between maintenance treatment status, χ2 =6.5, pG0.05

a

Residential status % Own residence Not in maintenance In maintenance Rent residence/room Not in maintenance In maintenance 8.0 10.8 6.1 63.2 61.5 66.3

6.3 7.7 5.7 56.6 42.3 67.9

9.8 9.2 10.1 56.2 58.5 55.1

5.1 5.7 0.0 56.6 37.1 67.5cc

within groupings by the original medication assignment and then by actual participation in maintenance treatment at the time of interview (Tables 2 and 3). Some results that were not statistically significant are presented in the results text to allow for comparison to other literature findings (e.g., levels of cocaine use). Outcomes at 18 months A summary of 18-month status at interview is presented in Table 1; the 18-month time point was 6 months after subsidization of treatment had ended, a contextual change that should be kept in mind in interpreting 18-month results. Other 18-month status results (events in past 6 months) are presented in Tables 2 and 3, which also present past 6 months behavioral results for each of the other interview points. Interview rates at 18 months were high (90.0% for LAAM and 91.5% for MM). By the 12-month interview, one in each medication group had died (one because of cancer and the other from an overdose). At 18 months, an additional three LAAM participants and one MM participant had died (two overdoses, one coronary, one undetermined). Cumulative deaths subsequent to intake were four in the LAAM group and two in MM, which is in accordance with the original 2:1 assignment ratio. Incarceration When possible, participants who were incarcerated at the time of follow-up were interviewed at the correctional facility. Some 7.7% of LAAM and 8.5% of MM participants were incarcerated at the time of interview (ns). Treatment status Of the LAAM group, 36.7% were in maintenance treatment (12 were receiving LAAM; the remainder were receiving methadone) as were 42.3% of the MM group (all receiving MM; ns). Some 30.1% of LAAM and 21.5% of MM participants had been continuously in maintenance treatment over the study period (ns). Of those discharged from treatment at any point, 16.1% of LAAM and 20.9% of MM participants had reentered treatment for some period prior to the 18-month interview (ns). Other more disparate patterns of maintenance treatment history over the study period are presented in Table 1. Research urinalysis results At follow-up, non-incarcerated participants were requested to provide a voluntary urine sample to assess recent drug use. Compliance rates were not significantly different: 71.8% for LAAM and 78.4% for MM. A cross-tabulation of urinalysis results for opiates compared to self-reported opiate use indicated relatively good validity. Specifically, of those participants with negative urinalysis data, 77.8% reported no heroin use; 22.2% reported using heroin in the last 4 weeks, but this use was likely not detected in the relatively short timeframe that urinalysis captures. Of those with positive urinalysis data, only 10.3% self-reported no heroin use in the previous 4 weeks. However, for both medication groups, those not in treatment were significantly more likely to have refused a urine sample. For this group, 42.2% of LAAM participants and 38.6% of MM participants were in these categories; for those in maintenance treatment, the comparative figures were 16.3% and 9.4% (pG0.001). Accordingly, both urine research data and self-reported heroin use for those not in maintenance treatment must be considered as underreported. Opiate use Results from a multilayer chi-square analysis (medication group, treatment status at 18 months, opiate urinalysis result, controlled for race/ethnicity) revealed no significant difference in the percentage of participants with positive urinalysis results (see Table 2). Each group used at about the same rate (59–68%). Codeine use was seen at follow-up at about the same rates observed throughout the study period (20–25%). Results of a 2×2×3 (medication group × treatment status × race/ethnicity) between-subjects ANOVA revealed no significant differences in the mean number of days of self-reported heroin use

276

The Journal of Behavioral Health Services & Research

36:2

April 2009

in the 4 weeks prior to the 18-month follow-up interview (Table 2). Participants in LAAM reported 9.3 (SD=11.2) mean days of heroin use versus 9.1 (SD=11.3) for those in MM. Combining medication groups, where no between-group significant differences were observed, those in maintenance treatment reported 7.9 (SD=10.2) days of heroin use versus 10.7 (SD=12.9) means days for those not in treatment. There were no significant differences in days of reported codeine use (2 to 5 days). Cocaine use Results from a multilayer chi-square analysis (medication group, treatment status, urinalysis result) revealed no significant differences between cocaine urinalysis results by medication group (about 22%) even when further grouped by concurrent treatment status. Results of a 2×2×3 (medication group × treatment status at 18 months × race/ethnicity) between-subjects ANOVA revealed no significant differences in the self-reported percentage of those using in the prior 30 days or the mean number of days of cocaine use in this period. Participants in LAAM reported a use rate of 32.8% and 3.0 (SD=7.5) mean days of cocaine use versus 25.3% and 2.4 mean days of use (SD=6.8) for those in MM. Combining treatment status groups, those in maintenance treatment had rates of use between 30.0% and 34.8% and reported 3.2 (SD=7.8) days of cocaine use. Those not in treatment had rates between 20.0% and 30.8% and reported 2.4 (SD=6.7) days of use.

Behavior over time contingent on medication group Tables 2 and 3 show the drug use and behavioral results by medication group and by maintenance treatment status across the study period: intake, 6, 12, and 18 months. Significant results are noted with footnotes reporting the statistical test results. Two patterns are revealed by this array of results: differential effects over time between the medication groups and the “overlay” of participation in maintenance treatment at each of the follow-up points. Probability levels up to 0.10 are noted to ensure that patterns of behavior that emerge are not overlooked as candidates for further research. This criterion increases the likelihood of type 1 error, especially in light of the number of statistical tests conducted. However, precise probability levels are provided in both text and tables, allowing an independent determination of the strength of individual findings. Results without significant differences are reported in the text to give the base levels of the reported behaviors for this sample. Interview rates were consistently high at all follow-up points, ranging from 90% to 94%. As would be expected, about 90% of each group had heroin-positive urine samples at intake, but rates dropped significantly on entry into either maintenance treatment. As previously noted,5,6 participants in LAAM had a lower frequency of heroin drug use during treatment than did participants in MM. However, at the 12-month interview, group differences between research urine samples attenuated and remained stable at the 18-month interview. Codeine use was also relatively high at intake (36% to 45%), dropping by about one half after treatment admission and remaining relatively stable thereafter. Cocaine use as represented by urine tests fluctuated modestly between 19% and 24% throughout the study period. The rate of self-reported use was about twice that detected by urinalysis and did not decline from intake levels (about 37%), ranging from 25% to 45% thereafter; no significant differences were observed between medication groups. Only a few behavioral differences were observed in other domains by medication group (see Table 3). All participants had high lifetime rates of arrests, convictions, and incarcerations prior to study intake. Although there was a consistent pattern of LAAM participants having lower involvement with the criminal justice system at each follow-up point, these differences were not significant. Percentages of those with any self-reported illegal income ranged from 32% to 42% at intake. Thereafter, these rates ranged from 17% to 23% for LAAM participants and 17% to

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

277

29% for MM participants. Examining reports of any illegal income from dealing and property crime showed reductions from intake levels, but no group differences. In terms of HIV risk, other than the reduction of injection drug risk during treatment, no group differences were observed on injection hygiene practices, number of sexual partners or sexual risk behaviors (results fluctuated between 50% and 60% reporting at least one sexual risk behavior and between 24% and 42% for two or more risk behaviors). There was, however, a significant (pG0.05) racial/ethnic difference in sexual risk behaviors, with a greater proportion of Hispanic/Latinos reporting one or more risks than other groups (LAAM: non-Hispanic white, 58.8; Hispanic, 63.2, African American, 59.3; MM: non-Hispanic white, 53.8; Hispanic, 79.4, African American, 66.7). Employment rates were significantly higher (pG0.05) for LAAM participants in the first 6 months of treatment, a difference that attenuated at 12 and 18 months. While low at under 10%, LAAM participants reported living in their own homes significantly more frequently at 6 months, a trend that continued thereafter, although not at a significant level. A greater variability in this variable was seen for MM participants. Most participants, regardless of medication group, were living in rented spaces (about 60%), with family or friends (about 20%), or in group home environments (about 10%). Less than 5% reported being homeless at each time point. Behavior over time contingent on active treatment status More dispersed, but more frequently significant, results emerged from the analyses of active treatment status at each follow-up point. The earlier papers from this study emphasized that improved performance was primarily contingent on actively receiving maintenance treatment, as these results also support. Urinalysis results and self-reported recent (past week) participation in LAAM and MM treatment were combined to determine treatment status. However, in the few cases where self-report was not confirmed by the presence of LAAM or methadone in the urinalysis report, the participant was considered not to be in active treatment. At each follow-up point, an increased percentage of participants were not in treatment: about 25% at 6 months, nearly 50% at 12 months, and 60% to 65% by 18 months. Those in treatment at 6 months had significantly lower rates of heroin-positive urine samples than those not in treatment, a pattern that continued at 12 months (significantly for LAAM participants, not significantly for MM participants) but attenuated by 18 months. To a lesser degree, this pattern was also observed for codeine-positive urine samples and significantly so at 6 and 12 months for MM. No relationships were observed in the pattern of cocaine-positive urine samples, which varied around 20% at all points. The percentage of participants who reported at least 1 day of cocaine use in the past 30 days varied around 35% for the LAAM group and 30% for the MM group. In terms of criminal justice involvement and illegal behavior, those in treatment at each followup point were less likely to have been arrested, convicted, or incarcerated in the prior 6 months, significantly so for those LAAM participants but not for MM participants. Moreover, except for the MM participants at 18 months, the percent reporting any illegal income in the prior 6 months was consistently lower for those in treatment, significantly so at some time points. When disaggregated into components, the percent (these data not shown for brevity of presentation) reporting dealing and property crime income followed a similar pattern. Differences related to active treatment status were observed in HIV risk behaviors for frequency of use by injection and in needle hygiene. Numbers of sexual partners were consistently lower during in-treatment status. Employment rates for LAAM participants were consistently higher during the study period, but not at the 18-month follow-up. Similarly, employment during intreatment status for those in LAAM was consistently higher except at 18 months. While there was an overall pattern of higher rates of residence in group homes or in unstable housing when participants were not in treatment (about 22%) than when they were (about 13%), these differences were not significant.

278

The Journal of Behavioral Health Services & Research

36:2

April 2009

Discussion Commensurate with literature reports, the results presented herein found LAAM outperforming MM in terms of treatment retention and greater suppression of during-treatment heroin and codeine use. Unlike most previous studies comparing the two medications, the present study examined a more diverse set of behaviors and contained a 6-month post-treatment follow-up that was conducted to assess the stability of behavioral change over time. It should be noted that the discontinuation of LAAM availability nationally occurred after the last follow-up of participants and thus had no external effect on study data or findings. The planned termination of subsidy at the 12-month point disrupted treatment continuity for some clients and may have affected their 18month interview data. Moreover, recall bias is inherent in all self-report data, but should have been equivalent over analytic subsamples, rendering relative comparisons valid even if actual behavioral reports may have some recall variation. About one third of the participants were in NRT at 18 months, a reduction of the rate at 12 months. The majority of participants were in MM since few other Los Angeles programs other than the study site offered LAAM. Apparently, staff and participant efforts at maintaining or reentering maintenance treatment were relatively successful, at least in the short term. The majority of those in treatment at 18 months (55.0%) had also been in treatment at 12 months, suggesting that the year of subsidized treatment may have prepared many of them for obtaining maintenance services on their own. Importantly, of those successful at maintaining at least 4 weeks of heroin abstinence prior to the 12-month interview, 65.2% were in treatment at 18 months. At each follow-up, there were reductions from baseline in illicit opiate use for those active in either LAAM or MM, with LAAM outperforming MM. This result is consistent with the existing literature comparing the two medications. Although based on self-report, which is subject to recall bias, there were also several less substantial improvements in criminal behavior and criminal justice contacts and in employment status, mostly related to those participants who were in active treatment. Similarly, reductions in injection HIV risk (lower frequency of injection, better needle hygiene) and in reduced number of sexual partners and high-risk sexual practices were also related to active treatment status. These results are concordant with those reported by Marsch2 in a metaanalysis of experimental studies of MM. In the absence of active NRT participation, however, nearly all behaviors rebounded to nearly pretreatment levels. Thus, these findings confirm that the advantages associated with NRT are primarily contingent on active treatment participation, which is, again, consistent with existing literature. Generally, the findings continue to support the contention that NRT is an effective form of treatment for heroin addiction but that its effects are optimally obtained only under continuous medication administration.

Policy issues related to maintenance funding While not a design element of the current study, the retention results illuminate the public policy paradox that exists in the USA in regard to funding for NRT, particularly MM, as opposed to all other forms of drug treatment. Private non-profit and for-profit client-paid fee-for-service mechanisms began replacing the original extensive public funding for NRT in the late 1970s, swept the nation by the mid-1980s, and remains, for the most part, in effect nationally.9 There was a limited remission of this policy when public concern for HIV transmission among injection drug users was high. NIDA, during the late 1980s and early 1990s, provided funding to pay for treatment services required by research intervention projects. Under such funding stability, it is not surprising that most researchers, especially those providing maintenance medications, reported excellent treatment retention.24 For example, Grella et al.19 reported that for a MM program targeting heroin addicts at high-risk for HIV contraction, treatment retention over a 2-year subsidized period was 70% at 1 year, and total retention time averaged 482 days over 2 years. In

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

279

contrast, of admissions to MM for all other MM programs in Los Angeles County during the same period (1991 through 1993), which are funded primarily as client-paid fee-for-service, only 17% were retained for 1 year. Moreover, the average time in treatment was significantly shorter, 178 days, over a 2-year period (analyses conducted by the authors utilizing data from the Los Angeles County Participant Reporting System).22 Although these comparisons were not statistically adjusted for client characteristics, there were no significant differences in client demographics for the research sample compared to the general countywide sample based on aggregate data. The current study, in which all treatment services were paid by NIDA for 12 months, had excellent retention rates for all participants at 6 months (about 77%) and differentially by medication group at 12 months (62.1%, LAAM; 49.5%, MM). Yet again, of all second or later admissions to MM (97% of study participants had previously been in MM at least once) during the study period (February 1997 through January 1999) in Los Angeles County, only 38.6% and 22.1% were retained at 6 and 12 months, respectively. Furthermore, the mean number of treatment days, 257.4 (SD=117) for LAAM and 248.7 (SD=110) for MM, exceeded time in treatment for the Los Angeles MM comparison group, which was 165.2 days (SD=77). As noted earlier, the study and countywide comparison samples’ demographics and drug histories did not statistically differ. The economic uncertainty that characterized the participants is a serious impediment to maintenance stability in the absence of public funding, which is typically allocated for all other forms of drug abuse treatment. Unless such policies change, fee-for-service NRT paid primarily by patients themselves will continue to demonstrate shorter durations in treatment and substantial counterproductive turnover in the client base than publicly funded services, limiting the potential benefits that accrue from treatment continuity. Policymakers, providers, and clients should make a concerted effort to induce suitable federal, state, and local policy changes regarding the public funding of NRT. Study limitations The study sample is restricted to NRT-seeking individuals addicted to heroin in the Los Angeles area; hence, generalizability may be limited. Issues of recall bias and the planned termination of subsidized treatment have been discussed earlier. Although the study was designed as a randomized clinical trial and the data were analyzed on an intent-to-treat basis, treatment termination was not likely a random event. Presumably, participants with a poorer prognosis terminated NRT more often than participants with a better prognosis. Accordingly, the in-treatment versus not-in-treatment differences may be inflated, although analyses controlled for basic demographic differences. Moreover, significance was reported at the 0.10 level to minimize the likelihood of overlooking results worthy of further study, although this was at some risk of inflating type I error.

Implications for Addiction Health Services The results from the current study indicate that a greater understanding of heroin addiction as a chronic illness that requires management through long-term treatment needs to be promulgated among policy makers, clinicians, and patients. This is particularly evident in the data from the 18month follow-up in which the type of medication was less important than active treatment status. Those individuals who were still in treatment at 18 months had significantly better treatment outcomes when compared to those that had discontinued treatment over the course of the study. Addiction researchers need to continue studying the best ways to improve treatment engagement and continuation, especially after research or other supportive funding is discontinued and participants become responsible to pay for further treatment.

280

The Journal of Behavioral Health Services & Research

36:2

April 2009

LAAM has consistently been shown to have several behavioral, clinical, and administrative advantages over MM. Within medicine, all drugs have the potential for adverse reactions. Such risks are weighed against the potential benefit to the patient. Morbidity and mortality rates for heroin addicts are several times that of the general population, especially so for those just released from incarceration, and overdose deaths predominate.25–27 Although immediate entry into treatment would ameliorate this mortality, it is difficult to determine the number of lives saved or the prolongation of life attributable to the benefits of LAAM and MM. Accordingly, the relatively few deaths in which LAAM was assessed as a contributing factor had a disproportionate effect on its regulation.6 Earlier research into potential patient conditions conferring vulnerability to LAAM administration and more careful medical monitoring of LAAM patients could have allowed it to become a safe and welcome addition to the treatment armamentarium for heroin addiction. Although LAAM is no longer available for NRT, the history of its development, approval, utilization, and market withdrawal illustrate common issues with the introduction of any new medication. Had LAAM been better researched in community-based clinics with findings widely conveyed to NRT programs and personnel, especially in regard to prudent induction procedures and medical monitoring, the unfortunate termination of a successful medication could have been avoided.

Acknowledgments Data collection and analysis were supported by grant R01 DA10422 from the National Institute on Drug Abuse. Dr. Anglin is also supported by NIDA Senior Research Scientist Award K05 DA00146. The authors wish to thank Gerhard Hellemann, Ph.D. for his assistance in statistical analysis.

References 1. Ball J, Ross A. The Effectiveness of Methadone Maintenance Treatment: Patients, Programs, Services, and Outcome. New York: Springer; 1991. 2. Marsch L. The efficacy of methadone maintenance interventions in reducing illicit opiate use, HIV risk behavior and criminality: a metaanalysis. Addiction. 1998;93:515–532, doi:10.1046/j.1360-0443.1998.9345157.x. 3. US Food and Drug Administration (FDA). LAAM Approved to Treat Drug Dependence. Talk Paper T93-36. August 3, 1993. Available at: http://www.fda.gov/bbs/topics/ANSWERS/ANS00517.html. Accessed February 14, 2008. 4. Vocci FJ, Acri J, Elkashef A. Medication development for addictive disorders: the state of the science. American Journal of Psychiatry. 2005;162:1432–1440, doi:10.1176/appi.ajp.162.8.1432. 5. Longshore D, Annon J, Anglin MD, et al. Levo-alpha-acetylmethadol (LAAM) versus methadone: treatment retention and opiate use. Addiction. 2005;100:1131–1139, doi:10.1111/j.1360-0443.2005.01122.x. 6. Anglin MD, Conner BT, Annon J, et al. Levo-alpha-acetylmethadol (LAAM) versus methadone maintenance: one-year treatment retention, outcomes, and status. Addiction. 2007;102:1432–1442, doi:10.1111/j.1360-0443.2007.01935.x. 7. Newcombe D, Bochner F, White J, et al. Evaluation of levo-alpha-acetylmethadol (LAAM) as an alternative treatment for methadone maintenance patients who regularly experience withdrawal: a pharmacokinetic and pharmacodynamic analysis. Drug and Alcohol Dependence. 2004;76:63–72, doi:10.1016/j.drugalcdep.2004.04.004. 8. Fischer B, Rehm J, Kim G, et al. Eyes wide shut: a conceptual and empirical critique of methadone maintenance treatment. European Addiction Research. 2005;11:1–14, doi:10.1159/000081410. 9. Anglin MD. Conversation with M. Douglas Anglin. Addiction. 2006;101:169–180, doi:10.1111/j.1360-0443.2006.01318.x. 10. Villafranca S, McKellar J, Trafton J, et al. Predictors of retention in methadone programs: a signal detection analysis. Drug and Alcohol Dependence. 2006;83:218–224, doi:10.1016/j.drugalcdep.2005.11.020. 11. Anglin MD, McGlothlin WH. Methadone maintenance in California: a decade’s experience. In: Brill L, Winick C, eds. The Yearbook of Substance Use and Abuse. Vol. III. New York: Humana Sciences; 1985:219–280. 12. Anglin MD, Perrochet B. Drug use and crime: a historical review of research conducted by the UCLA Drug Abuse Research Center. Substance Use and Misuse. 1998;33:1871–1914. 13. Anglin MD, Powers KI. Methadone treatment and legal supervision: individual and joint effects on the behavior of narcotics addicts. Journal of Applied Behavioral Science. 1991;27:515–531, doi:10.1177/0021886391274009. 14. Powers KI, Anglin MD. A differential assessment of the cumulative versus stabilizing effect of methadone maintenance treatment. Evaluation Review. 1998;22:175–206, doi:10.1177/0193841X9802200202. 15. Longshore DL, Hsieh S, Danila B, et al. Methadone maintenance and needle/syringe sharing. International Journal of the Addictions. 1993;28:983–996.

Longitudinal Effects of LAAM and MM on Heroin Addict Behavior

ANGLIN et al.

281

16. Grella CE, Anglin MD, Wugalter SW, et al. The effectiveness of the Los Angeles Enhanced Methadone Maintenance Project: reducing HIV risk among injection drug users. In: Tims FM, Inciardi JA, Fletcher W, et al, eds. The Effectiveness of Innovative Approaches in the Treatment of Drug Abuse. Westport, CT: Greenwood; 1997:17–32. 17. Lott D, Strain E, Bronner R, et al. HIV risk behavior during pharmacologic treatment for opioid dependence: a comparison of levomethadyl acetate, buprenorphine, and methadone. Journal of Substance Abuse Treatment. 2006;31:187–194, doi:10.1016/j. jsat.2006.04.005. 18. Longshore D, Hsieh S, Anglin MD. Reducing HIV risk behavior among injection drug users: effect of methadone maintenance treatment on number of sex partners. International Journal of the Addictions. 1994;29:741–757. 19. Grella CE, Anglin MD, Annon JJ. HIV risk behaviors among women in methadone maintenance treatment. Substance Use and Misuse. 1996;31:277–301, doi:10.3109/10826089609045813. 20. Jaffe JH. Can LAAM, like Lazarus, come back from the dead? Addiction. 2007;102(9):1342–1343, doi:10.1111/j.1360-0443.2007.01976.x. 21. Grella C, Anglin M, Rawson R, et al. What happens when a demonstration project ends: consequences for a clinic and its clients. Journal of Substance Abuse Treatment. 1996;12:249–256, doi:10.1016/S0740-5472(96)00055-4. 22. Los Angeles County Participant Reporting System (LACPRS). Los Angeles County Department of Public Health, Alcohol and Drug Program Administration, FYs 1996–97, 1997–98, 1998–99. Web site: http://www.lapublichealth.org/index.htm. 23. SPSS, Inc. SPSS for Windows. Release 14.0. Chicago: SPSS Inc., 2005. 24. Tims F, Inciardi J, Fletcher B, et al. The Effectiveness of Innovative Approaches in the Treatment of Drug Abuse. Westport, CT: Greenwood; 1997. 25. Binswanger I, Stern M, Deyo R, et al. Release from prison—a high risk of death for former inmates. New England Journal of Medicine. 2007;356(2):157–165, doi:10.1056/NEJMsa064115. 26. Christensen P, Brehm H, Smith E, et al. Mortality among Danish drug users released from prison. International Journal of Prisoner Health. 2006;2(1):13–19, doi:10.1080/17449200600743644. 27. Smyth B, Hoffman V, Fan J, et al. Years of potential life lost among heroin addicts 33 years after treatment. Preventive Medicine. 2007;44 (4):369–374, doi:10.1016/j.ypmed.2006.10.003.

282

The Journal of Behavioral Health Services & Research

36:2

April 2009