Transcript Long-Term Course of Opioid Addiction

Long-Term Course of
Opioid Addiction
Yih-Ing Hser, Ph.D.
UCLA Integrated Substance Abuse Programs
Addiction Seminar (Psychiatry 434)
Supported by the National Institute on Drug Abuse
(P30 DA016383)
Overview
 Background
 CALDAR
 This
topic
 Overview of morality and opioid abstinence in
long-term follow-up studies
 The 33-year follow-up study
 The START follow-up study
2
Center for Advancing Longitudinal
Drug Abuse Research (CALDAR)
Increase knowledge of patterns of drug
addiction & their interplay with
treatment and other service systems
Enhance scientific collaboration
through integration analysis, training,
consultation, dissemination
3
Examples of CALDAR’s
Long-term Follow-up Studies
1.
2.
3.
4.
5.
6.
The 33-year follow-up study of heroin addicts
A 12-year follow-up of a cocaine-dependent
sample
A 5-year follow-up of participants in the Amity
treatment program at a correction facility
Follow-up studies of methamphetamine patients
An 10-year follow-up of mothers and their
children
START follow-up study
(Starting Treatment with Agonist Replacement
Therapy—Randomization to Suboxone vs.
Methadone)
4
Longitudinal Research Design
In contrast to cross-sectional research
design—data are collected on one or
more variables for a single time period
Longitudinal research design—data
are collected on one or more variables
for two or more time periods
► Longitudinal
research design allows
measurement of change, and possibly
explanation of change
5
Goals of Longitudinal Analyses
Assess changes over time:
► How
does it change over time?
► What is the time trend?
► How does the time trend differ by group?
► Group differences at end of study (group
differences at end of study) minus (group
differences at baseline)
Investigate factors related to the
different patterns of changes
► Time
trends as functions of covariates
6
Longitudinal Drug Abuse Research
Persistence of drug use: Drug
addiction is a chronic condition
High relapse rates over long periods
of time
Non-compliance, require long-term
care management
Frequent encounters with social and
health service systems
7
Life Course Perspective on
Drug use
1.
2.
3.
4.
Life course theory recognizes the importance of time,
timing, and temporal processes in the study of
human behavior and experience over the life span,
characterized by trajectories, transitions, and turning
points
Persistence of drug use resembles chronic diseases:
high relapse rates, non-compliance, require longterm care/management
Critical life events often lead to or explain changes
Social capital, situated choice are additional key
concepts
8
Longitudinal Approach to Study
Drug Use over Time
Life-course Drug Use Career
Protective Factor
e.g. family support
Protective Factor
(occurrence of positive life events)
e.g. got married, got employed
Trajectories of drug use are heterogeneous among
individuals and can be classified as several distinctive
trajectory groups
Estimated trajectory of
Drug use
Risk Factors
e.g. crime
involvement
Age
Age
9
Global Burden of Disease

Approximately 16.5 millions people worldwide are
users of heroin or opium (UN World Drug Report
2013)

In the US, approximately 467,000 individuals with
heroin use disorder; 2,056,000 with prescription pain
relievers in 2012 (NSDUH)

Opioid dependence is the biggest contributor to
overdose deaths

Opioid dependence is the biggest contributor to global
burden of disease attributable to illicit drug use and
dependence
10
A 33-year Follow-up of HeroinDependent Sample

A cohort of 581 male heroin addicts admitted to
the California Civil Addict Program (CAP) in
1962-64 has been followed-up and interviewed
over more than 30 years

The CAP was the only major publicly-funded
drug treatment program available in California in
the 1960s

The CAP provided a combination of inpatient
and outpatient drug treatment to narcoticsdependent criminal offenders committed under
court order
11
Life Course of Heroin Addiction
Childhood/
Adolescence
Young
Adulthood
Adulthood
Middle-aged
Late-middleaged & Older
CAP
Admission
Follow-up
at 1974/75
Follow-up
at 1985/86
Follow-up
at 1996/97
Mean age = 50
Mean age = 60
28%
25%
12%
49%
23%
6%
Mean age = 25
Onset of Heroin
Mean age = 18
Mean age = 40
Death: 14%
Negative urine on heroin: 29%
Incarcerated: 18%
Influencing Factors
12
Hypothetical Drug Use
Trajectories
Incarcerated
Drug tx
Employment
Mental health tx
Criminally active
30
Days of use
25
20
15
10
5
0
1
24
36
48
60
84
96
108 120 132 180
Months
Person 1
Person 2
Person 3
13
The Natural History of Narcotics
Addiction Among CAP Sample
(N=581)
100
Daily Narcotic Use
Methadone Maintenance
Percent of Sample
80
Occasional Narcotic Use
60
Abstinent
40
Dead
Incarcerated
20
Unknown
0
56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96
Years 1956 through 1996
14
Identify Groups with Distinctive
Heroin Use Trajectories
 Growth Mixture Modeling
 First half of the observation (16 years)
since heroin initiation
 Two-part model (skewness)
 Linear and quadratic terms
 Three Distinctive Groups
 Standard statistical criteria: BIC,
entropy
15
Mean Number of Days Per Month
Using Heroin, 33 Year Follow-up
30
Decelerate
Days used
25
Stably High
20
Quitter
59%
15
32%
10
9%
5
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16
Years since first heroin use
16
Differences in Trajectory Groups:
Demographics
% 80
66
59
60
50
38
34
40
20
29
12
7
6
0
Late Decelerated
White
Stably High Use
Hispanic
Early Quitter
African American
< .01
Note: (no**p
difference
in education or age)
17
Differences in Trajectory Groups:
Mortality
80
50.3
Percent
60
38.1
40
25.0
20
0
Late Decelerated
Stably High Use
Early Quitter
**p < .01
18
Consistent with other
studies showing:

Some users did stop using

Many continued to use at high
levels, over a long period of time

At any given time, 40-60%
“relapsed”
19
What New?



Distinctive patterns of drug use
trajectory
Individual’s baseline not
necessarily determines future
Important to identify why the
different patterns of trajectory



Escalating
Decreasing
High vs. low vs. no use
20
What Have we Learned?



Cyclical patterns of abstinence and use
of different levels, protracting over a
long time
Long-term observation is necessary to
explicate addiction patterns and
trajectories. Otherwise, we may miss
the critical points or differences as well
as opportunities for intervening
If addiction is a chronic disease and
cumulative treatment effect exists, then
long-term care makes sense for these
21
individuals
21
Is stable long-term recovery
possible?
22
Rates of Abstinence by Years
Abstinent Prior to Last Interview
Percent Abstinent (1985/86 to
present)
(N = 242)
80
75%
72%
6-15
(n = 36)
15+
(n = 34)
60
40
20
15%
17%
0
(n = 85)
1-5
(n = 66)
0
Years Abstinent Prior to 1985/86
23
More than 5 Years of Abstinence:
Predicting lower depression
2
1.6
Mean Score
1.6
1.5
1.3
1.3
1.3
1.2
1.2
0.8
0.4
0
Depression *
No abstinence (N=121)
* p < .05
Anxiety
1-<5 Years (n=31)
5+ Years (n=69)
SCL58 Scale (1- 4) at the 33-year follow-up:
higher scores indicate greater symptom severity.
24
More than 5 Years of Abstinence:
Predicting better emotional well-being
100
Mean Score
80
74.0
78.0
83.0
65.0
59.0
53.0
60
40
20
0
Emotion*
No abstinence (N=121)
p < .05
Health
1-<5 Years (n=31)
5+ Years (n=69)
SF36 Scale (0-100) at the 33-year Follow-up: higher
scores indicate better a status
25
More than 5 Years of Abstinence:
Higher self-esteem and life satisfaction
30
Mean Score
21.5
20
22.8
19.1
12.4
8.3
10
9.7
0
Self-esteem **
No abstinence (N=121)
* p<.05; **p < .01
Life satisfication **
1-<5 Years (n=31)
5+ Years (n=69)
Self-Esteem (0-30) and Life Satisfaction (0-18) Scales at the
33-year : Higher scores indicate a better status
26
Alcohol, Tobacco and Illicit Drug Use
at the 33-year Follow-up
percent of subjects
100
80
78
64
60
56
45
40
43
35 32
20
14
9
65
61
48
18 16
6
0
Heroin **
Coca/Meth Other illict
*
drugs *
No abstinence (N=121)
* p<.05; **p < .01
Alcohol
1-<5 Years (n=31)
Tobacco **
5+ Years (n=69)
27
Employment at the 33-year Follow-up
percent of subjects
100
80
58
60
39
40
18
20
0
Employed **
No abstinence (N=121)
* p<.05; **p < .01
1-<5 Years (n=31)
5+ Years (n=69)
28
Summary of Findings
 Five years appear to be a good benchmark
Less future use
 Less CJS involvement
 Better emotional and social functioning

 Timing may be critical
Health is not much better
 Alcohol and tobacco still problematic

 Need to
 Understand the underlying mechanisms
 Promote recovery in early stages of
addiction
CTN START: Background

1267 opioid dependent users

Randomly assigned to Suboxone vs.
Methadone

Recruitment over the period of 2006 to
2009

Mortality status (date of death)
determined by 3/2012
START: Starting Treatment with Agonist Replacement Therapy
Suboxone (Buprenorphine+naloxone) vs. Methadone
30
Medications for Opioid Addiction

Methadone: agonist

Morphine
Tincture of opium

N
CH3
CH2 CH N
CH3 CH2
O
CH3
HO
CH3
HO

Naltrexone:antagonist

Depo-naltrexone

Buprenorphine: partial agonist
 Subutex, Suboxone, Probuphine


Clonidine: non-opioid
Lofexidine
O
O
START: Study Sites
8 sites (across 5 states)

California
 Bi-Valley Medical Clinic Inc., Sacramento (n=117/84; 201)
 BAART, Turk St. Clinic, San Francisco (n=109/78; 187)
 Matrix Institute, Los Angeles (n=78/50; 128)

Oregon
 CODA-Research, Portland (n=136/89; 225)

Washington
 Evergreen Treatment Services, Seattle (n=79/55; 134)

Connecticut
 CT Counseling Centers, Waterbury (n=71/52; 123)
 Hartford Dispensary, Hartford (n=101/71; 172)

Pennsylvania
 NET Steps, Philadelphia (n=48/49; 97)
32
START:
Treatment & Randomization

24 weeks (active phase), ending 36 weeks

739 to Suboxone vs. 528 to Methadone
2006
71 vs. 72
1:1
2007 207 vs. 197
1:1
2008 254 vs. 139
2:1
2009 192 vs. 99
2:1
33
Data Collection

Baseline (randomization)
 Demographics, substance use/urine, physical and
psychiatric history, quality of health

START treatment
 Suboxone vs. methadone, days in treatment, dose

3 waves of follow-up starting late 2011
34
Research Questions
Mortality
 Treatment retention
 Long-term use

35
Description of Sample at
Baseline: Demographics
Mean age
Female
Ethnicity
White
Black
Hispanics
37
32%
72%
8%
12%
36
Description of Sample at
Baseline: Opioid and Other Drugs
Urine
Use
Positive (%) Disorder (%)
Amphetamine
Cannabis
9
24
11
20
Cocaine
37
33
Opiates
97
100
37
Description of Sample at
Baseline: smoking and alcohol use
Current smoker
Alcohol use disorder
89%
23%
38
Description of Sample at
Baseline: psychiatric history
Schizophrenia
Major depressive disorder
Bipolar
Anxiety or panic disorder
2.5%
28%
12%
30%
39
Description of Sample at
Baseline: Quality of health1
Percentile 2
Physical
Mental health
49 (9)
39 (13)
1.
SF-36
2. Relative to the U.S. population with similar age & gender
40
Baseline differences between the
two treatment conditions
No differences in
Age, gender, ethnicity, injection, sites,
alcohol, amphetamine, cannabis, sedative
Physical and mental health quality
Exceptions: cocaine & smoking (higher in the
methadone group)
41
START Treatment
Treatment
condition
Days in treatment1
Suboxone
58%
Methadone
42%
99 (70)
138 (54)
(within 168 days)
Treatment completion2
46%
Average dose, mg
14 (8)
74%
68 (35)
1. The difference between the two treatment groups was significant at p < .01
2. The difference between the two treatment groups was significant at p < .01
42
Mortality
 Mortality status (date of death) determined by
3/2012
 Web archives: date of death
 CDC National Death Index: date and ICD-10 causes of death
 CDC has a 2-year lag time in their data
 Death certificates from local corner’s office
Figure 1. Survival Curves for Buprenorphine Versus Methadone
Buprenorphine (n=738)
1
Methadone (n=529)
0.6
0.4
0.2
168
160
140
120
100
80
60
40
20
0
0
Survival
0.8
Days in treatment during 24 weeks
1
17.1%
23.3%
8.7%
11.8%
5.8%
35.6%
Figure 3. Average Weekly Dose and Positive
Opiate over Weeks in Treatment (n=1,267)
Suboxone Dose (n=739)
Suboxone with Positive Urine (n=739)
Methadone Dose (n=528)
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
Methadone with Positive Urine
(n=528)
0
100
90
80
70
60
50
40
30
20
10
0
Week in Treatment
46
To improve retention, clinicians need to
1. use higher medication doses,
particularly for BUP,
2. address continued use of opiates and
other drugs, and
3. identify additional factors/strategies
influencing BUP retention, particularly
during the first 30 days of treatment.
The Future?
 Changing profiles of opioid addiction
 Evidence-based intervention, practice, &
principles
 Long-term care or management
 Service structure



Integration within treatment systems
Integration across systems
Affordable Care Act
 Technology
48
Longitudinal Studies and Analyses
 Two or more observations of the response variable taken
at different times are made on the same individuals
 Can be used to assess on-going/recurring behaviors &
events
 Adjust for correlated observations over time, and/or
 Allow examination of both within- and between-subjects
hypotheses, i. e. can separate
 differences within individuals (e.g. aging/drug career
progression), from
 differences among people (cohort effects)
 Allow complexity, depending on models chosen:
covariates (both time-variant and time-invariant), missing
data, clustering of observations, latent constructs,
temporal structuring
 More powerful for some hypotheses than cross-sectional
49
designs
Learn more about
longitudinal research
findings and modeling
techniques??
See CALDAR website (www.caldar.org) for new findings,
development, and workshops
50