NEPOD detoxification outcomes 1
This is the submitted version of this paper, which was published at:
Addict Behav. 2005 Mar;30(3):443-56. DOI: 10.1016/j.addbeh.2004.06.002
Short-Term Outcomes of Five Heroin Detoxification Methods in the Australian
Erol Digiusto a,f, Nicholas Lintzeris b, Courtney Breen a, Jo Kimber a, Richard P. Mattick a, James
Bell c, Robert Ali d, John B. Saunders e and the NEPOD Research Group g
a National Drug and Alcohol Research Centre, University of New South Wales, Sydney, NSW 2052, Australia.
b Turning Point Alcohol and Drug Centre Inc., 54-62 Gertrude Street, Fitzroy 3065, Melbourne, Australia.
c The Langton Centre, 591 South Dowling Street, Surry Hills, NSW 2010, Australia.
d Drug and Alcohol Services Council of South Australia, 161 Greenhill Road, Parkside, South Australia 5063, Australia.
e Department of Psychiatry, University of Queensland. Mental Health Centre, Royal Brisbane Hospital, Herston,
Queensland 4029, Australia.
f Western Sydney Area Drug and Alcohol Service, 4a Fleet Street, North Parramatta, NSW 2151 Australia.
g Please see Acknowledgements for list of names of NEPOD Research Group.
This study included 380 participants in five heroin detoxification trials whose data were pooled
to enable direct comparison of five detoxification methods in the Australian National Evaluation
of Pharmacotherapies for Opioid Dependence (NEPOD). Rapid detoxification achieved similar
initial abstinence rates with either anaesthesia or sedation (average 59%), which were higher
than was achieved by inpatient detoxification using clonidine plus other symptomatic
medications (24%), which in turn was higher than outpatient detoxification using either
buprenorphine (12%) or clonidine plus other symptomatic medications (4%). Older
participants and those using more illicit drugs were more likely to achieve abstinence. Entry
rates into ongoing post-detoxification treatment were: buprenorphine outpatient (65%),
sedation (63%), anaesthesia (42%), symptomatic outpatient (27%), and symptomatic inpatient
(12%). Post-detoxification treatment with buprenorphine or methadone were preferred over
naltrexone. Participants with more previous detoxification attempts were more likely to enter
post-detoxification treatment. Given that outpatient detoxification was more effective with
buprenorphine that with symptomatic medications and that rapid detoxification was more
effective than the symptomatic inpatient method, the roles of the symptomatic methods should
Keywords: heroin, detoxification, withdrawal, drug rehabilitation, methadone, naltrexone
NEPOD detoxification outcomes 2
Methods for assisting heroin-dependent individuals to detoxify from heroin have been
evaluated using a wide range of criteria in different studies. Each criterion has complexities
and limitations, making it more difficult to assess and directly compare the effectiveness of the
methods across studies. Published reports have nearly always included data regarding
withdrawal symptoms (Gowing, Ali, & White, 2000; Gowing, Ali, & White, 2001a,b,c; Gowing,
Farrell, Ali, & White, 2000), which are relevant when comparing methods that may differ in
effectiveness in managing symptoms. However, using symptom cessation as evidence of
successful detoxification (e.g., Backmund, Meyer, Eichenlaub, & Schultz, 2001; McGregor et
al., 2002; San, Cami, Peri, Mata, & Porta, 1990; White, Alcorn, & Feinmann, 2001) can be
problematic with both objective and subjective withdrawal measures (e.g., Handelsman et al.,
1987). Objective measures can be insensitive to patients’ experienced discomfort, whereas
subjective measures include symptoms that are not specific to opioid withdrawal and may be
experienced by anyone in a medical treatment environment, making it difficult to define a cutoff
score which reliably means that withdrawal symptoms have ended.
Data regarding the proportion of participants who complete detoxification are also
commonly reported, however, the operational definition of “completion” has taken many
different forms. Definitions have included administration of a first dose of naltrexone,
particularly in rapid detoxification studies (O’Connor et al., 1995; O’Connor et al., 1997;
Seoane et al., 1997; Gowing et al., 2001c); negative results on naloxone challenges or opiate-
free urine tests at a specified time-point or during a specified period (Backmund et al., 2001;
Gowing, Farrell, et al., 2000); compliance with medication (Cheskin, Fudala, & Johnson,
1994); attendance as an outpatient or inpatient for some minimum number of days (Backmund
et al., 2001; Gossop, Bradley, & Phillips, 1987; Lintzeris, Bell, Bammer, Jolley, & Rushworth,
2002; San et al., 1990; Umbricht et al., 1999); absence of withdrawal symptoms (Backmund et
al., 2001; McGregor et al., 2002; San et al., 1990); achievement of a comfortable opiate-free
state (White et al., 2001), and completion of an associated psychotherapy program (Gossop et
al., 1987). The use of completion rate as an evaluation criterion is also complicated by the fact
that clinical detoxification protocols are often flexible in nature to cater for individual patients’
needs. Furthermore, it can be hard to define when the clinical detoxification process itself
ends in the context of ongoing provision of medication or psychosocial care. Finally,
completion does not necessarily imply abstinence from opioids or absence of withdrawal
Another criterion involves achievement of some minimum initial period of abstinence
(Broers, Giner, Dumont, & Mino, 2000; O’Connor & Kosten, 1998; Gerra et al., 1995; Gerra et
al., 2000). Issues requiring consideration here include the length of the abstinence period,
whether objective verification of self-reported abstinence is required, whether to ignore
participants’ testing of naltrexone blockade by ineffectively using heroin during the
“abstinence” period, and whether prescribed opioids used in the detoxification procedure, such
NEPOD detoxification outcomes 3
as codeine, should be ignored. There is also the issue of when the defined period should
begin – at the start of self-reported abstinence (even if this is before the beginning of the
formal detoxification process), at the beginning of the process, or at the end of it. If it
commences when detoxification has been “completed”, how do we address the fact that the
actual period of abstinence involved will differ for different detoxification methods?
Finally, another criterion is the extent to which a detoxification method is effective in
engaging patients in subsequent treatment, given that patients who have undergone
detoxification-only have been found as likely to relapse as people who did not enter any
treatment (Gerstein & Harwood, 1990; Simpson, Joe, & Bracy, 1982). Although people may
enter a detoxification program without intending to engage in subsequent treatment, we
believe that detoxification processes should generally be provided in a context that facilitates
entry to ongoing treatment (Lintzeris et. al., 2002; O’Connor et al., 1995). However, recent
reviews have found that data regarding post-detoxification outcomes, including entry rates into
ongoing treatment, have rarely been reported (Gowing, Ali, & White, 2000; Gowing et al.,
2001a; O’Connor & Kosten, 1998), except in relation to naltrexone treatment following rapid
detoxification (Gowing et al., 2001c; Legarda & Gossop, 1994; O’Connor & Kosten, 1998;
Vining, Kosten, & Kleber, 1988).
Many published evaluations of heroin detoxification methods have been case-series or
single-group studies, and few of the remaining studies have involved more than two types of
method (Gowing et al., 2001a,b,c; Gowing, Farrell, et al., 2000; O’Connor & Kosten, 1998).
Our knowledge in terms of direct comparisons of the effectiveness of different methods is
consequently limited, particularly given the complexities of the evaluation criteria employed
and the fact that relevant methodological details are often omitted from published reports. The
present study was designed to compare five detoxification methods in terms of two
standardised, useful short-term outcome measures: their effectiveness in achieving an initial
period of abstinence and the rate at which their participants entered post-detoxification
treatment. In addition, as available evidence regarding predictors of detoxification outcomes is
limited, the study examined several potential predictor variables.
2.1. Study Context
The study was based on pooled data collected in five trials that were included in the
Australian National Evaluation of Pharmacotherapies for Opioid Dependence (NEPOD) project
(Digiusto et al., 2001). NEPOD was a co-ordinating project funded by the Australian
Commonwealth government in which a group of collaborating investigators who conducted
independent trials contributed relevant data to a pooled data set that was analysed centrally.
Each trial included in the study evaluated one or two detoxification methods. As a
NEPOD detoxification outcomes 4
consequence of the pooling process, the analyses and results reported herein do not duplicate
those in publications from the individual trials. One trial compared rapid detoxification under
anaesthesia versus conventional inpatient detoxification (McGregor et al., 2002). Another trial
compared rapid detoxification under anaesthesia versus sedation (Saunders et al.,
unpublished). A third trial was a single-group study of rapid detoxification under sedation (Bell
et al., 1999). Two trials evaluated outpatient methods. One of these compared the use of
buprenorphine versus symptomatic medications, including clonidine (Lintzeris et al., 2002).
The other outpatient trial compared the use of buprenorphine in specialist clinic versus general
practice settings; data from the two settings were pooled for the present study – there were no
significant outcome differences (Gibson et al., 2003).
2.2. Participant Inclusion and Exclusion Criteria
All trials involved detailed baseline clinical and research assessments of participants.
Major inclusion and exclusion criteria were similar across trials. The trials included participants
who were dependent on heroin (DSM-IV), aged over 18, and able and willing to give informed
consent. The trials excluded participants who were pregnant, had relevant serious medical or
psychiatric conditions, or were dependent on psychoactive drugs other than heroin. Two trials
also required participants to have a support person (McGregor et al., 2002; Saunders et al.,
unpublished), and one trial required that participants have a social environment suitable for
outpatient withdrawal (Lintzeris et al., 2002). Local ethics committees approved each trial
included in NEPOD, and informed consent was obtained from all participants. Any data
collected from participants who subsequently withdrew their consent were excluded.
2.3. Detoxification Methods Evaluated
Five detoxification methods were evaluated in the study. The descriptions of methods
provided here relate mainly to the first week of the detoxification episodes, which is the focus
of this report. All of the methods included administration of additional medications to
participants if needed to alleviate specific withdrawal symptoms, and provision of counselling.
Participants in conventional inpatient detoxification (McGregor et al., 2002) were given
clonidine and other medications as inpatients at a drug and alcohol facility for a mean of 3.5
days. At least three days after admission, the OOWS scale and naloxone challenges were
used to confirm completion of withdrawal, following which naltrexone treatment was
commenced. After discharge, participants were monitored daily (usually by telephone),
provided with support and intervention if needed, and were seen at the end of Week One.
Participants in conventional outpatient detoxification (Lintzeris et al., 2002) had daily
appointments with a medical officer or a counsellor at one of two drug and alcohol clinics
during the first week, and were given clonidine and other medications. Buprenorphine-based
outpatient detoxification was evaluated in two trials. One trial was conducted at two drug and
alcohol clinics (Lintzeris et al., 2002), whereas the other trial was conducted at one of those
clinics and also in the general practice setting by GPs who had been trained in the method
NEPOD detoxification outcomes 5
(Gibson et al., 2003). The same dosing protocol was used in both trials, with 4 –12mg of
buprenorphine being administered daily over a 5-day period, with clinical review and
symptomatic medications (if needed) thereafter. Participants in the two trials of sedation-
based rapid detoxification were inpatients in public hospital drug and alcohol wards for a mean
of 2.2 days. In one trial (Bell et al., 1999), participants underwent withdrawal accelerated by
naltrexone on the day of admission. Participants were usually discharged on Day 2 and were
thereafter seen daily. In the other sedation trial (Saunders et al., unpublished), participants
underwent withdrawal accelerated by naloxone and naltrexone on the day after admission.
Participants were usually discharged on Day 3 and were thereafter contacted daily by
telephone. Anaesthesia-based rapid detoxification was evaluated in two trials that involved a
mean inpatient stay of 2.3 days. Participants in one trial (Saunders et al., unpublished) were
initially admitted to a public hospital drug and alcohol ward, were transferred to a private
hospital on Day 2 to undergo withdrawal accelerated by naloxone and naltrexone. They were
then transferred back to the public hospital for monitoring and care until discharged and were
thereafter contacted by telephone daily. Participants in the other anaesthesia trial (McGregor
et al., 2002) were admitted directly to the intensive care unit of a public hospital to undergo
withdrawal accelerated by naloxone. Naltrexone was not administered until completion of
withdrawal had been confirmed using the OOWS scale and a naloxone challenge.
Participants were discharged on the day of admission if well enough, or following further
inpatient care if necessary. After discharge, participants were contacted daily by telephone
and were seen at the end of Week One.
2.4. Outcome Measures
Two dichotomous short-term outcome measures were defined for the present study: (a)
completion of detoxification plus achievement of an initial seven-day period of abstinence from
heroin use, and (b) initiation of post-detoxification pharmacotherapy (methadone,
buprenorphine, or naltrexone), regardless of whether initial abstinence was achieved. To
maximise within-trial accuracy, all available relevant data from each trial were used to code
these outcomes, subject to the coding rules outlined below. The individual trials are being
reported elsewhere by their investigators (e.g., Bell et al., 1999; Gibson et al., 2003; Lintzeris
et al., 2002; McGregor et al., 2002), generally using different outcome measures, or different
definitions to those used herein.
The completed-detoxification aspect of the first outcome measure was operationalised
somewhat differently for each method. This variability was not considered to be a significant
confounding factor given the additional requirement of seven days’ abstinence for the first
outcome measure, as explained later. For rapid detoxification, “completion” required initiation
of anaesthesia / sedation and administration of naltrexone. In one of the trials, it also required
an OOWS score less than 5 (Handelsman et al., 1987) and a negative result on a naloxone
challenge (McGregor et al., 2002). For conventional inpatient detoxification, “completion”
required a minimum three day inpatient admission, an OOWS score less than 5 (Handelsman
NEPOD detoxification outcomes 6
et al., 1987), a negative result on a naloxone challenge, and administration of naltrexone. For
conventional and buprenorphine-based outpatient detoxification (Gibson et al., 2003; Lintzeris
et al., 2002), “completion” required that participants did not miss two or more consecutive daily
clinic appointments in the eight day episode.
The seven-day heroin abstinence period began on the first day on which abstinence
was expected with each detoxification method (i.e., the day of admission for inpatient methods
and the first day of pharmacotherapy for outpatient methods). Abstinence outcomes were
generally based on participants’ self-reports regarding non-use of heroin or regular use of
naltrexone (which was assumed to prevent heroin use), but in some cases data regarding
naltrexone challenges or urine tests on Day 7 or 8 were available and were used. Self-reports
of heroin use have been found adequately valid for evaluating treatments (Darke, 1998;
Digiusto, Seres, Bibby, & Batey, 1996; Secades-Villa & Fernandez-Hermida, 2003).
Abstinence outcomes were coded conservatively in this study, as follows. Participants who
reported any heroin use were regarded as non-abstinent even if they had a negative urine
sample or reported regular naltrexone use. Participants who reported abstinence but provided
heroin-positive urine samples were regarded as non-abstinent. Participants were also
regarded as non-abstinent if follow-up data were unavailable because they had dropped out.
Opioids prescribed as part of the detoxification process were ignored.
Different types of appropriate ongoing pharmacotherapy were available following the
different detoxification methods. The rapid and conventional inpatient detoxification methods
aimed to induct participants into ongoing naltrexone treatment. The conventional outpatient
detoxification method provided participants with a choice of either naltrexone treatment or
methadone maintenance. The buprenorphine-based method provided participants with a
choice of either naltrexone, methadone or buprenorphine maintenance. Entry into non-
pharmacotherapy treatments (e.g., counselling or residential rehabilitation), re-entry into
detoxification, or entry to treatments provided outside the trials were not counted for this
2.5. Trial Monitoring and Data Quality Assurance
A systematic trial monitoring and data quality assurance process, supervised by an
independent monitoring company, was implemented across all the trials included in NEPOD.
Coding and entry of data were guided by manuals that specified conventions to be followed in
relation to missing data, not-applicable data, coding of ambiguous responses from
participants, etc. Trial monitoring staff verified the existence of participants and checked
adherence to trials’ inclusion and exclusion criteria. A minimum 10% sample of each trial’s
data were re-entered and cross-checked against the trial’s main data files using the SPSS
Data Entry package. Any errors or ambiguities found were communicated to trial staff for
NEPOD detoxification outcomes 7
2.6. Data Analysis
Differences between the detoxification methods were analysed using logistic
regression, with seven baseline participant characteristic variables used as covariates to
reduce any effects of baseline differences between the groups of participants who entered the
various detoxification methods. The variables used as covariates (and in separate analyses
as predictors of outcome) were: age, sex, the number of DSM-IV opioid dependence criteria
satisfied (Feingold & Rounsaville, 1995), number of drug classes used in the past 28 days
(“polydrug score”), heroin use in the past 28 days (coded as 0 = daily, 1 = less than daily),
employment status (coded as unemployed, part-time, or full-time), and number of previous
entries to detoxification treatment (coded as 0, 1, 2, 3, 4-6, 7+). Selection of predictor
variables was guided by a-priori plausibility and findings in previous studies (eg., Backmund et
al., 2001; Broers et al., 2000; Harrison & Asche, 1999; Lawental, 2000; White et al., 2001), but
was limited to variables that were available in the NEPOD core data set.
All participants who entered the trials were included in analyses on an intention-to-treat
basis (other than any who withdrew consent). Statistical significance was determined on the
basis of decision-wise tests with an alpha level of 0.05. Analyses were carried out using
Version 11 of SPSS for Windows.
The study included 380 participants, of whom 62% were male, and 37% were in some
form of employment or were students. On entry to the trials, they had a mean age of 30 years
(range 18 – 50), they had been using heroin for a mean of nine years, and all satisfied four or
more of the seven DSM-IV opioid dependence criteria (mean = 6.5). In the past 28 days, they
had used an average of 3.8 classes of illicit drugs, and 91% had used heroin every day. They
had previously undertaken inpatient or outpatient detoxification a mean of 4.2 times. Detailed
information regarding participants in each of the detoxification methods is shown in Table 1.
Significant differences were found between the groups that entered the various methods in
terms of their DSM-IV scores, previous detoxification attempts, and the percentages who were
employed or students.
Table 1 about here
3.1. Short-Term Outcomes
All of the heroin users who enrolled in the trials were included in the outcome analyses,
on an intention-to-treat basis. All who enrolled in the outpatient detoxification trials actually
began the detoxification procedure. Of whose who enrolled in the inpatient detoxification
trials, 33/40 (83%), 63/76 (83%) and 34/50 (68%) actually began the sedation, anaesthesia
and conventional inpatient methods. Table 2 shows the percentages of participants who
NEPOD detoxification outcomes 8
completed detoxification and achieved (at least) an initial seven day period of abstinence from
heroin use, and the percentages who entered post-detoxification pharmacotherapy.
Table 2 about here
The detoxification methods differed in terms of completion / initial abstinence rates,
ranging from 4% for conventional outpatient detoxification to 60% for sedation-based rapid
detoxification. About half of the rapid detoxification participants entered naltrexone treatment,
which was their only available option. Participants in the two outpatient methods were more
likely to enter post-detoxification treatment with buprenorphine when it was available (56%) or
methadone (23%), rather than with naltrexone (4-6%).
Table 3 shows results of comparisons between the abstinence rates achieved by the
various detoxification methods, using logistic regression with seven baseline variables as
covariates. Rapid detoxification under either anaesthesia or sedation achieved equivalent
abstinence rates (average 59%). Participants were more likely to achieve abstinence through
rapid detoxification than through either conventional inpatient detoxification (24%) or outpatient
detoxification using buprenorphine (12%). Conventional inpatient detoxification was more
effective in achieving initial abstinence than outpatient detoxification using buprenorphine.
Finally, outpatient detoxification was more effective using buprenorphine than when
conventional symptomatic medications were used (4%).
Table 3 about here
Table 4 shows the results of logistic regression analyses which compared the post-
detoxification pharmacotherapy entry rates observed with the various detoxification methods.
The difference between the entry rates in the anaesthesia and sedation groups was not
significant, but the averaged rate in those two groups (49%) was significantly higher than in
the conventional inpatient group (12%). The rate was highest in the buprenorphine group
(65%), which was significantly higher than in the combined anaesthesia and sedation groups,
the conventional outpatient group (27%), or the conventional inpatient group.
Table 4 about here
3.2. Predictors of Outcome
Tables 5 and 6 show the results of analyses of predictors of the two short-term
outcomes. These analyses were conducted across all detoxification methods, as the sample
sizes who entered most methods were too small to warrant separate analysis. Initial
abstinence was significantly more likely to be achieved by older participants and by those who
were using more drug classes at baseline. Participants who had previously undertaken
detoxification more often were significantly more likely to enter post-detoxification treatment.
NEPOD detoxification outcomes 9
When these three relationships were examined on a bivariate basis, they were found to be
weak (Spearman rho = 0.07, 0.11, 0.14, respectively).
Tables 5 and 6 about here
4.1. Achievement of Initial Abstinence From Heroin Use
The rapid detoxification techniques, using anaesthesia or sedation, led to the highest
short-term abstinence rates (average 59%). Previous studies have found rates of completion
of detoxification and induction onto naltrexone for RODS and RODA ranging from 75% to
100% (Cucchia, Monnat, Spagnoli, Ferrero, & Bertschy, 1998; Gerra et al., 2000; Gowing, Ali,
& White, 2000; O’Connor et al. 1995). However, those studies did not measure short-term
abstinence, which is more difficult for participants to achieve. The equivalence of the two rapid
methods in achieving abstinence was an important finding, given previous claims that
anaesthesia-based methods would yield better results than using sedation (Mattick et al.,
1997). It should be noted that the use of anaesthesia may nevertheless be warranted for
some patients, particularly those who are withdrawing from high levels of methadone. The
rapid detoxification methods were more effective in achieving initial abstinence than
conventional inpatient detoxification (24%). A key feature of the rapid methods is that they are
near-irreversible once initiated, such that most participants who commence them are inducted
onto naltrexone. In comparison, participants in conventional inpatient detoxification must
typically stay for longer, experiencing withdrawal symptoms, and are more able to exit prior to
completion. The rapid and conventional inpatient methods were all more effective in achieving
abstinence than the outpatient methods using buprenorphine (12%) or conventional
medications (4%). This result was consistent with findings in a comprehensive review that
inpatient methods are more likely to be completed than outpatient methods (Gowing, Ali, &
White, 2000). The greater effectiveness of inpatient methods may be due to several reasons,
including participant self-selection, separation from social contacts and environments that
trigger use, reduced ability to obtain heroin, more supervision and support in the inpatient
setting, and greater capacity to adjust symptomatic medication and safely administer high
doses of medication. The use of buprenorphine was more effective than conventional
symptomatic medications in achieving initial abstinence in the outpatient methods, although it
should be borne in mind that the former group were actually taking buprenorphine for most of
the seven day period. This was also an important finding, as only very limited relevant data
from randomised studies have been published. Other studies have found that using
buprenorphine led to higher completion rates than conventional medications, in both inpatient
(Cheskin et al., 1994) and outpatient settings (Lintzeris et al., 2002; O’Connor et al., 1997;
White et al., 2001), however only Lintzeris et al. (2002) reported short-term abstinence rates.
NEPOD detoxification outcomes 10
4.2. Entry Into Post-Detoxification Treatment
Only a small proportion of regular heroin users who detoxify are able to maintain
abstinence, or even limit their heroin use, without further treatment (Mattick & Hall, 1996).
Thus, detoxification is usually only the first step towards sustained control over heroin use, and
engagement in ongoing treatment is arguably more clinically important than achieving initial
abstinence. In spite of this, it has been rare for detoxification studies to report rates of entry
into subsequent ongoing treatment (Gowing, Ali, & White, 2000).
Rapid detoxification led to 42% of anaesthesia participants and 63% of sedation
participants entering naltrexone treatment. These outcomes were not significantly different,
and were similar to data reported from four previous studies, involving a total of 52 participants
of whom 54% entered naltrexone treatment (Gowing et al., 2001c). The overall rate of 49%
with the rapid methods was higher than the 12% entry rate into naltrexone treatment in the
conventional inpatient group, presumably for the same reasons as were given above regarding
abstinence rates. Trials of conventional inpatient detoxification have not previously had an
aim of engaging participants in naltrexone treatment, and there are consequently no other
published data available for comparison. The contrast between the high naltrexone treatment
uptake rates observed with the rapid methods and the lower rates among participants in the
outpatient detoxification methods would have been due to rapid detoxification’s irreversibility,
to the fact that naltrexone was the only available option following rapid detoxification, and
perhaps to differences in the type of people who entered the outpatient versus rapid methods
in the first place. It is conceivable that people who choose to enter rapid detoxification are less
ambivalent, at least initially, about abstaining from heroin use or entering naltrexone treatment.
Outpatient detoxification participants who entered post-detoxification
pharmacotherapies showed a clear preference for agonist or partial-agonist maintenance
rather than naltrexone treatment. These results may have reflected ambivalence about
abstaining completely from opioids, and perceived or actual inability to do so by these
participants, bearing in mind that few outpatient participants had achieved sustained
abstinence from heroin and therefore most could not transfer directly to naltrexone. The
preference for maintenance with buprenorphine rather than methadone among the
buprenorphine detoxification participants may have been due to several factors, including their
very recent experience of buprenorphine treatment, its novelty as a treatment option, its
unavailability at the time outside of clinical trials, a preference for its pharmacological effects
and less-frequent dosing, and possibly negative attitudes towards methadone treatment.
Some of these factors may also have contributed to the difference between the buprenorphine
and conventional outpatient methods in terms of entry to post-detoxification treatment (65%
versus 27%, respectively).
NEPOD detoxification outcomes 11
4.3. Predictors of Outcome
Few published reports have included data regarding predictors of detoxification
outcomes, and, as far as we are aware, none have done so regarding the two outcome
variables examined herein. Published studies have found that participants were more likely to
complete detoxification successfully if they had relatively low depression and anxiety scores
(Kosten, Rounsaville, & Kleber, 1984; Rounsaville, Kosten, & Kleber, 1985); were older, better
educated and did not have other drug-dependent people in their families (Backmund et al.,
2001); were male, did not live with a drug user, and were employed (White et al., 2001); and
engaged in less illegal activity and use of other illicit drugs (Broers et al., 2000). Other studies
have found that long-term abstinence was more likely to be achieved by participants with
higher education and employment status (Harrison & Asche, 1999; Lawental, 2000); who were
older and who engaged in less polydrug use (Harrison & Asche, 1999); or who were younger
and had lower DSM-IV dependence scores (McGregor et al., 2002). These findings generally
indicate that people who are relatively psychologically healthy, mature, and who are less
socially marginalised and less deeply involved in drug use tend to have better outcomes.
The present study investigated seven predictor variables, but failed to add much to our
knowledge, possibly in part because of limitations on the variables that had been measured in
all the trials. Two variables predicted initial abstinence. Consistent with previous findings,
older participants were more likely to achieve abstinence. Contrary to expectation, however,
participants who were using more illicit drug classes at baseline were also more likely to
achieve abstinence. The only post-hoc explanation that we can suggest for the latter finding is
that those who saw themselves in the most serious condition may have tried harder. The only
significant predictor of entry to post-detoxification pharmacotherapy was participants’ number
of previous detoxification attempts, such that those with more previous (presumably failed)
attempts were more likely to enter ongoing pharmacotherapy. It seems reasonable to surmise
that participants with more previous experience of failure to maintain abstinence may have
been more likely to believe that detoxification alone was unlikely to have lasting benefits. We
had anticipated that the two measures of dependence (DSM-IV opioid dependence score and
heroin use in past 28 days) would predict initial abstinence in particular, but they did not. In
retrospect, it may be that the seven days’ abstinence criterion was too short in this regard
(limited by availability of comparable data from all trials), as this was a period that involved
relatively intense support and, for nearly half of the sample, an average of about three days of
inpatient stay. The remaining two variables (sex and employment status) were also unrelated
to either outcome. It should be noted that even the statistically significant relationships found
were not strong (Spearman rho 0.07 – 0.14), and are unlikely to be very useful clinically.
4.4. Methodological Issues
Randomised trials enable comparisons between treatments to be based on the
assumption that the groups of participants who are offered the various treatments are similar
NEPOD detoxification outcomes 12
in all important respects, and that any outcome differences are due to the treatments
themselves. However, in the present study, data from five trials were pooled to enable quasi-
experimental comparisons between detoxification methods which were evaluated in different
trials. In this situation, it is possible that outcome differences were determined partly by pre-
existing differences between the groups of participants who entered the various methods. It is
important to note that a similar situation exists with any meta-analytic or descriptive review that
compares outcomes of treatments across studies. We do not believe that this issue materially
affected the findings of this study. Significant differences between groups were found on three
of the seven baseline variables examined: DSM-IV scores, employment status, and number of
previous detoxification attempts. The first two of those variables did not significantly predict
either outcome, and thus the differences are unlikely to have directly influenced the outcomes.
Number of previous detoxification attempts did predict entry to post-detoxification treatment,
but the relationship was quite weak (Spearman rho = 0.14). No significant differences
between groups were found in terms of the two variables that predicted initial abstinence, i.e.,
age and mean number of drug classes being used. Finally, any such problems should have
been reduced by the use of covariance analysis to adjust for baseline differences. It must be
acknowledged, however, that we were limited to the baseline variables that were available in
the NEPOD core data set, and that the possibility of a biasing effect of pre-existing differences
between groups cannot be entirely eliminated.
Another issue relates to the fact that two of the trials involved participants being
randomly allocated to either “conventional” detoxification or to a “new” method (conventional
inpatient versus rapid detoxification; conventional versus buprenorphine-based outpatient). It
is possible that some participants who were allocated to conventional methods may have
experienced “resentful demoralisation” regarding their detoxification episode if they had been
hoping for the new method. This situation may have contributed to somewhat worse
outcomes for the conventional methods than might be achieved by patients who choose to
enter those methods in routine clinical practice. This issue potentially affects any such non-
blinded randomised study.
Focusing on rates of short-term abstinence and entry to post-detoxification treatment
enabled direct comparison of the effectiveness of a range of detoxification methods in
themselves, independently of the effects of any subsequent ongoing treatment. A
comprehensive examination of the pros and cons of each method is beyond the scope of this
paper. However, to the extent that the outcomes examined herein are considered important,
this study’s findings make it difficult to justify ongoing use of the conventional (clonidine-based)
inpatient or outpatient detoxification methods. Better outcomes were achieved with the newer
detoxification methods, an encouraging result for clinicians, researchers and policymakers
who are interested in improving management options for heroin dependence. The highest
NEPOD detoxification outcomes 13
initial abstinence rates were achieved by rapid detoxification methods. Of these, the sedation-
based method would generally be preferred because of lower medical resource requirements
and thus lower cost to healthcare providers and/or patients (a detailed cost-effectiveness
analysis of heroin detoxification methods is being reported separately by the NEPOD group
(Shanahan et al., 2003).
Some heroin users enter detoxification for reasons other than an unequivocal desire to
achieve and sustain total abstinence from opioid use. In this context, a noteworthy finding was
the high proportion of participants who entered post-detoxification treatment with
buprenorphine or methadone, rather than naltrexone, following outpatient detoxification.
Although this outcome obviously does not constitute opioid abstinence, it is usually likely to
represent a relatively good outcome. Evaluations of naltrexone treatment have usually shown
poor retention, with most participants in rapid detoxification studies relapsing to regular heroin
use within two months after detoxification (Bell & Kimber, 2000). In comparison, retention in
methadone and buprenorphine maintenance treatment is typically better (Gossop, Marsden,
Stewart, & Treacy, 2001; Ling, Wesson, Charuvastra, & Klett, 1996; Pani, Maremmani,
Pirastu, Tagliamonte, & Gessa, 2000). Thus, in spite of the relatively low short-term
abstinence rate that it achieved, buprenorphine-based outpatient detoxification has a number
of advantages. Its properties and flexibility make it potentially useful in a “marketing” sense for
attracting ambivalent heroin users into the treatment system. It may also be clinically more
appropriate than naltrexone-based methods for heroin users who present for treatment
indicating that they want to abstain from opioid use, but who are considered likely to benefit
more from maintenance with an opioid agonist.
Finally, given the range of available detoxification methods, each with its own
advantages and disadvantages, investigation would be warranted into assessment strategies
for matching patients to the most appropriate method, with the aim of improving overall
effectiveness and cost-effectiveness of treatment services. To our knowledge, we don’t yet
have evidence-based clinical guidelines to assist the matching process.
Funding for the NEPOD Project was provided by the Commonwealth Department of Health and
Aged Care. The trials included herein were funded by The Victorian government’s Turning the
Tide Community Support Fund, and the Health Departments of New South Wales, South Australia
and Queensland. Buprenorphine tablets were supplied by ReckittBenckiser. The authors would
like to thank the investigators and staff of these trials and the trial participants.
The NEPOD Research group consists of (listed alphabetically): Robert Ali, Gabriele Bammer,
James Bell, Courtney Breen, Erol Digiusto, Chris Doran, Jenny Gates, Nicholas Glasgow, Elena
Gospodarevskaya, Anthony Harris, Lynn Hawken, Nicky Henderson, Jo Kimber, Nick Lintzeris, Phil
Marshall, Richard P. Mattick, Susannah O’Brien, Allan Quigley, Alison Ritter, John B. Saunders,
Anthony Shakeshaft, Marian Shanahan, James Shearer, Jason White.
NEPOD detoxification outcomes 14
Backmund, M., Meyer, K., Eichenlaub, D., & Schutz, C.G. (2001). Predictors for completing an
inpatient detoxification program among intravenous heroin users, methadone substituted
and codeine substituted patients. Drug and Alcohol Dependence, 64, 173-180.
Bell, J.R., & Kimber, J. (2000). Guidelines for rapid detoxification from opioids prepared for New
South Wales Health. Unpublished manuscript.
Bell, J., Young, M., Masterman, S., Morris, A., Mattick, R.P., & Bammer, G. (1999). A pilot study of
naltrexone-accelerated detoxification in opioid dependence. Medical Journal of Australia,
Broers, B., Giner, F., Dumont, P., & Mino, A. (2000). Inpatient opiate detoxification in Geneva:
follow-up at 1 and 6 months. Drug and Alcohol Dependence, 58, 85-92.
Cheskin, L.J., Fudala, P.J., & Johnson, R.E. (1994). A controlled comparison of buprenorphine and
clonidine for acute detoxification from opioids. Drug and Alcohol Dependence, 36, 115-
Cucchia, A., Monnat, M., Spagnoli, J., Ferrero, F., & Bertschy, G. (1998). Ultra-rapid opiate
detoxification using deep sedation with oral midazolam: short and long-term results. Drug
and Alcohol Dependence, 52, 243-250.
Darke, S. (1998). Self-report among injecting drug users: A review. Drug and Alcohol Dependence,
Digiusto, E., Mattick, R.P., Kimber, J., O’Brien, S., Doran, C., Henderson, N., & Breen, C. (2001).
National Evaluation of Pharmacotherapies for Opioid Dependence – why have we done it?
Drug and Alcohol Review, 20, 139-141.
Digiusto, E., Seres, V., Bibby, A., & Batey, R. (1996). Concordance between urinalysis results and
self-reported drug use by applicants for methadone maintenance in Australia. Addictive
Behaviors, 21, 319-329.
Feingold, A., & Rounsaville, B. (1995). Construct validity of the dependence syndrome as
measured by DSM-IV for different psychoactive substances. Addiction, 90, 1661-1669.
Gerra, G., Marcato, A., Caccavari, R., Fontanesi, B., Delsignore, R., Fertonani, G., Avanzine, P.,
Rustichelli, P., & Passeri, M. (1995). Clonidine and opiate receptor antagonists in the
treatment of heroin addiction. Journal of Substance Abuse Treatment, 12, 35-41.
Gerra, G., Zaimovic, A., Rustichelli, P., Fontanesi, B., Zambelli, U., Timpano, M., Bocchi, C., &
Delsignore, R. (2000). Rapid opiate detoxification in outpatient treatment: Relationship with
naltrexone compliance. Journal of Substance Abuse Treatment, 18, 185-191.
Gerstein, D.R., & Harwood, H.J. (Eds.). (1990). Treating drug problems. Volume 1. A study of the
evolution, effectiveness, and financing of public and private drug treatment systems.
Washington, D.C.: National Academy Press.
Gibson, A.E., Doran, C.M., Bell, J.R., Ryan, A., Lintzeris, N. (2003). A comparison of
buprenorphine treatment in clinic and primary care settings: a randomised trial. Medical
Journal of Australia, 179, 38-42.
Gossop, M., Bradley, B., & Phillips, G.T. (1987). An investigation of withdrawal symptoms shown
by opiate addicts during and subsequent to a 21-day inpatient methadone detoxification
procedure. Addictive Behaviors, 12, 1-6.
NEPOD detoxification outcomes 15
Gossop, M., Marsden, J., Stewart, D., & Treacy, S. (2001). Outcomes after methadone
maintenance and methadone reduction treatments: two-year follow-up results from the
National Treatment Outcome Research Study. Drug and Alcohol Dependence, 62, 255-
Gowing, L.R., Ali, R.L., & White, J.M. (2000). The management of opioid withdrawal. Drug and
Alcohol Review, 19, 309-318.
Gowing, L., Farrell, M., Ali, R., & White, J. (2000). Alpha-2 adrenergic agonists for the
management of opioid withdrawal (Cochrane Review). In: The Cochrane Library, Issue 2,
2000. Oxford: Update Software.
Gowing, L., Ali, R., & White, J. (2001a). Buprenorphine for the management of opioid withdrawal
(Cochrane Review). In: The Cochrane Library, Issue 4, 2001. Oxford: Update Software.
Gowing, L., Ali, R., & White, J. (2001b). Opioid antagonists and adrenergic agonists for the
management of opioid withdrawal (Cochrane Review). In: The Cochrane Library, Issue 4,
2001. Oxford: Update Software.
Gowing, L., Ali, R., & White, J. (2001c). Opioid antagonists under sedation or anaesthesia for
opioid withdrawal (Cochrane Review). In: The Cochrane Library, Issue 4, 2001. Oxford:
Handelsman, L., Cochrane K.J., Aronson, M.J., Ness, R., Rubinstein, K.J., & Kanof, P.D. (1987).
Two new rating scales for opiate withdrawal. American Journal of Drug and Alcohol Abuse,
Harrison, P.A., & Asche, S.E. (1999). Comparison of substance abuse treatment outcomes for
inpatients and outpatients. Journal of Substance Abuse Treatment, 17, 207-220.
Kosten, T.R., Rounsaville, B.J., & Kleber, H.D. (1984). Relationship of depression to clonidine
detoxification of opiate addicts. Comprehensive Psychiatry, 25, 503-508.
Lawental, E. (2000). Ultra rapid opiate detoxification as compared to 30-day inpatient
detoxification program – a retrospective follow-up study. Journal of Substance Abuse, 11,
Legarda, J.J., & Gossop, M. (1994). A 24-h inpatient detoxification treatment for heroin addicts: a
preliminary investigation. Drug and Alcohol Dependence, 35, 91-93.
Ling, W., Wesson, D.R., Charuvastra, C., & Klett, J. (1996). A controlled trial comparing
buprenorphine and methadone maintenance in opioid dependence. Archives of General
Psychiatry, 53, 401-407.
Lintzeris, N., Bell, J., Bammer, G., Jolley, D.J., & Rushworth, L. (2002). A randomized controlled
trial of buprenorphine in the management of short-term ambulatory heroin withdrawal.
Addiction, 97, 1395-1404.
Mattick, R.P., Bell, J., Daws, L.C., White, J., O'Brien, S., & Harris, S.J. (1997). Review of the
evidence on the effectiveness of antagonists in managing opioid dependence, National
Drug and Alcohol Research Centre Monograph Series No.34. Sydney: National Drug and
Alcohol Research Centre.
Mattick, R.P., & Hall, W. (1996). Are detoxification programmes effective? Lancet, 347, 97-100.
McGregor, C., Ali, R., White, J.M., Thomas, P., & Gowing, L. (2002). A comparison of antagonist-
precipitated withdrawal under anaesthesia to standard inpatient withdrawal as a precursor
to maintenance naltrexone treatment in heroin users: outcomes at 6 and 12 months. Drug
and Alcohol Dependence, 68, 5-14.
NEPOD detoxification outcomes 16
O'Connor, P.G., Carroll, K.M., Shi, J.M., Schottenfeld, R.S., Kosten, T.R., & Rounsaville, B. J.
(1997). Three methods of opioid detoxification in a primary care setting: A randomized trial.
Annals of Internal Medicine, 127, 526-530.
O'Connor, P.G., & Kosten, T. (1998). Rapid and ultrarapid opioid detoxification techniques. Journal
of the American Medical Association, 279, 229-234.
O’Connor, P.G., Waugh, M.E., Carroll, K.M., Rounsaville, B.J., Diagkogiannis, I.A., & Schottenfeld,
R.S. (1995). Primary care-based ambulatory opioid detoxification: the results of a clinical
trial. Journal of General Internal Medicine, 10, 255-260.
Pani, P.P., Maremmani, I., Pirastu, R., Tagliamonte, A., & Gessa, G.L. (2000). Buprenorphine: a
controlled clinical trial in the treatment of opioid dependence. Drug and Alcohol
Dependence, 60, 39-50.
Rounsaville, B.J., Kosten, T., & Kleber, H. (1985). Success and failure at outpatient opioid
detoxification: evaluating the process of clonidine and methadone assisted withdrawal.
Journal of Nervous and Mental Disease, 173, 103-110.
San, L., Camí, J., Peri, J.M., Mata, R., & Porta, M. (1990). Efficacy of clonidine, guanfacine and
methadone in the rapid detoxification of heroin addicts: A controlled clinical trial. British
Journal of Addiction, 85, 141-147.
Secades-Villa, R., & Fernandez-Hermida, J.R. (2003). The validity of self-reports in a follow-up
study with drug addicts. Addictive Behaviors, 28, 1175-1182.
Seoane, A., Carrasco, G., Cabre, L., Puiggros, A., Hernandez, E., Alvarez, M., Costa, J., Molina,
R., & Sobrepere, G. (1997) Efficacy and safety of two new methods of rapid intravenous
detoxification in heroin addicts previously treated without success. British Journal of
Psychiatry, 171, 340-345.
Shanahan, M.D., Doran, C.M., Digiusto, E., Lintzeris, N., White, J., Ali, R., Saunders, J.B., &
Mattick, R.P. (2003) The use of pharmacotherapies to aid detoxification from heroin: a cost
effectiveness analysis. Manuscript submitted for publication.
Simpson, D.D., Joe, G.W., & Bracy, S.A. (1982). Six-year follow-up of opioid addicts after
admission to treatment. Archives of General Psychiatry, 39, 1318-1323.
Umbricht, A., Montoya, I.D., Hoover, D.R., Demuth, K.L., Chiang, C.T., & Preston, K.L. (1999).
Naltrexone shortened opiate withdrawal with buprenorphine. Drug and Alcohol
Dependence, 56, 181-190.
Vining, E., Kosten, T.R., & Kleber, H.D. (1988). Clinical utility of rapid clonidine-naltrexone
detoxification for opioid abusers. British Journal of Addiction, 83, 567-575.
White, R., Alcorn, R., Feimann, C. (2001). Two methods of community detoxification from opiates:
an open-label comparison of lofexidine and buprenorphine. Drug and Alcohol Dependence,
NEPOD detoxification outcomes 17
Table 1. Baseline Characteristics of Participants in Each Detoxification Method
Characteristic Sedation Anaesthesia Buprenor-
Age - mean
of drug classes
out of 8 (SD)
% with no
days in past 28
out of 7 (SD)
% employed or
* Continuous variables tested with ANOVA, categorical variables tested with chi-square tests.
NEPOD detoxification outcomes 18
Table 2. Short-Term Outcomes of Detoxification
Detoxification Method (Number of
% Who Completed
Achieved Initial 7
Days of Heroin
Type of Post-Detoxification
(4%, 56%) **
(23%, n/a *) **
Sedation-based rapid (n = 40)
Anaesthesia-based rapid (n = 76)
Conventional inpatient (n = 50)
Buprenorphine outpatient (n = 158)
Conventional outpatient (n = 56) 4% 4%
* n/a = This post-detoxification treatment was not available for this group within the trial.
** Figures in brackets show percentages of participants who entered methadone and
buprenorphine maintenance, respectively.
Table 3. Comparisons Between Detoxification Methods in Terms of Initial Abstinence
Comparison * B Wald Signif. Exp (B) 95% C.I. for Exp
Anaesthesia vs. sedation
Rapid ** vs. conventional
Conventional inpatient vs.
Rapid ** vs. buprenorphine
Buprenorphine outpatient vs.
* These comparisons were made in five separate logistic regression analyses that each
included the seven baseline variables as covariates.
** Rapid refers to the anaesthesia and sedation groups combined.
- 1.11 4.89 .027 0.33 0.12 0.88
- 2.35 39.46 .000 0.10 0.05 0.20
- 1.64 3.99 .046 0.19 0.04 0.97
NEPOD detoxification outcomes 19
Table 4. Comparisons Between Detoxification Methods in Terms of Entry to Post-
Comparison * B Wald Signif. Exp (B) 95% C.I. for Exp
Anaesthesia vs. sedation
Rapid ** vs. conventional inpatient
Conventional inpatient vs.
Rapid ** vs. buprenorphine
Buprenorphine outpatient vs.
* These comparisons were made in five separate logistic regression analyses that each
included the seven baseline variables as covariates.
** Rapid refers to the anaesthesia and sedation groups combined.
Table 5. Baseline Predictors of Achievement of Initial Abstinence
0.86 8.42 .004 2.37 1.32 4.26
- 1.77 23.74 .000 0.17 0.08 0.35
Variable B Wald Signif. Exp (B) 95% C.I. for
No. of previous detoxifications
Heroin use in past 28 days
Number of drug classes
NEPOD detoxification outcomes 20 Download full-text
Table 6. Predictors of Entry to Post-Detoxification Treatment
Variable B Wald Signif. Exp (B) 95% C.I. for
No. of previous detoxifications
Heroin use in past 28 days
Number of drug classes