Ever since the pathfinding studies in New York of Dole and Nyswander 1 and colleagues methadone has been the mainstay of the medical treatment of opiate addiction of patients for whom detoxification and a drug free lifestyle appears to be beyond reach. Indeed the successes of agonist replacement therapy, first with methadone and more recently with buprenorphine, have made agonist replacement therapy something of a standard of care to be sought in the treatment of other chemical addictions for which treatment as yet either does not exist or is still in its infancy. There are however increasing concerns in relation to the immunosuppressive activities of opiate 23 and other addictive drugs 457, and there have long been recognized significant cytostatic activities of opiates 8 and other addictive agents 910 which oppose cell growth and stem cell renewal which although less well known are potentially also of significant medical concern. Moreover there is increasing concern in the community about the failure of methadone to lead to a long term stable drug free lifestyle including on-going drug use, drug dealing and drug affected driving 11 together with very high rates of psychiatric symptomatology 1213. For practical clinical purposes however there has been little alternative for patients for whom the lifestyle and health issues or the legal constraints related to opiate addiction make medical assistance mandatory.

Naltrexone is one alternative which has existed for several decades but is rarely used. First synthesized by Matossian at Endo laboratories in 1963 14, it was trialed clinically under sponsorship from the National Institute of Drug Abuse (NIDA) in the early 1970's and a flurry of papers were written documenting this experience. Naltrexone itself has many useful features to assist the management of opiate addiction including direct chemical blockade of opiate use for up to three days after a dose, rapid cessation of the symptoms of acute withdrawal syndrome once the initial treatment induction is accomplished, reduction of opiate craving in most patients 151617, increasing the time to relapse to dependent use up to three days from the last administration, and in the experience of this clinic, a great reduction in the frequency with which patients think about addiction. The most well known of the early papers is that by Hollister 18 which although encouraging in terms of reporting 20% treatment retention at 7 months compared to 13% in the placebo group with many other positive adjustments in patients' lives, was felt to demonstrate a disappointingly low rate of retention in treatment. Similarly most of this early literature is similarly unenthusiastic. The major reason for this appears to be a lack of patient compliance outside of select groups such as physicians and parolees, perhaps because it does not have the addictive attraction of agonist medications, but also related to concerns of overdose and mood depression management 1920. Another feature of this literature is that, following the agonist literature, much of it reports treatment retention rather than non-return to dependent opiate use. Whilst this is standard practice with studies of agonist medications, non-compliance with the oral medication is not necessarily the same as dependent drug use. Studies which report "opiate free success" generally give a much more robust view of antagonist based treatment than those which focus solely on the former index 212223.

It has been hoped that many of these compliance problems will be overcome by the recent development of depot formulations of long acting naltrexone injections and implants 24. This was recognized very early by the National Institute of Drug Abuse which as long ago as 1976 25 and 1981 26 published research monographs on the subject of depot and long acting formulations of the drug. Several long acting depot injections with activity over about one month have recently been developed 27 and marketed in the USA 28 mainly for alcohol 293031, but are also likely to find application in opiate dependent patients. More recently a longer acting device efficacious for 4–6 months 32 and used widely in Australia for opiate dependency has been developed and trialed, and is presently under consideration before the regulatory authorities in this country. Contrary to the classical view of the limited clinical role of naltrexone 33 several fascinating papers have emerged from researchers in Perth and elsewhere in relation to the use of the depot formulation in the pregnant heroin user with improved outcomes over agonist therapy 34; the addicted anesthetist 35; prolonged maintenance of therapeutic serum naltrexone levels above 1–2 ng/ml with sequential implants of 13–17 months 36; potentiation of the antiviral effect of combination therapy for Hepatitis C 37 and HIV 3839 possibly by an immunopotentiating mechanism and a dramatic truncation of overdose behaviour in high risk frequently overdosing heroin addicts 40 to the point where the earlier concerns in relation to the safety profile of naltrexone have been largely addressed 41424344. Other centres have a similar encouraging experience with the use of such preparations in opiate dependency 24284546.

We were therefore interested to compare our early experience with a one month depot device from the USA, to the "Go Medical" implant from Perth, and to compare this experience in turn with the concurrent and historical experience with the oral formulation of naltrexone. Some of these results, and important lessons for patient management which follow from them, are presented in the following report. Serum levels of naltrexone, serious adverse events, and the experience with buprenorphine in the office management of opiate dependent patients are presented in companion papers.

There were 102, 113 and 161 patients in each of the implant, tablet and historical groups respectively, a total of 376 patents. Various follow-up and demographic criteria are presented in the accompanying Table 1. The mean ages, mean duration of opiate use, sex distribution, and doses of heroin used were as indicated. There were no significant differences between the groups on these indicators. In the following discussion the consecutive ordered triplets refer to the implant, tablet and historical groups respectively.

When however the financial and social status of the three groups were compared significant differences were revealed between the various groups. These various ratios are given in the later part of Table 1. Note that the groups were not assigned randomly. Based on a numbers in the "Pensioner" category, the differences in financial means from the implant group were significant. There were more patients in the pensioner category in the tablet group (Chi Squ. = 68.48, df = 1, O.R. = 18.15, 95% C.I. = 7.85 – 43.20, P ≤ 0.001), and the historical' group (Chi Squ. = 12.54, df = 1, O.R. = 3.77, 95% C.I. = 1.66 – 8.77, P = 0.0004). The combined differences from the implant group were significant at (Chi Squ. = 37.52, df = 1, O.R. = 7.59, 95% C.I. = 3.53 – 16.83, P ≤ 0.001). The variables of social support were also distributed differently. The implant group appeared to have the largest number of patients in the "strong" category. Compared to the tablet group this difference was not significant (Chi Squ. = 2.87, df = 1, O.R. = 1.78 95% C.I. 0.87–3.68, P = 0.09), but compared to the larger historical group it was significant (Chi Squ. = 14.65, df = 1, O.R. = 3.84 95% C.I. 1.78–8.42, P = 0.001). When both groups were analyzed together the difference was significant (Chi Squ. = 11.29, df = 1, O.R. = 1.42, 95% C.I. 1.42–4.92, P = 0.0007). These baseline sociodemographic differences should be borne in mind in the consideration of all subsequent results.

The follow-up rate was 82% overall. The overall urine drug screen follow-up was 76%. The proportion of patients in each group whose urine tested positive for naltrexone was 70%, 52% and 39% respectively and 50.4% overall. The re-treatment rate of all patients attending this clinic is 1.4 (1301 treatments in 902 patients). The numbers of patients having various numbers of implants is indicated in Table 2. The re-implant rate in the present series was 24.5% (25 of 102 patients having more than one implant; and 102 patients having 141 implants).

Figure 1 shows the proportion of patients complying with treatment by the various treatment groups. Figure 1A shows the outcome by the three treatment groups. Figure 1B illustrates the respective outcomes by the two implant and two tablet groups conflated. Figure 1C shows the outcomes by the four treatment groups including the two different implant groups. When a statistical analysis was done of the type of implant used including no implant, there was a significant difference (Chi squared = 20.93, df = 3, P = 0.001). However when the two implants alone were compared the difference was not significant (WW = 90.00, Gehan's test statistic = 0.77, P = 0.43). Figure 1D shows the outcome by the presence or absence of naltrexone in the urine (WW = 1961.0, Gehan's test statistic = 2.4089, P = 0.1600).

Table 3 lists the urine positivity rates for various illicit drugs. The rate of obtaining a urine drug screen (UDS) for assessment in the three groups was 98%, 70% and 76% respectively. The dominant drug used was cannabis, followed by amphetamines. Amphetamine use was significantly greater in the implant group than either the tablet (Chi Squ. = 7.13, df = 1, O.R. = 2.78, 95% C.I. = 1.22–6.42, P = 0.0075) or historical (Chi Squ. = 11.14, df = 1, O.R. = 3.28, 95% C.I. 1.52–7.18, P < 0.001) groups. Amphetamine use was significantly elevated in the implant group post-treatment compared to the two oral groups combined (Chi Squ. = 13.83, df = 1, O.R. = 3.05, 95% C.I. 1.60–5.83; P = 0.0002).

Table 4 lists the improvement in work status before and after treatment for the three groups. There was a significant improvement in the implant and tablet groups, but not in the historical group. If the analysis is performed on all three groups as three groups the employment status improved from 35% prior to treatment to 54% afterwards (Chi Squ. = 27.57, df = 5, O.R. = 2.24, C.I. 1.65–3.09, P < 0.001).

Review of spiritual status of patients after treatment revealed that 65%, 26%, and 9% scored 1, 2 and 3 respectively, with no significant differences in between group distribution.

Table 5 lists detailed outcome data by various treatment parameters. Figure 2 shows the outcome for selected psychosocial variables, and Figure 3 illustrates outcomes for selected treatment variables. Table 6 lists the bivariate significance of these various relationships. Dichotomous and multiple response variables have been analyzed in closely related but differing statistical modules within the software giving rise to the varying statistics presented for each variable. The following variables are significant on bivariate analysis: post-treatment employment, the number of detox's, social support, the use of implants rather than tablets, and the number of implants administered (all P ≤ 0.001), the treatment group in which patients were, the type of implant (as discussed above), financial status, the presence of morphine on urine drug screen, spirituality, olanzepine on urine testing, and naltrexone urine positivity were all significant (P < 0.05). Other variables identified were not significantly related to outcome. It is noteworthy that the trend for the number of implants was of borderline significance (S-score = 8, Sens slope estimate 8.44, Mann-Kendall statistic = 1.959, P = 0.0500) and for the number of detox's was highly significant (S-score = 19, Sens slope estimate = 14.56, Mann-Kendall statistic 2.853, P = 0.0043) as illustrated for both variables in Figure 4.

Clearly however the central issue relates to relative importance of these various factors. A multivariate survival analysis was performed which demonstrated that social support, the number of detox's, work after treatment, type of implant, the number of implants and spirituality continue to be significant. When a similar analysis was performed limited to implant patients alone, social support, work prior to treatment and the presence of cannabis in the urine were significantly related to opiate free success.

This study has been performed as part of an ethically approved clinical audit of the opioid treatment program in this clinic and makes pertinent comparisons between rates of opiate free success in patients treated with implant and orally administered naltrexone as well as a variety of psychosocial and treatment parameters. Factors which were noted on bivariate analysis to be significant were in order were employment subsequent to treatment, the number of detox episodes, the level of social support, the number of implant episodes, whether implants or tablets were used, the use of multiple implants, the type of implant (including no implant), spirituality, the financial status of the family unit, the presence of morphine in the urine, the presence of naltrexone in the urine, negatively correlated with the presence of olanzepine in the urine and the semi-quantitative presence of naltrexone in the urine (Table 6). On multivariate regression the significant variables were the strength of social support, the number of detox episodes, work after treatment, the type of implant (including none), the use of multiple implant episodes, and spiritual belief. When the multiple regression was limited to implant patients alone the only variables of significance were social support, work prior to treatment and the presence of cannabis in the urine. In this group the number of implants was of borderline significance (P = 0.06).

With the increasing availability of long acting and depot preparations of antagonists onto the pharmaceutical market some guidelines for their application would be most useful. It must be noted in considering the outcome analysis of this study that the form of the present work was a retrospective chart review. Patients were not randomized and there were notable differences between the groups at enrolment in social support and financial status which were related to having to fund the cost of the treatment including implants. Given this important limitation there are a number of indications from this data which are useful for directing clinical practice, and may be tested in well designed randomized clinical trials. Strengths of this study included the relatively high overall follow up rate of 82%, the relatively large patient numbers reviewed, the breadth of treatment and psychosocial factors examined, and the various formulations of naltrexone used.

The major point of this study is the remarkable success of sustained implant therapy. 88% of the Perth implant patients and 78% of the USA implant patients were opiate free at 12 months. Our data clearly indicates that repeated episodes of detoxification and of implant insertion are important contributors to opiate free success and to the long term engagement with patients whose long term course is recognized to be marked by periodic relapse. This notable success rate suggests a potential alternative to other management streams for refractory opiate dependence such as rehabilitation where places are typically limited, and incarceration with all its attendant costs and problems. This latter option has been extensively discussed 5354.

We did not find a difference in the two types of implant used. Whilst this may be due to considerations of statistical power, the null hypothesis cannot be confidently excluded on the basis of the present data. It is noteworthy that the surgical procedure for implant insertion is straightforward. Since this study was completed the usual technique has changed in that rather than precipitate an acute detoxification syndrome under sedation using oral or parenteral antagonists as described here, a generally milder procedure is performed where the implant itself is used to detoxify the patient. As the serum levels achieved with the subcutaneous route are lower than those seen with alternative routes in well prepared patients, this technique is easier and much faster, and suited to ambulatory use.

We were pleased to note a 54% opiate free success rate for tablet administration which is higher than that reported from many other centres. However it should be noted that this figure includes supervised and assisted administration of the tablet by a significant other individual, opiate free success rates rather than simply retention only, and repeated episodes of detoxification and treatment initiation. Whilst this is a very different design from most naltrexone studies in the literature, it is nonetheless normal practice in both other addiction pharmacotherapies and most drug free rehabilitation programs. Careful involvement of the residual functional social network is an important buttress to pharmacological treatment, including supervision of the taking of the oral tablet, crushing the tablet and intensive medical supervision of minor side effects in the early post-treatment period, which all appear to be important to optimize results. The involvement of supervision might explain the classical rather paradoxical perception that naltrexone appears to work best for professionals and recent prisoners 55, which stereotypically represent the highest and lowest end of the social spectrum.

The literature experience with SSRI antidepressants appears mixed 5657. In the present project no useful effect was found. The presence of olanzepine in urine correlated negatively with opiate free success, which is not surprising given its use in psychotic and the most severe mood disorders. The elevated rate of amphetamine usage amongst the implant cohort is of concern, albeit not perhaps unexpected given the persistence of the ritual of intravenous drug use in a milieu in which opiate use is abruptly truncated. Unpublished data suggests that amphetamine use shows a steady decline after an initial surge after implant insertion (O'Neil personal communication) and our results would be consistent with this pattern. It is likely that one of the major triggers to stimulant use is the physical exhaustion which follows a an induced detoxification period but further work is required to delineate the natural history of such activity amongst implant patients, and this is understood to be an area of active investigation at this time.

The significant and dramatically increased rates of work involvement after treatment is gratifying. This is also of therapeutic importance in that it was shown to correlate with success in both univariate and multivariate regression. Privileged financial status (as defined in the Methods section) whilst having some bearing on success in a bivariate study, was shown not be significant when treatment success was controlled for social support and spirituality, which appeared to be the proximate underlying and determinative variables.

The prominence of spirituality in the statistical analysis of this study is of some interest. This area is increasingly becoming a focus of research interest in recent times 5859. Probably in common with most drug dependent cohorts, the major religious affiliation of these patients was "not involved/no belief". As this clinic had only four Muslim families and two Hindu in the present sample, the major religion or background of these patients was Christian which is in keeping with the major religious framework in Australia. The strong effects persisting on multivariate analysis are of interest. They may in part be related to the inclusion of a category of religious life practice and consistency in this rating in which drug use was a factor. The salience of this factor in this study is consistent with previous demonstrations of the centrality of an holistic understanding both of the psychopathology of addiction and of the road to recovery 606163, the acknowledged success of 12-step programs, with classic statements on the centrality of religion to character formation and development 64 and also with the enduring holistic definition of Health itself as proposed by WHO 65. Spiritual pursuits offer obvious mechanisms to cope with the some of the major issues of the psychological agenda for recovery including refuge and solace, peace, joy, coping with stress and guilt, values and virtue transformation, as well as an alternative source for the elusive paradise and esoteric pathway sought by many 66.

As mentioned above this paper does not discuss adverse outcomes which are addressed in a companion study. In general terms it can be said that this clinic has not encountered a mortality concern with these long acting implants in the present experience no patient has died related to causes which are attributable to the implant. A useful margin of safety however must form a pillar of successful adoption of a new treatment if it is to be widely deployed. The recent study of the Perth cohort by Hulse was noted above, and the overall mortality risk of naltrexone implant management was shown to be at least as low as that of methadone maintenance 44. A further and very reassuring data source of interest was a recent Australia wide study which examined all coronial reports in the whole country which mentioned naltrexone implants. Of the several thousand Perth implants which have been used in Australia, only a single case was identified where a naltrexone implant was plausibly temporally implicated in the mode of death in which metamphetamine and cocaine metabolites were also present and the level of naltrexone measured in the serum suggested a problematic assay technique.

There are several limitations of the above study. Some of these relate to the design limitations of a clinical audit review. These would include the non-randomized nature of the study, the use of historical controls, and the need of patients to provide both drug free carers to supervise their treatment, and to fund the costs of their treatment themselves. One limitation mentioned above was the smaller numbers of patients with Perth implants which may have lacked statistical power to define a difference between the two implant types. There are several indications that the experience reported herein is preliminary in that this clinic lacks associated services which may be found with larger programs such as counsellors, integrated Hepatitis C care, gynecological care 67 and assisted accommodation entry, all of which would ideally be part of a demonstration or pilot program, and may receive consideration for inclusion in a formal randomized trial.

A further physiological point of some interest is the general consideration of the health effects of naltrexone, which in general may be said to be opposite to those of opiates. The effect on tissue growth and cell division are of particular interest although they would appear to have received little research attention 68. In this regard the recent demonstration in four genome wide screens in three reports that the locus most strongly associated with coronary artery disease on the human genome encodes the two cyclin dependent kinases P16INK4A and P15INK4B is fascinating 697071. It is relevant to the present therapeutic study by virtue of the known potentiation of apoptosis 7273747576 and impairment of cell growth by addictive drugs 777879 which together suggest an effect to accelerate the ageing process, and by the demonstrated effect of naltrexone to reverse such effects on cell growth 80. Furthermore evidence of an acceleration of the atherogenesis by addictive drugs has been published in recent years from Johns Hopkins hospital 81, and the NIH intramural laboratories 82, from the National Drug and Alcohol Research Centre in Sydney 83 and has been confirmed in this clinical population (unpublished observations). Indeed biochemical, immune and haematological evidence for such a progeroid effect of addiction was recently published 84.

Evidence of quantitative circulating stem cell deficits in addiction 85, together with clinical evidence of stem cell deficiencies in hair 86, bone 87 and teeth 88 together with substantial impairment of body growth 89 have recently been reported. In the study of dental pathology 88 morphine and methadone dose were shown in multivariate regression to be significant correlates of the disorder of erosive periodontitis in which both immune and stem cell components are believed to play major roles. That naltrexone might reverse such an effect 9091, and that this may be demonstrable by assay of the classical molecules of cell growth 92and senescence 9394 in somatic and stem cell tissues is an intriguing possibility, well worth further investigation in view of the elevated rate of many pathologies in addiction 95 and the rapid age related accrual of hepatic, cardiac, pulmonary, renal and multi-system pathology amongst decedent IVDU's 83, not to mention the very high mortality associated with long term substance dependence 9697. A recent comment by the scientific leaders of addiction medicine that indeed the lead candidates involved in nicotine addiction were molecules involved in intracellular machinery is a further intriguing clue in this regard 98.

In summary, this study confirms extensive literature experience that both oral and implantable naltrexone can be used with gratifying results in a primary care setting. Open access to affordable treatment appears to be important to ensure maintained success rates. The observed 12 month opiate free success rate of 54% with tablets could possibly be raised to 82% by the use of subcutaneous implants. Whilst it seemed to the clinician involved that the longer lasting Perth implants may be superior to those which are presently in more widespread use, the present study evidently lacked sufficient statistical power to show this. Ready access to repeat treatment and particularly repeat depot implantation would appear to be key components of a program seriously endeavoring to deliver high quality outcomes, and would appear to have the capacity to dramatically raise the treatment potential and patient appeal of naltrexone therapeutics. It is important not to overlook likely benefits to health of such medium term opiate antagonism including immunostimulation and recovery of cellular growth and organismal dysfunction which is likely to have accrued over a period of drug dependency. Employment after treatment, cannabis use, and spiritual development were all shown to be significant. It would seem likely that in the future implants will become the route of administration of choice for naltrexone. Subsequent to the completion of successful trials in the USA and Perth, extension of these findings in opiate dependent populations in prospective randomized multicentred trials would appear to be the logical next step for research and development in this area. Long acting injections, and depot delivered antagonists for other drugs of addiction appear to potentially hold therapeutic promise and might prove to be the way of the future.

Dr. Reece reports that the naltrexone implants mentioned in this study were sold for use in patients from his clinic during the period of this study.

ASR conceived and designed the study, perfoemed the literature review, coordintaed the follow up, drafted the manuscript, perforemed the statistical analysis, and constituted the graphical illustrations. He has read and approved the final manuscript.