Executive Summary

Background: Naltrexone is an opiate antagonist, meaning that it blocks the actions of opioids such as heroin in the body. It is registered in Australia for use as an oral maintenance treatment for heroin dependence, and tablets are administered on a daily basis. Provided a person continues to take naltrexone daily, the effects of any opioids taken are blocked or substantially reduced.

However, one of the effects of naltrexone is that it effectively removes tolerance to opioid drugs. Tolerance is the effect that occurs after repeated administrations of a drug, and results in a given dose of a drug having less effect than previously, or a higher dose of a drug being required to get the same effect. Because naltrexone blocks the actions of opioids, naltrexone rapidly removes a person’s tolerance to opioids so that a given dose of opioids would have more effect than previously. The lack of naltrexone, not its presence, exposes a naltrexone-maintained patient to risk of opioid overdose. If naltrexone treatment is ceased, individuals may be at risk of opioid overdose if they choose to return to opioid use. There have been no attempts to quantify the mortality rate associated with naltrexone treatment.

Naltrexone implants are an unregistered form of naltrexone treatment which can be accessed through the Therapeutic Goods Administration “Special Access Scheme”. Their efficacy in the treatment of opioid dependence is yet to be supported by randomised controlled trial evidence and they have been associated with death and other serious adverse events, both in Australia and overseas.

Buprenorphine and methadone are also used as maintenance pharmacotherapies in the treatment of opioid dependence. These drugs have opioid agonist actions (as do heroin and morphine). These treatments have also been associated with death, primarily in the period soon after treatment commences. These deaths are similar to deaths from opioid overdose, in that methadone or buprenorphine (together with any other depressant drugs that may be present) cause death primarily by respiratory failure or the complications that develop in a coma related to central nervous system depression.

It is important to note at the outset that naltrexone-related deaths are more difficult to monitor than deaths associated with either buprenorphine or methadone. Not only is naltrexone typically not detected at autopsy, but coroners, police and medical professionals are also unlikely to be aware of the relevance of (or be informed about) a recently terminated episode of naltrexone treatment in the investigation of the circumstances surrounding a death. This report is a first attempt to quantify the mortality associated with oral naltrexone treatment for heroin dependence in Australia. It also compares the resulting rates with those associated with buprenorphine and methadone treatment.

It should be noted that mortality rates associated with implanted naltrexone treatment were not able to be estimated in this study due to lack of national data on the numbers of patients receiving naltrexone implant treatment. As a result, our picture of naltrexone-related death in Australia remains incomplete. We recommend that this lack of data be addressed.

Results:

Oral naltrexone-related deaths
Searches of the National Coronial Information System (NCIS) revealed 32 deaths related to the use of oral naltrexone in the period 2000-2003 in Australia. This number is an underestimate since the majority of known naltrexone-related deaths in Western Australia (WA) and Queensland (QLD) were not detected in our searches.

When expressed as deaths per number of treatment episodes, it was estimated that naltrexone had a mortality rate of 10.1 per 1000 treatment episodes. If the mean treatment retention in naltrexone treatment was estimated at 3 months (rather than two months, as assumed in the above estimate), the mortality rate for naltrexone treatment increased to 15.2 deaths per 1000 treatment episodes.

Naltrexone was associated with a mortality rate of 22.1 per 100 person years during the period of high risk (2 weeks post-treatment), and 1 per 100 person years during the period of low risk (during treatment).

While we did not specifically search for deaths related to naltrexone implants, two fatal cases were identified in the search period, one in Western Australia and one in Queensland. These cases were not included in the above naltrexone mortality rates.

Comparison with methadone- and buprenorphine-related deaths

NCIS searches revealed 1 buprenorphine-related death and 282 methadone-related deaths during the same time frame.

The mortality rate for naltrexone was four times higher than for methadone when calculated as deaths per number of episodes of treatment, and substantially higher than for buprenorphine. The estimated mortality rate was 0.02 per 1000 treatment episodes for buprenorphine and 2.7 per 1000 episodes for methadone.

When considering deaths per periods of high and low risk, the mortality related to naltrexone was approximately seven times that of methadone during the period of high risk and three times the rate during the period of low risk. Naltrexone treatment was associated with a mortality rate of 22.1 per 100 person years during the period of high risk (two weeks following treatment cessation) and 1 per 100 person years during the period of low risk (during treatment). Buprenorphine mortality rates were not expressed in terms of periods of high and low risk due to the low number of deaths detected with this search method.

Conclusions: Deaths related to oral naltrexone maintenance treatment have occurred in Australia. However, this study found that identifying naltrexone-related death was difficult, and it will remain so as long as coronial databases do not systematically receive and record treatment data in a detailed fashion. The estimates produced in this study are underestimates, since a significant number of known naltrexone deaths reported elsewhere were not detected in our NCIS searches. Because naltrexone-related deaths are not captured in a systematic way, consideration of our results must take into account the various assumptions made and their potential to bias estimates of mortality.

This study also found that the mortality related to oral naltrexone treatment was higher than that for buprenorphine and methadone, two of the most common forms of pharmacotherapy for opioid dependence in this country. Deaths were also related to buprenorphine and methadone treatment, but whether estimated as deaths per 1000 treatment episodes or per 100 person years of risk, the death rate for naltrexone was higher and we believe the estimate provided here is a conservative one.

These mortality rates are plausible given the pharmacology of these drugs. Naltrexone is a treatment that provides blockade of opioid effects during treatment and a sudden reduction in tolerance to all opioids. Buprenorphine and methadone, in contrast, provide tolerance to all other opioids during treatment. It is not surprising, then, that there is a higher potential for more deaths to occur post-treatment in the case of naltrexone. It is also not surprising that more deaths occur during treatment induction with buprenorphine and methadone (where opioid levels are rising), than in naltrexone (where there is an opioid blockade in place).

The mortality rates suggest that oral naltrexone treatment, as it is provided in Australia, can place recipients at significant risk of death, and at higher risk than buprenorphine and methadone. However, it should be noted that naltrexone treatment is a useful option in some well-motivated patient subgroups that form a minority of the opioid-dependent population.

Implant technologies have been proposed as alternative methods for delivering naltrexone. A number of potential issues also relate to this form of treatment, and rigorous research is certainly required to carefully examine the potential for this delivery system to represent a viable treatment option for opioid-dependent persons. Specifically, these issues are: the lack of randomised controlled trial evidence of naltrexone implant efficacy in the treatment of opioid dependence; considerable inter and intra-subject variability in the blood levels of naltrexone resulting from an implant (and so the level of opioid blockade); the lack of good monitoring of adverse events relating to the use of naltrexone implants; and the acceptability of the naltrexone implant preparation to patients and medical professionals. Due to lack of data on the number of people receiving naltrexone implants, this study was unable to include naltrexone implant deaths in estimates of naltrexone-related mortality. Our incidental discovery of two deaths related to naltrexone implants suggests that this formulation of naltrexone also carries with it a mortality risk. The current inability to measure naltrexone implant-related death is an issue that needs to be investigated as a matter of priority.

In comparing mortality rates associated with these pharmacotherapies, it is important to draw the reader’s attention to the rates of mortality for active heroin users. It has been estimated that mortality rates for heroin-dependent persons not in treatment are in the vicinity of 0.9 per 100 person years of risk, very similar to the mortality rate of a person in naltrexone treatment (during the period of low risk) calculated in this study. While maintained in methadone or buprenorphine treatment after the initial induction stages, opioid-dependent people are at lower risk of dying. Clearly, an important aspect of methadone and buprenorphine treatment for opioid dependence is the improvement of treatment retention rates.

The mortality risks associated with oral naltrexone treatment, particularly following treatment cessation, warrant serious attention. This is especially the case considering that the majority of unselected opioid-dependent persons will return to opioid use soon after leaving naltrexone treatment. It is recommended that future trials of all treatments for opioid dependence include monitoring of post-treatment mortality risk, as is estimating the rate of naltrexone implant-related mortality. In order to more effectively monitor the use of this drug for the treatment of opioid dependence, and because of the risk of mortality, it may be appropriate to consider naltrexone for scheduling.

Citation: Gibson, A. and Degenhardt, L. (2005) Mortality related to naltrexone in the treatment of opioid dependence: A comparative analysis, Sydney: National Drug and Alcohol Research Centre.