The past few decades have seen an explosion of research on opioid overdose, work that has dramatically recast our understanding of this phenomenon. Indeed, it is not stretching truth too far to opine that almost everything that we firmly believed to be the case regarding overdose was subsequently shown to be incorrect. As with many cherished beliefs, it was only in the harsh light of research that this was shown to be the case. Why should so much effort have been put into this field? Firstly, there is the sheer extent of mortality associated with the opioids. Of all commonly used drugs, opioids have by far the highest mortality risk 1-6. Indeed, by the age of 50, approximately half of any cohort of opioid users will have died, overdose being the most common cause 3-5, 7, 8. The harms associated with overdose, however, extend far beyond mortality: for every fatal overdose, we estimate that there are 25–50 ‘near misses’ 9-15. Apart from the immediate harm deriving from near misses (e.g. anoxic brain damage, injury through fall), a history of non-fatal overdose is a significant predictor of subsequent overdose and, more particularly, of a subsequent fatal overdose 4, 12. Despite the extensive body of work that has accrued, old myths have proven remarkably powerful, particularly in the media. In this piece, I will examine what was believed to be the case, what turned out to be true and how this has informed our ideas on overdose prevention. The literature on opioid overdose is now extensive and covers the circumstances, toxicology and organ pathology of fatal case series, laboratory studies of the physical effects of drug interactions, longitudinal cohort studies and cross-sectional structured interviews with opioid users on non-fatal overdose histories. In reviewing this literature, I will address four of what I regard as the most prominent myths that have bedevilled the field. Broadly speaking, these relate to misapprehensions as to who is dying and why they are dying: Myth 1. It is the young, inexperienced user who overdoses; Myth 2. It is variation in the purity of illicit opioids that is the major cause of overdose; Myth 3. It is the opioid that is crucial in overdose, not other drugs; and Myth 4. Impurities in illicit opioids are the major cause of overdose. When I commenced work on overdose in the early 1990s, it was a widely held opinion that it was the young, novice user who was most at risk. Indeed, the media representation of overdose cases would lead one to expect just this. This, of course, does make intuitive sense. We might well expect the younger user to be at greatest risk because of inexperience with dosing and lower tolerance to opioids. The literature, however, is quite clear and tells an entirely different story. Rather than being a young, inexperienced user, the typical overdose fatality is a long-term, dependent, drug-injecting user in their 30s or older 1-6, 16-18. Indeed, teenage decedents constitute a small minority of cases 16-18. Consistent with this, opioid users report their first non-fatal overdose as having occurred several years after first use 19. Recent years has also seen a major increase in the prescribing of opioids for chronic pain, most notably oxycodone, and a substantial increase in overdose fatalities attributable to these drugs 20-22. Again, these overdoses are most prominent among older groups, in this case typically aged in their 50s. Why should this be so? The natural history of heroin tolerance may partially explain the phenomenon. It has been suggested that tolerance to the hedonistic effects of opioids develops more rapidly than to the respiratory depressant effects, with the latter being an incomplete tolerance 23. As the use career progresses, the gap between the doses needed to obtain a hedonistic effect and the lethal dose decreases. Given this, it is the longer term user who would be most likely to suffer overdoses. Chronic pain patients are, by the nature of the disorders that cause such pain, more likely to be from an older demographic. Finally, we must also bear in mind the high levels of liver and heart disease seen among opioid users and older opioid users in particular 3, 24. It is known that the overall metabolic capacity of the liver is reduced with age because of factors, such as decreases in liver mass, hepatic enzyme activity and hepatic blood flow 25. In one study of opioid-related fatalities, we reported that a quarter of cases in their 40s or older had cirrhotic livers 25. Liver disease may have serious effects on hepatic metabolism. The ability of opioid users with serious liver disease to metabolise opioids must be questioned and may help explain the age demographics of overdose. Similarly, poor cardiac health among older opioid users (who are almost all heavy smokers) may well increase the risk of hypoxia-induced cardiac arrest and arrhythmia and contribute to the age demographics of toxicity cases. Whatever the reason for the demographic patterns of cases, the research informs us that the highest risk target group for intervention is the older user, not the neophyte. Curiously, a similar mythology is seen in the popular presentation of alcohol toxicity deaths as being of young people engaged in drinking binges, contests, etc. The reality is that the typical fatal alcohol toxicity case is a long-term, dependent drinker in their 40s and young deaths are rare 26-28. Perhaps the most potent myth regarding the causes of opioid overdose is that it is primarily due to variation in the purity of illicit opioids. This is the ‘killer heroin’ myth, so beloved on the media. Indeed, this belief is frequently used to support arguments for the legalisation of illicit opioids 29, 30. The very term ‘overdose’ (the consumption of a quantity, or purity, of a drug that is in excess of the person's tolerance) implies that this would indeed be the case although, of course, this would not apply to overdoses of pharmaceutical opioids. Despite the intuitive appeal of this view, evidence to support it is lacking. Two important points need to be made here. The first was made in the 1970s by Edward Brescher 31. He asked why so many overdoses appeared to be what he termed ‘underdoses’. The evidence since has remained consistent: large proportions of fatal heroin overdose cases have low blood morphine (the major metabolite of heroin) concentrations, in many cases below accepted toxic levels 16-18. Moreover, blood morphine concentrations in fatal cases are frequently below those of intoxicated heroin users, or users who died due to causes other than drug toxicity 32-34. The second point is that, contrary to expectation, fluctuations in heroin purity have at best only a modest relationship to the incidence of heroin-related death. Again, this has been documented across decades and countries. Desmond et al. 35 in the 1970s reported no correlation between heroin potency and overdose fatalities. Studies from the 1980s and 1990s found variations in heroin purity accounted for only a quarter of the variance in overdose fatalities 36, 37. Risser et al 38 in 2000 found no relationship between the purity of heroin seizures and heroin-related death. Myth 3 may be considered a corollary of the purity myth, to which it owes its longevity. I remember well in the early 1990s being confidently informed that alcohol and other drugs had nothing at all to do with opioid overdose. If the preceding decades of research have proven anything, it is that polydrug toxicity is the major factor in opioid overdose. In retrospect, it would have been extraordinary if the concomitant use of other central nervous system (CNS) depressants did not increase the risk of overdose. Any anaesthetist worth his or her salt could have told us this years ago. What does the research from around the world consistently tell us? It tells us that the overwhelming majority of opioid overdoses, both fatal and non-fatal, involve multiple CNS depressants, most notably alcohol and benzodiazepines 11-14, 16-18, 39, 40. Why does such concomitant use matter so greatly? Because the co-administration of other depressants substantially increases the combined likelihood of respiratory depression and death. Thus, we see a negative correlation between blood alcohol and morphine concentrations, suggesting a functional reduction in opioid tolerance in the presence of alcohol 39, 40. Similarly, benzodiazepines substantially decrease oxygen saturation when consumed in conjunction with opioids 41, 42. Importantly, we see this familiar pattern of polydrug toxicity in overdoses involving opioids other than heroin, including methadone, buprenorphine and oxycodone 17, 21, 43-45. Chronic pain patients are frequently also prescribed with benzodiazepines (e.g. diazepam for back spasm) and drink heavily. The new wave of pharmaceutical overdoses, whether among drug-injecting users or chronic pain patients, is presenting with the same polydrug toxicity that we have seen in previous overdose epidemics. The somewhat puzzling toxicology of low morphine concentrations in heroin overdose may well be explicable by the presence of other CNS depressants. What termed ‘opioid’ overdose are, in reality, multiple drug toxicity deaths. In such cases, the concentration of each individual substance may well be sub-toxic. While no one drug consumed alone would cause respiratory failure, the combined effects of multiple depressants may well do so. In my opinion, the evidence is cogent: concomitant drug use, not drug purity or impurities, is the primary factor in overdose. While Myths 2 and 3 emphasise the role of the opioid itself, Myth 4 downplays the drug itself and emphasises dangerous adulterants. As with the purity myth, the putative role of impurities in causing overdoses has been used as a basis for arguments supporting the legalisation of illicit opioids 29, 30. Despite its longevity as a myth, there is absolutely no evidence to suggest that contaminants play any significant role in overdose. The evidence from toxicological analyses of blood, drugs and used syringes is clear: harmful contaminants are rarely detected 16-18. If contaminants are an important vector for what are termed overdoses, we should most certainty be seeing them frequently among overdose cases. Indeed, when contaminants are detected, they are typically innocuous substances, such as caffeine and sucrose 46-48. In retrospect, this should not be surprising. At the most prosaic level, for a dealer, killing one's customers is never good business, and a bag of sugar for cutting the drug is cheap and easily available. Why search for dangerous adulterants? If we have learnt one thing over the past 25 years, it is that contaminants play little, if any, role in opioid overdose and absolutely no role whatsoever in pharmaceutical opioid overdose—an increasingly common event. This is important because if overdoses were primarily due to adulterants, their occurrence would be close to random and targeted interventions close to impossible. When I commenced working in the field, the general picture of overdose was of a young, inexperienced user who dies due to inexperience, fluctuations in purity and/or contaminants. The picture that emerged, however, was of an older, very experienced polydrug user who dies from multiple drug toxicity. Does any of this matter? In terms of addressing overdoses, very much indeed. Previously, overdoses were regarded as being close to random. The most important implication to emerge regarding overdose in the past few decades is that overdose is not a random, unpredictable event. If this were the case, then any overdose intervention would be doomed. Has the research from the preceding decades indicated anything else that might inform our ability to intervene to reduce overdose mortality and morbidity, in light of the above? The answer to this question is a resounding yes and relates to treatment, responses to overdose, high risk periods and polydrug use. Possibly the most important finding in drug treatment research in the past few decades has been that drug treatment substantially reduces mortality rates, primarily because of reductions in overdose 3, 5, 8, 12, 13, 43-45, 49-51. The demography of overdose is consistent with this, as heroin overdose cases are rarely in drug treatment at the time of fatal or non-fatal overdose 12, 13, 17-19. Indeed, even in cases of methadone or buprenorphine toxicity, half were not enrolled in treatment at the time of death 17, 52-57. Why does treatment make such an impact on overdose? Firstly, the use of opioids declines substantially while in treatment. Secondly, and equally important, the use of other CNS depressants also declines 58. There is simply less overdose risk exposure while enrolled. I should note that by ‘treatment’, I am referring specifically to long-term, stable treatment, such as residential rehabilitation or opioid substitution. Detoxification, in and of itself, does not constitute a treatment and, without subsequent treatment enrolment, may increase the risk of overdose (see below). Overall, the single biggest impact we can make on overdose rates is to have larger proportions of opioid users enrolled in long-term treatment programmes. Given their high overdose risk, older users are a particular target group for treatment engagement. Despite their widespread experience with overdose, responses by opioid users are particularly poor 59-63. Calling an ambulance is rarely the first thing done and is frequently avoided altogether. This is one area of overdose research that was not surprising. The reasons for this tardiness are prosaic: they fear police involvement. It has been shown that it is possible to improve the responses of users at overdose if health-care providers and the police cooperate on emergency call-outs, such that the police only attend where there is violence or death 60. It must be borne in mind that any delay not only increases the risk of death but also of hypoxic brain damage. Given the reluctance to call for help, a number of authors (including this one) have called for the distribution of naloxone hydrochloride (the narcotic antagonist used to reverse acute narcosis) directly to heroin users 62-65. Indeed, naloxone has been available as an over-the-counter drug in Italy since 1995, and pilot distribution schemes have been conducted in several countries 58, 65-69. All studies to date indicate that opioid users will take home naloxone, that they will use it appropriately and that adverse effects do not appear to be a problem. Given the epidemic of pharmaceutical opioid toxicity deaths seen in recent years, similar consideration should be given to the provision of naloxone to patients being prescribed with opioids for chronic pain management. As noted above, most overdoses are cases of polydrug toxicity among people with high levels of opioid tolerance. In recent years, research has emerged that clearly delineates very high risk periods for overdose in which tolerance does play a major role. The first of these is the period immediately following detoxification 3, 14, 70-72. Similarly, the two-week period post-prison release is also associated with a substantially elevated risk 13, 14, 70, 73-78. In both these situations, relapse to opioid use is extremely risky and, given what we now know, ideally such people would be channelled into a long-term drug treatment programme. The importance of polydrug toxicity as the cause of ‘opioid’ overdoses has been repeatedly emphasised throughout this piece. If we believe that concomitant CNS drug use is high-risk drug use, can we change these behaviours, particularly among those not enrolled in treatment? Educating opioid users on the risks of polydrug use may help reduce the frequency of overdose as may education on the risk of overdosing after periods of abstinence. On the positive side, campaigns to increase such awareness have been conducted in a number of countries, and it has been shown that awareness of risk factors can indeed be raised 6, 65. While we may have increased awareness of the risks of polydrug use, I am unaware of any evidence to suggest that we have successfully reduced such use patterns outside of the treatment setting. Furthermore, these campaigns are directed at illicit drug users, most commonly injectors, and not to the older chronic pain patient who may have no history of illicit drug use but is at risk of opioid toxicity. Again, it is the older user who is most at risk and is a priority target. The past few decades have shed considerable light upon the actuality of overdose. Much of what we believed to be the case was false. The research has highlighted who is dying, given clues as to why and provided information on how we may more effectively respond. The typical overdose case is an older, experienced opioid user, who is not enrolled in a drug treatment programme and who is an extensive polydrug user. It is on the basis of such facts that intervention must proceed. Drug use patterns, however, are never static and the opioids are no exception. New trends in the demography of opioid use are emerging, most prominently the use of pharmaceutical opioids such as oxycodone, which will require new interventions. At least we now know these events are not random. The fact that we can now identify who is likely to overdose and why substantially increases our chances of successful intervention to reduce the considerable harm attributable to overdose. In terms of overdose, the past few decades of research have been crucial. The National Drug and Alcohol Research Centre at the University of New South Wales is supported by funding from the Australian Government.
No takes yet. Share an insight, caveat, or question.
Shane Darke (2014) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: