Location,location,location: The variable geography of opioid use and misuse

Prior to the appearance of the COVID epidemic pandemic, the ‘opioid epidemic’ was the ‘public health emergency’ prominent in healthcare, law enforcement and social issues. Both of these epidemics have challenged our specialties of anaesthesia, intensive care and pain medicine. Both continue, modified by time, geography, and personal behaviours.

A long time has passed since the beginning of the opioid epidemic, during which much research has focused on all aspects of opioids and pain relief, alongside a shift in the understanding of and language surrounding what are now called substance use disorders (SUDs). Much of what we have learned is about risk evaluation and management, together with situations in which an opioid is unlikely to produce a benefit for the patient. ¹ Predictably, prescriber behaviour takes a long time to follow new knowledge. While turning off what was started remains difficult, the landscape has altered somewhat from the abuse of prescription medications to recreational designer drugs; for example, non-medicinal fentanyl analogues. There is a large geographical variation in the use patterns of these. This paper is centred around opioids prescribed with the aim of pain alleviation, rather than those in substitution programmes or non-medical drugs.

‘Opioid’ has diminished its meaning as an objective pharmacological descriptor, to become a word now often perceived as a descriptor for misdemeanours and/or illegality by any party: prescribers, manufacturers, recipients, and more. A highly emotional language ² has developed in both professional and lay media, with terms such as ‘opioid crisis’, ‘prescription opioid abuse’, ‘pain killers’ and even ‘civil war’. Objectively, opioids are a class of agents which interact with a series of biological receptors and pathways involved in an intricate system which modifies responses to noxious inputs. The utility of opioids comes to the fore in perioperative situations; it is no secret that they typically relieve severe acute pain, which is usually short lived. Another property is that tolerance with declining efficacy and dependency will inevitably develop if continued longer term, with significant potential harm if control is lost. Opioids are often referred to as if a homogeneous group, but members of the class include many agents, which exhibit vast differences – for example, metabolites.

There were triggers for the major swing towards over-liberalising opioid prescribing to treat pain. ‘Addiction rare in patients treated with narcotics’ was stated in the title of the first of two letters to editors published in high-impact journals in 1980 ³ and 1986, ⁴ from respected institutions and clinicians. The second letter came from palliative care authors, but its title ‘Chronic use of opioid analgesics in non-malignant pain: report of 38 cases’ shifted into non-cancer pain observations, albeit in small numbers. Ironically, 36 years later a further letter to the editor to one of those journals was published, but this time with collated data to illustrate errors and over-interpretation in papers subsequently citing them, and demonstrating over-adherence to a belief in the words of the title. ⁵

We should not forget that clinicians’ good intentions to relieve pain, in the absence of enough factual information, drove their over-enthusiastic supply of opioids. These coupled with aggressive pharmaceutical company marketing and even fraudulent misrepresentation of facts conspired to accelerate longer-term opioid prescribing. Now there are concerns being raised about the shift of balance in response to the ‘opioid crisis’, from overuse towards undertreatment of acute pain in particular, ⁶ and in directions such as opioid-free anaesthesia. Is this swing of the pendulum too much in the face of a highly developed endogenous opioidergic system modifying the biological organism’s response to noxious inputs?

There was an earlier history of promoting opioids, through the Single Convention on Narcotic Drugs (1961) ⁷ adopted by United Nations member countries. It contained the duality of ‘combat the spread of the illicit use of drugs’ alongside declaring that member countries should recognise ‘that the medical use of narcotic drugs continues to be indispensable for the relief of pain and suffering and that adequate provision must be made to ensure the availability of narcotic drugs for such purposes’. The World Health Organization (WHO) recognised opioid analgesics as essential medicines then, and still does. ⁸ The International Narcotics Control Board in 1995 put member countries on notice of anticipated increasing global medical need. ⁹ This early recognition of global need has hardly been a success, as in 2015–2017 approximately 90% of controlled opioid use was by only 10% of the world’s population, mainly in North America and Europe. ¹⁰

Although yearly United States (US) opioid sales began to fall around 2011, and by approximately 13% since 2015, global sales have increased and geographical imbalance remains.^11,12 Significantly, the US opioid death toll has accelerated despite that, unconnected with medicinal opioid sales anymore. The opioid epidemic is international, but arguably the USA is the most severely affected country, with its very different healthcare and regulatory systems compared with Australia and New Zealand (NZ). Different jurisdictions legislate and regulate vastly differing limits on quantities and durations for opioid prescription dispensing, and differ in whether real time monitoring by prescribers is possible. In the sad US statistics, it is noteworthy that in a 2006 survey more than half the persons who used prescription opioids non-medically in the previous 12 months did not have a personal opioid prescription. ¹³

Throughout the above-mentioned good intentions to alleviate pain, there was no stratification of pain severity was ever specified as to when opioids were considered justified. The 1986 WHO analgesic ladder ¹⁴ laid out a schema for escalating analgesic measures, but was specifically about cancer pain (i.e. context) and arguably for lower resourced locations because advanced analgesic techniques did not feature. Some depictions showed ‘freedom from cancer pain’ as the outcome after opioid therapy commenced. ¹⁵ No actual severity score was specified for each step, just a stepwise change of analgesic measures if previous measures had failed, with step three being an opioid. But what had earlier measures failed to do? What is reasonable pain relief? This question will be returned to.

When pain was declared the ‘fifth vital sign’,^16,17 and the relief of pain declared a fundamental human right (Declaration of Montreal, 2010),^18,19 these brought an expectation of relieving or eliminating it, and clinician reaction shifted to more opioid administration for lower pain severities in both acute and persistent pain conditions. However, by the time of the Declaration of Montreal the epidemic in opioid deaths was already well established, and close to the turning point for a downwards trend. The blunt unidimensional number rating scale (NRS11) with 11 discrete ordinal number choices 0–10 became the norm and, in many places, effectively a requirement, yet with little understanding of its psychometric properties. ²⁰ There were even punitive actions taken against some practitioners for failing to treat pain, which put further perceived pressure on the rest. It was not widely appreciated that large proportions of populations with persistent pain function quite adequately with pain scores in the lower end of that scale, and there is no convincing evidence of ratio properties inherent in that scale. Knowing what constitutes a clinically important change on the unidimensional NRS11 is difficult, given it has never been defined, particularly with persisting pain. What followed was a non-critical approach, which simply accepted the statement in the title of one of the above trigger papers, namely ‘Addiction rare in patients treated with narcotics’, without data. Sluggishly, over around 30 years since then we gained a refinement in language and severity classifications for dependency as SUDs, which also apply to other dependency-producing agents such as alcohol and tobacco.

A long-term conflict existed over the differentiation between substance dependence and substance abuse. It took six decades (1952–2013) to evolve from the pejorative nature of earlier addiction terminology to the latest (Diagnostic and Statistical Manual of Mental Disorders, version 5), ²¹ which focuses on 11 criteria (symptoms) that characterise the severity of an individual’s SUD. Although more consumer-oriented, a simplified resource illustrating this subject is available for the interested reader. ²²

Continuing the theme of research into opioids, this edition of the journal includes a pair of papers reporting research observations from the same research group. The authors studied the prevalence of opioid intake in the cohort as they entered orthopaedic surgery ²³ (predominantly major joint replacements), and again in the same cohort 90 days following surgery. ²⁴ A major observation in both papers is the geographical variation in the prevalence of opioid use.

A problem with published results is whether they are generalisable, or whether they apply only to a discrete population, location, and context. Both Australia and NZ have similar atlases of health care variation,^25,26 with evidence of wide variation in opioid prescribing across locations. However, in both countries’ versions, the data are ‘blunt’, because they measure only the number of prescriptions filled without indicating quantity or duration for each script. The new papers in this issue report data from only one part of Australia (New South Wales), and different locations within it. The authors wisely recommend caution in ‘understanding the local context’. These papers add further evidence of geographical variation, and discuss some possible explanations. Not discussed, however, is the fact that of necessity each prescription required a prescriber, who of course contributes to this variation. We remain in the dark about the attitudes and other drivers for prescriber behaviour. Is it the place or the people there?

There are a number of further interesting elements to examine in the two papers in this issue. Looking at the differences between those not taking opioids and those taking them, they report a mean difference of only one using the NRS11 scale (4.9 vs. 5.9; median scores might have been more meaningful), and despite statistical significance reported this is clinically insignificant. It means that despite taking opioids, there was no clinically significant difference between those taking or not taking them for this cohort and location. These authors also found that 15% of their followed-up cohort continued to take opioids at 90 days, which they called ‘long-term opioid use’. The commonest opioid was codeine; much less common was tramadol, with the most potent oxycodone sitting in between. Given the wide variety and more potent opioids available in Australia, does this already demonstrate prescriber restraint?

The authors make no judgement about the ‘badness’ of their finding other than recommending to understand the local context so as to ‘facilitate improved management of pain… in orthopaedic surgery’. Does that imply that they expect such surgery will always produce a no-pain result, without further analgesic responses? Conversely, there is no positive acknowledgement of the 85% who did not continue to take opioids; they do report none meeting high-risk severe SUD criteria. Surely it is reassuring to know that the large majority did not continue taking opioids? This contrasts with 41% using an opioid after total knee joint replacement in a US retrospective study ²⁷ quoted by these authors.

Going back full circle to the 1980s and the liberal opioid uptake for non-cancer pain in the belief it was safe, trying to meet perceived demand while believing that NRS11 pain scores could actually be brought to near zero, we have to ask what is enough pain relief? If we now know from studies that opioids provided no benefit over placebo in chronic osteoarthritis hip or knee pain, how long will education take to correct prescriber behaviours? The unidimensional NRS11 both then and now is not enough to inform this. Bedside behavioural observation of movement and ability to engage freely in conversation are possibly more informative, just harder to document.

Although the 1980 triggering letter did not refer to cancer pain, it extrapolated from opioid use in cancer pain to non-cancer pain. ³ Around that time this writer recalls that intractable pain in near-terminal cancer cases was appallingly high before opioid administration was considered justified. Having cancer-related pain in the 1980s, with a relatively short survival expectancy, was usually considered an ‘honourable’ justification for opioid use. But now we see pain in patients who had cancer but are ‘cured’ and survive for decades. Cancer surgery can leave a burden of non-opioid-responsive neuropathic pain as its legacy. So is a cancer label still justification for liberal ‘prn’ opioid therapy? No! Pain is not a single four-letter word entity. We know both non-pharmacologic and non-opioid strategies can bring relief, but not relief synonymous with a zero pain score. Many non-opioid measures are time consuming and expensive. Many are not accessible because they are poorly or not funded in many jurisdictions, thereby biasing towards the next pill prescription.

It is comparatively recently that clinicians grasped the differences between opioids in acute short-term pain and long-term pain. A popular but arbitrary distinction exists between justification for opioids in cancer pain compared with persistent non-cancer pain. Although both could involve long-term exposures, dependency in cancer pain was mostly regarded as inconsequential. Bearing in mind long-term cancer survivals, this distinction demands rethinking.

Fortunately, current research leaves us better informed about outcomes from opioid prescribing than existed before the ‘opioid crisis’ developed. Whenever we read such reports, it is important to ask how generalisable are the conclusions, and ‘do they apply in our location?’ Our professions are taking up opioid stewardship in pain management, including time-limited supplies, step-down tapering and having an exit plan if high severity SUD features emerge. An infographic time-line illustrates the crossover between declining US prescription opioids being replaced by deaths from illicit substances. ²⁸ We need to replace emotion with useful research questions, accurate research reporting and careful peer review before publication, education and advocacy, these being our challenges in continuing to reverse the 30 plus years of rising prescription opioid overdoses and deaths. Is the US downwards trend for methadone deaths in recent years one indication of increased stewardship there? This writer suspects a solution to the ‘third wave’ of the accelerating US death rate from illicit designer opioids, ²⁹ made mostly elsewhere, is almost certainly outside our clinical prescribing behaviours for pain.

A thoughtful summary which appeared in a 2014 legal treatise helps us conceptualise our future path in managing pain. ³⁰ Although researched in a NZ context, it was non-judgemental, and could equally transcend most geographical jurisdictions. It concluded: ‘A right to pain relief must be limited to the practical realities of current medical knowledge. This does not mean we should accept severe pain as something that is as certain as death and taxes, but practitioners and patients must continually seek to provide and receive adequate pain relief. This is a humanitarian obligation.’ Note particularly the terms ‘severe pain’ and ‘adequate pain relief'.