Concurrent use of opioids and stimulants and risk of fatal overdose: A cohort study

Illicit drug toxicity (i.e. overdose) death has been on the rise in recent years in North America. In Canada there were 26,690 opioid toxicity deaths between Jan 2016 and Sep 2021, and more than half (58%) of these deaths in 2021 also involved a stimulant [1]. Furthermore, among all stimulant toxicity deaths in 2021, nearly 90% also involved an opioid. These data reflect a population-level trend toward increasing concurrent use of opioids and stimulants. Though the public health emergency of drug poisoning deaths (overdose crisis) has impacted most provinces and territories in Canada, British Columbia (Canada’s third most populous province) has consistently reported the highest rates of illicit drug toxicity deaths in the country, more than double the national average in 2021 [1].

Provincial death data reveal an increasing prevalence of methamphetamine alongside opioids in toxicity records. For example, methamphetamine was detected in only 14% of deaths in 2012 however between 2019 and 2021 it was detected in approximately 42% of deaths [2].While the detection of cocaine in drug toxicity deaths has generally been declining in recent years, it remained involved in 44% of illicit drug toxicity deaths in 2020 [2]. This trend has also been detected among people accessing harm reduction sites across British Columbia (BC), where methamphetamine is the most commonly reported substance used among more than 70% of respondents [3]. This analysis also revealed that people who reported opioid use had three times the odds of concurrent methamphetamine use compared to people who did not report using opioids [3]. These patterns have also been observed outside BC [4]. For example, a recent national study of people with opioid use disorder (OUD) in the United States revealed a significant rise in reported methamphetamine use from 18.8% in 2011 to 34.2% in 2017 [5]. This rise in stimulant use among people who use opioids in North America has been referred to as “twin epidemics” [5].

While evidence-based treatments for people with OUD are available to reduce risk of illicit drug toxicity events, including opioid agonist treatment (OAT) [6], engagement in this treatment is often limited among people who use stimulants [7]. For example, prior studies have found that people who concurrently use cocaine or methamphetamine alongside opioids have reduced retention rates in OAT compared to people with OUD alone [8]. Given the protective effect of OAT on overdose, where stimulant use interferes with OAT engagement, overdose risk may be elevated [7, 9]. Furthermore, studies have suggested that the concurrent use of opioids and stimulants is common among people who access harm reduction sites in BC [10]. Motivations for co-use have been described in a variety of studies, and have ranged from social influences, to seeking to reduce opioid withdrawal, to improving functionality, energy, or wakefulness [11‐13]. One study has also revealed a dangerous misperception among people who co-use opioids and stimulants, that stimulants can be protective against opioid overdose [14]. This is untrue, in fact, studies have identified that polysubstance use elevates overdose risk especially where the quality and potency of the substance is unknown [15].

Evidence is emerging to support the safety and effectiveness of pharmacological treatments for stimulant use disorder, however their implementation has been limited, often leaving people who use both opioids and stimulants disconnected from care [16]. This is particularly concerning considering people who use both opioids and stimulants are also known to face an elevated burden of concurrent chronic health conditions [17], necessitating access to care in order to avoid premature morbidity and mortality, including from illicit drug toxicity (overdose).

Given recent trends indicating rising stimulant use at a population level, and in particular, its increasing detection in illicit drug toxicity deaths alongside fentanyl in BC, there is a need to investigate the association of stimulant use with illicit drug toxicity deaths. As such, this study aims to estimate the effect of stimulant use, on its own and with opioid use on risk of fatal overdose in BC.

The sample included 7460 people including people in the stimulants only (38.0%), opioids only (38.9%), or both (23.1%) groups between January 1, 2010 and December 31, 2018. There were significant differences in demographic, comorbidity, and medication access characteristics by substance use type. There were significantly more males among people who used stimulants (64.2%) compared to opioids (58.6%) and to people who used both (60.5%). There was a relatively consistent age distribution across substance use types, while people in the stimulant use only group were overrepresented among the < 30 age group and people in the opioid use only group were overrepresented among the 50+ age group. People who used both opioids and stimulants were more likely to have 3 or more comorbidities. People who used opioids only were more likely to receive benzodiazepines, z drugs, sedatives, and pain medications, as compared to people who used stimulants only, and people who use both opioids and stimulants (Table 1).

During follow-up, 272 people died of overdose (illicit drug toxicity death); 40% were in the opioid and stimulant group, 32% were in stimulants only group, and 27% were in the opioids only group. Approximately 70% of overall drug toxicity deaths were among males while 75% of the deaths in the stimulant use only group were in males. Among people who died, the presence of comorbidities was highest in people in the opioid and stimulant group. A higher proportion of people in the opioid only group had dispensations for sedatives and benzodiazepines (Table 2).

Data were missing for variables of interest among 2.5% (N = 185) of all participants (See supplement). The final cox proportional hazards model was run on the sample with no missing data, and revealed that people in the opioid and stimulant group had more than twice the hazard of fatal overdose (HR: 2.02, 95% CI: 1.47-2.78, p < 0.001) compared to people in the opioid only group, while there was no significant difference in the hazard for people in the stimulant only group compared to the opioid only group (HR: 1.05, 95%CI: 0.75-1.48, p = 0.7644). Females had approximately half the hazard of fatal overdose compared to males (HR: 0.53, 95%CI: 0.40-0.69, p < 0.001). After adjusting for all other variables, there were no significant differences in the hazard of fatal overdose by age, health authority, comorbidity index, or prescribed medications (Table 3).

The analysis of survival probability revealed a significant difference by substance use type (p-value < 0.001). The levels of survival were similar up to two years of follow-up across substance use types, with an increased risk of fatal overdose for participants in the opioid and stimulant use group over time (Fig. 1).

The risk of fatal overdose steadily increased from 2015 to 2018 for people in the opioid and stimulant use group (Table 4). A sensitivity analysis was conducted using a series of cox proportional hazards models with different combinations of covariates and the results remained consistent with those presented in Table 3. OAT is known to have a protective effect on overdose [27] therefore an additional sensitivity analysis was run to estimate the impact of OAT (in the prior 5 years, prior 30 days, and on day of death) on fatal overdose using an adjusted cox model with the same covariates outlined in Table 3. Analyses were repeated by sex and age group. Results in the overall sample were mirrored in females and males. In age stratified analyses, people in the stimulant only group had half the hazard of death compared to those in the opioid only group when considering people aged less than 40. When considering those aged 40 or older, this relationship was reversed, with the stimulant only group having twice the hazard of death compared to the opioid only group. Full model estimates are included in the Supplement.

This study found that people who used both opioids and stimulants had more than twice the hazard of fatal overdose compared to people using opioids only. This finding is consistent with prior studies, whereby the compounded impact of concurrent stimulant and opioid use poses an increased risk of poor outcomes, including discontinuation of OAT [7]. To our knowledge, this is the first Canadian population-level study examining the association between both opioid and stimulant use and illicit drug toxicity (overdose) death. Much of what is known about the use of both opioid and stimulants and overdose is derived from detection of these substance in post-mortem toxicology. In this study, we use a prospective cohort design to identify cases of illicit drug toxicity death following contact with health services for opioid and/or stimulant use.

The descriptive findings suggest that when compared to people who use opioids only, people who use both drugs were significantly less likely to receive all prescribed medication types, including benzodiazepines, z drugs, sedatives, and opioids for pain. This was true despite people who used both opioids and stimulants having significantly more comorbidities. This suggests increased barriers in access to and engagement with health services among this population. Stigma poses a known barrier to service engagement among people who use substances [28] and has been found to be more severe towards people with dual diagnoses [29]. Prior studies have suggested the need for education to reduce provider stigma toward this population and to better understand how to meet service needs [29]. This is particularly important for people who use stimulants, for whom the implementation of harm reduction and treatment interventions remain limited [30].

While much of the framing of overdose risk in North American has focused on opioids, often referred to as the “opioid crisis” [31], our analysis revealed that people who use stimulants have a similar risk of overdose death as compared to people who use opioids. Services remain limited for people who use stimulants in BC, and internationally. Traditionally, interventions available for stimulant use include psychosocial treatments, which have been provided with limited and short-term efficacy [17]. Given the rise of stimulant use and the escalation of overdose deaths in recent years, there has been a call for a new treatment paradigm for stimulant use, paralleling the OUD treatment framework, including prescribed stimulant medications coupled with other health care interventions [30, 32]. In BC and Canada, alternatives to illicit stimulants have been prescribed in a small number of settings and programs, with positive effects. For example, dextroamphetamine has been provided in a clinic in Vancouver since 2016 with observed reductions in illicit stimulant use and improved health outcomes [33, 34]. In 2021, residents of a COVID-19 isolation hotel in Halifax who were using illicit stimulants were provided dextroamphetamine and methylphenidate with no reported cases of overdose or adverse events [35]. Nevertheless, the range of options and reach of this prescribing remains limited, often available only to people with clinical diagnoses of stimulant use disorder and prescribed for daily dispensation at community pharmacies. A wider diversity of alternatives to the illicit drug supply are required to support people with diverse motivations for, patterns of, and goals around their opioid and stimulant use, which may or may not include abstinence [36].

The analysis of survival over time revealed a similar survival probability in the first two years of follow-up across substance use groups, followed by a subsequent decline in survival, the greatest of which was observed in the group using both opioids and stimulants. This finding aligns well with recent data from the US, indicating rising rates of concurrent use among overdose cases over time. For example, between 2012 and 2018, the rate of fatal overdoses that involved cocaine more than tripled and those involving methamphetamine increased almost five-fold [37]. Furthermore, overdose mortality in the US involving both stimulants and opioids has increased more sharply than deaths involving stimulants alone. For example, by 2019, more than 75% of deaths involving cocaine, and approximately half of those involving methamphetamine also involved opioids [38].

The decline in survival over time in BC among people using both opioids and stimulants may reflect a number of missed opportunities for intervention in the time immediately following contact with care for substance use. One such intervention that can be provided for people using opioids and stimulants concurrently is OAT. While widely available in BC, prior studies have demonstrated that concurrent stimulant use is associated with poorer long-term OAT engagement [7]. Our sensitivity analyses revealed that when OAT was added to the model, the hazard of fatal overdose among people who used stimulants and opioids was lower than in the model where OAT was not included. Furthermore, the impact of OAT on overdose mortality did not differ between people using opioids only compared to people using both opioids and stimulants. Studies have shown that where people who use stimulants have been retained in OAT, reductions in stimulant use have been observed over time [8]. This confirms the importance of providing evidence-based effective interventions for OUD to people who use both opioids and stimulants.

Nevertheless, OAT retention rates in BC have been declining in recent years [6]. This decline may reflect that current treatment options are not meeting the preferences of many people who use opioids in BC. For example, more than 70% of people who use opioids and accessed harm reduction sites in BC in 2019 reported smoking opioids [39], and approximately 60% reported heroin as their preferred opioid [36]. These preferences for substances and routes of administration are not met by available treatment options. As such, treatment interventions must be introduced to address the diverse preferences for routes of administration of people who use opioids and stimulants in BC, to include smokable options, such as inhalable diacetylmorphine [40, 41]. Furthermore, it is important to acknowledge that treatment is not always suitable or desired. Scaling up of treatment interventions therefore must be complemented by a robust range of harm reduction services. For example, in BC there are several overdose preventions sites and safe consumption services where people can inject, snort, or swallow drugs, however spaces for safe inhalation remain limited [42]. Interventions such as oral OAT will not be sufficient to reduce contacts with the illicit drug supply and reduce overdose risk when they do not meet patients' goals and preferences [43]. In this context, efforts must be made to expand OAT to include injectable and inhalable (e.g. hydromorphone and diacetylmorphine) options to meet a wider range of patient preferences [40, 44, 45]. This expansion must be coupled with an accessible and acceptable safe supply [46] of regulated drugs in order to reduce the risk of illicit drug toxicity death resulting from a toxic drug crisis which remains the deadliest public health crisis facing BC and Canada [1, 2].

There are a number of important limitations to this study. First, we report on sex as male or female. This binary variable reflects biological sex assigned to a person at birth, and does not provide information on gender identity. While BC has introduced a third “X” option to its birth certificates, it is not universally accepted and is not yet available in our datasets. Drug types involved in each of the reported deaths are only available for cases deemed as “closed” by the Coroners Service (See supplement). Nevertheless, analysis of illicit drug toxicity deaths in BC has consistently revealed the relevance of both opioids and stimulants to illicit drug toxicity deaths [2]. For example, analyses of all illicit drug toxicity deaths in BC between 2015 and 17 revealed that one or more opioid was found to be relevant to death in 85.5% of cases, and one or more stimulants were found to be relevant to death in 70.6% of cases [47].

In this study, we rely on the use of ICD9/10 codes to identify people with health care visits related to opioid or stimulant use. As such, misclassification of exposure is possible. For example, someone may use opioids and stimulants, but may only have diagnostic codes for one of these substances if both were not recorded during their health care visit. Furthermore, our study does not represent people who may use opioids and or stimulants but who are not in contact with health care. A recent death review panel in BC identified that many people who died of illicit drug toxicity death had not accessed substance use services prior to their death [48]. BC-ODC data show that approximately 51% of deaths between 2010 and 18 had no associated diagnostic codes for substance use disorders. This suggests that the risk of death is present among people who use occasionally or infrequently or who use regularly but do not have a SUD diagnosis. As such, it is possible that the hazard of fatal overdose is underestimated among people who use opioids and stimulants in BC. Nevertheless, given our study is based on a random general population sample, findings may be generalizable to population-level study samples in other North American settings where similar trends in rising opioid and stimulant use have been observed.

This study has emphasized the elevated risk of fatal overdose facing people who use opioids and stimulants. Expanding access to and increasing support for people who use stimulants is urgently needed in order to reduce risk of overdose mortality in BC. This includes approaches to improving accessibility of treatment such as OAT for people who use stimulants and policy directives to scale up the implementation of evidence-based pharmaceutical alternatives to illicit stimulants. Lessons can be applied more broadly in a North American context where similar trends of opioid and stimulant polysubstance use are being observed.