We demonstrate in this study that, notwithstanding changes in the last 20 years, trends for declared opioid-related ARs have increased since 1965. This retrospective population-wide investigation of opioid-related ARs is, to our knowledge, the first of its kind and demonstrates the vast potential of using the Canada Vigilance database to examine AR patterns in Canada.
In recent years, manufacturers have been the most prominent reporting body. This is consistent with the Canadian homologation procedure , which requires manufacturers to disclose any ARs they become aware of, especially in the first few years after a new product is commercialized. However, it was not mandatory for non-manufacturers to disclose ARs during the period covered by this study.
Increasing and Fluctuating Trends of Opioid-Related AR Declarations
As expected, we saw an upward trend in opioid-related AR declarations over time (4.2% [95% CI 3.1–5.2]). This is consistent with our hypotheses, including the effects of the increasing incidence of cancer in Canada , along with the aging of the population .
The rise in opioid-related AR declarations is also consistent with a report released by Health Canada in 2018 , which revealed the total number of AR declarations (not rates) for any type of substance (including opioids) from 2008 to 2017. However, in our study, we were able to fine tune the methodology to produce rates rather than just absolute numbers. Furthermore, rather than a 10-year span, we looked at trends over the last 50 years.
In the last 20 years, reporting rates have fluctuated. Many other factors were changing at the same time, which could have led to some of the fluctuations. These could be explained by, first, the increased availability of various surgical procedures for the general public ; the availability of long-acting, but no less strong, opioid therapy choices over time (Table 1) ; and the growth of the medical world . Perhaps it is the increased knowledge and other factors involving individuals that have changed over time, such as patient features, comorbidities, and so on.
Purdue Pharma’s oxycodone, for example, was taken off the market in Canada in 2012 . The onset of this withdrawal appears to coincide with the peak in opioid-related AR rates in 2012 (Fig. 2). Perhaps the heightened awareness and vigilance in the community and health professionals increased their willingness to report opioid-related ARs. Additionally, between 2011 (n = 454) and 2012 (n = 698), oxycodone was the only opioid for which there was a significant rise in AR declarations.
The averted opioid crisis of the mid-2000s could possibly explain, at least in part, the fluctuations in the trends (Fig. 2). In fact, the community of Stand Off in Alberta, Canada, raised the alarm for the first time in 2015, when family physicians on the Blood Tribe reserve were treating two to three fentanyl-related overdoses per work shift . This was discovered to be just the tip of the iceberg, whether for prescription or illegal opioids. As a result, the government of Canada introduced steps that may have aided in raising awareness of severe opioid-related ARs.
Indeed, the Minister of Health convened a meeting of health partners to commit to a collaborative action aimed at minimizing the detrimental effects of opioids in Canada , as it is globally . Similarly, Health Canada dropped the medical prescription requirement for obtaining naloxone nasal spray in 2016, making it a more accessible treatment for opioid-related overdoses . With all of the media attention given to the opioid problem and naloxone availability, we believe that public awareness of the need to report opioid-related ARs to Health Canada has grown.
The unexpected decrease in the rates of reported opioid-related ARs in 2016 is intriguing. According to our data, hydromorphone (83 ARs in 2015; 18 ARs in 2016), codeine (144 in 2015; 73 in 2016), and buprenorphine (186 in 2015; 105 in 2016) were all declared lower in 2016. This decline could not, to our knowledge, be explained by any removal from the market of these compounds, let alone shortages. Notably, since 2016, Health Canada has mandated that medication makers disclose actual and predicted drug shortages, as well as discontinuations, to Drug Shortages Canada . It is not unlikely that this drop could be caused by the data extraction problems we encountered. For instance, we were unable to retrieve data on fentanyl (Table S1 in the ESM) after November 2015. Fortunately, we believe this information bias does not render the entire trend study incorrect, as the findings of our conservative sensitivity tests were positive (rates increased over time).
Sex and Gender Differences in Trends of Opioid-Related AR Declarations
Although the overall increases in AR rates for both sexes were statistically similar, female patients had overall higher rates than male patients. This is in line with research that claims ARs have a greater impact on women. Furthermore, the Canadian Tobacco, Alcohol and Drugs survey revealed in 2015  and 2017  that females consumed more prescribed opioids than did males (13.9 vs. 12.1%) . Male trends, on the other hand, were higher between 2008 and 2013. Unfortunately, stratified analyses by gender were not possible because of a lack of this information in our database (we only had access to the biological sex, not the social construct of it). In future studies, sex and gender disparities in opioid-related ARs should be explored and comprehended more thoroughly.
Strengths and Limitations
Although various descriptive studies on ARs have previously been published [2, 3, 9, 29,30,31, 35], a key strength of our work was the use of national datasets that encompass practically the whole population and opioid-related AR rates since 1965. As a result, selection bias was reduced and generalizability improved. The data for this study were publicly available through the Canada Vigilance database, and we share how to extract and analyze it in a way that is both sustainable and democratic. In terms of transparency, our final dataset is available to the public. Our group has previously worked on validating the methodologies for using Canada Vigilance data in research, boosting confidence in the quality of the information . Despite these strengths, some limitations were unavoidable. First, the reported trends in declared opioid-related ARs should be viewed as an underestimation of the true situation. This underestimation is because of a potential information bias, as some molecules were not extractable throughout the entire study period/reporter types, despite multiple interactions with the database owner (Table S2 in the ESM). We were able to conduct post-hoc sensitivity analyses, which yielded reassuring results. According to a meta-analysis of international data, the underestimating of self-reported ARs could be as high as 95% . Second, despite all efforts to avoid residual information bias , given the secondary data analysis nature of this vast database and the small number of non-essential missing data, we cannot rule out the chance of mistakes. Obviously, some external variables over the last 50 years may have influenced opioid-related AR reporting. For example, increased awareness of Purdue Pharma’s oxycontin when it was taken off the market in 2012 could have influenced the number of reports. Finally, we were unable to obtain statistics on the population’s drug prescriptions and usage over time. This would have aided in refining the AR rates on a smaller denominator rather than the entire population. Nonetheless, this field of study is rapidly growing, and we believe our work contributes to pharmacovigilance improvement for better patient care.
Although our findings should be interpreted with caution, they show that, compared with other health professionals, consumers, and other non-health people, pharmacists and physicians appear to declare the fewest opioid-related ARs. Patients take multiple doses of opioids every day, both in hospitals and in the community . In addition to being underreported to health authorities, opioid-related deaths are frequently preventable , as evidenced by the loss of many lives, some of whom are quite young [27, 28, 32]. Healthcare professionals provide a solution to increase population safety by reporting ARs to their local health authority as soon as they are discovered. Studying this field may aid clinicians to become more aware of the need to report ARs to health authorities. Developing a partnership of healthcare providers, as recommended by the World Health Organization, will help guarantee that drug safety choices and activities are truly patient centered and result in safe care .