At-a-glance – The role of opioid toxicity in suicide deaths in Alberta, 2000 to 2016

Correspondence: Larry Svenson, Alberta Health, 10025 - Jasper Avenue, Edmonton, AB  T5J 1S6; Tel: 780-422-4767; Email: Larry.Svenson@gov.ab.ca

Abstract

Given the current opioid crisis in Canada, there is interest in the role of opioid toxicity in suicide deaths, particularly in whether any observed patterns are similar to those of unintentional deaths. The present analysis examined characteristics of opioid-toxicity suicide, and its role in relation to other suicide methods, from 2000 to 2016 in Alberta. It does not appear that the opioid crisis has resulted in a disproportionately higher number of suicides in Alberta. Individuals who die from unintentional opioid toxicity and those who die by opioid-toxicity suicide are likely distinct populations, requiring nuanced public health responses for prevention.

Keywords: opioids, opioid analgesics, suicide, public health surveillance, Canada

Introduction

Canada is in the midst of a nation-wide opioid crisis, spurred in part by an increase in opioid prescribing frequency, dosage and potency from the 1990s onward,^{Footnote 1Footnote 2} with increasing rates of opioid-related harms and deaths. In 2016, there were over 2800 arguably preventable opioid-related deaths in Canada.^{Footnote 3Footnote 4} Although there is no long-term, national-level mortality data, opioid-related deaths in Ontario reportedly doubled between 1991 and 2004.^{Footnote 2} In the past 10 years, the rate of hospitalizations in Canada due to opioid poisoning increased by 53%.^{Footnote 4}

While the majority of opioid-related harms and deaths are unintentional, opioids may also be involved where a death is intentional.^{Footnote 5Footnote 6Footnote 7} Approximately 30% of hospitalizations for opioid poisonings in Canada in 2016 were the result of self-inflicted harm, including suicide attempts.^{Footnote 4} Further, 16% of all opioid-related deaths in Ontario from 2006 to 2008 were suicides.^{Footnote 5}

It is known that opioids are commonly used for suicide (e.g. a Toronto-based study found opioids to be in 30% of suicides by overdose—the most frequently detected drug class—between 1998 and 2007^{Footnote 8}), and that individuals with chronic pain, who historically have frequently received opioid prescriptions, may be more likely to die by suicide.^{Footnote 6Footnote 9Footnote 10} However, there is a lack of information on the use of opioids for suicide in the context of Canada’s opioid crisis, such as whether the number of suicides by opioid toxicity has increased and whether patterns in opioid use for suicide are similar to those for unintentional deaths due to opioids. The present analysis aimed to address this knowledge gap by examining the demographics and opioid types used for suicide by opioid toxicity, as well as the role of opioid toxicity in relation to other methods of suicide, in Alberta over a 17-year period.

Methods

We examined de-identified data from Alberta Vital Statistics for suicide deaths in Alberta from 2000 to 2016. We identified manner (e.g. intentional/suicide) and method (e.g. poisoning) of death using International Statistical Classification of Diseases and Related Health Problems 10^th Revision (ICD-10) codes. Data were extracted where the underlying cause of death indicated death by suicide (ICD-10 codes X60–X84, Y87.0). Suicides for which the underlying cause of death was coded as ICD-10 X60 to X64 were considered suicides by drug toxicity. We broke down drug toxicity further using the following ICD-10 codes: T40.0 (opium), T40.1 (heroin), T40.2 (other opioids), T40.3 (methadone), T40.4 (other synthetic narcotics) and T40.6 (other and unspecified narcotics). Deaths with these codes were considered suicide deaths by opioid toxicity, regardless of the presence of other nonopioid drugs. We calculated descriptive statistics and performed logistic regression analyses with sex and age as predictors, for suicide by drug toxicity and suicide by opioid toxicity. We examined trends over time for patterns and tested for significance using chi-square and ANOVA tests.

Results

Demographics and trends over time

Between 2000 and 2016, there were 8344 deaths by suicide in Alberta; 2049 (24.5%) were females and 6295 (75.4%) were males. While the majority of suicide deaths (80%) were nondrug suicides, 20% were suicides by drug toxicity, of which 52.9% occurred in females. Approximately 22% of suicides by drug toxicity involved opioid toxicity; 51.4% of these opioid-toxicity suicides occurred among females. In a model fitting sex and age, females had 5.7 times the odds (95% confidence interval [CI]: 5.02–6.36) of suicide by drug toxicity compared to males (p < .0001). However, sex did not significantly predict opioid toxicity among suicides by drug toxicity.

Individuals who died by drug-toxicity suicide were older than individuals who completed suicide by nondrug means (average age 47.4 [standard deviation (SD) 14.2] vs. 42.0 [SD 17.0] years, respectively). In a model adjusting for sex, the odds of suicide by drug toxicity versus nondrug means increased by 2.4% (95% CI: 2.0–2.7%) for every one-year increase in age (p < .0001). The sex-adjusted odds of opioid-toxicity death among suicides by drug toxicity also increased by 0.8% (95% CI: 0–1.7%) for every additional year of age (p < .05). The mean age for individuals dying by opioid-toxicity suicide was 48.8 (SD 14.4) years.

The proportion of females to males for overall suicide (p = .35), suicide by drug toxicity (p = .71) and suicide by opioid toxicity (p = .62) remained fairly constant from 2000 to 2016. Mean age for overall suicide (p = .063), suicide by drug toxicity (p = .36) and suicide by opioid toxicity (p = .43) were stable over the 17-year period examined.

Methods of suicide over time

Suicide by nondrug causes was generally more common than suicide by drug toxicity in Alberta from 2000 to 2016. Among females, nondrug methods accounted for 57% of annual suicides, on average, and represented a higher proportion of suicides than drug toxicity (using any drug) in every year except 2004 and 2011, when drug toxicity accounted for 53% and 55% of suicides, respectively. The average annual rate of suicide by nondrug methods in females was 3.9 per 100 000 person-years. In males, nondrug methods were the primary means of suicide death, accounting for an average of 88% of annual suicides and an average annual rate of 17.9 per 100 000 person-years. The average rate of suicide by nonopioid drug toxicity was slightly higher for females compared to males (2.3 vs. 1.9 per 100 000 person-years) but nonopioid drug toxicity accounted for a much higher average percentage of suicide deaths in females than males (34% vs. 10%) (Figure 1). Similarly, the average rate of suicide by opioid toxicity per 100 000 person-years was 0.61 for females and 0.57 for males, while the average annual percentage of suicides attributed to opioid toxicity was 9% in females versus 3% in males.

Figure 1A Females

Text Description

Year	% Suicide by non-drug method	% Suicide by non-opioid drug toxicity	% Suicide by opioid toxicity	Rate of suicide by non-drug method	Rate of suicide by non-opioid drug toxicity	Rate of suicide by opioid toxicity	Overall suicide rate
2000	60.74766355	32.71028037	6.54205607	4.371882	2.35409	0.470818	7.19679
2001	60.57692308	28.84615385	10.5769231	4.165295	1.983474	0.727274	6.876042
2002	57.77777778	34.44444444	7.77777778	3.364156	2.005555	0.452867	5.822578
2003	55.35714286	34.82142857	9.82142857	3.942746	2.480115	0.699519	7.12238
2004	46.66666667	35.23809524	18.0952381	3.064958	2.314356	1.188453	6.567767
2005	62.96296296	33.33333333	3.7037037	4.149132	2.196599	0.244067	6.589798
2006	55.75221239	36.28318584	7.96460177	3.739851	2.433871	0.534264	6.707987
2007	52.98507463	40.29850746	6.71641791	4.111523	3.127074	0.521179	7.759776
2008	51.72413793	40.51724138	7.75862069	3.397836	2.661638	0.509675	6.569149
2009	55.88235294	35.29411765	8.82352941	3.154433	1.992273	0.498068	5.644774
2010	53.54330709	35.43307087	11.023622	3.70118	2.44931	0.762008	6.912498
2011	45.45454545	51.23966942	3.30578512	2.94264	3.317158	0.21401	6.473808
2012	65.04065041	28.45528455	6.50406504	4.182668	1.829917	0.418267	6.430851
2013	55.31914894	39.0070922	5.67375887	3.966002	2.79654	0.406769	7.169312
2014	62.67605634	28.16901408	9.15492958	4.408057	1.981149	0.643874	7.03308
2015	63.15789474	21.05263158	15.7894737	5.255935	1.751978	1.313984	8.321897
2016	68.42105263	18.79699248	12.7819549	4.357641	1.197154	0.814065	6.368859

Figure 1B Males

Text Description

Year	% Suicide by non-drug method	% Suicide by non-opioid drug toxicity	% Suicide by opioid toxicity	Rate of suicide by non-drug method	Rate of suicide by non-opioid drug toxicity	Rate of suicide by opioid toxicity	Overall suicide rate
2000	86.92810458	8.823529	4.248366	17.52971	1.779331	0.856715	86.92810458
2001	90.47619048	6.349206	3.174603	22.12753	1.552809	0.776405	90.47619048
2002	90.6779661	6.779661	2.542373	20.28366	1.516535	0.568701	90.6779661
2003	87.17201166	10.78717	2.040816	18.56746	2.297646	0.43469	87.17201166
2004	89.88439306	6.936416	3.179191	18.96723	1.463709	0.670867	89.88439306
2005	84.5659164	12.54019	2.893891	15.62925	2.317646	0.534841	84.5659164
2006	83.90092879	12.6935	3.405573	15.60339	2.360661	0.633348	83.90092879
2007	84.03614458	14.75904	1.204819	15.61121	2.741754	0.223817	84.03614458
2008	90.19073569	7.901907	1.907357	18.08817	1.584764	0.382529	90.19073569
2009	85.67639257	11.67109	2.65252	17.25325	2.350288	0.534156	85.67639257
2010	86.64987406	10.32746	3.02267	18.14995	2.163221	0.633138	86.64987406
2011	86.25336927	11.32075	2.425876	16.65692	2.186221	0.468476	86.25336927
2012	91.62303665	7.329843	1.04712	17.78363	1.42269	0.203241	91.62303665
2013	86.54822335	10.91371	2.538071	16.78779	2.116936	0.492311	86.54822335
2014	88.11881188	8.910891	2.970297	17.03847	1.722992	0.574331	88.11881188
2015	86.77354709	8.617234	4.609218	20.39842	2.02571	1.083519	86.77354709
2016	89.7810219	7.055961	3.163017	17.178	1.350033	0.605187	89.7810219

Method of suicide fluctuated more for females than males, while the overall suicide rate fluctuated more in males (Figure 1). In females, from 2014 onwards, there was a downward trend in suicide by nonopioid drug toxicity to previously unobserved levels, and concurrent upward trends in nondrug and opioid-toxicity suicides. In males, rate and percentage contribution of suicide by opioid toxicity also rose, peaking in 2015.

Types of opioids used in opioid-related suicides over time

In both females and males, the most commonly used opioids for suicide by opioid toxicity included morphine, codeine, hydromorphone and oxycodone (Figure 2). The second most common opioids recorded were “other synthetic narcotics,” which include fentanyl and U-47700. The early 2000s (for both sexes), 2013 (for males) and 2014 (for females) all saw an increased percentage of “other synthetic narcotics” use, and a decreased percentage of “other opioids” use for suicide.

Figure 2A. Females

Text Description

Year	T40.1 Heroin	T40.2 Other opioids	T40.3 Methadone	T40.4 Other synthetic narcotics	T40.6 Other and unspecified narcotics
2000	0	85.71428571	0	14.28571429	0
2001	0	63.63636364	0	36.36363636	0
2002	0	42.85714286	0	57.14285714	0
2003	0	90.90909091	9.090909091	18.18181818	0
2004	0	78.94736842	5.263157895	15.78947368	0
2005	0	75	0	25	0
2006	0	100	0	0	0
2007	0	100	0	0	0
2008	0	77.77777778	22.22222222	0	0
2009	0	77.77777778	0	22.22222222	0
2010	0	78.57142857	0	28.57142857	0
2011	0	75	0	25	0
2012	0	75	0	25	0
2013	0	87.5	0	12.5	12.5
2014	7.692307692	53.84615385	0	38.46153846	0
2015	0	81.48148148	0	25.92592593	7.407407407
2016	5.882352941	70.58823529	5.882352941	23.52941176	0

Figure 2B. Males

Text Description

Year	T40.1 Heroin	T40.2 Other opioids	T40.3 Methadone	T40.4 Other synthetic narcotics	T40.6 Other and unspecified narcotics
2000	0	61.53846	0	38.46154	0
2001	0	91.66667	0	8.333333	0
2002	0	66.66667	0	22.22222	11.11111
2003	0	85.71429	0	14.28571	0
2004	0	72.72727	27.27273	0	0
2005	0	88.88889	0	11.11111	0
2006	0	90.90909	0	9.090909	0
2007	0	100	0	0	0
2008	0	85.71429	14.28571	14.28571	0
2009	0	100	0	0	0
2010	0	83.33333	8.333333	8.333333	0
2011	0	100	0	0	0
2012	0	100	0	0	0
2013	0	70	0	40	0
2014	0	66.66667	8.333333	25	0
2015	4.347826	78.26087	0	26.08696	0
2016	0	84.61538	7.692308	23.07692	0

Notes: International Statistical Classification of Diseases and Related Health Problems 10^th Revision (ICD-10) codes were used to identify opioid types. T40.0 “Opium” was not used for suicide in Alberta from 2000 to 2016. T40.2 “Other opioids” include morphine, codeine, hydromorphone and oxycodone. T40.4 “Other synthetic narcotics” include synthetic opioids such as fentanyl and U-47700.

Discussion

Individuals in Alberta who died from unintentional opioid toxicity deaths in 2016 were predominantly male (73%) and between 30 and 39 years of age.^{Footnote 11} In comparison, suicide deaths by opioid toxicity were observed in slightly more females, with a mean age of 49 years. These data suggest that unintentional and suicide opioid-toxicity deaths occur in distinct populations, and may thereby represent different public health concerns. Previous study results have also shown more opioid-toxicity suicides among females and more opioid-misuse toxicity deaths among males.^{Footnote 12}

There was no substantial increase in rates of suicide by opioid toxicity from 2000 to 2016 like that observed in unintentional deaths in Alberta between 2011 and 2016,^{Footnote 7} nor was there a significant change in suicide decedent demographics over time. The varied proportions of suicide method and stable overall suicide rate in females suggest a substitution of suicide methods rather than changes in frequency of suicide. The increased proportion of suicide by opioid toxicity around 2014 and 2015 may be a result of individuals utilizing an increased availability of synthetic opioids such as fentanyl.

Suicide deaths by opioid toxicity show a different opioid-type usage pattern compared to unintentional opioid toxicity deaths; fentanyl is the main contributor to unintentional opioid toxicity deaths,^{Footnote 7} but other opioids (morphine, codeine, hydromorphone and oxycodone) are predominantly used for suicides. Combined with the differences in demographics, it is likely that opioid-related suicides and unintentional deaths largely occur in different population groups in Alberta, with differing access to various types of opioids.

Strengths and limitations

Strengths of this study include the use of Alberta Vital Statistics data to capture all suicide deaths in Alberta over a 17-year span. However, only suicides for which the main cause of death was drug toxicity were considered drug-toxicity suicides; suicides for which the main cause of death was another method (e.g. hanging, firearm), but that also involved drug toxicity, were excluded. In addition, the use of ICD-10 codes prevented identification of the contribution of specific opioids, limiting the level of detail to that inherent in the code. Further, the contribution of opioids to drug toxicity deaths may also be underestimated, especially in earlier years, as coding practices occasionally do not identify specific drugs contributing to drug toxicity. There is potential for confounding in using administrative data over a long time period, whereby observed differences could reflect changes in coding practices over time and in response to known crises rather than true differences.

Conclusion

There are differences in population demographics, observed patterns and the types of opioids used for opioid-related unintentional and suicide deaths in Alberta, which suggests that these are separate phenomena within the opioid crisis. Prevention of suicide by opioid toxicity likely requires a public health response separate to the current response used to prevent predominantly illicit fentanyl-driven and unintentional opioid-toxicity deaths.

Acknowledgements

The authors would like to thank Sean L’Heureux and Alberta Vital Statistics for their support.

Conflicts of interest

The authors have no conflicts of interest to report.

Authors’ contributions and statement

EC, BM and LS conceived and designed the study. BM extracted and de-identified the data for analysis, and EC conducted the analysis. All authors were involved in drafting the manuscript and interpreting the data. All authors read and approved the final manuscript.

The content and views expressed in this article are those of the authors and do not necessarily reflect those of the Government of Canada.

References