Unintentional cannabis exposures in children pre- and post-legalization: A retrospective review from a Canadian paediatric hospital

Abstract

Objectives

Canada legalized recreational cannabis in October 2018. Cannabis is increasingly available in numerous forms—especially edibles—that make children vulnerable to unintentional intoxication. We sought to: determine the frequency of visits due to cannabis intoxication pre- and post-legalization; characterize the clinical features and circumstances of cannabis intoxication in the paediatric population; and create greater awareness among healthcare providers about this issue.

Methods

We performed a retrospective chart review of Emergency Department visits at the Children’s Hospital of Eastern Ontario (Ottawa, ON) between March 2013 and September 2020. Inclusion criteria were: age <18 years; unintentional cannabis ingestion, identified by ICD-10 codes T40.7 and X42. We assessed basic demographics, clinical signs and symptoms, exposure details, investigations, and patient disposition.

Results

A total of 37 patients (22 male) met inclusion criteria, mean age 5.9±3.8 years. Most visits (32; 86%) occurred in the 2-year period after legalization. Altered levels of consciousness, lethargy/somnolence, tachycardia, and vomiting were the most common presenting signs and symptoms. The majority of exposures were to edibles (28; 76%) in the home setting (30; 81%). Poison control and child protective services were involved in 19 (51%) and 22 (59%) of cases, respectively. Twelve patients (32%) required admission to the hospital, the majority of whom stayed <24 h.

Conclusions

Our data confirm increased paediatric hospital visits related to unintentional cannabis exposures post-legalization. Consideration of this clinical presentation is critical for acute care providers. Advocacy for safe storage strategies and appropriate enforcement of marketing/packaging legislation are imperative for public health policymakers.

Cannabis is one of the most widely used recreational substances in Canada and around the world, with an estimated 25% to 27% of Canadian adults reporting consumption each year (1). Over the last several decades, many countries and jurisdictions have moved towards decriminalizing and even legalizing cannabis. Canada legalized the purchasing and selling of recreational cannabis in October 2018, and more specifically edibles in October 2019, through the Cannabis Act (2). While the laws regarding cannabis are fairly similar across Canada’s provinces and territories, there are a few regional differences to note, especially in Québec, where there are stricter regulations regarding the type of edibles allowable for sale (3,4). Ontario has the highest density of legal cannabis retailers in the country, with over 1000 retailers as of August 2021, and more than 100 retail applications in progress (5). Cannabis can be purchased in retail stores or online in Canada.

Cannabis is available in numerous forms, including inhalation (smoking, vaping), topical applications, and a wide range of ingestible forms (baked goods, candies, gummies, etc.). Compounding the effect of increasing availability, the palatability and packaging of cannabis makes it much more appealing for children (6,7). Further, there has been a documented rise in the concentration of tetrahydrocannabinol (THC)—the psychoactive component of cannabis—in many of these products (8). Cannabis-related toxicity in paediatrics can present with a range of clinical signs and symptoms, including elevated heart and respiratory rates, nausea, vomiting, slurred speech, agitation or drowsiness, tremor, seizures, and even coma (6,7,9–12).

Several American studies have shown an increase in unintentional paediatric ingestions of cannabis, and a rise in children presenting to medical attention with acute CNS toxicity secondary to unintentional cannabis exposure (6,7,9,13). These studies were conducted in jurisdictions that had recently legalized recreational marijuana (e.g. Colorado, Washington) (13,14). There are limited data in the Canadian context (15–17).

This study investigated unintentional cannabis ingestions in a high-volume tertiary care paediatric hospital in Ontario, Canada. We sought to: (i) determine the frequency of paediatric visits due to cannabis intoxication pre- and post-legalization in a tertiary paediatric hospital; (ii) characterize the clinical features and circumstances of cannabis intoxication in the paediatric population; and (iii) create greater awareness among caregivers and healthcare providers regarding unintentional cannabis exposure.

METHODS

We conducted a retrospective chart review of cases of unintentional cannabis exposure who had presented to the Children’s Hospital of Eastern Ontario (CHEO; Ottawa, ON, Canada) from March 1, 2013 to September 30, 2020. CHEO has a catchment population of approximately 1.5 million and serves the children and families of greater Ottawa, Eastern Ontario, Western Quebec, as well as those referred from Northern Ontario and Nunavut (18). CHEO has one of the busiest paediatric emergency departments in the country, with upwards of 75,000 annual visits (19).

The study was approved by the CHEO Research Ethics Board protocol #21/13X. The need for parental consent was waived. We obtained data via electronic health record (Epic™) or hard copy medical chart review (for two charts).

We included any patient <18 years of age who had visited CHEO during the specified date range, and who had one of two ICD-10 (International Statistical Classification of Diseases and Related Health Problems—10th revision) discharge codes: (i) T40.7—‘poisoning by cannabis [derivatives]’ and/or (ii) X42—‘unintentional poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified’. We excluded any patients with ICD-10 codes associated with intentional exposures (e.g. F12.0, F12.1) (20); however, in order to rule out the possibility of an unintentional exposure in the younger paediatric range (age <13 years), we also reviewed any patient in this age group assigned codes F12.0 and F12.1 to confirm that the exposure was indeed not unintentional. Our search strategy is outlined in Figure 1.

Figure 1.

Search strategy.

T40.7—poisoning by cannabis [derivatives]; X42—accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified; Y12—poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified, undetermined intent; F12.0—mental and behavioural disorders due to use of cannabinoids, acute intoxication; F12.1—mental and behavioural disorders due to use of cannabinoids, harmful use.

Each chart was reviewed independently by both co-authors (A.C., A.R.-L.), and data were tabulated in a de-identified and coded data collection form (see Figure 2). Any discrepancies were identified and resolved with consensus discussion. Data were analysed using descriptive statistical methods.

Figure 2.

Data collection form.

RESULTS

A total of 71 charts met the initial inclusion criteria. Of these, 30 were duplicate entries from both the Emergency Department and in-patient searches and were thus excluded. An additional four cases were excluded based on the following rationale: (i) an infant with significant comorbid medical illness (whom we excluded as his course was significantly complicated by confounding issues and cannabis exposure was speculative and not documented with testing); (ii) a patient receiving long-term medical cannabis in the setting of malignancy who had been inadvertently overdosed by his caregiver; and (iii) two patients <13 years who were assigned code F12.0 were ultimately found to have had a clearly intentional ingestion. Thus, a total of 37 patients were included in the final case review.

There were a total of 22 males (59%) and 15 females (41%) included in the study, with a mean age of 5.9±3.8 years (range 0.9 to 17.4 years). Nearly half of the patients (18; 49%) arrived at our hospital by car; another 18 (49%) arrived by ambulance. Of the latter, 11 (30%) arrived by ambulance directly, and 7 (19%) via transfer from another medical facility. In terms of geographic origin, the vast majority (32; 86%) came from the province of Ontario and the remaining 5 (14%) were from Québec. We saw a total of 32 presentations (86% of total) in the 2-year study period following legalization with the Cannabis Act (October 2018 to September 2020), compared to only five presentations (14%) in the 5-year study period preceding it (March 2013 to October 2018). This is displayed in Figure 3.

Figure 3.

Accidental cannabis presentations to CHEO (by year).

Table 1 provides details surrounding the cannabis exposure, clinical features, as well as patient outcomes and dispositions. The majority of accidental exposures (28; 76%) were to some form of edible product, including gummies, candies, chocolates, or baked products. In just over half of the cases (22; 59%), an estimated quantity of cannabis ingested was provided on history. On average, based on parental reports, this amounted to approximately 180 mg of THC component (range of 17.5 to 1000 mg). Nearly half of the patients (16; 43%) presented within 2 h of ingestion. The majority of exposures took place in the home setting (30; 81%), with the patient’s parent(s) being the source person(s). In the majority of cases (31; 84%), the cannabis ingestion was disclosed upon presentation to the hospital. The most common abnormal vital sign among our patients was tachycardia (18; 49%), and the majority of patients had alterations in their Glasgow Coma Scale (GCS): GCS 13 to 14 (n=14; 37%), GCS 9 to 12 (n=6; 17%), GCS ≤8 (n=3; 9%). The most frequent presenting signs and symptoms were altered level of consciousness (28; 76%), lethargy/somnolence (22; 59%), and vomiting (11; 30%). Cannabinoid toxicology—done via urine—was performed in fewer than half of the study population (18; 49%). Other commonly performed investigations or interventions included blood work (19; 51%), electrocardiography (16; 43%), and intravenous fluid administration (14; 38%). The majority of patients were ultimately discharged from the Emergency Department (25; 68%); eleven patients (30%) required a medical ward admission; and one patient was admitted to the Intensive Care Unit. Of the patients admitted to the hospital, the majority (7/12; 58%) had a length of stay shorter than 24 h.

Table 1.

Exposure details, clinical presentation, and patient disposition

Exposure details
Type of cannabis product	Gummies, candies=17 (46%) Chocolates, baked goods=11 (30%) Medicinal (nabilone)=1 (3%) Cannabis leaves=1 (3%) Unknown=7 (19%)
Quantity of cannabis product	Estimated/known=22 (59%)   • Estimated=5 cases (14%)   • Known=17 cases (46%)      ○ Average quantity=180 mg±223 [range 17.5–1000] Unknown=15 (41%)
Disclosure of cannabis exposure at presentation?	Yes=31 (84%) No/unknown=6 (16%)
Timing of exposure (prior to Emergency Department presentation)	≤2 h=16 (43%) >2 h=14 (38%) Unknown=7 (19%)
Source person and location	Parent/relative=30 (81%) Home setting=30 (81%)
Documented inquiry into parental cannabis use	Yes=16 (43%) No/not specified=21 (57%)
Vital Signs & Glasgow Coma Scale
Abnormal vital signs at triage	Tachycardia (18/37; 49%) Bradycardia (1/37; 3%) Tachypnea (3/37; 8%) Hypertension (5/37; 14%) Hypotension (1/37; 3%) Hyperthermia (1/37; 3%) Hypothermia (1/37; 3%)
Glasgow Coma Scale (GCS) at triage	GCS 15=14 (37%) GCS 13–14=14 (37%) GCS 9–12=6 (17%) GCS ≤8=3 (9%)
Clinical features
Cannabinoid urine toxicology results performed?	Yes=18 (49%) No=19 (51%)
Documented signs/symptoms	Altered level of consciousness=28 (76%) Lethargy, somnolence=22 (59%) Vomiting=11 (30%) Ataxia=10 (27%) Seizures or abnormal movements=8 (22%)
Other investigations and procedures performed	Blood work=19 (51%) Electrocardiogram±echocardiogram=16 (43%) Head imaging=2 (5%) Intravenous fluids administered=14 (38%)
Additional details
Disposition	Observation in the Emergency Dept.=25 (68%) In-patient admission=11 (30%) Intensive care unit admission=1 (3%)
Length of stay (LOS) in hospital, in hours	Observation in the Emergency Department:   • Mean LOS=7.0±4.5 (range 2–21)      ○ <12 h=19 (51%)      ○ 12–23 h=6 (16%) Admission to hospital (ward or intensive care):   • Mean LOS=31.3±14.6 (range 16–63)      o 12–23 h=7 (19%)      o 24–35 h=2 (5%)      o 36–47 h=2 (5%)      o 48–71 h=1 (3%)
Other services involved	Poison centre=18 (51%) Child protective services=22 (59%)

INTERPRETATION

The issues surrounding cannabis decriminalization and legalization are complex and involve nuanced health, economic, and sociocultural considerations. Studies have repeatedly shown that children and youth face unique and significant health risks when it comes to cannabis, whether through intentional use or accidental exposure (21–23).

In our study, accidental exposures to cannabis occurred primarily due to edibles located in the home setting, usually of a known or estimated quantity, and tended to affect primarily young children (mean age ~6 years). Our findings also reveal a marked increase in unintentional paediatric cannabis exposures in the period since federal legalization in Canada, as demonstrated earlier in Figure 3. While one cannot prove a causative link, it is a concerning trend certainly worth highlighting in the context of greater availability, increased recreational use, and social acceptance of cannabis (24,25). Of particular note, our study partially overlapped with the COVID-19 pandemic period. The January 2021 National Cannabis Survey (NCS)—which collects self-reported data on Canadians’ cannabis consumption—found that among those Canadians who had previously used cannabis, 34% reported that their consumption had increased compared with the pre-pandemic period (26). There are higher rates of consumption overall, with 20% of Canadians having used cannabis in the last 3 months of 2020 (compared with 14% using at least once every 3-month pre-legalization and 18% post-legalization but before the COVID-19 pandemic). This increasing use of cannabis among adults could be related to increased unintentional accessibility for children. At the same time, the public health regulations during the pandemic—especially lockdown and work-from-home measures—may have led to increased parental presence and supervision in the home setting. As a result, it may be reasonable to speculate that the number of accidental exposures was made lower than it would have been otherwise (i.e. without the added time spent at home by many parents).

We describe the clinical features and circumstances surrounding unintentional paediatric cannabis exposures. Most presentations included a change in level of consciousness, which appeared to be relatively short-lived. A review of the clinical features of our patients reveals a non-specific presentation of altered level of consciousness, somnolence, and vomiting. Similar results were found in case series conducted in the USA. One review of cases from the Alaska/Oregon poison centre—which looked at both accidental and non-accidental exposures across a range of ages—found reduced levels of consciousness to be the most common clinical sign among children and adolescents (45.1% and 33.3%, respectively), followed by vomiting (9.9% and 9.5%, respectively) (16). This study also documented a total of eight patients who required admission to an intensive care unit; of these, four were infants and young children (age ≤4 years) with accidental exposures, as well as one teenager with an intentional ingestion. In a case series conducted in San Francisco by Vo et al. (27), tachycardia (83%), lethargy (50%), and vomiting (42%) were also commonly documented among paediatric patients. Interestingly, of the twelve children included in this study, three (25%) required admission to intensive care, though all were discharged home within 12 h. The remaining patients had short stays either in the Emergency Department or on a medical floor, ranging from 1.5 to 12 h.

We noted inconsistent practices with regard to a physician’s decision to involve child protective services, whether through the Children’s Aid Society (CAS) or our hospital’s Child and Youth Protection Team. It is difficult to appreciate or fully understand through chart review and documentation the particular reasons that the treating physician(s) chose to involve or not involve these services, though we suspect it was a combination of physician discomfort, the age of the child, the severity of the presentation, and potential perceptions around parental/guardian responsibility (or lack thereof) surrounding the exposure. There were also inconsistent practices in the utilization of urine toxicology studies, presumably done for investigation, confirmation, or determination of co-ingestants; these were performed in approximately half of the cases. To us, this number was surprisingly low. In a systematic review of case series and case reports of accidental paediatric cannabis exposures by Richards et al. (28), the authors document a far higher rate of toxicology studies being performed. Of the 114 children included in the analysis, toxicology was performed in 107 (94%) cases.

It is also interesting to note that 84% of parents declared the cannabis exposure at triage or at first assessment by a treating physician. The rest were undeclared, and the diagnostic clarity came only with investigation or later in the encounter; the clinical importance of recognizing the presentation of cannabis intoxication is thus critical.

This study has a few key limitations to note, primarily its small sample size. Despite the single-centre data collection, our numbers are comparable to previously published case series (29–31). Given the relatively mild nature of symptoms in most cases (as demonstrated by our review) and the potential for parental fear of reprimand, the sample in question could represent a substantially underreported underestimate of the true number of unintentional cannabis exposures among infants, children, and youth, if the caregiver chose not to seek medical assistance or was unaware of the exposure. Additionally, some of the charts were missing data, especially if the patient was transferred from an external facility (n=7; 19%); this made it more challenging to capture the initial presentation of the patient, which may have evolved significantly by the time of arrival at our centre. Given the case review methodology, we were not able to appreciate whether inquiry surrounding safe storage and/or parental pattern of substance use may have been had (but not documented) in the patient chart. Finally, there are legislative differences across the different regions of our hospital’s catchment areas, particularly regarding edibles in the provinces of Ontario (where edibles are legal as of October 2019) versus Québec (‘An edible cannabis product offered in Québec may not be sweets, confectionery, dessert, chocolate or any other product attractive to persons under 21 years of age’) (4). Our data reflect this, given that the majority of presentations (32; 86%) were in children living in Ontario. Moreover, as our hospital’s catchment area spans these two jurisdictions, it would be interesting to see how data compare to other hospitals in the region.

This work has important public health implications. We found an increased rate of unintentional cannabis ingestion—particularly edibles—in children after Canadian legalization. An important next step would be broader data collection at the national level to determine the reproducibility of this trend. Professional development and education among healthcare providers may be required, particularly when it comes to recognizing and treating paediatric patients with cannabis intoxication; making decisions about involving (or not involving) child protective services; and educating caregivers about the safe use and storage of recreational cannabis products. Finally, given that most of the reported ingestions were of edible products appealing to children (namely, cannabis-infused sweets), we would strongly advocate that federal authorities better enforce legislation regarding the labelling, packaging, and marketing of recreational cannabis products—whether in retail stores or online—to better protect our children.

Funding

There are no funders to report for this submission.

Potential Conflicts of Interest

All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.