Contribution of Addictovigilance data to assess adverse‐events linked to psychoactive substances in children and adolescents

Abstract

Aims

We sought to characterize adverse events and deaths associated with the use of psychoactive substances in children and adolescents.

Methods

Two French Addictovigilance databases were analysed: spontaneous reports and deaths over the period 2016–2021, in subjects aged 10–<18 years. An unsupervised classification was implemented on consumption data (medications or nondrug substances [NDS]) to identify subject clusters.

Results

A total of 1544 spontaneous reports were analysed, comprising mainly boys (65.6%), aged on average 16 ± 1 years. Four clusters were identified: The cannabinoids users cluster (n = 597) was typified by the use of cannabis or/and synthetic cannabinoids (95.1%), with psychiatric (67.7%) and digestive disorders (16.7%). The medications/solvents/cannabidiol users cluster (n = 699) was distinguished by the use of medications or NDS including nitrous oxide/cannabidiol, with mainly neurological disorders (46.5%). The polydrug users cluster (n = 177) includes polyusers (98.3%) of NDS and medications. These users mainly have substance use disorders (63.8%). The psychotropic medications users cluster (n = 71) was characterized by the use of psychotropic medications. This cluster appeared to be correlated with psychiatric and organic disorders. The death database recorded 44 deaths, mainly in boys (61.4%) aged over 15 years. The main substances involved in the deaths were NDS (70.5%) and methadone. In 68.2% of cases, a single substance was responsible for the death.

Conclusion

The adverse events related to the abuse of psychoactive substances identified in children and adolescents and the emerging signals show the need for increased surveillance and the implementation of prevention campaigns adapted to each group of consumers.

What is already known about this subject

• The consumption of psychoactive substances in minors remains a public health problem.

• Most studies evaluate consumption in this population, but not the associated adverse events.

What this study adds

• The analysis of these data in real life provides up‐to‐date and recent data on the main psychoactive substances consumed and the adverse events observed.

• Cluster analysis makes it possible to identify groups of subjects justifying specific approaches.

• Health authorities can use our results to implement prevention and information campaigns.

1. INTRODUCTION

Worldwide, 1 in 6 people are aged 10–19 years old. ¹ Adolescence is a time of life characterized by complex physical, psychological and social changes influenced by family, school and the wider community environment. ¹ This period is a phase marked by curiosity, risk‐taking and challenges during which young people seek new experiences. The consumption of psychoactive substances (PS) forms part of such behaviour. However, the use of PS has particularly harmful effects on adolescents. The health and social consequences of this usage represent a major public health concern. ²

Adolescence is a neurodevelopmental trajectory marked by heightened vulnerability whereby drug consumption may present unique developmentally contextualized risks. Research suggests that early alcohol use, prior to the age of 18 years, is associated with susceptibility to deficits in neuropsychological performance. ³ Likewise, early drug use is an important indicator of later emergence of lifelong drug abuse. ³

Large‐scale national surveillance studies have been conducted to assess prevalence of use and reported the following information: in French adolescents in 2019, the most widely consumed substances were alcohol (53% of 15–16‐year‐olds), tobacco (22%) and cannabis (13%). ⁴ This form of usage remains high among teenagers despite regulations aimed at limiting access and prevention campaigns. Moreover, 6.1% of young people in France at this age described having already taken at least 1 other illicit substance (mainly 3,4‐methylenedioxy‐N‐methylamphetamine [MDMA], 3.4%; cocaine 2.8%; amphetamines 2.3%). ⁴ Conversely, recent data show a decline in all levels of PS use (tobacco, alcohol, cannabis and other illicit substances) between 2017 and 2022 in adolescents with an average age of 17. ⁵

All these studies on usage and prevalence among adolescents means changes can be followed up and prevention programmes implemented. However, the adverse events related to these uses are more rarely documented.

In France, the network of Addictovigilance Centres ensures the careful monitoring of abuse and dependence cases secondary to the consumption of PS. ⁶ , ⁷ This network aims to detect safety signals by analysing the characteristics of spontaneous reports (SRs) related to PS, psychiatric and somatic adverse events (including deaths), and practices (routes of administration, context of use). ⁶

Lately, an increase in reports of adverse events linked to the use of PS among children and adolescents has been noted by the French Addictovigilance Network (FAN).

Consequently, the objective of our study was to better identify the patterns of PS consumption and reported adverse outcomes among children and adolescents aged 10–<18 years.

2. METHODS

2.1. Presentation of the FAN

The FAN is made up of 13 centres located in French administrative regions and aims to monitor risks and consequences of PS with abuse potential including prescription drugs and other legal and illegal substances (except for alcohol and tobacco). ⁶ , ⁷ , ⁸

This surveillance system, under the authority of the French Medicines Agency, is based on: (i) spontaneous reporting by healthcare professionals and patients, which is regulated by law (medical personnel have a legal obligation to declare all serious cases of abuse or dependence to their regional Addictovigilance Centre); and (ii) specific pharmacoepidemiological monitoring programmes to complement the data collected by SRs and to counteract the main limitation of spontaneous reporting‐based surveillance, the underreporting, and to facilitate the identification of emergent signals of psychoactive substance misuse. ⁹ , ¹⁰

1. The SRs database. All SRs correspond to adverse events related to consumption of PS or criterion of Diagnostic and Statistical Manual of Mental Disorders substance use disorder (SUD). The link between the consumption of psychoactive substances and the observed adverse effects is established by the healthcare professional who makes the declaration to the Addictovigilance Centre. In this study, all SRs identified by the FAN over the period 2016–2021 and involving subjects aged from 10 to <18 years‐old were included in our analysis. For the analysis, we identified 2 age groups 10–<15 years and ≥15–17 years. We chose these age groups to distinguish adolescents enrolled in middle school from those enrolled in high school. Finally, in France, paediatric services go up to the age of 15 years and 3 months, which also explains the choice of our age groups. This study was based on anonymous routinely collected data. For each case, the following data were collected: year of report, age, sex, family situation, housing, school status, concomitant medical and substance abuse history, substance use, psychoactive substance use patterns (route of administration, method of obtainment), adverse events, and associated severity criteria. According to the French Public Health Code, serious cases are defined by lethal or life endangering, disabling or incapacitating impact or which may result in hospitalization or prolonged hospitalization. ⁸

2. The DRAMES (Décès en Relation avec l'Abus de Médicaments et de Substances: Deaths related to the Abuse of Drugs and Psychoactive Substances) survey database. This survey collects and analyses all cases of death related to substance use (psychoactive drugs and illicit substances) among drug users. ¹¹ The aim of this survey is to identify the psychoactive substances involved, assess their dangerousness, and estimate the trends in the number of these deaths among drug users. In cases of violent death and/or suspicion of death caused by substance use, a request to the judicial authorities allows forensic analysis with substance identification/quantification in the biological fluids of the deceased. These analyses are carried out in approved toxicology laboratories. Drug‐related deaths are classified as directly attributed to drug use and as indirectly linked to drug use when caused, for instance, by falls, drowning, injuries or asphyxia.

Psychoactive substances have been categorized into 2 groups: (i) drugs (morphine, benzodiazepines, tramadol etc.); and (ii) nondrug substances (NDS), which include illegal substances (cocaine etc.) and certain legal substances that are not drugs (nitrous oxide, cannabidiol).

Polydrug use is defined by the use of 2 or more drugs (except alcohol and tobacco) by an individual, often at the same time or sequentially and usually with the intention of enhancing, potentiating or counteracting the effects of another drug (World Health Organization lexicon). ¹²

2.2. Ethics procedure

The study protocol was approved by the Local Ethics Committee (approval number from the Institutional Review Board [IRB], Montpellier University Hospital: IRB‐MTP_2023_03_202301354).

2.3. Statistical analysis

Characteristics of SRs were described using mean and standard deviation, minimal and maximal values for quantitative variables, and number and percentage for categorical variables.

Evolution of the proportions of minors in SRs over time was studied using the Cochran‐Armitage trend test as well as the proportion of severe cases within these minor populations.

Clusters of young consumers were identified on consumption data (drugs or NDS). A hierarchical clustering on principal components, which combines multiple correspondence analysis (MCA) followed by Ward's hierarchical clustering method, was implemented. The MCA method was used as a preprocessing to convert categorical consumption variables into continuous ones. The first components of the MCA were used to perform the agglomerative hierarchical clustering. Ward's method was used to minimize the total within‐cluster variance by merging clusters that minimize the increase in the total sum of squared distances. The number of clusters was then determined using the hierarchical tree and the elbow method. Clusters were described using construction variables.

Then, sociodemographic, psychiatric and consumption characteristics were detailed and compared between these clusters using Kruskal Wallis test for quantitative variables and chi‐squared or Fisher tests for categorical ones. When global comparison was significant, comparisons between clusters were implemented and P‐values were adjusted by Holm's correction. Significant pairwise comparisons were reported.

2.4. Missing data were not replaced

The analyses were performed by the Clinical Research and Epidemiology Unit at Montpellier University Hospital using SAS software (Enterprise Guide, Version 9.04, SAS Institute, NC, USA).

2.5. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22 (Alexander et al., 2021).

3. RESULTS

3.1. SRs database

3.1.1. Overall population

During the study period, 1544 SRs were identified and analysed; they concerned mainly boys (n = 1009, 65.6%). The mean age of included subjects was 15.6 ± 1.4 years. The 15–17‐year‐old age group was predominantly represented (79.4%).

Regarding the living conditions of the subjects included, 65% have family support, 45.8% stable housing and 61.1% are in school. The other data are summarized in Table 1.

TABLE 1.

Sociodemographic, clinical and consumption data (n = 1544).

	N	Study population
Age, mean ± SD [range]	1538	15.6 ± 1.4 [10–17]
Male sex, n (%)	1539	1009 (65.6)
Familial situation, n (%)	320
Parents/family/relationship		208 (65.0)
Migrant context		38 (11.9)
Isolated: alone, precarious minor		56 (17.5)
Foster home		14 (4.4)
Not stable		4 (1.3)
Housing, n (%)	299
Stable: family, friends, institution		137 (45.8)
Hotel, foster home, specialized centre		108 (36.1)
Homeless/squat		48 (16.0)
Not stable: family/friends		6 (2.0)
Activity, n (%)	332
School		203 (61.1)
Out of school/not in school		88 (26.5)
No activity		33 (9.9)
Other activity (working)		8 (2.4)
Severe SRs, n (%)	1400	579 (41.4)
Detail of severe SRs	419
Hospitalization/prolonged		309 (73.8)
Hospitalization		91 (21.7)
Other medically serious condition		9 (2.2)
Life‐endangering		8 (1.9)
Death		2 (0.5)
Disabling or incapacitating
History of psychiatric diseases, n (%)	1544	170 (11.0)
History of psychoactive substances, n (%)
Cannabis		250 (16.2)
Tobacco		142 (9.2)
Alcohol		110 (7.1)
Substance user*, n (%)	1544
Poly drug user		211 (13.7)
Antidepressant user		28 (1.8)
Antipsychotics		41 (2.7)
Benzodiazepines		290 (18.8)
Gabapentinoids		151 (9.8)
Opioids		200 (13.0)
Methylphenidate		10 (0.7)
Antihistamines‐H1		40 (2.6)
Solvents		116 (7.5)
Cannabinoids		102 (6.6)
Cannabis or synthetic cannabinoids		764 (49.5)
Hallucinogens		82 (5.31)
Heroin		21 (1.4)
Plants		15 (1.0)
Stimulants		293 (19.0)
Alcohol		229 (14.8)
Tobacco		4 (0.3)

^*Variables used to identify user clusters.

Abbreviation: SD, standard deviation.

Over the study period, we observed an uptick in the number of SRs on this group age and the total number of SRs received by the FAN (Figure 1). The proportion of these SRs on this group age also tended to rise over time (P < .001). According to the criteria specified above, 579 SRs (41.4%) were considered as severe, and these figures appeared to increase over the years too (P < .001; Figure 2).

FIGURE 1.

Repartition of spontaneous reports (SRs) reported to the French Addictovigilance network: young people (aged 10–<18 years) compared to those aged >18 years.

FIGURE 2.

Repartition of severe and nonsevere spontaneous reports (SRs) among the 1544 SRs analysed.

3.1.2. Psychoactive substances

The 1544 subjects included in the study consumed 2444 PS (apart from alcohol and tobacco), 916 medications (37% of all substances, mostly benzodiazepines or opioids) and 1528 NDS (63% of all substances and largely cannabinoids or stimulants). The PS are reported in the Supplementary Data.

The evolution of PS reported in the SRs is presented in Figure 3. There has been an overall increase in the number of reports, particularly involving cannabis, stimulants and benzodiazepines or Z drugs. Several safety signals should be highlighted: the misuse of pregabalin and nitrous oxide from 2019, and a peak in reports involving a product sold as cannabidiol in 2019.

FIGURE 3.

Characteristics of user clusters on clustering variables.

Polydrug use (>2 substances) was reported in 211 subjects (13.7%).

Associated alcohol consumption was present in 229 SRs (14.8%).

3.1.3. Adverse events related to psychoactive substances use

The main adverse events associated with PS consumption were psychiatric problems (n = 978, 63.3%), SUD (n = 803, 52.0%) and withdrawal syndrome (n = 85, 5.4%). Somatic adverse events accounted for 56.4% of cases including mainly neurological disorders (n = 633, 41.0%), digestive diseases (n = 230, 14.9%) and cardiovascular problems (n = 169, 11.0%).

3.1.4. Clustering results

According to the clustering method implemented on the consumption data, clusters of patients were defined. The 4 firsts component of the MCA, which represents 37% of inertia, were used to implement the hierarchical clustering. According to the dendrogram analysis and the elbow method, 4 clusters of patients were determined. The variables used and the characteristics of these user clusters are presented in Figure 4.

FIGURE 4.

Evolution of main psychoactive substances reported over the study period.

The cannabinoids users cluster (n = 597) was composed of patients consuming cannabis (87.10%) and synthetic cannabinoids (SCs; 8%), whether or not associated with other NDS (hallucinogens and stimulants).

The medications/solvents/cannabidiol users cluster (n = 699) comprised users of medications (opioids, mostly codeine/tramadol [19.7%], benzodiazepines [19%], gabapentinoids [12.7%]) or NDS (stimulants [13%], solvents [12.7%] including nitrous oxide and cannabidiol [14.4%]).

The polydrug users cluster (n = 177) was formed by polyusers (98.3%) of NDS (stimulants [67.8%, mainly cocaine], heroin 7.9%, hallucinogens 17.5%) and medications (benzodiazepines 57.1%, gabapentinoids 34.5%, opioids 26%).

The psychotropic medications users cluster (n = 71) was represented users of psychotropic medications: antidepressants 39.4%, benzodiazepines 28.2%, methylphenidate 14.1%, antihistamines‐H1 53.5%. This cluster refers to prescription drugs that may be used off‐label (from a prescribing physician), unauthorized (i.e., the prescription was issued to someone else, including family members) or obtained illicitly.

The user groups were compared in terms of sociodemographic characteristics and adverse events and reported in Tables 2 and 3, respectively.

TABLE 2.

Sociodemographic characteristics of user clusters.

Variable	(C1)	(C2)	(C3)	(C4)	Global P‐value	Pairwise comparison
	Cannabinoids users	Medications/solvents/Cannabidiol users	Polydrug users 3	Psychotropic medications users
	n = 597	n = 699	n = 177	n = 71
Age					< .01KW	C1 ≠ C2, C4
Mean ± SD	15.8 ± 1.3	15.5 ± 1.4	15.8 ± 1.2	15.1 ± 1.5		C3 ≠ C2, C4
[range]	[10–17]	[10–17.07]	[11–17]	[10–17]
Age class n (% col)	n = 597	n = 697	n = 177	n = 71	<.01C	C1 ≠ C2, C4 C3 ≠ C4
[10–< 15 years]	105 (17.6)	160 (22.96)	30 (16.95)	23 (32.4)
[≥15–17 years]	492 (82.4)	537 (77.04)	147 (83.05)	48 (67.6)
Sex, n (% col)	n = 595	n = 696	n = 177	n = 71	.96C
F	209 (35.1)	237 (34.05)	60 (33.90)	24 (33.80)
M	386 (64.9)	459 (65.95)	117 (66.10)	47 (66.20)
Age class in girls (years)	n = 209	n = 237	n = 60	n = 24	< .01C	C1 ≠ C2, C4
[10–<15]	41 (19.6)	74 (32.2)	16 (26.7)	14 (58.3)		C4 ≠ C2, C3
[≥15–17]	168 (80.4)	163 (68.8)	44 (73.3)	10 (41.7)
Age class in boys (years)	n = 386	n = 457	n = 117	n = 47	.31C
[10–<15]	63 (16.3)	85 (18.60) 372 (81.40)	14 (11.97) 103	9 (19.15) 38 (80.85)
[≥15–17]	323 (83.7)		(88.03)
Familial status, stable and secure, n (% col)	n = 149	n = 93	n = 66	n = 12	.01F	C1 ≠ C2, C1 C4 ≠ C2, C3
No	28 (18.79)	51 (54.84)	32 (48.48)	1 (8.33)
Yes	121 (81.21)	42 (45.16)	34 (51.52)	11 (91.7)
In education or activity, n (% col)	n = 171	n = 92	n = 56	n = 13	< .01F	C3 ≠ C1,C2,C4
No	56 (32.75)	28 (30.43)	36 (64.29)	1 (7.69)
Yes	115 (67.25)	64 (69.57)	20 (35.71)	12 (92.3)
Severity of SRs, n (%)	n = 554	n = 612	n = 168	n = 66	< .01C	C3 ≠ C1,C2
No	334 (60.3)	375 (61.3)	73 (43.5)	39 (59.1)
Yes	220 (39.71)	237 (38.73)	95 (56.55)	27 (40.91)

Abbreviations: C, χ² test; F, Fisher test used for global comparison; KW: Kruskal–Wallis test; SD, standard deviation.

TABLE 3.

Main adverse events in users.

Adverse events	Cannabinoids users	Medications/solvents/cannabidiol users	Polydrug users	Psychotropic medications users	Global P‐value	Pairwise comparison
Psychiatric, Yes	404 (67.67)	400 (57.22)	131 (74.01)	43 (60.56)	< .01C	C2 ≠ C1, C3
Psychotic, Yes	73 (12.23)	71 (10.16)	17 (9.60)	7 (9.86)	.60C
Mania‐inducing, Yes	65 (10.89)	56 (8.01)	18 (10.17)	4 (5.63)	.22C
Behaviour troubles, Yes	110 (18.43)	51 (7.30)	28 (15.82)	4 (5.63)	< .01C	C2 ≠ C1, C3 C1 ≠ C4
Suicidal behaviour, Yes	47 (7.87)	12 (1.72)	8 (4.52)	4 (5.63)	< .01F	C2 ≠ C1
Insomnia, Yes	17 (2.85)	8 (1.14)	3 (1.69)	1 (1.41)	.15F
Mood and Anxiety Disorders, Yes	84 (14.07)	39 (5.58)	11 (6.21)	5 (7.04)	< .01C	C1 ≠ C2, C3
Substance Use Disorders, Yes	314 (52.60)	339 (48.50)	113 (63.84)	37 (52.11)	< .01C	C3 ≠ C1, C2
Withdrawal Syndrome, Yes	41 (6.87)	25 (3.58)	15 (8.47)	2 (2.82)	< .01F	C2 ≠ C1, C3
Somatic, Yes	315 (52.76)	424 (60.66)	96 (54.24)	35 (49.30)	.02C	C1 ≠ C2
Neurological, Yes	215 (36.01)	319 (45.64)	72 (40.68)	27 (38.03)	< .01C	C1 ≠ C2
Hepatic, Yes	1 (0.17)	5 (0.72)	3 (1.69)	0 (0.00)	.12F
Respiratory, Yes	25 (4.19)	27 (3.86)	11 (6.21)	4 (5.63)	.46F
Digestive, Yes	100 (16.75)	110 (15.74)	15 (8.47)	5 (7.04)	.01C	C1 ≠ C3
Ocular, Yes	38 (6.37)	70 (10.01)	22 (12.43)	4 (5.63)	.02C	C1 ≠ C3
Cardiac, Yes	54 (9.05)	82 (11.73)	21 (11.86)	12 (16.90)	.15C
Metabolic, Yes	13 (2.18)	17 (2.43)	8 (4.52)	0 (0.00)	.19F
Dermatological, Yes	1 (0.17)	11 (1.57)	0 (0.00)	0 (0.00)	.02F	C1 ≠ C2

Abbreviations: C, χ² test; F, Fisher test used for global comparison; KW: Kruskal–Wallis test.

The severity of SRs differed significantly within the clusters with a higher proportion of severe SRs in polydrug users (56.6%).

In terms of adverse events, the cannabis users group was mainly characterized by psychiatric complications (psychotic symptoms 12.2%, mania‐inducing 10.9%, behavioural disorders 18.4%, mood disorders 14.1%, suicidal behaviour 7.9%) and digestive disease (16.8%). The polydrug users cluster was associated with organic or addiction‐related complications since it listed the highest proportion of SUD at 63.8%.

3.2. DRAMES surveys database

During the study period, 44 deaths were collected in the DRAMES Surveys involving young people with a mean age of 16.4 ± 0.7 years (range: 15–17 years) and predominantly male (61.4%). History of substance abuse is reported for 24 subjects (54.5%), mainly cannabis (n = 16), MDMA (n = 8), cocaine (n = 6), opioids (n = 6), hallucinogens (n = 2), and amphetamine‐related (n = 1).

When available, the place of death was at home (n = 19), in hospital (n = 11), on public roads (n = 3) and at a festive venue (n = 1).

Indirect deaths accounted for 8 cases (18.2%; e.g. drowning, fall and public road accident). The remaining cases (direct deaths, n = 36, 81.8%) were deaths directly linked to the consumption of PS, mainly MDMA, cocaine, methadone and heroin. NDS were reported in 70.5% of deaths. All the PS responsible for direct deaths are reported in Figure 5. In 68.2% of deaths, only 1 substance was found to be responsible for the death.

FIGURE 5.

Main psychoactive substances involved in direct deaths in the DRAMES survey.

4. DISCUSSION

In this study, we analysed adverse events and deaths related to the consumption of PS among adolescents. These cases were routinely collected via a multisource approach including spontaneous Addictovigilance reports and deaths recorded by forensic doctors and/or toxicology laboratories. ⁹ , ¹¹ Multiple detection systems seems the current best practice to assess misuse, abuse and diversion on psychoactive substances and identify signals, in various populations as adolescents. ¹⁰ Over the study period, we observed an increase in the proportion of severe cases and a rise in the respective number of SRs linked to most substances from 2018 to 2019 (Figures 1, 2, 3). This is of major interest to us because, at the same time, a French study demonstrated that all levels of PS use declined between 2017 and 2022 in adolescents with an average age of 17 years. ⁵ It is possible that only the most serious cases (justifying hospitalization) and attributed to the use of substances recently identified as risky (pregabalin, nitrous oxide) are preferentially reported to the FAN, explaining these results.

4.1. Adverse events

The main adverse events identified through this analysis were psychiatric problems (63.3%), and SUD (n = 803, 52.0%). Somatic adverse events accounted for 56.4% of cases including mainly neurological disorders (41.0%), digestive diseases (14.9%) and cardiovascular problems (11.0%).

The psychiatric disorders identified in this study are at higher levels than those identified by the World Health Organization, which estimates that 14% of adolescents suffer from mental disorders. Of these, 0.1% suffer from schizophrenia, between 1.4 and 3.5% from depression, between 4.4 and 5.5% from anxiety disorders and between 1.9 and 3.5% from behavioural disorders. ¹

4.2. Clusters analysis

The original analysis of SR data into user clusters made it possible to identify 4 distinct clusters and to highlight particularities.

The first cluster mainly concerned consumers of cannabinoids: cannabis and almost all consumers of SCs. Even if cannabis use can induce severe complications (20% of users), SCs are emerging drugs of abuse with desired and adverse effects considered to be more intense than those observed with cannabis smoking, which is partly explained by the full agonist activity and higher affinity for cannabinoid receptors. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ In our results, a centre effect is observed for SCs since cases originate from the north of France or overseas territories (Mayotte Island) with a substance called chimique that contains SCs. Over the last 10 years, use of this substance has emerged among adolescents and young adults in precarious situations living on Mayotte Island and is associated with neurological and psychiatric disorders. ¹⁷ In our study, the first adverse events identified were psychiatric and neurological. Young individuals are particularly sensitive to the psychological effect of PS due to their still developing brains with immature neural circuits. ¹⁸ Mental health issues, such as anxiety and psychological distress, appear to be important correlates of cannabis use disorder and should be addressed in substance misuse prevention programming. ¹⁹

In this cluster, digestive symptoms could be linked to cannabis hyperemesis syndrome, associating violent abdominal pain and uncontrollable vomiting in a context of heavy cannabis consumption. ²⁰

The second cluster involves adverse events related to consumption of medications and some particular NDS such as nitrous oxide and cannabidiol. As shown in Figure 3, SRs related to usage of the main reported medications have been rising since 2019, especially benzodiazepines (including Z‐drugs) and gabapentinoids (almost exclusively pregabalin). Over the COVID period, benzodiazepines and Z‐drugs were expected to be highly consumed due to social isolation and psychological disorders. ²¹ Two types of users were typically identified: those consuming these drugs for therapeutic purposes and those taking them solely for recreational effects. ²² , ²³ More recently, a signal of misuse was identified with pregabalin, widely utilized for epilepsy, neuropathic pain and generalized anxiety. ²¹ , ²⁴ Pregabalin is a γ‐aminobutyric acid (GABA) analogue with no action on its receptor. ²⁴ Its mechanism of action is through binding to α2 voltage‐dependent calcium channels in the central nervous system, thereby reducing central neuronal excitability. ²⁴ , ²⁵ In our study, pregabalin was associated with benzodiazepine abuse, mainly clonazepam. Increased recreational use of pregabalin in adolescents was observed in another study, mainly in homeless adolescents or those living in migrant shelters (81%) as per our study. ²⁶

Other medications include opioids, mainly codeine and tramadol, 2 weak prescription opioids whose consumption is increasing and which seems to be trivialized. In France, codeine misuse has been increasingly occurring within the context of the purple drank phenomenon among adolescents and young adults. In 2017, the reporting of serious intoxications (including 2 deaths in minors) prompted the Ministry of Health to designate codeine, as prescription‐only drugs on 12 July 2017. ²⁷ In many countries including France, the nonmedical use of tramadol for its psychoactive effects is associated with major public health problems, and can in some cases lead to severe tramadol use disorders and deaths from overdose. ²⁸ In the data analysed in the DRAMES Survey, tramadol was attributed to 1 death, without other associated substances. In the entire DRAMES survey, tramadol is the second opioid antalgic implicated in deaths after morphine. ²⁸

In this second group, 2 substances pose questions or are especially harmful. The first drug involved is cannabidiol (CBD) accounting for 14.4% of substance use in this group, without any other associated substance. The main complications related to this use were psychiatric, neurological or general (dizziness, malaise). Contrary to some beliefs, CBD is a psychoactive substance via a pleiotropic action, particularly on serotonin receptors, on dopamine D2 receptors, on GABA receptors, on the transient receptor potential channel, as well as on glutamate systems, ¹⁶ , ²⁵ which could explain certain neurological and psychoactive effects (sedation, drowsiness, anticonvulsant effect). ²⁹ The consumption of CBD is prevalent among French adolescents. Indeed, 17.1% of 17‐year‐olds in 2022 said they had already tried it and 14.0% had used it in the last 12 months who were mainly boys (15.8 vs. 12.2% in girls) as in our study (82.2% of boys). ⁵ Cannabidiol is increasingly used by vaping, which can democratize its use despite the risks. ³⁰ Addictovigilance data on vaped substances (excluding nicotine) shows a discrepancy between labelling and content with the sale of SCs under the name CBD. ³⁰ In case of doubt, a specific toxicological analysis can specify the product contained and allow a risk reduction approach.

Other substances widely taken by adolescents are volatile solvents, including a recent problematic substance, nonmedical nitrous oxide. ³¹ Solvents are cheap and present in legal, affordable and easily accessible industrial and household products, including deodorants, gasoline and glues. These substances can be intentionally inhaled to achieve psychoactive effects and their repeated administration has deleterious effects on mood and cognitive functions. ³¹ Presently, in France as in many other countries, the abuse of nonmedical nitrous oxide, particularly in very young people has become a public health issue over the last few years: 5.4% of high school students surveyed had experimented with nitrous oxide in 2022. ³² This misuse is associated with a wide variety of sometimes‐dramatic complications including SUD, neurological disorders, psychiatric symptoms and cardiovascular events. ³³ These adverse events may be linked to specific pharmacological properties of nitrous oxide such as: (i) functional inhibition of vitamin B12, which induces neurological deficits, risks of thrombosis (via the secondary increase in homocysteine); and (ii) an increase in dopamine levels (via the GABAergic effect), which can lead to psychiatric disorders. ³⁴ In our study, adverse events related to nitrous oxide are mainly neurological disorders and have been observed since 2019. In 2021, because of this problematic misuse, the French health authorities decided to ban the sale of nitrous oxide cartridges to minors: our study timeframe does not allow us to investigate the effects of this measure on the SRs received by the FAN.

The third cluster identified in our study is, in our opinion, the most problematic as it regards young polydrug users (98% of subjects in this cluster) and the consumption of illicit PS, mainly cocaine and MDMA that are combined with medications and a higher frequency of SUD (63%). Although levels of illicit drug use show a notable decrease compared to 2017, MDMA remains the substance most commonly used by 17‐year‐olds, with a lifetime use rate of 2.0%, followed by cocaine (1.4%). ⁵ MDMA is popular among young users for its positive effects, which include euphoria, excitement, relaxation, increased sociability and closeness to others. Via increased release of serotonin and dopamine, MDMA can also induce undesirable effects potentially being life‐threatening (serotonin syndrome, cerebrovascular accidents, acute anxiety and panic disorder) or even leading to death. ³⁵ , ³⁶ , ³⁷ These data are supported by results of the DRAMES Survey in which MDMA is notably attributed to deaths and especially mono‐intoxication (8/10 cases). It is important to prevent the use of this substance in social party environments (e.g. techno music festivals) by creating risk reduction strategies in order to limit these severe complications and deaths.

This group is particularly worrying because it includes both polydrug users, therefore at greater risk of poisoning and/or death, but also subjects with identified SUDs. Despite their young ages, they are no longer used recreationally but present a real pathology requiring specific treatments (pharmacological and/or psychotherapeutic).

Finally, the fourth cluster affects younger subjects who almost exclusively misused psychotropic medications with psychiatric and somatic disorders. This cluster, which may seem a priori less problematic since it is more integrated, highlights the problem of easy access to psychotropic medications by young people and more particularly, young girls, whether by prescription, access to the family pharmacy or by friends. ³⁸ In studies, the reasons put forward to explain these uses are the management of stress, sleep, managing physical and/or psychological suffering and for stimulants (a few cases of methylphenidate in our study) to enhance academic performance. ³⁹ The misuse of medications makes some of the young people affected more vulnerable to adopting risky sexual practices, attempting suicide or using illicit drugs. ³⁹ These changes in behaviour can have multiple explanations. A recent French study highlighted an increase in depressive syndromes, particularly among women in the 15–18 years age group: this may explain the abusive use of benzodiazepines or other psychotropic medications. ⁴⁰ The misuse of benzodiazepines in a depressive context can induce a worsening of depression and suicidal thoughts. Furthermore, benzodiazepines, due to their sedative and amnesic effects (GABAergic effect), ¹⁶ can increase the vulnerability of young users to a risk of chemical submission. Finally, repeated use of psychotropic drugs (benzodiazepines, opioids or, more rarely, antidepressants) can induce an SUD via the activation of dopaminergic pathways and the reward circuit.

4.3. DRAMES survey

The analysis of data from the DRAMES survey completes the data from SRs. The 44 deaths recorded in DRAMES in minors represent 1.46% of all deaths analysed by the DRAMES survey over the period (n = 2999). ¹¹ Our results highlighted some signals of risks associated with PS. In addition to the signal of risk on MDMA or tramadol discussed above, other risks concern deaths linked to methadone (10 deaths including 7 for which methadone was the only substance found). This trend is also found in the global DRAMES Survey, with methadone being the substance most frequently attributable to deaths. ¹¹ Methadone is a narcotic, full μ‐receptor agonist used as an opioid maintenance therapy. ¹⁶ In opioid‐intolerant patients, 1 mg/kg can be lethal. In the DRAMES Survey, the history of methadone consumption is rarely reported: however, we can think that given the age of the users, it was more likely to constitute recreational methadone use. ⁴¹

The deaths identified in the survey do not constitute all of the deaths recorded in France in this population as not all deaths are investigated by medicolegal toxicology. Nonetheless, they represent unacceptable premature deaths and constitute an important warning on the uses of these highly active substances in this population. Emphasis must be placed on prevention among these young populations who are not always aware of the risks involved and on reduction of risks in social festival environments, for example. Most deaths are unintentional and could potentially be prevented through implementation of a range of harm reduction strategies, including mobile medical care, drug checking services, diffusion of take‐home naloxone, and increased consumer education and awareness. ⁴² , ⁴³

4.4. Strengths and limitations

The major strength of this study lies in the use of different tools to highlight the risks associated with the use of psychoactive substances among minors: however, these data sources have some limitations. As for all studies relying on spontaneous notifications, underreporting and limited information are the main drawbacks when analysing SRs. Moreover, a control group for mental disorders in adolescents were not available to compare with the results from the current study. Finally, the exposure data were not available in the SRS and consequently, no conclusion can be drawn with regard to the implications of chronic vs. sporadic use of the psychoactive substances; it was not possible to establish a threshold beyond which the consumption becomes deleterious.

The main restricting factor of the DRAMES Survey is that it is based on the willingness of expert toxicologists to report the cases and does not therefore provide a complete description of all deaths related to psychoactive substances in France. Furthermore, only the judicial authorities can decide to carry out toxicological investigations.

5. CONCLUSION

This study made it possible to highlight the main adverse events associated with the consumption of psychoactive substances and to characterize 4 groups of young consumers. Given the small number of subjects younger than 12 years included, our data focuses on those over 15 years in which this problem of consumption of psychoactive substances seem to be more worrying.

Despite a significant decline in the consumption of the main PS among adolescents observed in French studies, the adverse events identified in Addictovigilance are worrying and highlight problematic uses. The emergence of risk signals for certain substances shows the need for increased monitoring and the implementation of prevention campaigns adapted to each group of consumers. Recreational drug use has become an increasing societal norm within drug‐using peer groups and cannot be completely avoided. Developing knowledge and reducing the risks (e.g. those associated with the use of cannabinoids and nitrous oxide due to recent safety signals identified), providing information on the use of nonprescription medications, reducing the stigma associated with drugs and psychiatric illnesses are all avenues of prevention of substance abuse consumption among adolescents.

AUTHOR CONTRIBUTORS

H.P. and A.B.: conception of the work, acquisition, analysis and interpretation of data, writing—original draft and writing—review; N.F.S., A.D., C.F.C., S.D., C.C., C.L., V.G., A.A., S.P. and E.J.: acquisition and analysis of data, writing—review; M.C.P. and E.N.: methodology, analysis of data, writing—review. All authors reviewed and approved the final version to be published.

ACKNOWLEDGMENTS

The Authors thank the MILDECA (Mission interministérielle de lutte contre les conduites addictives) for their financial support.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Peyrière H, Batisse A, Sam‐Laï NF, et al. Contribution of Addictovigilance data to assess adverse‐events linked to psychoactive substances in children and adolescents. Br J Clin Pharmacol. 2025;91(10):2957‐2968. doi: 10.1002/bcp.70133

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.