Antipsychotic Abuse, Dependence, and Withdrawal in the Pediatric Population: A Real-World Disproportionality Analysis

Antipsychotic drugs (APs) aim to treat schizophrenia, bipolar mania, and behavioral symptoms. In child psychiatry, despite limited evidence regarding their efficacy and safety, APs are increasingly subject to off-label use. Studies investigating addictology-related symptoms in young people being scarce, we aimed to characterize the different patterns of AP misuse and withdrawal in children and adolescents relying on the WHO pharmacovigilance database (VigiBase®, Uppsala Monitoring Centre, Sweden). Using the standardized MedDRA Query ‘drug abuse, dependence and withdrawal’, disproportionality for each AP was assessed with the reporting odds ratio and the information component. A signal was detected when the lower end of the 95% confidence interval of the information component was positive. Results revealed mainly withdrawal symptoms in infants (under 2 years), intentional misuse in children (2 to 11 years), and abuse in adolescents (12 to 17 years). Olanzapine, risperidone, aripiprazole, and quetiapine were disproportionately reported in all age groups, with quetiapine being subject to a specific abuse signal in adolescents. Thus, in adolescents, the evocation of possible recreational consumption may lead to addiction-appropriate care. Further, in young patients with a history of AP treatment, a careful anamnesis may allow one to identify misuse and its role in the case of new-onset symptoms.


Introduction
Since the discovery of chlorpromazine in France in the 1950s [1], antipsychotic drugs (APs) have aimed to relieve disorganized thoughts and behaviors, hallucinations, and delusions. Accordingly, they are used to treat schizophrenia, bipolar mania, and behavioral symptoms, inter alia. The common mechanism underlying the efficacy of most APs is considered to be the antagonism of brain dopamine D2 receptors [2]. In contrast with older drugs, referred to as 'typical (or first-generation) antipsychotics' (TAPs), the more recent ones ('atypical (or second-generation) antipsychotics', AAPs) are characterized by a stronger binding affinity to the serotonin 5HT2 receptor, compared to dopamine D2 receptors [3].
In child psychiatry, despite limited evidence regarding their efficacy and safety profiles, APs are increasingly subject to off-label use [4][5][6]. This phenomenon stems from a global increase in the prescribing of APs [7,8], a restricted pattern of marketing authorizations, as well as the lack of guidelines for their use in this population [9,10]. Aside from drug tolerance (which frequently leads to an increase in dosage), patients treated with APs are at risk of drug dependency [11][12][13]. Whether resulting from prescription or recreative purposes (such as seeking euphoria or relaxation), this consumption may lead to abuse, intentional misuse, but also withdrawal phenomena [14,15]. In younger populations, 'pharming', which involves the non-medical use and misuse of, mainly psychoactive, medication [16], may be favored in the growing role of the Internet and social media, especially regarding the easier accessibility of drugs [17,18]. Accordingly, the 'psychonauts' [14,16], 'pharming' users that try various psychoactive drugs and then share their experiences on social media, appeal to an ever-expanding audience, potentially influencing children and adolescents.
Studies investigating addictology-related symptoms in young people being scarce, beyond reports of accidental overdosage, we tried to characterize the different patterns of antipsychotic misuse and withdrawal, relying on an analysis of the World Health Organization (WHO) safety database (VigiBase ® , Uppsala Monitoring Centre, Sweden) [19]. While withdrawal cases in infants and abuse cases in adolescents might be expected, there is still a grey area regarding middle-aged children. As antipsychotics are more and more subject to off-label use and illicit consumption in children and youth, we also aimed to identify potential drug safety signals regarding antipsychotic-related abuse, dependence, and withdrawal in this population. Further, we tried to shed some light on the most involved drugs, while suspecting the existence of different consumption patterns for each one, depending on the age group.

Data Source
The WHO mandates the Uppsala Monitoring Centre (UMC) to oversee drug safety [20]. This independent center aims to gather evidence about adverse drug reactions (ADRs), therefore leading to the identification of safety signals [21]. Indeed, VigiBase ® (UMC, Sweden), the WHO safety database, collects Individual Case Safety Reports issued by more than 172 national pharmacovigilance network members, along with pharmaceutical companies. The preservation of the anonymity of patients and notifiers is ensured by VigiBase ® (UMC, Sweden) [19].
Each individual Case Safety Report features sociodemographic characteristics of the patients (age, sex), administrative information (country, reporter qualification), suspected drug (indication, start and cessation dates, dose), concomitant drug(s), and characteristics of the ADR(s)' (effect(s), seriousness, onset, outcome). In pharmacovigilance, an ADR is considered to be serious if it justified a hospitalization or its prolongation, caused a congenital malformation, resulted in persistent or significant disability or incapacity, was life threatening, resulted in death, or required significant medical intervention to prevent one of these outcomes [22,23].

Query
In the Medical Dictionary for Regulatory Activities (MedDRA, version 25.0 [24]), a Standardised MedDRA Query (SMQ) is an exhaustive, validated, predetermined collection of Preferred Terms (PTs) intended to help in investigating drug safety issues in pharmacovigilance [25]. A PT expresses a single medical concept in the most clinically accurate way [24].
Then, queried reports involving patients aged under 18 were classified into 3 age groups: under 23 months, 2 to 11 years, and 12 to 17 years. For each group, records were aggregated depending on the PTs, into one 1 of 3 categories: misuse, abuse, or withdrawal [18].
'Misuse' can be defined as the intentional and inappropriate use of a product other than as prescribed or not in accordance with validated drug labels. It included the PTs 'Intentional overdose' and 'Intentional product misuse'.
'Withdrawal' describes symptoms or signs related to the cessation of a drug, either physiological withdrawal reaction to the drug or exacerbation of the underlying disease itself [26]. It included the PTs 'Drug withdrawal convulsions', 'Drug withdrawal headache', 'Drug withdrawal syndrome', and 'Drug withdrawal syndrome neonatal'.

Statistical Analyses
Quantitative variables were described in terms of means with standard deviations (±SD). Qualitative variables were described with proportions. Statistical analyses were performed using GraphPad Prism version 8.0.2. Thereupon, we performed a disproportionality analysis, to partially mitigate the impact of potential confounding factors [27]. Potential pharmacovigilance signals were sought for all antipsychotics (as described above, see Section 2.2. Query), for each age group:

•
Using the narrow SMQ 'Drug abuse, dependence, and withdrawal' when the chosen drug accounted for ≥2 cases and >1% of the reports of the respective age group; • Using the most represented combination of PTs (in each age group), when the chosen drug accounted for ≥10% of the reports and/or was quetiapine, olanzapine, risperidone, and aripiprazole. This particular analysis aimed to clarify the specific ADRs driving any possible signal arising from the disproportionality analysis of the SMQ.
The disproportionality analysis relied on the reporting odds ratio (ROR) and the information component (IC).
As an approximate of the odds ratio (used in case-control studies), the ROR is estimated, in case-non-case studies, to assess the strength of disproportionality. An ROR equal to 1 indicates the absence of signal: the ADR is equally reported with the drug of interest as with the other drugs. In contrast, an ROR greater than 1 suggests the existence of a signal, as cases appear to be more frequently reported with the drug of interest than with the others. The higher the ROR, the stronger the association. A 95% confidence interval (95% CI) reflects the precision of the approximate ROR. Therefore, an ROR is deemed to be statistically significant when the lower bound of its 95% CI is greater than 1 [28].
The IC allows one to compare observed and expected values for the combination of a given drug and an ADR, to check for a potential association. It favors the reduction in the risk of false-positive signals, especially if the chosen ADR has a very low expected frequency in the database (therefore, mechanically increasing the ROR). The positivity of the IC reflects that the number of observed reports is higher than expected. The IC025, which is the bottom end of the 95% CI of the information component, is required to statistically confirm the detection of a signal in VigiBase ® (UMC, Sweden) [19,29].
In this disproportionality analysis, potential drug-ADR associations were selected by using IC025. Then, the ROR of each drug-effect association allowed us to assess the strength of respective suggested signals. For this purpose, we used Microsoft ® Excel ® 2019 Version 2210.

Drug Abuse, Dependence, and Withdrawal
As of 4 August 2022, 16,054 reports belonging to the narrow SMQ 'Drug abuse, dependence and withdrawal' and involving consumers of antipsychotics were collected in VigiBase ® (UMC, Sweden). Among these reports, 1023 (6.4%) involved patients below 18 years of age, mostly belonging to the 12-to-17-year group (732, 71.6%). Records with patients aged under 18 mostly originated from the United States (398, 38.9%). Healthcare professionals issued 84.8% of the cases, with a majority of physicians (617, 60.3%). Details regarding the characteristics of the reports are provided in Table 1.  Four antipsychotics accounted for 839 (82.0%) reports of abuse, dependence, and withdrawal: quetiapine (368, 36.0%), risperidone (224, 21.9%), aripiprazole (129, 12.6%), and olanzapine (118, 11.5%). The number of reports for all ADRs in patients aged under 18 involving each of these antipsychotics is displayed in Table S1.
In patients aged under 18, antipsychotics were subject to a disproportionate reporting for the SMQ 'Drug abuse, dependence and withdrawal' (ROR 5.5; IC025 2.2).
The most often involved antipsychotics (absolute number of reports) were disproportionately reported with the SMQ 'Drug abuse, dependence, and withdrawal': quetiapine ( Figure 1 summarizes the disproportionality analysis for the main antipsychotics involved in reports of abuse, dependence, or withdrawal, depending on the age range. Upon closer inspection, in the narrow SMQ 'Drug abuse, dependence, and withdrawal', most reports involving infants aged between 0 days and 23 months referred to PTs related to withdrawal (192, 96.7%), as displayed in Figure 2 and  Table 3.
The most frequently involved antipsychotics were disproportionately reported with the narrow SMQ 'Drug abuse, dependence and withdrawal': quetiapine (ROR 8.9; 95% CI: Most of these reports referred to PTs related to intentional misuse (545, 74.5%) and abuse (168, 23.0%), as displayed in Figure 2 and Table S7.

Discussion
Our analysis of the WHO pharmacovigilance database brings to light varying profiles for abuse, dependence, and withdrawal related to antipsychotics. Indeed, in infants aged under 2 years, almost all ADRs were understandably related to withdrawal symptoms. In children aged between 2 and 11 years, intentional misuse was at the forefront, with fewer cases of withdrawal, frequently complicated by extrapyramidal symptoms. Lastly, adolescents aged from 12 to 17 years were subject to intentional misuse, but also to abuse issues.
In infants aged from 0 days to 23 months, withdrawal syndrome reflects prenatal maternal exposure to APs. AAPs (quetiapine, risperidone, aripiprazole, olanzapine) prevailed, both in terms of absolute number of cases and disproportionality, which is consistent with existing literature regarding their respective safety profiles and prescribing trends during pregnancy [30,31]. However, their drug labels indicate a risk of withdrawal in exposed neonates and recommend avoiding these drugs during pregnancy, unless their use is an absolute necessity [32][33][34][35]. Quetiapine showed the strongest signal in this population. Indeed, in women previously stabilized with quetiapine, it often remains prescribed during pregnancy [36,37]. At birth, the newborn being severed from the maternal blood supply and quetiapine being rapidly dissociated from the D2 receptors [38,39], its abrupt discontinuation may lead to greater risks of withdrawal at birth [40]. The frequent maternal cotreatment with antidepressants (e.g., mood disorders) that we observed may also reinforce withdrawal symptoms in infants [40][41][42].
In children aged between 2 and 11 years, the age and gender distribution regarding treatment with antipsychotics in the literature concurs with our findings [43][44][45]. Indeed, significant disproportionality was found for AAP only, subject to a dramatic increase in consumption in children, circumscribed by the rise of off-label use in this population [43,44,46,47]. Regarding risperidone, which is the leading drug in terms of absolute number of reports, both the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) have granted it marketing authorization in irritability occurring in patients with an autistic disorder from the age of 5 [35,48]. Aripiprazole is authorized by the FDA from the age of 6 [34] in the same indication, but its use is only recommended from the age of 13 by the EMA (manic episodes of bipolar I disorder) [49]. Lastly, marketing authorization for olanzapine and for quetiapine is granted by the FDA in acute mixed or manic episodes of bipolar I disorder from the age of 10 [32,33]. However, the EMA stated that their safety and efficacy in children under the age of 18 have not been established yet and, therefore, that they should not be used in this group until further data become available [50,51]. In more than one-quarter of the reports, abuse, dependence, or withdrawal occurred concomitantly with dystonia and/or dyskinesia. These effects may reflect either overdosing following intentional misuse or withdrawal syndrome [52,53], subsequent to dopamine receptor hypersensitivity, GABA insufficiency, or cellular degeneration (neurotoxicity) [54,55]. In cases of withdrawal, the strongest disproportionality signals concerned risperidone and aripiprazole. This may indicate abrupt treatment discontinuations that could be favored by the urge of appeasing irritability and/or aggressive symptoms via the initiation of a new medication (e.g., in patients suffering from autistic disorders). In addition, possible impulsive and voluntary intoxications may also occur, especially in children presenting with behavioral disorders.
Adolescents aged from 12 to 17 years accounted for the vast majority of the reports belonging to the SMQ 'drug abuse, dependence and withdrawal'. Paracetamol was quite frequently co-reported as suspect or interacting, which may reflect the fact that suicide attempts accounted for nearly one-third of the reports in this age range [56]. Further, ingestion of co-reported substances, such as antidepressants and benzodiazepines, might also concur with phenomena of intentional misuse or abuse or designate a population of patients receiving long-term treatment with psychotropic drugs. In this context, the AAPs quetiapine, olanzapine, risperidone, and aripiprazole were subject to disproportionate reporting, to a greater degree than TAPs, such as promazine, chlorprothixene, and cyamemazine. Regarding TAPs, cyamemazine has been granted marketing authorization in France (Agence Nationale du Médicament et des produits de santé-ANSM) for behavioral disorders with psychomotor agitation and aggressivity for children aged 3 years or older [57]. Promazine and chlorprothixene were withdrawn from the market in the United States [58,59]. However, commonly prescribed AAPs may be ingested for recreative purposes, as Novel Psychoactive Substances (NPSs) [12,18]. According to the United Nations Office on Drugs and Crime (UNODC), NPSs are 'substances of abuse, either in a pure form or a preparation, that are not controlled by the 1961 Single Convention on Narcotic Drugs or the 1961 Convention on Psychotropic Substances, but which may pose a public health threat' [60]. Compared to 'conventional' illicit substances, NPSs are distinguished by greater affordability and ease of online purchase, but also lower detectability and social stigma [61]. In younger populations, both appeasement of other consumption symptoms and search of emotional anesthesia are becoming more and more accessible, as these phenomena are further exacerbated by the 'pharming' and 'psychonauts' trends. As abuse and intentional misuse have been described for risperidone and aripiprazole [13,62], we extend this potential signal to adolescent populations. In line with previous findings (irrespective of the age range) [13,15], olanzapine was disproportionately reported for intentional misuse. Olanzapine can be consumed as a psychedelic or for its sedative effect, with rewarding properties involving glutamatergic stimulation of dopaminergic neurons of the ventral tegmental area [63]. Our findings confirm that quetiapine, already involved in pharmacovigilance signals of intentional misuse and abuse, raises also issues in adolescents. Indeed, it is the leading drug in terms of absolute number of reports. Quetiapine has multiple street names, such as 'quell', 'Susie-Q', 'baby heroin', and 'Q-ball' (when associated with cocaine) [64]. Consumed for anxiolytic properties, its popularity as an NPS may be linked to H1 and α1 receptor antagonism. The same rationale may apply to olanzapine, more and more available on the black market [13,64,65].
While other studies aimed to assess the importance of antipsychotic abuse and misuse in different pharmacovigilance databases [13,15,66], our analysis was the first to focus on children and youth. It highlighted different profiles, depending on the age of the patients. Nevertheless, this study is hindered by the inherent flaws of spontaneous reporting systems and post-marketing pharmacovigilance approaches, such as incomplete data and reporting bias. The strict definitions of the terms, from a pharmacovigilance perspective, might be unclear for notifiers, which may have led to a substantial coding heterogeneity, especially regarding intentional misuse and abuse. However, most reports were notified by healthcare professionals, which limits the risk of coding errors. Further, lack of follow-up and underreporting, although usual in pharmacovigilance [67], may have been heightened by the extent of overall consumption and the acute nature of manifestations (especially related to withdrawal or intentional overdose). Then, associated medications were recorded (coreported active ingredients), but we were not able to retrieve data on the consumption of other illegal substances, as this is outside the scope of VigiBase ® (UMC, Sweden) Lastly, pharmaco-epidemiological studies aim to raise awareness about possible drug safety signals and no definite causality can be drawn from our findings.
Regarding implications for practice and research, different areas have to be taken into account. During pregnancy, as recommended by the safety agencies [32-35, [48][49][50][51], each antipsychotic treatment must be introduced or maintained following a careful assessment of the benefit/risk ratio. In children, off-label prescription should be properly substantiated, considering the scarcity of data regarding safety in efficacy in this population [50,51]. In adolescents, the risk of antipsychotic abuse and its characteristics should be further investigated. In addition, particular heed should be paid to specialized social networks, underlying new consumption patterns [14,16].

Conclusions
In this study, relying on a comprehensive analysis of drug abuse, dependence, and withdrawal reports with APs from the WHO safety database, we suggest potential pharmacovigilance signals involving several AAPs, especially quetiapine, olanzapine, risperidone, and aripiprazole. Defining three profiles according to the patients' age ranges, we highlighted the potential role of these drugs in withdrawal phenomena among infants, but also in intentional misuse behaviors among children and adolescents. We confirmed existing signals of intentional misuse of APs and the growing importance of quetiapine as an NPS. In young patients with a history of AP treatment, a careful anamnesis may allow one to identify the role of these medications in the case of new-onset symptoms. Indeed, given the variety in their manifestations, the hypothesis of an antipsychotic-related ADR should be systematically evoked. In adolescents, the evocation of a possible recreational consumption may lead to addiction-appropriate care, therefore, reducing further morbidity.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines10112972/s1, Table S1: Reports of any adverse drug reaction involving risperidone, aripiprazole, quetiapine, and olanzapine in children and adolescents (<18 years old); Table S2: Reported suspected antipsychotics in patients aged between 0 days and 23 months with abuse, dependence, or withdrawal; Table S3: Main Preferred Terms (PTs) reported in cases of abuse, dependence, or withdrawal in patients aged between 0 days and 23 months; Table S4: Reported suspected antipsychotics in patients aged between 2 and 11 years with abuse, dependence, or withdrawal; Table S5: Main PTs reported in cases of abuse, dependence, or withdrawal in patients aged between 2 years and 11 years; Table S6: Reported suspected antipsychotics in patients aged between 12 and 17 years with abuse, dependence, or withdrawal; Table S7: Main PTs reported in cases of abuse, dependence, or withdrawal in patients aged between 12 years and 17 years.

Institutional Review Board Statement:
In accordance with European regulation, this observational study did not need approval from an institutional review board/independent ethics committee. The views expressed in this article are the authors' personal views and may not be understood or quoted as being made on behalf of or reflect the position of the ANSM, the EMA, or one of their committees or working parties.

Data Availability Statement:
The data that support the findings of this study are available from Uppsala Monitoring Center (UMC) but restrictions apply to the availability of these data, which were used under license for the current study and so are not publicly available. Access to VigiBase ® is available without fees to Fanny Rocher. Data are, however, available from the authors upon reasonable request and with permission of UMC.