Trends in antipsychotic drug utilisation and indication in the paediatric population: a nationwide cohort study in Sweden Authors: Karl-Hermann S Kamgang, Bjorn Pasternak, Johan Reutfors, Richeek Pradhan, Viktor Wintzell Published: 2026 ABSTRACT AbstractBackgroundData on antipsychotic use for the Swedish paediatric population is limited.ObjectiveWe assessed the incidence, likely indication and extent of off-label antipsychotic use among children and adolescents in Sweden.MethodsIn this nationwide register-based study, individuals younger than 18 years with at least one antipsychotic dispensing between 1 January 2008 and 31 December 2021 were identified. Yearly age-standardised and sex-standardised incidence per 1000 persons was estimated and temporal trends were assessed using Poisson log-linear regression. We also assessed the likely indication and the proportion of on-label and off-label use.FindingsAmong 4 108 171 individuals who were children or adolescents at some point during the study period, 24 742 initiated antipsychotic treatment at least once, contributing 25 576 incident dispensings during the study period. Incidence increased from 0.48 to 1.28 per 1000 persons (risk ratio, RR per year 1.08, 95% CI 1.05 to 1.11; cumulative increase 180%), with a more pronounced rise in females (0.43 to 1.51 per 1000; RR 1.11, 95% CI 1.10 to 1.12; 284%) than males (0.53 to 1.06 per 1000; RR 1.06, 95% CI 1.05 to 1.07; 113%), driven by adolescents aged 12–17 years. The largest drug-specific increases were observed for quetiapine (0.04 to 0.33 per 1000; RR per year 1.18, 95% CI 1.17 to 1.19; cumulative increase 804%) and aripiprazole (0.03 to 0.36 per 1000; RR per year 1.16, 95% CI 1.14 to 1.18; cumulative increase 608%). A likely indication was identified in 64.3% (n=16 455) of incident dispensings, most commonly autism spectrum disorder (15.4%), attention-deficit hyperactivity disorder (10.9%) and anxiety disorders (9.6%). Among those with an identifiable indication, 81% were dispensed off label, mainly due to non-approved indications (64.8%) and most frequently among adolescent females.ConclusionsIn the paediatric population in Sweden, incidence of antipsychotic use increased between 2008 and 2021, especially among adolescent females and most incident dispensing with an identifiable indication were off label, underscoring the need to evaluate the long-term effectiveness and safety of antipsychotics in children and adolescents.Clinical implicationsThese patterns highlight the need to evaluate the long-term effectiveness and safety of antipsychotics in children and adolescents.Keywords: Adolescent, Psychotic Disorders WHAT IS ALREADY KNOWN ON THIS TOPIC Antipsychotic use in children and adolescents has increased substantially in many countries, with a large proportion prescribed off-label for conditions such as attention-deficit hyperactivity disorder, mood disorders and anxiety despite limited evidence and safety concerns. In Sweden, previous nationwide studies have mainly described trends in prevalent use up to 2017 and have not examined national trends in incident use, indication patterns at initiation or the extent of off-label prescribing in this population. Antipsychotic use in children and adolescents has increased substantially in many countries, with a large proportion prescribed off-label for conditions such as attention-deficit hyperactivity disorder, mood disorders and anxiety despite limited evidence and safety concerns. In Sweden, previous nationwide studies have mainly described trends in prevalent use up to 2017 and have not examined national trends in incident use, indication patterns at initiation or the extent of off-label prescribing in this population. WHAT THIS STUDY ADDS In a nationwide cohort of Swedish children and adolescents (0–17 years), the incidence of antipsychotic use increased from 0.48 to 1.28 per 1000 persons between 2008 and 2021, particularly among adolescent females, driven mainly by an increase in aripiprazole, quetiapine and olanzapine use. We identified marked sex-specific indication patterns: males were predominantly treated for autism spectrum disorder, conduct disorder and attention deficit- hyperactivity disorder while females were primarily treated for anxiety, mood and eating disorders. For incident dispensings with an identifiable indication, off-label use accounted for 81% of dispensings, reflecting the lack of approved treatments for conditions. In a nationwide cohort of Swedish children and adolescents (0–17 years), the incidence of antipsychotic use increased from 0.48 to 1.28 per 1000 persons between 2008 and 2021, particularly among adolescent females, driven mainly by an increase in aripiprazole, quetiapine and olanzapine use. We identified marked sex-specific indication patterns: males were predominantly treated for autism spectrum disorder, conduct disorder and attention deficit- hyperactivity disorder while females were primarily treated for anxiety, mood and eating disorders. For incident dispensings with an identifiable indication, off-label use accounted for 81% of dispensings, reflecting the lack of approved treatments for conditions. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY The rising incidence of antipsychotic use, particularly off-label use among adolescent females, underscores the need for careful risk-benefit assessment and updated clinical guidelines addressing the expanding scope of paediatric antipsychotic prescribing. The growing use of quetiapine and olanzapine for anxiety, mood and sleep disorders, despite limited evidence in the paediatric population, highlights the need for further research on their long-term safety and effectiveness. The rising incidence of antipsychotic use, particularly off-label use among adolescent females, underscores the need for careful risk-benefit assessment and updated clinical guidelines addressing the expanding scope of paediatric antipsychotic prescribing. The growing use of quetiapine and olanzapine for anxiety, mood and sleep disorders, despite limited evidence in the paediatric population, highlights the need for further research on their long-term safety and effectiveness. Introduction Antipsychotics are the treatment of choice for psychotic disorders, including schizophrenia and bipolar mania,1 2 with some individual antipsychotics also approved for selected indications such as behavioural disturbances in autism spectrum disorder (ASD) and conduct disorders.3 4 Beyond their approved indications, the broad pharmacological action of antipsychotics on dopamine, serotonin and other receptor systems has driven expanding use for irritability, impulsivity and mood dysregulation across a range of psychiatric conditions.5 Antipsychotic use among children and adolescents has risen substantially over recent decades. A multinational study including data from 16 countries found that prevalence increased in most settings between 2005 and 2014, ranging from a 3% increase in the publicly insured US population (prevalence 20.5 per 1000 by 2012) to a 220% increase in Finland (8.0 per 1000 by 2014).6 In the Nordic region, the same study found that prevalence increased by 69% in Denmark (4.0 per 1000 by 2014) and 33% in Iceland (13.1 per 1000 by 2014) over the same period. More recent data from Nordic countries show a 58% increase in Sweden between 2008 and 2017 (from 1.8 to 2.8 per 1000) and a 45% increase in Norway (from 2.2 to 3.2 per 1000),7 while the incidence among children in Finland rose by 138% over the same period (from 2.1 to 3.8 per 1000).8 Much of this increase has been attributed to use for indications for which antipsychotics have not been approved (off-label), including attention-deficit hyperactivity disorder (ADHD), mood disorders and anxiety disorders.9 10 Despite their widespread use, robust evidence supporting efficacy of antipsychotics for many off-label indications is limited.11 Although off-label antipsychotic use is well-documented,9 few studies have investigated the indications driving such use in the paediatric population.10 12 Swedish data on paediatric antipsychotic use remains limited, with the most recent data extending only through 2017.6 7 13 There is limited nationwide evidence characterising trends in incident antipsychotic use, the underlying psychiatric indications at treatment initiation and the extent of off-label prescribing among Swedish children and adolescents. Considering the gaps, this study aimed to describe these trends among Swedish children and adolescents (aged 0–17 years) between 2008 and 2021. Methods Data sources This was a cohort study using linked data from multiple Swedish national registers with nationwide coverage. Registers were linked based on the unique personal identification number assigned to each individual in Sweden.14 The Total Population Register includes all individuals registered in Sweden, with data on date of birth, sex and migration status.15 The Multi-Generation Register linked children and adolescents to parents.16 The National Patient Register contains information on all inpatient care and specialist outpatient care, including dates of visit, primary and secondary diagnoses (International Statistical Classification of Diseases, V.10 (ICD-10)) and surgical procedures.17 The Prescribed Drug Register covers all dispensed prescription drugs nationwide, with information on dispensing date and Anatomical Therapeutic Chemical (ATC) codes.18 This register covers prescriptions from primary and specialised outpatient care settings. In Sweden, any licensed medical doctor can prescribe antipsychotic drugs. However, prescriptions for children and adolescents are mainly issued by child and adolescent psychiatrists and other physicians working in specialist psychiatric services. The study was approved by the Regional Ethics Committee in Stockholm, Sweden. Source population and procedures The study population included children and adolescents aged 0–17 years residing in Sweden between 2008 and 2021, identified annually using the Total Population Register within each annual cohort, individuals with antipsychotic dispensings were identified using the Prescribed Drug Register. Prevalent use in a given calendar year was defined as having at least one antipsychotic drug dispensed during that year. For individuals with multiple antipsychotic dispensings in the same year, only the first dispensing was classified as the prevalent dispensing. Thus, an individual could contribute at most one prevalent dispensing in any given calendar year but could contribute prevalent dispensings in several different years. Incident use in a given calendar year was defined as having an antipsychotic dispensing in that year with no antipsychotic dispensed in the 24 months prior to the dispensing date. The same individual could contribute at most one incident dispensing in each calendar year that met this criterion. The ATC codes used to identify antipsychotic drugs are listed in online supplemental eTable 1. The specific indication for prescription of antipsychotics is not available in the Prescribed Drug Register. We instead determined the likely indication for incident antipsychotic dispensing based on the psychiatric diagnoses assigned to patients prior to incident dispensing. All inpatient and specialist outpatient visits within 30 days prior to the date of incident dispensing were retrieved from the National Patient Register. The validity of psychiatric diagnoses in the Swedish National Patient Register, including specialised outpatient care, is generally high, with positive predictive values of approximately 80–90% for several psychiatric disorders, including schizophrenia spectrum and bipolar disorder.19 From all identified visits, the primary and the first two secondary ICD-10 diagnosis codes were retrieved. An algorithm (online supplemental eTable 2) was applied to identify the most likely indication based on the relevant psychiatric diagnosis codes. For incident dispensings with no healthcare visit or no healthcare visit with a relevant psychiatric diagnosis in the 30 days prior, the likely indication was classified as non-identifiable. Among incident dispensings with an identifiable likely indication we determined if the use was on-label or off-label according to the approved indications and age ranges specified in the official Summary of Product Characteristics from the European Medicines Agency20 and the Swedish Pharmaceutical Specialties,21 using product information current as of 2023. An incident dispensing was considered off-label for indication if the identified indication lacked regulatory approval for the dispensed antipsychotic or off-label for age if the individual’s age was outside the approved range (online supplemental eTable 2). Statistical analysis We estimated yearly incidence proportions as the number of incident dispensings divided by the number of individuals aged 0–17 years residing in Sweden in the corresponding calendar year, expressed per 1000 persons. Analysis was stratified by sex, sex by age group (0–4, 5–11, 12– 17 years) and drug type (risperidone, aripiprazole, olanzapine, quetiapine, other atypical and typical antipsychotics), drug type by sex and drug type by sex-age group. Yearly prevalence per 1000 persons was estimated as the number of prevalent dispensings divided by the corresponding annual population (stratified by sex, sex by age group and drug type). This was estimated in a secondary analysis and presented in the online supplemental file 1. To account for changes in the age and sex distribution of the paediatric population over time, direct standardisation was applied using six sex-by-age strata (females and males each within 0–4, 5–11 and 12–17 years). Overall incidence was standardised to the mean sex-by-age group distribution across all study years; sex-specific estimates were age-standardised within females and males separately, using the mean age-group distribution for that sex across all study years. Drug-specific incidence proportions were presented without further standardisation. Temporal trends were assessed using Poisson log-linear regression, with the yearly number of incident users as the outcome and the log of the annual population size as an offset, so that the exponentiated coefficient represents the average annual relative change in the incidence proportion (risk ratio, RR). The overall trend model was adjusted for sex and age group; sex-specific models were adjusted for age group. Baseline characteristics of individuals at incident dispensing were described, including socio-demographic variables, concurrent medication use and healthcare utilisation (online supplemental eTable 3). We reported the yearly number of incident dispensings by likely indication (overall, sex and sex by age group), and by on-label and off-label use (overall, sex and sex by age group). To assess the influence of methodological choices, we repeated the identification of likely indication using alternative look-back windows (60, 90 and 180 days before the index date). This allowed us to examine how the look-back period affected the distribution of the likely indication and the proportions of on-label and off-label use. All analyses were performed in SAS Enterprise Guide 7 and data visualisation was performed in R. Results Study population We identified 4 108 171 individuals who lived in Sweden and were children or adolescents at some point during the study period (1 January 2008 to 31 December 2021). We identified 26 265 children and adolescents who were dispensed at least 1 antipsychotic drug during this period, contributing 57 537 prevalent dispensings. Of these dispensings, 25 576 were incident, contributed by 24 742 children and adolescents (online supplemental eFigure 1). Among these, 13 493 (52.7%) were dispensed to males, the mean age was 13.7 years (SD 3.2) and the age group 12–17 years accounted for 79.1% of the incident dispensings (table 1). Risperidone was the most initiated antipsychotic (n=10 787; 42.2%), followed by aripiprazole (n=5631; 22.0%), quetiapine (n=4354; 17.0%) and olanzapine (n=3892; 15.2%). In the 12 months prior to antipsychotic initiation, the most frequently dispensed drugs were melatonin (39.0%), selective serotonin reuptake inhibitor (36.0%) and ADHD medications (32.1%). Table 1Baseline characteristics of Swedish children and adolescents who initiated an antipsychotic, 2008–2021Incident dispensings(N=25 576)N (%)Age, mean (SD)13.2 (3.2)Age, median (Q1, Q3)14 (11,16)Age group 0–4 years184 (0.7) 5–11 years6671 (26.1) 12–17 years18 721 (73.2)Sex Female12 083 (47.3) Male13 493 (52.7)Antipsychotic drug Risperidone10 787 (42.2) Aripiprazole5631 (22.0) Quetiapine4354 (17.0) Olanzapine3892 (15.2) Other atypical53 (0.2) Typical848 (3.3)Birth country of user Nordic22 991 (89.9) Africa457 (1.8) Asia1141 (4.5) Rest of Europe760 (3.0) Other225 (0.9) Missing<3 (<0.1)Concurrent medication use in past 12 months SSRIs9204 (36.0) Other antidepressants1891 (7.4) ADHD medications8218 (32.1) Lithium0 (0.0) Benzodiazepines879 (3.4) Melatonin9951 (39.0) Z-drugs547 (2.1)No. of outpatient visits for psychiatric conditions in past 12 months 05846 (22.9) 14563 (17.8) 2–36877 (26.9) 4+8290 (32.4)No. of inpatient admissions for psychiatric conditions in past 12 months 020 371 (79.6) 13300 (12.9) 2–31516 (5.9) 4+389 (1.5)Note: The incident use cohort consists of 24 742 children and adolescents contributing 25 576 incident antipsychotic dispensings during the study period.ADHD, attention-deficit hyperactivity disorder; SSRI, selective serotonin reuptake inhibitors; Z-drugs, non-benzodiazepine hypnotics used for insomnia. Baseline characteristics of Swedish children and adolescents who initiated an antipsychotic, 2008–2021 Note: The incident use cohort consists of 24 742 children and adolescents contributing 25 576 incident antipsychotic dispensings during the study period. ADHD, attention-deficit hyperactivity disorder; SSRI, selective serotonin reuptake inhibitors; Z-drugs, non-benzodiazepine hypnotics used for insomnia. Incidence of antipsychotic use Between 2008 and 2021, the number of incident antipsychotic dispensings for children and adolescents increased from 1262 to 3032 dispensings (online supplemental eTable 4). Overall age and sex-standardised incidence increased from 0.48 to 1.28 per 1000 persons, with a Poisson estimated average annual increase of 8.2% and a cumulative increase of 180% over the study period (RR per year 1.08, 95% CI 1.05 to 1.11; figure 1). Among females, incidence increased from 0.43 to 1.51 per 1000, with an average annual increase of 10.9% and a cumulative increase of 284% (RR per year 1.11, 95% CI 1.10 to 1.12; figure 1) driven predominantly by adolescents aged 12–17 years, for whom incidence increased from 0.99 to 3.60 per 1000 (RR per year 1.11, 95% CI 1.11 to 1.12; cumulative increase 264%). Trends in the incidence of antipsychotic drug use for Swedish children and adolescents between 2008 and 2021, presented as the number of users per 1000 persons, by sex, drug type and age. Note: The four most common atypical antipsychotic drugs are presented separately from other atypical antipsychotics which are presented as a group (other atypical), as are typical antipsychotics. Among males a more modest increase in incidence was observed rising from 0.53 to 1.06 per 1000, with an estimated average annual increase of 6.0% (RR per year 1.06, 95% CI 1.05 to 1.07; cumulative increase of 113%). This increase was driven by adolescents aged 12–17 years (0.88 to 1.76 per 1,000; RR per year 1.06, 95% CI 1.06 to 1.06; cumulative increase 114%). Yearly counts of incident users for all age groups by sex are presented in online supplemental eTable 4. For the analysis stratified by antipsychotic drug, the largest estimated average yearly increases were observed for quetiapine (0.04 to 0.33 per 1000; RR per year 1.18, 95% CI 1.17 to 1.19; cumulative increase 804%) and aripiprazole (0.03 to 0.36 per 1000; RR per year 1.16, 95% CI 1.14 to 1.18; cumulative increase 608%), and olanzapine (0.06 to 0.26 per 1000; RR per year 1.11, 95% CI 1.09 to 1.14; cumulative increase 304%). Risperidone initiation remained relatively stable (0.34 to 0.34 per 1000; RR per year 1.01, 95% CI 1.00 to 1.02; cumulative increase 13%), while typical antipsychotic initiation decreased from 0.05 to 0.02 per 1000 (RR per year 0.94, 95% CI 0.91 to 0.97; cumulative decrease 58%). Among females, quetiapine showed the largest average yearly increase (0.05 to 0.50 per 1000; RR per year 1.20, 95% CI 1.18 to 1.21; cumulative increase 960%), followed by aripiprazole (0.04 to 0.38 per 1000; RR per year 1.17, 95% CI 1.15 to 1.19; cumulative increase 677%) and olanzapine (0.09 to 0.42 per 1000; RR per year 1.13, 95% CI 1.10 to 1.16; cumulative increase 378%), with these three antipsychotics surpassing risperidone as the most initiated antipsychotic by 2018 (online supplemental eFigure 3). Risperidone showed no meaningful trend among females (0.25 to 0.22 per 1000; cumulative increase 13%; RR per year 1.01, 95% CI 0.99 to 1.03). Among males, risperidone remained the most initiated drug throughout, though no notable increase in incidence was observed (0.43 to 0.45 per 1000; RR per year 1.01, 95% CI 1.00 to 1.02; cumulative increase 14%). The largest average yearly increase among males was observed for aripiprazole (0.03 to 0.34 per 1000; RR per year 1.16, 95% CI 1.12 to 1.19; cumulative increase 558%) and quetiapine (0.02 to 0.17 per 1000; RR per year 1.15, 95% CI 1.13 to 1.18; cumulative increase 549%), while typical antipsychotics declined from 0.05 to 0.02 per 1000 (RR per year 0.92, 95% CI 0.89 to 0.95; cumulative decrease 67%). Results stratified by drug and age group are presented in online supplemental eFigure 4. Prevalence of antipsychotic use Prevalence of antipsychotic use showed similar patterns to incidence, increasing from 1.16 to 2.81 users per 1000 persons between 2008 and 2021, with an estimated average annual increase of 7.8% (RR per year 1.08, 95% CI 1.06 to 1.10; cumulative increase 165%). Similar to incidence trends, the most pronounced increase in prevalent use was observed among females aged 12–17 years (online supplemental eFigure 2 and eTable 5). Likely indications for incident use A likely indication was identified for 16 455 (64.3%) incident dispensings; the remaining 9121 (35.7%) had no relevant psychiatric diagnosis recorded in the 30 days prior to initiation and were classified as non-identifiable (table 2). Among those with a likely psychiatric indication, the most common identified indications overall were ASD (15.4%), followed by ADHD (10.9%), anxiety disorders (9.6%), mood and depressive disorders (6.2%) and eating disorders (4.5%). Sex-specific patterns were observed (table 2). Among females, the most common likely indication was anxiety disorders (14.9%), followed by ASD (10.9%), eating disorders (9.1%), mood/depressive disorders (8.9%) and ADHD (9.0%). Among males, ASD was the most common likely indication (19.4%), followed by ADHD (12.6%), conduct disorder (6%), anxiety disorders (5.0%) and mood/depressive disorders (3.7%). Trends in incidence by likely indication for females, males and specific antipsychotic drugs are presented as yearly number of users in the online supplemental eFigures 6 and 7. Table 2Likely indication associated with incident antipsychotic use in Swedish children and adolescents, 2008–2021Psychiatric diseaseTotal (N=25 576)Females (n=12 083)Males (n=13 493)Non-identifiable9121 (35.7)3817 (31.6)5304 (39.3)ASD3935 (15.4)1321 (10.9)2614 (19.4)ADHD2780 (10.9)1085 (9.0)1695 (12.6)Anxiety disorder2465 (9.6)1797 (14.9)668 (5.0)Mood and depressive disorder1579 (6.2)1081 (8.9)498 (3.7)Eating disorder1163 (4.5)1097 (9.1)66 (0.5)Conduct disorder1091 (4.3)283 (2.3)808 (6.0)Other psychiatric disorder950 (3.7)457 (3.8)493 (3.7)Schizophrenic and delusional disorders652 (2.5)166 (1.4)486 (3.6)Tic disorder649 (2.5)270 (2.2)379 (2.8)Manic episodes and bipolar disorder602 (2.4)412 (3.4)190 (1.4)Obsessive-compulsive disorder519 (2.0)271 (2.2)248 (1.8)Other developmental disorder70 (0.3)26 (0.2)44 (0.3)Note: For each incident dispensing, we retrieved all inpatient and specialist outpatient visits recorded in the National Patient Register within 30 days prior to incident antipsychotic drug dispensing. We recorded the primary diagnosis and the two subsequent secondary diagnoses for each visit identified during the look-back period. The identified psychiatric diagnosis codes were used in an algorithm (online supplemental eTable 2) to identify the likely indication for each incident dispensing.Non-identifiable: refers to incident dispensings for which no relevant psychiatric diagnosis was identified within the 30-day look-back period. This includes dispensing with no recorded inpatient or specialist outpatient visit during the look-back window, or with visits but without any ICD-10 code corresponding to psychiatric conditions included in the indication algorithm.ADHD, attention-deficit hyperactivity disorder; ASD, autism spectrum disorder; ICD, International Statistical Classification. Likely indication associated with incident antipsychotic use in Swedish children and adolescents, 2008–2021 Note: For each incident dispensing, we retrieved all inpatient and specialist outpatient visits recorded in the National Patient Register within 30 days prior to incident antipsychotic drug dispensing. We recorded the primary diagnosis and the two subsequent secondary diagnoses for each visit identified during the look-back period. The identified psychiatric diagnosis codes were used in an algorithm (online supplemental eTable 2) to identify the likely indication for each incident dispensing. Non-identifiable: refers to incident dispensings for which no relevant psychiatric diagnosis was identified within the 30-day look-back period. This includes dispensing with no recorded inpatient or specialist outpatient visit during the look-back window, or with visits but without any ICD-10 code corresponding to psychiatric conditions included in the indication algorithm. ADHD, attention-deficit hyperactivity disorder; ASD, autism spectrum disorder; ICD, International Statistical Classification. On-label and off-label use In the subset of incident dispensings with an identifiable indication (n=16 455), we found that the proportion of off-label use increased from 76% to 88% during the study period (online supplemental eTable 6). Off-label use by indication accounted for most of this increase: it remained the most common reason for incident dispensing (rising from 61% to 73% of initiations with an identifiable indication) with incidence increasing from 0.14 to 0.62 per 1 000 persons (RR per year 1.13, 95% CI 1.10 to 1.16; cumulative increase 386%; figure 2). Trends in the incidence of on-label and off-label antipsychotic drug use for Swedish children and adolescents between 2008 and 2021, by sex. Note: Only the 16 455 (64.3%) dispensing with an identifiable psychiatric disorder indication associated with incident antipsychotic use are accounted for in this analysis. The increase in off-label initiation by indication was more pronounced among females (0.16 to 0.90 per 1000; RR per year 1.16, 95% CI 1.14 to 1.17; cumulative increase 560%; online supplemental eFigure 5) than males (0.13 to 0.36 per 1000; RR 1.09, 95% CI 1.05 to 1.14; cumulative increase 216%). For both females and males, the increase in off-label initiation for indication was most pronounced for adolescents aged 12–17 years. Among female adolescents aged 12–17 years, incidence increased from 0.34 to 2.20 per 1000 (RR per year 1.16, 95% CI 1.15 to 1.17; cumulative increase 584%; online supplemental eFigure 5), for adolescent males a more modest increase in incidence was observed (0.20 to 0.72 per 1000; RR per year 1.10, 95% CI 1.09 to 1.12; cumulative increase 252%). Sensitivity analysis Extending the look-back period to 60, 90 and 180 days did not materially change the distribution of the likely indication, with ASD, ADHD and anxiety disorders remaining the most common (online supplemental eTable 7). Similarly, the proportions of on-label and off-label use were stable across different look-back periods (online supplemental eTable 8). Discussion In this Swedish nationwide study on children and adolescents younger than 18 years, we observed substantial increases in the incidence of antipsychotic use between the years 2008 and 2021. The increase was most pronounced in adolescent females. Risperidone remained the most common antipsychotic initiated with aripiprazole and quetiapine having the largest increases in incidence during the study period. We identified ASD, ADHD and anxiety disorders as the most common likely indications for incident use. Off-label use of antipsychotics was common, particularly among adolescent females and was primarily driven by prescribing for non-approved indications. Previous studies on antipsychotic drug utilisation in the Swedish paediatric population have focused on the trends of prevalent use. A multinational study with Swedish data reported a 57% increase in prevalence between 2005 and 2014 among individuals aged up to 19 years.6 Gómez-Lumbreras et al reported an increase in prevalence in Sweden from 1.8 to 2.8 users per 1000 persons between 2008 and 2017 among individuals aged 0–19 years.7 Restricting our analysis to the period 2008 and 2017, we observed attenuated estimates for prevalence proportions (from 1.2 to 2.0 users per 1000 persons). This lower estimate likely reflects the younger age range (0–17 years) in our study compared with prior work including individuals up to age 19 years. Consistent with Gómez-Lumbreras et al, we found that antipsychotic use remained more prevalent among males throughout the study period. However, our data, which included more recent trends, showed that the prevalence among females had reached parity with males by 2021, with female incidence surpassing that of males by 2019. In line with prior findings,12 22 we identified ASD and ADHD as the most common likely indications overall. Notably, psychotic disorders accounted for only 2.5% of incident dispensings. This aligns with Spanish data indicating that prescribing for psychotic conditions is uncommon in paediatric populations.23 These findings likely reflect both the low prevalence of primary psychotic disorders in children and adolescents and the widespread off-label use of antipsychotics to manage behavioural and emotional symptoms associated with neurodevelopmental and other psychiatric disorders. Additionally, the sex differences in the likely indications identified were consistent with prior findings showing that ASD, conduct disorder and ADHD were more common indications for incident antipsychotic dispensing among males, as compared with mood, depressive and anxiety disorders for females.10 12 In Sweden, risperidone and aripiprazole are indicated for aggression and irritability in ASD and conduct disorders, conditions more frequently diagnosed in adolescent boys than girls.24 This likely contributed to risperidone being the most commonly initiated antipsychotic and to the marked increase in aripiprazole use observed among males during the study period (online supplemental eFigure 3). In contrast, females, who were more often treated for unapproved conditions, such as anxiety, mood or eating disorders, instead initiated quetiapine, aripiprazole and olanzapine, surpassing the incidence of risperidone by the year 2018 (online supplemental eFigure 3). Among incident dispensings with an identifiable psychiatric indication, 81% initiated an antipsychotic off-label, consistent with prior international estimates ranging from 36% to 93%.9 22 25 Much of this off-label prescribing was attributed to use for non-approved indications (64.8%) suggesting that antipsychotics are frequently initiated for psychiatric conditions lacking paediatric regulatory approval. During the study period we observed that off-label use, particularly off-label by indication was especially common among adolescent females. This is likely driven by the increased quetiapine and olanzapine use among adolescent females (online supplemental eFigure 4), since they are not approved for use in children and adolescents in Sweden for many of the conditions they are likely prescribed for (eg, anxiety, mood and sleep disorders). A growing use of quetiapine has been reported for children and adolescents in many Nordic countries.6 13 A previous study in Finnish children and adolescents found a more pronounced increase in quetiapine use among females than males,8 possibly reflecting greater prescribing for anxiety and mood disorders in this group.26 Due to its sedative properties, low dose quetiapine is commonly used for insomnia across Nordic countries.13 27 Our findings support the possibility of use for insomnia, as 39% of incident dispensings had prior melatonin use, indicating underlying sleep disturbances, which often accompany, for example, ADHD, anxiety and depressive disorders.28 It was however not possible to capture insomnia as such, as it is rarely recorded as a diagnosis unless it is the main reason for a healthcare contact. Our findings have several clinical implications. The rising incidence of antipsychotic use, particularly off-label use among adolescent females, underscores the need for careful risk-benefit assessment and updated clinical guidelines addressing the expanding scope of paediatric antipsychotic prescribing. The growing use of quetiapine and olanzapine for anxiety, mood and sleep disorders, despite limited evidence in the paediatric population, highlights the need for further research on their long-term safety and effectiveness. This is particularly important given that children and adolescents are especially vulnerable to cardiometabolic and endocrine adverse effects of antipsychotics including weight gain, dyslipidaemia and hormonal disturbances which are more likely with long-term use.29 Given that most antipsychotic initiations in our study were off-label, and in light of these safety concerns, regular cardiometabolic monitoring is crucial, yet available evidence indicates that such monitoring is inconsistently implemented in routine clinical practice.30 A strength of our study is the use of nationwide registers to provide an unselected and comprehensive assessment focusing on children and adolescents younger than 18 years. Our study is not without limitations. We defined drug exposure based on antipsychotic drug dispensings, which do not always translate to actual use by the patients. Hence, our results might to some degree be overestimated in relation to actual use, potentially leading to some overestimation of true drug exposure. Another limitation is that the Swedish Prescribed Drug Register does not capture medications administered exclusively during inpatient care; however, inpatient antipsychotic initiation without subsequent outpatient dispensing is rare and unlikely to materially impact incidence trends observed. Treatment duration and persistence following initiation were not evaluated, preventing distinction between short-term and longer-term use. As prolonged antipsychotic exposure in children and adolescents has been associated with cardiometabolic adverse effects,29 information on treatment persistence is important for interpreting the potential risk-benefit profile of prescribing in this population. Future studies examining treatment duration and patterns of continued use would provide important additional insight. Also, specific indications were not recorded in the prescription data, and we used an algorithm to capture psychiatric diagnoses from the National Patient Register to identify the likely indication for antipsychotic drug use. We could not identify a likely indication for approximately 36% of the incident users; an issue that has also been observed in prior studies from the UK presenting likely indications.12 Even though implementing different look-back periods did not materially change the distribution of likely indications, (online supplemental eTable 7) the potential misclassification of indication remains a limitation of our data. This may reflect real-world clinical practice where diagnoses may be provisional or assigned after treatment initiation and therefore not captured at or before initiation. In Sweden, such delays in establishing a definitive psychiatric diagnosis prior to antipsychotic treatment initiation are not uncommon among children and adolescents in routine clinical practice. During this period, clinicians may record no diagnosis or use non-specific ICD-10 diagnosis codes (Z-codes and R-codes). These clinical nuances likely explain why a psychiatric indication could not be identified for a subset of incident dispensing by the algorithm used. Conclusion This Swedish nationwide study on children and adolescents showed a substantial increase in antipsychotic use between 2008 and 2021, particularly among adolescent females. Among incident users, ASD, ADHD and anxiety disorders were the most common likely indications, with marked sex differences. We highlight the substantial increase in antipsychotic off-label use, especially among females and encourage future studies to evaluate the long-term effectiveness and safety of these treatments in the paediatric population. Supplementary material Footnotes Funding: The study was supported by grants from the Jane and Dan Olsson Foundations, the Swedish Society for Medical Research (Grant number: PG-24-0418), the Frimurare Barnhuset Foundation and the Erik and Edith Fernström Foundation (Grant number: 2025-00609). BP was supported by a consolidator investigator grant from Karolinska Institute. The funders had no role in the study design; in the collection, analysis and interpretation of the data; in the writing of the report; or in the decision to submit the article for publication. Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: Ethical approval was obtained from the Stockholm review board (administrative reference number: 2017/715–-31). In Sweden, informed consent is not required in large scale register-based studies, so long as ethical approval is obtained from the ethical committee. Data availability free text: Individual-level data from the Swedish national health and population registers cannot be shared by the authors under Swedish law. Data availability statement Data may be obtained from a third party and are not publicly available. References Associated Data Supplementary Materials Data Availability Statement Data may be obtained from a third party and are not publicly available.