Antipsychotic Treatment Duration in Children and Adolescents: A Register-Based Nationwide Study

Introduction

During the past decade, the use of antipsychotic drugs in children and adolescents has increased globally. The prevalence of antipsychotic use in these groups has been especially pronounced in many European countries, but the prevalence in European countries is altogether lower than the numbers reported in the United States (Olfson et al. 2015; Verdoux et al. 2015; Kalverdijk et al. 2017; Kaguelidou et al. 2020; Varimo et al. 2020). Recently, we and others have reported that over the past 10 years the incidence of new antipsychotic users in children and adolescents has ∼1.5- to 2-folded, and the increase has been more profound in adolescent girls (Kloosterboer et al. 2018; Varimo et al. 2020).

In children and adolescents, second-generation antipsychotics (SGAs) have emerged as a more favorable treatment modality, with fewer adverse effects than first-generation antipsychotics, resulting in increased prescribing of SGAs in many countries (Vitiello et al. 2009; Patten et al. 2012). In Finnish children and adolescents, the four most commonly used SGAs are risperidone, quetiapine, aripiprazole, and olanzapine, accounting for ∼96% of all SGA prescriptions (Varimo et al. 2020). Notably, we previously reported a profound increase in the use of quetiapine in recent years, especially in girls between 2015 and 2017 (Varimo et al. 2020). Despite the wide use of SGAs in children and adolescents, there is little evidence for the efficacy and long-term effects of these drugs in this vulnerable population (Ben Amor 2012; Loy et al. 2012).

The use of SGAs is associated with a risk of short- and long-term adverse effects on lipid metabolism, endocrine functions, weight gain, and blood pressure and can cause abnormal involuntary movements and daytime sedation (Laita et al. 2007; Calarge et al. 2009; Pringsheim et al. 2011; Cohen et al. 2012; Patten et al. 2012; Bobo et al. 2013; Arango et al. 2014; Cicala et al. 2020). Children and adolescents are more at risk of sustaining adverse effects of SGAs than adults (Vitiello et al. 2009), and thus, SGA treatments should be more closely monitored and ineffective treatments discontinued.

In children and adolescents, SGAs are used to treat target symptoms such as inattention/hyperactivity, disruptive behaviors, anxiety, and insomnia (Olfson et al. 2015; Rettew et al. 2015; Schroder et al. 2017; Dinnissen et al. 2020). Clinical experience suggests that SGAs are used even for long periods in these patient groups mainly without authorized treatment indications (i.e., off-label). It is noteworthy that studies investigating the treatment duration of SGAs in children and adolescents are scarce, and, to the best of our knowledge, no nationwide study has evaluated the factors associated with treatment duration.

The aim of this register study was to describe treatment duration in new antipsychotic users aged 1–17 years between 2008 and 2017 and to evaluate the influence of age, sex, starting year and season, and the SGA drug selected with the Cox proportional hazard models on the discontinuation of SGA treatment.

Patients and Methods

In Finland, the National Health Insurance (NHI) covers all Finnish inhabitants (∼5.5 million). The Social Insurance Institution of Finland maintains the Finnish Prescription Registry in which drug reimbursements covered by the NHI are recorded. The register contains detailed information on patients and prescriptions such as patients’ age, sex, sum of reimbursement received, purchase date, active substance of the dispensed medication, package size (number of dosing units, e.g., tablets, capsules, milliliters of the medication), and number of dispensed packages. Detailed information on the register has been reported elsewhere (Varimo et al. 2020).

The study population comprised all Finnish children and adolescents aged 1–17 years who had received antipsychotic drug reimbursements between January 1, 2008, and December 31, 2017 (n = 70,012). A new antipsychotic user was defined as a person who was dispensed an antipsychotic but who had not filled a prescription for antipsychotics in the previous 730 days. In subjects aged younger than 2 years, the prior use of SGAs was searched from the birth date. To cover the history of SGA use for possible new users in 2008, the database was searched from January 1, 2006, to December 31, 2007. Similarly, the new users of the year 2017 were excluded to allow an adequate follow-up period for all subjects. The index date was defined as the date of the initial purchase, and subjects’ age at treatment initiation was calculated as the difference between birth date and index date.

In the prescription register, all medications are coded according to the Anatomical Therapeutic Chemical (ATC) classification system. Data on subjects receiving reimbursements for antipsychotic medication (all medications belonging to ATC level 3 class [N05A]) were collected. These data included all reimbursed antipsychotic prescriptions, also small packages. We included the four most commonly used SGAs (risperidone, quetiapine, olanzapine, and aripiprazole) in the analyses since these drugs account for 96.1% of all SGAs used in Finnish children and adolescents (Varimo et al. 2020). Additionally, risperidone, quetiapine, and aripiprazole are the commonly used SGAs in Europe and the United States (Rettew et al. 2015; Halfdanarson et al. 2017; Kalverdijk et al. 2017; Korno and Aagaard 2018; Cicala et al. 2020; Varimo et al. 2020).

These four drugs have only a few authorized indications for patients aged younger than 18 years in Finland (risperidone: 6 weeks of use for aggressive behavior in intellectually disabled children aged older than 5 years; aripiprazole: schizophrenia [age older than 15 years] or bipolar I disorder for less than 12 weeks [age older than 13 years]; quetiapine and olanzapine: no authorized indications). The indications were obtained from the summaries of product characteristics accepted by the medical authorities (Finnish Medicines Agency, European Medicines Agency). Furthermore, the U.S. Food and Drug Administration (FDA) has approved the use of these as treatments for pediatric bipolar disorder (aged 10–17 years) and schizophrenia (aged 13–17 years), and risperidone and aripiprazole also for the treatment of irritability in children with autism spectrum disorders.

Next, we calculated the duration of antipsychotic treatment for new users. We decided not to use defined daily doses (DDD) in defining treatment duration since DDDs of antipsychotics are defined by using adult doses for the main indication in adults (i.e., schizophrenia), which are significantly higher than the doses prescribed in children and adolescents (WHO Collaborating Centre for Drug Statistics Methodology, 2020). To this end, we defined the treatment duration for new user as the time between their initial medicine purchase date (index date) and the date until which the drugs purchased at their last purchase would be expected to last (end date).

We first determined the treatment duration between the first purchase date and the last purchase date. Then, we defined the treatment duration for the last purchase in days using the dosage of the drug (Supplementary Fig. S1). The dosage of SGA use was calculated by using assumed daily doses. Because the often used assumption of one tablet per day in defining drug exposure is not valid in antipsychotics (Helin-Salmivaara et al. 2010; Rikala et al. 2013; Saastamoinen et al. 2012; Tanskanen et al. 2017), we used the assumed doses for drug exposure, which were based on the summaries of product characteristics accepted by medical authorities, clinical experience, and previously published psychopharmacology studies in children and adolescents (Pappadopulos et al. 2003; Caccia 2013; Whitney et al. 2015).

The assumed daily doses for the four most common antipsychotics were as follows. In subjects younger than 13 years, risperidone 0.5 mg twice a day, aripiprazole 10 mg once a day, quetiapine 25 mg twice a day (tablet), quetiapine depot 100 mg once a day, and olanzapine 5 mg once a day. In subjects older than 13 years, risperidone 1 mg twice a day, aripiprazole 15 mg once a day, quetiapine 25 mg three times a day (tablet), quetiapine depot 300 mg once a day, and olanzapine 10 mg once a day. The daily dose of liquid risperidone was assumed to be the same as with the tablet form (i.e., 0.5 mg twice a day in subjects younger than 13 years and 1 mg twice a day in subjects older than 13 years). When the subject became 13 years of age, the daily dose for younger than 13-year-olds was continued until the last purchased package of antipsychotic was finished. Thereafter, the daily dose was changed to correspond to the new age. We also assumed that no tablets were halved.

Thus, by using the assumed daily doses, the age of the patient, the strength of the medicine, and the number of tablets purchased we were able to estimate the length of the treatment after the last purchase (Supplementary Fig. S1). In the base case, the treatment was regarded as discontinued if the time between the end date and the next purchase date was more than 270 days (i.e., a treatment-free gap occurred). In Finland, up to 90 days (3 months) of supply of a drug can be reimbursed at a single dispense, thus we selected the gap of 270 days to represent a stable condition (90 day plus 180 days of grace period).

We divided the study population into four age groups: 1–6, 7–12, 13–15, and 16–17 years. The increased burden of schoolwork and social demands (such as time spent in social networks) may vary between the four seasons and influence treatment duration, thus, we determined the season of treatment initiation and categorized it into summer (June–August), autumn (September–November), winter (December–February), or spring (March–May) based on the index date.

In Finland, drug reimbursements of antipsychotics are divided into the Basic Refund Category (reimbursement rate of 40%) or the Special Refund Category (reimbursement rate of 100%). Medicines prescribed for intellectually disabled subjects with aggressive behavior and for children and adolescents diagnosed with psychoses or depressive disorders with mania or psychosis can be reimbursed in the Special Refund Category. In addition, in Finland, children and adolescents with a severe somatic or psychiatric disorder (i.e., mental and behavioral disorders) that causes disability and those who need regular care, attention, and rehabilitation are entitled to a disability allowance. Information on the disability allowance is recorded in the Register of Disability Benefits of the Social Insurance Institution. The allowance is intended to provide monetary support to families of children with disabilities. Information on Special Refund status of the purchased drug and possible disability allowance of the subject was included in the analysis.

Additionally, the adherence of SGA use was assessed with Medication Possession Ratio (MPR). MPR was calculated by the total days of medication dispensed between index date and the last refill divided by the treatment duration multiplied by 100 (maximum adherence was limited to 100%).

The study was based solely on register data, and thus, according to the Finnish Medical Research Act, no ethics committee approval was required. The Social Insurance Institution approved the study protocol.

Results

Between 2008 and 2017, altogether 20,932 children and adolescents aged 1–17 years started an SGA (risperidone, quetiapine, aripiprazole, or olanzapine). The mean treatment duration was 509 days (95% CI: 500–517 days), and the treatment duration was shorter in girls than in boys (Table 1) (p < 0.001). The median of SGA treatment duration was 317 days (95% CI: 306–325 days). From 2008 to 2017, the mean treatment duration shortened every year of the follow-up (Table 1). Interestingly, in 35.1% of the treatments, the duration was less than 50 days, whereas 16.0% used antipsychotics for more than 600 days (Table 1). The shortest treatment duration was in subjects whose treatment was initiated with quetiapine, followed by olanzapine, aripiprazole, and risperidone (p < 0.001) (Table 1). Additionally, treatment duration and used substance in subjects who had received a disability allowance based on the four most common mental and behavioral disorders are shown in Supplementary Table S1.

Next, we investigated factors associated with antipsychotic treatment discontinuation. In the unadjusted analyses, girls had an increased HR for treatment discontinuation (HR: 1.22 [95% CI: 1.17–1.26]), whereas in the adjusted model sex did not influence treatment duration significantly (HR: 0.97 [95% CI: 0.93–1.02]) (Table 2). Children and adolescents aged 7–12 years and 13–15 years were associated with shorter treatment duration than subjects aged 1–6 years (reference) in the adjusted model (HR: 1.23 [95% CI: 1.11–1.36]; HR: 1.35 [95% CI: 1.21–1.51], respectively) (Table 2). Of the SGA drugs, quetiapine increased the HR for treatment cessation versus risperidone (reference) in both models (unadjusted: HR: 1.45 [95% CI: 1.39–1.50]; adjusted: HR: 1.18 [95% CI: 1.12–1.25]), whereas olanzapine had the highest HR for discontinuation compared with risperidone in the adjusted model (HR: 1.66 [95% CI: 1.46–1.88]) (Table 2). Initiating SGA treatment in winter months compared with summer months (reference) increased the HR for treatment discontinuation in the adjusted model (HR: 1.08 [95% CI: 1.02–1.14]) (Table 2). Switching SGA drug within risperidone, quetiapine, aripiprazole, or olanzapine during treatment was associated with longer treatment duration (adjusted model: HR: 0.40 [95% CI: 0.38–0.43]) (Table 2). Similarly, a granted special refund or disability allowance was associated with longer treatment duration (HR: 0.34 [95% CI: 0.31–0.37]; HR: 0.56 [95% CI: 0.53–0.59], respectively) (Table 2).

Then, we measured adherence of the new SGA users. The mean MPR of all subjects were 69.5% (95% CI: 69.2–69.9%). The MPR was more than 80% in 42.0% of all subjects, and in 46.1% of girls and in 37.6% of boys. In girls, the mean MPR was 71.3% (95% CI: 70.8–71.8%) and in boys 67.7% (95% CI: 67.2–68.2%) (Supplementary Table S2).

In sensitivity analyses, the treatment-free gap was shortened in 30-day periods from 270 to 60 days (Table 3). The reduction in the treatment-free gap resulted in similar findings to those reported with the 270-day gap (Table 2). In the adjusted models, the association between sex and treatment discontinuation was not statistically significant. Children and adolescents aged 7–15 years were associated with an increased HR for treatment cessation with all treatment-free gaps, except for 60 days. Quetiapine was associated with treatment discontinuation compared with risperidone with all gaps, whereas olanzapine increased the risk of treatment cessation in gaps of 240, 210, 180, 150, and 120 days (Table 3). Switching of SGA after treatment initiation, being granted disability allowance, or belonging to the special refund category were associated with longer treatment duration in all sensitivity analyses (Table 3).

Discussion

In this nationwide study, we found that the mean and median durations of SGA treatment in children and adolescents were 509 days (17 months) and 317 days (11 months), respectively. The durations were shorter in girls than in boys. Age groups of 7–12 and 13–15 years were associated with treatment discontinuation, whereas subjects aged 1–6 years used SGAs for longer periods. Initiating treatment with quetiapine was associated with shorter treatment duration, whereas risperidone use and switching of antipsychotic drug during therapy increased the risk of longer SGA treatment.

Our results for SGA treatment duration (mean 17 months, median 11 months) are consistent with findings reported in the Netherlands (mean 19 months, median 6 months) and in the United States (median 6 months) (Burcu et al. 2014; Kloosterboer et al. 2018), but longer than reported in France (median 1 month) (Verdoux et al. 2015). In our study, girls used SGA for shorter periods than boys, and a similar trend has been reported by others (Kloosterboer et al. 2018). In 35.1% of the treatments, the duration was less than 50 days, which supports the observations in clinical practice that SGAs are prescribed off-label as an acute medication to treat such symptoms as insomnia and anxiety (Schroder et al. 2017; Menard et al. 2019; Dinnissen et al. 2020). Conversely, in 54.3% of the treatments, the duration was more than 100 days, and 16.0% used SGAs for more than 600 days. These durations are not especially long given that the authorized treatment indications of SGAs are for schizophrenia and bipolar disorder. However, these conditions are very rare, affecting only 1% of Finnish children and adolescents (Mäki et al. 2010), and entitle the subject to receive reimbursement for the medication under the Special Refund Category. In the present study, the Special Refund Category was granted to 9% of cases and was associated with longer SGA use. Longer use of SGAs has been shown to be associated with increased risk of adverse effects (Laita et al. 2007; Bobo et al. 2013). Consequently, before initiating SGA treatment in children and adolescents, it is essential to confirm that other treatment options (e.g., more favorable medications, psychosocial treatments) have been utilized. Possible contraindications and risk factors associated with SGAs would be considered carefully and the potential benefits and risks weighed. In addition, treatment monitoring should be coordinated to result in good treatment adherence, and useless SGA treatment should be discontinued.

This study is, to the best of our knowledge, the first to investigate multiple factors associated with SGA treatment duration in children and adolescents. The age groups 7–12 and 13–15 years were associated with shorter treatment duration than the age group 1–6 years. This could be result of several factors. First, SGAs are probably prescribed for different indications in younger children than in older adolescents, and younger children may have a more severe psychiatric condition requiring longer SGA treatment. Second, the shorter treatment duration in older adolescents may be a result of noncompliance since the stronger parental supervision in younger children may influence positively on the treatment compliance. This finding is consistent with the results reported by others (Kloosterboer et al. 2018) and may reflect the psychiatric diagnoses predominant in the age group. In fact, SGAs are prescribed to treat aggressive and impulsive and hyperkinetic behavior in attention-deficit/hyperactivity disorder in pervasive developmental disorders, and in autism, all of which emerge in early childhood and are lifelong conditions (Kalverdijk et al. 2017). Additionally, in Finnish children aged 1–6 years using antipsychotics, the frequency of disability allowance, a surrogate marker for a more severe psychiatric diagnostic status, has been reported to be up to 74% (Varimo et al. 2020). It seems that symptoms such as aggression are treated with SGAs in young children having severe and lifelong neurodevelopmental or psychiatric disorders (Fung et al. 2016). In children and adolescents, risperidone is often used to treat aggressive behavior related to autism spectrum disorders, and in these conditions, the mean treatment duration has been reported to be up to 2.5 years (Wink et al. 2014; Yoon et al. 2016). In this study, risperidone was associated with longer treatment duration, although it has only one authorized indication in Finland (6 weeks of use for aggressive behavior in intellectually disabled children aged older than 5 years). This is noteworthy since prolonged use of SGAs increases the risk of adverse effects, such as weight gain and metabolic changes, which may have a detrimental influence on well-being over time (Laita et al. 2007; Bobo et al. 2013). The long treatment duration of SGA in young children highlights the importance of careful follow-up of efficacy and adverse effects of SGA treatment, especially in subjects with disabling neurodevelopmental disorders (e.g., autism disorder) who are even more susceptible to adverse effects of SGAs (Alfageh et al. 2019). Additionally, SGA treatment should be combined with psychosocial treatments such as parent management training and training of social skills and executive functions, which can be tailored based on the severity of the conditions and the needs of the family. Further studies should investigate polypharmacy in new SGA users and the opinions of parents on their child's SGA treatment initiation, monitoring, and actualization.

We found that initiating SGA treatment with risperidone was associated with longer treatment duration, whereas quetiapine treatment increased the HR for treatment discontinuation. Of all SGA treatments, 50.8% were initiated with quetiapine, and the mean duration of quetiapine treatment was less than half of the mean duration of risperidone. It is tempting to speculate that quetiapine is prescribed as a short-term treatment for anxiety disorders, insomnia, agitation, or symptoms that do not meet the criteria of a mental disorder (Chow et al. 2017). Quetiapine has no authorized indications in subjects aged younger than 18 years in Finland, supporting the view of high off-label use. The incidence of anxiety and depressive disorders has been reported to increase in adolescent girls, and based on our clinical experience, these conditions often present with acute symptoms, such as insomnia and anxiety, requiring short-term SGA monotherapy (e.g., quetiapine) or cotreatment with an antidepressant at treatment initiation. Additionally, the shorter treatment duration with quetiapine may reflect the prescribing habits observed in adult populations (Duncan et al. 2016) or may signify treatment discontinuation due to lack of effectiveness or significant adverse effects. Supporting the latter, adverse effects of quetiapine have been reported in up to 50% of users with pervasive developmental disorders, and in 15% of these, the adverse effects resulted in treatment discontinuation (Corson et al. 2004). In children and adolescents, even short-term quetiapine treatments are associated with adverse effects such as weight gain and increased levels of cholesterol and triglycerides (Cohen et al. 2012). Long quetiapine treatment can cause abnormalities in thyroid hormone levels (Lambert et al. 2016) or even hypothyroidism (Dobbs et al. 2004). Thus, the acute and long-term benefits of SGA treatment should be carefully weighed at the individual level before treatment initiation and proper treatment monitoring should be organized. The increased off-label use of quetiapine in Finland warrants administrative guidelines at clinics to use approved product for such indications as anxiety and insomnia as a first-line treatment, as well as, further large trials to assess benefits and risks of quetiapine in a pediatric population.

We reported, for the first time, that switching an antipsychotic drug after treatment initiation is associated with longer treatment duration. The changing of a drug after treatment initiation may result from the first-line medication being ineffective for the treated symptom or causing a significant adverse effect, the expectations of the effects of SGA treatment being unrealistic, or the overall treatment being insufficient for the severity of the condition. In support of the latter, the granted disability allowance and the special refund category were associated with longer SGA treatment duration in all the age groups. Disability allowance and special refund were granted to subjects suffering from more severe mental illnesses, such as hyperkinetic disorders and pervasive developmental disorders, which are known to be associated with longer SGA use (Wink et al. 2014; Yoon et al. 2016).

This study includes some strengths and limitations that warrant mention. Finnish Prescription Registry is a nationwide database, which includes all reimbursed outpatient drug prescriptions in children and adolescents aged 1–17 years. Treatment duration was carefully determined, adherence was calculated, and the data included information on users who switched medication after initiation of SGA treatment; however, the calculation of treatment duration was based on assumed and previously reported daily doses. The study did not include information of the possible adverse effects of reported during SGA treatment, and the follow-up period was shorter in subjects who initiated treatment in 2016 than in those who started SGA in 2008. Additionally, we did not have access to the indication of the prescribed drug, and we assumed that all the purchased medications were used. These limitations could be resolved in future studies by combining the data of the prescription register with other registers such as medical records.

Conclusions

The mean treatment duration of SGAs in children and adolescents was consistent with durations reported in other countries. Older age was associated with treatment discontinuation, whereas younger children used SGAs for longer periods. The longer use of SGA in younger children probably stems from the fact that neurodevelopmental disorders are predominant at this age but also serves as a reminder to clinicians to utilize psychosocial treatments (e.g., parental management training) alongside SGA treatment. Risperidone was associated with longer SGA use, whereas quetiapine was used for shorter periods. Both short and long treatments of SGA are associated with the risk of adverse effects. Thus, the harms and benefits of SGA treatment should be carefully weighed before treatment initiation, the treatment actualization should be monitored, and any useless SGA treatments should be promptly discontinued.

Clinical Significance

Our results suggest that SGAs are used for longer periods in younger children and that initiating treatment with risperidone is associated with longer use. Use of quetiapine is associated with treatment discontinuation. These findings suggest that the harms and benefits of SGA treatment should be carefully weighed before treatment initiation.