PERS consist of three related clinical studies, a short-term efficacy and safety study, a study on maintenance effects, and an open-label observational study. The three study designs including details regarding clinical and/or laboratory parameters are outlined in detail as part of the Schedule of Events (SOE; Annex 1).
Short-term efficacy of risperidone in CD: conduct disorder in children and adolescents (CONCA)
CONCA is a multicentre, randomised, double-blind, parallel, placebo-controlled trial. This RCT will examine whether risperidone, given orally in a dose of 0.25–3.0 mg/day for 12 weeks, is superior to placebo in reducing disruptive behavioural symptoms of children and adolescents with CD without mental retardation. CONCA comprises three study periods and aims to enrol 264 patients (50 % children 5- to 11-years old and 50 % adolescents 12–17.5 years). The sample size calculations were based on an assumption of an effect size of 0.4 and 90 % power with a 1:1 randomization scheme to risperidone or placebo. The clinical diagnosis of CD according to DSM-IV-TR will be confirmed by a structured interview (Kiddie-SADS, conduct disorder module  and a score ≥27 on the ODD/CD composite score of the Nisonger CBRF will be required at Visits 2 or 3 to meet criteria of sufficiently severe CD). Children and adolescents with ADHD and those on stable stimulant medication, with no planned changes during the trial, will not be excluded (see supplementary material in Annex 1 for details of inclusion and exclusion criteria).
Study Period I is a 2-week screening and washout period; during this period, patients will be screened for study eligibility. Randomization to risperidone or placebo will occur at Visit 3. Subjects will randomly be assigned to risperidone or placebo in a 1:1 ratio and randomisation will not be stratified. Study Period II is a 12-week acute treatment period (Visits 3–9). In order to safeguard patients in the study, patients can be either in outpatient or inpatient settings. For patients in the risperidone treatment group, medication will be given in the evening and dosing will begin at either 0.25 (<50 kg body weight) or 0.5 (≥50 kg), mg/day (tablets), depending on the patient’s weight. Doses will be increased by 0.25 mg/day or 0.5 mg/day increments each week to maximum doses that vary by patient weight (<50 kg or ≥50 kg), based on the investigator’s assessment of efficacy and tolerability/safety. Study Period III is a 1-week double-blind period, a progressive withdrawal from study medication (Visits 9–10).
Efficacy is measured by the last observation carried forward (LOCF) mean change from baseline to endpoint on the pivotal scale, the NCBRF–Typical IQ Version . The primary measure is the ODD/CD Disruptive Behaviour Composite Total score using investigator ratings based on all available information. The decision to use an investigator-rated NCBRF in the PERS studies is due to regulatory reasons (i.e. because in registration trials there is a strong preference for a primary outcome based on professional/clinical assessment rather than an assessment by parents, patients or teachers).
Secondary objectives of the study are to examine whether treatment with risperidone is associated with lower symptoms on the Clinical Global Impression: Improvement and Severity scales (CGI-I and CGI-S , the Modified Overt Aggression Scale (M-OAS [36, 37]), and with better functioning on the Child-Global Assessment Scale (C-GAS ), the CHIP-CE Parent Report Form  and the Positive Social subscale of the Nisonger CBRF Typical IQ version . We will also assess the effect of risperidone on concurrent ADHD symptoms using the ADHD-Rating Scale  and examine whether ADHD and its treatment influence the risperidone effect on disruptive behaviours.
We plan to further assess the effect of risperidone compared to placebo on cognitive functioning (e.g. due to possible sedative effects) using attentional and set-shifting tests from the Amsterdam Neuropsychological Test battery (ANT) . Safety and tolerability will be assessed extensively and include spontaneously reported and investigator-rated treatment-emergent adverse events based on information from both patient and parents. Measures comprise potential extrapyramidal symptoms (Barnes Akathisia Scale , Simpson-Angus Scale , Abnormal Involuntary Movement Scale (AIMS) ); changes in ECG, vital signs, body temperature, and laboratory analyses; changes in growth, assessed using height, weight, and body mass index (BMI); and changes in suicidal ideation and suicidal behaviour, as assessed by the investigator-rated Columbia Suicide Severity rating Scale (C-SSRS) .
Maintenance effects in CD: discontinuation for conduct disorder in children and adolescents (DIS-CONCA)
DIS-CONCA is a multicentre, randomised, double-blind, parallel, placebo-controlled discontinuation (relapse prevention) trial in children and adolescents, aged 6–17.5, with CD and without mental retardation. The study incorporates a screening phase, an open-label period of treatment with risperidone followed by randomization to continued risperidone or placebo substitution over a 12-week period. This allows for an evaluation of rates of relapse following response, and it is an accepted method for determining the value of differing lengths of continuation treatment following an initial response.
DIS-CONCA will examine whether risperidone given orally in a dose of 0.25–3.0 mg/day is superior to placebo in preventing relapse of symptoms of CD, as assessed through a 12-week, double-blind discontinuation trial. Inclusion criteria are similar to those for CONCA, and the pivotal scale is as in CONCA the NCBRF–Typical IQ Version-ODD/CD disruptive behaviour Composite Total score  using investigator ratings based on all available information. Since compliance is often poor in CD children and adolescents, significant non-compliance [missing more than two consecutive days of study medication (full doses), or failure to take at least 70 % of prescribed doses of study medication (full doses) during two or more visit intervals] will be discussed with patient and parent and the patient will be discontinued only when, in the opinion of the investigator, the patient is deemed unlikely to become compliant and data obtained from the patient judged unreliable.
Clinical response is defined as >25 % reduction from baseline score on the NCBRF-TIQ D-Total subscale at endpoint and a score of 1 or 2 (“much” or very much” improved) on the CGI-Improvement scale. The primary efficacy measure will be comparison of the number of days from randomization to relapse for each treatment group using the Kaplan–Meier product limit estimator. Relapse is defined as a deterioration of >2 points on the CGI-Severity scale and a 25 % increase in the score on the Nisonger pivotal scale, compared to start of Study Period II (average of Visits 7 and 8), for at least two consecutive visits, 15 days apart.
Secondary objectives of DIS-CONCA are to establish the long-term efficacy of treatment with risperidone on secondary measures as the CGI-S, C-GAS, and Overt Aggression Scale (OAS) and several measures of functioning as the Child-Global Assessment Scale (C-GAS;), the CHIP-CE Parent Report Form  and the ‘Positive Social’ subscale of the Nisonger CBRF Typical IQ version . Further, we will assess the long-term effects of risperidone on cognitive functioning and on ADHD symptoms, and collect extensive data on safety and tolerability.
Participants will undergo four study periods. After screening (Study Period I), Study Period II will be a 16-week titration and open-label treatment period (7 visits). This acute treatment period was chosen to increase the chance to reflect the change based on medication treatment. All patients who complete Study Period II and maintain clinical response are eligible to participate in Study Period III. Two consecutive positive evaluations that meet the criteria for response are required to enter into the randomised discontinuation phase. Study Period III will be a 12-week double-blind, randomised, placebo-controlled discontinuation period (8 visits). At the start of Study Period III, patients will be randomised to continued risperidone or placebo (1:1 ratio). In the placebo-discontinuation group, risperidone gradual discontinuation will be started during the first 4 weeks of Study Period III, but the specific timing of starting the discontinuation will be unknown to patients, their families and the investigators. Patients who meet relapse criteria during Study Period III will be considered completers and may be discontinued and given the opportunity of entering the observational prospective study. Therefore, no interim analysis to stop the study prematurely has been planned in the discontinuation phase. After 24 weeks, medication will be discontinued in 2 weeks in all the non-relapsing patients (Study Period IV); patients will have their final clinical assessment after further 2 weeks, with the opportunity of entering, if appropriate, the observational prospective study.
Based on power calculations on previous studies with risperidone in CD with mild mental retardation, 150 (50 % children 6–11, 50 % adolescents 12–17.5) patients should be enrolled in Study Period II. A sample size of 150 using a 1:1 randomization over risperidone and placebo will have 80 % power to detect an increase in relapse rate from 33.3 % in the risperidone group to 66.6 % in the placebo group, using a proportional hazards analysis with a censoring of data following a uniform distribution between the 6th and the 12th week (simulations, functionspower, libraryHmisc, R software 2.14). Considering that 25 % of patients enrolled in Study Period II will not meet inclusion criteria to continue in Study Period III, 200 patients should be enrolled in Study Period II.
Observational pharmacovigilance study
There is a lack of knowledge about which factors make some children and adolescents more vulnerable than others to the short-term and long-term adverse effects of risperidone [8, 45–47]. Dose may be one obvious mediator, but many available studies did not obtain risperidone plasma concentrations. Some recent evidence has suggested that genetic factors that may also influence the occurrence of metabolic or endocrine side effects such as weight gain and hyperprolactinemia, but much remains to be investigated in this area [48, 49]. Finally, little is known about the degree of reversibility of these adverse reactions upon treatment discontinuation. The pharmacovigilance study within PERS is aimed at filling in the safety knowledge gap, through conducting an observational open-label study in which 600 children and adolescents aged between 5 and 16 years who have been prescribed risperidone will be followed over a treatment period of up to 2 years and comparing them to 250 controls. If there has been previous use of risperidone, children can also enter the study as long as there has been no use of risperidone for the past 6 months. The study will be able to detect a difference in the rates of clinically significant weight gain over the first year of usage, defined as an increase in BMI standardised score of 0.5 or greater. We plan to follow 600 patients under medication and this will allow detection of an increase from 3 per 100 to 8 per 100 with 90 % power. The study has further 95 % power to detect rare adverse events, i.e. with an incidence rate as low as 0.5 per cent. All participating patients will be regular referrals to child and adolescent psychiatry centres. There will be no restrictions for the reasons of prescribing risperidone, both labelled and unlabelled indications will be included. Also, any possible concomitant medication will be allowed. The pharmacovigilance study will be fully embedded within regular clinical practice and will solely collect data from patients’ medical records, without imposing any impact on choice and duration of treatment. Participating clinical centres have therefore agreed on systematic safety monitoring, in accordance with international recommendations .
Main objectives of the PERS pharmacovigilance study are to follow the course of BMI, waist circumference and biological parameters and to systematically collect the frequency, nature, and course of subjectively reported possible side effects. Subjects will be enrolled at the start of risperidone treatment. The PERS pharmacovigilance study includes clinical and/or laboratory parameters prior to and after 1, 2, 3, 6, 12, 18, and 24 months of treatment will be collected. Subjectively reported possible side effects, including those related to suicidality and sexuality will be collected. Blood laboratory parameters include glycosylated haemoglobin A1, liver enzymes, fasting lipids (triglycerides, total cholesterol, HDL cholesterol, LDL cholesterol); fasting blood glucose; fasting insulin; and prolactin (see supplementary information for details about Schedule of Events).
We will also investigate moderating and/or mediating factors of the short- and long-term adverse events and changes in weight and waist circumference. Specifically, we will examine the influence of (1) medication factors (average daily dose; cumulative dosage of risperidone; pre-treatment history of medication; use of concomitant medication; duration of treatment; discontinuation versus continued use); (2) patient characteristics (age; Tanner stage; initial BMI/waist circumference; socioeconomic status; clinical efficacy; clinical indication), and (3) lifestyle factors including dietary factors, levels of physical activity, and illicit drug use.
A hierarchical linear regression model will be used to compare the BMI standardised scores over time with adjustment for the correlations between assessments within the same individual. Similar models will be used to identify patterns that may be associated with drug usage for the secondary outcomes.
Secondary and exploratory studies on moderating factors
In addition to examining main treatment effects, it is important to also examine the impact of potential moderating factors of efficacy and safety. Moderators are characteristics of subjects before treatment that have an impact on efficacy or tolerability.
An important topic is the subtyping of aggressive and antisocial behaviour. Among children and adolescents who develop severe patterns of aggressive and antisocial behaviour, there are certain subgroups that may be subject to distinct causal processes that result in their problem behaviour. Specifically, callous-unemotional (CU) traits (e.g. lack of guilt, absence of empathy, callous use of others) are relatively stable and are associated with more severe conduct problems, delinquency, or aggression. Children and adolescents with CD with and without CU traits also differ in their emotional, cognitive, and personality characteristics . Another important distinction has been made between reactive forms of aggression (e.g. in response to perceived provocation or treat) and instrumental aggression (e.g. premeditated aggression for some gain or reward) . Post hoc, we will examine whether risperidone differentially affects CD with and without CU traits, and instrumental versus affective aggression in both CONCA and DIS-CONCA trials. We will also examine the role of co-occuring ADHD, because children with CD and with concurrent ADHD have an earlier onset of CD and a worse outcome compared to children with CD without ADHD . Moderators of the short- and long-term adverse events occurring during risperidone treatment will also be analysed in the observational pharmacovigilance study.