Cognitive Function with Long-Term Risperidone in Children and Adolescents with Disruptive Behavior Disorder

Abstract

Objective:

The aim of this study was to evaluate cognitive effects from long-term risperidone treatment for disruptive behavior disorders (DBDs) in children and adolescents.

Methods:

Patients 5–17 years old with DBDs and an intelligence quotient (IQ) ≥54 were randomized to flexibly dosed risperidone or placebo in a 6-month recurrence prevention trial. Cognitive function was assessed with a modified California Verbal Learning Test for Children (MVLT-C) and Continuous Performance Test (CPT), which assessed vigilance through computer testing with both an easy and a hard test. Somnolence was also evaluated throughout treatment. Clinically meaningful treatment effects were assessed as changes of ≥0.5 or ≥1.0 standard deviation (SD) from baseline.

Results:

A total of 284 subjects participating in 6-month maintenance treatment had both baseline and end point cognition assessments and were included in this analysis. Significant improvements from baseline occurred in risperidone-treated subjects for CPT hard hit rates and discrimination ability (Pr) (p < 0.05 for both), and in placebo subjects for CPT easy false alarms rates (p < 0.001) and hard Pr (p < 0.05). Both the easy and hard CPTs correct mean response time worsened with placebo. The MVLT-C short-delay free recall improved significantly for both risperidone and placebo. After adjusting for country, somnolence, age, IQ, and baseline scores, no significant differences were noted in cognition between treatment groups. Clinically meaningful changes were generally similar for risperidone and placebo patients. Mild to moderate somnolence occurred in only 2% of patients treated with either risperidone or placebo. The change in cognitive testing was not different in subjects experiencing somnolence as an adverse event (AE) compared with subjects not experiencing somnolence.

Conclusions:

Risperidone treatment resulted in no decline in cognitive function among children and adolescents. These results extend on previous results from risperidone studies in DBD in patients with lower IQ.

Introduction

D isruptive behavioral disorders (DBDs) affect 7–11% of children and adolescents in the general population, with most affected children diagnosed with conduct or oppositional disorder (Angold et al. 2002; Canino et al. 2004). Among children requiring medication management for disruptive and aggressive behavior, risperidone has been recommended as a first-line therapy by an international expert panel, due to its favorable efficacy and tolerability (Kutcher et al. 2004). Risperidone has consistently demonstrated symptomatic improvement in DBDs in both short- and long-term open-label and blinded clinical trials (Jensen et al. 2007). Most children with DBDs require long-term, multimodal maintenance treatment, with symptom resolution by mid to late adolescence occurring in less than 15% of children (Lahey et al. 2002).

The decision to initiate risperidone treatment in pediatric patients may be influenced by concerns about effects on cognitive function. Some of these concerns may be related to the potential adverse effects (AEs), such as somnolence or fatigue, that may occur with antipsychotic treatment. A recent review of pediatric clinical trials with antipsychotics reported that cognitive AEs occurred infrequently with atypical antipsychotics in pediatric populations (Jensen et al. 2007). Somnolence was frequently reported, although typically mild to moderate in severity and improved by changing dosing. Most prospective studies of antipsychotic treatment and cognitive outcome are conducted in patients with schizophrenia and are suggestive of either no change or benefit (Keefe et al 2007). Cognitive complaints are often noted secondary to potential side effects of sedation with higher doses of antipsychotics, and thus it would be important to determine if the sedative effect played a causal role in cognitive complaints. Furthermore, cognitive deficits in adult schizophrenia have been shown to improve with atypical, but not conventional, antipsychotics (Jensen et al. 2007). Overall, the use of antipsychotics has increased in children in both the United States and Europe, with use during early and middle childhood years typically for indications other than psychosis (Rani et al. 2008). This increased use has led experts to recommend studies evaluating long-term efficacy and safety issues in pediatric populations (Rani et al. 2008).

In controlled studies of adults with schizophrenia, cognitive functioning (particularly executive functioning and memory) has been demonstrated to improve with risperidone treatment (Harvey et al. 2003; Weickert et al. 2003), although there are some contrary reports from open studies (Daban et al. 2005). Pandina and colleagues evaluated data from five studies treating children with DBDs and subnormal intellect (intelligence quotient [IQ] ≥36 to ≤84) with risperidone (Pandina et al. 2007). Data were analyzed from 228 patients in two 6-week, double-blind risperidone trials, and 688 patients in three 1-year, open-label studies. There was no short- or long-term decline in attention or verbal memory with either risperidone or placebo. Short-term treatment with both risperidone and placebo was associated with cognitive improvement, as was long-term risperidone treatment. Treatment effects were not mediated by treatment-emergent somnolence, nor did somnolence appear to adversely impact cognitive function, either in the short- or long-term. The authors commented that it appeared that many subjects made gains over the long term, consistent with expected development, but these gains could not be directly attributed to drug treatment, as there was no placebo group comparator. In addition, these data were also limited by including only patients with baseline subnormal intellect.

A recent, long-term, controlled study demonstrated efficacy of risperidone in children and adolescents with DBDs and either normal or subaverage intellect (IQ ≥55); cognitive tests identical to those reported in Pandina et al. (2007) were also performed. In this recurrence prevention study, 527 children with a DBD ages 5–17 without moderate or severe cognitive impairment were treated with risperidone (Reyes et al. 2006). Patients responding to risperidone (mean daily dose: 0.02 mg/kg) during 12 weeks of initial risperidone treatment were randomized to 6 months of treatment with the previously effective dose of risperidone (n = 172) or placebo (n = 163). Symptom recurrence occurred in 27% treated with risperidone versus 42% with placebo (p = 0.002). Recurrence occurred in 25% of patients after 119 days with risperidone versus 37 days with placebo (p < 0.001). The current report describes cognitive functioning in patients treated in this recurrence prevention study (Reyes et al. 2006). This report expands upon cognitive function data from earlier studies (Pandina et al. 2007) by providing a long-term, blinded, placebo-treated comparator group, and it includes data from children and adolescents with average intellect.

Methods

This double-blind, international, multicenter study was conducted in eight countries (Belgium, Germany, United Kingdom, Israel, Poland, South Africa, Spain, and The Netherlands) and 25 centers. Before study initiation, the protocol and written informed consent were approved by local Independent Ethics Committees. This study was conducted in accordance with good clinical practice and the Declaration of Helsinki. Informed consent was obtained from the parents or guardians of every subject prior to study enrollment. The current study reports detailed cognitive testing data from this trial. Full efficacy and safety data have been reported previously (Reyes et al. 2006).

Subjects

Children and adolescents (ages 5–17 years) without moderate or severe intellectual impairment (IQ ≥54 obtained at screening or during the preceding 3 years using the Wechsler Intellect Scale for Children or the Stanford–Binet Intellect Scale) were eligible for enrollment if they met criteria for a Diagnostic and Statistical Manual of Mental Disorders, 4^th edition (DSM-IV) (American Psychiatric Association 1994) diagnosis of conduct disorder, oppositional defiant disorder, or DBD not otherwise specified (NOS). The DBD diagnosis was confirmed with the Schedule for Affective Disorders and Schizophrenia for School-Age Children–Present and Lifetime Version (K-SADS). Inclusion required the conduct problem to be serious enough to warrant clinical treatment with risperidone and be associated with a Nisonger Child Behavior Rating Form–Parent Version (N-CBRF) (Aman et al. 1996; Tassé et al. 1996) Conduct Problem subscale score ≥24 at both screening and treatment initiation. Children and adolescents with other serious medical or psychiatric conditions (except for attention deficit/hyperactivity disorder) were excluded.

Study design

All subjects were evaluated with standardized cognitive testing with the Continuous Performance Test (CPT) and a modified version of a Verbal Learning Test–Children's Version (MVLT-C) at double-blind baseline and double-blind end point. These same measures have been used in previous open-label and double-blind trials evaluating cognition in children with DBDs treated with risperidone (Pandina et al. 2007). These tasks were selected because they have been employed across the entire program of DBD studies spanning over a decade to establish if there were any changes in cognitive function, either positive or negative, in these two domains with risperidone treatment. The CPT is an attention test, assessing vigilance through computer testing with both an easy test and a hard test. In the easy test, stimuli were a witch and a princess appearing on a blank screen. Stimuli were presented for 2 seconds at regular intervals and were presented randomly at approximately 1:5 ratio, with no two targets presented consecutively. Subjects were required to respond whenever the princess appeared. In the hard test, stimuli were identical to the easy condition, but the interval varied (from 1 to 4 seconds) and stimulus duration was 3/10 of a second. Targets were also presented randomly at an approximately 1:5 ratio, with no two targets presented consecutively. Parameters were analyzed separately for the easy CPT and the hard CPT tests. Computer-generated scores were calculated for hit rates, false alarm rates, correct mean response time, and false mean response time.

The MVLT-C involves repeated presentation of word lists with requests to freely recall and discriminatively recognize presented words (Delis et al. 1987). Two modified list-learning test versions were used in this study: The MVLT-C-10 was used for patients with a chronological or mental age <6 years and the MVLT-C-15 for patients with a chronological or mental age ≥6 years. The MVLT-C consisted of two parts: (1) short-delay free recall and (2) long-delay free recall and recognition. A list of 15 words was presented (orally or by pictures). For the short- and long-delay free recall studies, subjects were asked to enumerate the words they recalled.

Subjects were treated with risperidone 1 mg/mL oral solution, dosed once or twice daily, as tolerated. Treatment was divided into three sequential phases: Acute treatment (6 weeks open label), continuation treatment (6 weeks single blind), and maintenance (6 months double blind). During acute treatment, subjects received risperidone for 6 weeks using clinically determined flexible doses. In subjects weighing <50 kg, treatment was initiated at 0.25 mg, with dosage increased to 0.50 mg on treatment day 3 and 0.75 mg on treatment day 5, if needed. Maintenance dosage was 0.25–0.75 mg/day. In subjects weighing ≥50 kg, treatment was initiated at 0.50 mg, with dosage increased to 1.0 mg on treatment day 3 and 1.5 mg on treatment day 5, if needed. Maintenance dosage was 0.5 to 1.5 mg/day based on clinical response and tolerability.

Somnolence was evaluated as an AE biweekly during acute and continuation treatment and monthly during maintenance treatment.

Data analysis

A total of 335 patients were randomized into the double-blind maintenance treatment period. The sample size for this study was estimated from a power calculation based on the prespecified primary parameter (time to symptom recurrence) as described in the earlier report (Reyes et al. 2006). The analysis presented here is based on the 284 patients with both baseline and end point cognitive assessments in the maintenance period.

Descriptive statistics were used to evaluate demographics, cognitive testing, and the occurrence, severity, and duration of somnolence. Analyses were performed for all patients who had completed baseline and at least one postbaseline MVLT-C or CPT. Descriptive statistics of raw scores and changes from double-blind baseline were calculated for total short-delay free recall and long-delay free recall for the MVLT-C, and for hit rates, false alarms rates, correct mean response time, and false mean response time for the CPT. Hit rates and false alarms rates were expressed as a proportion of all targets. Items of “false mean response time” and “correct mean response time” were analyzed with log transformation. Values equaling zero were excluded. The probability of correct discrimination was determined using the two-high threshold theory (Snodgrass and Corwin 1988). CPT discrimination ability (Pr) was calculated as the proportion of total hits minus the proportion of false alarms. CPT response bias (Br) was calculated as the proportion of false alarms divided by 1 − Pr. MVLT-C-PR was calculated as (NHIT − number of false alarms (NFA))/16, where NHIT equaled the number of correctly recognized words out of 15. NFA was calculated as 15 − NHIT. MVLT-C Br was calculated as (NFA + 0.5)/16 divided by 1 − Pr. Pr scores could range from −1 (perfect inaccurate discrimination) to 1 (perfect discrimination). Br scores could range from 0 (very conservative) to 1 (liberal bias). Changes from double-blind baseline were compared within treatment groups using paired t-tests and between treatment groups using an analysis of covariance (ANCOVA), controlling for country, absence or presence of somnolence as an adverse event, age (<12 vs. >12), IQ (<84 vs. >84), and baseline score. In addition, clinically meaningful patient worsening and improvement for both CPT and MVLT-C were evaluated based on change from baseline of >0.5 or >1.0 SD of change. Comparison of outcome between treatment groups was analyzed using a Fisher exact test.

A possible relationship between somnolence and change in cognitive testing was evaluated by comparing change in CPT and MVLT in patients with an AE of somnolence and those without somnolence during open-label treatment.

Results

Subjects

Approximately 85% (n = 284) of the 335 subjects who participated in the 6-month, double-blind maintenance treatment had both baseline and end point (6 months) cognition assessments. Demographics were similar for those randomized to risperidone versus placebo (Table 1). The demographics and IQ for these 284 subjects were similar to the full sample of 335 patients (Reyes et al. 2006). Most subjects were males with conduct disorder, with a mean IQ of 92 (range = 54–144). IQ distribution was similar for patients treated with either risperidone or placebo. Mean mode dose in risperidone patients was 0.91 ± 0.38 mg/day. Treatment was completed by 67.8% (n = 97) treated with risperidone and 41.8% (n = 59) treated with placebo (Table 2).

Table 1.

Subject Demographics

		Maintenance phase
	All patients n = 284	Risperidone n = 143	Placebo n = 14
Age, years
Mean (SD)	10.8 (2.9)	10.8 (2.8)	10.8 (2.9)
Gender, n (%)
Male	248 (87.3)	117 (81.8)	131 (92.9)
Female	36 (12.7)	26 (18.2)	10 (7.1)
Diagnosis, n (%)
CD	102 (35.9)	52 (36.4)	50 (35.5)
ODD	174 (61.3)	88 (61.5)	86 (61.0)
DBD NOS	8 (2.8)	3 (2.1)	5 (3.5)
IQ, mean SD	91.9 (18.4)	91.9 (17.8)	91.9 (19.0)

Abbreviations: SD = standard deviation; CD = conduct disorder; ODD = oppositional defiant disorder; DBD = disruptive behavior disorder; NOS = not otherwise specified; IQ = intelligence quotient.

Table 2.

Subject Disposition

Disposition	Risperidone n = 143	Placebo n = 141
Completed study	97	59
Discontinued study	46	82
Relapse during maintenance	36	69
Withdrew consent	3	5
Adverse event	3	2
Lost to follow up	0	2
Noncompliance	2	2
Other	2	2

Continuous performance test

Changes in CPT scores during double-blind maintenance are presented in Table 3. Within-group improvement effects were noted for both easy and hard CPT for both placebo- and risperidone-treated subjects. Significant improvements from baseline occurred in risperidone-treated subjects with hard hit rates and Pr (p < 0.05 for both), and in placebo subjects with easy false alarms rates (p < 0.001) and hard Pr (p < 0.05). Both easy (p < 0.05) and hard (p < 0.01) CPT correct mean response time increased significantly from baseline with placebo. An ANCOVA model including treatment group, country, somnolence, age (≤12 vs. >12), IQ (≤84 vs. >84) and baseline value was performed. Age group was a significant factor for easy false alarms rates (p = 0.002), easy correct mean response time (p = 0.03), easy Br (p = 0.01), and hard Br (p = 0.03). IQ (≤84 vs. >84) was not a significant factor for any CPT parameter.

Table 3.

Continuous Performance Test Values at Baseline and End Point

	Open label			Double blind
	Risperidone (n = 429)			Risperidone (n = 131)			Placebo (n = 127)
	Baseline mean (SD)	End point mean (SD)	Within-group effect size	Baseline mean (SD)	End point mean (SD)	Within-group effect size	Baseline mean (SD)	End point mean (SD)	Within-group effect size
“Easy” version
Hit rates	0.57 (0.23)	0.60 (0.24)^**	0.13	0.57 (0.24)	0.59 (0.24)	0.08	0.61 (0.23)	0.63 (0.25)	0.11
False alarms rates	0.09 (0.09)	0.06 (0.08)^***	–0.26	0.07 (0.10)	0.06 (0.09)	–0.08^a	0.07 (0.07)	0.05 (0.06)^***	–0.27
False mean response time (msec)	6.04 (0.31)	6.02 (0.31)	–0.04	6.06 (0.32)	6.01 (0.70)	–0.18	6.05 (0.31)	6.10 (0.36)	0.16
Correct mean response time (msec)	6.29 (0.32)	6.30 (0.28)	0.02	6.31 (0.32)	6.33 (0.25)	0.06	6.31 (0.27)	6.36 (0.28)^*	0.18
Br	0.16 (0.12)	0.13 (0.14)^**	–0.20	0.13 (0.14)	0.11 (0.12)	–0.15	0.14 (0.13)	0.11 (0.14)^*	–0.26
Pr	0.48 (0.28)	0.53 (0.29)^***	0.19	0.50 (0.28)	0.53 (0.29)	0.10	0.54 (0.28)	0.58 (0.28)	0.16

				Risperidone (n = 127)			Placebo (n = 131)
“Hard” version
Hit rates	0.47 (0.40)	0.49 (0.39)	0.07	0.45 (0.39)	0.52 (0.39)^*	0.17	0.50 (0.39)	0.55 (0.38)	0.15
False alarms rates	0.08 (0.10)	0.08 (0.10)	–0.06	0.08 (0.11)	0.08 (0.11)	–0.04	0.08 (0.09)	0.06 (0.08)	–0.14
False mean response time (msec)	6.29 (0.30)	6.28 (0.30)	–0.05	6.31 (0.31)	6.32 (0.29)	0.03^b	6.27 (0.30)	6.36 (0.26)	0.29
Correct mean response time (msec)	6.24 (0.29)	6.24 (0.24)	0.01	6.27 (0.26)	6.28 (0.26)	0.06	6.23 (0.23)	6.29 (0.24)^**	0.26
Br	0.25 (0.26)	0.24 (0.25)	–0.06	0.23 (0.25)	0.24 (0.25)	0.06	0.24 (0.26)	0.24 (0.24)	–0.02
Pr	0.38 (0.36)	0.42 (0.36)^*	0.10	0.37 (0.35)	0.44 (0.36)^*	0.20	0.42 (0.36)	0.49 (0.36)^*	0.20

Pr was calculated as the proportion of Hit Rate (phit) − proportion of False alarm rate (pfa); Br was calculated as pfa/(1 − Pr).

False alarms rates = lower score at end point indicates improvement; Response time = lower score at end point indicates improvement; Hit rates = higher score at end point indicates improvement; Pr = higher score at end point indicates improvement; Br = higher score at end point indicates a more liberal response bias level.

***

p < 0.001; ^** p < 0.01; ^* p < 0.05 for change from baseline.

p = 0.09.

p = 0.08 for difference between treatment groups in adjusted mean change from baseline in ANCOVA model including treatment group, country, somnolence, age (≤12 vs. >12), IQ (≤84 vs. >84)and baseline value as covariates.

Abbreviations: Pr = Probability of correct discrimination; Br = probability of response bias; ANCOVA = analysis of covariance; IQ = intelligence quotient.

Modified verbal learning test

The majority of patients in the study qualified to complete the MVLT-C-15 (chronological or mental age ≥6 years) and did not complete the MVLT-C-10. As a result, the small subpopulation of patients with chronological or mental age <6 years (only 9 patients treated with risperidone and 7 with placebo) was excluded, and only MVLT-C-15 data were analyzed. The MVLT-C short-delay free recall improved significantly among all subjects during this 6-month maintenance study (Table 4). Within-group effect sizes, calculated by mean change divided by baseline standard deviation (SD), were generally small to modest. Long-delay free recall decreased numerically for both risperidone and placebo subjects, although this difference is trivial and suggests relatively stable performance over time. After adjusting for country, somnolence, age, IQ, and baseline scores, there were no significant differences between the treatment groups. Using an ANCOVA model, there were no significant MVLT-C differences between patients with subnormal versus normal intellect, or between children (≤12 years) versus adolescents (>12 years).

Table 4.

MVLT-C 15 Values at Baseline and End Point

MVLT-C test	Open-label treatment			Double-blind treatment
	Risperidone (n = 429)			Risperidone (n = 136)			Placebo (n = 136)
	Baseline mean (SD)	End point mean (SD)	Within-group effect size	Baseline mean (SD)	End point mean (SD)	Within-group effect size	Baseline mean (SD)	End point mean (SD)	Within-group effect size
Learning trial total
Trial 1	5.55 (1.96)	7.48 (2.47)^***	0.99	7.18 (2.44)	8.04 (2.48)^***	0.35^a	7.18 (2.04)	8.43 (2.34)^***	0.61
Trial 2	7.81 (2.43)	9.25 (2.67)^***	0.59	9.03 (2.67)	9.71 (2.61)^**	0.25	9.03 (2.45)	9.86 (2.76)^***	0.34
Trial 3	9.30 (2.88)	10.23 (2.83)^***	0.32	10.03 (2.76)	10.83 (2.52)^***	0.29	10.10 (2.61)	10.85 (2.83)^***	0.28
Trial 4	10.25 (3.02)	10.80 (2.83)^***	0.18	10.65 (2.86)	11.13 (2.75)^*	0.17	10.82 (2.58)	11.35 (2.98)^**	0.20
Trial 5	10.69 (3.33)	11.37 (2.83)^***	0.20	11.18 (2.96)	11.76 (2.51)^**	0.20	11.40 (2.63)	11.63 (3.22)	0.08
Total short-delay free recall	43.61 (11.39)	49.00 (11.62)^***	0.47	48.07 (11.47)	51.47 (11.21)^***	0.30	48.54 (10.46)	52.12 (12.10)^***	0.34
Long-delay free recall	10.12 (2.98)	10.63 (3.25)^**	0.17	10.66 (2.87)	10.28 (4.32)	–0.13	10.79 (2.79)	10.34 (4.52)	–0.16
Pr	0.72 (0.29)	0.77 (0.23)^**	0.16	0.74 (0.21)	0.72 (0.30)	–0.09	0.74 (0.21)	0.75 (021)	0.04
Br	0.47 (0.14)	0.46 (0.14)	−0.06	0.47 (0.14)	0.47 (0.14)	0	0.46 (0.12)	0.48 (0.13)	0.18

Free recall = higher score at end point indicates improvement.

***

p < 0.001; ^** p < 0.01; ^* p < 0.05 for change from baseline.

p = 0.05 for difference between treatment groups in adjusted mean change from baseline in ANCOVA model including treatment group, country, somnolence, age (≤12 vs. >12), IQ (≤84 vs. >84) and baseline value as covariates.

Abbreviations: MVLT-C = modified version of the California Verbal Learning Test–Children's Version; SD = standard deviation; Pr = Probability of correct discrimination; Br = probability of response bias; ANCOVA = analysis of covariance; IQ = intelligence quotient.

Potentially clinically meaningful cognitive changes

Potentially clinically meaningful worsening and improvement, defined as a decrease or increase of ≥0.5 SD from baseline, respectively, are shown in Tables 5 and 6. Between-group comparisons of the percentage of subjects meeting these criteria were made with Fisher exact tests. With both cognitive assessment measures, patients were more likely to improve than worsen; this was true regardless of treatment condition. The majority of CPT tests showed similar changes for subjects treated with either risperidone or placebo. Overall, there were few statistically significant differences. Significantly more improvement was noted with placebo on the easy condition of the CPT for false alarms and hard correct mean response time (p < 0.05 for both). MVLT-C testing likewise showed similar clinical changes between subjects treated with risperidone or placebo for both learning and memory. There were no significant between-group differences in MVLT-C clinically meaningful changes. Data from both CPT and MVLT-C testing, therefore, support that treatment with risperidone does not result in cognitive impairment, with minimal between-group differences for patient treated with risperidone or placebo.

Table 5.

Percentage of Subjects with Clinical Worsening or Improvement in Continuous Performance Test Scores at End Point

	Risperidone (n = 127 )				Placebo (n = 131)
	Worsening (%)		Improvement (%)		Worsening (%)		Improvement (%)
	≥0.5 SD	≥1.0 SD	≥0.5 SD	≥1.0 SD	≥0.5 SD	≥1.0 SD	≥0.5 SD	≥1.0 SD
“Easy” version
Hit rates	18.11	5.51	26.77	8.66	19.85	8.40	29.01	8.40
False alarms rates	10.24	6.3	18.11	4.72	7.63	4.58	29.77^a	11.45
False mean response time (msec)	28.00	13.00	27.00	14.00	30.43	17.39	19.57	13.04
Correct mean response time (msec)	22.05	8.66	18.11	6.30	32.06	14.50	16.03	7.63
Pr	17.32	5.51	26.77	9.45	17.56	6.87	28.24	9.16
“Hard” version
Hit rates	13.39	9.45	22.05	18.11	12.21	6.11	19.85	16.79
False alarms rates	18.90	7.87	18.90	12.60	19.08	11.45	23.66	16.03
False mean response time (msec)	29.41	10.29	25.00	13.24	38.67	17.33	16.00	6.67
Correct mean response time (msec)	20.90	8.96	19.40	5.97	40.54^a	18.92	21.62	4.05
Pr	13.39	8.66	22.83	8.11	12.98	6.11	20.61	16.79

p < 0.05 for difference between treatment groups.

Abbreviations: SD = Standard deviation; Pr = probability of correct discrimination.

Table 6.

Percentage of Subjects with Clinical Worsening or Improvement in MVLT-C Scores at End Point

	Risperidone (n = 125)				Placebo (n = 133)
	Worsening (%)		Improvement (%)		Worsening (%)		Improvement (%)
	≥0.5 SD	≥1.0 SD	≥0.5 SD	≥1.0 SD	≥0.5 SD	≥1.0 SD	≥0.5 SD	≥1.0 SD
Learning trial total, trial 5	16.30	8.89	30.37	20.74	12.78	7.52	31.58	15.04
Total short-delay free recall	12.50	4.41	41.91	13.24	11.0.3	4.41	43.38	16.18
Long-delay free recall	11.57	5.79	36.86	22.31	11.67	3.33	35.00	15.83

Abbreviations: MVLT-C = modified California Verbal Learning Test for Children; SD = standard deviation.

Somnolence

Adverse Events of Somnolence

Somnolence was recorded as a treatment-emergent AE for 61 patients (11.6%) during acute, 6-week, open-label risperidone treatment and 10 patients (2.3%) during the single-blind, 6-week continuation phase. Incidence and severity of treatment-emergent somnolence decreased with continued risperidone treatment. Most early cases of AEs of somnolence were considered to be either probably or very likely related to risperidone. During double-blind maintenance treatment, treatment-emergent somnolence was reported by only 3 risperidone (2.1%) and 3 placebo (2.1%) patients. Two risperidone and all placebo cases were considered mild, with 1 risperidone case rated as moderate. There were no cases of severe sedation. Two of these cases (1 risperidone and 1 placebo) were considered to be probably or very likely related to study drug. Among those 6 patients reporting somnolence, mean time to first report of somnolence was 51.3 days with risperidone and 154.8 days with placebo. The mean duration of somnolence was 34.3 ± 42.0 days with risperidone and 42.3 ± 50.0 days with placebo.

Because the presence of somnolence as an AE can potentially influence cognitive testing results, a subgroup analysis of change in cognitive from open-label baseline to open-label end point was performed comparing outcome in patients with and without somnolence as an AE. There were no significant differences in the change in any CPT or MVLT-C values based on the presence or absence of somnolence as an AE.

Discussion

The current study expands on previously published data showing lack of negative cognitive effects of short- and long-term risperidone in children with DBD and subnormal intellect (Pandina et al. 2007) by reporting comparable results in a sample that included children and adolescents with average intellect. In the current study, cognitive testing results after 6 months were similar in patients who initially responded to short-term risperidone and were subsequently continued on maintenance risperidone or discontinued risperidone by using placebo. These data support that continuation of previously effective risperidone does not result in meaningful cognitive decline and there is no cognitive advantage to a drug holiday. Covariate testing of cognitive changes was not significantly affected by baseline IQ, suggesting similar results for patients with either subnormal or normal intellect, irrespective of age (i.e., for both children and adolescents). In the Pandina et al. (2007) sample, patients >12 years old were only included in one of the three open-label studies and neither double-blind study. In the current study, most cognitive changes were not significantly affected by age. These data suggest that long-term risperidone use in children or adolescents of normal or subaverage intellect will generally not result in significant medication-related cognitive decline or sedation.

Sustained attention did not appear to be impaired with risperidone treatment in the current study. With both easy and hard CPT testing, within-group treatment effects were generally small to modest, with few significant changes. Overall, changes were similar for subjects assigned to risperidone or placebo. Verbal learning and memory significantly improved for both risperidone- and placebo-treated patients. Changes from baseline were similar between treatment groups, suggesting a lack of effect of long-term risperidone on verbal learning and memory.

Somnolence was most likely to occur during the initial 6-week risperidone treatment, with treatment-emergent somnolence incidence and severity decreasing with persistent therapy. Treatment-emergent somnolence occurred in 12% of patients with acute risperidone treatment and was infrequently reported with continuation and maintenance therapy. These data support the findings in a systematic literature review of atypical antipsychotics in DBDs, which concluded that the somnolence associated with antipsychotic treatment is typically mild and of transient nature (Jensen et al. 2007). Changes in cognitive testing were not significantly affected by the presence or absence of somnolence as an AE.

Finally, risperidone did not appear to have a negative impact on cognition compared with placebo as assessed by potentially clinically meaningful changes in cognitive test performance.

Limitations

There are a number of limitations to both the design and the results of current study. First, the dose range used in the current study was narrow and the overall mean modal dose (0.91 ± 0.38 mg/day) was low, despite the flexible dose option. Thus, no conclusions can be drawn regarding higher doses. Due to the narrow dose range selected by treating clinicians for patients in the current study based on efficacy, an evaluation of the influence of higher versus lower dosing was not possible.

Only responders to risperidone treatment were eligible for double-blind treatment, so those who had initial intolerance or lack of efficacy would have dropped from the trial. This is tempered by the fact that there was a high response rate overall in the trial (67.8% of subjects treated with risperidone), suggesting that the results may be generalizable. The design of this study does mirror clinical practice, however, because only patients who respond and tolerate the drug are continued on treatment over the long term, this design represents an opportunity to see if patients removed from the drug show differences in their cognitive functioning over time.

In addition, there is no normal comparison group to determine the extent of disability associated with cognitive scores in this study. Some of the apparent cognitive impairment identified in this study may reflect excessive caution among subjects. For example, a hit rate of about 50% on CPT testing suggests a poor ability to detect targets; however, the false alarm rates were very low (<10%), suggesting a possible tendency for these subjects to under respond. This hypothesis is supported by the relatively low Br scores, also suggesting subject caution. Similarly, comparison to a normal group was also not available for MVLT-C testing to determine extent of deficiency at baseline, although increasing performance on the learning trials to remembering an average of 11 of 15 words demonstrates learning.

Another limitation of the relapse prevention design is that, due to the earlier and more frequent discontinuation in the placebo group, the placebo group had fewer completers and a shorter duration of treatment. Furthermore, this study was not designed to compare attributes of those individuals who might have worsened versus those who did not. Response rate was not analyzed to assess the effect of covariates or subgroups as possible predictors of cognitive change. Last, only two cognitive measures were used in the study, and the versions used did not completely assess the full range of cognitive function in these domains.

Conclusion

In summary, this double-blind study shows that long-term treatment with risperidone in children and adolescents with a DBD and normal intellect is generally not associated with potential medication-related negative cognitive effects or sedation. Changes in attention, verbal learning, and memory testing were similar between risperidone and placebo groups and did not appear to be affected by somnolence. These results are consistent with those observed in earlier long-term studies of children with subaverage intellect.

Footnotes

Disclosures

Drs. Pandina and Zhu are full-time employees of Johnson and Johnson Pharmaceutical Research and Development. Dr. Cornblatt is a current consultant to Xenoport.

This study was supported by Johnson and Johnson Pharmaceutical Research and Development.

References

Aman

, Tassé

, Rojahn

, Hammer

. The Nisonger CBRF: A child behavior rating form for children with developmental disabilities. Res Dev Disabilities, 17:41–57. 1996.

American Psychiatric Association. Diagnostic, Statistical Manual of Mental Disorders 4^th edition (DSM-IV) Washington (DC): American Psychiatric Asosociation, 1994.

Angold

, Erkanli

, Farmer

, Fairbank

, Burns

, Keller

, Costello

. Psychiatric disorder, impairment, and service use in rural African American and white youth. Arch Gen Psychiatry, 59:893–901. 2002.

Canino

, Shrout

, Rubio-Stipec

, Bird

, Bravo

, Ramirez

, Chavez

, Alegria

, Bauermeister

, Hohmann

, Ribera

, Garcia

, Martinez-Taboas

. The DSM-IV rates of child and adolescent disorders in Puerto Rico: Prevalence, correlates, service use, and the effects of impairment. Arch Gen Psychiatry, 61:85–93. 2004.

Daban

, Amado

, Bourdel

, Loo

, Olié

, Poirier

, Krebs

. Cognitive dysfunctions in medicated and unmedicated patients with recent-onset schizophrenia. J Psych Res, 39:391–398. 2005.

Delis

, Kramer

, Kaplan

, Ober

. California Verbal Learning Test Children's version, manual. Orlando (Florida): The Psychological Corp/Harcourt, Brace and Co., 1987.

Harvey

, Green

, McGurk

, Meltzer

. Changes in cognitive functioning with risperidone and olanzapine treatment: A large-scale, double-blind, randomized study. Psychopharmacology, 169:404–411. 2003.

Jensen

, Buitelaar

, Pandina

, Binder

, Haas

. Management of psychiatric disorders in children and adolescents with atypical antipsychotics. A systematic review of published clinical trials. Eur Child Adolesc Psychiatry, 16:104–120. 2007.

Keefe

, Bilder

, Davis

et al. Neurocognitive effects of antipsychotic medications in patients with chronic schizophrenia in the CATIE Trial. Arch Gen Psychiatry, 64:633–647. 2007.

10.

Kutcher

, Aman

, Brooks

, Buitelaar

, van Daalen

, Fegert

, Findling

, Fisman

, Greenhill

, Huss

, Kusumakar

, Pine

, Taylor

, Tyano

. International consensus statement on attention-deficit/hyperactivity disorder (ADHD) and disruptive behavior disorders (DBDs): Clinical implications and treatment practice suggestions. Eur Neuropsychopharmacol, 14:11–28. 2004.

11.

Jensen

, Buitelaar

, Pandina

, Binder

, Haas

. Management of psychiatric disorders in children and adolescents with atypical antipsychotics: A systematic review of published clinical trials. Eur Child Adolesc Psychiatry, 16:104–120. 2007.

12.

Lahey

, Loeber

, Burke

, Rathouz

. Adolescent outcomes of childhood conduct disorder among clinic-referred boys: Predictors of improvement. J Abnorm Child Psychol, 30:333–348. 2002.

13.

Pandina

, Bilder

, Harvey

, Keefe

, Aman

, Gharabawi

. Risperidone and cognitive function in children with disruptive behavior disorders. Biol Psychiatry, 62:226–234. 2007.

14.

Rani

, Murray

, Byrne

, Wong

. Epidemiologic features of antipsychotic prescribing to children and adolescents in primary care in the United Kingdom. Pediatrics, 121:1002–1009. 2008.

15.

Reyes

, Buitelaar

, Toren

, Augustyns

, Eerdenkens

. A randomized, double-blind, placebo-controlled study of risperidone maintenance treatment in children and adolescents with disruptive behavior disorders. Am J Psychiatry, 163:402–410. 2006.

16.

Snodgrass

, Corwin

. Pragmatics of measuring recognition memory: Applications to dementia and amnesia. J Exp Psychol, 117:34–50. 1988.

17.

Tassé

, Aman

, Rojahn

. The Nisonger Child Behavior rating form: Age and Gender effects and Norms. Res Dev Disabilities, 17:59–75. 1996.

18.

Weickert

, Goldberg

, Marenco

, Bigelo

, Egan

, Weinberger

. Comparison of cognitive performances during a placebo period and an atypical antipsychotic treatment period in schizophrenia: Critical examination of confounds. Neuropsychopharmacology, 28:1491–1500. 2003.