Weight-Change Trajectories of Pediatric Outpatients Treated with Risperidone or Aripiprazole in a Naturalistic Setting

Abstract

Objectives:

Second-generation antipsychotics (SGAs) increase appetite and weight, leading toward a metabolic syndrome. Risperidone and aripiprazole, the most widely used pediatric SGAs, have been studied predominantly in short-term clinical trials, where risperidone leads to a rapid weight increase and aripiprazole to a slower one, while long-term effects are not yet elucidated. Factors that may influence weight gain are likewise not clarified, although baseline weight, previous SGA exposure, pubertal status, and type of SGA have been suggested as moderators. We analyzed weight gain induced by risperidone and aripiprazole in a sample of pediatric outpatients enrolled into a 2-year observational study.

Methods:

We assessed at several time points their body mass index (BMI)-Z scores (age and sex-corrected and referred to national norms). We used hierarchical mixed-effects modeling to design BMI-Z trajectories and observed the effects of several variables on determining them.

Results:

The study group comprised of 127 patients, predominantly males (79%), of 12.6 years on average, treated with risperidone (81%) and aripiprazole (19%) for disruptive behavioral symptoms in patients with and without neurodevelopmental disorders. Overall, BMI-Z was 1.2 at first and 1.4 at last visit (no significant change). We could design four weight-change trajectories, determined by the factors: drug (risperidone/aripiprazole) and age status (children/adolescent). Additional factors not retained in the model but possibly explanatory include the previous duration of SGA treatment and a progressive patient-selection effect due to dropouts in this observational study. Risperidone treatment was associated with trends of BMI-Z increase in children and decrease in adolescents. Aripiprazole treatment was associated with significant BMI-Z increase, higher in children than in adolescents. Results are probably due to longer previous drug exposure in adolescents.

Conclusions:

Children were at risk of weight gain more than adolescents, for both risperidone and, of note, aripiprazole. Adolescents and patients with long previous drug exposure tend to reach stable BMI-Z, although in the range between excessive weight and obesity.

Introduction

Almost all second-generation antipsychotics (SGAs) present with a risk of inducing appetite increase and weight gain, together with features related to the metabolic syndrome: weight gain, central obesity, dyslipidemia, hyperglycemia/hyperinsulinemia, and hypertension (Correll et al. 2009; De Hert et al. 2011; Maayan and Correll 2011; Ebdrup et al. 2014; Pisano et al. 2013, 2016; Galling et al. 2016; Zhang et al. 2016).

Considering the pediatric use of risperidone and aripiprazole, the most compelling evidence comes from short-term studies. Pediatric patients using risperidone described in these studies have been consistently reported to gain weight during clinical trials, although the time span of observations was limited to 2–3 months on average (Geller et al. 2012). Risperidone displayed a consistent short-term effect, replicated in several studies; for example, the body mass index (BMI) of drug-naïve adolescents increased 3.5-fold as compared with that of controls, after 1 month of treatment with risperidone (Correll et al. 2009). Aripiprazole also displayed a weight-increasing effect in these patients. For example, in autistic patients treated for irritability, weight increased significantly during 2 months of aripiprazole therapy, more pronouncedly in drug-naïve patients as compared with those already under treatment (1.2 kg vs. 0.9 kg on average) (Mankoski et al. 2013). Together with previous exposure to SGAs, the pre-/postpubertal age has been suggested as a relevant factor for weight gain. Part of the effect of age may be mediated by the different duration of previous antipsychotic exposure, where younger patients experience a significantly greater increase in weight and blood glucose levels as compared with older patients (Safer 2004; Mankoski et al. 2013; Arango et al. 2014).

Studies describing longer observation times produced heterogeneous results; a study reported weight gain as maximal during the first 6 months of therapy, and then lessening (Calarge et al. 2012), while others reported a constant weight gain (Calarge et al. 2014). A 6-month-long study investigated naïve patients, finding that risperidone was associated with a significantly higher proportion of patients gaining +0.5 BMI-Z as compared with another SGA such as quetiapine (Calarge et al. 2012; Masi et al. 2015). A discontinuation study was conducted on patients previously treated with risperidone for 18 months, showing that the effect of risperidone was not reversible, at least during the 6 months of perspective study follow-up (Noguera et al. 2013). A 24-month comparative risk study demonstrated a smaller increase in BMI-Z at 6 months for patients treated with risperidone, rather than olanzapine and quetiapine, while no differences among SGAs were found at 24 months (Pérez-Iglesias et al. 2014). Another 2-year-long study found similar weight increases for risperidone (+0.5 BMI-Z per year) and aripiprazole (+0.6), questioning the assumption that aripiprazole might be less risky than risperidone (Simon et al. 2009), an assumption based on the observation of a weaker effect of aripiprazole on weight gain in short-term clinical trials (Correll et al. 2009; De Hert et al. 2011; Maayan and Correll 2011; Pisano et al. 2016). One 3-year-long study reported almost all SGAs having the same overall effect, although with different time courses depending upon the SGA (Simon et al. 2009; Pérez-Iglesias et al. 2014), an observation that needs confirmation in larger long-term studies.

Although an overall effect of promoting weight gain is demonstrated for SGAs and covariates that influence weight gain have been studied in samples of mixed-age patients, the trajectories of weight gain over time and the factors that influence trajectories are still unclear. Indeed, when considering a population of developmental age, the excessive weight gain due to treatment must be separated from the physiological weight gain due to growth. One practical way is to analyze the weight gain trajectories. We here analyzed a sample of pediatric outpatients followed through a 2-year observational study, to (1) design trajectories possibly different between risperidone and aripiprazole; (2) identify putative factors, such as age and naive status, which may influence the trajectories. We focused on BMI-Z measures at the study baseline and during follow-up, and we designed a statistical model to describe BMI-Z change trajectories. Estimates from the model were then used to report on BMI-Z at the baseline and on mean differences between the baseline and prespecified follow-up times (3, 6, and 12 months).

Methods

This study is based on data collected within an active pharmacovigilance project on the safety and efficacy of four SGAs (risperidone, aripiprazole, olanzapine, and quetiapine) in children and adolescents. It was approved by the Ethics Committees of the participating structures. The project was carried out in Italy between March 2012 and March 2014 by two pharmacology units (for monitoring and laboratory examinations) and three tertiary-care neuropsychiatry departments (for recruitment and clinical evaluation). These clinical centers follow only outpatients, undergoing routine neuropsychiatric rehabilitation (no inpatients or emergency admissions) for autism spectrum disorders, intellectual disability, disruptive behavioral disorders, tic disorders, and associated comorbidities. Regarding the treatment of disruptive behaviors in Italy, risperidone is labeled for the short-term (6 weeks) treatment of patients with subaverage IQ, while no other SGA is labeled. Frequent off-label uses for disruptive behavioral disorders comprise the extension of risperidone therapies in time and to patients without intellectual disability, the use of aripiprazole in second line, and the use of olanzapine or quetiapine in third line.

This study sample consisted of 166 patients, aged 6–17 (median age 12.6 years, 1st–3rd quartiles 9.6–14.6 years), treated with antipsychotic monotherapy (risperidone or aripiprazole) for disruptive behavioral disorders, with or without autism spectrum disorders and/or intellectual disability, followed for up to 2 years (see Supplementary Table S1 for more details on the patient sample). There were no explicit exclusion criteria regarding syndromes; we only included one syndromic patient with Down syndrome in the project sample. No patient was affected by Prader–Willi syndrome. Moreover, to our knowledge, no patient was affected by concomitant conditions that could influence eating habits. Relevant data of eligible subjects were reanalyzed, which are obtained from a previous study (Pozzi et al. 2016a, 2016b; Rafaniello et al. 2016, 2018). Throughout the whole follow-up period, weight and height data were recorded during visits, separated by indicative intervals of 3 months, as allowed by the clinical routine. A data set was assembled, comprising only patients who did not switch SGA monotherapy and who had at least two data recordings; the data set consisted of 127 patients with 492 total observations. BMI scores were calculated using weight and height, then BMI-Z scores were calculated confronting each patient at each observation with Italian normative data (Cacciari et al. 2006), taking age and sex into account. Patients' essential characteristics were presented as medians with first and third quartile (follow-up time; age at baseline; duration of previous SGA therapy; BMI; BMI-Z; clinical global impression score-severity [CGI-S]; and drug dose), and as frequency counts and percentages for categorical variables (sex, age category, and drug used). The CGI-S (Guy 1976) is a Likert scale to be filled out by the clinician based on his/her impression. It consists of one item only; its scoring ranges from 1 to 7. The scores mean that the patient is 1 “Normal, not at all ill”; 2 “Borderline mentally ill”; 3 “Mildly ill”; 4 “Moderately ill”; 5 “Markedly ill”; 6 “Severely ill”; and 7 “Among the most extremely ill patients.” This is a widespread quick tool to provide an overview of patient's severity and direct the path of care. Based on previous observations from the literature, the variables considered a priori as potentially associated with changes of BMI-Z scores during therapy were as follows: drug used (risperidone or aripiprazole), age at baseline (continuous or categorized as: children ≤13 years or adolescents >13 years), and previous antipsychotic therapy (duration, or yes or no). We did not collect objective data on the puberty status; thus, the cutoff age of 13.0 years was chosen to clearly distinguish children from adolescents, as International Conference for Harmonisation guidelines define 12 years as the beginning of adolescence (ICH 2000).

Follow-up time was defined as the time elapsed, in months, between the patient's first visit and the following ones.

To describe the sample with respect to BMI-Z variations, a table was used to report the rounded-off BMI-Z score numbers of patients at the beginning and end of the follow-up.

To analyze the trajectories of BMI-Z scores, hierarchical mixed-effects modeling techniques were used (Raudenbush and Bryk 2002; Gueorguieva and Krystal 2004). These techniques are designed for flexible modeling of changes over time, and accounting for the correlation among repeated measurements taken on each subject. In a preliminary step, a model including only follow-up time as covariate was fitted, to estimate the mean overall trend of BMI-Z scores. Then, to assess the association of Z-score trajectories with the variables of interest, a model was estimated, including as covariates: follow-up time, treatment (aripiprazole or risperidone), age category (child or adolescent), and previous antipsychotic therapy. The association between each covariate and BMI-Z score trajectories was assessed through likelihood ratio (LR) tests: in this task, the associations with a p-value <0.10 were deemed significant. In the second step, the covariates without a significant association were not included in the model. Then, estimated trajectories of mean BMI-Z scores and 95% confidence intervals (CIs) were reported. Furthermore, to evaluate weight gain, estimated mean changes of BMI-Z scores at 3, 6, and 12 months were reported with 95% CIs. The analysis was performed using the R software version 3.3.2 (R Core Team 2016) with the nlme package (Pinheiro et al. 2016).

Results

Patient characteristics

This study is based on the analysis of data collected from 166 pediatric outpatients treated with SGAs and followed through an observational study (Pozzi et al. 2016a, 2016b; Rafaniello et al. 2016, 2018), from which 127 patients were selected, who had at least two measurements of height and weight, and were treated with antipsychotic monotherapies. All patients who had measurements taken under different therapies, that is, who switched treatment (n = 9) or had less than two measurements (n = 30), were excluded. Of the 127 included patients, 69 (54.3%) were drug naïve, and 76 (59.8%) were antipsychotic naïve. The seven patients who used other drugs before beginning treatment with SGAs used melatonin for sleep disorders (5) and valproic acid for seizure prophylaxis (2). Among patients, 103 (81.1%) were treated with risperidone and 24 (18.9%) with aripiprazole. Patients who were not SGA naïve had a heterogeneous duration of previous exposure. The median (1st–3rd quartile) follow-up period of observation for these patients was 6.8 (4.6–11.1) months, with a median interval between observations of 66 (42–106) days, with 22.8% patients followed for >1 year, 39.4%> 6 months, and 26.0%> 3 months. The median number of visits per patient was 3 (2–5). Relevant characteristics of patients are summarized in Table 1. Overall, the median BMI was 20.9 (18.2–25.5) at the first visit and 21.6 (19.5–26.7) at the last visit; Table 2 reports the rounded-off BMI-Z scores of patients at the beginning and at the end of follow-up.

Table 1.

Summary of Patient Characteristics

	Risperidone	Risperidone	Aripiprazole	Aripiprazole
	Children	Adolescents	Children	Adolescents	Total
Patients, n (%)	68 (53.5)	35 (27.6)	7 (5.5)	17 (13.4)	127
Age at baseline (years)	10.0 (8.4, 11.2)	14.8 (13.7, 15.3)	11.5 (10.8, 12.6)	15.0 (13.3, 16.1)	12.6 (9.6, 14.6)
Sex: females, n (%)	11 (16.2)	5 (14.3)	3^* (45.8)	8^* (47)	27 (21.3)
Drug-naïve patients, n (%)	41 (60.3)	17 (48.6)	2 (28.6)	9 (52.9)	69 (54.3)
Duration of previous SGA exposure (days)	0 (0, 47)	0 (0, 470)	55 (0, 142)	0 (0, 76)	0 (0, 137)
BMI at baseline	20.1 (17.2, 23.7)	22.5 (19.9, 22.5)	20.6 (16.6, 22.7)	25.4 (19.5, 31.0)	20.9 (18.2, 25.5)
BMI-Z at baseline	1.57 (0.39, 2.54)	1.14 (−0.09, 2.11)	0.99 (−0.98, 1.96)	1.17 (0.14, 2.53)	1.34 (0.16, 2.33)
CGI-S at baseline	4 (3, 5)	4 (3, 5)	4 (4, 5.5)	4 (4, 5.25)	4 (3, 5)
Drug dose (mg/kg per day)	0.026 (0.015, 0.048)	0.024 (0.016, 0.039)	0.11 (0.06, 0.24)	0.13 (0.09, 0.19)	—

Patients and sex are described as numbers with percentages; continuous variables are reported as medians, with first and third quartile. The frequency of discrete variables was compared across groups by chi-square tests with subgroup comparisons.

Indicate values significantly different from those without ^*(p < 0.05). The distribution of continuous variables was compared across groups by Mann–Whitney tests; no significant difference was found.

BMI, body mass index; CGI-S, clinical global impression score-severity; SGA, second-generation antipsychotic.

Table 2.

Description of Body Mass Index-Z Score Classes at the First and Last Study Visit

BMI-Z score	At first visit	At last visit
−3	1	1
−2	5	4
−1	7	6
0	29	25
+1	26	31
+2	33	29
+3	21	25
+4	3	4
+5	2	2

Numbers represent the number of patients with each BMI-Z score (rounded off to the nearest unit) at the beginning and at the end of study follow-up.

BMI, body mass index.

The estimates of the overall trend of BMI-Z scores indicate an average increase of 0.013 points per month, with a 95% CI of −0.0002, 0.027 points per month (Table 3). Since the lower limit of the CI is negative, statistical significance cannot be claimed.

Table 3.

Estimates of Linear Trend Coefficients and Body Mass Index-Z Changes

Subgroup	Linear trend coefficients: estimates (95% CI)	Mean BMI-Z values: estimates (95% CI)	BMI-Z change from baseline: estimates (95% CI)
Overall	Intercept: 1.246 (0.985 to 1.506) Slope: 0.013 (−0.0002 to 0.027)	At 0 months: 1.246 (0.985 to 1.506) At 3 months: 1.285 (1.031 to 1.540) At 6 months: 1.325 (1.070 to 1.579) At 12 months: 1.403 (1.131 to 1.676)	At 3 months: +0.039 (−0.0006 to +0.079) At 6 months: +0.079 (−0.001 to +0.159) At 12 months: +0.158 (−0.002 to +0.317)
Risperidone Children	Intercept: 1.425 (1.083 to 1.767) Slope: 0.016 (−0.001 to 0.032)	At 0 months: 1.425 (1.083 to 1.767) At 3 months: 1.472 (1.139 to 1.805) At 6 months: 1.519 (1.188 to 1.851) At 12 months: 1.617 (1.286 to 1.948)	At 3 months: +0.048 (−0.003 to +0.096) At 6 months: +0.096 (−0.006 to +0.192) At 12 months: +0.192 (−0.012 to +0.374)
Risperidone Adolescents	Intercept: 1.084 (0.622 to 1.546) Slope: −0.012 (−0.034 to 0.011)	At 0 months: 1.084 (0.622 to 1.546) At 3 months: 1.049 (0.599 to 1.499) At 6 months: 1.015 (0.567 to 1.462) At 12 months: 0.945 (0.472 to 1.418)	At 3 months: −0.036 (−0.102 to +0.033) At 6 months: −0.072 (−0.204 to +0.066) At 12 months: −0.144 (−0.408 to +0.132)
Aripiprazole Children	Intercept: 1.206 (0.497 to 1.915) Slope: 0.066 (0.029 to 0.102)	At 0 months: 1.206 (0.497 to 1.915) At 3 months: 1.403 (0.712 to 2.094) At 6 months: 1.601 (0.911 to 2.290) At 12 months: NE	At 3 months: +0.198 (+0.087 to +0.306) At 6 months: +0.396 (+0.174 to +0.612) At 12 months: NE
Aripiprazole Adolescents	Intercept: 0.865 (0.251 to 1.479) Slope: 0.039 (0.008 to 0.069)	At 0 months: 0.865 (0.251 to 1.479) At 3 months: 0.980 (0.382 to 1.579) At 6 months: 1.096 (0.499 to 1.693) At 12 months: 1.327 (0.692 to 1.962)	At 3 months: +0.117 (+0.024 to +0.207) At 6 months: +0.234 (+0.048 to +0.414) At 12 months: +0.468 (+0.096 to +0.828)

Reported in the table are, both for the overall sample (first row) and for each subgroup of patients: estimates of intercepts and slopes of mean BMI-Z score trajectories (second column); estimated mean values of BMI-Z scores at 0, 3, 6, and 12 months (third column); estimated differences between mean BMI-Z scores at 3, 6, and 9 months and at first visit (fourth column). The intercept corresponds to the value of mean BMI-Z score at time = 0 (first visit). The slope is equal to the size of increase or decrease in mean BMI-Z score over a 1-month period (independent of the initial time).

BMI, body mass index; CI, confidence interval; NE, not estimated to avoid extrapolation.

Model proposed to describe BMI-Z change trajectories

Details on the steps that led to the final proposed model are available in Supplementary Table S2. Within the final model, only the type of drug (aripiprazole or risperidone; LR test: p < 0.01) and the patients' age category at baseline visit (children or adolescents; LR test: p = 0.05) were significantly associated with the variations of BMI-Z scores, while the association with previous therapy was not significant (p = 0.21). The final model represents the trajectories of average BMI-Z scores of the four subgroups of patients, according to the combinations of treatment and age: children treated with risperidone, adolescents treated with risperidone, children treated with aripiprazole, and adolescents treated with aripiprazole.

BMI-Z changes

As shown in Figure 1, each BMI-Z score trajectory consists of a straight line; that is, a linear trend. The corresponding coefficients for intercept and slope are reported in Table 3: the intercept corresponds to the mean value of BMI-Z scores at the baseline visit. The slope represents the amount of change (increase or decrease) in mean BMI-Z scores over a period of 1 month, independent of the starting time point considered. The estimates of mean BMI-Z scores at the baseline visit ranged between +1.425 and +0.865, with no significant difference between groups (Table 3). The estimates of the increase of mean BMI-Z score (slopes) for each group of patients are also reported in Table 3. For patients who received aripiprazole, they indicate that the BMI-Z scores increased significantly over time, both for children (+0.066 per month) and for adolescents (+0.039 per month). The estimated difference between the two slopes was 0.027, with a significant 95% CI of 0.0003, 0.054; this suggests a steeper increase in children than in adolescents. For patients treated with risperidone, the trend of BMI-Z change tended in the direction of a BMI-Z increase (+0.016 per month) although not reaching statistical significance (the lower 95% CI was −0.001 per month) in the children group; no change of BMI-Z scores in the adolescent age group was instead observed.

FIG. 1.

Estimated trajectories of average BMI-Z score according to treatment and age. Solid lines: estimates; dashed lines: 95% confidence intervals. BMI, body mass index.

Discussion

Increased appetite and weight gain are among the most frequent causes of loss of adherence to treatment and reduced persistence in therapy, affecting negatively patients' health (Pozzi 2016b). Pediatric patients exposed to SGAs also demonstrate higher incidence of diabetes in the adulthood (Galling et al. 2016). Weight gain increasingly becomes a limiting factor with long-term use of SGA, as in rehabilitation settings (Pozzi 2016b; Rafaniello et al. 2016). A long-term analysis may thus clarify the role of possible risk factors and suggest better management strategies. Up to date, younger age and higher baseline BMI were associated with less weight gain in pediatric bipolar patients (Baeza et al. 2017; Patel et al. 2017). Conversely, another study found that BMI scores increased significantly more in children <7 years of age, as compared with older ones (Scahill et al. 2016). The role of age, BMI at baseline, drug-naïve status, and drug dose in shaping weight trajectories still needs to be investigated further to allow proper patient management (Pringsheim et al. 2011; Rafaniello et al. 2016), a task that we addressed in this work.

Proposed model and possible confounders

In this analysis of observational data, we focused on the changes in BMI-Z scores of child and adolescent patients treated with risperidone or aripiprazole in a naturalistic neuropsychiatric rehabilitation setting. We found that risperidone, and unexpectedly even more aripiprazole, tilt the weight-change trajectory toward excessive gain until a plateau is reached, in between excessive weight and obesity, after which weight gain reduction, or for risperidone even weight loss, ensue. We chose to investigate BMI-Z scores, rather than uncorrected BMI, for two reasons: the first is that in the developmental age, the BMI value alone could be misleading, as the definition of overweight changes across the years, because of the redistribution of weight across body districts. For example, the 50th percentile of BMI can vary from 15 at 5 years to 22 at 16 years of age. The second reason is that the perception of excessive weight is culturally influenced, and this was taken into account by using Italian BMI-Z normative data. A stepwise model was built to describe potential influences on BMI-Z scores from clinical and demographical variables, and the model automatically discriminated which factors were significant at each step. The effect of drugs was probed in the model in two distinct ways. Both previous and current therapies were considered as potential continuous or categorical variables. Regarding previous therapies, time was retained as a categorical variable (previous SGA: yes or no) in the second step, but this factor was eliminated in the third. Regarding current therapies, time was retained as a continuous factor until the final model. This procedure led us into finding significant contributions only from the time elapsed (meaning that there was a change over time), from the drug used (aripiprazole or risperidone), and from patient's age category (as child or adolescent). In detail, the drug effect was that patients treated with aripiprazole gained more weight than those treated with risperidone; the age effect was that adolescent patients gained less BMI-Z as compared with drug-matched children. Clinical and demographical factors can explain these results.

Underlying role of the duration of past treatment with SGAs

In our study, patients had already been in therapy with risperidone for an average of 4 months (children) or even 10 months (adolescents), while those on aripiprazole had an average duration of previous aripiprazole therapies of 3 months (children) and 4.5 months (adolescents). Previous therapies not based on SGAs regarded melatonin (5) and valproic acid (2), which on their own should not have a sizable effect on weight gain. A number of clinical trials have shown that weight gain due to antipsychotics, in particular to risperidone, occurs predominantly during the first 4–6 months of treatment, after which a progressive stabilization occurs (Simon et al. 2009; Calarge et al. 2012; Noguera et al. 2013; Pérez-Iglesias et al. 2014). This is in line with our finding that the group of patients with shortest previous exposure to SGAs showed the highest increase in BMI-Z (children with aripiprazole) and the group with longest previous exposure to SGAs showed the highest decrease (adolescents with risperidone). This may support the existence of a plateau effect due to the time of exposure to SGAs, especially for risperidone.

In a naturalistic setting, factors contributing to this effect may be mostly connected with clinical management strategies. As shown in previous reports on the population of this observational study, the occurrence of adverse drug reactions was high (Rafaniello et al. 2016) and weight gain was the commonest treatment-limiting adverse reaction (Pozzi et al. 2016b). We had also observed that a large number of patients were withdrawn from treatment by caregivers, because of subclinical adverse effects, which also involved weight gain (Pozzi et al. 2016b). This consideration highlights a possible selection effect: patients who gain too much weight in the context of routine clinical practice can either quit (or switch) treatment or receive a dose reduction, ideally with a counseling on feeding habits and physical activity. The population followed in this observational study, however, is composed predominantly of patients with disruptive behavioral disorders associated with intellectual disability or autism spectrum disorders. It is difficult for caregivers to regulate the eating habits and physical activity of these children; therefore, patients who maintain drug treatment are most likely those with a low propensity of gaining excessive weight. This may explain the BMI-Z trend we observed in adolescents treated with risperidone, who had been already able to tolerate ∼10 months of risperidone therapy: this patient group may have been naturally selected for tolerance to weight gain; that is, they did not quit or switch therapy. Interestingly, the other three groups of patients that show increasing trends of BMI-Z (significant for aripiprazole, not significant for children on risperidone) all had a lower and comparable duration of previous exposure to SGAs.

Underlying roles of age, BMI, BMI-Z, and baseline scores

The group of children treated with risperidone was ∼2 years younger than children treated with aripiprazole, suggesting a stronger age effect on the conversion from BMI to BMI-Z (Cacciari et al. 2006). This may contribute to explain the differences in BMI-Z scores observed at baseline in the risperidone- vs. aripiprazole-treated patients. Children treated with risperidone started from a high score (1.5) and did not show a significant trend of BMI-Z increase. Conversely, children treated with aripiprazole had baseline BMI-Z scores of 0.8, indicating that they were less than overweight on average and were observed to gain BMI-Z significantly. Although this demands confirmation, it is interesting to speculate that a plateau effect may also be determined by a physical upper limit to BMI, imposed by the younger age of children. Patients under risperidone may have already gained much of the possible weight excess and may be beginning a trend of normalization (Safer 2004; Mankoski et al. 2013; Noguera et al. 2013). Indeed in a similar clinical setting involving only SGA-naïve patients, with much lower BMI-Z scores at baseline as compared with the present patients (Wink et al. 2014), BMI-Z has been shown to increase by ∼0.5 per year. That result was nearly similar to the one we observed in our patients treated with aripiprazole, who also had a short previous exposure and a lower BMI-Z at baseline as compared with our patients who used risperidone, with longer previous exposure and higher BMI-Z at baseline. The finding that patients receiving aripiprazole gained more BMI-Z than those treated with risperidone, in the time frame of our follow-up, may be in line with the hypothesis that they did not yet reach a plateau of weight gain, and with the different rapidity of risperidone- and aripiprazole-induced weight gain. These hypotheses may be supported by previous observations from the literature, where clinical trials comparing risperidone and aripiprazole show a similar end-point on weight gain, although the effect of aripiprazole seems to be delayed in time (Safer 2004; Mankoski et al. 2013; Arango et al. 2014). However, as we observed in previous investigations on this patient sample, it is possible that the use of aripiprazole paradoxically worsens the adverse effects profile as compared with risperidone. This observation may be explained by the fact that risperidone is better known for the treatment of pediatric patients, and so are its adverse profile and management strategies; that is, clinicians expect weight gain from risperidone and know how to manage it by tapering or switching drugs. Conversely, Italian clinicians may have less experience in using aripiprazole, especially off-label for irritability, and may be inappropriately expecting a better safety profile, losing track of weight gain. We already reported this possibility for the risk of adverse drug reactions (Rafaniello et al. 2016) and of inefficacy (Pozzi et al. 2016b).

Limitations

The main limitation of this study is its observational nature: results have to be interpreted as a description of what happened in a clinically relevant cohort of pediatric patients treated with SGAs, rather than as an accurate prediction of drug effects. What we observed is the result of both drug treatment and clinical practice. Consistent with this limitation, we did not include drug dose in the analysis as it was repeatedly adjusted, following clinical judgment. Even at the baseline, patients' characteristics determine drug dose, which is an opposite situation as compared with a clinical trial. Current and previous results showed that drug dose is also adjusted to contain the occurrence of adverse drug reactions, including excessive weight gain. Moderate weight gain can instead be tolerated as far as drug treatment yields a satisfactory clinical improvement. Further investigation would be required to assess the impact of weight gain on drug dose reduction, therapy discontinuation, and loss of compliance. Data on adherence were collected during interviews with caregivers, and total compliance was always reported, although drug consumption was not measured because of the observational nature of the study. Another technical limitation of this study is that the data set was assembled from available observations, which were not systematically scheduled or taken in the same number for all patients, limiting study accuracy at detecting changes. The density and number of available observations were also limiting factors for the choice of variables to be investigated: we omitted to consider several potentially relevant confounders; for example, familial and social characteristics, cognitive levels, concomitant lifestyle interventions, and others. In this respect, for future studies, it would be crucial to also consider the potential role of genetic and pharmacogenetic markers in determining different weight-change trajectories (Nurmi et al. 2013), an aspect we have previously investigated in part, finding an association between CA/AA ABCG2 genotypes and a higher occurrence of adverse drug reactions reporting “excessive weight gain” (Rafaniello et al. 2018).

Conclusions

In this study, we observed that (1) children previously treated with risperidone for an average of 4 months displayed a baseline BMI-Z status at the middle of excessive weight and obesity, and during further 8 months of (average) follow-up demonstrated a tendency to further worsening; (2) adolescents previously treated with risperidone for an average of 10 months displayed a baseline BMI-Z status of frank excessive weight, with a tendency of progressive reduction, possibly due to adaptation strategies involving both drug and lifestyle and due to progressive patient selection; and (3) patients previously treated with aripiprazole for an average of 4 months displayed a baseline BMI-Z status of excessive weight and showed a significant continued increase, more prominent in children.

Clinical Significance

This work provides a naturalistic long-term assessment of the weight-gain effects of risperidone and aripiprazole used in pediatric patients in rehabilitation for irritability associated to neurodevelopmental disorders. The use of risperidone and aripiprazole is connected with excessive weight gain in children, which tends to become a stable status of excessive weight in adolescents. Progressive patient selection may contribute to the stabilization of weight gain trends. Drug switch may constitute an additional source of weight gain, as it may prolong an SGA therapy after the previous drug was withdrawn for its adverse effect, as in case of aripiprazole causing more weight gain than risperidone, possibly due to its use in second line. Patients treated with SGAs tend to stabilize mostly within the range of excessive weight to obesity, which may imply severe future health consequences.

Footnotes

Acknowledgments

This work was supported by the Regional Centre of Pharmacovigilance of Lombardy (to E.C.) the Italian Medicines Agency, Agenzia Italiana del Farmaco (AIFA, to E.C.), and by the Italian Ministry of Health (Ricerca Corrente 2018, to M.P., M.N., and M.M.), which are gratefully acknowledged. The funding public institutions had no role in any part of the work.

Disclosures

No competing financial interests exist.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

References

Arango

, Giráldez

, Merchán-Naranjo

, Baeza

, Castro-Fornieles

, Alda

, Martínez-Cantarero

, Moreno

, de Andrés

, Cuerda

, de la Serna

, Correll

, Fraguas

, Parellada

: Second-generation antipsychotic use in children and adolescents: A six-month prospective cohort study in drug-naïve patients. J Am Acad Child Adolesc Psychiatry, 53:1179–1190, 2014.

Baeza

, Vigo

, de la Serna

, Calvo-Escalona

, Merchán-Naranjo

, Rodríguez-Latorre

, Arango

, Castro-Fornieles

: The effects of antipsychotics on weight gain, weight-related hormones and homocysteine in children and adolescents: A 1-year follow-up study. Eur Child Adolesc Psychiatry, 26:35–46, 2017.

Cacciari

, Milani

, Balsamo

, Spada

, Bona

, Cavallo

, Cerutti

, Gargantini

, Greggio

, Tonini

, Cicognani

: Italian cross-sectional growth charts for height, weight and BMI (2 to 20 yr). J Endocrinol Invest, 29:581–593, 2006.

Calarge

, Nicol

, Schlechte

, Burns

: Cardiometabolic outcomes in children and adolescents following discontinuation of long-term Risperidone treatment. J Child Adolesc Psychopharmacol, 24:120–129, 2014.

Calarge

, Nicol

, Xie

, Zimmerman

: Correlates of weight gain during long-term risperidone treatment in children and adolescents. Child Adolesc Psychiatry Ment Health, 6:21, 2012.

Correll

, Manu

, Olshanskiy

, Napolitano

, Kane

, Malhotra

: Cardiometabolic risk of second-generation antipsychotic medications during first-time use in children and adolescents. JAMA, 302:1765–1773, 2009.

De Hert

, Dobbelaere

, Sheridan

, Cohen

, Correll

: Metabolic and endocrine adverse effects of second generation antipsychotics in children and adolescents: A systematic review of randomized, placebo controlled trials and guidelines for clinical practice. Eur Psychiatry, 26:144–158, 2011.

Ebdrup

, Knop

, Madsen

, Mortensen

, Søgaard

, Holst

, Szecsi

, Lublin

: Glucometabolic hormones and cardiovascular risk markers in antipsychotic-treated patients. J Clin Psychiatry, 75:e899–e905, 2014.

Galling

, Roldán

, Nielsen

, Gerhard

, Carbon

, Stubbs

, Vancampfort

, De Hert

, Olfson

, Kahl

, Martin

, Guo

, Lane

, Sung

, Liao

, Arango

, Correll

: Type 2 diabetes mellitus in youth exposed to antipsychotics: A systematic review and meta-analysis. JAMA Psychiatry, 73:247–259, 2016.

10.

Geller

, Luby

, Joshi

, Wagner

, Emslie

, Walkup

, Axelson

, Bolhofner

, Robb

, Wolf

, Riddle

, Birmaher

, Nusrat

, Ryan

, Vitiello

, Tillman

, Lavori

: A randomized controlled trial of Risperidone, lithium, or divalproex sodium for initial treatment of bipolar I disorder, manic or mixed phase, in children and adolescents. Arch Gen Psychiatry, 69:515–528, 2012.

11.

Gueorguieva

, Krystal

: Move over ANOVA: Progress in analyzing repeated-measures data and its reflection in papers published in the archives of general psychiatry. Arch Gen Psychiatry, 61:310–317, 2004.

12.

Guy

: Clinical Global impressions. In: ECDEU Assessment Manual for Psychopharmacology Revised. Rockville, MD: Department of Health, Education, and Welfare; Public Health Service, Alcohol; Drug Abuse, and Mental Health Administration; National Institute of Mental Health; Psychopharmacology Research Branch; Division of Extramural Research Programs, 1976, pp. 218–222.

13.

ICH harmonised tripartite guideline—Clinical investigation of medicinal products in the pediatric population—E11-Current Step 4 version. 2000. Available at: https://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E11/Step4/E11_Guideline.pdf Accessed February 12, 2018 .

14.

Maayan

, Correll

: Weight gain and metabolic risks associated with antipsychotic medications in children and adolescents. J Child Adolesc Psychopharmacol, 21:517–535, 2011.

15.

Mankoski

, Stockton

, Manos

, Marler

, McQuade

, Forbes

, Marcus

: Aripiprazole treatment of irritability associated with autistic disorder and the relationship between prior antipsychotic exposure, adverse events, and weight change. J Child Adolesc Psychopharmacol, 23:572–576, 2013.

16.

Masi

, Milone

, Veltri

, Iuliano

, Pfanner

, Pisano

: Use of quetiapine in children and adolescents. Paediatr Drugs, 17:125–140, 2015.

17.

Noguera

, Ballesta

, Baeza

, Arango

, de la Serna

, González-Pinto

, Parellada

, Graell

, Moreno

, Otero

, Castro-Fornieles

: Twenty-four months of antipsychotic treatment in children and adolescents with first psychotic episode: Discontinuation and tolerability. J Clin Psychopharmacol, 33:463–471, 2013.

18.

Nurmi

, Spilman

, Whelan

, Scahill

, Aman

, McDougle

, Arnold

, Handen

, Johnson

, Sukhodolsky

, Posey

, Lecavalier

, Stigler

, Ritz

, Tierney

, Vitiello

, McCracken

; Research Units on Pediatric Psychopharmacology Autism Network: Moderation of antipsychotic-induced weight gain by energy balance gene variants in the RUPP autism network risperidone studies. Transl Psychiatry, 3:e274, 2013.

19.

Patel

, Chan

, Aparasu

, Ochoa-Perez

, Sherer

, Medhekar

, Chen

: Effect of psychopharmacotherapy on body mass index among children and adolescents with bipolar disorders. J Child Adolesc Psychopharmacol, 27:349–358, 2017.

20.

Pérez-Iglesias

, Martínez-García

, Pardo-Garcia

, Amado

, Garcia-Unzueta

, Tabares-Seisdedos

, Crespo-Facorro

: Course of weight gain and metabolic abnormalities in first treated episode of psychosis: The first year is a critical period for development of cardiovascular risk factors. Int J Neuropsychopharmacol, 17:41–51, 2014.

21.

Pinheiro

, Bates

, DebRoy

S J

, Sarkar

, EISPACK authors, Heisterkamp

, Van Willigen

; R Core Team: NLME: Linear and Nonlinear Mixed Effects Models. 2016. R package version 3.1–128. Available at: http://CRAN.R-project.org/package=nlme Accessed September 27, 2017 .

22.

Pisano

, Catone

, Veltri

, Lanzara

, Pozzi

, Clementi

, Iuliano

, Riccio

, Radice

, Molteni

, Capuano

, Gritti

, Coppola

, Milone

, Bravaccio

, Masi

: Update on the safety of second generation antipsychotics in youths: A call for collaboration among paediatricians and child psychiatrists. Ital J Pediatr, 42:51, 2016.

23.

Pisano

, Gritti

, Catone

, Pascotto

: Antipsychotic-induced dyslipidemia treated with omega 3 fatty acid supplement in an 11-year-old psychotic child: A 1-year follow-up. J Child Adolesc Psychopharmacol, 23:139–141, 2013.

24.

Pozzi

, Cattaneo

, Baldelli

, Fucile

, Capuano

, Bravaccio

, Sportiello

, Bertella

, Auricchio

, Bernardini

, Ferrajolo

, Guastella

, Mani

, Carnovale

, Pisano

, Rafaniello

, Riccio

, Rizzo

, Scuderi

, Sperandeo

, Villa

, Pascotto

, Molteni

, Rossi

, Radice

, Clementi

: Therapeutic drug monitoring of second-generation antipsychotics in pediatric patients: An observational study in real-life settings. Eur J Clin Pharmacol, 72:285–293, 2016a.

25.

Pozzi

, Pisano

, Bertella

, Capuano

, Rizzo

, Antoniazzi

, Auricchio

, Carnovale

, Cattaneo

, Ferrajolo

, Gentili

, Guastella

, Mani

, Rafaniello

, Riccio

, Scuderi

, Sperandeo

, Sportiello

, Villa

, Radice

, Clementi

, Rossi

, Pascotto

, Bernardini

, Molteni

, Bravaccio

: Persistence in therapy with risperidone and aripiprazole in pediatric outpatients: A 2-year naturalistic comparison. J Clin Psychiatry, 77:e1601–e1609, 2016b.

26.

Pringsheim

, Panagiotopoulos

, Davidson

, Ho

; CAMESA Guideline Group: Evidence-based recommendations for monitoring safety of second generation antipsychotics in children and youth. J Can Acad Child Adolesc Psychiatry, 20:218–233, 2011.

27.

R Core Team: R: A language and environment for statistical computing. 2016. R Foundation for Statistical Computing, Vienna, Austria. Available at: https://www.R-project.org Accessed September 27, 2017 .

28.

Rafaniello

, Pozzi

, Pisano

, Ferrajolo

, Bertella

, Sportiello

, Carnovale

, Sullo

, Cattaneo

, Gentili

, Rizzo

, Pascotto

, Mani

, Villa

, Riccio

, Sperandeo

, Bernardini

, Bravaccio

, Clementi

, Molteni

, Rossi

, Radice

, Capuano

: Second generation antipsychotics in “real-life” paediatric patients. Adverse drug reactions and clinical outcomes of drug switch. Expert Opin Drug Saf, 15(Suppl. 2):1–8, 2016.

29.

Rafaniello

, Sessa

, Bernardi

, Pozzi

, Cheli

, Cattaneo

, Baldelli

, Molteni

, Bernardini

, Rossi

, Clementi

, Bravaccio

, Radice

, Capuano

: The predictive value of ABCB1, ABCG2, CYP3A4/5 and CYP2D6 polymorphisms for risperidone and aripiprazole plasma concentrations and the occurrence of adverse drug reactions. Pharmacogenomics J, 18:422–430, 2018.

30.

Raudenbush

, Bryk

: Hierarchical Linear Models: Applications and Data Analysis Methods (2nd ed.). Thousand Oaks (California), Sage Publications, 2002.

31.

Safer

: A comparison of Risperidone-induced weight gain across the age span. J Clin Psychopharmacol, 24:429–436, 2004.

32.

Scahill

, Jeon

, Boorin

, McDougle

, Aman

, Dziura

, McCracken

, Caprio

, Arnold

, Nicol

, Deng

, Challa

, Vitiello

: Weight gain and metabolic consequences of Risperidone in young children with autism spectrum disorder. J Am Acad Child Adolesc Psychiatry, 55:415–423, 2016.

33.

Simon

, van Winkel

, De Hert

: Are weight gain and metabolic side effects of atypical antipsychotics dose dependent?. A literature review. J Clin Psychiatry, 70:1041–1050, 2009.

34.

Wink

, Early

, Schaefer

, Pottenger

, Horn

, McDougle

, Erickson

: Body mass index change in autism spectrum disorders: Comparison of treatment with risperidone and aripiprazole. J Child Adolesc Psychopharmacol, 24:78–82, 2014.

35.

Zhang

, Lencz

, Zhang

, Nitta

, Maayan

, John

, Robinson

, Fleischhacker

, Kahn

, Ophoff

, Kane

, Malhotra

, Correll

: Pharmacogenetic associations of antipsychotic drug-related weight gain: A systematic review and meta-analysis. Schizophr Bull, 42:1418–1437, 2016.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.03 MB