ORIGINAL CONTRIBUTION

Cardiometabolic Risk of Second-Generation

Antipsychotic Medications During First-Time
Use in Children and Adolescents
Christoph U. Correll, MD Context Cardiometabolic effects of second-generation antipsychotic medications are
Peter Manu, MD concerning but have not been sufficiently studied in pediatric and adolescent patients
Vladimir Olshanskiy, MD naive to antipsychotic medication.

Barbara Napolitano, MA Objective To study the association of second-generation antipsychotic medications

with body composition and metabolic parameters in patients without prior antipsy-
John M. Kane, MD chotic medication exposure.
Anil K. Malhotra, MD Design, Setting, and Patients Nonrandomized Second-Generation Antipsy-
chotic Treatment Indications, Effectiveness and Tolerability in Youth (SATIETY) co-

S
ECOND-GENERATION ANTIPSY-
hort study, conducted between December 2001 and September 2007 at semi-urban,
chotic medications are com-
tertiary care, academic inpatient and outpatient clinics in Queens, New York, with a
monly and increasingly pre- catchment area of 4.5-million individuals. Of 505 youth aged 4 to 19 years with 1
scribed to children and week or less of antipsychotic medication exposure, 338 were enrolled (66.9%). Of
adolescents in the United States as first- these patients, 272 had at least 1 postbaseline assessment (80.5%), and 205 patients
line treatment for psychotic disorders, who completed the study (60.7%). Patients had mood spectrum (n=130; 47.8%),
bipolar disorder, and nonpsychotic schizophrenia spectrum (n=82; 30.1%), and disruptive or aggressive behavior spec-
mental disorders.1 Increasingly, the car- trum (n=60; 22.1%) disorders. Fifteen patients who refused participation or were non-
diometabolic effects of second- adherent served as a comparison group.
generation antipsychotic medications Intervention Treatment with aripiprazole, olanzapine, quetiapine, or risperidone for
have raised concern. 2 Cardiometa- 12 weeks.
bolic adverse effects, such as age- Main Outcome Measures Weight gain and changes in lipid and metabolic parameters.
inappropriate weight gain, obesity, hy- Results After a median of 10.8 weeks (interquartile range, 10.5-11.2 weeks) of treat-
pertension, and lipid and glucose ment, weight increased by 8.5 kg (95% confidence interval [CI], 7.4 to 9.7 kg) with
abnormalities, are particularly prob- olanzapine (n=45), by 6.1 kg (95% CI, 4.9 to 7.2 kg) with quetiapine (n=36), by 5.3
lematic during development because kg (95% CI, 4.8 to 5.9 kg) with risperidone (n=135), and by 4.4 kg (95% CI, 3.7 to
they predict adult obesity, the meta- 5.2 kg) with aripiprazole (n=41) compared with the minimal weight change of 0.2 kg
bolic syndrome, cardiovascular mor- (95% CI, −1.0 to 1.4 kg) in the untreated comparison group (n=15). With olanzapine
bidity, and malignancy.3-6 and quetiapine, respectively, mean levels increased significantly for total cholesterol
Emerging findings indicate that (15.6 mg/dL [95% CI, 6.9 to 24.3 mg/dL] P⬍.001 and 9.1 mg/dL [95% CI, 0.4 to
17.7 mg/dL] P=.046), triglycerides (24.3 mg/dL [95% CI, 9.8 to 38.9 mg/dL] P=.002
youth are especially vulnerable to an- and 37.0 mg/dL [95% CI, 10.1 to 63.8 mg/dL] P=.01), non–high-density lipoprotein
tipsychotic medication−induced weight (HDL) cholesterol (16.8 mg/dL [95% CI, 9.3 to 24.3 mg/dL] P⬍.001 and 9.9 mg/dL
gain,7-10 but limited prospective, pedi- [95% CI, 1.4 to 18.4 mg/dL] P=.03), and ratio of triglycerides to HDL cholesterol (0.6
atric data suggest minimal or no meta- [95% CI, 0.2 to 0.9] P=.002 and (1.2 [95% CI, 0.4 to 2.0] P=.004). With risperi-
bolic liabilities, except for olanza- done, triglycerides increased significantly (mean level, 9.7 mg/dL [95% CI, 0.5 to 19.0
pine.9,10 However, the interpretation of mg/dL]; P=.04). Metabolic baseline-to-end-point changes were not significant with
the data is hampered by variable prior aripiprazole or in the untreated comparison group.
antipsychotic medication exposure, Conclusions First-time second-generation antipsychotic medication use was asso-
which can obscure cardiometabolic ef- ciated with significant weight gain with each medication. Metabolic changes varied
fects. Therefore, data are needed in pa- among the 4 antipsychotic medications.
tients with minimal antipsychotic medi- JAMA. 2009;302(16):1765-1773 www.jama.com
cation exposure. Such data are lacking
in youth and are limited to small samples in adults. Furthermore, be- Author Affiliations are listed at the end of this article.
Corresponding Author: Christoph U. Correll, MD,
cause isolated studies in chronic pa- Zucker Hillside Hospital, Psychiatry Research, 75-59
For editorial comment see p 1811.
tients have implicated age11 and anti- 263rd St, Glen Oaks, NY 11004 ([email protected]).

©2009 American Medical Association. All rights reserved. (Reprinted with Corrections) JAMA, October 28, 2009—Vol 302, No. 16 1765

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

psychotic medication dose 1 2 in sessments obtained within 7 days of obesity with BMI ⱖ95th percentile,
cardiometabolic changes, data are antipsychotic medication initiation. blood pressure ⬎90th percentile, level
needed in patients naive to antipsy- Exclusion criteria were treatment of triglycerides ⬎110 mg/dL [to con-
chotic medication. with more than 1 antipsychotic medi- vert to mmol/L, multiply by 0.0113],
To assess the cardiometabolic pro- cation; active or past eating disorder; HDL cholesterol level ⬍40 mg/dL [to
files of the 4 most commonly used biochemical evidence of thyroid dys- convert to mmol/L, multiply by 0.0259],
second-generation antipsychotic function; acute medical disorders; preg- and glucose level ⱖ100 mg/dL [to con-
medications not confounded by car- nancy or breastfeeding; wards of the vert to mmol/L, multiply by 0.0555]).14
ryover effects from prior treatment state (because research consent by a Individuals were assessed after 8 or
with antipsychotic medication, we public agency representative within 1 more hours of overnight fasting at base-
conducted a prospective study of week was unlikely); and leaving the line and weeks 4, 8, and 12. Height was
weight and metabolic changes in a catchment area within 4 weeks. measured 3 times using the Seca 214
large cohort of pediatric patients naive Psychiatric diagnoses and past treat- stadiometer (Seca, Hamburg, Ger-
to antipsychotic medication. We ment history were assessed by chart many). Weight, BMI, and fat mass were
hypothesized that 12 weeks of treat- review, discussion with treatment cli- assessed by impedantiometry with the
ment with aripiprazole, olanzapine, nicians, and clinical interview of the Tanita Body Composition Analyzer
quetiapine, or risperidone would patient or caregiver. Postpubertal sta- TBF-310 (Tanita Corp, Arlington
result in rapid and significant worsen- tus (Tanner stage of 3-5) was deter- Heights, Illinois). Patients were weighed
ing in body composition and meta- mined through inspection and inter- clothed, with emptied pockets and
bolic parameters, and that these view of the patient and/or caregiver. without shoes or socks, using the fol-
would be strongly correlated. Based on the literature in the general lowing subtraction schedule: −1.3 kg for
population,13 we obtained information those taller than 150 cm, wearing long
on race and ethnicity as a potential pre- trousers, and long-sleeve shirts or
METHODS dictor for cardiometabolic outcomes. sweatshirts; −1.1 kg for those wearing
Data were collected as part of the non- Patients received antipsychotic treat- 1 of the 2 items with short sleeves; −0.7
randomized Second-Generation Anti- ment based on the clinician’s choice. In- kg for those wearing short pants or
psychotic Treatment Indications, Ef- formed consent or assent was obtained short-sleeve or light shirts; and −0.5 kg
fectiveness and Tolerability in Youth after the antipsychotic medication choice for those wearing just underwear.
(SATIETY) study, a cohort study of an- was made. Dosing, co-medications, and For individuals measuring less than
tipsychotic medications in pediatric treatment changes were based on clini- 150 cm but 120 cm or more, an addi-
psychotic, mood, or aggressive spec- cal necessity. Although 6 patients naive tional 0.2 kg was subtracted from the
trum disorders. Between December to antipsychotic medication started tak- formula above. For individuals mea-
2001 and September 2007, patients ing ziprasidone, they were excluded from suring less than 120 cm, an additional
were recruited from pediatric inpa- the analyses due to the small sample size. 0.45 kg was subtracted. Waist circum-
tient and outpatient clinics. Primary outcomes were absolute and ference was measured at the level of
Caregivers of all minors aged 4 to 17 relative weight change. Secondary out- both superior iliac crests and umbili-
years and individuals aged 18 to 19 comes included change in additional cus, using the point of largest abdomi-
years signed informed consent. Addi- body composition parameters (body nal circumference. Fasting blood was
tionally, minors aged 9 to 17 years mass index [BMI; calculated as weight drawn between 7 and 11 AM, prior to
signed informed assent. This study was in kilograms divided by height in me- taking morning antipsychotic medica-
approved by the institutional review ters squared], BMI percentiles and z tions. Plasma levels were obtained at
board of the North Shore-Long Island scores, fat mass, and waist circumfer- each postbaseline visit. Families were
Jewish Health System. Data for this re- ence), change in fasting metabolic para- called before the visit and reminded of
port are restricted to youth naive to an- meters (total cholesterol, low-density the overnight fast. At the visit, pa-
tipsychotic medication and a psychi- lipoprotein [LDL] cholesterol, high- tients or their caregivers were asked
atric comparison group consisting of density lipoprotein [HDL] cholesterol, about adherence to fasting. The fast-
patients who refused or discontinued triglycerides, ratio of triglycerides to ing blood work was rescheduled if pa-
taking antipsychotic medications within HDL cholesterol, glucose, insulin, and tients had not fasted, and repeated if the
4 weeks of starting. the homeostasis model assessment of in- glucose level was 100 mg/dL or higher
Inclusion criteria were age of 4 to 19 sulin resistance [HOMA-IR]), and inci- or insulin increased by more than 100%
years and 1 week or less of lifetime an- dence rates of weight gain of 7% or from the last assessment. Glucose and
tipsychotic treatment; psychiatric ill- higher, individual metabolic para- lipid levels were analyzed at the North
ness prompting antipsychotic medica- meters, dyslipidemia, and the meta- Shore University Hospital Core Labo-
tion initiation; and consent, or baseline bolic syndrome (defined by the pres- ratory (Manhasset, New York) with the
anthropometric and biochemical as- ence of ⱖ3 of the following 5 criteria: Roche Hitachi 747 chemistry analyzer
1766 JAMA, October 28, 2009—Vol 302, No. 16 (Reprinted with Corrections) ©2009 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

(Roche Diagnostics, Montclair, New group using mixed-models repeated- ased results due to missing data, mul-
Jersey) and insulin level was analyzed measures analysis of variance in which tiple imputation was applied to the end
via Roche Elecsys 2010 immunochem- the repeated (within subjects) factor point continuous variables and cat-
istry analyzer (Roche Diagnostics). was time relative to baseline at 4, 8, and egorical outcomes. These results did not
Plasma levels were measured with liq- 12 weeks. Summary statistics for mixed- differ appreciably from the analyses per-
uid chromatography at the Cooper models repeated-measures analysis of formed without multiple imputation.
Laboratory (Nathan Kline Institute, Or- variance are expressed as adjusted least- Therefore, we conducted the analyses
angeburg, New York). squares means and 95% confidence in- without multiple imputation. Analy-
Patients with 1 or more postbase- tervals (CIs). The incidence rates for di- ses were repeated in patients with and
line assessments comprised the intent- chotomous outcomes were analyzed without co-medications known to affect
to-treat sample. Sex- and age-adjusted using last observation carried for- weight (weight neutral: benzodiaz-
BMI z scores were calculated using a ward. The Pearson ␹2 test was used to epines, anticholinergics, ␣-agonists,
Web-based calculator (http://www compare categorical outcomes across escitalopram and citalopram, fluvox-
.kidsnutrition.org/bodycomp/bmiz2 antipsychotic medications, with corre- amine, sertraline, venlafaxine).
.html). Insulin resistance was deter- sponding baseline values as fixed co- For exploratory analyses of the effect
mined with HOMA-IR (fasting insulin variates, controlling for significantly dif- of patients’ age on changes in body com-
µmol ⫻ glucose mmol/22.5).15 The ferent baseline variables. position and metabolic parameters, pa-
HOMA-IR values higher than 4.39 were Given the large body weight changes, tients were dichotomized into postpu-
diagnostic for insulin resistance.16 post hoc analyses were performed for bertal status (n = 191; mean age, 15.8
Baseline values were compared across the prespecified categorical change in years [95% CI, 15.5-16.1 years]) vs pre-
groups with the ␹2 and Fisher exact tests weight (ⱖ14% and ⱖ21%) and BMI z pubertal or peripubertal status (n=81;
for categorical variables and the score (ⱖ1.0). To confirm that mixed- mean age, 9.5 years [95% CI, 8.9-10.0
Kruskal-Wallis test for continuous vari- models repeated-measures analysis of years]). For the exploration of a dose
ables. Change in continuous variables variance and last-observation-carried- effect, we dichotomized the data using
was analyzed within each treatment forward analyses were not yielding bi- a median split of the maximum (in most

Figure. Flow of Patients Through Study

853 Pediatric patients started

taking an atypical antipsychotic

348 Excluded because of lifetime

antipsychotic exposure >7 d

505 Antipsychotic naive

167 Excluded
76 Declined
44 Unable to contact or enroll
within 7 d
18 Had an eating disorder
18 Ward of the state
7 Medical reasons
4 Expected transfer out of
catchment area in <4 wk

338 Enrolled

6 Excluded after starting ziprasidone

47 Received aripiprazole 52 Received olanzapine 45 Received quetiapine 168 Received risperidone 20 Served as comparison group

33 Completed 12 wk 36 Completed 12 wk 27 Completed 12 wk 96 Completed 12 wk 13 Completed 12 wk

14 Discontinued aripiprazole 16 Discontinued olanzapine 18 Discontinued quetiapine 72 Discontinued risperidone 6 Lost to follow-up
4 Lack of efficacy 1 Lack of efficacy 5 Lack of efficacy 9 Lack of efficacy 1 Restarted antipsychotic
3 Lack of efficacy and 1 Lack of efficacy and 2 Lack of efficacy and 3 Lack of efficacy and
intolerability intolerability intolerability intolerability
2 Intolerability 4 Intolerability 4 Intolerability 20 Intolerability
2 Nonadherence 4 Nonadherence 5 Nonadherence 16 Nonadherence
3 Other reasons 6 Other reasons 2 Other reasons 24 Other reasons

41 Included in analysis 45 Included in analysis 36 Included in analysis 135 Included in analysis 15 Included in analysis
6 Excluded (had baseline 7 Excluded (had baseline 9 Excluded (had baseline 33 Excluded (had baseline 5 Excluded (had baseline
visit only) visit only) visit only) visit only) visit only)

©2009 American Medical Association. All rights reserved. (Reprinted with Corrections) JAMA, October 28, 2009—Vol 302, No. 16 1767

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

cases final) antipsychotic medication (mean exposure, 12.4 days; 95% CI, ing aripiprazole, olanzapine, or queti-
dose (aripiprazole = 10 mg/d; olanza- 10.8-14.0 days) but had 8- or 12-week apine. With risperidone, doses greater
pine=10 mg/d; quetiapine = 275 mg/d; assessments. than 1.5 mg/d were associated with sig-
risperidone=1.5 mg/d). Analyses were After a median of 10.8 weeks (inter- nificantly greater increases in weight,
2-sided with an ␣ level of less than .05 quartile range, 10.5-11.2 weeks) of waist circumference, fat mass, and BMI
and were conducted using SAS statis- treatment, weight increased by 8.5 kg z score. The metabolic effects of arip-
tical software version 9.1 (SAS Insti- (95% CI, 7.4 to 9.7 kg) with olanza- iprazole or quetiapine did not differ be-
tute Inc, Cary, North Carolina). pine (n=45), by 6.1 kg (95% CI, 4.9 to tween dosage groups. Conversely, pa-
For this observational cohort study, 7.2 kg) with quetiapine (n=36), by 5.3 tients treated with doses of greater than
we conducted a generic power analy- kg (95% CI, 4.8 to 5.9 kg) with risperi- 10 mg/d of olanzapine and patients
sis for a mean change from baseline to done (n =135), and by 4.4 kg (95% CI, treated with greater than 1.5 mg/d of
12 weeks per 1 standard deviation using 3.7 to 5.2 kg) with aripiprazole (n=41) risperidone experienced significantly
a paired t test. Except for the compari- compared with minimal weight change greater increases in total cholesterol and
son group in which only a large effect of 0.2 kg (95% CI, −1.0 to 1.4 kg) in non-HDL cholesterol.
size of 0.78 could be detected, we had the untreated comparison group
80% power to show significant differ- (n = 15) (TABLE 2). The proportions of
ences corresponding to a moderate, patients gaining 7% or greater weight COMMENT
clinically meaningful effect size of 0.43 are presented in eTable 3. In this short-term study of youth naive
for olanzapine, 0.45 for aripiprazole, Each antipsychotic medication was to antipsychotic medications, aripipra-
and 0.48 for quetiapine, and a small associated with significantly in- zole, olanzapine, quetiapine, and ris-
effect size of 0.24 for risperidone. creased fat mass and waist circumfer- peridone were each associated with
ence (Table 2; eTable 2 contains data rapid and significant increases in body
RESULTS for weeks 0-4 and 0-8) and shifts to composition, whereas metabolic
Of 505 pediatric patients naive to an- overweight (ⱖ85th-⬍95th BMI per- changes were less uniform. Effect sizes
tipsychotic medication, 338 were en- centile) or obese (ⱖ95th BMI percen- for body composition changes were
rolled (66.9%). Six patients initially tile) status. Using increases of 14% or large (eTable 2). Altogether, 10% to
given ziprasidone were excluded and greater and 21% or greater of unad- 36% of patients transitioned to over-
60 (17.9%) did not undergo postbase- justed body weight gain and BMI z score weight or obese status within 11
line assessment, yielding 272 (81.0%) standard deviations of 0.5 or greater and weeks. The lack of significant changes
analyzed patients with confirmed an- 1.0 or greater as the pathological thresh- in weight and metabolic parameters in
tipsychotic medication adherence old, the same ranking order emerged psychiatric comparison patients and
(FIGURE). The 173 individuals who re- (eTable 3). short inpatient stays (10-18 days is
fused to participate in the study or who Adverse baseline-to-end-point equal to 14%-25% of treatment time)
were ineligible were not different from changes reached statistical signifi- indicates that the observed alterations
consenting patients except for having cance for olanzapine and quetiapine for are not likely due to a newly develop-
less autism-spectrum disorders (1.9% total cholesterol, triglycerides, non- ing or worsening psychiatric disorder
vs 8.1%; P=.009), substance abuse co- HDL cholesterol, and ratio of triglyc- or hospitalization. The results are con-
morbidity (8.4% vs 16.5%; P=.02), and erides to HDL cholesterol (TABLE 3; cerning because they include fat mass
mixed ethnicity (3.7% vs 12.5%; eTable 4 contains data for weeks 0-4 and waist circumference, which are
P = .002) in the excluded group (in and 0-8). With risperidone, levels of tri- associated with the metabolic syn-
whom substance abuse and ethnicity glycerides increased significantly. Meta- drome17 in adults treated with antipsy-
were assessed solely via chart review bolic baseline-to-end-point changes chotic medications and heart disease
compared with a formal interview in the were not significant with aripiprazole in the general population.18 Moreover,
included patients). There were no sig- or in the untreated comparison group. abnormal childhood weight and
nificant differences in any variable in- Patients receiving quetiapine had metabolic status adversely affect adult
cluded in TABLE 1 (eTable 1 contains modestly higher incidence rates of hy- cardiovascular outcomes3-6 via con-
data on fasting metabolic characteris- perglycemia and the metabolic syn- tinuation of these risk factors 19 or
tics and treatment characteristics drome and patients receiving olanza- independent or accelerated mecha-
and is available at http://www.jama pine experienced the highest incidence nisms.20
.com) between the 272 analyzed pa- rates (eTable 3). Pubertal status was un- It has been argued that youth are
tients and the 60 youth without post- related to metabolic changes in any an- more vulnerable than adults to weight
baseline assessments. The compari- tipsychotic medication group. gain induced by antipsychotic medica-
son group was composed of 15 patients Antipsychotic medication dose was tions. A comparison of our findings
who refused or stopped taking an an- not associated with body composition with prior studies does not support this.
tipsychotic medication within 4 weeks parameter changes in patients receiv- Rather, it appears that the greater weight
1768 JAMA, October 28, 2009—Vol 302, No. 16 (Reprinted with Corrections) ©2009 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

Table 1. Baseline Demographic and Clinical Characteristics a

Comparison
Total Aripiprazole Olanzapine Quetiapine Risperidone Group P
(N = 272) (n = 41) (n = 45) (n = 36) (n = 135) (n = 15) Value
Demographic characteristics, No. (%)
Age, mean (SD) [range], y 13.9 (3.6) 13.4 (3.1) 14.7 (3.2) 14.0 (3.1) 13.6 (4.0) 15.5 (2.0) .13
[4.3-19.9] [7.0-19.7] [6.6-18.6] [6.1-19.4] [4.3-19.9] [12.2-18.1]
Postpubertal status 191 (70.2) 25 (61.0) 38 (84.4) 26 (72.2) 88 (65.2) 14 (93.3) .03
Male sex 155 (57.0) 23 (56.1) 29 (64.4) 13 (36.1) 84 (62.2) 6 (40.0) .03
Race/ethnicity
White 131 (48.5) 25 (62.5) 21 (46.7) 18 (50.0) 62 (46.3) 5 (33.3)
Black 70 (25.9) 8 (20.0) 13 (28.9) 7 (19.4) 38 (28.4) 4 (26.7)
Hispanic 24 (8.9) 3 (7.5) 1 (2.2) 3 (8.3) 15 (11.2) 2 (13.3) .32
Asian 11 (4.1) 2 (5.0) 2 (4.4) 4 (11.1) 2 (1.5) 1 (6.7)
Mixed 34 (12.5) 2 (5.0) 8 (17.8) 4 (11.1) 17 (12.7) 3 (18.7)
Socioeconomic status, mean (95% CI) b 2.8 2.6 2.8 2.7 2.8 2.8 .91
(2.6 to 2.9) (2.3 to 3.0) (2.5 to 3.2) (2.4 to 3.1) (2.6 to 3.0) (1.7 to 3.9)
Family history in first-degree relative, No. (%)
Obesity 107 (44.4) 19 (57.6) 25 (58.1) 10 (32.3) 71 (58.7) 9 (69.2) .08
Diabetes 38 (16.0) 5 (16.1) 6 (14.3) 5 (16.1) 19 (15.7) 3 (25.0) .94
Dyslipidemia 83 (35.5) 10 (32.3) 15 (35.7) 14 (45.2) 42 (35.9) 2 (15.4) .45
Hypertension 70 (29.5) 9 (28.1) 9 (20.9) 8 (26.7) 39 (33.3) 5 (33.3) .63
Primary psychiatric diagnosis, No. (%) c
Schizophrenia spectrum disorder 82 (30.1) 14 (31.1) 14 (35.6) 6 (16.7) 46 (34.1) 1 (6.7) .07
Psychosis NOS 54 (19.9) 11 (26.8) 5 (11.1) 4 (11.1) 33 (24.4) 1 (6.7) .08
Schizophrenia or schizoaffective disorder 27 (9.9) 3 (7.3) 9 (20.0) 2 (5.6) 13 (9.6) 0 .10
Mood disorder spectrum disorder 130 (47.8) 11 (27.5) 16 (35.6) 9 (25.0) 55 (40.7) 5 (33.3) .006
MDD or depressive disorder NOS 49 (18.0) 10 (24.4) 8 (17.8) 8 (22.2) 19 (14.1) 4 (27.7) .44
Bipolar I and II disorder, bipolar disorder NOS 44 (16.2) 3 (7.3) 9 (20.0) 10 (27.8) 17 (12.6) 5 (33.3) .03
Mood disorder NOS 37 (13.6) 5 (12.2) 4 (8.9) 6 (16.7) 19 (14.1) 3 (20.0) .78
Disruptive or aggressive behavior 60 (22.1) 9 (22.2) 9 (20.0) 6 (16.7) 34 (25.2) 2 (13.3) .71
spectrum disorder
ODD, CD, IED, or ICD 39 (14.3) 5 (12.2) 7 (15.6) 4 (11.1) 21 (15.6) 0 .96
Autism spectrum disorder 21 (7.7) 4 (9.8) 2 (4.4) 2 (5.6) 13 (9.6) 2 (13.3) .79
CGIS score, mean (95% CI) 5.4 5.3 5.7 5.3 5.4 5.1 .14
(5.3 to 5.5) (5.1 to 5.5) (5.4 to 6.0) (5.0 to 5.6) (5.3 to 5.6) (4.6 to 5.6)
GAF score, mean (95% CI) 36.4 38.5 33.1 37.8 36.5 37.6 .02
(35.5 to 37.4) (36.5 to 40.6) (30.3 to 35.9) (35.3 to 40.3) (35.1 to 37.9) (33.3 to 42.0)
Body composition, mean (95% CI)
Weight, kg 53.5 55.0 54.0 59.6 50.8 58.5 .10
(51.5 to 56.2) (50.2 to 63.5) (48.1 to 59.8) (52.3 to 67.0) (47.6 to 54.0) (51.7 to 65.3)
Fat mass, kg 12.5 14.5 10.7 17.4 11.1 13.1 .004
(11.5 to 14.0) (11.6 to 20.5) (8.1 to 13.3) (13.2 to 21.6) (9.5 to 12.6) (8.9 to 17.2)
Waist circumference, cm 77.1 80.7 75.5 83.1 74.7 79.7 .01
(75.5 to 79.1) (76.6 to 87.3) (71.4 to 79.6) (77.5 to 88.7) (72.3 to 77.1) (73.6 to 85.8)
BMI d 21.3 22.4 20.4 23.3 20.6 22.1 .02
(20.7 to 22.1) (21.0 to 25.2) (19.1 to 21.7) (21.2 to 25.4) (19.7 to 21.4) (20.2 to 24.1)
BMI z score 0.24 0.66 −0.08 0.71 0.09 0.33 .01
(0.09 to 0.41) (0.29 to 1.14) (−0.48 to 0.32) (0.29 to 1.14) (−0.13 to 0.31) (−0.32 to 0.98)
BMI percentile 57.4 68.6 49.9 67.7 53.6 60.0 .01
(53.7 to 61.6) (59.2 to 78.9) (40.0 to 60.0) (56.8 to 78.6) (48.1 to 59.2) (41.9 to 78.1)
Weight status, No. (%) e
Underweight (⬍5th) 21 (7.7) 2 (4.9) 6 (13.3) 2 (5.6) 10 (7.4) 1 (6.7) .60
Normal (ⱖ5th-⬍85th) 168 (61.7) 20 (48.8) 30 (66.7) 18 (50.0) 90 (66.7) 10 (66.7) .14
Overweight (ⱖ85th-⬍95th) 37 (13.6) 6 (14.6) 5 (11.1) 6 (16.7) 19 (14.1) 1 (6.7) .89
Obese (ⱖ95th) 46 (16.9) 13 (31.7) 4 (8.9) 10 (27.8) 16 (11.8) 3 (20.0) .007
Abbreviations: BMI, body mass index; CD, conduct disorder; CGIS, Clinical Global Impression Severity; CI, confidence interval; GAF, Global Assessment of Functioning; ICD, im-
pulse control disorder; IED, intermittent explosive disorder; MDD, major depressive disorder; NOS, not otherwise specified; ODD, oppositional defiant disorder.
a For some adopted patients, reliable information regarding ethnic or racial origin and family history could not be obtained so the denominators for these categories are different.
b Based on Hollingshead scale, which has a range of 1 to 5.
c Assessed with the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition).
d Calculated as weight in kilograms divided by height in meters squared.
e By BMI percentile category.

©2009 American Medical Association. All rights reserved. (Reprinted with Corrections) JAMA, October 28, 2009—Vol 302, No. 16 1769

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

gain in youth treated with antipsy- exposure compared with most adult greater than in adults with chronic
chotic medication is related to less samples. As in previous pediatric stud- schizophrenia.21 It was also greater than
frequent antipsychotic medication ies,7-9 the weight gain in our study was in adults with first-episode schizophre-
nia (24% naive to antipsychotic
medication),22 in which weight gain of
Table 2. Change in Body Composition Parameters Over Time 7% or greater was similar only after 1
Weeks 0-12 year of treatment. Our observed weight
Outcome Variable Mean (95% CI) P Value gain was also considerably greater
Weight, kg compared with recent, short-term,
Aripiprazole 4.44 (3.71 to 5.18) ⬍.001 placebo-controlled trials in pediatric
Olanzapine 8.54 (7.38 to 9.69) ⬍.001 schizophrenia and bipolar disorder10
Quetiapine 6.06 (4.90 to 7.21) ⬍.001 (mean absolute weight gain and pro-
Risperidone 5.34 (4.81 to 5.87) ⬍.001 portion of patients gaining ⱖ7% of
Untreated 0.19 (−1.04 to 1.43) .77 weight with aripiprazole: 0-0.9 kg and
Weight % of baseline 4.0%-12.3%; quetiapine: 1.7 kg and
Aripiprazole 8.14 (6.97 to 9.49) ⬍.001
9.9%-14.5%; risperidone: 1.4-1.9 kg
Olanzapine 15.20 (13.24 to 17.16) ⬍.001
and 15%-16%; olanzapine: 3.7-4.7 kg
Quetiapine 10.42 (8.54 to 12.30) ⬍.001
and 41.9%-45.8%). The weight gain was
Risperidone 10.37 (9.41 to 11.33) ⬍.001
also greater than in pediatric studies
Untreated 0.65 (−1.33 to 2.63) .53
Fat mass, kg
comparing olanzapine and risperi-
Aripiprazole 2.43 (1.81 to 3.04) ⬍.001 done with only 36%8 and 33%,9 respec-
Olanzapine 4.12 (3.29 to 4.95) ⬍.001 tively, youth naive to antipsychotic
Quetiapine 2.82 (2.01 to 3.63) ⬍.001 medication. The gains in BMI z score
Risperidone 2.45 (2.07 to 2.84) ⬍.001 that adjust for baseline sample differ-
Untreated 0.35 (−0.43 to 1.12) .39 ences were more than double com-
BMI a pared with the 8-week Treatment of
Aripiprazole 1.67 (1.36 to 1.98) ⬍.001 Early-Onset Schizophrenia Spectrum
Olanzapine 3.01 (2.60 to 3.42) ⬍.001 Disorder (TEOSS) study9 (olanzapine:
Quetiapine 2.12 (1.71 to 2.53) ⬍.001
0.93 vs 0.39; risperidone: 0.60 vs 0.23).
Risperidone 1.92 (1.72 to 2.12) ⬍.001
By contrast, our absolute and relative
Untreated −0.003 (−0.45 to 0.45) .99
weight findings (especially important for
BMI % of baseline
Aripiprazole 7.20 (5.83 to 8.57) ⬍.001 a comparison with adults who gener-
Olanzapine 14.04 (12.08 to 16.00) ⬍.001 ally have higher baseline weights) are
Quetiapine 9.29 (7.43 to 11.15) ⬍.001 similar to a 3-month adolescent queti-
Risperidone 9.12 (8.18 to 10.06) ⬍.001 apine study (77% naive to antipsy-
Untreated 0.05 (−1.89 to 1.99) .96 chotic medication)23 and a 3-month24 and
BMI z score 4-month25 first-episode adult schizo-
Aripiprazole 0.37 (0.25 to 0.49) ⬍.001 phrenia study in which 100% were na-
Olanzapine 0.93 (0.75 to 1.11) ⬍.001 ive to antipsychotic medication24 or 91%
Quetiapine 0.44 (0.28 to 0.60) ⬍.001 had 7 or fewer days of antipsychotic
Risperidone 0.60 (0.52 to 0.68) ⬍.001 medication exposure.23 This weight gain
Untreated −0.003 (−0.12 to 0.11) .96 similarity (despite ⬎10-year higher age)
BMI percentile suggests that prior treatment may be
Aripiprazole 8.44 (5.26 to 11.62) ⬍.001
more relevant than age and developmen-
Olanzapine 24.22 (19.32 to 29.12) ⬍.001
Quetiapine 10.86 (7.35 to 14.37) ⬍.001
tal differences.
Risperidone 16.42 (13.64 to 19.20) ⬍.001
Despite significant body composi-
Untreated 0.69 (−2.58 to 3.96) .69
tion changes with each antipsychotic
Waist, cm
medication, metabolic risk profiles var-
Aripiprazole 5.40 (2.87 to 7.93) .001 ied, lipid abnormalities predominated
Olanzapine 8.55 (7.43 to 9.67) ⬍.001 over glucose abnormalities after short-
Quetiapine 5.27 (4.07 to 6.47) ⬍.001 term exposure, and the metabolic syn-
Risperidone 5.10 (4.49 to 5.71) ⬍.001 drome and diabetes developed rarely.
Untreated 0.70 (−0.87 to 2.27) .40 Olanzapine had the largest weight ef-
Abbreviations: BMI, body mass index; CI, confidence interval.
a Calculated as weight in kilograms divided by height in meters squared.
fects and also significantly worsened all
glucose and lipid parameters, except
1770 JAMA, October 28, 2009—Vol 302, No. 16 (Reprinted with Corrections) ©2009 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

HDL cholesterol, which is more re-

Table 3. Change in Metabolic Parameters Over Time
lated to physical activity.26
Weeks 0-12
Quetiapine and risperidone signifi-
cantly increased triglycerides, but did Outcome Variable Mean (95% CI) P Value
not produce significant abnormalities Glucose, mg/dL
Aripiprazole 0.54 (−2.85 to 3.93) .76
in glucose homeostasis. Despite simi-
Olanzapine 3.14 (0.69 to 5.59) .02
lar body composition changes com- Quetiapine 2.64 (−0.65 to 5.93) .12
pared with risperidone, quetiapine was Risperidone 1.14 (−0.84 to 3.12) .26
additionally associated with signifi- Untreated 0.69 (−4.84 to 6.22) .81
cantly increased total cholesterol, non- Insulin, µIU/mL
HDL cholesterol, and ratio of triglyc- Aripiprazole 2.61 (−2.09 to 7.31) .28
erides to HDL cholesterol, indicating Olanzapine 2.71 (0.42 to 5.00) .02
broader metabolic effects, as sug- Quetiapine 1.08 (−2.80 to 4.96) .59
Risperidone 0.69 (−0.86 to 2.24) .39
gested recently in youth27 and adults.28,29
Untreated −0.47 (−4.31 to 3.37) .81
The TEOSS trial 9 reported signifi-
HOMA-IR
cantly increased levels of total and LDL Aripiprazole 0.55 (−0.49 to 1.59) .31
cholesterol only with olanzapine and no Olanzapine 0.62 (0.07 to 1.17) .03
triglyceride signal with olanzapine and Quetiapine 0.35 (−0.57 to 1.27) .46
risperidone, a difference possibly due Risperidone 0.20 (−0.15 to 0.55) .28
to carryover effects from prior treat- Untreated −0.09 (−0.85 to 0.67) .82
ment or fasting sample size limita- Ratio of triglycerides to HDL cholesterol, mg/dL
Aripiprazole −0.19 (−0.66 to 0.28) .42
tions. Similar reasons may account for Olanzapine 0.59 (0.24 to 0.94) .002
the lack of a metabolic signal in large- Quetiapine 1.22 (0.44 to 2.00) .004
scale, pediatric second-generation an- Risperidone 0.20 (0 to 0.40) .05
tipsychotic medication registration Untreated −0.31 (−0.94 to 0.32) .35
trials—except for olanzapine.10 Total cholesterol, mg/dL
Despite significant worsening in all Aripiprazole 3.75 (−3.85 to 11.35) .34
body composition parameters, arip- Olanzapine 15.58 (6.88 to 24.28) ⬍.001
Quetiapine 9.05 (0.41 to 17.69) .046
iprazole was not associated with sig-
Risperidone 3.46 (−1.44 to 8.36) .17
nificantly worsened metabolic indices
Untreated 2.38 (−7.69 to 12.45) .82
(except for an isolated, near signifi- LDL cholesterol, mg/dL
cant increase in LDL cholesterol level). Aripiprazole 7.38 (0.77 to 13.99) .05
Reasons for this apparent dissociation Olanzapine 11.54 (3.97 to 19.11) .004
are unclear, but could be related to a Quetiapine 3.88 (−3.37 to 11.13) .30
lower effect size that was greater than Risperidone 0.21 (−4.14 to 4.56) .92
50% for increased waist circumfer- Untreated 2.99 (−5.18 to 11.16) .49
ence compared with quetiapine and ris- HDL cholesterol, mg/dL
Aripiprazole 0.29 (−2.32 to 2.90) .83
peridone despite similar effect sizes for Olanzapine −1.27 (−3.80 to 1.26) .33
all other body composition para- Quetiapine −1.47 (−5.06 to 2.12) .43
meters (eTable 2). However, due to the Risperidone 0.33 (−1.26 to 1.92) .68
relatively small aripiprazole sample, we Untreated 1.49 (−3.10 to 6.08) .53
cannot exclude a type II error for lipid Triglycerides, mg/dL
parameters (effect sizes, 0.15-0.35; Aripiprazole −2.40 (−19.71 to 14.91) .79
Olanzapine 24.34 (9.80 to 38.88) .002
eTable 4), which is not likely for tri-
Quetiapine 36.96 (10.13 to 63.79) .01
glycerides and ratio of triglycerides to
Risperidone 9.74 (0.45 to 19.03) .04
HDL cholesterol that decreased and Untreated −11.84 (−41.55 to 17.87) .45
HDL cholesterol that increased. Non-HDL cholesterol, mg/dL
The same caveat applies to the non- Aripiprazole 4.41 (−2.10 to 10.92) .19
significant glucose homeostasis changes Olanzapine 16.81 (9.30 to 24.32) ⬍.001
with aripiprazole, quetiapine, and ris- Quetiapine 9.93 (1.42 to 18.44) .03
peridone (effect sizes, 0.05-0.26; Risperidone 3.02 (−1.45 to 7.49) .19
eTable 4). However, our findings of less Untreated 0.52 (−8.65 to 9.69) .91
Abbreviations: CI, confidence interval; HOMA-IR, homeostasis model assessment of insulin resistance.
lipid abnormalities with aripiprazole are SI conversion factors: To convert HDL, LDL, and total cholesterol to mmol/L, multiply values by 0.0259; glucose to
supported by early, short-term metabo- mmol/L, multiply values by 0.0555; insulin to pmol/L, multiply values by 6.945; triglycerides to mmol/L, multiply val-
ues by 0.0113.
lomic studies.30 In view of a significant
©2009 American Medical Association. All rights reserved. (Reprinted with Corrections) JAMA, October 28, 2009—Vol 302, No. 16 1771

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

association between a stable BMI and position changes were dose related only Author Affiliations: Zucker Hillside Hospital, Psychia-
try Research, North Shore-Long Island Jewish Health
metabolic health in young adults from with risperidone supports weight- System, Glen Oaks, New York (Drs Correll, Manu, Ol-
the general population followed up for independent metabolic effects with shanskiy, Kane, and Malhotra, and Ms Napolitano);
Albert Einstein College of Medicine, Bronx, New York
15 years31 and of significant weight gain olanzapine.33 However, fixed-dose, ran- (Drs Correll, Manu, Kane, and Malhotra); Feinstein In-
with all studied antipsychotic medica- domized studies and blood level assess- stitute for Medical Research, Manhasset, New York
(Drs Correll, Kane, Malhotra, and Ms Napolitano); and
tions in our study, longer-term assess- ments are needed to further examine North Shore University Hospital, Manhasset, New York
ments are needed to clarify the trajec- antipsychotic medication−dose rela- (Ms Napolitano).
tory of metabolic changes with specific tionships. Author Contributions: Dr Correll had full access to all
of the data in the study and takes responsibility for
antipsychotic medications. Such stud- The results from this study need to be the integrity of the data and the accuracy of the data
ies should evaluate the importance of interpreted within its limitations, which analysis. Drs Kane and Malhotra contributed equally
to the article.
weight change vs end point BMI for include the nonrandomized, observa- Study concept and design: Correll, Kane, Malhotra.
metabolic abnormalities because emerg- tional design, baseline differences pre- Acquisition of data: Correll, Olshanskiy.
ing data suggest a potentially greater im- cluding rigorous group comparisons, Analysis and interpretation of data: Correll, Manu,
Napolitano, Kane, Malhotra.
portance of the latter.32 flexible dosing, allowance of co- Drafting of the manuscript: Correll, Manu, Napolitano,
More research is also needed to de- medications, relatively short treatment Kane, Malhotra.
Critical revision of the manuscript for important in-
termine the time course and magnitude duration, and a small comparison group. tellectual content: Correll, Manu, Olshanskiy, Kane,
of developing diabetes or the metabolic Moreover, we did not include a first- Malhotra.
syndrome and to uncover the mecha- generation antipsychotic medication Statistical analysis: Correll, Napolitano.
Obtained funding: Correll, Kane, Malhotra.
nisms underlying the apparent delay in comparator. In the TEOSS study,9 mo- Administrative, technical or material support: Manu,
acquiring the metabolic syndrome and lindone was found to be weight neu- Olshanskiy, Kane, Malhotra.
insulin resistance with rapid weight gain tral, but many patients lost weight, sug- Study supervision: Kane.
Finanical Disclosures: Dr Correll reported being a con-
during childhood. This phenomenon, gesting prior treatment effects. sultant or receiving honoraria from AstraZeneca, Bristol-
also suggested in the general pediatric Despite these caveats, this is the larg- Myers Squibb, Cephalon, Eli Lilly, Intra-Cellular Thera-
peutics, Medicure, OrthoMcNeill-Janssen, Otsuka,
population,6 seems to exclude olanza- est study focusing on changes in weight Organon, Pfizer, Schering-Plough, Solvay, Supernus,
pine. Reasons for this could be the mag- and metabolic parameters in pediatric Vanda, and Wyeth, and serving on the speaker’s bu-
reau of AstraZeneca, Bristol-Myers Squibb/Otsuka, and
nitude of body composition changes or patients naive to antipsychotic medi- Pfizer. Dr Manu reported being a consultant or re-
weight-independent effects.33 Of note, tri- cation, using strictly reinforced fast- ceiving honoraria from Bristol-Meyers Squibb and
glycerides and the ratio of triglycerides ing assessments and verifying medica- Pfizer, and serving on the speaker’s bureau of Bristol-
Myers Squibb/Otsuka and Pfizer. Dr Kane reported
to HDL cholesterol, which are sug- tion adherence via interview and blood being a consultant to or receiving honoraria from Ab-
gested markers in adults,34 seem to be levels. This design enabled us to en- bott, Astra-Zeneca, Bristol-Myers Squibb, Cephalon,
Dainippon Sumitomo, Eli Lilly, Intra-Cellular Thera-
more sensitive than glucose and insulin roll a fairly large group of patients na- peutics, Janssen Pharmaceutica, Johnson & Johnson,
for the early identification of worsening ive to antipsychotic medication and Lundbeck, NuPathe, Otsuka, Pfizer Inc, PgXHealth,
Proteus, Schering, Shire, Solvay, Vanda, and Wyeth,
insulin resistance. Triglyceride changes treated under real-life conditions, emit- serving on the speaker’s bureau of AstraZeneca, Bristol-
reflect early insulin resistance at the ting a larger signal for body composi- Myers Squibb/Otsuka, and Eli Lilly, and being a share-
holder of MedAvante. Dr Malhotra reported being a
muscle cell level, while changes at the he- tion and, especially, metabolic abnor- consultant to or receiving honoraria from Bristol-
patic level seem to occur later, giving rise malities compared with prior studies. Myers Squibb, Otsuka, Pfizer, and Vanda, and serv-
to delayed glucose, insulin, and Our results, together with data from ing on the speaker’s bureau of Bristol-Myers Squibb/
Otsuka and Pfizer. No other authors reported financial
HOMA-IR signals.35 first-episode studies, suggest that guide- disclosures.
Not surprisingly, some absolute body lines for antipsychotic medication expo- Funding/Support: Supported in parts by grant
MH01760 to Dr Malhotra from the National Insti-
composition changes were greater in sure for vulnerable pediatric and ado- tutes of Health, a National Alliance for Research in
postpubertal patients who also were lescent patients naive to antipsychotic Schizophrenia and Depression Independent Investi-
gator Award to Dr Malhotra, grant MH 074543-01
heavier at baseline. However, the lack medication should consider more fre- to Dr Kane for the Zucker Hillside Hospital National
of a moderating effect of pubertal sta- quent (eg, biannual 36 ) cardiometa- Institute of Mental Health Advanced Center for In-
tus on age- and sex-adjusted BMI z bolic monitoring after the first 3 months tervention and Services Research for the Study of
Schizophrenia, and by funding from the Feinstein In-
scores and any metabolic parameter in- of treatment.2 Finally, in view of poor stitute for Medical Research North Shore-Long Is-
dicates that the same caution is re- physical health outcomes37 and subop- land Jewish Health System General Clinical Research
Center, and grant M01 RR018535 from the National
quired when treating younger chil- timal metabolic monitoring38 in the Center for Research Resources, which is a compo-
dren and adolescents. Our data support severely mentally ill, the benefits of sec- nent of the National Institutes of Health.
Role of the Sponsors: None of these noncommercial
recent findings that higher doses of ond-generation antipsychotic medica- funding organizations had any role in the design and
olanzapine (⬎10 mg/d) are associated tions must be balanced against their car- conduct of the study; collection, management, analy-
with greater metabolic abnormali- diometabolic risks through a careful sis, and interpretation of the data; and preparation,
review, or approval of the article.
ties.12 While data for risperidone were assessment of the indications for their Disclaimer: The content of this article is solely the re-
inconclusive,12 our data suggest a dose- use, consideration of lower-risk alter- sponsibility of the authors and does not necessarily rep-
resent the official view of National Center for Re-
response relationship at doses higher natives, and proactive adverse effect search Resources, National Institutes of Health, or the
than 1.5 mg/d. The fact that body com- monitoring and management.39 National Institute of Mental Health.

1772 JAMA, October 28, 2009—Vol 302, No. 16 (Reprinted with Corrections) ©2009 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 10/19/2020

 CARDIOMETABOLIC RISK OF ATYPICAL ANTIPSYCHOTICS IN YOUTH

Additional Information: eTables 1-4 are available at of treatment on weight gain and metabolic abnor- Randomized comparison of olanzapine versus risperi-
http://www.jama.com. malities in patients with first-episode psychosis. Aust done for the treatment of first-episode schizophrenia.
Additional Contributions: We thank Martin Lesser, N Z J Psychiatry. 2009;43(9):812-817. Am J Psychiatry. 2006;163(12):2096-2102.
PhD, and Meredith Ackerman, BS, for their statistical 12. Simon V, van Winkel R, De Hert M. Are weight 26. Kelley GA, Kelley KS. Aerobic exercise and HDL2-C.
support (which was provided as part of their institu- gain and metabolic side effects of atypical antipsy- Atherosclerosis. 2006;184(1):207-215.
tional affiliation, without additional financial compen- chotics dose dependent? J Clin Psychiatry. 2009; 27. Fraguas D, Merchán-Naranjo J, Laita P, et al. Meta-
sation); the medical and nursing staff of the child and 70(7):1041-1050. bolic and hormonal side effects in children and ado-
adolescent psychiatry programs at Schneider Chil- 13. Winkleby MA, Robinson TN, Sundquist J, et al. lescents treated with second-generation antipsychotics.
dren’s Hospital and Zucker Hillside Hospital for their Ethnic variation in cardiovascular disease risk factors J Clin Psychiatry. 2008;69(7):1166-1175.
help with identifying eligible patients; and all of the among children and young adults. JAMA. 1999; 28. Meyer JM, Davis VG, McEvoy JP, et al. Impact
patients and their families for their study participa- 281(11):1006-1013. of antipsychotic treatment on nonfasting triglycer-
tion and donation of their time during difficult peri- 14. Cook S, Weitzman M, Auinger P, et al. Preva- ides in the CATIE Schizophrenia Trial phase 1. Schizophr
ods in their lives. lence of a metabolic syndrome phenotype in Res. 2008;103(1-3):104-109.
adolescents. Arch Pediatr Adolesc Med. 2003; 29. Daumit GL, Goff DC, Meyer JM, et al. Antipsy-
157(8):821-827. chotic effects on estimated 10-year coronary heart dis-
REFERENCES 15. Matthews DR, Hosker JP, Rudenski AS, et al. Ho- ease risk in the CATIE schizophrenia study. Schizophr
meostasis model assessment. Diabetologia. 1985; Res. 2008;105(1-3):175-187.
1. Olfson M, Blanco C, Liu L, et al. National trends in 28(7):412-419. 30. Kaddurah-Daouk R, McEvoy J, Baillie RA, et al.
the outpatient treatment of children and adolescents 16. Lee JM, Okumura MJ, Davis MM, Herman WH, Metabolomic mapping of atypical antipsychotic ef-
with antipsychotic drugs. Arch Gen Psychiatry. 2006; Gurney JG. Prevalence and determinants of insulin re- fects in schizophrenia. Mol Psychiatry. 2007;12
63(6):679-685. sistance among US adolescents. Diabetes Care. 2006; (10):934-945.
2. American Diabetes Association. Consensus devel- 29(11):2427-2432. 31. Lloyd-Jones DM, Liu K, Colangelo LA, et al. Con-
opment conference on antipsychotic drugs and obe- 17. Straker D, Correll CU, Kramer-Ginsberg E, et al. sistently stable or decreased body mass index in young
sity and diabetes. Diabetes Care. 2004;27(2):596- Cost-effective screening for the metabolic syndrome adulthood and longitudinal changes in metabolic syn-
601. in patients treated with second-generation antipsy- drome components. Circulation. 2007;115(8):
3. Srinivasan SR, Myers L, Berenson GS. Predictabil- chotic medications. Am J Psychiatry. 2005;162 1004-1011.
ity of childhood adiposity and insulin for developing (6):1217-1221. 32. Calarge CA, Acion L, Kuperman S, et al. Weight
insulin resistance syndrome (syndrome X) in young 18. De Michele M, Panico S, Iannuzzi A, et al. Asso- gain and metabolic abnormalities during extended ris-
adulthood. Diabetes. 2002;51(1):204-209. ciation of obesity and central fat distribution with ca- peridone treatment in children and adolescents. J Child
4. Sinaiko AR, Donahue RP, Jacobs DR Jr, et al. Re- rotid artery wall thickening in middle-aged women. Adolesc Psychopharmacol. 2009;19(2):101-109.
lation of weight and rate of increase in weight during Stroke. 2002;33(12):2923-2928. 33. Henderson DC, Cagliero E, Copeland PM, et al.
childhood and adolescence to body size, blood pres- 19. Juonala M, Raitakari M, S A Viikari J, et al. Obe- Glucose metabolism in patients with schizophrenia
sure, fasting insulin, and lipids in young adults. sity in youth is not an independent predictor of ca- treated with second-generation antipsychotic agents.
Circulation. 1999;99(11):1471-1476. rotid IMT in adulthood. Atherosclerosis. 2006; Arch Gen Psychiatry. 2005;62(1):19-28.
5. Bhargava SK, Sachdev HS, Fall CH, et al. Relation 185(2):388-393. 34. McLaughlin T, Reaven G, Abbasi F, et al. Is there
of serial changes in childhood body-mass index to im- 20. Raitakari OT, Juonala M, Kahonen M, et al. Car- a simple way to identify insulin-resistant individuals
paired glucose tolerance in young adulthood. N Engl diovascular risk factors in childhood and carotid ar- at increased risk of cardiovascular disease? Am J
J Med. 2004;350(9):865-875. tery intima-media thickness in adulthood. JAMA. 2003; Cardiol. 2005;96(3):399-404.
6. Baker JL, Olsen LW, Sørensen TI. Childhood body- 290(17):2277-2283. 35. Hoffman RP. Indices of insulin action calculated
mass index and the risk of coronary heart disease in 21. Allison DB, Mentore JL, Heo M, et al. Antipsychotic- from fasting glucose and insulin reflect hepatic, not
adulthood. N Engl J Med. 2007;357(23):2329- induced weight gain: a comprehensive research peripheral, insulin sensitivity in African-American and
2337. synthesis. Am J Psychiatry. 1999;156(11):1686- Caucasian adolescents. Pediatr Diabetes. 2008;
7. Safer DJ. A comparison of risperidone-induced 1696. 9(3 pt 2):57-61.
weight gain across the age span. J Clin 22. McEvoy JP, Lieberman JA, Perkins DO, et al. Ef- 36. Correll CU. Monitoring and management of an-
Psychopharmacol. 2004;24(4):429-436. ficacy and tolerability of olanzapine, quetiapine, and tipsychotic-related metabolic and endocrine adverse
8. Sikich L, Hamer RM, Bashford RA, et al. A pilot study risperidone in the treatment of early psychosis. Am J events in pediatric patients. Int Rev Psychiatry. 2008;
of risperidone, olanzapine, and haloperidol in psy- Psychiatry. 2007;164(7):1050-1060. 20(2):195-201.
chotic youth. Neuropsychopharmacology. 2004; 23. Schimmelmann BG, Mehler-Wex C, Lambert M, 37. Fleischhacker WW, Cetkovich-Bakmas M, De Hert
29(1):133-145. et al. A prospective 12-week study of quetiapine in M, et al. Comorbid somatic illnesses in patients with
9. Sikich L, Frazier JA, McClellan J, et al. Double- adolescents with schizophrenia spectrum disorders. severe mental disorders. J Clin Psychiatry. 2008;
blind comparison of first- and second-generation an- J Child Adolesc Psychopharmacol. 2007;17(6): 69(4):514-519.
tipsychotics in early-onset schizophrenia and schizo- 768-778. 38. Morrato EH, Newcomer JW, Allen RR, et al. Preva-
affective disorder. Am J Psychiatry. 2008;165(11): 24. Perez-Iglesias R, Crespo-Facorro B, Amado JA, et al. lence of baseline serum glucose and lipid testing in us-
1420-1431. A 12-week randomized clinical trial to evaluate meta- ers of second-generation antipsychotic drugs. J Clin
10. Correll CU. Assessing and maximizing the safety bolic changes in drug-naive, first-episode psychosis pa- Psychiatry. 2008;69(2):316-322.
and tolerability of antipsychotics used in the treat- tients treated with haloperidol, olanzapine, or 39. Kryzhanovskaya LA, Robertson-Plouch CK, Xu
ment of children and adolescents. J Clin Psychiatry. risperidone. J Clin Psychiatry. 2007;68(11):1733- W, et al. The safety of olanzapine in adolescents with
2008;69(suppl 4):26-36. 1740. schizophrenia or bipolar I disorder. J Clin Psychiatry.
11. Verma S, Liew A, Subramaniam M, et al. Effect 25. Robinson DG, Woerner M, Napolitano B, et al. 2009;70(2):247-258.

©2009 American Medical Association. All rights reserved. (Reprinted with Corrections) JAMA, October 28, 2009—Vol 302, No. 16 1773

Downloaded From: https://jamanetwork.com/ on 10/19/2020