The n e w e ng l a n d j o u r na l of
original article
From the Veterans Affairs (VA) New Eng-
land Mental Illness, Research Education
and Clinical Center, VA Connecticut
Healthcare System, West Haven, and the
Yale School of Medicine, New Haven, CT
(R.A.R., J.H.K.); the Massachusetts Veter-
ans Epidemiology and Research Informa-
tion Center VA Cooperative Studies Pro-
gram Coordinating Center, Boston (R.L.,
L.F., D.V., S.S.T., M.H.L.); the VA Health
Economics Resource Center, Menlo Park,
CA (P.G.B.); and the VA Cooperative
Studies Program Clinical Research Phar-
macy Coordinating Center, Albuquerque,
NM (J.E.V.). Address reprint requests to
Dr. Rosenheck at the VA New England
Mental Illness, Research Education and
Clinical Center, VA Connecticut Health-
care System/151D, 950 Campbell Ave.,
West Haven, CT 06516, or at robert
.rosenheck@va.gov.
* The Cooperative Studies Program (CSP)
555 Research Group investigators are
listed in the Supplementary Appendix,
available at NEJM.org.
This article (10.1056/NEJMoa1005987)
was updated on March 7, 2011, at NEJM
.org.
N Engl J Med 2011;364:842-51.
Copyright 2011 Massachusetts Medical Society.
842
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
m e dic i n e
Long-Acting Risperidone and Oral
Antipsychotics in Unstable Schizophrenia
Robert A. Rosenheck, M.D., John H. Krystal, M.D., Robert Lew, Ph.D.,
Paul G. Barnett, Ph.D., Louis Fiore, M.D., M.P.H., Danielle Valley, M.P.H.,
Soe Soe Thwin, Ph.D., Julia E. Vertrees, Pharm.D.,
and Matthew H. Liang, M.D., M.P.H., for the CSP555 Research Group*
A bs t r ac t
Background
Long-acting injectable risperidone, a second-generation antipsychotic agent, may
improve adherence to treatment and outcomes in schizophrenia, but it has not been
tested in a long-term randomized trial involving patients with unstable disease.
Methods
We randomly assigned patients in the Veterans Affairs (VA) system who had schizo-
phrenia or schizoaffective disorder and who had been hospitalized within the pre-
vious 2 years or were at imminent risk for hospitalization to 25 to 50 mg of long-
acting injectable risperidone every two weeks or to a psychiatrists choice of an oral
antipsychotic. All patients were followed for up to 2 years. The primary end point
was hospitalization in a VA or non-VA psychiatric hospital. Symptoms, quality of life,
and functioning were assessed in blinded videoconference interviews.
Results
Of 369 participants, 40% were hospitalized at randomization, 55% were hospital-
ized within the previous 2 years, and 5% were at risk for hospitalization. The rate of
hospitalization after randomization was not significantly lower among patients who
received long-acting injectable risperidone than among those who received oral anti-
psychotics (39% after 10.8 months vs. 45% after 11.3 months; hazard ratio, 0.87;
95% confidence interval, 0.63 to 1.20). Psychiatric symptoms, quality of life, scores
on the Personal and Social Performance scale of global functioning, and neuro-
logic side effects were not significantly improved with long-acting injectable ris-
peridone as compared with control treatments. Patients who received long-acting
injectable risperidone reported more adverse events at the injection site and more
extrapyramidal symptoms.
Conclusions
Long-acting injectable risperidone was not superior to a psychiatrists choice of oral
treatment in patients with schizophrenia and schizoaffective disorder who were
hospitalized or at high risk for hospitalization, and it was associated with more local
injection-site and extrapyramidal adverse effects. (Supported by the VA Cooperative
Studies Program and Ortho-McNeil Janssen Scientific Affairs; ClinicalTrials.gov
number, NCT00132314.)
n engl j med 364;9 nejm.org march 3, 2011
The New England Journal of Medicine
Copyright 2011 Massachusetts Medical Society. All rights reserved.
 Long-Acting Risperidone in schizophrenia
T
he most common and potentially
remediable cause of treatment failure in
patients with schizophrenia is lack of ad-
herence to prescribed oral medications.1,2 By en-
suring sustained levels of drug in the blood,
long-acting injectable delivery may improve adher-
ence and symptom control and reduce the rate of
relapse and hospitalization.2-5
In the United States, the first second-generation
antipsychotic agent to be made available in a long-
acting injectable delivery system was risperidone
(Risperdal Consta, Ortho-McNeil Janssen). Long-
acting injectable risperidone may cause fewer
extrapyramidal symptoms than the long-acting
injectable first-generation antipsychotic agents.6
A randomized trial showed the efficacy of
long-acting injectable risperidone over placebo
in patients with schizophrenia,7 and before-and-
after studies have shown tolerability in switching
from oral to long-acting injectable risperidone,
with improved symptoms and reduced hospital
use.8-11 These studies involved clinically stable
patients and lacked randomized control groups.
Three randomized trials that also involved pa-
tients with stable disease showed no advantage
of long-acting injectable risperidone therapy over
oral treatment.12-14
In this trial involving patients with unstable
disease, we hypothesized that long-acting inject-
able risperidone would be superior in reducing
the risk of hospitalization for up to 2 years as
compared with a psychiatrists choice of an oral
antipsychotic.
Me thods
Participants
Patients were eligible to participate in the study
if they were 18 years of age or older, had a diag-
nosis of schizophrenia or schizoaffective disor-
der as assessed with the use of the Structured
Clinical Interview based on the fourth edition of
the Diagnostic and Statistical Manual of Mental Disor-
ders,15 and were at risk for psychiatric hospital-
ization as evidenced by current psychiatric hospi-
talization, hospitalization in the previous 2 years,
or increased use of mental health services to pre-
vent relapse as adjudicated by the study chairper-
sons (the first two authors). The original entry
criteria required hospitalization in the previous
year but were extended to enhance recruitment
(see the study protocol, available with the full text
of this article at NEJM.org).
n engl j med 364;9 nejm.org march 3, 2011
The New England Journal of Medicine
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
Randomization began in September 2006, and
data collection continued for 3 years, with 209
of 369 patients (56.6%) randomly assigned in the
first year, 140 patients (37.9%) assigned in the
second year, and 20 patients (5.4%) assigned
during the first 3 months of the third year.
Follow-up continued for up to 2 years.
Exclusion criteria were the following: detoxi-
fication in the previous month; reported past
intolerance to risperidone or intramuscular injec-
tions; current treatment with long-acting inject-
able antipsychotics, oral clozapine, warfarin, or
a combination of these agents; serious medical
conditions; unstable living arrangements; and a
history of assault or suicidal behavior requiring
urgent intervention.
The patients decisional capacity was assessed
with the use of the MacArthur Competence As-
sessment Tool.16 Guardian consent was allowed.
Subjects received payment for their travel ex-
penses and time: $25 for monthly and injection-
only visits and $45 for extended quarterly assess-
ment visits. The injectable-risperidone group thus
had more planned paid visits than the oral-anti-
psychotic group. After written informed consent
had been obtained from the patient or guardian,
testing for allergic reactions was performed with
an oral test dose of 1 mg of risperidone. Long-
acting injectable risperidone was provided free
of charge by Ortho-McNeil Janssen Scientific Af-
fairs, which had no role in the study.
The study and consent forms were approved
by the institutional review boards of the 19 col-
laborating centers. The analyses were conducted
at the Veterans Affairs (VA) Cooperative Studies
Program Coordinating Center, Boston, and the VA
Health Economics Resource Center, Menlo Park,
California. All authors designed the trial, inter-
preted the findings, agreed to publication of the
manuscript, and reviewed and approved the manu-
script. The first author wrote the first draft of
the manuscript. All authors vouch for the com-
pleteness and accuracy of the data, the data
analyses, and the fidelity of this report to the
study protocol.
Randomization
Randomization was conducted centrally and strat-
ified according to site because of potential prac-
tice differences. Randomization was conducted
with the use of randomly permuted blocks of
variable size to ensure an approximate balance
over time.
843
 The n e w e ng l a n d j o u r na l
Treatment Groups
Patients randomly assigned to long-acting inject-
able risperidone were seen clinically by a study
nurse every 2 weeks for the first month and then
monthly. All patients were seen monthly by their
psychiatrist and by the nurse. On the basis of
consensus guidelines,17 long-acting injectable ris-
peridone was administered intramuscularly at an
initial dose of 25 mg every 2 weeks. Dosage in-
crements of 12.5 mg were permitted every 4 weeks
at the discretion of the treating psychiatrist, up
to the maximum approved dose of 50 mg.
Steady-state drug levels are reached 6 to 8 weeks
after initiation of treatment with long-acting
injectable risperidone,17 and efforts to reduce
the use of oral antipsychotics subsequently were
encouraged in the injectable-risperidone group.
Previous oral antipsychotics were to be continued
for at least 3 weeks. Treatment interruptions
among patients randomly assigned to long-acting
injectable risperidone were addressed by restart-
ing the intramuscular medication and providing
oral medication for 3 weeks if the interruption
occurred before the steady state was reached, or
if the interruption was longer than 6 weeks.
Concomitant psychotropic medications (i.e.,
antianxiety agents, antidepressants, and oral anti-
psychotics and mood stabilizers) and anticho-
linergic medications were allowed.
Control-group participants continued to re-
ceive oral antipsychotic therapy as prescribed by
their treating physician. Treating psychiatrists
were given a summary of optimal dosage ranges
for oral antipsychotic and anticholinergic agents,
based on published recommendations.18
Concomitant Psychosocial Treatment
To ensure that no patient was randomly assigned
to less than standard best practice an ethical
imperative a short checklist of potentially use-
ful ancillary psychosocial services available at the
participating centers was provided to all partici-
pants during follow-up visits.19
Measures
Blinded videoconference assessments were com-
pleted every 3 months on measures of symptoms,
quality of life, and functioning.
At a monthly unblinded meeting with the
study nurse, the Clinical Global Impressions
(CGI) scale20 was used to assess the patients
global mental health status and the change from
844 n engl j med 364;9 nejm.org march 3, 2011
The New England Journal of Medicine
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
baseline (on a scale of 1 to 7, with higher scores
indicating poorer functioning or less improve-
ment). Satisfaction with medication was measured
with the use of the Drug Attitude Inventory (on
a scale of 1 to 20, with higher scores indicating
greater satisfaction).21
Retention on the assigned drug was measured
according to the number of days until discontinu-
ation of the assigned treatment or, among partici-
pants assigned to the oral medication, days to
crossover to any new oral or long-acting inject-
able treatment. The use of long-acting injectable
risperidone was documented according to study
prescribing records, and the use of oral medica-
tion was documented according to patient inter-
views. Efforts were made to ensure that patients
continued to receive the medications selected by
their doctor if they discontinued the study drug.
Symptoms of schizophrenia were measured
according to the total score on the Positive and
Negative Syndrome Scale (PANSS, which ranges
from 30 to 210, with higher scores indicating
more symptoms), and its positive, negative, and
general subscales.22 PANSS ratings were obtained
from standardized videoconferences conducted
by trained raters from MedAvante who were un-
aware of the patients study-drug assignments.
Psychiatric assessments by video conference are
reliable in patients with schizophrenia and are
well received.23
Subjective psychological distress was measured
with the use of the depression and anxiety sub-
scales of the Brief Symptom Index (on a scale of
0 to 4, with higher scores indicating greater
distress).24
Quality of Life and Social Functioning
Quality of life was measured with the use of the
HeinrichsCarpenter Quality of Life Scale (rang-
ing from 0 to 120, with higher scores indicating
better quality of life)25 and the Personal and So-
cial Performance scale (ranging from 1 to 100,
with higher scores reflecting better function-
ing),26 the latter providing a global assessment of
social functioning. Both were administered by
videoconference assessors who were unaware of
the study-drug assignments.
Health-related quality of life was assessed
with the use of the Quality of Well-Being scale
(ranging from 0 to 1, with higher scores indicat-
ing greater well-being),27 which has been vali-
dated for use in schizophrenia.28
 Long-Acting Risperidone in schizophrenia
Substance Use
At screening, physicians and patients were asked
whether substance abuse was a problem. Alcohol
and drug use in the previous 30 days was as-
sessed with the use of the alcohol and drug com-
posite indexes from the Addiction Severity Index
(on a scale of 0 to 1, with higher scores indicat-
ing more severe problems).29
Side Effects
Neurologic side effects were measured with the
use of three scales.30-32 Sexual dysfunction was
measured with items from the Novel Antipsychot-
ic Medication Experience Scale (ranging from
0 to 4, with higher scores indicating worse side
effects) (Ames D: personal communication).
Hospitalization and Use of Other Medical Services
Administrative data on service use, including hos-
pitalizations, were available for all VA health ser-
vices. Psychiatric inpatient admissions were iden-
tified through the VAs Patient Treatment File.
Non-VA admissions were identified according to
discharge summaries validated as psychiatric by
a physician who was unaware of the patients
study-drug assignments.
The primary outcome measure was the time
from randomization to psychiatric hospitaliza-
tion (in both VA and non-VA hospitals) or, in the
case of patients who were hospitalized at ran-
domization, the time from the date of discharge
from the initial stay to subsequent hospitaliza-
tion. The key secondary outcome measure was the
change in the PANSS total score at 12 months.
Secondary analyses compared outcomes at all
time points up to 18 months, rather than the
difference between follow-up scores and base-
line scores at one specific time point.
Statistical Analysis
The planned sample size of 450 patients (the
original sample size of 600 was resized because
of recruitment difficulties) provided 90% power
for analyses of our primary outcome and second-
ary outcome, each with a two-sided test and a
type I error of 2.5% (i.e., 1.25% in each tail). First,
a time-to-event analysis, with the use of the log-
rank test, compared the hazard ratios associated
with the time to the first psychiatric hospitaliza-
tion. With a null hypothesis that the hazard ratio
would equal 1, the alternative hypothesis was that
for long-acting injectable risperidone versus oral
n engl j med 364;9 nejm.org march 3, 2011
The New England Journal of Medicine
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
agents, the hazard ratio was greater than or equal
to 1.65 or less than or equal to 0.60. This hypoth-
esis was derived from an assumption based on
three studies in which baseline rates of relapse
were approximately 41% in the oral-antipsychotic
group and approximately 25% in the intramuscu-
lar-medication group (i.e., a rate ratio of 1.64
[4125] corresponding to a difference of 16 per-
centage points [41%25%] in the annual rate of
a first psychiatric hospitalization).2,33,34 The fol-
low-up period for this outcome was up to 2 years,
terminating with hospitalization or discontinua-
tion of the assigned study medication.
Confirmatory Cox proportional-hazards analy-
ses controlled for potential confounding factors.
These factors included prior use of risperidone,
history of substance abuse, and hospitalization
at the time of enrollment.
A repeated-measures mixed-effects model was
used to compare the mean change from baseline
to 12 months in the PANSS score for injectable
and oral treatments. With a null hypothesis of
no difference, the alternative hypothesis was
that the difference was greater than or equal to
5 units or less than or equal to 5 units. The
model had fixed effects for treatment group and
time (a categorical variable); the interaction of
treatment with time, site, and individual patients
were treated as random effects. A first-order
autocorrelation structure was used. The baseline
PANSS score was added to the model to assess
its effect on changes from baseline. Confirma-
tory mixed models were run with the PANSS
score.
Further descriptive analysis of outcome and
side-effect measures used mixed models of all
outcome data up to 18 months because of exten-
sive sample attrition after that time. Because of
the skewed distribution of service use, the sig-
nificance of differences was tested with the
Wilcoxon rank-sum test.
R e sult s
Study Participants
Altogether, 1045 patients were screened at 19 VA
medical centers between 2006 and 2009, yielding
a final analytic sample of 369 patients (Fig. 1).
Five sites discontinued the study because of in-
sufficient recruitment. Participants were hospi-
talized at the time of randomization (40%), had
been hospitalized within the previous 2 years
845
 The n e w e ng l a n d j o u r na l of m e dic i n e

1045 Patients were assessed for eligibility

514 Were ineligible

110 Patients or physicians declined participation
39 Patients or physicians never called back

382 Underwent randomization

192 Were assigned to receive 190 Were assigned to receive

oral treatment injectable risperidone

7 Declined participation 2 Declined participation

185 Received oral treatment 188 Received injectable risperidone

120 Completed study 117 Completed study
65 Were lost to follow-up or discon- 71 Were lost to follow-up or discon-
tinued intervention tinued intervention

3 Were excluded because

patient did not have a 1 Was excluded because
Social Security number or of lack of baseline data
did not have baseline data

182 Were included in analysis 187 Were included in analysis

Figure 1. Enrollment, Randomization, and Follow-up of the Study Patients.

(55%), or had recent increased service use indi-
cating a risk of hospitalization (5%). At screening,
problems with medication adherence were report-
ed for 64% of the patients; 43% of the patients
reported problems by themselves and in 60% of
the cases, problems were reported by physicians.
Active problems with alcohol or drug use were
reported for 37% of the patients; 25% were re-
ported by the participants and 36% were report-
ed by their physicians. There were no significant
differences between groups at baseline in this
sample of older male veterans (Table 1 in the
Supplementary Appendix, available at NEJM.org).
Treatment and Follow-Up Assessments
For patients assigned to and receiving long-acting
injectable risperidone, at 6 weeks, 86% of injec-
tion doses were 25 mg, 11% were 37.5 mg, and
3% were 50 mg, with an average of 1.8 injections
per month. During the remainder of the trial,
17% of doses were 25 mg, 31% were 37.5 mg, and
50% were 50 mg, with an average of 1.5 injec-
tions per month (the percentages do not sum to
100 because of rounding). During the first 6 weeks,
40% of patients receiving long-acting injectable
risperidone received concomitant oral antipsy-
chotics. During the remainder of the trial, 32%
of injections were accompanied by prescriptions
for oral antipsychotics during the same month.
The follow-up interview rates in the intention-
to-treat analysis were as follows: 60% (223 pa-
tients) at 1 year, 46% (170) at 18 months, and
29% (107) at 24 months, with no significant dif-
ferences between groups at these time points
(P=0.42 to 0.99). The mean (SD) duration of
participation was 474235 days for long-acting
injectable risperidone versus 502226 days for
oral antipsychotics (P=0.22).

846 n engl j med 364;9 nejm.org march 3, 2011

The New England Journal of Medicine

Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
 Long-Acting Risperidone in schizophrenia

Outcomes
Long-acting injectable treatment was not superi- 1.0 P=0.39 by the log-rank test
or to oral treatment in the duration of adherence

Freedom from Hospitalization

0.8
to the randomized treatment (P=0.19) (Fig. 1 in
the Supplementary Appendix). Among partici- Injectable risperidone
0.6
pants receiving oral treatment, however, 21 of 182
(12%) switched to long-acting injectable risperi- 0.4
Oral antipsychotic
done an average of 153203 days after randomiza-
tion. There were no significant differences with 0.2
respect to the initiation of concomitant psycho
tropic medications (Fig. 2 in the Supplementary 0.0
0 3 6 9 12 15 18 21 24
Appendix).
Months after Randomization
A total of 237 of 369 patients (64%) continued
No. at Risk
to receive the study drug throughout their par- Oral antipsychotic 182 136 116 96 84 71 58 49 28
ticipation in the study. Reasons for medication Injectable 187 136 110 92 82 65 53 45 37
risperidone
discontinuation were not significantly different
between groups (Table 2 in the Supplementary Figure 2. Time to Hospitalization after Randomization.
Appendix). In this analysis, data on patients who withdrew from the study were cen-
With a mean follow-up of 11.3 and 10.8 sored at the time of withdrawal from the study.
months, respectively, 81 of 182 (45%) patients
receiving oral medication and 72 of 187 (39%)
receiving long-acting injectable risperidone were favored long-acting injectable risperidone (P=0.02).
hospitalized. Long-acting injectable risperidone Although there was no superiority of long-acting
was not superior to oral treatment with respect injectable risperidone on the unblinded assess-
to the time to hospitalization (P=0.39 by the ment of illness severity at each time point, the
log-rank test; hazard ratio, 0.87, 95% confidence unblinded CGI improvement score, representing
interval [CI], 0.63 to 1.20) (Fig. 2). An analysis the rater-perceived change from baseline, favored
that excluded the 21 subjects who switched from long-acting injectable risperidone (P<0.001).
an oral antipsychotic to long-acting injectable Analysis of adverse events (Table 3 in the Sup-
risperidone provided similar results (hazard ratio, plementary Appendix) showed that patients who
1.00; 95% CI, 0.71 to 1.40), as did an analysis received long-acting injectable risperidone had
that was adjusted for covariates (hazard ratio, more general disorders and administration site
0.82; 95% CI, 0.59 to 1.13). conditions (injection-related pain or induration)
The mixed-model analysis of the change from (P=0.04) and nervous system disorders (head-
baseline to 12 months in the PANSS total score ache and extrapyramidal signs and symptoms)
did not show superiority of long-acting inject- (P<0.001). There were four deaths. In the inject-
able risperidone (P=0.72). able-risperidone group, one patient died in his
Further outcome comparisons across all time sleep from an unknown cause and another com-
points up to 18 months showed no significant mitted suicide. In the oral-antipsychotic group, one
between-group differences in the PANSS total patient died from chronic obstructive pulmonary
score or subscales (Table 1, and Fig. 3 in the disease, and another from accidental drowning.
Supplementary Appendix). No significant superi-
ority of long-acting injectable risperidone was Use of Services
observed on the blindly rated HeinrichsCarpenter A larger proportion of patients receiving long-
Quality of Life Scale or its subscales, the Per- acting injectable risperidone were hospitalized at
sonal and Social Performance scale or the self- the time of randomization and they were hospi-
reported Quality of Well-Being scale, the current talized for more days during the period before
CGI functioning measure, or the Addiction Sever- randomization (Table 2). After randomization,
ity Index composite drug scores (Table 1). The there were no significant differences between
composite alcohol index of the Addiction Sever- groups with respect to VA service use (Table 2) or
ity Index was higher in the oral-antipsychotic non-VA service use (Table 4 in the Supplementary
group (P=0.04) and the Drug Attitude Inventory Appendix), including the number of hospital days.

n engl j med 364;9 nejm.org march 3, 2011 847

The New England Journal of Medicine
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 1. Follow-up Assessment Outcomes Based on Mixed Models with the Use of All Available Data over All Time
Points up to 18 Months.*

Variable Oral Antipsychotic Injectable Risperidone Mean Difference P Value

PANSS
Total score 74.690.92 74.100.91 0.591.27 0.65
General symptoms 37.170.46 36.890.45 0.270.64 0.67
Positive symptoms 18.840.38 18.120.38 0.720.48 0.13
Negative symptoms 18.690.35 19.030.35 0.350.37 0.36
HeinrichsCarpenter Quality of Life Scale
Total score 2.860.06 2.780.06 0.080.07 0.28
Interpersonal relations 2.550.08 2.460.08 0.090.10 0.36
Instrumental functioning 2.660.05 2.650.05 0.01 0.06 0.81
Intrapsychic foundations 3.240.06 3.140.06 0.10 0.08 0.18
Personal and Social Performance Scale 53.830.78 53.640.78 0.180.90 0.84
Body-mass index 30.690.50 30.070.51 0.620.72 0.39
Clinical Global Impressions

Severity of illness 4.190.13 4.220.13 0.030.09 0.34

Change in condition 3.520.08 3.220.08 0.300.06 <0.001
Addiction Severity Index**
Alcohol use 0.130.03 0.070.03 0.060.03 0.04
Drug use 0.0120.003 0.0180.003 0.0060.004 0.13
Brief Symptom Index 0.670.62 0.640.62 0.030.06 0.55
Quality of Well-Being scale 0.660.02 0.670.02 0.010.01 0.63
Abnormal Involuntary Movement Scale 0.260.04 0.210.04 0.060.03 0.11
SimpsonAngus Scale 0.230.05 0.220.05 0.010.03 0.60
Barnes Akathisia Scale 0.440.09 0.450.09 0.010.06 0.80
NAMES***
Sexual interest 1.060.10 1.010.10 0.050.10 0.61
Sexual activities 1.100.11 0.930.11 0.170.11 0.13
Drug Attitude Inventory 7.960.13 8.270.13 0.310.13 0.02

* Plusminus values are means SE. For all outcomes, the treatment comparison was a linear contrast based on a
mixed-effects model with three fixed effects (time, treatment, and timetreatment interaction), with site as a ran-
dom effect and with autocorrelated repeated measures over time.
Scores on the Positive and Negative Syndrome Scale (PANSS) range from 30 to 210, with higher scores indicating
more symptoms.
Scores on the HeinrichsCarpenter Quality of Life Scale range from 0 to 120, with higher scores indicating better
quality of life.
Scores on the Personal and Social Performance Scale range from 1 to 100, with higher scores reflecting better functioning.
Body-mass index is the weight in kilograms divided by the square of the height in meters.

Scores on the Clinical Global Impressions scale range from 1 to 7, with higher scores indicating poorer functioning
or less improvement.
** Scores on the Addiction Severity Index range from 0 to 1, with higher scores indicating more severe problems.
Scores on the Brief Symptom Index range from 0 to 4, with higher scores indicating greater distress.
Scores on the Quality of Well-Being scale range from 0 to 1, with higher scores indicating better well-being.
Scores on the Abnormal Involuntary Movement Scale range from 0 to 4, with higher scores indicating more severe
tardive dyskinesia.

Scores on the SimpsonAngus Scale range from 0 to 4, with higher scores indicating more severe extrapyramidal
symptoms.
Scores on the Barnes Akathisia Scale range from 0 to 5, with higher scores indicating more severe akathisia.
*** Scores on the Novel Antipsychotic Medication Experience Scale (NAMES) range from 0 to 4, with higher scores in-
dicating worse side effects.
Scores on the Drug Attitude Inventory range from 1 to 20, with higher scores indicating greater satisfaction.

848 n engl j med 364;9 nejm.org march 3, 2011

The New England Journal of Medicine

Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
 Long-Acting Risperidone in schizophrenia

Discussion ena, headache, and extrapyramidal signs and

This randomized, controlled trial showed that in
high-risk patients with schizophrenia or schizo
affective disorder, long-acting injectable risperi-
done was not superior to oral antipsychotics with
respect to the primary outcome of time to hospi-
talization, or multiple standard measures of
symptoms, quality of life, side effects, or service
use. Greater numbers of adverse events were re-
ported by the injectable-risperidone group. These
events primarily included injection-site phenom-
symptoms, suggesting that patients receiving oral
medication may flexibly adjust their medication
use to avoid such adverse effects. The duration of
treatment with long-acting injectable risperidone
was not significantly longer than the duration of
treatment with oral antipsychotics.
The findings were not modified by the addi-
tion of covariates or the exclusion of crossover
observations (for participants who switched from
oral to long-acting injectable treatment). Differ-
ences in the alcohol composite index of the Ad-

Table 2. Use of Health Services Provided by the Veterans Affairs System.*

Oral Injectable
Antipsychotic Risperidone
Type of Use (N=182) (N=187) P Value
Inpatient care
Acute medical or surgical hospital stays
Days 1.04.1 1.04.0 0.95
Patients with any hospitalization (%) 15.4 15.0 0.91
Total acute psychiatric hospital stays after randomization
Total days 20.343.4 19.259.7 0.80
Patients with any hospitalization (%) 62.1 64.7 0.60
Hospitalization at time of randomization
Patients hospitalized (%) 35.2 45.5 0.04
Days from hospitalization at randomization to discharge 2.77.4 8.453.0 0.02
Hospitalizations subsequent to the original stay
Patients with new hospitalization after randomization (%)* 42.9 36.4 0.20
Days in subsequent stays 17.641.1 10.828.0 0.21
No. of subsequent stays among patients with any stays 2.62.5 2.21.5 0.60
Residential treatment, nonhospital
Patients with any residential treatment admission (%) 23.6 19.3 0.31
Days 26.486.4 18.171.3 0.49
Outpatient care
Outpatient visits after randomization (no.)
Individual psychiatry 58.965.8 52.056.2 0.67
Group psychiatry 30.163.5 24.556.6 0.36
Vocational rehabilitation 5.415.4 3.815.3 0.25
Telephone psychiatry 3.66.6 2.44.8 0.05
Other psychiatry 1.02.9 0.62.0 0.33
Medical and surgical 15.115.9 16.124.3 0.22
Other ancillary care 22.433.3 23.140.5 0.77
Total outpatient visits 136.5137.0 122.4130.9 0.26
Visits to administer long-acting injectable risperidone (no.) 1.24.9 19.714.7 <0.001

* These data pertain to hospitalizations at Veterans Affairs (VA) hospitals only and thus the percentages are somewhat
smaller than the total proportion of patients who were hospitalized (i.e., at either VA or non-VA hospitals).

n engl j med 364;9 nejm.org march 3, 2011 849

The New England Journal of Medicine
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

diction Severity Index and the Drug Attitude In-
ventory were not significant after adjustment for
multiple comparisons. Although the current CGI
scores assigned by raters who were aware of the
patients study-drug assignments did not differ
between groups, the CGI improvement scores as-
signed by these raters indicated significantly
greater improvement in the group of patients who
received long-acting injectable risperidone, sug-
gesting an unblinded rater bias favoring long-
acting injectable risperidone.
Taken together, these findings are consistent
with three efficacy trials that also showed no
superiority of long-acting injectable risperidone
over oral regimens in patients with stable schizo-
phrenia.12-14 Two studies have suggested that
unintended intramuscular injections into fat tis-
sue may decrease pharmacologic effectiveness,
but this was not assessed in our study.35,36
Our study had several limitations. First, 12%
of control patients received long-acting injectable
risperidone treatment an average of 5 months
into the trial. This may have biased the results
in favor of oral treatment in the intention-to-
treat analysis. Replication of the analyses of hos-
pitalization risk and blinded outcomes excluding
observations after these crossovers or discontinu-
ation of long-acting injectable risperidone yield-
ed no significant findings favoring long-acting
injectable treatment.
Second, the dose of long-acting injectable ris-
peridone may have been inadequate in some pa-
tients, and some injections were missed, but this
reflects the real-world practice that was the focus
of this effectiveness study.
Third, decisions regarding hospitalization
were unblinded, and the direction of any bias is
unknown. If physicians thought there was less
need to hospitalize patients, knowing that they
were receiving ample medication, the bias could
favor long-acting injectable risperidone. On the
other hand, if admitting physicians knew that
patients receiving long-acting risperidone were
symptomatic in spite of being adequately medi-
cated, the bias could favor oral treatment.
Fourth, this sample involved older, primarily
male veterans, and results may not be generaliz-
able to other populations.
Finally, although our revised target sample
was 450 subjects, we enrolled only 382 subjects,
and data were available for only 369 because of
early dropouts. Dropout patterns and sample
sizes were similar to those of previous schizo-
phrenia trials.36,37 Our study did not show the
superiority of long-acting injectable risperidone,
but the confidence intervals for the time to hos-
pitalization were fairly wide (hazard ratio, 0.87;
95% CI, 0.63 to 1.20), and the study was not
large enough to exclude modest differences
between the groups.
Supported by the Veterans Affairs Cooperative Studies Pro-
gram and an unrestricted grant from Ortho-McNeil Janssen
Scientific Affairs.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.

References
1. Thieda P, Beard S, Richter A, Kane J.
An economic review of compliance with
medication in the treatment of schizo-
phrenia. Psychiatr Serv 2003;54:508-16.
2. Weiden PJ, Olfson M. Cost of relapse
in schizophrenia. Schizophr Bull 1995;21:
419-29.
3. Davis JM, Matalon L, Watanabe MD,
Blake L, Matalon L. Depot antipsychotic
drugs: place in therapy. Drugs 1994;47:741-
73.
4. Adams CE, Fenton MK, Quraishi S,
David AS. Systematic meta-review of de-
pot antipsychotics for people with schizo-
phrenia. Br J Psychiatry 2001;179:290-9.
5. Hogarty GE, Schooler NR, Ulrich R,
Mussare F, Ferro P, Herron E. Fluphena
zine and social therapy in the aftercare of
schizophrenic patients: relapse analyses
of a two-year controlled study of fluphen-
azine decanoate and fluphenazine hydro-
chloride. Arch Gen Psychiatry 1979;36:
1283-94.
6. Miyamoto S, Duncan GE, Goff DC,
Lieberman JA. Therapeutics of schizo-
phrenia. In: Davis KL, Charney D, Coyle
JT, Nemeroff C, eds. Neuropsychophar-
macology: the fifth generation of prog-
ress. Philadelphia: Lippincott Williams
& Wilkins, 2002:775-807.
7. Kane JM, Eerdekens M, Lindenmayer
JP, Keith SJ, Lesem M, Karcher K. Long-
acting injectable risperidone: efficacy and
safety of the first long-acting atypical
antipsychotic. Am J Psychiatry 2003;160:
1125-32.
8. Mller HJ, Llorca PM, Sacchetti E,
Martin SD, Medori R, Parellada E. Efficacy
and safety of direct transition to risperidone
long-acting injectable in patients treated
with various antipsychotic therapies. Int
Clin Psychopharmacol 2005;20:121-30.
9. Fleischhacker WW, Eerdekens M,
Karcher K, et al. Treatment of schizophre-
nia with long-acting injectable risperi-
done: a 12-month open-label trial of the
first long-acting second-generation anti-
psychotic. J Clin Psychiatry 2003;64:1250-
7.
10. Lindenmayer J-P, Eerdekens E, Berry
SA, Eerdekens M. Safety and efficacy of
long-acting risperidone in schizophrenia:
a 12-week, multicenter, open-label study
in stable patients switched from typical
and atypical oral antipsychotics. J Clin
Psychiatry 2004;65:1084-9.
11. Fleischhacker WW. Second-generation
antipsychotic long-acting injections: sys-
tematic review. Br J Psychiatry Suppl 2009;
52:S29-S36.
12. Chue P, Eerdekens M, Augustyns I,
et al. Comparative efficacy and safety of
long-acting risperidone and risperidone
oral tablets. Eur Neuropsychopharmacol
2005;15:111-7.
13. Bai YM, Ting Chen T, Chen JY, et al.
Equivalent switching dose from oral ris-
peridone to risperidone long-acting injec-
tion: a 48-week randomized, prospective,

850 n engl j med 364;9 nejm.org march 3, 2011

The New England Journal of Medicine

Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.
 Long-Acting Risperidone in schizophrenia

single-blind pharmacokinetic study. J Clin
Psychiatry 2007;68:1218-25.
14. Keks NA, Ingham M, Khan A, Karcher
K. Long-acting injectable risperidone v.
olanzapine tablets for schizophrenia or
schizoaffective disorder: randomised, con-
trolled, open-label study. Br J Psychiatry
2007;191:131-9.
15. First MB, Spitzer RL, Gibbon MB,
Williams JB. Structured Clinical Interview
for Axes I and II DSM IV Disorders Pa-
tient Edition (SCID-I/P). New York: Bio-
metrics Research Institute, New York State
Psychiatric Institute, 1996.
16. Grisso T, Appelbaum PS. MacArthur
Competence Assessment Tool Treat-
ment. Worcester: University of Massachu-
setts Medical Center, 1995.
17. Keith SJ, Pani L, Nick B, et al. Practi-
cal application of pharmacotherapy with
long-acting risperidone for patients with
schizophrenia. Psychiatr Serv 2004;55:
997-1005.
18. Lehman AF, Kreyenbuhl J, Buchanan
RW, et al. The Schizophrenia Patient Out-
comes Research Team (PORT): updated
treatment recommendations 2003. Schizo
phr Bull 2004;30:193-217.
19. Rosenheck R, Tekall J, Peters J, et al.
Does participation in psychosocial treat-
ment augment the benefit of clozapine?
Arch Gen Psychiatry 1998;55:618-25.
20. Guy W. Clinical Global Impressions
(CGI). In: ECDEU assessment manual for
psychopharmacology. Rockville, MD: Na-
tional Institute of Mental Health, 1976:
534-7.
21. Hogan TP, Awad AG, Eastwood R.
A self-report scale predictive of drug com-
pliance in schizophrenics: reliability and
discriminative validity. Psychol Med 1983;
13:177-83.
22. Kay SR, Fiszbein A, Opler LA. The
Positive and Negative Syndrome Scale
(PANSS) for schizophrenia. Schizophr Bull
1987;13:261-76.
23. Ruskin PE, Reed S, Kumar R, et al.
Reliability and acceptability of psychiatric
diagnosis via telecommunication and au-
diovisual technology. Psychiatr Serv 1998;
49:1086-8.
24. Derogatis LR, Spencer N. The Brief
Symptom Index: administration, scoring
and procedure manual. Baltimore: Johns
Hopkins University Press, 1982.
25. Heinrichs DW, Hanlon ET, Carpenter
WT Jr. The Quality of Life Scale: an instru-
ment for rating the schizophrenic deficit
syndrome. Schizophr Bull 1984;10:388-98.
26. Morosini PL, Magliano L, Brambilla L,
Ugolini S, Pioli R. Development, reliability
and acceptability of a new version of the
DSM-IV Social and Occupational Func-
tioning Assessment Scale (SOFAS) to as-
sess routine social functioning. Acta Psy-
chiatr Scand 2000;101:323-9.
27. Kaplan RM, Anderson JP. The general
health policy model: an integrated ap-
proach. In: Spilker B, ed. Quality of life
and pharmacoeconomics in clinical trials.
2nd ed. Philadelphia: Lippincott-Raven,
1996:309-22.
28. Pyne JM, Sullivan G, Kaplan R, Wil-
liams DK. Comparing the sensitivity of
generic effectiveness measures with symp-
tom improvement in persons with schizo-
phrenia. Med Care 2003;41:208-17.
29. McLellan AT, Luborsky L, Woody GE,
OBrien CP. An improved diagnostic evalu-
ation instrument for substance abuse pa-
tients: the Addiction Severity Index. J Nerv
Ment Dis 1980;168:26-33.
30. Barnes TR. A rating scale for drug in-
duced akathisia. Br J Psychiatry 1989;154:
672-6.
31. Guy W. Abnormal Involuntary Move-
ment Scale (AIMS). In: ECDEU assessment
manual for psychopharmacology. Rock-
ville, MD: National Institute of Mental
Health, 1976:534-7.
32. Simpson GM, Angus JW. A rating
scale for extrapyramidal side effects. Acta
Psychiatr Scand Suppl 1970;212:11-9.
33. Herz MI, Glazer WM, Mostert MA,
et al. Intermittent vs maintenance medi-
cation in schizophrenia: two-year results.
Arch Gen Psychiatry 1991;48:333-9.
34. Young JL, Zonana HV, Shepler L. Med-
ication noncompliance in schizophrenia:
codification and update. Bull Am Acad
Psychiatry Law 1986;14:105-22.
35. Nesvg R, Tanum L. Therapeutic drug
monitoring of patients on risperidone de-
pot. Nord J Psychiatry 2005;59:51-5.
36. Marder SR, Meibach RC. Risperidone
in the treatment of schizophrenia. Am J
Psychiatry 1994;151:825-35.
37. Csernansky JG, Mahmoud R, Brenner
R. A comparison of risperidone and halo-
peridol for the prevention of relapse in
patients with schizophrenia N Engl J Med
2002;346:16-22. [Erratum, N Engl J Med
2002;346:1424.]
Copyright 2011 Massachusetts Medical Society.

new nejm application for iphone

The NEJM Image Challenge app brings a popular online feature to the smartphone.
Optimized for viewing on the iPhone and iPod Touch, the Image Challenge app lets
you test your diagnostic skills anytime, anywhere. The Image Challenge app
randomly selects from 300 challenging clinical photos published in NEJM,
with a new image added each week. View an image, choose your answer,
get immediate feedback, and see how others answered.
The Image Challenge app is available at the iTunes App Store.

n engl j med 364;9 nejm.org march 3, 2011 851

The New England Journal of Medicine
Downloaded from nejm.org on April 27, 2017. For personal use only. No other uses without permission.
Copyright 2011 Massachusetts Medical Society. All rights reserved.