How Long to Wait for a Response to Clozapine:
A Comparison of Tune Course of Response to
Clozapine and Conventional Antipsychotic
Medication in Refractory Schizophrenia
by Robert Rosenheck, Denise Evans, Lawrence Herz, Joyce Cramer,
Weichun Xu, Jonathan Thomas, William Henderson, and Dennis Charney
This study compared the time course to clinical
improvement with clozapine and with conventional
antipsychotic medications. A double-blind trial compared clozapine and haloperidol in patients with schizophrenia who were refractory to conventional antipsychotic medication and were hospitalized for 30 to 364
days at 15 Veteran Affairs medical centers during the
year before study entry. Patients in the original study
were randomly assigned to haloperidol or clozapine
and followed for 12 months, at maximum tolerable
doses. Patients who completed a full year of treatment
with clozapine (n = 122), or with either haloperidol or
another conventional antipsychotic medication (n =
123) and who also completed the 9- or 12-month
assessment were included. Response to treatment was
defined as 20 percent improvement on standard scales
of symptoms and quality of life at the latter of the 9- or
12-month interviews. More patients assigned to clozapine achieved 20 percent improvement in symptoms at
each followup. Among patients who did not improve
at 6 weeks, 3 months, or 6 months, there were no significant differences between clozapine and comparison
patients in outcomes at 1 year. Among patients who
did improve, maintenance of that improvement also
did not differ between the groups at 1 year on symptom measures. Maintenance of improvement in quality
of life at 1 year was significantly greater for clozapine
patients who had improved at 6 months (p < 0.04).
Significant differential symptom response to clozapine
occurred exclusively during the first 6 weeks of treatment.
Key words: Schizophrenia, psychopharmacology,
clozapine.
Schizophrenia Bulletin, 25(4):709-719,1999.
cious than conventional pharmacotherapies in the treatment of refractory schizophrenia—with benefits in both
positive and negative symptoms and in quality of life
(Kane et al. 1988; Meltzer et al. 1990; Pickar et al. 1992;
Breier at al. 1993, 1994; Lieberman et al. 1994; Buchanan
1995; Carpenter et al. 1995; Rosenheck et al. 1997).
However, clozapine is expensive, and because of the associated risk of potentially fatal agranulocytosis, clozapine
treatment requires inconvenient weekly blood monitoring.
Treatment with clozapine costs approximately $4,500 per
year, plus an additional $1,000 for blood monitoring—11
times the typical annual cost of conventional antipsychotic medications (Meltzer and Cola 1994). Although
some studies have indicated that these higher costs can be
offset by reduced hospital use among high-cost patients
(Revicki et al. 1991; Meltzer et al. 1993; Reid et al. 1994;
Essock et al. 1996; Rosenheck et al. 1997), there is no
evidence that such savings can be realized among refractory outpatients with limited hospital utilization, and
potential savings in such patients are likely to be limited
(Rosenheck et al. 1993).
The cost-effectiveness of clozapine treatment in typical clinical practice would be vastly improved if it were
possible to rapidly and accurately identify clozapine
"responders" and differentiate them from "nonresponders." Such differentiation would allow treatment resources
for clozapine to be targeted precisely at those patients
who would gain the greatest benefit. While ongoing clinical assessment allows ready identification of patients who
have improved with treatment, it is not possible to determine whether the observed improvement is the result of
clozapine treatment, and therefore warrants continued
clozapine therapy, or would have' occurred with less
costly, conventional antipsychotic treatment. Although
clozapine is specifically indicated for refractory (i.e., nonresponsive) schizophrenia patients, virtually every con-
Controlled studies have shown that the atypical antipsychotic medication clozapine is significantly more effica-
Reprint requests should be sent to Dr. R. Rosenheck, VA Connecticut
Healthcare System, 950 Campbell Ave., West Haven, CT 06516-2770;
e-mail: [email protected].
Abstract
709
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
R. Rosenheck et al.
trolled study of clozapine has reported improvement
among some refractory patients treated with conventional
antipsychotic medications and nonimprovement among
some patients treated with clozapine. It is thus virtually
impossible in clinical practice to identify individual clozapine responders who would not have responded to less
costly medications. Further information is needed to guide
longitudinal clinical decisionmaking.
One approach to this problem is to identify a targeted
time course for response to clozapine. In such an approach patients who have not shown significant improvement after a specified time would be regarded as nonresponders, and their clozapine treatment would be
discontinued. The challenge is to identify the appropriate
timeframe for evaluating responsiveness. If the timeframe
is too short, some patients who would benefit from clozapine would be excluded from long-term treatment. If the
timeframe is too long, some nonresponsive patients will
receive sustained clozapine treatment without benefiting
from it. In addition to costing less, specifying a timeframe
would reduce the risk of agranulocytosis and seizures
associated with clozapine treatment.
Most of these studies addressing the time course have
demonstrated early response to clozapine, as soon as 1
week after initiating treatment (Pickar et al. 1992; Stern et
al. 1994), while others have reported new responses
occurring up to 1 year after treatment initiation (Wilson
1996). Lieberman et al. (1994) suggested that the optimal
trial was from 3 to 6 months, although Meltzer and colleagues (1990; Meltzer 1992) indicated a need for trials
lasting from 6 to 12 months. Our understanding of the
course of response to clozapine is further complicated by
an unreplicated study showing that in 25 percent of clozapine responders, the initial improvement on the drug is
not sustained by 1 year (Zito et al. 1993). A recent review
of published studies concluded that the evidence for late
response to clozapine is limited (Carpenter et al. 1995).
A major limitation of previous studies of time to
clozapine response is that they have not systematically
compared the time course of response to clozapine with
the time course of response to treatment with standard
antipsychotic medications in a sample of comparable
patients. Although there is substantial evidence that (1)
refractory schizophrenia patients are more likely to show
significant improvement on clozapine than on conventional antipsychotic medications, and (2) that response to
clozapine often occurs up to 6 months after treatment initiation, there is no evidence that a differential response to
clozapine eventually appears among patients who have
shown no initial response to treatment.
In this paper we use data from a 12-month study of
clozapine and haloperidol in a sample of patients with
710
refractory schizophrenia that compared treatment
response at 6 weeks, 3 months, 6 months, and 1 year. We
define treatment response as a 20 percent improvement on
either Positive and Negative Syndrome Scale (PANSS;
Kay et al. 1987) or the Heinrichs-Carpenter Quality of
Life Scale (Heinrichs et al. 1984). We further examine 1year medication responses among two subgroups of these
patients: (1) patients who had not responded to treatment
at various time points (i.e., time-specific nonresponders)
and (2) patients who had responded at the various time
points (i.e., time-specific responders). Through these
analyses we attempt to elucidate differential clozapine
versus haloperidol response rates as well as maintenance
rates among subgroups of treatment responders and nonresponders. We hope these data will provide useful information to guide longitudinal clinical decisionmaking.
Methods
Data for this study were obtained from a prospective, double-blind study, in which schizophrenia patients refractory
to antipsychotic medication were recruited over 2 years at
15 Veterans Affairs (VA) medical centers. Patients were
randomly assigned to clozapine or haloperidol and treated
for 12 months. The protocol was approved by a human
rights committee at the Hines VA Cooperative Studies
Program Coordinating Center and by human rights committees at each participating medical center. All patients
gave written informed consent.
Entry Criteria
Hospital use criteria. The study was targeted at a
subgroup of currently hospitalized treatment-refractory
schizophrenia patients with a history of high inpatient service use, defined as at least 30 days' hospitalization for
schizophrenia during the previous year, and no more than
364 days.
Clinical criteria. Clinical eligibility criteria modeled on those of Kane et al. (1988) included (1) a DSMIII-R (American Psychiatric Association 1987) diagnosis
of schizophrenia on the Semi-Structured Clinical Interview for Diagnosis (Spitzer and Endicott 1990); (2)
refractoriness defined as persisting psychotic symptoms
despite two documented adequate treatment trials; (3)
severe symptoms indicated by scores on the Brief
Psychiatric Rating Scale (Overall and Gorham 1962) and
the Clinical Global Impression Scale (Guy 1976); and (4)
serious social dysfunction for the previous 2 years.
Exclusion criteria. Patients were excluded if they
were unable to give informed consent or had a previous
trial of clozapine; had a current myeloproliferative disorder; or were pregnant.
Time Course of Clozapine Response
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
After providing written informed consent to participate in the study and completing baseline assessments,
patients were randomly assigned, within centers, to a
treatment group.
of efficacy or adverse effects and switched to other treatments. Thus some clozapine patients received standard
antipsychotic medications (including haloperidol), and
some haloperidol patients received clozapine. Altogether,
83 (40%) of 205 patients assigned to clozapine discontinued it by the 48th week of the trial and crossed to a standard antipsychotic medication (including haloperidol).
Although 157 (72%) of the 218 haloperidol patients were
unblinded, only 49 (22%) received clozapine treatment
for 4 or more weeks during the trial; they were excluded
from the analyses.
The inclusion of these "crossover" patients in our
analyses would have distorted our evaluation of the timing of medication response, so all the analyses presented
here were conducted with crossover cases excluded.
Patients in the clozapine group were treated with clozapine for at least 48 weeks, while patients in the control
group received conventional antipsychotic medications
for the entire period.
Analyses were conducted to determine how the
crossover patients differed from those who did not cross
over. Examination of 25 sociodemographic and clinical
status measures showed that crossover patients were not
significantly different from those who did not cross over
on any baseline clinical or demographic measure, except
that they had better quality of life on the HeinrichsCarpenter Quality of Life Scale (p = 0.01). However,
within the subgroup of patients who did not cross over,
there were no significant differences between treatment
groups at baseline. Thus, although the original study was
based on random assignment, the subsamples of each
treatment group examined in this study were not randomly
assigned. Although no significant differences were noted
between treatment groups, the analyses presented here
must be regarded as coming from a nonequivalent control
group design since group assignment was not completely
random. It should also be noted again that only 61 of the
haloperidol patients were blind throughout the trial.
Pharmacotherapy. Random assignment was made to
double-blind treatment with clozapine (100-900 mg/day)
or haloperidol (5-30 mg/day). Dose adjustments were
made as clinically indicated, using 12 fixed dosage levels.
Haloperidol-treated patients also received benztropine
mesylate (2-10 mg/day) for extrapyramidal side effects
while clozapine patients received a matching benztropine
placebo. Haloperidol patients participated in weekly
blood counts as required for clozapine treatment.
Psychosodal Treatment. To assess the potential effectiveness of clozapine in typical clinical practice, a predefined program of adjunctive psychotherapeutic and rehabilitative treatments was offered through a structured
treatment planning module based on a comprehensive
menu of locally available services. This system ensured
that all patients were encouraged to use psychosocial services that could be of therapeutic value to them.
Assessment of Outcomes and Reliability of Ratings.
Symptom outcomes were assessed with the structured
clinical.interview of the PANSS (Kay et al. 1987)
(range = 30-210, with higher scores reflecting more
severe symptoms). Social functioning and quality of life
were evaluated with the Heinrichs-Carpenter Quality of
Life Scale, a clinician-rated scale of social functioning
and severe behavioral deficits (Heinrichs et al. 1984)
(range = 0-126 with higher scores reflecting higher functioning). Interviews were conducted at 6 weeks and at 3,
6, 9, and 12 months after random assignment. Outcome
data for the present study were based on the latter of the
9- and 12-month interviews, and patients for whom these
interviews were missing were excluded from the analyses.
Improvement of 20 percent on these scales was regarded
as a clinically significant response.
All interviewers were formally trained and received
annual reassessment of interrater reliability using ratings
from videotaped demonstration interviews. A statistical
method was developed to assess the reliability of multiple
assessments in a single patient or taped interview
(Cicchetti et al. 1997). Reliability across interviewers (n =
39), as assessed by this method, was excellent for both the
PANSS (96% agreement; range = 94-96% across subscales) and the Heinrichs-Carpenter Quality of Life Scale
(95% agreement; range = 90-98% across subscales).
Methods of Analysis. Chi-square tests were used to
determine the significance of differences between groups
in the proportion who had achieved 20 percent improvement at each time point. Further analyses, limited to 1year outcomes among patients who had improved or had
not improved at each assessment point, also used the chisquare test of statistical significance.
Results
Sample Characteristics. Table 1 presents sociodemographic and clinical characteristics of the sample. There
were no significant differences between the two groups on
any measure. During the first half of the trial (weeks 1-
Crossovers. During the 12 months of followup, some
patients stopped taking study medication because of lack
711
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
R. Rosenheck et al.
Table 1. Comparison of treatment groups at baseline1
Age, mean (SD)
Gender, n (%)
Male
Female
Clozapine
Haloperidol
Chi-square
or t test
43.31 (7.1)
44.46 (8.4)
1.14
P
0.25
121
1
(99.2)
(0.8)
122
1
(99.1)
(0.8)
0.00
1.00
Race, n (%)
White
Black
Hispanic
Other
90
26
5
1
(73.8)
(21.3)
(4.1)
(0.8)
82
39
2
0
(66.7)
(31.7)
(1.6)
(0.0)
5.25
0.15
Marital Status, n (%)
Married
Never married
Separated/Divorced
Widowed
14
71
32
5
(11.5)
(58,2)
(26.2)
(4.1)
4
66
41
11
(4.1)
(53.7)
(33.3)
(8.9)
7.89
0.10
119
(97.5)
119
(96.8)
0.1
0.71
Receives disability (VA or SSA), n (%)
Past-Month Comorbidity, n (%)
Alcohol abuse
Drug abuse
9
8
(7.4)
(6.6)
10
9
(8.1)
(7-3) .
0.0
0.1
0.83
0.82
Age at, mean (SD)
Onset of schizophrenia
First hospitalization
22.1
23.4
(4.7)
(4.8)
22.6
23.8
(4.6)
(5.3)
0.78
0.52
0.44
0.60
Heinrichs-Carpenter Total, mean (SD)
42.0
(17.5)
38.6
(16.1)
1.54
0.12
PANSS Total
90.8
(13.8)
90.5
(13.3)
0.17
0.86
Global Assessment Scale
35.0
(8.2)
36.2
(7.1)
1.18
0.24
2.4
(2.5)
2.4
(2.6)
0.04
0.97
AIMS (TD), mean (SD)
5.2 (4.2)
0.29
0.77
Simpson Angus (EPS)
5.0 (4.8)
Note.—SD = standard deviation; VA = Veterans Affairs; SSA = Social Security Administration; PANSS = Positive and Negative Syndrome
Scale; AIMS (TD) = Abnormal Involuntary Movement Scale (Tardive Dyskinesia); EPS = extrapyramidal symptoms.
1
No group differences reached statistical significance.
26), average doses were 431 mg/day for clozapine and
22.9 mg/day for haloperidol. During the second half of
the trial (weeks 27-53), average doses were 628 mg/day
for clozapine and 28.2 mg/day for haloperidol.
symptoms at 6 weeks (X2 = 11.1, df = \,p < 0.001) and at
6 months (X2 = 9.87, df=l,p0.002). Trends suggested
higher percentages of improvement among clozapine
patients at 3 months (X2 = 2.83, df= \,p = 0.08) and at 1
year (X2 = 3.18, df= l,/? = 0.08).
On the Quality of Life Scale (figure 2), improvement
in the clozapine group was also steady with 29 percent
and 43 percent showing improvement at 6 weeks and 3
months, and smaller increases thereafter. The comparison
group showed the same likelihood of improvement in
quality of life at 6 weeks, but smaller increases thereafter.
Statistical analysis showed trends suggesting significant
differences between the clozapine and haloperidol groups
at 6 months (X2 = 2.51, df= 1, p = 0.11) and 1 year (X2 =
4.22, df=\,p = 0.04).
Improvement Over Time: Entire Sample. Figure 1
shows that rates of improvement among patients assigned
to clozapine increased steadily, but the incremental
increase from one time point to the next was far greater
during the first 6 weeks (29%) than in subsequent time
periods (8%, 3%, and 1% respectively). Improvement in
the group treated with haloperidol was also substantial but
occurred more slowly.
A significantly higher percentage of patients assigned
to clozapine had achieved a 20 percent improvement in
712
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
Time Course of Clozapine Response
Figure 1. Proportion of patients who achieved 20 percent improvement over baseline level on symptoms of schizophrenia measured on the PANSS. Statistical significance determined by chi-square
tests
H Clozapine
• Haloperidol
:**
12*
26**
Weeks
*p<0.10;**p<0.01
Attaining Improvement Among Time-Specific Nonresponders. Table 2 compares 1-year outcome responses
among specific subgroups who did not show a 20 percent
response to treatment at each time point during the trial.
The first comparison, using the PANSS, shows that
among patients whose symptoms had not responded to
treatment by 6 weeks, 27.1 percent of clozapine patients
and virtually the same percentage of haloperidol patients
(25.7%) responded by 1 year. Results were quite similar
for each group of time-specific nonresponders, with an
average response rate of 25 percent of clozapine patients
and 21 percent of control patients.
Results for change in quality of life are also generally
similar across nonresponding groups, although among 6week nonresponders, a marginally significant difference is
observed at 1 year with 42 percent of clozapine patients
responding by 1 year versus 29 percent of those assigned
to haloperidol (X2 = 2.97, df= l,p = 0.09). There were no
significant differences between treatment groups in 1-year
quality of life outcomes among 3-month and 6-month
nonresponders.
Maintaining Improvement Among Responders. Table
3 presents data on 1-year symptom and quality of life status among patients who had previously shown 20 percent
improvement, thus reflecting the maintenance of gains
through sustained treatment.
There were no significant differences between treatment groups on maintenance of symptom improvement
(PANSS), regardless of when their improvement was
observed. Across all groups, an average of 69 percent of
clozapine patients and 68 percent of comparison patients
maintained their 20 percent improvement in symptoms.
On the quality of life measure, in contrast, an average
of 83 percent of clozapine patients who had improved at
each time point maintained their improvement at 1 year,
compared with an average of 69 percent of comparison
patients. Groups were differentiated by nonsignificant
trends (p < 0.10) at 3 months and by a significant difference at 6 months (p < 0.05).
Medication Dose Levels. Comparison of average daily
clozapine doses during the first 6 months of the study
showed no significant difference between those who
713
R. Rosenheck et al.
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
Figure 2. Proportion of patients who achieved 20 percent improvement over baseline level on the
Heinrichs-Carpenter Quality of Life Scale. Statistical significance determined by chi-square tests
Clozapine
Haloperidol
12
36*
Weeks
*p<0.10
Table 2. Achievement of 20 percent improvement at final assessment among patients who had not
achieved 20 percent improvement at earlier assessments (crossovers excluded)
H-C Quality of Life Scale
PANSS
Improved at 1 yr.
Improved at 1 yr.
Not improved at
n
%
Not improved at
n
%
6 weeks
Clozapine, n = 85
23
27.1
6 weeks
Clozapine, n = 84
35
41.7
Haloperidol, n= 105
27
25.7
Haloperidol, n = 83
24
28.9
3 months
Clozapine, n = 68
23
33.8
Haloperidol, n = 80
21
26.3
2
2
X = 0.04, d f = 1 , p = 0.84
X = 2.97, cff=1,p=0.09
3 months
Clozapine, n = 77
18
23.4
Haloperidol, n = 87
17
19.5
2
X = 0.36, df=\,p
X2 = 1.01, d f = 1 , p = 0.32
= 0.55
6 months
Clozapine, n = 73
18
24.7
6 months
Clozapine, n = 67
17
25.4
Haloperidol, n = 94
17
18.1
Haloperidol, n = 75
16
21.3
2
2
X = 0.38, c # = 1 , p = 0 . 5 4
X = 1.07, d f = 1 , p = 0.30
Note.—PANSS = Positive and Negative Syndrome Scale; H-C = Heinrichs-Carpenter.
714
Time Course of Clozapine Response
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
Table 3. Maintenance of improvement at time of final assessment (9 months or 1 year; crossovers
excluded)
PANSS
H-C Quality of Life Scale
Improved at 1 yr.
Improved at
n
%
6 weeks
Clozapine, n = 35
25
71.4
Haloperidol, n= 14
10
71.4
Improved at 1 yr.
Improved at
n
%
6 weeks
Clozapine, n = 36
29
80.6
Haloperidol, n = 34
23
67.6
X2 = 2.15, < # = 1 , p = 0.14
2
X = 0.00, ctf=1,p=1.00
3 months
Clozapine, n = 44
31
70.5
3 months
Clozapine, n = 51
41
80.4
Haloperidol, n = 31
19
61.3
Haolperidol, n = 37
25
67.6
2
2
X = 1.88, c # = 1 , p = 0.17
X = 0.69, df=\, p = 0.41
6 months
Clozapine, n = 48
32
66.7
6 months
Clozapine, n = 55
48
87.3
Haloperidol, n = 25
18
72.0
Haloperidol, n = 41
29
70.7
2
2
X = 4.05, cff=1,p=0.04
X = 0.22, df = 1 , p = 0.64
/Vote.—PANSS = Positive and Negative Syndrome Scale; H-C = Heinrichs-Carpenter.
showed 20 percent improvement on the PANSS at 6
months (mean average dose = 413 mg/day, SD =161) and
those who did not (mean average dose = 439 mg/day,
SD = 156) (t = 0.91, df= 98.3, p = 0.36). There was also
no difference between those who showed 20 percent
improvement on the Quality of Life Scale at 6 months
(mean average dose = 427 mg/day, SD = 159) and those
who did not (mean average dose = 431 mg/day, SD = 157)
(t = 1.17; df= 47.7, p = 0.24). Significant dosage differences were also absent between haloperidol improvers
and nonimprovers and between groups sorted by response
type during the last 6 months of the study.
Discussion
This study compared time course of response to treatment
among comparable samples of refractory schizophrenia
patients who were treated for a full year with either clozapine or conventional antipsychotic medications. Patients
selected for these analyses were those who had not
crossed over to the other treatment condition during a randomized clinical trial, and who had completed a formal
clinical assessment at either 9 or 12 months. Although the
study was derived from a randomized clinical trial, since
crossover cases were excluded, group assignment was not
completely random and only 61 control cases remained
blinded throughout the trial.
The largest proportion of patients who obtained substantial clinical improvement in both the clozapine and
conventional antipsychotic groups did so during the first 6
715
weeks of treatment. While differences between treatment
groups in the proportion of responders was greatest on the
measure of symptoms at 6 weeks, group differences in
quality of life were greatest at 1 year. Thus while clinical
improvement in both symptoms and quality of life
occurred most frequently in both treatment groups, during
the first 6 weeks of the trial, differences between the
groups, which invariably favored the clozapine group,
emerged at different times—early in the trial for symptom
measures, but late in the trial in the case of quality of life
improvements. These findings were not explained by differences in medication dosage.
Rates of overall improvement are based on achieving
significant levels of improvement when it has not previously been achieved, and maintaining improvement once
it has been achieved. A particular strength of this study is
that it tracked the time course of both of these processes
for both the clozapine group and the comparison group.
Examination of symptom outcomes showed significant
differences in improvement during the first 6 weeks of the
trial, but no significant differences in either achieving or
retaining improvement after the first 6 weeks of treatment.
Thus both absolute and relative rates of symptom
improvement are concentrated during the first 6 weeks of
treatment.
In contrast, differences in achieving quality of life
outcomes at 1 year were marginally significantly different
among patients who had not achieved improvement by 3
months, although not among subsequent groups of nonresponders. Patients treated with clozapine, however, were
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
R. Rosenheck et al.
more likely to maintain quality of life improvements
achieved during the first 6 months of the trial through to
the 1-year interview. This study thus replicated the findings of Zito et al. (1993). The probability of retaining
improvement is what distinguishes clozapine's long-term
impact on quality of life. As with symptom improvement,
long-term clozapine responders achieve their improvement early in the trial, suggesting that decisions about
continuing treatment with clozapine can be made on the
basis of the early response. The differential benefit of
clozapine largely results from the greater likelihood of
retaining that improvement among those who responded.
the current data suggest that patients who have not
responded to clozapine by 6 weeks are not likely to have a
better response to clozapine as compared with conventional antipsychotic medications over the remainder of the
year of treatment. Patients who have responded to clozapine, in contrast, were more likely to maintain their gains,
at least in the domain of social functioning, with continued clozapine treatment.
Third, this study considered only outcomes in symptoms and quality of life. Clozapine has a dramatically
superior extrapyramidal symptom profile compared with
standard antipsychotic medications at all time points
(Buchanan 1995; Rosenheck et al. 1997). For some
patients, this may be reason enough to use clozapine as a
medication of choice (Kane et al. 1988; Lieberman et al.
1994), although it also adds a small risk of agranulocytosis. In addition, by reducing hospital use clozapine has significant potential cost savings (see below). Furthermore,
since the PANSS measures factors other than psychotic
symptoms, it is possible that clozapine response in some
individuals was diluted by changes in these other areas.
Before we draw conclusions for clinical practice from
these data, several potential methodological limitations
must be noted. First, the early responders may have been
those with the best prognoses at baseline. The similar
rates of incremental improvement between groups at later
time points may be due to the fact that clozapine patients
who had not responded by 6 weeks were sicker at baseline
than haloperidol nonresponders and therefore had poorer
prognoses. To evaluate this possibility we conducted an
additional analysis in which we compared 6-week nonresponders on all baseline measures listed in table 1. These
analyses showed no significant differences between the 6week symptom and quality of life nonresponder groups on
any of the sociodemographic or clinical variables.
Finally, our inclusion of patients who were refractory
to conventional antipsychotic medication (many of whom
probably had failed previous trials of haloperidol) but not
patients with a previous trial of clozapine may have
biased the results in favor of clozapine to some degree.
Given the limited use of clozapine at the time this study
was conducted, this bias was unavoidable.
If the findings reported here are supported by subsequent studies, clozapine pharmacotherapy could be provided through standardized 6-week trials accompanied by
systematic outcome monitoring with standardized instruments. Responders at 6 weeks would be continued on
clozapine, while nonresponders could be switched to a
standard neuroleptic. Table 4 presents cost data using
methods from the original study (Rosenheck et al. 1997)
Second, the lack of subsequent differential response
to clozapine and conventional antipsychotic medications
among 6-week treatment nonresponders does not unambiguously demonstrate that clozapine patients would have
responded just as positively if they had been treated with
conventional antipsychotic medications. A study design
that re-randomized treatment nonresponders in both
groups at 6 weeks to either haloperidol or clozapine
would better address this question, but this would require
a much larger sample size. In the absence of such a study,
Table 4. Health care cost impact of 6-week trial strategy of clozapine pharmacotherapy, assuming
29 percent clozapine response rate at 6 weeks, $
Full Year of Treatment
6-Week Clozapine Trial
Difference
1
Clozapine2
Haloperidol
6,444
1,798
4,646
3,284
74
3,210
Difference 1
Clozapine
Haloperidol
Outpatient treatment
Pharmacotherapy
Outpatient services3'4
11,779
4,709
7,070
3,284
74
3,210
8,495
4,635
3,860
Inpatient treatment
47,835
52,895
(5,060)
47,835
52,895
(5,060)
Total health care costs
59,614
56,179
3,435
54,279
56,179
(1,900)
1
3,160
1,724
1,436
Average cost associated with clozapine treatment less cost of haloperidol treatment.
Estimate based on assumption that nonresponders (71% of the sample) would shift to haloperidol treatment after 6 weeks of treatment
with clozapine.
3
Outpatient services include professional services and laboratory visits including those required by the study protocol.
4
Costs for the haloperidol group have been adjusted for services exclusively required for the study protocol to 73 percent of their actual
value.
2
716
Time Course of Clozapine Response
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
that illustrate the economic consequences of such a strategy. In the intention-to-treat analysis in the original study,
health care costs associated with clozapine were nonsignificantly lower than haloperidol costs by $2,441
because of reductions in hospital utilization. In this study
sample, average 1-year health care costs were $3,435
higher because crossover cases, who took clozapine for
only part of the year, were excluded. Outpatient costs
(including medication and laboratory costs) associated
with clozapine pharmacotherapy were $8,495 greater than
costs associated with haloperidol treatment, and inpatient
savings of $5,060 were not sufficient to offset this greater
cost. However, when costs are reestimated assuming that
(1) there is no loss of drug efficacy if 6-week nonresponders are changed to standard antipsychotic medication, and
(2) 71 percent of patients do not respond to clozapine
within 6 weeks (as was found in this study), then we
would expect outpatient clozapine therapy to cost only
$3,160 more. Total health care costs would then be $1,900
lower with clozapine than with haloperidol. By lowering
the cost of treatment, this strategy would support clozapine trials with increased numbers of patients and would
allow more efficient targeting of the drug.
Young Kwon, M.D., Carl Faust, Kristin Brown, Marty
Manuel (Brecksville); Sidney Chang, M.D. (PI), Joanne
D. Wojcik, R.N., Carla D. Kohberger (Brockton);
Lawrence Dunn, M.D. (PI), Mara Evans, R.N., Al Bush,
Scott Bennett (Durham); John C. Crayton, M.D. (PI),
Kathleen F. Foley, R.N., Shonagh Neafsey, Sarah
Sideman (Hines); William B. Lawson, M.D., Ph.D. (PI),
Pat Arnold, R.N., Michael Edmison, R.N., Marilyn M.
Storey (Little Rock); Yeon Choe, M.D. (PI), Doreen
Broad, R.N., Eileen Waterbury, Elizabeth Cabezon
(Lyons); Richard Douyon, M.D. (PI), Marci Miller, R.N.,
Janet Hamel, Sean Kerr, Denice Feenane (Miami);
Edward Allen, M.D. (PI), Kelly Wainwright, R.N.,
Deborah Widmer (Montrose); John Lauriello, M.D. (PI),
Sheryl Whistance, R.N., Jeananne Breen, Maria Getty
(Palo Alto); Michael Peszke, M.D. (PI), Marcia
Castiglione, R.N., James Frock, Ph.D. (Perry Point);
Jeffrey L. Peters, M.D. (PI), Nadine Snyder, R.N.,
Dorothy Stefanik, Karen Henry (Pittsburgh); Janet Tekell,
M.D. (PI), Brenda Tobey, R.N., Deborah Hall, Edna Kyle,
R.N. (San Antonio); and Joseph Erdos, M.D., Ph.D. (PI),
Deborah Miles, R.N., Alicia Genovese, Dorothy Soliwoda
(West Haven).
We are indebted to the Data Monitoring Board (Alan
Breier, M.D., Howard Goldman, M.D., James Klett,
Ph.D., and David Pickar, M.D.) and the Executive
Committee (Boris Astrachan, M.D., John Crayton, M.D.,
Linda Frisman, Ph.D., Carol Fye, R.Ph., M.S., William
Hargreaves, Ph.D., and William Lawson, M.D.) for their
careful overview of the progress of the trial.
Conclusion
Data on the time course of treatment response to clozapine presented here are substantially more informative
than those of previous studies because data on a parallel
comparison group treated with conventional antipsychotic
medications are included. Because of the methodological
limitations of the trial design these data are not strong
enough to obviate the need for clozapine trials in ordinary
practice of greater duration than 6 weeks. They do suggest
that incremental benefits in symptoms and quality of life,
beyond those achievable with conventional medications,
are most likely to be realized among patients who respond
to clozapine during the first 6 weeks of treatment. If differential response to clozapine can be determined from a
6-week trial, cost savings could be substantial, and greater
numbers of patients could have access to the drug.
References
American Psychiatric Association.
DSM-III-R:
Diagnostic and Statistical Manual of Mental Disorders.
3rd ed., revised. Washington, DC: The Association, 1987.
Breier, A.; Buchanan, R.W.; Irish, D.; and Carpenter, W.T.
Clozapine treatment of outpatients with schizophrenia:
Outcome and long-term response patterns. Hospital and
Community Psychiatry, 44:1145-1149, 1993.
Breier, A.; Buchanan, R.W.; and Kirkpatrick, B.
Clozapine in schizophrenic outpatients: Effects on positive and negative symptoms. American Journal of
Psychiatry, 151:20-26, 1994.
Appendix
The following were responsible for the conduct of the
study at their respective VA facilities: John Grabowski,
M.D. (PI), Lawrence Alphs, M.D., Albert Pizzuti, R.N.,
Robert Wancha (Allen Park); Denise Evans, M.D. (PI),
Patsy Y. Puczkowski, R.N., James Martin, Mike
Brandsma (Augusta); Lawrence Herz, M.D. (PI), Marion
Avtges, R.N., Roberta Smith, Ted Phillips, Jennifer Di
Vicenzo (Bedford); George Jurjus, M.D. (PI), Kong
Buchanan, R.W. Clozapine: Efficacy and safety. Schizophrenia Bulletin, 21:579-591, 1995.
Carpenter, W.T.; Conley, R.R.; Buchanan, R.W.; Breier,
A.; and Tamminga, C.A. Patient response and resource
management: Another view of clozapine treatment of
schizophrenia. American Journal of Psychiatry,
152:827-832, 1995.
717
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
R. Rosenheck et al.
Cicchetti, D.; Showalter, D.; and Rosenheck, R. A new
method for assessing interexaminer agreement when multiple ratings are made on a single subject: Applications to
the assessment of neuropsychiatric symptomatology.
Psychiatry Research, 72:51-63, 1997.
Essock, S.M.; Hargreaves, W.A.; Covell, N.H.; and
Goethe, J. Clozapine's effectiveness for patients in state
hospitals: Results from a randomized trial. Psychoparmacology Bulletin, 32:683-697, 1996.
Guy, W. Clinical Global Impression. In: Guy, W., ed.
ECDEU Assessment Manual for Psychopharmacology
(DHEW No. ADM 76-338). Rockville, MD: National
Institute of Mental Health, 1976.
Heinrichs, D.W.; Hanlon, E.T.; and Carpenter, W.T. The
Quality of Life Scale: An instrument for rating the schizophrenic deficit syndrome. Schizophrenia
Bulletin,
10:388-398, 1984.
Reid, W.H.; Mason, J.; and Toprac, M. Savings in hospital
bed-days related to treatment with clozapine. Hospital
and Community Psychiatry, 45:261-268, 1994.
Revicki, D.A.; Luce, B.R.; Weschler, J.M.; Browen, R.E.;
and Adler, M.A. Cost-effectiveness of clozapine for treatment-resistant schizophrenic patients. Hospital and
Community Psychiatry, 41:850-854, 1991.
Rosenheck, R.A.; Cramer, J.; Xu, W.; Thomas, J.;
Henderson, W.; Frisman, L.K.; Fye, C ; and Charney, D.,
for the Department of Veterans Affairs Cooperative Study
Group on Clozapine in Refractory Schizophrenia. A comparison of clozapine and haloperidol in the treatment of
hospitalized patients with refractory schizophrenia. New
England Journal of Medicine, 337:809-815, 1997.
Rosenheck, R.A.; Massari, L.; and Frisman, L. Who
should receive high cost mental health treatment and for
how long? Issues in the rationing of mental health care.
Schizophrenia Bulletin, 19:843-852,1993.
Kane, J.M.; Honigfeld, G.; Singer, J.; Meltzer, H.Y.; and
the Clozaril Collaborative Study Group. Clozapine for the
treatment-resistant schizophrenic: A double-blind comparison with chlorpromazine. Archives of General
Psychiatry, 45:789-796, 1988.
Spitzer, R.S., and Endicott, J.E. The Semi-Structured
Clinical Interview for Diagnosis. Washington, DC:
American Psychiatric Press, 1990.
Stern, R.G.; Kahn, R.S.; Davidson, M.; Nora, R.M.; and
Davis, K.L. Early response to clozapine in schizophrenia.
American Journal of Psychiatry, 151:1817-1818, 1994.
Kay, S.R.; Fiszbein, A.; and Opler, L.A. The Positive and
Negative Syndrome Scale (PANSS) for schizophrenia.
Schizophrenia Bulletin, 13:261-276, 1987.
Wilson, W.H. Time required for initial improvement during clozapine treatment of refractory schizophrenia.
American Journal of Psychiatry, 153:951-952, 1996.
Lieberman, J.A.; Safferman, A.Z.; and Pollack, S. Clinical
effects of clozapine in chronic schizophrenia: Response to
treatment and predictors of outcome. American Journal of
Psychiatry, 151:1744-1752, 1994.
Zito, J.M.; Volavka, J.; Craig, T.J.; Czabor, P.; Banks, S.;
and Vitrai, J. Pharmacoepidemiology of clozapine in 202
inpatients with schizophrenia. Annals of Pharmacotherapy, 27:1262-1269, 1993.
Meltzer, H.Y. Treatment of the neuroleptic-nonresponsive
schizophrenic patient. Schizophrenia
Bulletin,
18:515-542, 1992.
Meltzer, H.Y.; Burnett, S.; Bastani, B.; and Ramirez, L.F.
Effects of six months of clozapine treatment on the quality of life of chronic schizophrenic patients. Hospital and
Community Psychiatry, 41:892-897, 1990.
Acknowledgments
This work was supported by the Department of Veterans
Affairs Health Services Research & Development
Service. Clozapine, benztropine, and matching haloperidol and placebo benztropine were generously provided by
Sandoz Pharmaceutical Corporation.
We would like to thank the Department of Veterans
Affairs Research Office for their support (John Feussner,
M.D., Daniel Deykin, M.D., Shirley Meehan, Ph.D.,
Charles Welch, Ph.D., Joseph Gough, Janet Gold, and
Ping Huang, Ph.D.) and also Lois Ucas of the Chairman's
Office; Amy Smith, Jeff Parker, Michael Kelley, and
Shenglin Wang of the Cooperative Studies in Health
Services Coordinating Center; Carol Fye, R.Ph., M.S.,
Bill Gagne, Loretta Guidarelli, and Mike Sather of the
Pharmacy Coordinating Center; and David Garver, M.D.,
Gary Ripper, and Stephanie Todd of the National VA
Clozapine Coordinating Center.
Meltzer, H.Y., and Cola, P.A. The pharmacoeconomics of
clozapine: A review. Journal of Clinical Psychiatry, 55
(SupplB): 161-165, 1994.
Meltzer, H.Y.; Cola, P.; Way, L.; Thompson, P.A.; Bastani,
B.; Davies, M.A.; and Snitz, B. Cost-effectiveness of
clozapine in neuroleptic resistant schizophrenia. American
Journal of Psychiatry, 150:1630-1638, 1993.
Overall, J.E., and Gorham, D.R. The Brief Psychiatric
Rating Scale. Psychological Reports, 10:799-812, 1962.
Pickar, D.; Ownen, R.R.; Litman, R.E.; Konicki, E.;
Gutierrez, R.; and Rapaport, M.H. Clinical and biologic
response to clozapine in patients with schizophrenia:
Crossover comparison with fluphenazine. Archives of
General Psychiatry, 49:345-353, 1992.
718
Time Course of Clozapine Response
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
Team 1 at the Augusta VA Medical Center, Augusta, GA,
and Assistant Professor of Psychiatry at the Medical
College of Georgia. Lawrence Herz, M.D., is Chief
General Psychiatry sub-service line, Edith Norse Rogers
Memorial Veterans Hospital, Bedford, MA, and Assistant
Professor of Psychiatry, Boston University School of
Medicine. Weichun Xu, Ph.D., is Biostatistician; Jonathan
Thomas is Research Scientist; and William Henderson,
Ph.D., is Director of the Cooperative Studies
Coordinating Center, Hines VA Medical Center, Hines,
IL.
The Authors
Robert Rosenheck, M.D., is Director of VA's Northeast
Evaluation Center and Professor of Psychiatry and Public
Health; Joyce Cramer is Health Services Researcher and
Lecturer in Psychiatry; and Dennis Chamey, M.D., is former Chief of Psychiatry and Professor of Psychiatry and
Vice Chairman for Academic Affairs, VA Connecticut
Healthcare System, West Haven, CT, and Yale School of
Medicine Department of Psychiatry, New Haven, CT.
Denise Evans, M.D., is Director of General Psychiatry
719
Schizophrenia Bulletin, Vol. 25, No. 4, 1999
PARIS 2000:
PSYCHOSOCIAL REHABILITATION
The 7th World Congress of the World Association for Psychosocial
Rehabilitation, titled "Psychosocial Rehabilitation: Promoting
Diversity and Ensuring Equality/' will convene May 7-10, 2000, in
Paris, France, at La Villette, Cite des Sciences et de l'lndustrie. The
congress will focus on the following questions:
•
What are the different approaches to rehabilitation
throughout the world?
•
How do these methods alleviate suffering and contribute
to recovery?
•
How can they improve integration in the areas of work,
housing, social relations, and access to culture?
•
How do these methods work together with biological
and psychotherapeutic treatments?
•
How can users, professionals, families, and policymakers
overcome the division between those seen as "contributors" to society and those viewed as a "burden"?
•
What interventions and services can users provide to
help individuals in the areas of work, housing, and social
relations?
For information On the Congress, including registration and
books of abstracts, please contact:
Comite d'Organisation Paris 2000
R. Onteniente
NHA Communication
3, rue de la Boetie
75008 Paris, France
720
Tel: 01-42-66-46-46
Fax: 01-42-66-45-45
E-mail: [email protected]
Web: http://www.rehabilite.org