Download Depressive Factors and Their Relationships With

Depressive Factors and Their Relationships
With Other Symptom Domains in
Schizophrenia, Schizoaffective Disorder,
and Psychotic Depression
by M.atihias J. Miiller, Armin Szegedi, Hermann Wetzel, and Otto Benkert
Abstract
chotic features. The fuzzy boundaries between diagnostic
groups are partly due to the nosological nonspecificity of
frequently coexisting positive, negative, and affective
symptoms (Mundt et al. 1989; Gerbaldo et al. 1995;
Maziade et al. 1995; Toomey et al. 1998). Within nosological groups, the overlap of certain symptom domains
(e.g., depressive and negative symptoms) additionally
obscures the picture (Prosser et al. 1987; Whiteford et al.
1987; Siris et al. 1988; Lewine 1990; Barnes and
McPhillips 1995; Collins et al. 1996). To disentangle
these sometimes confounded symptoms (Kitamura and
Suga 1991) and to investigate the mutual relationship of
empirically derived symptom domains could help establish meaningful diagnostic categories and differential
treatment implications (Sax et al. 1996).
As with negative symptoms, which have been disregarded for a very long time, the negligence of affective
symptoms in schizophrenia, foremost depression, in
schizophrenia is apparently over-simplistic. There is
growing evidence suggesting that depressive symptoms
and related mood disturbances are an integral part of
schizophrenic disorders (Hirsch 1982; Siris 1995).
Several previous studies have investigated the relationship
between negative and depressive symptoms, but most of
them treated depression as a homogenous construct. A
sophisticated analysis in patients not restricted to the diagnosis of schizophrenia suggested that depressive and negative symptoms are frequently strongly associated but
should nevertheless be regarded as distinct concepts
(Kitamura and Suga 1991). However, akinetic symptoms
due to treatment with conventional neuroleptics were not
taken into account in this study, despite the possible overlap of akinesia with either depressive or negative symptomatology (Van Putten and May 1978; Harrow et al. 1994).
When considering the relationship between depressive
and positive symptoms in psychotic disorders, only a few
Relationships among different symptom domains were
investigated in patients with acute exacerbation of
schizophrenia with depressive symptoms, psychotic
depression, or schizoaffective disorder, depressive subtype. Scores for depression and depressive factors
were correlated with positive, negative, and extrapyramidal symptoms within diagnostic categories. No
between-group differences in the relationship of different symptom domains could be found, and no substantial relationship between depression and positive
symptoms could be revealed in any diagnostic subgroup. Only the retardation factor of depression
showed a significant overlap with negative symptoms;
depressive core symptoms did not. Core symptoms of
depression were independent from other symptoms in
all investigated diagnostic groups. Depression seems to
represent a heterogeneous symptom domain with
unique relationships of components to positive and
negative symptoms across nosological borders. A more
differentiated assessment, analysis, and treatment of
depressive symptoms is therefore recommended for
patients with combined depressive and psychotic
symptoms.
Keywords: Schizophrenia, major depression,
schizoaffective disorder, symptom domains.
Schizophrenia Bulletin, 27(l):19-28, 2001.
The heterogeneity of symptomatology within the group of
schizophrenias is still a major obstacle for defining clinically useful subgroups of these disorders. Furthermore,
the problem expands to the nosological borders of schizophrenia on the one hand and other major psychiatric disorders on the other. It is not only difficult to define subgroups of schizophrenia; it is also comparably
troublesome to find a clear distinction between the group
of schizophrenic disorders, the depressive subtype of
schizoaffective disorder, and major depression with psy-
Send reprint requests to Dr. M.J. Miiller, Dept. of Psychiatry,
University of Mainz, Untere Zahlbacher Str. 8, D-55131 Mainz,
Germany; e-mail: [email protected].
19
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
M.J. Miilleretal.
empirical reports, with inconclusive findings, are available
so far. A negative correlation was found in patients with
acute schizophrenia (Dollfus et al. 1993), whereas a positive relationship between positive and depressive symptoms was found during the chronic course of schizophrenia
(Barnes et al. 1989; Norman and Malla 1994). Sax et al.
(1996) investigated patients with either schizophrenia and
substantial concurrent depression (« = 17) or psychotic
depression (n = 25). They revealed an association of
depressive and positive symptoms as measured by the
Hamilton Depression Rating Scale (HAMD) (Hamilton
1960) and the Scale for the Assessment of Positive
Symptoms (SAPS, Andreasen 1982) in only patients with
schizophrenia, not patients with psychotic depression. The
authors suggested that different levels of depression severity could account for this difference in correlations.
However, it has been previously shown (Goldman 1992)
that depression as assessed by the HAMD cannot be
assumed to be homogenous in psychotic patients.
Furthermore, the HAMD sum score might not be an appropriate index of depression severity (Gibbons et al. 1993).
Our confirmatory factor analysis (CFA) studies using
the Bech-Rafaelsen Melancholia Scale (BRMES, Bech et
al. 1979) yielded a three-dimensional structure comprising dimensions of "retardation," "depressive core symptoms," and "accessory depressive symptoms" (Miiller and
Wetzel 1998). The three-factor model (table 1) provided
the best empirical data fit in patients with acute schizophrenia, although the BRMES had shown unidimensionality in major depression without psychotic features
(Maier and Philipp 1985). Therefore, this heterogeneity of
depression in psychotic disorders should be taken into
account when analyzing correlations with negative and
positive symptoms, as differential relationships between
latent depressive factors and negative symptoms could
emerge. Moreover, different diagnostic groups presenting
with concurrent depressive and psychotic symptoms could
differ in the structure and composition of depressive factors. Such differences may influence the relationship of
depression with other symptom domains.
We investigated the relationship between depressive
symptoms and other symptom domains (i.e., positive and
negative symptoms as assessed by the Positive and
Negative Syndrome Scale [PANSS, Kay et al. 1987] as
well as akinetic symptoms measured with the
Extrapyramidal Symptom Rating Scale [ESRS, Chouinard
et al. 1980]). For this purpose, we utilized data from a
previously published study (Miiller-Siecheneder et al.
1998). The patients were included according to an operationally defined criterion of concurrent substantial depressive and psychotic symptoms. Three diagnostic groups of
patients representing major depression with psychotic features, schizoaffective disorder, and schizophrenia with
substantial depression could be separated according to
nosological criteria (DSM-III-R, American Psychological
Association 1987). In a preceding methodological study,
we successfully replicated the aforementioned threedimensional model of depression in this sample (Miiller et
al. 1999fo). In the present analysis, we focused on
transnosological differences in depressive subfactors and
on relationships of depression with positive, negative, and
extrapyramidal symptoms.
Materials and Methods
Patients. The present analysis used baseline data from
patients participating in a multicenter clinical trial in 14 psy-
Table 1. Three-dimensional factor model of the BRMES
BRMES item
1
2
3
4
5
6
7
8
9
10
11
Motor retardation
Verbal retardation
Intellectual retardation
Psychic anxiety
Suicidality
Depressed mood
Self-depreciation
Emotional retardation
General insomnia
Tiredness and pains
Work and activities
Factor 1
Retardation
Factor 2
Depressive core
symptoms
Factor 3
Accessory depressive
symptoms
X
X
X
X
X
X
X
X
X
X
X
Note.—BRMES = Bech-Rafaelsen Melancholia Scale.
Muller and Wetzel 1998; Muller et al. 1999b
1
20
Depressive Factors
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
chiatric departments in Germany and Austria. The study was
conducted in accordance with the current version of the
Declaration of Helsinki, and the protocol was approved by
local ethics committees. All patients had given their written
informed consent; patients under legal guardianship were not
enrolled. Some results of the trial have been recently published (Miiller-Siecheneder et al. 1998). Patients were
included if they presented with coexisting substantial depressive and psychotic symptoms. The patients' diagnoses were
based on structured clinical interviews (the Structured
Clinical Interview for DSM-III-R, Spitzer et al. 1990), but
inclusion criteria were operationally defined as a total score
of at least 60 points on the PANSS, a score of at least 4
points on at least two items of the PANSS positive symptoms
subscale, a total score of at least 15 points on the BRMES,
and a score of at least 3 points on the BRMES depression
item. Included were 123 consecutively admitted patients
(aged 18-65 years) with depressive and psychotic symptoms. Thirty-eight patients had major depression with psychotic features (MDE, DSM-III-R 296.24/34), 19 patients
had a nonresidual schizophrenia or schizophreniform disorder with features of a major depression without fulfilling
syndromal MDE criteria (SCHIZ, DSM-III-R 295.X), and
66 patients had a schizoaffective disorder (DSM-III-R
295.7) of depressive (SAD, n = 64) or bipolar (« = 2) subtype. The two patients presenting with schizoaffective disorder, bipolar type, were excluded from all analyses. Axis II
disorder comorbidity (n = 17, 14%) and pretreatment with
antipsychotics (n = 73, 59%), antidepressants (n = 53,43%),
or benzodiazepines (n = 29, 24%) showed no differences
between groups according to DSM-III-R Axis I diagnosis.
Symptom Assessments. Psychopathometric baseline data
of 119 patients were available for the present analysis (table
2). All participating patients underwent at least a 3-day
washout period, during which no antipsychotic or antidepressant drugs were allowed. After 3 days, the baseline data
were assessed. All diagnostic assessments and ratings were
carried out by experienced clinicians. One rater did all the
ratings on a single patient. Rater training was performed
before the study; indexes of interrater reliability were not
calculated.
The assessment comprised the PANSS (Kay et al.
1987), the BRMES (Bech et al. 1979), and the ESRS
(Chouinard et al. 1980). The PANSS consists of three subscales measuring positive symptoms (POS, 7 items), neg-
Table 2. Sample characteristics at baseline
MDE group
(n = 37)
SAD group
(n = 63)
% female
Age, mean (SD)
44.7%
47.4(12.3)
75.0%
44.8(10.9)
52.6%
38.7 (10.3)
(9.7')
3.8
POS, mean (SD)
20.3 (3.3)
21.2(4.1)
26.1 (3.8)
15.5
NEG, mean (SD)
GEN, mean (SD)
TOTAL, mean (SD)
BRMES, mean (SD)
BRMES F1, mean (SD)
BRMES F2, mean (SD)
27.9 (5.7)
55.9 (7.9)
104.2(13.8)
26.9 (4.2)
9.3 (2.2)
9.8 (2.4)
27.0 (5.8)
51.7(8.2)
99.8(15.0)
25.6 (4.5)
9.2 (2.4)
8.8 (2.2)
28.8 (5.5)
55.6 (7.8)
110.5(13.8)
25.4 (3.7)
9.3 (2.7)
7.9 (2.2)
0.8
4.0
4.2
1.3
0.05
4.4
0.44
0.027
0.018
0.28
0.95
0.015
BRMES F3, mean (SD)
ESRS, mean (SD)
% ESRS = 0 4
7.9(1.6)
4.5 (6.2)
37.8%
7.7(1.6)
7.9 (9.5)
28.6%
8.2(1.4)
7.6(13.4)
36.8%
0.8
0.47
0.24
0.58
Characteristic
SCHIZ group
(n = 19)
'2,116
1.43
(1.1 1 )
P
0.0064
0.025
< 0.001
Single comparisons
(P<0.05)
MDE < SAD*
SCHIZ < MDE
SCHIZ < SAD
MDE < SCHIZ
SAD < SCHIZ
—
SAD < MDE
SAD < SCHIZ
—
—
SCHIZ < MDE
SAD < MDE
—
—
—
Note.— BRMES = Bech-Rafaelsen Melancholia Scale; ESRS = Extrapyramidal Symptom Rating Scale (Parkinsonism and akinesia); F1 =
Factor 1, retardation; F2 = Factor 2, depressive core symptoms; F3 = Factor 3, accessory depressive symptoms; GEN = PANSS general psychopathology subscale; MDE = major depression with psychotic features; NEG = PANSS negative subscale; PANSS = Positive and Negative
Syndrome Scale; POS = PANSS positive subscale; SAD = schizoaffective disorder, depressive subtype; SCHIZ = nonresidual schizophrenia
with depressive symptoms; SD = standard deviation; TOTAL = PANSS total score.
3
4
Nonparametric comparison by Kruskal-Wallis ANOVA revealed x 2 = 2.3, df = 2, p = 0.31.
Proportion of patients without extrapyramidal symptoms (ESRS score = 0) at the baseline assessment.
21
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
M.J. Miilleretal.
ative symptoms (NEG, 7 items), and general psychopathological symptoms (16 items). All 30 items are
rated on a seven-point scale from 1 (absent) to 7
(extreme), following well-defined anchor points.
The BRMES measures symptoms of depression with
high interrater reliability (Bech et al. 1979). Its 11 items
are rated from 0 (absent) to 4 (severe). Content, concurrent, and external validity of the BRMES with reference
to other observer ratings has been repeatedly reported
(Maier et al. 1988; Smolka and Stieglitz 1999). In a previous methodological study on the dimensionality of the
BRMES in 132 patients with acute schizophrenia, we
could show that a three-dimensional model was most
appropriate. The three dimensions of retardation, depressive core symptoms, and accessory depressive symptoms
were established by CFA. This three-dimensional model
was originally derived from an exploratory factor analysis and was shown to be superior to one-, two-, and fourdimensional models. The three-dimensional model was
successfully replicated for the total group of patients of
the present study (n = 121) with a goodness-of-fit index
of 0.91 for the total sample. A comparison of parameters
of the same model derived from the two independent
samples revealed a highly satisfactory agreement (comparative fit index = 0.87) (Miiller and Wetzel 1998;
Miiller et al., 19996). Furthermore, the correlation pattern
of the 11 BRMES items was very similar in the three
nosological subgroups of patients with MDE, SCHIZ, or
SAD (Miiller et al. 1999a). Table 1 shows the threedimensional model of the BRMES corroborated in two
studies (Miiller and Wetzel 1998; Miiller et al. 1999ft).
Summary scores for the three factors were derived by
summing up the corresponding item scores and were
used in the present evaluation of the relationship between
symptom domains.
Extrapyramidal symptoms were rated with the ESRS
(Chouinard et al. 1980), comprising a Parkinsonism selfrating questionnaire and scales for the objective assessment of Parkinsonism and akinesia, of dystonia, and of
dyskinesia. The present analysis is restricted to the sum
score consisting mainly of Parkinsonism and akinesia
symptoms.
computed. As parametric and nonparametric coefficients
did not show any substantial differences, only Pearson's
correlations were reported. Power calculations revealed
that in samples with a size of n - 20, 40, and 65, a significant correlation coefficient (p < 0.05) of r > 0.58, 0.42,
and 0.35, respectively, can be revealed with a power of
about 0.80 (Dallal 1985). The global level of significance
was set at a = 0.05. The Bonferroni procedure was used
for adjustment of multiple testing (adjusted level of significance a' = 0.05).
Results
Sample characteristics are presented in table 2. There
were significant differences in gender and age between
the groups, with a significantly higher proportion of
women with SAD than with MDE (p < 0.05). The patients
with SCHIZ were younger than patients of the other two
groups (p < 0.05). Despite the fact that substantial positive and depressive symptoms were inclusion criteria for
the study independent from nosological diagnosis, there
were notable differences between the nosological groups
with respect to psychopathology. Patients with SCHIZ
had significantly higher scores on global psychopathology
severity as measured by the PANSS (p < 0.05).
Nonspecific, general psychopathology (GEN) was higher
in patients with MDE compared to SAD (p < 0.05).
Moreover, patients with MDE presented with significantly
higher scores on the core depression subscore of the
BRMES scores (p < 0.05), whereas retardation and accessory depressive symptoms, as well as BRMES total
scores, were comparable between groups. Negative symptoms (PANSS negative subscale) did not differ significantly between groups and were rather high on average,
whereas extrapyramidal symptoms assessed by the ESRS
were nearly entirely constituted by akinetic symptoms
(Parkinsonism) and were comparably low in all nosological subgroups. Additionally, the proportion of patients
without extrapyramidal symptoms at baseline assessment
(according to an ESRS total score of zero, 32.2% of the
total group) was not significantly different between nosological groups.
The correlation coefficients for the four domains of
positive, negative, depressive, and extrapyramidal symptoms are reported in table 3 separately for the three diagnostic groups. The upper part of the table shows the relationships of PANSS positive and negative subscales as
well as BRMES and ESRS total scores. In all three diagnostic groups, only nonsignificant (p > 0.05) correlation
coefficients were revealed for the association of positive
symptoms with negative, depressive, and extrapyramidal
symptoms. With respect to the diagnostic groups, the
interrelations were of very similar magnitude and in the
Statistical Analyses. Means and standard deviations are
presented in table 2. Global group differences in continuous parameters were analyzed by one-way analysis of
variance (ANOVA) comparing groups of patients with
MDE, SCHIZ, and SAD. For single comparisons between
groups, unpaired t tests were used in case of significant
ANOVA results. Group differences in gender were examined by x2 tests. Correlations were analyzed by means of
Pearson product moment coefficient (r) after inspection of
all possible bivariate scatterplots. Additionally, nonparametric Spearman rank correlation coefficients (rs) were
22
Depressive Factors
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
Table 3. Correlation between symptom domains in patients with different diagnoses
POS
SAD
—
—
BRMES
NEG
POS
MDE
SCHIZ
MDE
SAD
SCHIZ
—
0.23
0.20
0.27
0.31
—
—
—
0.54*
—
0.64*
0.90*
0.70*
NEG
BRMES
MDE
SAD
ESRS
SCHIZ
MDE
SAD
SCHIZ
0.23
0.31
-0.23
-0.29
-0.25
0.52*
0.70*
0.33
0.41*
0.20
—
—
0.33
0.07
0.10
0.75*
0.68*
0.27
0.23
0.29
BRMES F1
0.27
0.03
0.27
0.59*
BRMES F2
0.25
0.21
0.03
0.31
0.21
-0.17
0.79*
0.76*
0.43
0.14
-0.02
-0.33
BRMES F3
0.08
0.31
0.25
0.14
0.22
0.37
0.49*
0.65*
0.64*
0.29
-0.12
0.21
Note.— BRMES = Bech-Rafaelsen Melancholia Scale; ESRS = Extrapyramidal Symptoms Rating Scale (Parkinsonism and akinesia); F1
= Factor 1, retardation; F2 = Factor 2, depressive core symptoms; F3 = Factor 3, accessory depressive symptoms; MDE = major depression with psychotic features (n = 37); NEG = PANSS negative subscale; PANSS = Positive and Negative Syndrome Scale; POS =
PANSS positive subscale; SAD = schizoatfective disorder, depressive subtype (n = 63); SCHIZ = nonresidual schizophrenia with depressive symptoms (n = 19).
*ps0.0027(s0.05/18)
same direction. Relationships between positive and
extrapyramidal symptoms were consistently negative,
whereas positive and depressive symptoms showed a low
but positive interrelationship. Moreover, negative symptoms showed a consistent pattern of positive, nonsignificant correlations with positive symptoms across the three
diagnostic groups as well as positive relations with ESRS
total scores, reaching significance in the group of SAD
patients (p < 0.05). Negative symptoms were statistically
significantly related to the BRMES total score in all subgroups (p < 0.05). The lower part of table 2 provides further information on the relationship between negative and
depressive symptoms. It was found that only the retardation subfactor of the BRMES (factor 1) showed substantial and significant (p < 0.05) correlations with the PANSS
negative subscore. Again, all three diagnostic subgroups
showed a similar correlation pattern. Neither depressive
core symptoms (BRMES factor 2) nor accessory depressive symptoms (BRMES factor 3) showed a substantial
overlap with PANSS negative symptoms (p > 0.05, figure
1). None of the depression subfactors showed a significant
correlation with either positive or extrapyramidal symptomatology.
With respect to interrelations among the depressive
subfactors, no considerable dissimilarities were yielded
between nosological groups (table 4). Within the three
diagnostic groups, the BRMES factors were almost unrelated, in accordance with an underlying latent factor
model considering orthogonal dimensions. Core depression was significantly related to accessory depressive
symptoms only in the SAD group (p < 0.05). When
descriptively comparing the correlation coefficients
between groups (Fisher's z transformation), no significant
differences between corresponding correlation coefficients emerged. Additionally, no significant differences
between female and male patients were found with
respect to the interrelation of BRMES depression factors
in the SAD group.
Discussion
Relationships Between Symptom Domains in Different
Diagnostic Groups. In a study of acutely exacerbated
patients with functionally defined concurrent depressive
and psychotic symptoms, the correlation pattern between
different symptom domains was very similar in patients
Table 4. Correlation between BRMES factors in patients with different diagnoses
Factor 3
Factor 2
Factor 1 Retardation
Factor 2 Depressive core symptoms
Factor 3 Accessory depressive symptoms
MDE
SAD
SCHIZ
MDE
SAD
SCHIZ
0.29
0.26
-0.27
0.01
0.20
0.24
0.37*
0.30
0.08
Note.— BRMES = Bech-Rafaelsen Melancholia Scale; MDE = major depression with psychotic features (n = 37); SAD = schizoaffective
disorder, depressive subtype (n = 63); SCHIZ = nonresidual schizophrenia with depressive symptoms (n = 19).
*pa0.017(s0.05/3)
23
M.J. Miiller et al.
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
with different nosological diagnoses. In patients with
either schizophrenia, schizoaffective disorder, or psychotic
depression diagnosed according to DSM-III-R, positive
symptoms (PANSS) were not significantly related to negative (PANSS), depressive (BRMES), or akinetic (ESRS)
symptoms. Negative and depressive symptoms were substantially related in the three subgroups of patients, and a
slight positive association emerged between negative and
extrapyramidal symptoms. These findings strongly support
the hypothesis that relations among different symptom
domains in patients with both psychotic and depressive
symptoms are not substantially different between diagnostic groups when assessed with the same standardized rating instrument. Different nosological categories seem to
differ in their particular composition of latent dimensional
factors and in the severity of specific symptoms, but not in
the presence and relationship of more global symptom
domains. Accordingly, global severity of psychopathology
as measured by the total scores of the PANSS was higher
in SCHIZ patients. The BRMES factor comprising depressive core symptoms (psychic anxiety, suicidality,
depressed mood, and self-depreciation) could significantly
separate MDE patients from patients with SAD and
SCHIZ, although total BRMES scores and scores of the
other depressive dimensions (retardation and accessory
depressive symptoms) were comparable.
The results reflect the ongoing problem in defining
nosological borders between different schizophrenic subtypes as well as in establishing a boundary between schizophrenia and related psychotic disorders, mainly schizoaffective disorder and psychotic depression. These results
also corroborate a recent factor-analytic study on 221
schizophrenia and 189 nonschizophrenia patients, which
proposes that the symptom structure described in schizophrenia is not specific for that diagnosis (Ratakonda et al.
1998). However, that study focused on positive, negative,
and disorganized symptoms and did not include depressive or akinetic symptoms.
Figure 1. Correlation between negative symptoms
and BRMES depression factors in patients with
different diagnoses
Negative Symptoms (PANSS)
•
O
•
O
I
I
o
T
I
I
I
I
I
I
I
I
I
Depression and Negative Symptoms. A particular issue
of the present study was a differentiated evaluation of the
relationship between depressive factors and other symptom domains. BRMES total scores were substantially correlated (r > 0.50) with the PANSS negative subscale in all
diagnostic groups. The overlap in variance (r2) ranged
from 27 percent to 49 percent and is in line with several
previous reports using different assessment instruments
and investigating patients with various clinical features
(Whiteford et al. 1987; Kitamura and Suga 1991; Sax et
al. 1996). This result suggests a lack of independence
between these two symptom domains. However, when
analyzing latent factors of depression, negative symptoms
were substantially related exclusively to the retardation
factor of the BRMES in all nosological groups, with a
8 m
•3 M
H
Note.— BRMES = Bech-Rafaelsen Melancholia Scale; MDE
major depression with psychotic features (n = 37); SAD
schizoaffective disorder, depressive subtype (n = 63); SCHIZ
nonresidual schizophrenia with depressive symptoms (n = 19).
24
Depressive Factors
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
shared variance (r2) of 35 percent to 81 percent, thus providing an explanation for the significant correlation
between negative symptoms and BRMES total scores in
the three subgroups of patients. Particularly, the BRMES
factor of depressive core symptoms was not substantially
correlated with negative symptoms in any diagnostic
group (r2 < 10%). The correlation of accessory depressive
symptoms with negative symptoms also showed low values in all investigated patient groups (r2 < 14%).
As previously claimed (Sax et al. 1996), the distinction between depressive and negative symptoms is clinically important in the investigation of underlying biological mechanisms and treatment. Previous studies using
the HAMD (e.g., Sax et al. 1996) are likely to have confused negative and depressive symptoms. The HAMD
was criticized for its nonspecificity (Goldman 1992),
heterogeneity (Gibbons et al. 1993), and deviation from
unidimensionality (Maier and Philipp 1985). The
BRMES was suggested to be superior in this regard
(Maier and Philipp 1985). However, there is still a lack
of dimensional homogeneity of the BRMES in psychotic
disorders, and a three-dimensional model of depressive
symptoms seems to fit empirical data best (Miiller and
Wetzel 1998; Miiller et al. 19996). The recently developed Calgary Depression Rating Scale for Schizophrenia
(CDSS, Addington et al. 1990; Miiller et al. 1999a)
could overcome such deficiencies of traditionally used
depression rating instruments in psychotic disorders
(Collins 1996). This scale has been reported to show
only a weak overlap with negative symptoms in schizophrenia patients (Addington et al. 1994, Addington et al.
1996). It is interesting that three of four symptoms of the
BRMES core depression factor (depressed mood, suicidality, and self-depreciation) are covered by the nineitem CDSS.
A confounding variable in earlier studies was the
propensity of conventional neuroleptics to induce
depressionlike symptoms ("akinetic depression," van
Putten and May 1978) or signs and symptoms that
mimic negative symptoms (Harrow et al. 1994). In the
present study, BRMES total score and BRMES subscores were not related to extrapyramidal symptoms,
whereas a slight association emerged between negative
and extrapyramidal symptoms (r2 = 4% to 17%) (shared
variance between 4% and 17%), which reached significance only in patients with SAD. Thus, the relationship
between depressive and negative symptoms does not
seem confounded by extrapyramidal symptoms in the
present study. Additional subgroup analyses in patients
with or without baseline extrapyramidal symptoms
(ESRS score > or = 0, respectively) revealed no substantial influence of extrapyramidal symptoms on group
means or on relationships of the remaining symptom
domains (data not shown). As clozapine and other newer
antipsychotics have less propensity for inducing
extrapyramidal symptoms and seem to show particular
antidepressant properties in addition to their antipsychotic efficacy, prospective studies with these drugs
using a refined assessment of depression should be carried out. Retrospective statistical analyses, including
path analysis (Tollefson et al. 1998), cannot replace
prospectively designed studies to untangle overlapping
psychopathological assessments. Furthermore, empirically derived and clinically meaningful symptom classifications are more appropriate to test hypotheses of
underlying biological mechanisms (e.g., by means of
functional imaging and receptor occupancy studies).
Depression and Positive Symptoms. Neither the
BRMES total score nor any of the depressive factors were
substantially related to positive symptoms in any of the
diagnostic subgroups. These results could not replicate the
findings of Sax et al. (1996) reporting a significant positive correlation between positive and depressive symptoms in schizophrenia spectrum disorder but not in psychotic depression. The authors have claimed that either
the higher severity of depression in delusional depression
or a different status of depression in both diseases could
explain their results. In our study, patients in the MDE
and SCHIZ groups had comparable BRMES total scores
but differed significantly in the level of depressive core
symptoms. Nonetheless, all diagnostic subgroups showed
comparable nonsignificant relationships between positive
and depressive symptoms.
When comparing our study with that of Sax et al.
(1996), some differences are apparent. Different assessment scales were used, and the HAMD may be rather
unsuitable for specific assessment of depression in people
with schizophrenia, as 65 percent of the HAMD items
cover symptoms that are not directly mood-related (Snaith
1993). In addition, Sax et al. (1996) investigated only
patients without psychopharmacological treatment prior
to the study; patients may have received medication
within the 72 hours before assessment. In the present
study, about 50 percent of patients had received neuroleptics or antidepressants prior to the study, but all underwent
washout for 3 days before assessment. It is unclear
whether previous medication or washout periods have
short- or long-term effects on the relationship of symptom
domains. Another study has found a positive association
between dysphoric mood and positive symptoms in schizophrenia patients during a period of 12 to 29 months
(Norman and Malla 1994), whereas Dollfus et al. (1993)
reported a negative association between depressive and
positive symptoms during the acute state of schizophrenic
illness and a lack of association between the symptom
domains at the postacute and residual states. Our study
therefore does not corroborate either finding. Nonetheless,
25
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
M.J. Miiller et al.
toms and dimensions than traditional investigations.
However, we did not analyze the relationship of single
symptoms, although such a strategy has been successfully
applied (Barnes et al. 1989; Costello 1992; Norman and
Malla 1994). Our endeavor may be regarded as intermediate between an entirely symptom-focusing and a completely nosological approach.
The high prevalence of depressive symptoms in
schizophrenia is increasingly recognized (Siris 1995;
Addington et al. 1998). The elevated suicide risk for psychotic patients seems to be related to depressive mood
states (Addington and Addington 1992), whereas—at
least in chronic schizophrenia—higher scores on negative
symptoms were associated with a lower risk for suicide
(Fenton et al. 1997). With respect to these data, it seems
clinically very important to disentangle depressive and
negative symptoms. According to the present results, negative symptoms and depressive core symptoms can be
separated across nosological borders. For this purpose,
accurate assessment and differentiated analysis are indispensable. When a traditional depression rating scale is
used, the partial overlap of some depressive symptoms
with negative symptoms has to be taken into account.
Depression in psychotic disorders should not be regarded
as unidimensional, and there seems to exist a depressive
core syndrome independent from positive, negative, and
akinetic extrapyramidal symptoms in schizophrenia, in
schizoaffective disorder, and in psychotic depression.
The findings should encourage further studies on the
underlying biological mechanisms, neuropsychological
correlates, and specific treatment of symptom domains
and factors in patients suffering from that serious combination of psychotic and depressive symptoms, which puts
them at the highest potential risk for suicide.
our results seem to be intriguing as they were very consistent among three diagnostic subgroups and our evaluation
strategy seems to be appropriate for separating negative
and depressive symptoms.
Limitations and Conclusions. A methodological limitation of the present study is the partly restricted and
unequal sample size, due to the inclusion criteria not
focusing on equal-sized subgroups. A power analysis
revealed that significant nonzero correlations could be
revealed with sufficient power in the analyzed subgroups.
In addition, the consistency of findings and adjustment for
multiple tests substantiate the present results. With respect
to the significantly greater number of women with SAD in
the present sample, we did not find substantial sex differences in the BRMES factor structure. However, given the
generally higher prevalence of depression in women, further studies addressing possible sex differences in the
relationship between different symptom domains are
needed to clarify this issue.
All patients of the present study were newly admitted to
the hospital due to an acute exacerbation. Hence, interpretation should not be incautiously extended to the posthospitalization period or to the chronic course of illness.
Another problem might be that patients were drug-free
for only 3 days. For studying basic psychopathological phenomena, untreated patients are certainly most suitable; however, totally drug-naive patients with acute psychosis and
depression are—fortunately, from a clinical viewpoint—the
exception in university hospitals. A washout period of 3 days
is about as long as could practically and ethically be done.
Nevertheless, some drug effects may still have been present
due to the different half-lives of different medication in different individuals. Such effects could have contributed to the
findings on the relationships between different symptom
domains. However, there was no significant difference
between the diagnostic groups regarding pretreatment with
neuroleptics, antidepressants, or benzodiazepines. Therefore,
pretreatment should not have influenced the findings of the
present analyses differentially in the three diagnostic groups.
When looking at the interrelations of depressive dimensions as assessed by the BRMES, no considerable differences emerged with respect to the nosological groups. The
associations were generally low, corroborating a CFA model
of three independent latent dimensions (Miiller and Wetzel
1998; Muller et al. 1999b). This finding does not support the
possibility of differences in the internal structure of depressive symptoms between diagnostic subgroups.
In the present analysis, patients with substantial concurrent psychotic and depressive symptoms were investigated. The patients were prospectively included according
to their functionally defined symptomatology and not on
the basis of a nosological diagnosis. Such a functional
approach (Benkert 1990) is a priori more related to symp-
References
Addington, D.; Addington, J.; and Atkinson, M. A psychometric comparison of the Calgary Depression Scale for
Schizophrenia and the Hamilton Depression Rating Scale.
Schizophrenia Research, 19(2-3):205-212, 1996.
Addington, D.; Addington, J.; and Maticka-Tyndale, E.
Specificity of the Calgary Depression Scale for schizophrenics. Schizophrenia Research, ll(3):239-244, 1994.
Addington, D.; Addington, J.; and Patten, S. Depression
in people with first-episode schizophrenia. British Journal
of Psychiatry, 172(33):90-92, 1998.
Addington, D.; Addington, J.; and Schissel, B. A depression rating scale for schizophrenics. Schizophrenia
Research, 3(4):247-251, 1990.
Addington, D., and Addington, J.M. Attempted suicide
and depression in schizophrenia. Ada Psychiatrica
Scandinavica, 85(4):288-291, 1992.
26
Depressive Factors
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
American Psychiatric Association.
DSM-HI-R:
Diagnostic and Statistical Manual of Mental Disorders.
3rd ed., revised. Washington, DC: APA, 1987.
Gibbons, R.D.; Clark, D.C.; and Kupfer, D.J. What
exactly does the Hamilton Depression Rating Scale measure? Journal of Psychiatric Research, 27(3):259-273,
1993.
Andreasen, N.C. Scale for the Assessment of Positive
Symptoms. Iowa City, LA: University of Iowa College of
Medicine, 1982.
Goldman, R.S.; Tandon, R.; Liberzon, I.; and Greden, J.F.
Measurement of depression and negative symptoms in
schizophrenia. Psychopathology, 25(l):49-56, 1992.
Barnes, T.R.; Curson, D.A.; Liddle, PR; and Patel, M.
The nature and prevalence of depression in chronic schizophrenic in-patients. British Journal of Psychiatry,
154:486-491, 1989.
Hamilton, M. A rating scale for depression. Journal of
Neurology, Neurosurgery, and Psychiatry, 23:56-62,
1960.
Harrow, M.; Yonan, C.A.; Sands, J.R.; and Marengo, J.
Depression in schizophrenia: Are neuroleptics, akinesia,
or anhedonia involved? Schizophrenia
Bulletin,
20(2):327-338, 1994.
Barnes, T.R., and McPhillips, M.A. How to distinguish
between the neuroleptic-induced deficit syndrome, depression, and disease-related negative symptoms in schizophrenia. International Clinical Psychopharmacology, 10(Suppl
3):115-121, 1995.
Hirsch, S.R. Depression "revealed" in schizophrenia.
British Journal of Psychiatry, 140:421^23, 1982.
Bech, P.; Bolwig, T.G.; Kramp, P.; and Rafaelsen, O.J.
The Bech-Rafaelsen Melancholia Scale and the Hamilton
Depression Scale. Ada Psychiatrica
Scandinavica,
59(4):420-430, 1979.
Kay, S.R.; Fiszbein, A.; and Opler, L.A. The Positive and
Negative Syndrome Scale (PANSS) for schizophrenia.
Schizophrenia Bulletin, 13(2):261-276, 1987.
Benkert, O. Functional classification and response to
psychotropic drugs. In: Benkert, O.; Maier, W.; and
Rickels, K., eds. Methodology of the Evaluation of
Psychotropic Drugs. Berlin, Germany: Springer, 1990.
pp. 155-163.
Kitamura, T., and Suga, R. Depressive and negative
symptoms in major psychiatric disorders. Comprehensive
Psychiatry, 32(l):88-94, 1991.
Lewine, R.R. A discriminant validity study of negative
symptoms with a special focus on depression and antipsychotic medication. American Journal of Psychiatry,
147(11): 1463-1466, 1990.
Chouinard, G.; Ross-Chouinard, A.; Annable, L.; and
Jones, B.D. The Extrapyramidal Symptom Rating Scale.
Canadian Journal of Neurology and
Psychiatry,
7:233-243, 1980.
Maier, W.; Heuser, I.; Philipp, M.; Frommberger, U.; and
Demuth, W. Improving depression severity assessment: n.
Content, concurrent, and external validity of three
observer depression scales. Journal of Psychiatric
Research, 22(1):13-19,1988.
Collins, A.A.; Remington, G.; Coulter, K.; and Birkett, K.
Depression in schizophrenia: A comparison of three measures. Schizophrenia Research, 20(l-2):205-209, 1996.
Costello, C.G. Research on symptoms versus research on
syndromes: Arguments in favor of allocating more
research time to the study of symptoms. British Journal of
Psychiatry, 160:304-308, 1992.
Maier, W., and Philipp, M. Comparative analysis of
observer depression scales. Ada
Psychiatrica
Scandinavica, 72(3):239-245, 1985.
Maziade, M.; Roy, M.A.; Martinez, M.; Cliche, D.;
Fournier, J.P.; Garneau, Y.; Nicole, L.; Montgrain, N.;
Dion, C ; and Ponton, A.M. Negative, psychoticism, and
disorganized dimensions in patients with familial schizophrenia or bipolar disorder: Continuity and discontinuity
between the major psychoses. American Journal of
Psychiatry, 152(10): 1458-1463, 1995.
Dallal, G.E. PC-SIZE: A Program for Sample Size
Determinations, Version 2.0. Boston, MA: USDA Human
Nutrition Research Center on Aging at Tufts University,
1985.
Dollfus, S.; Petit, M.; and Menard, J.F. Relationship
between depressive and positive symptoms in schizophrenia. Journal of Affective Disorders, 28(l):61-69, 1993.
Fenton, W.S.; McGlashan, T.H.; Victor, B.J.; and Blyler,
C.R. Symptoms, subtype, and suicidality in patients with
schizophrenia spectrum disorders. American Journal of
Psychiatry, 154(2): 199-204, 1997.
Miiller, M.J.; Marx-Dannigkeit, P.; Schlosser, R.; Wetzel,
H.; Addington, D.; and Benkert, O. The Calgary
Depression Rating Scale for Schizophrenia: Development
and interrater reliability of a German version (CDSS-G).
Journal of Psychiatric Research, 33(5):433-443, 1999a.
Gerbaldo, H.; Fickinger, M.P.; Wetzel, H.; Helisch, A.;
Philipp, M.; and Benkert, O. Primary enduring negative
symptoms in schizophrenia and major depression. Journal
of Psychiatric Research, 29(4): 297-302, 1995.
Miiller, M.J., and Wetzel, H. Dimensionality of depression
in acute schizophrenia: A methodological study using the
Bech-Rafaelsen Melancholia Scale (BRMES). Journal of
Psychiatric Research, 32(6):369-378, 1998.
27
Schizophrenia Bulletin, Vol. 27, No. 1, 2001
M.J. Mulleretal.
Miiller, M.J.; Wetzel, H.; Szegedi, A.; and Benkert, O.
Three dimensions of depression in patients with acute
psychotic disorders: A replication study. Comprehensive
Psychiatry, 40(6):449-457, 199%.
Smolka, M., and Stieglitz, R.D. On the validity of the
Bech-Rafaelsen Melancholia Scale. Journal of Affective
Disorders, 54(1-2): 119-128, 1999.
Snaith, P. What do depression scales measure? British
Journal of Psychiatry, 163:293-298, 1993.
Muller-Siecheneder, R; Miiller, M.J.; Hillert, A.; Szegedi, A.;
Wetzel, H.; and Benkert, O. Risperidone versus haloperidol
and amitriptyline in the treatment of patients with a combined psychotic and depressive syndrome. Journal of
Clinical Psychopharmacology, 18(2): 111-120,1998.
Spitzer, R.L.; Williams, J.B.W.; Gibbons, M.; and First,
M.B. The Structured Clinical Interview for DSM-III-R,
Patient Edition (SCID-P, Version 0). Washington, DC:
American Psychiatric Press, 1990.
Mundt, C ; Kasper, S.; and Huerkamp, M. The diagnostic
specificity of negative symptoms and their psychopathological context. British Journal of Psychiatry, l55(Suppl
7):32-36, 1989.
Tollefson, G.D.; Sanger, T.M.; Lu, Y.; and Thieme, M.E.
Depressive signs and symptoms in schizophrenia: A
prospective blinded trial of olanzapine and haloperidol.
Archives of General Psychiatry, 55(3):250-258, 1998.
Norman, R.M., and Malla, A.K. Correlations over time
between dysphoric mood and symptomatology in schizophrenia. Comprehensive Psychiatry, 35(l):34-38, 1994.
Toomey, R.; Faraone, S.V.; Simpson, J.C.; and Tsuang,
M.T. Negative, positive, and disorganized symptom
dimensions in schizophrenia, major depression, and bipolar disorder. Journal of Nervous and Mental Disease,
186(8):470-476, 1998.
Prosser, E.S.; Csernansky, J.G.; Kaplan, J.; Thiemann, S.;
Becker, T.J.; and Hollister, L.E. Depression, Parkinsonian
symptoms, and negative symptoms in schizophrenics
treated with neuroleptics. Journal of Nervous and Mental
Disease, 175(2): 100-105, 1987.
Van Putten, T., and May, R.P. "Akinetic depression" in
schizophrenia. Archives of General
Psychiatry,
35(9): 1101-1107, 1978.
Ratakonda, S.; Gorman, J.M.; Yale, S.A.; and Amador,
X.F. Characterization of psychotic conditions: Use of the
domains of psychopathology model. Archives of General
Psychiatry, 55(1):75-81, 1998.
Whiteford, H.A.; Riney, S.J.; and Csernansky, J.G.
Distinguishing depressive and negative symptoms in
chronic schizophrenia. Psychopathology, 20(5-6):234—236,
1987.
Sax, K.W.; Strakowski, S.M.; Keck, P.E., Jr.; Upadhyaya,
V.H.; West, S.A.; and McElroy, S.L. Relationships among
negative, positive, and depressive symptoms in schizophrenia and psychotic depression. British Journal of
Psychiatry, 168(1):68-71, 1996.
The Authors
Matthias J. Miiller , M.D., is Senior Psychiatrist and
Researcher at the Department of Psychiatry, University of
Mainz, Mainz, Germany; Armin Szegedi, M.D., is Senior
Psychiatrist and Researcher at the Department of
Psychiatry, University of Mainz, Mainz, Germany;
Hermann Wetzel, M.D., is Senior Psychiatrist and
Researcher at the Department of Psychiatry, University of
Mainz, Mainz, Germany; Otto Benkert, M.D., is Professor
of Psychiatry and Director of the Department of
Psychiatry, University of Mainz, Mainz, Germany.
Siris, S.G. Depression and schizophrenia. In: Hirsch, S.R.,
and Weinberger, D.R., eds. Schizophrenia. Oxford,
England: Blackwell Science, 1995. pp. 128-145.
Siris, S.G.; Adan, F.; Cohen, M.; Mandeli, J.; Aronson, A.;
and Casey, E. Postpsychotic depression and negative
symptoms: An investigation of syndromal overlap.
American Journal of Psychiatry, 145( 12): 1532-1537,
1988.
28