”will! W 11111" ‘1

 

3 129 93 10735 7240 “I’M-4" ‘
Michigan State
University |

 

.—— fl

This is to certify that the

dissertation entitled

THE USE OF THE DEXAMETHASONE SUPPRESSION
TEST TO IDENTIFY A DISTINCT
SUBTYPE 0F DEPRESSION

presented by

Charles L'Engle West

has been accepted towards fulfillment
of the requirements for }

Ph.D. (kgfimin Counseling Psychology

 

 

W W}WZ’\

Major professor

Date April 19, 1984

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

MSU

 

 

 

RETURNING MATERIALS:
Place in book drop to

 

usumas remove this checkout from
“ your record. ‘FINES Will
be charged if book is
returned after the date
( stamped below.
Ornts 75’
33937 “33?

 

 

THE USE OF THE DEXAMETHASONE SUPPRESSION
TEST TO IDENTIFY A DISTINCT
SUBTYPE OF DEPRESSION

by

Charles L'Engle West

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

College of Education

Department of Counseling, Educational Psychology
and Special Education

1984

Charles L. West

ABSTRACT

The major purpose of this study was to explore the
relationship between a positive response to the
dexamethasone suppression test (DST) and clinical features.
Both 'traditionally' endogenous land :neurotic symptomatology
were included in the investigation. Confirmation of a
hypothesized norepinephrine deficit associated with DST
nonsuppression. was also sought through. placing DST
nonsuppressors and patients diagnosed endogenously depressed
on the noradrenergic medication, desipramine.

A sample of 107 depressed men and women was obtained
from a consortium of mental health facilities as well as
from self-referral. The Schedule for Affective Disorders and
Schizophrenia (SADS) was administered to each subject along
with the Differential Diagnostic Depression Scale (DDDS) and
the DST. 48 subjects who met Research Diagnostic Criteria
(RDC) for major endogenous depressive disorder, definite or
probable, and/or demonstrated DST nonsuppression were placed
on a 5-week trial with desipramine. Thirty-four patients
completed the trial; 14 dropped out because of medication
side-effects or because of referral to another facility.

A blood cortisol value of.z4.1 ug/dl was the criterion

for DST nonsuppression. Ratings of 46 SADS items and DST

suppression/nonsuppressiom. were used. to address the main
research questions. T-tests yielded 12 individual SADS items
for which DST nonsuppressors had significantly higher
ratings: subjective feeling of severity, psychic anxiety,
initial insomnia, terminal. insomnia, insomnia (severity),
appetite loss, weight loss, indecisiveness, dim
concentration, psychomotor agitation, lack of reactivity,
and functional impairment.

Two discriminant analyses were also performed on the 46
SADS items and DST response. The first analysis, using the
entire sample, created. a: discriminant function which
correctly classified 90.65% of the subjects. In order to
have the opportunity to immediately cross-validate a derived
discriminant function, a second analysis was performed
employing half of the sample stratified by diagnosis and DST
response; it correctly classified 98.15% of the subjects
included. Cross-validation of this second discriminant
function on the withheld half of the sample yielded a
percentage of correct classification that was significantly
better than chance.

Analysis of ‘variance demonstrated that DST
nonsuppressors had a significantly larger change score on
the Hamilton Depression Rating Scale (HDRS) than suppressors
at the end of a 5-week trial of desipramine. However,
analysis of covariance , using baseline HDRS scores as the
covariate, did not uphold the significance of DST

nonsuppression in predicting HDRS change scores. Rather, a

high baseline HDRS score was most predictive of a large

change score.

DEDICATION
To My Parents

and Frances

ii

ACKNOWLEDGMENTS

I wish to acknowledge the very large role that others
had in the original design and execution of the research
project of which my dissertation is but a part. I am most
indebted to the staff of the Clinical Center Affective
Disorders Clinic who were accepting of my desire to become
involved as a research team member.

Gregory Holmes: Aside from. his role as principal
interviewer and trainer for the endogenous depression
project, Greg served as a ready and able consultant for all
project concerns. His easy-going personality made work
around the clinic much more enjoyable than it might have
been. I certainly cannot underestimate the value of the
support he offered me during the rough times with the
research.

Robert Bielski: Bob was very considerate of my needs
throughout my participation in the research project. His
having granted me a graduate assistantship not only provided
me with financial support but also gave me the opportunity
to take on new and previously untried responsibilities. In
his role as a committee member, he was a fair critic who
offered a clear and insightful perspective on research

issues I had to confront.

iii

William Farquhar: Bill was a source of support my
entire four years at Michigan State. He taught me as both
professor and clinical supervisor and helped. me develop
professional direction. During the dissertation process, he
offered valuable suggestions on tuwv to clearly communicate
my ideas on paper.

Christine Shafer: Chris was the main physician in
charge of the drug treatment phase of the study. She. was
always accessible as a consultant and was there as a
listening ear when I needed to talk about a patient. I
appreciated the respect and warmth she showed me.

Andrew Porter: Dr. Porter handled the design and
analysis questions I had about my dissertation with
precision and offered solutions that were manageable for my
level of understanding of statistics. I greatly valued his
input.

Gerald Osborn: Jerry challenged me to develop a
research focus that would have some practical application to
clinical work.

The dedication of Jay Terbush and Shelley Smithson in
their role as interviewers added to the integrity of the
project.

Karen Brainard offered some humorous interludes during

the all-too-serious project work.

iv

LIST OF
LIST OF
Chapter

I.

II.

III.

IV.

TABLE OF CONTENTS

TABLES O O O O O O O O O O O O O O O

FIGURES O O O O O O O O O O O I O O 0

THE PROBLEM . . . . . . . . . . . .

Need for the Study. . .
Purpose . . . . . . . .
Hypotheses. . . . . . .

Theory. . . . . . . .
Summary and Overview.

REVIEW OF THE LITERATURE. . . . . .

Clinical Features of Depressive Subtypes.

Psychoneuroendocrinological Evidence for

Depressive Subtypes . . . . . .
Drug Treatment Response and the Sn
Depression. . . . . . . . . . .
Summary and Conclusions . . . . .

METHODOLOGY OF THE STUDY. . . . . .

Instruments . . . . . . .
Subjects. . . . . . . . .
Procedure . . . . . .
Design of the Study . . . . . .
Operational Definition of ypothes
Data Analysis . . . . . . . . . .

Summary . . . . . . . . . . . . .

mo 0 o 0

RESULTS OF DATA ANALYSIS . . . . . .
Research Hypotheses . . . . . . .

T-Tests . . . . . . . . . . . .
Mann-Whitney U Tests. . . . . .
Discriminant Analyses . . . . .
Analysis of Secondary Hypotheses

Exploratory Hypothesis. . . . . .
Analysis of Variance. . . . . .

Analysis of Covariance. . . . .
Analysis of Secondary Hypotheses

e

S

btyping

Page
vii

viii

\lU‘luD-obw I—'

oo

21

36
45

47

47
52
53
59
59
60
65

67
67
67
71
72
79
81
81

85

V.

sumary O O O O O O O O O O O O O O O O O O O 0
SUMMARY AND CONCLUSIONS 0 O O O O O O O O O O O 0

Summary . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . .
A Review of Recent Findings in the Literature .
Discussion. . . . . . . . . . . . . . . . . .

The Timing of DST Sensitivity for a Depressed
Patient. . . . . . . . . . . . . . . . . . .
DST Nonsuppression and the Traditional Concept
of Endogenous Depression . . . . . . . . . .
Stress and DST Nonsuppression. . . . . . . . .
Weight Loss and DST Nonsuppression . . . . . .
Severity of Illness and DST Nonsuppression . .
Toward a Revised Concept of Endogenous
Depression . . . . . . . . . . . . . . . .
Prognostic Value of DST Nonsuppression . . .
DST Sensitivity and Diagnostic Classification
Systems. . . . . . . . . . . . . . . . . . .

Suggestions for Future Research . . . . . . . .

APPENDICES O O O O O O O O O O O O O O O O O O O O O O

"EIFJU DE!

A. CONSENT FORM A O O C O O C C O O C O O C O O

. CONSENT FORM B . . . . . . . . .
. SCHEDULE FOR AFFECTIVE DISORDERS AND
SCHIZOPHRENIA . . . . . . . . . . . . . .
RESEARCH DIAGNOSTIC CRITERIA. . . . . . . .
HAMILTON DEPRESSION SCALE . . . . . . . .
. COMPARISON OF RDC AND DSMIII CRITERIA . . .

REFERENCES 0 0 O O O O O O O O O O O O O O O O O O 0

vi

87
9O
90
95
97
104
105
107
109
110
111

112
113

114
115
119

119
120

125
129
132
133

134

LIST OF TABLES

Table Page
4.1 T-Test Results for SADS Items Using Entire
sample 0 O O O O O O O I O O O O O O O O O 0 0 68
4.2 Results of Mann-Whitney U Tests (Corrected
for TieS) O O O O O O O O O O O O O O I O O O I 71
4.3 Standardized Discriminant Function Coefficients
for Discriminant Analysis Using Entire Sample. 73
4.4 Discriminant Function Classification Results
for Discriminant Analysis Using Entire Sample. 74
4.5 Standardized Discriminant Function Coefficients
for Discriminant Analysis Using Split/Stratified
sample 0 O O O O O O I I O O O O O O O I O O I 76
4.6 Discriminant Function Classification Results for
Discriminant Analysis Using Split/Stratified
sample 0 O O O O O O O O O O O O O O O O O O O 77
4.7 Interaction of Time of Greatest Severity with
DST Response . . . . . . . . . . . . . . . . . 80
4.8 T-Test Results for Severity of Illness Based on

HDRS scores I O O O O O O I O O O O O O O C O O 81
4.8 ANOVA Results for DST and Dependent Measures . . 82

4.9 ANCOVA of DST with Posttest Using Pretest as
covariate O O O O O O O O O O O O O O O O I O O 8 2

4.10 ANCOVA of DST with Change Scores Using Pretest
as Covariate . . . . . . . . . . . . . . . . . 82

4.11 Correlation Matrix for DST Response and Dependent
Measures 0 O O O O O O O O O O O O O O O O O 0 84

4.12 Interaction of Desipramine Responsiveness with
DST Response 0 O O O O O O O O O O O O I O O O 8 6

4.13 Interaction of Psychotherapy with Response to
Desipramine. . . . . . . . . . . . . . . . . . 86

5.1 Association of Symptoms with Endogenous
Depression . . . . . . . . . . . . . . . . . . 108

vii

LIST OF FIGURES
Page
Symptoms and Characteristics of Depressive
Illness Chosen from the Schedule for Affective
Disorder and Schizophrenia . . . . . . . . . . . 49
Psychiatric and Medical Exclusion Criteria. . . . 54
Subject Progression Through Study Phases. . . . . 56
Distibution of DST Values . . . . . . . . . . . . 62

RDC Diagnosis and DST Response. . . . . . . . . . 63

viii

CHAPTER ONE

THE PROBLEM

Researchers of depressive illness have long sought to
identify distinct subgroups among the diverse phenomena
labeled depression. Historically, efforts to differentiate
depression have involved some form of clinical observation
from whidh a pool of features could be generated.
Statistical analysis was then used to identify correlated
groups of symptoms, groups of patients with common features,
or the frequency of certain symptoms in a given group. Many
of these studies were successful in deriving a clinical
distinction between endogenous and neurotic depression and
the subtypes "neurotic" and "endogenous" were adopted as
diagnostic entities in the Diagnostic and Statistical Manual
of Mental Disorders, Second Edition (American Psychiatric
Association, 1968). However, a limitation of these early
studies is that in using factor or cluster analysis, they
identified correlated groups of symptoms or groups of
patients with common features but did not look to any
external criterion, non-clinical in nature, which would aid
in the validation of such groups as existing in the actual
patient population. Statistical methods alone cannot provide
such validation: they only assist the intellect in
separating and categorizing phenomena and do so irrespective

of the empirical context within which the phenomena appear.

The need for external criteria to validate clinical
subgroups has recently given rise to new strategies in
depression research using biological markers and
anti-depressant drug treatment response (Carroll, 1982). One
of the current biological indicators receiving much research
attention is hypothalamo-pituitary-adrenal-axis (HPA)
dysfunction as indirectly measured by the 'dexamethasone
suppression test (DST). The DST, when positive, serves on
the average as a 96% accurate confirmation of the diagnosis
of endogenous depression. That is, only 4% of patients who
respond positively to the DST and who have been screened
with regard to specific psychiatric and medical exclusion
criteria (Carroll, 1981) would be expected to be
false-positive responders. Unfortunately, it also has been
reported to have an approximately 50% false-negative rate
among endogenous patients. Carroll (1982) has attributed the
large number of false- negative DST's among endogenous
depressives to the possibly heterogeneous biological
dysfunction behind endogenous depression. It has been
presumed by several researchers (e.g., Carroll, 1982, Brown
& Shuey, 1980) that patients with identical clinical
profiles could respond differently to the DST because their
underlying biological abnormalities are different. Such
clinical homogeneity coupled with biological heterogeneity
thus: one, would confound any attempt to differentiate the

.clinical features of depressed patients through DST response

and two, would preclude the possibility of eliminating a

certain percentage of false-negative DST results.

Need

While biochemical heterogeneity within endogenous
depression has received some research validation (e.g.,
Hollister, 1978), there does exist. an equally plausible
alternative hypothesis to account for the 50% false-negative
rate of the DST among patients diagnosed as endogenously
depressed. That rival hypothesis involves the specific set
of clinical criteria used for the diagnosis of endogenous
depression. If a patient is a normal DST suppressor and
does not meet one's criteria for endogenous depression, one
is reasonably confident of the clinical implication; a
clinical uncertainty, on the other hand, arises when the
patient has met the chosen criteria but is a suppressor.
Adhering to one's criteria, this latter patient is then
defined as a false-negative responder to the DST. But what
of the possible insufficenoy of the cflinical criteria one
has employed? There may exist other clinical features which,
along with those identified in previous research, may be
crucial to a profile of endogenous depression as externally
validated through the DST. This seems especially likely in
view of Klein's (1973) assertion that as a clinical entity,
"endogenomorphic" depression need not be exclusive of

certain neurotic signs and symptoms, e.g., a clearly

identifiable precipitant. Thus, in order to explore the
value of the DST in differentiating subtypes of depression,
one needs to expand the range of symptomatology so that it
is inclusive not only of the traditional features of
endogenous depression, but also of those symptoms usually

considered neurotic/reactive.

P11139036

The purpose of the current study is to investigate the
value of the dexamethasone suppression test (DST) in
l) isolating a distinct clinical profile of depression and

2) predicting anti—depressant drug treatment response.

Hypotheses

 

Research Hypothesis: Positive and negative responders to
the dexamethasone suppression test
will demonstrate different clinical
features.

Exploratory Hypothesis: Among endogenous depressives,
positive responders to the
dexamethasone suppression
test will have a better clinical
response to the noradrenergic drug,
desipramine, than will negative

responders.

Theory

The assertion that biochemical dysfunction may be
implicated in certain types of depressive disorder is not
new. The Greek physician Hippocrates believed almost 2500
years ago that the human body contained four "humors" -
blood, black bile, yellow bile, and phlegm. He stated that
the balance of these (apparently) physiological processes
was essential to normal brain functioning and that if cme
became predominant over the others, physical or mental
disease resulted. An overabundance of black bile led,
according to Hippocrates, to a deep sadness and hopelessness
he termed "melancholia" (Coleman, 1976).

Advances over the last thirty years in our
understanding of the biochemistry of depression have led to
considerably’ more complex theorizing about. metabolic
inbalance in depressive disorders. Biochemical
heterogeneity, within "melancholia" (endogenous depression)
has been postulated (Shildkraut. et al, 1978; Hollister,
1978) from :research; findings ‘which. indicate differential
pharmacologic response of patients who demonstrate low or
high levels of specific neurotransmitter metabolites, with
metabolites of norephinephrine, serotonin, and dopamine
being the major ones under current investigation. Hollister
(1978) asserts that at least six biochemical types of
depression could exist based on either increased or

decreased excretion of these metabolites. Brown, Haier, and

Qualls (1980) believe, for example, that the differential
response of DST nonsuppressors in their study to
noradrenergic medications suggests that cortisol
hypersecretion is associated with a decrease in levels of
norepinephrine.

While cortisol hypersecretion may be tied to
norepinephrine deficiency and may thus be suggestive of one
biological subtype of depression, researchers are not in
agreement regarding efforts to identify consistent clinical
characteristics of this subtype. Carroll (1982) has
investigated the sensitivity (true-positive rate, or
proportion of endogenous patients in whom an abnormal
response occurs) and specificity (true-negative rate, or
proportion of nonendogenous patients in whom a normal
response occurs) of the DST in the context of the Research
Diagnostic Criteria (Spitzer and Endicot, 1978) category of
endogenous depression. and. the Diagnostic and Statistical
Manual for Mental Disorders III (American Psychiatric
Association, 1980) diagnosis of major depression with
melancholia. He has asserted that no consistent clinical
profile emerges in association with DST nonsuppression that
would assist the researcher in increasing the DST's average
50% sensitivity rate. He states further that because the DST
is a dynamic challenge test and not an unobtrusive measure
of current neuroendocrine function, one may always expect a
certain percentage of endogenously depressed patients not to

have an abnormal response to that specific challenge even

though they could have some other form of neuroendocrine
inbalance which would qualify them as "endogenously"
depressed. However, attempts to cflinically define a
depressive subtype specifically linked to DST nonsuppression
may have thus far been obscured precisely because such
research has been limited to the clinical context of RDC and
DSM III diagnoses. No research has to date investigated
clinical symptomatology outside of the parameters of these
diagnostic categories which might be reliably associated

with DST nonsuppression.

Summary and Overview

 

This chapter described both the purpose and need for
this study: to explore the depressive subtyping capability
of the dexamethasone suppression test (DST) without regard
for the restrictions of symptom range imposed by current
diagnostic classification systems. Additionally, current
theory was presented about biological heterogeneity in
depression, along with discussion of the limitations of the
DST as a marker of a specific depressive subtype.

Chapter II is a review of the literature that addresses
the nature of depressive subtypes as defined by clinical,
neuroendocrinological, and drug treatment response criteria.
Chapter III contains a description of the study sample,
instruments, design, procedures, (and analysis plan. In

Chapter IV, the data analysis is presented. Chapter V

contains a summary of the study, conclusions, a review and
of recent findings in the literature, a discussion of

results and recommendations for future research.

CHAPTER TWO

REVIEW OF THE LITERATURE

At the beginning of the twentieth century, Freud
suggested that depressive disorders were a heterogeneous
group, noting that some may be physiologic in origin while
others more distinctly psychogenic: "Even in descriptive
psychiatry the definition of melancholia is uncertain; it
takes on various clinical forms (some of them suggesting
somatic rather than psychogenic affections) that do not seem
definitely to warrant reduction to a unity" (Freud, 1917,
p.124).

Since Freud's observations, many researchers (e.g.,
Lewis, 1938; Eysenck, 1970; and Kendell, 1976) have
attempted to differentiate the collective phenomena of
depressive illness into psychogenic and somatic subtypes,
more recently referred 11) as the "neurotic" versus
"endogenous" distinction. There have been four basic
approaches to the separation of these subtypes: one, the use
of clinical features alone; two, clinical features and a
biochemical criterion; three, clinical features and
treatment response; and four, a biochemical criterion and
treatment response. In the review to follow, these
approaches to the subtyping of depression will be examined,
with heaviest emphasis placed on the literature dealing with

clinical features and a biochemical criterion.

Clinical Features of Depressive Subtypes

In) until the last. decade, most studies. designed to
investigate the neurotic/endogenous distinction relied on
data gathered through clinical observation. Typical of such
studies is one by Kiloh and Garside (1963). They conducted a
factor analytic study with 53 endogenous and 61 neurotic
depressives diagnosed with "reasonable confidence" through a
psychiatric interview. Product moment correlations were
calculated between 35 clinical features and a summation
factor analysis was carried out. Two factors were derived,
one a general depression factor, the second a bipolar factor
whose loadings were very similar to the correlation
coefficients between diagnosis and each feature. The bipolar
factor thus provided a clear differentiation between
neurotic and endogenous depression. The following were
clinical features which correlated significantly (P<.05)
with diagnosis, listed here in order of magnitude of
correlation and according to diagnosis suggested by presence
of feature. Endogenous depression: early awakening,
depression worse in morning, distinct quality of depression,
retardation (used inclusively' to describe the subjective
experience of slowness of thought and action and objective
psychomotor slowing), duration one year or less, age 40 or
over, depth of depression (undefined), failure of
concentration, weight loss of 7 pounds or more, and previous

-attacks. Neurotic depression: reactivity of depression,

10

11

presence of precipitants, self-pity, variability of illness,
hysterical features, inadequacy, initial insomnia,
depression, worse in evening, sudden onset, irritability,
hypochondriasis, and obsessionality. It should be noted that
the authors do not report to what degree the absence of any
particular feature under one diagnostic category is
suggestive of the other diagnosis, nor do they provide a
composite correlation of all endogenous or neurotic features
with diagnosis.

The concept of endogenous depression was reviewed by
Rosenthal and Klerman (1966) and found to consist
historically of three elements: a particular clinical
pattern of symptoms and signs, a relative absence of
environmental precipitation and environmental influence on
the course of illness, and a relatively well-adjusted
premorbid personality. In another study, Rosenthal and
Gudeman (1967) conducted their own clinical investigation of
depressive symptom patterns and specifically tested the
concept of an endogenous pattern of depression. The patient
sample consisted of 100 acutely depressed women between the
ages of 25 and 65, about half of whom were inpatients.
Patients evidencing signs of schizophrenia, organic brain
disease, sociopathy, or alcoholism were excluded. Thirty two
symptom areas, historical background and personality traits
were assessed. The resulting endogenous symptom cluster
included lack of reactivity to the environment, feelings of

worthlessness, retardation, difficulty in concentrating,

12

sadness, guilt, visceral symptoms, agitation,
middle-of-the-night insomnia, loss of interest, and a
subjective quality of depression as different from normal
experience. Patients with this group of symptoms tended not
to show environmental precipitants (- 0.23 correlation) to
their depression but did have certain premorbid
characteristics, including obsessional and depressive
personality traits, lack of emotional reactivity, and a
history of previous depressive and manic episodes. While
this study replicates the results of several previous factor
analytic studies (Rosenthal and Klerman 1966; Kiloh and
Garside 1963; and Hamilton and White 1959) and is
representative of them, Rosenthal and Gudeman caution that
factor analysis describes patterns of symptoms Ibut says
nothing about groups of patients, i.e., the symptom pattern
is an intellectual construct which may or may not exist in
any actual group of patients. Additionally, the authors
comment that the term "endogenous" has been misused in the
United States, often applied to the presence or absence of a
precipitant regardless of the symptom pattern. They maintain
that the endogenous pattern does not have the total absence
of an environmental precipitant as its most important
criterion and that "endogenous" should more correctly refer
to the tendency of the illness to run its course once it has
fully developed, with a lack of reactivity to the
environment. They thus suggest that- the descriptor

"autonomous" might be preferred over "endogenous".

l3

Klein (1973), too, has noted that the term "endogenous"
has had different meanings for different investigators. He
therefore advocates that a more inclusive term,
"endogenomorphic", be used to label patients who exhibit the
traditional symptom pattern of endogenous depression,
whether or not precipitants are present. He does not thereby
suggest that the precipitant factor be simply glossed over,
but rather hypothesizes that the precipitant or reactive
component is conceptually independent of the endogenous
profile in depressive illness. That is, as Klein remarks,
"... the endogenomorphic depressions are conceptually
divided into endogenous depressions and. precipitated
endogenomorphic depressions" (p. 449), the former without a
clear precipitant and the latter with a pmecipitant while
still characterized by traditional endogenous features.

Akiskal (1983) asserts that closer scrutiny needs to be
given to the long-terni profile of depressive disorders.
Major depressive disorders like endogenous depression most
often occur within the context of chronic, low-grade
dysthymia and neurotic features. Among his proposed chronic
depressives subtypes, Akiskal postulates that episodes of
primary major depression can either have a late onset
pattern of major depression with residual chronicity, or an
early, insidious onset (usually before age 25) of
lower-grade depression leading to episodes of major
depression followed by a fluctuating course of low-grade

depression and major depressive episodes. While commenting

14

that endogenous depression can ".... serve as ea breeding
ground for neurotic personality developments", Akiskal
appears to relegate neurotic/reactive components of
endogenous depression to the status of concurrent features
distinct from the core profile of affective illness.
However, given the complexity of the nosological framework
of depression which Akiskal has himself constructed, with
all of its admixture of affective and non-affective
components, and given the lack of long-term empirical
evidence to verify his hypotheses, he may be premature in
discounting the usefulness of "neurotic" symptomatology in
deriving subtypes of affective illness.

Nelson and Charney (1981) provide an extensive review
of the major literature on endogenous depression. They cover
four major types of studies: multivariate analysis, symptom
frequency, instrumental measures, and treatment response. Of
particular significance to the examination of a clinical
profile of endogenous depression are the multivariate
analysis studies, which include the statistical methods of
factor analysis, cluster analysis, and discriminant function
analysis. In their discussion of 20 factor analytic studies
which they reviewed, the authors comment that while these
studies vary in item definition, rating methods, and sample
size and thus do not have comparable symptom loadings, the
consistent association. of’ a symptom.‘with the endogenous
factor under these varied conditions in fact increases the

generalizability of the finding. Symptoms with strong factor

15

loadings ().50) in the majority of the studies reviewed
were psychomotor retardation, lack of reactivity and
severity of depressed. mood; symptoms 'with strong factor
loadings in 50% of the studies and moderate loadings (.49 to
.30) in the other half were loss of interest, declusional
thinking, distinct quality of mood, guilt, agitation, and
morning worsening. Low factor loadings (< .30) were found
for a group of symptoms which have traditionally been used
as criteria for the diagnosis of major depressive illness-
suicidal thinking or attempts, weight loss, difficulty
falling asleep, and midnight awakening. Loss of appetite and
loss of energy, presumed by the Research Diagnostic Criteria
(Spitzer, Endicott and Robbins, 1978) and the Diagnostic and
Statistical Manual of Mental Disorders, Third Edition
(American Psychiatric Association, 1980) to be criteria of
major depressive illness, have not received much attention
in factor analytic studies (only 2 studies used these
criteria).

Nelson and Charney (1981) found. 9 cluster analytic
studies of major depressive illness, 8 of which identified
an endogenous cluster. The ninth study, to be noted,
involved a highly select sample of chronic depressives of
advanced age- variables not usually associated with
endogenous depression. As with the factor analytic studies,
severe depressed mood, retardation and lack of reactivity
had strong association with the endogenous depressive group

in the cluster analytic studies. Three symptoms of moderate

16

loading in.1flua factor analytic studies- guilt, agitation,
and delusional thinking- received similar support in the
cluster studies, as did, although to a lesser extent, the
symptoms of loss of interest, early morning awakening,
morning worsening, difficulty concentrating and mid-night
awakening; distinct quality (which received a moderate
loading in the factor studies) and weight loss (which had a
low loading) received support in only two studies. Review of
discriminant function analysis and symptom frequency studies
did not yield consistent findings. The four discriminant
function studies examined all dealt with a
psychotic/neurotic distinction rather than a non-psychotic
endogenous/neurotic differentiation. The only consistent
finding in these studies was that the symptom of psychomotor
change (retarded or agitated) had a heavy loading for
distinquishing psychotic from neurotic depressives. The data
presented by Nelson and Charney on four symptom frequency
studies revealed no consistent symptom frequencies across
all four studies. In summary, the authors conclude from the
multivariate and symptom frequency studies that psychomotor
change is the symptom having the single strongest
association with endogenous depression. Strong association
across several type of studies were also found for severity
of depressed mood, lack of reactivity, depressive delusions,
self-reproach and loss of interest. Distinct quality of
mood, diurnal morning' worsening, and difficulty

concentrating received moderate support, although it is

17

noted that further research is needed before these symptoms
can be deemed as appropriate for inclusion in an endogenous
profile as the other symptoms mentioned above.

Nelson and Charney (1981) report three major findings
in their review of objective instrumental measures of
depressive symptomatology: one, sleepl awakening' does not
appear useful as a criterion for distinquishing endogenous
from neurotic depression; two, telemetric recording of
depressed patients' motor activity indicates a psychomotor
change (either retarded or agitated), although clear
differences between endogenous and reactive patients have
not yet been found; and three, decreased ability to
concentrate, as measured by psychological testing, has not
been supportive of presumed differences among subtypes of
depression.

Endicott and Spitzer (1977) conducted a two-year
follow-up study of 33 patients rated endogenous major
depressive disorder and 21 patients diagnosed non-endogenous
major depressive disorder using the Research Diagnostic
Criteria (RDC). The Research Diagnostic Criteria, developed
by Spitzer, Endicott and Robbins (1978), delineate sets of
specific inclusion. and exclusion criteria for functional
psychiatric disorders of various types, with greatest
attention paid to the subtyping of affective disorders like
depression (see Appendix F for a comparison of RDC Major
Depressive Disorder, Endogenous Subtype and DSM III Major

Depressive Disorder with Melancholia). The authors tested

18

the hypothesis that endogenous depressives have less
residual symptomatology after a major episode than those
nonendogenous. Although not statistically significant, the
data trends indicated this hypothesis to be false. 81% of
nonendogenous vs. 72% of endogenous patients showed Global
Assessment Scale (GAS) ratings above 60, indicative of
minimal symptomatology or impairment in functioning. 39% of
endogenous vs. 24% of nonendogenous patients were still too
impaired to work at the time of follow-up. Endicott and
Spitzer acknowledge the tentative nature of their findings
given the small sample they used but comment that the
clinical lore suggesting that the presence of endogenous
features is a good prognostic sign should be more thoroughly
examined.

Matussek, Soldner and Nagel (1981) purport that the
validity of a diagnostic syndrome like endogenous depression
is questionable if it can only be diagnosed by clinical
methods. They attempted to confirm the existence of an
endogenous syndrome by ‘using' both symptom frequency and
cluster analyses on a sample of 198 subjects who had
previously been hospitalized for depression. Subjects were
selected according to fbur criteria: 1) they were between
the ages of 50 and 65 at the time of the interview; 2) they
had no signs of organic brain damage; 3) their depression
was not related to alcohol or drug use; and 4) they
exhibited no symptoms of schizophrenia. The 198 subjects,

27% male and 73% female, were classified by the RDC into two

19

categories: endogenous depression (57%) and neurotic
depression (29%); 14%(29) remained unclassified because they
did not fit the criteria for either group. Using a catalogue
of 38 symptoms and characteristics of the course of illness,
two or more evaluators rated retrospectively a subject's
most recent depressive episode.

Only 4 of 37 symptoms differed significantly between
the sexes: males were more openly aggressive, had more of a
delayed insomnia, more often had a sudden onset to their
depression, and did not have as much weight loss as women.
Comparison of the symptom frequency between the endogenous
and neurotic depressive groups yielded eight symptoms as
significantly more related to endogenous depression. These
included morning worsening, non-reactivity, short duration,
distinct quality of mood, psychomotor retardation,
indecisiveness, sudden onset and delusions. The neurotic
depression group had two significantly higher symptoms,
sadness and neuroticism (as measured by the Maudsley
Personality Inventory).

The cluster analysis determined eight symptoms as being
characteristic of the endogenous depressive syndrome:
distinct quality of the depressive mood, loss of reactivity,
withdrawal from social contact, inhibition, disturbance of
the circadian rhythm, physiological disturbances (appetite
loss, weight loss or sleep disturbances), typical
characteristics of the course (sudden onset, relatively

short duration, remission in the interval), and absence of a

20

precipitant. Matussek, Soldner and Nagel note that while
absence of a precipitant is not a necessary criterion for
the diagnosis of endogenous depression, when no precipitant
is in fact indicated, that absence is significant (P .001).
They conclude that retrospective use of a statistical
procedure like cluster analysis can detect an endogenous
depressive syndrome, although that syndrome does not appear
to be defined by the presence or absence of single items,
nor by a syndrome with distinct boundaries. The authors
suggest that the construct of an endogenous "component" with
varying levels of strength might be applicable to all
depressive illness.

Feinberg and Carroll (1982) have derived a
discriminant index to classify depressed patients as
endogenous or non-endogenous. One hundred sixty-five
patients were initially separated into endogenous and
non-endogenous groups using Carroll's (1980) diagnostic
criteria drawn from the Schedule for Affective Disorders and
Schizophrenia (Endicott. and. Spitzer, 1977), clinical
interviews, the l7-item. Hamilton. Depression. Rating' Scale
(Hamilton, 1960) and response to treatment. The Schedule for
Affective Disorders and Schizophrenia (SADS) is a clinical
interview instrument from which RDC, DSM III, or Hamilton
ratings can be derived. Only those patients with a Hamilton
score of 10 or more were selected for the discriminant
function analysis group to validate the clinical diagnostic

classifications (used during the initial assessment phase of

21

the study). To insure that severity of illness did not
account for the differences between groups found in the
discriminant function, the authors regressed total score of
the Hamilton, a severity index, on each clinical variable
and used the residuals after regression as the clinical
variables in a new discriminant analysis. They acknowledged
that this method of adjustment was not correct since some
clinical variables are items in the Hamilton, but stated
that the error introduced was toward reducing the
contribution of the clinical features in the discriminant
function and that, therefore, their method was acceptable.
The item weights and cutting scores from the discriminant
functions were then converted from adjusted data
(coefficients) to integers (multiplying raw data by 10 and
rounding to the nearest whole number). The index derived
for distinquishing unipolar endogenous from non-endogenous
has eight items, each followed by its respective weight and
scoring range: decreased appetite (9,0-2), guilt (6,0-4),
agitation (4,0-4), (affective) delusions (3,0-8), work and
interests (3,0-4), retardation (2,0-4), loss of pleasure
(2,0-2) and precipitants present (-6,0-l). Cross-validation
of the discriminant function index was conducted with a
separate group of 52 patients, each meeting the same
diagnostic criteria used with the orginal analysis group.
All clinicians involved with the second group were blind to
the discriminant function which had been derived. Correct

classification through the discriminant index (DI) of the

22

analysis group and the cross-validation group was high: 82%
and 81%, respectively. The DST was not used in this study as
a discriminant but only to verify that the DI group had
frequencies of non-suppresion (sensitivity) similar to that

of groups with clinical diagnostic classification.

Psychoneuroendocrinological Evidence for Depressive Subtypes

 

Efforts to specify a consistent profile of endogenous
depression through clinical features alone have been
hampered by the lack of an Archimedean point that all
researchers could agree upon. The currently most promising
source for such an external criterion is the field of
psychoneuroendocrinology. Ettigi and Brown (1977) have
reviewed the recent literature on the neuroendocrine
abnormalities involved in affective disorders. Research
relevant to depressive disorders has revealed many potential
candidates for biological dysfunction which could contribute
to affective and. Ibehavioral symptomatology. Cortisol
hypersecretion linked to hypothalamo-pituitary-adrenal axis
(HPA) dysfunction is one of the areas which has received the
most attention. Previously thought to be the result of such
factors as stress, anxiety, or depressive decompensation,
cortisol hypersecretion has in more recent studies been
found to occur in apathetic patients or even in patients
while asleep, and thus is indicative of a more fundamental

dysfunction. 'Most investigation. of HPA. abnormalities

23

reflected in cortisol hypersecretion has involved the
dexamethasone suppresion test (DST). Dexamethasone is a

synthetic corticosteroid which when adminstered to a normal

subject leads to the suppression of pituitary
adrenocorticotropic hormone (ACTH). Depressed subjects,
however, fail to ShOW’ normal suppression following the

administration of dexamethasone and have significantly
higher post dexamethasone cortisol blood plasma levels than
normals.

Another group of corticosteroids which have been
researched in connection with depression are the
17-hydroxycorticosteroids (17-OHCS). Early studies showed
high levels of 17-OHCS in the urinary excretion of
depressives. However, the studies reviewed by Ettigi and
Brown (1977) indicate that 17-OHCS levels: are: one, not
consistently correlated. with severity' of' depression; and
two, do not correlate highly with level of cortisol
secretion. Similarly, 3 methoxy-4-hydroxyphenylglycol
(MHPG), the principle metabolite of brain norephenephrine,
has thus far proved to be an inconsistent correlate with
depressive disorders. A more recent study by Hollister et a1
(1980) not included in the Ettigi and Brown review
highlights the problems with MHPG levels and depression. In
their study of nortriptyline response in patients with low
or normal-high excretion levels of MHPG, they were unable to
find.51 significant relationship between improvement during

nortriptyline treatment and initial MHPG levels.

24

Additionally, the researchers note that the collection of
urine for MHPG testing requires the most careful supervision
in an inpatient setting and that the excretion level of MHPG
varies considerably within patients.

A variety of other measures of neuroendocrine
disturbance have been investigated- e.g., growth hormone,
thyroid-stimulating hormone, prolactin and luteinizing
hormone- but will not be examined here because of the
relatively' few' carefully controlled studies conducted in
these areas.

Finally, Ettigi and Brown (1977) comment that recent
research evidence taken together has served to increase the
complexity of a biochemical model of affective disorders.
The often cited catecholamine hypothesis of depression
proposes that a relative deficiency of, or imbalance
between, catecholamines and indoleamines (notably
norepinephrine and serotonin) is linked to depression.
Ettigi and Brown assert that this hypothesis is a gross
oversimplification and that the "... simultaneous effects of
other biogenic amines, hormones, and ionic changes ‘will
ultimately be included in any comprehensive formulation of
the biochemistry of affective disorders" (p. 498).
Nevertheless, the authors do not discount the practical
value of attempting a reclassification of depression based
on the metabolic activity of specific biogenic amines,
(e.g., ACTH as measured by the dexamethasone suppression

test and MHPG as determined by urinary excretion levels),

25

although the clinical utility of this latter procedure is
still open to question.

In a recent study using 122 depressed patients
classified by RDC criteria, Schatzberg et al (1983) found no
differences in mean urinary MHPG levels between unipolar
depressed patients and control subjects. Patients with
unipolar depression did appear, however, to have a wide
range of MHPG levels, from low to intermediate to very high.
The authors suggest that differences in patient sampling in
previous studies may have accounted for the variation in
MHPG levels found from study to study. They further
hypothesize that different levels of MHPG may correspond to
at least three subtypes of unipolar depression and that "
... specific clinical characteristics may be associated with
these biological differences" (p. 473).

Carroll, Feinberg, Greden et al (1981) administered the
dexamethasone suppression test to 438 subjects in an attempt
to standardize the test for the diagnosis of major
depressive disorder with melancholia (see Appendix E for a
comparison of Criteria for RDC "endogenous" and DSM III
"melancholia"). Using DSM. III criteria derived from the
Schedule for Affective Disorders and Schizophrenia, 215
patients (both outpatients and inpatients) were diagnosed
major depressive disorder with melancholia, 100 with
nonendogenous depression, 53 with other psychiatric
disorders (e.g., schizophrenia, personality disorders) and

70 as normal. Severity of depression was clinician rated

26

with the Hamilton Depression Rating Scale (HDRS) and
self-rated through the Carroll Rating Scale for Depression
(CRSD).

Using marginal and interval probability analyses, the
authors compared the diagnostic performance associated with
plasma cortisol criterion values of 3, 4, 5, and 6 ug/dl for
the 368 patients with psychiatric diagnoses. 8 a.m., 4 p.m.,
and 11 p.m. postdexamethasone blood samples were taken with
inpatients, while only a 4 p.m. sample was drawn for the
outpatient series. Reviewing the DST results, the
researchers propose that a plasma cortisal value of 5 ug/dl
be used for the diagnosis of melancholia. This criterion
gave an overall test sensitivity (true-positive rate) of 43%
and a specificity (true-negative rate) of 96%. The
diagnostic confidence (proportion of abnormal test results
that are true-positive) for melancholia with a greater than
5 ug/dl blood cortisol criterion was 94%.

Carroll et a1 comment that, because of some variability
across patients in the cycling of cortsol secretion, the
inpatient serial blood tests at 8 a.m., 4 p.m., and 11 p.m.,
can always be expected to generate a greater DST sensitivity
to cortisol hypersecretion and that the 4 p.m. blood
sampling alone used with outpatients is a practical
compromise. A rather dramatic difference in sensitivity
apparently occurs if one administers a l-mg as opposed to a

2-mg dosage of dexamethasone: there is a 72% and 188% gain

27

in sensitivity for inpatients and outpatients, respectively,
when using the l-mg dose.

The researchers found that approximately 50% of the
patients with melancholia had plasma cortisol concentrations
equivalent to those of patients with other psychiatric
diagnoses and of normals. They thus remark that while a
positive DST result can be used with high confidence to
support a diagnosis of melancholia, a negative DST result
should not be considered a criterion to rule out
melancholia. No differences were found between DST
suppressors and non-suppressors in HDRS scores, in CRSD
ratings, in age (mean age 48), in sex, or in recent history
of psychotropic drug intake. Carroll et a1 theorize that the
only 50% hit-rate of the DST for patients diagnosed as
melancholic/endogenous (DSMIII/RDC) may be indicative of a
certain neuroendocrine heterogeneity within melancholia that
needs to be more fully explored.

Carroll (1982) has recently reviewed eight studies
which attempted to differentiate endogenous depression
(variously defined) from non-endogenous depression or from
other psychiatric diagnoses labeled "miscellaneous" (by
Carroll). An additional two studies ‘were examined. which
compared 1) primary unipolar depression with secondary
depression and. 2) primary unipolar depression. with
miscellaneous comparison patients. The 10 studies with a
total of 573 subjects yielded an average DST sensitivity of

45% and an average specificity of 96%. The author notes that

28

variations in clinical diagnostic criteria could have
affected the results with DST sensitivity and advocates a
greater uniformity in the use of criteria for the diagnosis
of melancholia or endogenous depression.

Further reviewing his and others' research with the
DST, Carroll remarks that the predictive use of an abnormal
DST result for good response to anti-depressant medication
is still unclear. The preferential effectiveness of certain
anti-depressants in patients with abnormal DST results is
also still being explored. However, there is already an
indication that the DST can serve as valuable confirmation
of a positive clinical outcome. In a study of
electroconvulsive therapy (ECT) with melancholia patients,
Albala, Greden, Tarika and Carroll (1981) report that a
pre—ECT abnormal DST response will convert to normal,
post-ECT, before clinical improvement is noticed. Similarly,
an initially abnormal DST response that has not converted to
normal after pharmacotherapy may indicate that the patient
is at a serious risk of early relapse.

In. their' review? of an intensive single case study,
Rothschild and Schatzberg (1982) caution that while the DST
was useful in determing the biological response to ECT and
tricylcic antidepressant treatment, normalization of a DST
does not necessarily mean that a patient is on the road to
full recovery. Their study patient, in fact, relapsed twice,
one time reverting to a nonsuppression pattern within two

weeks after showing suppression, the other after four weeks.

29

In view of their experience, the authors suggest that serial
DSTs should be conducted for several weeks after initial
re-conversion to DST suppression.

Carroll's reporting of the sensitivity and specificity
of the DST has not been without recent disconfirmatory
evidence. Coryell, Goffney, and Buchardt (1982) conducted a
study of 65 inpatients and their response to the DST, the
patients falling within one of 3 categories: 17 (26.2%) had
major depression without melancholia, 34 (52.3%) had major
depression with melancholia, and 14 (21.5%) had depression
with psychotic features. A second rater gave 50 (76.9%)
patients the diagnosis primary depression and 15 (23.1%)
secondary depression (primary depression following the DSM
III definition of no other type of psychiatric illness
evident prior to the depressive disorder). Of the 50
patients with primary depression, 26 (52.0%) had melancholia
and 13 (26%) had psychotic features; the secondary group
contained 8 (53.3%) patients with melancholia and 1 (6.7%)
with psychotic features. Subjects were given a 1-mg dose of
dexamethasone at 11 p.m. and had their blood drawn the
following day at 8 a.m. and/or 4 p.m.; most patients
reportedly had plasma taken at both times. Any patient with
a post DST level greater than 5 ug/dl at either time period
was considered a nonsuppressor.

Although not statistically significant, the rate of
nonsuppression was actually higher among patients without

melancholia than it was for patients with melancholia. 6 of

30

17 (37.3%) patients without melancholia were nonsuppressors,
while 9 of .34 (26.5%) patients with melancholia were
nonsuppressors, as were 7 of 14 (50.0%) patients with
psychotic features. A striking contrast was noted in
frequency of abnormal DST results among the patients when
rated according to the primary-secondary distinction. 22
(44.0%) of the patients with primary depression were
nonsuppressors but none of those with secondary depression
were nonsuppressors. The difference, computed in a
chi-square statistic, was significant (p< .005). In their
discussion of their findings, Coryell et a1 comment that
Carroll's (1982) review article cited studies that used a
global conceptualization of melancholia rather than one
characterized by specific, operationalized criteria like
those in the DSM III. While not accusing Carroll and his
.group of such inprecision, the authors conclude that the
vagueness of many prior studies limits the application of
their own results to these studies.

While the Coryell et a1 study may present findings
genuinely discrepant with those of Carroll, recent research
conducted by Amsterdam, Winokur, Caroff and Conn (1982) may
offer challenging results of less credibility because of its
serious methodological flaws. In their study, 46 women and
18 men fulfilling RDC and Feighner and associates criteria
for primary affective disorder were given the DST. 41
patients were diagnosed as primary unipolar depression and

23 as bipolar illness. All depressed patients had a Hamilton

31

Depression Rating Scale score of 16 or above on the 17-item
scale. The authors emphazise that subjects were drug free
for at least. 7 days prior to the DST; patients taking
neuroleptics had been stopped for 2 weeks. Standard medical
conditions and illness complications which may potentially
offset the DST result were carefully screened out. The DST
itself involved a 1-mg dose of dexamethasone taken orally at
10 p.m., with blood sampling the following day at 8 a.m. and
4 p.m. Cortisol was measured by means of a single anti-body
technique with a radioimmunoassay, using post-DST cortisol
levels greater than 5.0 ug/dl as the criterion for
nonsuppression.

Results indicated that patients and controls did not
differ significantly in mean serum cortisol levels at either
8 a.m. (n: 4 p.m. postdexamethasone. No significant
intergroup differences were found in the distribution of
suppressors and nonsuppressors ( X2 = .6, p>>.20), e.g., in
the areas of age, sex, length of illness or (for depressed
patients) severity of illness. Nonsuppressors included 10
(24.4%) of 41 unipolar depressed patients, 5 (29.4%) of 17
bipolar depressed patients and 1 (16.7%) of 6 bipolar
patients in. a hypomanic phase. The overall rate of 31
nonsuppresion at 4 p.m. among depressed patients was 25.9%,
compared with 15.1% of the "healthy" volunteers. Since
Amsterdam et a1 did not subgroup the primary unipolar
depressives as endogenous/nonendogenous, it is impossible to

say how their relatively low DST sensitivity for depressed

32

patients would compare with the average sensitivity of 45%
reported by Carroll (1981) for the DST with melancholia. The
authors, however, do make the mistake of comparing their
rate of sensitivity when using a 1-mg dose of dexamethasone
with the findings of studies using a 2-mg dose, apparently
attempting to justify the low frequency of nonsuppression
among their sample. As previously mentioned, Carroll (1981)
has found that the dose of dexamethasone has a strong effect
on the sensitivity of the DST. For the specific DSM III
diagnosis of major depressive disorder with melancholia, he
reports that among inpatients the sensitivity of the DST was
67% with 1 mg and 39% with 2 mg, significantly different by
X2 analysis at the p (.005 level. Carroll observed, by
constrast, that the specificity of the DST was not affected
by the dose of dexamethasone. This latter finding again
seriously calls into question the sampling or the
experimental methodology employed by Amsterdam et a1. They
had a 15.1% nonsuppression rate among their control group of
normals, far in excess of the average 4% of normal persons
noted by Carroll as having false-positive DST results.
Peselow, Goldring, et a1 (1983) recently conducted a
DST study with depressed outpatients, the results of which
do not confirm the findings of Amsterdam et a1. Eighty-eight
outpatients with primary affective disorder (unipolar or
bipolar), meeting RDC criteria for major depressive episode

and having Hamilton Depression Rating Scale scores of 16 or

greater participated in the study. Patients were further

33

divided into definite endogenous (53 patients) and
non-endogenous (35 patients) groups following the RDC, along
with a control group of 49 normals (35 men, 14 women).
Observors (presumably in the subject's family) were assigned
to verify that the subject took 1 mg of dexamethasone at 11
p.m., with blood drawn the following day at 4 p.m. to
determine cortisol levels. The authors comment that it was
impractical and unwise for 41 of the study patients to
discontinue their medications and be drug free a minimum of
5 days before testing. Those taking medications were taking
lithium or antidepressants or a combination of the two. The
cortisol levels of all subjects were measured against three
criterion values- 3 ug/dl, 4 ug/dl and 5 ug/dl- to further
explore the question. of the appropriate plasma cortisol
criterion for a positive DST with outpatients. Results
indicated that at all cortisol criterion values there was a
greater frequency' of .non-suppression. among' patients ‘with

2 ranged from p (.02 to

primary affective disorder (X
p (.001) and that patients in the endogenous subgroup had
significantly higher mean cortisol levels than controls
(p<:.02). No significant difference in DST sensitivity rates
was .found between those subjects on medication or off,
although this potentially biasing factor was not examined
with respect to a patient's subgroup membership
(endogenous/nonendogenous). In addition, unlike the findings

of Amsterdam et a1 (1982), the current study's data indicate

that only 4% of normals were nonsuppresSors at the 5 ug/dl

34

or above criterion (consistent with Carroll). However, the
authors mention only in passing the very significant finding
that no differences were found. in mean cortisol levels
between the endogenous and nonendogenous subgroups or
between the nonendogenous subgroups and controls.

Brown and Shuey (1980) conducted a study to assess the
depression subtyping capability of the dexamethasone
depression test. Forty-eight hospitalized patients meeting
RDC criteria for major depressive disorder were selected. In
addition to the DST, subjects were administered the Hamilton
Rating Scale for Depression (HRS), the Zung Self-Rating
Depression Scale (SRS) and the Profile of Mood States
(POMS). Ratings of tension-anxiety and depression-withdrawal
were made for blood draws at 8 a.m., 4 p.m., and 11:30 p.m.
following a midnight ingestion of 2 mg of dexamethasone and
baseline blood sampling of the previous night. Results of
study showed that 9 patients (50%) with primary depression
(following RDC criteria) were dexamethasone nonsuppressors
while only 6% of those with secondary depression were
nonsuppressors; 5 patients or (35%) of the nonsuppressors
but only 2 (or 5%) of the suppressors met RDC criteria for
endogenous depression. (Percentage figures are drawn from a
group total. of’ 18 jprimary' depressives, and. 29 secondary
depressives). An analysis of the clinical characteristics of
suppressors and nonsuppressors revealed that, as drawn from
HRS scores, diurnal variation was greater in the suppressors

and nonsuppressors showed greater helplessness. In addition,

35

five (45%) of the nonsuppressors were unable to complete the
self rating forms while only two (5%) of the suppressors
were unable to complete these forms. Suppressors and
nonsuppressors did not differ significantly in any other
clinical characteristics, e.g., anxiety, agitation,
retardation, insomnia, frequency of physical illness, nor in
anxiety-tension or depression-withdrawal ratings made at the
three of each blood sampling.

In the Brown and Shuey study, both suppressors and
non-suppressors were treated with a variety of medication,
most receiving tricyclic antidepressants. Nine (82%) of the
nonsuppressors were rated as having' a good response to
treatment (marked improvement, or return to premorbid
functioning) while only 14 (39%) of the suppressors had a
good response to treatment. Poor and good responders did not
differ systematically in treatment regimen, nor did response
treatment correlate with length of hospitalization. The
authors note that the results of treatment are Open to
question because the present research was not designed as a
response study and consequently did not control for
extraneous influences on drug treatment response. Separation
of clinical differences between DST positives
(nonsuppressors) and DST .negatives (suppressors) was
confounded by the variable primary/secondary; this latter
variable needs to be held constent if one wishes to explore
the effect of the variable, DST response. In addition, given

the small sample of nonsuppressors (9 primary, 2 secondary),

36

any observations about no differences between suppressors
and non-suppressors are inconclusive.

In a study by Schatzberg et a1 (1983), 88 patients (77
inpatient, 11 outpatient) meeting any of five possible DSM
III diagnoses- major depressive disorder, bipolar depressive
disorder, dysthymic disorder, situational disorder with
depressed mood and borderline personality organization- were
examined for differential cortisol levels following
adminstration of the DST. A control group of 31 "medically
and psychiatrically healthy" subjects free of medications
interactive with the DST was included in the study design.
With a criterion for nonsuppression of 5 ug/dl or more, it
was found that only 1 of 31 controls (3%) failed to
suppress, compared with 41 of the 88 identified patients
(47%). Among patients diagnosed with major depressive
disorders, the frequency of nonsuppression was somewhat
higher in mood-congruent psychotics (10 of 14 or 71.4%) as
opposed to nonpsychotics (18 of 31, or 58.1%). 7 of 9
psychotic major depressives had post-dexamethasone cortisol
levels of 15 ug/dl or more, suggestive of a distinct
depressive subgroup. Additional DST level clusterings at 2
ug/dl and 10 ug/dl indicated two other possible biological
subgroups. The authors comment, however, that further
research is needed to determine if these subgroups can be
discriminated on the basis of other biological measures,
clinical features (including severity of illness), and

response to treatment.

37

Drug Treatment Response and the Subtyping of Depression

A further source of biochemical differentiation for
depressive subtypes is the area of drug treatment response.
In an extensive review of the literature on tricylcic
antidepressant response, Bielski and. Friedel (1976) note
that while the pharmacologic evidence weighs in favor of
depression being divided into endogenous and neurotic
subtypes, support also exists for Klein's (1973)
endogenomorphic group, inclusive of some neurotic features,
as being tricylcic responsive. The authors comment that
clinical "lore" has often claimed that amitryptyline, which
blocks serotonin uptake, is most effective with agitated and
severely depressed patients; imipramine, metabolized as
desipramine and blocking noradrenaline uptake, on the other
hand, is presumed more effective with clients exhibiting
psychomotor retardation. They found, however, that the
clinical features associated with positive response to
amitryptyline and imipramine greatly overlapped: insidious
onsets weight. loss, :middle/late insomnia. and. psychomotor
retardation with imipramine response; anorexia, middle/late
insomnia, psychomotor retardation and psychomotor agitation
with amitryptyline response. Because they found that the
studies they reviewed lacked uniform diagnostic criteria,
had numerous methodologic flaws, and presented contradictory
evidence, Bielski and Friedel caution that their summary of

these studies be viewed as suggestive but inconclusive.

38

Nelson and Charney (1981) provide evidence in their
review of treatment response studies which suggests that
there may be at least two endogenous or "autonomous"
depressive states: a retarded anhedonic group and an
agitated delusional type. Retarded depression is most
responsive to tricyclics but may worsen with administration
of an antipsychotic medication (e.g., thioridozine,
perphenazine). Agitated delusional depression appears to
require both an antidepressant and an antipsychotic to
produce beneficial results; agitated depressives respond
poorly to antidepressants alone.

Prusoff et a1 (1980) conducted a study to test the
usefulness of the RDC subtypes in the prediction of
differential response to amitryptyline and short-term
interpersonal psychotherapy (IPT) in a 16-week controlled,
clinical trial using 81 ambulatory depressed patients. The
study design involved an evaluation of the efficacy of IPT
and amitryptyline each alone, in combination , and compared
with a nonscheduled treatment control group. Both patients
with a situational depression and those with an endogenous
depression responded to combined treatment. Patients with an
endogenous depression did not respond to IPT alone, whereas
those with a situational depression responded to IPT or
tricyclic medication in isolation. A methodological problem
arises, however, with the authors having dichotomized their
patient sample as endogenous or reactive. Studies by

Rosenthal and Gudeman (1967), Klein (1975), and Kendall and

39

Gourlay (1970) do not support the notion that psychosocial
stress provides a sufficient criterion for separating
"reactive" from endogenous depression. Interestingly,
Prusoff et a1 themselves note that in their study depressive
illness could be classified as both situational and
endogenous.

Stewart, Quitkin et a1 (1983) have recently tested the
value of the RDC and severity of illness (Hamilton
Depression Rating Scale- HRDS) in predicting differential
response to desipramine and placebo among mildly to
moderately depressed outpatients. 103 subjects between the
ages of 18 and 64 years, with HDRS scores between 4 and 18,
and meeting RDC criteria for major, minor or intermittent
depressive disorder were selected for the study. Exclusion
criteria, while covering the standard areas, also included
previously adequate treatment with any tricyclic
antidepressant for two weeks during the current depressive
episode. Three assessment instruments were administered at
baseline and after a 10-day period of placebo wash-out: the
Hamilton (HDRS), the Clinical Global Impression Scale (CGI)
and the self-rated Symptom Checklist (SOL-90). A CGI global
improvement rating of 1 or 2 ("very much improved" or "much
improved") was the criterion for defining a patient as a
responder.

Those patients showing nonresponsiveness at the end of
placebo therapy were randomly assigned to double-blind

treatment with desipramine or placebo. Sixteen patients were

40

placebo responders and another 23 dropped out, six before
randomization and 17 after; there were no significant
differences in droupout frequency between groups.
Desipramine dosage levels started at 50 ug daily and built
to a maximum of 300 ug daily by study day 25 and were
continued another 17 days (for a total of six weeks
monitored treatment). CGI ratings were made weekly and on
the final study day the HDRS and SCL-lO were also completed.

Results showed that patients improved significantly
more frequently with desipramine than with placebo,
(X2=4.65, p .05). Again, treatment response was defined as
a 1 or 2 rating on the CGI. The more subjective ratings of
the CGI were corroborated by responders showing
significantly lower SCL-90 scores and HDRS scores than
nonresponders. When examining Research Diagnostic Criteria
(RDC) categories for differential response to placebo and
desipramine, it was found that patients with major
depressive disorder demonstrated a significant difference
(p‘<.005), while those with intermittent depressive disorder
did not; patients with minor depressive disorder were too
few to run statistical computations. No subtype under major
depressive disorder (MDD) showed significant differences
between desipramine and placebo responsiveness, including
the endogenous subtype.

When focusing on HDRS scores, however, Stewart, Quitkin
et al found that severity of illness was significantly

related to desipramine response in patients belonging to any

41

MDD subtype. Their data reveal that a HDRS score of 14 or
above (after' a placebo wash-out period) coupled. with a
diagnosis of MDD (irrespective of subtype) is a better
predictor of desipramine response than diagnostic
classification by itself. Specific desipramine effect (drug
response minus placebo response rate) was 79% for MDD plus
HDRS of 14 or above, while only 44% for MDD alone. The
authors hypothesize that the higher specific drug treatment
effect created. when including severity of illness as a
classification variable may reflect a relatively homogenous
population whose underlying biochemical dysfunction may also
be homogenous.

Brown, Haier and Qualls (1980) examined the value of
the DST in predicting differential tricyclic response.
Nineteen patients who met RDC criteria for primary major
depressive disorder underwent the dexamethasone suppression
test and then were randomly assigned to one of two drug
treatment groups: desipramine/impramine or
amitryptyline/clomipramine. The authors hypothesized that
because some evidence exists to associate cortisol
hypersecretion with noradrenaline deficiency, one would
expect DST nonsuppressors to show a more favarable response
to impramine/desipramine (noradrenergic medications) and not
to amitryptyline/clomipramine (serotonergic medications);
suppresors, by contrast, would show an opposite pattern of
drug responsiveness. Before treatment and after one and two

weeks of treatment, patients completed the Beck Depression

42

Inventory and 'were rated on a modified version of the
Hamilton Depression Rating Scale. The authors state that DST
nonsuppressors treated with desipramine and imipramine (4
patients) showed "considerable improvement" as assessed on a
global rating scale (developed by the researchers), while
nonsuppressors (4 patients) showed no improvement (p=.029,
Mann-Whitney U.); suppressors (5 patients), on the other
hand, improved. with. amitryptyline and. clomipramine ‘while
non-suppressors (6 patients) did not (p< .026). However,
results of this study are questionable not only because of
the small number of patients used but also because of two
basic methodological flaws. One, global treatment response
was rated on a continuous scale from -2 (considerable
decline) to a +2 (considerable improvement). Yet all
suppressors were given rounded whole number ratings (from -2
to +2) while 5 of 11 nonsuppressors were given decimal
values (e.g., .6, .7, 1.4). One would suspect that raters
were either trained differently on the use of the scale or
had different rating styles with actual cases. Two, drug
treatment response was gauged at the end of two weeks when
three weeks is the more accepted (minimum) duration for
antidepressant drug treatment. Finally, Brown et al
themselves note in passing that changes in the Beck and
Hamilton ratings from baseline to week one and from baseline
to week two were not significantly different between

treatment groups.

43

Nelson, Orr, Stevenson, and Shane (1982) conducted a
study to explore the usefulness of
hypothalamic-pituitary-adrenal axis variables in predicting
antidepressant response. The subjects of the study were 28
inpatients (24 women, 4 men) meeting RDC criteria for major
depressive. disorder, endogenous subtype. Within. two <days
after hospital admission, a detailed psychiatric history was
taken and a battery of four rating scales were administered:
the 21-item Hamilton Depression Rating Scale, the Roskin
Three Area Assessment, the Brief Psychiatric Rating Scale,
and a seven-point global severity of illness scale. Ratings
were repeated four weeks later on an outpatient basis. The
HPA activity evaluation conducted when patients were in a:
drug-free state involved three components: one, a 24-urine
specimen was taken for measurement of urinary free cortisol
excretion (UFC); two, blood samples were at 8 a.m. and 11:30
p.m. pre-dexamethasone; and three, a 2 mg dose of oral
dexamethasone was administered after the 11:30 blood sample,
with post-dexamethasone blood samples drawn the following
day at 8 a.m., 4 p.m. and 11:30 p.m. Cortisol level of
urinary and serum samples were assayed in duplicate using a
double-antibody iodine 125 radioimmunoassay.
Postdexamethasone nonsuppression was defined as a cortisol
value of more than 5 ug/dl in any of the three samples taken
post-DST. After psychiatric and emdocrinologic evaluation,
all patients were randomly assigned to one of two

antidepressant treatments: 13 received. imipramine and 15

44

amitriptyline. All patients received 150 mg daily of the
chosen drug for four weeks; no drug serum level was obtained
at the end of that period.

Nelson et al computed their results statistically using
the following procedures: between group comparisons by the
two-tailed Mann-Whitney U test; within groups correlations
by the two-tailed Spearman rank-order method. While they
report that 57% of patients had at least one abnormally high
cortisol value, either baseline or after dexamethasone
administration, only five patients (18%) were
post-dexamethasone nonsuppressors. Carroll (1981) has stated
that baseline (pre-dexamethasone) cortisol values are highly
variable and are therefore unreliable for differentiating
patients from one another, unlike post-DST values which do
have a very high specificity (96%). There were almost no
significant pretreatment differences between normal and
abnormal groups on psychiatric ratings. Elevated (abnormal)
UFC excretion values did have significantly more diurnal
variation of mood (Z=2.09, p‘<.05) and less insight (Z=2.14,
p < .05) as measured on the subscales of the Hamilton, and
durinal variation was also significantly greater (z=2.57,
p< .05) in patients with elevated baseline serum cortisol
levels (either at 8 a.m. or 11:30 p.m.).

Treatment response was evaluated through comparison of
scores on the four rating scales for depression before and
after treatment. The only significant finding which emerged

was that improvement on the Brief Psychiatric Rating Scale

45

Summary and Conclusions

 

Many' clinical studies attempting' to distinguish
subtypes of depression have been sucessful in statistically
separating an endogenous depression subtype from neurotic or
reactive depression subtypes. However, not all of the
symptoms commonly associated with an endogenous profile
(i.e., in the RDC or DSM III) have received equal research
validation in the literature. Symptoms receiving consistent
support include psychomotor change, severity of depressed
mood, lack. of reactivity, depressive delusions,
self-reproach, and loss of interest.

Research into neuroendocrinological correlates with
depression has yielded some promising results. A positive
response to the dexamethasone suppression test, for example,
appears to offer highly specific confirmation of the
diagnosis of endogenous depression. But. while the DST's
true-positive rate is high (96%), its false-negative rate
among patients diagnosed endogenously depressed is also
relatively high (an average 50%). Carroll (1982) and others
have asserted that patients with clinically homogenous
profiles may respond differently to the DST because of the
biological heterogeneity underlying endogenous depression.
One may thus always expect a certain number of patients
clinically diagnosed. as endogenous depressive to show a

"false-negative" DST .resulta A. fundamental question. that

46

arises, however, is whether the DST positive and negative
endogenous depressives are, in fact, clinically homogeneous.
The assertion of no clinical differences between these two
groups has been made within the context of: l) the limited
range of symptomatology covered by RDC and DSM III
diagnostic categories and 2) research projects with very
small numbers of subjects. No study' to date has
re-investigated the possible clinical differences between
DST positive and negative groups outside of the breadth of
symptoms considered "core" endogenous, i.e., differences
inclusive of those symptoms that were considered neurotic
depressive in previous statistical derivations of an
endogenous depressive profile. Finally, while some evidence
exists to suggest that depressive subtypes respond
differentially to antidepressants, it is still unclear
whether any component of HPA activity or any profile of
clinical features can serve as a reliable predictor of drug

treatment response.

CHAPTER THREE

METHODOLOGY OF THE STUDY

In Chapter Three, a description is provided of the
clinical research instruments and the biological test
employed in this study. A presentation follows of the

research sample, procedures, and data analysis.

Instruments

 

Spitzer and Endicott (1978) and other participants in
the National Institute of Mental Health Clinical Research
Branch Collaborative Program on the Psychobiology of
Depression have developed the Schedule for Affective
Disorders and Schizophrenia (SADS), a clinical interview
instrument which they believe reduces information variance
(between research groups) in both descriptive and diagnostic
evaluation of a subject. The SADS, while designed for
companion use with the Research Diagnostic Criteria (RDC)
for formulating clinical diagnoses, is also comprehensive
enough in its coverage of symptomatology to be used to
derive Diagnostic and Statistical Manual III for
Mental Disorders diagnoses (Nelson, Charney, and Quinlan,
1981). The current study only includes Part I of the
SADS, which describes the features of the current episode
of illness from two perspectives: when the symptoms were
the most severe during the current episode and, for

many items, the severity of the symptoms for the week

47

48

prior to the interview (which may or may not also be the
time of greatest severity).

Ratings of 46 items of the SADS (see Figure I) were
chosen for analysis with a subject's response to the
dexamethasone suppression test (DST). These items comprised
all of the scalar items rated during the SADS interviews.
Items included those from the entire SADS section "Dysphoric
Mood and Related Symptoms" as well as two items regarding
general level of functioning. Analysis of all items with a
time differential was based on ratings of the patient's
clinical state during the week prior to the interview. This
restiction was made because of the state-dependent quality
of the DST: it can only serve as a marker of biological
dysfunction underlying depression when the patient is in the
midst of a depressive episode. The relationship between DST
response and clinical variables would thus be most
meaningful when ratings of the most recent clinical state
are employed. Post-DST plasma cortisol levels were assayed
using a Gammacoat (125I) radioimmunoassay kit (Clinical
Assays). This assay procedure is a more specific one than
that used by Carroll and associates (1981). The St. Lawrence
Hospital Department of Pathology (1982) has recently
conducted research which indicates that a cortisol level of

125I) is equivalent to 5.0

4.1 ug/dl using the Gammacoat (
ug/dl as derived from competitive protein binding (employed

by Carroll). Therefore, in the current study a serum

49

Figure 3-1

 

Symptoms and Characteristics
of Depressive Illness Chosen from the
Schedule for Affective Disorders and Schizophrenia

(see Appendix C)

 

Symptoms and Characteristics Sgalg
Classification of Current Condition 1-5
Period of Greatest Severity 1-5
Subjective Feelings of Depression 1-7
Distinct Quality of Mood 1-4
Association with Specific Concerns 1-4
Worrying 1-6
Self-Reproach 1-6
Negative Evaluation of Self 1-6
Discouragement about Future 1-6
Suicidal Tendencies 1-7
Panic Attacks 1-3
Somatic Anxiety 1-6
Psychic Anxiety 1-6
Phobia 1-6
Obsession or Compulsions 1-6
Insomnia (severity) 1-6
Middle Insomnia 041
Initial Insomnia 0-1
Terminal Insomnia 0-1

Sleeps More Than Usual 1-6

50

Subjective Feeling of Lack of Energy
Appetite

Weight Loss

Increase in Appetite

Weight Gain

Somatic Preoccupation
Indecisiveness

Difficulty Concentrating
Loss of Interest

Social Withdrawl
Depersonalization
Subjective Feeling of Anger
Overt Irritability
Agitation

Psychomotor Retardation
Reactivity

Mood Worse in Morning

Mood Worse in Evening
Alcohol Abuse

Libido

Drug Abuse

Antisocial Behavior
Suspiciousness
Non-Delusional Ideas of Reference
Memory Disturbance

FunCtional Impairment

1-6
1-6

1-6

51

cortisol level greater than 4.1 ug/dl was adopted as the
criterion for cortisol nonsuppression.

For those subjects diagnosed major depressive disorder,
endogenous subtype following Research Diagnostic Criteria
(RDC), response to a five-week trial of desipramine was
compared with response to the DST at initial evaluation.
Drug treatment response will be defined as the change score
between the Hamilton Depression Rating Scale score extracted
from the SADS at initial evaluation and the Hamilton score
at. the end. of ‘the five-week. desipramine trial (using a
regular 18-item Hamilton). Endicott, Cohen et al (1981) have
reported that no loss occurs when using the extracted
Hamilton from the SADS instead of the real Hamilton and
that, in fact, the extracted Hamilton is slightly more
reliable because of the use of a six-point rating of
severity for each item of the SADS. The authors provide a
conversion table which indicates how SADS item ratings are

to be collapsed to calculate a comparable Hamilton rating.

 

 

 

SADS Variable SADS Hamilton HDRS Variable
Psychomotor 0-2 0 Retardation
Retardation 3 1

4 2

5 3

6 4

The extracted Hamilton total score is then obtained by

summing the calculted scores and adding a constant of +2.

52

Subjects

The current study is essentially a post hoc examination
of data gathered through a research project entitled "Self
Reported Symptomatology in Major Depressive Illness" led by
Gregory Holmes. The author served as one of the clinical
interviewers with this project whose primary objective was
to ‘validate the Differential Diagnostic Depression. Scale
(DDDS), a self-report measure designed to distinquish
endogenous from non-endogenous forms of depression.

Subjects were recruited for the study through two main
sources: 1) referral from area inpatient psychiatric units,
psychiatric outpatient clinics or private mental health
professionals; 2) self-referral in response to direct public
solicitation in the local paper for study participants.

Initial recruitment of subjects continued until 50
individuals with endogenous depression and 50 individuals
with non-endogenous depression were obtained; an additional
seven subjects were seen and added to the original pool of
one hundred in order to allow for an increase in the number
of DST positives in the study (n=27). These diagnoses were
determined by the research team. using a combination of
criteria described in the next section (Holmes, 1982).

Subjects were between the ages of 18 and 65 and were
free of medical illnesses that might invalidate the results
of the laboratory test described in the procedures section.

Subjects were excluded if their depression was secondary to

53

other psychiatric illness. These exclusion criteria are
listed in Figure 3-2. Depressed subjects qualified for
inclusion in this study if, during the initial screening or
intake they presented with subjective complaints of
depression, dysphoria, or experience a loss of interest or
pleasure in their usual activities. Patients did not qualify
as subjects if depression was suspected by the clinician but

denied by the patient (Holmes, 1982).

Procedure

 

Initial screening of the subjects for this study
Voccurred at either time of admission to a psychiatric unit
or at time of intake at a psychiatric outpatient clinic.
Several clinical facilities in the Lansing area were
selected as primary patient sources based on their
willingness to collaborate in the study. Initial screening
of patients from these sites was conducted by the unit or
clinic staff’ as jpart. of the routine procedure used for
evaluation at the time of admission or intake (Holmes,
1982).

During the admission or intake procedure, or as soon as
it. was deemed clinically appropriate by unit. or clinic
staff, those patients who did qualify for inclusion were
informed by the staff as to the nature of the study and
asked if they would be willing to participate. At that time,
they were told that their decision to participate did not

influence their eligibility for treatment. If they were

54

Figure 3-2

 

Psychiatric and Medical Exclusion Criteria

 

10.
ll.
12.
13.

Psychiatric

l. Schizophrenia

2. Bipolar Depression

3. Organic Brain Syndrome

4. Alcoholism

5. Anorexia Nervosa

Medical*

1. Pregnancy, high dose estrogen therapy other than oral

contraceptives.

Cushing's disease or syndrome

Severe weight loss where body weight 80% of ideal weight
Hepatic enzyme induction (phenytoin sodium,
barbiturates, meprobamate)

Uncontrolled diabetes mellitus (hypoglycemia, acidosis)
Major physical illness; trauma; fever; dehydration;
nausea

Temporal lobe epilepsy; use of reserpine or narcotics
Addison's disease

Corticosteroid therapy

Hypopituitarism

High dose benzodiazepines ( 25mg/day of diazepam)

Other endocrine disease

Spironolactone therapy

*
Adapted from B. J. Carroll, M. Feinberg, J. F. Greden, et

al, 1981

55

willing to participate, consent forms were obtained for
each subject (Appendices A and B) (Holmes, 1982).

The clinic or unit staff contacted a designated member
of the research team when a qualified subject indicted their
willingness to participate in the study. With inpatients,
arrangements were made as soon as possible for a convenient
time for research interviewers to meet with the patient at
the unit. Patients initially screened during an intake at an
outpatient facility were referred directly to the Psychiatry
Clinics of the Michigan State University Clinical Center for
evaluation as a study patient. Self-referrals screened by
phone contact with Gregory Holmes were also seen for
evaluation at the Psychiatry Clinics (Holmes, 1982).

Figure 3-3 presents a flowchart of a study
participant's progression through each phase of the study.
Participants were asked to complete three procedures as soon
as possible after' the intake/admission. These ‘procedures
were as follows:

1. Each depressed subject (who passed initial
screening) was asked to complete the Differential
Diagnostic Depression Scale (DDDS).

2. Each subject was interviewed by a member of the
research team following the semi-structured format
of the Schedule for Affective Disorders and
SchiZOphrenia (Endicott and Spitzer, 1978). A
research diagnosis was then derived from the SADS

using the Research Diagnostic Criteria (RDC)

56

Figure 3-3

Subject Progression Through

 

[ Discard l

 

 

no

    
 

 
 
  

evaluation,
does patient
present symptoms
of some form

 

yes

Study Phases*

 

.l Discard '

 

I

no

  

  
        

the subject
meet RDC for
other subtype of
depression?

no

 

 

Patient qualifies for

study and is administered:

1) Differential Diagnostic
Depression Scale (DDDS)

2) Dexamethasone Suppression

Test (DST)
3) SADS/RDC

 

Does
the subject
meet RDC for
Endogenous Major
Depressive Disorder,
definite or
probable?

    
  

DST

Positive? n°

 

 

 

 

 

 

 

 

 

Subjects retained for SADS/DST
analysis and placed on a 5-week
desipramine trial for later

desipramine/DST analysis.

  

 

yes yes yes

 

 

 

Subjects retained of SADS/DST l

analysis only.

 

* Adapted from Holmes (1982)

57

formulated by Spitzer and Endicott (1978).

3. Each depressed subject was given the dexamethasone
suppression test (DST), a laboratory test for plasma
cortisol level. If the subject was on an inpatient
unit, 1 mg. of dexamethasone was administered at 11
p.m. by unit staff. A blood sample was drawn by
hospital staff the following day at 4 p.m. If the
subject was seen at the Psychiatry Clinics, they
were given the 1 mg for self-administration at 11
p.m. and instructed to report to the laboratory of
St. Lawrence Hospital for the blood sampling the
following day at 4 p.m.

All information obtained through the above three
procedures was considered confidential and held in a secure
location by the research team leader. Because procedures #2
and #3 have significant implications for clinical diagnosis
and treatment of depression, this information was released
by the study site if so authorized by the individual subject
(Holmes, 1982).

For those subjects whose RDC diagnosis was major
depressive disorder, endogenous subtype (probable or
definite), or whose DST response was positive (2.4.1 ug/dl) ,
a five- week trial response to desipramine was added as a
fourth study procedure. Approximately 50 , subjects
qualified for this phase of the study and were treated at
their initial site, either inpatient (n: outpatient.

Treatment consisted of an initial dosage of 50 mg at bedtime

58

with progressive increase to 150 mg withn 2 weeks according
to subject tolerance. The 150 mg dosage level was maintained
during the third week. Subjects who responded positively at
the end of the three week period continued at 150 mg for an
additional two weeks. Those patients who did not respond at
the end of three weeks had their dosage increased to 200-250
mg each night, depending upon response and side effects.
Care was taken that any medications which affect desipramine
plasma levels were discontinued during the five-week drug
treatment trial. At the end of week five, plasma was drawn
for desipramine levels (Bielski, Schafer and Holmes, 1982).
Both inpatients and outpatients were evaluated at weeks
one, three, and five following the institution. of
medication. Subjects were seen more often when clinically
indicated. Evaluation consisted of the Hamilton Depression
Rating Scale (see Appendix E) being completed by a trained
rater; the Zung Depression Rating Scale was also completed
by the patient him/herself at week zero when the medication
trial was begun. Positive response to medication was defined
as a Hamilton score of 7 or less, or 50% or less of the
week zero derived Hamilton score. Subjects who met this
criterion for drug treatment response within the first week
were to be excluded from the data analysis as placebo
responders. This rationale was adopted in view of the
consistent finding in the literature that the usual response
time for tricyclic drug treatment is approximately 3 weeks

(Bielski and Friedel, 1975; Bielski, Schafer and Holmes,

59

1982). No subject in the current study, however, was found

to be a placebo responder.

Design

The design for the study follows a correlational model.
First, the degree of association was measured between
individual clinical variables and a dichotomized biochemical
variable. Second, the dichotomized biochemical variable was
used to define group membership of study subjects. An
attempt was then made to identify the relative weightings of
clinical variables associated with membership in group one
as distinguished from group two. Lastly, for a subset of
study subjects the relationship between the dichotomized
biochemical variable and a change in a composite clinical
variable was examined after introduction of a pharmocologic

variable.

Research and Exploratory Hypotheses

A. Research Hypotheses:

1. DST suppressors and nonsuppressors will demonstrate
differences on individual clinical items drawn from
the Schedule for Affective Disorders and
Schizophrenia.

2. A linear combination of clinical items taken from the

Schedule for Affective Disorders and Schizophrenia

60

will separate DST suppressors from nonsuppressors.
B. Exploratory Hypotheses:
l. Probable and definite endogenous DST nonsuppressors
placed on a desipramine trial will demonstrate greater
improvement on the Hamilton Depression Rating Scale

than DST suppressors.

Analysis

A separate analysis plan was conducted for each
hypothesis. For the research hypothesis, analysis was
performed on the 46 items drawn from the Schedule for
Affective Disorders and Schizophrenia (SADS) in relation to
values on the dexamethasone suppression test (DST). Three
separate statistical analyses were conducted to examine the
relationship between SADS items and the DST. First,
Student's t-tests were performed on each SADS item with
regard to dichotomized DST values. Because the homogeneity
of variance assumption was found to be in violation between
DST response groups, Mann-Whitney U tests were also
conducted. An item by item analysis was thus conducted to
expand the range of symptoms examined beyond those included
in RDC and DSM III diagnostic categories. DST values were
dichotomized because: 1) 2. 4.1 ug/dl is the established
criterion for a positive response to the DST; and 2) the
distribution of DST values (see Figure 3-4) is such that

anything other than a dichotomy would result in insufficient

61

numbers of subjects in any smaller established interval,
e.g., 4.1-10 ug/dl, 10-15 ug/dl, 15-20 ug/dl, and 20 ug/dl.

The continuous DST values (2-26 ug/dl) were represented as

 

 

follows:
DST Values Analysis Value
< 4.1 ug/dl 0
24.1 ug/dl 1

Second, a discriminant analysis was conducted on the 46
SADS items using DST response as the designation of group
membership. This statistical. method. will be employed to
determine which linear combination of clinical item best
separated DST suppressors from nonsuppressors. Third, a
discriminant analysis was conducted on the 46 SADS items
using only one half of the study sample drawn randomly from
the total sample stratified by DST response
(suppression/nonsuppression) and RDC diagnosis (definite
endogenous, probable endogenous, and nonendogenous) . This
stratification procedure was used to insure that the
subsample created had the same distribution characteristics
as the total sample. The remaining half of the entire sample
was then withheld as a cross-validation group for the
discriminant function derived from the first half of the

sample.

Figure 3-4

Distribution of DST Values

I
”I 3'1 I'!’
l
' .....

 

W
{.321

r‘«.-

 

62

720.1

“C

t-‘O

com

t-‘O

0..

N00

HO

\ON

do

the

F‘O

<r1n

v-‘O

me-

do

NM

<2.0

'3' l" ":1 C-"'-J v.—-‘.

53553 90 HEEMHN

[HP/dl]

-

u—
I

u-l

(

L U E

n'.-.

H

DST '1

63

The analysis for the exploratory hypothesis was
conducted on Hamilton Depression Rating Scale (HDRS) total
scores derived from SADS ratings at the time of the
interview and on HDRS total scores at week five of the
desipramine trial.

It should be noted that only two of the cells in a 2 X
2 matrix addressing RDC diagnosis and DST response were

considered in this analysis.

 

Figure 3-5

RDC Diagnosis

END* NEND#
+ 24 3
DST
- 35 45

* endogenous

# nonendogenous

 

The analysis examined the distribution of Hamilton
change scores (response to treatment) for cells 1 and 3,

comprising a total of 34 subjects. Cell 2 was excluded

64

because it contained only 3 subjects, two of whom had
questionable diagnoses. One subject met the criteria for
endogenous symptomatology but not the criterion of mininum
duration (2 weeks); the other may have had endocrinological
complications which could have produced a false-positive
DST. Cell 4 was excluded from the analysis because
nonendogenous, negative DST subjects were not given a
desipramine trial in this study. Not including cells 2 and 4
of the above matrix compromised the quality of the analysis
because it was then impossible to evaluate the differential
contribution of diagnosis and DST response toward the
prediction of treatment response. The current design allowed
only for the analysis of DST response as a predictor of
treatment response exclusively within the subtype of
endogenous depression. Analysis of variance was performed to
estimate the value of the DST in predicting response to
desipramine as measured by the change in HDRS scores between
week zero and week five. Analysis of covariance was then
conducted to determine the predictive value of the DST in
conjunction with HDRS baseline (week zero) scores, i.e.,
HDRS scores at week zero were used as a covariate to be
regressed on HDRS mean change scores for DST suppressors and

nonsuppressors.

65

Summary

The Schedule for Affective Disorders and Schizophrenia
(SADS), a semi-structured interview format, was employed as
the primary clinical research instrument with this study.
Data from the SADS were used in conjunction with response to
the dexamethasone suppression test (DST) to explore the
clinical features of a possible subtype associated with DST
nonsuppression. A sample of 107 depressed men and women was
obtained through outside referral from an inpatient unit,
outpatient clinics, and private practitioners, as well as
through self-referral in response to a publicized
description of the study. The SADS was administered to each
subject along with the Differential Diagnostic Depression
Scale (DDDS) and the DST. Forty-eight subjects who met
either Research Diagnostic Criteria (RDC) for major
endogenous depressive disorder, definite or probable, or who
demonstrated a positive response to the DST were placed on a
five-week trial of the noradrenergic drug, desipramine.
Thirty-four patients completed the trial; fourteen dropped
out of the study either because of medication side-effects
or because of referral to another facility.

An analysis plan was presented for both research and
exploratory’ hypotheses. Two jparametric statistical
procedures were then used to examine the association between
clinical features and DST response: one, student's t-tests

with individual SADS items and DST response; two,

66

discriminant analysis to isolate a linear combination of
SADS items which might best separate DST suppressors from
nonsuppressors. A non-parametric statistic, the Mann-Whitney
U test, was used as a validity check for the t-tests. To
test the exploratory hypothesis of a correlation between DST
nonsuppression and response to a desipramine trial, analysis
of variance was performed. In addition, analysis of
covariance was conducted to explore the value of the DST
toward predicting desipramine response when adjusting for

HDRS baseline scores.

CHAPTER FOUR

RESULTS OF THE DATA ANALYSIS

In Chapter Four, the results of the statistical
analyses performed for both research and exploratory
hypotheses are presented. The first section is devoted to
the presentation of the results of the Student's t-tests,
Mann-Whitney U Tests, and discriminant analyses conducted on
the clinical features of the two groups identified as
positive or negative responders to the dexamethasone
suppression test. The second section presents results of
analysis of variance and analysis of covariance procedures
conducted on Hamilton scores to determine the value of the
DST in predicting response to the noradrenergic drug,

desipramine.

T-Tests:

The original analysis plan included the use of multiple
t-tests to investigate the relationship between individual
SADS items and DST response. The results appear in Table
4.1. DST“ nonsuppressors had significantly' higher ratings
(p=.05 or less) than suppressors on thirteen items:
subjective feeling about the severity of depression

(p=.002), psychic anxiety (.029), initial insomnia (p=.001),

£7

68

 

 

Hoooo. No.41 me.om Ho.H m~.m oo. Hm.H mmoq prone:
moo. Ho.m1 Ho.om om. oo.H 4H. o~.H mmoq muaumaam
oeo. Ho.H1 HH.om om. oe.m mm.H e~.m soemem mo xoeq
moo. ~m.a o~.~m o~.H oe.H Hm.H mH.~ one: mommam
So. of? RS 24 to $4 84. 39998. mEBmE
ooo. mo.m1 Hm.om om. mm. Ne. mm. eneeomeH Hmeaeume
So. oo.m1 omee mo. 8. om. me. 3985 6332
Hoo. so.m1 oo.oe as. so. me. om. HeeeeomeH HeHueeH
mom. mo; 3.3 84 $4 me; So .HQUUBeoEmmmeo
«me. on. eo.mm mo.H o~.H m~.H oo.H manage
omo. mm.m1 o~.mo HH.~ om.m so.H mo.m aueexea oasesme
NMH. mm.H1 ~o.ee H~.H m~.m o~.H oo.~ sumaxea oeemeom
mam. m~.H1 mo.om 4o. mo.H mo. oo.H meomeea enema
eom. oo. 1 om.om ~m.~ No.~ 4~.H mm.m mmaoemeemo Heoeoaom
ooe. oo. 1 oo.ee oo.H oe.m oH.H Hm.m bemewoeesoomeo
4mm. om. om.om Hm.o mm.m m~.H ~m.m acauesaeom msaueooz
moo. om. 1 oH.ee e~.H mm.m om.H m~.m aomoummm1oamm
oeo. mo. 1 mm.oo mo. Hm.m om.H Hm.m manages:
omo. mo. mm.oo mo. oo.~ mo. om.m meumoeoo oneeowow
mom. o~.H1 oe.me 4H.H ~o.~ oH.H oo.~ suaaeoo boeaumao
moo. Hm.m1 me.Hm oH.H oo.e . o~.H Ho.m weeum>mm
moo. om. 1 oH.me om. oo.H Ho. No.3 eoHHme wees
mom. H~.H H~.mo ee.a oH.~ me.H em.m oceueoeeemmeao

no 32 no ea
moemoaeaeoem msam>1e an N ozone H ozone eeuoeom
A1emoo Aeamoo

 

Asuaaoemeoee eeaaea1ozeo
£98m 338.. and: Sufi mode “so madame 989% z .4 Same.

69

 

 

moo. m~.m1 ~o.mo mo. mm.~ oo. oo.n .nneoEH neeOnnoeso
8m. em. 1 No.3 om. a: on. em; 8:39»an ones:
mom. o~.n1 mm.mm om. oo.n oo. oo. eoemnmeem no women
oom. no. oo.oo mo.n mo.n oo.n no.n momemoOnonomom
eoe. no. 1 mo.no no. m~.n mo. on.n non>eamm noncomnnea
own. om.n1 oo.om oo. om.n mm. nn.n emcee mono
moo. mm. 1 oe.om mo. m~.n oo. oo. meson noncone
ooo. em. 1 8.2. em; eon on; won 92883 So
ooo. oo. 1 oo.oe mo.n oo.~ on.n oo.n oenemmnoz_ee
moo. no.~1 oo.oo oo.n no.m em.n on.m onn>nnoeem
ooo. eo.n1 om.oo no. on.n oo. oo. oenenn
Non. mo.n1 oo.om on.n mm.~ oo. oo.n cenneenenem nonoe\omo
noo. om.~1 oo.mo n~.n mm.~ oo.n oo.n cannennoa
4mm. oo. 1 oo.mo o~.n ee.~ om.n nm.m onnnneennnnn nnoso
omn. mo.n1 om.nm oo.n om.m om.n no.m nmoea m>nnomoosm
ooo. no.n1 oo.nm om.n oo.n oo. em.n cennennneeomneomo
nom. oo. om.oo ne.n oo.m om.n oo.~ nzeneennz,nenoom
ooo. no.n1 on.oe oe.n oo.m mm.n oo.m nmmnenen mo moon
moo. om.~1 on.nm on.n oo.m om.n on.m connennemoeoe ego
moo. om.m1 me.oo on.n mm.m n~.n me.~ mmeem>nmnoeeen
ooo. oo. 1 om.me on.n oo.n om.n oo.n .oooooono unneeoo
nooo. eo.o oe.no om. oo.n mm.n oo.n enee neonmz
moo. on.~ mm.oo oo. on.n o~.n no.n enee ennnmoom
no on no 2
eoemononeone mone>1a no N ozone n oaone eenoeom
T Ho. 1+ .88

 

Ennnnnebono madden.

£96m ennnfi mime Amen eoam nee 338m 5819 :4 29a.

7O

terminal insomnia (p=.006), insomnia- severity (p=.001),
appetite loss (p=.002), weight loss (p=.00001),
indecisiveness (p=.025), dim concentration (p=.027),
agitation (p=.021), reactivity (p=.005), and functional
impairment (p=.002). Two items had a significant association
with DST suppression: appetite gain (p=.033) and weight gain
(p=.00001). It was, however, discovered that with all but
two of these items, the homogenity of variance assumption
did not hold. It was thus decided that a non-parametric
statistic, the Mann-Whitney U test, would be employed to

check the validity of the Student's t-tests.

Mann-Whitney U Tests:

 

The Mann-Whitney test is considered the non-parametric
analog of the two independent sample t-test (Pfaffenberger
and Patterson, 1977) and, as such, is a test that is
sensitive to both the central tendency and distibution of
scores. It is designed to determine whether two random
samples have been drawn from the same or different
populations. The results (see Table 4.2) of the Mann-Whitney
U tests conducted on the 46 SADS items and DST response
closely mirror the findings of the Student's t-tests. This
outcome suggests that there are significant differences
between DST suppressors and nonsuppressors cuiia univariate
level and that the results of the t-tests are not an
artifact resulting from. violation of the homogeneity of

variance assumption.

71

Table 4.2: Results of Mann-Whitney U Tests

(Corrected for Ties)

 

 

Symptom g 2 Significance
Severity 680.1 -2.96 .0331
Psychic Anxiety 748.0 -2.47 .0133
Initial Insomnia 698.5 -3.16 .0016
Middle Insomnia 846.0 -l.93 .0526
Terminal Insomnia 736.5 -3.05 .0023
Insomnia (severity) 683.5 -2.91 .0035
Appetite Loss 644.0 -3.33 .0008
Weight Loss 438.5 -5.00 .00001
Indecisiveness 771.0 -2.28 .0223
Dim Concentration 810.0 -1.99 .0460
Agitation 757.5 —2.45 .0142
Lack of Reactivity 718.0 -2.77 .0056
Functional Impairment 683.0 -2.99 .0028
Increased Appetite 683.0 2.00 .0448
Weight Gain 786.0 2.60 .0091

 

72

Discriminant Analyses:

 

It will be recalled that the homogeneity of variance
assumption was not upheld for individual SADS items and the
groups DST +, DST-. The assumption of equal covariance
matrices was consequently also in violation. Because
discriminant analysis is relatively robust with respect to
this violation and because no nonparametric equivalent
exists for discriminant analysis, discriminant analyses were
still performed on the 46 SADS items to determine which
linear combination of clinical items, with respective
weightings, would. best separarte DST :nonsuppressors from
suppressors. Two «discriminant. analyses following' the RAO
stepwise procedure were run on the Michigan State University
computer using the Statistical Package for the Social
Sciences (SPSS). Both used DST suppression or nonsuppression
as the specification of group membership. The first
discriminant function constructed involved the use of all 27
nonsuppressors in the study, with none held back for
cross-validation purposes. The standardized canonical
discriminant function coefficients and classification
results for this analysis are presented in Tables 4.3 and
4.4. The rationale for conducting a discriminant analysis
based on all subjects, without the immediate possiblity of
cross-validating the function constructed, was that the

nonsuppression group contained only 27 subjects and may,

73

Table 4.3: Standarized Discriminant Function Coefficients

for Discriminant Analysis Using Entire Sample

 

 

Variable Coefficient
Weight Loss .64154
Psychic Anxiety .59174
Indecisiveness .48633
Lack of Reactivity .42449
Depersonalization .40760
Functional Impairment .35372
Non-Delusional Ideas of Reference .33453
Classification .31166
Severity of Depression .28551
Terminal Insomnia .27944
Depression/Concerns -.18640
Drug Abuse -.26658
Concern with Bodily Functioning -.20960
Discouragement -.23920
Suicidal Tendency -.24956
Diurnal Mood Variation AM -.25659
Psychomotor Retardation -.31386
Weight Gain -.36279
Phobia -.41175
Negative Evaluation of Self -.51090
Obsessions/Compulsions —.53056

74

Table 4.4: Discriminant Function Classification Results

for Discriminant Analysis Using Entire Sample

 

 

Number Predicted
Actual of Group
Group Cases Membership
1 .2.
Group 1 (DST-) 80 74 6
Percentage 92.5 7.5
Group 2 (DST+) 27 4 23

Percentage 14.8 85.2

 

75

because of its small size, already be inadequate for
estimating the within-group variance of nonsuppressors. The
stepwise procedure eliminated 25 of the original 46 items as
having too small a unique contribution to be included in a
discriminant function to separarte DST suppressors from
nonsuppressors. Clinical items whose canonical discriminant
function coefficients were closest to the group centroid
associated with DST nonsuppression (2.07667) were weight
loss, psychic anxiety, indecisiveness, and lack of
reactivity. Those items closest to the group centroid of DST
suppressors (-.70088) were obsessions/compulsions, negative
evaluation of self, phobia, and weight gain. Seven of the
ten items with weighting toward the DST nonsuppression group
— weight loss, psychic anxiety, indecisiveness, lack. of
reactivity, functional impairment, severity of illness, and
terminal insomnia - are consistent with the results of the
univariate analyses (Student's t-tests and Mann-Whitney U
tests). The percentage of grouped cases correctly classified
for this sample was 90.65 (X2=85.498, p=.00001).

The second analysis used half of the subjects from the
total sample stratified by diagnosis. Standardized
canonical discriminant function coefficients and
classification results for the stratified sample analysis
are presented in Tables 4.5 and 4.6. A discriminant function
composed of 21 items resulted from the stepwise procedure,

which was then cross-validated on the withheld half

76
Table 4.5: Standardized Discriminant Function Coefficients

for Discriminant Analysis Using Split/Stratified

 

 

Sample
Variable Coefficient
Psychic Anxiety 1.51449
Diurnal Mood Variation AM 1.39740
Psychomotor Agitation 1.33162
Middle Insomnia 1.22449
Sleeps More 1.20609
Functional Impairment .90527
Antisocial Behavior .81160
Weight Loss .79548
Appetite Loss .76403
Diurnal Mood Variation PM .43896
Non-delusional Ideas of Reference - .31502
Panic Attacks - .47272
Phobia - .53064
Terminal Insomnia - .55904
Discouragement - .60586
Alcohol Abuse — .64859
Negative Evaluation of Self - .71696
Dim Concentration - .71962
Self-Reproach - .73842
Indecisivenes - .80436

Worrying -1.44933

 

m.hw m.mm m.nm m.m mmmuceonem

 

 

em on oo om n oe A1emo. m ozone
o.mm m.oo o o.oon moonzoonoo
o o on o on on .+emoo n ozone
M m M .n.
onemneQEez memoo ofleenonsez meeou
moono mo moono mo
oouonoenm nonesz envenoeno nebsoz
moonw cowumooam>1mmonu moono mflmwamcd amouflcH

 

eHmEmm omHmoumnum\uHHmm meow: efle>aoc¢ usmcflfionomoo

now measmom GoducoomammmHu cofluocom uCMCAEHnomHQ "one canoe

 

78

of the stratified sample. The group centroids for the
nonsuppression and suppression groups were 3.56213 and
-1.24675, respectively. Clinical items most closely
associated with the nonsuppression group were psychic
anxiety, psychomotor agitation, diurnal mood worsening, and
middle insomnia; worrying, indecisiveness, self-reproach,
and dim concentration, on the other hand, were related to
suppression. The percentage of cases correctly classified
for the initial analysis group ‘was 98.15 (X2=71.583, p:
.00001). For the cross-validation group, correct
classification was 67.93%. Classification accuracy with this
latter group ‘was still significantly better than chance
(p<.005).

It shall be noted that the item composition is quite
different for the discriminant function generated for the
entire sample compared with that for the
split/stratification sample. Only weight loss and psychic
anxiety maintained an association with the DST
nonsuppression group across discriminant functions. Diurnal
mood variation AM moved from identification with the
suppression group in the entire sample discriminant function
to identification with the nonsuppression group in the
split/stratification sample. Indecisiveness and
nondelusional ideas of reference were weighted toward the
nonsuppression group in the entire sample discriminant
function but emerged as associated with suppression in the

split/stratification. sample. The 1differences in item

79

composition between the two discriminant functions were
largely an artifact of the larger within-group variance of
the DST nonsuppression group (N=14) used in the
split/stratification discriminant analysis. That is, given
the greater likelihood that DST nonsuppressors in the
smaller split/stratification sample would differ as much
from each other as they would from members in the
suppression group, it is highly improbable that the same
clinical items would be weighted toward the nonsuppression
group as were found in the discriminant function of the
entire sample. Thus the most significant finding of the
discriminant analysis based on the split/stratification
sample remains the fact that the function derived, of
whatever item composition, could still separate a
cross-validation sample of DST suppressors and
nonsuppressors better than would be expected by chance.

It will be recalled that the DST is a state-dependent
biological measure and is usually only sensitive when an
individual is in the midst of a depessive episode. A
chi-square analysis was therefore conducted to examine the
possible relationship between time of greatest severity in
the present depressive episode and DST response (see Table
4.7). Time of greatest severity was measured through the
patient's subjective rating (SADS item "time period") of
whether his/her illness was most severe during the two weeks
up until the study evaluation or prior to that time. No

significant association was found (X2=.47247, p=.4919).

80

Table 4.7: Interaction of Time of Greatest Severity with

DST Response

 

DST Response

 

DST— DST+

Past

Week 16 14

Time

Prior

to 19 10
Past

Week

 

A. second. additional analysis ‘was conducted. on ‘the
subhypothesis that DST nonsuppressors experience more severe
symptomatology than suppressors, i.e., these former would
score higher on a severity of depression measure like the
Hamilton Depression Rating Scale (HDRS). A Student's t-test
was performed on HDRS baseline scores and DST response.
Results revealed that DST nonsuppressors in this study's
sample did have an overall more severe illness than

suppressors (see Table 4-8).

81

Table 4.8: T-Test Results for Severity of Illness

Based on HDRS Scores

 

 

DST- DST+

Group 1 Group 2 DE T-Value Significance
a .92 £4. a
16.42 7.06 22.26 7.52 105 -3.65 .001

 

The exploratory hypothesis of the study involving the
relationship between DST response and response to a
desipramine trial (as measured by a change in Hamilton
scores) was examined for the 34 subjects who completed the
trial using analysis of variance (ANOVA) and analysis of
covariance (ANCOVA) procedures. A one-way ANOVA was
performed on DST response (+,-) and each of three sets of
Hamilton (HDRS) scores:

1) Pretest HDRS scores.
2) Posttest HDRS scores.
3) HDRS change scores.
The results of these analyses are presented in Tables 4.9,

4.10, and 4.11.

82

Table 4.9: ANOVA Results for DST and Dependent Measures

 

 

MS F Significance
Pretest 36.813 1.261 .270
Postest 81.715 2.108 .156
Change 228.222 4.132 .050

 

Table 4.10: ANCOVA of DST with Posttest Using Pretest as

 

 

 

 

 

Covariate
MS F Significance
Pretest 22.82 .591 .418
DST 103.313 2.678 .112
Explained 63.067 1.634 .211
(variance)
Table 4.11: ANCOVA of DST with Change Scores Using
Pretest as Covariate
MS F Significance
Pretest 696.084 18.040 .001
DST 103.313 2.678 .112
Explained 799.398 10.359 .001

(variance)

 

83

Only the HDRS change scores had a significant
univariate relationship (p=.050) with DST response, i.e.,
the magnitude of change in HDRS scores between week zero and
week five was greater on the average for DST nonsuppressors
than suppressors. An analysis of covariance conducted on
pre- and posttest HDRS scores did not yield a significant
result. A second ANCOVA performed using pretest scores as
the covariate with change scores as the dependent variable
did result in significant findings. However, DST response no
longer constituted a significant factor in a linear equation
designed to predict a favorable response to desipramine.
Rather, the pretest scores' strong negative correlation with
HDRS change scores (P=-.5906) served as the most significant
component (p=.001) of the predictive equation. A correlation
matrix including DST response and the three dependent
measures pretest, posttest, and change score appears in

Table 4.12.

Table 4.12: Correlation Matrix for DST Response and

84

Dependent Measures

 

Posttest

Change

Pretest

DST

Posttest

1.0000

p=.00

.7224

p=.001

.1314

p=.229

-.2486

p=.078

Change

.7224

p=.001

-.5906

p=.001

-.3382

p=.025

Pretest

.1314
p=.229

-.5906

p=.001

.1947

p=.135

DST

-.2486
p=.078

-.3382

p=.025

.1947

p=.135

1.000

p=.00

 

85

Two hypotheses subordinate to the :major exploratory
hypothesis were also investigated. First, the question was
posed that if DST nonsuppressors demonstrate a greater
magnitude of response 1x: a desipramine trial, there might
also be a proportionately larger number of desipramine
responders among nonsuppressors than among suppressors.
Patients were categorized responders if their Hamilton.
change score was 7 or less, or 50% or less of their baseline
(week zero) Hamilton score. A chi-square test of
significance was then performed on the dichotomized
variables responder/nonresponder and DST
suppression/nonsuppression. Results (see Table 4.13) do not
show a significantly higher frequency of desipramine
response within the DST nonsuppression group as compared
with the suppression group (X2=1.10436, p=.2933).

Finally, it has been argued recently in the literature
(Smith, Glass, and Miller, 1982) that psychotherapy
conducted concurrently with drug treatment does not produce
an interactive effect superior to either intervention in
isolation. The hypothesis that DST nonsuppressors receiving
psychotherapy and desipramine at the same time would have a
higher frequency of response to desipramine (i.e., as
reflected in HDRS scores) was tested by chi-square analysis.
The results (see Table 4.14) do not indicate a positive
interactive effect (X2=1.5909, p=.2817) between
psychotherapy and desipramine for the 7 patients in this

study who received both forms of treatment .

86

Table 4.13: Interaction of Desipramine Responsiveness

with DST Response (N=34)

 

 

Responder
No Yes
- 9 11
DST
+ 3 11

 

x2=1.104, p=.293

Table 4.14: Interaction of Psychotherapy
with DST Response (N=34)

 

Responder
No Yes
No 8 9
Psycho-
therapy
Yes 4 ‘ 3

 

x2=1.59o, p=.281

87

Summary

The major hypothesis in this study involved the
relationship between response to the dexamethasone
suppression test and clinical features. This hypothesis was
examined from two perspectives. One, the association between
individual clinical items and DST response was explored
through the calculation of Student's t-tests. DST
nonsuppressors had significantly higher ratings (p<.05) on
twelve SADS items. These items were severity, psychic
anxiety, initial insomnia, terminal insomnia, insomnia
(severity), appetite loss, weight loss, indecisiveness, dim
concentration, agitation, lack of reactivity, and functional
impairment. DST suppressors had significantly higher ratings
for the items weight gain and appetite gain. While it was
ascertained that the homogeneity of variance assumption
underlying the valid use of the t-test was in violation,
execution of Mann-Whitney U tests, the nonparametric analog
to the t-test, yielded the same clinical items as
significantly related to DST nonsuppression and at close to
the same levels of significance.

Two, the relationship between DST response and a linear
combination of clinical variables was examined through two
separate discriminant analyses. The first discriminant
analysis involved the use of the entire sample, with no
subjects withheld for cross-validation. A discriminant

function was derived which correctly classified 90.85% of

88

2 =85.498,p=.00001). Discriminant function

the cases (X
canonical coefficients with the highest loadings toward DST
nonsuppression were weight loss, psychic anxiety,
indecisiveness and lack of reactivity; those toward DST
suppression were obsessions/compulsions, negative evaluation
of self, phobia, and weight gain. A second discriminant
analysis was conducted on one half of the entire sample
stratified by diagnosis, with the other half held back for
cross-validation of the function derived from. the first
half. The discriminant function calculated for the
split/stratified sample correctly classified 98.15% of the
cases employed (X2=71.583,p=.00001). Discriminant function
canonical coefficients with the highest loadings toward the
DST nonsuppression group in this split sample were psychic
anxiety, diurnal mood worsening AM, psychomotor agitation,
and middle insomnia; those toward the suppression group were
worrying, indecisiveness, self-reproach, and dim
concentration. Cross-validation of this function on the
withheld half of the stratified sample yielded 67.93 percent
correct classification. Although a 30.22 percent drop in
classification accuracy occurred between initial derivation
of this discriminant function and its cross-validation, the
cross-validation percentage of correct classification was
still significantly better than chance (p< .005) .

Two additional subhypotheses were explored. The first
involved the possible association between time of greatest

severity in a subject's current depressive episode and

89

response to the DST. A chi-square test performed did not
establish a significant association (X2=.47247,p=.4919).
The second examined the hypothesis that DST nonsuppressors
would score significantly higher than suppressors on a
measure of severity of illness. A student's t-test conducted
on derived HDRS scores at week zero did demonstrate that DST
nonsuppressors experienced more severe symptomatology than
suppressors (p=.019).

The exploratory hypothesis focused on the relationship
between DST response and response to a five-week desipramine
trial as measured by a change in HDRS scores. Univariate
analyses were first conducted on DST response and HDRS
pretest, posttest, and (post minus pre) change scores. Only
change scores emerged as having a significant association
(p=.050) with DST nonsuppression. Secondly, analysis of
covariance (ANCOVA) was performed using DST response as the
independent variable, HDRS pretest scores as the covariate
and HDRS posttest scores as the dependent variable; a second
ANCOVA was executed using HDRS change scores as the
dependent variable. DST response did not have a significant
loading in a linear equation to predict HDRS posttest scores
when using pretest scores as the covariate. Rather, a large
pretest score emerged as being significantly predictive of a

large change score.

CHAPTER FIVE

SUMMARY AND CONCLUSIONS

Summary

This investigation was an attempt to explore the
relationship between response to the dexamethasone
suppression test (DST) and clinical and pharmacologic
variables. The need for this study arose from the
observation that. no previous research had. examined this
relationship outside of the restricted symptom coverage of
already established diagnostic categories, i.e., those
defined by the Research Diagnostic Criteria or by the
Diagnostic and Statistical Manual of Mental Disorders III.
It was suggested that an abnormal response to the DST may
serve as a discriminating variable for defining a distinct
biological subtype of depression

Until the last decade, the literature on depression
relied almost exclusively on statistical analysis of
clinical features to derive depressive subtypes. The
endogenous/neurotic distinction among depressives was thus
created. With the introduction of biological measures of
depressive states like MHPG levels and the DST, researchers
attempted to establish the depressive subtype membership of
patients through the existence of specific forms of
biological dysfunction. While Carroll (1982) has commented

that the DST, as an indirect marker of HPA disorder, has an

on

91

unreliable association with specific clinical features,
neither he nor any other researcher has reported the
systematic investigation of the DST as a discriminant for a
clinical depressive subtype. The little research which has
focused on the subtyping capability of the DST has been
marred by methodological flaws or small sample sizes. The
literature reporting on the use of the DST to predict
anti-depressant medication response is also inconclusive.
Although several researchers (e.g., Brown, Haier, and
Qualls, 1980) have hypothesized a differential response of
DST nonsuppressors 1x3 noradrenergic tricyclics, these
studies have involved small samples and thus can offer only
tentative conclusions.

The current study aimed to overcome the shortcomings of
earlier research by: one, gathering initial clinical data
through the Schedule for Affective Disorders and
Schizophrenia (SADS), a standardized, semi-structured
interview format which can be used reliably by other
researchers; two, by further testing the hypothesis that DST
nonsuppressors respond more favorably to noradrenergic
medications (e.g., desipramine) than. do suppressors; and
three, by using a large enough sample of subjects to make
statistical analysis meaningful.

A sample of 107 depressed patients (77 women, 30 men)
was obtained from a consortium of mental health facilities
in the Lansing area as well as from self-referral in

response to a newspaper article about the study. For those

92

patients who passed the initial screening, the SADS was
administered, along with the Differential Diagnostic
Depression Scale (DDDS) and the DST. Forty-eight subjects
who met either Research Diagnostic Criteria (RDC) for major
endogenous depressive disorder, definite or probable, or who
were DST nonsuppressors, were begun on a trial of the
noradrenergic medication, desipramine. Thirty-four patients
completed the five-week trial.

Ratings of forty-six SADS items and response to the DST
were used to address the first study question: Is there a
relationship between individual clinical variables and
response to the DST? Both Student's t-tests and Mann-Whitney
U tests revealed that DST nonsuppressors had significantly
higher ratings on twelve items. These were: subjective
feeling of severity, psychic anxiety, initial insomnia,
middle insomnia, terminal insomnia, insomnia (severity),
appetite loss, weight loss, indecisiveness, dim
concentration, psychomotor agitation, lack of reactivity,
and functional impairment. DST suppressors had higher
ratings on two SADS items: increased appetite and weight
gain.

The same forty-six SADS items were employed with DST
response to examine the second question: Can a linear
combination of clinical variables be generated which would
separate the group of DST nonsuppressors from the group of
suppressors? Two1 discriminant analyses ‘were conducted to

respond tx> this question. First, an analysis was executed

93

using the entire study sample (n=107). A discriminant
function was derived which correctly classified 90.65% of
the subjects. Those clinical items most strongly associated
with the DST nonsuppression group were: weight loss, psychic
anxiety, indecisiveness, and lack of reactivity; those
associated with suppression were obsessions/compulsions,
negative evaluation of self, and phobia. Secondly, a
discriminant analysis was performed using one half of the
entire sample stratified by diagnosis. The discriminant
function derived correctly classified 98.15% of the subjects
included. Items with the highest loadings toward the DST
nonsuppression group in this sample were psychic anxiety,
diurnal mood worsening AM, psychomotor agitation, and middle
insomnia; those toward the suppression group were worrying,
indecisiveness, self-reproach, and dim concentration.
Although a 30.22% drOp in classification accuracy occurred
with the cross-validation group in comparison with the
initial derivation group, the percentage of correct
classification. with. the cross-validation. group1 was still
significantly better than chance.

Two questions ancillary' to the :major research
hypotheses were posed. One, in view of the DST's
state-dependent quality, do more DST nonsuppressors report
their depression being most severe at the time of their
study evaluation (and DST administration) than at some
earlier time during the current episode? A chi-square

analysis revealed that. they' do not. Two, do DST

94

nonsuppressors experience more severe symptomatology than
suppressors? A t-test conducted on DST response and total
derived Hamilton scores at week zero of the study did
demonstrate that DST nonsuppressors have a more severe
depressive illness.

Derived total Hamilton scores at week zero and direct
Hamilton ratings at week five of a desipramine trial served
as data for the major exploratory hypothesis: Does a patient
who fails to suppress normally’ on the DST .respond. more
favorably to a noradrenergic medication like desipramine
than does a DST suppressor? Results of univariate analyses
indicated a relationship between DST nonsuppression and a
significantly larger Hamilton change score (week five minus
week zero scores) than with suppression. However, an
analysis of covariance using Hamilton pretest scores as the
covariate did not uphold the significance of the DST as a
variable predicting desipramine response. Rather, a large
Hamilton pretest score emerged as the most significant
factor toward prediction of a large change score. Finally,
two secondary questions were addressed. One, is there a
higher frequency of response to desipramine among DST
nonsuppressors than suppressors? Hamilton change scores were
dichotomized as responder/nonresponder following an
established criterion of response. Chi-square analysis did
not demonstrate a significantly greater proportion of
responders falling within the nonsuppression group. Two, do

patients who are in psychotherapy concurrently with drug

95

treatment more often show clinical responsiveness than do

patients who are only on medication? Results of a chi-square

analysis did not suggest such an interactive effect.

Conclusions

 

Conclusions Drawn From a Review of the Literature:

1)

2)

3)

Previous studies have demonstrated that the
dexamethasone suppression test has an average
specificity of 96% and sensitivity of between 40 and 70%
when using a l-mg dose of dexamethasone.

The few studies which have attempted to clinically
separate DST suppressors and nonsuppressors either used
a limited range of symptoms on which to base such a
differential or were methodologically flawed.

There have been discrepant findings regarding support of
the hypothesis that DST nonsuppression is associated
with deficiency of the neurotransmitter, norepinephrine.
Noradrenergic medications like desipramine have not been
consistently shown to be more effective with DST

nonsuppressors than suppressors.

Conclusions From the Current Study:

1)

The current study had the following nonsuppression rates

across Research Diagnostic Criteria categories: definite

2)

3)

4)

5)

96

endogenous- 57.1% (16/28); probable endogenous- 25.8%
(8/31); and nonendogenous- 6.3% (3/48); the specificity
for endogenous depression was thus 93.7%. These rates
are comparable to those reported in the literature.

DST nonsuppressors had significantly higher ratings on
the individual clinical items subjective feeling of
severity, psychic anxiety, initial insomnia,

terminal insomnia, insomnia (severity),

appetite loss, weight loss, indecisiveness,

dim concentration, psychomotor agitation, lack of
reactivity, and functional impairment were associated
with DST nonsuppression. DST suppressors had higher
ratings on the items increased appetite and weight

gain.

A discriminant analysis performed on SADS items and
membership in the DST suppression or nonsuppression
groups generated a function which correctly classified
90.65% of the cases in the total sample.

A second discriminant function derived after analysis of
one half of the sample stratified by diagnosis correctly
classified 98.15% of those cases. The second half of the
stratified sample held back for cross-validation

dropped to 67.93 percent correct classification.
Cross-validation classification accuracy was, however,
still better than chance.

A patient's report of greatest severity of illness

occurring at the time of evaluation was not related to a

97

higher frequency of DST nonsuppression.

6) DST nonsuppressors did demonstrate more severe
depressive symtomatology than suppressors at the time of
evaluation.

7) DST nonsuppressors had a better response to desipramine
than suppressors as reflected in change scores on a
clinical measure.

8) When DST response and baseline scores on a clinical
measure were considered together in predicting
desipramine response (defined in terms of the same
measure), DST response was no longer a significant
predictive factor. Rather, a patient's having a large
baseline score on the clinical measure was most
predictive of a large change score (pre minus post).

9) There was no significant difference in the proportion of
DST suppressors and nonsuppressors who responded to
desipramine.

10) Psychotherapy did not make a significant interactive
contribution to a patient's response to desipramine (for
the small number of patients who were receiving both

forms of treatment in this study, N=7).

A Review of Recent Findings in the Literature

 

Hirschfeld, Koslow, and Kupfer (1983) published a
summary of a recent conference on the DST paneled by leading

researchers in neuroendocrinology, psychopathOlOng and

98

general clinical psychiatry (e.g., B. Carroll, W. Brown, D.
Klein, and R. Spitzer). It was the consensus among members
of the conference that while the DST should continue to be
used as a: research instrument, until nonsuppression rates
are determined for the normal population and for specific
diagnostic groups (e.g., schizophrenia and schizo-affective
disorder), the DST should not be used as a routine screening
device or in differential diagnosis. Content of the
conference relevant to the current study included the
observations that: the clinical differences between DST
suppressors and nonsuppressors should be further
investigated using new strategies, i.e., ones that do not
rely on the DSMIII or RDC categories; the relationship
between severity of illness and DST nonsuppression has
received inconsistent support; and there is preliminary
evidence to indicate that nonsuppressors respond more
favorably to noradrenergic tricyclic antidepressants.

The current study yielded specificity and sensitivity
rates for the DST and endogenous depression comparable to
those previously cited by Carroll (1982) and others. Recent
findings in the literature, however, can no longer lead one
to conclude that a concensus exists regarding the DST's
specificity to endogenous depression. Several researchers
(Coppen. et al, 1983; Castrc> et al, 1983) have provided
evidence that not only certain groups of non-depressive

patients (e.g., schiZOphrenics and alcoholics) show

99

dexamethasone nonsuppression but that a significant
proportion of nonendogenous depressives do as well.

While significant rates of DST nonsuppression among
non-depressive psychiatric disorders may be borne out with
further investigation, reports of the DST's lack of
specificity to endogenous depression within the more focused
area of the affective disorders are, upon close examination,
not as credible. Coppen and associates (1983) found a 49%
rate of nonsupression in patients diagnosed as nonendogenous
major depressive disorder, as well as a 40% rate of
nonsuppression among their sample of neurotic depressives.
These figures, taken out of context of how the diagnositc
categoreis nonendogenous and neurotic were defined by these
researchers, are largely misleading. Although the
International Classification of Diseases (ICD) was used to
assign patients to the category "major depressive disorder",
the subcategories endogenous and nonendogenous were
speficied by a patient's score on the Newcastle index. It is
erroneous to assume that patients found to be nonendogenous
in this fashion would be similarly rated under any of the
more frequently used diagnostic systems, i.e., many
Newcastle nonendogenous patients could conceivable fall into
at least the probable major depressive category in the RDC
(Kasper and Beckman, 1983). Additionallyy it. is unclear
whether Carroll's (1981) exclusion criteria (medical,
pharmacologic, and clinical) were followed by Coppen and his

colleagues in their assignment of patients to the ICD

category
given the
Kasp
nonsuppre
diagnosti
index. Si
upon hosp
dexametha
single b:
above 4 L
Results
unipolar
disorder,
nonsuppre
disorder
9%, respe
its memb
severity
features
Rating S1
were mor
gastroint
insight
direct a
of illne

than a d;

that the

100

eurotic depressive disorder who would later be

>ST.
and. Beckman (1983) investigated DST
sion within the context of three different
systems: the ICD, the RDC, and the Newcastle

ty-seven depressed men and women were evaluated

tal admission and were given the DST. One mg of

me was administered at 11 p.m. followed by a

od draw at 8 a.m. the next day. A blood level

/dl was used as the criterion for nonsuppression.

ndicated 55% of ICD endogenous depressives,

51% of RDC definite major depressive

subtype,

and 53% of Newcastle endogenous were DST

sors. While ICD neurotic and RDC minor depressive

emonstrated rates of nonsuppression of only 6% and

tively, the Newcastle neurotic category had 23% of
rs to show nonsuppression. Differences in the
of symptomatology and in individual clinical

'ere examined by means of the Hamilton Depression

regardless of diagnosis,

ile. DST nonsuppressors,

severely depressed, had more loss of interest,

stinal symptoms, somatic symptoms, and lack «of

1an suppressors. Kasper and Beckman comment that a
sociation between DST nonsuppression and severity
3 suggests a state variable in operation rather

stinct depressive subtype. They further hypothesize

:elationship between DST nonsuppression and somatic

101

complaints supports the interpretation of an "unspecific"
stress response pattern in nonsuppressors related to
pituitary-adrenal. cortical. systenl activationn Lastly, the
authors point to the rate of nonsuppression among neurotic
depressives diagnosed by the Newcastle index as a clear
indication of how specificity rates for the DST vary because
of discrepant criteria for diagnostic classification.
Although Kasper and Beckman appear to have been careful in
most respects with their research design, their having
instituted an 8 a.m. blood draw for the DST constitutes a
serious methodological abberation. Carroll et a1 (1981), in
an article on the standardization of the DST, reported that
only 24% of their total number of abnormal test results were
detectable based on blood samples taken at 8 a.m. Blood is
usually drawn at 8 a.m., 4 p.m., and 11 p.m. when a patient
is hospitalized, and at 4 p.m. when only a single blood draw
is possible. One may therefore hypothesize that while Kasper
and Beckman's reported rates of DST nonsuppression (51% for
RDC major depressive disorder) were not unusual, a certain
percentage of those patients who suppreSsed abnormally at 8
a.m. may not have at 4 p.m. and vice versa. Thus the focus
by these authors on eliminating threats to the DST's
specificity imposed by diagnostic inconsistency may have led
them to overlook a crucial factor in the accurate
identification of DST nonsuppression.

The evidence for the value of the DST in predicting a

patient's response to tricyclic anti-depressants is still

102

contradictory. Two recent studies highlight the opposing
claims. Ettigi and his colleagues (1983) conducted a study
to investigate the separate and combined significance of the
DST and an amphetamine challenge test toward prediction of a
positive response tx>aa desipramine trial. They report that
in their sample of eleven DST nonsuppressors who met
criteria for DSMIII major depressive illness, ten (91%) had
a positive response to desipramine, while only three of
seven suppressors (43%) responded favorably. Positive
responders were defined as patients who, after a 4-week
desipramine trial, were "able to function and make plans for
discharge". A. question, however, arises with this study
regarding how "able to function" was measured, i.e., no
specific target symptoms were mentioned for determining
response to desipramine. Extein, Kirstein, Pottark, and Gold
(1983) retrospectively examined the differential
effectiveness of the DST and a thyrotropinjreleasing hormone
test in predicting response to tricyclics and/or
electroconvulsive therapy. Nineteen patients who met
criteria for RDC major depressive disorder were DST
nonsuppressors. Eight of the ten DST nonsuppressors and six
of the nine suppressors responded to tricyclics, resulting
in no significant difference in frequency of response
between the two groups. However, this study has two basic
flaws: one, both noradrenergic (desipramine/imipramine) and
serotonergic (nortriptyline/amitriptyline) tricyclics *were

administered, thus confounding the possible differential

103

effects of each type; two, scores on the self—administered
Zung Depression Rating Scale were used as one of the primary
determinants for' a patient's favorable response to
treatment, when the Zung’ has long' been thought. to Ihave
questionable validity (Carroll, 1982).

Kline and Beeber (1983) conducted a study to
investigate the relationship between DST nonsuppression and
weight loss. The records of twenty-seven hospitalized
depressed patients were retrospectively reviewed. Thirteen
of fourteen DST nonsuppressors identified among this sample
were found to have weight loss of between 0.9 to 5.4 kg;
none had weight loss exceding 10% of normal body weight. The
authors hypothesize that since weight loss is one of the
criteria for the DSMIII diagnosis major depressive disorder
with melancholia, weight loss may be the key variable for
the specificity of the DST to melancholia with any
particular sample of depressed patients. Two other studies
appear to support Kline and Beeber's hypothesis. Berger et
al (1983) reported that with their sample of nine
nonsuppressors diagnosed endogenous or neurotic depressive,
seven had an average weight loss of 0.91 i 0.76 kg during
the week prior to the administration of the DST. The
implication of this finding is that melancholic patients who
do not have weight loss of marked or anorectic proportions,
and thus are not diagnosed melancholic through a positive
rating on the clinical feature weight loss, could still

demonstrate DST nonsuppression from even mild downward

104

fluctuations in normal body weight. Berger et al also
examined the relationship Ibetween DST :nonsuppression and
weight loss in normals. Nine (37.5%) of 24 subjects placed
on a diet of 1,000 to 1,3000 kcal per day for two weeks
demonstrated DST nonsuppression after suppressing normally
during two baseline DST's. Edelstein et al (1983) placed
eighteen healthy, depression free, obese subjects on a
protein-sparing diet who suppressed normally on a basesline
DST. After eight to twelve weeks of fasting, five (27.5%) of
the group of eighteen showed DST nonsuppression. These five
subjects weighed less initially and lost a significantly
greater percentage of their ideal body weight than the
suppressors (average loss of 13.5 kg, 17.6% of total body
weight). While Edelstein et al acknowledge that the design
of their study did not permit the separation of weight loss
from administered diet as the cause of abnormal HPA
activity, they caution that the DST should not be relied
upon for confirmation of the diagnosis of endogenous
depression with patients who have weight loss in the range
of 9-22.5 kg (the range of weight loss among their

subjects).

Discussion

 

The following discussion has been divided into seven
subsections which address a variety of issues raised by the

findings of the study.

105

The Timing of DST Sensitivity for a Depressed Patient

 

It is necessary, before interpreting the results of the
current study, to review the characteristics of the key
variable examined, response to the dexamethasone
supppression test (DST). With the exclusion of nondepressive
disorders in this study, the DST was a state biological
marker whose sensitivity to endogenous depression was
dependent on a patient being tested in the midst of a major
depressive episode. The state-dependency of the DST ‘was
examined by stating the hypothesis that those RDC probable
or definite endogenous subjects tested at the time of
greatest severity of their depression (i.e., severity as
reported at the time of the SADS interview) would have a
significantly higher frequency of nonsuppression than would
subjects who reported their depression being at its worst at
least two weeks prior to their evaluation with the study.
The data did not confirm this hypothesis. Nonsuppressors
were just as likely to report the greatest severity of their
current depressive episode occurring over two weeks prior to
the study evaluation as they were to report feeling the
worst at the time of the evaluation. This finding suggests
that the state-sensitvity of the DST varies across
endogenously depressed patients because of individually
variable lengths to the period of HPA dysfunction detectable
by the DST and current assay procedures. Such variability in

DST sensitivity' periods could confound attempts to

 

106

clinically separate DST suppressors and nonsuppressors if
one includes in the analysis group only subjects who are
diagnosed endogenous depressive (probable or definite). The
risk is then run that a certain percentage of the
suppressors among this group may have not suppressed
normally at another time during their current episode and
may not present a significantly different clinical profile
than those subjects who were negative responders to the DST
within a few days of the interview. There is no research
evidence yet available on the average time lapse between
normalization of HPA functioning and clinical improvement.
It is therefore conceivable that a depressed patient who had
HPA dysfunction detectable by the DST could revert to normal
cortisol suppression well before clinical change would be
noted.

It should now be clear that while creating a subset of
the entire sample in this study (i.e., only RDC probable and
definite endogenous) would have brought the ratio of DST
suppressors to nonsuppressors down from 3:1 to about 1.4:1
and would thus have created a more even balance in group
size for statistical analysis, it would, at the same time,
have been difficult to assess whether an endogenous
depressive who demonstrated DST suppression at the time of
the evaluation would not have at some previous or later
point in his/her episode. Having included, on the other
hand, both endogenous and nonendogenous patients in the

analysis pool provided the opportunity for the DST

107

suppression group to be more firmly weighted with
individuals who most probably would not fail to suppress
normally on the DST at any point during their depressive
episode. That is, given the rigorous exclusion criteria of
the present study, it was reasonable to assume that the
nonsuppression rate of nonendogenous patients in the sample
reflected the low rate of nonsuppression for RDC
nonendogenous reported elsewhere (Carroll, 1982; Schatzberg

et al, 1983).

DST Nonsuppression and the Traditional Concept of Endogenous

Depression

 

Specific clinical differences between the entire sample
of DST suppressors and nonsuppressors did emerge. Many of
the clinical features found to be significantly related to
DST nonsuppression have their analogs in the profile of
"endogenous depression" isolated by previous research using
largely statistical analysis of clinical features without
the inclusion of psychobiological correlates. In Table 5-1,
adapted from Nelson and Charney's (1981) analysis of a large
body of previous research on endogenous depression, a
summary is presented of the relative importance of specific
symptomatology in a pmofile of endogenous depression. Five
of the twelve symptoms which emerged as significant in the
univariate analyses of the current study - lack of

reactivity, severity of depressed mood, terminal insomnia,

108

ecoz enamoams .oumnoooz .mconum

 

 

o o *
mumEouum

2 no \mnzozozn neononzm

z 2 z 2 z mcoflmoaoo o>flmmonmoo

am am am 2 am MAGEOmCo HocoEneB

am am am nonsense dance:

2 nonsense HowuflcH

Hm Am mm mmoH unmfloz

am 2 Hm mcflcomnoz @002 gm

2 am am cooumnucmocoo nozoq

2 He 2 onnnmzo noznnmno

z z z z unenoucfl mo mmoq

2 2 am pm 2 nomonmenlmaom

s be be onnno>om

2 no no onn>nnooom

2 am um 2 am cofiucufimfi

um oz um um mum zonnoononom
omcooeom mocoooenm coauocom

ucofiumone EouQE>m ucmcfififlnomfla neumsau nouomm Ecumfimm

wooum mo mews

 

coflmmmnmmo msocomoocm Spas mEoumE>m mo aeouchOmmd "H.m canoe

109

psychomotor agitation, and dim concentration - are
consistent with the weightings of symptoms reviewed by
Nelson and Charney. However, the symptom weight loss, while
having the strongest association with DST nonsuppression in
the current study as indicated by both the univariate
analyses and by the discriminant analysis using the entire
sample, has been given only a "slight" association with
endogenous depression by previous researchers. In addition,
functional impairment, indecisiveness, appetite loss,
initial insomnia, and insomnia (severity), all significant
in the findings of the current study, apparently have not
been included as possibly associated symptoms in prior
studies. The symptoms psychic anxiety and initial insomnia
emerging as significant in the current findings are the most
discrepant with previous research of endogenous depression.
Both of these symptoms have traditionally been regarded as
indicative of neurotic depression (Kiloh and Garside, 1963;

Rosenthal and Gudeman, 1967).

Stress and DST Nonsuppression

 

The association of relatively high levels of psychic
anxiety with DST nonsuppression also appears to grant stress
some role in the psychobiological dysfunction behind
endogenous depression. However, the boundaries between such
terms as "stress", "anxiety", and "arousal" remains unclear

(Sweeney and Maas, 1979), leaving it possible to confer on

110

stress significance both as an insidious factor underlying a
symptomatological pattern and as symptom (anxiety)
consciously experienced. The symptom status of anxiety in
the context of endogenous depression is supported by
Akiskal's (1983) observation that it is not uncommon for
patients with affective illness to develop neurotic
personality traits. Thus DST nonsuppressors may become
anxious in response to the core affective illness rather

than depressed as a consequence of chronic anxiety (stress).

Weight Loss and DST Nonsuppression

 

Recent reports in the literature have presented data
which indicates an association between DST nonsuppression
and weight loss in normals as well as depressives. DST
nonsuppression in normals following the loss of less than
20% of normal body’ weight requires further research to
determine the causative factors involved. The significance
of weight loss for depressives with DST nonsuppression is
also not clear. Research thus far has investigated the
relationship between the individual variables weight loss
and DST response across diagnostic categories and not within
the context of other individual symptoms. The results of the
current study offer support for a strong association between
failure to suppress normally on the DST and weight loss but
also point to other clinical features having a concurrently

significant association with DST nonsuppression.

111

Severity of Illness and DST Nonsuppression

An association between DST nonsuppression and global
severity of illness has received inconsistent support, with
most studies reporting no significant relationship (Carroll,
1982; Brown and Shuey, 1980; Mendlewicz et al, 1982). Kasper
and Beckman (1983) interpreted their finding of DST
nonsuppression associated with severity of illness as
indicating that DST nonsuppression is a state variable tied
to an "unspecific" stress response pattern. Such an
interpretation. was based largely' on "somatic complaints"
having been the only cluster of symptoms significantly
related to DST nonsuppression in their sample. However, as
outlined above, the results of the current study are not
supportive of the view that DST nonsuppression is either due
to any single somatic symptom (e.g., weight loss) or
nonspecific in its implications.

While global severity’ of depression alone does not
account for the differential response to dexamethasone among
patients in the current study, it should be pointed out that
most symptom differences found between DST suppressors and
nonsuppressors from. SADS ratings were not based on the
presence or absense of a symptom but on the level of
severity of that symptom. The assessment of severity with
DST nonsuppressors thus moves from being important globally
to having significance in specifically defined areas of

symptomatology.

112
Toward a Revised Concept of Endogenous Depression

This study provides support for the DST as a functional
index (Carroll, 1981) which may serve to further specify the
clinical classification of endogenous depression. Symptoms
for which DST nonsuppressors reported higher severity levels
included many of the symptoms identified in prior research
as "endogenous". However, the current findings place in the
forefront certain "vegetative" features: weight loss,
appetite loss, lack of reactivity, and insomnia. It is
hypothesized that cortisol hypersecretion reflected in DST
nonsuppression may be associated with these specific
vegetative signs.

It is not clear from the results of the discriminant
analyses in this study what specific weighting would be
'given the clinical features in a redefined classification of
endogenous depression based on DST nonsuppression. While it
is promising that a discriminant function derived from the
small split/stratification sample still classified a
cross-validation sample (of equal size) significantly better
than chance, there is no expectation that a similar linear
combination of clinical features separating DST suppressors
from nonsuppressors would be generated with an independent
sample; the variation possible in composition and loading of
this linear combination has already been demonstrated in the
differences between the split/stratification and entire

sample discriminant functions computed in this study. The

113

point to be made here is that it is possible to
statistically separate DST suppressors from nonsuppressors
within the context of a multiple-variable profile.
Discriminant analysis will, however, have to be performed on
a much larger sample of DST nonsuppressors before a reliable
function will be isolated. A discriminant function index
could then be formulated following the technique outlined by
Feinberg and Carroll (1982) so that the clinician in the
field could readily identify a patient who would most likely

fail to suppress normally on the DST.

Prognostic Value of DST Nonsuppression

 

If one already concedes the future construction of a:
reliable discriminant function index for predicting DST
nonsuppression, what prognostic significance ‘would a
"positive" index finding have? Is identifying a potential
nonsuppressor anything more than satisfaction of nosological
curiosity? Certainly, the findings of the current study
indicate that DST response has less prognostic value than a
clinical index of severity of illness. While there were no
significant differences in baseline severity of illness for
those DST suppressors and nonsuppressors who began a
desipramine trial (34 of the total 107 patients in the
sample), a large change score on the clinical measure used
was significantly associated with a high (severe) baseline

score, regardless of DST response. Thus the magnitude of a

114

patient's clinical response to desipramine had more to do
with how severely ill he/she was initially than with DST
nonsuppression. This association. is .neither' difficult to
understand nor particularly meaningful: the more severely
ill patient has a larger interval for clinical improvement
over a five-week period than does a less severely ill
patient. A much more meaningful statistical finding would
have been that a higher proportion of DST nonsuppressors
than suppressors responded favorably to desipramine. In that
case the nosological significance of DST nonsuppression
would have translated into prognostic significance. To
summarize, based on the findings of the current study, the
prognostic value of DST response has yet to be determined.
The hypothesized association between DST nonsuppression and
a norepinephrine deficit differentially responsive to a
noradrenergic tricyclic currently lacks support and must be

further investigated.

DST Sensitivity and Diagnostic Classification Systems

Finally, relating the results of the current study to
the diagnostic categories major endogenous depressive
disorder of the RDC and. major depressive disorder with
melancholia of the DSMIII, one finds that RDC definite
endogenous has a DST sensitivity rate of 55% compared with
29% for DSMIII melancholia. This discrepancy in sensitivity

rates can be attributed to the more exclusive diagnostic

115

criteria with the DSMIII: two symptoms, loss of pleasure in
all or almost all activities and lack of reactivity, must be
present for DSMIII melancholia, as Opposed to the criterion
of at least one of the first group of four symptoms-
distinct quality of depressed mood, lack of reactivity, AM
mood worsening, and pervasive loss of interest- for RDC
endogenous (see Appendix F). Carroll's (1982) assertion that
the RDC category is overinclusive, while literally true in
comparison with the DSMIII, appears to be more of an
attribute than a shortcoming when one encounters the
heterogeneity of the depressive population, i.e, a more
exclusive diagnostic category is of little utility if it has
not been first demonstrated to match a real clinical

syndrome or disorder.

Suggestions for Future Research

 

1) One of the crucial issues surrounding the use of the
dexamethasone suppression test to isolate a unique
depressive subtype is the accurate identification of DST
nonsuppressors. As previously discussed, there may be
some degree of variability between endogenously depressed
patients both in the length of the period of HPA
dysfunction detectable by the DST and in the amount of
time which elapses between normalization of HPA
functioning and significant clinical improvement.

Speculation was offered that some DST suppressors who

116

presented a clinical profile similar to that of
nonsuppressors could have, in fact, responded abnormally
to the DST at some other juncture in their current
depressive episode. It is therefore suggested that a
group of patients be followed over the course of at least
one entire depressive episode, with the serial
administration of the DST, to determine the accuracy of-a
patient's designation as never having been a DST
nonsuppressor. Only then will sufficient safeguards be in
place to ensure that future discriminant analysis using
DST response is based on a valid and reliable assignment
of initial group membership. In addition, an increase in
the sample size of DST nonsuppressors would better insure

the reliable estimation of within-group variance.

2) While most antidepressants have not been found to alter
cortisol activity (Carroll, 1981), these medications do,
of course, change a patient's clinical state. This
potentially confounding factor in efforts to clinically
separate DST suppressors from nonsuppressors proved to be
approximately equal across response groups in the current
study: 22% (6/27) of nonsuppressors were on a tricyclic
antidepressant for one week or longer at the time of
evaluation, compared with 20% (16/80) of suppressors.
Ideally, subjects should remain at a drug-free, clinical
baseline prior to the administration of the DST.

Discontinuance of tricyclics one to two weeks before the

3)

4)

117

DST after a therapeutic regimen had already been
instituted could not be expected to return a patient to a

true clinical baseline.

It is not recommended that future research into clinical
differences between DST suppressors and nonsuppressors
involve univariate analysis. Univariate analysis was done
in the current study because the number of nonsuppressors
in the sample was marginal for running multivariate
analyses and because there was interest in comparing the
findings of individual clinical differences in this study
with individual symptoms identified by previous research
as being associated with endogenous depression. With a
larger sample, procedures like discriminant analysis are
the most meaningful for the objective of clinical
subtyping: to identify a group of symptoms which covary
simultaneously and have different weight with respect to

one another.

Finally, additional research into the use of the DST as a
predictor of response to drug treatment should follow a
design that calls for the inclusion of both nonendogenous
and endogenous patients. With this more complete design,
such questions as the following might be addressed: Do
endogenous DST nonsuppressors with a high initial
severity rating more frequently respond favorably to drug

treatment than nonendogenous DST suppressors with a

118

similar severity rating? Do endogenous DST nonsuppressors
with a low initial severity rating respond favorably less
frequently to drug treatment than nonendogenous DST

suppressors with high initial severity ratings?

APPENDICES

APPENDIX A

CONSENT FORM A

f . 119
APPENDIX A

Michigan State University
Consent Fonn (Form A)

Study: Self Reported Symptomatology
in Major Depressive Illness
Investigator: Gregory Alan Holmes, M.A.

Doctoral Candidate. Counseling Psychology
Robert J. Bielski, M.D.

Department of Psychiatry
I, , have had the above named study

explained to my satisfaction and I freely consent to participate in the study. I

have been informed that my decision to participate in this study will in no way

alter the treatment I will receive for my depression.
I understand that I have been asked to complete three-procedures during this
study. These three procedures are as follows:

1. I will be interviewed by members of the research team. The interview will
last for approximately 1% hours, during which time I will be asked questions
about my depression.

2. I will be given the Dexamethasone Suppression Test, a blood test for plasma
cortisol concentration. I will be given 1 mg of Dexamethasone in a tablet
form and be asked to take the tablet at 11:00 p.m. I understand that there is
minimal risk in taking this medication. I will then have a small sample of
my blood drawn the following day at 4:00 p.m.

3. I will be given the Differential Diagnostic Depression Scale (0003), a
163 item questionnaire.

I agree to participate in the procedures described above.: I understand

that the amount of risk and discomfort involved in this study is very small. being

no greater than that usually involved in drawing a small blood sample. I under-

stand that the benefits to me from participating in the study will be a special

evaluation of my depression.

" . 120

I further understand that I may ask questions before signing this consent
form, or anytime thereafter, that my participation in this study is voluntary
and that I am free to withdraw at any point without penalty.-

I further understand that the results from these procedures are confidential

and can only be released to others with my written permission.

Signed:

 

(Subject) (Date)

 

(witness) ' (Date)

Copies to: Subject
File

APPENDIX B

CONSENT FORM B

121
APPENDIX B «

CONSENT FORM

Dexamethasone Suppression Test and Desipramine Response
in Depressed Patients

Investigators: Robert J. Bielski, M.D.
Christine L. Shafer, M.D.
Department of Psychiatry
Michigan State University
Gregory Alan Holmes. M.A.

Department Qf_Counseling Psychology
Michigan State University

I agree to participate
in this study comparing the treatment response to the antidepressant
medication, desipramine, to the results of the dexamethasone sup-
pression test determined prior to entering treatment. I understand
that I will be evaluated in the first, third and fifth week after
entering treatment in sessions that last approximately one hour.
I will be seen by the study physician and also evaluated in interview.
I will receive desipramine, a commonly prescribed antidepressant.
I understand that the dosage of desipramine may be increased each
week, as is standard practice, and that the study physician will
follow my progress closely in order to insure safety. At the end of
five weeks of treatment, the amount of medication in my blood will
be determined. I will undergo venipuncture by trained, licensed
personnel and 10cc (about 2 teaspoons) of blood will be drawn.

If I begin treatment as an inpatient, I will be treated in
the hospital until discharge. After discharge I agree to par-
ticipate in the remaining sessions as an outpatient at the MSU
Clinical Center. If I am treated as an outpatient, each of the

three sessions will take place at the MSU Clinical Center. If I

122

CONSENT FORM - 2

desire, arrangement for continuing treatment of my depression

will be offered to me at the conclusion of my five week participa-
tion in this study.

I agree to participate in the procedures described above. I
understand that taking any additional medications during the study
might result in adverse effects. I agree to abstain from medica-
tions other than prescription medications approved by the study
physician for the duration of the study. I also agree to limit
my alcohol intake to either 1 ounce of liquor, 6 ounces of wine,
or 12 ounces of beer per day.

I further understand that the amount of risk and discomfort
involved in this study is very small, being no greater than that
usually involved in drawing of a small blood sample and in the
taking of this medication. I understand that symptoms such as
dry mouth, drowsiness, blurred vision, and delay in starting urine
stream, are common side effects of antidepressants and that I may
experience one or more of these symptoms.

I further understand that I should be careful while driving
or performing any act requiring dexterity or concentration because
of the possible side effect of drowsiness occasionally noted with
these medications. I understand that my response to alcoholic
beverages may be exaggerated while taking this medication.

I understand that safe use of this medication during pregnancy
and lactation has not been established. If female, I am not
currently breastfeeding. If female, of childbearing potential,

I do not suspect that I am pregnant, I have menstruated within
the last month, and, if I am sexually active, I will use contracep-

tion while participating in this study. If male, I will use

123
CONSENT FORM - 3
contraception while participating in this study.

To the best of my knowledge. I am not allergic to the
medication in this study or similar substances. I understand
that no beneficial effects are guaranteed.

I further understand that the alternatives to treatment
with this type of medication is treatment with another medication
(MAO inhibitor antidepressant) which has a slightly higher risk
of side effects. Another alternative is talking about my problems
without medication.

I further understand that I may ask questions before consenting,
or anytime thereafter, that my participation in this study is
voluntary and that I am free to withdraw at any point without
penalty.

I understand that the benefits to me from participating in this
study is a special evaluation of my depression. Additionally, my
response to desipramine and the desipramine plasma level which
will be drawn free of charge may be useful in treatment of subsequent
depression, should one occur. Also, if I am an outpatient, I will
receive the medication free of charge and a free screening physical
examination at the discretion of the study physician.

I understand that all the information gathered in this study
will be treated with strict confidence and I will remain anonymous
in any publications resulting from this data. Information con-
cerning the treatment of my depression may be specifically released
if I so authorize.

I understand that the MSU Clinical Center will gather some

information including my name, age, address, and telephone number

124
CONSENT FORM - L
for accounting purposes only and that this information will remain
confidential. I further understand that beyond this, my identity
will also remain confidential. I understand that I will not be
paid for my participation. I understand that if I am hospitalized
I or my insurance carrier will be billed according to hospital
policy. If I am covered by insurance, my insurance carrier will
be billed for any outpatient treatment-study visits. If I am not
covered by insurance a fee will be negotiated on a sliding scale
basis for outpatient visits. Upon request, I will be informed of

the results of this study.

If I experience an adverse reaction to the antidepressant
medication, I understand that I am to report this to my treating
physician if I am an inpatient. If I'am an outpatient I should
report these adverse effects to my treating physician at the M.S.U.
Affective Disorders Clinic, 353-3070, during regular working hours
Monday through Friday. Outside of regular working hours, if I am
an outpatient, I should contact the Ingham Community Mental Health
Center Emergency Service at 374-8000, and they will contact my treating
physician for me.

The study and procedures have been explained to my satisfaction.

 

 

Date Participant Signature

 

 

Date Investigator Signature

xc: participant file

APPENDIX C

SCHEDULE FOR AFFECTIVE DISORDERS AND SCHIZOPRENIA

n- ..
‘ i'
0
Q‘
a
Q
2
3
4

Third Edhlon 125

.. Jan Its - 333c l ()1 [-

SCHt UULI- F01! At FECTIVl: Uz‘LOHOLH's mu) sum/(WNW NM lS/~I)Sl
SCOIILSHIZLT

Card No. l
(l—Zl
Name v‘ll ‘3ut..¢tt

ugmq Interviewed ......

g

It Relative 0' Motel Sulnccl.
Name of index Sumcc1 .- _

 

 

Still|l3C1.S'Ui-.H

 

 

 

Study No. Razor's Name _ , llatcr's No, __
til-12H )— (ta-141°
Rater is: lntervucwer - 1 Age 0' Sex of Status at Time ol
Observer —2 Sublet“ Sublect: Male - 1 Evaluation: inpatient - l
(15) ' 116-171 Female ~ 2 _ Outpatient - 2
(16) Other patient - 3
Normal-cm - d
Other - 5
(19)
Date ol ‘ '1§’°°° Type of - Hospital - Form No: 83
Evaluationz.__ r--~'i ‘ Code: ,__._ Evaluation: Admission - 1 ID No. 09-80)?
(ZG-ZS) {26-27) ‘ Follow-up - 2 (29-36)
Other - 3
(23)

Instructions for use of SADS Scoreshcet: Some investigators may prefer to use a scoresneet rather than one protocol per patient.
In such cases the rater must refer to the SADS protocol for all definitions and instructions tor waging the items, checking or
writing in information. The item abbreviations used here are not adequate lor comnlctmq the term.

SADS - PART!

CURRENT CONDITION

 

 

 

 

 

Classification .......... O 1 2 3 E ........................... 213
L__~,slup to 223

Duration ................ 0 l 7 2 3 4 5 6 7 8 9 ...... 214

Onset ..................... 0 1 2 3 4‘ 5 6 7 ‘i 9 ...... 21$

Judgment/stress ..... NA 0 1 2 3 4 S 6 .............. 216

Assocnted with ......
Prc;nancy ......... YES ....... “.... ..................................... 217
Menopause ........ YES i it 213
Medication ........ YES a 219
Physically ill ..... YES i 220
Death ................ YES - 221
Stressful event... 0 1 '2 3: a 222

TREATMENT '

Outpatient....... ...... o 1 J» 3 .............. Sb.” ...................... 223
Psychiatric Hosp.... El 2 .............................................. 22.:
__-.-—) skip to 226
No. 0! Hospitaliz.... 1 2 3 4 5 6 7 8r ............ 22$

 

 

t Kcypunch: Duplicate on all cards.

'Il it is a reliability study, duplicate Rater No. in 77—78
On all cards.

0
o
a
O.

 

 

 

 

 

 

 

 

 

Treatment received
Lithium ............. -———u.., .................................................
Antidepressants. 4.,— J O .f ........ ' ‘7‘
Min Tranqwl ...... ...—......” ... ..... 7
Mai Tranqml ...... a—'—- .......
Other Rx ........... , ,n i " —'
Psychoth/cOuns.. -
Time period .......... 1 2 3 4 5
OYSPHORlC MOOD AND RELATED SYMPTOMS

 

 

 

Depression; ......... ..ID 1 2 3 C» 5 6 7 ................
Pastweek ............... a0 1 2 3 lg ,_ 5 ’5 7 ..................
Quality at mood o 1 ix" 3 a

Dept/concerns ........ O l (L. 3 d ......
Worrying ................ Lo 1 2‘ i z s 6...: ....................
Past week ............... O 1 2 7 4 S 6 ........................
Selbreproach .......... 0 l f. l 4 S 6 ........................

a

11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l'ast week .. .. . D | 3 3 -t -, i, 3,;
Ncqiiiw out i) i 2 J .i i. ., ' '
i’ast week 0 1 2 J .'. 'i i. 3.: i
llisi'iiiimgcmcnl. ._ 0 t 2 3 it '- I. ..... 2.33-
lust i'JCL'h ti 1 2 J 4 t. o ,. . 2r.»
5U|thlill tciiti. .7.... 0 l 2 3 it 5 i. I 2.10
l’ast week ............. .. 0 t 2 3 it :i 6 / 241
No. 01 gestures ....... [EA 1- 2 3 a s 6 7 6 9o 2:3
-—_.‘. slur) to 2‘31
sciioiiincii............. o l 2 ’3 a ‘i 6 ....................... 249
Medical lethality.... 0 l 2 l a 5 6 ....................... 250
Panic attacks .......... O 'i 2 _ 3 .......................................... 251
1 skip to 263
Shortness ol brth ......................................................... 252
Palpitations ............ —. ..................... 253.
Chest pain ............... _. 254
Smothering ............ —— 255 .
Dizziness. 256
Tingling -- 257
Faintness 253
Sweating 259
Trembling 250
Fear of dying 261
No. of consec.wks... 1 2 3 4 S 6 7 8 9e . . 262
Somat-c anxiety ..... O 1 2 3 4 5 6 ...................... 263
Past week ............... 0 1 2 3 4 S 6 ...................... 264
Psychic anxiety ...... O _ 1 2 3 4 5 6 ........ 7 ...........-.... 2675
Past week .............. 0 1 2 3 4 5 . 6 ....................... 266
Phobu ................. ‘o 1 2i 3 4 s 6 ...................... 267
e._l
Past week. ........... r o 1 2 3 4 s 6 ...................... 263
Type ot phobia.... .. 1 2 3 4 .................................... 269
Ohm/camoui ...... o t . 2 a 4 5 6 ...................... 270
Past week ............ . G 1 2 3 4 5 6 ...................... 271
Insomnia ................ 3 4 S 6 ...................... 272
‘ . —_—7xskip to 313
Initial insomnia ...... 273
Middle insomnia ..... — .................... 274
Terminal insomnia.. -— 275
Past week ............... O l 2 3 4 5 6 ....................... 313

 

519i.-

17

 

 

 

 

 

 

my“, .......- ,. l) I 2 3 .i s 6 ..................... 3

LJLh iil rim-iiy ll 1 3 l 4 'J 6 . . . 3.
Past week . . . . l.) l 2 3 4 ‘J (i ....................... J

Pour auiietite. . O l 3 .1 4 5 6 ..................... 3.
Past week ... .. i) .1 2 3 4 S 6 ....................... 3.
Weight loss ., O l 2 3 4 5 6 .................... 3.
Ille. appetite ...... U l 2 3 4 5 6 ........................ 3:
Weight gain............ 0 1 2 3 4 5 o ........................ 32
Cone/bouily lunc... 0 1 2 3 4 5 6 ........................ 33
Past week ............... 0 1 2 3 4 5 6 ........................ 32
IndeCisivehess ......... 0 1 2 3 4 5 6 ........................ 3:
Dim contentration. O t 2 3 4 S 6 ........................ 3:
Loss at .nicieiim. o t 2 3 s s 6 ....................... a:
Past week ............... 0 1 2 3 4 5 6 ........................ 32
Social withdrwl ...... 0 ° 1 2 3 47 S 6 ........................ 3:
Depersonalization" O 1 2 3 4 5' 6 ........................ 32
SUbieC£|VC anger ..... 0 1 72 3 4 S 6 ........................ 3:-
Past week ............... 0 1 2 3 4 5 6 ........................ 33
Overt irritblty......... 0 1 2 3 4 5 6 ....................... . 33
Past week ............... O 1 2 3 4 5 6 .......... 7 ............. 3 2
Agitation ................ IO 1 2 3 4 S 6 ........................ 3.:

;__i Skip to 341

Unable to sit still.... __ 33
Pacing .................... ._ 33
Hand-WTIOQIOQ ......... __ 33
Putting, etc ............. _ 3:
Outbursts ............... — ...... 33
Talks on and on ..... — 34
Past week ............... O 1 2 3 4 5 6 ..................... 34

Psychomotor ret....- 3 4 5 6 .................... 34.

 

 

 

 

 

Slowed speech 34
Inc pauses 3.:
Low speech ................. 34
Mute/deer. amt 3.:
Slawed movemt ..... ' 34
past week ............... o 7 1 2 3 4 s 6 ....................... a:

 

18 Characteristics ........ l 0] 1 l2 3 41.....- 34
* 0

skip to 353
Normal for age and martial status

1 Mildly decreased drive and satisfactio:
2 Definate loss of desire: functional
impotence

page

20

24

Heaclintv....

Diurnal mu » ..in _

Diurnal mil - pm

U

0

MA NlC svriCflOuE

Elevated "lead...”
Past week .........
Less sleep ..............
Past wee-i..........
More energetic ......
Past week ..............
lncr actIVIty ..........
Past week............ ..
Grandiosity ...........

9.151 week ..............

O

0

O

0

O

0

0

0

. 0

O

1

1

'J

'4

I‘d

2

2

3

3

4

4

‘1‘

5

5

...........................

.......................

CHARACTERISTICS OF BEHAVIOR AND ICEATION
DURING A PERIOD WHICH MIGHT BE "MANIC"

 

 

 

 

 

 

 

 

 

 

 

No other arid. ck. here C] skip to 415
21 Overt irritability......0 1 2 3 4 S 6 ........................
Motor hyperactiv... O 1 2 3 4 5 6 ........................
Acceleratsoeecm... 0 1 2 3 4 5 6 ........................
Accelerated thkg.... O 1 2 3
22 Intrusive ................ ——
Public dismay ........
Fin indiscretns ....... ...—
Assumes tasks ........ 7
Antisocial behav....
Sexual excesses ...... ——
Drunkennesi
Grossly bizarre .......
poorjudgrhent ....... O 1 2 3 4 S 6 ........................
. Duration ................ O 1 2 3 4
23 Alcohol abuse ........ O 1 2 3 4 5 6 ........................
Drug abuse ............. 0 1 2 3 4 S 6 ........................
Antisocial Denvru", o t 2 3 4 5 6 ........................
Distrustlulness ........ 0 l 2 3 4 5 6 7 ..................
r’ast week ............... O 1 2 3 4 5 6 7 ....................
Non-del ideas at ref 0 1 2 3

 

 

357

358

359

360

361

362

413

414

 

0.356

t...‘
U"

27

28

29

3O

DELUSIONS

Nu mud. Ck hctc

Del 01 releterice .....

Ccinq controlled...

People rcac mind ..

‘i'ht broati‘CJsting....

Tht :nsertion .........

Ttit Withdra~~l........
PerseCutory ............
JeaIOusy .................
Guilt or S'flm..........
Grandiose ...............

Somatic ..................

0

O

O

0

$76 in

l

1

11/7‘) ' page Jul Ill

(044‘. ..........................
2 3

2 3

2 3 ............................
2. 3.

2 3 ................................

2 3 ......................................
2 3 ..........................................
2 3 ...........................................
2 3 ..........................................
2 a ..........................................
2 3.. .. .-.

b'.

CHARACTERISTICS or! DELUSICNS ascmmass or Tva

Severity .................. 2 a 4 s 6 ...................... 43.-

Past wegk ............... O
Sensorium.............. D
Other delusions...... O
Consis w/rnood ...... 0
Bizarre duality ........ 0
Multiple Oelu .......... O
Fragmentary ........... 0
HALLUCINATIONS

No evid. ck. here C7:

Auditory ................
Running commtry..
2 or more voices .....
Non-alt verbal ........
Visual .....................

Olfactory ...............

Tactile ...................

0

0

0

0

7 O

0

O

1

1

1

Skip to 453

l

1

1

1

Grandio/pers. type. 0 1

CHARACTERISTICS OF HALLUCINATIONS - ANY TYPE

2

2

2

2

2

2

3

3 .

3

3

456 ........................

 

 

 

 

 

 

 

 

 

..........................................

 

Severity .................. -2 3 4 s 6 ........................

L———.\l skip to 455

Past week ...............

0

1

2

3

456 ........................

 

'Note: Some items are punched 1 when blank. o'I.

‘Note: Some items are punched I when plants. these are noted by b - I, when editing, circle I for these items.

b=I

43.1

441‘

4.12
6-1

443
on

444
12'!

445
{3'1

4:5
on
4.17
ml
4.18
ml
a.:9
[3'1

450
D'l
as;
0'1

d
31 Consis wimuuii .. (I I 2 1 , 4'3:
Ifiaqiiiciiiaiy . .. I) l 2 J 4'31
'ieiiiiiiiiiiii , I) I 2 1 , , .. 4‘14
«I -i-.i-.: I m I) l 27 1 .. 455
At least I mo. ..... 0 I 2 3 "55
Ni)n-allect..... 0 1 2 3 ......................................... 457
32 Ilitarrc pchvr..... ..... 0 I 2 3 4 5 Ci ........................ 458
Catatonic stupor ..... , ..................................................... 459
Catalan-c rigidity... _ ..................................................... 460
Wa-y flexibility..." __ .. 461
Catatonic ciictmt ......................................................... 462
Catalan-c postur.... __ ..................................................... 463
Memory disturb..." 0 I 2 3 4 5 6 ........................ 464
33 Funct.imorml ......... 0 1 2 3 4 S 6 ........................ 465
Past week ............... O 1' 2 3 4 S 6 ........................ 466
ADDITIONAL BEHAVIORAL ITEMS
Flight of ideas ........ 0 I 2 3 4 S 6 ........................ 467
34 Inappro allect ........ O 1 2 3 4 S 6 ........................ 46a
Stunted altect ........ 0 I 2 3 4 S 6 ....................... 469
Distractibility ......... 0 1 2 3 4 S 6 ........................ 470
Self-pity ................. 0 I 2 3 4 5 6 ....................... . 471
Dernandmgness ...... 0 1 2 73 4 5 6 ........................ 472
35 Depressed appear... 0 1 2 3 4 S 6 ........................ 473
FORMAL THOUGHT DISORDER . '
'Understandability.. o "I 2 3 4 S 6 ........................ 474
36 Loosening . asso ..... gfi__1__2 3 4 5 6~....r.-.....-.-.-....-:.-.-::.“475
.——-——-—---—- - - .
°Udrst.past week ..... 0 I 2 3 4 S 6 ........................ 476
”logical Ihkg .......... o I 2 3 4 s 6 ........................ 513
Paverty ot contnt... O 1 2 3 4 5 6 ....................... 514
Neologisms. ........... 0 I 2 3 515

37

128

 

SPECIAL ITEMS RELATED TO DIAGNOSIS OR SUB-
CLASSIFICATIONS OF SCHIZO-AFFECTIVE DISORDERS

 

 

 

Schizo t» attect- ..... m 2 516
Skip to 525

Dew/hallo .............. 0 I 2 ............... 517

Formal tht dis ........ 0 1 2 - 513

Delta/MIN .............. 0 I 2 ................................................ 519

”HOG. chU/hailu.... 0 I 2 ................................................ 520

 

’Formerly this Item was called Incoherence.
Data can be merged Irorh the ditterent versions.

 

page

3‘)

40

‘5
...

42

43

11/75 "7 page -I 0110

17t)lf1\.)111\1(||\,l."\ 0 1 2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mainly Scliiru .. I) I 2 . 3.
mainly Minx-vi.- 0 .1 3 . MWL ‘i.
Our.Schi.'u feat. . D I 3 3 4 ‘a .. .. 'i.‘
GLOBAL ASSESSMENT SCALE
Current worst peroop .......................................... 525-5“
Wk prior to .Idin __ ......................................... 52 7- 22‘
Past week ........................................................ 529-53
Rel/completeness... I 2 3 4 5 ................................... 53~
Mi-cd IcaI -S/A. O I 2 3 4 5 ............................... 53‘
Our. S/A Ep-s ................................................... 533.53~
Mixed features, ..... O I 2 3 4 5 .............................. 53‘
Our. Allcct. epic” __ ..... $36-53
Family Hx of prep:
Family Hx of ETDH
SADS :- PARTII 7 7
BACKGROUND
Highest grade......... 0 1 2 3 4 .5 6 7 .................. 538
A601. triendshlp ..... O 1 2 3 4 .S 6 ....................... 539
Marital iiaius.-......... -o I 2 3 4 5 Sat
Work last 5 yrs ....... O 1 2 3 4 S 6 7 8 9 ...... 541
Outpatient trmt ..... O 1 2 3 4 542
Age 1st OP care...... 543-544
No. psych. hosp ..... 545-546
Age Is! now .......... ......- ;.54 7:548
.Tdtal‘time big-5:... O 1 2 7 3 4 S 6 7 .................. 549
EPISODES OF MANIC SYNDROME
Criterion I .............. lo I 2 3 550
skip to 644

NO
Criterion II INFO NO 7 YES
Active .................... X 1 2 551
Talkativc ................ x .‘ I 2 $52
Thqunts race........ X 1 . 2 553
Grand-osity ............ X 1 2 .............................. 554
Less sleep ............... X 1 2 555
Distractibility ........ X 1 2 .............................. 556

APPENDIX D

RESEARCH DIAGNOSTIC CRITERIA

Third Edition 129 2/1/78

RESEARCH DIAGP'JCSTIC CRITERIA (RDC) SUi‘Ji‘vtARY DATA SHEET

 

Card No. col. 1 & 2 Date;

Name of Subject
Being Intewicwcd

 

Subject's ID Number

 

 

 

 

 

 

 

 

 

 

 

 

Sex:
If “Native OI Index Subject. ,
Name of Index Subject Age:
(3 - 10M
Shady No. Ralcr's Name Rater's No. Hospital ID ._____ Service: _
(It-12p II3-I4)+
Occasron: Intake-1; Update (Discharge or 2 mos)-2,' Update (SD'ICi'y reason) -3. 01", -4
(15)
Type of Evaluation. SADS- I; Si‘uCS-L-Z: Unstr. CI. Int-3; Cate Rec. Only -4; Other (Specify) -5
I16)
This scoresheet is designed to summarize judgments made in using the RDC. Many investigators prefer to use
one R 3C per subject so that they can nore which aspects of the specific criteria were met for inclusion and ex-
clusion for given diagnoses. The summary data is then transferred to the Summary Data Sheet for data proces~
sing. Other investigators prefer to re-use the RDC pr0tocol and are interested in recording only the summary
judgments n0ted on this scoresheet. If the Summary Data Sheet is the sole record for the subject, the rater
shputd refer to the RDC definition of the ratings being made.
Instructions: Circle the appropriate number for 91 diagnoses. Refer to the RDC definitions and instructions
prior to completing the Summary Data Sheet.
Diag. Duration Age at Diag.
Present Pres. Epis. First Previ0us Ever Met
Episode in Weeks Episode Episode Criteria
1, ~hizophrenia ................................... I 2 3":- __ 11:12- 423.4. I 2 325 I 2 326:
a. Course“ I - Acute 2 - Subac. 3 - Subchr. 4 — Chronic mm
b- Phenomenology I — Para. 2 - Disorg. 3 - Catat. 4 - Undiff. 5 — ReSid. .1.
_ + r
2. Schizo-affecuve, Manic ...................... I 2 3 429i H3021 ”33-4! I 2 3.135. I 2 3436.
a. CourseM I — Acute 2 - Subac. 3 - Subchr. 4 - Chronic II37.
b. Features I — Mainly Schiz. 2 — Mainly affect. 3 - Otheriizai
c. Onset I = ( 2days: 2' ( I week; 3= LI mo; 48 <2mos; 5= > 2mos.'139i
3. Schizo~affective, Depressed ................ I 2+ 3+H401 I‘M-3' ”“51 1 2 3:146. I 2 3mm
3. Course++ I — Acute 2 — Subac. 3 — Subchr. 4 - Chronicnaai
b. Features I — Mainly Schiz. 2 — Mainly affect 3 — Other nag.
c. Onset I= -2days: = L‘Iweek; 3= (Imp; 4= <2mps; 5=>2mos.i~.soi
4. Depressive Synd. Spimp. on Res.
Schizo. (Sec. Depression) ................. I 2 Bus” “524! ”556) I 2 3i157i I 2 3 '158I
Manic Disorder ................................ 1 2T 31:159I nee-25 iieaai I 2 35'65I I 2 3 «165'

+KeypunCh: Duplicate on all cards.

«Course. II the CDurse is best charaCIerized by chronic or subcnroniewitn an exacerbation. note chronic or subchronic here
and see item 439 40.
HI the Current illness invalves Cycling of the affective syndrome see item; 1+3 I 4434 .
III CVCIed during the present episOde see items 227‘230 -

/

 

130

 

 

 

 

 

 

 

Diag. Duration Age at Diag.
Preset! Pres. Epis. First Previous Ever met
Episode in weeks Episode Episode Criteria
6. Hypomanic Disorder ....................... I 21, 31:67. meant.“ __ int-2t I 2 3-l73i I 2 3 rim
Mipolar Depression with Mania (Bipolar I) .............................................................. I ..................... I 2 3 (I75!
8. Bipolar Depression with Hypomania (Bipolar 2) ............................................ -. ............................... I 2 3 me
‘- T S9‘I179-80)
9. Major Depressive Disorder ............... I 2 3 019- ‘22’3—2) «2234i 1 2 3 (225! I 2 3 lzzri
Pres. Affect. Epis. I - M/D: 2 7 WM; 3 - Changes, now M; 4 — Changes, now D; 5 - Mixed :277i
Tatal Duration of Affective Episode ......................................................................... (weeks) I2:~.-ir‘il
a. Primary ................................... I 2 3 '23“ ................................................ 1 2 3 mm I 2 3 mil
b. Secondary’ ............................. I 2 3 I234) I 2 3 (23‘3". I 2 3 mm
c. Recurrent (Unipolar) ............................................................................................................... I 2 3 (237)
d. Psychotic ................................. I 2 3 I238: .................................................. I 2 3 (239) I 2 3 :7an
e. Incapacitating .......................... I 2 3 I241; .................................................. I 2 3 (242) I 2 3 (243:
f. Endogenous ............................. I 2 3 (244'
g. Agitated .................................. ' I 2 3‘24“»
h. Retarded ................................. I 2 3 046»
i. Situational .............................. I 2 3 mm
j. Simple .................................... I 2 3 =24?
k. Predom. Mood I- Dep; 2 - Dep 8i Euph; 3 — Anx; 4 — Anx & Dep; 5 - HOStile;
6 - Apathetic; 7 - Other (249)
IO. Minor Depressive Disorder ............. I 2 3i250i (251-3) (254-5) I 2 3 (256) I 2 3 (257)
a. With Anxiety I 2 '3 (258)
II. Intermittent Depressive Dis ........... I 2 3 I259) ..‘__ (260-2)
12. Panic Disorder .............................. I 2 3 i263) r2545) __ 062.8) I 2 3 l269l I 2 3 (21m
59 = (279-81))
13. Generalized Anxiety Disorder ...... I 2 3 1319; (3202) __ (323-4! I 2 3 (325) I 2 3 (326)
a. With Depression I 2 3 (327)
I4. Cyclothymic Personality .................................................................................................................. I 2 3 I328:
15. Labile Personality ...................................................................................................... I 2 3 I339)
‘. Briquet's Disorder (Somatization Disorder) ............................................. g ........................................ I 2 3 330;
I7. Antisocial Personality ...................................................................................................................... I 2 3 mm

 

'lf Secondary see item 430. . .
1- If cycled during the present ePISOde see Items 227-230.

131

 

 

 

 

3

Diag. Duration Age at Diag.

Present Pres. Epis. First Previous Ever met

Episode in weeks Episode Episode Criteria
18. Alcoholism .......................................... I 2 3i352i ___i3335i (335.7r I 2 3 (338) I 2 3 (339)
19. Dfug USE Disorder .............................. I 2 3 Jag; ____I3413j (3.4.1.5) I 2 3 (3:6) I 2 3 I347)
20. ODSESSIVQ CODIDUISIVQ DISOrder ......... I 2 3(343) ___(34951, __(352,3, 1 2 3 (354, 1 2 3 (355..
21. Phobic Disorder ............................... 1 2 3i356i (3579) (3601) I 2 3 I362) 1 2 3 (363)

a. Subtypes I — Agoraphobia; 2 - Social phobia." 3 -- Simple phobia; 4 - Mixed iaeai

22, unspecified FunctionaIPsyct-iosistt I 2 3i365i ____jaseei £36970” 2 3i37ii I 2 3i372l

 

59. 7 .
23. Other Psychiatric Disorderrtii ......... I 2 3isi9i _____J4202I ___:423.4i I 2 3i‘425) I 2 (33:32:
24. Schizmypat Features ....................................................................................................................... I 2 3(4271
25. Currently Not Mentally Ill ............... I 2 3mm:
26. Never Mentally Ill. ................................... . ..................................................................................... I 2 3I429)

‘lf Secondary Major Depresswe Disorder, make sure the prior condition as listed on page 18 of the RDC is
also noted either above or here.

 

 

I - Preferential homosexuality; 2 - Anorexia Nervosa; 3 — Transsexualism; 4 -— 08$ “39’
(SDECIIy)
tFres. Schizo-aff. Ep. 1 - MD; 2 - D/M, 3 -- Changes, now M; 4 - Changes, now D; 5 - Mixed .43”
T0tal Duration of Schizoaff. Episode .............................................................................. (weeks) (432-34)

++ Diagnosis of SchizOphrenia or Schizo-affective Disorder is given and course is best described as an (Sup 435.38)
exacerbation in a subject With a chronic or subchronic illness.
Total Duration of exacerbation ........................................................................................ (weeks) __ (439-40)

H If Unspecified Functional Psychosis, describe heregusing RDC and DSM-III terms 59 (479 80
= - )

 

"

Ht If Other Psychiatric Disorder, describe here using RDC and DSM-Ill terms.

 

 

Alternative diagnosis for current episode if diagnosis noted above is questioned:

 

Note reason:

 

Narrative Summary Relevant to Psychiatric History:

APPENDIX E

HAMILTON DEPRESSION RATING SCALE

 

132
HAMILTON DEPRESSION SCALE - M.S.U. AFFECTIVE DISORDERS CLINIC

 

 

 

DATE
PATIENT MEDICATION - DOSAGE
'CI’CIP one "UMP" Per item. Score all items. 9- AGITATION (may coexist niitdly with “may...“
I. DEPRESSED MOOD (Sad. blue. gloomy. wecpv. pessimistic. .0. Mm"

 

helpless. hopeless. worthless)
0 Not depressed
I Feeling states elicited only on questioning
2 Occasional weeping. Spontaneously reports feellng states

3 Frequent weeping. Obvious behavioral evidence in faciea.
posture. voice. Speaks mostly about feeling states

4 Exhibits virtually onlv these feeling states verbally and non-
verbally. May have "gone beyond weeping"

GUILT FEELINGS AND DELUSIONS

0 Absent .

I Self-reproach. feels he/she has let people down
2 Expresses guilt regarding past errors or misdeeds
3

Present illness is deserved punishment. Ruminations over past
errors and sins

4 Severe self-reproach. Guilty delusmns. e.g.. is making other
people ill. Deserves to die. May have accusatory or denouncing
auditory or visual hallucinations

SUICIDE
0 Absent

I Feels life is empty. not worth living
Recurrent thoughts or wishes about death of self

Active suicidal thoughts. threats. gestures

bun

Serious suicide attempt
INITIAL INSOMNIA (as part of present illness)
0 Absent '

 

I Mild. infrequent: more than ‘6 hour occasionally
2 Obvious and severe: more than Va hour usually
MIDDLE INSOMNIA

0 Absent (Rate I if hypnotic is being used.)

 

I Complains of feeling restless and disturbed during night

2 Wakes during the night: any reading or smoking in bed or
getting out of bed except to void

DELAYED INSOMNIA
0 Absent
I Wakes earlier than usual but goes back to sleep

 

2 Wakes I - 3 hours before usual; unable to sleep again

WORK AND INTERESTS (Apathy: loss of interest in work.

hobbies. sacral life. Anhedonia: unable to feel pleasure)
0 No disturbance

Feels incapable. listless. less efficient. (Rate fatigue. loss of
energy under item 13)

2 Has to push self to work or play. No active interests. gets
little satisfaction. feels listless. indecisive

3 Clearly decreased efficiency. Spends less time at usual work.
In hospital. rate 3 if no spontaneous activity or marked loss
of personal tidiness

4 Stopped working because of present illness. Doesn't shave.
bathe. etc. Avaids ward activities: works only with urging

RETARDATION (Psychomotor slowing of thought. speech.
and movement. May vary diurnally)

Absent
Slightly flattened affect. fixed facial expression
Monotonous veice. delayed answering. sits motionless

Interview difficult and prolonged. Moves slowly

bun-O

Depressive stupor. Interview impussible
r

 

- l 'Fidgety. Clenching lists or chair arm. kicking fret

2: Wringing hands. pulling hair. picking at hands ur CIHIIIC\
Restless on the ward. some pacing

3 Can’t sit still: much pacing on ward

4 Interview conducted “on the run“. Constant pacing. l'ullin; nfl
clorhes. tearing at hair. picking at face

 

 

IO. PSYCHIC ANXIETY (as part of present illness. NOT part iil
previous disposition. Includes feeling tense. irritable. anprehrn~
sive. fearful. phobic or panic attacks)

0 Absent
1 Minimal distress. admitted only on direct questioning
2 Spontaneously expresses discomfort; worries over Irina
3 Obviously apprehensive in face and speech
4 Severely anxious. panicky. forgetful
II. SOMATIC ANXIETY (physiological concomitants of answis
0 Absent such as: fainting. tinnitus. blurred
vision. headache. tremor. sweating.
I Trivial flushing. hyperventilation. palpitaiiiins.
. indigestion. belching. diarrhea. urinars
2 Mud frequency)
3 Moderate
4 Severe
12. m

13.

16.

I7.

0 Normal appetite
I Eats spontaneously but without relish or pleasure

I Marked reduction of appetite and food intake. Eats onts “tin
urging. Requests or requires laxatives

SOMATIC ENERGY
0 Normal
I Occasional. mild fatigue. easy tiring. aching

2 Obviously low in energy. tired all the time; frequent backache).
headaches. heavy feelings in limbs

. LIBIDO (Rate only definite change with illness)

0 ‘ Normal for age and marital status
I Mildly decreased drive and satisfaction
2 Definite loss of desire: functional impotence

 

HYPOCHONDRIASIS

0 Absent '

I Mildly preoccupied with bodily functions and physical
symptoms
Moderately concerned with physical health
Morbid convictions of organic disease. e.g.. brain tumor. cancer
Bizarre delusions (often with guilty associations) e.g.. worms
eating head. rotting inside. bowels blocked. terrible odor

LOSS OF INSIGHT

 

0 Acknowledges being depressed and ill

I Acknowledges illness but attributes cause to bad food.“
climate. overwork. virus. need for rest

2 Denies being ill at all
WEIGHT LOSS (Rate either A or 8)
A. When rated by history
0 No weight loss
I Probable weight loss associated with present illness
2 Definite weight loss
8. When rated by weekly weight measure
0 Less than I lb. during past week
I Greater than 1 lb. during past week

2 Greater than 2 lb. during past week

16.01URNAL M000 VARIATION

 

O 1 2
0 1 2

Horse in a.m.
Horse in p.m.

APPENDIX F

COMPARISON OF RDC AND DSMIII CRITERIA

133

82 8.2 ..u .o .eBmfiemmz £0.23 .32

 

 

 

BER. .memeeoma 352 Mo Ease: .Nmeetmm .xfleflommm 3850 Co 8365
Eunmflfim es... uflmocmflo £03883 sumac unmoemflo soee$mm .m 5:323
ugmfloxmm :muflmad EoHu. :mxme « mono .h .HBQBCM ..h. .m .umNfiQm Sold 2935 ...

mac/flaw Hosxmm wmmomuomo

 

 

 

 

“Ham Baaoummmfi no m>fimwoxm E no 33.0.9,qu on DO: 085 mmflgflom
33 ӣ963 no mflxmuocm ucgmflm $3 How: 5 0.3983 mo pmmumflfi mo mmoq 6v
confine... eo mmofi Ease: 5
coflmcufimu MBOEocome 697.8: a. madman? Moon 3;
Amémxmzm coflmfimm no :QDmcumuou HBOEono>mm AC
mo menu How: 30an muse; 025 "29:85 @835 Ho mamxozm 9350.: 3.8m Am.
ummmfi um. mficoxmzm 05.95.: Egon A8 #30 Bgmoummmfi mo
093.88 93 5 o>fimooxm Ho noooummu flow wo mmcfiwmm A: .m
0.303 hangout.“ mfi cofimmumon one .8 .mwoum mm 5 $9 @802
@008 Ummmmummt mo >Hmbsv Dogma Am. ousmmflm no umoumpfi mo mmoH afiwmeflom .3
553ng m5 m0 mmufi ummmfl “E .U 9950.: $5 5 @953 fiumgmou ma @002 Am.
fidfiflm mmmcmzo
flnmusmoma 339m: B hie/Doomu mo xomq .m Econ—cogcm B baa/Howdy WHO xomq AN.
mmflfl/flomgm umopfim Mo do 5 whammmma mo mmoq .d @008 fimwwmumoc B \flnfiwsv 893mg A: :4
mflonocmflz 5H3 uwvuomE memoume hem: Hovuomwo m>Hmmmummo uoflmz msocomoocm
UH amaze: HQDWDBm can uflmocwflo mflmfiuu oflmocmflo noncomom
k. a”

 

amigo 52mm new mom “.5 comuemcoo

134

References

Akiskal, H. Dysthymic disorder: psychopathology of proposed
chronic depressive subtypes. American Journal of
Psychiatry, 140:11-20, 1983.

 

 

Albala, A.A., Greden, J. F., Tarika, J., and Carroll, B. J.
Changes in serial dexamethasone suppression tests among
unipolar depressives receiving electro-convulsive
treatment. Biological Psychiatry, 16, 551-60, 1981.

 

American Psychiatric Association. Diagnostic and Statistical
Manual for Mental Disorders, Second Edition,
Washington, D. C., APA, 1968.

American Psychiatric Association. Diagnostic and Statistical
Manual for Mental Disorders, Third Edition, Washington,
D. C., APA, 1980.

Amsterdam, J. D., Winokur, A., Caroff, S. N. and Conn, E. J.
The dexamethasone suppression test in outpatients with
primary affective disorder and healthy control
subjects. American Journal of Psychiatry, 139:3,
287-291, 1982.

 

Berger, M., Pirke, K., Doerr, P. Krieg, C., and Zerssen, D.
Influence of weight loss on the dexamethasone
suppression test. Archives of General Psychiatry
40:585-586, 1983.

 

Bielski, R. J. and Friedel, R. 0. Prediction of tricyclic
antidepressant response: a critical review. Archives of
General Psychiatry, 33:1479-1489, 1976

 

 

Bielski, R. J., Shafer, C. L., and Holmes, G. A.
Dexamethasone suppression test and desipramine response
in depressed patients. Unpublished manuscript, 1982.

Brown, W. A., Johnston, R. and Mayfield, D. The 24-hour
dexamethasone suppression test in a clinical setting,
relationship to diagnosis, symptoms, and response to
treatmnent. American Journal of Psychiatry,
136:543-547,1974.

 

Brown, W. A., Haier, R. J., and Qualls, C. B. Dexamethasone
suppression test identifies subtypes of depression
which respond to different antidepressants. The Lancet,
1979, i:928-929.

 

Brown, W. A. and Shuey, J. Response to dexamethasone and
subtype of depression. Archives of General Psychiatry,
37:747-751, 1980.

 

REFERENCES

135

Bunney, W. E. and Garland, B. L. Recent advances in the
biology of depression. In Stephanis, C.N. (Ed.) Recent
Advances in Depression, New York: Pergamon Press, 1983.

 

Carroll, B. J. Implications of biological research for the
diagnosis of depression. In New Advances in the
Diagnosis and Treatment of Depression (ed. J.
Mendlewiez) Amsterdam: Excerpta Medica, 1977, p.

 

85-107.

Carrroll, B. J. Neuroendocrine aspects of depression:
theoretical and practical significance. In The
Psychobiology of Depressive Disorders, Basel: Karger,
1980.

 

 

Carroll, B. J. Dexamethasone suppression test and diagnosis
of melancholia, reply to G. E. Murphy. Archives of
General Psychiatry, 38:1067-1068, 1981.

 

 

Carroll, B. J., Feinberg, M., Greden, J. P., Tarika, J.,
Albala, A. A., Haskett, R. J., James, N. M., Kronfol,
Z., Lohr, N., Steiner, M., de Vigne, J. P., and Young,
E. A specific laboratory test for the diagnosis of
melancholia: standardization, validation, and clinical
utility. Archives of General Psychiatry, 38:15-22,
1981.

Carroll, B. J. The dexamethasone suppression test for
melancholia. British Journal of Psychiatry,
140:292-304, 1982.

 

Castro, P., Lemaire, M., and Toscano-Aguilar, M. Abnormal
DST results in patients with chronic schizophrenia.
American Journal of Psychiatry, 140:1261, 1983.

 

Coleman, J. C. Abnormal Psychology and Modern Life,
Glenview, Illinois: Scott, Foresman and Company, 1976.

 

Coppen, A. Abou-Saleh, M., Milln, P., Metcalfe, M., Harwood,
J. and Bailey, J. Dexamethasone suppression test in
depression and other psychiatric illness. British
Journal of Psychiatry, 142:498-504, 1983.

 

Coryell, W., Goffney, G. and Burkhardt, P. DSM III
melancholia and the primary-secondary distinction: A
comparison by means of the dexamethasone suppression
test. American Journal of Psychiatry, 139:120-122,
1982.

 

Edelstein, C., Roy-Byrne, P., Fawzy, I. F., and Dornfeld, L.
Effects of weight loss on the dexamethasone suppression
test. American Journal of Psychiatry, 140:338-341,
1983.

 

136

Endicott J. and Spitzer, R. L. Use of Research Diagnostic
Criteria and the Schedule for Affective and
Schizophrenia to study affective disorders. American
Journal of Psychiatry, 136:52—56, 1977.

 

Endicott J. and Spitzer R. L. A diagnostic interview: the
Schedule for Affective Disorders and Schizophrenia.
Archives of General Psychiatry, 35:837-844, 1978.

 

Endicott, J., Cohen, J., Nee, J., Fleiss, J. and Sarantakos,
8. Hamilton Depression Rating Scale: extracted from
regular and change versions of the Schedule for
Affective Disorders and Schizophrenia. Archives of
General Psychiatry, 38:98-103, 1981.

 

Ettigi, P. G. and Brown, G. M. Psychoneuroendocrinology of
affective disorders: an overview. American Journal of
Psychiatry, 134:493-501, 1977.

 

 

Ettigi, P. G., Hayes, P. E., Narasimhachari, N., Hamer, R.
M., Goldberg, 8., and Secord, G. J. d-Amphetamine
response and dexamethasone suppression test as
predictors of treament outcome in unipoloar depression.
Biological Psychiatry, 18:499-502, 1983.

 

Exstein, I., Kirstein, L. S., Pottash, A. L., and Gold, M.
S. The dexamethasone suppression and
thyrotropin-releasing hormone tests and response to
treatment in unipolar depression. International Journal
of Psychiatry in Medicine, 12:267-275, 1982-83.

 

Eysenck, H. J. The classification of depressive illness.
British Journal of Psychiatry, 117:241-250, 1970.

Feinberg, M. and Carroll, B. J. Separation of Subtypes of
depression using discriminant analysis (I. Separation
of unipolar endogenous depression from non-endogenous
depression). British Journal of Psychiatry,
140:384-391, 1982.; _

Freud, S. Mourning and Melancholia (1917) in Collected
Papers. London: Hogarth Press Ltd., Vol. 4, 1956.

Gunaga, K. P., Zichus, P. A., and Bowman, H. E. Suppression
Test. St. Lawrence Hospital Newsletter, 1982.

Hamilton, M. and White, J. M. Clinical symdrome in
depression states. Journal of Mental Science,
105:985-997, 1959.

Hamilton, M. A rating scale for depression. Journal of
Neurological Neurosurgery and Psychiatry, 23:56-62,
1960.

137

Hirschfeld, R. A., Koslow, S. H. and Kupfer, D. J. The
clinical utility of the dexamethasone suppression test:
summary of a National Institute of Mental Health
workshop. Journal of the American Medical Association,
250:2172-2174, 1983.

 

Hollister, L. The amine hypothesis revisited: a clinician's
View. In Depression: Biologyy Psychodynamics, and
Treatment, New York: Pergamon Press, 1979.

 

 

Hollister, L. E., Davis, K. L., and Berger, P. A. Subtypes
of depression based on excretion of MHPG and response
to nortryptyline. Archives of Genreal Psychiatry,
37:1107-1110, 1980.

 

Holmes, G. A. Self-reported symptomatology in major
depressive illness: a pilot study. Unpublished
dissertation, 1982.

Kasper, S. and Beckman, H. Dexamethasone suppression test in
a pluridiagnostic approach: its relaionship to
psychopathological and clinical variables. Acta
Psychiatrica Scandinavica, 68:31—37, 1983.

 

Kendell, R. E. The classification of depression: a review of
contemporary confusion. British Journal of Psychiatry,
129:15-28, 1976.

 

Kendell, R. E. and Gourlay, J. The clinical distinction
between psychotic and neurotic depressions. British
Journal of Psychiatry, 117:257-266, 1970.

 

Kiloh, L. G. and Garside, R. F. The independence of neurotic
depression and endogenous depression. British Journal
9E Psychiatry, 109:415-463., 1963.

 

 

Klecka, W. R. Discriminant Analysis. Beverely Hills: Sage
Publications, 1980.

 

Klein, D. F. Endogenomorphic depression. Archives of General
Psychiatry, 31:447-454, 1974.

 

 

Kline, M. D. and Beeber, A. R. Weight loss and the
dexamethasone suppression test. Archives of General
Psychiatry, 40:1034-1035, 1983.

 

 

Lewis, A. J. States of depression: their clinical and
aetiological differentiation. British Medical Journal,
2:875-878, 1938.

 

Matussek, P., Soldner, M. and Nagel, D. Identification of
the endogenous depressive syndrome based on the
symptoms and the characteristics of the course. British
Journal of Psychiatry, 138:361-372, 1981.

 

138

Mendlewicz, J., Charles, G., and Franckson, J. M. The
dexamethasone suppression test in affective disorder:
relationship to clinical and genetic subgroups. British
Journal of Psychiatry, 141:464-470, 1982.

Nelson, J. C. and Charney, D. The symptoms of major
depressive illness. American Journal of Psychiatry,
138:1-13, 1981.

Nelson, W. H., Orr, W. W., Stevenson, J. M. and Shane, S. R.
Hypothalamic-pituitary-adrenal axis activity and
tricyclic response in major depression. Archives of
General Psychiatry, 39:1033-1036, 1982.

 

Peselow, E. D., Goldring, T., Ngaere, S., Fieve, R. R. and
Wright, R. The dexamethasone suppression test in
depressed outpatients and normal control subjects.
American Journal of Psychiatry, 140:245-247, 1983.

Pfaffenberger, R. C. and Patterson, J. H. Statistical
Methods for Business and Economics, Homewood: Richard
Do Imin' InCo' 1977.

 

Prusoff, A. A., Weissman, M. M., Klerman, G. L. and
Rounsaville, B. J. Research Diagnostic Criteria
subtypes of depression. Archives of General Psychiatry,
37:796-801, 1980.

Rosenthal, S. H. and Klerman, G. L. Content and consistency
in the endogenous depressive pattern. British Journal
of Psychiatry, 112:471-484, 1966.

 

Rosenthal, S. H. and Gudeman, J. E. The endogenous
depressive pattern: an empirical investigation.
Archives of General Psychiatry, 16:241-249, 1967.

Rothschild, A. J. and Schatzberg, A. F. Fluctuating
postdexamethasone cortisol levels in a patient with
melancholia. American Journal of Psychiatry,
139:129-130, 1982.

Schatzberg, A. P., Orsulak, P. J., Rosenbaum, A. H.
Toshihiko, M., Kruger, E. R., Cole, J. O. and
Schildkraut, J. J. Toward a biochemical classification
of depressive disorders, V: heterogeneity of unipolar
depressions. American Journal of Psychiatry,
139:471-475, 1982.

Schatzberg, A. F., Rothschild, A. J., Stahl, J. 8., Bond, T.
C., Rosenbaum, A. H., Lofgren, S. B., MacLaughlin, R.
A., Sullivan, M. A. and Cole, J. O. The dexamethasone
suppression test: identification of subtypes of
depression. American Journal of Psychiatry, 140:88-91,
1983.

139

Smith, M. L., Glass, G. V., and Miller, T. I. The Benefits
9: Psychotherapy. Baltimore: The Johns Hopkins Press,
1982.

 

 

Spitzer, R. L., Endicott, J. and Robbins, E. Research
Diagnostic Criteria. Archives of General Psychiatry,
35:773-782, 1978.

Stewart, W. J., Quitkin, F. M., Liebowitz, M. R., McGrath,
P. J., Harrison, W. M., and Klein, D. F. Efficacy of
desipramine with depressed outpatients. Archives of
General Psychiatry, 40:202-207. 1983.

 

Sweeney, D. R. and Maas, J. W. Stress and noradrenergic
function in depression. In (Depue, R. A., Ed.) 3r;
Psychobiology of Depressive Disorders: Implications for
the Effects of Stress. New York: Academic Press, Inc.,
1979.