See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/14008193

Reversed neurovegetative symptoms of depression: A community study of Ontario

Article  in  American Journal of Psychiatry · August 1997

DOI: 10.1176/ajp.154.7.934 · Source: PubMed

CITATIONS READS

83 1,542

5 authors, including:

Robert D Levitan Alain D Lesage

Centre for Addiction and Mental Health Université de Montréal
161 PUBLICATIONS   8,083 CITATIONS    397 PUBLICATIONS   9,767 CITATIONS   

SEE PROFILE SEE PROFILE

Sagar V Parikh Paula Goering

University of Michigan University of Toronto
238 PUBLICATIONS   11,169 CITATIONS    215 PUBLICATIONS   8,666 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

At Home/Chez Soi View project

Signature Bank View project

All content following this page was uploaded by Sidney H Kennedy on 22 May 2014.

The user has requested enhancement of the downloaded file.

 LEVITAN,
REVERSED
Am J Psychiatry
LESAGE,
DEPRESSIVE
154:7,
PARIKH,
SYMPTOMS
July 1997
ET AL.

Reversed Neurovegetative Symptoms of Depression:

A Community Study of Ontario

Robert D. Levitan, M.D., Alain Lesage, M.D., Sagar V. Parikh, M.D.,

Paula Goering, Ph.D., and Sidney H. Kennedy, M.D.

Objective: Most research on depression with reversed neurovegetative features (hypersom-

nia, hyperphagia, and weight gain) has been based on site-specific clinic-based samples. The
goal of this study was to delineate the epidemiology of reversed symptoms in a large community
sample and to use other symptom patterns for comparison. Method: Interviewers assessed
8,116 subjects across Ontario, aged 15–64 years, by using the World Health Organization
Composite International Diagnostic Interview. Individuals who met the DSM-III-R criteria
for major depression, current or lifetime, were classified into four groups on the basis of
lifetime neurovegetative symptoms: episodes of typical symptoms only, episodes of reversed
symptoms only, neither type, or both types (fluctuating-symptom group). The groups were
compared on demographic characteristics, comorbidity, disability, and health care utilization.
Results: Of the 653 individuals with lifetime major depression, 11.3% had episodes of reversed
symptoms only, and another 5.8% were classified as fluctuating. Most of the differences
among the four groups were due to the unique characteristics of the groups with neither type
of episode or a fluctuating pattern; individuals who had experienced only reversed symptoms
were remarkably similar to those who had had only typical symptoms. The fluctuating-symp-
tom group had high rates of comorbidity, substance abuse, and health care utilization. Con-
clusions: Several popular beliefs about depression with reversed features did not hold true for
this community sample. Identifying individuals who fluctuate between reversed and typical
episodes may be important in studies of major depression, in particular when reversed
neurovegetative symptoms are a consideration.
(Am J Psychiatry 1997; 154:934–940)

F or decades, clinicians and investigators have at-

tempted to link clusters of depressive symptoms to
pathogenesis and treatment response (1–3). Most work
in depression has focused on patients with the typical
symptoms of insomnia, decreased appetite, and weight
loss. More research has begun to address depression
characterized by reversed neurovegetative changes,
such as hypersomnia, hyperphagia, and weight gain.
These latter symptoms are characteristic of “atypical
depression” (3), seasonal affective disorder (4), bulimia
nervosa (5), premenstrual disorders (6), and “anergic
depression” (7, 8).
One of the problems in this area of research has been
a lack of agreement on how to define the population of
individuals with reversed neurovegetative symptoms,
Received Aug. 8, 1996; revision received Jan. 27, 1997; accepted
Feb. 27, 1997. From the Clarke Institute of Psychiatry and the De-
partment of Psychiatry, Faculty of Medicine, University of Toronto.
Address reprint requests to Dr. Levitan, Rm. 1135, Clarke Institute
of Psychiatry, 250 College St., Toronto, Ont. M5T 1R8, Canada;
levitanr@cs.clarke-inst.on.ca (e-mail).
with individual centers tending to focus on particular
groups of patients. Quitkin, Liebowitz, and colleagues
(3, 9) have focused on the group of patients with mood
reactivity and various combinations of reversed neuro-
vegetative features, leaden paralysis, and sensitivity to
rejection. Himmelhoch et al. (7) and Thase et al. (8)
have defined a group characterized by reversed neuro-
vegetative symptoms, fatigue, psychomotor slowing,
and lack of mood reactivity. The goal of these investi-
gators has been to identify individuals who might re-
spond to particular treatments. Another line of research
has looked at possible biological markers that may dis-
tinguish depressed patients with reversed neuroveg-
etative features (5, 10). To enhance the validity of these
lines of work, however, more preliminary research is
needed to delineate the overall epidemiology of de-
pression with reversed neurovegetative features in
the community. So far as we are aware, only one group
has reported on the epidemiology of reversed neuro-
vegetative symptom patterns in a nonclinical commu-
nity sample. Horwath et al. (11), using data from the
Epidemiologic Catchment Area study (12, 13), found

934 Am J Psychiatry 154:7, July 1997


that 16% of patients with lifetime major depression
had both hypersomnia and hyperphagia while de-
pressed. Compared to all other patients with lifetime
major depression, this group had a younger age at
onset of depression, more psychomotor slowing, and
more comorbid panic disorder, drug abuse or depend-
ence, and somatization disorder. The authors con-
cluded that this group may represent a distinct subtype
of major depression.
We have noted that many clinic patients have epi-
sodes of depression that have neither the typical nor the
reversed pattern. Other patients report having had both
typical and reversed episodes over time. This suggests
that the distinction made by Horwath et al. of only two
categories based on neurovegetative symptoms, i.e.,
typical or reversed, may not have reflected the full spec-
trum of syndrome presentations possible. We thus de-
cided to extend the work of Horwath et al. by studying
a second large community sample and by delineating
four, as opposed to two, depressive groups based on
neurovegetative symptoms, i.e., typical symptoms only,
reversed symptoms only, neither type, or both types
(fluctuating-symptom group). It was hypothesized that
this further analysis might have theoretical and clinical
relevance.
METHOD
This study is a secondary analysis of data from the Ontario Health
Supplement, a community survey of 9,953 Ontario residents con-
ducted in 1990–1991. The Ontario Health Supplement was designed
to assess the prevalence of major psychiatric disorders and associated
risk factors, formal and informal utilization of health care services,
and disability. A stratified, multistage sampling design was used and
is described elsewhere (14). In summary, the province was divided
into a number of enumeration areas, and households were sampled
from each area. One individual aged 15 years or older was randomly
selected from each household. Specifically excluded were individuals
living in institutions or on native reserves. A response rate of 76.5%
was achieved, resulting in a sample of 9,953 individuals. Subjects over
the age of 64 were assessed with a shortened version of the original
questionnaire and were not included in the current study. The overall
sample from which our groups were derived thus included 8,116 re-
spondents aged 15–64. Because this project was funded by the pro-
vincial government and implemented by Statistics Canada (equivalent
to the U.S. Census Bureau), standard provincial and federal proce-
dures were followed for informing participants about the study. Be-
fore interviewer contact, an official letter was sent to each prospective
respondent to describe the nature of the project, state that participa-
tion was voluntary, and describe the federal legislation that protected
the confidentiality of any volunteered information.
Instruments
The questionnaire used in the Ontario Health Supplement is a
composite of a diagnostic instrument, measures to assess disability,
and questions to measure service utilization; it begins with a probe of
basic demographic information. Individuals were interviewed for 1 to
2 hours in face-to-face encounters with specially trained lay inter-
viewers. The diagnostic instrument used, the World Health Organi-
zation’s Composite International Diagnostic Interview, is a struc-
tured interview based on the Diagnostic Interview Schedule and the
Present State Examination, and it has good reliability and validity
(15). The version used here had been modified and further field tested
by Kessler and Wittchen (16). The major diagnoses surveyed included
Am J Psychiatry 154:7, July 1997
LEVITAN, LESAGE, PARIKH, ET AL.
anxiety disorders, affective disorders, eating disorders, psychotic dis-
orders, substance abuse, and antisocial personality disorder, which
were diagnosed according to DSM-III-R criteria. The Composite In-
ternational Diagnostic Interview systematically records the symptoms
of each potential disorder, and both current and lifetime diagnoses
are delineated.
Disability was determined by a variety of questions on functioning
in a number of domains. The subjects were asked to estimate the
number of days they were 1) totally unable to function, 2) partially
unable to function, and 3) functioning only with extreme effort. The
subjects could endorse more than one type of dysfunction. Function-
ing was broadly defined to include performance at a job, performance
in housework, and performance of leisure activities.
Utilization of health care services was defined broadly; the queries
covered use of a wide variety of professional and nonprofessional
support for mental health reasons across many different treatment
settings. Outpatient use of general physicians, psychiatrists, and other
health care providers (psychologists, social workers, nurses, and oc-
cupational therapists) in the past year was determined.
Sample
Only subjects aged 15–64 years and meeting the criteria for current
or lifetime major depression were included in the current study. Each
subject was assigned to one of four groups as follows.
1. Episodes of typical symptoms. Subjects in this group positively
endorsed each of the three items “decreased appetite,” “weight loss,”
and “insomnia” (initial, middle, or late) on the Composite Interna-
tional Diagnostic Interview section for major depression (current or
lifetime).
2. Episodes of reversed symptoms. Subjects in this group positively
endorsed each of the three items “increased appetite,” “weight gain,”
and “hypersomnia” on the Composite International Diagnostic Inter-
view section for major depression (current or lifetime).
3. Neither type of episode. Subjects in this group had current or
lifetime major depressive episodes but did not meet the criteria for
either of the preceding symptom types.
4. Fluctuating episodes. Subjects in this group had experienced
both episodes of typical symptoms and episodes of reversed symp-
toms over time and met the criteria for both typical and reversed
symptoms.
Statistical Procedures
The four patient groups were compared along four domains of
study—demographic characteristics, comorbidity, disability, and
health care utilization. For variables in each of these domains, two
types of comparisons were made: 1) a single initial comparison of
differences across all four groups and 2) a comparison of just the
typical-symptom and reversed-symptom groups. The initial tests of
differences across all four groups were done by using chi-square
analysis for categorical variables and analysis of variance (ANOVA)
for continuous variables. Where the distribution of values for con-
tinuous variables was skewed, nonparametric (Kruskal-Wallis) tests
were used. For the comparisons of the typical-symptom and reversed-
symptom groups, the Mann-Whitney test was used for skewed distri-
butions of results for continuous variables. Statistical significance was
set at p=0.01.
RESULTS
Of the sample of 8,116 individuals aged 15–64, a to-
tal of 653 (8.0%) met the criteria for current or lifetime
major depression. Of these, 346 (53.0%) had had nei-
ther reversed nor typical episodes, 195 (29.9%) had
had only typical episodes, 74 (11.3%) had had only re-
versed episodes, and 38 (5.8%) had had both types of
episodes.
935
REVERSED DEPRESSIVE SYMPTOMS

TABLE 1. Demographic Characteristics of 653 Persons With Lifetime Major Depression in a Community Sample, by Presence of Reversed
Neurovegetative Symptoms
Type of Depressive Episodes Over Lifetimea
Group Comparisons
Neither Both
Typical nor Typical Reversed Typical and Typical Only
Reversed Only Only Reversed Versus Reversed
Characteristic (N=346) (N=195) (N=74) (N=38) All Groups Only
2
N % N % N % N % χ df p χ2 df p

Female sex 221 63.9 151 77.4 56 75.7 27 71.1 12.4 3 0.01 0.1 1 n.s.
Marital status 31.9 6 0.0001 6.2 2 n.s.
Currently married 214 61.8 84 43.1 39 52.7 16 42.1
Never married 85 24.6 50 25.6 23 31.1 11 28.9
Separated, divorced,
or widowed 47 13.6 61 31.3 12 16.2 11 28.9
Social status
Low income 53 15.6 44 22.7 12 16.4 9 23.7 5.0 3 n.s. 1.2 1 n.s.
Employed or student 299 87.2 168 87.0 59 79.7 30 78.9 4.5 3 n.s. 3.0 1 n.s.
Education 13.2 6 n.s. 4.7 2 n.s.
Some secondary 121 35.2 74 37.9 27 37.0 12 32.4
Secondary only 135 39.2 80 41.0 22 30.1 22 59.5
Postsecondary 88 25.6 41 21.0 24 32.9 3 8.1

Mean SD Mean SD Mean SD Mean SD F df p F df p

Age (years) 37.0 12.6 36.1 12.7 36.1 11.5 36.6 12.3 0.27 3, 649 n.s. 0.00 1, 267 n.s.
aDiscrepancies in percentages are due to removal of subjects with incomplete data from corresponding denominator.

In the following presentation, we report the most sa-
lient findings that 1) differentiated the four study
groups on each of the four domains and 2) differen-
tiated the typical-symptom and reversed-symptom
groups only.
Demographic Characteristics
As shown in table 1, there was no significant differ-
ence across the four groups with respect to age, social
status, or education. There was a significant difference
in sex distribution, with the neither-symptom group
standing out as having a lower proportion of women
than the other groups. There was also a significant dif-
ference in marital status; the most notable finding was
the higher rates of separated/divorced/widowed status
in the typical- and fluctuating-symptom groups than in
the neither- and reversed-symptom groups.
When only the typical- and reversed-symptom groups
were compared, no significant differences were found.
Comorbidity
Lifetime comorbidity was common for each of the four
groups. As shown in table 2, there were significant differ-
ences in comorbidity across the four groups for all diag-
nostic categories other than dysthymia and generalized
anxiety disorder. The neither-symptom group had the
lowest absolute comorbidity rates for the remaining six
diagnostic categories. The fluctuating-symptom group
had the highest absolute rates of comorbidity for mania,
panic disorder, simple phobia, bulimia nervosa, and an-
tisocial personality. The typical- and reversed-symptom
groups tended to have intermediate comorbidity rates for
these diagnostic categories.
The only statistically significant difference between
the typical- and reversed-symptom groups was a much
higher rate of bulimia nervosa in the reversed-symptom
group.
Nonparametric tests were used to compare the num-
bers of comorbid disorders (from the Composite Inter-
national Diagnostic Interview) experienced in the past
year across groups. This revealed a significant differ-
ence across the four groups, with the reversed-symptom
group having the most comorbid disorders and the nei-
ther-symptom group the least comorbidity. The re-
versed-symptom group was also significantly different
from the typical-symptom group.
Substance Abuse
As shown in table 3, the patterns of substance abuse
were fairly consistent. For substances other than alco-
hol, the fluctuating-symptom group consistently had
the highest frequency of substance abuse and the nei-
ther-symptom group the lowest absolute frequency.
The differences across the four groups met statistical
significance for sedative abuse and stimulant abuse.
The typical- and reversed-symptom groups did not dif-
fer significantly on any of these items.
Disability
There was no significant difference in number of days
in the past 30 days on which individuals were unable to
work (table 4). There was a significant difference across

936 Am J Psychiatry 154:7, July 1997

 LEVITAN, LESAGE, PARIKH, ET AL.

TABLE 2. Lifetime Comorbidity for 653 Persons With Lifetime Major Depression in a Community Sample, by Presence of Reversed Neurovege-
tative Symptoms
Type of Depressive Episodes Over Lifetimeb Group Comparisons
Neither Typical Both Typical Typical Only
nor Reversed Typical Only Reversed Only and Reversed Versus Reversed
Diagnostic Variablea (N=346) (N=195) (N=74) (N=38) All Groups Only
2
N % N % N % N % χ (df=3) p χ2 (df=1) p
Lifetime diagnoses
Mania 18 5.5 24 12.7 12 16.9 9 24.3 21.0 0.001 0.8 n.s.
Dysthymia 83 24.2 51 26.2 24 32.4 8 21.1 2.6 n.s. 1.1 n.s.
Generalized anxiety
disorder 42 12.4 34 17.9 17 23.3 5 13.2 6.9 n.s. 1.0 n.s.
Panic disorder 29 8.7 28 14.5 15 21.4 11 28.9 18.7 0.001 1.8 n.s.
Social phobia 87 26.1 75 39.3 37 53.6 15 40.5 24.1 0.001 4.3 n.s.
Simple phobia 88 27.8 71 39.2 29 43.9 21 58.3 19.5 0.001 0.4 n.s.
Bulimia nervosa 3 0.9 4 2.1 9 12.2 6 15.8 42.8 0.0001 11.8 0.001
Antisocial personality 14 4.4 16 8.6 5 7.0 9 23.7 20.0 0.001 0.2 n.s.

95% 95% 95% 95% Kruskal-

Confidence Confidence Confidence Confidence Wallis χ2 Mann-
Mean Interval Mean Interval Mean Interval Mean Interval (df=3) p Whitney z p
Number of disorders in
past year 1.3 1.1–1.4 1.8 1.6–2.1 2.5 2.0–2.9 2.2 1.5–2.9 36.4 0.0001 –2.6 0.01
aSubjectswere assessed with a modified version of the Composite International Diagnostic Interview (15), and diagnoses were made according
to DSM-III-R criteria.
bDiscrepancies in percentages are due to removal of subjects with incomplete data from corresponding denominator.

TABLE 3. Lifetime Prevalence of Substance Abuse Among 653 Persons With Lifetime Major Depression in a Community Sample, by Presence
of Reversed Neurovegetative Symptoms

Type of Depressive Episodes Over Lifetimea Group Comparisons

Neither Typical Only

Typical nor Typical Reversed Both Typical Versus
Reversed Only Only and Reversed All Groups Reversed Only
(N=346) (N=195) (N=74) (N=38) 2
χ χ2
Substance N % N % N % N % (df=3) p (df=1) p
Sedatives 14 4.1 18 9.3 9 12.2 10 26.3 27.2 0.0001 1.2 n.s.
Tranquilizers 34 9.9 22 11.4 8 10.8 10 26.3 9.2 n.s. 0.8 n.s.
Stimulants 25 7.3 18 9.4 9 12.2 15 39.5 38.6 0.0001 1.6 n.s.
Analgesics 58 17.0 49 25.7 20 27.0 12 32.4 10.1 n.s. 1.6 n.s.
Cocaine 29 8.4 28 14.6 9 12.3 9 23.7 10.5 n.s. 0.2 n.s.
Alcohol
Abuse 27 7.8 16 8.3 7 9.5 3 7.9 0.2 n.s. 0.1 n.s.
Dependence 40 11.6 28 14.5 13 17.6 9 23.7 5.4 n.s. 0.4 n.s.
aDiscrepancies in percentages are due to removal of subjects with incomplete data from corresponding denominator.

groups in the number of days of limited activity during
the past 30 days. A similar pattern was found for num-
ber of days during the past 30 on which effort was re-
quired to perform activities. The typical- and reversed-
symptom groups did not differ significantly from one
another on any of the disability variables.
rates. In particular, the fluctuating-symptom group re-
ported the highest absolute number of visits to a health
care professional in the past year and had a high rate of
lifetime use of a psychiatrist. The typical- and reversed-
symptom groups were not different from one another
on any of the health care utilization variables.

Health Care Utilization

As shown in table 5, the proportion of individuals
who had seen a family doctor either during their life-
times or in the past year did not differ significantly
among groups. Otherwise, the neither-symptom group
had the lowest absolute rates of health care utilization,
and the fluctuating-symptom group had the highest
DISCUSSION
The goal of the current project was to extend prior
work by further examining the epidemiology of re-
versed neurovegetative symptoms of depression in a
large, unbiased community sample. On the basis of ob-
servations derived from our clinic, patients with major

Am J Psychiatry 154:7, July 1997 937

REVERSED DEPRESSIVE SYMPTOMS

TABLE 4. Recent Disability of 653 Persons With Lifetime Major Depression in a Community Sample, by Presence of Reversed Neurovegetative
Symptoms
Group Comparisons
Type of Depressive Episodes Over Lifetime
Typical
Neither Typical Both Typical Only Versus
nor Reversed Typical Only Reversed Only and Reversed Reversed
(N=346) (N=195) (N=74) (N=38) All Groups Only
95% 95% 95% 95% Kruskal- Mann-
Measure of Disability in Confidence Confidence Confidence Confidence Wallis Whitney
Past 30 Days Mean Interval Mean Interval Mean Interval Mean Interval χ2 (df=3) p z p
Number of days on which
person was unable to work 2.7 1.6–3.9 4.3 2.4–6.2 2.2 0.9–3.5 2.5 0.5–4.6 8.0 n.s. –0.1 n.s.
Number of days of limited
activity 3.2 2.1–4.4 4.9 3.2–6.7 6.1 3.0–9.2 5.0 2.1–7.9 13.2 0.01 –1.8 n.s.
Number of days it took an
effort to perform activities 3.6 2.3–4.9 4.2 2.5–5.9 4.5 2.5–6.4 4.2 1.5–6.8 11.5 0.01 –2.3 n.s.

TABLE 5. Health Care Utilization by 653 Persons With Lifetime Major Depression in a Community Sample, by Presence of Reversed Neurovege-
tative Symptoms
Type of Depressive Episodes Over Lifetimea Group Comparisons
Neither Typical Both Typical Typical Only
nor Reversed Typical Only Reversed Only and Reversed Versus
Utilization Variable (N=346) (N=195) (N=74) (N=38) All Groups Reversed Only
N % N % N % N % χ2 (df=3) p χ2 (df=1) p
Seen family doctor
Lifetime 160 46.2 94 48.2 47 63.5 23 60.5 9.3 n.s. 5.0 n.s.
Past year 60 17.3 46 23.6 23 31.1 12 31.6 10.3 n.s. 1.0 n.s.
Seen psychiatrist
Lifetime 92 26.6 54 27.7 32 43.2 21 55.3 19.9 0.001 6.0 n.s.
Past year 21 6.1 22 11.3 12 16.2 8 21.1 15.0 0.01 1.2 n.s.

95% 95% 95% 95% Kruskal-

Confidence Confidence Confidence Confidence Wallis χ2 Mann-
Mean Interval Mean Interval Mean Interval Mean Interval (df=3) p Whitney z p
Number of visits to
health care profes-
sional in past year 2.7 1.8–3.7 6.0 3.5–8.4 7.4 4.3–10.4 9.4 3.9–15.0 21.7 0.0001 –1.7 n.s.
aDiscrepancies in percentages are due to removal of subjects with incomplete data from corresponding denominator.

depression who had never experienced an episode of
typical or reversed symptoms and those who had expe-
rienced both types of episodes over time were included
in our analysis.
A striking and unexpected finding in the current
study was that most of the identified differences among
groups were attributable to the distinct nature of the
neither- and fluctuating-symptom groups; the typical-
and reversed-symptom groups were markedly similar
overall. This suggests that our strategy of identifying
four as opposed to two depressive groups on the basis
of neurovegetative patterns may be highly pertinent;
most work on major depression has not included pa-
tients with neither or fluctuating neurovegetative pat-
terns. Our data also suggest that these four patterns of
neurovegetative symptoms may reflect varying degrees
of affective dysregulation and overall psychopathology
and that they may be influenced by gender. In general,
subjects in the neither-symptom group had less overall
comorbidity and disability, while those in the fluctuat-
ing-symptom group had the highest comorbidity for de-
bilitating psychiatric disorders such as mania and bu-
limia nervosa, had the most substance abuse and anti-
social characteristics, and had the highest rates of visits
to psychiatrists and other health care professionals.
With respect to gender, individuals with neither re-
versed nor typical episodes had proportionately more
men than did the other groups. This suggests that in
women the mechanisms underlying mood regulation,
eating behavior, and sleep may be more closely linked
than in men.
A novel finding in the current study was the identifi-
cation of individuals who had had both reversed and
typical depressive episodes at different periods in their
lifetimes. Given their high rates of comorbidity, sub-
stance abuse, antisocial traits, and health care utiliza-
tion, identification of these individuals may be an im-
portant consideration in the interpretation and design
of studies of major depression (particularly when re-
versed neurovegetative symptoms are a focus). Al-
though it was not formally assessed in the current
study, many of these individuals might fit into Akiskal’s

938 Am J Psychiatry 154:7, July 1997


“borderline” realm of subaffective disorders character-
ized by affective dysregulation and impulsivity (17). It
has been proposed historically that depression with re-
versed features is more closely related to the personality
disorders than it is to the affective disorders (18). On
the basis of the current findings, we speculate that this
might be true of many individuals who fluctuate be-
tween having reversed and typical lifetime episodes, but
not of the group with reversed episodes only. Research
on personality disorders in the four identified groups
would be necessary to investigate this issue directly.
While persons with both types of lifetime depressive
episodes account for a small proportion of major de-
pression in the Ontario community sample (5.8%),
they represent about one-third of the individuals with
lifetime histories of reversed symptoms. In epidemio-
logic work, inclusion of these individuals among de-
pressed persons classified as having pure reversed
symptoms would likely contaminate the results by sug-
gesting higher rates of severe psychopathology than ac-
tually exist in subjects who have experienced only re-
versed symptoms over the course of their depressive
disorders. This may have been a factor in the high rates
of panic disorder and substance abuse found by Hor-
wath et al. (11) in their subjects with “atypical” (re-
versed) symptoms. These authors reported that 16% of
depressed patients had experienced a depressive episode
with reversed features at some point in their lifetimes,
and this proportion corresponds closely with the total
for the reversed- and fluctuating-symptom groups in
our study (17.2% of the depressed population). The
current results suggest that high rates of panic disorder
and substance abuse are not characteristic of individu-
als with only reversed lifetime symptoms, at least in re-
lation to patients with only typical symptoms.
Another important finding in the current study was
that over 50% of the individuals with major depression
over their lifetimes had never had an episode with either
a typical or a reversed pattern. This group differs from
the other groups in having lower absolute rates of co-
morbidity, disability, and health care utilization. In
many prior studies of major depression, this neither-
symptom group would likely have been included, by
default, in either a typical- or a reversed-symptom
group. Such inclusion would tend to alter the pheno-
menological, biological, and treatment response pro-
files of the groups under consideration. In the study by
Horwath et al. (11), for example, it is possible that
many patients with neither typical nor reversed symp-
toms were included in their “nonatypical” depressed
group, and this classification might have contaminated
the results by yielding an overestimation of the relative
psychopathology of the “atypical” group. Overall, the
large size and distinct characteristics of the neither-
symptom group support its inclusion in future research
on major depression.
The reversed- and typical-symptom groups were
strikingly similar on measures of demography, comor-
bidity, disability, and health care utilization. Particu-
larly surprising were the similar rates of dysthymia,
Am J Psychiatry 154:7, July 1997
LEVITAN, LESAGE, PARIKH, ET AL.
panic disorder, and other anxiety disorders in these two
groups. Zisook et al. (19) reported similar findings in a
clinic sample of patients with major depression. A re-
cent community study of female twin pairs (20) showed
no association between depression with reversed
neurovegetative features and anxiety disorders. These
various results undermine the results of previous studies
of depression with reversed neurovegetative features,
which indicate a high rate of anxiety disorders, in par-
ticular panic disorder, in this group (21, 22). The cur-
rent results also point to similar degrees of disability in
depressions with reversed and with typical features,
suggesting that depressions with reversed symptoms are
not uniformly mild in nature.
While our use of a community-based sample may help
avoid the bias inherent in clinic-based research, a number
of limitations of the current study merit consideration.
All the information gathered was based on retrospective
reporting, a potential problem for lifetime diagnoses in
particular. Furthermore, the combination of lifetime and
current symptom reporting may be problematic in that
recall for specific neurovegetative symptoms over time
may be inaccurate. Notwithstanding these potential dif-
ficulties, there is no obvious reason to expect differential
recall or reporting of symptoms across the four identified
groups, with the possible exception of the fluctuating-
symptom group, given their severe psychopathology and
antisocial characteristics. It is also possible that the pro-
portion of individuals who refused to undergo an inter-
view in the first place differed across groups; if the fluc-
tuating-symptom group in particular had a high rate of
refusal, our data would underestimate their true preva-
lence in the community.
The choice of diagnostic instrument and use of lay
interviewers may be problematic. There is evidence that
lay interviewers may overdiagnose major depression
(23), which could have led in particular to an overesti-
mate of the number of persons with neither symptom
pattern; it is possible that some of these individuals
would have been diagnosed with adjustment disorders
or minor depressions by skilled clinicians.
The current results support a novel approach to the
study of depressive subtypes based on neurovegetative
features. In particular, our data suggest that the identi-
fication of four rather than two groups (i.e., inclusion
of a group with both typical and reversed episodes and
a group with neither type of episode) may be an impor-
tant consideration in studies of major depression. Strik-
ingly, most of the differences found among the four
groups were due to the unique characteristics of the
fluctuating- and neither-symptom groups rather than
differences between the typical- and reversed-symptom
groups. The fluctuating-symptom group appeared to
suffer from affective dysregulation and impulsivity, and
it had the highest rates of severe psychopathology, sub-
stance abuse, and use of psychiatrists and other health
professionals, while the neither-symptom group ac-
counted for the majority of individuals and was charac-
terized by low overall morbidity.
The finding that about one in three individuals with
939
 REVERSED DEPRESSIVE SYMPTOMS
reversed symptoms also experiences typical symptoms
at some point suggests that reversed symptoms are less
robust over time than are typical symptoms; in our sam-
ple, fewer than one in six individuals with typical symp-
toms had experienced a depressive episode with re-
versed symptoms in their lifetimes. Our results also
suggest that some popularly held beliefs about depres-
sion with reversed features, such as a high rate of panic
disorder and low disability in this group, may not hold
true when only patients with pure reversed symptoms
are considered.
REFERENCES
1. Kraepelin E: Psychiatrie, ein Lehrbuch fur Studierande und Arzte
(Psychiatry: A Textbook for Students and Practitioners). Leipzig,
Germany, Barth, 1896
2. Robins E, Guze SB: Classification of affective disorders: the pri-
mary-secondary, the endogenous-reactive and the neurotic-psy-
chotic, in Recent Advances in the Psychobiology of Depressive
Illnesses. Edited by Williams TA, Katz MM, Shield JA. Washing-
ton, DC, US Government Printing Office, 1972
3. Quitkin FM, Stewart JW, McGrath PJ, Liebowitz MR, Harrison
WM, Tricamo E, Klein DF, Rabkin JG, Markowitz JS, Wager
SG: Phenelzine versus imipramine in the treatment of probable
atypical depression: defining symptom boundaries of selective
MAOI responders. Am J Psychiatry 1988; 145:306–311
4. Rosenthal NE, Sack DA, Gillin JC, Lewy AJ, Goodwin FK, Dav-
enport Y, Mueller PS, Newsome DA, Wehr TA: Seasonal affec-
tive disorder: a description of the syndrome and preliminary
findings with light therapy. Arch Gen Psychiatry 1984; 41:72–80
5. Levitan RL, Kaplan AS, Brown GM, Joffe RT, Levitt AJ: Low
plasma cortisol in bulimic patients with reversed neurovegetative
symptoms of depression. Biol Psychiatry 1997; 41:366–368
6. Rubinow DR, Roy-Byrne P: Premenstrual syndromes: overview
from a methodologic perspective. Am J Psychiatry 1984; 141:
163–172
7. Himmelhoch JM, Fuchs CZ, Symons BJ: A double-blind study of
tranylcypromine treatment of major anergic depression. J Nerv
Ment Dis 1982; 170:628–634
8. Thase ME, Carpenter L, Kupfer DJ, Frank E: Atypical depres-
sion: diagnostic and pharmacologic controversies. Psychophar-
macol Bull 1991; 27:17–22
9. Liebowitz MR, Quitkin FM, Stewart J, McGrath PJ, Harrison
W, Rabkin J, Tricamo E, Markowitz JS, Klein DS: Phenelzine vs
imipramine in atypical depression. Arch Gen Psychiatry 1984;
41:669–677
940
View publication stats
10. Vanderpool J, Rosenthal NE, Chrousos GP, Wehr TA, Gold PW:
Evidence for hypothalamic CRH deficiency in patients with sea-
sonal affective disorder. J Clin Endocrinol Metab 1991; 72:
1382–1387
11. Horwath E, Johnson J, Weissman MM, Hornig CD: The validity
of major depression with atypical features based on a community
study. J Affect Disord 1992; 26:117–126
12. Regier DA, Myers JK, Kramer M, Robins LN, Blazer DG, Hough
RL, Eaton WW, Locke BZ: The NIMH Epidemiologic Catch-
ment Area Program: historical context, major objectives, and
study population characteristics. Arch Gen Psychiatry 1984; 41:
934–941
13. Eaton WW, Kessler LG: Epidemiologic Field Methods in Psy-
chiatry: The NIMH Epidemiologic Catchment Area Program.
Orlando, Fla, Academic Press, 1985
14. Parikh SV, Wasylenki D, Goering P, Wong J: Mood disorders:
rural/urban differences in prevalence, health care utilization and
disability in Ontario. J Affect Disord 1996; 38:57–65
15. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M,
Eshleman S, Wittchen H-U, Kendler KS: Lifetime and 12 month
prevalence of DSM-III-R psychiatric disorders in the United
States. Arch Gen Psychiatry 1994; 51:8–19
16. Wittchen H-U: Reliability and validity studies of the WHO-
Composite International Diagnostic Interview (CIDI): a critical
review. J Psychiatr Res 1994; 28:57–84
17. Akiskal HS: Subaffective disorders: dysthymic, cyclothymic and
bipolar II disorders in the “borderline” realm. Psychiatr Clin
North Am 1981; 4:26–46
18. Schwartz A, Schwartz RM: The formal diagnosis of depression,
in Depression: Theories and Treatments. New York, Columbia
University Press, 1993, pp 23–33
19. Zisook S, Shuchter SR, Gallagher T, Sledge P: Atypical depres-
sion in an outpatient psychiatric population. Depression 1993;
1:268–274
20. Kendler KS, Eaves LJ, Walters EE, Neale MC, Heath AC, Kessler
RC: The identification and validation of distinct depressive syn-
dromes in a population-based sample of female twins. Arch Gen
Psychiatry 1996; 53:391–399
21. Sheehan DV, Sheehan KH: The classification of anxiety and hys-
terical states, part I: historical review and empirical delineation.
J Clin Psychopharmacol 1982; 2:235–244
22. Liebowitz MR, Quitkin FM, Stewart JW, McGrath PJ, Harrison
WM, Markowitz JS, Rabkin JG, Tricamo E, Goetz DM, Klein
DF: Antidepressant specificity in atypical depression. Arch Gen
Psychiatry 1988; 45:129–139
23. Romanoski AJ, Folstein MF, Nestadt G, Chahal R, Merchant A,
Brown CH, Gruenberg EM, McHugh PR: The epidemiology of
psychiatrist-ascertained depression in DSM-III depressive disor-
ders: results from the Eastern Baltimore Mental Health Survey
Clinical Reappraisal. Psychol Med 1992; 22:629–655
Am J Psychiatry 154:7, July 1997