Vous êtes sur la page 1sur 7

Acta Psychiatr Scand 2007: 115 (Suppl.

433): 14–20 Copyright  Blackwell Munksgaard 2007

All rights reserved ACTA PSYCHIATRICA

Resurrecting melancholia
Fink M, Taylor MA. Resurrecting melancholia. M. Fink1, M. A. Taylor2
Department of Psychiatry and Behavioral Science,
Objective: To define melancholia as a distinct mood disorder, SUNY at Stony Brook School of Medicine, St James, NY
identified by unremitting depressed mood, vegetative dysfunction, and and 2Department of Psychiatry at the University of
psychomotor disturbances, verifiable by neuroendocrine tests, and Michigan, Ann Arbor, MI, USA
treatable by electroconvulsive therapy and tricyclic antidepressants.
Method: A review of the literature of two centuries finds descriptions Professors Fink and Taylor declare no conflict of
of severe mood disorders, either depression or mania or circular,
defined as Ômelancholia.Õ In the 1980 diagnostic revision (DSM-III),
melancholia was relegated to a features specifier only.
Results: DSM classification criteria develop heterogeneous patient
samples that are neither guides to prognosis nor to treatment response,
and confound studies of pathophysiology. Within the large population
of mood disorders, a syndrome of melancholia is identifiable by specific
behaviors, vegetative signs, and validated by neuroendocrine Key words: melancholia; classification;
abnormalities (cortisolemia). Populations so identified are clinically pathophysiology; hypercortisolemia; tricyclic
homogeneous and have improved treatment responses. Patients antidepressant; electroconvulsive therapy
meeting criteria for melancholia are now identified as psychotic Max Fink, Department of Psychiatry and Behavioral
depressed, geriatric depressed, postpartum psychosis, and Science, SUNY at Stony Brook School of Medicine,
pharmacotherapy resistant. PO Box 457, St James, NY 11780, USA.
Conclusion: The review supports the establishment of melancholia by E-mail: mafink@attglobal.net
empirically derived criteria rather than by a checklist is an alternative Presented at the Conference: ÔMelancholia: Beyond
to the major depression choice and offers an improved model for DSM, Beyond NeurotransmittersÕ, May 2–4, 2006,
psychiatric classification. Copenhagen, Denmark.

chotic unipolar depression were first treated with

citalopram, and 30% responded. The remaining
The DSM-III classification of psychiatric disorders patients were offered one of three antidepressant
of 1980 was a hoped-for improvement in nosology, agents, to which an additional one-quarter respon-
as it sought an operationally defined system based ded, leaving half the subjects without symptomatic
on overt behaviors. In this and subsequent itera- relief in repeated trials. The study does not offer
tions, the core mood disorder became a monolithic effective guidelines for subject selection for these
Ômajor depressive disorderÕ with modifiers for medications.
differences in severity (e.g. psychotic or not), and The out-patient population sample was hetero-
for circumstances presumed to have clinical signi- geneous. Although the severity of the illness was
ficance (e.g. abnormal bereavement, postpartum rated as moderate (HDRS (Hamilton Depression
depression). The traditional concept of Ômanic- Rating Scale)17 score 19 ± 7.3), 17% reported
depressionÕ was separated into unipolar and bipo- suicide attempts, 76% suffered recurrent illnesses,
lar categories. The well-established concept of the duration of the present episode averaged
melancholia was discarded, and laboratory tests 30 months, the lifelong duration of illness was
were excluded from the criteria (1). 17 years, and the subjects reported an average of
The diagnostic criteria that were formulated, seven episodes of illness. Features of anxious
however, are imprecise, inadequate for treatment depression were recorded in 44% and atypical
decisions, and do not assure homogeneous popu- depression in 20%. The criteria for Ômajor depres-
lations in clinical trials, thereby missing opportun- sionÕ failed to distinguish those with characterolog-
ities to define the specific actions that are sought in ical and social influences on their illness from those
new treatments. A cogent example is the evaluation with a distinctive biological basis. The heterogeneity
of antidepressants in the STAR*D study (2). In a of the sample is one explanation for the weak results.
US government-supported multi-site collaborative Guidelines to establish diagnostic validity for a
effort, out-patients meeting criteria for non-psy- clinical condition, and therefore maximize homo-

Resurrecting melancholia

geneity, were offered by Robins and Guze (3). They Melancholia has been ignored in each subse-
defined a clinical syndrome as the delineation of quent DSM revision. To acknowledge the many
characteristic signs and symptoms, supported by conditions that could be described, DSM-III listed
laboratory tests or biologic markers, with a char- 265 disorders, increased to 292 in DSM-IIIR, and
acteristic clinical course and response to treat- to 295 in DSM-IV (20). A similar increase in the
ments. We applied these principles in an earlier number of putative disorders is found in the ICD
assessment of catatonia. We described the charac- classification system (21).
teristic motor signs in examination and rating
scales, validated the syndrome by a lorazepam
From major depression to melancholia
response test, and successfully treated patients with
sedative anticonvulsant medications and convul- In our present system of diagnosis, we identify
sive therapy (4, 5). patients who meet criteria for major depression by
Using a similar method, we propose the cross-sectional, unweighted features. Persons who
replacement of the present potpourri of mood feel saddened by life events or who are disgruntled
disorders in DSM-III and DSM-IV that serves and unhappy by virtue of life’s vicissitudes or by
our patients poorly. We demarcate a prototypic their personality traits meet criteria for major
syndrome of melancholia by specific psychopa- depression if these moods are associated with
thology, laboratory tests, and effective treatments decreased energy or interest. Their inclusion
(1). The remaining mood disorders are categor- within the class of major depression confounds
ized in a non-homogeneous class of non-melan- the results of clinical treatment trials and laborat-
cholic disorders, awaiting the definition of other ory searches for an understanding of mood disor-
distinctive syndromes using the same principles. ders. The large numbers of placebo responders in
An example is offered by the analysis of atypical such samples attest to the need for a higher
depression by Stewart et al. (6) in this issue (see threshold to identify depressed persons that war-
pp. 58–71). rant somatic treatment (1, 2).
Quality of mood (unremitting apprehension and
sadness), psychomotor disturbance (agitation or
The concept of melancholia
bradykinesia), multiple vegetative signs, and
A syndrome of ÔmadnessÕ with Ôbodily causesÕ has psychosis are essential elements in defining melan-
been clinically recognized for centuries. Except for cholia. The present psychiatric classification does
two periods in Western history – the Middle Ages not require these features in the diagnosis of a
in Europe when church teachings dominated depressive illness.
Western thought and in the 20th century when Abnormal neuroendocrine tests and disturbed
psychoanalytic notions usurped psychiatric think- sleep EEG measures are common in melancholia.
ing – melancholia was identified as a distinct Hypercortisolemia is prominent, particularly when
periodic mood disorder with both manic and the melancholic patient is agitated or psychotic.
depressive phases and without progressive deteri- When measured by the dexamethasone suppression
oration (1, 7–12). In 20th century classifications, test (DST), cortisol functions are abnormal when
psychoanalysts defined psychiatric illnesses as patients are ill, normalize with remission, and
being of Ômental,Õ not brain origin, with depressive became abnormal again in relapse (1, 22, 23).
and manic periods as ÔreactionsÕ to personal Similar findings, although less robust, are seen in
experience. The classifications in DSM-I and thyroid function tests and sleep EEG measures (1).
DSM-II were based on this philosophy. Although developed heuristically and without a
By the late 1960s, however, it was no longer central theory, these tests buttress the melancholia
considered acceptable to treat depressed patients as diagnosis. They are equally useful (e.g. similar in
if they were all suffering from a single condition, as sensitivity and specificity) to the EEG, brain
treatments were found to be effective for some imaging, and serum prolactin levels in defining a
patients and not for others (13–16). An assessment seizure disorder.
of 33 studies of medication treatments for depres- In the psychopathological literature, melancho-
sion, for example, could not find a diagnostic lia is consistently described as a severe illness of
formulation that had predictive strength (17). The acute onset with unremitting moods of apprehen-
poor reliability of the existing classifications was sion and gloom, psychomotor disturbance, and
clearly demonstrated in the strikingly different vegetative signs. Psychosis, intermittent mania, and
diagnostic conclusions reached by American and suicide intent are prominent features. A compelling
British psychiatrists examining the same patients picture of melancholia is offered by Falret in
(18, 19). describing a circular insanity:

Fink and Taylor

At the commencement of this phase… the patients begin Table 2. Proposed inclusions in melancholia
to withdraw and now speak only rarely. Sometimes they
Melancholia Non-melancholic mood disorder
express remorse over their previous condition…
remaining all alone and motionless… they are now Psychotic depression Characterological depression
meek, and their humility may go so far as for them to Manic-depression Reactive depressive disorders
Puerperal depression Premenstrual dysphoria
refuse treatment in the belief that they do not deserve it. Abnormal bereavement
This despondency becomes more pronounced daily… Depression with stupor or catatonia
[and] the patient is transformed into a statue… were he
not coaxed to eat, the patient would not bother to seek
Verification by psychopathology
The thought processes and his movements are very slow; Almost all empirical studies over four decades
rarely this may result in complete cessation of all intel- report Ômajor depressionÕ as encompassing different
lectual activity…. The face is pale; the features sag, patient groupings with between 20% and 70% of
suggesting dejection rather than anxiety…. Appetite is the samples described as psychotic, melancholic, or
decreased, and the patient eats little; digestion is equally endogenously depressed. The less homogeneous
slow and defecation is laborious. groups are younger, less severely ill, exhibit
abnormal trait behavior, and meet criteria for
Nevertheless, there are a certain number of patients who anxiety disorder or atypical depression (1).
present with specific preoccupations, among which we The strongest ÔnegativeÕ study is that of Kendell
have noticed ideas of humility, of ruin, of being poi- (25, 26) who used data from hospital discharge
soned, or of guilt. (24) forms collected by trainee psychiatrists at the
We propose to reinstitute the definition of Maudsley Hospital between 1949 and 1963. Ken-
melancholia, set a duration criterion of two dell failed to find a bimodal distribution of patient
weeks, and add as secondary criteria the associated characteristics and concluded that depressive dis-
laboratory findings of dexamethasone non-sup- orders were best viewed as a continuum. Although
pression of cortisol, high nighttime cortisol levels, the Maudsley report form did not include items
or decreased REM latency or other characteristic needed to delineate melancholia and the inter-rater
sleep abnormalities (Table 1). reliability of the raters is unknown, we looked
Applying such criteria to the entities described in carefully at the two factors that were delineated
DSM-IV finds that patients with diverse syn- (Table 3). Kendell clearly identified a Ôpsychotic
dromes as presently defined meet the criteria for depressionÕ group that closely matches criteria for
melancholia (Table 2). Patients with mood disor- the melancholia syndrome.
ders who do not meet criteria for melancholia Over 70 other studies in the past quarter century
comprise a large heterogeneous population best refute Kendell’s conclusion and support the finding
labeled for the present as within the non-melan- of a melancholia syndrome. For example,
cholic mood disorders. Matussek et al. (27), Maes et al. (28), Parker and
The proposed criteria for melancholia have face his coworkers (12, 29), Sullivan et al. (30) and
validity for high specificity. They make false- Ambrosini et al. (31) identify clusters of depressed
positive diagnoses unlikely, and maximize the patients with distinct qualities of mood, loss of
identification of homogeneous population samples reactivity, social withdrawal, loss of energy,
for clinical trials and studies of pathophysiology. psychomotor abnormalities of paucity of speech
They meet the three principal criteria for a clinical and movement, vegetative signs and disturbances
syndrome delineated by Robins and Guze (3).
Table 3. Kendell's two-depression factors

Table 1. Proposed diagnostic criteria for melancholia (1) (all must be present) Neurotic depression Psychotic depression

A. An episode of illness with reduced functioning characterized by an unremitting Previous anxiety symptoms Gross disturbance in food intake
mood of apprehension and gloom that compromises normal daily activities and Previous subjective tension Gross disturbance in weight
persists for at least 2 weeks Previous ÔhystericalÕ symptoms Severe insomnia
B. Psychomotor disturbance as agitation, retardation (including stupor and cata- Childhood neurotic Retarded activity
tonia), or both traits/suicidal feelings Abnormal quantity of speech
C. Vegetative signs (at least two). Previous obsessional symptoms Social withdrawal
D. At least one of the following: Delusions (guilt, self-reproach,
Abnormal dexamethasone suppression or dexamethasone-suppressed cortico- worthlessness)
tropin-releasing hormone test (DEX/CRH); high nighttime cortisol levels Ideas of reference, suspiciousness,
Decreased REM latency or other sleep abnormalities persecutory delusions

Resurrecting melancholia

in circadian rhythms, and suicidal thoughts. Such Doing the test properly, however, is difficult. Each
patients have a poor prognosis for marginally laboratory needs to establish local normative stand-
effective interventions and require intensive bio- ards. Monitoring plasma dexamethasone levels is
logically based treatments (1, 11, 27–31). necessary as the systemic response is sensitive to
dexamethasone dosage and serum levels. Preg-
nancy, end-stage dementia, substantial weight loss,
Verification by laboratory tests
and interference from other agents (e.g. steroids,
Few laboratory tests are associated with abnormal some anticonvulsants) affect the test and are neces-
behaviors. Aside from serological tests for syphilis sarily considered in assessing its results.
and systemic examinations for infection or meta- The DST was widely recognized as a state marker,
bolic error, the EEG has been the most helpful test but it was also examined for its correlation with
of a behavior syndrome. Single inter-ictal record- clinical diagnoses. The DST is commonly abnormal
ings exhibit seizure rhythms in 50–55% of patients in patients with delusional (psychotic) depression. A
with well-defined epilepsy. Two recordings increase meta-analysis of 14 studies comparing DST results
sensitivity to 92% (32). It is difficult to conceive of in psychotic and non-psychotic depressed patients
a medical examination that considers epilepsy found the non-suppression rate to be substantially
without examining EEG recordings. This standard higher in psychotic depressed patients (36). Patients
is useful in discussing the role of laboratory tests in with schizophrenia typically do not show abnormal
melancholia. cortisol levels, arguing that the high levels are
Hypercortisolemia is a feature of pituitary and characteristic not of psychosis, but of mood disorder
adrenal pathology. In the search for normative (1).
standards, highly abnormal levels were found in The frequency of non-suppression in manic
melancholic depressed patients. The finding stimu- patients varied from 0% to 70% with the highest
lated an interest in cortisol physiology and led to frequency in mixed bipolar disorder (1).
the development of the DST (22). Depressed patients with abnormal DST have a
Elevated levels of cortisol, loss of diurnal rhyth- higher risk for suicide. In a 15-year follow-up study
micity, and impaired suppression of cortisol by the suicide risk in those with an abnormal DST was
dexamethasone were well documented in severely 27% compared with 3% among patients with a
depressed patients, especially in those with weight normal DST. A review of 101 patients re-examined
loss, agitation, suicide risk, and psychosis. With over 2 years confirmed the higher risk for suicide
successful antidepressant treatment, plasma corti- and indicated hospitalization for suicide risk in
sol levels fell to the levels of normal subjects. Carroll those with abnormal DST. Patients with abnormal
concluded that: ÔThe suppression test distinguished cortisol metabolism are more likely to make suicide
clearly between the depressed and the control attempts (1).
patients who could not be distinguished simply on DSM-III was introduced at the time that the
the basis of their diurnal plasma cortisol levelsÕ (33). DST was being assessed, and its role as a diagnos-
By 1976, the DST was established as a measure tic test was examined. By 1985, reviewers conclu-
of severe depression. Greater Hypothalmic-Pitui- ded that:
tary-Adrenal (HPA) activity before dexamethasone
The sensitivity of the DST in major depression is limited
and less complete HPA suppression following its
(about 44% in over 5000 cases) but is higher in psychotic
administration was reported in 42 patients with
affective disorders and mixed manic-depressive states
Ôendogenous depressionÕ compared with 42 patients
(67% to 78%). The high specificity of the DST vs con-
with other psychiatric disorders. Patients with two
trol subjects (over 90%) is not maintained vs other
or more abnormal cortisol values after dexameth-
psychiatric disorders (77% specificity overall), and acute
asone were correctly identified as being endogen-
ÔdistressÕ may contribute to non-suppression of cortisol
ously depressed (34).
By 1981, the DST was validated Ô…as a test for
the diagnosis of melancholia (endogenous depres- These authors concluded:
sion)Õ in 438 patients using a cut-off level of 5 lg/dl
The test may have power in differentiating severe
plasma cortisol after 1 mg dexamethasone admin-
melancholic depression, mania, or acute psychosis from
istration (35). Two blood samples taken at 16.00
chronic psychosis (87% specificity) or dysthymia (77%
and 21.00 h after dexamethasone administered at
21.00 h the night before Ôdetected 98% of the
abnormal test results. This version of the DST The DST was discarded from the clinic and in
identified melancholic patients with a sensitivity of psychiatric research. Yet, as recently as 1996, Rush
67% and a specificity of 96%Õ. et al. (38), in a large study of patients meeting

Fink and Taylor

criteria for both unipolar and bipolar depression ders, and almost all patients in a melancholic or
by RDC criteria, reported cortisol non-suppression manic episode exhibit some features of catatonia
in 35% of endogenously ill out-patients and in 9% (e.g. automatic obedience, ambitendency) (3, 4).
of non-endogenously ill. Among in-patients, non- Furthermore, the broad action of penicillin in
suppression was found in 62% and 19% respect- bacterial and spirochaetal infections does not
ively. The sensitivity was assessed as 40%, the invalidate its specificity for either condition.
specificity at 90%. In placebo responders, 10% are The multi-site collaborative study of continu-
DST non-suppressors and 70% are suppressors (1). ation ECT and continuation pharmacotherapy
The weight of the evidence indicates that the DST (known as CORE) used bilateral ECT in patients
is a valid test of the presence of a severe depressive meeting rigorous criteria for unipolar major
mood disorder, especially in those meeting criteria depression. It reported an overall remission rate
for psychotic and melancholic depression. Like the of 87% among treatment completers, with an
EEG in seizure disorders, measures of cortisol astonishing 95% rate for the 30% of patients
physiology are not simple tests, but offer a useful identified as psychotic depressed based on Struc-
criterion in defining an illness. A positive test assures tured Clinical Interview for DSM-IV (SCID)
the diagnosis; a negative test encourages further criteria (41, 42). Although characterized as meeting
assessment. Additional measures as the thyroid criteria for major depression, the major part of the
stimulating hormone (TSH) response to thyrotro- CORE sample meets the criteria for melancholia
pin-releasing hormone (TRH) and the sleep EEG based on severity of the mood disorder, high
offer further validation criteria (1). incidence of psychosis and suicide risk, and failure
The dexamethasone-suppressed corticotropin- to respond to multiple medication trials. A post-
releasing hormone (DEX/CRH) measure of cort- hoc analysis of the items rated on the Hamilton
isol physiology improves sensitivity (1). It took Depression Rating Scale (HAMD)24, SCID, and
more than half a century of experience from the Inventory of Deppressive Symptomatology (IDS)
descriptions of the human EEG by Hans Berger in scales finds that more than 60% meet symptom
the 1930s to achieve the quantitative digital com- criteria for melancholia.
puter and video-assisted EEG diagnoses of today. Lesser remission rates of 50–60% with ECT in
Improvements in measures of cortisol and neuro- hospitalized unipolar depressed patients are still
endocrine physiology offer a similar opportunity to reported, but are understood to result from inef-
design better laboratory test criteria for melancho- ficient treatment methods (43). Successful ECT
lia. also reverses the hormone imbalances seen in
Although less well codified, studies of EEG sleep depressive illness. Carroll (33) reported a normal-
characteristics consistently find greater abnormal- ized DST in five patients treated with ECT, with
ities in hospitalized, more severely ill patients reversion to abnormality heralding relapse. The
exhibiting features of endogenous depression. same reversal in other hormone tests has been
Sleep studies, are difficult to apply clinically, but observed and these experiences are the basis for a
have heuristic research value in defining a melan- neuroendocrine explanation of the mode of action
cholic depression (1). of induced seizures in melancholia (39, 44).
Broad pharmacodynamic spectrum antidepres-
sants are more effective in melancholic than in non-
Verification by treatment
melancholic patients. Lithium moderates abnormal
The development of highly efficient antibiotics for mood and reduces suicidal drive, and is the most
bacterial infections, and vitamin and hormone effective augmenting agent in the treatment of
replacements for dietary and hormone deficiencies acute depressive illness when the patient is melan-
verified a clinical diagnosis. The astonishing effic- cholic. It is efficient as continuation therapy for
acy of ECT in remitting melancholia also supports melancholic patients, especially when combined
the diagnosis. Melancholia remits with greater than with the tricyclic nortriptyline (1, 11, 42).
90% efficiency within 3 weeks with bilateral ECT Tricyclic antidepressants effectively relieve mel-
(39, 40). Critics will note that ECT is as effective in ancholic depression but do so less efficiently than
delirious mania and catatonia as it is in melancho- ECT. The imbalance is seen in the relative ineffi-
lia, and will argue that these actions are too broad cacy of TCA in psychotic melancholic patients
to verify a diagnosis. Patients with mania, how- where remission rates are low (1, 11).
ever, almost always have episodes of depressive Although the more recently introduced selective-
illness. Over their life-time, they experience depres- serotonin reuptake inhibitor and similar agents are
sion more often than episodes of mania. Catatonia widely recommended as the first agents in treat-
is commonly defined in patients with mood disor- ment algorithms for major depression, their overall

Resurrecting melancholia

30–40% remission rates differ only minimally from In conclusion, the present classification of mood
placebo (45). They are much less effective in disorders is arbitrary, divided into a multitude of
hospitalized patients, those more likely to be presumed disorders with imprecise criteria. The
suffering from melancholic depression. These formulations do not encourage effective therapy or
experiences speak against their specificity in the definitive research. Patients meeting criteria for
pathophysiology of melancholia. By contrast, single major depression are said to respond to the
the extraordinary efficacy of ECT in inducing same treatment algorithm, with the more severe
remission, reducing suicide risk, and reversing needing higher doses or drug combinations, but
endocrine abnormalities marks the mechanism of not different approaches. The putative unipolar
action of induced seizures as an essential target for and bipolar disorders are said to require different
studies of the pathophysiology of melancholia (44). algorithms.
The evidence is the opposite: melancholia can be
distinguished from other depressive mood disorders
Unipolar–bipolar distinction
by its characteristic psychopathology. The unipolar/
Patients with mood disorder are now dichotomized bipolar dichotomy is not supported when the criteria
by a history of manic or hypomanic episodes into for melancholia are used to classify depressive
recurrent depressive or bipolar disorder. When illness. The treatment algorithm for melancholia
depressive illness is grouped by criteria of melan- differs from that of non-melancholic depressions,
cholia, however, the unipolar/bipolar dichotomy but the treatment algorithms for melancholia with
fails. No characteristic psychopathology separates or without mania are similar. We find compelling
the melancholic patient with a history of mania or evidence that hypercortisolemia is characteristic of
hypomania from one without such a history. The severely ill depressed patients, and tests of this
illnesses are commonly recurrent and in succeeding hormone (and other hormones) should be consid-
episodes 70% of patients initially classified as ered in the classification and the management of
bipolar disorder meet criteria for major depression; patients with depressive mood disorders. Various
and, in the course of a major depressive illness, treatments – lithium, tricyclic antidepressants, and
10% exhibit episodes of mania and over half ECT – effectively relieve both the depressive and
exhibit features of mania when depressed (1). manic phases of the syndrome. The expert treatment
In an extensive analysis of unipolar and bipolar algorithms thus require substantial revision if they
depression examining the course, epidemiology, are to claim to be Ôevidence basedÕ.
family history, physiological measures, incidence We propose the demarcation of a syndrome of
of clinical signs of anxiety, anger, psychomotor melancholia as the best opportunity to develop
agitation, mood lability, sleep time, and pain homogeneous study populations, more effective
sensitivity Goodwin and Jamison (11) were treatment algorithms, better understanding of the
unable to identify distinguishing characteristics. pathophysiology of mood disorders, and reducing
They found the breadth of differences and hetero- the social burden of Ôtherapy-resistantÕ depressed
geneity impressive, and yet they concluded: patients.
Taken together, the data suggest that they (mania,
depression) are best considered as two subgroups of References
manic-depressive illness rather than separate and dis- 1. Taylor MA, Fink M. Melancholia: the diagnosis, patho-
tinct illnesses. The available data also support a con- physiology and treatment of depressive illness. Cambridge,
tinuum model, with ÔpureÕ bipolar illness at one end and UK: Cambridge University Press, 2006.
unipolar illness at the other. 2. Rush AJ, Trivedi MH, Wisniewski SR et al. Bupropion-SR,
sertraline, or venlafaxine-XR after failure of SSRIs for
Others also fail to find psychopathological separ- depression. N Engl J Med 2006;354:1231–1242.
ation points for the two conditions (46–48). Histor- 3. Robins E, Guze SB. Establishment of diagnostic validity in
psychiatric illness: its application to schizophrenia. Am J
ically, the two conditions were identified as facets of Psychiatry 1970;126:983–987.
a condition of circular insanity by French psycho- 4. Fink M, Taylor MA. Catatonia: a clinician’s guide to
pathologists Falret (24) and Baillarger (49). Their diagnosis and treatment. Cambridge, UK. Cambridge
descriptions were adopted by Kraepelin (50) in the University Press, 2003.
concept of manic-depressive psychosis that has been 5. Taylor MA, Fink M. Catatonia in psychiatric classification:
a home of its own. Am J Psychiatry 2003;160:1233–1241.
a mainstay of classification until DSM-III. Demar- 6. Stewart J, Klein DF, McGrath P, Quitkin F. Heterogeneity:
cating melancholia as a defined depressive mood the hobgoblin of psychiatric diagnosis; dissecting atypical
disorder encourages the earlier views of a single depression: a roadmap to homogeneity? Acta Psychiatr
psychopathology for mania and depression. Scand Suppl 2007;433:58–71.

Fink and Taylor

7. Hopewell-Ash EL. Melancholia in everyday practice. Lon- 31. Ambrosini PJ, Bennett DS, Cleland CM, Haslam N. Taxo-
don: John Bale Sons & Danielsson, 1934. nicity of adolescent melancholia: a categorical or dimen-
8. Lewis AJ. Melancholia: a historical review. J Ment Sci sional construct? J Psychiatr Res 2002;36:247–256.
1934;80:1–42. 32. Parra J, Augustijn PB, Geeerts Y, Van Emde Boas W.
9. Schmidt-Degenhard M. Melancholie und depression. Stutt- Classification of epileptic seizures: a comparison of two
gart: Verlag W. Kohlhammer, 1983. systems. Epilepsia 2001;42:476–482.
10. Jackson SW. Melancholia and depression from Hippocra- 33. Carroll B. Control of plasma cortisol levels in depression:
tes to modern times. New Haven, CT: Yale University studies with the dexamethasone suppression test. In: Davies
Press, 1986. B, Carroll BJ, Mowbray RM, eds. Depressive illness: some
11. Goodwin FK, Jamison KR. Manic-depressive illness. New research studies, Vol. 5. Springfield, IL: C.C. Thomas,
York: Oxford University Press, 1990; 227–244. 1972:87–149.
12. Parker G, Hadzi-Pavlovic D. Melancholia: a disorder of 34. Carroll BJ, Curtis GC, Mendels J. Neuroendocrine regu-
movement and mood. Cambridge, UK: Cambridge Uni- lation in depression, II: discrimination of depressed from
versity Press, 1996. nondepressed patients. Arch Gen Psychiatry 1976;33:1051–
13. Klein DF, Fink M. Psychiatric reaction patterns to imipr- 1058.
amine. Am J Psychiatry 1962;119:432–438. 35. Carroll BJ, Feinberg M, Greden JF et al. A specific labor-
14. Klein DF, Fink M. Behavioral reaction patterns with phe- atory test for the diagnosis of melancholia. Arch Gen
nothiazines. Arch Gen Psychiatry 1962;7:449–459. Psychiatry 1981;38:15–22.
15. Fink M, Klein DF, Kramer J. Clinical efficacy of chlor- 36. Nelson JC, Davis JM. DST studies in psychotic depression:
promazine-procyclidine combination, imipramine and a meta-analysis. Am J Psychiatry 1997;154:1497–1503.
placebo in depressive disorders. Psychopharmacologia 37. Arana GW, Baldessarini RJ, Ornsteen M. The dexametha-
1965;7:27–36. sone suppression test for diagnosis and prognosis in psy-
16. Klein DF. Delineation of two-drug-responsive anxiety chiatry. Arch Gen Psychiatry 1985;42:1193–1204.
syndromes. Psychopharmacologia 1964;5:397–408. 38. Rush AJ, Giles DE, Schlesser MA et al. The dexamethasone
17. Nelson JC, Charney DS. The symptoms of major depressive suppression test in patients with mood disorders. J Clin
illness. Am J Psychiatry 1981;138:1–13. Psychiatry 1996;57:470–484.
18. Sandifer MG, Hordern A, Timbury GC, Green LM. Simi- 39. Fink M. Convulsive therapy: theory and practice. New
larities and differences in patient evaluation by U.S. and York: Raven Press, 1979.
U.K. psychiatrists. Am J Psychiatry 1969;126:206–212. 40. Abrams R. Electroconvulsive therapy, 4th edn. New York:
19. Spitzer RL, Fleiss JL. A re-analysis of the reliability of Oxford University Press, 2002.
psychiatric diagnosis. Brit J Psychiatry 1974;125:341–347. 41. Petrides G, Fink M, Husain MM et al. ECT remission rates
20. Shorter E. A history of psychiatry from the era of the in psychotic versus non-psychotic depressed patients: a
asylum to the age of Prozac. New York: John Wiley & report from CORE. Journal of ECT 2001;17:244–253.
Sons, 1997. 42. Kellner CH, Knapp RG, Petrides G et al. Continuation ECT
21. World Health Organization. The ICD-10 classification of versus pharmacotherapy for relapse prevention in major
mental and behavioral disorders. Geneva, Switzerland: depression: a multi-site study from CORE. Arch Gen
WHO, 1992. Psychiatry 2006;63:1337–44.
22. Davies BJ, Carroll BJ, Mowbray RM. Depressive illness: 43. Sackeim HA, Prudic J, Devanand DP et al. A prospective,
some research studies. Springfield, IL: C.C. Thomas, 1972. randomized, double-blind comparison of bilateral and
23. Fink M. Should the dexamethasone suppression test be right unilateral electroconvulsive therapy at different sti-
resurrected? Acta Psychiatr Scand 2005;112:245–249. mulus intensities. Arch Gen Psychiatry 2000;57:425–434.
24. Falret JP. Mémoire sur la folie circulaire. Bulletin de 44. Fink M. Electroshock revisited. Am Sci 2000;88:162–167.
l’Academic nationale de Médicine 1854;19:382–415. 45. Khan A, Kolts RL, Rapaport MH, Krishnan KR, Brodhead
25. Kendell RE. The classification of depressive illness. Lon- AE, Brown WA. Magnitude of placebo response and drug-
don: Oxford University Press, 1968. placebo differences across psychiatric disorders. Psychol
26. Kendell RE. The classification of depressions: a review Med 2005;35:743–749.
of contemporary confusion. Br J Psychiatry 1976;129: 46. Akiskal HS, Benazzi F. Validating Kraepelin’s two types of
15–28. depressive mixed states: Ôdepression with flight of ideasÕ
27. Matussek P, Soldner M, Nagel D. Identification of the and Ôexcited depressionÕ. World J Biol Psychiatry
endogenous depressive syndrome based on the symptoms 2004;5:107–113.
and the characteristics of the course. Br J Psychiatry 47. Benazzi F. Melancholic outpatient depression in bipolar-II
1981;138:361–372. vs. unipolar. Prog Neuro-psychopharm Biol Psychiatry
28. Maes M, Cosyns P, Maes L, D’Hondt P, Schotte C. Clinical 2004;28:481–485.
subtypes of unipolar depression, Part I: a validation of the 48. Benazzi F. Intra-episode hypomanic symptoms during
vital and nonvital clusters. Psychiatry Res 1990;34:29–41. major depression and their correlates. Psychiatry Clin
29. Parker G, Hadzi-Pavlovic D, Wilhelm K et al. Defining Neurosci 2004;58:289–294.
melancholia: properties of a refined sign-based measure. Br 49. Baillarger J. De la folie à double forme. Ann Med Psychol
J Psychiatry 1994;164:316–326. 1854;6:367–391.
30. Sullivan PF, Prescott CA, Kendler KS. The subtypes of 50. Kraepelin E. Manic-depressive insanity and paranoia.
major depression in a twin registry. J Affect Disord Barclay RM, translator, Robertson GM, ed. Edinburgh:
2002;68:273–284 E&S Livingstone, 1921; repr. New York: Arno Press, 1976.