Should the Diagnosis of Melancholia Be Revived?

June 1, 2006

An appreciation of melancholia as the principal definable mood disorderoffers a better guide to diagnosis and treatment than does DSM-identified major depression.

Melancholy (n) – c.1303, “condition characterizedby sullenness, gloom, irritability,” from Gk.melankholia“sadness,” lit. “black bile,” from melas“black” + kholé “bile.”1

The treatment of depressive mooddisorders is in disarray. Despite aplethora of medications, novel psychotherapies,and expert treatment algorithms,one therapeutic trial afteranother fails, patients suffer endlessly,are labeled “therapy-resistant,” andoverwhelm our treatment facilities.

A powerful example is the report ofthe large government-supported multisiteassessment of antidepressantmedications known as STAR*D.2 Afterrelief was accorded to 30% of depressedpatients in a primary treatment coursewith the selective serotonin reuptakeinhibitor (SSRI) citalopram (Celexa),a subsequent trial in the remainder withthe antidepressants bupropion(Wellbutrin), venlafaxine (Effexor), andsertraline (Zoloft) elicited improvementin 21%, 18%, and 25% of thepatients, respectively. Sadly, half thepatients failed to obtain relief in therepeated trials. The study populationmet DSM criteria for nonpsychoticmajor depression.

What contributed to the low responserates? The weak antidepressant efficacyof the medications is a factor. From 20%to 40% of similar patient populationsimprove with placebo; these rates areequivalent to those associated with themedications studied.1,3

Heterogeneity of the populationensures a low response rate. The DSMcriteria for major depression are impreciseand fail to distinguish betweenthose with psychological and socialcauses for their illness and those witha biologic (endogenous) basis. Theseverity of the illness in the STAR*D study was moderate (HamiltonDepression Rating Scale-17 score of19 ± 7.3). Despite this modest degreeof severity reflected in depression scaleratings, 17% of the patients reportedsuicide attempts, 76% reported recurrentillnesses, the average duration ofthe present episode was 30 months, thelifelong duration of illness was 17years, and the patients reported an averageof 7 episodes of illness. Anxiousfeatures were recorded in 44% of thepatients and atypical features in 20%.Yet, of the patients in this outpatientpopulation, 53% were employed.

The STAR*D experience reflectsour difficulty in assessing treatments fora heterogeneous population of majordepressed patients. The difficulty issimilar to that of assessing the effectivenessof an antibiotic in a populationof city dwellers who, during the winterseason, complain of chest pain, cough,rhinorrhea, fever, and chills. Heterogeneityof bacterial and viral causeswould offer the researchers ambiguousresults for an effective antibiotic.

Is there an alternative to the DSMcriteria for major depression that wouldensure more homogeneous populationsamples in clinical trials?

In the 1970s, studies that monitoredserum blood levels found a cluster ofdepressed patients who failed torespond to the tricyclic antidepressantimipramine. The patients' illness did,however, remit quickly with electroconvulsivetherapy (ECT).4 The presenceof psychosis was the distinguishingcharacteristic of the medication nonresponders.The importance of psychosis as a discriminator for treatment choiceand for clinical outcome in depressionwas quickly confirmed and set the stagefor our understanding that patients withpsychotic depression warrant differenttreatment algorithms from those effectivein nonpsychotic patients. Theyrespond best to ECT or to combinationsof a tricyclic antidepressant anda conventional neuroleptic.5

The syndrome of melancholia

Rooting diagnosis in a definablepsychopathology is essential to replicatethis success. A syndrome of melancholia,like that of psychosis, is a readilydistinguishable mood disorder thatoffers a useful discriminator in depressivemood disorders.1,6

For centuries, physicians and laymendescribed an illness of mood andthought, often of sudden onset, and sosevere as to threaten life by suicide anddebilitation.1 Medieval philosophers,struggling with concepts of bodyhumors controlling health and disease,imagined that an excess of black bilewas the basis for the illness and describedthe condition as “melancholie.” Thesyndrome's pathophysiology, rating instruments,laboratory test abnormalities,and effective treatment algorithms arewell delineated by many authors.

What is melancholia?

Melancholia is the quintessentialdepressive mood disorder. Movementand thought are slowed; mood is sadand gloomy; insomnia, anorexia, poorbody care, weight loss, disinterest in sex,and fears of impoverishment, infidelity,and hopelessness dominate daily living.Suicide becomes a logical solution aspain and distress overwhelm daily life.The agony and helplessness are excruciatinglydescribed by such recent writersas Endler (1982), Styron (1990), Manning (1994),Rosenberg (2002),and Nuland (2003).7

Melancholia ismeasurable by assessingsymptoms,for which ratingscales, especiallythe Hamilton andthe Montgomery-Asberg depressionrating scales, are useful. These standardizedinstruments estimate the severityof the syndrome and are usefulguides to management.

Hypercortisolemia is one distinctionof melancholia from other mood disorders.It is a unique pathophysiologicabnormality that reflects the severityof the illness and varies with remissionand relapse.8 Hypercortisolemia wasfirst recognized in patients with pituitaryand adrenal tumors, but evenhigher elevations were found in severelyill psychiatric patients. The finding wasinitially a curiosity, but systematic studiesby Carroll and coworkers9 inMelbourne in the 1970s showed thatserum cortisol levels were elevated inpatients with International Classificationof Diseases-defined melancholicdepression. Patients lacked the normaldiurnal rhythmicity of cortisol, and itslevels were not suppressed with theadministration of the corticosteroiddexamethasone. These findings wereformalized in the dexamethasonesuppression test (DST).9

In patients referred for ECT, an abnormaltest result was quickly recognizedas a measure of severity of illness. Thetest results normalized with remissionand became abnormal again with relapse.Despite its credibility as a monitor ofclinical status, the DST was rejected bypsychiatric leaders in the mid-1980s onthe grounds that it was not a valid measureof DSM-III diagnostic criteria.10,11

Although they are less well established,sleep electroencephalographyand thyroid function tests are additionallaboratory tools for measuring thesyndrome's severity.1

Relevance for research assessment

Researchers already acknowledgedifferent response rates for psychotic andnonpsychotic depressed patients andexclude those with psychotic featuresfrom clinical trials. Drug dependencecomplicates compliance, and concomitantuse of medications for medicaldisorders risks mixing the systemic effects of medications with those ofpsychotropic agents. Patients with drugdependence and those who are takingdrugs for medical illness are usuallyexcluded from experimental trials (leavingclinicians who have to treat such populations with unstudied risks and anabsence of guidelines). Stratifying populationsin depression research studiesby the presence of melancholia willensure a more homogeneous populationfor evaluating treatment efficacy.

Relevance for clinical practice1

Few studies specifically assess the efficacyof modern treatments for patientswith melancholia, so we are limited tointerpreting experiences before theintroduction of DSM-III. When convulsivetherapy was introduced in the1930s, its ease of use encouraged widespreadtesting. Within a few years, itsefficacy in patients with manic-depressiveillness, endogenous depression,and high suicide risk was recognized.The introduction of chlorpromazine andimipramine followed, with both treatmentsbeing found effective for thesepatients. Lithium, anxiolytics, anticon anticonvulsants,and the spate of SSRI antidepressantshave been poorly assessed inmelancholic patients, and we are thereforeinsecure in their use.1

A useful algorithm starts with consideringthe possibility of melancholiaand establishing a symptom baselinewith a depression rating scale. A formalDST or a 4 PM cortisol level marks theseverity of the illness, and repeatedtests guide decisions as to remission andrelapse. Treatment with a tricyclic antidepressanthas better evidence for efficacythan treatment with more recently introduced medications1 the formaltreatment algorithm for melancholiacalls for dosages equivalent to 200 to300 mg/d of imipramine, established asrapidly as can be tolerated.

In severely suicidal patients whorequire continuing observation or whoare in such a state of inanition that theircondition is considered to be life-threatening,ECT is the preferred treatment.It is also the treatment of choice whenmedications fail. Periodic monitoringof cortisol levels and use of ratingscales offer clinicians a guide to thecourse of treatment and the scheduleof continuation treatments.


DSM-III and DSM-IV diagnostic criteriaare too imprecise to define homogeneouspopulations in treatmentassessments. Discarding serum cortisolmeasurements served our patients andour science poorly. A parallel exampleis the DSM's consideration of catatonia,a syndrome that is often seen inpatients with mood disorders or as atoxic response to medications. In theDSM, catatonia is defined only as asubtype of schizophrenia, a narrow viewthat has handicapped the recognition ofits many forms and inhibited the developmentof effective therapeutics. Thisdeficiency has recently been redressedin 3 useful texts that offer effective guidelinesto diagnosis and treatment.12-14

An appreciation of melancholia asthe principal definable mood disorderoffers a better guide to diagnosis andtreatment than does DSM-identified“major depression.” It offers morehomogeneous populations of depressivepatients, avoiding the difficulties oftreatment assessments in mixed populations,and ensuring better outcomesfor more patients even with the limitedmeans that are now at hand.

Dr Fink is professor of psychiatry and neurologyat the State University of New York atStony Brook. He is the author of Electroshock:Restoring the Mind (Oxford University Press),founding editor of The Journal of ECT, and coauthorof Catatonia: A Clinician's Guide toDiagnosis and Treatment and Melancholia: TheDiagnosis, Pathophysiology, and Treatment ofDepressive Illness (both from CambridgeUniversity Press).



1. Taylor MA, Fink M. Melancholia: The Diagnosis,Pathophysiology and Treatment of DepressiveIllness. Cambridge, England: Cambridge UniversityPress; 2006.
2. Rush AJ, Trivedi MH, Wisniewski SR, et al; STAR*DStudy Team. Bupropion-SR, sertraline, or venlafaxine-XR after failure of SSRIs for depression. N EnglJ Med. 2006;354:1231-1242.
3. Khan A, Kolts RL, Rapaport MH, et al. Magnitudeof placebo response and drug-placebo differencesacross psychiatric disorders. Psychol Med.2005;35:743-749.
4. Glassman AH, Kantor SJ, Shostak M. Depression,delusions and drug response. Am J Psychiatry.
5. Crismon ML, Trivedi M, Pigott TA, et al. The TexasMedication Algorithm Project: report of the TexasConsensus Conference Panel on MedicationTreatment of Major Depressive Disorder. J ClinPsychiatry. 1999;60:142-156.
6. Parker G, Hadzi-Pavlovic D. Melancholia: A Disorderof Movement and Mood. Cambridge, England:Cambridge University Press; 1996.
7. Fink M. A new appreciation of ECT. Psychiatr Times.2004;21(4):13-14.
8. Carroll BJ. The hypothalamus-pituitary-adrenal axisin depression. In: Burrows GD, ed. Handbook ofStudies on Depression. Amsterdam: ExerptaMedica;1977:325-342.
9. Carroll BJ, Feinberg M, Greden JF, et al. A specificlaboratory test for the diagnosis of melancholia.Standardization, validation, and clinical utility. ArchGen Psychiatry. 1981;38:15-22.
10. Fink M. Should the dexamethasone suppressiontest be resurrected? Acta Psychiatr Scand.2005;112:245-249.
11. Fink M. The DST riddle. Psychiatr Times.2006;23:25-26.
12. Fink M, Taylor MA. Catatonia: A Clinician’s Guideto Diagnosis and Treatment. Cambridge, England:Cambridge University Press; 2003.
13. Caroff SN, Mann SC, Francis A, Fricchione GL,eds. Catatonia. From Psychopathology toNeurobiology.Washington, DC: American PsychiatricPublishing; 2004.
14. Dhossche D, Wing L, Ohta M, Neumärker K-J,eds. Catatonia in Autism Spectrum Disorders.Amsterdam: Elsevier; 2006.1975;132:716-719.