early- and late-onset depression in the older: no differences found within the melancholic subtype

Early- and late-onset depression in the older: no differencesfound within the melancholic subtype

Pilar Alvarez1, Mikel Urretavizcaya2, Luisa Benlloch3, Julio Vallejo2 and Jose Manuel Menchon2

1Department of Psychiatry, Centre Forum, Parc de Salut Mar, Barcelona, Spain2Department of Psychiatry, Hospital de Bellvitge, Barcelona, Spain3Department of Psychiatry, Consorci Sanitari de Terrassa, Barcelona, SpainCorrespondence to: P. Alvarez, E-mail: [email protected]

Objective: Several studies have reported clinical and biological differences between early- and late-onset(EO and LO) depression, which suggest different underlying aetiological processes. The aim of thepresent study is to examine whether there are differences between EO and LO depressed patients withmelancholy, controlling for current age, with regard to clinical variables, vascular risk factors and familyhistory of affective disorders or suicide.

Methods: One hundred and twenty-one melancholic patients were divided into three groups: patientswith current age and onset earlier than 60 (N¼ 60), patients aged 60 or over and with onset at 60 or later(N¼ 30) and patients aged 60 or over and with onset before the age of 60 (N¼ 31). Systematic clinicaldata were collected with the structured interview ’The Schedule for Affective Disorders and Schizo-phrenia’. Symptom ratings at admission and at discharge were assessed by means of the 21-itemHamilton Depression Rating Scale, the Hamilton Anxiety Scale and the Widlocher DepressionRetardation Scale. Family history of affective disorders or suicide was obtained using the Family HistoryResearch Diagnostic Criteria. Vascular risk factors were also recorded.

Results:The only symptoms that differed across the groups were feelings of anger and irritability, whichscored lower in the LO older group. No other significant differences were found in the variables studied.

Conclusion: According to this study, LO depression with melancholia should not be considered as adistinct entity. Further studies on EO and LO-depression should consider this diagnostic subtype, amongothers, as a key variable. Copyright # 2010 John Wiley & Sons, Ltd.

Key words: depression; melancholia; onset; geriatrics; risk factorsHistory: Received 11 January 2010; Accepted 12 May 2010; Published online 29 December 2010 in Wiley Online Library(wileyonlinelibrary.com).DOI: 10.1002/gps.2571

Introduction

Several studies have reported clinical and biologicaldifferences between early- and late-onset (EO and LO)depression (Alexopoulos et al., 1988; Caine et al.,1994) which suggest different underlying aetiologicalprocesses (Brodaty et al., 2001). For instance, positivefamily history of affective disorders, greater personalitydysfunction and substance abuse have been relatedwith EO depression (Charney et al., 1981; Mendlewiczand Baron, 1981; Brown et al., 1984; Burvill et al.,1989; Conwell et al., 1989; Kupfer et al., 1989; Musettiet al., 1989; Brodaty et al., 1991, 2001; Abrams et al.,

1994; Baldwin and Tomenson, 1995; Klein et al., 1999;Devanand et al., 2004), while systemic and neurode-generative diseases and personal losses have been morefrequently considered as contributing factors to LOdepression (Alexopoulos et al., 1988; Caine et al.,1994).Results from some clinical studies suggest that LO

depressed patients present a higher degree of severity(Kessing, 2006), are more likely to show a melancholicsubtype (Burvill et al., 1989) and present morehypochondriasis (Brown et al., 1984; Brodaty et al.,1991), psychomotor disturbances (Avery and Silver-man, 1984), psychotic features (Meyers et al., 1984;

RESEARCH ARTICLE

Copyright # 2010 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2011; 26: 615–621.

Meyers and Greenberg, 1986; Kessing, 2006), suicidalbehaviour (Lyness et al., 1999b), intellectual impair-ment (Alexoupoulos et al., 1991; Dillon et al., 2009),generalized anxiety (Brown et al., 1984; Meyers andGreenberg, 1986) and precipitating stressors (Musettiet al., 1989; Alexoupoulos et al., 1991). On the otherhand, feelings of guilt and other depressive thoughtsand personality disorders seem less common in olderthan in adult depressed patients (Brown et al., 1984;Conwell et al., 1989; Fava et al., 1996; Heun et al.,2000; Brodaty et al., 2001).Biological markers suggested for general depression

have also been examined in geriatric patients. The mostsignificant findings in biological research come fromneuroimaging studies. Evidence of more lateralventricular enlargement, white matter hyperintensitiesand basal ganglia lesions have been reported in LOdepression, suggesting the importance of structuralchanges in its aetiology (Coffey et al., 1990; Lesseret al., 1996; Salloway et al., 1996; Hickie et al., 1997;Krishnan et al., 1997; Dahabra et al., 1998). Also, inregards to neuropsychology, one recent study found aworse neuropsychological performance in LO depressivepatients in comparison with EO patients (Thomas et al.,2009). These findings may be due to cerebrovasculardisease (Drevets, 1998; Sackheim et al., 2000), in whichthe increase of vascular risk factors may be implicated(Awad et al., 1986; Coffey et al., 1990; Breteler et al.,1994; Brodaty and Luscombe, 1996).However, other studies have not replicated these

results and have found no differences between EOand LO depression (Caine et al., 1994). Similaritiesbetween the two age groups have been reported inglobal severity of depression (Greenwald and Kramer-Ginsberg, 1988; Conwell et al., 1989; Baldwin, 1990;Brodaty et al., 2001), somatization (Musetti et al.,1989), presence of a melancholic subtype (Brodatyet al., 2001; Kessing, 2006), psychotic symptoms(Nelson and Bowers, 1978; Charney and Nelson, 1981;Glassman and Roose, 1981; Nelson et al., 1989),poor neuropsychological performance (Brodaty et al.,2001), cardiovascular risk factors/vascular pathology(Lyness et al., 1999a; Janssen et al., 2006), concomitantmedical illness (Greenwald and Kramer-Ginsberg,1988; Burvill et al., 1989; Conwell et al., 1989),personality factors (Grace and O’Brien, 2003) psycho-social stressors (Burvill et al., 1989; Baldwin, 1990;Janssen et al., 2006) and family history of affectivedisorders (Greenwald and Kramer-Ginsberg, 1988;Burvill et al., 1989; Baldwin, 1990; Janssen et al., 2006).Even so, in a recent study, contrary to expectations,higher levels of total cholesterol were more frequentlyobserved in EO depressed patients (Dillon et al., 2009).

These controversial results may be due to theheterogeneity of the depressed samples studied andthe use of different ages to divide EO and LO patients.The concept of major depression may encompass anumber of different depressive conditions, in particu-lar melancholia (Hickie, 1996) which, some authorshave suggested, may constitute a more homogeneousclinical entity (Parker et al., 1996; Fink and Taylor,2007; Shorter, 2007).In this study we try to examine whether some

differential factors between EO and LO major depress-ive disorder—namely clinical data, family historyof affective disorders or suicide or vascular riskfactors—could also be found in a highly homogeneoussample—melancholic depressed patients. As mentaldisorders are defined in DSM-IV as politethic concepts,different combinations of the listed symptoms couldlead to the same diagnosis. Regarding melancholia, thediagnostic criteria involve 8 symptoms and a thresholdof only four symptoms for the diagnosis. As a result, twodifferent diagnosed cases could meet the criteria ofmelancholiawithouthaving any symptoms incommon.In summary, even in a homogeneous sample asmelancholia, clinical variability could not be excluded.In order to differentiate between the effects of the

ageing process and the age of the onset of depression,we compared three groups: adult depressed patients,older depressed patients with EO and older depressedpatients with LO. Based on previous literature reports,the cutoff point selected for adult versus older patientsand for EO versus LO was 60 years.

Methods

Subjects

The sample comprised 121 consecutive in-patientsadmitted to a university general hospital, all ofwhom met DSM-IV criteria for major depressionwith melancholia with or without psychotic features.Patients with another axis I diagnosis, neurological orserious systemic diseases were excluded. The study wasdescribed to the subjects in detail and written informedconsent was obtained in 121 of the 130 candidates forinclusion.According to the current age and the age of onset of

depression the sample was divided into three groups:

1. Adult EO patients (EO-adult): current age between18 and 60 years (n¼ 60).

2. Older EO patients (EO-older) patients: current age60 or over, with onset before 60 (n¼ 31).


616 P. Alvarez et al.

3. Older LO patients (LO-older): current age 60 orover with onset after 60 (n¼ 30).

Procedure

Diagnoses were made by agreement between threedifferent psychiatrists. None of the patients showedphysical, chemical or historical evidence of any neuro-logical or other major illnesses. Systematic clinicaldata were collected with the structured interview ‘TheSchedule for Affective Disorders and Schizophrenia’(SADS) (Endicott and Spitzer, 1978).Symptom ratings at admission and at discharge

were assessed by means of the following ratingscales: the 21-item Hamilton Depression Rating Scale(HDRS) (Hamilton, 1960), the Hamilton AnxietyScale (HAS) (Hamilton, 1959) and the WidlocherDepression Retardation Scale (WDRS) (Widlocher,1980).Family history of affective disorders or suicide was

obtained from the patient and close relatives or keyinformants through the FH-RDC (Andreasen et al.,1977). Two cases in which no reliable psychiatric datacould be obtained about relatives were recorded asmissing and were excluded from this analysis.Every vascular risk factor (hypertension, dyslipide-

mias, diabetes mellitus, smoking, obesity and cardi-ovascular diseases) was recorded and rated as follows:

� 0, absent.� 1, one mild risk factor:

– hypertension or dyslipidemias or diabetes mellitusor cardiovascular diseases well controlled andwithout organic complications;

– light or moderate smoking (01–20 cigarettes/day);– overweight (body mass index, BMI¼ 25.0–29.9).

� 2, two or more mild risk factors or one severe riskfactor:

– hypertension or dyslipidemias or diabetes mellitusor cardiovascular diseases not well controlled orwith organic complications;

– heavy smoking (21 or more cigarettes/day);– obesity (BMI> or¼ 30).

Statistical analysis

Differences between the three groups were analyzedusing an analysis of variance (least square differencesprocedure for pair comparisons). When differencescould have been due to the length of illness, an analysisof covariance was carried out with this variable as the

covariant. Categorical variables were analyzed usingthe chi square test. The level of significance wasestablished at p< 0.05. All analyses are two-tailed.

Results

Demographic and diagnostic characteristics

Current age, age at onset and sex distribution for eachgroup are shown in Table 1. There were no differencesbetween groups in sex distribution. As expected, theEO-adult group was significantly younger than theother two groups (F¼ 97.72, d.f.¼ 2, 118, p< 0.001)and the age at onset of the LO-older group wassignificantly higher than in the other two groups(F¼ 64.20, d.f.¼ 2, 118, p< 0.001).Psychotic melancholic depression was diagnosed in

26 (43%) patients in the EO-adult group, in 21 (68%)in the EO-older group and in 17 (57%) in the LO-oldergroup. Although no significant differences were found,there was a trend towards more delusional patientsin the EO-older group (x2¼ 5.12, d.f.¼ 2, p¼ 0.07).When each pair of groups was compared, the onlysignificant difference (x2¼ 3.95, d.f.¼ 2, p¼ 0.047,continuity correction) was found between the EO-adult and the EO-older groups.

Course of depression

The EO-older group reported significantly moreprevious episodes (F¼ 7.46, d.f.¼ 2, 118, p< 0.001)and previous psychiatric admissions (F¼ 5.23, d.f.¼ 2,118, p¼ 0.007), as well as a longer time of evolution

Table 1 Sample characteristics

EO-adult(N¼60)

EO-older(N¼ 31)

LO-older(N¼30)

Gendera N (%) N (%) N (%)Male 16 (26.7) 12 (38.7) 9 (30.0)Female 44 (73.3) 19 (61.3) 21 (70.0)Current ageb (y.) Mean�SD Mean�SD Mean�SD

51.4�7.3 65.9�5.0 68.8�5.3Age at onsetc (y.) Mean�SD Mean�SD Mean�SD

45.1�10.1 44.9� 10.1 66.9�5.3

EO-adult: current adult patients, early-onset.

EO-older: current older patients, early-onset.

LO-older: current older patients, late-onset.aNo significant differences between the three groups.bEO-adult was significantly younger (p< 0.001) than the other two groups.cLO-older was significantly older (p< 0.001) than the other two groups.


Early- and late-onset melancholic depression 617

(F¼ 41.64, d.f.¼ 2, 118, p< 0.001) (Table 2). How-ever, when an analysis of covariance with time of lengthof the disease as a covariant was carried out, thesignificant differences disappeared.

Clinical symptomatology

There were no significant differences between groupsin the total scores on the HDRS (F¼ 0.97, d.f.¼ 2, 118,p¼ 0.384), HAS (F¼ 1.03, d.f.¼ 2, 118, p¼ 0.358)or WDRS (F¼ 0.98, d.f.¼ 2, 118, p¼ 0.378) scales(Table 2). Nor were significant differences found wheneach specific item on HDRS and on WDRS wascompared across the three groups.Specific mood symptoms recorded on SADS were

also compared across the groups. The only symptomsthat differed across the groups were feelings ofanger (F¼ 3.61, d.f.¼ 2, 118, p¼ 0.030) and irrit-ability (F¼ 5.83, d.f.¼ 2, 118, p¼ 0.004), which scoredlowest in the LO-older group (Table 2). No othersymptoms presented significant differences.

Family history and vascular risk factors

The rates of positive family history for affectivedisorder or suicide in every group were: 23 (38%)patients in the EO-adult group, 11 (37%) in the EO-older group and 14 (48%) in the LO-older group. No

significant differences were found between the threegroups (x2¼ 0.89, d.f.¼ 2, p¼ 0.64).Vascular risk factors were present in 32 (53.2%)

patients in the EO-adult group, in 18 (58.1%) inthe EO-older and in 12 (40%) in the LO-older group.No significant differences were found between thethree groups—not even between the mild risk groupand the no risk group or the severe risk group.

Discussion

The main strength of this study is the recruitmentto date, as far as we know, of the largest number ofmelancholic patients with EO and LO depression.Several study limitations must be acknowledged.

First, from an epidemiological point of view, the scarcenumber of patients included in this study and, ingeneral, in studies developed in clinical settings, couldmake it difficult to find subtle differences betweengroups. Secondly, although biological research benefitsfrom homogeneous samples, its use limits thegeneralization of results. Hence, our findings maynot be extrapolated to other populations: non-melancholics, less severe melancholics, melancholicoutpatients, etc. Lastly, it must also be noted that arelatively high current median age of the EO-adultgroup could minimize the differences between groups.Also, this factor makes it difficult to compare theresults with earlier studies using an age 50 cut off, in

Table 2 Clinical characteristics

EO-adult EO-older LO-older pN¼ 60 N¼31 N¼30

Mean�SD Mean�SD Mean�SD

Gl global scalesHDRS at admission 33.0� 7.1 32.8�7.8 30.9�5.7 0.384HAS at admission 28.0� 8.0 26.0�8.1 25.9�7.2 0.358WDRS at admission 28.3� 13.4 32.2�12.0 29.1�12.0 0.378

SymptomsFeelings of angera 2.3� 1.4 2.1�1.6 1.5�0.8 0.030b

Irritabilitya 2.1� 1.1 2.1�1.5 1.3�0.5 0.004c

CoursePrevious episodes (nu.) 1.4� 2.2 2.4�2.1 0.5�0.8 <0.001d

Previous admissions (nu.) 0.8� 1.4 1.3�2.1 0.1�0.3 0.007e

Length of illness (y.) 6.3� 8.3 21.0�12.5 1.9�3.4 <0.001d



LO-older: current older patients, late-onset.aAccording to SADS.bSignificant differences were found between LO-older and EO-adult (oneway ANOVA).cSignificant differences were found between LO-older and the other two groups (oneway ANOVA).dSignificant differences were found among the three groups (oneway ANOVA).eSignificant differences were found between EO-older and LO-older (oneway ANOVA).



which some patients might have been included in theolder group.The main result of this study is the absence of

differences in clinical symptoms, family history andvascular risk-factors between EO and LO depression indepressed melancholic patients.Hypochondriasis, psychomotor disturbances, anxiety

and other symptoms which have been more frequentlyreported in LO-older patients in some studies (Averyand Silverman, 1984; Brown et al., 1984; Meyers et al.,1984; Meyers and Greenberg, 1986; Burvill et al., 1989;Alexoupoulos et al., 1991; Brodaty et al., 1991; Kleinet al., 1999; Lyness et al., 1999b), were distributedequally across groups. The only differences were inirritability and feelings of anger on the SADS.Significant differences were found for delusions

between EO-adult and EO-older groups, which suggestthat they could be more associated to the length of theillness than to age of onset. However, age of onsetand current older age may also have some influence ondelusionality since no differences were found betweenEO-older (who have a longer time course of thedisorder) and LO-older. Some studies have found arelationship between later onset and the presence ofdelusions, but those samples included non-melan-cholic depressions (Meyers and Greenberg, 1986;Kessing, 2006). However, another study by one ofthese groups (Meyers et al., 1984), in which the samplecomprised only primary endogenous depressives,found that the EO group who were delusional duringthe index episode tended to have a higher mean ageat this time, a finding that may be related to thelonger course of illness. Finally, another point to beconsidered in the relationship between delusions andage of onset is whether delusional depression is adifferent subtype of depression (Schatzberg andRothschild, 1992; Parker et al., 2000), which, in turn,may have a slightly different mean age of onset buta wide variability. This may, perhaps, account for thedifferent findings reported by studies.

Although EO-older patients showed more previousepisodes and admissions, they differed not only fromLO-older but from EO-adult patients as well. While theage of onset was not related to this finding, the lengthof illness may have accounted for this result since thedifferences did not persist when adjustment for thisvariable was made in the analysis of covariance. Thisdoes not mean that with the increase in the length ofillness there is a greater risk of suicide attempts orhospitalization, merely that in a longer period theseevents will occur more often.To our knowledge, only two similar clinical studies

with such a homogeneous sample have been carriedout in recent years. Brown et al. (1984) found highersomatization and hypochondriasis and less loss oflibido, guilt and suicide attempts in LO melancholicpatients. However, the study was retrospective and thecut-off point was 50, not 60. Those authors suggest thatthere is insufficient evidence to view LO melancholiaas a separate clinical entity. Our results support thefindings of the study by Blazer et al. (1987) whoreported clinical similarity between middle-aged andolder admitted melancholic depressed patients.Like clinical variables, family history of affective

disorder was not able to differentiate between EOand LO melancholic depression. As expected, a highpercentage of patients had a family history of affectivedisorders or suicide.A higher percentage of vascular risk factors was

found in the EO-older group, also the group with ahigher percentage of delusional depression. This resultis in agreement with a seminal paper suggesting anassociation between delusional depression and vascularrisk factors (O’Brien et al., 1997).The absence of significant differences in vascular risk

factors between EO and LO depression corroboratesthe findings of Lyness et al. (1999a) in primary-caredepressive patients and those of Devanand et al. (2004)in a Late Life Depression Clinic. The lack of associationbetween a family history of affective disorders orsuicide and EO onset depression is in accordance withthe study by Brodaty et al. (2001) but not with theresults of two other recent studies (Devanand et al.,2004; Gallagher et al., 2009). The significantly largerproportion of melancholic patients in the EO group inthe Devanand study and the recruitment of com-munity dwelling older adults in the Gallagher study,could account for this discrepancy.Since melancholic depression may be more frequent

in older patients (Burvill et al., 1989) and sincesymptoms described as characteristic of LO depressionare prominent symptoms in melancholia, the differ-ences reported in some previous studies may be due in

Table 3 Vascular risk factors

Vascularrisk factors

EO-adult(N¼60)

EO-older(N¼ 31)

LO-older(N¼30)

N (%) N (%) N (%)None 28 (46.6) 13 (41.9) 18 (60.0)Mild 16 (26.6) 7 (22.6) 6 (20.0)Severe 16 (26.6) 11 (35.5) 6 (20.0)



LO-older: current older patients, late-onset.

No significant differences were found between the groups.



part to a greater percentage of melancholic depressedpatients in older groups. The inconsistencies in theresults in several studies may also be due, at leastpartially, to the probable heterogeneity of the groupsof depressed patients since the diagnosis of majordepression mixes both melancholic and non-melan-cholic cases.

Conclusion

According to the variables studied, melancholicdepression is similar regardless the age of onset. So,our results suggest that LOdepressionwithmelancholiashould not be considered as a distinct entity. Furtherstudies on early- and late-depression should considerthis diagnostic subtype, among others, as a key variable.

Conflict of interest

None declared.

REFERENCES

Abrams RC, Rosendahl E, Card CH, Alexopoulos S. 1994. Personality disordercorrelates of late and early onset depression. J Am Geriatr Soc 42: 727–731.

Alexopoulos GS, Young RC, Meyers VS, Abrams RC, Shmoian CA. 1988. Late-onsetdepression. Psychiatr Clin North Am 11: 101–115.

Alexoupoulos GS, Meyers BS, Young RC. 1991. Disability and environment in late-onset depression (NR643). In New Research Program and Abstracts: AmericanPsychiatric Association 144th Annual Meeting, APA: Washington, DC; 202.

Andreasen NC, Endicott J, Spitzer RL, Winokur G. 1977. The family history methodusing diagnostic criteria. Arch Gen Psychiatry 34: 1229–1235.

Avery D, Silverman J. 1984. Psychomotor retardation and agitation in depression:relationship to age, sex and response to treatment. J Affect Disord 7: 67–76.

Awad IA, Spetzler RF, Hodak JA, Awad CA, Carey R. 1986. Incidental subcorticallesions identified on magnetic resonance imaging in the elderly. I. Correlation withage and cerebrovascular risk factors. Stroke 17: 1084–1089.

Baldwin B. 1990. Age of onset of depression in the elderly. Br J Psychiatry 156: 445–446.Baldwin RC, Tomenson B. 1995. Depression in later life: a comparison of symptoms

and risk factors in early and late onset cases. Br J Psychiatry 167: 649–652.Blazer D, Bachar JR, Hughes DC. 1987. Major depression with melancholia: a comparison

of middle-aged and elderly adults. J Am Geriatr Soc 35: 927–932.

Breteler MM, van Swieten JC, Bots ML, et al., 1994. Cerebral white matter lesions,vascular risk factors, and cognitive function in a population-based study: theRotterdam Study. Neurology 44: 1246–1252.

Brodaty H, Luscombe G. 1996. Depression in persons with dementia. Int Psychogeriatr8: 609–622.

Brodaty H, Peters K, Boyce P, et al., 1991. Age and depression. J Affect Disord 23: 137–149.

Brodaty H, Luscombe G, Parker G, et al., 2001. Early and late onset depression in oldage: different aetiologies, same phenomenology. J Affect Disord 66: 225–236.

Brown RP, Sweeney J, Loutsch E, Kocsis J, Frances A. 1984. Involutional melancholiarevisited. Am J Psychiatry 141: 24–28.

Burvill PW, Hall WD, Stampfer HG, Emmerson JP. 1989. A comparison of early onsetand late-onset depressive illness in the elderly. Br J Psychiatry 155: 673–679.

Caine ED, Lyness JM, King DA. 1994. Clinical and etiological heterogeneity of mooddisorders in elderly patients. In Diagnosis and Treatment of Depression in Late Life,Schneider L et al. (eds.). American Psychiatric Press: Washington; 3–53.

Charney DS, Nelson JC. 1981. Delusional and nondelusional unipolar depression:further evidence for distinc subtypes. Am J Psychiatry 138: 328–333.

Charney DS, Nelson JC, Quinlan DM. 1981. Personality traits and disorder indepression. Am J Psychiatry 138: 1601–1604.

Coffey CE, Figiel GS, Djang WT, Weiner RD. 1990. Subcortical hyperintensity onmagnetic resonance imaging: a comparison of normal and depressed elderlysubjects. Am J Psychiatry 147: 187–189.

Conwell Y, Nelson C, Kim KM, Mazure CM. 1989. Depression in late life: age of onsetas marker of a subtype. J Affect Disord 17: 189–195.

Dahabra S, Ashton CH, Bahrainian M, et al., 1998. Structural and functionalabnormalities in elderly patients clinically recovered from early- and late-onsetdepression. Biol Psychiatry 44: 34–46.

Devanand DP, Adorno E, Cheng J, et al., 2004. Late onset dysthymic disorder andmajor depression differ from early onset dysthymic disorder and major depressionin elderly outpatients. J Affect Disord 78: 259–267.

Dillon C, Allegri RF, Serrano CM, et al., 2009. Late- versus early-onset geriatricdepression in a memory reseach center. Neuropsychiatr Dis Treat 5: 517–526.

Drevets WC. 1998. Functional neuroimaging studies of depression: the anatomy ofmelancholia. Annu Rev Med 49: 341–361.

Endicott J, Spitzer RL. 1978. A diagnostic interview: the schedule for affective disordersand schizophrenia. Arch Gen Psychiatry 35: 773–779.

Fava M, Alpert JE, Borus JS, et al., 1996. Patterns of personality disorder comorbidityin early-onset versus late-onset major depression. Am J Psychiatry 153: 1308–1312.

Fink M, Taylor MA. 2007. Resurrecting melancholia. Acta Psychiatr Scand 115(suppl443): 14–20.

Gallagher D,Mhaolain AN, Greene E, et al., 2009. Late life depression: a comparison ofrisk factors and symptoms according to age of onset in community dwelling olderadults. Int J Geriatr Psychiatry. Dec 8 [Epub ahead of print].

Glassman AB, Roose SP. 1981. Delusional depression: a distinct clinical entity? ArchGen Psychiatry 38: 424–427.

Grace J, O’Brien JT. 2003. Association of life events and psychosocial factors with earlybut not late onset depression in the elderly: implications for possible differences inaetiology. Int J Geriatr Psychiatry 18: 473–478.

Greenwald BS, Kramer-Ginsberg E. 1988. Age at onset of geriatric depression:relationship to clinical variables. J Affect Disord 15: 61–68.

Hamilton M. 1959. The assessment of anxiety states by rating. Br J Med Psychol 32:50–55.

Hamilton M. 1960. A rating scale for depression. J Neurol Neurosurg Psychiatry 23:56–62.

Heun R, Kockler M, Papassotiropoulos A. 2000. Distinction of early- and late-onsetdepression in the elderly by their lifetime symptomatology. Int J Geriatr Psychiatry15: 1138–1142.

Hickie I. 1996. Issues in classification: III. Utilising behavioural constructs in mel-ancholia research. In Melancholia: A Disorder of Movement and Mood, Parker Get al. (eds.). Cambridge University Press: Cambridge; 38–56.

Hickie I, Scott E, Wilhelm K, Brodaty H. 1997. Subcortical hyperintensities onmagnetic resonance imaging in patients with severe depression—a longitudinalevaluation. Biol Psychiatry 42: 367–374.

Janssen J, Beekman AT, Comijs HC, Deeg DJ, Heeren TJ. 2006. Late-life depression:the differences between early-and late-onset illness in a community-based sample.Int J Geriatr Psychiatry 21: 86–93.

Kessing LV. 2006. Differences in diagnostic subtypes among patients with late andearly onset of a single depressive episode. Int J Geriatr Psychiatry 21: 1127–1131.

Klein DN, Schatzberg AF, McCullogh JP, et al., 1999. Age of onset in chronic majordepression: relation to demographic and clinical variables, family history, andtreatment response. J Affect Disord 55: 149–157.

Krishnan KR, Hays JC, Blazer DG. 1997. MRI-defined vascular depression. Am JPsychiatry 154: 497–501.

Kupfer KJ, Frank E, Carpenter LL, Neiswanger K. 1989. Family history in recurrentdepression. J Affect Disord 17: 113–119.

Lesser IM, Boone KB, Mehringer CM, et al., 1996. Cognition and white matterhyperintensities in older depressed patients. Am J Psychiatry 153: 1280–1287.

Key Points

� Clinical and biological differences between early-and late-onset (EO and LO) depression have beenreported.

� EO and LO hospitalized melancholic depressedpatients do not differ with respect to clinicalsymptoms, family history and vascular risk-factors.

� LO melancholic depression should not beconsidered as a distinct entity.



Lyness JM, Caine ED, King DA, et al., 1999. Cerebrovascular risk factors anddepression in older primary care patients. Testing a vascular brain disease modelof depression. Am J Geriatr Psychiatry 7: 252–258.

Lyness JM, Conwell Y, Nelson JC. 1999. Suicide attempts in elderly psychiatricinpatients. J Am Geriatr Soc 40: 320–324.

Mendlewicz J, Baron M. 1981. Morbidity risks in subtypes of unipolar depressiveillness: differences between early and late onset forms. Br J Psychiatry 139: 463–466.

Meyers BS, Greenberg R. 1986. Late-life delusional depression. J Affect Disord 11: 133–137.

Meyers BS, Kalayam B, Mei-Tal V. 1984. Late-onset delusional depression: a distinctclinical entity? J Clin Psychiatry 45: 347–349.

Musetti L, Perugi G, Soriani A, et al., 1989. Depression before and after age 65: a re-examination. Br J Psychiatry 155: 330–336.

Nelson JC, Bowers MB. 1978. Delusional unipolar depression: description and drugresponse. Arch Gen Psychiatry 35: 1321–1328.

Nelson JC, Conwell Y, Kim K, Mazure C. 1989. Age at onset in late-life delusionaldepression. Am J Psychiatry 146: 785–786.

O’Brien JT, Ames D, Schweitzer I, et al., 1997. Clinical, magnetic resonance imagingand endocrinological differences between delusional and non-delusional depres-sion in the elderly. Int J Geriatr Psychiatry 12: 211–218.

Parker G, Hadzi-Pavlovic D, Boyce P. 1996. Issues in classification: II. Classifyingmelancholia. In Melancholia: A Disorder of Movement and Mood, Parker GG et al.(eds.). Cambridge University Press: Cambridge; 20–37.

ParkerG, Roy K,Hadzi-Pavlovic D, et al., 2000. Subtyping depression by clinical features:the Australasian database. Acta Psychiatr Scand 101: 21–28.

Sackheim HA, Lisanby SH, Nobler MS. 2000. MRI hyperintensities and the vascularorigins of late life depression. In Progress in Psychiatry, Andrade C (ed.). OxfordUniversity Press: New York; 73–116.

Salloway S, Malloy P, Kohn R, et al., 1996. MRI and neuropsychological differences inearly- and late-life-onset geriatric depression. Neurology 46: 1567–1574.

Schatzberg AF, Rothschild AJ. 1992. Psychotic (delusional) major depression: shouldit be included as a distinct syndrome in DSM-IV? Am J Psychiatry 149: 733–745.

Shorter E. 2007. The doctrine of the two depressions in historical perspective. ActaPsychiatr Scand 115(suppl 443): 5–13.

Thomas AJ, Gallagher P, Robinson LJ, et al., 2009. A comparison of neurocognitiveimpairment in younger and older adults with major depression. Psychol Med 39:725–733.

Widlocher D. 1980. L’evaluation quantitative du ralentissement psychomoteur dansles etats depressifs. Psychol Med 12: 2725–2739.



early- and late-onset depression in the older: no differences found within the melancholic subtype

Documents