dysthymia among the community-dwelling elderly

DYSTHYMIA AMONG THE COMMUNITY-DWELLING ELDERLY

MICHAEL KIRBY1*, IRENE BRUCE2, DAVIS COAKLEY3 AND BRIAN A. LAWLOR4

1Research Registrar, Mercer's Institute for Research on Ageing and Dept of Psychiatry for the Elderly,St James's Hospital, Dublin, Ireland

2Research Nurse, Mercer's Institute for Research on Ageing, St James's Hospital, Dublin, Ireland3Professor in Geriatric Medicine, Department of Medicine for the Elderly and Mercer's Institute for Research

on Ageing, St James's Hospital, Dublin, Ireland4Professor in Psychiatry for the Elderly, Department of Psychiatry for the Elderly and Mercer's Institute for Research

on Ageing, St James's Hospital, Dublin, Ireland

SUMMARY

There are few data on the clinical features of dysthymia among the community-dwelling elderly. Forty elderlyindividuals with dysthymic disorder were identi®ed following screening in the community with GMS±AGECAT. Adetailed clinical history was obtained and DSM-IV checklists and standardized scales were used, at a secondinterview. Comparisons were made with a group of 630 non-depressed elderly from the same community. Dysthymiawas predominately of late onset (93%) and associated with a major stressor in 65% of cases. Comorbid axis 1disorders were present in 15% of dysthymics and an axis 2 disorder in 10%. The dysthymic group had signi®cantlyhigher degrees of physical impairment than the non-depressed elderly. The symptom pro®le demonstrated prominentanxiety and functional features. Eighty-three per cent of the elderly with dysthymia had presented to their GP withanxiety/depressive symptoms at some stage during the dysthymic disorder. The presentation of dysthymia in olderpeople di�ers from that in earlier life. Late life dysthymia is less associated with axes 1 and 2 comorbidity but isassociated with signi®cant degrees of physical impairment. Dysthymia in older people presents to primary care, ratherthan specialist services, and interventions must be delivered at this level. Copyright # 1999 John Wiley & Sons, Ltd.

KEY WORDSÐdysthymia; depression; elderly; community; primary care; comorbidity; treatment

The concept of chronic depression, and thediagnostic category of dysthymic disorder inparticular, has been controversial. Individualswith long-standing depressive disorders have beenregarded at times as su�ering from an a�ectiveillness, and at other times a personality disorder.DSM-111 introduced the term `dysthymia' todescribe a chronic depressive state with symptomsof less severity than major depression. However, ofmore consequence was the inclusion of dysthymiain the a�ective disorders section, implying aconceptual shift as regards aetiology, treatmentand prognosis of chronic depressive states. Re®ne-ments were made in DSM-111-R to create a clearerdemarcation between dysthymia and major depres-sion, with a change in symptom criteria and the

exclusion of chronic depressive states whichemerged from an initial unresolved major depress-ive episode. Dysthymia was subclassi®ed as early(less than 21 years) or late onset. The DSM-IV ®eldtrials produced an alternative criteria set fordysthymia in an attempt to further delineate thedisorder, but this has been placed in the DSM-IVresearch appendix rather than in the o�cialnomenclature. A diagnosis of dysthymia inDSM-IV currently requires the presence of depres-sed mood for at least 2 years, with two of sixadditional symptoms (disturbed appetite and sleep,low energy and self-esteem, poor concentration,hopelessness).

Most studies of dysthymia in the literatureconcentrate on young adult samples. The major®ndings include the presence of a comorbid axis 1disorder in 77% of cases (Weissman et al., 1988),concomitant personality disorder in up to 85%(Markowitz et al., 1992) and a predominance ofearly onset (before 21 years) subtype (Markowitz

CCC 0885±6230/99/060440±06$17.50 Received 9 September 1998Copyright # 1999 John Wiley & Sons, Ltd. Revised 12 October 1998

Accepted 4 November 1998

INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int. J. Geriat. Psychiatry 14, 440±445 (1999)

*Correspondence to: Dr M. Kirby, Mercer's Institute forResearch on Ageing, St James's Hospital, Dublin 8, Ireland.Tel: �353 1 4537941. Ext. 2640/1. Fax: �353 1 4541796.

Contract/grant sponsor: Health Research Board, Dublin.

et al., 1992; Klein et al., 1988). Lifetime comorbid-ity for major depression is common in young adultdysthymia and was present in 68% of casesdescribed by Markowitz et al. (1992).

The only study which has assessed the clinicalfeatures of dysthymic disorder in the elderly foundmarked di�erences to those described in youngadult dysthymia (Devenand et al., 1994). In thisstudy of 40 elderly individuals with dysthymia whowere recruited to a research clinic through mediaadvertisements, only one (2.5%) had an early onsetaccording to DSM-111-R criteria, four (10%)patients had a concomitant personality disorder,®ve (12.5%) had a comorbid axis 1 disorder andseven (17.5%) had an episode of major depressionearlier in the dysthymic illness. Major stressorspreceded the onset of dysthymia in 30 (75%)individuals. However, the large majority of elderlydepressed individuals either do not present to anyservice or remain within primary care(MacDonald, 1986), and Devenand et al. (1994)point out that the dysthymic patients in their studymay not be typical of those residing in thecommunity. Therefore, a community-based sampleis essential to ascertain an accurate picture ofdysthymic disorder in the elderly. The aim of thisarticle is to describe the clinical features of late lifedysthymia as it presents in older people in thecommunity.

METHODS

The subjects were 40 individuals of 65 years andover with a diagnosis of dysthymic disorder,according to DSM-IV criteria, and identi®edduring the course of a large naturalistic study ofmental disorders among the community-dwellingelderly in Dublin (Kirby et al., 1997). This studyinvolved the screening of individuals of 65 yearsand over on the practice lists of ®ve generalpractices, using the Geriatric Mental State (GMS)(Copeland et al., 1976) semi-structured interview.The data collected by the GMS are applied to theAGECAT computerized diagnostic system (Cope-land et al., 1986) to generate standardized diag-noses. GMS±AGECAT generated diagnoses havebeen compared to those derived from DSM-111criteria and have demonstrated good agreement forcases of depression against combined majordepression and dysthymia (Copeland et al., 1990).The subjects who were GMS±AGECAT diagnosticcases of depression had a second assessment, if

agreeable, conducted by one psychiatrist (MK)using DSM-IV checklists to identify those withdysthymic disorder. Those with `double depression'(major depression at the time of assessment,superimposed on dysthymic disorder) were class-i®ed as major depression and not included asdysthymia.

A detailed symptom and course of illness historywas obtained to determine the age of onset andduration of dysthymic disorder, the occurrence ofmajor stressors immediately preceding the onset ofdysthymia and the presence of previous episodes ofmajor depression. Treatment for depressive symp-toms during the course of dysthymia was recorded.Severity of depression and symptom pro®le wasassessed using the 21-item Hamilton DepressionRating Scale (HAM-D) (Hamilton, 1960). TheMini-Mental State Examination (Folstein et al.,1975), `WORLD' version, was performed as ameasure of cognitive function. DSM-IV checklistswere used to establish comorbid axis 1 and axis 2disorders. A collateral history was obtained froman informant, usually a close relative, withparticular emphasis on assessment for a comorbidpersonality disorder. At the end of the interview, aglobal observer-rated assessment of physical im-pairment, adapted from the OARS (Duke Univer-sity Center for the Study of Aging and HumanDevelopment, 1978) was made.

A group of 630 older people who did not havecase-level depression (or dementia) were used forcomparison of physical impairment and as anindicator of the frequency of life stressors. Thesesubjects were also screened with GMS±AGECATas part of the larger study, and are a sample of thenon-depressed population (we had not collectedthe additional data on all non-depressed individ-uals in the earlier part of the larger study). The lifeevent list contained nine items covering issues ofhealth, bereavement, loss/theft, ®nancial di�culty,house move and a general item on `any otherserious upset' (over the preceding 2 years) and wasadapted from the List of Threatening Events(Brugha et al., 1985; Brugha and Cragg, 1990) foruse among the community-dwelling elderly.

RESULTS

The mean age of the 40 elderly with dysthymia was74.4 years (SD 6.6, range 65±93 years). There were27 females (68%) and 13 males and the meannumber of years of education was 10.6 (SD 1.7).

Copyright # 1999 John Wiley & Sons, Ltd. Int. J. Geriat. Psychiatry 14, 440±445 (1999)

DYSTHYMIA AMONG THE COMMUNITY-DWELLING ELDERLY 441

The group of 630 non-depressed older people didnot di�er from the dysthymic group in age (mean74.8 years, SD 6.8, range 65±98 years), gender (419females, 67%) or educational status (mean 10.9years, SD 1.8).

Table 1 describes features of the onset andcourse of dysthymia in this group. There wereonly three cases (7.5%) of DSM-IV early onsetdysthymia (before 21 years) and only nine (22.5%)had an age at onset of less than 50 years. The mostcommon stressors reported as being associatedwith the onset of the depressive disorder werebereavement (9/40) and physical illness (6/40).However, major life events would be expected tobe common in this age group, and at least onenegative life event in the previous 2 years wasreported by 445 (70%) of the non-depressedindividuals, with bereavement of a close friend orrelative being the most frequently reported event(252/630, 40%).

Of the six individuals with dysthymia and acomorbid axis 1 disorder, three had generalizedanxiety disorder and three had agoraphobia with-out panic disorder. Two subjects had comorbiddependent personality disorder and two hadavoidant personality disorder. Current reportedalcohol intake was low with a mean of 2.7 units perweek (SD 4.8, range 0±20 units) and 18 (45%) werecurrent non-drinkers. The reported mean alcoholconsumption among the non-depressed group (datain 617 cases) was similar at 3.7 units (SD 7.8, range0±56) (t � 0.795, df � 653, p � 0.43) and 236(38%) were current non-drinkers. Thirty-three(83%) elderly with dysthymia were judged to haveat least mild physical impairment. In comparison, asigni®cantly lower proportion of the non-depressedgroup (49% vs 83%) had some degree of physicalimpairment (w2 � 17, df � 1, p5 0.0001).

The mean HAM-D score was 14.7 (SD 3.1, range9±22). A�ective and cognitive symptoms occurredwith much greater frequency than vegetativesymptoms (Table 2). The severity of symptomswas usually mild or moderate, rating 1 or 2 on theHAM-D items. With regard to the suicide item, 19(47.5%) felt that life was not worth living, seven(17.5%) expressed a death wish and three (7.5%)had suicidal ideation. The mean MMSE score ofthose with dysthymia was 24.9 (SD 3.8), comparedto 26.9 (SD 2.7) in the non-depressed group(t � 4.29, df � 667, p4 0.001).

The treatment history during the course ofdysthymia is shown in Table 3.

DISCUSSION

Methodological issues

Studies on young adult dysthymia have includedsubjects with current `double depression', primarilybecause `pure dysthymia' (dysthymia withoutcurrent major depression) was rare and, therefore,di�cult to study (Devenand et al., 1994). Therelatively low rate of major depression comorbidity

Table 1. Onset and course of dysthymia in 40 community-dwelling elderly

Mean age at onset of dysthymia 60.5 yr (SD 19)

Mean duration of dysthymia 14 yr (SD 17.1)

Stressor associated with onset of dysthymia 26 (65%)

Comorbid axis 1 disorder 6 (15%)

History of a major depressive episode 10 (25%)

During dysthymic disorder 7 (17.5%)

Prior to onset of dysthymic disorder 5 (12.5%)

Cormorbid personality disorder 4 (10%)

Current alcohol dependence 0

Past history of alcohol dependence 3 (7.5%)

Table 2. The frequency of positive scores on each item ofthe HAM-D

Ham-D item N (%)

Depressed 40 (100%)

Anxiety (psychic symptoms) 39 (97.5%)

Work and interests 36 (90%)

General somatic symptoms 36 (90%)

Anxiety (somatic symptoms) 31 (77.5%)

Suicide 29 (72.5%)

Middle insomnia 27 (67.5%)

Hypochondriasis 25 (62.5%)

Initial insomnia 24 (60%)

Late insomnia 24 (60%)

Diurnal variation 23 (25.5%)

Morning 11 (27.5%)

Evening 12 (30%)

Agitation 19 (47.5%)

Gastrointestinal symptoms 15 (37.5%)

Obsessional symptoms 11 (27.5%)

Guilt 10 (25%)

Retardation 5 (12.5%)

Loss of weight 5 (12.5%)

Loss of libido 3 (7.5%)

Loss of insight 2 (5%)

Derealization and depersonalization 1 (2.5%)

Paranoid symptoms 0


442 M. KIRBY, I. BRUCE, D. COAKLEY AND B. A. LAWLOR

in the elderly with dysthymia (Devenand et al.,1994) allowed us to study pure dysthymia and toexamine clinical features, including symptom pro-®le, without the confounding problem of a super-imposed major depressive episode at the time ofassessment. This does cause some limitations whencomparisons are made with studies on young adultdysthymia as the inclusion of double depression inour study would, for instance, have increased therate of major depression comorbidity. However,pure dysthymia represents the large bulk of late lifedysthymia (Devenand et al., 1994) and we feel it isthe most appropriate approach to the study ofdysthymia in older people.

The classi®cation of chronic depressive disordersrequires a detailed history of the onset of course ofthe illness. Consequently, clinical diagnosis ofdysthymia and research studies on dysthymia relyheavily on patients' self-reports, as does this study.While cognitive distortions associated with depres-sion could result in negative retrospective biases,data from a number of studies support the accuracyof self-report historical data in chronic depression(Kandel and Davies, 1986; Billings and Moos,1984; Beekman et al., 1997). However, personalityassessment using self-report methodologies is likelyto be in¯uenced by the subject's a�ective state, andespecially so in chronic depression. We relied oninformant history, particularly relating to thesubject's premorbid functioning, to make axis 2diagnoses according to DSM-IV criteria. Thisapproach was facilitated by the late onset ofdysthymia in most of our cases, as compared toearly onset dysthymia where there is a relatively

short period of adolescence and adulthood withoutthe chronic depressive state. We feel that thisapproach is appropriate, but it is possible that itmay have resulted in a lower reported frequency ofpersonality disorder.

The prevalence of GMS±AGECAT depressionin the earlier Dublin study (Kirby et al., 1997) was10.3%. This is not a prevalence study of dysthymiaas not all GMS±AGECAT depression cases agreedto the second interview, where the DSM-IVdiagnoses were made, and our primary aim wasto recruit a group of 40 cases of dysthymia forstudy of the clinical presentation in late life.

Findings

It is not clear whether the ®ndings which relateto early onset dysthymia can be applied to lateonset dysthymia (WPA Dysthymia WorkingGroup, 1995) and, therefore, the relevance ofstudies on young adult dysthymics to the elderlymay be limited. Our ®ndings indicate that dysthy-mia in the elderly is predominately of late onset,unlike young adult dysthymia (Markowitz et al.,1992; Klein et al., 1988). Many of the stressorswhich our elderly subjects associated with the onsetof dysthymia, such as bereavements and the onsetof physical illness, are frequently issues of later lifeand are less common among young adults. It is notpossible to accurately assess the causal in¯uence ofreported life stressors in a cross-sectional study,and adverse life events were common in the lives ofnon-depressed older people. However, dysthymiaamong the community-dwelling elderly wouldappear to have its origins in later life as opposedto being a persistent state since adolescence oryoung adulthood, and the stressors which wereidenti®ed with the onset of dysthymia may indicatepotential areas for intervention.

The high frequency of superimposed episodes ofmajor depression, and personality disorder comor-bidity, appear to adversely a�ect outcome in youngadult dysthymia (Wells et al., 1992; WPA Dysthy-mia Working Group, 1995) and contribute to thetherapeutic nihilism which the dysthymic diagnosiscan generate. In contrast, dysthymia in the com-munity-dwelling elderly shows low rates of axes 1and 2 comorbidity and the signi®cant associationwith physical impairment may present a furtheropportunity for intervention.

There has been much debate as to whether dys-thymia is a distinct clinical entity or merely repre-sents mild, chronic major depression (Weissman

Table 3. Treatment history of 40 elderly dysthymicpatients

Treatment N (%)

Treatment by GP during dysthymic illness 33 (82.5%)

Antidepressant 17 (42.5%)

Benzodiazepine alone 12 (30%)

Counselling alone 4 (10%)

Currently taking psychotropic drug 23 (57.5%)

Antidepressant 9 (22.5%)

Benzodiazepine 19 (47.5%)

Benzodiazepine alone 14 (35%)

Treatment by psychiatric service during dysthymia 9 (22.5%)

Antidepressant 9

Currently attending psychiatric service 3 (7.5%)

Antidepressant 3



et al., 1988). The mean HAM-D score of our group(14.7), which was similar to that in Devenandet al.'s (1994) study on late life dysthymia (14.3),was lower than one would expect in majordepression, with the item severity scores predomin-ately in the mild to moderate range. However, thesymptom breakdown illustrates the predominanceof anxiety, somatic and functional symptoms overvegetative symptoms, and supports similar ®ndingsin young adult dysthymia (Shores et al., 1992;Keller et al., 1995) which have led to the alternativecriterion set in the appendix of DSM-IV (whichgives greater prominence to a�ective and cognitivesymptoms than vegetative symptoms). While thedi�erent symptom pro®le is not conclusive evid-ence of the validity of dysthymia as a distinctdisorder, it highlights the utility of the concept ofdysthymia as a depressive disorder with a presenta-tion that di�ers from typical major depression andwhich otherwise might not be recognized, part-icularly in primary care. The high positive score onthe life not worth living/suicidal feelings item(72.5%) emphasizes the impact of dysthymia onthe individual's perception of life and the futureand, together with the mean duration of 14.3 years,belies the view that dysthymia in the elderly is aminor or low-grade disorder.

Only three of the 40 individuals with dysthymiawere, at the time of interview, attending thepsychiatric services and this highlights the import-ance of a community, rather than an outpatient,based sample in the study of dysthymia in theelderly. While there may not yet be a consensus onthe most appropriate interventions in `non-major'depression, there is evidence for the e�cacy ofantidepressants in dysthymia (Vanelle et al., 1997)and certainly they are to be preferred to benzo-diazepines, which were most commonly prescribed.The prominence of anxiety-related symptoms inthe symptom pro®le of dysthymia provides someexplanation for the frequent precribing of benzo-diazepines. Many depressed elderly do not presentto primary care, those who do present may not bediagnosed, and even if the diagnosis is made theymay not be treated (MacDonald, 1986). However,it would appear that elderly dysthymics do presentto their GP during the course of their dysthymicdisorder and receive some treatment (82.5%),including antidepressants (42.5%), probably as aconsequence of the long duration of the disorder(mean duration of 14 years). In spite of this, onlysix (15%) were currently taking an antidepressantat a potentially therapeutic doseÐtricyclic

antidepressants at 75 mg per day (Old AgeDepression Interest Group, 1993), newer anti-depressants according to data sheets. It may bethat antidepressant drugs are not being prescribedat adequate doses or for a su�ciently long periodto achieve a therapeutic response in these patients.Recent evidence suggests that antidepressantresponse in dysthymia may take up to 6 monthsand may require higher doses if initial doses areine�ective (Vanelle et al., 1997).

This study replicates many of the ®ndings ofDevenand et al. (1994) and lends support to hissuggestion that dysthymia may be di�erent in theelderly and that `elderly dysthymic patients do notappear to be young dysthymic patients who simplygrew older'. The value of the study derives from thesample used, which is representative of the com-munity-dwelling elderly as opposed to the smallproportion of elderly dysthymics who attendspecialist psychiatric clinics. Therapeutic interven-tions must be delivered predominately fromprimary care and the high rate of presentation ofsymptoms by the elderly person to the GP at somestage of the dysthymic disorder a�ords such anopportunity for treatment. Community-basedinterventions, implemented at the level of primarycare, have been shown to be e�ective in late lifedepression (Blanchard et al., 1995) and furtherintervention trials on dysthymia among the elderlyin primary care, as opposed to merely thoseattending psychiatric services, are warranted.

ACKNOWLEDGEMENTS

The authors would like to thank our generalpractitioner colleagues for their cooperation, andthe Health Research Board for contributing to thefunding of this project.

REFERENCES

Beekman, A. T. F., Deeg, D. G. H. and Braam, A. W.et al. (1997) Consequences of major and minordepression in later life: A study of disability, well-being and service utilisation. Psychol. Med. 27,1397±1409.

Billings, A. G. and Moos, R. H. (1984) Chronic andnonchronic unipolar depression: The di�erential roleof environmental stressors and resources. J. Nerv.Ment. Dis. 172, 65±75.

Blanchard, M. R., Waterreus, A. and Mann, A. H.(1995) The e�ect of primary care nurse intervention


444 M. KIRBY, I. BRUCE, D. COAKLEY AND B. A. LAWLOR

upon older people screened as depressed. Int. J.Geriatr. Psychiat. 10, 289±298.

Brugha, T. S., Bebbington, P. and Tennant, C. et al.(1985) The list of threatening experiences: A subset of12 life event categories with considerable long-termcontextual threat. Psychol. Med. 15, 189±194.

Brugha, T. S. and Cragg, D. (1990) The list of threat-ening experiences: The reliability and validity of a brieflife events questionnaire. Acta Psychiatr. Scand. 82,77±81.

Copeland, R. J. R., Dewey, M. E. and Gri�ths-Jones,H. M. (1986) A computerized psychiatric diagnosticsystem and case nomenclature for elderly subjects:GMS and AGECAT. Psychol. Med. 16, 89±99.

Copeland, J. R. M., Dewey, M. E. and Gri�th-Jones,H. M. (1990) Dementia and depression in elderlypersons: AGECAT compared with DSM-III andpervasive illness. Int. Geriatr. Psychiat. 5, 47±51.

Copeland, J. R. M., Kelleher, M. J. and Kellett, J. M.et al. (1976) A semi-structured clinical interview for theassessment of diagnosis and mental state in the elderly:The Geriatric Mental State, 1: Development andreliability. Psychol. Med. 6, 439±449.

Devanand, D. P., Nobler, M. S. and Singer, T. et al.(1994) Is dysthymia a di�erent disorder in the elderly?Am. J. Psychiat. 151, 1592±1599.

Duke University Center for the Study of Aging andHuman Development (1978) Multidimensional Func-tional Assessment: The OARS Methodology. DukeUniversity, Duke, North Carolina.

Folstein, M. F., Folstein, S. E. andMcHugh, P. R. (1975)`Mini-Mental State'. A practical method for gradingthe cognitive state of patients for the clinician.J. Psychiatr. Res. 12, 189±198.

Hamilton, M. A. (1960) A rating scale for depression.J. Neurol. Neurosurg. Psychiat. 23, 56.

Kandel, D. B. and Davies, M. (1986) Adult sequelae ofadolescent depressive symptoms. Arch. Gen. Psychiat.43, 255±262.

Keller, M. B., Klein, D. N. andHirschfeld, R.M. A. et al.(1995) Results of the DSM-IV Mood Disorders FieldTrial. Am. J. Psychiat. 152, 843±849.

Kirby, M., Radic, A. and Bruce, I. et al. (1997) Mentaldisorders among the community-dwelling elderly. Brit.J. Psychiat. 171, 369±372.

Klein, D. N., Taylor, E. B. and Dickstein, S. et al. (1988)The early±late onset distinction in DSM-III-R dys-thymia. J. A�ect. Disord. 14, 25±33.

MacDonald, A. J. D. (1986) Do general practitioners`miss' depression in elderly patients? Brit. Med. J. 292,1365±1367.

Markowitz, J. C., Moran, M. E. and Kocsis, J. H. et al.(1992) Prevalence and comorbidity of dysthymicdisorder among psychiatric outpatients. J. A�ect.Disord. 24, 63±71.

Old Age Depression Interest Group (OADIG) (1993)How long should the elderly take antidepressants: Adouble-blind placebo-controlled study of continu-ation/prophylaxis therapy with dothepin. Brit. J.Psychiat. 162, 175±182.

Shores, M. M., Glubin, T. and Cowley, D. S. et al. (1992)The relationship between anxiety and depression: Aclinical comparison of generalised anxiety disorder,dysthymic disorder, panic disorder, and majordepressive disorder. Compr. Psychiat. 33, 237±244.

Vanelle, J. M., Attar-Levy, D. and Poirier, M. F. et al.(1997) Controlled e�cacy study of ¯uoxetine indysthymia. Brit. J. Psychiat. 170, 345±350.

Weissman, M. M., Leaf, P. J. and Bruce, M. L. et al.(1988) The epidemiology of dysthymia in ®ve commu-nities: Rates, risks, comorbidity, and treatment. Am. J.Psychiat. 145, 815±819.

Wells, K. B., Burnam, M. A. and Rogers, W. et al. (1992)The course of depression in adult outpatients: Resultsfrom the medical outcomes study. Arch. Gen. Psychiat.49, 788±794.

WPA Dysthymia Working Group (1995) Dysthymia inclinical practice. Brit. J. Psychiat. 166, 174±183.



dysthymia among the community-dwelling elderly

Documents