Depression in older people with mild stroke, carotid stenosis and peripheral vascular disease: a comparison with healthy controls

DEPRESSION IN OLDER PEOPLE WITH MILDSTROKE, CAROTID STENOSIS AND PERIPHERAL

VASCULAR DISEASE: A COMPARISON WITHHEALTHY CONTROLS

RAHUL RAO1*, STEPHEN JACKSON

2AND ROBERT HOWARD

3

1Senior Registrar, Department of Old Age Psychiatry, Maudsley Hospital and Institute of Psychiatry, London, UK2Professor, Department of Health Care of the Elderly, King's College School of Medicine and Dentistry, Dulwich

Hospital, London, UK3Senior Lecturer in Old Age Psychiatry, Department of Old Age Psychiatry, Institute of Psychiatry, London, UK

ABSTRACT

Background. Although depression has a recognised association with stroke, the role of `silent' cerebrovascularpathology associated with carotid stenosis and peripheral vascular disease remains unexplored.Methods. Four groups of 25 community residents aged 65 and over were recruited, comprising ®rst anterior

circulation stroke, carotid stenosis accompanied by transient ischaemic attack, peripheral vascular disease and a non-vascular control group. All participants were interviewed using the Hamilton Rating Scale for Depression [HRSD](including a modi®ed version) and Geriatric Depression Scale. DSM IV criteria for major depression and measures ofhandicap, social support and physical illness were also administered. Head computerised tomography (CT) scanswere performed on stroke patients to examine the relationship between lesion location and depression.Results. One hundred patients were interviewed. Stroke patients were more likely to live in a nursing home and had

less social support than other groups. Mean scores on the modi®ed Hamilton and Geriatric Depression Scales werehigher in stroke and carotid stenosis groups than controls. Patients with stroke did not show a higher prevalence ofDSM IV major depressive disorder than those with carotid stenosis. There was no relationship between the presenceof lesions a�ecting the frontal/subcortical system and prevalence/severity of depression.Limitations. Small numbers, mortality of stroke patients in hospital, possible selection bias in the control group and

use of a previously unvalidated depression rating scale all limit the study.Conclusions. A possible role for carotid stenosis in the pathogenesis of depressive disorder is suggested. Larger

studies incorporating brain imaging may be required to examine the mechanism of this association more closely. Theuse of a shorter version of the HRSD in older people with cerebrovascular disease may warrant further exploration.Copyright # 2001 John Wiley & Sons, Ltd.

KEY WORDSÐdepression; stroke; carotid stenosis; peripheral vascular disease; brain imaging; rating scales

INTRODUCTION

Interest in the relationship between cerebrovasculardisease and depression has grown since the early19th century, when attention was ®rst drawn totheir association (Durand-Fardel, 1843).Notwithstanding changing classi®catory systems,this observation has remained consistent over thepast two decades, with depression recognised as afeature of multi-infarct dementia (Hachinski et al.,

1974). The concept of `vascular depression' nowencompasses a narrower de®nition of depression inthe presence of vascular risk factors, accompanyingneuropsychological de®cit and distinct localisedbrain pathology seen on structural imaging(Alexopoulos et al., 1997).

Depression is more likely after stroke than inmedical disorders with similar levels of disability(Folstein et al., 1977) and compared with age andsex-matched populations without stroke (Dam etal., 1989; House et al., 1991; Andersen et al., 1994).The prevalence of operationally de®ned poststrokedepression during the ®rst month ranges from 11 to50% and declines over the two years after stroke

Copyright # 2001 John Wiley & Sons, Ltd. Received 24 January 2000Accepted 24 May 2000

INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int. J. Geriatr. Psychiatry 16, 175±183 (2001)

* Correspondence to: Dr R. Rao, Job Ward, Thomas GuyHouse, Guy's Hospital, London, SE1 9RT, UK. Tel: �44-20-7232-0148. Fax: �44-20-7394-1097. E-mail: [email protected]

(Robinson et al., 1984, 1987; Sharpe et al., 1990;House et al., 1991; Astrom et al., 1993; Shima et al.,1994; Burvill et al., 1995). Only two studies haveused community stroke registers to assess post-stroke depression (House et al., 1991; Burvill et al.,1995), with only one (House et al., 1991) studyingpatients following their ®rst stroke.

There still remains some uncertainty regardingthe relationship between proximity to the frontalpole and severity of depression, particularly as themajority of studies ®nding such an association havebeen con®ned to the same patient cohort (Rao,2000).

Apart from established stroke, the role of `silent'cerebrovascular disease in depression is a growingarea of research. This has been given impetus bythe ®nding of asymptomatic infarction associatedwith depression (Colantonio et al., 1992; Fujikawaet al., 1994).

This study compared the prevalence of depres-sion in people over 65 following their ®rst strokecompared to those with transient ischaemiaattack(s) (TIA), peripheral vascular disease(PVD) and a control group matched (with thestroke group) for age and sex. The relationshipbetween stroke lesion location and prevalence/severity of depression was also examined. Theprimary outcome variables were categorical/di-mensional measures of depression according togroup and (in the stroke group) lesion location.Although not subject to primary outcome analysis,the relationship between percentage stenosis/timesince ®rst TIA/side of carotid lesion and depressionwas also studied.

METHODS

Sample

Four groups of 25 patients were selected. Thenumber calculated for each group was based onstudies using the Hamilton Rating Scale forDepression and 15-item Geriatric Depressionscale in people with and without stroke. For studiesemploying the HRSD, the di�erence in meanHRSD scores between these two populations isthree or greater (Folstein et al., 1977; Andersen etal., 1994; Cummings et al., 1987; Sultzer et al.,1983).

Only one study has compared the 15-item GDSscores in older people with and without stroke;®nding the di�erence in mean GDS score to bethree points in community residents over 80 (Rao

et al., 1997). In order to detect at least a three-pointdi�erence in the GDS and modi®ed version of theHRSD (see below) for a power of 80% andsigni®cance level of 0.05, at least 19 patients wererequired in each group.

A more detailed description of patient selectionhas been described elsewhere (Rao et al., 1999a).Groups comprised patients with ®rst anteriorcirculation stroke (who had undergone a periodof rehabilitation and had been admitted to hospitalbetween six months and a year prior to interview),patients on the waiting list for carotid endarter-ectomy (su�ering one or more transient ischaemicattacks) and patients with peripheral vasculardisease awaiting femoro-popliteal bypass. Giventhe known in¯uence of amputation on mood state(Kashani et al., 1983), amputees were also ex-cluded.

The control group was matched with the strokegroup for age and sex. Since osteoarthritis iscommonly accompanied by disability and handi-cap (Pinals, 1996) and is likely to a�ect multiplejoints, the control group comprised patients over65 who had undergone elective total hip or kneereplacement for osteoarthritis, and who had under-gone a period of rehabilitation. Patients who hadundergone elective surgery were chosen rather thanall those who had sustained fractures. This decisionwas taken in view of the possible selection biasoperating in fractures secondary to falls, as falls areknown to be over-represented in those withcognitive impairment, particularly in Alzheimer'sdisease (Buchner and Larson, 1987; Morris et al.,1987; Mossey, 1985).

Procedure

A fuller description is provided elsewhere (Raoet al., 1999a). After contacting patients' generalpractitioners (GPs) by letter, documentation wasmade of patients whose GP refused for the patientto be interviewed, as well as those patientsconsidered by their GP to be too frail/cognitivelyimpaired/uncommunicative. Patients who had diedor moved out of the hospital catchment area sincethe study had begun were also excluded. Allpatients were then approached by means of a letterand patient information sheet; a week later,patients were approached for interview. Peopleagreeing to participate were excluded if they wereconsidered by the interviewer to be too frail,severely cognitively impaired, aphasic or lackedsu�cient comprehension of the English language

176 R. RAO ET AL.

Copyright # 2001 John Wiley & Sons, Ltd. Int. J. Geriatr. Psychiatry 16, 175±183 (2001)

so as to be unsuitable for interview.Documentation was made of patients in eachgroup who had been excluded from the study as aresult of the above criteria, after initial patientselection. Severe cognitive impairment was de®nedaccording to both clinical impression and a score of10 or less on the mini-mental state examination. Allparticipants were interviewed by the same assessor(RR). The study was approved by the King'sCollege Hospital Local Research EthicsCommittee.

Data analysis

Data was analysed using parametric and non-parametric tests using the Statistical Package forthe Social Sciences (SPSS/PC) version 7.5 with theapplication of Yates' correction for contingencytables with cell values of less than ®ve. ANOVAtests used post hoc analyses with Tukey A andBonferroni corrections. Correlations were analysedusing the Spearman correlation coe�cient (two-tailed signi®cance). Statistical advice was obtainedfrom the Department and Biostatistics, Institute ofPsychiatry, London.

Assessment

Hamilton Rating Scale for Depression (HRSD).A 17 item rating scale with each item measured oneither a three or ®ve point scale was used (Hamilton1960). The scale has a maximum possible score of52. Given the fact that this maximum scale scorewould not allow a three-point di�erence to bedetected between groups containing 25 patients forthe power and signi®cance levels required, ashortened version of the scale was used. Theitems chosen were drawn from the original paperby Hamilton outlining the HRSD (Hamilton,1960), in which all items were subject to factoranalysis. The factor cluster that most closelyresembled a clinical diagnosis of depressive dis-order comprised depressed mood, guilt, retar-dation, loss of insight, work and interest andsuicide: the maximum score attainable for thismodi®ed scale is 22.

Fifteen-item Geriatric Depression Scale.Originally this was a 30-item scale designed forassessing depression in the elderly (Yesavage andBrink, 1983). The shorter 15-item version includesitems that showed the highest correlation withdepressive symptoms in validation studies for theoriginal scale (Sheik and Yesavage, 1986).

Other scales/instruments. Other scales comprisedthe London handicap scale (Harwood et al., 1994)and social support scale (Livingston et al., 1990). Afuller description of these scales is provided else-where (Rao et al., 1999a). Sociodemographic data,blood pressure and a physical illness checklist(Livingston et al., 1990) were also incorporated.

Other mood-related variables comprised pre-sence/absence of DSM-IV major depressive dis-order, family history of depression, wish to die andsuicidal ideation within the past year.

Brain imaging. Head CT scans had already beenperformed on all stroke patients on admission tothe acute stroke unit using an IGE high-speedscanner, with 10 axial slices taken through thebrain at 1 cm intervals. Scans were reported blindto psychiatric diagnosis by a consultant neuro-radiologist. The presence of infarct/haemorrhagewas classi®ed according to site, age and whether ita�ected frontal/subcortical or other areas.

RESULTS

For all groups, 27 patients ful®lled diagnosticcriteria for DSM-IV major depressive disorderand 11 patients reported a history of depressionmore than a year before interview. In strokepatients, the mean time between stroke and inter-view was nine months. In the case of stroke patientswith DSM-IV depression, none of these patientsreported a depressive episode to have been presentat the time of stroke. Ten patients reported a familyhistory of depression in female relatives and four inmale relatives. Thirty patients reported a wish todie and nine reported a wish to end their livesduring the past year.

Although no measure of stroke severity wasrecorded at the time of interview, the description of`mild' stroke refers to the fact that all wereconsidered for rehabilitation. None of the patientshad dysphasia su�cient to interfere with interview-ing/administration of rating scales.

Comparison of groups for sociodemographicvariables

Comparisons across sociodemographic variableshave been reported elsewhere (Rao et al., 1999b).Brie¯y, stroke patients and controls were alreadymatched for age and sex and there were nodi�erences in mean age, educational status, racialorigin or handedness between groups. Stroke

DEPRESSION IN OLDER PEOPLE WITH CAROTID STENOSIS 177


patients were more likely to live in a nursing homeand had less social support than other groups, butshowed a similar level of handicap [mean (SD)score 521 (91)] compared with controls [mean (SD)score 542 (88)] (Mann±Whitney Ù'�250.5,p�0.3). PVD patients were more likely to bemarried and less likely to be widowed than strokepatients and controls. Stroke patients also had alonger mean duration of hospital stay thancontrols. Using the Spearman partial correlationcoe�cient, there was no relationship between totalHRSD score and handicap scale score(PCC�ÿ0.4, p�0.1), social support scale score(PCC�ÿ0.4, p�0.1), chronic physical illness scalescore (PCC�0.3, p�0.2) or type of accommo-dation (PCC�0.2, p�0.4) in the stroke group,when controlling for age, sex, social class andsystolic blood pressure.

Group comparisons for mood-related variables

Comparison of depression scale scores betweengroups is detailed in Table 1. Stroke and TIAgroups showed higher mean scores on both themodi®ed HRSD and GDS compared with con-trols. Table 2 shows group comparisons forcategorical mood-related variables. Stroke and

TIA had a higher prevalence of DSM-IV MajorDepression than the control group, none of whomful®lled criteria for Major Depression. Wish to dieshowed a higher prevalence in the stroke groupsthan all other groups.

Other associations examined in TIA patients

The relationship between side of carotid lesion(on duplex ultrasonography) and mood-relatedvariables was explored in patients with TIA.Owing to the comparatively small number ofcases involving bilateral lesions, the analysis wascon®ned to a comparison of right (N�12) versusleft sided (N�10) lesions. There was no relation-ship between lesion laterality and DSM-IV depress-ive disorder (w2�0.0, df�1, p�1.0). The timebetween interview and ®rst transient ischaemicattack (in years) was also studied in its relationshipto the above variables. This was classi®ed accord-ing to whether the ®rst TIA ®rst occurred more(N�20) or less (N�5) than 5 years before inter-view. Although two-thirds (67%) of patientsexperiencing their ®rst TIA within one year ofinterview met criteria for DSM-IV depressivedisorder compared with six out of 19 (32%) ofpatients experiencing their ®rst TIA more than one

Table 1. Comparison of groups for depression rating scales

Variable Mean (SD) score F value p� 95% Con®dence interval

Stroke TIA PVD Controls

Modi®ed HRSD Scale 7.5 4.5 2.4 0.9 17.8 50.0001{ TIA 3.1±5.9; Control 0.4±1.4

Geriatric Depression Scale 7.4 4.1 3.3 1.2 19.1 50.0001{ TIA 2.6±5.5; Control 0.9±1.6

Table 2. Comparison of groups across other mood-related variables

Variable Stroke TIA PVD Control Pearson w2 value� df p 95% Con®dence intervals for

depressive disorder

Major depression 14/25 10/25{ 3/25 0/25 24.9 3 50.0001 Stroke: 35±77; TIA: 19±61;

PVD: 0±26

History of depression

(412 months before interview)

3/25 3/25 2/25 3/25 0.3 3 1.0

Family history of depression 5/25 5/25 2/25 2/25 2.9 3 0.4

Suicidal ideation 6/25 2/25 1/25 0/25 10.1 3 0.2

Wish to die 14/25 7/25 5/25 4/25 11.6 3 0.01

�After Yates' correction.{p�0.01 for Group 2 versus Group 4 (Fisher's Exact Test).

178 R. RAO ET AL.


year previously, this was not signi®cant (w2�1.1,df�1, p�0.3). Lastly, the relationship betweenpercentage stenosis (on doppler ultrasound) andthe above variables was examined. This was de®nedaccording to whether the stenosis was greater(N�16) or less than 80% (N�9). Eight out ofsixteen (50%) patients with greater than 80%stenosis met criteria for DSM-IV depressivedisorder compared with only two out of nine(22%) with less than 80% stenosis, but this was notsigni®cant (w2�0.3, df�1, p�0.9).

Depression and lesion location

Computerised tomography (CT) head scans had

been performed on all stroke patients, but two

could not be located, leaving 23 scans to analyse.

Table 3 details the location of lesions visualised on

CT scans for each stroke patient. All lesions were

within the boundaries of areas supplied by cortical

or central branches of anterior or middle cerebral

arteries. All cortical lesions were single, with only

one patient having a single subcortical infarct

(a�ecting the frontal white matter).

All other subcortical lesions involved more than

one brain area: corpus striatum, white matter of the

frontal lobe or the internal capsule. There were no

scans in which frontal/subcortical lesions and

lesions in other brain areas co-existed. Seven

scans showed cortical atrophy.

Three patients showed bilateral infarcts, with

lesions in the contralateral (to the most recent

stroke) hemisphere present in the centrum semi-

ovale, internal capsule and lenticular nucleus.

Patients with unilateral infarcts were then divided

into two groups: those with frontal or subcortical

lesions and those with other lesions. According to

this classi®cation, 13 patients were classi®ed as

frontal/subcortical and seven as other lesions (three

parietal, four insular). There were no di�erences

between the two groups for mean scores on the

modi®ed HRSD (Mann±Whitney Ù'�44.5,p�0.9) or GDS (Mann±Whitney Ù'�54.0,p�0.9). The prevalence of depressive disorder

also showed no di�erence between groups

(w2�0.2, df�1, p�0.7).

Table 3. Location of lesion(s) in stroke patients

Right Number Left Number Total

Subcortical Subcortical

Anterior limb internal capsule 8 Anterior limb internal capsule 1 9

Centrum semiovale 5 Centrum semiovale 3 8

Lentiform nucleus 4 Lentiform nucleus 1 5

Caudate 2 Caudate 1 3

Thalamus 1 Thalamus 0 1

Cortical Cortical

Insula 2 Insula 2 4

Frontal lobe 2 Frontal lobe 1 3

Parietal lobe 2 Parietal lobe 1 3

Table 4. Distribution of scores on the Modi®ed HRSDaccording to HRSD-de®ned depression

HRSD513 HRSD513 Total

Modi®ed HRSD score

0 0 25 25

1 18 0 18

2 7 1 8

3 7 0 7

4 6 3 9

5 2 2 4

6 2 4 6

7 0 6 6

8 0 0 0

9 0 5 5

10 0 4 4

11 0 1 1

12 0 3 3

13 0 2 2

14 0 1 1

15 0 1 1

Total 67 33 100



Validation of the modi®ed Hamilton Rating Scalefor Depression

Using a conventional cut-o� point of 13 or morefor de®ning depression (Andersen et al., 1994), thedistribution of scores on the modi®ed HRSD wasexamined in all interviewees (Table 4). A cut-o�point for this modi®ed scale of 6 or more,discriminated well between people de®ned asdepressed according to the HRSD (w2�63.0,df�1, p50.0001).

Sensitivity and speci®city for this cut-o� were93% and 92%, respectively, with positive andnegative predictive values of 82% and 97%.Furthermore, the same cut-o� point on themodi®ed HRSD also discriminated betweendepressed and non-depressed individuals usingDSM-IV (w2�60.0, df�1, p50.0001).

LIMITATIONS

The rating of depression in this study was made bya single rater, thereby introducing possible ob-server bias. There may also have been selection biasthrough di�erential mortality of stroke patients inhospital. The ®nding that none of the controlgroup met diagnostic criteria for depressive dis-order is surprising and may have represented abiased sample, possibly arising from the selectionof people with elective orthopaedic procedures.

The TIA group represented a homogeneousgroup in terms of the severity of vascular pathol-ogy, also having exactly the same proportion ofright-sided, left-sided and bilateral lesions as thestroke group. The distribution of cerebrovascular(anterior circulation) pathology was also similar tothat of the stroke group. It is uncertain as towhether previous TIAs may have been associatedwith undetected strokes, as subcortical infarcts areknown to be associated with clinically `silent'strokes (Weisberg, 1982). The possibility of carotiddisease in the PVD group cannot be discounted, asPVD may be associated with coincidental carotidpathology (Ahn et al., 1991; Alexandrova et al.,1996). As no CT scans had been performed ineither TIA or PVD groups, the contribution ofcerebrovascular pathology to mood disorder isdi�cult to examine. All vascular groups had ahigher mean systolic blood pressure than thecontrol group. Hypertension has previously beenraised as a possible correlate of depression(Krishnan et al., 1994).

Group di�erences in level of handicap and typeof residence existed between groups, but it isstriking that the TIA group showed signi®cantlyless handicap than the control group, yet exhibitedhigher mean scores on both depression scales thanthis group.

Stroke patients also showed a similar degree ofhandicap to controls. Residing in a nursing home isknown to be associated with depression after strokefor up to one year (Burvill et al., 1997). However,this study showed no relationship between nursinghome residence or handicap and HRSD score inthe stroke group. The likelihood of a past or familyhistory of depression was comparable in all groups.

It is possible that patients with TIA consideredthemselves more at risk of stroke than PVDpatients; this may have accounted for di�erencesin the prevalence/severity of depression betweenthese two groups.

Use of the modi®ed Hamilton Rating Scale forDepression

As the original version of the HRSD may haveresulted in type 2 errors in group comparisons, theoriginal power calculations were for a modi®edversion of this scale. The use of this modi®edversion of the HRSD appears to be the ®rst of itskind.

Given that similar di�erences were observedbetween vascular and non-vascular groups usingthe modi®ed HRSD and GDS, together with thefact that these scales avoid an undue emphasis onbiological symptoms of depression, it is possiblethat the overlap between chronic physical symp-toms and depressive symptoms may be small. Thismay give more weight to the concept of depressionand physical illness as being phenomenologicallyseparate, but nevertheless having a possible aetio-logical association. The modi®ed version hassatisfactory sensitivity and speci®city when vali-dated against the original scale in this study.

All patients reported their ®rst stroke (veri®ed byscrutiny of their case notes) and a high proportion(92%) of their CT scans were located. However,there was no quanti®cation in terms of lesion sizeor white matter pathology.

CONCLUSIONS

The study has not elucidated any new ®ndingsrelated to mood disorder accompanying stroke;

180 R. RAO ET AL.


however, it represents one of few studies to havecompared depression in elderly people followingtheir ®rst stroke with a control group matched forage, sex and handicap. Only one study hascompared HRSD scores in stroke and orthopaedicpopulations (Folstein et al., 1977). Both groupswere interviewed 30 days after admission tohospital and were comparable in age, sex andphysical disability (measured by activities of dailyliving). The mean HRSD score for stroke patientswas 12, as opposed to eight for the orthopaedicgroup. This di�erence is considerably less than thepresent study, in which mean HRSD scores instroke and orthopaedic (control) groups were 16and 2, respectively. However, the timing of assess-ment was at least 5 months earlier than the currentstudy.

Only one study has reported the prevalence ofpoststroke depression using the GDS (Diamond etal., 1995). All patients were interviewed using the30-item GDS within three months of their stroke.Using a score of more than 10 for de®ningdepression, the prevalence of depression was 29%at admission. In the present study, using a cut-o�point of ®ve or more for the shortened (15-item)version, 19 (76%) were classi®ed as depressedcompared with none of the controls. Again, anycomparison is made di�cult by di�erences in thetime interval between stroke and assessment ofdepression compared with the present study.

There have been no other published studiesemploying the 15-item GDS to compare patientswith and without stroke.

The current study found the prevalence of DSM-IV major depressive disorder to be present in 56%of stroke patients compared with none of theorthopaedic controls. In a follow-up study of 103patients admitted to a stroke unit over a one yearperiod, Robinson et al. found a prevalence of 34%for DSM-III major depression at 6 months and14% at one year (Robinson et al., 1984). Astrom etal. (1993) found a similar prevalence (16%) ofDSM-III depression one year after stroke. Using apopulation taken from a community stroke data-base. House et al. (1991) found the prevalence ofDSM-III major depression at 6 months and oneyear to be 9% and 5%, respectively. The abovestudies reported a considerably lower prevalence ofdepression than the present study, in spite ofsimilar diagnostic criteria.

There is no evidence from this study to suggestan association between the presence/severity ofdepression and lesions con®ned to frontal/subcor-

tical brain areas. The controversy surrounding therelationship between such lesions and depressionseverity continues, but there appears to be a lack ofevidence for a consistent association (Rao, 2000).However, given the lack of quanti®cation in lesionsize/location in the present study, the absence of asigni®cant association should be treated with somecaution.

Peripheral vascular disease and depression

Although the prevalence of DSM-IV depressivedisorder was not higher in the peripheral vasculardisease (PVD) group than in controls in the presentstudy, mean scores on both depression rating scaleswere signi®cantly higher in the PVD group.

The only study to date to have examined therelationship between PVD and mood comparedpatients with PVD, stroke and an age/education-matched control group. Di�erences in `globalpsychological distress' showed no di�erences be-tween groups, but higher mean scores on the BeckDepression Inventory were observed in the PVDgroup when compared with controls. This di�er-ence was seen in both amputees and non-amputees(Phillips and Mate-Cole, 1997). Given the paucityof research into mood disorder in patients withPVD, this is an area requiring further exploration.

Transient cerebral ischaemia and depression

In the light of the ®nding from this study thatmajor depressive disorder accompanies anteriorcirculation stroke in over 50% of patients, it maynot be unreasonable to propose that ischaemicdamage within the same territory may be associ-ated with a similar prevalence of mood disorder.There has been a dearth of studies examining mooddisorder in patients with and without clinicallysigni®cant carotid stenosis. Early descriptions of a`depressive reaction' accompanying this pathology(Sands and Meredith, 1992) were followed bystudies incorporating some aspects of moodassessment into their study design. Kelly et al.(1980) found no di�erences between the two groupsin depression scale scores using the MinnesotaMultiphasic Personality Inventory. In a study ofCE patients assessed pre-operatively by Parker etal. (1983), no di�erences were seen for `depression-dejection' using the Pro®le of Mood States scale,



when compared with a control group of non-vascular surgical controls. One other study used arating scale for depression with greater validity andreliability [Beck Depression Inventory (BDI)].

None of the CE patients or healthy age/sex/education-matched controls was classi®ed asdepressed (Iddon et al., 1997). However, thestudy neither described a cut-o� point for dis-tinguishing depressed from non-depressed subjects,nor did it report mean BDI scores for the twogroups.

The present study appears to be the ®rst to useboth operational and dimensional approaches tothe assessment of depression in patients withcarotid stenosis and in a matched control group.It also appears to be the ®rst study to use an olderpatient group for the assessment of depression inpatients with a well-de®ned circumscribed degreeof stenosis.

Given the lack of di�erence in the prevalence ofDSM-IV depression between the stroke and TIAgroup, prospective studies may also seek to clarifythe role of transient ischaemia in the developmentof depression, also examining the in¯uence of ageand blood pressure.

In view of continued interest in the prevention ofstroke in TIA patients through drug trials, includ-ing measures of depression may complement thesestudies. There appears to be some limited evidencefrom this study that transient ischaemic attacksmay be related to the development of mooddisorder, but the precise aetiopathological mech-anism is unclear. There is clearly much scope forreplicating of the ®ndings of this study in bothcross-sectional and cohort studies, with a viewto re-examining the relationship between bothsymptomatic and asymptomatic cerebrovascularpathology and mood in patients withoutstroke.

ACKNOWLEDGEMENTS

We would like to thank Mr Paul Baskerville andDr Steven Novac for giving permission to inter-view patients under their care, as well as themedical records department at King's Collegeand Dulwich Hospitals, London. MargaretTompkins provided valuable assistance intracing patient records. The study was carried outas part of a University of London MD Thesiscarried out by RR on the Maudsley HigherTraining Scheme.

KEY POINTS

. Depressive symptoms were more severe inolder people with symptomatic carotid ste-nosis compared with peripheral vasculardisease and control groups using the modi-®ed Hamilton Scale for Depression andGeriatric Depression Scale.

. The modi®ed Hamilton Scale has acceptablevalidity against the original scale.

. Forty per cent of older people with sympto-matic carotid stenosis had major depression;this was not signi®cantly di�erent from 56%in the stroke group.

. There was no demonstrable relationshipbetween frontal/subcortical lesion locationand severity/prevalence of depression.

REFERENCES

Ahn SS, Baker JD, Walden K, Moore WS. 1991. Whichasymptomatic patients should undergo routine screen-ing carotid duplex scan? Am J Surg 162: 180±184.

Alexandrova NA, Gibson WC, Norris JW, MaggisanoR. 1996. Carotid artery stenosis in peripheral vasculardisease. J Vasc Surg 23: 645±649.

Alexopoulos GS, Meyers BS, Young RC, Campbell S,Silbersweig D, Charison M. 1997. `Vascular depres-sion' hypothesis. Arch Gen Psychiatry 54: 915±922.

Andersen G, Vestergaard K, Riis J, Lauritzen L. 1994.Incidence of post-stroke depression during the ®rstyear in a large unselected stroke population deter-mined using a valid standardized rating scale. ActaPsychiatr Scand 90: 190±195.

Astrom M, Adolfsson R, Asplund K. 1993. Majordepression in stroke patients. 3-year longitudinalstudy. Stroke 24: 976±982.

Buchner DM, Larson EB. 1987. Falls and fractures inpatients with Alzheimer-type dementia. J Am Assoc257: 1492±1495.

Burvill PW, Johnson GA, Jamrozik KD, Anderson CS,Stewart-Wynne EG, Chakera TM. 1995. Prevalence ofdepression after stroke: the Perth Community StrokeStudy. Br J Psychiatry 166: 320±327.

Burvill P, Johnson G, Jamrozik K, Anderson C, Stewart-Wynne E. 1997. Risk factors for post-stroke depres-sion. Int J Geriatr Psychiatry 12: 219±226.

Colantonio A, Kasi SV, Ostfeld AM. 1992. Depressiveand other psychosocial factors as predictors of strokein the elderly. Am J Epidemiol 136: 884±894.

Cummings JL, Miller B, Hill MA, Neshkes R. 1987.Neuropsychiatric aspects of multi-infarct dementiaand dementia of the Alzheimer type. Arch Neurol 44:389±393.

182 R. RAO ET AL.


Dam H, Pedersen HE, Ahlgren P. 1989. Depressionamong patients with stroke. Acta Psychiatr Scand 80:118±124.

Diamond PT, Holroyd S, Macciocchi SN, Felsenthal G.1995. Prevalence of depression and outcome on thegeriatric rehabilitation unit. Am J Phys Med Rehab 74:214±217.

Durand-Fardel M. 1843. Traite du Ramollisement daCerveau. Balliere: Paris.

Folstein MF, Maiberger R, McHugh PR. 1977. Mooddisorder as a speci®c complication of stroke. J NeurolNeurosurg Psychiatry 40: 1018±1020.

Fujikawa T, Yamawaki S, Touhouda Y. 1994. Back-ground factors and clinical symptoms of majordepression with silent cerebral infarction. Stroke 25:798±801.

Hachinski VC, Lassen NA, Marshall J. 1974. Multi-infarct dementia: a cause of mental deterioration in theelderly. Lancet 21: 207±209.

Hamilton M. 1960. A rating scale for depression. JNeurol Neurosurg Psychiatry 23: 56±62.

Harwood RH, Gompertz P, Ebrahim S. 1994. Handicapone year after stroke: validity of a new scale. J NeurolNeurosurg Psychiatry 57: 825±829.

House A, Dennis M,Mogridge L,Warlow C, Hawton K,Jones L. 1991. Mood disorders in the year after ®rststroke. Br J Psychiatry 158: 83±92.

Iddon JL, Sahakian BJ, Kirkpatrick PJ. 1997. Uncom-plicated carotid endarterectomy is not associated withneuropsychological impairment. Pharmacol BiochemBehav 56: 781±787.

Kashani JH, Frank RG, Kashani SR, Wonderlich SA,Reid JC. 1983. Depression among amputees. J ClinPsychiatry 44: 256±258.

Kelly MP, Garron DC, Javid H. 1980. Carotid arterydisease, carotid endarterectomy and behaviour. ArchNeurol 37: 743±748.

Krishnan KR, Tupler LA, McDonald WM. 1994.Background factors and clinical symptoms of majordepression with silent cerebral infarction. Stroke 25:2507.

Livingston G, Hawkins A, Graham N, Blizard B, MannA. 1990. The Gospel Oak Study: prevalence rates ofdementia, depression and activity limitation amongelderly residents in inner London. Psychol Med 20:137±146.

Morris JC, Rubin EH, Morris EJ, Mandel SA. 1987.Senile dementia of the Alzheimer's type: an importantrisk factor for serious falls. J Gerontol 42: 412±417.

Mossey JM. 1985. Social and psychologic factors relatedto falls among the elderly. Clin Geriatr Med 1: 541±555.

Parker JC, Granberg BW, Nichols WK, Jones JG,Hewett JE. 1983. Mental status outcome following

carotid endarterectomy: a six month analysis. J ClinNeuropsychol 5: 345±353.

Phillips NA, Mate-Kole C. 1997. Cognitive de®cits inperipheral vascular disease. A comparison of mildstroke patients and normal control subjects. Stroke 28:777±784.

Pinals RS. 1996. Mechanisms of joint destruction, painand disability in osteoarthritis. Drugs 52: 14±20.

Rao R, Dening TR, Brayne C, Huppert FA. 1997.Suicidal thinking in community residents over eighty.Int J Geriatr Psychiatry 12: 337±343.

Rao R, Jackson S, Howard R. 1999a. Primitive re¯exesin cerebrovascular disease: a community study of olderpeople with stroke and carotid stenosis. Int J GeriatrPsychiatry 14: 964±972.

Rao R, Howard R, Jackson S. 1999b. Neuropsycholo-gical impairment in stroke, carotid stenosis, andperipheral vascular disease: a comparison with healthycommunity residents. Stroke 30: 2167±2173.

Rao R. 2000. Cerebrovascular disease and late lifedepression: an age old association revisited. Int JGeriatr Psychiatry 15: 419±433.

Robinson RG, Starr LB, Price TR. 1984. A two yearlongitudinal study of mood disorders following stroke.Prevalence and duration at six months follow-up. Br JPsychiatry 144: 256±262.

Robinson RG, Bolduc PL, Price TR. 1987. Two-yearlongitudinal study of poststroke mood disorders:diagnosis and outcome at one and two years. Stroke18: 579±584.

Sands LP, Meredith W. 1992. Blood pressure andintellectual functioning in late midlife. J. Gerontol 47:81±84.

Sharpe M, Hawton K, House A, Molyneux A, Sander-cock P, Bamford J, Warlow C. 1990. Mood disordersin long-term survivors of stroke: associations withbrain lesion location and volume. Psychol Med 20:815±828.

Sheik JI, Yesavage JA. 1986. Geriatric Depression Scale(GDS): recent evidence and development of a shorterversion. Clin Gerontol 5: 165±173.

Shima S, Kitagawa Y, Kitamura T, Fujinawa A,Watanabe Y. 1994. Poststroke depression. Gen HospPsychiatry 16: 286±289.

Sultzer DL, Levin HS, Mahler ME, High WM,Cummings JL. 1983. A comparison of psychiatricsymptoms in vascular dementia and in Alzheimer'sdisease. Am J Psychiatry 150: 1806±1812.

Weisberg LA. 1982. Lacunar infarcts: clinical andcomputed tomographic correlations. Arch Neurol 39:37±40.

Yesavage J, Brink D. 1983. Development and validationof a Geriatric Depression Screening Scale: A prelimi-nary report. J Psychiatr Res 17: 39±49.



Documents

Depression in older people with mild stroke, carotid stenosis and peripheral vascular disease: a comparison with healthy controls