School of Neurology, Neurobiology and Psychiatry, University of Psychiatry, University Newcastle, Royal Victoria Infirmary, Newcastle upon Tyne
Institute for Ageing and Health, Newcastle General Hospital, Newcastle upon Tyne
Regional Neurosciences Centre, Newcastle General Hospital, Newcastle upon Tyne
School of Neurology, Neurobiology and Psychiatry, University of Newcastle, Royal Victoria Infirmary, Newcastle upon Tyne
Wolfson Research Centre, Institute for Ageing and Health, University of Newcastle, Newcastle General Hospital, Newcastle upon Tyne, UK
Correspondence: Dr A. J. Lloyd, School of Neurology, Neurobiology and Psychiatry, University of Newcastle, Leazes Wing, Royal Victoria Infirmary, Newcastle uponTyne NE1 4LP, UK. E-mail: a.j.lloyd{at}ncl.ac.uk
Declaration of interest Funding from the Wellcome Trust and the Stanley Foundation.
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Aims To compare hippocampal and white matter structural change in demographically matched controls and participants with early-onset and late-onset depression.
Method High-resolution volumetric magnetic resonance imaging (MRI) and rating of MRI hyperintensities.
Results Atotal of 51 people with depression and 39 control participants were included. Participants with late-onset depression had bilateral hippocampal atrophy compared with those with early-onset depression and controls. Hippocampal volumes did not differ between control participants and those with early-onset depression. Age of depression onset correlated (negatively) with hippocampal volume but lifetime duration of depression did not. Hyperintensity ratings did not differ between groups.
Conclusions Results suggest that acquired biological factors are of greater importance in late-than in early-onset illness and that pathological processes other than exposure to hypercortisolaemia of depression underlie hippocampal atrophy in depression of late life.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
METHOD |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Study participants
Individuals with depression were recruited from old age and adult
psychiatry in-patient and out-patient services. Informed consent was obtained.
Healthy controls were recruited from among patients spouses, voluntary
organisations and a list of individuals who had expressed an interest in
research. All subjects were aged 60 years or over. The following assessment
schedules were used: the MontgomeryÅsberg Depression Rating Scale
(MADRS; Montgomery & Åsberg,
1979), the Geriatric Depression Scale (GDS;
Yesavage et al,
1983), the National Adult Reading Test (NART;
Nelson, 1982) and the
Cambridge Cognitive Examination (CAMCOG;
Roth et al, 1986)
within which is the Mini-Mental State Examination (MMSE;
Folstein et al, 1975). Social class was assessed using the UK Office of Population Censuses and
Surveys (1991) definitions.
All patients met the DSMIV criteria for major depression
(American Psychiatric Association,
1994) and scored at least 20 on the MADRS. Exclusion criteria
applied to patients were: history or clinical evidence of cognitive
impairment; score of <75 on CAMCOG; electroconvulsive therapy in the 3
months preceding assessment; evidence or any history of other psychiatric
disorder, drug miuse or regular alcohol consumption above recommended maximum
levels (more than 28 units/week for males and 21 units/week for females, as
judged from screening history and review of notes); neurological disease;
insulin-dependent diabetes; other serious physical illness or any medical
condition likely to affect HPA axis function; sedative medication likely to
affect cognitive function (low-dose benzodiazepines as a hypnotic only were
allowed); and current or past use of steroids, with the exception of
occasional brief courses for obstructive airways disease, at least 3 months
prior to the study. Exclusion criteria for controls were identical, with the
addition that the maximum allowable MADRS score was 8. Age of illness onset
and lifetime duration of depression were determined by interview and
inspection of hospital and general practitioner records. Early-onset illness
was defined as having the first episode before age 60 years; late-onset
illness was defined as having the first episode at age 60 years or above.
Vascular risk factors were rated using the Framingham stroke risk scale
(Wolf et al, 1991),
those participants in the current study with ages above the range of the
Framingham study being scored as though in the Framingham maximum age
band.
Magnetic resonance imaging data
Data were acquired on a 1.0 Tesla Siemens Magnetom Impact/Expert MRI
scanner (Siemens Medical Solutions, Siemans AG, Erlangen, Germany). The same
experienced radiographer performed all scans with the use of standard head
positioning. Axial pilot scans were used to optimise initial head position.
Axial slices were aligned parallel to a plane joining the inferior-most
portions of the genu and splenium of the corpus callosum this
approximates closely to the anterior commissureposterior commissure
line. The following were acquired: a sagittal three-dimensional
magnetisation-prepared rapid acquisition gradient echo (3D-MPRAGE) turbo fast
low angle shot whole-brain volumetric T1-weighted sequence (repetition time
11.4 ms, echo time 4.4 ms, inversion time 400 ms, time to delay 50 ms, flip
angle 15°, field of view 256 mm, matrix 256x256, slice thickness 1
mm, no interslice gap) with truly isotropic voxels of side 1 mm; and rapid
acquisition with relaxation enhancement technique fast-spin echo axial dual
echo proton density and T2-weighted sequences (repetition time 2800 ms, echo
time (proton density) 14 ms/(T2) 85 ms, field of view 230 mm, matrix
256x256, slice thickness 5 mm, interslice gap 0.5 mm) to give pixel size
0.92x0.92 mm and acquisition time 4 min 13 s.
Images were stored on optical disc and then transferred via digital magnetic tape to a Sun Ultra 10 workstation running Solaris 2.7 (Sun Microsystems, Mountain View, CA, USA). All volumetric image processing was carried out on the three-dimensional T1 data-set by the same operator (A.J.L.), masked to participant identity and diagnosis, using AnalyzeAVW-3.0 software (AnalyzeDirect.com, Lenexa, KS, USA). Image volumes were cropped to remove unwanted voxels, the inferior axial level being at the foramen magnum. Volumes were digitally reorientated to bring the long axis of the hippocampi into the axial plane. This reorientated file was used for all subsequent analyses. Reduced image intensity due to field inhomogeneity was observed only inferiorly in tissues of the neck and was not observed in the scan volume containing the brain. Tissue intensity ranges for cerebrospinal fluid, white matter and grey matter were defined for each subject by sampling multiple areas of each tissue remote from the boundary with any other tissue class, thereby excluding pixels of partial volume intensity at tissue boundaries.
An index of intracranial volume was calculated from the midsagittal area of
the intracranial cavity and its maximum axial width. Whole brain volume was
measured using a semi-automated iterative process of erosion and
region-growing within Analyze-AVW. Ventricular volumes were measured to give
an additional index of brain atrophy, including localised temporal lobe
change, and were normalised for head size by dividing by the intracranial
volume. Hippocampal volumes were normalised for brain size and any overall
atrophic effect by dividing by the whole brain volume. Ventricular volumes
were calculated by setting tissue thresholds as defined previously to expand a
boundary from a seed placed in each area of ventricular cerebrospinal fluid on
coronal slices. Limits were added manually as needed to separate lateral,
third and temporal horn ventricular volumes. The hippocampus was measured on
each side separately on coronal images with reference to standard
neuroanatomical works (Jackson &
Duncan, 1996; Talairach &
Tournoux, 1998). The posterior limit on each side was the slice in
which the fornix was visible in its longest length. The anterior limit was
taken as the last slice in which the head of the hippocampus was visible
according to a priori defined boundaries and/or by visualisation of a
small, but distinct, bulge of the hippocampal head into the medial aspect of
the temporal horn. Full definitions for hippocampal boundaries are given in a
data supplement to the online version of this paper, and are available from
the authors on request. Boundaries were traced manually with a mouse-driven
cursor on interpolated, triple-sized images in the following order: lateral,
inferior, medial, superior; starting at the posterior limit and moving through
slices in a postero-anterior direction. The alveus was identifiable anteriorly
in most participants. Hippocampal volume was calculated automatically as the
sum of grey matter voxels of the areas defined. Intra-rater reliability
(intraclass correlation coefficient, ) assessed on ten participants
selected randomly and rated twice on separate occasions was excellent for all
volumetric measures, reflecting the use of operationalised definitions of
tissue intensity ranges for all tissues in each participant and minimal need
for operator-dependent boundary-drawing for cerebrospinal fluid volumes
(intracranial volume,
=0.98; brain volume,
=0.99; left
hippocampus,
=0.97; right hippocampus,
=0.99; right lateral
ventricle,
=1.0; left lateral ventricle,
=0.96; third ventricle,
=1.0; right temporal horn,
=1.0; left temporal horn,
=1.0).
White matter lesions were rated using the Scheltens rating scale (Scheltens et al, 1993) on hard copies of dual echo (T2 and proton density) axial scans. This scale has been shown previously to have good inter- and intra-rater reliability (Scheltens et al, 1993; Kapeller et al, 2003). Two raters (R.B. and J.T.OB.) independently scored each participant. Any differences in score were resolved by discussion and arrival at a consensus rating.
Statistical analysis
Analyses were conducted with SPSS for Windows, release 11. Normally
distributed continuous variables were analysed with independent
t-tests. One-way analysis of variance (ANOVA) was used for multiple
comparisons, with the conservative post hoc Scheffes test used
when significant between-group differences were identified. Based on the
Levene test, results are reported with modified values for t, d.f.
and P as appropriate when variances were unequal. Non-normally
distributed data were transformed using natural logarithms (ln) and square
roots and re-examined for fit to the normal distribution. When non-normality
persisted, non-parametric tests (MannWhitney U and
KruskallWallis with post hoc 2 as appropriate)
were used. Categorical data were compared with
2 or, when the
expected cell frequency fell below 5, Fishers exact test. Correlations
were assessed with Pearsons r or Spearmans r
for parametric and non-parametric variables, respectively.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Demographic characteristics of participants are summarised in Table 1. The mean MADRS score for the whole depression group was 30.5 (s.d.=7.2) and for controls was 2.3 (s.d.=2.1). The GDS scores for depression and control groups were, respectively, 19.9 (s.d.=7.6) and 4.2 (s.d. =3.3). Neither of these ratings differed significantly between early-onset (MADRS, 29.7 (s.d.=8.1); GDS, 20.6 (s.d.=6.5)) and late-onset (MADRS, 31.2 (s.d.=6.3); GDS, 19.4 (s.d.=8.5)) depression. A significant difference in age at onset of depression was confirmed between the depression subgroups: early-onset, 38.7 years (95% CI 33.244.1), late-onset 72.0 years (95% CI 69.774.2), U=0.00, P<0.001. Differences were also found in the number of depressive episodes (early-onset, 5.1, 95% CI 3.76.6; late-onset, 2.0, 95% CI 1.42.5; U=78.0, P<0.001) and cumulative lifetime weeks depressed (untransformed means: early-onset, 88.3; late-onset, 24.3; ln-transformed means: early-onset, 4.3 (s.d.=0.7); late-onset, 2.8 (s.d.=0.9); t=6.33, d.f.=49, P<0.001). Early-onset and late-onset illness did not differ significantly with regard to DSMIV diagnosis of melancholic depression (early-onset, 52%; late-onset, 64%) or history of previous electroconvulsive therapy (early-onset, 39%; late-onset, 21%). The presence of DSMIV psychotic depression did not reach significance (early-onset, 4%; late-onset, 21%; Fishers exact test: P=0.085). Participants with depression performed significantly less well than controls on the CAMCOG total score (depression mean of 88.6 (s.d.=7.9) and control mean of 96.8 (s.d.=4.8); t=6.10, d.f. 84, P<0.001) and memory sub-score (depression, 23.7 (s.d.=2.9); control, 26.7 (s.d.=1.9); t=5.85, d.f.=87, P<0.001). A similar relationship was found with MMSE (depression, 26.4 (s.d.=2.5); control, 26.7 (s.d.=1.9); t=2.80, d.f.=88, P=0.006). Significant between-group differences were observed on all three of these measures on comparing controls and the two depression sub-groups (CAMCOG total: F= 17.8, P<0.001; early-onset (mean=90.1, s.d.=7.5) <control, P=0.001; late-onset (mean=87.3, s.d.=8.2) <control, P<0.001; early-onset v. late-onset, no significant difference (NS); CAMCOG memory: F=16.3, P<0.001; early-onset (mean=24.2, s.d.=2.6) <control, P= 0.001; late-onset (mean=23.2, s.d.=3.2) <control, P<0.001; early-onset v. late-onset, NS; MMSE: F=5.49, P=0.006; late-onset (mean=27.7, s.d.=1.9) <control, P=0.006; early-onset v. control and early-onset v. late-onset, NS.)
|
Framingham scores for vascular risk factors did not differ in comparisons between any of the groups. Many age bands of both the control group and the participants with depression had Framingham scores above those of the original Framingham study. This was particularly so for women. Details are given in Table 2.
|
Volumetric analyses
Raw volumetric data are given in Table
3, detailing absolute values for reference. Male control
participants had greater mean values than females for intracranial volume
(male: 1468.4 cm3, s.d.=133.2; female: 1351.4 cm3,
s.d.=99.0; t=2.94, d.f.=37, P=0.006). This comparison did
not reach significance in participants with depression. Whole brain volume did
not differ between genders. All subsequent statistical analyses refer to
normalised volumes (as described in the Method section) rather than absolute
values.
|
Normalised hippocampal volumes are expressed as 103x ratio of hippocampal volume to whole brain volume. Left/right asymmetry was not found in any group for normalised ventricular volume but was evident in hippocampal volumes, the right being greater than the left in all cases. (Control left: 2.9, s.d.=0.3; right: 3.1, s.d.=0.3; t=-4.74; d.f.=38, P<0.001. All depression left: 2.8, s.d.=0.3; right: 2.9, s.d.=0.4; t=-3.75, d.f.=50, P<0.001. Early-onset left: 3.0, s.d.=0.3; right: 3.1, s.d.=0.4; t=-3.47, d.f.=22, P=0.002. Late-onset left: 2.7, s.d.=0.3; right: 2.8, s.d.=0.4; t=-2.19, d.f.=27, P=0.037.) Within the control and depression groups there was no difference in normalised hippocampal volume on the basis of gender. (Control, left males: 2.9, s.d.=0.2; females: 2.9, s.d.=0.4; t=0.42, d.f.=37, P=0.680. Control, right males: 3.1, s.d.=0.3; females: 3.1, s.d.=0.4; t= -0.08, d.f.=37, P=0.940. Depression, left males: 2.8, s.d.=0.4; females: 2.8, s.d.=0.3; t=-0.28, d.f.=49, P=0.783. Depression, right males: 2.8, s.d.=0.4; females: 3.0, s.d.=0.4; t=-0.83, d.f.=49, P=0.409.)
Normalised ventricular volumes (ln-transformed) did not differ significantly between control participants (see Table 4) and the whole depression group (left temporal horn: 3.5, s.d.=1.8; right temporal horn: 3.7, s.d.=1.8; left lateral ventricle: 8.6, s.d.=0.6; right lateral ventricle: 8.5, s.d.=0.6; third ventricle: 6.2, s.d.=0.5). Comparison of control participants and early-onset and late-onset depression groups is given in Table 4, revealing a between-group difference in the right lateral ventricle volume that was shown to be significant between the control and late-onset depression groups. Individual normalised hippocampal data are illustrated in Fig. 1. Comparisons between groups are detailed in Table 4. Both the left and right hippocampal volumes were reduced by 7% (P=0.047) and 10% (P=0.003), respectively, in participants with late-onset depression compared with controls. The reduction in late-onset compared with early-onset depression was: left, 10% (P=0.013); right, 11% (P=0.011). Hippocampal volume did not differ significantly between individuals with early-onset depression and controls.
|
|
Hyperintensity ratings
No difference was found between any of the study groups with regard to
total hyper-intensity load, total deep white matter lesion score or total
periventricular lesion score. Specific regional comparisons of deep white
matter lesion score for frontal lobes, thalami and basal ganglia similarly
showed no differences (see Table
5).
|
Correlations
The significance level for correlations was set at P=0.01. Age
correlated significantly with the following variables:
The cumulative duration of depression did not correlate with hippocampal measures, whereas age at onset of illness correlated negatively with hippocampal volume (r=-0.41, P=0.003) (see Fig. 2). Partial correlation covarying for age in this relationship gave r=-0.33 (P=0.018). An initial negative correlation of ln-transformed Framingham score with hippocampal volume in the depression group (r=-0.41, P=0.008) was lost once age was controlled for (r=-0.23, P=0.145). These two variables showed no significant correlation in control participants. The CAMCOG memory and total scores did not correlate with hippocampal volume either in individuals with depression or in control participants. Framingham risk score did not correlate with hyperintensity load in any group, nor did the CAMCOG total score.
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
These results were at variance with the expected finding of greater hippocampal atrophy in subjects with early-onset illness. Groups were generally well-matched in terms of demographic factors although the early-onset group had only one male out of 23 subjects. We are unable to define a reason for this gender mismatch other than chance. Because there was no gender effect on the measures of normalised hippocampal volume, it is unlikely that this confounded the findings. The reduction in NART score in the whole depression and late-onset groups compared with controls is difficult to ascribe causally to depression and was not matched by any difference in years of education. No significant differences were found in vascular risk (Framingham scores) between groups, which is at odds with previous work identifying clear links between vascular risk factors and depression (Everson et al, 1998).
Methodological issues
High-resolution imaging and operationalised definitions of tissue intensity
ranges minimised the degree of error due to partial volume effects and
inter-subject intensity differences in MRI data. The finding of close matching
between groups on both vascular risk factors and MRI hyperintensities,
although unexpected, adds strength to the hippocampal volumetric findings,
which are unlikely to be accounted for by the effects of generalised
cerebrovascular change. Calculation of lifetime duration of depression was
retrospective and thus may have included a degree of inaccuracy, but this was
minimised by the use of interview combined with examination of available
medical records. The majority of participants were taking medication, the main
difference between those with and without depression being the use of
antidepressants (but with sedative tricyclic medications an exclusion
criterion) with their inherent potential for hypotension. The absence of
evidence for excess vascular change (hyperintensities) in participants with
depression would argue against this being of significance in this study.
Volumetric findings
Greater ventricular volumes in males compared with females is consistent
with previous work in late life, as is the negative correlation of whole brain
volume with age in controls. Larger right than left hippocampal volume, as in
this study, fits with the majority of previous findings
(Sheline et al,
1996). Although the hypothesised reduction in hippocampal volume
in depression is sustained, the finding of greater atrophy in late-onset than
in early-onset depression and the robust lack of any significant difference
between participants with early-onset depression and controls was precisely
the opposite of that expected on the basis of the glucocorticoid cascade
hypothesis. This finding contrasts with negative results in other work that
dichotomised subjects by age at illness onset
(Steffens et al,
2000) and begins to argue against the hypothesis of more prolonged
exposure to depression (and thus hypercortisolaemia) leading to greater
hippocampal atrophy. This finding cannot be accounted for by group demographic
differences and, although the NART score differed between depression and
control groups, it did not differ between early-onset and late-onset
depression groups. It is therefore unlikely that the IQ rating results from or
is causative in this finding.
The contrasting gender mix of the early-onset and late-onset groups does not explain this difference because hippocampal volumes were normalised to brain volume to account for gender differences in both premorbid head/brain size and rate of atrophy with ageing. In contrast to work by Sheline et al (1996, 1999), we found no correlation between cumulative duration of depression and hippocampal volume. The most important difference between that study and the current work is the female-only sample and the wider age range (2386 years) in Sheline et als studies. Thus, the latter work included a whole subset of subjects less affected by ageing. There has been no direct replication of Sheline et als results, although some indirect support comes from a negative correlation described between time since onset of depression and hippocampal volume (Bell-McGinty et al, 2002). Axelson et al (1993) found a negative but statistically non-significant correlation of depression duration and amygdalo-hippocampal complex volume but also described a negative correlation with age at onset of illness when the posterior part (mostly hippocampus) of this structure was considered. Steffens et al (2000) described a strong trend towards negative correlation between age at onset and hippocampal volume that agrees with our findings of a correlation in the same direction between these variables. Along with the finding of reduced hippocampal volume in late-onset rather than early-onset depression, this further strengthens the argument against the glucocorticoid cascade hypothesis as the primary mechanism underlying hippocampal damage in this group of individuals. Cortisol data were collected on some of these people but also on others who did not have MRI and will be reported elsewhere (OBrien et al, 2004).
Brain hyperintensities
The lack of difference between any of the study groups with respect to
hyperintense lesions is surprising, given the large body of evidence
supporting increased lesion load in depression
(OBrien et al,
1996; Salloway et al,
1996). Higher levels of vascular risk factors in both participants
with depression and controls in the current study than in previous similar
work and in the Framingham study may have been a confounding factor. Also, the
number of participants is around the minimum required for such studies to be
adequately powered and therefore type II error cannot be excluded. These
results thus should not be taken as strong evidence against findings to the
contrary of previous work.
Potential mechanisms of hippocampal atrophy
In addition to the cumulative effect of exposure to cortisol alone, as
proposed by the glucocorticoid cascade hypothesis, an additional factor may
enhance vulnerability to such damage in late life. Dehydroepiandrosterone
(DHEA) has an antiglucocorticoid action
(Kalimi et al, 1994).
The cortisol:DHEA concentration ratio across years of the lifespan is a
U-shaped curve, representing periods of potentially enhanced vulnerability to
glucocorticoid neuronal damage towards the extremes of the age spectrum
(Yen & Laughlin, 1998). A
relatively high cortisol:DHEA ratio in the first years of life may be needed
for successful brain maturation but in old age this results in a prolonged
period of susceptibility to dendritic and neuronal loss. An increased
cortisol:DHEA ratio has been described recently in depression
(Young et al, 2002)
in a younger group, thus the normal increase in this ratio in late life may be
exaggerated in individuals with depression. This does not, however, offer
explanation as to the differential hippocampal volume between age-matched
participants with late-onset and early-onset illness as found in this study.
What mechanisms, then, could account for reduced hippocampal volume in people
with late-onset depression when the glucocorticoid cascade indicates that
those with early-onset illness should suffer the greatest hippocampal neuronal
damage?
Overall effects of vulnerability to damage from cortisol and the first of the three potential mechanisms proposed above thus appear to be the more likely candidates and need not be mutually exclusive. Indeed, it is perhaps the interaction of increased vulner-ability to damage with acquired biological factors that is key in the genesis of depression of onset late in life, and in the associated hippocampal volume loss observed in these individuals. Our results do not indicate a significant difference in basic cognitive function between the early-onset and late-onset depression groups, thus we do not have evidence of dementing processes in these individuals. However, the CAM-COG and MMSE may not be sufficiently sensitive to reveal subtle, early differences. Longitudinal follow-up, therefore, will be required to confirm whether or not the late-onset group contains a disproportionate number of people destined to develop dementia, their depression being an early manifestation of this, as discussed above.
It thus appears that individuals with late-onset illness have factors operating that override the glucocorticoid cascade mechanism that is supported in studies including younger adults (Sheline et al, 1996, 1999). Some important questions are raised. Is the hippocampal atrophy in late-onset illness consequent not only upon the presence of elevated cortisol in depression but also on the reduced protective factor of DHEA at the time of illness onset? Do the findings as described represent undeclared Alzheimers disease presenting initially with depressive symptomatology? Are mechanisms such as oxidative stress implicated in depression of late onset? Although volumetric imaging can define hippocampal atrophy in late-life depression, it does not indicate the nature of this (dendritic, neuronal or glial changes) and recent neuropathological and neurochemical evidence in humans has indicated only limited neuronal loss in the hippocampus consequent upon depression (Lucassen et al, 2001).
In order to address these issues there is a need for studies to follow subjects longitudinally with regard to imaging, neuroendocrine and neuropsychological function and, if possible, ultimately neuropathological investigation.
![]() |
Clinical Implications and Limitations |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
LIMITATIONS
![]() |
ACKNOWLEDGMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Axelson, D. A., Doraiswamy, P. M., McDonald,W. M., et al (1993) Hypercortisolaemia and hippocampal changes in depression. Psychiatry Research, 47, 163 -173.[CrossRef][Medline]
Bell-McGinty, S., Butters, M. A., Meltzer, C. C., et al
(2002) Brain morphometric abnormalities in geriatric
depression: long-term neurobiological effects of illness duration.
American Journal of Psychiatry,
159, 1424
-1427.
De Kloet, E. R.,Vreugdenhil, E., Oitzl, M. S., et al
(1998) Brain corticosteroid receptor balance in health and
disease. Endocrine Reviews,
19, 269
-301.
Everson, S. A., Roberts, R. E., Goldberg, D. E., et al
(1998) Depressive symptoms and increased risk of stroke
mortality over a 29-year period. Archives of Internal
Medicine, 158, 1133
-1138.
Folstein, M. F., Folstein, S. E. & McHugh, P. R. (1975) Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189 -198.[CrossRef][Medline]
Jackson, G. D. & Duncan, J. S. (1996) MRI Neuroanatomy: A New Angle on the Brain. Edinburgh: Churchill Livingstone.
Jorm, A. F., van Duijn, C. M., Chandra,V., et al (1991) Psychiatric history and related exposures as risk factors for Alzheimers disease: a collaborative re-analysis of case-control studies. EURODEM Risk Factors Research Group. International Journal of Epidemiology, 20, S43 -S47.[Abstract]
Kalimi, M., Shafagoj,Y., Loria, R., et al (1994) Anti-glucocorticoid effects of dehydroepiandrosterone (DHEA). Molecular and Cellular Biochemistry, 131, 99 -104.[Medline]
Kapeller, P., Barber, R.,Vermeulen, R. J., et al
(2003) Visual rating of age-related white matter changes on
magnetic resonance imaging: scale comparison, interrater agreement, and
correlations with quantitative measurements. Stroke,
34, 441
-445.
Lucassen, P. J., Muller, M. B., Holsboer, F., et al
(2001) Hippocampal apoptosis in major depression is a minor
event and absent from subareas at risk for glucocorticoid overexposure.
American Journal of Pathology,
158, 453
-468.
McEwen, B. S., De Kloet, E. R. & Rostene,W.
(1986) Adrenal steroid receptors and actions in the nervous
system. Physiological Reviews,
66, 1121
-1188.
Mervaala, E., Fohr, J., Kononen, M., et al (2000) Quantitative MRI of the hippocampus and amygdala in severe depression. Psychological Medicine, 30, 117 -125.[CrossRef][Medline]
Montgomery, S. A. & sberg, M. (1979) A new depression scale designed to be sensitive to change. British Journal of Psychiatry, 134, 382 -389.[Abstract]
Nelson, H. E. (1982) National Adult Reading Test. Windsor: NFER-Nelson.
OBrien, J. T. (1997) Theglucocorticoid cascade hypothesis in man. Prolonged stress may cause permanent brain damage. British Journal of Psychiatry, 170, 199 -201.
OBrien, J.T., Ames, D. & Schweitzer, I. (1996) White matter changes in depression and Alzheimers disease: a review of magnetic resonance imaging studies. International Journal of Geriatric Psychiatry, 11, 681 -694.[CrossRef]
OBrien, J.T., Lloyd, A. J.,McKeith,I.G., et al (2004) A longitudinal study of hippocampal volume, cortisol levels and cognition in older depressed subjects. American Journal of Psychiatry, in press.
Office of Population Censuses and Surveys (1991) Standard Occupational Classification. London: HMSO.
Roth, M., Tym, E., Mountjoy, C. Q., et al (1986) CAMDEX: a standardised instrument for the diagnosis of mental disorder in the elderly with special reference to the early detection of dementia. British Journal of Psychiatry, 149, 698 -709.[Abstract]
Salloway, S., Malloy, P., Kohn, R., et al (1996) MRI and neuropsychological differences in early- and late-life-onset geriatric depression. Neurology, 46, 1567 -1574.[Abstract]
Sapolsky, R. M. (2000) The possibility of neurotoxicity in the hippocampus in major depression: a primer on neuron death. Biological Psychiatry, 48, 755 -765.[CrossRef][Medline]
Sapolsky, R. M., Krey, L. C. & McEwen, B. S. (1986) The neuroendocrinology of stress and aging: the glucocorticoid cascade hypothesis. Endocrine Reviews, 7, 284-301.[Medline]
Scheltens, P., Barkhof, F., Leys, D., et al (1993) A semiquantitative rating scale for the assessment of signal hyperintensities on magnetic resonance imaging. Journal of the Neurological Sciences, 114, 7 -12.[CrossRef][Medline]
Sheline,Y. I.,Wang, P.W., Gado, M. H., et al
(1996) Hippocampal atrophy in recurrent major depression.
Proceedings of the National Academy of Sciences of the
USA, 93, 3908
-3913.
Sheline,Y. L., Sanghavi, M., Mintun, M. A., et al
(1999) Depression duration but not age predicts hippocampal
volume loss in medically healthy women with recurrent major depression.
Journal of Neuroscience,
19, 5034
-5043.
Steffens, D. C., Byrum, C. E., McQuoid, D. R., et al (2000) Hippocampal volume in geriatric depression. Biological Psychiatry, 48, 301 -309.[CrossRef][Medline]
Talairach, J. & Tournoux, P. (1998) Co-Planar Stereotoxic Atlas of the Human Brain (trans. M. Rayport). Stuttgart: Thieme.
Thomas, A. J., OBrien, J. T., Davis, S., et al
(2002) Ischemic basis for deep white matter hyperintensities
in major depression: a neuropathological study. Archives of General
Psychiatry, 59, 785
-92.
Wolf, P. A., DAgostino, R. B., Belanger, A. J., et al (1991) Probability of stroke: a risk profile from the Framingham Study. Stroke, 22, 312 -318.[Abstract]
Yen, S. S. & Laughlin, G. A. (1998) Aging and the adrenal cortex. Experimental Gerontology, 33, 897 -910.[CrossRef][Medline]
Yesavage, J. A., Brink, T. L., Rose, T. L., et al (1983) Development and validation of a geriatric depression screening scale: a preliminary report. Journal of Psychiatric Research, 17, 37 -39.[CrossRef]
Young, A. H., Gallagher, P. & Porter, R. J.
(2002) Elevation of the cortisol:dehydroepiandrosterone ratio
in drug-free depressed patients. American Journal of
Psychiatry, 159, 1237
-1239.
Received for publication September 8, 2003. Revision received January 19, 2004. Accepted for publication February 1, 2004.
Related articles in BJP: