Abstract
‘Vascular depression’ has been proposed as an important disease concept largely on the basis of neuroimaging and neuropsychological studies in ageing clinical cohorts [1–3]. These studies suggest that neurologically silent cerebrovascular disease plays a key aetiological role, particularly in those older persons who present with depression for the first time after 50 years of age. Specifically, the vascular depression hypothesis proposes that small vessel and/or covert cerebrovascular disease (typically resulting from hypertension and/or diabetes) results in disruption of those fronto-subcortical circuits that regulate mood, cognition and movement [3–6]. Further, such patients are held to have a characteristic set of clinical features (psychomotor change, significant cognitive impairment), neuroimaging findings (subcortical hyperintensities on magnetic resonance imaging [MRI]) and longitudinal course (poorer response to conventional antidepressant therapies, residual cognitive impairment, progression to chronic depression, ongoing disability and subsequent cerebrovascular disease) [4, 5, 7, 8].
While not all studies support the vascular depression hypothesis [9], and some highlight the relevance of other non-vascular medical risk factors [10], a considerable number offer indirect support [4, 5, 8, 11–13]. Importantly, Steffens et al. [14] described the first major epidemiological study (n < 3300) combining crosssectional determination of MRI-defined subcortical lesions and self-reports of depressed mood. They reported a significant relationship between the number of basal ganglia lesions (but not white matter changes) and symptoms of depression. In this prospective study, we sought to determine the relationship between vascular risk factors, cerebrovascular symptoms and late-onset depressive symptoms in a community study of older persons [15]. The vascular depression hypothesis proposes that people with minor cerebrovascular symptoms and/or other clear risk factors to small vessel disease, such as hypertension, would be at increased risk longitudinally of late-onset depression.
Method
The initial study cohort, recruited between 1988 and 1989, comprised 2805 non-institutionalized residents of a regional centre (Dubbo, Australia) born before 1930 [15]. At baseline, histories of diabetes mellitus and prior cardiovascular disease were determined and measures of anthropometry, cholesterol, triglyceride, glucose, blood pressure and resting electrocardiogram were obtained. Alcohol and tobacco use, physical disability, self-rated health, general medical history and antidepressant and antihypertensive medication use were also assessed.
Ten years later (in 1998), subjects were reviewed to determine if they were alive, residing in the community and were younger than 75 (males) or 80 (females) years of age in the initial study. Those who had had a clinical stroke in the intervening decade were not considered further. Nineteen subjects were diagnosed as having had a transient ischaemic attack (TIA) since 1988–1989. This was determined from hospital discharge coding, after which each record was reviewed to confirm that the patient met the usual criteria (i.e. temporary neurological deficit, fully resolved in less than 24 h). At least two normotensive and two hypertensive control subjects were then randomly selected for each TIA case on the basis of blood pressure recordings in 1988–1989 and medication history. Hypertensive controls were using hypertensive medication at baseline. Normotensive controls were required to have systolic blood pressure < 140 mmHg and diastolic blood pressure < 90 mmHg. In 1998, 16 out of 19 TIA, 38 out of 44 hypertensive and 40 out of 45 normotensive subjects were available for clinical review. All control subjects were free of prior cardiovascular disease. Subjects gave informed written consent prior to participation.
Depression measures
In 1988, subjects were interviewed with a modified 16-item Center for Epidemiological Studies Depression Scale (CES-D) [16]. Scores were divided into tertiles for analytical purposes and subjects in the upper third were considered as possible cases of depression. At followup, a screening interview was conducted for current cognitive impairment and depressive symptoms [17]. Subjects were asked about any symptoms of depression in the last five years. If subjects gave any response indicative of a possible depressive episode, the interview then focused on whether they had experienced at least one of the two central features of major depression (i.e. persistent depressed mood and/or persistent anhedonia). If subjects reported either of these features, the full DSM-criteria set was reviewed.
It is important to note that there is ongoing debate about the appropriateness of the current DSM-IV symptom threshold for the diagnosis of major depression in older persons [18–20]. It is likely that cases of depression in older persons are underrecognized as a consequence of the need to report a large number of somatic and/or psychological symptoms that are not attributed to another medical or psychological disorder. This apparent underreporting of cases of depression in older persons may be particularly relevant to older persons who are residing in the community. Therefore, for the purposes of this study subjects were rated as a case of depression if they had experienced at least one of the two key DSM-IV features for at least two weeks.
Given the prevalence of minor depressive syndromes in older persons, subjects were also rated for the presence of 10 features of mixed depression and anxiety (MDAS; 1–10 score: irritability, waking up tired, feeling unhappy and depressed, under strain, poor sleep, easily annoyed, overwhelmed, poor concentration, loss of confidence, unable to overcome difficulties). Those subjects with at least two such symptoms were considered cases of mixed depression and anxiety. Syndromes characterized by mixed symptoms of depression and anxiety are among the commonest forms of psychiatric morbidity encountered in community, primary care and other general medical settings [21, 22].
Results
In 1988–1989, there were no significant differences between groups on measures of disability, self-rated health, alcohol consumption, marital status or blood lipid levels. At 10-year follow-up, only two subjects met full DSM-IV criteria for major depression and only five were receiving any antidepressant therapy. In 1988, CES-D rated depressive symptoms were more evident in the two control groups than in the TIA subjects (Table 1), while at 10-year follow-up, cases of depression were more prevalent in the TIA group (as compared with all control subjects; 6/16 versus 10/78: χ2 = 5.72, df = 1, p < 0.05). All of the cases of depression at follow-up in the TIA group were of late-onset. However, there was no evidence that those in the treated hypertension group were at increased risk of depression. For subjects with mixed depression and anxiety symptoms, no clear association with TIAs was evident.
Table 1. Baseline and follow-up characteristics†
Discussion
In a 10-year prospective study of ageing persons residing in the community, depressive syndromes (but not minor depressive symptoms) were associated with minor incidents of cerebrovascular disease. That is, subjects who had had a TIA without significant neurological residual disability in the period between initial assessment and follow-up were at increased risk of developing a new depressive episode. As in other similar studies, the rate of formal DSM-IV defined major depression in the entire cohort at 10-year follow-up was low (2%) and few subjects were receiving antidepressant therapy. The presence of at least one of the key features of major depression, however, was more prevalent (17%), while the rate of more non-specific depressive and anxiety symptoms was consistent with that observed in other studies (29%).
For the whole cohort, significant depressive syndromes were largely of new onset (13/16), and in the patients who reported TIAs in the follow-up period, all six cases of depression were new. The onset of such cases was not, however, predicted by the presence of treated hypertension (a key risk factor to small vessel disease). There was no clear association between more limited symptoms of depression and anxiety and cerebrovascular symptoms or treated hypertension.
To date, the vascular depression hypothesis has been based on clinical, neuroimaging and neuropsychological studies of patients with severe depressive disorders evaluated in specialist settings. The results of this study with regard to the patients with TIAs developing new depressive episodes, provide some intrinsic epidemiological support for the validity of the concept. However, the vascular depression hypothesis would also predict higher rates of depression in patients with hypertension. The absence of such a finding in this study may be due to the limited number of subjects examined, the exclusion of patients who may have already sustained severe vascular events (deaths, strokes, cardiac events and/or residential care), and the fact that hypertension was being treated at entry to the study. Due to the longitudinal nature of this study, many other illness-related factors (e.g. cholesterol, late-onset diabetes) may have been operative, and could also have influenced outcome. Clearly, consideration of such issues warrant evaluation in a larger epidemiologic study.
Footnotes
Acknowledgements
This study was funded by NHMRC Program Grant 953208.