Abstract
Factors associated with reduced survival in dementia include male sex, more severe cognitive impairment, older age and longer duration of illness [1–4]. Poor behavioural functioning and higher dependency have also been associated with a worse outcome [4–6].
However, little is known about the influence of social factors on prognosis. An extensive social network seems to be protective against dementia [7]. People with dementia have been found to be sensitive to social changes in their routine and environment [8], [9]. Relocation can lead to disorientation and behavioural disturbance [10] and an increase in mortality [11]. Life events in people with dementia have been linked with acute deterioration and presentation to services [12]. Since life events have been associated with decline in physical health and death [13–15] it is possible that they adversely affect outcome for people with and without dementia. Understanding the relationship between life events, physical health and the outcome of dementia, informs predictions of prognosis. It may also offer insights into the clinical course of dementia and its management. The primary objective of this study was to examine the relationship between life events and survival for people with dementia and to see if this relationship differs from that for people without dementia. The secondary objective was to identify which psychiatric and social factors are associated with survival in people with dementia.
Method
The recruitment and evaluation of patients and controls is reported in more detail in an initial paper that examined the relationship between life events and acute deterioration in dementia [12]. The sample was 120 people with dementia and 50 fit elderly controls. The people with dementia met criteria for primary degenerative dementia including consultant diagnosis and diagnosis using the Geriatric Mental State Schedule [16]. All had an available informant. They were recruited from three sources: two local day centres (n = 50); inpatient admissions to a psychogeriatric assessment unit (n = 40); and day-patient admissions to a psychogeriatric assessment unit (n = 30). Fit, elderly controls were recruited randomly from the register of a local primary care group practice (n = 50). The practice served a large socially mixed population, in an area similar to and overlapping the catchment area for the dementia group. Patients were also excluded if they had a physical illness that might affect cognitive function (e.g. stroke, liver failure).
Initial evaluation
Information was gathered from informant and subject interviews. Demographic details included age, gender, social class (scale 1–5: 1 = professional, 5 = unskilled manual) and age at leaving full-time education. Home help, meals on wheels and support from relatives were rated as present or absent. Physical health was rated on a scale for recent physical disability. Level of functioning was measured using the Clifton Assessment Procedure for the Elderly – Behaviour Rating Scale (CAPE-BRS) which is a valid and reliable tool for assessing a wide variety of areas of behavioural functioning in the elderly [17]. Cases of dementia were identified and graded using the Geriatric Mental State Schedule (GMSS). This is a semistructured clinical interview for the assessment of diagnosis and mental state in the elderly [16]. It was also used to rate depression and anxiety as present or absent.
Life events
Life events were evaluated using the Life Events and Difficulties Schedule (LEDS) [18]. This is a semistructured interview and panelrating procedure for eliciting and rating life events and difficulties. The life event interview was conducted with a close informant. Life events were rated by a panel of trained raters according to the degree of ‘threat’ and ‘independence’ obtained by the contextual rating method of Brown and Harris. Threatening life events were those with a high level of stress and negative impact such as accidents, financial changes, illness in spouse, marital separation, bereavement or being charged with a crime. The threat of each life event was rated on a fourpoint scale: 1 indicating a great threat; 2 a moderate threat; 3 some threat; and 4 little or no threat. In this paper, the focus of life events was not addressed in rating the severity [19]. Events of great or moderate threat are therefore denoted as ‘severe’ and refer to subjects who had experienced independent life events in the 6 months prior to their index admission.
Follow-up evaluation
Follow-up was carried out 5 years after the initial evaluation. Follow-up information was gathered by contacting patients, relatives or staff at residential homes and other facilities. Hospital and GP notes were obtained, and local health authorities were contacted. Data were recorded on standardized data sheets and included details of present health or date of death. The length of follow-up was defined as the final date when all the follow-up details were collected, less the date of initial admission. Length of survival was calculated as the time between the date of initial admission and the date of death.
Statistics
The chi-squared test with Yates' correction was used to compare characteristics between the dementia and control groups. The relationships between the exposure variables and survival were examined using Cox's proportional hazards model [20]. This is a regression model that assumes that the effect of a prognostic variable on the death rate is constant over time.
Results
The follow-up survival status was found for 166 (98%) of the 170 subjects. The people with dementia and the control group did not significantly differ for male sex: 39/116 (34%) versus 14/50 (28%); age over 80, 61/116 (53%) versus 29/50 (58%); or poor physical health, 60/116 (52%) versus 23/50 (46%). The only demographic variable where there was a significant difference between the two groups was the number in social classes 4 and 5: 40/113 (35%) versus 6/50 (12%), p = 0.004.
Table 1 shows associations of severe life events with survival in the dementia group and controls. Overall, the analysis of the relationship between life events and survival was not significant for the dementia group (p = 0.2) but was for the control group (p = 0.001). An analysis of the interaction between dementia status and life event status was not significant (p = 0.052) though the hazard ratios suggest that having two or more severe life events had a worse impact on survival amongst the controls than it did for subjects with dementia. Depression was the only psychiatric or social factor associated with reduced survival (hazard ratio 1.97, 95% CI = 1.12–3.47, p = 0.03). Demographic and dementia variables associated with reduced survival were: male sex (hazard ratio 1.68, 95% CI = 1.10–2.56, p = 0.02); older age (hazard ratio 2.83, 95% CI = 1.12–7.12, p = 0.04); worse physical health (hazard ratio 2.83, 95% CI = 1.85–4.31, p < 0.001); older age of onset (hazard ratio 2.55, 95% CI = 1.34–4.86, p = 0.01); and poorer functioning as measured by the CAPE score (hazard ratio 3.79, 95% CI = 2.05–7.01, p < 0.001).
Associations of severe life events in the past 6 months with survival in the dementia group and controls
Discussion
The finding that two or more severe life events may be associated with reduced survival in people with dementia and elderly controls challenges the assumption that dementia mitigates the effects of stressful and negative events through cognitive impairment. Instead, the finding could be consistent with evidence that relocation can lead to an increase in mortality in people with dementia [11] and that emotional memories may be relatively maintained in dementia [21]. This finding lends extra weight to the belief that services for older people including those with dementia should try to offer continuity of care and environment where possible to prevent unnecessary stress.
The mechanisms by which threatening life events affect survival may be similar to those thought to link depression with poor survival. Possible pathways include the direct effect of depression on platelet function, autonomic regulation, cardiac conduction, or the immune system, or indirect effects through health behaviour [22]. Burns et al. [2] also found that depression in dementia was associated with reduced survival. In this sample, depression was associated with reduced survival but the numbers were too small to test statistically any interaction between severe life events and depression. However, Brown and Harris' original finding in younger adults was that only those events rated as severely threatening were causally related to the onset of depression [15]. More recent studies have established a link between life events and depression but it is a relationship complicated by the severity and duration of stressors and the type of depression [23]. It may be that people with dementia are poor at coping with stressful events, resulting in neuroendocrine changes and neuronal damage [19], [24].
This study has the strength that it examined life events and the outcome of dementias with a control group. However, the sample was relatively small and very few patients experienced two or more life events. Furthermore, the study was not large enough to adjust for confounders such as poor physical health or poor functioning. Female sex, lower age, good physical health, younger age of onset and less dependency were protective factors (consistent with other studies). Good physical health was associated with increased length of survival, even though the original selection process excluded patients with a physical illness that might affect cognitive function. The CAPE-BRS, a wellvalidated measure of dependency and behaviour, previously known to be associated with outcome [6], was also strongly associated with outcome in this study. This reconfirms the usefulness of CAPE-BRS as an outcome measure. Geerlings et al. [25] in a community sample of people with dementia also found no association between education and survival. Cognitive impairment is an indicator of severity of dementia and is likely to be linked with prognosis [2], [3], [26]. However, Reisberg et al. [4] did not find such an association. In this study, the limited scale used to measure cognitive function (GMSS organic scale) would have reduced the likelihood of a significant association. Causes of death were not ascertained in this study but Keene and colleagues [27] found that pneumonia and cardiovascular diseases were the main immediate causes of death in their cohort of people with dementia followed up for 11 years.
Social factors, such as life events, which may have occurred during follow-up were not taken into account and the effects of the original events may decrease over time. The informants for the people with dementia may not have known about all the life events that had occurred in the preceding 6 months although every effort was made to include all relevant life events. Future research could prospectively measure life events over a longer period of time in people with dementia. This study has identified that two or more adverse life events may be associated with reduced survival in dementia but a larger study is needed to investigate things further. Depression may be related to outcome and may reflect decompensation and adaptation difficulties in the face of stressful new experiences. There appears to be a complex interaction between depression, life events and dementia.