Cognitive scores, even within the normal range, predict death and institutionalisation.
St John PD, Montgomery PR, Kristjansson B and McDowell I;
Commented by , 28 Oct 2002
Aim of the Study
To determine if variation at the high end of cognitive test scores, in the absence of dementia, predicts mortality or institutionalisation, independent of age, functional status, and health status.
Introduction
Many older people, who do not fulfil diagnostic criteria for dementia, have significant cognitive impairment, including patients with
- low intellect/learning disabilities
- delirium
- other psychiatric disorders
- terminal medical illness
- incipient dementia
The boundary between mild cognitive impairment and frank dementia remains blurred.
Method
Secondary analysis of an existing data set was used; The Canadian Study of Health and Ageing is a longitudinal, population-based study of over 65s in Canada. The study began in 1991, with follow-up in 1996.
The Sample: Representative samples of over 65s were drawn from 10 Canadian provinces on 31.10.1990, comprising 9,008 people from the community and 1,255 from long term care institutions. Participants underwent structured interviews in their own homes comprising general health assessment, functional and social status, and cognitive assessment. Complete data was collected on 72% of the sample.
The data set was repeated in 1996.
Outcome Measures: Education was measured as numbers of years in school. Functional status was measured by the 7-item Activities of Daily Living (ADL and IADL) scales. Self-rated health was assessed by a five point scale. Cognitive status was measured by the MMSE. Outcome in 1996 in terms of mortality and institutionalisation was recorded.
The date was analysed using Student’s t-test, Chi-squared, and logistic regression.
Results
60% of the sample were women. At follow-up in 1996, 24.3% had died and 12.4% were institutionalised. (4.4% of these had been institutionalised first and subsequently died).
The main finding was a clear inverse proportionality between MMSE score and subsequent mortality. This applied equally to people at the top end of MMSE scores (ie, people scoring 25 had substantially higher risk of mortality than those scoring 27). For each point increase on the MMSE, the unadjusted odds of mortality was 0.85.
After controlling for age, gender, education and self-rated health, the association between cognition and mortality remained statistically significant : the adjusted odds ratio of mortality was 0.95 (95% Confidence Intervals, 0.93, 0.97). Similar results were found for institutionalisation.
Discussion
This study found a strong association between cognition and adverse outcomes, across the entire range of MMSE scores, including the upper end. This persisted after controlling for age, gender, and health status. This association has been previously reported for mortality, but not for institutionalisation.
The authors suggest that increased mortality amongst non-demented people with cognitive impairment is unlikely to have occurred due to development of dementia during the study period; incident cases of dementia are statistically unlikely to have died in the five-year study interval.
They conclude that interventions should be targeted not just at those fulfilling diagnostic criteria for dementia, but at those older people with low normal cognition too.
This study has limitations, some of which are acknowledged by the authors. The study ‘reworks’ existing data (ie, was not prospective). There may be other confounding variables such as socio-economic status, smoking and specific disease status, or bias in the original sample and in those dropping out.
Taken overall, however, the results suggest poor outcome of an under-researched group – those with mild cognitive impairment.