Background: disease often confounds the identification of risk factors for age-associated cognitive decline in elderly subjects. If the cognitive effects of ageing are to be distinguished from those of disease, healthy people need to be studied.
Methods: we examined the effects of incident disease and drug prescription on cognitive change in a sample of initially healthy old people in a longitudinal study and related these to age, education, social class and blood pressure. We screened general practice case notes of 10,000 patients aged 70 years and over resident in Edinburgh to identify potentially healthy subjects. We visited 1467 potential subjects at home and enquired directly about health problems and medications, administered the Mini-Mental State Examination (MMSE) and National Adult Reading Test and recorded educational attainment, occupation and blood pressure.
Results: 603 subjects (237 male, 366 female), mean age 75.7 years (range 70-88 years), reported no health problems and were taking no regular medications. Four years after the initial visit we determined the outcome of all 603 subjects and retested available survivors. Psychometric tests were then administered to the 429 (71.1%) available survivors after a median period of 4.2 years (69 subjects were dead, 15 were too unwell, 12 had moved away and 78 either refused or failed to reply). Forty-two subjects had significant sensory impairment or interrupted testing, 195 remained in good health, 29 reported or had documented disease but were on no regular medication and 163 were on regular medication for diseases diagnosed during the follow-up period. MMSE score declined by 0.3 points in the healthy group (P < 0.048). However, once a single outlier whose MMSE score fell from 29 to 22 was excluded, the mean decline for the remainder was non-significant at 0.2 points (P = 0.079). There was no significant difference in cognitive decline between those who had and those who had not started medication (P = 0.59).
Conclusions: the study fails to support the hypothesis that cognitive decline can be attributed to age alone in healthy old people. If such a decline exists, we consider that it is unlikely to account for loss of more than 0.1 MMSE point per year.