Mild cognitive impairment: prevalence and incidence according to different diagnostic criteria.
Busse A, Bischkopf J, Reidel-Heller SG and Angermeyer MC;
Commented by , 26 May 2003
Aim of the Study
To identify the age-specific prevalence, incidence and predictive validities for four diagnostic concepts of mild cognitive impairment.
Method
Participants were derived from 1,692 over 75s, in the Leipzig Longitudinal Study of the Aged; these individuals were identified from the Community by age-ordered random sampling, and included those in residential care.
Clinical interviews incorporating neuropsychological testing were conducted with 1,265 participants. 220 were diagnosed with dementia (DSM-IV criteria) and were excluded from this study.
Structured clinical interviews on the remaining 1,045 were conducted in participants’ homes by trained Psychologists and Physicians, using a comprehensive battery of tests. Corroborative history was obtained on cognitive and psychosocial functioning. Baseline assessments were repeated after 18 and 36 months.
Mild Cognitive Impairment (MCI) was defined in four ways:
- MCI – amnestic (after Petersen, 2001)
- Modified MCI – no subjective complaint of memory loss
- Age-Associated Cognitive Decline (after Levy, 1994) – AACD
- Modified AACD – no subjective complaint of memory loss.
Results
Of the original sample of 1,692, 242 declined to participate, 57 died and 15 were not traceable. The remaining 1,265 did not differ from the total sample in age, gender or marital status. Having then excluded a further 220 with dementia at baseline, and a further 116 meeting other exclusion criteria, data was analysed on 929 over 75s.
Prevalence
At baseline, the diagnoses were:
- MCI 3.1 %
- MCI-modified 5.1 %
- AACD 8.8 %
- AACD – modified 19.7 %
The prevalence of AACD – but not MCI – increased with age.
Incidence
The annual incidence rates varied from 8 (MCI) to 77 (AACD-modified) per 1,000 person years.
Conversion to Dementia
89 people developed dementia in 2.6 years. The conversion rates were similar for all four categories (23 % - 47 %); compared to 5 % of the original sample who did not fulfil diagnostic criteria for cognitive impairment at baseline who then developed dementia.
A receiver operating characteristic analysis suggested a much better predictive value for developing dementia for AACD and AACD-modified than MCI or MCI-modified;
AACD-modified criteria showed 62 % sensitivity compared to 10 % sensitivity for MCI.
Discussion
“Mild cognitive impairment” is a term gaining increasing credence in the literature, and the acronym “MCI” is increasingly used by clinicians. This study highlights the importance of using tightly defined diagnostic criteria, and suggests that the older term “age-associated cognitive decline” (possibly modified to exclude the unreliable subjective complaint of memory loss) may be a better term to use for two reasons:
- Its prevalence increases with age (in contrast to MCI)
- It is more predictive of subsequent conversion to dementia