Mild cognitive impairment in the oldest old: a comparison of two approaches.

The main purpose of the current investigation was to examine the predictive utility of two sets of criteria for mild cognitive impairment (MCI). A second purpose was to determine the incidence of MCI in a longitudinal, population-based sample of the oldest old in Sweden. The participants were from the larger 'Origins of Variance in the Old-Old: Octogenarian Twins' (OCTO-Twin) study, identified through the Swedish Twin Registry. Participants were initially aged 80 or above with a mean age of 83 and were re-examined after two, three, six and eight years. The sample for this study consisted of 263 elders that were randomly selected singletons from twin pairs. Mild cognitive impairment was rated at baseline using both Petersen'sand Ritchie's criteria. Petersen's criteria emphasize memory and memory complaint whereas Ritchie's criteria use a broader set of cognitive indicators. The incidence rate was comparable to that of other studies, however, neither set of criteria predicted subsequent dementia. The failure to confirm subsequent dementia suggests that there may be many sources of MCI in very late life besides incipient dementia. The presence of a key informant, as well as the weighting of performance across domains by an experienced clinician, both of which are hard to quantify, may be the key elements that give predictive values to MCI in prior studies.