Memory improvement in African Americans with amnestic mild cognitive impairment.

Objective: Amnestic mild cognitive impairment (aMCI) has an uncertain course. Valid methods to evaluate memory change will best identify predictors of course. This issue is especially relevant to older persons in minority groups, who may have encountered life course factors that adversely affect cognition.

Methods/design: Growth curve mixture models were used to identify trajectories of memory test scores obtained every 6 months over 2 years in 221 African Americans with aMCI.

Results: Participants sorted into two classes, with clinically and statistically significant differences in memory scores over time. Class 1 (n = 28 [14.7%]) had sustained improved scores. Class 2 (n = 162 [85.3%]) scores remained low, fluctuated, or declined. Class 1 had better baseline cognition and daily function than class 2.

Conclusions: The observed rate of improved memory is lower than reported reversion rates from aMCI to normal cognition. Evaluating trajectories of memory test scores rather than changes in categorical diagnoses of aMCI, which may depend on recalling (or not recalling) one or two words, may yield a more valid indicator of cognitive change. These approaches require further study in minority groups.