Longitudinal changes in verbal memory in older adults: distinguishing the effects of age from repeat testing.

OBJECTIVE: To characterize the relationship between age-related memory change and repeat testing using serial administrations of the California Verbal Learning Test (CVLT) in 385 nondemented Baltimore Longitudinal Study of Aging participants aged 55 and older with two or more memory assessments.
METHODS: The authors investigated longitudinal change and the effects of age, sex, education, and repeat testing on new learning and recall by analyzing measures of learning and interference, short- and long-delay free and cued recall, and recognition hits in separate mixed-effects regressions.
RESULTS: The authors found cross-sectional effects of age (p < or = 0.001) and sex (p < or = 0.05) on learning and interference, regardless of baseline performance. Younger adults outperformed older adults, and women outperformed men. Longitudinal age changes were documented across total learning and long-delay free and cued recall regardless of baseline scores (p < or = 0.05). In addition, controlling for baseline scores enhanced the sensitivity for detection of longitudinal age changes on Trial 5, short-delay cued recall and recognition hits (p < or = 0.05). The influence of repeated administrations changed with advancing baseline age for total learning and short- and long-delay recall such that younger baseline age was associated with improvement over time whereas older baseline age was associated with decline over time.
CONCLUSION: Longitudinal declines in CVLT performance exist in normal aging and are influenced by baseline age. Furthermore, failure to account for the influence of repeat testing with aging may decrease sensitivity to detect pathologic decline.