Objective: We assessed the validity of the Free and Cued Selective Reminding Test (FCSRT) in predicting dementia 2 and 5 years after initial evaluation in a population-based cohort over age 65 years participating in the French 3 Cities (3C) study.
Methods: The FCSRT was administered at baseline along with demographics, cognitive and functional scales, and a memory-complaint questionnaire. Two and 5 years later, subjects were assessed for dementia using a 2-stage design. We calculated the sensitivity, specificity, and positive and negative predictive values of various cutoff FCSRT scores in dementia prediction.
Results: Among 1,761 participants at the Bordeaux 3C center, 98 were excluded because of existing dementia, leading to 1,663 candidate participants. Among them, 1,464 underwent FCSRT and other mental status testing. Of these, 1,299 subjects were reassessed at 2 years, and 1,160 were seen 5 years later. At 2 years, 23 subjects had developed dementia; by 5 years, 60 additional subjects showed dementia. With optimized cutoff scores, the subtests of free and total recall from the FCSRT showed good sensitivity and fair specificity for the diagnosis of dementia, and the negative predictive values were high. However, positive predictive values were low: many subjects with low scores for free and total recall on FCSRT remained free of dementia at 5 years.
Conclusions: In a population-based setting, despite good sensitivity, low Free and Cued Selective Reminding Test (FCSRT) scores must be interpreted with caution. Those subjects, who often do not have memory complaints, may have unrecognized poor memory status. High FCSRT scores are useful to rule out dementia.