Free and cued selective reminding test predicts progression to Alzheimer's disease in people with mild cognitive impairment

Introduction To assess the diagnostic accuracy of the free and cued selective reminding test (FCSRT) for the development of Alzheimer's disease (AD) in people with mild cognitive impairment (MCI). Methods We enrolled 187 consecutive MCI outpatients from a memory clinic that were evaluated at baseline and every 6 to 12 months through an extensive clinical and neuropsychological protocol. For each test, measures of diagnostic accuracy were obtained. To improve the overall specificity of the neuropsychological battery, we also used the diagnostic tests in parallel combination. The association between FCSRT indexes and AD was tested through proportional hazard regression models with other dementia subtypes as competing event. Laplace regression was used to model time-to-AD diagnosis as a function of FCSRT indexes. Results The area under the curve of the FCSRT indexes ranged from 0.69 (95% CI: 0.62-0.76) to 0.76 (95% CI: 0.70-0.82). The specificity peaked up to 100% when we combined the category fluency test with the delayed total recall index of the FCSRT. Participants who tested positive at the FCSRT, as compared with those with negative tests, presented a twofold to fivefold higher risk of developing AD (median follow-up time 2.5 years; p < 0.001) and were diagnosed with AD 2-3 years earlier (p < 0.001). Discussion The FCSRT assessment suite shows the best predictive performance in detecting AD in people with MCI. These findings might help to reliably and timely identify people at higher risk of AD that is crucial both for properly selecting participants to clinical trials and to fine tune an effective and patient-centered care.