FREE AND CUED SELECTIVE RECALL REMINDING TEST (FCSRT): STANDARDIZZAZIONE ITALIANA ED APPLICAZIONE IN UNA COORTE DI SOGGETTI CON MILD COGNITIVE IMPAIRMENT (MCI)

Abstract Background The presence of episodic memory impairment is required for the diagnosis of Alzheimer’s dementia by all current diagnostic criteria. The new research criteria proposed by Dubois et al. (Lancet Neurol 6:734–746, 2007) require that the impairment should not improve significantly with cueing, recognition testing nor after the control of effective encoding. This is considered to be the core deficit of ‘‘prodromal Alzheimer’s disease’’. The Free and Cued Selective Reminding Test (FCSRT) is a memory test that allows in assessing these specific features of memory impairment. FCSRT has been proposed as a sensitive marker of the “hippocampal memory deficit” of early Alzheimer’s Disease (AD) (B. Dubois et al. Lancet Neurol 2010; 9:1118-1127). The aims of this research are: 1) To develop an italian version of the FCSRT and standardize normative data in a sample composed by 227 adult italians homogeneously distributed by age, sex and education (Part 1). 2) To assess the construct validity of FCSRT, in terms of convergent and divergent validity (Part 2). 3) To assess diagnostic accurancy of the FCSRT in a sample composed by Mild Cognitive Impairment (MCI) affected subjects (Part 3). Part 1 (P. Frasson et al. Neurological Sciences 2011; DOI 10.1007/s10072-011-0607-3) We report normative data for an Italian version of the FCSRT. The test is based on the 12 pictorial stimuli, 6 belonging to the living domain, and 6 to the nonliving domain. Six scores were derived from the performance of 227 healthy Italian adults, with age, sex and education homogenously distributed across subgroups: immediate free recall (IFR), immediate total recall (ITR), delayed-free recall (DFR), delayed total recall (DTR), Index of Sensitivity of Cueing (ISC), number of intrusions. In multiple regression analyses, age emerged as an influencing factor for both IFR and DFR, with older people obtaining lower scores. Education and gender appear to influence only IFR, with better performance by more educated subjects and females. Adjusted scores were used to determine inferential cutoff scores and to compute equivalent scores. Part 2 To assess its construct validity, the FCSRT was administered to 146 community-dwelling (females 57.5% , age 73 ± 7.8 yrs; education 8.8 ± 4.2 yrs) subjects with memory complaints attending two memory clinics (from the L. Sacco Hospital and the IRCCS S. Raffaele). Following both clinical and neuropsychological examination, the study population consisted of 15 patients with mild AD, 80 subjects with MCI, including 12 amnestic MCI, 27 non-amnestic MCI, 41 multiple domain MCI, and 51 subjects with Subjective Memory Impairment. To assess convergent validity five FCSRT scores (IFR, ITR, DFR, DTR) and ISC) were correlated with two episodic memory tests: Story Recall (SR) and Rey Auditory Verbal Learning test immediate (I-RAVLT) and delayed (D-RAVLT) scores. To assess divergent validity a factor analysis was performed including, in addition to the above mentioned memory tasks, the following tests: Raven Coloured Progressive Matrices, both semantic and phonemic fluencies, Rey complex figure copy, Stroop test, and trail making test (part A and B). All FCSRT scores were correlated with SR, I-RAVLT and D-RAVLT scores (Pearson’s r equal to or major of 0.45 in all cases, with p < 0.0001). Free immediate (IFR vs. I-RAVLT) and delayed (DFR vs. D-RAVLT) recall measures were highly correlated (respectively r = 0.60 and r = 0.71). The factor analysis (principal component analysis with varimax rotation) identified three factors (memory, verbal fluency, executive functions), with all the FCSRT scores loading only on the memory factor. To our knowledge, this is the first study specifically designed to assess the construct validity of the FCSRT. Our results demonstrate that FCSRT is highly correlated with other routinely used memory tests and that it is independent from language and executive tasks. Overall these findings support the construct validity of the FCSRT as a memory measure. Part 3 To examine the diagnostic accuracy of FCSRT in predicting the progression of MCI to dementia (DSM IV criteria) 108 MCI consecutive patients from the L. Sacco Hospital were followed-up for an average of 14.9  4.42 months. Twenty-six (24%) out of 108 MCI subjects progressed to dementia, including 19 AD cases. These subjects were older at baseline (76.65 vs. 73.33 yrs; p= 0.037) and had lower Minimental State Examination (MMSE) score (24.62 ± 2.64 vs. 25.98 ± 2.42; p=0.016) as compared to non-converters. Moreover the proportion of subjected with spared Instrumental Actvities of Daily Living (IADL) was lower among converters as compared to non-converters (42.3% vs. 69.1%; p= 0.016). At baseline all the FCSRT subscores (IFR, ITR, DFR, DTR,ISC) were statistically significant lower in the MCI group who progressed to dementia, as compared to non-converters (26 vs 82). The ROC analysis showed that the sensitivity and specificity of the FCSRT cut-off subscores were as follow: IFR 69% and 60% , ITR 65% and 66%, DFR 81% and 66%, DTR 62% and 80%, ISC 65% and 68%. The multivariate analysis, adjusted for demographic variables, MMSE score, IADL score and MCI subtypes showed that the risk of progression from MCI to dementia was associated with the pathological score of DFR (OR 8.68; 95% CI 1.26-59.96), loss of autonomy measured by the IADL (OR 4.56; 95% CI 1.35-15.47), increasing age (OR 1.13: 95% CI 1.01-1.25) and the decline of the ISC score (OR 0.001; 95% CI 0.00-0.47). We conclude that the FCSRT is useful tool in predicting progression of MCI to dementia.