Prevalence of Mild Cognitive Impairment and Dementia in Patients Admitted to Non-neurological Rehabilitation Wards

Rationale. Cognitive impairment is highly prevalent among hospitalized older adults, with reported rates ranging from approximately 20% to 40% depending on setting and assessment method, yet it remains substantially underdiagnosed in general medical and rehabilitation settings, where clinical priorities are directed toward physical recovery. Undetected cognitive dysfunction has been associated with longer hospital stays, higher rates of medical complications, reduced functional recovery, and increased likelihood of institutionalization. Several conditions commonly managed in non-neurological rehabilitation units - cardiovascular disease, chronic respiratory disease, and depression - are independently associated with an increased risk of cognitive decline, further supporting the need for systematic cognitive screening in these populations.

Objectives. The primary objective is to determine the prevalence of cognitive impairment among patients aged ≥45 years admitted for non-neurological conditions to cardiac, pulmonary, or orthopedic intensive rehabilitation units, or to intermediate care units, of the Fondazione Don Carlo Gnocchi. Secondary objectives: 1) to determine the rate of cognitive impairment not previously documented in the clinical history of patients admitted to the rehabilitation wards of the Don Carlo Gnocchi Foundation for non-neurological conditions; 2) to characterize the clinical and demographic profile of patients in whom MCI or cognitive dysfunction is identified during hospitalization; 3) to evaluate modifications of the clinical profile and the progression of cognitive dysfunction at the 12-month follow-up in patients with confirmed cognitive impairment at baseline; 4) to investigate the correlation between plasma biomarkers of neurodegeneration, APOE ε4 allele dose, and cognitive profile at baseline (T0) in patients with confirmed cognitive impairment; 5) to assess longitudinal changes in plasma biomarkers of neurodegeneration and their correlation with the progression of cognitive impairment in patients presenting cognitive dysfunction at baseline (T0); 6) to evaluate longitudinal changes in resting-state EEG indices, and their correlation with the progression of cognitive impairment, in patients with cognitive dysfunction at baseline (T0); 7) to examine the correlation between baseline neurophysiological measures, including auditory P300 parameters and sleep-wake measures derived from actigraphy and polysomnography, cognitive profile at T0 and cognitive impairment progression at follow-up (T1).

Design and setting. COGinREHAB is a multicenter, prospective, longitudinal interventional study (single-group, diagnostic purpose) conducted over 36 months (January 2026-December 2028) across three sites of the Fondazione Don Carlo Gnocchi in Italy: IRCCS Fondazione Don Gnocchi (Florence); Fondazione Don Gnocchi, Istituto Palazzolo (Milan); IRCCS Fondazione Don Carlo Gnocchi ONLUS, Santa Maria Nascente (Milan).

Procedures. The study protocol comprises two assessment time points (T0 baseline and T1 12-month follow-up), structured as sequential, contingent visits.

T0 - Screening Visit (within 10 days of admission, approximately 60 minutes): review of medical records for prior documentation on cognitive status; collection of sociodemographic and clinical data (age, sex, education, diabetes, current pharmacological treatment, Cumulative Illness Rating Scale \[CIRS\], Halm's Clinical Instability Scale \[SIC\]); administration of the Montreal Cognitive Assessment (MoCA), Hospital Anxiety and Depression Scale (HADS), Lifestyle for Brain Health index (LIBRA), Cognitive Reserve Index Questionnaire-Short Version (S-CRIq), Motor Reserve Index Questionnaire (MRIq), and Current Physical Activity Questionnaire (cPAq). Patients with a MoCA Equivalent Score of 0 or 1 (overall or in at least one cognitive domain, including the Memory Index Score) proceed to the Baseline Neuropsychological Assessment. Patients scoring more than 2 standard deviations below normative values (corrected MoCA total score \<15.485) are excluded from further baseline assessment, as this is considered indicative of severe cognitive impairment.

T0 - Baseline Neuropsychological Assessment Visit: forward/backward Digit Span and Corsi Span; Rey Auditory Verbal Learning Test (immediate recall, delayed recall, intrusions, recognition, Memory Efficiency Index); Rey-Osterrieth Complex Figure (copy, immediate and delayed recall); Naming subtest of the Screening for Aphasia in NeuroDegeneration (SAND); Trail Making Test A and B; Stroop Colour-Word Test; phonemic, semantic, and alternate verbal fluency. Cognitive impairment is confirmed using neuropsychological criteria adapted from the approach of Jak and Bondi (Jak et al., 2009; Bondi et al., 2014), which require objective evidence of low performance on formal testing rather than reliance on a single test score. Consistent with this approach, impairment on an individual test is defined as an Equivalent Score (ES) ≤1 (Capitani \& Laiacona, 1997), corresponding to performance below the 20th percentile of the demographically adjusted normative distribution and considered comparable to the original Jak/Bondi threshold of \>1 SD below normative means. Cognitive impairment is confirmed when a participant obtains an ES ≤1 on at least two tests within a single cognitive domain, or on at least one test in each of three cognitive domains: (i) memory (Digit Span, Corsi Span, RAVLT, and Rey-Osterrieth Complex Figure recall); (ii) language (SAND naming subtest and semantic verbal fluency); and (iii) attention/executive functioning (Trail Making Test A and B, Stroop Colour-Word Test, and phonemic and alternate verbal fluency).

T0 - Baseline laboratory, neurophysiological, and imaging assessments performed only in patients with confirmed cognitive impairment, as defined above: venous blood sampling for APOE genotyping, including determination of APOE ε4 allele dose (0, 1, or 2 copies) and allelic frequency distribution (ε2, ε3, ε4), and quantification of plasma biomarkers of neurodegeneration (GFAP, NfL, and pTau-217, all expressed in pg/mL); non-contrast brain CT scan; multimodal neurophysiological assessment, including resting-state EEG (quantified as relative spectral power in the delta, theta, alpha, and beta bands), auditory oddball P300 event-related potentials (quantified as latency and amplitude), actigraphy-based sleep-wake monitoring (quantified as Sleep Efficiency), and overnight polysomnography (quantified as percentage of total sleep time in each sleep stage: N1, N2, N3, and REM), in a subset of patients.

T1 - 12-month Follow-up Visit (patients with confirmed cognitive impairment at T0 only): update of pharmacological treatment and CIRS; re-administration of the MoCA and the same neuropsychological battery used at T0 (alternative forms where available); repeat blood sampling for GFAP, NfL, and pTau-217; repeat resting-state EEG.

Statistical analysis. Analyses will be performed using Jamovi software. Data distribution will be assessed with the Kolmogorov-Smirnov test; continuous variables will be summarized as mean ± SD or median (IQR) and categorical variables as frequencies/percentages. The prevalence of cognitive impairment (primary outcome) and the prevalence of previously undiagnosed cognitive impairment will be estimated with 95% confidence intervals using the binomial distribution with normal approximation. Group comparisons across MoCA-defined subgroups will use the Kruskal-Wallis test or chi-square test as appropriate. Longitudinal (T0 vs T1) comparisons will use the Wilcoxon signed-rank test for continuous variables and the McNemar test for dichotomous variables. Correlations between cognitive profile, plasma biomarkers, APOE ε4 allele dose, and neurophysiological indices will be assessed using linear regression models adjusted for sociodemographic and clinical confounders, including length of hospital stay.

Ethics. The study is conducted in accordance with the Declaration of Helsinki and was approved by the Comitato Etico Regione Toscana - Area Vasta Centro (protocol no. 28251\_spe, approved 23 September 2025). Written informed consent is obtained from all participants.

• Rate of previously undiagnosed cognitive impairment [Time frame: Baseline (T0), within approximately 10 days of hospital admission]
• Clinical and demographic characteristics associated with cognitive status at baseline [Time frame: Baseline (T0), within approximately 10 days of hospital admission]
• Change in Montreal Cognitive Assessment (MoCA) Total and Domain Scores [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Digit Span Performance [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Corsi Block-Tapping Span Performance [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Rey Auditory Verbal Learning Test (RAVLT) Recall Performance and Memory Efficiency Index [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Rey-Osterrieth Complex Figure (ROCF) Performance [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Screening for Aphasia in NeuroDegeneration (SAND) Naming Subtest Score [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Trail Making Test (TMT) A and B Completion Time [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Stroop Colour-Word Test Time and Error Interference Score [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Verbal Fluency Performance [Time frame: Baseline (T0) to 12-month follow-up (T1)]
• Change in Cumulative Illness Rating Scale (CIRS) Scores [Time frame: Baseline (T0) to 12-month follow-up (T1)]