Safety and Immunogenicity of Repeated Doses of ABvac40 in Patients With a-MCI or Vm-AD

Subject Discontinuations Due to TEAEs

Number of withdrawn subjects due to treatment-emergent adverse events (TEAEs) during the whole study.

Time frame: Entire study duration (Week 0 to Week 104 in Part A, and Week 0 to Week 77 in Part B)

Number of Subjects With Clinically Significant Abnormalities in Physical Examination

Clinically significant (CS) abnormalities in physical examination reported during the study.

Time frame: Entire study duration (Week 0 to Week 104 in Part A, and Week 0 to Week 77 in Part B)

Number of Subjects With Clinically Significant Abnormalities in Neurological Examination

Clinically significant (CS) abnormalities in neurological examination reported during the study.

Time frame: Entire study duration (Week 0 to Week 104 in Part A, and Week 0 to Week 77 in Part B)

Number of Subjects With Clinically Significant Abnormalities in Analytical Hematology

Clinically significant (CS) abnormalities in hematology parameters reported during the study.

Time frame: Entire study duration (Week 0 to Week 104 in Part A, and Week 0 to Week 77 in Part B)

Number of Subjects With Clinically Significant Abnormalities in Analytical Biochemistry

Clinically significant (CS) abnormalities in biochemistry parameters reported during the study.

Time frame: Entire study duration (Week 0 to Week 104 in Part A, and Week 0 to Week 77 in Part B)

Number of Subjects With Clinically Significant Abnormalities in Coagulation

Clinically significant (CS) abnormalities in coagulation parameters reported during the study.

Time frame: Entire study duration (Week 0 to Week 104 in Part A, and Week 0 to Week 77 in Part B)

Level of Anti-Aβ40 Antibodies in CSF

The change in levels of anti-Aβ40 antibodies in cerebrospinal fluid (CSF) from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included the recorded outcome value as the dependent variable; treatment, protocol specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Level of Anti-Aβ40 Antibodies in Plasma

The change in levels of anti-Aβ40 antibodies in plasma from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included the recorded outcome value as the dependent variable; treatment, protocol specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline age as covariates; and measures within-patient at each visit as a repeated measure. A compound symmetric variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 2A, Week 6A, Week 10A, Week 14A, Week 18A, Week 24A, Week 40A, Week 44A, Week 50A, Week 77A, and Week 104A)

Level of Antibody-secreting Cells

The change in levels of antibody-secreting cells from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included the recorded outcome value as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 2A, Week 6A, Week 10A, Week 14A, Week 18A, Week 24A, Week 40A, Week 44A, Week 50A, Week 77A, and Week 104A)

Level of Aβ40 Peptides in Plasma - ABtest-IA

The change in levels of anti-Aβ40 peptides in plasma (ABtest-IA) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 2A, Week 6A, Week 10A, Week 14A, Week 18A, Week 24A, Week 40A, Week 44A, Week 50A, Week 77A, and Week 104A)

Level of Aβ42 Peptides in Plasma - ABtest-IA

The change in levels of anti-Aβ42 peptides in plasma (ABtest-IA) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 2A, Week 6A, Week 10A, Week 14A, Week 18A, Week 24A, Week 40A, Week 44A, Week 50A, Week 77A, and Week 104A)

Level of Aβ40 Peptides in Plasma - ABtest-MS

The change in levels of anti-Aβ40 peptides in plasma (ABtest-MS) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. A compound symmetric variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 2A, Week 6A, Week 10A, Week 14A, Week 18A, Week 24A, Week 40A, Week 44A, Week 50A, Week 77A, and Week 104A)

Level of Aβ42 Peptides in Plasma - ABtest-MS

The change in levels of anti-Aβ42 peptides in plasma (ABtest-MS) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 2A, Week 6A, Week 10A, Week 14A, Week 18A, Week 24A, Week 40A, Week 44A, Week 50A, Week 77A, and Week 104A)

Cortical Fibrillary Amyloid Deposition Assessed by a-PET Scans

The change in amyloid-PET (a-PET) standard centiloid global cortical area (reference Pons) from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Percentage of Change in Brain Volume

The percent change in brain volume from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, and Week 104A)

Percentage of Change in Hippocampal Volume

The percent change in right and left hippocampal volume from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, Week 104A)

Percentage of Change in Ventricular Volume

The percent change in ventricular volume from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, and Week 104A)

Level of Aβ42 Peptides in CSF

The change in levels of Aβ42 peptides in cerebrospinal fluid (CSF) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Level of Aβ40 Peptides in CSF

The change in levels of Aβ40 peptides in cerebrospinal fluid (CSF) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Aβ42/Aβ40 Ratio in CSF

The change in Aβ42/Aβ40 ratio in cerebrospinal fluid (CSF) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Level of Total Tau in CSF

The change in levels of total Tau in cerebrospinal fluid (CSF) from baseline to each applicable postbaseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Level of p-Tau 181 in CSF

The change in levels of p-Tau 181 in cerebrospinal fluid (CSF) from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Level of Neurofilament Light in CSF

The change in levels of neurofilament light in cerebrospinal fluid (CSF) from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Level of Neurogranin in CSF

The change in levels of neurogranin in cerebrospinal fluid (CSF) from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

Mini Mental State Examination (MMSE) Score

The change in MMSE score from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a MMRM and the ITT analysis set. MMSE is an 11-question measure that tests 5 areas of cognitive function: orientation, registration, attention and calculation, recall, and language. MMSE score ranges: 0-30, with lower scores indicating worst cognition. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (baseline, and post-baseline at Week 24A, Week 50A, Week 77A, and Week 104A)

Clinical Dementia Rating-Sum of Boxes (CDR-SB) Score

The change in CDR-SB score from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a MMRM and the ITT analysis set. CDR-SB assesses 6 cognitive and functional domains: Memory, Orientation, Judgment \& Problem Solving, Community Affairs, Home \& Hobbies, Personal Care. CDR-SB score ranges: 0-18. The higher scores mean a worst outcome. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, Week 77A, and Week 104A)

Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) Score

The change in RBANS total score from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a MMRM and the ITT analysis set. RBANS assesses 5 cognitive domains: Immediate Memory, Visuospatial/constructional, Language, Attention, Delayed Memory. Total score (range 40-160) sums the 5 domain scores. The higher scores mean a better outcome. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, Week 77A, and Week 104A)

Alzheimer's Disease Cooperative Study - Activities of Daily Living, Mild Cognitive Impairment (ADCS-ADL MCI) Score

The change in ADCS-ADL MCI total score from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a MMRM and the ITT analysis set. ADCS-ADL MCI is a 24-item scale that includes 6 basic activities of daily living (ADL) items and 16 instrumental ADL items that provide a total score: 0-78, with a lower score indicating greater severity. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, Week 77A, and Week 104A)

Trail Making Test (TMT) Scores

Change in TMT score from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a MMRM and the ITT analysis set. TMT has 2 parts in which the patient connects 25 dots in order as quickly as possible. In TMT-A, targets are numbers 1-25; in TMT-B, targets are numbers 1-13 interleaved with letters A-L. Lower timings indicate better outcome. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, Week 77A, and Week 104A)

Investigator Global Evaluation (IGE) Score

Change in IGE from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using MMRM and ITT analysis set. IGE at baseline:1-Good general status;2-Slight deterioration;3-Moderate deterioration;4-Bad general status. IGE after baseline:1-Marked improvement;2-Moderate improvement;3-Slight improvement;4-No change;5-Slight worsening;6-Moderate worsening;7-Marked worsening. MMRM included IGE after baseline as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix is used. Following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly significantly associated with response measure (p \< 0.15).

Time frame: Part A (Week 24A, Week 50A, and Week 104A)

Columbia Suicide Severity Rating Scale

Subjects with suicidal ideation or suicidal behavior since last visit.

Time frame: Part A (Week 24A, Week 50A, Week 77A, and Week 104A)

EuroQol 5 Dimensions 5 Levels (EQ-5D-5L) Overall Severity Index Score

Change in EQ-5D-5L overall severity index from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using MMRM and ITT analysis set. EQ-5D-5L has 5 dimensions: mobility, self-care, usual activities, pain/discomfort, anxiety/depression; rated: 1=no problems, 2=slight problems, 3=moderate problems, 4=severe problems, and 5=extreme problems. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)

EuroQol 5 Dimensions 5 Levels - Visual Analogue Scale (EQ-5D-5L - VAS) Score

The change in EQ-5D-5L - VAS score from baseline to each applicable post-baseline efficacy visit (Part A) was analyzed using a Mixed-Model Repeated Measures (MMRM), and the ITT analysis set. VAS records the patient's self-rated health on a vertical scale, ranging from 100 = 'Best imaginable health state' down to 0 = 'Worst imaginable health state'. The MMRM included change from baseline in the efficacy parameter as the dependent variable; treatment, protocol-specified visits, treatment-by-visit interaction, and amyloid positivity as the fixed effects; baseline efficacy parameter and baseline age as covariates; and measures within-patient at each visit as a repeated measure. An unstructured variance-covariance matrix was used. The following factors may also have been included in the model: ApoE carrier status, baseline use of AD symptomatic medication and clinical subgroup - MCI or vmAD, if found to be significantly associated with the response measure (p \< 0.15).

Time frame: Part A (Week 50A and Week 104A)