The ADAS-cog is a structured scale (approximately 40 minutes to complete) that evaluates memory, orientation, attention, reasoning, language and constructional praxis (Rosen, 1984). The ADAS-cog scoring range for the version used in this study is from 0 to 70, with higher scores indicating greater cognitive impairment.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. B-Study excludes data from visits occurring after A-Study results announcement on 09 Apr 2018. ADAS-cog was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

The CDR scale is used as a global measure of dementia and is completed by a clinician in the setting of detailed knowledge of the individual patient collected from interviews with the patient and caregiver (Berg, 1988). The CDR describes 5 degrees of impairment in performance on each of 6 categories including memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care. CDR ratings are 0 for healthy individuals, 0.5 for questionable dementia and 1, 2 and 3 for mild, moderate and severe dementia as defined in the CDR scale. The scores for each category can also be summed and this is known as the sum of box score (CSR-SB). Sum of box scores range from 0 to 18 with higher scores indicating greater cognitive impairment.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. B-Study excludes data from visits occurring after A-Study results announcement on 09 Apr 2018. CDR-sb was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

Percent of Total Hippocampus Atrophy to Intracranial Volume

Number of subjects with paired Baseline and Month 18 MRIs

Extent and severity of brain hypometabolism was assessed centrally at Baseline and Month 18. SUVR PET was designed to make use of FreeSurfer-based segmentations of the brain obtained using the 3DT1 MRI. Following the methods published by Landau and Jagust (Landau SM, Annals of Neurology 2012) and described on the ADNI website (http://adni.loni.usc.edu/methods/pet-analysis-method/), regions were defined in native patient space on the 3DT1 MRI acquired at the Baseline visit and at Month 18 visit. An SUVR measure was computed regionally over five sub-regions (anterior/posterior cingulate, temporal, parietal, frontal and hippocampal areas), normalized to activity in the cerebral white matter. These sub-regions were selected to optimize sensitivity in longitudinal studies. This outcome measure presents the change from baseline in the normalized mean composite SUVR of the 5 regions. A negative change from baseline indicates a decrease (worsening) in brain glucose metabolism/utilization.

Analyzed as part of an optional FDG-PET substudy

The ADCS-ADL is an activity of daily living inventory developed by the ADCS to assess functional performance in participants with AD (Galasko et al., 1997). Informants are queried via a structured interview format as to whether participants attempted each item in the inventory during the preceding 4 weeks, as well as their level of performance. Scores range from 0-78 with lower scores indicating greater functional impairment.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. B-Study excludes data from visits occurring after A-Study results announcement on 09 Apr 2018. ADCS-ADL was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

The MMSE is a brief 30-point test that is used to assess cognition (Folstein, 1975). It is commonly used to screen for dementia. In the time span of about 10 minutes, it samples various functions, including arithmetic, memory and orientation. Scores range from 0-30 with lower scores indicating greater cognitive impairment.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. B-Study excludes data from visits occurring after A-Study results announcement on 09 Apr 2018. MMSE was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

The NPI is a well-validated, reliable, multi-item instrument to assess psychopathology in AD based on an interview with the caregiver (Cummings et al, 1994). It evaluates both the frequency and severity of 12 behavioral areas including delusions, hallucinations, dysphoria (depression) anxiety, agitation/aggression, euphoria, disinhibition, irritability, lability, apathy, aberrant motor behavior, appetite and eating changes and night-time behaviors.

Frequency assessments range from 1 (occasionally, less than once per week) to 4 (very frequently, once or more per day or continuously) as well as severity (1= mild, 2 = moderate, 3 = severe). Distress is rated by the study partner or caregiver and ranges from 0 (no distress) to 5 (very severe or extreme). The overall score and the score for each subscale are the product of severity and frequency. Scores range from 0-144 with higher scores indicating a greater presence of neuropsychiatric symptoms.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. B-Study excludes data from visits occurring after A-Study results announcement on 09 Apr 2018. NPI was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

The DEMQOL-Proxy questionnaire is a validated and reliable questionnaire that is interview administered and completed by the caregiver about the patient's health related quality of life (Smith et al, 2005). It consists of 31 items representing 5 domains (daily activities and looking after yourself, health and well-being, cognitive functioning, social relationships, and self-concept) and takes approximately 20 minutes to complete. Scores range 31-124 with higher scores indicate better health related quality of life.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. DEMQOL was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

The COWAT is a measure of verbal fluency in which the participant is asked to generate orally as many words as possible that begin with the letters "F", "A", and "S", excluding proper names and different forms of the same word. (Borkowski, 1967, Loonstra 2001) For each letter, the participant is allowed one minute to generate the words. Performance is measured by the total number of correct words produced summed across the three letters. Perseverations (i.e., repetitions of a correct word) and intrusions (i.e., words not beginning with the designated letter) are noted.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. NPI was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.

Study participants are given one minute to provide exemplars of the category 'animals'.

Full Analysis Set (all randomized subjects who receive study medication and have at least one post-baseline efficacy assessment) with Month 18 (A-Study) or Month 12 (B-Study) data. NPI was not reported / not valid for some subject visits; therefore, the number analyzed is smaller than overall number.