This Phase II randomized clinical trial aims to test if supplementation with high dose oral vitamin D will successfully correct vitamin D insufficiency, compared to treatment with standard (RDA) dose vitamin D in a diverse community-based elderly cohort. The effect of high-dose vs. standard-dose vitamin D on altering cognitive trajectories will also be assessed and data will be expected to be used in designing a potential definitive Phase III trial in elderly groups at risk for dementia. A total of 180 elderly persons with longitudinal biomarkers, neuropsychological testing and brain MRI scans will be enrolled, with 152 (~50 with MCI, 50 with mild AD and 50 with no cognitive impairment) expected to complete the 3½-year study. One-half of each diagnostic group will be randomized to treatment with high-dose vitamin D3 (4,000 IU daily) or to standard dose Vitamin D (600 IU capsule daily + ~200 IU dietary = ~800 IU total/day). Longitudinal MRI analyses will provide an estimate of the treatment effect size on brain atrophy rate. Vitamin D receptor genotype polymorphisms and their impact on response to oral supplementation will also be examined. If vitamin D supplementation improves cognitive outcome, this could have a large impact on the public health, since low vitamin D status is a common, readably treatable condition which may provide a novel window to prevent dementia and AD. Furthermore, the higher prevalence of AD and dementia in African Americans and Latinos could be partially attributable to vitamin D insufficiency.

1. General Design This Phase 2 randomized clinical trial of high-dose vs. standard dose Vitamin D (VitD) supplementation aims to prove feasibility of the intervention in a diverse community-recruited cohort and to provide an estimate of the effect sizes of treatment on cognitive change and brain MRI volumetric measures as the main outcomes. We will study 180 elderly participants (approximately 60 with MCI, 60 with mild AD and 60 with No Cognitive Impairment) with longitudinal neuropsychological testing and brain MRI scans over a 3½ year study period. One-half of each diagnostic group will be randomized to treatment with high-dose (4,000 IU p.o. daily) vitamin D3 (n=90) or to standard dose vitamin D3 treatment (n=90 who will receive 600 IU vitamin D3 daily). The choice of the low-dose amount of vitamin D3 was determined by the goal to meet the Institute of Medicine (IOM) RDA recommendation of 800 IU/day for older adults (age >71y) from all sources, including both diet and supplements. The estimated average dietary intake of VitD by older adults is ~200 IU/day. Thus, the combined intake of VitD from diet and the low-dose supplement in our study will be approximately the RDA (i.e. 800 IU/day). Blood will be collected every 6 months (except month 30) for VitD levels along with other clinical labs.

   This study aims to test if supplementation with high dose oral VitD (4,000 IU) will successfully correct VitD insufficiency, compared to treatment with standard RDA dose VitD (~800 IU total intake, as recommended by the IOM for those >age 71) in a diverse community-based cohort with serum VitD levels <20 ng/ml at study entry. The primary aims are to prove feasibility of the intervention, and the effectiveness of high-dose VitD in correcting VitD insufficiency in this diverse old cohort. Our primary outcome will be correction of VitD insufficiency in all subjects treated with 4,000 IU daily. Secondary aims are to provide an estimate of the effect sizes of treatment on cognitive change (executive function and global), and to gather preliminary data relevant to the evidence for moving forward with a potential definitive Phase III study in elderly groups at risk for dementia. Additional analyses will test if correction of VitD insufficiency correlates with changes in key biomarkers measured in blood and urine. Longitudinal MRI analyses will provide an estimate of the treatment effect size on brain atrophy rate. We will also examine VitD receptor genotype polymorphisms and their impact on response to oral supplementation.

2. Schedule of Visits Screening Visit (Visit 1). Elderly participants with a prior diagnosis of either No Cognitive Impairment, MCI, or mild AD, will be screened for potential study entry. Informed consent will be obtained before any study procedures are conducted. The following assessments will be done at screen (before Visit 2): MOCA, GDS, CDR, screening labs, then brain MRI for those who remain eligible. Blood and urine screening lab results will be reviewed by the investigator or designee for assessment of eligibility. Patients who do not meet all inclusion criteria or who meet any exclusion criteria will be discontinued from the study. Up to 45 days are allowed for completion of Visit 1 screening procedures, assessments, and evaluation of results from laboratory tests and Brain MRI.

   Baseline Visit (Visit 2). The treatment period is a double-blind phase beginning at Visit 2, with treatment duration of 3½ years from baseline to the final 42 month visit. Patients who meet entry criteria will be enrolled and randomized to high-dose VitD 4,000 IU p.o. daily or standard dose VitD (600 IU daily). Randomization will be stratified by diagnosis and race, such that 50% of each diagnostic subgroup will be randomized to the high-dose treatment arm.

   Assessments to be performed at the Baseline visits are: additional neuropsychological tests including the SENAS and ADAS-Cog, the ECog which has sections completed by the participant and the informant. Blood samples drawn at the Screening or Baseline visit will be sent for measurements of parathyroid hormone (PTH) level and genotyping to include polymorphisms of the VitD receptor for all enrolled subjects.

   Telephone Check-Ins at Months 1 and 30. Brief telephone calls will be made to all enrolled participants at Month 1 & Month 30 to check on compliance with the study medication.

   Treatment Period Visits at Months 6 and 18 (Visits 3 & 5). Brief visits will be performed at Month 6 and Month 18, with venipuncture for blood VitD level and the ADAS-Cog. In addition, pill counts and study drug compliance will be reviewed and documented.

   Treatment Period Visits at Months 12, 24, 36 and 42 (Visits 4, 6, 8 & 9). These visits will provide the main outcome measures for the trial. Assessments to be performed at these visits are: Neuropsychological testing with the SENAS, MOCA and the ADAS-Cog. Functional assessment instruments (collected from the participant and informant) are the ECog and CDR. The GDS will also be collected. Blood samples will be drawn for measurement of VitD level, and both plasma and serum will be stored for additional studies. The final assessment visit at Month 42 (Visit 9) will take place on the day of the patient's last VitD dose. A volumetric brain MRI study will also be obtained at this visit.

   Early Termination (ET) Visit (if applicable). For participants who do not complete this 42 month study, an early termination visit will be done whenever possible. For the early termination visits, we plan to obtain all the measures usually collected at Month 42.

3. Protocol Procedures. Medical History, Physical, and Neurological Exam: A complete medical history with review of systems, and recent medication history will be performed at screening. All study patients will receive a comprehensive physical examination at screening and annually. A detailed neurological exam will be performed either at screening or baseline. Any deficits identified will be documented using a standardized form. The physician also takes a functional history, and a psychiatric/behavioral history. Participants will be closely monitored during the treatment period (from month 1 through 42) to assess side effects or adverse events.

   Blood draws and Laboratory Tests. Blood samples will be collected into serum separator tubes (to collect serum) containing EDTA. One EDTA tube will be sent directly to the UC Davis Medical Center (UCDMC) clinical lab for a complete blood count and measurement of hemoglobin A1C. A second EDTA tube will be centrifuged at 40C to collect plasma. The buffy coat of the second tube will be collected and frozen for subsequent isolation of genomic DNA. Serum and plasma samples will be aliquoted in volumes specific for each assay, and frozen at -800C until analysis.

   Laboratory Assays VitD (25-OHD) level will be performed at screen and every 6 months after (serum/blood, except for the month 30 Telephone check-in visit). Clinical Labs (screening): Comprehensive Metabolic Panel (CMP), CBC, UA, vitamin B12 at screen. PT/PTT will be drawn for any participant with history of a coagulopathy or excessive bleeding. If none of the above screening labs are exclusionary, we will add homocysteine, PTH, high-sensitivity CRP, Interleukin-6 (IL-6), IL-10, TNF-alpha, Monocyte chemoattractant protein-1 (MCP-1), blood and urine Isoprostane F 2-α levels. Inflammatory markers: annual CRP blood levels and cytokine panel (IL-6, IL-10, TNF-alpha, and MCP-1) will be measured at Baseline and month 42 (or ET visit). Oxidative damage markers: Isoprostane F 2-α levels in blood and urine will be measured at Baseline and month 42 (or ET visit). Genetic analysis of the VitD receptor will be carried out for SNPs previously associated with cognitive decline or AD: Apa1, Taq1, CdX, and Bsml. For safety purposes, the following Labs will be done as needed: Urine & serum Calcium PRN renal stones, cultures (e.g. urine, blood) and Microbiology tests if with high wbc count. We will repeat CMP annually (months 12, 24, 36 & 42), and PTH level at month 42 (or ET visit).

   Neuropsychological Testing. All subjects will undergo neuropsychological testing with the MOCA at screening and at the 12, 24, 36 and 42 month visits. The Spanish English Neuropsychological Assessment Scale (SENAS) will be administered at baseline and at the 12, 24, 36 and 42 month office visits. The ADAS-Cog will be administered at baseline and every 6 months (except month 30).

   Functional Assessments. All subjects will undergo functional assessments with the ECog (informant based measure of daily function) and CDR at screen/baseline and at months 12, 24, 36, and 42.

   Dietary Survey. Dietary data will be collected to assess overall intake of VitD from food sources. Participants or their caregivers will be interviewed using a validated VitD Questionnaire.

   Other Clinical Instruments/Assessments. All subjects will complete the Geriatric Depression Scale (GDS) at screening and at the 12, 24 & 42 month assessments.

   Brain MRI Procedure & Analysis. Structural Brain MRI will be acquired at Screening and at study end (42 months, or at time of ET visit) for all enrolled participants. Subjects will be scanned at the UC Davis Imaging Research Center using a 3 Tesla Siemens Tim Trio machine and ADNI-compatible pulse sequences.

   MRI acquisition and analysis. MRI acquisition will include T1-weighted, fast T2, fluid-attenuated inversion recovery (FLAIR), and diffusion tensor imaging (DTI). MRI data will be analyzed for quantification of regional brain, gray, white and CSF volumes. All subjects will have baseline and longitudinal quantification of regional WMH, DTI and gray matter measures.

4. Statistical Analysis Plan (SAP), Hypothesis Testing & Statistical Considerations.

   A brief description of hypothesis testing strategies and power calculations are provided separately by aim below. All power calculations assumed a two-sided test and alpha=0.05 unless otherwise specified.

   Aim 1: To compare the effectiveness of high-dose versus standard-dose oral VitD in correcting low VitD status in a diverse community-based cohort.

   Hypothesis 1.1: Rate of correction of VitD insufficiency is higher after high-dose versus standard dose oral VitD supplementation.

   Analytic Plan. The primary analyses for this aim will involve two parts: 1) a descriptive summary of the percentage of individuals randomized to each group whose VitD levels reach and continue in the adequate range (>20 ng/ml); and 2) a comparison of the percentage of individuals whose levels reach the adequate range between the high-dose and standard dose groups. We will compute the percentage of individuals who attain adequate VitD levels by six months, the percentage who maintain these levels from months 6 through 42, and 95% confidence intervals. A stratified exact test of proportions will be used to compare percentages between groups. Secondary analyses will characterize the trajectory of VitD levels in the two groups over time using mixed effects models to assess patterns of change, and in particular, whether levels plateau or continue to increase.

   Power calculations. With 76-90 individuals per group, we will have over 80% power to detect a difference in percentage that reach or maintain adequate levels between groups assuming the percentage in the high-dose group is at least 90% and the percentage in the standard-dose group is no more than 72%.

   Aim 2: To assess the effect of high-dose vs. standard-dose VitD on altering cognitive trajectories, and gather preliminary data relevant to the design of a potential Phase III trial in elderly at-risk groups.

   Hypothesis 2.1: High dose VitD supplementation is associated with less cognitive decline (better cognitive trajectory) in executive function over 3 ½ years than in those receiving the standard dose.

   Rationale. Published data from our observational study showed a significant difference in rates of decline between those with insufficient or deficient serum VitD levels and those whose levels are in the adequate range. If high-dose VitD supplementation results in a slowing of cognitive decline, similar to that seen between the decline in the inadequate or deficient group and the adequate group, in addition to evidence of effectiveness (Aim 1) and biological mechanism (Aim 3), such a result would support moving on to a Phase III clinical trial for a definitive demonstration of clinical efficacy of high-dose VitD.

   Analytic Plan. The primary outcome for this aim is the SENAS executive function score, assessed annually over the course of the trial. A secondary outcome is the ADAS-Cog (global cognition). The analytic strategy is the same for each outcome and will utilize an intent-to-treat approach in which everyone randomized to the high dose group will be compared to those randomized to the standard dose regardless of adherence to treatment; secondary as-treated analyses will utilize continuous serum VitD levels as a measure of exposure to VitD to assess a "dose-response" relationship. Mixed effects regression models, similar to those described above for the secondary analyses for VitD trajectories will be used to assess differences in rates of decline. Models will further include age, education, ethnicity, ApoE4 status, BMI, and season of blood draw as covariates. Secondary analyses will compare those with adequate serum VitD levels by six months (who sustain levels throughout the treatment period) to those with continued VitD insufficiency. Further secondary analyses will utilize the continuous serum VitD levels as time-varying variables in the models to assess how levels over time are associated with level and change in cognitive function.

   Power calculations. Because this Phase II trial will assess preliminary evidence of efficacy, the type I error rate (alpha) will be set at 0.05 and power will be based on a one-tailed test as is standard for Phase II trials to assess whether rates of decline are slower in the high-dose group than in the standard-dose group. Using estimates of between and within-person variability in change in executive function from preliminary data, with 76 individuals (completers) per group, we will have 80% power to detect a difference in the rate of executive function decline, based on the effects seen in our observational study. We allow for up to a 15.5% drop out rate with dropout uncorrelated to missing data.

   Aim 3 (exploratory): To compare the effects of high-dose VitD versus standard-dose VitD on key brain, blood and urine biomarkers relevant to healthy brain aging.

   Hypothesis 3.1: High dose VitD supplementation is associated with less brain atrophy (total gray + white matter volume) and less loss of hippocampal volume over 3 ½ years than in those on standard dose VitD.

   Hypothesis 3.2: Blood inflammatory markers are lower in persons on high-dose VitD supplementation compared to persons on standard dose VitD.

   Hypothesis 3.3: Blood and urine markers of oxidative stress are lower in persons on high-dose VitD supplementation compared to persons on standard dose VitD.

   Rationale. MRI, blood and urine markers provide information on the potential biological mechanisms of VitD supplementation, and in particular, correction of VitD insufficiency. These exploratory hypotheses will provide further evidence of the benefit of VitD supplementation in at risk elderly individuals.

   Analytic Plan The analysis strategy will be similar for each of these hypotheses. MRI and measures from blood (e.g. CRP, IL-6, IL-10, TNF-alpha), and urine (e.g. isoprostane) will be acquired at baseline and at month 42 (or the ET). The primary outcome will be the value from the final visit. Because this is a randomized trial, it is expected that the baseline means will be similar in the two groups, so that a primary comparison of the measure at the final visit provides an assessment of the impact of high- vs standard dose VitD supplementation. Linear regression, using the baseline measure as a covariate will be used. Other covariates, including age, and ApoE4 status will be included in the models. To account for differences in the timing of the final visit (since some may be terminal visits prior to the end of the trial), time since baseline will be included. Model assumptions of normality of the residuals, constant variance, and linearity will be checked and transformations or nonlinear models will be used if needed. Secondary analyses will compare those individuals who have sustained corrected VitD levels versus those who remain insufficient. Hypotheses will be supported if the average levels at the terminal visit are improved in the high-dose group relative to the standard dose group.

   Power calculations. With 152 individuals, we will have 80% power to detect a difference between groups as small as 0.46 standard deviations (SD). Using data from the Alzheimer's Disease Neuroimaging Initiative, brain atrophy and hippocampal atrophy over 3 ½ years between Normal and MCI individuals ranged from 0.53-0.69 SD. Therefore, we will have sufficient power to detect a slowing in atrophy rates comparable to an improvement of the MCI subjects to the rate of cognitively normal individuals through VitD supplementation.

   Additional exploratory analyses. Additional analyses will examine the effects of high-dose VitD on measures of mood (GDS), using the GDS as the outcome in mixed effects regression models. Because response to VitD supplementation may differ by VitD receptor genotype, the association between genotype and cognitive decline and brain atrophy will be explored by adding genotype to the models used for Aim 2 and Hypothesis 3.1. For planning of the Phase III clinical trial, differences in effects by diagnosis will be explored. We will biobank serum and plasma making these materials available for further analyses which could include metabolomics and further genetic characterization.

5. Safety analyses. The Data Safety Monitoring Board for this study will review Adverse Events and all safety data collected every six months, commencing with the first patient enrolled. The DSMB will be particularly concerned with falls and fractures, renal stones, hepatic/renal insufficiency, and achieving extremely high levels of VitD. Frequencies of these events will be recorded and compared between the high- and standard dose groups. Overall, we expect the tolerability of VitD supplementation to be excellent in this population, but will test if safety and aspects of health are superior/inferior in either of the treatment arms.

Inclusion Criteria:

• Inclusion criteria for this study are: 1) Age 65-90; 2) Low Vitamin D status, defined by a serum 25-OHD <20 ng/ml, as measured by competitive immunoassay at the Screening visit; 3) Fluent in either English or Spanish; 4) Willingness to be randomized to high-dose vs. standard dose vitamin D; 5) Education adjusted scores between 12-30 on the Montreal Cognitive Assessment (MOCA) at baseline; 6) A consensus clinical diagnosis of No Cognitive Impairment, Mild Cognitive Impairment, or mild AD dementia.
• In order to be included in the No Cognitive Impairment (NC) subgroup, an individual must show no significant cognitive impairment on the baseline neuropsychological tests. Diagnoses are made by a comprehensive case conference review, including investigators from both sites, resulting in a consensus diagnosis made according to current research criteria. We will also require a minimum MOCA score of 23 or above for those with education >12 y, or MOCA >20 (uncorrected score) for those with education <12 years. These subjects are expected to mostly have CDR global scores of 0, but we will not exclude CDR=0.5, as long as Peterson criteria (Petersen, Journal of Internal Medicine 2004) for MCI (amnestic or non-amnestic) are not met and the CDR-Sum of Boxes is <1.0. Similarly, we will not exclude elderly with subjective memory complaints from the NC group.
• In order to be included in the Mild Cognitive Impairment (MCI) subgroup, a participant will need to meet research criteria for amnestic MCI, either single-domain or multiple-domain (McKhann et al, Neurology 1984). Thus, participants with amnestic MCI will have standardized memory scores >1.5 SDs below average, and if cognitive scores in other cognitive domains are also >1.5 SDs below average they will be classified as multiple-domain amnestic MCI. All MCI participants will be required to have a global CDR=0.5. In addition, we will require a minimum MOCA score of >20 for those with education >12 y, or MOCA >17 (uncorrected score) for those with education <12 years.
• In order to be included in the Mild AD dementia subgroup, a participant will need to meet research criteria for probable or possible AD (McKhann et al, Neurology 1984). A global CDR score of 1 will be required (mild dementia). In addition, we will require a minimum MOCA score of >15 at entry for those with education >12 y, or MOCA >12 (uncorrected score) for those with education <12 y. AD therapies will be allowed (e.g. donepezil, memantine) as long as doses have been stable for >6 weeks, and no changes in doses or CNS active medications are planned while participating in this trial.

Exclusion Criteria:

• Exclusion criteria (for all participants) are: 1) Lacks adequate vision, hearing, or literacy to complete the required psychometric tests (e.g. severe bilateral deafness despite use of hearing aids, or visual acuity poorer than 20/100 in both eyes); 2) Hepatic insufficiency, defined by either albumin <3.3 g/dL or by a value >2X the upper limit of normal (ULN) in either alanine aminotransaminase (ALT/SGPT) or bilirubin, or >3X the ULN for aspartate aminotransaminase (AST/SGOT); 3) Renal insufficiency, defined by either serum creatinine >1.7 mg/dL or glomerular filtration rate <40 mL/min/1.73 m2; (calculated per CKD-EPI formula). 4) Hypercalcemia, defined by serum calcium level >2 standard deviations above the mean. Corrected Calcium mg/dL = [0.8*(4.0g/dL - Patients Albumin g/dL)] + Serum Calcium mg/dL (X of 8.6-10.5 is in normal range; formula accurate only if Albumin is in 3.2-4.6 range). 5) Current serious or unstable medical illnesses including cardiovascular (e.g. unstable ischemic cardiovascular disease), hepatic, renal, gastroenterologic, respiratory, endocrinologic, neurologic, psychiatric, immunologic, or hematologic disease and other conditions that, in the investigator's opinion, could interfere with the participant being able to safely take high-dose vitamin D for the 3.5 year study duration; or has a life expectancy of <4 years; 6) History of recurrent renal stones; 7) Unable to undergo MRI scanning (e.g. pacemaker, metallic implants, severe claustrophobia); 8) Subjects with a history of osteoporosis will be excluded if the Screening serum 25-OHD level is < 12 ng/ml. 9) History of chronic psychiatric illness (e.g. schizophrenia, bipolar disorder), any episode of major depression within last 2 years, or current GDS > 6, any recent suicide attempts or suicidal ideation; 10) history within the last 5 years of a serious infectious disease affecting the brain (including neurosyphilis, meningitis, or encephalitis), or head trauma resulting in protracted loss of consciousness (>10 minutes) or any persistent cognitive deficit; 11) History of chronic alcohol or drug abuse/dependence as defined by DSM-IV, within the past 5 years; 12) History within the last 5 years of a primary or recurrent malignant disease with the exceptions of resected cutaneous squamous cell carcinoma in situ, basal cell carcinoma, cervical carcinoma in situ, or non-metastatic prostate cancer with a normal prostate-specific antigen post-treatment; 13) Does not have good venous access, such that multiple blood draws would be precluded; 14) Regular use of any of these CNS active medications: benzodiazepines, antipsychotics, narcotics, cholinesterase inhibitors, memantine or anti-epileptic drugs. Stable doses of SSRIs or SNRI anti-depressants will be allowed, and included persons will be discouraged to change the doses of any potentially CNS active medication throughout the 3.5-year study. 15) Those who plan to change their dosage of any vitamin supplement during the duration of the study may be discontinued from the clinical trial. Changes in vitamin supplement dosing (e.g. vitamin B12) will only be allowed if a specific deficiency has been found. If participants are taking vitamin D supplementation (e.g. in a daily multi-vitamin) at the time of screening, they should continue on this same dose of vitamin supplementation throughout the duration of the randomized clinical trial. 16) Current participation in any clinical trial involving experimental AD therapies (anti-amyloid, anti-tau, etc.). 17) Female subjects who are pregnant or plan to become pregnant during participation in this trial. 18) Inability to swallow oral capsules.
• We will not exclude subjects with stable coronary artery disease or vascular risk factors such as diabetes or hypertension, who otherwise meet our inclusion/exclusion criteria. We will allow inclusion of participants with prior TIA or prior history of one stroke, but will exclude those with past history of multiple strokes. We will allow participants with up to two incidental infarcts on structural MRI, because the presence of cerebrovascular disease is very common in any representative sample of U.S. elderly persons, and may be an important contributor to dementia and age-associated cognitive decline. We will not exclude those with severe white matter (WM) hyperintensities, for similar reasons.

It is not yet known if there will be a plan to make IPD available

Participants receive identically appearing capsules which contain either 4,000 IU (high-dose) or 600 IU (standard-dose) of vitamin D3.