Association of Antioxidant Supplement Use and Dementia in the Prevention of Alzheimer’s Disease by Vitamin E and Selenium Trial (PREADViSE)

Key Points

Question  Can vitamin E or selenium prevent dementia in asymptomatic older men?

Findings  The Prevention of Alzheimer’s Disease by Vitamin E and Selenium trial initially enrolled 7540 elderly men who were exposed to the supplements for an average of 5.4 years; a subset of 3786 men agreed to be observed for up to 6 additional years. Dementia incidence (4.4%) did not differ among the 4 study arms.

Meaning  Neither supplement is recommended as a preventive agent for dementia.

Importance  Oxidative stress is an established dementia pathway, but it is unknown if the use of antioxidant supplements can prevent dementia.

Objective  To determine if antioxidant supplements (vitamin E or selenium) used alone or in combination can prevent dementia in asymptomatic older men.

Design, Setting, and Participants  The Prevention of Alzheimer’s Disease by Vitamin E and Selenium (PREADViSE) trial began as a double-blind randomized clinical trial in May 2002, which transformed into a cohort study from September 2009 to May 2015. The PREADViSE trial was ancillary to the Selenium and Vitamin E Cancer Prevention Trial (SELECT), a randomized clinical trial of the same antioxidant supplements for preventing prostate cancer, which closed in 2009 owing to findings from a futility analysis. The PREADViSE trial recruited 7540 men, of whom 3786 continued into the cohort study. Participants were at least 60 years old at study entry and were enrolled at 130 SELECT sites, and Cox proportional hazards models were used in a modified intent-to-treat analysis to compare hazard rates among the study arms.

Interventions  Participants were randomized to vitamin E, selenium, vitamin E and selenium, or placebo. While taking study supplements, enrolled men visited their SELECT site and were evaluated for dementia using a 2-stage screen. During the cohort study, men were contacted by telephone and assessed using an enhanced 2-stage cognitive screen. In both phases, men were encouraged to visit their physician if the screen results indicated possible cognitive impairment.

Main Outcomes and Measures  Dementia case ascertainment relied on a consensus review of the cognitive screens and medical records for men with suspected dementia who visited their physician for an evaluation or by review of all available information, including a functional assessment screen.

Results  The mean (SD) baseline age of the 7540 participants was 67.5 (5.3) years, with 3936 (52.2%) reporting a college education or better, 754 (10.0%) reporting black race, and 505 (6.7%) reporting Hispanic ethnicity. Dementia incidence (325 of 7338 men [4.4%]) was not different among the 4 study arms. A Cox model, which adjusted incidence for participant demographic information and baseline self-reported comorbidities, yielded hazard ratios of 0.88 (95% CI, 0.64-1.20) for vitamin E, 0.83 (0.60-1.13) for selenium, and 1.00 (0.75-1.35) for the combination compared with placebo.

Conclusions and Relevance  Neither supplement prevented dementia. To our knowledge, this is the first study to investigate the long-term association of antioxidant supplement use and dementia incidence among asymptomatic men.

In the United States, an estimated 5 million elderly persons have Alzheimer disease (AD), and the number of people who develop AD is expected to increase substantially by 2050.¹ A 2014 review² of clinical trials from 1984 to 2014 showed a focus on enrolling patients with mild to moderate dementia in many trials, with no real progress on identifying disease-modifying treatments. As a result, there has been a shift in focus to clinical trials emphasizing the prevention of cognitive decline, especially through secondary prevention trials targeting high-risk groups^3,4 and large trials that promote lifestyle changes to address modifiable risk factors for AD.^5,6 The usual primary end points of these trials are cognitive decline or composites of biomarkers and cognitive measures.⁷ The gold standard of prevention is disease incidence, but to our knowledge, few current trials have this as their primary end point because of the time required to observe reductions in disease incidence.

Multiple mechanisms associated with disease onset and progression have been described,⁸ and one key mechanism implicated in AD is oxidative stress,⁹ which may be modifiable through diet and/or antioxidant supplements.¹⁰ Antioxidant use as a potential treatment for cognitive impairment or dementia has been of interest for many years. The use of vitamin E has had mixed results in treatment trials; in patients with moderate dementia treated for 2 years, its use slowed disease progression,¹¹ but more recently, when used in an antioxidant cocktail, its use failed to improve cognition in patients with AD with mild to moderate dementia.¹² It also failed as a preventive agent for dementia progression in persons with mild cognitive impairment,¹³ although early in the trial, its use did somewhat improve cognition. A review¹⁴ of controlled trials and case-control studies on the use of selenium in arresting the progression of AD also yielded mixed results. Observational studies correlate cognitive decline with decreased plasma selenium over time.^15,16 However, to our knowledge, nothing is known about long-term use of these supplements as preventive agents in asymptomatic individuals.

This article reports the main results of the Prevention of Alzheimer’s Disease by Vitamin E and Selenium (PREADViSE) primary prevention trial. The PREADViSE trial, the largest primary prevention trial in AD to date, began in 2002 as an ancillary trial within the Selenium and Vitamin E Cancer Prevention Trial (SELECT), a double-blind randomized clinical prostate cancer prevention trial.¹⁷ When SELECT ended prematurely in 2009 owing to results of a futility analysis, the PREADViSE trial continued as a cohort study for a subset of its enrollees.¹⁸ Extended follow-up over a 7-year period (blinded) allowed for case ascertainment, yielding a comparison of the study arms for effectiveness in preventing dementia. We present the results of this large primary prevention study of antioxidant supplements as a method to modify oxidative stress, one mechanism in the evolution of AD.⁹

Study design,¹⁹ participant recruitment, and the conversion of the trial into a cohort study¹⁸ are described in earlier publications. Briefly, the parent study, SELECT, was designed as a double-blind, 4-arm randomized clinical trial (RCT) and initiated enrollment in 2001.¹⁷ The primary aim of SELECT was to determine the effectiveness of antioxidant supplements vitamin E (400 IU/d) and selenium (200 µg/d) alone or in combination in preventing prostate cancer. The PREADViSE trial recruited a subsample of SELECT participants 60 years and older at 130 participating clinical sites in the United States, Canada, and Puerto Rico. The PREADViSE trial enrolled 7540 men in SELECT without dementia between 2002 and 2008. The trial protocol can be found in the Supplement. The RCT was powered to detect a hazard rate of 0.60 with 80% power, assuming a targeted enrollment of 10 400 men; this was then lowered to a detectable hazard rate of 0.50 once the actual enrollment finished at 7540 men.

Study eligibility required SELECT enrollment at a participating site as well as the absence of dementia, active neurologic and neuropsychiatric conditions that affect cognition, or a history of serious head injury (>30-minute loss of consciousness within the last 5 years prior to enrollment) and substance abuse. The RCT was scheduled to continue supplements until 2012, but in September 2008, the SELECT Data and Safety Monitoring Committee recommended that supplements be discontinued owing to lack of efficacy on prostate cancer incidence. Study sites closed over the next 2 years, during which time both the PREADViSE trial and SELECT transitioned into cohort studies.^18,20 Participants were asked to continue in the cohort study, and 4271 of the original 7540 men (56.6%) enrolled in the PREADViSE trial consented to continue annual memory screenings, which were then conducted by telephone. All study activities were approved by the University of Kentucky Institutional Review Board and by the institutional review boards at each SELECT site. All participants provided written informed consent.

The Memory Impairment Screen (MIS)²¹ was used as the primary screening instrument in both the RCT and cohort study. Participants who scored below cutoffs for intact cognition received a secondary screening instrument. The modified Telephone Interview for Cognitive Status (TICS-m)²² was used during the cohort study, replacing the expanded Consortium to Establish a Registry in Alzheimer Disease (CERAD) neuropsychological battery²³ used during the RCT. In addition, subgroups of participants who did not fail the MIS also completed the CERAD battery and the TICS-m for assessing diagnostic accuracy.²⁴ Annual screenings were completed in May 2014, and case ascertainment follow-up continued until Fall 2015. All participants in the PREADViSE trial who completed at least 1 follow-up visit were included in the current analyses, whether they participated in just the RCT or both the RCT and cohort studies.

Dementia incidence, the primary end point, was determined by 1 of 2 methods. First, if participants failed both the first tier of the screen (MIS score ≤ 5 of 8) and the second tier (T Score ≤ 35 on the CERAD battery; total score ≤35 on the TICS-m), then they were encouraged to obtain a memory workup from their local clinician and share medical records with PREADViSE trial investigators. Medical records were reviewed by a team of 2 to 3 expert neurologists and 2 to 3 expert neuropsychologists to determine consensus diagnoses. Participants who did not obtain the workup were assessed by additional longitudinal measures collected during the study. These included the Ascertain Dementia 8-Item Informant Questionnaire (AD8) Dementia Screening Interview,²⁵ self-reported medical history, self-reported medication use, and cognitive scores, including the MIS, CERAD T Score, New York University Paragraph Delayed Recall,²⁶ and TICS-m. An AD8 of 1 or greater (at any time during follow-up) as well as a self-reported dementia diagnosis, use of a memory-enhancing prescription drug (eg, donepezil, rivastigmine, galantamine, or memantine), or cognitive score 1.5 SDs or more below expected performance yielded a dementia diagnosis. The diagnosis date was assigned to the earliest event.

Groups were compared using χ² statistics for categorical variables and 2-sample t tests or analyses of variance for interval-level variables. Cox proportional hazards models were used in a modified intent-to-treat (mITT) analysis to compare hazard rates among the study arms. Hazard ratios (HRs) were adjusted for baseline age, black race, APOE ε4 carrier status (present or absent), college education, baseline MIS score, and the presence or absence of the following self-reported comorbidities at PREADViSE trial baseline: coronary artery bypass graft, congestive heart failure, diabetes, hypertension, stroke, sleep apnea, and memory change or problem. In this analysis, survival time represents the time from PREADViSE trial baseline to ascertainment of a case (event) or last cognitive assessment (right censored). All dropouts owing to death, personal withdrawals from the study, and administrative withdrawals (eg, SELECT sites not reconsenting men for the cohort study) were right censored. The proportional hazards assumption required by the Cox model was checked using martingale residuals and Schoenfeld residuals. The latter analysis indicated that proportionality was met for the indicators of the treatment arms and all covariates except for the indicators of baseline memory problems and baseline MIS score less than 7 or 8. A stratified analysis using the combination of these latter 2 variables yielded the same results for the effect of the treatment arms. Therefore, the original Cox model is reported here. Missing APOE ε4 carrier status values (352 of 7338 participants [4.8%]) were imputed; a reanalysis limited to known APOE ε4 carriers yielded virtually the same results as those reported here.

Three additional analyses were conducted. First, the observations were weighted by treatment compliance. This weight equaled the length of time the man could take the supplements (ie, time between enrollment and the date the supplements were stopped) multiplied by the proportion of visits at which the man was compliant with assigned supplement 1 (ie, vitamin E or placebo) or compliant with supplement 2 (ie, selenium or placebo), whichever was greater. Compliance information was based on returned pill counts. For example, suppose a randomly selected study participant had 3 years receiving treatment, and he was found to be 50% compliant with assigned treatment. The weight assigned to that man’s record was 1.5 (3 years × 50% compliance, normed to make the sum of all weights equal the sample size for this analysis, given the 7289 men with compliance information). Second, the Cox models were refitted using only medical records–based dementia as the outcome (in other words, patients identified without medical records were right censored).

Table 1 shows that the 3786 participants evaluated with at least 1 memory screen in the cohort study were similar to the complete PREADViSE trial enrollment, except possibly for less education at the college level or higher (52.2% vs 60%), fewer black participants (8.4% vs 10.0%), and fewer Hispanic participants (2.5% vs 6.9%). Table 2 lists participant characteristics at the time of PREADViSE trial enrollment by study arm. Based on the means and percentages, despite randomization occurring at SELECT baseline rather than PREADViSE trial enrollment, there were no perceivable differences between study arms in terms of medical history, APOE ε4 genotype, or initial MIS score.

Incident dementia cases were defined as above. There were 325 of 7540 men (4.3%) with incident dementia in the study, and of these, 121 of 325 men (37.2%) provided medical records (Table 3). Unadjusted cumulative dementia incidence varied among the study arms; there were 71 men (4.0%) in the vitamin E arm, 78 men (4.2%) in the selenium arm, 85 men (4.6%) in the placebo arm, and 91 (5.0%) in the combination arm (Table 3).

To compare the hazard rates for dementia among the 4 study arms, mITT analysis based on a Cox proportional hazards model was fitted to the data. The mITT analysis excluded 201 participants who had only a baseline visit. Hence, the Cox model included 7338 men with at least 1 follow-up visit. None of the active study arms had a significantly lower adjusted hazard rate for incident dementia compared with the placebo arm; the vitamin E arm had an HR of 0.88 (95% CI, 0.64-1.20), while the selenium arm had an HR of 0.83 vs placebo (95% CI, 0.61-1.13) (Table 4). The combined supplements arm was indistinguishable from the placebo arm and had an HR of 1.00 (95% CI, 0.74-1.35) (Figure).

Because this was a truncated RCT (owing to the futility analysis for SELECT), the Cox model was refitted by weighting each man’s survival time to dementia by the length of time that man was exposed (time receiving supplement) multiplied by the estimated proportion of this time that the man was compliant with the assigned treatment (ie, compliance). The weighted analysis showed results similar to the unweighted mITT analysis (Table 4).

Additional sensitivity analyses were conducted. First, the analysis was restricted to the 3786 participants who volunteered for the cohort study and were screened by telephone at least once; here, the number of incident dementia cases was 228. The reported effects were an HR of 0.97 (95% CI, 0.65-1.43) for the vitamin E arm, an HR of 0.92 (95% CI, 0.63-1.34) for the selenium arm, and an HR of 1.18 (95% CI, 0.82-1.68) for both. Second, the analysis used all 7338 participants with at least 1 annual follow-up screen but only classified a participant as having dementia if confirmed by medical records. This reduced the number of events to 121, yielding an HR of 0.80 (95% CI, 0.59-1.62) for vitamin E, an HR of 0.70 (95% CI, 0.41-1.19) for selenium, and an HR of 0.97 (95% CI, 0.60-1.58) for both.

The PREADViSE trial was a double-blind RCT conducted as an ancillary study to a cancer prevention trial (SELECT), both of which evolved into observational cohort studies. This trial investigated whether the supplements vitamin E and selenium used alone or in combination would prevent new AD or dementia onset. The results showed that neither vitamin E nor selenium (with a mean [SD] time of supplement use of 5.4 [1.2] years) had a significant preventive effect on incidence. One possible explanation for the negative findings is that the trial met only 75% of its planned accrual; however, the results also show that the effect sizes observed for either supplement are likely much lower than the projected HR of 0.60.¹⁹ Nevertheless, to our knowledge, this is the first large-scale primary prevention trial to investigate the association of antioxidant supplements with reducing dementia incidence.

The PREADViSE trial is a member of the first generation of AD prevention trials,^27-30 all of which failed in their primary goal. One common reason for the failure of those trials was the low incidence of AD and dementia observed during follow-up. This was attributed partially to selection bias, since participants had higher levels of education than the general population and perhaps more cognitive reserve. Dementia incidence among PREADViSE trial participants was very low for several additional reasons. Case ascertainment could not use modern diagnostic procedures because those with suspected dementia did not, as a rule, visit dementia specialists for diagnosis. Instead, ascertainment relied on a consensus review of medical records derived from a variety of care professionals. Case ascertainment was also affected by continual staff turnover at study sites in the RCT component of the study, which was associated with a low cognitive screen failure rate (less than 1%). The failure rate increased substantially once centralized follow-up began.¹⁸

The rationale for this trial was supported by results from the basic sciences as well as observational and prospective studies in humans suggesting that the use of antioxidants improved cognition and reduced dementia incidence. The studies preceding the trial focused more on the benefits of vitamin E because it was being evaluated in a number of therapeutic trials at that time, including a trial that studied its effect on the progression of mild cognitive impairment to dementia.¹³ As in the mild cognitive impairment trial, vitamin E had no effect on dementia incidence in the asymptomatic cohort in the PREADViSE trial.

A 2014 review of studies on selenium³¹ argues that selenium is vital for central nervous system function and that brain selenium levels are maintained at the expense of other tissues. A study¹⁶ of 2000 rural Chinese individuals 64 years and older showed that participants in the lowest quintile of nail selenium levels had significantly lower scores on all instruments in the CERAD battery except for animal fluency. It is well known that high levels of selenium are toxic, inducing a severe selenosis by pro-oxidant action and glial activation, leading to neuronal death.³¹

For the PREADViSE trial, dose decisions were made by a committee in the parent SELECT study.³² The absence of biomarkers for target engagement of the supplements makes it difficult to translate basic science findings into prevention trials in a rigorous manner.

Data monitoring in SELECT showed that selenium appeared to elevate levels of type 2 diabetes, although this elevated rate subsequently decreased with additional follow-up, and that vitamin E appeared to increase prostate cancer incidence.¹⁷The supplements had no effect on mortality, other cancers, cardiovascular events, nausea, fatigue, or nail changes. Selenium was associated with a significant increase in alopecia and grades 1 and 2 dermatitis¹⁷ (Table 4).

Our study had some limitations. The transitioning process cost the study about half of its participants for long-term follow-up; this was unavoidable because the study had to rely on SELECT sites to enroll the participants for the cohort study. Fortunately, Table 1 shows that the major cost was sample size (reduction in person-years of follow-up), as no new large selection biases were introduced. Aside from transitioning to a cohort study from an RCT, the PREADViSE trial had other limitations. Publicity about the negative effect of supplements may have affected the conduct of both the RCT and cohort study. The Selenium and Vitamin E Cancer Prevention Trial reports on the potentially harmful effects of vitamin E (increased prostate cancer) and selenium (potentially increased diabetes), coupled with outside negative reports on vitamin E (ie, increased mortality³³) created issues to be addressed.³⁴ Many individuals who failed the screening instruments refused to see clinicians for further testing; recent research suggests that this could be because of the stigma attached to such an event or living alone.³⁵ This complicated case ascertainment. Diagnoses of AD that rely on the choice of cut points on neuropsychological tests have been shown to fail to identify individuals in the initial symptomatic stages of the disease.³⁶ Hence, our methods for case ascertainment could have missed individuals that would have been identified by more rigorous in-clinic examinations. Future trials may benefit from electronic medical records. The reluctance to visit physicians prompted the introduction of the AD8 during the follow-up phase of the study, which could have induced bias because of the inability of participants or informants to distinguish anosognosia from dementia. This is a likely a small bias because examining those with medical records showed a high level of agreement with the AD8, as also evidenced in another study.³⁷ The reliability of telephone assessments vs in-clinic assessments used in the 2 phases of the study was less of an issue owing to high intraclass correlations, which ranged between 0.92 and 0.98.³⁸ In addition, the TICS-m has been shown to be a reliable discriminator between cognitively intact individuals and those with amnestic mild cognitive impairment.³⁹

The supplemental use of vitamin E and selenium did not forestall dementia and are not recommended as preventive agents. This conclusion is tempered by the underpowered study, inclusion of only men, a short supplement exposure time, dosage considerations, and methodologic limitations in relying on real-world reporting of incident cases.

Corresponding Author: Richard J. Kryscio, PhD, Sanders Brown Center on Aging, University of Kentucky, 800 S Limestone St, Lexington, KY 40536 (kryscio@email.uky.edu).

Accepted for Publication: March 20, 2017.

Published Online: March 20, 2017. doi:10.1001/jamaneurol.2016.5778

Author Contributions: Drs Kryscio and Abner had full access to the data and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Kryscio, Abner, Crowley, Schmitt.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Kryscio, Schmitt.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kryscio, Abner.

Obtained funding: Kryscio, Schmitt.

Administrative, technical, or material support: Caban-Holt, Lovell, Goodman, Darke, Yee, Crowley.

Supervision: Kryscio, Crowley, Schmitt.

Conflict of Interest Disclosures: None reported.

Funding/Support: Additional support for the current study comes from grants R01 AG038651 and P30 AG028383 from the National Institute on Aging. The Selenium and Vitamin E Cancer Prevention Trial was supported by grants CA37429 and UM1 CA182883 from the National Cancer Institute.

Role of the Funder/Sponsor: The National Cancer Institute was involved in the design of Selenium and Vitamin E Cancer Prevention Trial. Otherwise, the sponsors had no role in the design and conduct of the current study; in the collection, analysis, and interpretation of data; in the preparation of the manuscript; or in the review or approval of the manuscript.

Additional Contributions: We gratefully acknowledge all Prevention of Alzheimer’s Disease by Vitamin E and Selenium trial volunteers for their many years of participation and the support staff of Prevention of Alzheimer’s Disease by Vitamin E and Selenium at the University of Kentucky as well as the SWOG Statistical Center at Cancer Research and Biostatistics for assistance with study procedures and data management. We also acknowledge Gregory Jicha, MD, and John Ranseen, PhD (University of Kentucky, Lexington), as well as Gregory Cooper, MD (Baptist Health Medical Group Neurology, Lexington, Kentucky), for reviewing medical records on suspected cases; Steven Estus, PhD, and Donna Wilcock, PhD (Department of Physiology, University of Kentucky, Lexington), for APOE genotyping; and Marta Mendiondo, PhD (retired), for expertise in data management. We thank Herman Buschke, MD (Department of Neurology and Neuroscience, Albert Einstein College of Medicine, New York, New York), for recommending the Memory Impairment Screen as the first-stage screening instrument. The project also acknowledges the late William Markesbery, MD, the original principal investigator of the Prevention of Alzheimer’s Disease by Vitamin E and Selenium trial. All contributors except Dr Buschke were compensated for their work.