Effect of a 24-Month Physical Activity Intervention vs Health Education on Cognitive Outcomes in Sedentary Older Adults: The LIFE Randomized Trial

Importance  Epidemiological evidence suggests that physical activity benefits cognition, but results from randomized trials are limited and mixed.

Objective  To determine whether a 24-month physical activity program results in better cognitive function, lower risk of mild cognitive impairment (MCI) or dementia, or both, compared with a health education program.

Design, Setting, and Participants  A randomized clinical trial, the Lifestyle Interventions and Independence for Elders (LIFE) study, enrolled 1635 community-living participants at 8 US centers from February 2010 until December 2011. Participants were sedentary adults aged 70 to 89 years who were at risk for mobility disability but able to walk 400 m.

Interventions  A structured, moderate-intensity physical activity program (n = 818) that included walking, resistance training, and flexibility exercises or a health education program (n = 817) of educational workshops and upper-extremity stretching.

Main Outcomes and Measures  Prespecified secondary outcomes of the LIFE study included cognitive function measured by the Digit Symbol Coding (DSC) task subtest of the Wechsler Adult Intelligence Scale (score range: 0-133; higher scores indicate better function) and the revised Hopkins Verbal Learning Test (HVLT-R; 12-item word list recall task) assessed in 1476 participants (90.3%). Tertiary outcomes included global and executive cognitive function and incident MCI or dementia at 24 months.

Results  At 24 months, DSC task and HVLT-R scores (adjusted for clinic site, sex, and baseline values) were not different between groups. The mean DSC task scores were 46.26 points for the physical activity group vs 46.28 for the health education group (mean difference, −0.01 points [95% CI, −0.80 to 0.77 points], P = .97). The mean HVLT-R delayed recall scores were 7.22 for the physical activity group vs 7.25 for the health education group (mean difference, −0.03 words [95% CI, −0.29 to 0.24 words], P = .84). No differences for any other cognitive or composite measures were observed. Participants in the physical activity group who were 80 years or older (n = 307) and those with poorer baseline physical performance (n = 328) had better changes in executive function composite scores compared with the health education group (P = .01 for interaction for both comparisons). Incident MCI or dementia occurred in 98 participants (13.2%) in the physical activity group and 91 participants (12.1%) in the health education group (odds ratio, 1.08 [95% CI, 0.80 to 1.46]).

Conclusions and Relevance  Among sedentary older adults, a 24-month moderate-intensity physical activity program compared with a health education program did not result in improvements in global or domain-specific cognitive function.

Trial Registration  clinicaltrials.gov Identifier: NCT01072500

Epidemiological evidence suggests that physical activity is associated with lower rates of cognitive decline. Exercise is associated with improved cerebral blood flow and neuronal connectivity,¹ maintenance or improvement in brain volume,^2,3 and favorable changes in brain-derived neurotrophic factor and neurogenesis.^4,5 In transgenic Alzheimer mouse models, exercise reduces β-amyloid deposition.⁶

Randomized clinical trials (RCTs) assessing the effect of physical activity on cognitive function are equivocal,^7-9 perhaps due to small sample sizes, short intervention periods, and differences in cohorts and protocols, particularly intensity of physical activity.⁷ Two small RCTs of physical activity^10,11 found no benefit from a structured physical activity program vs no intervention or cognitive training in older adults without dementia, but with cognitive complaints or at risk for cognitive decline. However, a 6-month RCT of a home-based physical activity program vs usual care in participants with memory complaints or mild cognitive impairment (MCI) found a modest cognitive benefit.¹² The Lifestyle Interventions and Independence for Elders (LIFE) pilot study showed a correlation between changes in physical and cognitive performance during a 12-month exercise intervention.¹³

We report the prespecified secondary cognitive outcomes of the LIFE study, the largest and longest RCT to assess the effect of a standardized physical activity intervention on cognitive function and impairment in sedentary older adults at risk for mobility disability.¹⁴ We hypothesized that compared with health education, physical activity for 24 months would result in better cognitive function and lower risk of incident all-cause MCI or dementia.

The LIFE study was a single-blinded RCT of a physical activity intervention compared with a health education control conducted at 8 US field centers; participants were from rural and urban communities. Details of the LIFE study design and results have been published^14,15 and the trial protocol appears in the Supplement. The study included sedentary men and women aged 70 to 89 years who were at high risk for mobility disability based on objectively assessed lower-extremity functional limitations defined by a Short Physical Performance Battery¹⁶ score of 9 or less (of 12 points), but who could walk 400 m (without assistance) within 15 minutes at baseline.

Eligible participants had no diagnosis of dementia or significant cognitive impairment on the Modified Mini-Mental State Examination¹⁷ (3MSE) based on education- and race-specific norms. Participants with less than 9 years of education were excluded if the screening 3MSE score was less than 70 for black individuals and native Spanish speakers or less than 76 for English-speaking nonblack individuals. Participants with 9 or more years of education were excluded if their 3MSE score was less than 76 for black individuals and less than 80 for native Spanish speakers and English-speaking nonblack individuals. Race/ethnicity was self-reported and collected as required by the National Institutes of Health.

Recruitment was predominantly by mass mailing to age-eligible residents. Additional strategies included newspaper, radio, and television advertisements and presentations at health fairs, senior centers, medical clinics, and churches.

The LIFE study was approved by the institutional review boards at all 8 sites and monitored by a data and safety monitoring board appointed by the National Institute on Aging. Written informed consent was obtained from all participants.

Participants were randomly assigned using a secure web-based data management system (permuted block algorithm with random block lengths) with equal probability to either a physical activity intervention or a successful aging health education program, stratifying by field center and sex. The physical activity intervention focused on walking, strength, flexibility, and balance training. Participants were expected to attended 2 center-based visits per week and perform home-based activity 3 to 4 times per week. The physical activity sessions progressed toward a goal of 30 minutes of walking at moderate intensity, 10 minutes of primarily lower-extremity strength training with ankle weights, and 10 minutes of balance training and large muscle group flexibility exercises.

The health education group attended weekly health education workshops during the first 26 weeks of the intervention and at least monthly sessions thereafter. Sessions lasted 60 to 90 minutes and consisted of interactive and didactic presentations, facilitator demonstrations, guest speakers, or field trips. Sessions included approximately 10 minutes of group discussion and interaction and 5 to 10 minutes of upper-extremity stretching and flexibility exercises. Example topics included travel safety, age-appropriate preventive services, legal and financial issues, and nutrition. The intervention committee ensured that health education activities were consistent across sites and unlikely to increase physical activity.

Assessments were conducted every 6 months in person by staff masked to treatment group assignment. Home, telephone, and proxy assessments were attempted if participants could not attend clinic visits. Information on demographics, medical and hospitalization history, medication inventory, quality of well-being, and functional limitation was based on self-report.¹⁸ Usual physical activity was assessed by self-report using the Community Healthy Activities Model Program for Seniors questionnaire¹⁹ to measure total weekly minutes of walking and performing strength training exercises and objectively using an Actigraph accelerometer to measure total minutes of at least moderate activity (>760 counts/min) over 7 days.¹⁴

A previously described neuropsychological battery of tests was administered by trained and certified examiners at baseline and at 24 months after randomization.²⁰ Three computerized tasks were administered at baseline and at either 18 or 30 months, depending on when the participant was enrolled.²⁰

Cognitive tests at baseline included (1) the 3MSE,¹⁷ which is a 100-point test of global cognitive function, (2) the Digit Symbol Coding (DSC) task subtest of the Wechsler Adult Intelligence Scale Third Edition,²¹ which is a test of psychomotor speed, attention, and working memory, (3) the revised Hopkins Verbal Learning Test (HVLT-R),²² which is a 12-item word list learning and recall task, and (4) a modified version of the Rey-Osterrieth Complex Figure, which assesses visuospatial function (copy) and figural memory (immediate recall). At 24 months, these measures were repeated along with (1) the Boston Naming Test, which is a measure of language,²³ (2) the Trail Making Test²⁴ part A, which is a measure of attention, concentration, and psychomotor speed, and part B, which is a measure of executive function, and (3) the category fluency test for animals, which is a measure of executive function. In all tests, except parts A and B of the Trail Making Test, higher scores indicate better performance.

Using a laptop computer, participants were administered 3 tasks that were chosen for added sensitivity in assessing speed of processing and executive function: the n-back task,²⁵ the Eriksen Flanker task,²⁶ and a task switching exercise.²⁷

At baseline and 24 months after randomization, all participants were assigned 1 of the following cognitive classifications: no cognitive impairment, MCI, or dementia. Participants who scored 88 points or less on the 3MSE were sent for central adjudication by a panel (blinded to treatment assignment) of 8 clinical experts in the diagnosis of late-life cognitive impairment.²⁰ Each case was assigned to 2 independent adjudicators; disagreements were resolved by the full panel.

Adjudicators reviewed data from the neuropsychological battery, medical history, medications, discharge diagnoses for hospitalizations during the trial, Center for Epidemiology Studies-Depression scores,²⁸ self-reported disability, and informant-reported functional status (Functional Assessment Questionnaire; FAQ).²⁹ The FAQ is a 10-item interviewer-administered questionnaire assessing degree of dependence in cognitively challenging activities of daily living, such as preparing balanced meals, traveling outside the neighborhood, and managing finances. For all those who had a 3MSE score of 88 or less, the FAQ was administered either at baseline or at 24 months, or both, to the participant’s proxy. Based on 2011 criteria from the National Institute on Aging and the Alzheimer’s Association,^30,31 MCI and dementia were adjudicated.

The LIFE protocol specified DSC (total score) and HVLT-R (immediate and delayed recall subscales mean) as the 2 primary cognitive outcomes for assessing cognitive decline. Outcomes were tested according to the intention-to-treat principle with analysis of covariance using 24-month data and covariate adjustment for field center, sex, and the baseline value. Additional prespecified cognitive outcomes were based on scores from the computerized battery. Raw scores from this battery were first winsorized to limit the influence of extreme values; this was done by replacing scores less than the first percentile of the cohortwide distribution with the value of the first percentile and replacing scores greater than the 99th percentile with the 99th percentile value. The z scores were formed for each cognitive test score by dividing their difference from the baseline mean by the baseline standard deviation.

Composite scores for the HVLT-R (immediate and delayed recall scores), n-back (1- and 2-back scores), task switching (no switch and switch reaction times), and Flanker tasks (congruent and incongruent reaction times) were formed by averaging the z scores for their 2 individual components. The global cognitive function score was the average of scores from these composites and the z-transformed DSC, renormalized to have a mean of 0 and an SD of 1 at baseline. The executive function composite score was the renormalized average of scores from the n-back, task switching, and Flanker tasks. In creating these composite scores, averages were taken of all available data (ie, missing data if participants did not complete the full battery were ignored). Supporting analyses were conducted using multiple imputation for which missing measures and examination scores were imputed to create 5 databases that were analyzed in parallel.³²

Subgroup comparisons using interaction terms were prespecified for sex and baseline Short Physical Performance Battery score (<8 vs ≥8), 3MSE score (<90 vs ≥90), and age (70-79 years vs ≥80 years). Associations between changes in cognitive function and changes in objective and subjective physical activity level were assessed using linear regression and tests of interactions.

Progression in cognitive impairment (ie, from baseline normal cognitive function to either MCI or dementia or from baseline MCI to dementia) was a tertiary outcome. Logistic regression was used to compare progression rates between intervention groups. Participants with prevalent MCI (n = 141) at baseline were not included in the incidence of MCI, but were included in the incident dementia outcome if they progressed to dementia at 24 months. Seven participants were adjudicated to have dementia at the baseline visit (despite otherwise meeting LIFE study entry criteria). These participants were excluded from the incident dementia outcome analysis.

Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc). Two-sided inferences with P < .05 were considered statistically significant. The targeted sample size of 1600 was expected to provide 87% power to detect mean differences between groups of 0.15 SD for cognitive tests. This was projected to correspond to mean differences of 1.8 units for DSC scores and 0.8 units for HVLT-R immediate memory scores.

From February 2010 until December 2011, 1635 participants were randomized (818 to physical activity group and 817 to health education group; Figure 1). Analyses are limited to 1476 participants (90.3%) with cognitive data during follow-up. Compared with participants without cognitive follow-up data, the participants included in the analyses had faster gait speeds (P < .001). The 24-month retention rates were 89.8% for the physical activity group and 90.7% for the health education group (P = .56).

Characteristics of the participants appear in Table 1. The mean (SD) age was 78.9 (5.2) years, 68% were women, and 67% had a college education. The mean (SD) 3MSE score was 91.7 (5.4) (range, 71-100). There were more black participants in the physical activity group vs the health education group.

Based on accelerometry data, the physical activity group maintained moderate to vigorous physical activity levels between baseline and 24-month follow-up (mean difference, −2.1 min/wk [95% CI, −9.7 to 13.9 min/wk]) compared with the health education group (mean difference, −40.4 min/wk [95% CI, −29.4 to −51.4 min/wk]; P < .001). Based on data from the Community Healthy Activities Model Program for Seniors questionnaire, the physical activity group had a greater increase in self-reported physical activity level from baseline to 24 months (mean difference, 130.4 min/wk [95% CI, 116.7 to 144.1 min/wk]) compared with the health education group (mean difference, 30.5 min/wk [95% CI, 18.9 to 42.1 min/wk]; P < .001). The median attendance at physical activity sessions was 71%, excluding medical leave.

At baseline, interviewer-administered cognitive assessments were collected on all participants. Computer-based assessments were collected on 85.5% (2-back task) to 96.2% (Flanker task) of participants. There were no differences between groups on any cognitive tests at baseline.

Table 2 presents the raw scores and z-transformed cognitive outcomes, adjusting for clinic site, sex, and baseline values. The z scores are interpreted as the change from baseline in standard deviations. The adjusted mean raw DSC task scores (score range, 0-133) at follow-up were not different between the 2 groups (46.26 points [95% CI, 45.75 to 46.82 points] in the physical activity group vs 46.28 points [95% CI, 45.72 to 46.83 points] in the health education group; mean difference, −0.01 points [95% CI, −0.80 to 0.77 points]; P = .97). Similarly, the adjusted scores for mean HVLT-R delayed word recall (score range, 0-12) were not different between groups (7.22 words [95% CI, 7.03 to 7.41 words] for the physical activity group vs 7.25 words [95% CI, 7.06 to 7.44 words] for the health education group; mean difference, −0.03 words [95% CI, −0.29 to 0.24 words]; P = .84). There were no between-group differences in the executive function composite z score (P = .59) or the mean global composite z score (P = .40). Additional adjustment for race/ethnicity and education did not change the results.

The results of prespecified subgroup comparisons appear in Figure 2. Intervention effects did not vary by sex or baseline 3MSE score. However, for participants with a baseline Short Physical Performance Battery score of less than 8 or age of 80 years or older, there was heterogeneity in the intervention effects for the executive function composite, suggesting benefit in executive function associated with physical activity (P = .01 for interaction).

The 24-month changes in the 4 cognitive function measures were not correlated with changes in moderate physical activity as measured by accelerometry (P > .30) among the 697 participants with 24-month data. The 24-month changes in weekly walking and strength training from the Community Healthy Activities Model Program for Seniors questionnaire were modestly associated with global cognitive function (r = 0.07; P = .006) and executive function (r = 0.06; P = .04). These relationships were not different between the 2 groups (P > .70 for interaction). Results were unchanged when using 12-month change in level of physical activity.

There was no significant difference between groups in the incidence of MCI, dementia, or both combined; 13.2% of the physical activity group developed MCI or dementia by 24 months compared with 12.1% of the health education group (unadjusted odds ratio, 1.08 [95% CI, 0.80-1.46]; P = .61; Table 3). There were no between-group differences within MCI subtypes (incident amnestic MCI was 5.5% for the physical activity group vs 5.7% for the health education group [P = .85] and nonamnestic MCI was 4.6% and 3.2%, respectively [P = .16]).

The LIFE study’s structured, 24-month moderate-intensity physical activity intervention did not result in better global or domain-specific cognition compared with a health education program in older, sedentary adults. There was also no difference between groups in the incidence of MCI or dementia, although this was an exploratory outcome with limited statistical power. However, participants in the physical activity group who were 80 years or older and those with lower baseline physical functioning levels experienced benefits in executive functioning compared with participants in the health education group. Cognitive function remained stable over 2 years for all participants. We cannot rule out that both interventions were successful at maintaining cognitive function.

Despite epidemiological evidence supporting the benefits of exercise and physical activity on cognition, the results of the LIFE study are consistent with some other randomized trials.⁷ In the Mental Activity and eXercise trial,¹⁰ a structured aerobic physical activity intervention was not superior to a stretching exercise control or mental activity control in sedentary older adults. The Action for Health in Diabetes trial³³ found no benefit of diet plus physical activity on cognitive function over 8 years. A large trial of a multifactorial intervention including diet, physical activity, cognitive training, social activity, and management of metabolic and vascular risk factors showed a small, statistically significant benefit on global and executive cognitive function at 2 years.³⁴ However, it is difficult to compare this trial with the LIFE study because the population was 10 years younger, physically active at baseline, and had a multifactorial intervention.

Possible explanations for the lack of cognitive benefit of the physical activity intervention include (1) the assigned level of physical activity may have been insufficient to produce changes in the cognitive measures despite its effect on physical function¹⁴; (2) improvements in cognitive function in some shorter clinical trials, including the LIFE pilot study,¹³ may dissipate by 24 months and thus may have been missed, especially if adherence to the physical activity intervention wanes over time¹⁴; (3) the study population was not specifically selected for cognitive vulnerability, although poor physical function, especially gait speed, has been shown to be a risk for cognitive decline^35,36; (4) the participants were well educated (>two-thirds went to college), and high cognitive reserve may have protected against cognitive decline over 2 years³⁷; and (5) the health education intervention may have benefited cognition.^10,38 The health education group attended interactive seminars providing both cognitive and social stimulation. Both cognitive and social stimulation have been shown to preserve cognition in older adults.^10,38

The dose-response relationship between physical activity and cognition is not well understood.^7,39 The physical activity intervention was designed to provide moderate-intensity aerobic walking activity and was consistent with American College of Sports Medicine recommendations. However, we recruited a population with limited physical ability. Impaired lower-extremity functioning and the high prevalence of comorbidities may have limited participants’ ability to exercise at sustained levels sufficient to improve cognition. Nonetheless, the physical activity group had significantly greater physical activity levels than the health education group, and a more intensive, sustained intervention that could be translatable at the population level would be difficult to achieve.

Despite the lack of overall benefit, our prespecified subgroup analyses of participants aged 80 years or older and those with lower baseline physical performance demonstrated that the physical activity group had better performance on executive function tasks than those in the health education group at 24 months. This finding is important because executive function is the most sensitive cognitive domain to exercise interventions,⁴⁰ and preserving it is required for independence in instrumental activities of daily living. Future physical activity interventions, particularly in vulnerable older adult groups (eg, ≥80 years of age and those with especially diminished physical functioning levels), may be warranted.

To our knowledge, the LIFE study is the largest, longest RCT of a physical activity intervention in sedentary older adults at increased risk for mobility disability. Other strengths include high retention rates, without differential loss to follow-up in the 2 groups; comprehensive standardized, well-validated cognitive assessments; and blinded adjudication of MCI and dementia.

However, there are several limitations. First, even though cognitive function and incident MCI and dementia were a priori outcomes for the LIFE study, our study was not specifically powered for these outcomes and may have been too short to affect incident events. Second, the intensity of the physical activity intervention was moderate by design. Although the physical activity intervention was sufficient to increase physical activity level and reduce incident mobility disability,¹⁴ it may have been insufficient to produce cognitive effects. Third, the components of the health education intervention, including the cognitive and social components, may have improved or prevented cognitive decline. Fourth, we did not measure changes in mechanistic surrogate outcomes, such as brain volumes or cerebrospinal fluid β-amyloid levels.

Among sedentary older adults, a 24-month moderate-intensity physical activity program compared with a health education program did not result in improvements in global or domain-specific cognitive function.

Corresponding Author: Kaycee M. Sink, MD, MAS, Wake Forest University School of Medicine, Sticht Center on Aging, One Medical Center Boulevard, Winston-Salem, NC 27157 (kmsink@wakehealth.edu).

Author Contributions: Dr Espeland had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Sink, Espeland, Church, Guralnik, Hendrie, Jennings, Katula, McDermott, Pahor, Reid, Verghese, Rapp, Williamson.

Acquisition, analysis, or interpretation of data: Sink, Espeland, Castro, Church, Cohen, Dodson, Guralnik, Hendrie, Jennings, Katula, Lopez, McDermott, Pahor, Reid, Rushing, Verghese, Williamson.

Drafting of the manuscript: Sink, Espeland, Cohen, Hendrie, Katula, Lopez, Reid, Rushing, Williamson.

Critical revision of the manuscript for important intellectual content: Sink, Espeland, Castro, Church, Cohen, Dodson, Guralnik, Hendrie, Jennings, Katula, McDermott, Pahor, Reid, Verghese, Rapp, Williamson.

Statistical analysis: Espeland, Jennings, Rushing.

Obtained funding: Espeland, Church, Guralnik, McDermott, Pahor, Verghese, Williamson.

Administrative, technical, or material support: Cohen, Guralnik, Jennings, Katula, Pahor, Reid, Verghese, Rapp, Williamson.

Study supervision: Sink, Church, Guralnik, Hendrie, Katula, Lopez, McDermott, Pahor, Reid, Williamson.

Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Sink reported receiving grant funding from the National Institute on Aging for several Alzheimer’s Disease Cooperative studies; and grant funding from Navidea while serving as a primary investigator. Dr Cohen reported receiving an endowment from the McKnight Brain Research Foundation. Dr Lopez reported serving as a consultant for Baxter, Lilly, Grifols, and Lundbeck. Dr McDermott reported receiving grant funding from Regeneron. Dr Reid reported serving as a consultant for Eli Lilly and Co and Bay Cove Human Services. No other disclosures were reported.

Funding/Support: The Lifestyle Interventions and Independence for Elders (LIFE) study was funded by cooperative agreement UO1 AG22376 from the National Institutes of Health, National Institute on Aging and 3U01AG022376-05A2S from the National Heart, Lung, and Blood Institute; and sponsored in part by the Intramural Research Program, National Institute on Aging, National Institutes of Health. This research was also supported by the following grants awarded to Claude D. Pepper Older Americans Independence Centers: Boston, Massachusetts (1P30AG031679), Florida (1 P30 AG028740), Pittsburgh, Pennsylvania (P30 AG024827), Wake Forest University (1 P30 AG21332), and Yale University (P30AG021342). Dr Reid’s contribution was partially supported by the US Department of Agriculture under agreement 58-1950-7-707.

Role of the Funder/Sponsor: The National Institutes of Health was a voting member (1 of 12 votes) of the LIFE study steering committee, which approved the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The investigators for the LIFE study were Mary M. McDermott, MD (field center principal investigator), Bonnie Spring, PhD (field center co-principal investigator), Joshua Hauser, MD, Martha Gulati, MD, Sanjiv Shah, MD, Kathryn Domanchuk, Rex Graff, Kasia Kadela, and Alvito Rego (Northwestern University, Chicago, Illinois); Timothy Church, MD, PhD, MPH (field center principal investigator), Steven Blair, PED, Valerie Myers, PhD, Ron Monce, PA-C, Nathan Britt, NP, Melissa Nauta, BS, Ami Parks, MPA, BS, Ruben Rodarte, MBA, MS, BS, Heidi Millet, BS, Catrine Tudor-Locke, PhD, FACSM, Ben Butitta, BS, and Sheletta Donatto, MS, RD, LDN, CDE (Pennington Biomedical Research Center, Baton Rouge, Louisiana); Denise Bonds, MD (National Heart, Lung, and Blood Institute, Bethesda, Maryland); Evan C. Hadley, MD, Kushang V. Patel, PhD, and Sergei Romashkan, MD, PhD (National Institute on Aging, Bethesda, Maryland); Robert M. Kaplan, PhD (Office of Behavioral and Social Sciences Research, National Institutes of Health, Bethesda, Maryland); Abby C. King, PhD (field center principal investigator), Cynthia M. Castro, PhD, William L. Haskell, PhD, Randall S. Stafford, MD, PhD, Veronica Yank, MD, Leslie A. Pruitt, PhD, Kathy Berra, MSN, NP-C, FAAN, Carol Bell, NP, Rosita Thiessen, Kate P. Youngman, MA, Selene B. Virgen, BAS, Kristina N. Tarin, MS, CSCS, Heather Klaftenegger, Carolyn A. Prosak, RD, Ines Campero, BA, Dulce M. Garcia, José Soto, BA, Linda Chio, and David Hoskins, MS (Stanford University, Palo Alto, California); Roger Fielding, PhD (field center principal investigator), Miriam Nelson, PhD, Sara Folta, PhD, Edward Phillips, MD, Christine Liu, MD, Erica Cifarelli, MS, Kieran Reid, MSc, MPH, Paige Lacasse, BS, Dylan Kirn, BS, Evan Pasha, BS, Sonnly Ribourg, BS, Karen Ruais, NP, Won Kim, BS, and Gregory Cloutier, BS (Tufts University, Boston, Massachusetts); Todd Manini, PhD (field center principal investigator), Marco Pahor, MD (field center co-principal investigator), Stephen Anton, PhD, Thomas Buford, PhD, Susan Nayfield, MD, MSc, Joe Nocera, PhD, Michael Marsiske, PhD, Bhanu Sandesara, MD, Jocelyn Lee, PhD, Deborah Hiatt-Jensen, ARNP, MPH, Mieniecia Black, MS, Kimberly Case, PhD, Katherine Herring, BS, Amber Schwier-Delvisco, BS, Bill Burk, Brian Hoover, BS, Jeffrey Knaggs, BS, and Allison Martin, MS, RCEP (University of Florida, Gainesville); Chonglun Xie, MD, Holly Morris, RN, Flo Singletary, MS, CCC-SLP, Jackie Causer, Susan Yonce, RN, Christine Del Boccio, Charles Gay, DC, and Tangerica Peavy Alexander (Brooks Center for Rehabilitation Studies, satellite site); Marco Pahor, MD (principal investigator), Stephen Anton, PhD, Thomas Buford, PhD, Christiaan Leeuwenburgh, PhD, Susan Nayfield, MD, MSc, Todd Manini, PhD, Connie Caudle, Lauren Crump, MPH, Latonia Holmes, Jocelyn Lee, PhD, Ronald Lester, PhD, MBA, Ching-ju Lu, MPH, and Ryan O’Mara (Administrative Coordinating Center); Carl J. Pepine MD, MACC, Mario Ariet, PhD, Eileen Handberg, PhD, ARNP, Daniel Deluca, BS, and James Hill, MD, MS, FACC (Electrocardiogram Reading Center); Jack M. Guralnik, MD, PhD (co-principal investigator; University of Maryland School of Medicine, Baltimore); Anne B. Newman, MD, MPH (field center principal investigator), Stephanie A. Studenski, MD, MPH (field center co-principal investigator), Bret H. Goodpaster, PhD, Oscar Lopez, MD, Nancy W. Glynn, PhD, Janet T. Bonk, MPH, RN, Jennifer Rush, MPH, Piera Kost, BA, Allison Gerger, BS, Pamela Vincent, CMA, LaTisha Davis, BS, Mark A. Newman, PhD, George Grove, MS, and Kathy Williams, RN, BSEd, MHSA (University of Pittsburgh, Pittsburgh, Pennsylvania); Erik J. Groessl, PhD (VA San Diego Healthcare System and University of California, San Diego); Stephen B. Kritchevsky, PhD (field center principal investigator), Anthony Marsh, PhD (field center co-principal investigator), Tina Brinkley, PhD, Jamehl Demons, MD, Kaycee Sink, MD, MAS, Kimberly Kennedy, BA, CCRC, Rachel Shertzer-Skinner, MA, CCRC, Abbie Wrights, MS, Rose Fries, RN, CCRC, and Deborah Barr, MA, RHEd, CHES (Wake Forest University, Winston-Salem, North Carolina); Jeff Williamson, MD, MHS (center principal investigator), Kaycee M. Sink, MD, MAS (center co-principal investigator), Hugh C. Hendrie, MB, ChB, DSc, Stephen R. Rapp, PhD, Joe Verghese, MBBS, Nancy Woolard, and Valerie K. Wilson, MD (Cognition Reading Center); Michael E. Miller, PhD (principal investigator), Mark Espeland, PhD (co-principal investigator), Walter Ambrosius, PhD, Don Babcock, BSEE, PE, Daniel Beavers, PhD, MS, Robert P. Byington, PhD, MPH, FAHA, Delilah Cook, Curt Furberg, MD, PhD, Candace Goode, Jason Griffin, BS, Lea Harvin, BS, Leora Henkin, MPH, Med, John Hepler, BFA, Fang-Chi Hsu, PhD, Kathy Joyce, Laura Lovato, MS, Wesley Roberson, BSBA, Julia Robertson, BS, Julia Rushing, BSPH, MStat, Scott Rushing, BS, Cynthia L. Stowe, MPM, Michael P. Walkup, MS, Don Hire, BS, Jack Rejeski, PhD, Jeff Katula, PhD, MA, Peter H. Brubaker, PhD, Shannon Mihalko, PhD, Janine M. Jennings, PhD, and Kathy Lane, BA (Data Management, Analysis and Quality Control Center); Thomas M. Gill, MD (field center principal investigator), Robert S. Axtell, PhD, FACSM (field center co-principal investigator), Susan S. Kashaf, MD, Nathalie de Renekeire, MD, Joanne M. McGloin, MDiv, MS, MBA, Raeleen Mautner, PhD, Sharon M. Huie-White, MPH, Luann Bianco, BA, Janice Zocher, Denise M. Shepard, RN, MBA, Barbara Fennelly, MA, RN, Sean Halpin, MA, Theresa Barnett, MS, APRN, Karen C. Wu, RN, Lynne P. Iannone, MS, Julie A. Bugaj, MS, and Christine Bailey, MA (Yale University, New Haven, Connecticut); and Geoffrey Chupp, MD, Gail Flynn, RCP, CRFT, Thomas M. Gill, MD, John Hankinson, PhD, and Carlos A. Vaz Fragoso, MD (Spirometry Reading Center).

Disclaimer: The opinions, findings, conclusion, and recommendations expressed are those of the authors and do not necessarily reflect the view of the US Department of Agriculture. Dr McDermott, JAMA Senior Editor, was not involved in the review of or decision to publish this article.

Additional Contributions: We thank Valerie K. Wilson, MD (Wake Forest University School of Medicine, Winston-Salem, North Carolina, and Veterans Affairs, Johnson City, Tennessee), who additionally contributed as an adjudicator of cognitive outcomes and received compensation.