Analyses were adjusted for covariates (age, sex, race, educational level, weight, hypertension, coronary heart disease, stroke, bilateral hippocampal volume, and white matter hyperintensities on magnetic resonance imaging normalized to intracranial volume) not including apolipoprotein E ε4 (APOE ε4) carrier status (left) and including APOE ε4 carrier status (right). Regions are arranged in descending order of magnitude of the regression coefficient. Error bars indicate 95% CIs. PiB indicates Pittsburgh Compound B.
eTable 1. Association Between APOE (Independent Variable) and Gait Speed (Dependent Variable) in the Whole Dementia-Free Sample and in the CN Subsample
eTable 2. Stratified Analysis by APOE ε4 Carrier Status Showing the Association Between Global PiB SUVR and Gait Speed in the Whole Dementia-Free Sample and in the CN Subsample Adjusted for Covariates
Customize your JAMA Network experience by selecting one or more topics from the list below.
Nadkarni NK, Perera S, Snitz BE, et al. Association of Brain Amyloid-β With Slow Gait in Elderly Individuals Without DementiaInfluence of Cognition and Apolipoprotein E ε4 Genotype. JAMA Neurol. 2017;74(1):82–90. doi:10.1001/jamaneurol.2016.3474
Copyright 2016 American Medical Association. All Rights Reserved.
Is brain amyloid-β (Aβ) associated with gait speed, and do cognition and apolipoprotein E ε4 (APOE ε4) status influence this association in elderly individuals?
In this secondary analysis of a cross-sectional study, we found that, in 183 elderly individuals without dementia, gait speed was associated with Aβ deposition independent of cardiac risk, hippocampal volume, and small-vessel disease burden. This association was attenuated by cognition and rendered statistically nonsignificant when accounting for APOE ε4 carrier status.
Cognitive functions and APOE ε4 status influence the association between brain Aβ and motor function in elderly individuals without dementia.
Motor slowing appears in preclinical Alzheimer disease (AD), progresses with AD progression, and is associated with AD pathologic findings at autopsy. Whether amyloid-β (Aβ) is associated with gait speed in elderly individuals without dementia and whether cognition and apolipoprotein E ε4 (APOE ε4) influence this association remain unknown.
To examine the association between Aβ and gait speed in elderly individuals without dementia and to study the influence of cognition and APOE ε4 status on this association.
Design, Setting, and Participants
This cross-sectional analysis included 183 elderly individuals without dementia, including a cognitively normal (CN) subsample of 144 adults, enrolled in the Ginkgo Evaluation of Memory study at a university center from January 1, 2000, through December 31, 2009, and enrolled in a follow-up substudy a mean (SD) of 10 (3) months after the initial study closeout. Data analysis was performed from October 1, 2015, to June 1, 2016.
Main Outcomes and Measures
We assessed cerebral Aβ on Pittsburgh Compound B (PiB) positron emission tomography, gait speed over 4.57 m (15 ft), and cognition on the Mini-Mental State Examination and Trail Making Test Parts A and B. We grouped participants into high Aβ (PiB+) and low Aβ (PiB–) groups on standardized global PiB cutoffs and examined group differences. We studied the influence of cognition and APOE ε4 on the global and regional associations between gait speed and Aβ in the whole sample and the CN subsample.
Among the 183 study participants, mean (SD) age was 85.5 (3) years, 76 were women (41.5%), and 177 were white (96.7%). The PiB+ individuals were comparable to the PiB– individuals on demographics, comorbidities, cognition, hippocampal volume, and small-vessel disease but not on gait speed (0.85 vs 0.92 m/s, P = .01) or proportion of APOE ε4 carriers (29 [29.0%] vs 5 [6.0%], P < .001). In the whole sample and the CN subsample, the association between global PiB retention and slower gait withstood adjustment for covariates (β = −0.068, P = .03 and β = −0.074, P = .04, respectively); however, this association was attenuated by Mini-Mental State Examination and Trail Making Test Parts A and B and was rendered statistically nonsignificant by APOE ε4 in both samples (β = −0.055 and β = −0.058, respectively; both P ≥ .10). Several regional associations between gait speed and PiB uptake withstood relevant adjustments; however, APOE ε4 rendered only the medial (β = −0.22, P = .03) and lateral (β = −0.08, P = .03) temporal regions, subcortical white matter (β = −0.13, P = .02), and occipital regions (β = −0.15, P = .03) in the whole sample and the occipital regions (β = −0.21, P = .01) in the CN subsample statistically significant.
Conclusions and Relevance
Cerebral Aβ deposition is associated with slower gait speed in elderly individuals without dementia; however, this association is weaker in those who are CN. Cognition and APOE ε4 carrier status influence the association between Aβ and gait speed in elderly individuals without dementia.
Motor slowing appears during the preclinical stages of Alzheimer disease (AD),1 accelerates in those who subsequently develop mild cognitive impairment (MCI) and AD,2,3 and is associated with severity of AD pathologic findings at death.4 In elderly individuals without dementia, elevated levels of fibrillar amyloid-β (Aβ) are seen in 30% to 65% of adults between 80 and 88 years of age.5-8 Recent evidence suggests that high levels of Aβ increase the risk of falls in these populations.9 Slowing of gait speed is an important determinant of falls and mortality in older adults10 and is associated with cerebral small-vessel disease and cortical atrophy.11 However, cerebral Aβ may also play a role in gait slowing in older adults without dementia, but there is little research to support this association.
Cognitive processes influence gait in cognitively normal (CN) older adults and in those with MCI.11,12 Deficits in these cognitive processes are associated with greater Aβ deposition in these populations.13,14 In addition, the apolipoprotein E ε4 (APOE ε4) genotype is associated with poor mobility and physical function in older adults15 and in those with MCI,16 and the presence of an APOE ε4 allele is linked to accelerated motor decline in aging,17,18 with exceptions.19,20 Besides age, APOE ε4 is the largest risk factor for AD, and the presence of an APOE ε4 allele increases Aβ accumulation in CN older adults5,21-24 and in prodromal and clinical AD.24-26 Therefore, cognition and APOE ε4 genotype may influence the association between Aβ and gait speed in elderly individuals without dementia.
We investigated the association between cortical and regional Aβ deposition and gait speed in elderly individuals without dementia and assessed whether cognition and APOE ε4 status influence this association. We further examined these associations in a subsample of older adults deemed CN in the parent study5 to understand whether the association between Aβ deposition and gait in the entire cohort was driven by those with MCI. We hypothesized that greater cortical Aβ deposition is associated with slower gait in the whole sample, but its magnitude and statistical significance would be weaker in those without MCI (ie, the CN individuals). In addition, we posited that cognition and APOE ε4 would independently influence the global and regional association between Aβ deposition and gait speed beyond that explained by demographic factors, cardiac risk, cortical atrophy, and small-vessel disease, factors known to play a role in age-related motor slowing.11,12
The Gingko Evaluation of Memory (GEM) study (clinicaltrials.gov Identifier: NCT00010803), a double-blind, placebo-controlled, randomized clinical trial of Ginkgo biloba targeted to prevent dementia, particularly AD, recruited CN elderly individuals and elderly individuals with MCI.27,28 Participants without dementia at study entry were followed up annually from January 1, 2000, through December 31, 2009, in the GEM study. A mean (SD) of 10 (3) months after the GEM study closeout visit, 194 participants without dementia enrolled at the University of Pittsburgh site were recruited for the GEM imaging substudy that included brain magnetic resonance imaging and Pittsburgh Compound B (PiB) positron emission tomography (PiB-PET).5,29 Eligibility criteria for the GEM imaging substudy were described previously.5,29 Eleven participants were not included in this analysis because of technical issues relating to their PiB-PET (n = 3) and magnetic resonance imaging (MRI) (n = 8). This study analyzed data from 183 GEM study participants who had complete brain MRI and PiB-PET data along with physical performance measures. The GEM parent study and neuroimaging substudy protocols were approved by local institutional review boards of the University of Pittsburgh, Pittsburgh, Pennsylvania. Written informed consent was obtained from all participants for the parent study and the imaging substudy. Data analysis was performed from October 1, 2015, to June 1, 2016.
All participants underwent detailed neuropsychological assessments annually as part of the GEM study,27,28 a subset of whom were included in the GEM imaging substudy.5,29 We assessed global cognitive function on the Mini-Mental State Examination (MMSE) and attention and executive function on the Trail Making Test Part A (TMT-A) and Part B (TMT-B).28 Adjudication of MCI was conducted by the GEM Cognitive Diagnostic Center, taking into account all neuropsychological assessments from the GEM parent study and the GEM imaging substudy. Criteria for MCI included a cutoff of 1.5 SDs below the age- and education-adjusted norm on 2 to 3 tests.5
Time to walk 4.57 m (15 ft) was measured using a protocol similar to the one used to assess gait speed in the Short Physical Performance Battery30 and has been described previously.31 Briefly, a 4.57-m-long traverse was demarcated with tape, and participants were instructed to begin walking from standing position at the start line and continue walking past the end line. Time was measured using a stopwatch, which was started after the prompt when one foot started to move across the start line and was stopped when the first foot crossed the 4.57-m end mark. Two consecutive 4.57-m walks were obtained, the first at the usual pace and the second at a rapid pace.31 We used the usual-paced timed walk measure to derive gait speed in meters per second. The 4.57-m walk test is a well-validated measure of gait speed in older adults with and without dementia with an excellent test-retest reliability (intraclass correlation coefficient = 0.973-0.977).32 Mean duration between gait speed assessment and brain imaging was 16 months (range, 10-25 months).
The PiB-PET method was reported previously.5,29 In brief, carbon 11–labeled PiB ligand (approximately 15 mCi) was injected for 20 seconds, and a 10-minute transmission scan was acquired for attenuation correction followed by 20-minute PiB-PET (4 × 5-minute frames) acquired 50 minutes after injection. The PiB retention was assessed in the resection-normalized PET image in the regions of interest (ROIs) that encompassed the following bilateral regions33: anterior cingulate gyrus (ACG) (pregenual and subgenual), anteroventral striatum (AVS) (anterior caudate and putamen), frontal cortex (FRC) (dorsal and ventral), lateral temporal cortex (LTC), parietal cortex (PAR), precuneus cortex (PRC), mesial temporal cortex (MTC) (amygdala and hippocampus), occipital cortex (OCC) (primary visual cortex), occipital pole (OCP), pons, sensory-motor cortex (SMC), subcortical white matter (SWM), and thalamus. We grouped participants into high Aβ (PiB+) and low Aβ (PiB–) groups on standardized global PiB cutoffs of Aβ deposition and examined group differences. An iterative outlier cutoff method was used to define individuals as PiB+ if the atrophy-corrected PiB standardized uptake value ratio (SUVR) (referenced to the cerebellar value) was greater than 1.57 averaged from the PiB SUVR of the ACG, AVS, FRC, LTC, PAR, and PRC.29 A continuous measure of global PiB SUVR represented values in these 6 ROIs.
APOE genotyping was performed using polymerase chain reaction on DNA isolated from whole-blood samples as described previously.34 Participants with at least 1 APOE ε4 allele (ε2/ε4, ε3/ε4, or ε4/ε4) were identified as APOE ε4 carriers.
Several age-related changes are linked to gait and cognitive performance in older adults.11,12 Among these, cardiac risk factors are also associated with cerebral Aβ deposition29 and APOE ε4 carrier status.35 Hence, analyses were adjusted for covariates that included demographics (age, sex, race, and educational level), body weight, hypertension, coronary heart disease (presence of self-reported or medical record review–based diagnoses of angina, myocardial infarction, angioplasty, bypass surgery, pacemaker, valve replacement, or heart failure), stroke, and MRI measures of cortical atrophy and small-vessel disease (bilateral hippocampal volume and total volume of white matter hyperintensities, both presented as a proportion of intracranial volume).29
We compared PiB+ and PiB− groups on demographic, health, and key brain measures using unpaired, 2-tailed, independent-samples t tests. We examined the association between global PiB retention and gait speed using multiple regression adjusting for the above covariates. We included cognitive measures (MMSE, TMT-A, and TMT-B) and APOE ε4 carrier status separately as additional independent variables in the unadjusted and adjusted models. We repeated the analyses using a subsample of participants deemed CN after excluding those with MCI. We included an interaction term in the models to examine whether the association between PiB SUVR and gait speed was different in APOE ε4 carriers and noncarriers based on observations in other settings.36 However, recognizing that interaction terms have low statistical power, we a priori planned an exploratory analysis stratified by APOE ε4 carrier status. Finally, we performed another exploratory analysis of regional association between PiB SUVR and gait speed, initially adjusting for all covariates and then further adjusting for APOE ε4 status. To assess whether the time between gait assessment and PiB-PET influenced the overall results, we performed a sensitivity analysis by including the duration of time between gait assessment and PiB-PET as a covariate. We set α to .05 to identify statistical significance.
In the whole sample (183 adults; mean [SD] age, 85.5  years; 76 women [41.5%]; 177 white [96.7%]), the mean MMSE score was 28. There were 34 (18.6%) APOE ε4 carriers, and 100 (54.6%) were designated as PiB+. The characteristics of the whole group and subgroups divided on global PiB SUVR are given in Table 1. Both PiB+ and PiB− subgroups were similar in terms of their demographics, body weight, comorbidities, MMSE scores, TMT-A scores, TMT-B scores, physical performance measures, number of falls, and hippocampal and white matter hyperintensity volumes. The PiB+ group had a greater proportion of APOE ε4 carriers compared with the PiB− group (29 [29.0%] vs 5 [6.0%], P < .001). In addition, we found that the PiB+ group had a slower gait speed than the PiB− group (0.85 vs 0.92 m/s, P = .01).
Sample characteristics of the CN subsample (144 adults; mean [SD] age, 85.4 [2.9] years) are given in Table 2. The PiB+ and PiB− groups were similar on all the above characteristics except for prevalence of hypertension (31 [41.9%] vs 15 [22.4%], P = .01), the proportion of APOE ε4 carriers (22 [29.3%] vs 5 [7.2%], P < .001), and gait speed (0.87 vs 0.94 m/s, P = .04).
No differences were found in gait speed between the placebo and Gingko biloba arms of the study in the whole sample (0.89 vs 0.87 m/s, P = .60) or in the CN subsample (0.89 vs 0.92 m/s, P = .40).
Table 3 gives the association between global PiB binding and gait speed. In the whole sample, greater global PiB SUVR was associated with slower gait (regression coefficient [β] = −0.086, P = .005), and this association remained significant after adjustment for above covariates (β = −0.068, P = .03). The MMSE score correlated with gait speed (r = 0.24, P = .002); the TMT-A and TMT-B scores correlated with both global PiB SUVR (r = 0.3, P = .005 and r = 0.18, P = .02, respectively) and gait speed (r = −0.3, P < .001 and r = −0.19, P = .01, respectively). The association between PiB SUVR and gait speed was attenuated but tended to persist after adjusting for MMSE (β = −0.07, P = .02), TMT-A (β = −0.06, P = .04), and TMT-B (β = −0.07, P = .02) scores. Accounting for APOE ε4 in the model rendered the association between PiB SUVR and gait speed nonsignificant and contributed to approximately 16% of the additional explained variance in the entire sample (Table 3).
In the CN elderly individuals, greater global PiB SUVR was associated with slower gait speed (β = −0.072, P = .04) even after adjusting for the above covariates (β = −0.074, P = .04); however, this association was no longer significant after additional adjustments for MMSE (β = −0.059, P = .08), TMT-A (β = −0.063, P = .07), or TMT-B (β = −0.068, P = .06) scores or APOE ε4 status (β = −0.06, P = .10). APOE ε4 explained approximately 10% of the additional explained variance in the association between PiB SUVR and gait speed (Table 3).
We did not find a statistically significant interaction with APOE ε4 and global PiB with respect to gait speed. We found no significant associations between APOE ε4 carrier status and gait speed (eTable 1 in the Supplement). The stratified analysis by APOE ε4 carrier status suggested that the associations between gait speed and PiB SUVR, MMSE scores, TMT-A scores, and TMT-B scores were stronger in the APOE ε4 noncarriers than in carriers in both samples (eTable 2 in the Supplement).
We performed a sensitivity analysis to examine whether the duration between gait assessment and PiB-PET had any bearing on the association between Aβ and gait speed. With the period between MRI and gait assessment as a covariate in the regression analysis, the strength of the unadjusted association between global PiB SUVR and gait speed was unchanged (whole sample: β = −0.086, P = .005; CN subsample: β = −0.072, P = .04).
The Figure depicts an exploratory analysis showing the coefficient of the association between regional PiB SUVR for global PiB SUVR and regional PiB SUVR for the whole group and for the subsample limited to CN individuals adjusted for demographics, body weight, hypertension, coronary heart disease, stroke, MMSE score, and normalized hippocampal and white matter hyperintensity volumes but not adjusted for APOE ε4. In the whole sample, slower gait was significantly associated with regional PiB SUVR in the AVS (β = −0.07, P = .03), LTC (β = −0.09, P = .01), MTC (β = −0.24, P = .02), PAR (β = −0.07, P = .03), PRC (β = −0.06, P = .02), SMC (β = −0.12, P = .02), SWM (β = −0.11, P = .03), and OCP (β = −0.14, P = .046). However, in CN elderly individuals, slower gait was significantly associated with greater regional PiB SUVR in the AVS (β = −0.08, P = .04), LTC (β = −0.10, P = .02), PRC (β = −0.07, P = .02), and SMC (β = −0.08, P = .02) (Figure). Gait speed was not significantly associated with regional PiB SUVR in any other ROIs in the whole group or in the subsample of CN elderly individuals.
The Figure also shows the associations between regional PiB SUVR and gait speed adjusted for APOE ε4 status. In the models adjusted for the above covariates (Figure), we found that additional inclusion of APOE ε4 rendered the regional association between gait speed and PiB SUVR in the FRC, PRC, AVS, and ACG nonsignificant (Figure) in the whole sample and in the subsample of individuals without MCI (the CN individuals). The statistically significant associations between gait speed and PiB SUVR in the MTC (β = −0.22, P = .03), LTC (β = −0.08, P = .03), OCC (β = −0.15, P = .03), OCP (β = −0.15, P = .03), and SWM (β = −0.13, P = .02) were retained in the whole sample; however, in the sample without MCI, this association was limited to the OCC (β = −0.21, P = .01).
In this sample of elderly individuals without dementia in the category that we term oldest old, brain Aβ deposition was present at high levels in 54.6% and was associated with slower gait speed independent of demographics, cardiac risk, hippocampal volume, and small-vessel disease. The association between brain Aβ deposition and gait speed was influenced by global cognitive and executive function capabilities and APOE ε4 carrier status. Gait speed was associated with regional Aβ in the AVS, MTC, LTC, PAR, PRC, SMC, and SWM. However, APOE ε4 attenuated the associations between gait speed and global Aβ and the Aβ deposition in several anterior brain regions, particularly in the CN sample. To our knowledge, this is one of the first reports to highlight the influence of cognition and APOE ε4 on the association between global and regional Aβ deposition and gait speed in elderly individuals without dementia and in CN elderly individuals.
Prior studies4,37 have linked Aβ pathologic findings to physical performance measures in aging and neurodegenerative diseases. In population studies, Aβ plaques and neurofibrillary tangles, hallmarks of AD pathologic findings, are associated more strongly with motor performance measures than are other changes in the aging brain, such as small cerebral infarcts or Lewy body pathologic findings.17 Although both Aβ plaques and neurofibrillary tangles are associated weakly with gait speed,4 they are associated more strongly with gait slowing during 6.4 years of longitudinal assessments.37 In elderly individuals without dementia, elevated levels of Aβ are associated with a 5-fold increase in falls.9 These studies4,9,37,38 support our findings that greater Aβ deposition in the brain is associated with mobility decline in elderly individuals.
APOE ε4 increases Aβ deposition in individuals with preclinical AD and CN individuals.21,24,35,39APOE ε4 also has other effects on the brain, such as facilitating tau hyperphosphorylation.35 Moreover, APOE ε4 carriers have more severe gait impairments15 and worsened gait speed decline17,18 than APOE ε4 noncarriers. APOE ε4 status is also associated with gait speed in MCI, although not with other physical performance measures, such as grip strength, chair stands, or cardiorespiratory status.16 This body of literature suggests that APOE ε4 may influence cortical control of gait by influencing both Aβ-related and Aβ-unrelated processes, which may explain why controlling for APOE ε4 in the statistical analyses led to diminution of the magnitude of the association between Aβ and gait speed. The association between Aβ and gait speed appears to have been driven by 34 and 27 APOE ε4 carriers in the whole sample and the CN subsample, respectively; this small sample of APOE ε4 carriers may have also precluded any interaction between APOE ε4 and interpretation of the association between global PiB retention and gait speed within APOE ε4 subgroups. However, our findings complement the increasing body of literature that indicates that APOE ε4 status may play a role in the association between global cortical Aβ deposition and gait speed in elderly individuals without dementia and CN individuals.
Cognition plays an important role in motor planning and gait control in older adults, including those without dementia.40 In CN elderly individuals, slower gait is associated with worse attention, executive function, visuospatial processing, and memory.40-42 In elderly individuals without dementia, Aβ is associated with global cognitive function,13,14,43 memory,13,43-45 attention/executive function,14,46 and visual-spatial processing.14,43 We found that global cognitive function and executive function measures attenuated the association between global PiB SUVR and gait speed in the whole sample and rendered the association between Aβ and gait speed statistically nonsignificant in the CN subsample, suggesting that Aβ may influence higher-level cognitive processes that play an important role in gait control in these populations. Furthermore, APOE ε4 modulates the association between global Aβ and global cognition, memory, and visual-spatial processing,14,43 albeit with exceptions.13 Therefore, our findings suggest that APOE ε4 may influence the cognitive processes involved in the control of gait and influence gait slowing in elderly individuals.
The exploratory regional analysis in the entire cohort, including the 21.3% with MCI, revealed that Aβ deposition in the AVS, ACG, MTC, LTC, PAR, SWM, SMC, and PRC was associated with gait speed, whereas in the CN sample, the regional associations were limited to the SMC, AVS, PRC, and LTC. These findings are supported by another recent report38 on the regional associations between Aβ and gait speed. The SMC, AVS, PAR, PRC, and related networks play an important role in gait control,47,48 and our data suggest that Aβ in these areas may affect gait speed in older adults. However, we also found that regional associations in the ACG, AVS, PAR, and PRC were not significant after correction for APOE ε4. The APOE ε4 allele influences Aβ deposition in the FRC, ACG, AVS, PAR, and PRC regions.21,49 In PiB+ elderly individuals without dementia, regional Aβ distribution is similar to that of patients with AD, including nonspecific binding in the SWM,8,49,50 and is associated with cognition–medial temporal Aβ with memory44,45 and frontal, temporal, and parietal Aβ with global cognition.14 This finding may explain why APOE ε4 rendered these predominantly anterior regional associations nonsignificant in the entire sample and the CN subsample. Our findings differ slightly from the recent study38 on regional Aβ deposition and gait associations in a heterogeneous sample of older adults selected on the basis of memory problems (99.2%), slow gait (11.7%), and impaired instrumental activities of daily living (6.3%) that reported that greater regional Aβ deposition in the anterior cingulate, precuneus, putamen, and occipital cortex regions was related to slower gait; these analyses were adjusted for APOE ε4 status but not for cardiac risk, white matter hyperintensity volume, or cortical atrophy.38 The differences in our findings from this study38 could relate to varying inclusion criteria, delineation of ROIs, differences in the study samples, and the statistical adjustments used.
Our findings indicate that Aβ is not strongly associated with gait speed in CN individuals, suggesting that Aβ is not the main driver of slow gait speed in aging or AD. Aβ may coexist and contribute to other AD-related brain changes, such as inflammation, tau aggregation, and neurofibrillary tangle pathologic findings, which spread to the neocortex and coincide with onset of AD symptoms,51 and may include gait slowing.2,3 In PiB+ CN individuals, Aβ deposition may be an early event in the AD process and may be weakly associated with gait speed, nevertheless modified by the APOE ε4 allele that favors Aβ deposition over tau aggregation in CN aging.24 Changes in the brain in CN older adults may be independent of Aβ.52 Given the lack of research in this area, we speculate that gait speed in elderly individuals may be affected by Aβ-dependent and Aβ-independent pathways influenced by APOE ε4.
Our study has several limitations. The study was an exploratory secondary analysis of data of well-characterized elderly individuals who had available PiB-PET and physical performance measure data. The smaller sample size, especially in the APOE ε4 subgroup analyses, resulted in lower statistical power that is required to indicate meaningful conclusions. We cannot exclude the possibility of other AD-related conditions, such as tau, contributing to gait slowing in our sample. The timing of gait speed assessment was not concurrent with PiB-PET; however, we performed a sensitivity analysis that found that controlling for the duration of time between gait assessment and PiB-PET did not affect the overall results. Last, this was a cross-sectional analysis that included an exploratory analysis of regional associations on a well-characterized sample; therefore, although our findings are hypotheses generating, we cannot address causality or directionality of these associations.
This study reveals that, in elderly individuals without dementia, gait speed was modestly associated with Aβ deposition independent of cardiac risk, hippocampal volume, and small-vessel disease burden. In addition, the association between Aβ deposition and gait speed was attenuated by APOE ε4 and cognition.
Corresponding Author: Neelesh K. Nadkarni, MD, PhD, FRCPC, Division of Geriatric Medicine and Gerontology, Department of Medicine, University of Pittsburgh, 3471 Fifth Ave, Ste 500, Pittsburgh, PA 15213 (email@example.com).
Accepted for Publication: July 13, 2016.
Published Online: November 14, 2016. doi:10.1001/jamaneurol.2016.3474
Author Contributions: Dr Perera had full access to all 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: Nadkarni, Williamson, DeKosky, Lopez.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Nadkarni, Perera, Price, Lopez.
Critical revision of the manuscript for important intellectual content: Nadkarni, Perera, Snitz, Mathis, Williamson, DeKosky, Klunk, Lopez.
Statistical analysis: Nadkarni, Perera.
Obtained funding: Nadkarni, Perera, Williamson, DeKosky, Mathis, Klunk, Lopez.
Administrative, technical, or material support: Perera, Snitz, Mathis, Price, Williamson, DeKosky.
Conflict of Interest Disclosures: GE Healthcare holds a license agreement with the University of Pittsburgh. Drs Klunk and Mathis reported being coinventors of PiB and, as such, having a financial interest in this license agreement. No other disclosures were reported.
Funding/Support: This study was supported by grant U01 AT000162 from the National Center for Complementary and Alternative Medicine and the Office of Dietary Supplements (Drs DeKosky and Williamson) and grants K23 AG049945 (Dr Nadkarni), P30 AG024827 (Dr Perera), P50 AG005133 (Drs Lopez, Snitz, and Mathis), and AG047266 (Dr DeKosky) from the National Institute on Aging.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Additional Contributions: Tao Jiang, PhD, Division of Geriatric Medicine, Department of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, assisted with statistical analysis and creating the figures and was financially compensated for these contributions.
Create a personal account or sign in to: