REVIEW Annals of Internal Medicine

Physical Activity Interventions in Preventing Cognitive Decline and

Alzheimer-Type Dementia
A Systematic Review
Michelle Brasure, PhD, MSPH, MLIS; Priyanka Desai, MSPH; Heather Davila, MPA; Victoria A. Nelson, MSc; Collin Calvert, MPH;
Eric Jutkowitz, PhD; Mary Butler, PhD, MBA; Howard A. Fink, MD, MPH; Edward Ratner, MD; Laura S. Hemmy, PhD;
J. Riley McCarten, MD; Terry R. Barclay, PhD; and Robert L. Kane, MD†

Background: The prevalence of cognitive impairment and de- about the effectiveness of aerobic training, resistance training, or
mentia is expected to increase dramatically as the population tai chi for improving cognition. Low-strength evidence showed
ages, creating burdens on families and health care systems. that multicomponent physical activity interventions had no effect
on cognitive function. Low-strength evidence showed that a
Purpose: To assess the effectiveness of physical activity inter- multidomain intervention comprising physical activity, diet, and
ventions in slowing cognitive decline and delaying the onset of cognitive training improved several cognitive outcomes. Evi-
cognitive impairment and dementia in adults without diagnosed dence regarding effects on dementia prevention was insufficient
cognitive impairments. for all physical activity interventions.
Data Sources: Several electronic databases from January 2009 Limitation: Heterogeneous interventions and cognitive test
to July 2017 and bibliographies of systematic reviews. measures, small and underpowered studies, and inability to as-
Study Selection: Trials published in English that lasted 6 sess the clinical significance of cognitive test outcomes.
months or longer, enrolled adults without clinically diagnosed Conclusion: Evidence that short-term, single-component phys-
cognitive impairments, and compared cognitive and dementia ical activity interventions promote cognitive function and prevent
outcomes between physical activity interventions and inactive cognitive decline or dementia in older adults is largely insuffi-
controls. cient. A multidomain intervention showed a delay in cognitive
Data Extraction: Extraction by 1 reviewer and confirmed by a decline (low-strength evidence).
second; dual-reviewer assessment of risk of bias; consensus de- Primary Funding Source: Agency for Healthcare Research and
termination of strength of evidence. Quality.
Data Synthesis: Of 32 eligible trials, 16 with low to moderate Ann Intern Med. 2018;168:30-38. doi:10.7326/M17-1528 Annals.org
risk of bias compared a physical activity intervention with an in- For author affiliations, see end of text.
active control. Most trials had 6-month follow-up; a few had 1- or This article was published at Annals.org on 19 December 2017.
2-year follow-up. Evidence was insufficient to draw conclusions † Deceased.

F orty-seven million people worldwide live with de-

mentia (1), and this number is expected to triple by
2050 (2). Despite evidence that the overall incidence of
gest that physical activity may reduce or delay the
development of potential modifiable risk factors for
cognitive decline, such as obesity, diabetes, and hyper-
dementia has declined in the United States (3, 4), the tension (8 –13). However, the relationships among phys-
number of U.S. adults older than 70 years with demen- ical activity, other risk factors, and cognitive decline are
tia or mild cognitive impairment (MCI) increases as our complex and interrelated. Findings of associations from
population ages (5, 6). Dementia severely erodes func- cohort studies alone cannot clarify whether physical
tioning and quality of life, creates burden and stress on activity affects cognitive decline directly, indirectly
families, and leads to institutionalization. Dementia- through the reduction of medical risk factors, or both.
related costs exceed those of heart disease and cancer Previous systematic reviews of randomized controlled
and often are paid directly by families (7). Therefore, trials report some cognitive benefits of physical activity
preventing dementia is an urgent public health priority. interventions, although the certainty and clinical impor-
Many believe that an active lifestyle may prevent tance of these findings have not always been clear (14,
cognitive decline and dementia. Findings of several re- 15). This systematic review reports a synthesis of the
views, primarily those looking at cohort studies, sug- evidence assessing the effectiveness of physical activity
interventions in slowing cognitive decline and delaying
the onset of cognitive impairment and dementia in
See also: adults without diagnosed cognitive impairments.
Related articles. . . . . . . . . . . . . . . . . . . . . . . 39, 52, 63
Editorial comment . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Web-Only
METHODS
Supplement We developed and followed a standard protocol
(16). Our full technical report (17) contains details on
CME/MOC activity
methods and findings, an analysis of studies address-
30 Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 Annals.org

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 Physical Activity to Prevent Cognitive Decline and Dementia REVIEW
ing secondary prevention in adults with MCI, and an Data Synthesis and Analysis
evaluation of comparative effectiveness. We grouped studies by type of physical activity in-
tervention and analyzed results by direction of effect
Data Sources and Searches and statistical significance. We found it impossible to
We searched bibliographic databases, including assess the clinical significance of findings of the inter-
MEDLINE, EMBASE, and PsycINFO via Ovid, as well as mediate outcomes across all studies, because many dif-
the Cochrane Library, to identify controlled trials pub- ferent instruments were used and we did not always
lished in any language from January 2009 through July find information on the degree of change in specific
2017. (See Part A of the Supplement, available at instrument scores or subscores that would indicate clin-
Annals.org, for search strategies.) We identified studies ical importance. (Part B of the Supplement shows the
published before 2009 by citation searching relevant information we did find about clinically important
systematic reviews. changes in specific instrument scores.) In addition, re-
Study Selection sults were measured, analyzed, and reported in many
Two investigators independently reviewed titles different ways.
and abstracts of search results and screened the full When sufficient data were available (from more
text of potentially eligible references. Disagreements than 1 study or 1 study with ≥ 500 participants), 1 in-
about eligibility were resolved by consensus. We in- vestigator assessed the strength of evidence for unique
cluded randomized controlled trials of physical activity comparisons. These assessments were confirmed
interventions with any sample size and large (n > 500) through consensus. We assessed strength of evidence
prospective quasi-experimental cohort studies with by using 5 required domains: study limitations (risk of
comparator groups if they enrolled adults without diag- bias of eligible studies for a given comparison), direct-
nosed cognitive impairments, had follow-up of at least ness (single, direct link between intervention and out-
6 months, were published in English, and reported 1 of come), consistency (similarity of effect direction and
our preselected primary or intermediate outcomes. We size), precision (degree of certainty around an estimate
excluded trials enrolling pure subgroups of patients that includes attention to small sample sizes with power
with major medical conditions or conditions that may to detect only large differences), and reporting bias
explain changes in cognitive function (namely stroke, (19). On the basis of these factors, the overall strength
Parkinson disease, cancer, and traumatic brain injury). of evidence for each outcome from a given intervention
Our main outcomes of interest were MCI or de- was rated as high, moderate, low, or insufficient.
mentia. Intermediate outcomes included measures of
cognitive function assessed by instruments that tested Role of the Funding Source
cognition across several domains or those that specifi- This review was funded by the National Institute on
cally tested executive function, attention, and process- Aging and AHRQ. These agencies and members of the
ing speed, or memory. Intermediate outcomes were National Academies Committee on Preventing Demen-
categorized as follows: broad measures intended to tia and Cognitive Impairment helped refine the scope
capture several cognitive domains that were either and reviewed a draft report of findings. The authors are
brief cognitive tests (category 1) or more comprehen- solely responsible for the content preparation, writing
sive multidomain neuropsychological tests (category of the manuscript, and decision to submit the manu-
2) and domain-specific neuropsychological tests or script for publication.
subscales of broader instruments that assessed execu-
tive function, attention, and processing speed (cate-
gory 3) or memory (category 4). Part B of the Supple-
ment shows a list of the intermediate outcomes
RESULTS
reported from the studies and our categorization of Of 32 eligible studies that compared interventions
those outcomes. using physical activity components with an inactive con-
trol in adults without a cognitive impairment diagnosis
Data Extraction and Quality Assessment (20 –51), 16 —all of which were randomized trials—were
One reviewer extracted the study population, treat- considered to have low to medium risk of bias (20, 23,
ment characteristics, and funding source from all eligi- 25, 29, 30, 33, 35–37, 39, 40, 43, 45, 46, 48, 50). Inac-
ble studies. Risk of bias was assessed independently by tive controls in the trials with low to medium risk of bias
2 investigators using an instrument developed with included waitlist, usual care, no-intervention, and atten-
guidance from the Agency for Healthcare Research and tion (that is, education and information) groups. Most
Quality (AHRQ) (18). Risk of bias for each reported out- trials were government funded. Most studies enrolled
come was rated as low, medium, or high on the basis of older adults; some limited enrollment to men or
adequacy of randomization and allocation conceal- women. Total sample sizes ranged from 42 to 1635
ment, masking, attrition, use of intention-to-treat analy- participants. Trials rarely reported adverse effects;
ses, selectiveness of outcome reporting, and confi- those that did showed no differences between groups,
dence that results were believable given limitations. with 1 exception. Intervention components, frequency,
Outcomes and adverse effects were extracted from el- and duration varied. (Part C of the Supplement con-
igible trials with low or moderate risk of bias, and a tains the literature flow diagram; part D contains evi-
second investigator checked the extraction. dence tables.)
Annals.org Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 31

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 REVIEW Physical Activity to Prevent Cognitive Decline and Dementia

The Table shows overall conclusions and strength- Lautenschlager and colleagues (30) (n = 170) re-
of-evidence ratings. Details of studies considered to ported dementia diagnoses and showed that partici-
have low to medium risk of bias are described later. For pants assigned to aerobic training were less likely to
any cognitive outcome, evidence was insufficient to receive a dementia diagnosis than those assigned to
draw conclusions about most interventions (aerobic the attention control group at 18-month follow-up
training, resistance training, tai chi, physical activity with (change from baseline at 18 months in Clinical Demen-
diet, and physical activity with a cognitive component). tia Rating Sum of Boxes scores: ⫺0.33 [CI, ⫺0.46 to
Low-strength evidence showed that multicomponent ⫺0.2] vs. ⫺0.20 [CI, ⫺0.33 to ⫺0.03] point; P = 0.050).
physical activity interventions of 1 to 2 years did not Eleven of 35 reported results for intermediate out-
improve multidomain neurologic performance; execu- comes from the 6 trials showed a statistically significant
tive function, attention, and processing speed; or mem- benefit with aerobic training compared with an atten-
ory compared with an attention control. Low-strength tion control, whereas 24 of 35 showed no statistically
evidence showed that an intervention combining phys- significant differences between groups.
ical activity, diet, and cognitive training benefited mul-
tidomain neuropsychological test performance and
executive function, attention, and processing speed Resistance Training
compared with an attention control; however, evidence Three trials (n = 315) examined the effectiveness of
was insufficient to draw conclusions about the efficacy resistance training in preventing cognitive decline (23,
of this intervention on memory. Moderate-strength ev- 29, 48). Studies enrolled prefrail and frail adults older
idence showed that more participants in the interven- than 65 years (48), sedentary men aged 65 to 75 years
tion than the control groups had musculoskeletal pain. with a minimum MMSE score of 24 points (23), and
sedentary older adults with at least 1 disability (29).
Physical Activity Interventions Mean age of participants was in the early 70s. In an
Multicomponent Physical Activity analysis of combined data from 2 separate trials, van de
Four trials (n = 1885) with low to medium risk of Rest and colleagues (48) compared supervised resis-
bias examined multicomponent physical activity inter- tance training twice weekly with usual care. Both trials
ventions. Components included flexibility, strength, randomly assigned participants to receive protein sup-
balance, endurance, and aerobic training (36, 45, 46, plements or placebo; 1 trial also incorporated resis-
50). Enrollment criteria varied by trial. Sink and col- tance training. Both trials lasted for 24 weeks. Cassilhas
leagues (45) and Williamson and colleagues (50) en- and colleagues (23) randomly assigned participants to
rolled sedentary adults older than 70 years, most of 1 of 3 groups: attention control, high-resistance train-
whom were white women. Mean Modified Mini-Mental ing, and low-resistance training. Lachman and col-
State Examination (MMSE) scores were higher than 90 leagues (29) randomly assigned participants to the
points (on a scale of 0 to 100 points). Taylor-Piliae and Strong for Life program (home-based, video-directed
colleagues (46) enrolled adults, mostly white college- resistance training) or a waitlist control group.
educated women, older than 60 years. Napoli and col- No trial reported diagnostic outcomes. Less than
leagues (36) enrolled frail, obese older adults, most of one third of the results for executive function, attention,
whom were white women; mean Modified MMSE score and processing speed (32%) and for memory (27%) fa-
was 96 points. Interventions during the trials lasted vored the intervention. Most positive results arose from
from 6 months to 2 years. 1 small 3-group trial (23).
Sink and colleagues (45) (n = 1635) reported diag-
nostic outcomes that showed no difference in the inci-
Tai Chi
dence of MCI (odds ratio, 1.14 [95% CI, 0.79 to 1.62]) or
One small trial (n = 93) with moderate risk of bias
dementia (odds ratio, 0.96 [CI, 0.57 to 1.63]) between
compared tai chi with an attention control in adults
groups at 2 years. A wide range of neuropsychological
aged 60 to 79 years with education-adjusted Chinese
tests were used to assess intermediate outcomes
MMSE scores higher than 26 points (46). Diagnostic
across the 4 trials. Only 3 of 25 comparisons showed a
outcomes and adverse events were not reported. One
statistically significant benefit with multicomponent
of 2 outcomes for executive function, attention, and
physical activity interventions compared with the atten-
processing speed showed a benefit with tai chi com-
tion controls.
pared with the attention control, whereas the other did
not.
Aerobic Activity Physical Activity Combined With Other
Six trials (n = 531) with low to medium risk of bias Interventions
compared aerobic training with an attention control Physical Activity and Diet
(20, 30, 35, 39, 40, 43). Four studies enrolled healthy Two trials (n = 79) compared interventions combin-
older adults (35, 39, 40, 43), whereas 1 trial enrolled ing physical activity and diet with attention controls (33,
sedentary older men (20) and another enrolled adults 36). Both studies enrolled overweight or obese adults.
with MMSE scores of at least 24 points (on a scale of 0 Napoli and colleagues (36) randomly assigned obese,
to 30 points) who also reported memory challenges sedentary adults with a minimum MMSE score of 24
(30). points to a calorie-restricted diet plus multicomponent
32 Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 Annals.org

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 Physical Activity to Prevent Cognitive Decline and Dementia REVIEW

Table. Conclusions: Physical Activity Versus Inactive Comparisons in Adults

Outcome Conclusion Strength of Evidence (Justification)

Multicomponent physical activity vs. attention control
(k ⴝ 4; n ⴝ 1885)
Dementia Unable to draw conclusion Insufficient (medium study limitations,
imprecise, unknown consistency)
MCI Unable to draw conclusion Insufficient (medium study limitations,
imprecise, unknown consistency)
Brief cognitive test performance No benefit (n = 155; 6 mo–1 y) Insufficient (medium study limitations,
indirect, imprecise, inconsistent)
Multidomain neuropsychological performance No benefit (n = 1635; 2 y) Low (medium study limitations,
indirect, unknown consistency)
Executive function/attention/processing speed No benefit (n = 1885; 6 mo–1 y) Low (medium study limitations,
indirect, imprecise)
Memory No benefit (n = 1836; 6 mo–1 y) Low (medium study limitations,
indirect, imprecise)

Aerobic training vs. attention control (k ⴝ 6; n ⴝ 531)

Dementia Limited data Insufficient (limited data)
MCI No data Insufficient (no data)
Brief cognitive test performance No benefit (n = 162; 6 mo–1 y) Insufficient (medium study limitations,
indirect, imprecise)
Multidomain neuropsychological performance Limited data Insufficient (limited data)
Executive function/attention/processing speed Unable to draw conclusion Insufficient (medium study limitations,
indirect, imprecise, inconsistent)
Memory Unable to draw conclusion Insufficient (medium study limitations,
indirect, imprecise, inconsistent)

Resistance training vs. attention control (k ⴝ 3; n ⴝ 315)

Dementia No data Insufficient (no data)
MCI No data Insufficient (no data)
Brief cognitive test performance No data Insufficient (no data)
Multidomain neuropsychological performance No data Insufficient (no data)
Executive function/attention/processing speed No benefit (n = 120; 6 mo) Insufficient (medium study limitations,
indirect, imprecise, inconsistent)
Memory No benefit (n = 172; 6 mo) Insufficient (medium study limitations,
indirect, imprecise, inconsistent)

Tai chi vs. attention control (k ⴝ 1; n ⴝ 93)

Dementia No data Insufficient (no data)
MCI No data Insufficient (no data)
Brief cognitive test performance No data Insufficient (no data)
Multidomain neuropsychological performance No data Insufficient (no data)
Executive function/attention/processing speed Limited data Insufficient (limited data)
Memory No data Insufficient (no data)

Physical activity plus diet vs. attention control

(k ⴝ 2; n ⴝ 79)
Dementia No data Insufficient (no data)
MCI No data Insufficient (no data)
Brief cognitive test performance Limited data Insufficient (limited data)
Multidomain neuropsychological performance No data Insufficient (no data)
Executive function/attention/processing speed Unable to draw conclusion Insufficient (medium study limitations,
indirect, imprecise, inconsistent)
Memory Limited data Insufficient (limited data)

Physical activity plus protein supplementation vs.

attention control (k ⴝ 1; n ⴝ 58)
Dementia No data Insufficient (no data)
MCI No data Insufficient (no data)
Brief cognitive test performance No data Insufficient (no data)
Multidomain neuropsychological performance No data Insufficient (no data)
Executive function/attention/processing speed Limited data Insufficient (limited data)
Memory Limited data Insufficient (limited data)

Physical activity and cognitive training (k ⴝ 1; n ⴝ 134)

Dementia No data Insufficient (no data)
MCI No data Insufficient (no data)
Brief cognitive test performance Limited data Insufficient (limited data)
Multidomain neuropsychological performance Limited data Insufficient (limited data)
Executive function/attention/processing speed No data Insufficient (no data)
Memory No data Insufficient (no data)
Continued on following page

Annals.org Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 33

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 REVIEW Physical Activity to Prevent Cognitive Decline and Dementia

Table—Continued

Outcome Conclusion Strength of Evidence (Justification)

Physical activity, diet, and cognitive training vs. attention
control (k ⴝ 1; n ⴝ 1260)
Dementia No data Insufficient (no data)
MCI No data Insufficient (no data)
Brief cognitive test performance No data Insufficient (no data)
Multidomain neuropsychological performance Intervention comprising Low (indirect, unknown consistency)
physical activity, diet, and
cognitive training improves
multidomain neuropsycho-
logical test performance
(n = 1260; 2 y)
Executive function/attention/processing speed Intervention comprising Low (indirect, unknown consistency)
physical activity, diet, and
cognitive training improves
executive function/attention/
processing speed
(n = 1260; 2 y)
Memory Unable to draw conclusion Insufficient (indirect, imprecise,
(n = 1260; 2 y) inconsistent)
Musculoskeletal pain Musculoskeletal pain is higher Moderate (unknown consistency)
with intervention comprising
physical activity, diet, and
cognitive training
MCI = mild cognitive impairment.

exercise for 90 minutes, 3 times weekly, for 1 year. Mar- followed instructions for pattern changes according to
tin and colleagues (33) randomly assigned overweight cues in the music as the cognitive training component.
adults aged 25 to 50 years to a calorie-restricted diet Diagnostic outcomes were not reported. Brief cog-
plus group aerobic training for 6 months. Neither trial nitive test performance was not statistically different be-
reported diagnostic outcomes. Only 2 of the 22 cogni- tween groups after the intervention. Results from 1 of 2
tive tests reported showed a benefit with combined tests of executive function, attention, and processing
physical activity plus diet versus the attention controls. speed showed a postintervention difference between
groups favoring the intervention, whereas the other
Physical Activity and Protein Supplementation
test did not.
In an analysis of combined data from 2 separate
trials, van de Rest and colleagues (48) compared phys- Physical Activity, Diet, and Cognitive Training
ical activity and protein supplementation with usual
In the FINGER (Finnish Geriatric Intervention Study
care (n = 58). The original trials did not report out-
to Prevent Cognitive Impairment and Disability) trial,
comes pertinent to our review (52, 53). Both trials ran-
Ngandu and colleagues (37) (n = 1260) compared
domly assigned participants to receive protein supple-
an intervention program comprising components ad-
ments or placebo; 1 trial also incorporated resistance
dressing several risk factors simultaneously with an at-
training. In both trials, the interventions lasted 24
tention control. Adults aged 60 to 77 years with CAIDE
weeks. Participants were older than 65 years and pre-
(Cardiovascular Risk Factors, Aging and Dementia) de-
frail or frail according to measures of physical function.
mentia risk scores of at least 6 points and cognition
Trial inclusion criteria did not specifically address cog-
near or slightly below that expected for age were ran-
nitive status; mean MMSE scores at baseline were
domly assigned to receive nutritional counseling, mul-
higher than 27 points (52, 53).
ticomponent exercise, and cognitive training or an
One of 11 tests of executive function, attention,
attention control (general health advice). The interven-
and processing speed and none of 6 memory tests
tion involved individual and group sessions to foster
showed a statistically significant difference favoring
tailored dietary changes, 1 to 3 aerobic exercise and 2
physical activity and protein supplementation after the
to 5 resistance training sessions per week, and group
intervention.
and individual cognitive training. Intention-to-treat
analysis at 2 years included 1190 participants (94% of
Physical Activity Plus Cognitive Training those randomly assigned).
Hars and colleagues (25) (n = 134) compared phys- Diagnostic outcomes were not reported. Improve-
ical activity plus cognitive training with a waitlist control. ment in multidomain neuropsychological test perfor-
Adults aged 65 years and older with an increased fall mance was 25% greater with the intervention versus the
risk were randomly assigned to a structured, music- control at 2 years. The mean change in 2 measures of
based exercise program or a waitlist group (25). The executive function, attention, and processing speed
intervention consisted of 6 months of weekly 60-minute also was greater with the intervention versus the con-
multitasking exercise classes during which participants trol. Mean improvement in executive function, atten-
34 Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 Annals.org

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 Physical Activity to Prevent Cognitive Decline and Dementia REVIEW
tion, and processing speed was 83% greater with 1 However, our analysis and conclusions may not be as
measure and 150% higher with the other. No statisti- optimistic as others. This difference seems largely to be
cally significant difference was observed between the the result of interpretation. We saw many results with
changes in memory outcomes. Several adverse events no statistically significant differences compared with in-
were reported, including musculoskeletal pain, stress, active controls. That minority of statistically significant
challenges related to time commitment, death, and results often was interpreted as evidence of effective-
“other.” However, only musculoskeletal pain affected as ness in other reviews. Our review also had stricter inclu-
many as 5% of the participants in either group. Results sion criteria, requiring a minimum follow-up of 6
showed a difference between groups, with participants months, so many other reviews included studies not
receiving the intervention reporting a significantly eligible for this one.
higher rate of pain (5%) than those in the control group Our review has several limitations. Most trials were
(0%). small and did not follow participants long term. The
type, frequency, intensity, and duration of physical ac-
tivity interventions varied widely. Most trials enrolled
DISCUSSION sedentary adults, yet how sedentary status was defined
This systematic review found that evidence for the or measured was not always clear. Trials used many
effectiveness of single-component physical activity in- different cognitive tests and often repeated tests more
terventions in preventing cognitive decline is largely in- frequently than appropriate. Many studies measured
sufficient. The only exception is low-strength evidence cognitive outcomes with various instruments and did
of no effect for multicomponent physical activity in pre- not correct for multiple comparisons. Ideally, our anal-
venting cognitive decline. The intervention with suffi- ysis would assess effect size and the point at which in-
cient evidence of benefit on cognitive function com- terventions demonstrated a statistically significant ef-
prised physical activity, diet, and cognitive training, fect. However, because many studies did not use
addressing several risk factors simultaneously. instruments with demonstrated clinical thresholds or
Although our results predominantly showed no provide sufficient information to calculate standard ef-
cognitive benefit from interventions comprised solely fect sizes, we relied on statistical significance. Often
of physical activity, several cognitive results favored when results showed no significant differences, it was
physical activity over controls for aerobic training, resis- unclear whether the insignificance was the result of a
tance training, and multicomponent physical activity in- lack of effectiveness or of underpowered studies. In
terventions. We believe that those findings provide a these cases, the imprecision of the estimates led us to
signal that physical activity offers cognitive benefit but assess the evidence as insufficient. Checks for treat-
that the studies conducted were not long enough or ment fidelity rarely were mentioned, and participant
sufficiently powered to show the true long-term effect adherence was not always reported. In addition, the
of a physically active lifestyle, which many cohort stud- question of publication bias was challenging to assess
ies suggest is beneficial. To be effective, regular phys- for this topic. Trials of physical activity interventions of-
ical activity may need to begin earlier in life and be ten measured cognitive outcomes as secondary to oth-
sustained as a lifestyle. Short-term interventions begun ers, such as cardiovascular disease outcomes. Studies
after decades of high-risk behavior likely are insufficient of physical activity interventions, especially older trials,
to reduce dementia incidence. are not consistently registered. We did not identify un-
Physical activity may slow cognitive decline directly published trials at ClinicalTrials.gov potentially eligible
through increased blood flow or indirectly by reducing for this review.
other risk factors (namely, cardiovascular disease, obe- Future research should seek to identify population
sity, and diabetes). The cognitive benefits demon- traits (such as age at which physical activity began, in-
strated in the FINGER trial are intriguing. Because the tervention duration, and baseline activity level) and
intervention included components addressing several intervention characteristics (such as type, frequency,
risk factors, isolating the benefits attributable to the duration, and intensity) that trend toward positive out-
physical activity component was not possible. However, comes among these trials. Studies then could be de-
given the complex relationship between identified and signed appropriately to test identified characteristics
interrelated risk and factors that protect against cogni- for research and knowledge purposes. Long-term trials
tive decline and dementia, interventions addressing enrolling younger adults with interventions sustained
many risk factors at once might offer the best approach for longer periods would benefit the field and provide
to successful prevention, especially if initiated in older important insight on prevention. On the basis of the
adults already at higher risk for MCI and dementia. Al- evidence from this review, it seems that preventing
though evidence of effectiveness was low strength, cognitive decline in older adults may be best accom-
harms were not serious, and benefits from physical ac- plished by addressing several lifestyle risk factors si-
tivity, diet, and cognitive training for other health con- multaneously, as demonstrated in the FINGER trial.
ditions and functioning are more firmly established. Although a physically active lifestyle often is pro-
Therefore, clinical approaches to incorporate similar in- posed as a way to prevent cognitive decline and re-
terventions likely are warranted. duce risk for cognitive impairment and dementia, we
Our results are consistent with those of previous found little evidence to determine whether single-
systematic reviews identified in our searches (8 –15). component physical activity interventions offer cogni-
Annals.org Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 35

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 REVIEW Physical Activity to Prevent Cognitive Decline and Dementia

tive benefits. However, clinical practice largely encour- Heart Study. N Engl J Med. 2016;374:523-32. [PMID: 26863354] doi:
ages physical activity to prevent or manage other 10.1056/NEJMoa1504327
chronic conditions, and this practice should continue, 5. Plassman BL, Langa KM, Fisher GG, Heeringa SG, Weir DR, Of-
stedal MB, et al. Prevalence of cognitive impairment without demen-
because it may offer benefits for preventing cognitive
tia in the United States. Ann Intern Med. 2008;148:427-34. [PMID:
decline as well. Recommending multidomain interven- 18347351]
tions may prove even more beneficial. 6. Williams JW, Plassman BL, Burke J, Holsinger T, Benjamin S. Pre-
venting Alzheimer's Disease and Cognitive Decline. (Prepared by
From University of Minnesota, Minneapolis, Minnesota; Brown the Duke Evidence-based Practice Center under contract no. HHSA
University, Providence, Rhode Island; Minneapolis VA Health 290-2007-10066-I.) Rockville: Agency for Healthcare Research and
Care System, Minneapolis, Minnesota; and HealthPartners, Quality; 2010.
Minneapolis, Minnesota. 7. Kelley AS, McGarry K, Gorges R, Skinner JS. The burden of health
care costs for patients with dementia in the last 5 years of life. Ann
Intern Med. 2015;163:729-36. [PMID: 26502320] doi:10.7326/M15-
Disclaimer: Findings and conclusions are those of the authors, 0381
who are responsible for the article's contents; findings and 8. Blondell SJ, Hammersley-Mather R, Veerman JL. Does physical
conclusions do not necessarily represent the views of AHRQ. activity prevent cognitive decline and dementia?: A systematic re-
No statement in this report should be construed as an official view and meta-analysis of longitudinal studies. BMC Public Health.
position of AHRQ or the U.S. Department of Health and Hu- 2014;14:510. [PMID: 24885250] doi:10.1186/1471-2458-14-510
man Services. 9. Beckett MW, Ardern CI, Rotondi MA. A meta-analysis of prospec-
tive studies on the role of physical activity and the prevention of
Alzheimer's disease in older adults. BMC Geriatr. 2015;15:9. [PMID:
Financial Support: This manuscript is based on research con-
25887627] doi:10.1186/s12877-015-0007-2
ducted by the Minnesota Evidence-based Practice Center un-
10. Bellou V, Belbasis L, Tzoulaki I, Middleton LT, Ioannidis JP, Evan-
der AHRQ contract 290-2015-00008-I. gelou E. Systematic evaluation of the associations between environ-
mental risk factors and dementia: an umbrella review of systematic
Disclosures: Dr. Jutkowitz reports grants from AHRQ outside reviews and meta-analyses. Alzheimers Dement. 2017;13:406-18.
the submitted work. Drs. Ratner, Hemmy, and Barclay report [PMID: 27599208] doi:10.1016/j.jalz.2016.07.152
grant support from AHRQ during the conduct of the study. 11. Beydoun MA, Beydoun HA, Gamaldo AA, Teel A, Zonderman
Authors not named here have disclosed no conflicts of AB, Wang Y. Epidemiologic studies of modifiable factors associated
interest. Disclosures can also be viewed at www.acponline with cognition and dementia: systematic review and meta-analysis.
.org/authors/icmje/ConflictOfInterestForms.do?msNum=M17 BMC Public Health. 2014;14:643. [PMID: 24962204] doi:10
.1186/1471-2458-14-643
-1528.
12. Hamer M, Chida Y. Physical activity and risk of neurodege-
nerative disease: a systematic review of prospective evidence.
Reproducible Research Statement: Study protocol: Avail- Psychol Med. 2009;39:3-11. [PMID: 18570697] doi:10.1017
able at https://effectivehealthcare.ahrq.gov/search-for /S0033291708003681
-guides-reviews-and-reports/?pageaction=displayproduct& 13. Williams JW, Plassman BL, Burke J, Benjamin S. Preventing Alz-
productID=2202. Statistical code: Not applicable. Data set: heimer's disease and cognitive decline. Evid Rep Technol Assess
See Systematic Review Data Repository at https://srdr.ahrq (Full Rep). 2010;193:1-727.
.gov/. 14. Angevaren M, Aufdemkampe G, Verhaar HJ, Aleman A, Van-
hees L. Physical activity and enhanced fitness to improve cognitive
function in older people without known cognitive impairment.
Requests for Single Reprints: Michelle Brasure, PhD, MSPH, Cochrane Database Syst Rev. 2008:CD005381. [PMID: 18646126]
MLIS, Division of Health Policy and Management, University of doi:10.1002/14651858.CD005381.pub3
Minnesota, 420 Delaware Street Southeast, Mayo Memorial 15. Smith PJ, Blumenthal JA, Hoffman BM, Cooper H, Strauman TA,
Building D351, Minneapolis, MN 55455; e-mail, brasu001 Welsh-Bohmer K, et al. Aerobic exercise and neurocognitive perfor-
@umn.edu. mance: a meta-analytic review of randomized controlled trials. Psy-
chosom Med. 2010;72:239-52. [PMID: 20223924] doi:10.1097/PSY
.0b013e3181d14633
Current author addresses and author contributions are avail-
16. Evidence-based Practice Center Systematic Review Protocol: In-
able at Annals.org.
terventions for Preventing Cognitive Decline, Mild Cognitive Impair-
ment, and Alzheimer's Disease. Rockville: Agency for Healthcare Re-
search and Quality; 2015.
References 17. Kane R, Butler M, Fink H, Brasure M, Davila H, Desai P, et al.
1. Prince M, Ali GC, Guerchet M, Prina AM, Albanese E, Wu YT. Interventions to Prevent Age-Related Cognitive Decline, Mild Cogni-
Recent global trends in the prevalence and incidence of dementia, tive Impairment, and Clinical Alzheimer's-Type Dementia. Compara-
and survival with dementia. Alzheimers Res Ther. 2016;8:23. [PMID: tive Effectiveness Review no. 188. (Prepared by the Minnesota
27473681] doi:10.1186/s13195-016-0188-8 Evidence-based Practice Center under contract no. 290-2015-
2. Prince M, Bryce R, Albanese E, Wimo A, Ribeiro W, Ferri CP. The 00008-I.) AHRQ publication no. 17-EHC008-EF. Rockville: Agency for
global prevalence of dementia: a systematic review and metaanaly- Healthcare Research and Quality; February 2017. Accessed at https:
sis. Alzheimers Dement. 2013;9:63-75. [PMID: 23305823] doi:10 //ahrq-ehc-application.s3.amazonaws.com/media/pdf/cognitive
.1016/j.jalz.2012.11.007 -decline_research-2017.pdf on 1 November 2017.
3. Langa KM, Larson EB, Crimmins EM, Faul JD, Levine DA, Kabeto 18. Viswanathan M, Ansari M, Berkman N, Chang S, Hartling L,
MU, et al. A comparison of the prevalence of dementia in the United McPheeters LM, et al. Assessing the risk of bias of individual studies
States in 2000 and 2012. JAMA Intern Med. 2017;177:51-8. [PMID: in systematic reviews of health care interventions. In: Agency for
27893041] doi:10.1001/jamainternmed.2016.6807 Healthcare Research and Quality Methods Guide for Comparative
4. Satizabal CL, Beiser AS, Chouraki V, Chêne G, Dufouil C, Seshadri Effectiveness Reviews. Rockville: Agency for Healthcare Research
S. Incidence of dementia over three decades in the Framingham and Quality; March 2012.

36 Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 Annals.org

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 Physical Activity to Prevent Cognitive Decline and Dementia REVIEW
19. Berkman ND, Lohr KN, Ansari MT, Balk EM, Kane R, McDonagh 35. Muscari A, Giannoni C, Pierpaoli L, Berzigotti A, Maietta P, Foschi
M, et al. Grading the strength of a body of evidence when assessing E, et al. Chronic endurance exercise training prevents aging-related
health care interventions: an EPC update. J Clin Epidemiol. 2015;68: cognitive decline in healthy older adults: a randomized controlled
1312-24. [PMID: 25721570] doi:10.1016/j.jclinepi.2014.11.023 trial. Int J Geriatr Psychiatry. 2010;25:1055-64. [PMID: 20033904] doi:
20. Antunes HK, De Mello MT, Santos-Galduróz RF, Galduróz JC, 10.1002/gps.2462
Lemos VA, Tufik S, et al. Effects of a physical fitness program on 36. Napoli N, Shah K, Waters DL, Sinacore DR, Qualls C, Villareal DT.
memory and blood viscosity in sedentary elderly men. Braz J Med Effect of weight loss, exercise, or both on cognition and quality of life
Biol Res. 2015;48:805-12. [PMID: 26222648] doi:10.1590/1414 in obese older adults. Am J Clin Nutr. 2014;100:189-98. [PMID:
-431X20154529 24787497] doi:10.3945/ajcn.113.082883
21. Bun S, Ikejima C, Kida J, Yoshimura A, Lebowitz AJ, Kakuma T, 37. Ngandu T, Lehtisalo J, Solomon A, Levälahti E, Ahtiluoto S, An-
et al. A combination of supplements may reduce the risk of Alzhei- tikainen R, et al. A 2 year multidomain intervention of diet, exercise,
mer's disease in elderly Japanese with normal cognition. J Alzhei- cognitive training, and vascular risk monitoring versus control to pre-
mers Dis. 2015;45:15-25. [PMID: 25524956] doi:10.3233/JAD vent cognitive decline in at-risk elderly people (FINGER): a ran-
-142232 domised controlled trial. Lancet. 2015;385:2255-63. [PMID:
22. Carlson MC, Saczynski JS, Rebok GW, Seeman T, Glass TA, 25771249] doi:10.1016/S0140-6736(15)60461-5
McGill S, et al. Exploring the effects of an “everyday” activity program
38. Nguyen MH, Kruse A. A randomized controlled trial of Tai chi for
on executive function and memory in older adults: Experience
balance, sleep quality and cognitive performance in elderly Vietnam-
Corps. Gerontologist. 2008;48:793-801. [PMID: 19139252]
ese. Clin Interv Aging. 2012;7:185-90. [PMID: 22807627] doi:10
23. Cassilhas RC, Viana VA, Grassmann V, Santos RT, Santos RF,
.2147/CIA.S32600
Tufik S, et al. The impact of resistance exercise on the cognitive func-
39. Oken BS, Zajdel D, Kishiyama S, Flegal K, Dehen C, Haas M,
tion of the elderly. Med Sci Sports Exerc. 2007;39:1401-7. [PMID:
17762374] et al. Randomized, controlled, six-month trial of yoga in healthy se-
24. Evers A, Klusmann V, Schwarzer R, Heuser I. Improving cognition niors: effects on cognition and quality of life. Altern Ther Health Med.
by adherence to physical or mental exercise: a moderated mediation 2006;12:40-7. [PMID: 16454146]
analysis. Aging Ment Health. 2011;15:446-55. [PMID: 21500011] doi: 40. Okumiya K, Matsubayashi K, Wada T, Kimura S, Doi Y, Ozawa T.
10.1080/13607863.2010.543657 Effects of exercise on neurobehavioral function in community-
25. Hars M, Herrmann FR, Gold G, Rizzoli R, Trombetti A. Effect of dwelling older people more than 75 years of age. J Am Geriatr Soc.
music-based multitask training on cognition and mood in older 1996;44:569-72. [PMID: 8617907]
adults. Age Ageing. 2014;43:196-200. [PMID: 24212920] doi:10 41. Oswald WD, Gunzelmann T, Rupprecht R, Hagen B. Differential
.1093/ageing/aft163 effects of single versus combined cognitive and physical training
26. Hötting K, Reich B, Holzschneider K, Kauschke K, Schmidt T, with older adults: the SimA study in a 5-year perspective. Eur J Age-
Reer R, et al. Differential cognitive effects of cycling versus stretch- ing. 2006;3:179. [PMID: 28794762] doi:10.1007/s10433-006-0035-z
ing/coordination training in middle-aged adults. Health Psychol. 42. Rosano C, Venkatraman VK, Guralnik J, Newman AB, Glynn NW,
2012;31:145-55. [PMID: 21895371] doi:10.1037/a0025371 Launer L, et al. Psychomotor speed and functional brain MRI 2 years
27. Klusmann V, Evers A, Schwarzer R, Schlattmann P, Reischies FM, after completing a physical activity treatment. J Gerontol A Biol Sci
Heuser I, et al. Complex mental and physical activity in older women Med Sci. 2010;65:639-47. [PMID: 20348185] doi:10.1093/gerona
and cognitive performance: a 6-month randomized controlled trial. /glq038
J Gerontol A Biol Sci Med Sci. 2010;65:680-8. [PMID: 20418350] 43. Ruscheweyh R, Willemer C, Krüger K, Duning T, Warnecke T,
doi:10.1093/gerona/glq053 Sommer J, et al. Physical activity and memory functions: an interven-
28. Komulainen P, Kivipelto M, Lakka T, Savonen K, Hassinen M, tional study. Neurobiol Aging. 2011;32:1304-19. [PMID: 19716631]
Kiviniemi V, et al. Exercise, fitness and cognition—a randomised con- doi:10.1016/j.neurobiolaging.2009.08.001
trolled trial in older individuals: the DR's EXTRA study. Eur Geriatr 44. Satoh M, Ogawa J, Tokita T, Nakaguchi N, Nakao K, Kida H, et al.
Med. 2010;1:266-72. The effects of physical exercise with music on cognitive function of
29. Lachman ME, Neupert SD, Bertrand R, Jette AM. The effects of elderly people: Mihama-Kiho project. PLoS One. 2014;9:e95230.
strength training on memory in older adults. J Aging Phys Act. 2006; [PMID: 24769624] doi:10.1371/journal.pone.0095230
14:59-73. [PMID: 16648652] 45. Sink KM, Espeland MA, Castro CM, Church T, Cohen R, Dodson
30. Lautenschlager NT, Cox KL, Flicker L, Foster JK, van Bockxmeer JA, et al; LIFE Study Investigators. Effect of a 24-month physical ac-
FM, Xiao J, et al. Effect of physical activity on cognitive function in tivity intervention vs health education on cognitive outcomes in sed-
older adults at risk for Alzheimer disease: a randomized trial. JAMA. entary older adults: The LIFE randomized trial. JAMA. 2015;314:781-
2008;300:1027-37. [PMID: 18768414] doi:10.1001/jama.300.9.1027
90. [PMID: 26305648] doi:10.1001/jama.2015.9617
31. Lehtisalo J, Lindström J, Ngandu T, Kivipelto M, Ahtiluoto S,
46. Taylor-Piliae RE, Newell KA, Cherin R, Lee MJ, King AC, Haskell
Ilanne-Parikka P, et al; Finnish Diabetes Prevention Study. Associa-
WL. Effects of Tai Chi and Western exercise on physical and cogni-
tion of long-term dietary fat intake, exercise, and weight with later
tive functioning in healthy community-dwelling older adults. J Aging
cognitive function in the Finnish Diabetes Prevention Study. J Nutr
Phys Act. 2010;18:261-79. [PMID: 20651414]
Health Aging. 2016;20:146-54. [PMID: 26812510] doi:10.1007
/s12603-015-0565-1 47. Tesky VA, Thiel C, Banzer W, Pantel J. Effects of a group program
32. Liu-Ambrose T, Donaldson MG, Ahamed Y, Graf P, Cook WL, to increase cognitive performance through cognitively stimulating
Close J, et al. Otago home-based strength and balance retraining leisure activities in healthy older subjects. GeroPsych. 2011;24:83-
improves executive functioning in older fallers: a randomized con- 92.
trolled trial. J Am Geriatr Soc. 2008;56:1821-30. [PMID: 18795987] 48. van de Rest O, van der Zwaluw NL, Tieland M, Adam JJ, Hiddink
doi:10.1111/j.1532-5415.2008.01931.x GJ, van Loon LJ, et al. Effect of resistance-type exercise training with
33. Martin CK, Anton SD, Han H, York-Crowe E, Redman LM, Ravus- or without protein supplementation on cognitive functioning in frail
sin E, et al. Examination of cognitive function during six months of and pre-frail elderly: secondary analysis of a randomized, double-
calorie restriction: results of a randomized controlled trial. Rejuvena- blind, placebo-controlled trial. Mech Ageing Dev. 2014;136-137:85-
tion Res. 2007;10:179-90. [PMID: 17518698] 93. [PMID: 24374288] doi:10.1016/j.mad.2013.12.005
34. Mortimer JA, Ding D, Borenstein AR, DeCarli C, Guo Q, Wu Y, 49. Williams P, Lord SR. Effects of group exercise on cognitive func-
et al. Changes in brain volume and cognition in a randomized trial of tioning and mood in older women. Aust N Z J Public Health. 1997;
exercise and social interaction in a community-based sample of non- 21:45-52. [PMID: 9141729]
demented Chinese elders. J Alzheimers Dis. 2012;30:757-66. [PMID: 50. Williamson JD, Espeland M, Kritchevsky SB, Newman AB, King
22451320] doi:10.3233/JAD-2012-120079 AC, Pahor M, et al; LIFE Study Investigators. Changes in cognitive

Annals.org Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 37

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 REVIEW Physical Activity to Prevent Cognitive Decline and Dementia

function in a randomized trial of physical activity: results of the life- 52. Tieland M, Borgonjen-Van den Berg KJ, van Loon LJ, de
style interventions and independence for elders pilot study. J Geron- Groot LC. Dietary protein intake in community-dwelling, frail,
tol A Biol Sci Med Sci. 2009;64:688-94. [PMID: 19244157] doi:10 and institutionalized elderly people: scope for improvement. Eur J
.1093/gerona/glp014 Nutr. 2012;51:173-9. [PMID: 21562887] doi:10.1007/s00394-011
51. Andrieu S, Guyonnet S, Coley N, Cantet C, Bonnefoy M, Bordes -0203-6
S, et al; MAPT Study Group. Effect of long-term omega 3 polyunsat- 53. Tieland M, Dirks ML, van der Zwaluw N, Verdijk LB, van de Rest
urated fatty acid supplementation with or without multidomain inter- O, de Groot LC, et al. Protein supplementation increases muscle
vention on cognitive function in elderly adults with memory com- mass gain during prolonged resistance-type exercise training in frail
plaints (MAPT): a randomised, placebo-controlled trial. Lancet elderly people: a randomized, double-blind, placebo-controlled
Neurol. 2017;16:377-89. [PMID: 28359749] doi:10.1016/S1474- trial. J Am Med Dir Assoc. 2012;13:713-9. [PMID: 22770932]
4422(17)30040-6 doi:10.1016/j.jamda.2012.05.020

ANNALS VIRTUAL PATIENTS

Annals Virtual Patients are online interactive cases that mirror real patient
care decisions and their consequences, while earning you CME credit and
MOC points.

Visit www.annals.org/virtualpatients to try a free case to see where your

decisions take you.

38 Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018 Annals.org

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018

 Current Author Addresses: Drs. Brasure and Butler, Ms. Desai, Author Contributions: Conception and design: M. Brasure, M.
Ms. Davila, and Ms. Nelson: Division of Health Policy and Man- Butler, H.A. Fink, L.S. Hemmy, J.R. McCarten, R.L. Kane.
agement, University of Minnesota, 420 Delaware Street South- Analysis and interpretation of the data: M. Brasure, P. Desai,
east, Mayo Memorial Building D351, Minneapolis, MN 55455. H. Davila, V.A. Nelson, E. Jutkowitz, M. Butler, H.A. Fink, E.
Mr. Calvert: Division of Epidemiology, University of Minne- Ratner, L.S. Hemmy, J.R. McCarten, T.R. Barclay.
sota, 1300 South 2nd Street, Room 300 West Bank Office Drafting of the article: M. Brasure, P. Desai, C. Calvert, E. Jut-
Building, Minneapolis, MN 55454. kowitz, M. Butler, L.S. Hemmy, T.R. Barclay, R.L. Kane.
Dr. Jutkowitz: Brown University, School of Public Health, Box Critical revision for important intellectual content: M. Brasure,
G-S121-6, 121 South Main Street, 6th Floor, Providence, RI C. Calvert, M. Butler, H.A. Fink, E. Ratner, L.S. Hemmy, J.R.
McCarten, T.R. Barclay.
02912.
Final approval of the article: M. Brasure, P. Desai, H. Davila,
Drs. Fink, Ratner, Hemmy, and McCarten: Geriatric Research
V.A. Nelson, C. Calvert, E. Jutkowitz, M. Butler, H.A. Fink, E.
Education and Clinical Center, VA Health Care System, One
Ratner, L.S. Hemmy, J.R. McCarten, T.R. Barclay.
Veterans Drive, 11-G, Minneapolis, MN 55417.
Provision of study materials or patients: M. Brasure.
Dr. Barclay: Department of Neurology, University of Minne- Statistical expertise: M. Brasure.
sota, 295 Phalen Boulevard, Mailstop 41203C, St. Paul, MN Obtaining of funding: M. Brasure, M. Butler, R.L. Kane.
55130. Administrative, technical, or logistic support: M. Brasure, P.
Desai, H. Davila, V.A. Nelson.
Collection and assembly of data: M. Brasure, P. Desai, H. Da-
vila, V.A. Nelson, C. Calvert, E. Jutkowitz, M. Butler.

Annals.org Annals of Internal Medicine • Vol. 168 No. 1 • 2 January 2018

Downloaded From: https://annals.org/ by a Kaohsiung Medica University User on 07/26/2018