Get Newsletter
AlzRisk Risk Factor OverView
Back

Risk Factor:
Risk Factor Type: Behavior
Current Understanding:
The tables below summarize results from a series of observational studies of cognitive activity in relation to Alzheimer's disease dementia (AD) risk. Overall, these data are consistent with an association between participation in cognitive activities and lower risk of both clinical AD and all-cause dementia. Confounding and reverse causation could have biased some of these results, meaning that they should be interpreted with caution. Nonetheless, our bias sensitivity analysis, shown in an expanded paper, suggests that confounding per se is unlikely to explain the entirety of the findings, although reverse causation remains a concern (Sajeev G, et al. Epidemiology. 2016). In addition, the lack of an agreed-upon definition of cognitive activity as a construct and of its relevant dimensions, along with differences in its operationalization across studies also complicate comparison and synthesis of results. Thus far, the only large randomized trials of cognitive training interventions have shown modest, typically transient benefits on targeted cognitive domains, and it is unclear to what extent these relate to long-term pursuit of personally selected leisure activities. Further, whereas recommendations to remain cognitively engaged (“use it or lose it!”) are unlikely to be harmful, little is known about dose-response relationships, and how issues such as type, duration, intensity and timing of activity influence AD risk. Overall, engaging in activities enjoyable for their own sake seems a safe approach while further information is gathered. For a review of the putative mechanisms by which cognitive activity may influence AD risk and detailed commentary on interpreting these findings in a broader context, please view the Discussion. An expanded review, which includes studies reporting only on all-cause dementia as well, and bias sensitivity analyses, see our published paper, Sajeev G, Weuve J, Jackson JW, VanderWeele TJ, Bennett DA, Grodstein F, Blacker D. Epidemiology. 2016 Sep;27(5):732-42.
Literature Extraction: Search strategy  * New *
Last Search Completed: 16 June 2016


Table 1:   Cognitive activity participation - categorical
Notes These studies evaluate the association between AD risk and participation in leisure-time cognitive activity, considered as a categorical variable. The studies varied in the leisure activities they considered to be "cognitive", and in how they defined and categorized participation.  
  Alzheimer Disease Total Dementia  
Paper Cohort Study Type # Subjects
(% Female)
Average Follow-up Time Exposure Distribution
# of Cases Effect Size 95% CI P-value # of Cases Effect Size 95% CI P-value Ethnicity Age at Start of Follow-up:
Mean (SD)
(Range)
Diagnostic Assessment Covariates & Analysis Comment Paper
Akbaraly, 2009 3C Incidence study reporting hazard ratios (HRs) 5698
(61%)
-
(detail)
Stimulating activities
Low: 38%
Mild: 32%
High: 30%
(detail)

-
-
-
Total: 97

1.00
0.45
0.39

Ref.
0.26-0.77
0.21-0.71

Ref.
0.004
0.002
*

-
-
-
Total: 147

1.00
0.49
0.49

Ref.
0.32-0.75
0.31-0.79

Ref.
0.001
0.003
*
 (detail) 74 (5)
(65 - )
Screening: IST, MMSE

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, ADLI, APOE4, MMSE, DEP, DM, HC, HTN, MS, OS, SP, VD‡
(detail)
Akbaraly, 2009
Akbaraly, 2009 3C Incidence study reporting hazard ratios (HRs) 5698
(61%)
-
(detail)
Passive activities
Low: 39%
Mild: 23%
High: 38%
(detail)

-
-
-
Total: 97

1.00
1.02
0.68

Ref.
0.62-1.69
0.41-1.13

Ref.
0.94
0.14
*

-
-
-
Total: 147

1.00
0.89
0.75

Ref.
0.59-1.36
0.50-1.12

Ref.
0.58
0.16
*
 (detail) 74 (5)
(65 - )
Screening: IST, MMSE

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, ADLI, APOE4, MMSE, DEP, DM, HC, HTN, MS, OS, SP, VD‡
(detail)
Akbaraly, 2009
Crowe, 2003 STR Nested case control study with cumulative incidence sampling reporting odds ratios (ORs) 214
(66%)
-
(detail)
Activity before age 40
Less active:         
More active:         
(detail)

-
-
Total: 67

1.00
0.55

Ref.
0.28-1.08

Ref.
0.08
*

-
-
Total: 107

1.00
0.80

Ref.
0.49-1.31

Ref.
0.37
*
 (detail) 57 (6)
( - )
(detail)
Screening: MMSE, TELE

AD Diagnosis: Death Records, Medical History, NINCDS ADRDA
(detail)
A, E, G‡
(detail)
Crowe, 2003
* Derived value.
‡ Covariates: "A" (age), "E" (education), "G" (gender), "ADLI" (activities of daily living impairment), "APOE4" (APOE e4 genotype), "MMSE" (baseline MMSE), "DEP" (depression), "DM" (diabetes mellitus), "HC" (high cholesterol), "HTN" (hypertension), "MS" (marital status), "OS" (Occupational status), "SP" (study population), "VD" (vascular disease)
 
Table 2:   Cognitive activity participation - continuous
Notes These studies evaluate the association between AD risk and participation in leisure-time cognitive activity, considered as a continuous variable. The studies varied in the leisure activities they considered to be "cognitive". Studies also differ in their measures of participation frequency, as noted in the exposure subheadings for each report.  
  Alzheimer Disease Total Dementia  
Paper Cohort Study Type # Subjects
(% Female)
Average Follow-up Time
Mean (SD)
(Range)
# of Cases Effect Size 95% CI P-value # of Cases Effect Size 95% CI P-value Ethnicity Age at Start of Follow-up:
Mean (SD)
(Range)
Diagnostic Assessment Covariates & Analysis Comment Paper
Heser, 2014 German AgeCoDe Study Incidence study reporting hazard ratios (HRs) 2300
(65%)
-
(detail)
Participation score
1.4 (0)
(0 - 4)
(detail)

150

0.62

0.44-0.87

0.006

213

0.66

0.50-0.87

0.004
  82 (3)
(75 - )
(detail)
AD Diagnosis: DSM IV
(detail)
A, E, G, APOE4, MMSE, IADLI‡
(detail)
Heser, 2014
Verghese, 2003 Bronx Aging Study Incidence study reporting hazard ratios (HRs) 469
(64%)
5.8 y
*

(detail)
activity days per week
9.8 (6)
(0 - 42)
(detail)

61

0.93

0.88-0.98

0.008
*

124

0.93

0.90-0.97

0.0001
*
Caucasian
(detail)
79 (3)
(75 - )
Screening: Blessed, Other

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, Blessed, CDUB‡ Verghese, 2003
Wilson, 2002a ROS Incidence study reporting hazard ratios (HRs) 733
(67%)
4.5 y
Participation score
3.6 (1)
(1 - 5)
(detail)

111

0.67

0.49-0.92

0.01
*

 

 

 

 
Caucasian, Other
(detail)
75 (6)
(65 - )
Screening: CERAD

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G‡ Wilson, 2002a
Wilson, 2002b CHAP Cumulative incidence study reporting odds ratios (ORs) 835
(59%)
4.1 y
Participation score
3.3 (1)
(1 - 5)
(detail)

139

0.39

0.21-0.73

0.003
*

 

 

 

 
Caucasian, African-American (Black)
(detail)
76 (6)
(65 - )
Screening: CERAD

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, FUT, OS, RE, SS‡
(detail)
Wilson, 2002b
Wilson, 2007 MAP Incidence study reporting hazard ratios (HRs) 775
(75%)
2.5 y
*

(detail)
Participation score
3.2 (1)
(1 - 5)
(detail)

90

0.53

0.32-0.86

0.01
*

 

 

 

 
Caucasian
(detail)
80 (7)
( - )
Screening: CERAD

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, INC, CA-Past, PA, SA, SES‡ Wilson, 2007
* Derived value.
‡ Covariates: "A" (age), "E" (education), "G" (gender), "APOE4" (APOE e4 genotype), "Blessed" (baseline Blessed score), "MMSE" (baseline MMSE), "CDUB" (chronic disease up to baseline), "FUT" (follow up time), "INC" (Income), "IADLI" (instrumental activities of daily living impairment), "OS" (Occupational status), "CA-Past" (Past Cognitive Activity ), "PA" (physical activity), "RE" (race/ethnicity), "SA" (social Activity), "SES" (socioeconomic status), "SS" (stratified sampling)