Get Newsletter
AlzRisk Risk Factor OverView
Back

Risk Factor:
Risk Factor Type: Nutrition and supplements
Current Understanding:
Current Understanding The studies reviewed here generally provide support for an inverse relation between Alzheimer's disease and a healthy dietary pattern, defined loosely as a diet high in fruits and vegetables, low in red and processed meats, and favoring mono- and polyunsaturated fats over saturated fats. Although the studies varied somewhat in their definitions of healthy dietary pattern and used different methods to quantify adherence to that pattern, results were largely consistent. However, these studies have several methodological limitations that complicate their interpretation. Some studies adjusted for putative confounders like diabetes and cardiovascular disease that may in fact be causal intermediates, although effect estimates were similar with and without adjustment for these factors. No studies measured long-term dietary pattern and few measured dietary intake in mid-life, which are likely to be both more biologically relevant and less susceptible to reverse causation. Additional prospective studies addressing these limitations and evidence from randomized trials will help clarify the issue. In the meantime, a healthy dietary pattern, recommended for lowering risk of cardiovascular disease, may also lower risk of AD, particularly as part of a broader intervention including adequate physical activity and other lifestyle changes. For a review of the putative mechanisms by which dietary pattern may influence AD risk and detailed commentary on interpreting the findings below in a broader context, please view the Discussion.
Literature Extraction: Search strategy  * New *
Last Search Completed: 28 November 2016


Table 1:   Healthy dietary pattern - categorical
Notes These studies examine the association between adherence to a healthy dietary pattern and AD risk. The studies measure adherence to a healthy dietary pattern using either data-driven methods (reduced rank regression) or hypothesis-driven dietary patterns based on previous literature. Definitions of “healthy dietary pattern” vary, but are not explicitly described as Mediterranean (see tables 3 to 5).  
  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
Barberger-Gateau, 2007 3C Incidence study reporting hazard ratios (HRs) 8085
(61%)
3.5 y
Unhealthy dietary pattern: 8%
Healthy dietary pattern: 92%
(detail)
-
-
Total: 183
1.00
0.61
Ref.
0.39-0.96
Ref.
0.03
*
39
242
Total: 281
1.00
0.66
Ref.
0.45-0.96
Ref.
0.03
*
French
(detail)
74 (5)
(65 - )
AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, CC, INC, MS‡
(detail)
Barberger-Gateau, 2007
Eskelinen, 2011 NKP-FINMONICA-CAIDE Cumulative incidence study reporting odds ratios (ORs) 385
(62%)
14 y
Unhealthy dietary pattern: 55%
Healthy dietary pattern: 45%
(detail)
14
2
Total: 16
1.00
0.08†
Ref.
0.01-0.89
Ref.
0.03
*
15
3
Total: 18
1.00
0.12†
Ref.
0.02-0.85
Ref.
0.03
*
Caucasian
(detail)
57 (4)
( - )
Screening: MMSE

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, BMI, DM, FUT, MI, PA, RES, SM, SH, SBP, TC‡
(detail)
Eskelinen, 2011
Gelber, 2012 HAAS Nested case control study with cumulative incidence sampling reporting odds ratios (ORs) 3468
(0%)
25 y
Unhealthy dietary pattern: 60%
Healthy dietary pattern: 40%
(detail)
75
42
Total: 117
1.00
0.74
Ref.
0.48-1.12
Ref.
0.16
*
138
85
Total: 223
1.00
0.83
Ref.
0.61-1.14
Ref.
0.24
*
Japanese-American
52 (-)
( - )
Screening: CASI, HDS, MMSE, 3MSE

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, APOE4, CVD, CYJPN, DM, HC, HXHTN, OS‡ Gelber, 2012
Gu, 2010 WHICAP Incidence study reporting hazard ratios (HRs) 2148
(68%)
4.0 y
Reduced rank regression scores
Lowest tertile: 35%
Second tertile: 33%
Highest tertile: 32%
(detail)

117
86
50
Total: 253

1.00
0.81
0.62

Ref.
0.59-1.12
0.43-0.89

Ref.
0.2
0.01

 
 
 

 

 

 
Caucasian, Other, Hispanic, African-American (Black)
77 (7)
(65 - )
Screening: BDRS, Other

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, BMI, Kcal, CI, ORGCH, RE, SM‡
(detail)
Gu, 2010
Morris, 2015 MAP Incidence study reporting hazard ratios (HRs) 923
(76%)
4.5 y
(detail)
MIND diet score
Lowest tertile (2.5-6.5): 33%
Second tertile (7-8): 33%
Highest tertile (8.5-12.5): 33%
(detail)

-
-
-
Total: 144

1.00
0.64
0.48

Ref.
0.42-0.97
0.29-0.79

Ref.
0.04
0.004
*

 
 
 

 

 

 
Caucasian
(detail)
81 (-)
(58 - 98)
AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, Kcal, CVRF, CA-Past, PA‡ Morris, 2015
Morris, 2015 MAP Incidence study reporting hazard ratios (HRs) 923
(76%)
4.5 y
(detail)
DASH diet score
Lowest tertile (1.0-3.5): 33%
Second tertile (4.0-4.5): 33%
Highest tertile (5.0-8.5): 33%
(detail)

-
-
-
Total: 144

1.00
0.98
0.60

Ref.
0.64-1.46
0.37-0.96

Ref.
0.92
0.04
*

 
 
 

 

 

 
Caucasian
(detail)
81 (-)
(58 - 98)
AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, Kcal, CVRF, CA-Past, PA‡ Morris, 2015
Olsson, 2015 ULSAM Incidence study reporting hazard ratios (HRs) 1038
(0%)
12 y
HDI score
Low adherent (-1 to 1):         
Medium Adherent (2 to 5):         
High Adherent (6 to 8):         
(detail)

-
-
-
Total: 84

1.00
1.25
0.99

Ref.
0.61-2.53
0.39-2.53

Ref.
0.54
0.98
*

-
-
-
Total: 143

1.00
1.26
0.95

Ref.
0.73-2.18
0.46-1.99

Ref.
0.41
0.89
*
Caucasian
(detail)
71 (1)
( - )
AD Diagnosis: DSM IV, NINCDS ADRDA
(detail)
E, APOE4, Kcal, LS, PA, SM‡ Olsson, 2015
Olsson, 2015 ULSAM Incidence study reporting hazard ratios (HRs) 1038
(0%)
12 y
LCHP score
Low Adherent (≤6):         
Medium Adherent (7 to 15):         
High Adherent (≥16):         
(detail)

-
-
-
Total: 84

1.00
1.33
1.60

Ref.
0.61-2.27
0.74-3.44

Ref.
0.64
0.23
*

-
-
-
Total: 143

1.00
1.30
1.59

Ref.
0.77-2.21
0.85-2.96

Ref.
0.33
0.15
*
Caucasian
(detail)
71 (1)
( - )
AD Diagnosis: DSM IV, NINCDS ADRDA
(detail)
E, APOE4, Kcal, LS, PA, SM‡ Olsson, 2015
Ozawa, 2013 Hisayama Study Incidence study reporting hazard ratios (HRs) 1006
(57%)
15 y
Reduced rank regression scores
Lowest quartile (<-0.82): 25%
Second quartile (-0.82 to -0.05): 25%
Third quartile (-0.06 to 0.83): 25%
Highest quartile (≥0.83): 25%
(detail)

36
31
37
40
Total: 144

1.00
0.64
0.74
0.65

Ref.
0.39-1.04
0.46-1.18
0.40-1.06

Ref.
0.07
0.21
0.08
*

69
72
63
67
Total: 271

1.00
0.85
0.72
0.66

Ref.
0.61-1.19
0.50-1.02
0.46-0.95

Ref.
0.34
0.07
0.02
*
Japanese
(detail)
68 (-)
(60 - 79)
Screening: HDS, HDS-R, MMSE

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, BMI, Kcal, DM, HTN, PA, SMKH, SH, TC‡ Ozawa, 2013
* Derived value.
† Five or fewer cases exist.
‡ Covariates: "A" (age), "E" (education), "G" (gender), "APOE4" (APOE e4 genotype), "BMI" (body mass index), "Kcal" (caloric intake), "CVD" (cardiovascular disease), "CVRF" (cardiovascular risk factors), "CYJPN" (childhood years spent in Japan), "CC" (city center), "CI" (comorbidity index), "DM" (diabetes mellitus), "FUT" (follow up time), "HC" (high cholesterol), "HXHTN" (history of hypertension), "HTN" (hypertension), "INC" (Income), "LS" (Living situation), "MS" (marital status), "ORGCH" (member of the original cohort), "MI" (mycardial infarction history), "OS" (Occupational status), "CA-Past" (Past Cognitive Activity ), "PA" (physical activity), "RE" (race/ethnicity), "RES" (residential township), "SMKH" (smoking habits), "SM" (smoking status), "SH" (stroke history), "SBP" (systolic blood pressure), "TC" (total cholesterol)
 
Table 2:   Healthy dietary pattern - continuous, per 1 SD increase
Notes These studies examine the association between adherence to a healthy dietary pattern and AD risk. The effect estimate is the relative risk of AD (or total dementia) corresponding to a 1 SD increase in the score for the healthy dietary pattern. Definitions of “healthy dietary pattern” vary, but are not explicitly described as Mediterranean (see tables 3 to 5).  
  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
Olsson, 2015 ULSAM Incidence study reporting hazard ratios (HRs) 1038
(0%)
12 y
HDI score
- (-)
(-1 - 8)
(detail)

84

0.95

0.75-1.22

0.68
*

143

0.94

0.78-1.14

0.52
*
Caucasian
(detail)
71 (1)
( - )
AD Diagnosis: DSM IV, NINCDS ADRDA
(detail)
E, APOE4, Kcal, LS, PA, SM‡ Olsson, 2015
Olsson, 2015 ULSAM Incidence study reporting hazard ratios (HRs) 1038
(0%)
12 y
LCHP score
- (-)
(2 - 20)
(detail)

84

1.16

0.95-1.43

0.15
*

143

1.16

0.99-1.37

-
Caucasian
(detail)
71 (1)
( - )
AD Diagnosis: DSM IV, NINCDS ADRDA
(detail)
E, APOE4, Kcal, LS, PA, SM‡ Olsson, 2015
* Derived value.
‡ Covariates: "E" (education), "APOE4" (APOE e4 genotype), "Kcal" (caloric intake), "LS" (Living situation), "PA" (physical activity), "SM" (smoking status)
 
Table 3:   Mediterranean dietary pattern - categorical
Notes These studies examine the association between adherence to a Mediterranean dietary pattern and AD risk. The studies use the Mediterranean diet score (range: 0-9). Please see the Discussion for details of an additional randomized controlled trial, PREDIMED-NAVARRA, that showed benefits of a Mediterranean diet over a low-fat diet in prevention of all-cause dementia.  
  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
Feart, 2009 3C Incidence study reporting hazard ratios (HRs) 1410
(63%)
5.0 y
Mediterranean diet score
Low score (0-3): 30%
Middle score (4-5): 44%
High score (6-9): 26%
(detail)

-
-
-
Total: 66

1.00
0.99
0.86

Ref.
0.51-1.94
0.39-1.88

Ref.
0.98
0.71

-
-
-
Total: 99

1.00
1.11
-

Ref.
0.63-1.94
0.60-2.10

Ref.
0.71
0.72
French
(detail)
76 (-)
(65 - )
Screening: BVRT, IST, MMSE, Other

AD Diagnosis: DSM IV
(detail)
E, G, APOE4, BMI, Kcal, DEP, DM, HC, HTN, MS, MEDS, PA, SM, SH‡
(detail)
Feart, 2009
Morris, 2015 MAP Incidence study reporting hazard ratios (HRs) 923
(76%)
4.5 y
(detail)
MedDiet score
Lowest tertile (18-29): 33%
Second tertile (30-34): 33%
Highest tertile (35-46): 33%
(detail)

-
-
-
Total: 144

1.00
0.81
0.49

Ref.
0.53-1.21
0.29-0.85

Ref.
0.32
0.009
*

 
 
 

 

 

 
Caucasian
(detail)
81 (-)
(58 - 98)
AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, Kcal, CVRF, CA-Past, PA‡ Morris, 2015
Olsson, 2015 ULSAM Incidence study reporting hazard ratios (HRs) 1038
(0%)
12 y
mMDS score
Low Adherent (≤2):         
Medium Adherent (3 to 5):         
High Adherent (≥6):         
(detail)

-
-
-
Total: 84

1.00
1.16
0.99

Ref.
0.64-2.10
0.44-2.26

Ref.
0.62
0.98
*

-
-
-
Total: 143

1.00
1.05
0.85

Ref.
0.67-1.66
0.44-1.62

Ref.
0.83
0.63
*
Caucasian
(detail)
71 (1)
( - )
AD Diagnosis: DSM IV, NINCDS ADRDA
(detail)
E, APOE4, Kcal, LS, PA, SM‡ Olsson, 2015
Scarmeas, 2006 WHICAP Incidence study reporting hazard ratios (HRs) 2258
(68%)
4.0 y
Mediterranean diet score
Lowest tertile (0-3): 32%
Second tertile (4-5): 42%
Highest tertile (6-9): 26%
(detail)

96
107
59
Total: 262

1.00
0.85
0.60

Ref.
0.63-1.16
0.42-0.87

Ref.
0.3
0.006
*

 
 
 

 

 

 
Caucasian, Other, Hispanic, African-American (Black)
77 (7)
( - )
Screening: BDRS, CDR, Neuropsych Testing

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, BMI, Kcal, CHRT, CI, RE, SM‡
(detail)
Scarmeas, 2006
* Derived value.
‡ Covariates: "A" (age), "E" (education), "G" (gender), "APOE4" (APOE e4 genotype), "BMI" (body mass index), "Kcal" (caloric intake), "CVRF" (cardiovascular risk factors), "CHRT" (cohort), "CI" (comorbidity index), "DEP" (depression), "DM" (diabetes mellitus), "HC" (high cholesterol), "HTN" (hypertension), "LS" (Living situation), "MS" (marital status), "MEDS" (medication use), "CA-Past" (Past Cognitive Activity ), "PA" (physical activity), "RE" (race/ethnicity), "SM" (smoking status), "SH" (stroke history)
 
Table 4:   Mediterranean dietary pattern - continuous, per 1 point increase
Notes These studies examine the association between adherence to a Mediterranean dietary pattern and AD risk. The effect estimate is the relative risk of AD (or total dementia) corresponding to a 1-point increase in the Mediterranean diet score (range: 0-9).  
  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
Feart, 2009 3C Incidence study reporting hazard ratios (HRs) 1410
(63%)
5.0 y
Mediterranean diet score
4.4 (2)
(0 - 8)
(detail)

66

1.00

0.85-1.19

0.96

99

1.06

0.92-1.21

0.43
French
(detail)
76 (-)
(65 - )
Screening: BVRT, IST, MMSE, Other

AD Diagnosis: DSM IV
(detail)
E, G, APOE4, BMI, Kcal, DEP, DM, HC, HTN, MS, MEDS, PA, SM, SH‡
(detail)
Feart, 2009
Scarmeas, 2006 WHICAP Incidence study reporting hazard ratios (HRs) 2258
(68%)
4.0 y
Mediterranean diet score
4.3 (2)
( - )
(detail)

262

0.91

0.83-0.98

0.02

 

 

 

 
Caucasian, Other, Hispanic, African-American (Black)
77 (7)
( - )
Screening: BDRS, CDR, Neuropsych Testing

AD Diagnosis: NINCDS ADRDA
(detail)
A, E, G, APOE4, BMI, Kcal, CHRT, CI, RE, SM‡
(detail)
Scarmeas, 2006
‡ Covariates: "A" (age), "E" (education), "G" (gender), "APOE4" (APOE e4 genotype), "BMI" (body mass index), "Kcal" (caloric intake), "CHRT" (cohort), "CI" (comorbidity index), "DEP" (depression), "DM" (diabetes mellitus), "HC" (high cholesterol), "HTN" (hypertension), "MS" (marital status), "MEDS" (medication use), "PA" (physical activity), "RE" (race/ethnicity), "SM" (smoking status), "SH" (stroke history)
 
Table 5:   Mediterranean dietary pattern - continuous, per 1 SD increase
Notes These studies examine the association between adherence to a Mediterranean dietary pattern and AD risk. The effect estimate is the relative risk of AD (or total dementia) corresponding to a 1 SD increase in the modified Mediterranean diet score (range: 0-8).  
  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
Olsson, 2015 ULSAM Incidence study reporting hazard ratios (HRs) 1038
(0%)
12 y
mMDS score
- (-)
(0 - 8)
(detail)

84

1.00

0.75-1.33

0.99
*

143

0.94

0.75-1.18

0.59
*
Caucasian
(detail)
71 (1)
( - )
AD Diagnosis: DSM IV, NINCDS ADRDA
(detail)
E, APOE4, Kcal, LS, PA, SM‡ Olsson, 2015
* Derived value.
‡ Covariates: "E" (education), "APOE4" (APOE e4 genotype), "Kcal" (caloric intake), "LS" (Living situation), "PA" (physical activity), "SM" (smoking status)