Skip navigation

Diet and dementia

Clinical bottom line: There is no convincing evidence of a link between dietary factors and dementia. Most studies showed conflicting results, though there was some indication of modest reductions in dementia with consumption of fish and wine, but an increased risk for high fat intake.

Systematic review:

Ernst E. Diet and dementia, is there a link? A systematic review. Nutritional Nueroscience 1999; 2: 1-6.

Date review completed: End of 1997

Number of trials included: 11

Number of patients: Retrospective studies: 289 cases, 334 controls; Prospective studies: 50 cases, 12469 controls)

Control groups: individuals described as healthy, matched for age or matched for age, gender and origin.

Main outcomes: total energy intake, vitamins and mineral intake, protein intake; fat intake.

Inclusion criteria were published cross-sectional or longitudinal study which assessed the relationship of dietary factors and dementia (multiple-infarct or Alzheimer's).

MEDLINE, EMBASE, the Cochrane Library (all years to end 1997), the reviewer's own database and bibliographies of retrieved reports were searched for relevant reports. No language restrictions were made. Experts were contacted for further information. Data were extracted in a pre-defined fashion. A descriptive analysis was conducted.


Case-control studies:

Six case-control studies were included. Cases were institutionalised or noninstitutionalised patients with dementia or Alzheimer's. All case-control studies were small and produced conflicting results. Four found no significant differences between cases and controls. One reported higher intakes of energy, proteins, vitamins and minerals in cases, and one reported the opposite.

Longitudinal studies:

There were two prospective case-control studies with 12 and 28 months follow up. These studies were small and used nondemented or community controls. One found higher nutrient intake in cases than in controls, the other found no difference between groups.

The three cohort studies had between 2.1-6 years of follow up. One assessed wine consumption only, the others assessed dietary factors including heavy meat consumption in one. Wine consumption of 3-4 glasses per day in 3777 nondemented French individuals resulted in a reduced incidence of dementia and Alzheimer's; adjusted odds ratios OR = 0.19 and OR = 0.28 respectively. Consumption of 1-2 glasses per day resulted in more modest reductions. There was no convincing evidence of relationship between heavy meat consumption and dementia, but one study found increased risk with total fat (relative risk (RR) = 2.4), saturated fat (RR = 1.9), and cholesterol intake (RR = 1.7) and lower risk with fish intake (RR = 0.4). Unfortunately 95% confidence intervals were not provided around the ORs or RRs and there was insufficient information provided to calculate them.

Adverse effects

Adverse effects were not mentioned.



The reviewer stated that most of the included studies were methodologically flawed. All of the case-control studies were small and, therefore, were not of adequate power to detect true differences between groups. Retrospective case-control studies and cohort studies are open to various confounding factors and bias. The studies used different diagnostic criteria for dementia. All of these factors may explain the inconsistency of the results.

Further reading

Related topics


AT056 - 5898 DIET AND DEMENTIA: Aug-2000