Dementia risk factors

The Lancet Commission's 2024 model lists 14 potentially modifiable risk factors: less education, hearing loss, high blood pressure, smoking, obesity, depression, physical inactivity, diabetes, excessive alcohol consumption, traumatic brain injury, air pollution, social isolation, untreated vision loss and high LDL cholesterol (low-density lipoprotein, the cholesterol carried in particles that build up in artery walls) [1]. The last two were added in 2024; the 2020 model contained the other 12 and attributed about 40% of cases to them. The Commission adjusted its total for the overlap between factors, which tend to cluster in the same people [1].

Hearing loss has the most consistent evidence among the sensory factors. In the Baltimore Longitudinal Study of Aging, dementia risk rose by 27% for every 10 decibels of hearing loss [3], and the Commission's own pooled analysis of 666,370 people found 37% higher risk with hearing loss than with normal hearing [1]. Hearing and vision loss are described in a separate article.

Several factors act through the blood vessels of the brain. High blood pressure in midlife, diabetes, smoking, obesity and high LDL cholesterol all damage small and large arteries, and stroke itself roughly doubles later dementia risk (risk ratio 2.18 after an incident stroke) [7]. Midlife obesity (a body mass index of 30 or more between ages 35 and 65) was associated with 33% higher dementia risk across 19 studies of 589,649 people, while overweight was not clearly associated [8]. Current smokers had 30% higher risk than never-smokers across 37 studies, and former smokers had no measurable excess risk [9]. Heavy alcohol use shows the strongest association with dementia that begins before age 65: in over 1.1 million French hospital patients with dementia, alcohol use disorders were the strongest modifiable risk factor, with a hazard ratio (the risk in one group relative to another) of about 3.3 in both sexes [10].

Social and psychological factors form a second group. In a pooled analysis of 19 long-term studies, less frequent social contact was associated with 57% higher dementia risk and loneliness with 58% higher risk [11]; social connection is covered in its own article. Depression, traumatic brain injury and physical inactivity are each associated with higher risk in the Commission's review, and physical activity was associated with 20% lower risk across 58 studies of 257,983 people [1]. The evidence on air pollution is weaker: a 2023 pooled analysis of 14 studies found a 4% higher risk per 2 µg/m³ of fine particulate matter (PM2.5), an estimate whose uncertainty range included no effect [12].

Full deep dive: Dementia risk factors.

Alzheimer's disease is defined by two abnormal proteins in the brain: amyloid-beta, which forms plaques between nerve cells, and tau, which forms tangles inside them. The disease develops over decades. In a study of 128 people from families carrying genes for inherited Alzheimer's disease, amyloid levels in the spinal fluid began falling about 25 years before expected symptoms, amyloid deposits were visible on brain scans and tau levels rose about 15 years before, memory began to decline about 10 years before, and dementia was diagnosed on average 3 years after expected symptom onset [5]. The authors noted that the timeline may not apply to the far more common non-inherited form.

The APOE gene has the largest common genetic effect. In 7,351 people with Alzheimer's disease and 10,132 controls of European ancestry, lifetime risk at age 85 was 11% for men and 14% for women without regard to genotype, 23 to 30% for people with one copy of APOE4 alongside the common E3 variant and 51 to 60% for people with two copies [2]. Many carriers of APOE4 therefore never develop the disease. Having a close relative with dementia also raises risk, through both shared genes and shared habits.

Sleep has been studied as a possible influence on amyloid. In 20 healthy adults, one night without sleep increased amyloid in the hippocampus (the brain's memory centre) and thalamus on brain scans [14]. In 7,959 participants of the Whitehall II study followed for 25 years, sleeping 6 hours or less at age 50 and 60 was associated with 22% and 37% higher dementia risk than sleeping 7 hours, and persistent short sleep across ages 50 to 70 with 30% higher risk [15]. The biology of sleep is described under sleep regulation and sleep stages. Exercise may act on the same brain regions: in a trial of 120 older adults, a year of aerobic exercise increased hippocampal volume by 2%, reversing about 1 to 2 years of age-related loss [16]; the wider evidence is covered under exercise and longevity.

The first drugs that remove amyloid slow the disease modestly in its early stages. In a trial of 1,795 people with early Alzheimer's disease, lecanemab reduced the 18-month worsening on a standard dementia rating scale from 1.66 to 1.21 points, about 27% less decline than placebo [17]. In a trial of 1,736 people, donanemab slowed decline on a combined cognition and daily-function scale by 3.25 points over 76 weeks in those with low or medium tau levels, about 35% of the decline seen with placebo [18]. Both drugs caused amyloid-related imaging abnormalities, brain swelling or fluid visible on scans, in 12.6% of lecanemab and 24.0% of donanemab recipients [17, 18].

Full deep dive: Alzheimer's disease risk.

Diet is not one of the 14 factors in the Lancet Commission model [1], although it affects several of them, including blood pressure, cholesterol, diabetes and weight. The most studied pattern is the Mediterranean diet, which is high in vegetables, legumes, fruit, nuts, olive oil and fish and low in processed and red meat. In 60,298 UK Biobank participants followed for 9.1 years, the third with the highest Mediterranean diet score had 23% lower dementia risk than the lowest third, independent of genetic risk [6]. An earlier pooled analysis of five studies found 33% lower risk of mild cognitive impairment or Alzheimer's disease in the highest third [19]. Broader evidence on eating patterns is covered under dietary patterns.

The MIND diet combines the Mediterranean and DASH (Dietary Approaches to Stop Hypertension) diets with extra emphasis on leafy greens and berries. In 923 older adults in Chicago followed for 4.5 years, the highest third of MIND scores had 53% lower risk of Alzheimer's disease and the middle third 35% lower risk [20]. A three-year randomised trial of 604 older adults, however, found that cognition improved by a similar amount in the MIND group and in a control group, both of which received weight-loss counselling [21].

The quality of plant foods appears to matter more than the absence of meat. In 180,532 UK Biobank participants, an overall plant-based diet score was not associated with dementia (hazard ratio 1.03), while the top fifth of a healthful plant-based score (whole grains, vegetables, fruit, nuts, legumes, tea and coffee) had 18% lower risk and the top fifth of an unhealthful score (juices, refined grains, potatoes, sweets) 29% higher risk [22]. In 5,710 Taiwanese adults with 121 dementia cases, vegetarians had about a third lower risk than non-vegetarians after adjustment [23].

Among single foods, processed meat has the clearest association. In 493,888 UK Biobank participants, each additional 25 g a day of processed meat was associated with 44% higher dementia risk, while each additional 50 g a day of unprocessed red meat was associated with 19% lower risk [24]. Each 10% of the diet by weight from ultra-processed food was associated with 25% higher dementia risk in 72,083 older adults [25]. Across 21 studies of 181,580 people, each additional weekly serving of fish was associated with 5% lower dementia risk [26].

Full deep dive: Plant-based diets and dementia.

No trial has yet shown that changing a single risk factor prevents dementia diagnoses, partly because dementia develops slowly and trials are short. The strongest trial evidence concerns blood pressure. In SPRINT MIND, 9,361 adults with high blood pressure were assigned to a systolic target below 120 mm Hg or below 140 mm Hg; the intensive target reduced mild cognitive impairment (hazard ratio 0.81) and the combined outcome of mild cognitive impairment or dementia (0.85), while the reduction in probable dementia alone (0.83) was not statistically significant because the trial stopped early with fewer cases than expected [4].

The FINGER trial assigned 1,260 Finnish adults aged 60 to 77 at increased risk to two years of combined diet, exercise, cognitive training and vascular risk monitoring, or to general health advice; the intervention group's cognitive test scores improved slightly more [27]. The ACHIEVE trial assigned 977 adults aged 70 to 84 with untreated hearing loss to hearing aids and counselling or to health education. It found no difference in three-year cognitive decline overall, but a difference between its two recruited populations [28]: in the older, higher-risk group, hearing aids reduced decline by 48% [1]. The diet trials are summarised in the section above.

Most dementia risk estimates come from observational studies, which cannot fully separate a factor from the characteristics of the people who have it. Reverse causation is a particular problem because brain changes begin years before diagnosis: people in the early stages of dementia may become less active, more isolated, more depressed or lose weight, making those changes appear to cause the disease. Studies that measure risk factors in midlife and follow people for decades reduce, but do not remove, this bias.

The Commission's figure of 45.3% is a theoretical estimate that assumes each association is causal and that a factor could be removed entirely [1]. The individual relative risks it uses come from studies of varying quality, and the Commission notes that some associations may be only partly causal. Diet studies usually record food intake once, by questionnaire, and several of the largest results come from a single cohort, the UK Biobank.