How to escape dementia

Untreated midlife hearing loss carries the largest single attributable risk in the Lancet Commission model (about 7% of cases). The likely mechanisms: an under-stimulated cortex, higher listening effort, and social withdrawal. In the ACHIEVE randomized trial, hearing aids cut cognitive decline by about 48% over three years in higher-risk adults [2]. The boundary matters: this is about acquired loss in people who grew up hearing — being Deaf from birth or early childhood with full language access (signed or otherwise) is not the exposure these studies measured, and it is not on this list.

Midlife hypertension damages the small vessels that feed white matter, and it is the best-proven medical target: SPRINT MIND [3] found intensive blood-pressure treatment reduced mild cognitive impairment by 19%, the first trial evidence that treating any risk factor protects cognition.

The headline addition in the 2024 Commission update: high midlife LDL cholesterol now carries about 7% of attributable risk, tied with hearing for the largest share. The mechanism is vascular — atherosclerosis in the vessels feeding the brain plus stroke, and stroke roughly doubles subsequent dementia risk (Kuźma et al. [4] meta-analysis).

Physically active people develop dementia substantially less often across dozens of cohorts, and exercise plausibly works through several channels at once: vascular health, insulin sensitivity, and brain-derived growth factors that support the hippocampus, the memory hub that shrinks first.

Social isolation raises dementia risk about 50% in meta-analyses of longitudinal studies [5, 6]. Conversation is heavy cognitive exercise: prediction, memory, emotion-reading, and turn-taking, all at once.

Type 2 diabetes raises dementia risk roughly 50%, and the longer the exposure the higher the risk. Insulin resistance impairs the brain vasculature and may impair amyloid clearance; some researchers half-jokingly call Alzheimer’s "type 3 diabetes".

Smokers develop dementia meaningfully more often, through vascular injury and oxidative stress, and the risk falls after quitting: long-term ex-smokers approach never-smoker risk in large cohorts. The brain forgives, slowly.

Midlife obesity raises later dementia risk about 33% in a meta-analysis of 39 cohort studies following 1.3 million adults [7, 4], working through the same vascular and metabolic channels as diabetes. The timing matters: it is the waistline in your 40s and 50s that predicts, not late-life weight, which often falls in the years before diagnosis.

More years of education and mentally demanding work associate with later dementia onset. The leading explanation is cognitive reserve: brains with denser, more redundant networks can absorb more pathology before symptoms appear. Challenge keeps building reserve in adulthood.

Depression, particularly in midlife and beyond, is a consistent dementia risk factor in the Commission model, through stress hormones, inflammation, and withdrawal from activity and people. Treatment is protective on all three fronts.

Traumatic brain injury, even the moderate kind from falls, crashes, and contact sports, raises later dementia risk, and repeated injuries compound. This is one of the most literal risk factors on the list.

Heavy drinking is strongly tied to early-onset dementia: in a study of over 1 million French hospital records, alcohol use disorders were the strongest modifiable factor for dementia before 65 [8]. Alcohol is directly neurotoxic and hits the same vessels as the other vascular factors on this list.

One of the two factors the 2024 Commission added: untreated vision loss associates with higher dementia risk, likely through the same under-stimulation and withdrawal channels as hearing. The encouraging flip side: adults who had cataract surgery went on to develop dementia about 30% less often than those who did not [9].

The other 2024 addition: long-term exposure to fine-particle air pollution (PM2.5) associates with higher dementia risk across cohorts and in a meta-analysis of 14 studies [10], plausibly via vascular injury and neuroinflammation.