The cognitive trajectory is not a single deterministic slope. It is a curve shaped by many inputs, some of which are modifiable. The clinical question — once prevention is being delivered, or once the inflection has been recognized — is which modifiers actually bend the slope, and by how much. Ranking modifiers by effect size matters because clinical time is finite and patient attention is more finite still. The patient who can change three things should change the three highest-leverage things, not three random things from a list.
The largest single modifier is treatment of hearing loss. The Lancet Commission identified hearing loss as the highest population-attributable-risk factor for dementia, larger than any other single modifiable variable. The ACHIEVE trial, published in 2023, showed that hearing aids reduced three-year cognitive decline by approximately 48% in higher-risk older adults, though the effect in lower-risk community-dwelling adults was smaller. The mechanism is biologically plausible: sensory deprivation, social withdrawal, and increased cognitive load on the brain processing impoverished input. The intervention is available, often covered, and chronically underused — fewer than one in five patients with hearing loss who would benefit from hearing aids actually wear them.
After hearing, the largest modifiers cluster around vascular health. Blood pressure control to under 130 systolic — the SPRINT-MIND trial demonstrated a reduction in MCI incidence with intensive blood pressure control. Lipid management, particularly apoB rather than just LDL-C, addresses the cardiovascular substrate that contributes to vascular cognitive impairment. Diabetes management — the metabolic-cognitive link runs through insulin resistance, glycemic variability, and the cerebrovascular disease that diabetes accelerates. Physical activity — particularly aerobic activity that maintains VO2max — is among the most consistently evidence-supported cognitive interventions across populations.
Depression treatment is a major modifier that is consistently underweighted. Untreated depression accelerates cognitive aging through multiple mechanisms: chronic systemic inflammation, HPA axis dysregulation with sustained cortisol exposure, hippocampal volume loss, reduced cognitive engagement and social withdrawal. The patient with treatment-resistant depression that has been refractory for years has accumulated cognitive cost that does not fully reverse with eventual treatment. The patient whose depression has been well managed with durable remission has avoided that cost. The implication for psychiatric practice is that aggressive, persistent, refractory-pursuit treatment of depression is also longevity-psychiatry work, not just symptom management.
Sleep is among the highest-leverage modifiers, particularly in the 50s-70s window. Sleep apnea is highly prevalent, substantially underdiagnosed in psychiatric populations, and chronically untreated. It produces intermittent nocturnal hypoxia and disrupts the glymphatic clearance that the brain depends on to wash out amyloid and other metabolic byproducts. CPAP treatment changes the cognitive trajectory of treated patients meaningfully — though adherence remains the central practical challenge. Chronic insomnia, separately from apnea, is itself associated with cognitive decline and responds to CBT-I and other interventions.
The lifestyle cluster modifies the slope in the aggregate. Physical activity, social engagement, cognitive enrichment, purpose, diet, alcohol restraint. No single lifestyle intervention dramatically changes the trajectory by itself. The cumulative effect of multiple modifiable factors, applied consistently across years, is what produces the large population-level findings reported in the Lancet Commission and elsewhere. The patient who exercises, eats reasonably, sleeps well, has meaningful social engagement, treats their psychiatric conditions, and manages their medical conditions has a meaningfully better trajectory than the patient who does none of these. The individual pieces are not glamorous. The aggregate is.
The pharmacological modifiers in clinical use as of 2026 — donepezil and the cholinesterase inhibitors, memantine, and the anti-amyloid antibodies lecanemab and donanemab — modify the slope of established dementia modestly. The cholinesterase inhibitors offer symptomatic improvement and possibly modest slope effect. The anti-amyloid antibodies offer measurable but modest slowing of clinical decline in early Alzheimer’s disease at the cost of significant infusion logistics, monitoring requirements, and ARIA risk. They are not prevention. They are slope-bending tools applied when the upstream work has either succeeded partially or failed. The largest modifiers remain upstream and unglamorous: hearing, vascular, depression, sleep, lifestyle. The drugs are downstream tools, used appropriately when needed.