What the research found

The LATTICE trial tested whether low-dose lithium could slow cognitive decline in older adults at risk of Alzheimer's dementia. Despite years of suggestive evidence from observational studies and earlier smaller trials, this randomized controlled trial found no statistically significant benefit on its six primary cognitive endpoints.

However, the negative result doesn't necessarily close the book on lithium. A prior trial by Forlienza and colleagues, which followed participants for years after treatment ended, showed slowing of cognitive decline during active lithium use and hints of sustained benefit afterward. That earlier work also documented changes in cerebrospinal fluid markers—elevated amyloid-42 (suggesting improved clearance from the brain) and reduced abnormal tau—alongside a roughly 50% reduction in an enzyme called GSK-3β that is abnormally active in Alzheimer's pathology.

The evidence base supporting lithium investigation spans observational data showing lower dementia rates in populations with higher drinking-water lithium exposure, and studies of bipolar disorder patients on long-term lithium who show less hippocampal atrophy and white matter damage than untreated controls. The LATTICE failure, combined with this earlier data, suggests the question remains open rather than settled.

Why it matters for you

If you're tracking cognitive biomarkers or concerned about dementia risk, LATTICE's negative result tempers enthusiasm for lithium supplementation as a straightforward protective strategy. The trial tested a specific dose in a specific population, so it doesn't prove lithium has no role—but it does mean you shouldn't expect a clear cognitive benefit from adding low-dose lithium on its own.

The mechanistic findings from earlier work are noteworthy if you're already measuring amyloid or tau in blood or CSF: lithium appeared to shift amyloid metabolism (not by reducing production, but by improving clearance), and modulated GSK-3β activity. These aren't endpoints most MyKine users track routinely, but they hint at a real biological effect even when clinical outcomes didn't shift.

The mismatch between LATTICE's null result and the earlier positive trial suggests study design matters—timing, dose, population selection (amnestic mild cognitive impairment versus prodromal AD), and follow-up duration all likely influence whether you'd detect a benefit. This is a reminder that negative trials in humans don't erase animal or mechanistic data, but they do narrow where future investigation should focus.

Caveats

  • LATTICE is a single human trial; negative results don't exclude benefit in other populations or dosing strategies
  • The earlier Forlienza trial was small (45 participants) and showed statistical hints of long-term benefit that weren't formally significant
  • Observational studies linking ambient lithium exposure to lower dementia rates cannot establish causation and are vulnerable to unmeasured confounding
  • Extrapolating from bipolar disorder patients (whose brain pathology differs from cognitively normal individuals) to prevention in unaffected populations relies on an unproven mechanism
  • The LATTICE population, design, and dosing may differ in important ways from what would be needed to detect benefit in a different group