Importance  Lithium deficiency may contribute to Alzheimer disease pathogenesis. No randomized clinical trial has examined lithium’s effects on cognition, neuroimaging, and plasma biomarkers in mild cognitive impairment (MCI).

Objective  To examine the feasibility, safety, and preliminary efficacy of lithium carbonate for delaying cognitive decline in older adults with MCI.

Design, Setting, and Participants  This single-site, randomized, double-blind, placebo-controlled pilot feasibility clinical trial was conducted at the University of Pittsburgh School of Medicine from February 2018 to August 2024, with 2-year follow-up. Analyses used linear mixed-effects models in the intention-to-treat population. Adults aged 60 years or older with MCI who were free of major psychiatric or neurologic illness and contraindications to lithium were included. Of 170 individuals assessed, 83 were randomized (41 lithium vs 42 placebo), with 80 starting treatment (41 lithium vs 39 placebo). Data were analyzed from August 2024 to December 2025.

Intervention  Daily low-dose lithium carbonate or placebo for 2 years.

Main Outcomes and Measures  Six prespecified coprimary outcomes included cognitive performance (California Verbal Learning Test-II [CVLT-II] delayed recall, Brief Visuospatial Memory Test-Revised, preclinical Alzheimer cognitive composite), hippocampal volume, cortical gray matter volume, and brain-derived neurotrophic factor.

Results  Among 80 participants (mean [SD] age, lithium: 72.93 [8.77] years; placebo: 71.22 [6.47] years; 56% female), none of the 6 coprimary outcomes met the prespecified significance threshold. Mean (SD) CVLT-II baseline scores were 7.95 (3.4) for lithium and 7.90 (3.9) for placebo; scores declined 1.42 points annually in the placebo group vs 0.73 points in the lithium group (difference, 0.69 points per year; 95% CI, 0.01-1.37; P = .05). Hippocampal and cortical volumes showed a decline over time in both groups, but no significant treatment × time interactions. Serious adverse events occurred in 12 of 41 (29%) receiving lithium vs 9 of 39 (23%) receiving placebo; none were definitely treatment related. One death occurred in the placebo group. Common adverse events included increased creatinine levels (12 of 41 [29%] with lithium vs 12 of 39 [31%] with placebo), diarrhea (12 of 41 [29%] vs 6 of 39 [15%]), tiredness (12 of 41 [29%] vs 6 of 39 [15%]), and tremor occurrence (10 of 41 [24%] vs 6 of 39 [15%]).

Conclusions and Relevance  This pilot randomized clinical trial established feasibility, confirmed safety and tolerability, and generated effect size estimates for future trials of low-dose lithium in MCI. None of the coprimary outcomes met the prespecified significance threshold.