Nunes et al (Reference Nunes, Forlenza and Gattaz2007) reported that the prevalence of Alzheimer's disease in a group of elderly patients with bipolar disorder who were on continuous lithium treatment was significantly less than in a similar group without recent lithium therapy. After controlling for age, lithium use remained associated with a smaller risk of Alzheimer's disease (age-adjusted OR=0.079, 95% CI 0.020–0.321). Conversely, Dunn et al (Reference Dunn, Holmes and Mullee2005) showed that patients who received lithium had a significantly higher risk of dementia than those who did not (age-adjusted OR=1.8, 95% CI 1.1–2.8).

Nunes et al (Reference Nunes, Forlenza and Gattaz2007) found no differences between the lithium and the comparison group in neuropsychological performance after excluding patients with Alzheimer's disease. This is in accordance with our study using Mini-Mental State Examination (MMSE) scores (Reference Terao, Nakano and InoueTerao et al, 2006). Our study, however, showed that patients with present and/or past history of lithium treatment had significantly better MMSE scores than patients without any history of lithium treatment (Reference Terao, Nakano and InoueTerao et al, 2006). It is important to further investigate lithium in the prevention of Alzheimer's dementia with a large number of patients in prospective studies.

If lithium has a preventive effect for Alzheimer's disease, there may be two possible mechanisms. First, it might indirectly prevent dementia via its prophylactic effects on mood disorders, because the rate of dementia increased 13% with every episode leading to admission for patients with depressive disorder and 6% for patients with bipolar disorder, when adjusted for differences in age and gender (Reference Kessing and AndersenKessing & Andersen, 2004). Second, lithium might directly prevent dementia via its inhibition of glycogen synthase kinase 3 (GSK-3) alpha (Reference Phiel, Wilson and LeePhiel et al, 2003) and GSK-3 beta (Reference Phiel and KleinPhiel & Klein, 2001). Although Nunes et al (Reference Nunes, Forlenza and Gattaz2007) found no significant differences in the number of previous depressive and manic episodes between the lithium and comparison groups, at present both possibilities should be borne in mind.