BACKGROUND: Mood disorders are common, disabling and tend to be recurrent. They carry a high risk of suicide. Maintenance treatment, aimed at the prevention of relapse, is therefore of vital importance. Lithium has been used for some years as the mainstay of maintenance treatment in bipolar affective disorder, and to a lesser extent in unipolar disorder. However, the efficacy and effectiveness of prophylactic lithium therapy has been disputed. Low suicide rates in lithium-treated patients have led to claims that lithium has a specific anti-suicidal effect. If so, this is of considerable importance as treatments for mental disorders in general have not been shown convincingly to be effective in suicide prevention. OBJECTIVES: 1. To investigate the efficacy of lithium treatment in the prevention of relapse in recurrent mood disorders. 2. To examine the effect of lithium treatment on consumers' general health and social functioning, its acceptability to consumers, and the side-effects of treatment. 3. To investigate the hypothesis that lithium has a specific effect in reducing the incidence of suicide and deliberate self-harm in persons with mood disorders. SEARCH STRATEGY: The Cochrane Collaboration Depression, Anxiety and Neurosis Controlled Trials Register (CCDANCTR) and The Cochrane Controlled Clinical Trials Register (CCTR) were searched. Reference lists of relevant papers and major text books of mood disorder were examined. Authors, other experts in the field and pharmaceutical companies were contacted for knowledge of suitable trials, published or unpublished. Specialist journals concerning lithium were hand searched. SELECTION CRITERIA: Randomised controlled trials comparing lithium with placebo, where the stated intent of treatment was maintenance or prophylaxis. Participants were males and females of all ages with diagnoses of mood disorder. Discontinuation studies (in which all participants had been stable on lithium for some time before being randomised to either continued lithium treatment or placebo substitution) were excluded. DATA COLLECTION AND ANALYSIS: Data were extracted from the original reports independently by two reviewers. The main outcomes studied were related to the objectives stated above. Data were analysed for all diagnoses of mood disorder and for bipolar and unipolar disorder separately. Data were analysed using Review Manager version 4.0. MAIN RESULTS: Nine studies were included in the review, reporting on 825 participants randomly allocated to lithium or placebo. Lithium was found to be more effective than placebo in preventing relapse in mood disorder overall, and in bipolar disorder. The most consistent effect was found in bipolar disorder (random effects OR 0.29; 95% CI 0.09 to 0.93 ). In unipolar disorder, the direction of effect was in favour of lithium, but the result (when heterogeneity between studies was allowed for) did not reach statistical significance. Considerable heterogeneity was found between studies in all groups of patients. The direction of effect was the same in all studies; no study found a negative effect for lithium. Heterogeneity may have been due to differences in selection of participants, and to differing exposures to lithium in the pre-study phase resulting in variable influence of a discontinuation effect. There was little reported data on overall health and social functioning of participants under the different treatment conditions, or on the participants' own views of their treatment. Descriptive analysis showed that assessments of general health and social functioning generally favoured lithium. Small absolute numbers of deaths and suicides, and the absence of data on non-fatal suicidal behaviours, made it impossible to draw meaningful conclusions about the place of lithium therapy in suicide prevention. REVIEWER'S CONCLUSIONS: This systematic review indicates that lithium is an efficacious maintenance treatment for bipolar disorder. In unipolar disorder the evidence of efficacy is less robust. This review does not cover the relative efficacy of lithium compared with other maintenance treatments, which is at present unclear. There is no definitive evidence from this review as to whether or not lithium has an anti-suicidal effect. Systematic reviews and large scale randomised studies comparing lithium with other maintenance treatments (e.g. anti-convulsants, antidepressants) are necessary. Outcomes relating to death and suicidal behaviour should be included in all future maintenance studies of mood disorder.