Lithium for Suicidal Behavior in Mood Disorders

Objective: To test the hypothesis that lithium augmentation of enhanced usual care will reduce the rate of repeated episodes of suicidal self-directed violence (repeated suicide attempts, interrupted attempts, hospitalizations specifically to prevent suicide, and deaths from suicide) in participants with bipolar disorder or depression who have survived a recent event. Background: The hypothesis that lithium can prevent suicide in patients with bipolar disorder and depression is based on data from observational studies and randomized clinical trials conducted to evaluate other outcomes. The question about the effectiveness of lithium for suicide prevention is one of major scientific, clinical, and public health significance. There have been no adequately powered clinical trials conducted specifically to evaluate suicide behaviors as an outcome. Two recent randomized clinical trials failed to recruit adequate numbers of subjects to be conclusive. The VHA, as a large national healthcare system with an established program for identifying new suicide attempts, evaluating patients for underlying mental health and medical conditions, providing needed services, connecting Veterans to state-of-the-art suicide risk management, and monitoring outcomes is uniquely able to conduct a large scale clinical trial of lithium for suicide prevention. The rationale for the study is based on the following: * Data from observational studies and double-blind randomized clinical trials suggest that lithium can prevent suicide-related behaviors in patients with bipolar disorder and major depression. * The high risk of suicide in Veterans receiving health care services from VHA has persisted despite extensive improvements in mental health services and in programs for suicide prevention. * Each month, there are over 1,100 unique VHA patients with bipolar disorder or depression who attempt suicide and survive. * Surviving a suicide attempt is the most powerful known risk factor for death from suicide in VA and elsewhere. * Approximately 15% of VA survivors reattempt or die from suicide within one year. * Evaluating rates of reattempts in those who have survived attempts is an established and effective method for testing interventions that may prevent suicide. * Experimental treatment in CSP-590 would supplement usual care for major depression or bipolar disorder. * Study procedures for the management of suicide risk would meet or exceed VA standards and requirements. * Study procedures optimize the safety of lithium, including the potential risk of overdoses, and meet or exceed all published practice standards. The trial will utilize multiple strategies to minimize risks including frequent monitoring and assessment, determination of lithium levels during titration and at steady state, and dispensing medications in limited quantities in blister packs. * The investigator's survey of VA psychiatrists indicates that the question is clinically important and compelling and that a clinical trial that demonstrated the hypothesized effect would transform the clinical management of suicidality. Design: Randomized, double-blind, placebo-controlled clinical trial of lithium versus placebo augmentation of enhanced usual care. Patient population: VHA patients with bipolar disorder or depression who have survived a recent episode of suicidal self-directed violence. Primary outcome: Time to the first repeated episode of suicidal self-directed violence, including suicide attempts, interrupted attempts, hospitalizations specifically to prevent suicide, and deaths from suicide Duration: Total study duration will be 4.5 years. Recruitment will occur over 3 years. Participants will be followed for one year. Sample size calculations and number of sites required: The design of the study is based on testing for a 37% reduction in the rate of repeated suicidal self-directed violence, a figure based on an effect size of approximately 43% observed in recent studies and then allowing for attenuation due to non-adherence. Adjusting for potential data loss due to attrition, 90% statistical power to detect a significant 37% reduction in reattempt rates at 5% overall type I error would require 1862 subjects. With recruitment of 20% of eligible subjects over a three year period, this would require approximately 9310 potentially eligible subjects. Based on current suicide surveillance data, this could be achieved with 29 sites.