ological concerns have been raised (Bowden et al., 2000; Calabrese et al., 2001a; Goodwin, 1999). Some also caution that although the data is mostly positive, the anti-suicidal effect of lithium may not be as strong as originally thought (Bowden et al., 2000).
One of the confounding issues in these studies is the time course of psychopharmacological treatment. Decreased rates of suicide are most pronounced when lithium has been used for a minimum of 2 years (Baldessarini and Tondo, 1999). Rates were reduced only while the patients took lithium; following discontinuation of lithium, the rates began to rise to levels similar to those seen prior to the commencement of lithium. Rapid discontinuation of lithium may lead to a more dramatic increase in rates of suicidal behavior as compared to more gradual discontinuation. Early studies, because of their abrupt discontinuation of lithium, may have increased placebo relapse figures (Bowden et al., 2000). Tondo, Baldessarini and colleagues (Baldessarini and Tondo, 1999; Tondo and Baldessarini, 2000; Baldessarini et al., 1999; Tondo et al., 1997) noted that the latency from onset of bipolar disorder to lithium maintenance in their patients averaged 8.3 years, but that half of the suicidal acts had occurred in the first 7.5 years. Thus, it may be of crucial importance to commence lithium treatment as early as possible in the course of bipolar disorder for patients thought to be at risk for suicidal behavior. It is noteworthy that lithium and clozapine (see below) are both effective in reducing suicidal behavior and both require regular clinic visits and blood tests. This suggests a benefit from regular clinic monitoring.
Nonadherence with medication, particularly lithium, is a critical issue for individuals with bipolar disorder and one of the primary reasons for poor treatment response (Goodwin and Jamison, 1990). Since lithium treatment is associated with an almost 8-fold decreased suicide rate (Tondo and Baldessarini, 2000), this has a serious impact on suicide risk. Research has shown that almost one-half of patients with bipolar disorder are non-adherent to lithium treatment at some point in their lives, and one-third are non-adherent two or more times (Jamison and Akiskal, 1983; Jamison et al., 1979). Younger males within the first year of lithium treatment and those patients who have elevated moods and a history of euphoric manias, especially those who complain about missing the “highs” of their illness, are more likely to be nonadherent (Goodwin and Jamison, 1990). Many people stop taking their medication after being released from the hospital, one of the factors causing significantly increased risk of suicide during this period (see below). Furthermore, clinical research with bipolar populations is very difficult due to poor treatment adherence (Goodwin and Jamison, 1990; Goodwin, 1999; Jamison and Akiskal, 1983; Jamison et al., 1979), and the poor adherence rate makes interpreting results more difficult and the conclusions less powerful in many studies.