Data Question Claims That Lithium Reduces Suicide

For years, lithium has been thought to have a protective effect against suicide in patients with bipolar disorder. Recent data show, however, that in one sample of patients with the mood disorder, this is not the case.

Researchers conducting a new data analysis of the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) study compared subjects with bipolar disorder who took lithium in the six months prior to a suicide attempt or completion with subjects who did not make such an attempt and failed to find an association between lithium and suicide attempts or completions (p = .41).

The findings were published in the April 8 online edition of the Journal of Clinical Psychiatry.

Lead author lauren Marangell, M.D., told Psychiatric News that for clinicians treating patients with bipolar disorder, suicide prevention is an important concern and that there is reason to challenge an automatic assumption that lithium protects patients against suicide based on past studies. “Until we have further data,” she noted, “the clinician must weigh the risks and benefits for each individual patient.”

At the time of the study analysis, Marangell was a professor of psychiatry at Baylor College of Medicine and codirector of the National Institute of Mental Health's Bipolar Trials Network. She is currently a distinguished scholar at Eli Lilly and Company.

The STEP-BD study is a multisite study designed to evaluate clinical outcomes in patients with bipolar disorder. Researchers recruited 4,360 patients with bipolar disorder from September 1998 through November 2004. They examined the association between eight cases of suicide and 262 suicide attempts and medications prescribed to the patients in the six months preceding the suicidal behavior.

According to the report, there a number of published studies supporting the use of lithium as a suicide-protective agent in patients with bipolar disorder. One 2003 study conducted by Ross Baldessarini, M.D., also published in the Journal of Clinical Psychiatry, showed a nearly 15-fold risk reduction associated with lithium treatment, for example.

Marangell noted, however, that Baldessarini's analysis and other studies showing lithium's suicide-protective effects are mostly retrospective and not randomized, so they may reflect a bias toward those who respond positively to lithium.

In her analysis, Marangell matched 93 cases of STEP-BD participants who attempted or completed suicide with 93 participants who did not based on gender, age, and a propensity score that accounted for bipolar subtype, marital status, age at illness onset, and history of psychosis. She categorized each group by whether they were continually prescribed lithium, had only partial exposure to lithium, or were not prescribed lithium in the six months prior to the suicide or attempt; she also did so for controls, also looking at the same six-month period, as indicated by clinical-monitoring forms completed throughout the study.

Of the 53 subjects who had a suicide attempt or completion with no lithium in the six months prior to the event, 32 matched controls also had no lithium in the six months prior, eight had some lithium use, and 13 had steady lithium exposure.

Of the 18 subjects who had some lithium use and who attempted or completed suicide, five matched controls had no lithium use, six had some, and seven had continual lithium exposure.

Finally, of the 22 cases of suicide or attempt with steady lithium use, nine had matched controls with no lithium use, four with some lithium use, and nine with steady lithium use.

The findings do not support an association between lithium use and reduced suicide attempts or completions, according to the report, and there was no association between suicide events and use of valproate, carbamazepine, lamogatrine, or atypical antipsychotic medications.

Marangell also found that in 23 cases of a suicide or attempt, there was steady exposure to a selective serotonin reuptake inhibitor (SSRI) in the six months prior to the event. Within the matched controls, 20 had no exposure to an SSRI, one had some, and two had exposure throughout the six-month period.“ These findings suggest that exposure to an SSRI medication is associated with a suicide event over a six-month period prior to the event (p<.0001).”

Marangell emphasized, however, that “this study was not designed to assess the association between SSRIs and suicide risk”; thus, these findings must be interpreted with caution.

She also pointed out that her analysis on the association between lithium and suicide had certain limitations: for example, the overall rate of suicides and suicide attempts was low, she said. “Mostly the events included were suicide attempts, with only a few completed suicides. The risk factors for completed suicides may be different from those of suicide attempts, but I think both are clinically important.”

To further elucidate her findings, Marangell would like to conduct a prospective, randomized clinical trial in a population at high risk for suicide.

The study was funded by the American Foundation for Suicide Prevention and National Alliance for Research in Schizophrenia and Depression. The STEP-BD Project was funded by the National Institute of Mental Health.

“Case-Control Analyses of the Impact of Pharmacotherapy on Prospectively Observed Suicide Attempts and Completed Suicides in Bipolar Disorder: Findings From the STEP-BD” is posted at<http://www.psychiatrist.com/privatepdf/oap/ej07m03621.pdf>.▪