The estimated lifetime risk of suicide in persons with bipolar disorder ranges from 8 to 20 percent, a rate that is 10 to 20 times the rate in the U.S. general population. Lithium and the anticonvulsant agent lamotrigine are the only drugs for which long-term efficacy in the treatment of bipolar disorder has been reliably established. Suicide rates in patients taking lithium have been shown to be lower than rates in patients not taking lithium. Nonetheless, lithium use in the United States had declined over recent years in favor of anticonvulsant agents, including divalproex. In this study, Goodwin and colleagues examined the risk of attempted and completed suicide in patients treated for bipolar disorder, to identify, among other things, periods of exposure to lithium, divalproex, and carbamazepine.

Participants in this retrospective cohort study included those 14 years or older with a record of outpatient treatment for bipolar disorder, who filled at least one prescription for lithium, divalproex, or carbamazepine during the study period. Outcome measures were suicide attempts requiring or not requiring hospitalization and suicide deaths. Treatment exposure was estimated on the basis of prescriptions filled, and confounding factors considered included a variety of organic diseases and mental health disorders that could have influenced the risk of suicide or the choice of mood stabilizer prescribed.

The authors identified 20,638 health plan members who fit the study’s inclusion and exclusion criteria. There were 53 suicides, 338 attempts resulting in hospitalization, and 642 attempts identified in emergency departments. Groups were subdivided into patients with no exposure to the studied drugs, patients exposed to lithium alone, to divalproex alone, to carbamazepine alone, or to a combination of these agents.

For all outcomes measured, rates of suicide and suicide attempts were substantially greater when patients were exposed to divalproex than to lithium; hazard ratios were 2.7, 1.7, and 1.8 for suicide death, attempt resulting in hospitalization, and attempt ascertained in the emergency department, respectively. Risks were also considerably higher, at 2.2, 1.6, and 1.4, for these three outcomes when patients taking no mood stabilizers were compared with those taking lithium alone.

Overall, the risk of suicide attempt or suicide death was one and one half to three times higher in patients treated with divalproex than with lithium. This finding held even after controlling for characteristics that might indicate greater severity of illness or suicide risk and thus influence prescription choice. The authors emphasize that further research is needed to determine whether suicide risk is also lowered with lithium-anticonvulsant combinations, because the number of patients using combination treatment in this study was too small to allow accurate estimation of suicide risk for that subgroup. Current prescribing patterns, which have moved away from lithium in recent years, should be reevaluated.