Lithium and Suicidality Revisited

To the Editor:

Andrea Cipriani, M.D., et al. (1) investigated the efficacy of lithium in the prevention of suicidal behavior in mood disorders, presenting a state-of-the-art meta-analysis of 32 randomized trials of lithium versus other compounds. They reported significantly lower suicide numbers with lithium (two suicides) than with other compounds (11 suicides) and concluded that patients who received lithium were less likely to die by suicide. The methods appeared sound and results were as expected. I would also like to join the majority and believe in an antisuicidal effect of lithium. Unfortunately, the results by Dr. Cipriani et al. (1) strongly depended on work by Greil et al. (2, 3), which may have been misunderstood. The meta-analysis treats the work by Greil et al. (2, 3) as reporting five suicides with carbamazepine and one suicide with lithium. Actually, however, Greil et al. (2) clearly stated, “During the treatment period, there was no suicide.” The misunderstanding may have been provoked by Kleindienst and Greil (4) themselves, who reported five suicide attempts and one completed suicide in the carbamazepine group, wrongly summarizing the results by writing, “All six suicides occurred in the carbamazepine group.” Unhappily, the mistaken results by Greil et al. (2, 3) are by far the strongest argument for an antisuicidal effect of lithium, wrongfully receiving a weight of 61.4% in the meta-analysis. The correct numbers would have yielded a much lower, if any, superiority of lithium. The story of lithium and suicidality might not be over after all.