Le traitement au long cours des troubles de l’humeur conduit à une diminution marquée de la mortalité, notamment par suicide, de ces patients dont le risque suicidaire est majeur. Les données s’accumulent depuis 40ans pour proposer que le lithium bénéficie d’un effet suicide, lui conférant une place à part dans l’arsenal thérapeutique. En effet, le lithium présente un effet préventif des conduites suicidaires dans le trouble bipolaire comme dans le trouble unipolaire. Cet effet semble indépendant de son efficacité thymorégulatrice et supérieur à ce qui est observé avec les anticonvulsivants. Les mécanismes d’action du lithium continuent d’être étudiés tant ils sont complexes. Les pistes se multiplient pour envisager que le lithium agisse sur différents traits de vulnérabilité suicidaire ou qu’il puisse corriger des anomalies cérébrales associées au suicide. Les travaux scientifiques, de la molécule à l’épidémiologie, concourent à proposer le lithium en première ligne pour lutter contre le risque suicidaire des patients souffrant de troubles thymiques. Il est alors envisageable de recourir au lithium dans une perspective dimensionnelle, pour traiter le risque suicidaire, l’impulsivité et l’agressivité, et ce, indépendamment du traitement thymorégulateur choisi pour le patient.

Affective disorders are associated with an increased risk of suicidal behavior. For instance, 18% of bipolar males having a past history of suicide attempt die by suicide. In comparison to untreated patients, those receiving a continuous pharmacological treatment for bipolar disorder have a reduced surmortality, including by suicide (Angst et al., 2002). Evidence shows that lithium has specific antisuicidal properties. Meta-analyses including long-term randomized controlled trials have shown a 80% decrease of suicidal behaviors in patients receiving lithium for 18months or more. In a recent meta-analysis including controlled studies comparing lithium vs. placebo or active drugs in both unipolar and bipolar disorders, Cipriani et al., 2013 have shown significant reduction of suicide mortality in patients on lithium. In addition, lithium is more efficient than anticonvulsants to prevent suicidal behaviors. In a large Danish study (Sondergard et al., 2009), although the rate of suicide was higher during periods when patients purchased anticonvulsants than during periods with lithium, the suicide rate decreased with the number of prescriptions in a rather similar way for patients first treated with lithium and patients first treated with anticonvulsants. But switch to or augmentation with lithium to patients initiated on anticonvulsants was associated with a significantly reduced rate of suicide whereas a switch to or augmentation with anticonvulsants to patients first started on lithium showed no additional effect on the suicide rate. Antisuicidal effect of lithium is now admitted in bipolar disorder as well as in unipolar depression and seems to be independent of the effect on mood. Ahrens et al., 2001, have investigated a group of high-risk patients with recurrent affective disorders who had committed suicide attempts before the start of lithium. According to their recurrence-related response to long-term lithium prophylaxis, nearly 50% of poor responders did not exhibit any further suicidal behavior during lithium treatment. Khan et al., 2011, have conducted an exploratory proof of concept-randomized trial comparing suicidality in depressed patients assigned to citalopram+lithium or to citalopram+placebo for 4weeks. Lithium when used in therapeutic doses (>0.5mEq/L) was associated with prevention of suicidal thoughts and behaviors. It is important to know that lithium has to be discontinuated very slowly (by 10% a month). After the discontinuation of lithium, suicidal acts are more frequent in the first year than at later times or before start of lithium treatment (Tondo et al., 1998). The mechanism of action of lithium remains poorly understood. It may be due to an action of lithium on suicidal physiopathology. Indeed, lithium reduces aggressive and impulsive behaviors in both humans and animals, and possibly modulates suicidal vulnerability traits. Indeed, decision-making alterations are associated with suicidal phenotype and rely on orbitofrontal cortex dysfunction. In euthymic bipolar patients, decision-making performances have been correlated with lithium dose (Adida et al., submitted). In addition, main molecular targets of lithium, Wnt - GSK3 – β-catenin and IP-3 pathways, have been involved in suicidal vulnerability through post-mortem and genetic studies. In addition, SAT-1 pathway could be involved in antisuicidal effect of lithium. To summarize, for a therapeutic dose, lithium has well demonstrated preventive effect on suicide behaviours in mood disorders, independently of its action on mood. It improves clinical dimensions associated with suicidal risk (impulsivity, aggression), raising the question of dimensional prescription. Finally, further studies are needed to elucidate the antisuicidal mechanism of action of lithium. However, it could be suggested that lithium acts on suicidal vulnerability factors at a macroscopic level (decision-making, cerebral volumes) as well as microscopic level (molecular pathways).