Restoring physical function through intensive visual simulation in a ballet dancer with complex regional pain syndrome.

Lithium is an effective therapy for the acute treatment of mania and long-term prevention of relapses in bipolar disorder, as well as an effective agent in treatment-resistant depression. Because lithium may stabilise the mood and decrease aggression and impulsivity, it might conceivably reduce the risk of suicide. More than 10 randomised trials reported highly imprecise estimates of the effect of lithium on suicide deaths (most trials had 0 events) and almost all reported null estimates of the effect of lithium on non-lethal suicidal behaviours. The largest trial (CSP-590) found little evidence of a substantial effect of lithium on suicide-related events over 1 year. Meta-analyses of randomised trials show conflicting results depending on different analytical choices, such as alternative inclusion criteria or methodological approaches, including how trials with zero events were handled statistically. In contrast, more than 20 observational studies found a strong association between lithium and a lower risk of suicide deaths. The discrepancy between observational and randomised evidence could be due to biases in observational studies (eg, incomplete adjustment for suicide risk factors) or the use of a surrogate outcome for suicide in trials. We emulated target trials of lithium using electronic health records of the VA Healthcare (the largest integrated healthcare system in the USA) which were linked to a database that included suicide risk information. The first target trial was similar to the CSP-590 trial and was emulated for benchmarking purposes. Replicating CSP-590’s short-term estimates may increase confidence in our effect estimates for the 10-year risk of suicide deaths. We emulated a second target trial to estimate the effect of lithium on the 10-year risk of, separately, suicide deaths and non-lethal suicidal behaviours. The estimated 10-year risk ratio (95% CI) for lithium versus no lithium was 1.00 (0.86 to 1.15) for suicide deaths and 0.96 (0.90 to 1.02) for non-lethal suicidal behaviours. While lithium remains a first-line treatment for affective disorders, our findings suggest no substantial effect of lithium on suicide and, at most, a small protective effect on non-lethal suicidal behaviours, when added to the treatment of mood disorder patients already receiving pharmacological therapy. Explicitly emulating a target trial using a large database containing suicide risk information reduced the differences between effect estimates from observational studies and randomised trials and helps complement the results from the latter for suicide prevention research.