Genetic Predictors of Lithium Response in Bipolar Disorder

The long term focus of this research program has been identification of genes for bipolar disorder. The investigators have recently obtained evidence from several lines of investigation to support the role of the gene for G protein receptor Kinase 3(GRK3) in bipolar disorder. Work to replicate and extend these results is continuing under NIH funding. In this clinical the investigators will extend the investigators' work into Pharmacogenetics to attempt to identify genes that are associated with medication response in bipolar disorder. Lithium is the first mood stabilizer medication and remains a mainstay of treatment. Many patients have an excellent response to lithium, tolerate it well, and are stabilized for years, while others do not. The reasons for this difference in response are unclear, but it is likely that genetic factors make a substantial contribution. The lack of good predictors of response frequently result in a time consuming trial and error clinical process to find the best medication. Such a trial and error process can take months with prolongation of patient suffering. Hence, there is a strong clinical need for predictors. The investigators have conducted a preliminary study with 92 lithium responders and 92 non-responders identified through retrospective detailed history and chart review. These subjects have been genotyped at 88 single nucleotide polymorphism (SNP) markers in 9 candidates genes relevant to lithium presumed mechanism of action for bipolar disorder. Four SNP markers in three genes showed nominally significant association to lithium response. One of the SNPs in the gene for neurotrophic receptor tyrosine kinase 2 (NTRK2), the receptor for brain-derived neurotrophic factor (BDNF), showed a strong association in patients who had predominantly euphoric a opposed to dysphoric mania (p=0.0005). Many data argue for the role of BDNF in the mechanism of antidepressants and mood stabilized action as well as susceptibility to bipolar disorder. No association was observed in those with dysphoric mania. This suggests that variations in this gene may operate in a clinically and genetically distinct subset of patients. It also argues for the importance of incorporating clinical subtypes into such analyses. These pilot results are preliminary but suggest the feasibility of such an approach. The investigators will conduct a prospective trial of lithium monotherapy in 100 patients with bipolar disorder. 200 patients who are unstable, mildly to moderately ill and not on lithium will be screened and then entered into 16-week stabilization phase where they will be treated and switched to lithium monotherapy. Patient stable on lithium will also be entered and other mediations withdrawn. After stabilization patients will be followed for one year or until a mood episode requires intervention. It is expected that 50% of patients will be stabilized and therefore 100 patients will enter the maintenance phase. Time to relapse and pharmacological intervention will be the primary outcome measure. This prospective sample will be used to replicate previous results at the NTRK2 and other genes. Analyses will be conducted to test for differences in survival curves between different genotypic group. Genomic control methods will be employed to detect or correct for possible stratification and heterogeneity. Clinical features of illness such as dysphoric mania, family history and rapid cycling will be employed as co-variates. Multivariate methods will also be employed in order to attempt to develop a multi-gene predictor of lithium response.