Lithium is still the first drug of choice for treatment of bipolar disorder. However, it is also evident that a certain proportion of patients with bipolar disorder do not respond to maintenance treatment with lithium. If predictors of lithium response could be identified before initiation of maintenance treatment, it would help in selecting an appropriate mood stabilizer to minimize the risk of relapse. Many clinical predictors of positive lithium response have been reported. These include pure classical mania, psychotic features, small number of episodes, positive family history, lower scores of neuroticism, and patterns of course-of-illness (Carroll, 1979; Maj et al., 1989, for review). However, these variables still cannot completely predict lithium response. Few pharmacological or biochemical predictors of lithium response have been reported. High red blood cell :plasma lithium ratio, which reflects compliance, has been reported to be associated with positive response (Harvey and Kay, 1991). A possible genetic marker for lithium response, the C973A polymorphism in the inositol polyphosphate 1-phosphatase gene, has been reported (Steen et al., 1998), although it has not yet been replicated.