• Three separate lithium chloride doses, calculated according to body weight, and a placebo were administered under doubleblind conditions to 68 manic inpatients. The relationship of lithium chloride treatment dose to steady-state serum lithium levels (day 7 to 10 of treatment) and clinical response were examined.

High (0.72 mEq/kg/day) and medium (0.5 mEq/kg/day) lithium chloride doses were more efficacious than placebo (P <.001 and p <.05, respectively), as determined by decrements in global mania ratings (day 7 to 10 of treatment). A low dose (0.24 mEq/kg/day) was not found to be more efficacious than placebo.

The proportion of patients with improved manic ratings increased markedly as a function of increased steady-state serum lithium level (x2 for trend in proportions, 17.91; p <.001).