• bipolar I and II disorders;
  • episodes;
  • kindling;
  • lithium;
  • maintenance treatment;
  • outcome;
  • treatment latency

Objectives: To test whether longer treatment-delays or more pretreatment illness episodes are followed by diminished response to lithium maintenance. Methods: In 360 DSM-IV bipolar I (n=220) or II (n=140) patients, effects of latency from illness onset to starting lithium and number of pretreatment episodes were evaluated by survival analysis based on the number of months stable before a first recurrence on lithium. Factors associated with treatment latency were identified by regression modeling. Relationships of time, episode number, and morbidity before treatment to the overall proportion of time ill on lithium were also tested by nonparametric correlation. Results: Latency to first lifetime lithium maintenance averaged 8.3 years, with 9.3 episodes/subject. Time stable before a first recurrence on lithium averaged 29.6 months and was unrelated to treatment latency (in terciles) or to a high (≥ten), intermediate (four–nine), or low (<four) number of prior episodes. Overall morbidity during 4.6 years of lithium maintenance was also unrelated to these pretreatment factors. More episodes/year and percentage of time ill before treatment anticipated shorter treatment latency, with greater relative improvement, but were unrelated to morbidity during treatment. Conclusions: Treatment latency and prelithium episode number were unrelated to morbidity during treatment. Although multiple untreated episodes can lead to severe disability, lithium evidently can greatly limit morbidity, even after years of delay and multiple episodes of bipolar illness.