The association between course of illness and subsequent morbidity in bipolar I disorder.

OBJECTIVE: We examined the relationship between certain bipolar I disorder clinical course variables over 5 years with outcome over the subsequent 5-year period.
METHODS: Prospective observational follow-up data of 123 bipolar I subjects were analyzed. Predictive clinical variables included the frequency and direction of switches, and the quantity, polarity and length of affective periods. Outcome variables were an affective burden index (ABI) accounting for week-by-week severity and weeks hospitalized. Bivariate analyses guided the selection of predictors for multivariable analyses against the outcome variables.
RESULTS: Affective burden index: while the number and direction of switches, the number of polyphasic episodes, weeks in hypomania/mania/mixed state, weeks in minor/major depression, weeks in at least marked affective syndrome, and weeks in any affective syndrome all had bivariate correlation (p<0.01) with the ABI, only weeks in hypomania/mania/mixed state and weeks in minor/major depression made significant contributions in the multivariable analysis (p<0.01) with the ABI. Weeks hospitalized: weeks in at least marked affective syndrome were significantly correlated with weeks hospitalized in bivariate analysis (p<0.01), and maintained a contribution to weeks hospitalized in the multivariable analysis (p<0.01).
CONCLUSIONS: The quantity and severity of weeks in symptomatic affective states are possibly greater predictors of affective burden in bipolar I patients than the quantity and direction of affective switches.