• antipsychotic medications;
  • bipolar disorder;
  • confirmatory modeling;
  • factor analysis;
  • symptom structure;
  • treatment response

Background:  While the factor structure of clinical symptoms in schizophrenia has been examined and provided crucial information about the illness, there is much less information available in bipolar disorder. This study examined the structure of symptoms of bipolar disorder at an unmedicated baseline assessment and after double-blind treatment with ziprasidone, haloperidol, or placebo. We hypothesized, consistent with recent studies of schizophrenia, that the factor structure after treatment would be similar to the structure of untreated symptoms at study baseline.

Methods:  Hospitalized patients with manic (n = 363) or mixed (n = 71) bipolar episodes were rated with the Hamilton Depression Rating Scale (HAM-D) and the Mania Rating Scale [(MRS); derived from the Schedule for Affective Disorders and Schizophrenia – Change Bipolar Scale]. After 21 days of double-blind treatment, all patients were again rated with the MRS and HAM-D.

Results:  Exploratory orthogonal factor analysis (varimax rotation) including both HAM-D total scores and the MRS items found different five-factor solutions for mixed and manic patients at the unmedicated baseline assessment. Confirmatory modeling indicated that these models, with some modifications, fit the data well. At the endpoint, however, a single-factor solution was found for mixed and manic groups.

Implications:  Symptomatology in bipolar disorder is multifactorial in an acute and unmedicated state, with slightly different factor structures for mixed and manic episodes. Following treatment, a single severity dimension is detected. These results suggest that symptom dimensions in mania may be different from those seen in schizophrenia, where different elements of symptoms have been proven to have different functional correlates and treatment response.