Medication use patterns and 2-year outcome in first-admission bipolar disorder with psychotic features


  • None of the authors have a current commercial conflict of interest in connection with this manuscript. Dr Gibson had been an employee of Eli Lilly during the writing of the manuscript but is no longer so employed.

Thomas J. Craig, MD, MPH, 6804 Wild Rose Court, Springfield, VA 22152, USA.
Fax: (202) 273 9097;


Objective:  This study examined the patterns and predictors of medication use and 24-month course/outcome in first-admission patients with bipolar disorder (BD) with psychotic features.

Method:  An epidemiologic sample of 155 first-admission patients with research diagnoses of BD with psychotic features received intensive clinical assessments at baseline, 6- and 24-month follow-ups and telephone assessments at 3-month intervals. Use of antipsychotics, antimanic agents (lithium and anticonvulsants), and antidepressants was determined from self-reports corroborated by external sources where possible. Outcome was assessed with the Global Assessment of Functioning (GAF), consensus evaluations of illness course, and time in remission.

Results:  More patients received antipsychotics (80.0%) than antimanics (52.3%) at discharge. At the 24-month point, 44.6% reported using no medications; 19.4 and 38.8% received antipsychotics and antimanics, respectively. Health insurance and early regular antimanic use were the strongest predictors of regular antimanic use during the 6–24-month follow-up. Early and later regular use of antimanics and less use of antipsychotics were associated with higher GAF scores and greater time in remission. Conversely, failure to use antimanics regularly was associated with the worst outcome on these two measures. Early functioning as measured by the GAF strongly predicted outcome for all three measures. Younger and non-White subjects were significantly less likely to attain complete remission during the 6–24-month follow-up.

Conclusions:  The findings confirm clinical trial data that for some outcome measures use of antimanics is associated with good outcome in bipolar populations while failure to use these medications regularly was common among subjects with the worst outcomes. In addition, our findings that higher educational attainment and having health insurance predicted regular antimanic use and (for the latter) better outcome underscore the effect of socioeconomic influences while achieving a complete remission seems unrelated to medication use but strongly predicted by age and ethnicity. Finally, the fact that early regular treatment and early high functioning strongly predicted better outcome supports the need for the early diagnosis and treatment of patients with this disorder.