OBJECTIVE: In unipolar depressed patients participating in trials on antidepressants, we investigated if illness characteristics at baseline could predict conversion to bipolar disorder.
METHOD: A long-term register-based follow-up study of 290 unipolar depressed patients with a mean age of 50.8 years (SD=11.9) participating in three randomized trials on antidepressants conducted in the period 1985-1994. The independent effects of explanatory variables were examined by applying Cox regression analyses.
RESULTS: The overall risk of conversion was 20.7%, with a mean follow-up time of 15.2 years per patient. The risk of conversion was associated with an increasing number of previous depressive episodes at baseline, [HR 1.18, 95% CI (1.10-1.26)]. No association with gender, age, age at first depressive episode, duration of baseline episode, subtype of depression or any of the investigated HAM-D subscales included was found.
LIMITATIONS: The patients were followed-up through the Danish Psychiatric Central Research Register, which resulted in inherent limitations such as possible misclassification of outcome.
CONCLUSION: In a sample of middle-aged hospitalized unipolar depressed patients participating in trials on antidepressants, the risk of conversion was associated with the number of previous depressive episodes. Therefore, this study emphasizes that unipolar depressed patients experiencing a relatively high number of recurrences should be followed more closely, or at least be informed about the possible increased risk of conversion.