Bipolar outcome in the course of depressive illness. Phenomenologic, familial, and pharmacologic predictors.

Twenty percent of a cohort of 206 outpatient depressives with no past bipolar history switched during prospective observation. These 41 probands developed manic periods on the average of 6.4 years (median 4, range 1-25) after their first depressive episode. The change in polarity occurred throughout the life span, but was most common in adolescence and early adulthood. The following variables were found useful in predicting this outcome: onset less than or equal to 25 years, bipolar family history, loaded pedigrees, precipitation by childbirth, hypersomnic-retarded phenomenology, and pharmacologically-mobilized hypomania. Although the respective sensitivities of these findings were relatively low (32-71%), their specificities ranged from 69% to 100% for bipolar outcome; the diagnostic specificity of any 3 of these variables when combined was 98%. When compared with nonbipolar depression, bipolar disorder was seldom chronologically secondary to nonaffective psychiatric disorders. These findings suggest that many young depressives with lethargy and oversleeping are not manifesting a "neurotic" disorder, but rather a precursor of primary bipolar affective disorder. Finally, a psychotically depressed adolescent or young adult with positive bipolar family history should be observed for eventual bipolar outcome, especially when the clinical presentation is that of stupor.