To the Editor: In a recent issue of the Journal, Maria A. Oquendo, M.D., et al. (1) attempted to characterize unipolar depression and determine the stability of 24 symptoms, five dimensional categories, and three subtypes over two episodes up to 20 years apart after adjusting for severity. They convincingly made the case that only the severity of anxiety and suicidal ideation were correlated and only to a moderate degree. They concluded that this finding would support a conception of unipolar depression as a single pleomorphic illness.

Our experience with the Systematic Treatment Enhancement Program for Bipolar Disorder Cohort raises a number of questions. We found that depression predicted suicidal ideation regardless of suicide attempt history but that anxiety and neuroticism predicted suicidal ideation only among those with a negative history of prior attempts. One of our collaborators, in examining anxiety comorbidity in this population, found that bipolar subjects with a lifetime history of anxiety disorder were more than twice as likely to have a lifetime history of suicide attempts as well. This raises the question of whether the two factors that Dr. Oquendo and colleagues found correlated across episodes might be confounded.

Dr. Oquendo et al. did not provide data on the frequency of prior attempts. Only 1.7% made an attempt during the study period. We found that among those with a history of prior attempts, poor role functioning and poor openness predicted suicidal ideation rather than anxiety and neuroticism and that extraversion was protective. It may be helpful in understanding the propensity for suicidal ideation to stratify by attempt history.

Joiner (2) postulated that attempting may lower the threshold for suicidal behaviors by sensitizing related cognitive structures and strengthening opponent processes. These, in turn, may contribute to the evolution of depressed states.

Dr. Oquendo et al. did not examine traits, such as neuroticism, that we find interacting with depression to influence suicidal ideation. Neuroticism, extraversion, and openness have been found to have both state and trait components, and we hope to examine the stability and contribution of such personality factors prospectively.

Finally, Dr. Oquendo et al. noted that these issues may have a bearing on the selection of acute and maintenance medication treatments. If anxiety and suicidal ideation are stable features, treatment must take account of that. We have suggested that distinguishing between negative and positive attempters provides a rationale for treatment selection in bipolar patients. Overall, aggressive treatment of depression appears warranted but also for attempters with the features mentioned, a focus on role functioning may minimize adverse experiences and reduce the risk of suicide. For nonattempters who fit this profile, more vigorous pharmacological treatment of anxiety and psychosocial treatment of neurotic thinking may be indicated.