Methods have been developed to assess and plot the longitudinal course of affective illness systematically in relationship to pharmacological interventions and psychosocial stressors. We have written a manual for life charting and have a computer program for data entry. Course of illness has been characterized systematically in more than 254 patients with primary affective disorders. These data redocument the early observations that the course of illness is often one of recurrences and a trend for progressive increases in frequency of cycling and severity of illness in refrac-tory patients. Ultra-rapid and ultradian cycling patterns have been identified and systematically documented for the first time. This cycle acceleration is occurring in the context of a decreas- ing incidence of psychosocial stressors precipitating affective episodes; i.e., as in sensitization and kindling, they are occurring autonomously. Pharmacological intervention must be considered against this backdrop. We have noted that one of the previous traditional interventions for bipolar depression, tricyclic antidepressants (TCAs), appear causally related to switches into mania in 35% of patients. In those individuals who show this induction, we have observed a pattern of increased rapidity of cycling in the year prior to NIMH admission and longer hospital stays at NIMH, suggesting that TCA-induced switches are a marker if not the cause of a more rapid cycling course of illness. The phenomenon of lithium dis-continuation-induced refractoriness has been described in 14% of our patients and now by others in the literature as well, providing an additional rationale for continuing patients on long-term maintenance treatment. We have begun to assess neurobiological correlates of course of illness using a variety of neurotransmitter and endocrine markers as well as peptides and proteins in the CSF. CSF N-CAMs are elevated in bipolar patients, as is CSF total protein, particularly in bipolar I males. In patients studied after a prolonged period of medication-free evaluation we observed that rapid cyclers had non-significantly higher T4 and free T4 levels compared with non-rapid cyclers. These data contrast with previous findings in the literature in medicated patients in whom rapid cycling was often associated with hypothyroidism. Rapid cyclers respond poorly to either lithium or carbamazepine monotherapy, but 50% respond well to the combination. Differential response to phar- macological agents as a function of course of illness will be explored further within the NIMH-Stanley Foundation Bipolar Research Network.
