Models of bipolar disease and their clinical implications

In assessing the current state of progress in understanding the pathogenesis of bipolar disease and seeking to point to future directions, we need to think of models of disease that provide the appropriate scaffolding. In this issue, four aspects of our clinical/basic neuroscience “toolbox” that may provide important leads in our desire to push the envelope of understanding are described. All of these areas, including neuropathology (postmortem studies), neuroimaging, genetics, and molecular biology, are relatively recent additions to the toolbox of the clinical researcher. Before I comment on each of these areas individually, it is noteworthy to point to those tools that will allow us to evaluate the success of our efforts. As clinical scientists, we are searching for appropriate surrogate markers to assess prognosis, treatment response, and risks for future difficulties. Any understanding of brain circuitry and the functional components of the brain and behavior will provide important leads, but we need to pay attention to the following: these action plans do not take into account that bipolar disease is a very early-onset disorder. Indeed, in a recent survey of over 2800 patients with bipolar disease, the median age of onset was 17.5 years (E. Frank and D.J. Kupfer, personal communication, 2000). The disease has a strong developmental component, so that the early identification of subsyndromal presentations will be a vital necessity in defining phenotypes for study. In addition to the complex phenotypic description of this disorder, the next important issue is that treatment response, unlike treatment response in recurrent depression, is a constantly moving target, with change rather than constancy being the rule. Using our neuroscience toolbox for biological assessment, we assume there is persistence in each clinical phase, but this is likely a false assumption. The numerous disease transitions represent an essential challenge to the interpretation of our neuroscience investigations. We will need to develop better clinical assessment measures that allow us to take into account these variations in clinical state. Furthermore, we will need to provide better measures of functional state as well as the statistical strategies to appropriately analyze these ongoing clinical and functional changes. Another key issue in the study of bipolar disorder is taking sufficient account of both psychiatric and other medical conditions. This is particularly important because it would appear that patients with bipolar disease are not only more liable to suffer from other psychiatric conditions such as panic disorder and alcohol and drug dependency, but also suffer more frequently from concurrent medical disease, particularly cardiovascular disease. Another complication of assessment relates to psychosis and the fact that mania can occur with and without psychotic features. Because of the extensive need for improved assessment (with clear nosologic implications that would broaden current DSM-IV classification to include bipolar spectrum conditions), it is appropriate to recommend that studies utilizing any of our clinical neuroscience tools be based on standardized, systematic clinical assessment methods across sites.