Late-Life Generalized Anxiety: The Constraint of Categorization

Abstract

The American Journal of Geriatric Psychiatry published its first anxiety disorders theme issue 41⁄2 years ago. Since then, there has been a steady stream of articles related to late-life anxiety in the Journal. At least half of these articles have focused on generalized anxiety disorder (GAD). The current issue of the Journal is no exception: of the five anxiety-related articles in this issue, four pertain to GAD. The emphasis on GAD is understandable, given that it is one of the two most common Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV-defined anxiety disorders among older adults living in the community and is arguably the most common DSM-IV anxiety disorder among older persons in clinical settings. Before discussing the articles in this issue, it is useful to consider the evolution of the construct of GAD. For most of the 20th century, generalized anxiety was subsumed in the category of anxiety neurosis. Generalized anxiety was first described as a distinct disorder in 1980, when DSM-III separated it from panic disorder. However, GAD was primarily a residual disorder, diagnosed by exclusion. In DSMIII-R, the diagnostic criteria for GAD were substantially revised, in part because of poor reliability of the DSM-III construct. For the first time, GAD had its own defining characteristic: excessive or unrealistic worry unrelated to another Axis 1 disorder. In addition, the minimum duration of anxiety was changed from 1 month to 6 months, and the associated symptom criteria were revised to better reflect three clusters: motor tension, vigilance and scanning, and autonomic hyperactivity. Further major revisions occurred in DSM-IV. The core criterion of worry was refined, so that worry was not only excessive but also pervasive and difficult to control. Autonomic symptoms were removed, because they were considered to be less reliable than motor tension and vigilance and scanning in differentiating GAD from other anxiety disorders, especially panic. Finally, in common with other mental disorders, DSM-IV introduced an impairment criterion. What are the implications of these changes for the conceptualization of GAD? First, the substantial changes in criteria that took place between DSM-III and DSM-IV highlight the fact that symptoms of generalized anxiety, which are nonspecific and dimensional, are not easily categorized as a discrete disorder. Much has been written about the problems and limitations of attempting to categorize dimensional constructs; GAD is an example, par excellence, of this difficulty. Second, although the removal of autonomic symptoms from the DSM definition of GAD may have improved its reliability with respect to panic disorder, the definition of GAD moved from “panic-like” in DSM-III to “depression-like” in DSMIV. There is considerable symptomatic overlap between DSM-IV-defined GAD and major depression. This overlap, and the finding that GAD frequently evolves into major depression, poses a challenge to interpreting the findings of treatment studies of GAD, especially short-term studies that may not capture the dynamic longitudinal relationship between GAD and major depression. To what extent are these studies characterized by persons with a discrete anxiety disorder, versus a manifestation or variant of a