Treating anxiety by discontinuing antidepressants: a case series

When a patient has symptoms of anxiety, while taking an antidepressant for depression, is it possible that the antidepressant is part of the problem? Can antidepressants cause anxiety? If this were so, even if relatively rare, it would have widespread implications because of the broad use of antidepressants. However, antidepressants are widely used as a treatment for anxiety. Therefore, unless suggestive evidence were to emerge to implicate them as a potential exacerbating factor, broad use of antidepressants would likely continue for patients whose depression has improved but whose anxiety has not responded, or worsened. In that context we present 12 patients whose anxiety diminished substantially when antidepressants were tapered off, as reflected in Clinical Global Improvement Scale scores assigned by their respective clinicians. Mean duration of antidepressant taper was 17 weeks (range 0-48), as suggested by limited prior evidence supporting very slow taper rates for this purpose. Alternative treatments for depression were often used for these patients as antidepressants were tapered, particularly lithium and lamotrigine, but none of the alternatives used are generally regarded as having anti-anxiety effects. Patients with bipolar disorder diagnoses, including schizoaffective disorder, were specifically excluded. In many of these cases, other medications that might have anti-anxiety effects (including buspirone, quetiapine, olanzapine, gabapentin, and diphenhydramine) were also tapered off. Results suggest that antidepressants may actually cause anxiety in some patients with unipolar depression. Alternatively, lamotrigine or lithium may have more anti-anxiety effects than generally recognized; or these patients may have had subtle bipolar disorder despite the absence of symptoms meeting formal criteria, supporting the "bipolar spectrum" perspective on mood disorder diagnosis. This study is limited by the outcome measure used, which assesses anxiety only indirectly in the context of global improvement. However, in view of the broad implications of the findings, these preliminary observations warrant further consideration. Some patients with anxiety may be treatable not by adding medications, but rather by tapering off existing ones.