Should Anxiety and Insomnia Influence Antidepressant Selection: A Randomized Comparison of Fluoxetine and Imipramine

Background: The more sedating antidepressantsare often recommended for patients presenting with anxiety orinsomnia. We examine whether baseline anxiety or insomniasymptoms (1) show differential response to fluoxetine orimipramine or (2) predict differences between drugs in overallclinical response or likelihood of medication discontinuation.

Method: 336 health maintenance organizationprimary care patients beginning antidepressant treatment fordepression were randomly assigned to an initial prescription forfluoxetine or imipramine. All subsequent care (medication dosage,change, or discontinuation) was managed as usual by the primarycare physician. The 17-item Hamilton Rating Scale for Depression(HAM-D) and the Hopkins Symptom Checklist (SCL) anxiety anddepression subscales were administered prior to randomization and1 month later.

Results: Rates of improvement in insomnia (HAM-Dinsomnia items), agitation (HAM-D agitation item), and anxiety(SCL anxiety subscale) were essentially identical in the twotreatment groups. Baseline level of insomnia did not predictsignificant differences between randomization groups inimprovement in overall HAM-D score (p=.44) or SCL depressionsubscale (p=.44). Similarly, baseline level of anxiety did notpredict significant differences in improvement in HAM-D (p=.19)or SCL depression subscale (p=.31). Patients assigned tofluoxetine were significantly less likely to change ordiscontinue antidepressant medication during the first month, butthis difference did not vary according to baseline level ofinsomnia (p=.68) or anxiety (p=.25).

Conclusion: Among patients with moderatedepression, baseline levels of insomnia or anxiety should notinfluence the choice of fluoxetine or imipramine as an initialantidepressant.