The relationship between hypochondriacal concerns, as assessed by the Illness Attitude Scales, and depressive symptoms was examined in a sample of 100 drug-free outpatients with major depressive disorder. These patients were treated with fluoxetine for 8 weeks, and the effect of treatment on hypochondriacal symptoms was examined. All patients were administered the Structured Clinical Interview for DSM-III-R, the Hamilton Depression Rating Scale, the Symptom Questionnaire, and the Personality Disorders Questionnaire-Revised. We found little relationship between severity of depressive symptoms and hypochondriacal concerns. Measures of anxiety, somatic symptoms, and psychological distress were more consistently related to these concerns. Similarly, patients with either histrionic personality disorder or a lifetime history of panic disorder had greater hypochondriacal concerns than patients without these diagnoses. After open treatment with fluoxetine, the degree of hypochondriacal concerns showed statistically significant decreases, which were only partly related to the degree of change in depression and anxiety severity. Our findings suggest that the presence of hypochondriacal concerns among depressed outpatients is more closely related to the presence of anxiety than depressive symptoms. The relatively small impact of an acute course of antidepressant treatment on hypochondriacal concerns in our sample suggests that these concerns may be enduring characteristics modulated only to a limited extent by short term pharmacological alterations of affective state.