The relationships of longitudinal biological measures to longer-term outcome in depressed patients have not been well explored. This study was designed to investigate whether in a sample of depressed patients: (a) symptomatic and functional outcome at 1 year was significantly different in psychotic major depressed (PMD) patients as compared with nonpsychotic major depressed (NPMD) patients and (b) high urinary or plasma cortisol levels at baseline or 1 year were associated with poorer outcomes at 1 year. Forty-two depressed patients (9 psychotic, 33 nonpsychotic) were evaluated at baseline and at 1 year using a battery of clinical ratings and measures of cortisol. A group of normal, healthy control subjects were similarly evaluated at baseline. At 1-year follow-up, PMD patients did not differ from NPMD patients in their Hamilton Depression Rating Scale (HDRS) and Brief Psychiatric Rating Scale scores (BPRS), but PMD patients demonstrated significantly poorer social and occupational functioning. Significant correlations were observed ( n = 18) between higher levels of urinary and plasma cortisol at 1 year and poorer social and occupational functioning at 1 year, independent of the degree of residual depression. In contrast, baseline measures of urinary and plasma cortisol did not predict social and occupational functioning at 1 year.