Background
Mild to moderate depressive symptoms are common during treatment for non-metastatic breast cancer. The goal of this secondary analysis was to determine if depressive symptoms predict clinical outcomes at long-term follow-up.
Methods
From 1998-2005, we interviewed 231 women with the Hamilton Rating Scale for Depression who were participating in a psychosocial study 2-10 weeks post-surgery for nonmetastatic breast cancer (Stage 0-IIIb). We conducted Kaplan Meier (K-M) curves and Cox proportional hazards (PH) models to examine associations between depressive symptoms, overall survival, and disease-free survival at 8-15 year follow-up.
Results
A total of 95 women (41.1%) scored in the mild-moderately depressed range. Non-depressed women had longer overall survival (M=13.56 years; SE=.26) than those in the mild/moderate depressed group (M=11.45 years; SE=.40), Log-rank χ2(1)=4.41, p=.036. Cox PH models, adjusting for covariates, showed comparable results: mild/moderate depressive symptoms hazard ratio=2.56, [95% CI, 1.11 to 5.91], p=.027. Similar results were observed in a subsample with invasive disease (n=191). Depression category did not predict disease-free survival in the overall or invasive sample.
Conclusions
Screening and referrals for treatment of depressive symptoms, even at subclinical levels, is important early in treatment. A randomized trial is warranted to determine effects of depressive symptoms on clinical outcomes.