The association of baseline anxiety and depression with overall survival (OS) in colorectal cancer (CRC).
706 Background: Patient-reported measures such as psychosocial wellbeing and quality of life have been reported to correlate with cancer survival, but the specific factors within these constructs that associate with outcomes remains poorly defined. Our aim was to examine the effect of baseline anxiety and depression on OS in a population-based cohort of CRC patients. Methods: Patients diagnosed with CRC in 2008, referred to any 1 of 5 regional cancer centers in British Columbia, and who completed a baseline questionnaire to evaluate anxiety and depression were reviewed. We developed a scoring system based on these patient-reported survey data to characterize intensity and subsequently categorized patients as having no, mild, moderate or severe anxiety and/or depression. Vital statistics were captured by medical records. Unadjusted Kaplan-Meier methods and multivariate Cox regression models that controlled for additional socio-demographic and clinical characteristics were constructed to correlate anxiety and depression with OS. Results: A total of 449 patients were included: median age was 66 years (range 26-92), 52% were men, and 69% were white. A significant proportion of respondents reported anxiety (58%) or depression (45%) at baseline. Among them, the majority indicated their anxiety and depressive symptoms were severe (43% and 43%, respectively). Women and those who were single or unmarried were more likely to have anxiety and depression (p<0.001 and p=0.02, respectively). In unadjusted analyses, we observed a trend where increased severity of anxiety (p=0.07) and depression (p=0.04) correlated with worse outcomes. In Cox regression that considered other prognostic factors, however, this observation failed to persist, and neither anxiety nor depression associated with OS (Table). Conclusions: The impact of psychosocial wellbeing and quality of life on CRC outcomes that have been previously described in the literature are unlikely to be mediated through baseline anxiety or depression. [Table: see text]