BACKGROUND AND PURPOSE: Clinical status of subjects with knee osteoarthritis (KOA) is influenced by a complex interaction of several biopsychosocial factors. The use of patient-reported measures (PROM) is considered the gold standard for their evaluation. However, considering that 1 in 5 subjects with KOA present with depressive symptoms, it is necessary to analyse how this psychological domain may influence the subjective perception of PROM. The objective was to study the impact that depressive symptoms have on functional outcome assessments, according to the degree of objectivity of diverse outcome measures. METHODS: Cross-sectional study. Subjects with severe KOA, verified with clinical and radiological symptoms, were assessed with patient-reported (Oxford Knee Score), clinician-reported (knee range of motion), and performance-based (Timed up and go test) measures. The existence of depressive symptoms was assessed with the Yesavage scale, and participants were classified for having no-, mild- or severe-symptoms. Linear correlations (r) and one-way analysis of variance compared groups (95% CI). RESULTS: 244 participants were analysed, of which 75 (30.7%) had depressive symptoms. These symptoms had an inverse moderate association with Oxford Knee Score (r = -0.387). However, the correlation with the Timed up and go test was low (r=0.176), while there was no correlation with knee ROM (r = -0.087). CONCLUSIONS: This study supports that patient-reported questionnaires may offer biased information on the clinical status of patients with severe knee osteoarthritis who present with depressive symptoms. Consideration of such symptoms may be critical to ensure data collected to accurately reflect patients' capacities and perceptions.