The aim of the current study was to assess subjective cognitive dysfunction in elderly depressed patients and the impact of socio-demographic and clinical factors on its presentation.
 All patients aged ≥ 65 years, which were hospitalized in the Psychiatric Clinic of “Alexandrovska” University Hospital for a two-year period and met the ICD-10 criteria for depression, were included. They were evaluated by Hamilton Depression Rating Scale 24 items (HDRS24), its emotional-cognitive subscale (HDRScog) and its suicidal item (HDRSsuic), Geriatric Depression Scale 15 items (GDS15), its suicide subscale (GDSsuic), Perceived Deficit Questionnaire – Depression 5 items (PDQ-D5) and University of California, Los Angeles Loneliness Scale (UCLA-LS). The PDQ-D5, UCLA-LS and GDS15 were also applied to a control group of 50 healthy elderlies. The data were processed with the IBM SPSS Statistics 25.0.
 Out of 131 patients, 29 had late-onset depression (LOD), 80 had recurrent depressive disorder (RDD) and 22 had bipolar depression (BD). Patients with BD had more episodes than those with RDD, and higher PDQ-D5 and HDRScog scores than those with LOD and RDD. The severity of PDQ-D5 correlated significantly with the severity of depression, loneliness, suicidal and HDRScog scores. After applying multiple linear regression analysis (R2 = 0.489, p <0.001) the UCLA-LS severity had the greatest impact on PDQ-D5, followed by HDRScog and depression severity, assessed by GDS15. The diagnosis of BD hadthe least impact.
 PDQ-D5 reflects the severity of the current late-life depression and correlates significantly with the loneliness, emotional-cognitive symptoms and self-assessment of depression.