Aim: This study aims to determine predictors of mortality among hospitalized geriatric patients with COVID-19 with a particular concern about the predictive value of blood urea nitrogen to albumin ratio (BAR). Methods: A retrospective cohort study included 114 geriatric patients (aged ≥ 60 years) admitted to a geriatric hospital for isolation. A review of the medical files of participants was conducted to extract data regarding age, sex, comorbidities, clinical manifestations of COVID-19 and laboratory data including haematology, biochemistry and inflammatory markers on admission. BAR and Charlson comorbidity index (CCI) was calculated. The primary outcome was in-hospital mortality. Appropriate statistical analyses were operated. RESULTS: 59 (51.8%) patients died at the hospital. Comorbidities associated with mortality included diabetes mellitus, old stroke, hypertension, and cardiac disease besides higher CCI. Reporting of fever, fatigue, sore throat and delirium /altered mental status were also associated with mortality. Significant biomarkers included lower haemoglobin and serum albumin, higher total leukocyte count, serum creatinine, blood urea nitrogen, lactate dehydrogenase, D-dimer, c-reactive protein (CRP) and BAR. BAR at a cutoff of 12.92 mg/gm had the best predictive performance for mortality. Serum BAR > 12.92 mg/gm, albumin < 3.05 g/dl and CRP > 108 mg/l were independent predictors of in-hospital mortality CONCLUSION: Serum BAR > 12.92 mg/gm, albumin < 3.05 g/dl and CRP > 108 mg/l are independent predictors of mortality among hospitalized geriatric patients with COVID-19. BAR is a convenient biomarker and has the best predictive performance to early pick up COVID-19 geriatric patients at high risk of mortality.